Thesis Kaplon by Nicole Nijhuis

SENTENCED TO SENESCENCE

SENTENCED TO SENESCENCE BRINGING REBELLIOUS CELLS TO JUSTICE

INVITATION to attend the public defence of the thesis

SENTENCED TO SENESCENCE BRINGING REBELLIOUS CELLS TO JUSTICE

by Joanna Kaplon Monday 26th January 2015 at 15:45 h in the aula of VU University Amsterdam The ceremony will be followed by a reception

Joanna Kaplon

Paranymphs:

Joanna Kaplon

Maarten Hoekstra mhhoekstra@gmail.com Katrin Meissl katrinmeissl@gmail.com

SENTENCED TO SENESCENCE - BRINGING REBELLIOUS CELLS TO JUSTICE

Joanna Kaplon

ISBN 978-90-75575-42-2 Cover design: Joanna Kaplon. Metaphorical illustration of senescence and the process leading to it. The research described in this thesis was carried out in the division of Molecular Oncology at the Netherlands Cancer Institute in Amsterdam, The Netherlands, and was supported by grants from The Netherlands Organization for Scientific Research (NWO) and an ERC Synergy grant. Copyright ÂŠ 2014 by Joanna Kaplon. All rights reserved Published by the Netherlands Cancer Institute/Antoni Van Leeuwenhoek Hospital. Layout and printing by Gildeprint - Enschede, with financial support from the Netherlands Cancer Institute.

VRIJE UNIVERSITEIT

SENTENCED TO SENESCENCE - BRINGING REBELLIOUS CELLS TO JUSTICE

ACADEMISCH PROEFSCHRIFT ter verkrijging van de graad Doctor aan de Vrije Universiteit Amsterdam, op gezag van de rector magnificus prof.dr. F.A. van der Duyn Schouten, in het openbaar te verdedigen ten overstaan van de promotiecommissie van de Faculteit der Geneeskunde op maandag 26 januari 2015 om 15.45 uur in de aula van de universiteit, De Boelelaan 1105

door Joanna Kaplon geboren te Pulawy, Polen

promotor:

prof.dr. D. S. Peeper

copromotor:

prof.dr. E. Gottlieb

TABLE OF CONTENTS

List of abbreviations

Chapter 1

General introduction

Chapter 2

Two-way communication between the metabolic and cell cycle machineries: the molecular basis

Chapter 3

Metabolic alterations in oncogene-induced senescence

Chapter 4

A critical role for the mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence

Chapter 5

Near-genomewide RNAi screening for regulators of BRAFV600E-induced 95 senescence identifies RASEF, a gene epigenetically silenced in melanoma

Chapter 6

Phosphoproteome dynamics in onset and maintenance of oncogene-induced senescence

117

Chapter 7

Signal transduction reaction monitoring deciphers site-specific PI3K-mTOR/MAPK pathway dynamics

143

Chapter 8

General discussion and perspective

161

Addendum

Summary Samenvatting Curriculum vitae List of publications PhD portfolio Acknowledgements

175 177 179 180 181 183

LIST OF ABBREVIATIONS 2 - HG - 2 - hydroxyglutarate 2PG - 3-phosphoglycerate 3PG - 3-phosphoglycerate 6PGD - 6-phosphogluconate dehydrogenase ACC - acetyl-CoA carboxylase AcCoA - acetyl-CoA ACLY - ATP-citrate lyase ACN - acetonitrile ALT - alanine aminotransferase AMPK - AMP-activated protein kinase APC/C - anaphase-promoting complex or cyclosome ASCT - Alanine/Serine/Cysteine/Threonine amino-acid transporter ATP - adenosine triphosphate BCA - bicinchoninic acid assay CDH1 - CDC20 homologue CDK - cyclin-dependent kinase CE - collision energy C/EBP - CCAAT/enhancer binding protein ChREBP - carbohydrate response element binding protein CKI - cyclin-dependent kinase inhibitor Cit - citrate COX/SCO - cytochrome c oxidase CPT1 - carnitine palmitoyltransferase 1 CS - citrate synthase Cycl - cycling cells DCA - dichloroacetic acid DDA - data-dependent acquisition DHAP - dihydroxyacetone phosphate DOX - doxocycline ECM - extracellular matrix ETD - electron transfer dissociation F-1,6-BP - fructose-1,6-bisphosphate F-2,6-BP - fructose-2,6-bisphosphate F6P - fructose-6-phosphate FA - fatty acid FAO - fatty acid oxidation FASN - fatty acid synthase

FDG-PET - 2-(18F)-fluoro-2-deoxy-D-glucose positron emission tomography FH - fumarate hydratase FOXO - forkhead box O transcription factors G3P - glyceraldehyde 3-phosphate G6P - glucose-6-phosphate G6PDH - glucose-6-phosphate dehydrogenase GAMT - guanidinoacetate methyltransferase GAPDH - glyceraldehyde 3-phosphate dehydrogenase GDH - glutamate dehydrogenase GLS - glutaminase GLUT - glucose transporter GO - gene ontology GPI - glucosephosphate isomerase GSH/GSSG - reduced:oxidized glutathione HCD - higher energy collision induced dissociation HDF - human diploid fibroblasts HK - hexokinase HNF4α - hepatocyte nuclear factor 4α IDH - isocitrate dehydrogenase ILs - interleukins IMAC - immobilized metal affinity chromatography αKG - α-ketoglutarate LC - liquid-chromatography LDH - lactate dehydrogenase Mal - malate MB - methylene blue MC - miscleavages MCT1 - monocarboxylate transporter 1 MDH - malate dehydrogenase ME - malic enzyme MOI - multiplicity of infection MS - mass spectrometry MRS - magnetic resonance spectroscopy NAT - natural peptides NF1 - neurofibromin 1 NRF1 - nuclear respiratory factor 1 OAA - oxaloacetate OCR - oxygen consumption rate

O-GlcNAc - O-linked N-acetylglucosamine OGAs - O-GlcNAcases OGTs - O-GlcNAc-transferases OIS - oncogene-induced senescence OISb - oncogene-induced senescence bypass OxPhos - oxidative phosphorylation PARP - poly ADP ribose polymerase PC - pyruvate carboxylase PDK - pyruvate dehydrogenase kinase PDH - pyruvate dehydrogenase PDP - pyruvate dehydrogenase phosphatase PEP - phosphoenolpyruvate PFK - phosphofructokinase PFKFB - 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase PGM - phosphoglycerate mutase PIN - prostate intraepithelial neoplasia PPAR - peroxisome proliferator-activated receptor PK (M1 or M2) - pyruvate kinase isoform M1 or M2 PPP - pentose phosphate pathway R5P - ribose-5-phosphate RASEF - RAS and EF-hand domain containing protein ROS - reactive oxygen species RRM- ribonucleotide reductase M RT - retention time SAC - spindle assembly checkpoint SA-Î˛-Gal - senescence-associated b-galactosidase activity SAHF - senescence-associated heterochromatin foci SASP - senescence-associated aecretory phenotype SCF - Skp1/cullin/F-box protein complex SCX - strong cation exchange SMS - senescence-messaging secretome SDH- succinate dehydrogenase SIS - stable isotope standard phosphopeptides SN - system N transporter SREBP - sterol regulatory element-binding protein SRM/MRM - selected or multiple reaction monitoring STREM - signal transduction reaction monitoring TCA cycle - tricarboxylic acid cycle

TFA - trifluoroacetic acid TIS - therapy-induced senescence TopImt - mitochondrial topoisomerase I UPS - ubiquitin proteasome system VDAC - voltage-dependent anion channel VHL - Hippel-Lindau protein WT1 - Wilms Tumor 1

足足足

CHAPTER 1 GENERAL INTRODUCTION

CELLULAR METABOLIC PATHWAYS Homeostasis of energy is the basis of all life: cells and organisms survive only because of the ability to give-and-take energy. Adenosine triphosphate (ATP) transports chemical energy within cells and is therefore an essential cellular currency providing energy for countless energy-consuming processes such as (macro) molecular biosynthesis, transport and motility. The main source of cellular ATP is the oxidation of carbons such as glucose and lipids1. Glucose enters the cell via glucose transporters (GLUTs) and is subsequently phosphorylated to glucose-6-phosphate (G6P) by hexokinases (HKs), which prevents its exit from the cell1. G6P can either proceed to glycolysis, producing ATP, NADH and pyruvate, or it can enter the pentose phosphate pathway (PPP, Figure 1). In addition to hexokinases, enzymes essential for the stimulation and suppression of glycolysis are 6-phosphofructo-2-kinase/fructose2,6-bisphosphatases (PFKFBs), phosphofructokinases (PFKs), phosphoglycerate mutase (PGM) and pyruvate kinases (PKs). Under anaerobic conditions, glucose-derived pyruvate is converted by lactate dehydrogenase (LDH) to lactic acid in a process called anaerobic glycolysis. In the presence of oxygen, the activity of latter enzyme is usually decreased and pyruvate is metabolized in the mitochondria to acetyl-CoA by pyruvate dehydrogenase (PDH), thereby fueling the tricarboxylic acid (TCA) cycle (Figure 1). Oxidation of acetyl-CoA in the TCA cycle converts NAD+ into NADH and FAD into FADH2, which subsequently generate ATP in a process called oxidative phosphorylation (OxPhos)2. Reactive oxygen species (ROS), by-products of OxPhos, can act as essential signaling molecules, but when present at high levels they can also cause irreversible oxidation of DNA, proteins and lipids1. Several glycolytic and TCA intermediates may be used for biosynthesis of macromolecules required for the cell cycle progression such as amino acids, lipids and nucleotides3. For fatty acid (FA) synthesis, the TCA intermediate citrate can be transported to the cytosol and converted into cytosolic acetyl-CoA by an enzyme called ATP-citrate lyase (ACLY)1,4. By the activity of acetyl-CoA carboxylase (ACC) and fatty acid synthase (FASN) enzymes, cytosolic acetyl-CoA can be metabolized into the FAs (Figure 1). In addition, the glycolysis intermediate dihydroxyacetone phosphate (DHAP) can fuel lipogenesis, the main method for storing glucose in adipocytes1,4. A major source of nucleotides is the PPP. It shunts G6P away from glycolysis by converting it into ribose-5-phosphate (R5P) in a reaction carried out by the rate-limiting G6P dehydrogenase (G6PDH) enzyme (Figure 1)5. By doing so, the PPP does not only provide nucleotide precursors, but is also the primary supplier of NADPH. NADPH is a cofactor required for reductive biosynthesis of FAs, amino acids and nucleotides, but also regulates the cellular redox status by reducing the antioxidant glutathione.

Chapter 1

GENERAL INTRODUCTION

GLUCOSE GLUT

PENTOSE PHOSPHATE PATHWAY

glucose HK

G6PDH

NADPH

GPI

PFKFB

F6P

NUCLEIC ACIDS

F-2,6-BP

PFK

F-1,6-BP

FATTY ACID SYNTHESIS

3PG

DHAP

PGM

FASN

malonyl-CoA

LIPIDS

2PG

AMINO ACIDS

ACC

PEP pyruvate

PDH

PDK

AMINO ACIDS

PDP ME

FATTY ACID OXIDATION

ACLY

LDH

lactate

CPT1

acetyl-CoA

acetyl-CoA malate

citrate

TCA CYCLE

ATP ROS

6PGD, RRM

R5P

isocitrate IDH

aKG

fumarate SDH

succinate

MITOCHONDRION

glutamate GLS

glutamine

CYTOSOL

GLUTAMINOLYSIS

ATP NADH

G LY C O LY S I S

G6P

ASCT SN

GLUTAMINE

Figure 1. Schematic representation of major metabolic pathways 2PG - 3-phosphoglycerate; 3PG - 3-phosphoglycerate; 6PGD - 6-phosphogluconate dehydrogenase; ACC - acetyl-CoA carboxylase; ACLY - ATP-citrate lyase; ASCT - Alanine/Serine/Cysteine/Threonine amino-acid transporter; ATP - adenosine triphosphate; CPT1 - carnitine palmitoyltransferase 1; DHAP - dihydroxyacetone phosphate; F-1,6-BP - fructose-1,6-bisphosphate; F-2,6-BP, fructose-2,6-bisphosphate; F6P - fructose6-phosphate; FASN - fatty acid synthase; FH - fumarate hydratase; G6P - glucose-6-phosphate; G6PDH glucose-6-phosphate dehydrogenase; GLS - glutaminase; GLUT - glucose transporter; GPI - glucosephosphate isomerase; HK - hexokinase; IDH - isocitrate dehydrogenase; αKG - α-ketoglutarate; LDH - lactate dehydrogenase; ME - malic enzyme; PDK - pyruvate dehydrogenase kinase; PDH - pyruvate dehydrogenase; PDP - pyruvate dehydrogenase phosphatase; PEP - phosphoenolpyruvate; PFK - phosphofructokinase; PFKFB - 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase; PGM - phosphoglycerate mutase; PK - pyruvate kinase; R5P - ribose-5-phosphate; RRM- ribonucleotide reductase M; ROS - reactive oxygen species; SDHsuccinate dehydrogenase; SN - system N transporter; TCA cycle - tricarboxylic acid cycle.

METABOLISM OF CANCER CELLS The most pronounced change in cancer metabolism was observed almost a century ago by Otto Warburg. Normal cells metabolize glucose in the mitochondrial TCA cycle. Only under low oxygen, glucose is converted into lactate. Warburg found that, in contrast, cancer cells metabolize glucose into lactate even in the presence of oxygen, which is referred to as aerobic glycolysis or “the Warburg effect” (Figure 2)8,9. This sustained aerobic glycolysis in cancer cells is linked to the activation of oncogenes or loss of tumor suppressors, with PI3K/ AKT/mTOR pathway, c-Myc, HIF-1a and AMPK as major players10-16. Interestingly, not only cancer cells but also highly proliferating nontransformed activated lymphocytes, thymocytes and embryonic cells metabolize glucose to lactate even in the presence of oxygen17-21. This indicates that the Warburg effect is a common feature of proliferation rather than a unique tumor characteristic. Yet, to achieve aerobic glycolysis, tumors upregulate cancer-specific enzyme isoforms, which may provide a therapeutic window for selective anti-tumor therapy. In order to compensate for the lower efficiency of energy generation in glycolysis (2 ATP compared to 36 ATP from mitochondrial respiration for each glucose molecule catabolized), cancer cells upregulate glucose transporters, particularly GLUT1 and GLUT3 to take up more glucose22. In fact, this dramatic increase in glucose uptake is exploited clinically to visualize tumors by 2-(18F)-fluoro-2-deoxy-D-glucose positron emission tomography (FDG-PET). In addition, cancer cells elevate the expression of several glycolytic enzyme isoforms, including HK2, PFKFB3 and M2 isoform of PK (PKM2, Figure 1). While HK1 is found in all mammalian tissues, HK2 is normally restricted to the skeletal muscle and adipocytes. Cancer cells show not only increased expression of HK2 but they also rely on this particular isoform23-26, indicating an additional role of HK2 in the context of malignancies. The observation that the mitochondria-associated form of HK influences the regulation of apoptosis may reflect one such activity27. PFKFB3 is an another example of a regulatory enzyme in glycolysis being upregulated in several cancers28. By generating fructose-2,6-bisphosphate (F-2,6-BP), PFKFB3 activates PFK1 to increase the flux through this step of glycolysis. Most isoforms of PFKFBs are bifunctional enzymes with both kinase and phosphatase activities, and can therefore also catalyze the destruction of F-2,6-BP and decrease PFK1 activity29. Notably, the PFKFB3

Chapter 1

Amino acids can be derived from the TCA intermediates citrate, α-ketoglutarate (αKG) and malate, as well as the glycolytic intermediate pyruvate and 3-phosphoglycerate (3PG; Figure 1). Glutamine, the most abundant of all amino acids, is another source of building blocks for macrosyntheses. Glutamine can be metabolized in the TCA cycle in a process called glutaminolysis, providing carbons for amino acids and lipids. Moreover it is important nitrogen source for the cells (Figure 1)6,7. In addition to glucose and glutamine, other amino acids can also provide TCA intermediates and lipids can provide acetyl-CoA through mitochondrial FA oxidation (FAO).

isozyme has low phosphatase activity and therefore activates PFK1 to trigger flux through glycolysis. Another important controller of the glycolytic flux in cancer is PK, an enzyme catalyzing the final irreversible step of glycolysis-conversion of phosphoenolpyruvate (PEP) to pyruvate with concomitant generation of ATP. Whereas differentiated tissues express the M1 isoform of PK (PKM1), all cancers express the M2 isoform (PKM2)30. PKM2 promotes aerobic glycolysis, and PKM2-expressing cells show a selective growth advantage in vivo relative to PKM1-expressing cells31. PKM1 is a constitutively active enzyme, whereas PKM2 enzyme activity is inhibited following its binding to tyrosineÂphosphorylated proteins downstream of cellular growth signals32. Interestingly, PKM2-depleted tumors continue to proliferate only in the absence of PKM1, suggesting that it is the inactivation of both PKMs (or their deficiencies) that underlies tumor growth33. The inhibition of PKMs might help to divert glucose-derived metabolites upstream of PK into biosynthetic pathways34. In this regard, accumulating G6P can fuel PPP to synthesize nucleotides and NADPH, and DHAP can be used for lipid synthesis and 3PG for amino acids (Figure 1). Paradoxically, PKM2expressing cells seem to generate more glucose-derived lactate relative to cells expressing PKM1. This process was suggested to be mediated by the upregulation of an alternative pathway that converts PEP to pyruvate without producing ATP, and therefore compensates for the decreased PKM2 activity in proliferating cells35. Moreover, LDH isoform A (LDHA), an enzyme responsible for the conversion of pyruvate to lactate, was shown to be upregulated in several tumors (Figure 1)36. The conversion of pyruvate to lactate is beneficial for cancer cells as it regenerates NAD+ to accelerate glycolysis and decreases ROS levels, thereby diminishing oxidative stress and promoting tumorsâ&#x20AC;&#x2122; survival37-39. Moreover, lactate secreted by one tumor cell can be taken up and oxidized in the TCA cycle of a well-oxygenated neighboring tumor cell40. To support aerobic glycolysis, cancer cells stimulate not only glycolytic enzymes, but also actively inhibit glucose-derived pyruvate oxidation in the TCA cycle. They do so by deactivating PDH, an enzyme that catalyzes pyruvate oxidation in the mitochondria and thus keeps it away from lactate production (Figure 1). Indeed, the expression of pyruvate dehydrogenase kinase 1 (PDK1), a negative regulator of PDH, is increased in several cancers41,42. Besides increasing lactate production, PDH inhibition diminishes OxPhos and ROS production, resulting in the decrease of oxidative stress. Conversely, reactivation of PDH upon PDK1 inhibition was shown to decrease lactate production and induce oxidative stress and cancer cell death43,44. Although oxidative metabolism of glucose is suppressed in cancer cells, the mitochondrial activity is normally conserved. Nonetheless, several genes encoding TCA cycle-related enzymes were found to contain mutations that support malignancy. For example, genomic analyses of tumors have implicated specific mutations of isocitrate dehydrogenases (IDH) isoform 1 and 2 as tumorigenic drivers45-48. IDHs are enzymes that normally catalyze the

Chapter 1

conversion of isocitrate to aKG with the concomitant production of NAD(P)H (Figure 1). The mutant IDH1/2 enzymes catalyze the normal IDH reaction very inefficiently. Instead, they convert aKG to the rare metabolite 2-hydroxyglutarate (2-HG) while oxidizing NAD(P) H to NAD(P)+49,50. Growing evidence suggests that 2-HG promotes tumorigenesis, at least in part, by the competitive inhibition of various ÎąKG-dependent dioxygenases that control gene expression epigenetically. They include the TET family of methylcytosine dioxygenases and histone lysine demethylases (KDM), enzymes that regulate the methylation state of DNA and histones, respectively51-53. Consistent with this notion, cells with IDH mutations were reported to have an altered methylation pattern and 2-HG-mediated inhibition of histone demethylation blocks the differentiation of untransformed cells54-56. Also, loss-offunction mutations in the TCA cycle enzymes succinate dehydrogenase (SDH) or fumarate hydratase (FH) have been linked to tumorigenesis (Figure 1). Succinate or fumarate, which accumulate in mitochondria owing to the inhibition of SDH or FH, overflow to the cytosol, where they inactivate prolyl hydroxylases (PHDs), a subset of the ÎąKG-dependent dioxygenase family57-59. Inactivation of PHDs leads to accumulation of the central regulator of oxygen-sensing pathway HIF-1a, which promotes metabolic reprogramming towards tumor-supporting aerobic glycolysis60,61. In addition to glycolysis, cancer cells also rely on glutaminolysis to meet their highly energetic demands for proliferation7. The products of glutaminolysis are very important in fueling the TCA cycle of tumors. In this way, glutamine provides intermediates for the synthesis of FA and amino acids, essential building blocks of lipids and proteins, respectively. In fact, cancer cells increase glutamine uptake by the upregulation of ASCT2 and SN2, two important glutamine transporters (Figure 1)62,63. Moreover, glutaminase 1 (GLS1), a key enzyme of glutaminolysis, is highly expressed in cancer cells and its silencing delays tumor growth (Figure 1)64-66. Likewise, cancer cells are highly sensitive to glutamine withdrawal67. Highly proliferating cancer cells are also in constant need for nucleotides to replicate DNA. By providing pentose phosphate for nucleotide synthesis, the PPP is important and, in agreement with this, is frequently upregulated in many types of tumors68. Accordingly, the activity and levels of G6PDH, a rate-limiting PPP enzyme, are increased in multiple types of cancer (Figure 1)69,70. Similarly, the levels of transketolase, another PPP enzyme, were shown to be elevated, too71. In addition to building blocks for nucleotides, the PPP promotes tumor survival by producing NADPH, a key tool for both preventing oxidative stress and biosynthesis68. Increased lipid turnover is another remarkable feature of cancer metabolism. Most normal cells preferentially use circulating FA for the synthesis of new structural phospholipids. Cancer cells, however, frequently upregulate de novo FA synthesis to satisfy high phospholipid needs for the generation of new membranes72. Notably, virtually all esterified FAs in tumor models are derived from de novo synthesis73. High rates of tumor lipogenesis

are reflected by increased activity and expression of several lipogenic enzymes. FASN, a key lipogenic enzyme responsible for the terminal step in FA synthesis, is elevated in several tumors and its inhibition dampens xenograft growth in mice (Figure 1)74-76. Similarly, ACLY, an enzyme necessary for exporting citrate from the mitochondria to the cytosol for de novo lipogenesis, is indispensable for tumor formation both in vitro and in vivo (Figure 1)77,78. Additionally, ACC is important for tumorigenesis, as its inhibition induces growth arrest and apoptosis of cancer cells (Figure 1)79,80. Notably, metabolism of choline, which, next to FAs, is an important precursor of cellular membrane phospholipids, is also altered in cancer cells. Several enzymes of choline metabolism are overexpressed and/or activated and therapeutic response of tumors can be monitored by magnetic resonance spectroscopy (MRS) of the total and phosphocholine levels81. In addition to the higher FA synthesis, upregulation of FAO has been linked to tumorigenesis. ATP production by FAO has been shown to prevent anoikis, a type of apoptotic cell death due to loss of attachment to the extracellular matrix82. Poly ADP ribose polymerase (PARP)-mediated activation of FAO correlates with poor survival of breast cancer patients83. Along these lines, the expression of carnitine palmitoyltransferase-1 isoform C (CPT1), an enzyme stimulating FAO, is higher in cancer cells, leading to increased FAO-derived ATP production and resistance to glucose deprivation (Figure 1)84. At the same time, depletion of CPT1C decreases tumor growth in vivo and increases sensitivity to metabolic stress. In short, cancer cells display a specific lipogenic phenotype essential for the malignant phenotype and this is therefore, in addition to the anaerobic glycolysis, an important hallmark of cancer. SENESCENCE AS A POTENT ANTITUMOR MECHANISM Several mechanisms have evolved to prevent malignant transformation. In response to a tumorigenic insult, cells can undergo various cell death programs such as apoptosis85 or autophagy86. Alternatively, they can stop proliferating and enter a program called senescence. Cellular senescence is a state of irreversible arrest of the cell cycle that involves an activation of tumor suppressor pathways, including p16INK4A-Rb and ARF-p5387-90. Conversely, expression of cell-cycle activators cyclin A, cyclin B and PCNA is downregulated91. Senescence is also accompanied by morphological changes, the induction of senescenceassociated b-galactosidase activity (SA-b-Gal)92, chromatin condensation into senescenceassociated heterochromatin foci (SAHF)93,94 and the activation of a secretory program called senescence-messaging secretome (SMS) or Senescence-Associated Secretory Phenotype (SASP)95-101. Senescence was originally identified as a process associated with exhaustion of replicative potential102. This type of irreversible cell cycle arrest, termed â&#x20AC;&#x2DC;replicative senescenceâ&#x20AC;&#x2122;, has been suggested to represent a failsafe mechanism preventing the expansion of aged cells103. Later, it was shown that this process involved the erosion of telomeres; and that senescence

Chapter 1

can also be induced prematurely (that is, in the presence of normal functional telomeres) by various stress signals, including DNA-damaging insults104,105 and oncogenic signaling triggered by the activation of an oncogene or the inactivation of a tumor-suppressor gene87. Research carried out in more recent years has clearly demonstrated that oncogene-induced senescence (OIS) is a powerful in vivo tumor-suppressing mechanism both in model systems and humans. For example, the finding that oncogenic RAS induces OIS in vitro was validated in vivo using mouse models. Expression of oncogenic KRAS from its endogenous promoter in mice results in the development of lung adenomas, as well as premalignant pancreatic intraductal neoplasia, which rarely progress to malignancy 106. These benign lesions show a low proliferative index and express several senescence markers, including the activation of SA-b-Gal and upregulation of p15INK4B and p16INK4A. Another RAS family member, HRAS, also induces senescence in vivo. Inducible expression of an oncogenic HRAS transgene in the mammary gland induces proliferation when the oncogene is present at low levels, but tumor cell senescence when the oncogene is highly expressed107. Moreover, expression of oncogenic HRAS, whether expressed from its endogenous promoter or directed to the bladder epithelium, leads to tumor cell senescence108,109. In addition, murine papillomas induced upon DMBA/TPA treatment (leading to RAS mutation), express several senescence markers106,110,111. Along these lines, senescence features are observed in mouse models with oncogenic BRAF, a proximal RAS downstream kinase. Conditional expression of oncogenic BRAF from its endogenous promoter in melanocytes causes the formation of nevuslike benign lesions112,113. These nevi express markers of senescence and remain stable for several months. Consistent with this, oncogenic BRAF expression in lungs initially induces proliferation, but is followed by proliferative arrest associated with characteristics of senescence114. In addition to RAS oncogenes and their downstream kinases, also distal effectors of the RAS pathway induce senescence in murine tumors. For example, the expression of an E2F3 transgene in the pituitary gland of mice causes an initial phase of proliferation, but cells successively stop dividing, acquire markers of senescence and fail to form tumors115. Also activation of PI3K/AKT/mTOR pathway triggers a senescence response. In fact, PTEN was the first example of a tumor suppressor gene whose loss triggers senescence in vivo. Conditional PTEN deletion in murine prostate cells results in the formation of high-grade prostate intraepithelial neoplasia (PIN), which displays characteristics of senescence. In combination with p53 loss, however, these lesions progress to malignant prostate carcinomas116. Similarly, the expression of AKT1 in the prostate results in the development of PIN lesions, which have features of cellular senescence117. Likewise, mice overexpressing RHEB, which connects AKT to mTOR, also triggers PIN lesions that are positive for senescence markers118. Oncogene inactivation may also induce cellular senescence: enforced loss of Myc in Mycinitiated hepatocellular carcinomas, lymphomas or osteosarcomas, causes tumor regression that is associated with the induction of senescence119.

Another example of cellular senescence in vivo comes from kidneys from von Hippel-Lindau (VHL) knockout mice, which display an increase in SA-b-Gal activity and levels of p27Kip1 and DcR2, a TNF family decoy receptor associated with senescence120. Also genetic loss of one allele of Rb drives formation of thyroid adenomas that have several senescence markers121. Cell senescence has also been reported in humans. The (benign) melanocytic nevus was the first human lesion for which the evidence was shown in favor of the idea that OIS prevents malignant progression122,123. In spite of the presence of oncogenic BRAF (or, in some cases, NRAS), nevi are commonly cell cycle arrested and exhibit evidence of senescence markers, such as SA-Î˛-Gal and p16INK4A. Lesions suffering from loss of the tumor suppressor neurofibromin 1 (NF1) are another example of cellular senescence in humans124. As NF1 is a negative regulator of RAS activity, loss of NF1 results in hyperactivated RAS signaling and the development of neoplastic lesions, known as neurofibromas, which show markers of senescence. Finally, premalignant human colon adenomas and PIN lesions also show features of senescence including increased SA-Î˛-Gal activity and induction of SMS components95,96. As described above, numerous studies have demonstrated that senescence is associated with pre-malignant stages of neoplastic transformation and has a crucial role in preventing tumor development. Hence, understanding the mechanisms controlling the senescence program is of vital importance as it may guide us to novel therapeutic targets in cancer. However, despite enormous efforts, we have only begun to uncover some of the underlying principles, and many questions remain to be answered. Among processes that remain largely unexplored in senescence, the ones controlling cellular metabolism are at the forefront. METABOLISM OF SENESCENT CELLS Proliferating and senescent cells are expected to have very different metabolic requirements. Considering that proliferating cells must duplicate their cellular biomass in order to divide, much of their metabolic energy is devoted to synthesis of DNA, proteins and lipids. Senescent cells appear to be relieved of this vast metabolic demand because they are not dividing. However, induction of senescence is accompanied by a multitude of specific morphologic and physiologic changes that involve energy-consuming processes, too. Accordingly, strongly induced production of, for example, senescence-associated cytokines must require shuffling of metabolic resources towards protein synthesis. Also, vesicular transport and secretion, both active in senescent cells, demand sufficient supply of the energy. Hence, while freed from the biosynthetic needs accompanied with creating daughter cells, senescent cells still must adopt their cellular metabolism to support energetic and anabolic requirements. Markedly, regardless of shared belief that senescent cells are metabolically active, relatively few studies have meticulously investigated senescence-associated changes of metabolism (Table 1).

Chapter 1

Table 1. Metabolic alterations in senescence Type of senescence

Metabolic alteration

Reference

replicative

strong decrease in glycolysis and production of glucosederived lactate

125

ectopic expression of glycolytic enzymes, PGM or GPI, increases glycolytic flux, decreases oxidative damage and abrogates senescence; downregulation of glycolysis upon depletion of PGM or GPI induces senescence

126

despite lower rate of de novo synthesis of FA, OIS cells show increased steady-state levels of free FA due to the higher FA oxidation

139

suppressed nucleotides metabolism due to the downregulation of RRM2; ectopic expression of RRM2 restores nucleotide metabolism and abrogates senescence

135

p53 induces the expression of TIGAR, an enzyme diminishing glycolytic activator F-2,6-BP; this inhibits glycolysis leading to cell cycle arrest

127

p53 represses the expression of TCA cycle-associated ME1 and ME2; depletion of ME1 and ME2 reciprocally activates p53 leading to a strong induction of senescence

132

chemotherapy-induced senescence associates with enhanced glucose utilization in the TCA cycle

129

growth arrest induced by inhibition of melanoma-driver BRAF is accompanied by several senescence features and increased oxidative metabolism

130,131

BRAF-induced

increased TCA cycle activity due to the activation of mitochondrial gatekeeper PDH; PDH is activated upon downregulation of its inhibitory kinase PDK1 and simultaneous upregulation of PDH-activating phosphatase PDP2; normalization of the levels of these enzymes inactivates PDH resulting in abrogation of OIS

133

replicative, DNA damage-induced and oncogeneinduced senescence

increased ratio of GPC to PC, two important components of choline metabolism

140

RAS-induced

p53-dependent

therapy-induced

GPC - glycerophosphocholine; GPI - glucosephosphate isomerase; FA - fatty acids; F-2,6-BP - fructose足2,6足 bisphosphate; ME - malic enzyme; OIS - oncogene-induced senescence; PC - phosphocholine; PDH - pyruvate dehydrogenase; PDK1 - pyruvate dehydrogenase kinase 1; PDP2 - pyruvate dehydrogenase phosphatase 2; PGM - phosphoglycerate mutase; RRM2- ribonucleotide reductase subunit M2; TCA cycle - tricarboxylic acid. 足

It has been reported that senescent cells adapt their metabolic profiles, which seem to oppose the ones seen in the cancer cells. In this regard, senescence manifests its antitumor function also at the level of metabolic (de)regulation. While tumor cells upregulate aerobic

glycolysis, replicative senescence is accompanied by strong decreases in glycolytic flux and production of glucose-derived lactate125. Downregulation of glycolysis induces metabolic imbalance, which is associated with decreased levels of ribonucleotide triphosphates, thereby leading to senescence-associated cell cycle arrest. Consistent with these observations, modulation of glycolytic enzymes controls the senescence response126. A function-based screen for immortalizing genes identified the glycolytic enzyme PGM to be crucial for abrogation of RAS-induced senescence (Figure 1). Analysis of the impact of other glycolytic enzymes on cell cycle arrest showed that glucosephosphate isomerase (GPI) can drive senescence escape (Figure 1)126. Along these lines, ectopic expression of PGM or GPI increases glycolytic flux, decreases the oxidative damage and extends the life span of primary fibroblasts. Conversely, downregulation of glycolysis upon depletion of PGM or GPI induces premature senescence126. Similarly, TIGAR, an enzyme diminishing glycolytic activator F-2,6-BP and thereby glycolysis, was reported to induce cell cycle arrest, a key characteristic of senescent cells (Figure 1)127,128. Growing evidence indicates that senescence provokes an anti-Warburg effect not only by downregulation of glycolysis but also by altering TCA cycle activity. In fact, therapyinduced senescence (TIS) has been associated with enhanced glucose utilization in the TCA cycle129. Along these lines, cell cycle arrest upon inhibition of the common melanoma driver oncogene BRAF is accompanied by the induction of several senescence features130 and an increase in oxidative metabolism131. Moreover, p53-dependent senescence has been linked to the regulation of TCA-associated malic enzymes (MEs, Figure 1)132. While p53 accumulation represses the expression of ME1 and ME2, depletion of ME1 and ME2 reciprocally activates p53, resulting in a strong induction of senescence. In addition, we have demonstrated that OIS cells show an increased TCA cycle activity due to the activation of PDH, an enzyme linking glycolysis to oxidative phosphorylation133. In OIS, PDH is activated upon downregulation of its inhibitory kinase PDK1 and simultaneous upregulation of PDHactivating phosphatase PDP2 (Figure 1). Normalization of the levels of these enzymes inactivates PDH, resulting in abrogation of OIS. Likewise, the analysis of changes in protein expression in OIS revealed upregulation of proteins involved in oxidative phosphorylation and downregulation of proteins involved in glycolysis, further supporting an important role of oxidative metabolism in senescence134. Regulation of nucleotides-producing PPP has also been reported to play a role in the senescence response. Ribonucleotide reductase subunit M2 (RRM2), a PPP enzyme crucial for nucleotide synthesis, is downregulated in senescence, leading to a sharp decrease in the number of nucleotides available for DNA synthesis required for cell cycle progression (Figure 1)135. The suppression of nucleotide metabolism represents a critical element underlying the establishment and maintenance of OIS. Accordingly, either addition of exogenous nucleosides or restoration of RRM2 abrogates RAS-induced senescence. This indicates that,

CONCLUDING REMARKS Research performed in recent years has made it clear that cancer cells rewire their metabolism to support growth and proliferation. This remodeling goes beyond upregulation of aerobic glycolysis but involves all major cellular metabolic pathways. The observation that several of the oncogenic and tumor suppressor pathways regulating metabolism in tumor cells are linked also to senescence, suggested that the latter program must, somehow, be mechanistically connected to metabolic (de)regulation. Indeed, a number of studies on metabolic changes in senescence have revealed that metabolic pathways altered in cancer cells are also important for the execution of the senescence program. In fact,

Chapter 1

in contrast to high PPP activity in cancer cells, the PPP must be shut down for senescence induction, as interference with this regulation abrogates the senescence program. Similarly, depletion of 6-phosphogluconate dehydrogenase (6PGD), the third enzyme in the PPP, inhibits proliferation of lung cancer cells in vitro and in mice due to the induction of senescence as evidenced by the upregulation of the senescence markers SA-b-Gal, p53, and p21Cip1 (Figure 1)136. As described above, cancer cells strongly depend on glutaminolysis for growth and proliferation. In contrast, senescence has been linked to the inhibition of glutamine metabolism. Endothelial cells rely heavily on glutaminolysis. The removal of this energy source by pharmacological inhibition of GLS, a key enzyme in glutaminolysis, drastically decreases cellular ATP levels, leading to the induction of cell cycle arrest, enhanced SAb-Gal activity and induction of p16INK4A and p21Cip1 protein levels in these cells137. Hence, downregulation of glutamine utilization likely represents a tumor-suppressive response. Lastly, several studies have recently linked senescence to changes in lipid metabolism. Replicative senescence is accompanied by reduced biosynthesis of FA and formation of phospholipids138. Also, OIS cells have been demonstrated to have a lower rate of de novo synthesis of FA139. Despite that, OIS cells show increased steady-state levels of free FA due to the higher FAO. Inhibiting the FAO, however, did not prevent senescence-associated cell cycle arrest, arguing against causal role of FA in senescence program139. In addition, senescent cells were recently reported to have altered choline metabolism. Analysis of metabolic changes in three types of senescence: replicative, DNA-damaged induced and OIS, showed that the ratio of glycerophosphocholine (GPC) to phosphocholine (PC) is increased independently of the senescent type140. Notably, this is diametrically opposite to the change in choline metabolism in tumor cells. During malignant transformation cells show â&#x20AC;&#x153;GPC to PCâ&#x20AC;? switch implying an increase of PC compared to GPC141-143. Conversely, chemotherapy-mediated induction of cell cycle arrest and apoptosis in cancer cells is accompanied by the increase in GPC levels144. This implicates that suppression of choline metabolism in senescence serves as an antitumor mechanism, too.

the metabolism in senescent cells is rewired in a way that counteracts changes required for malignant transformation. Hence, the antitumor function of senescence is manifested also at the level of metabolic regulation. Nevertheless, we have only begun to undercover the mechanism that controls metabolism in the senescence setting. Further research on metabolic alteration in senescence will be necessary to better understand how cancer cells disrupt the normal limitations of metabolism control. This will hopefully reveal new therapeutic metabolic cancer targets. REFERENCES 1. 2.

4. 5. 6.

9. 10.

11.

Berg, J. M. Biochemistry. (W H Freeman & Company, 2012). Krebs, H. A. The Pasteur effect and the relations between respiration and fermentation. Essays Biochem. 8, 1–34 (1972). Owen, O. E., Kalhan, S. C. & Hanson, R. W. The key role of anaplerosis and cataplerosis for citric acid cycle function. J Biol Chem 277, 30409–30412 (2002). Coleman, R. A. & Lee, D. P. Enzymes of triacylglycerol synthesis and their regulation. 43, 134–176 (2004). Stanton, R. C. Glucose-6-phosphate dehydrogenase, NADPH, and cell survival. IUBMB Life 64, 362–369 (2012). Kovacevic, Z. & McGivan, J. D. Mitochondrial metabolism of glutamine and glutamate and its physiological significance. Physiol. Rev. 63, 547–605 (1983). DeBerardinis, R. J. & Cheng, T. Q’s next: the diverse functions of glutamine in metabolism, cell biology and cancer. Oncogene 29, 313–324 (2010). Warburg, O., Wind, F. & Negelein, E. The metabolism of tumors in the body. The Journal of general physiology 8, 519–530 (1927). WARBURG, O. On the origin of cancer cells. Science 123, 309–314 (1956). Jones, R. G. & Thompson, C. B. Tumor suppressors and cell metabolism: a recipe for cancer growth. Genes Dev 23, 537–548 (2009). Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033 (2009).

12.

13.

14. 15. 16. 17.

18.

19.

20.

21.

22.

Levine, A. J. & Puzio-Kuter, A. M. The Control of the Metabolic Switch in Cancers by Oncogenes and Tumor Suppressor Genes. Science 330, 1340–1344 (2010). Koppenol, W. H., Bounds, P. L. & Dang, C. V. Otto Warburg’s contributions to current concepts of cancer metabolism. Nat Rev Cancer 11, 325–337 (2011). Dang, C. V. Links between metabolism and cancer. Genes Dev 26, 877–890 (2012). Cantor, J. R. & Sabatini, D. M. Cancer cell metabolism: one hallmark, many faces. Cancer Discov 2, 881–898 (2012). Cairns, R. A., Harris, I., McCracken, S. & Mak, T. W. Cancer Cell Metabolism. Cold Spring Harb. Symp. Quant. Biol. 76, 299–311 (2012). Hedeskov, C. J. Early effects of phytohaemagglutinin on glucose metabolism of normal human lymphocytes. Biochem J 110, 373–380 (1968). Brand, K. Glutamine and glucose metabolism during thymocyte proliferation. Pathways of glutamine and glutamate metabolism. Biochem J 228, 353–361 (1985). Wang, T., Marquardt, C. & Foker, J. Aerobic glycolysis during lymphocyte proliferation., Published online: 24 June 1976; | doi:10.1038/261702a0 261, 702–705 (1976). Steck, T. L., Kaufman, S. & Bader, J. P. Glycolysis in chick embryo cell cultures transformed by Rous sarcoma virus. Cancer Res 28, 1611–1619 (1968). Kondoh, H. et al. A high glycolytic flux supports the proliferative potential of murine embryonic stem cells. Antioxid Redox Signal 9, 293–299 (2007). Macheda, M. L., Rogers, S. & Best, J. D. Molecular and cellular regulation of glucose transporter (GLUT) proteins in cancer. J Cell Physiol 202, 654–662 (2005).

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Mathupala, S. P., Heese, C. & Pedersen, P. L. Glucose catabolism in cancer cells. The type II hexokinase promoter contains functionally active response elements for the tumor suppressor p53. J Biol Chem 272, 22776– 22780 (1997). Wolf, A. et al. Hexokinase 2 is a key mediator of aerobic glycolysis and promotes tumor growth in human glioblastoma multiforme. J Exp Med 208, 313–326 (2011). Chen, J., Zhang, S., Li, Y., Tang, Z. & Kong, W. Hexokinase 2 overexpression promotes the proliferation and survival of laryngeal squamous cell carcinoma. Tumour Biol. 35, 3743–3753 (2014). Ahn, K. J. et al. Evaluation of the role of hexokinase type II in cellular proliferation and apoptosis using human hepatocellular carcinoma cell lines. J. Nucl. Med. 50, 1525– 1532 (2009). Robey, R. B. & Hay, N. Mitochondrial hexokinases, novel mediators of the antiapoptotic effects of growth factors and Akt. Oncogene 25, 4683–4696 (2006). Atsumi, T. et al. High expression of inducible 6-phosphofructo-2-kinase/fructose-2,6bisphosphatase (iPFK-2; PFKFB3) in human cancers. Cancer Res 62, 5881–5887 (2002). Yalcin, A., Telang, S., Clem, B. & Chesney, J. Regulation of glucose metabolism by 6-phosphofructo-2-kinase/fructose-2,6bisphosphatases in cancer. Exp. Mol. Pathol. 86, 174–179 (2009). Mazurek, S. Pyruvate kinase type M2: a key regulator of the metabolic budget system in tumor cells. Int. J. Biochem. Cell Biol. 43, 969–980 (2011). Christofk, H. R. et al. The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 452, 230–233 (2008). Christofk, H. R., Vander Heiden, M. G., Wu, N., Asara, J. M. & Cantley, L. C. Pyruvate kinase M2 is a phosphotyrosine-binding protein. Nature 452, 181–186 (2008). Israelsen, W. J. et al. PKM2 Isoform-Specific Deletion Reveals a Differential Requirement for Pyruvate Kinase in Tumor Cells. Cell 155, 397–409 (2013). Lunt, S. Y. & Vander Heiden, M. G. Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu. Rev. Cell Dev. Biol. 27, 441–464 (2011).

35.

36.

37.

38.

39.

40. 41.

42.

43.

44.

45.

46.

Vander Heiden, M. G. et al. Evidence for an alternative glycolytic pathway in rapidly proliferating cells. Science 329, 1492–1499 (2010). Shim, H. et al. c-Myc transactivation of LDH-A: implications for tumor metabolism and growth. Proc Natl Acad Sci USA 94, 6658–6663 (1997). Fantin, V. R., St-Pierre, J. & Leder, P. Attenuation of LDH-A expression uncovers a link between glycolysis, mitochondrial physiology, and tumor maintenance. Cancer Cell 9, 425–434 (2006). Fan, J. et al. Tyrosine Phosphorylation of Lactate Dehydrogenase A Is Important for NADH/NAD+ Redox Homeostasis in Cancer Cells. Mol Cell Biol 31, 4938–4950 (2011). Wang, Z.-Y. et al. LDH-A silencing suppresses breast cancer tumorigenicity through induction of oxidative stress mediated mitochondrial pathway apoptosis. Breast Cancer Res. Treat. 131, 791–800 (2012). Semenza, G. L. Tumor metabolism: cancer cells give and take lactate. J. Clin. Invest. 118, 3835–3837 (2008). Kim, J.-W., Gao, P., Liu, Y.-C., Semenza, G. L. & Dang, C. V. Hypoxia-inducible factor 1 and dysregulated c-Myc cooperatively induce vascular endothelial growth factor and metabolic switches hexokinase 2 and pyruvate dehydrogenase kinase 1. Mol Cell Biol 27, 7381–7393 (2007). Papandreou, I., Cairns, R. A., Fontana, L., Lim, A. L. & Denko, N. C. HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab 3, 187–197 (2006). Bonnet, S. et al. A mitochondria-K+ channel axis is suppressed in cancer and its normalization promotes apoptosis and inhibits cancer growth. Cancer Cell 11, 37–51 (2007). Mcfate, T. et al. Pyruvate dehydrogenase complex activity controls metabolic and malignant phenotype in cancer cells. J Biol Chem 283, 22700–22708 (2008). Gross, S. et al. Cancer-associated metabolite 2-hydroxyglutarate accumulates in acute myelogenous leukemia with isocitrate dehydrogenase 1 and 2 mutations. J Exp Med 207, 339–344 (2010). Parsons, D. W. et al. An integrated genomic analysis of human glioblastoma multiforme. Science 321, 1807–1812 (2008).

Chapter 1

23.

47.

48. 49. 50.

51.

52.

53.

54. 55. 56.

57.

58.

59.

60.

Mardis, E. R. et al. Recurring mutations found by sequencing an acute myeloid leukemia genome. N Engl J Med 361, 1058– 1066 (2009). Yan, H. et al. IDH1 and IDH2 mutations in gliomas. N Engl J Med 360, 765–773 (2009). Dang, L. et al. Cancer-associated IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–744 (2009). Ward, P. S. et al. The common feature of leukemia-associated IDH1 and IDH2 mutations is a neomorphic enzyme activity converting alpha-ketoglutarate to 2-hydroxyglutarate. Cancer Cell 17, 225–234 (2010). Figueroa, M. E. et al. Leukemic IDH1 and IDH2 mutations result in a hypermethylation phenotype, disrupt TET2 function, and impair hematopoietic differentiation. Cancer Cell 18, 553–567 (2010). Chowdhury, R. et al. The oncometabolite 2-hydroxyglutarate inhibits histone lysine demethylases. EMBO Rep 12, 463–469 (2011). Xu, W. et al. Oncometabolite 2-hydroxyglutarate is a competitive inhibitor of α-ketoglutarate-dependent dioxygenases. Cancer Cell 19, 17–30 (2011). Lu, C. et al. IDH mutation impairs histone demethylation and results in a block to cell differentiation. Nature 483, 474–478 (2012). Turcan, S. et al. IDH1 mutation is sufficient to establish the glioma hypermethylator phenotype. Nature 483, 479–483 (2012). Sasaki, M. et al. IDH1(R132H) mutation increases murine haematopoietic progenitors and alters epigenetics. Nature 488, 656–659 (2012). Gottlieb, E. & Tomlinson, I. P. M. Mitochondrial tumour suppressors: a genetic and biochemical update. Nat Rev Cancer 5, 857–866 (2005). King, A., Selak, M. A. & Gottlieb, E. Succinate dehydrogenase and fumarate hydratase: linking mitochondrial dysfunction and cancer. Oncogene 25, 4675–4682 (2006). Bayley, J.-P. & Devilee, P. Warburg tumours and the mechanisms of mitochondrial tumour suppressor genes. Barking up the right tree? Curr Opin Genet Dev 20, 324–329 (2010). Kaelin, W. G. & Ratcliffe, P. J. Oxygen sensing by metazoans: the central role of the HIF hydroxylase pathway. Mol Cell 30, 393–402 (2008).

61. 62.

63. 64.

65.

66.

67. 68.

69.

70.

71. 72. 73.

74.

Semenza, G. L. HIF-1: upstream and downstream of cancer metabolism. Curr Opin Genet Dev 20, 51–56 (2010). Wise, D. R. et al. Myc regulates a transcriptional program that stimulates mitochondrial glutaminolysis and leads to glutamine addiction. PNAS 105, 18782– 18787 (2008). Nicklin, P. et al. Bidirectional Transport of Amino Acids Regulates mTOR and Autophagy. Cell 136, 521–534 (2009). Lobo, C. et al. Inhibition of glutaminase expression by antisense mRNA decreases growth and tumourigenicity of tumour cells. Biochem J 348 Pt 2, 257–261 (2000). Gao, P. et al. c-Myc suppression of miR23a/b enhances mitochondrial glutaminase expression and glutamine metabolism. Nature 458, 762–765 (2009). Wang, J.-B. et al. Targeting mitochondrial glutaminase activity inhibits oncogenic transformation. Cancer Cell 18, 207–219 (2010). Zamboni, N. Deficiency in glutamine but not glucose induces MYC-dependent apoptosis in human cells. 178, 93–105 (2007). Riganti, C., Gazzano, E., Polimeni, M., Aldieri, E. & Ghigo, D. The pentose phosphate pathway: an antioxidant defense and a crossroad in tumor cell fate. Free Radic Biol Med 53, 421–436 (2012). Jonas, S. K. et al. Increased activity of 6-phosphogluconate dehydrogenase and glucose-6-phosphate dehydrogenase in purified cell suspensions and single cells from the uterine cervix in cervical intraepithelial neoplasia. Br J Cancer 66, 185–191 (1992). Vizán, P. & Cascante, M. Modulation of pentose phosphate pathway during cell cycle progression in human colon adenocarcinoma cell line HT29. 124, 2789–2796 (2009). Zhao, J. & Zhong, C.-J. A review on research progress of transketolase. Neurosci Bull 25, 94–99 (2009). Santos, C. R. & Schulze, A. Lipid metabolism in cancer. FEBS J 279, 2610–2623 (2012). Medes, G., Thomas, A. & Weinhouse, S. Metabolism of neoplastic tissue. IV. A study of lipid synthesis in neoplastic tissue slices in vitro. Cancer Res 13, 27–29 (1953). Kuhajda, F. P. et al. Fatty acid synthesis: a potential selective target for antineoplastic therapy. Proc Natl Acad Sci USA 91, 6379– 6383 (1994).

76.

77.

78. 79.

80.

81.

82.

83. 84.

85. 86. 87.

88.

Menendez, J. A. & Lupu, R. Fatty acid synthase and the lipogenic phenotype in cancer pathogenesis. Nat Rev Cancer 7, 763–777 (2007). Flavin, R., Peluso, S., Nguyen, P. L. & Loda, M. Fatty acid synthase as a potential therapeutic target in cancer. Future Oncol 6, 551–562 (2010). Bauer, D. E., Hatzivassiliou, G., Zhao, F., Andreadis, C. & Thompson, C. B. ATP citrate lyase is an important component of cell growth and transformation. Oncogene 24, 6314–6322 (2005). Hatzivassiliou, G. et al. ATP citrate lyase inhibition can suppress tumor cell growth. Cancer Cell 8, 311–321 (2005). Chajès, V., Cambot, M., Moreau, K., Lenoir, G. M. & Joulin, V. Acetyl-CoA carboxylase alpha is essential to breast cancer cell survival. Cancer Res 66, 5287–5294 (2006). Beckers, A. et al. Chemical inhibition of acetyl-CoA carboxylase induces growth arrest and cytotoxicity selectively in cancer cells. Cancer Res 67, 8180–8187 (2007). Glunde, K., Bhujwalla, Z. M. & Ronen, S. M. Choline metabolism in malignant transformation. Nat Rev Cancer 11, 835–848 (2011). Schafer, Z. et al. Antioxidant and oncogene rescue of metabolic defects caused by loss of matrix attachment. Nature (2009). doi:10.1038/nature08268 Carracedo, A. et al. A metabolic prosurvival role for PML in breast cancer. J. Clin. Invest. 122, 3088–3100 (2012). Faubert, B., Berger, S. L., Jones, R. G., Thompson, C. B. & Mak, T. W. Carnitine palmitoyltransferase 1C promotes cell survival and tumor growth under conditions of metabolic stress. 25, 1041–1051 (2011). Lowe, S. W., Cepero, E. & Evan, G. Intrinsic tumour suppression. Nature 432, 307–315 (2004). Mathew, R., Karantza-Wadsworth, V. & White, E. Role of autophagy in cancer. Nat Rev Cancer 7, 961–967 (2007). Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88, 593–602 (1997). Sherr, C. J. Divorcing ARF and p53: an unsettled case. Nat Rev Cancer 6, 663–673 (2006).

89.

Campisi, J. Senescent cells, tumor suppression, and organismal aging: good citizens, bad neighbors. Cell 120, 513–522 (2005). 90. Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). 91. Stein, G. H., Drullinger, L. F., Robetorye, R. S., Pereira-Smith, O. M. & Smith, J. R. Senescent cells fail to express cdc2, cycA, and cycB in response to mitogen stimulation. Proc Natl Acad Sci USA 88, 11012–11016 (1991). 92. Dimri, G. P. et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci USA 92, 9363–9367 (1995). 93. Narita, M. et al. Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113, 703–716 (2003). 94. Zhang, R., Chen, W. & Adams, P. D. Molecular dissection of formation of senescenceassociated heterochromatin foci. Mol Cell Biol 27, 2343–2358 (2007). 95. Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). 96. Acosta, J. C. et al. Chemokine signaling via the CXCR2 receptor reinforces senescence. Cell 133, 1006–1018 (2008). 97. Wajapeyee, N., Serra, R. W., Zhu, X., Mahalingam, M. & Green, M. R. Oncogenic BRAF Induces Senescence and Apoptosis through Pathways Mediated by the Secreted Protein IGFBP7. Cell 132, 363–374 (2008). 98. Kuilman, T. & Peeper, D. S. Senescencemessaging secretome: SMS-ing cellular stress. Nat Rev Cancer 9, 81–94 (2009). 99. Coppé, J.-P. et al. Senescence-associated secretory phenotypes reveal cellnonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol. 6, 2853–2868 (2008). 100. Rodier, F. et al. Persistent DNA damage signalling triggers senescence-associated inflammatory cytokine secretion. Nat Cell Biol 11, 973–979 (2009). 101. Acosta, J. C. et al. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat Cell Biol 15, 978–990 (2013). 102. HAYFLICK, L. THE LIMITED IN VITRO LIFETIME OF HUMAN DIPLOID CELL STRAINS. Exp Cell Res 37, 614–636 (1965).

Chapter 1

75.

103. Campisi, J. & d’Adda di Fagagna, F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol 8, 729–740 (2007). 104. Di Leonardo, A., Linke, S. P., Clarkin, K. & Wahl, G. M. DNA damage triggers a prolonged p53-dependent G1 arrest and long-term induction of Cip1 in normal human fibroblasts. Genes Dev 8, 2540–2551 (1994). 105. Robles, S. J. & Adami, G. R. Agents that cause DNA double strand breaks lead to p16INK4a enrichment and the premature senescence of normal fibroblasts. Oncogene 16, 1113– 1123 (1998). 106. Collado, M. et al. Tumour biology: senescence in premalignant tumours. Nature 436, 642 (2005). 107. Sarkisian, C. J. et al. Dose-dependent oncogene-induced senescence in vivo and its evasion during mammary tumorigenesis. Nat Cell Biol 9, 493–505 (2007). 108. Mo, L. et al. Hyperactivation of Ha-ras oncogene, but not Ink4a/Arf deficiency, triggers bladder tumorigenesis. J. Clin. Invest. 117, 314–325 (2007). 109. Chen, X. et al. Endogenous expression of Hras(G12V) induces developmental defects and neoplasms with copy number imbalances of the oncogene. PNAS 106, 7979–7984 (2009). 110. Yamakoshi, K. et al. Real-time in vivo imaging of p16Ink4a reveals cross talk with p53. J Cell Biol 186, 393–407 (2009). 111. Sun, P. et al. PRAK is essential for ras-induced senescence and tumor suppression. Cell 128, 295–308 (2007). 112. Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 41, 544–552 (2009). 113. Dhomen, N. et al. Oncogenic Braf induces melanocyte senescence and melanoma in mice. Cancer Cell 15, 294–303 (2009). 114. Dankort, D. et al. A new mouse model to explore the initiation, progression, and therapy of BRAFV600E-induced lung tumors. Genes Dev 21, 379–384 (2007). 115. Lazzerini Denchi, E., Attwooll, C., Pasini, D. & Helin, K. Deregulated E2F activity induces hyperplasia and senescence-like features in the mouse pituitary gland. Mol Cell Biol 25, 2660–2672 (2005). 116. Chen, Z. et al. Crucial role of p53-dependent cellular senescence in suppression of Pten-

117.

118.

119.

120.

121.

122. 123. 124.

125.

126. 127. 128. 129. 130.

deficient tumorigenesis. Nature 436, 725– 730 (2005). Majumder, P. K. et al. A prostatic intraepithelial neoplasia-dependent p27 Kip1 checkpoint induces senescence and inhibits cell proliferation and cancer progression. Cancer Cell 14, 146–155 (2008). Nardella, C. et al. Aberrant Rhebmediated mTORC1 activation and Pten haploinsufficiency are cooperative oncogenic events. Genes Dev 22, 2172–2177 (2008). Wu, C.-H. et al. Cellular senescence is an important mechanism of tumor regression upon c-Myc inactivation. Proc Natl Acad Sci USA 104, 13028–13033 (2007). Signoretti, S. & Kaelin, W. G. VHL loss actuates a HIF-independent senescence programme mediated by Rb and p400. 10, 361–369 (2008). Shamma, A. et al. Rb Regulates DNA damage response and cellular senescence through E2F-dependent suppression of N-ras isoprenylation. Cancer Cell 15, 255–269 (2009). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Gray-Schopfer, V. C. et al. Cellular senescence in naevi and immortalisation in melanoma: a role for p16? Br J Cancer 95, 496–505 (2006). Courtois-Cox, S. et al. A negative feedback signaling network underlies oncogeneinduced senescence. Cancer Cell 10, 459– 472 (2006). Zwerschke, W. et al. Metabolic analysis of senescent human fibroblasts reveals a role for AMP in cellular senescence. Biochem J 376, 403–411 (2003). Kondoh, H. et al. Glycolytic enzymes can modulate cellular life span. Cancer Res 65, 177–185 (2005). Bensaad, K. et al. TIGAR, a p53-inducible regulator of glycolysis and apoptosis. Cell 126, 107–120 (2006). Madan, E. et al. TIGAR induces p53-mediated cell-cycle arrest by regulation of RB-E2F1 complex. Br J Cancer 107, 516–526 (2012). Dörr, J. R. et al. Synthetic lethal metabolic targeting of cellular senescence in cancer therapy. Nature 501, 421–425 (2013). Haferkamp, S. et al. Vemurafenib Induces Senescence Features in Melanoma Cells. J Investig Dermatol 133, 1601–1609 (2013).

144. Milkevitch, M. et al. Increases in NMR-visible lipid and glycerophosphocholine during phenylbutyrate-induced apoptosis in human prostate cancer cells. Biochim Biophys Acta 1734, 1–12 (2005).

Chapter 1

131. Haq, R. et al. Oncogenic BRAF Regulates Oxidative Metabolism via PGC1&alpha; and MITF. Cancer Cell 23, 302–315 (2013). 132. Jiang, P., Du, W., Mancuso, A., Wellen, K. E. & Yang, X. Reciprocal regulation of p53 and malic enzymes modulates metabolism and senescence. Nature 493, 689–693 (2013). 133. Kaplon, J. et al. A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112 (2013). 134. Li, M. et al. Oncogene-induced cellular senescence elicits an anti-Warburg effect. Proteomics 13, 2585–2596 (2013). 135. Aird, K. M. et al. Suppression of nucleotide metabolism underlies the establishment and maintenance of oncogene-induced senescence. CellReports 3, 1252–1265 (2013). 136. Sukhatme, V. P. & Chan, B. Glycolytic cancer cells lacking 6-phosphogluconate dehydrogenase metabolize glucose to induce senescence. FEBS Lett 586, 2389–2395 (2012). 137. Unterluggauer, H. et al. Premature senescence of human endothelial cells induced by inhibition of glutaminase. Biogerontology 9, 247–259 (2008). 138. Maeda, M., Scaglia, N. & Igal, R. A. Regulation of fatty acid synthesis and Delta9-desaturation in senescence of human fibroblasts. Life Sci. 84, 119–124 (2009). 139. Quijano, C. et al. Oncogene-induced senescence results in marked metabolic and bioenergetic alterations. cc 11, 1383–1392 (2012). 140. Gey, C. & Seeger, K. Metabolic changes during cellular senescence investigated by proton NMR-spectroscopy. Mechanisms of Ageing and Development 134, 130–138 (2013). 141. Aboagye, E. O. & Bhujwalla, Z. M. Malignant transformation alters membrane choline phospholipid metabolism of human mammary epithelial cells. Cancer Res 59, 80–84 (1999). 142. Iorio, E. et al. Alterations of choline phospholipid metabolism in ovarian tumor progression. Cancer Res 65, 9369–9376 (2005). 143. Glunde, K., Bhujwalla, Z. M. & Ronen, S. M. Choline metabolism in malignant transformation. Nat Rev Cancer 11, 835–848 (2011).

足足足

CHAPTER 2 TWO-WAY COMMUNICATION BETWEEN THE METABOLIC AND CELL CYCLE MACHINERIES: THE MOLECULAR BASIS

Manuscript in preparation

TWO-WAY COMMUNICATION BETWEEN THE METABOLIC AND CELL CYCLE MACHINERIES: THE MOLECULAR BASIS Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam,

The Netherlands. *These authors contributed equally to this work.

SUMMARY The relationship between cellular metabolism and the cell cycle machinery is by no means unidirectional. The ability of a cell to enter the cell cycle critically depends on the availability of metabolites. Conversely, the cell cycle machinery commits to regulating metabolic networks in order to support cell survival and proliferation. In this review, we will give an account of how proliferation pathways, the cell cycle machinery and metabolism are interconnected, focusing on the reciprocal regulation of the latter two. Acquiring information on how communication takes place among metabolic signaling networks and the cell cycle controllers is crucial to increase our understanding of the deregulation thereof in disease, including cancer. Resting cells require a basal level of catabolic metabolism to ensure energy homeostasis. Cells that commit to entering the cell cycle, however, differ greatly from resting cells in terms of their metabolic profile, as they will eventually have to double their cell content, that is, their DNA, membranes, organelles and other biomass. To support the energy-consuming processes needed for this program, cells increase the uptake of glucose and glutamine and shut down oxidative metabolism. In this way, glucose and glutamine-derived metabolic intermediates can be used for the biosynthesis of macromolecules required for the cell division. Highly proliferating cells, including cancer cells but also activated lymphocytes, thymocytes and embryonic cells, preferentially use glycolysis even in the presence of oxygen1-9. This phenomenon is called aerobic glycolysis or â&#x20AC;&#x153;the Warburg effectâ&#x20AC;? 10,11. In unicellular organisms, cell cycle progression is dependent on the availability of nutrients, which directly couples available resources to the generation of progeny. For example, stationary-phase yeast switches to a mitotic phenotype when exposed to glucose, but becomes quiescent or sporulates when no other nutrients are provided12. Under nutrientsteady growth conditions, cycling yeast cells display fluctuations in oxygen consumption, alternating between glycolysis and respiration. Their cell division is solely limited to the glycolytic phase, with DNA replication taking place only during that period13. Interestingly, many genes identified in classic screens for factors regulating the cell cycle in yeast, were later shown to have a function in metabolic regulation, too14-19. Also, transcriptome studies demonstrated that in yeast, genes involved in glycolysis, respiration, lipids and amino acid

Chapter 2

Joanna Kaplon1*, Loes van Dam1* and Daniel Peeper1#

synthesis are cyclically expressed as a function of the cell cycle20,21. Taken together, these observations show that in unicellular organisms, intimate connections between cell cycle and metabolism must exist. In contrast to single-cell eukaryotes, cells of multicellular organisms usually have an unlimited access to nutrients. However, they are not cell-autonomous for nutrient uptake but instead depend on proliferation pathways. Mitogen-mediated activation of these signaling routes triggers nutrient uptake and is the rate-limiting cue for cell cycle entry22. As a consequence, growth factor-stimulated cells initiate cell division in a fashion comparable to that of yeast exposed to a nutrient-rich medium23,24. Accordingly, in the absence of mitogens, even in a nutrient-rich environment, cells will not enter the cell cycle25. On the other hand, even in the presence of promitogenic cues, glucose deprivation will keep cells from proliferating, which is a widely used method for synchronizing mammalian cells26,27. The fact that signaling pathways coordinating cell cycle progression control and are controlled by changes in cellular metabolism28,29 shows that, also in multicellular organisms, there must be a crosstalk between these pathways, cell cycle and metabolism. Yet, the molecular basis that connects nutrient availability, biosynthetic intermediates and energetic balance to the core cell cycle machinery remains incompletely understood. Here, we will discuss how proliferation pathways, the cell cycle machinery and metabolism are interconnected, but we will focus on the reciprocal regulation of the latter two. PROLIFERATION PATHWAYS LINK CELL CYCLE WITH METABOLIC REGULATION When nutrients are abundant, proliferation pathway activation is the most important cue for the cellular decision whether to proliferate. Importantly, work from recent years clearly demonstrates that proliferation pathways regulate not only the cell cycle but also cellular metabolism. They do so by several ways and at the levels of glucose uptake and (aerobic) glycolysis, lipogenesis, protein and nucleotides synthesis, as well as glutamine utilization (Figure 1). PI3K-AKT signaling modulates all major metabolic pathways to promote the metabolic profile required by proliferating cells (Figure 1). AKT stimulates glucose uptake and glycolysis by the upregulation of the glucose transporter GLUT130-32 and glycolysis-stimulating enzymes hexokinase 2 (HK2) and phosphofructokinase 1 (PFK1)31,33-36. PI3K-AKT is also involved in the regulation of synthesis of biomass. As such, it activates sterol regulatory element-binding proteins (SREBPs) and consequently, lipid synthesis37,38. Moreover, AKT directly activates ATP citrate lyase (ACLY), an enzyme shuttling glucose-derived citrate out of the mitochondria for lipogenesis39-41. PI3K-AKT also promotes glutamine metabolism through the mTORC1mediated activation of glutamate dehydrogenase (GDH), an enzyme utilizing glutaminederived glutamate42. This is part of a positive feedback loop, in which glutamine itself

RTK

Chapter 2

stimulates mTORC143,44. Through the activation of mTORC1, PI3K-AKT signaling also directs available amino acids to protein synthesis3,45. In line with these findings, PTEN, the PI3K antagonist, has been shown to function in metabolism to oppose the PI3K-AKT46.

MAPK pathway

Ras

PI3K PTEN

Raf

AKT

PI3K-AKT pathway

GPCR

FOXO

MEK

ERK

c-Myc

GLYCOLYSIS

mTOR

OXIDATIVE METABOLISM

PROTEIN & NUCLEOTIDES SYNTHESIS LIPOGENESIS

GLUTAMINOLYSIS

Figure 1. Regulation of metabolism by PI3K-AKT and MAPK pathways A simplified illustration of the PI3K-AKT and MAPK pathways and their effect on cellular metabolism. The effects indicated by arrows may be direct or indirect.

Oncogenic transformation of cells upon activation of the MAPK pathway is also coupled to changes in metabolic profiles (Figure 1). Similar to PI3K-AKT signaling, activation of the MAPK pathway increases glucose uptake and glycolysis47-50. RAS stimulates glycolysis by increasing the levels of the glycolytic activator fructose-2,6-bisphosphate (F-2,6-BP), a product of 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase 3 (PFKFB3), which is required for RAS-induced transformation51,52. Moreover, BRAF, another key component of MAPK signaling, regulates the activity of pyruvate dehydrogenase (PDH), an enzyme responsible for directing glucose-derived pyruvate towards oxidation in the mitochondrial tricarboxylic acid (TCA) cycle53. Oncogenic transformation by RAS is also associated with lipid accumulation; RAS-driven tumors can be attenuated by inhibiting lipogenesis pathways54,55. Functioning downstream of the MAPK and PI3K-AKT pathways, the c-Myc transcription factor regulates genes involved in both cell cycle and metabolism and indeed provides a molecular link between the two (Figure 1). c-Myc stimulates glycolysis by direct regulation of glucose transporters and several glycolytic enzymes56-63. It also plays a role in maintaining the biogenesis and function of the mitochondria. Accordingly, inhibiting c-Myc in tumor cells causes mitochondrial dysfunction64-67. Moreover c-Myc induces several genes involved in nucleotide metabolism, thereby promoting DNA synthesis (Figure 1)56,68,69. Genes involved in glutamine metabolism are under the control of c-Myc, too70,71. By altering cellular

metabolism in several ways, c-Myc not only drives cells into cell cycle progression, but also fulfills their energetic and biosynthetic needs during cell proliferation. In addition to stimulation of c-Myc, PI3K-AKT signaling regulates cellular metabolism by inhibition of forkhead box O (FOXO) transcription factors (Figure 1)72. FOXOs are active under conditions of nutrient or growth factor deprivation and cellular stress. In these settings, FOXOs induce cell cycle arrest and activate a stress response, thereby contributing to maintaining cellular homeostasis. Along these lines, many genes regulated by FOXO are metabolic enzymes73. CELL CYCLE REGULATION OF METABOLISM As described above, proliferation pathways promote specific metabolic profiles required by proliferating cells. Besides, evidence is emerging in support of the coordinated temporal regulation of metabolism directly by the cell cycle modulators. A first indication for this came from the observation that in yeast, metabolites of nucleotides, proteins and lipid synthesis are cyclically fluctuating, as a function of cell cycle progression74. Indeed, it has been shown subsequently that the glycolysis-promoting enzyme PFKFB3 is subjected to cell cycle dependent temporal regulation by members of the ubiquitin proteasome system (UPS; Figure 2, upper panel)75,76. Since then, a number of mechanisms have been revealed that couple the cellular metabolic state to the cell cycle (Figure 3). Most somatic cells are differentiated and quiescent, that is, they reside in the G0 phase of the cell cycle. Following mitogenic stimulation, cells go through the preparatory G1 phase. Upon passage through the G1/S restriction point, cell enter S phase in which they double their DNA content, move on into the G2 phase and the final mitotic (M) phase, in which cellular contents are divided over two daughter cells (Figure 3). Key proteins for the tight regulation of the cell cycle are cyclin-dependent kinases (CDKs), which associate with one of different cyclins across the cell cycle to ensure accurate cell cycle progression77-80. CDKs are constitutively expressed and regulate cell cycle substrates by phosphorylation. In contrast, the activity of cyclins is predominantly determined by changes in their protein and mRNA expression levels in response to mitogens, whether directly or indirectly (Figure 2, lower panel). The kinase activity of cyclin-CDK complexes is tightly regulated by a plethora of CDK inhibitors (CKIs), which stop cell cycle progression in unfavorable circumstances81. D-type cyclins In the G1 phase, cells need to prepare a sufficient supply of metabolites, macromolecules and energy to initiate and complete the cell division. Therefore, factors controlling G1-S transition must also regulate metabolism. D-type cyclins, which act in G1, are involved in ensuring the necessary metabolic profile at that phase of the cell cycle. The role of D-type cyclins in metabolism was first demonstrated in cyclin D-deficient mice that display marked metabolic phenotypes. Cyclin D2-deficient mice show a diabetic

SCF complex

1/2

APC

6 K4/ -CD lin D

cyc

lin

cli

E-C

-C

cyc

abundance

cyclin B-CDK1

p53

/C CDC20

APC/CCDH1

Chapter 2

id sy d lip in an

glutam

prote

inolys is

(GLS1 )

nthe

sis

glyco lysis (PFK FB3) synth esis otide nucle

PPP

abundance

O-GlcNAcylation

Figure 2. Protein activities and metabolic events during the cell cycle A schematic representation of the temporal regulation of metabolic factors (upper panel) and the cell cycle machinery (lower panel). The represented protein levels are not relative, but rather indicate their relative timing of expression.

phenotype due to impaired pancreatic b-cell expansion and function, which is further enhanced by cyclin D1 co-depletion82,83. Cyclin D3-deficient mice display reduced adipocyte size and increased sensitivity to insulin, which is a consequence of the inactivation of peroxisome proliferator-activated receptor (PPAR)g, the master regulator of adipogenesis (Figure 3)84. Accordingly depletion of cyclin D3 diminishes PPARg activity and adipogenesis, while cyclin D3 overexpression has the opposing effect. Cyclin D1 is highly expressed in breast cancer cells; among other functions, it decreases the abundance and activity of the glycolytic enzyme hexokinase 2 (HK2; Figure 3)85. Correspondingly, cyclin D1 depletion in either normal or transformed breast cells leads to an increase in glycolytic enzyme pyruvate kinase (PK) and the lipogenic enzymes acetyl-CoA carboxylase (ACC) and fatty acid synthase levels (FASN; Figure 3)85. In hepatocytes, cyclin D1 inhibits the glucose-mediated induction of central lipogenic genes via repression of the carbohydrate response element binding protein (ChREBP) and hepatocyte nuclear factor 4Îą (HNF4Îą), which are important regulators

of glucose sensing and lipid metabolism (Figure 3)86. Hence, cyclin D1 inhibits both glycolysis and lipogenesis. Moreover, cyclin D1 hinders mitochondrial biogenesis and function through inhibition of the nuclear respiratory factor 1 (NRF1), which regulates nuclearencoded mitochondrial genes, and the mitochondrial voltage-dependent anion channel (VDAC), respectively (Figure 3)87,88. Finally, unbiased mass-spectrometry analysis of proteins interacting with cyclin D1 using cyclin D1 knockin mice revealed interactions between cyclin D1 and numerous metabolic proteins, such as lipogenic enzymes FASN and ACC, as well as the mitochondrial electron transport chain components cytochrome c oxidase (COX) and ATP synthase (Figure 3)89. These observations together demonstrate that cyclin D1, by hindering glycolysis, lipogenesis as well as mitochondrial activity, downregulates metabolic activity through several routes. Restraining the conversion of glucose to lipids for storage by D-type cyclins might allow cells to use glucose-derived metabolites for doubling their cell content for cell division to take place. Alternatively, cyclin D-mediated inhibition of metabolism might provide a negative feedback loop to assure unidirectionality of the cell cycle90. Cyclin D activity accumulates in G1 when metabolic activity is abundant (Figure 2)91. If metabolic activity increases cyclin D expression and, in turn, cyclin D inhibits metabolism, a rise in cyclin D in G1 would shut down metabolism, thereby preventing re-entry into G1 until the G1/S transition is complete. RB-E2F Cyclin-CDK couples regulate the phosphorylation state of the retinoblastoma tumor suppressor (RB); its phosphorylation abolishes inhibition of physically associated E2F/ DP transcription factors, thereby allowing the expression of genes required for DNA synthesis92-95. Thus, phosphorylation of RB in late G1 by cyclin D-CDK4/6 converts E2F/DP from transcriptional repressor into activator, thereby promoting S-phase entry (Figure 3). As E2F transcription factors are involved in cell cycle progression and survival, it is not surprising that roles for increased E2F activity have been shown in cancer95. Evidently, E2F promotes cancer cells proliferation through transcriptional activation of genes involved in cell cycle regulation and DNA synthesis, but a role for E2F in metabolism has also emerged. E2F1 as well as its upstream activator CDK4 have been shown to regulate adipogenesis by positively regulating PPARg (Figure 3)96,97. In accordance with that, RB represses PPARg at the early stages of adipocyte differentiation (Figure 3)98. On the other hand, RB acts positively on terminal adipocyte differentiation by binding directly to the transcription factor CCAAT/ enhancer binding protein (C/EBP)b, thereby facilitating its transactivation99. Besides adipogenesis, the RB-E2F pathway is involved in the glucose metabolism of pancreatic Î˛-cells. E2F1-depleted mice display a diabetic phenotype, resulting from glucose intolerance and deficient insulin secretion100. In line with this, E2F1 regulates Kir6.2, a key component of the KATP channel controlling glucose-induced insulin secretion (Figure 3)101. Expression of Kir6.2 is lost in the pancreas of E2F1-depleted mice, resulting in insulin secretion defects.

GLYCOLYSIS

Cell cycle machinery Metabolism

PPP PFKFB3

SCF

SAC

PKM2

APC/CCDC20

Cyclin B CDK1

p27

APC/CCDH1

GLUTAMINOLYSIS

RB G2

Cyclin D CDK4/6

CELL CYCLE

CIRCADIAN CLOCK GENES

GLS1

GLYCOLYSIS

Cyclin A CDK2

PPARϒ

p21 p27

Cyclin E CDK2

AMPK

SCF

E2F

FASN ACC HNF4α

LIPOGENESIS

PK HK2 ChREBP

GLYCOLYSIS

PDK2

p53

SIRT1

GLUT1&4 PGM MCT1

Kir6.2

OXPHOS

PFK2

CPT GAMT

LIPOGENESIS

SREBP1c FASN ACC

G6PDH RRM2

OXPHOS INSULIN RELEASE GLYCOLYSIS

DNA SYNTHESIS GLYCOLYSIS

TIGAR GLS2

ADIPOGENESIS

TopImt COX/ SCO2

FOXO3 NAD+/NADH ratio

OXPHOS

p21, p27

PDK4

low ATP/AMP

NRF1 VDAC COX ATP synthase

CIRCADIAN CLOCK GENES

E2F

Aldolase Enolase

EYA1

Chapter 2

CDC25C

GLUTAMINOLYSIS PPP

histone acetylation

histone glycosylation

GLYCOSYLATION acetyl-CoA

UDP-GlcNAc

OGA

OGT

Figure 3. Bidirectional regulation of the cell cycle machinery and metabolic enzymes Interactions between the cell cycle machinery and metabolic enzymes indicated by arrows may be either direct or indirect. Color codes represent proteins belonging to the cell cycle or to metabolism.

Furthermore, E2F1 was shown to induce expression of pyruvate dehydrogenase kinase 4 (PDK4). PDK4 is one of four kinases that inhibit PDH and thereby oxidative metabolism (Figure 3)102. Likewise, inactivation of RB by the oncogenic E1A adenoviral protein also triggers PDK4 expression (Figure 3). Induction of PDK4 by RB-E2F1 in myoblasts diverts pyruvate away from mitochondrial oxidation, thereby increasing lactate production and flux into biosynthetic pathways103. Furthermore, E2F1 stimulates glycolysis by upregulating the

expression of the PFK2 enzyme (Figure 3)104. Depletion of E2F1 increases the expression of regulators of mitochondrial biogenesis and function, such as mitochondrial topoisomerase I (TopImt)105. Correspondingly, E2F1 downregulates oxidative and increases glycolytic genes expression106. In this way, RB-E2F1 pathway promotes a switch from oxidative to glycolytic metabolism, thereby supporting the metabolic phenotype typical of proliferating cells. Altogether, the RB-E2F pathway provides yet another mechanism for the coordinated regulation of metabolism throughout the cell cycle by facilitating glycolysis and repressing oxidative phosphorylation (Figure 3). RB-E2F-mediated stimulation of adipogenesis and inhibition of oxidative metabolism are in line with the observations made for cyclin D1 and suggest that the latter acts on these metabolic pathways, at least to some extent, through RB-E2F. However, while RB-E2F promotes glycolysis, cyclin D1 suppresses it. This implies that cyclin D1 acts on glycolytic enzymes in RB-E2F-independent manner. CKIs As cyclin-CDK complexes regulate metabolism, and because the activities and functions of CDKcyclin complexes are regulated by the CKIs, the latter should also have a metabolic function. Indeed, p21Cip1 and p27Kip1 were demonstrated to modulate adipocyte differentiation, as loss of either of kinase inhibitor in mice induces adipocyte hyperplasia107. In accordance with this, combined disruption of p21Cip1 and p27Kip1 in mice induces an increase in the number of adipocytes and the development of hypercholesterolemia, glucose intolerance and insulin insensitivity, which are the features of obesity. Nevertheless, these observations require further investigation, as other studies have shown that p21Cip1 null mice are less prone to obesity induced by lipid-rich diet, whereas p27Kip1 null mice do not show hyperplasia of adipocytes but increased insulin secretion that prevents hyperglycemia in diabetic mice models108-110. The UPS and glycolysis Cyclins and CKIs are both subject to tight temporal control and degradation by the UPS. Two ubiquitin ligases are crucial in the cell cycle.Â The ligase anaphase-promoting complex or cyclosome (APC/C) regulates both the transition through G1 and the exit from M phase by degrading both S- and M-phase cyclins (cyclin A and B) and Securin (allowing chromosome separation). Directing of APC/C to correct substrates at specific time points in the cell cycle depends on one of two activators, CDC20 or CDC20 homologue (CDH1)111-113. The latter is expressed and active in mitosis, until it is replaced by CDH1 in late mitosis and the G1 phase (Figure 2 and 3)114. The ligase Skp1/cullin/F-box protein (SCF) complex controls the G1/Sphase and the G2/M transition. The F-box protein of the SCF complex regulates its subject recognition115,116. SCF-Skp2 mainly ubiquitinates and degrades CKIs such as p27Kip1 and p21Cip1, but also cyclin E, whereas SCF-Î˛-TrCP positively regulates APC/CCDC20 and CDK1 to ensure G2/M transition (Figure 3). APC/C and SCF control each other to regulate progression of the cell cycle117.

Chapter 2

A yeast F-box protein GRR1, a component of SCF complex regulating G1/S transition, was first identified as being essential for adaptation to nutrient availability. In response to extracellular glucose, SCFGRR1 blocks exit from the cell cycle and sporulation by targeting Ime2p kinase118. At the same time, SCFGRR1 represses genes required for the utilization of alternative carbon sources, and upregulates hexose transporters119-122. On the other hand, when glucose is removed, SCFGRR1-mediated degradation of Ime2p is abrogated, cells exit the cell cycle and sporulation proceeds118. Moreover, following glucose deprivation, SCFGRR1 inhibits glycolysis by degrading Pfk27, the yeast homolog of the glycolytic enzyme PFKFB3 (Figure 3)75. Thus, SCFGRR1 in yeasts regulates both cell cycle exit and metabolism as a function of glucose availability. Shortly after Pfk27 was identified as the target of SCFGRR1 in yeast, PFKFB3 was found to be a degraded by APC/CCDH1 in neurons (Figure 3)76. Constitutive breakdown of PFKFB3 helps terminally differentiated cortical neurons to preserve their low glycolytic state. It also prevents oxidative damage by redirecting glucose-derived metabolites into antioxidants-providing pentose phosphate pathway (PPP)123. In dividing cells, PFKFB3 is directed for degradation both by APC/CCDH1 during late mitosis and G1 and subsequently by SCFβ-TrCP in late S phase (Figure 2 and 3)124-127. Overexpression of the APC/C activator CDH1 leads to degradation of PFKFB3 and thereby restricts glycolysis. Accordingly, depletion of CDH1 promotes glycolysis in a PFKFB3-dependent manner and stimulates cells to enter S phase125,126. Inactivation of APC/CCDH1, resulting from a phosphorylation of CDH1 that occurs in late G1, leads to the accumulation of PFKFB3 and consequently promotes both proliferation and glycolysis (Figure 2). In the late S-phase, PFKFB3 levels drop again due to an increase in activity of SCFβTrCP , which specifically directs PFKFB3 for degradation (Figure 2)124,127. Restricted expression of PFKFB3 to the specific window of late G1 and early S phase generates a peak in anaerobic glycolysis during the G1/S transition and a PPP peak in S phase (Figure 2, upper panel). Thus, the joint action of the APC/C and SCF complexes on PFKFB3 coordinates metabolic activity and cell cycle progression. APC/CCDH1 was recently shown to regulate also phosphatase and transactivator EYA1128. The level of EYA1 protein oscillates in the cell cycle, peaking during mitosis and dropping radically as cells enter G1, when APC/CCDH1 reaches its peak level (Figure 2 and 3). While depletion of CDH1 stabilizes the EYA1 protein, overexpression of CDH1 reduces its levels. Thus, APC/CCDH1 degrades EYA1 precisely during M/G1 transition. EYA1 is required for proliferation during embryogenesis129-131 and its level is elevated in several cancers132-135. Interestingly, EYA1 is also known to reprogram the aerobic metabolism of slow-twitch muscle fibers, which depend on lipid oxidation, into the glycolytic phenotype of fast-twitch muscle fibers, which depend mostly on glycogen as an energy source. This is achieved through the upregulation of the glycolytic enzymes aldolase A and β-enolase (Figure 3)136. Although only correlative, this might provide another mechanism by which UPS regulates glycolysis.

The UPS and glutaminolysis Besides PFKFB3, the APC/CCDH1 also directs the glutaminolytic enzyme glutaminase 1 (GLS1) for degradation during mitotic exit and G1 (Figure 2 and 3)126,127. Glutamine is important for proliferation as it replenishes TCA cycle intermediates used for macrosyntheses of amino acids, lipids and nucleotides. GLS1 levels and the glutaminolysis rate rise after APC/CCDH1 activity declines in late G1. Accordingly, depletion of APC/C activator CDH1 leads to an increase in cellular GLS1 concentration and consequently glutamine metabolism. In contrast to PFKFB3, GLS1 is not a substrate for SCF and therefore it is not degraded through the S and G2 phases, but only when the cells progress to the G2/M transition (Figure 2 and 3). The distinct regulation of PFKFB3 and GLS1 proteins and therefore glycolysis and glutaminolysis, suggests the different functions of glucose and glutamine at particular phases of the cell cycle. Indeed, studies in synchronized cells show that both glucose and glutamine are required throughout G1, whereas only glutamine is needed to progress through S phase into the cell division phase127. The importance of regulating GLS1 for cell proliferation is further supported by the notion that GLS1 C, an isoform that is not targeted by the APC/CCDH1, is overexpressed in several tumors137. Taken together, while ubiquitination by the UPS was originally recognized for its major role in regulating the cell cycle machinery, it is now also acknowledged for its integrated regulation of metabolism and proliferation. The dual regulation by these two ubiquitin ligases, APC/C and the SCF complex, delineates a differential regulation of both aerobic glycolysis and glutaminolysis during distinct phases of the cell cycle. In this sense, the UPS plays a major role in the provision of a specific metabolic profile reminiscent of proliferating (cancer) cells, including the Warburg effect, upregulation of the PPP and increased glutamine utilization. p53 family members p53 is a major tumor suppressor protein, which is mutated in many types of cancer. The activation of p53 in response to stress provides the cell with two options: inhibition of the cell cycle at G1/S by inducing transcription of the CKI p21Cip1, or induction of pro-apoptotic signals138. This allows cells to either repair the damage before engaging in cellular division, or, if the damage is beyond repair, to prevent cells from proliferating ever again. Importantly, recent research shows that p53 is also a major regulator of cellular metabolism. Several functions of p53 have been demonstrated to silence glycolysis and promote oxidative metabolism. p53 hampers a flux through the various steps of glycolytic pathway by downregulating the expression of glucose transporters GLUT 1 and GLUT4, decreasing levels of the glycolytic enzyme phosphoglycerate mutase (PGM)139,140. It also induces TIGAR, an enzyme that lowers cellular F-2,6-BP and thereby inhibits glycolysis (Figure 3)141. Moreover, p53-mediated repression of monocarboxylate transporter 1 (MCT1) expression prevents the secretion of lactate under anaerobic conditions which also reduces glycolysis142. Inhibition of the glycolytic pathway would predict activation of PPP pathway. However, p53 evades this by

Chapter 2

binding and inhibiting the rate-limiting enzyme of PPP, glucose-6-phosphate dehydrogenase (G6PDH; Figure 3)143. p53 also downregulates the expression of the ribonucleotide reductase subunit M2 (RRM2), an enzyme controlling dNTP synthesis (Figure 3)144. Inhibition of glycolysis and PPP is paralleled by a role of p53 in promoting oxidative metabolism. Next to maintaining mitochondrial balance145,146, p53 stimulates oxidative phosphorylation by upregulating cytochrome c oxidase (mammalian COX/yeast SCO2), a component of mitochondrial electron transport chain (Figure 3)147,148. Moreover, p53 increases the TCA cycle rate by downregulating PDH-inhibitory kinase PDK2 and TCA cycle-associated malic enzymes (MEs; Figure 3)149,150. Additionally, p53 increases the expression of glutaminolytic enzyme glutaminase 2 (GLS2; Figure 3)151. GLS2 converts glutamine to glutamate, which can feed the TCA cycle, but can also participate in glutathione synthesis to control the redox state of the cell. Thus, by upregulating GLS2, p53 influences both glutamine metabolism and redox status. p53 also regulates the redox state upon serine withdrawal. It does so by allowing de novo serine to be channeled to glutathione synthesis, which occurs at the cost of nucleotide synthesis152. Finally, p53 was shown to negatively regulate lipogenesis (Figure 3). p53 inhibits fatty acid (FA) synthesis in mouse adipose tissue by suppressing the expression levels of SREBP1c transcription factor and FASN and ACLY enzymes (Figure 3)153. At the same time, p53 stimulates FA oxidation by inducing the expression of carnitine palmitoyltransferase (CPT), which is responsible for the transport of FA into the mitochondria and guanidinoacetate methyltransferase (GAMT), an enzyme involved in creatine synthesis (Figure 3)154,155. Hence, p53 promotes cell cycle arrest not only by acting on the cell cycle machinery, but also by counteracting metabolic profiles favorable for proliferation and by supporting oxidative metabolic reactions characteristics of resting cells. The other p53 family members p63 and p73, too, have metabolic functions, broadening the impact of the p53 family on cell metabolism. The Tp63 and Tp73 genes are transcribed from two separate promoters, encoding either full-length proteins that retain a full transactivation (TA) domain (TAp63 and TAp73) or N-terminally truncated isoforms (DNp63 and DNp73). Both TAp63 and TAp73 upregulate the expression of GLS2 and thereby increase glutaminolysis156,157. Moreover, TAp73 stimulates oxidative metabolism by upregulating COX subunit 4158. Correspondingly, the depletion of TAp73 results in a decrease in both respiration and ATP production. In contrast to p53, TAp73 activates the expression of G6PDH and the flux to PPP, therefore redirecting glucose for the synthesis of nucleotides and antioxidants159. TAp73 also regulates amino acids metabolism by increasing the levels of serine, glycine, and glutathione160. Similarly to p53, p63 was implicated in the regulation of lipid metabolism. Loss of TAp63 disrupts lipogenesis, FA synthesis and oxidation, and protects against insulin resistance in mice161. Likewise, DNp63 transcriptionally activates FASN162. Altogether, the p53 family provides yet another example of direct crosstalk between the cell cycle machinery and cellular metabolism.

METABOLIC REGULATION OF THE CELL CYCLE MACHINERY While it is important for a cell to provide sufficient building blocks to enable cell division, this mechanism also functions vice versa. It is equally important for a cell to adapt its cell cycle to the environment (i.e., changes in nutrient availability) and the state of metabolism. As early as in 1974 it was shown that cells in the absence of glucose arrest at the G1/S restriction point163. This demonstrated that glucose availability governs a metabolic cell cycle checkpoint. Since then, various mechanism through which cells synchronize their cell cycle with their metabolic state have been discovered, several of which are described below. PFKFB3 Beyond its metabolic activity, PFKFB3 has been described to localize to the nucleus and regulate the cell cycle machinery164. Overexpression of nuclear PFKFB3 is accompanied by increased expression of G1-promoting cyclin D3 (Figure 3). Moreover, nuclear PFKFB3 stimulates proliferation by increasing the expression of mitotic kinase CDK1 and M phasepromoting phosphatase Cdc25C and by decreasing the expression of the CDK1 inhibitor p27Kip1 (Figure 3). These effects are completely abrogated by mutating either the active site or nuclear localization residues of PFKFB3, demonstrating a requirement for nuclear delivery of F-2,6-BP in this setting. Consequently, addition of F-2,6-BP to cell lysates promotes CDK1â&#x20AC;? mediated p27Kip1 phosphorylation, which is a cue for p27Kip1 degradation. These data show that, PFKFB3 is not only subject to a tight regulation by the UPS, but it can also regulate the cell cycle machinery itself. PKM2 PKs catalyze the final rate-limiting step of glycolysis, generating ATP and pyruvate. PK isoform M2 (PKM2) is specifically enriched in highly proliferating cells, including cancer cells, where it regulates aerobic glycolysis165,166. Apart from its important metabolic function, PKM2 also has a non-metabolic role in the control of the cell cycle progression. In response to epidermal growth factor (EGF), PKM2 translocates to the nucleus where it binds to b-catenin and promotes its transcriptional activity167,168. PKM2-b-catenin complex subsequently localizes to the cyclin D1 (CCND1) and c-Myc promoters and enhances their expression through detachment of a histone deacetylase, HDAC3169. Thus, by inducing cyclin D1 expression, PKM2 regulates the G1-S phase transition (Figure 3). Besides PKM2â&#x20AC;&#x2122;s role in regulating the expression of cyclin D1, its direct involvement in the regulation of cell cycle progression has also been demonstrated. During mitosis, PKM2 binds to and phosphorylates the spindle assembly checkpoint (SAC) protein Bub3 (Figure 3)170. This phosphorylation is required for the formation of Bub3-Bub1-Blinkin complex and correct kinetochore microtubules attachment, allowing for proper chromosomal segregation. In this way, PKM2 regulates not only G1/S transition but also progression through mitosis. Acetyl-CoA Cytosolic and nuclear acetyl-CoA is not only an important intermediate for macrosyntheses,

Chapter 2

but also a precursor for the posttranslational modification of proteins by acetylation. For example, acetylation of histones is dependent on acetyl-CoA-producing ACLY. In various mammalian cell types, depletion of ACLY, and therefore a drop in acetyl-CoA levels, decreases histone acetylation171. On the other hand, in yeast, a rise in acetyl-CoA upon depletion of ACC increases the acetylation of histones172. Notably, acetylation of histones is an essential process in the release of DNA for replication and therefore for cell cycle progression173,174. Thus, by regulating the acetylation of histones, acetyl-CoA controls the cell cycle (Figure 3). In addition to acetylation, availability of acetyl-CoA regulates the glycosylation of proteins (Figure 3). O-linked N-acetylglucosamine (O-GlcNAc) modifications depend on UDP-GlcNAc, production of which is controlled by the availability of glucose, glutamine and acetylCoA175. O-GlcNAc-transferases (OGTs) transfer UDP-GlcNAc onto a target protein, while O-GlcNAcases (OGAs) remove it, permitting a dynamic regulation of O-GlcNAcylation levels. In recent years, numerous proteins have been identified as substrates for O足GlcNAc modification, including regulators of cell cycle progression. Moreover, O-GlcNAcylation levels have been found to vary along the cell cycle (Figure 2). Serum addition triggers G0/G1 transition by activating PI3K/AKT and MAPK pathways and transcription of cyclin D1 gene, a key regulator of G1 phase. Notably OGT levels are significantly increased following serumstimulation176 and OGT depletion delays G1 entry and prevents serum-induced cyclin D1 synthesis (Figure 3)177. Hence, OGT and therefore increased O-GlcNAcylation, are vital for entry into the cell cycle. Contrary to that, at the G1/S transition, global O-GlcNAcylation is decreased due to elevated OGA activity176. Interestingly, this leads to lower O-GlcNAcylation of histones, which, similarly to the histone acetylation mentioned above, promotes the relaxation of DNA, necessary for replication and cell cycle progression178,179. At the G2/M checkpoint, O-GlcNAcylation reaches its peak180-183. Accordingly, depletion of OGT decreases cyclin B1 expression and therefore impairs G2/M transition (Figure 3)180. Although we are only beginning to understand the importance of O-GlcNAcylation in regulating cell proliferation, the evidence discussed provides additional indications that the cell cycle machinery and metabolism are tightly intertwined. ATP/AMP ratio The ATP/AMP ratio reflects the energy status of the cell. The protein complex that plays a crucial role in regulating cellular energy is AMP-activated protein kinase (AMPK)184. Under conditions of energetic stress such as glucose deprivation and low ATP levels, activated AMPK negatively regulates energy-consuming metabolic processes such as protein and lipid synthesis, while at the same time it promotes energy-producing oxidative phosphorylation and FA oxidation (Figure 3)185. It has been recently suggested that AMPK can regulate energy levels also by direct control of the cell cycle machinery. Along these lines, activation of AMPK upon glucose deprivation or treatment with AMP analog AICAR, causes an arrest in the G1 phase186,187. This arrest is associated with phosphorylation of p53 at Ser15 and upregulation

of CKI p21Cip1 expression, a target of activated p53 (Figure 3). AMPK also phosphorylates the C-terminal residue of another CKI, p27Kip1, causing its stabilization (Figure 3)188. Likewise, methylene blue (MB)-mediated activation of AMPK reduces expression of cyclins A2, B1 and D1, leading to proliferative arrest189. Moreover, activation of AMPK due to mitochondrial dysfunction promotes p53-dependent transcription of the F‐box protein archipelago. Archipelago then recruits cyclin E to the SCF complex, resulting in cyclin E degradation and G1-S cell cycle arrest (Figure 3)190,191. As such, AMPK plays a pivotal role as a metabolic cell cycle checkpoint, preventing cell cycle entry in conditions of low nutrient availability. NAD+/NADH ratio The NAD+/ NADH ratio has an important role in the regulation of oxidative stress and is often considered to be a read out of the metabolic status. NAD+ is converted to NADH in catabolic reactions including glycolysis and the TCA cycle. To maintain a proper redox state, NADH is regenerated constantly via several mechanisms, such as oxidation in the mitochondrial respiratory chain and reduction of pyruvate to lactate in the last step of glycolysis38. NAD+ is a classical coenzyme mediating redox reactions38, but also plays an important role in regulation of NAD+ - consuming enzymes, including sirtuin family of NAD+-dependent deacetylases192. Notably, in recent years, several mechanisms linking sirtuins to the cell cycle machinery have been described. SIRT2 controls mitotic exit and acts as a checkpoint protein in cells treated with microtubule poisons193,194. Under genotoxic stress, SIRT1 deacetylates and hinders the activity of p53, thereby preventing p53-mediated transactivation of p21Cip1 and cell cycle arrest (Figure 3)195,196. SIRT1 also regulates the cell cycle by deacetylation of the FOXO3 transcription factor197. In response to oxidative stress, SIRT1 binds to and deacetylates FOXO3, resulting in increased levels of FOXO3 target p27Kip1 and induction of cell cycle arrest (Figure 3). Moreover, SIRT1 has been described to regulate the components of the circadian clock machinery198,199. Interestingly, among clock-controlled genes are those that have an essential role in cell cycle control, including cyclin D1 and inhibitor of CDK1cyclin B1 complex, Wee1200,201. Hence, by regulating the circadian clock, SIRT1 inhibits cells cycle progression in situations of stress (Figure 3). Taken together, sirtuins translate the NAD+/NADH ratio to the components of the cell cycle machinery. CONCLUDING REMARKS The combined regulation of metabolic events and components of the cell cycle machinery described here delineates an intricate relationship between the two. Their reciprocal regulation plays a pivotal role in the cellular decision to enter and progress through cell cycle and to adjust metabolic pathways. Cells strictly control the cell cycle machinery components, in order to provide themselves with the required macromolecules during specific stages of the cell cycle. However, sometimes the regulation is disturbed, which leads to pathologies including cancer as a prime example. Thus, studying the reciprocal regulatory networks

REFERENCES 1.

3. 4.

10.

11. 12.

Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033 (2009). Lunt, S. Y. & Vander Heiden, M. G. Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu. Rev. Cell Dev. Biol. 27, 441–464 (2011). Cantor, J. R. & Sabatini, D. M. Cancer cell metabolism: one hallmark, many faces. Cancer Discov 2, 881–898 (2012). Ward, P. S. & Thompson, C. B. Metabolic reprogramming: a cancer hallmark even warburg did not anticipate. Cancer Cell 21, 297–308 (2012). Hedeskov, C. J. Early effects of phytohaemagglutinin on glucose metabolism of normal human lymphocytes. Biochem J 110, 373–380 (1968). Brand, K. Glutamine and glucose metabolism during thymocyte proliferation. Pathways of glutamine and glutamate metabolism. Biochem J 228, 353–361 (1985). Wang, T., Marquardt, C. & Foker, J. Aerobic glycolysis during lymphocyte proliferation., Published online: 24 June 1976; | doi:10.1038/261702a0 261, 702–705 (1976). Steck, T. L., Kaufman, S. & Bader, J. P. Glycolysis in chick embryo cell cultures transformed by Rous sarcoma virus. Cancer Res 28, 1611–1619 (1968). Kondoh, H. et al. A high glycolytic flux supports the proliferative potential of murine embryonic stem cells. Antioxid Redox Signal 9, 293–299 (2007). Warburg, O., Wind, F. & Negelein, E. The metabolism of tumors in the body. The Journal of general physiology 8, 519–530 (1927). WARBURG, O. On the origin of cancer cells. Science 123, 309–314 (1956). Granot, D. & Snyder, M. Glucose induces cAMP-independent growth-related changes in stationary-phase cells of Saccharomyces

cerevisiae. Proc Natl Acad Sci USA 88, 5724– 5728 (1991). 13. Chen, Z., Odstrcil, E. A., Tu, B. P. & McKnight, S. L. Restriction of DNA replication to the reductive phase of the metabolic cycle protects genome integrity. Science 316, 1916–1919 (2007). 14. Hartwell, L. H., Culotti, J., Pringle, J. R. & Reid, B. J. Genetic control of the cell division cycle in yeast. Science 183, 46–51 (1974). 15. Jong, A. Y., Kuo, C. L. & Campbell, J. L. The CDC8 gene of yeast encodes thymidylate kinase. J Biol Chem 259, 11052–11059 (1984). 16. Maitra, P. K. & Lobo, Z. Pyruvate kinase mutants of Saccharomyces cerevisiae: biochemical and genetic characterisation. Mol Gen Genet 152, 193–200 (1977). 17. Game, J. C. Yeast cell-cycle mutant cdc21 is a temperature-sensitive thymidylate auxotroph. Mol Gen Genet 146, 313–315 (1976). 18. Dickinson, J. R. & Williams, A. S. The cdc30 mutation in Saccharomyces cerevisiae results in a temperature-sensitive isoenzyme of phosphoglucose isomerase. J. Gen. Microbiol. 133, 135–140 (1987). 19. Wrobel, C., Schmidt, E. V. & Polymenis, M. CDC64 encodes cytoplasmic alanyltRNA synthetase, Ala1p, of Saccharomyces cerevisiae. J. Bacteriol. 181, 7618–7620 (1999). 20. Cho, R. J. et al. A genome-wide transcriptional analysis of the mitotic cell cycle. Mol Cell 2, 65–73 (1998). 21. Spellman, P. T. et al. Comprehensive identification of cell cycle-regulated genes of the yeast Saccharomyces cerevisiae by microarray hybridization. Mol Biol Cell 9, 3273–3297 (1998). 22. Lloyd, A. C. The regulation of cell size. Cell 154, 1194–1205 (2013). 23. Boer, V. M., Crutchfield, C. A., Bradley, P. H., Botstein, D. & Rabinowitz, J. D. Growthlimiting intracellular metabolites in yeast

Chapter 2

linking the cell cycle and metabolism is likely to reveal potential targets for therapy. Precise mapping of the tight temporal regulation of metabolic events during the cell cycle events is an essential element for this.

24.

25.

26.

27.

28.

29.

30. 31. 32.

33.

34.

35.

growing under diverse nutrient limitations. Mol Biol Cell 21, 198–211 (2010). Conlon, I. & Raff, M. Differences in the way a mammalian cell and yeast cells coordinate cell growth and cell-cycle progression. J. Biol. 2, 7 (2003). Rathmell, J. C., Vander Heiden, M. G., Harris, M. H., Frauwirth, K. A. & Thompson, C. B. In the absence of extrinsic signals, nutrient utilization by lymphocytes is insufficient to maintain either cell size or viability. Mol Cell 6, 683–692 (2000). Kues, W. A. et al. Cell cycle synchronization of porcine fetal fibroblasts: effects of serum deprivation and reversible cell cycle inhibitors. Biol. Reprod. 62, 412–419 (2000). Langan, T. J. & Chou, R. C. Synchronization of mammalian cell cultures by serum deprivation. Methods Mol. Biol. 761, 75–83 (2011). Levine, A. J. & Puzio-Kuter, A. M. The Control of the Metabolic Switch in Cancers by Oncogenes and Tumor Suppressor Genes. Science 330, 1340–1344 (2010). Jones, R. G. & Thompson, C. B. Tumor suppressors and cell metabolism: a recipe for cancer growth. Genes Dev 23, 537–548 (2009). Barthel, A. et al. Regulation of GLUT1 gene transcription by the serine/threonine kinase Akt1. J Biol Chem 274, 20281–20286 (1999). Elstrom, R. L. et al. Akt stimulates aerobic glycolysis in cancer cells. Cancer Res 64, 3892–3899 (2004). Taha, C. et al. Opposite translational control of GLUT1 and GLUT4 glucose transporter mRNAs in response to insulin. Role of mammalian target of rapamycin, protein kinase b, and phosphatidylinositol 3-kinase in GLUT1 mRNA translation. J Biol Chem 274, 33085–33091 (1999). Deprez, J., Vertommen, D., Alessi, D. R., Hue, L. & Rider, M. H. Phosphorylation and activation of heart 6-phosphofructo-2-kinase by protein kinase B and other protein kinases of the insulin signaling cascades. J Biol Chem 272, 17269–17275 (1997). Gottlob, K. Inhibition of early apoptotic events by Akt/PKB is dependent on the first committed step of glycolysis and mitochondrial hexokinase. Genes Dev 15, 1406–1418 (2001). Majewski, N. et al. Hexokinase-Mitochondria Interaction Mediated by Akt Is Required to

36.

37.

38. 39.

40.

41.

42.

43. 44. 45.

46.

47.

48.

Inhibit Apoptosis in the Presence or Absence of Bax and Bak. Mol Cell 16, 819–830 (2004). Wieman, H. L., Wofford, J. A. & Rathmell, J. C. Cytokine stimulation promotes glucose uptake via phosphatidylinositol-3 kinase/Akt regulation of Glut1 activity and trafficking. Mol Biol Cell 18, 1437–1446 (2007). Porstmann, T. et al. PKB|[sol]|Akt induces transcription of enzymes involved in cholesterol and fatty acid biosynthesis via activation of SREBP. Oncogene 24, 6465– 6481 (2005). Berg, J. M. Biochemistry. (W H Freeman & Company, 2012). Krebs, H. A. The Pasteur effect and the relations between respiration and fermentation. Essays Biochem. 8, 1–34 (1972). Berwick, D. C., Hers, I., Heesom, K. J., Moule, S. K. & Tavaré, J. M. The identification of ATP-citrate lyase as a protein kinase B (Akt) substrate in primary adipocytes. J Biol Chem 277, 33895–33900 (2002). Bauer, D. E., Hatzivassiliou, G., Zhao, F., Andreadis, C. & Thompson, C. B. ATP citrate lyase is an important component of cell growth and transformation. Oncogene 24, 6314–6322 (2005). Csibi, A. et al. The mTORC1 pathway stimulates glutamine metabolism and cell proliferation by repressing SIRT4. Cell 153, 840–854 (2013). Nicklin, P. et al. Bidirectional Transport of Amino Acids Regulates mTOR and Autophagy. Cell 136, 521–534 (2009). Durán, R. V. et al. Glutaminolysis activates Rag-mTORC1 signaling. Mol Cell 47, 349–358 (2012). Kim, D.-H. et al. mTOR Interacts with Raptor to Form a Nutrient-Sensitive Complex that Signals to the Cell Growth Machinery. Cell 110, 163–175 (2002). Ortega-Molina, A. & Serrano, M. PTEN in cancer, metabolism, and aging. Trends in Endocrinology & Metabolism 24, 184–189 (2013). Ramanathan, A., Wang, C. & Schreiber, S. L. Perturbational profiling of a cell-line model of tumorigenesis by using metabolic measurements. Proc Natl Acad Sci USA 102, 5992–5997 (2005). Chiaradonna, F. et al. Ras-dependent carbon metabolism and transformation in mouse fibroblasts. Oncogene 25, 5391–5404 (2006).

50.

51.

52. 53.

54.

55.

56.

57.

58.

59.

60. 61.

Gaglio, D. et al. Oncogenic K-Ras decouples glucose and glutamine metabolism to support cancer cell growth. Mol Syst Biol 7, 523–523 (2011). Chesney, J. & Telang, S. Regulation of glycolytic and mitochondrial metabolism by ras. Curr Pharm Biotechnol 14, 251–260 (2013). Kole, H. K., Resnick, R. J., van Doren, M. & Racker, E. Regulation of 6-phosphofructo1-kinase activity in ras-transformed rat-1 fibroblasts. Arch Biochem Biophys 286, 586–590 (1991). Telang, S. et al. Ras transformation requires metabolic control by 6-phosphofructo-2kinase. Oncogene 25, 7225–7234 (2006). Kaplon, J. et al. A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112 (2013). Accioly, M. T. et al. Lipid bodies are reservoirs of cyclooxygenase-2 and sites of prostaglandin-E2 synthesis in colon cancer cells. Cancer Res 68, 1732–1740 (2008). Fritz, V. et al. Abrogation of de novo lipogenesis by stearoyl-CoA desaturase 1 inhibition interferes with oncogenic signaling and blocks prostate cancer progression in mice. Molecular Cancer Therapeutics 9, 1740–1754 (2010). Osthus, R. C. et al. Deregulation of glucose transporter 1 and glycolytic gene expression by c-Myc. J Biol Chem 275, 21797–21800 (2000). O’Connell, B. C. et al. A large scale genetic analysis of c-Myc-regulated gene expression patterns. J Biol Chem 278, 12563–12573 (2003). Menssen, A. & Hermeking, H. Characterization of the c-MYC-regulated transcriptome by SAGE: identification and analysis of c-MYC target genes. Proc Natl Acad Sci USA 99, 6274–6279 (2002). Kim, J.-W. et al. Evaluation of myc E-box phylogenetic footprints in glycolytic genes by chromatin immunoprecipitation assays. Mol Cell Biol 24, 5923–5936 (2004). Dang, C. V. MYC on the path to cancer. Cell 149, 22–35 (2012). Shim, H. et al. c-Myc transactivation of LDH-A: implications for tumor metabolism and growth. Proc Natl Acad Sci USA 94, 6658–6663 (1997).

62.

63. 64.

65.

66.

67.

68. 69.

70.

71.

72. 73.

74. 75.

David, C. J., Chen, M., Assanah, M., Canoll, P. & Manley, J. L. HnRNP proteins controlled by c-Myc deregulate pyruvate kinase mRNA splicing in cancer. Nature 463, 364–368 (2010). Dang, C. V. et al. The c-Myc target gene network. Semin Cancer Biol 16, 253–264 (2006). Li, F. et al. Myc stimulates nuclearly encoded mitochondrial genes and mitochondrial biogenesis. Mol Cell Biol 25, 6225–6234 (2005). Kim, J., Lee, J.-H. & Iyer, V. R. Global identification of Myc target genes reveals its direct role in mitochondrial biogenesis and its E-box usage in vivo. PLoS ONE 3, e1798 (2008). Graves, J. A. et al. Mitochondrial structure, function and dynamics are temporally controlled by c-Myc. PLoS ONE 7, e37699 (2012). Zirath, H. et al. MYC inhibition induces metabolic changes leading to accumulation of lipid droplets in tumor cells. PNAS 110, 10258–10263 (2013). Guo, Q. M. et al. Identification of c-myc responsive genes using rat cDNA microarray. Cancer Res 60, 5922–5928 (2000). Nikiforov, M. A. et al. A functional screen for Myc-responsive genes reveals serine hydroxymethyltransferase, a major source of the one-carbon unit for cell metabolism. Mol Cell Biol 22, 5793–5800 (2002). Wise, D. R. et al. Myc regulates a transcriptional program that stimulates mitochondrial glutaminolysis and leads to glutamine addiction. PNAS 105, 18782– 18787 (2008). Gao, P. et al. c-Myc suppression of miR23a/b enhances mitochondrial glutaminase expression and glutamine metabolism. Nature 458, 762–765 (2009). Manning, B. D. & Cantley, L. C. AKT/PKB signaling: navigating downstream. Cell 129, 1261–1274 (2007). Eijkelenboom, A. & Burgering, B. M. T. FOXOs: signalling integrators for homeostasis maintenance. Nat Rev Mol Cell Biol 14, 83–97 (2013). Tu, B. P. et al. Cyclic changes in metabolic state during the life of a yeast cell. Proc Natl Acad Sci USA 104, 16886–16891 (2007). Benanti, J. A., Cheung, S. K., Brady, M. C. & Toczyski, D. P. A proteomic screen reveals

Chapter 2

49.

76.

77. 78. 79. 80. 81.

82.

83. 84.

85. 86.

87.

88.

89.

SCFGrr1 targets that regulate the glycolyticgluconeogenic switch. Nat Cell Biol 9, 1184– 1191 (2007). Herrero-Mendez, A. et al. The bioenergetic and antioxidant status of neurons is controlled by continuous degradation of a key glycolytic enzyme by APC/C-Cdh1. Nat Cell Biol 11, 747–752 (2009). Hartwell, L. H. & Weinert, T. A. Checkpoints: controls that ensure the order of cell cycle events. Science 246, 629–634 (1989). Malumbres, M. & Barbacid, M. Cell cycle, CDKs and cancer: a changing paradigm. Nat Rev Cancer 9, 153–166 (2009). Sherr, C. J. G1 phase progression: cycling on cue. Cell 79, 551–555 (1994). Lim, S. & Kaldis, P. Cdks, cyclins and CKIs: roles beyond cell cycle regulation. Development 140, 3079–3093 (2013). Sherr, C. J. & Roberts, J. M. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev 13, 1501–1512 (1999). Georgia, S. & Bhushan, A. β cell replication is the primary mechanism for maintaining postnatal β cell mass. J. Clin. Invest. 114, 963–968 (2004). Kushner, J. A. et al. Cyclins D2 and D1 are essential for postnatal pancreatic beta-cell growth. Mol Cell Biol 25, 3752–3762 (2005). Sarruf, D. A. et al. Cyclin D3 promotes adipogenesis through activation of peroxisome proliferator-activated receptor gamma. Mol Cell Biol 25, 9985–9995 (2005). Sakamaki, T. et al. Cyclin D1 determines mitochondrial function in vivo. Mol Cell Biol 26, 5449–5469 (2006). Hanse, E. A. et al. Cyclin D1 inhibits hepatic lipogenesis via repression of carbohydrate response element binding protein and hepatocyte nuclear factor 4α. cc 11, 2681– 2690 (2012). Wang, C. et al. Cyclin D1 repression of nuclear respiratory factor 1 integrates nuclear DNA synthesis and mitochondrial function. Proc Natl Acad Sci USA 103, 11567–11572 (2006). Tchakarska, G., Roussel, M., Troussard, X. & Sola, B. Cyclin D1 inhibits mitochondrial activity in B cells. Cancer Res 71, 1690–1699 (2011). Bienvenu, F. et al. Transcriptional role of cyclin D1 in development revealed by a genetic-proteomic screen. Nature 463, 374– 378 (2010).

90. 91.

92. 93.

94. 95. 96. 97. 98.

99.

100.

101. 102.

103.

104.

Buchakjian, M. R. & Kornbluth, S. The engine driving the ship: metabolic steering of cell proliferation and death. 11, 715–727 (2010). Yang, K., Hitomi, M. & Stacey, D. W. Variations in cyclin D1 levels through the cell cycle determine the proliferative fate of a cell. Cell Division 1, 32 (2006). Trimarchi, J. M. & Lees, J. A. Sibling rivalry in the E2F family. Nat Rev Mol Cell Biol 3, 11–20 (2002). Dimova, D. K. & Dyson, N. J. The E2F transcriptional network: old acquaintances with new faces. Oncogene 24, 2810–2826 (2005). Cam, H. & Dynlacht, B. D. Emerging roles for E2F: beyond the G1/S transition and DNA replication. Cancer Cell 3, 311–316 (2003). Chen, H.-Z., Tsai, S.-Y. & Leone, G. Emerging roles of E2Fs in cancer: an exit from cell cycle control. Nat Rev Cancer 9, 785–797 (2009). Fajas, L. et al. E2Fs Regulate Adipocyte Differentiation. Dev Cell 3, 39–49 (2002). Abella, A. et al. Cdk4 promotes adipogenesis through PPARgamma activation. Cell Metab 2, 239–249 (2005). Fajas, L. et al. The Retinoblastoma-Histone Deacetylase 3 Complex Inhibits PPARγ and Adipocyte Differentiation. Dev Cell 3, 903– 910 (2002). Chen, P. L., Riley, D. J., Chen, Y. & Lee, W. H. Retinoblastoma protein positively regulates terminal adipocyte differentiation through direct interaction with C/EBPs. Genes Dev 10, 2794–2804 (1996). Fajas, L. et al. Impaired pancreatic growth, beta cell mass, and beta cell function in E2F1 (-/- )mice. J. Clin. Invest. 113, 1288–1295 (2004). Annicotte, J.-S. et al. The CDK4-pRB-E2F1 pathway controls insulin secretion. Nat Cell Biol 11, 1017–1023 (2009). Roche, T. E. & Hiromasa, Y. Pyruvate dehydrogenase kinase regulatory mechanisms and inhibition in treating diabetes, heart ischemia, and cancer. Cell. Mol. Life Sci. 64, 830–849 (2007). Hsieh, M. C. F., Das, D., Sambandam, N., Zhang, M. Q. & Nahlé, Z. Regulation of the PDK4 isozyme by the Rb-E2F1 complex. J Biol Chem 283, 27410–27417 (2008). Darville, M. I., Antoine, I. V., MertensStrijthagen, J. R., Dupriez, V. J. & Rousseau, G. G. An E2F-dependent late-serumresponse promoter in a gene that controls glycolysis. Oncogene 11, 1509–1517 (1995).

119. Carlson, M. Glucose repression in yeast. Curr. Opin. Microbiol. 2, 202–207 (1999). 120. Ozcan, S. & Johnston, M. Three different regulatory mechanisms enable yeast hexose transporter (HXT) genes to be induced by different levels of glucose. Mol Cell Biol 15, 1564–1572 (1995). 121. Spielewoy, N., Flick, K., Kalashnikova, T. I., Walker, J. R. & Wittenberg, C. Regulation and recognition of SCFGrr1 targets in the glucose and amino acid signaling pathways. Mol Cell Biol 24, 8994–9005 (2004). 122. Kim, J.-H., Brachet, V., Moriya, H. & Johnston, M. Integration of transcriptional and posttranslational regulation in a glucose signal transduction pathway in Saccharomyces cerevisiae. Eukaryotic Cell 5, 167–173 (2006). 123. Vizán, P. & Cascante, M. Modulation of pentose phosphate pathway during cell cycle progression in human colon adenocarcinoma cell line HT29. 124, 2789–2796 (2009). 124. Tudzarova, S. et al. Two ubiquitin ligases, APC/C-Cdh1 and SKP1-CUL1-F (SCF)-betaTrCP, sequentially regulate glycolysis during the cell cycle. PNAS 108, 5278–5283 (2011). 125. Almeida, A., Bolaños, J. P. & Moncada, S. E3 ubiquitin ligase APC/C-Cdh1 accounts for the Warburg effect by linking glycolysis to cell proliferation. PNAS 107, 738–741 (2010). 126. Colombo, S. L. et al. Anaphase-promoting complex/cyclosome-Cdh1 coordinates glycolysis and glutaminolysis with transition to S phase in human T lymphocytes. PNAS 107, 18868–18873 (2010). 127. Colombo, S. L. et al. Molecular basis for the differential use of glucose and glutamine in cell proliferation as revealed by synchronized HeLa cells. PNAS 108, 21069–21074 (2011). 128. Sun, J. et al. The phosphatase-transcription activator EYA1 is targeted by anaphasepromoting complex/Cdh1 for degradation at M-to-G1 transition. Mol Cell Biol 33, 927–936 (2013). 129. Xu, P. X. et al. Eya1-deficient mice lack ears and kidneys and show abnormal apoptosis of organ primordia. Nat Genet 23, 113–117 (1999). 130. Zou, D., Silvius, D., Fritzsch, B. & Xu, P.-X. Eya1 and Six1 are essential for early steps of sensory neurogenesis in mammalian cranial placodes. Development 131, 5561–5572 (2004). 131. Zou, D., Silvius, D., Rodrigo-Blomqvist, S., Enerbäck, S. & Xu, P.-X. Eya1 regulates the

Chapter 2

105. Goto, Y., Hayashi, R., Kang, D. & Yoshida, K. Acute loss of transcription factor E2F1 induces mitochondrial biogenesis in HeLa cells. J Cell Physiol 209, 923–934 (2006). 106. Blanchet, E. et al. E2F transcription factor-1 regulates oxidative metabolism. Nat Cell Biol 13, 1146–1152 (2011). 107. Naaz, A. et al. Loss of cyclin-dependent kinase inhibitors produces adipocyte hyperplasia and obesity. FASEB J 18, 1925– 1927 (2004). 108. Inoue, N. et al. Cyclin-dependent kinase inhibitor, p21WAF1/CIP1, is involved in adipocyte differentiation and hypertrophy, linking to obesity, and insulin resistance. J Biol Chem 283, 21220–21229 (2008). 109. Lin, J. et al. P27 knockout mice: reduced myostatin in muscle and altered adipogenesis. Biochemical and Biophysical Research Communications 300, 938–942 (2003). 110. Uchida, T. et al. Deletion of Cdkn1b ameliorates hyperglycemia by maintaining compensatory hyperinsulinemia in diabetic mice. Nat Med 11, 175–182 (2005). 111. Pagano, M. Cell cycle regulation by the ubiquitin pathway. FASEB J 11, 1067–1075 (1997). 112. Skaar, J. R. & Pagano, M. Cdh1: a master G0/G1 regulator. Nat Cell Biol 10, 755–757 (2008). 113. Pines, J. Cubism and the cell cycle: the many faces of the APC/C. Nat Rev Mol Cell Biol 12, 427–438 (2011). 114. Matyskiela, M. E., Rodrigo-Brenni, M. C. & Morgan, D. O. Mechanisms of ubiquitin transfer by the anaphase-promoting complex. J. Biol. 8, 92 (2009). 115. Skowyra, D., Craig, K. L., Tyers, M., Elledge, S. J. & Harper, J. W. F-box proteins are receptors that recruit phosphorylated substrates to the SCF ubiquitin-ligase complex. Cell 91, 209–219 (1997). 116. Skaar, J. R., Pagan, J. K. & Pagano, M. Mechanisms and function of substrate recruitment by F-box proteins. Nat Rev Mol Cell Biol 14, 369–381 (2013). 117. Vodermaier, H. C. APC/C and SCF: controlling each other and the cell cycle. Curr. Biol. 14, R787–96 (2004). 118. Purnapatre, K., Gray, M., Piccirillo, S. & Honigberg, S. M. Glucose inhibits meiotic DNA replication through SCFGrr1pdependent destruction of Ime2p kinase. Mol Cell Biol 25, 440–450 (2005).

132.

133.

134.

135.

136.

137.

138. 139.

140. 141. 142.

143.

growth of otic epithelium and interacts with Pax2 during the development of all sensory areas in the inner ear. Dev. Biol. 298, 430– 441 (2006). Zhang, L. et al. Transcriptional coactivator Drosophila eyes absent homologue 2 is up-regulated in epithelial ovarian cancer and promotes tumor growth. Cancer Res 65, 925–932 (2005). Pandey, R. N. et al. The Eyes Absent phosphatase-transactivator proteins promote proliferation, transformation, migration, and invasion of tumor cells. Oncogene 29, 3715–3722 (2010). Li, C.-M. et al. Gene expression in Wilms’ tumor mimics the earliest committed stage in the metanephric mesenchymal-epithelial transition. 160, 2181–2190 (2002). Miller, S. J. et al. Inhibition of Eyes Absent Homolog 4 expression induces malignant peripheral nerve sheath tumor necrosis. Oncogene 29, 368–379 (2010). Grifone, R. et al. Six1 and Eya1 expression can reprogram adult muscle from the slow-twitch phenotype into the fast-twitch phenotype. Mol Cell Biol 24, 6253–6267 (2004). Cassago, A. et al. Mitochondrial localization and structure-based phosphate activation mechanism of Glutaminase C with implications for cancer metabolism. PNAS 109, 1092–1097 (2012). Vousden, K. H. & Prives, C. Blinded by the Light: The Growing Complexity of p53. Cell 137, 413–431 (2009). Schwartzenberg-Bar-Yoseph, F., Armoni, M. & Karnieli, E. The tumor suppressor p53 down-regulates glucose transporters GLUT1 and GLUT4 gene expression. Cancer Res 64, 2627–2633 (2004). Kondoh, H. et al. Glycolytic enzymes can modulate cellular life span. Cancer Res 65, 177–185 (2005). Bensaad, K. et al. TIGAR, a p53-inducible regulator of glycolysis and apoptosis. Cell 126, 107–120 (2006). Boidot, R. et al. Regulation of monocarboxylate transporter MCT1 expression by p53 mediates inward and outward lactate fluxes in tumors. Cancer Res 72, 939–948 (2012). Jiang, P. et al. p53 regulates biosynthesis through direct inactivation of glucose-6phosphate dehydrogenase. Nat Cell Biol 13, 310–316 (2011).

144. Aird, K. M. et al. Suppression of nucleotide metabolism underlies the establishment and maintenance of oncogene-induced senescence. CellReports 3, 1252–1265 (2013). 145. Kulawiec, M., Ayyasamy, V. & Singh, K. K. p53 regulates mtDNA copy number and mitocheckpoint pathway. J Carcinog 8, 8 (2009). 146. Lebedeva, M. A., Eaton, J. S. & Shadel, G. S. Loss of p53 causes mitochondrial DNA depletion and altered mitochondrial reactive oxygen species homeostasis. Biochim Biophys Acta 1787, 328–334 (2009). 147. Matoba, S. et al. p53 regulates mitochondrial respiration. Science 312, 1650–1653 (2006). 148. Okamura, S. et al. Identification of seven genes regulated by wild-type p53 in a colon cancer cell line carrying a well-controlled wild-type p53 expression system. Oncol. Res. 11, 281–285 (1999). 149. Contractor, T. & Harris, C. R. p53 negatively regulates transcription of the pyruvate dehydrogenase kinase Pdk2. Cancer Res 72, 560–567 (2012). 150. Jiang, P., Du, W., Mancuso, A., Wellen, K. E. & Yang, X. Reciprocal regulation of p53 and malic enzymes modulates metabolism and senescence. Nature 493, 689–693 (2013). 151. Hu, W. et al. Glutaminase 2, a novel p53 target gene regulating energy metabolism and antioxidant function. PNAS 107, 7455– 7460 (2010). 152. Maddocks, O. D. K. et al. Serine starvation induces stress and p53-dependent metabolic remodelling in cancer cells. Nature 493, 542–546 (2013). 153. Yahagi, N. et al. p53 Activation in adipocytes of obese mice. J Biol Chem 278, 25395– 25400 (2003). 154. Faubert, B., Berger, S. L., Jones, R. G., Thompson, C. B. & Mak, T. W. Carnitine palmitoyltransferase 1C promotes cell survival and tumor growth under conditions of metabolic stress. 25, 1041–1051 (2011). 155. Ide, T. et al. GAMT, a p53-inducible modulator of apoptosis, is critical for the adaptive response to nutrient stress. Mol Cell 36, 379–392 (2009). 156. Velletri, T. et al. GLS2 is transcriptionally regulated by p73 and contributes to neuronal differentiation. cc 12, 3564–3573 (2013). 157. Giacobbe, A. et al. p63 regulates glutaminase 2 expression. cc 12, 1395–1405 (2013).

174. Berger, S. L. The complex language of chromatin regulation during transcription. Nature 447, 407–412 (2007). 175. Love, D. C. & Hanover, J. A. The hexosamine signaling pathway: deciphering the “O-GlcNAc code”. Sci STKE 2005, re13 (2005). 176. Drougat, L. et al. Characterization of O-GlcNAc cycling and proteomic identification of differentially O-GlcNAcylated proteins during G1/S transition. Biochim Biophys Acta 1820, 1839–1848 (2012). 177. Olivier-Van Stichelen, S. et al. Serumstimulated cell cycle entry promotes ncOGT synthesis required for cyclin D expression. Oncogenesis 1, e36 (2012). 178. Zhang, S., Roche, K., Nasheuer, H.-P. & Lowndes, N. F. Modification of histones by sugar β-N-acetylglucosamine (GlcNAc) occurs on multiple residues, including histone H3 serine 10, and is cell cycle-regulated. Journal of Biological Chemistry 286, 37483–37495 (2011). 179. Hanover, J. A., Krause, M. W. & Love, D. C. Bittersweet memories: linking metabolism to epigenetics through O-GlcNAcylation. Nat Rev Mol Cell Biol 13, 312–321 (2012). 180. Yang, Y. R. et al. O-GlcNAcase is essential for embryonic development and maintenance of genomic stability. Aging Cell 11, 439–448 (2012). 181. Dehennaut, V. et al. Identification of structural and functional O-linked N-acetylglucosamine-bearing proteins in Xenopus laevis oocyte. Mol. Cell Proteomics 7, 2229–2245 (2008). 182. Dehennaut, V. et al. Microinjection of recombinant O-GlcNAc transferase potentiates Xenopus oocytes M-phase entry. Biochemical and Biophysical Research Communications 369, 539–546 (2008). 183. Dehennaut, V. et al. Survey of O-GlcNAc level variations in Xenopus laevis from oogenesis to early development. Glycoconj. J. 26, 301– 311 (2009). 184. Hardie, D. G. AMP-activated/SNF1 protein kinases: conserved guardians of cellular energy. Nat Rev Mol Cell Biol 8, 774–785 (2007). 185. Hardie, D. G., Ross, F. A. & Hawley, S. A. AMPK: a nutrient and energy sensor that maintains energy homeostasis. Nat Rev Mol Cell Biol 13, 251–262 (2012). 186. Imamura, K., Ogura, T., Kishimoto, A., Kaminishi, M. & Esumi, H. Cell cycle

Chapter 2

158. Rufini, A. et al. TAp73 depletion accelerates aging through metabolic dysregulation. Genes Dev 26, 2009–2014 (2012). 159. Du, W. et al. TAp73 enhances the pentose phosphate pathway and supports cell proliferation. Nat Cell Biol 15, 991–1000 (2013). 160. Amelio, I. et al. p73 regulates serine biosynthesis in cancer. Oncogene (2013). doi:10.1038/onc.2013.456 161. Su, X. et al. TAp63 is a master transcriptional regulator of lipid and glucose metabolism. Cell Metab 16, 511–525 (2012). 162. Sabbisetti, V. et al. p63 promotes cell survival through fatty acid synthase. PLoS ONE 4, e5877 (2009). 163. Pardee, A. B. A restriction point for control of normal animal cell proliferation. Proc Natl Acad Sci USA 71, 1286–1290 (1974). 164. Yalcin, A. et al. Nuclear targeting of 6-phosphofructo-2-kinase (PFKFB3) increases proliferation via cyclin-dependent kinases. J Biol Chem 284, 24223–24232 (2009). 165. Mazurek, S., Boschek, C. B., Hugo, F. & Eigenbrodt, E. Pyruvate kinase type M2 and its role in tumor growth and spreading. Semin Cancer Biol 15, 300–308 (2005). 166. Christofk, H. R. et al. The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 452, 230–233 (2008). 167. Yang, W. et al. Nuclear PKM2 regulates β-catenin transactivation upon EGFR activation. Nature 480, 118–122 (2011). 168. Yang, W. et al. ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect. Nat Cell Biol 14, 1295–1304 (2012). 169. Yang, W. et al. PKM2 phosphorylates histone H3 and promotes gene transcription and tumorigenesis. Cell 150, 685–696 (2012). 170. Jiang, Y. et al. PKM2 regulates chromosome segregation and mitosis progression of tumor cells. Mol Cell 53, 75–87 (2014). 171. Wellen, K. E. et al. ATP-citrate lyase links cellular metabolism to histone acetylation. Science 324, 1076–1080 (2009). 172. Galdieri, L. & Vancura, A. Acetyl-CoA carboxylase regulates global histone acetylation. Journal of Biological Chemistry 287, 23865–23876 (2012). 173. Li, B., Carey, M. & Workman, J. L. The role of chromatin during transcription. Cell 128, 707–719 (2007).

187. 188.

189. 190.

191.

192.

193.

194.

195. 196. 197. 198.

regulation via p53 phosphorylation by a 5’-AMP activated protein kinase activator, 5-aminoimidazole- 4-carboxamide-1-betaD-ribofuranoside, in a human hepatocellular carcinoma cell line. Biochemical and Biophysical Research Communications 287, 562–567 (2001). Jones, R. G. et al. AMP-activated protein kinase induces a p53-dependent metabolic checkpoint. Mol Cell 18, 283–293 (2005). Liang, J. et al. The energy sensing LKB1AMPK pathway regulates p27(kip1) phosphorylation mediating the decision to enter autophagy or apoptosis. Nat Cell Biol 9, 218–224 (2007). Poteet, E. et al. Reversing the Warburg effect as a treatment for glioblastoma. Journal of Biological Chemistry 288, 9153–9164 (2013). Mandal, S., Freije, W. A., Guptan, P. & Banerjee, U. Metabolic control of G1-S transition: cyclin E degradation by p53induced activation of the ubiquitinproteasome system. J Cell Biol 188, 473–479 (2010). Owusu-Ansah, E., Yavari, A., Mandal, S. & Banerjee, U. Distinct mitochondrial retrograde signals control the G1-S cell cycle checkpoint. Nat Genet 40, 356–361 (2008). Imai, S., Armstrong, C. M., Kaeberlein, M. & Guarente, L. Transcriptional silencing and longevity protein Sir2 is an NAD-dependent histone deacetylase. Nature 403, 795–800 (2000). Dryden, S. C., Nahhas, F. A., Nowak, J. E., Goustin, A.-S. & Tainsky, M. A. Role for human SIRT2 NAD-dependent deacetylase activity in control of mitotic exit in the cell cycle. Mol Cell Biol 23, 3173–3185 (2003). Inoue, T. et al. SIRT2, a tubulin deacetylase, acts to block the entry to chromosome condensation in response to mitotic stress. Oncogene 26, 945–957 (2007). Vaziri, H. et al. hSIR2(SIRT1) functions as an NAD-dependent p53 deacetylase. Cell 107, 149–159 (2001). Luo, J. et al. Negative control of p53 by Sir2alpha promotes cell survival under stress. Cell 107, 137–148 (2001). Brunet, A. et al. Stress-dependent regulation of FOXO transcription factors by the SIRT1 deacetylase. Science 303, 2011–2015 (2004). Nakahata, Y. et al. The NAD+-dependent deacetylase SIRT1 modulates CLOCK-

mediated chromatin remodeling and circadian control. Cell 134, 329–340 (2008). 199. Asher, G. et al. SIRT1 regulates circadian clock gene expression through PER2 deacetylation. Cell 134, 317–328 (2008). 200. Matsuo, T. et al. Control mechanism of the circadian clock for timing of cell division in vivo. Science 302, 255–259 (2003). 201. Fu, L., Pelicano, H., Liu, J., Huang, P. & Lee, C. The circadian gene Period2 plays an important role in tumor suppression and DNA damage response in vivo. Cell 111, 41–50 (2002).

足足足

CHAPTER 3 METABOLIC ALTERATIONS IN ONCOGENE-INDUCED SENESCENCE

Manuscript in preparation

METABOLIC ALTERATIONS IN ONCOGENE-INDUCED SENESCENCE Joanna Kaplon1, Liang Zheng2, Vitaly A. Selivanov3,4, Marta Cascante3,4, Tomer Shlomi5, Eyal Gottlieb2 and Daniel S. Peeper1 Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam,

The Netherlands. 2Cancer Research UK, Beatson Institute for Cancer Research, Switchback Road, Glasgow de Barcelona, Av Diagonal 643, 08028 Barcelona, Spain. 4Institute of Biomedicine of Universitat de Barcelona (IBUB) and CSIC-Associated Unit, Spain. 5Computer Science Department, Technion, Israel Institute of Technology, Haifa, 32000, Israel.

SUMMARY Oncogene-induced senescence (OIS) represents a powerful pathophysiological mechanism suppressing cancer. While deregulation of metabolism is an important feature of cancer, little is known about the role of cellular metabolism in OIS. Here, we performed an unbiased analysis of the central carbon metabolism in cells undergoing OIS and found that senescence is associated with a distinct metabolic profile. OIS cells show an increased rate of pyruvate oxidation in mitochondria, less utilization of glutamine and an increased rate of fatty acid secretion. These metabolic changes directly oppose those observed in cancer, indicating that the antitumor function of OIS is manifested also on the level of metabolic regulation. INTRODUCTION Multicellular organisms have the capacity to renew, repair and regenerate tissues, allowing for the maintenance of proper tissue and organ functions. A downside of the ability to self-renew is the risk that this process goes wrong and culminates in uncontrolled cell proliferation, with cancer as a prime example. Different mechanisms have developed to prevent oncogenic transformation. Some cell types undergo various cell death programs such as apoptosis1 or autophagy2. Others follow an alternative tactic: they stop proliferating and enter a phase called senescence. Indeed, antitumor function of oncogene-induced senescence (OIS) has been clearly demonstrated in vivo both in mouse models and human samples3. Because of the substantial role for OIS in the prevention of cancer, great effort is being made to disclose that process. However, despite enormous progress, we are only beginning to reveal the molecular mechanism of OIS. Numerous studies have shown that deregulation of metabolism is closely linked to the proliferative potential of cells4-8. To support a high rate of proliferation, cancer cells commonly shift their metabolism towards biosynthesis, thereby providing the building

Chapter 3

G61 1BD, Scotland, UK. 3Department of Biochemistry and Molecular Biology, Faculty of Biology, Universitat

blocks necessary for tumor expansion. Metabolic pathways are therefore rewired such that anabolic processes are executed in parallel to programs generating energy production sufficient to sustain cell growth and survival9. In spite of the wide belief that senescent cells remain metabolically active10, few investigators have studied metabolic regulation in senescence in detail. Interestingly, several of the pathways regulating metabolism in tumor cells are linked to senescence. For example, p53, a tumor suppressor contributing to some types of OIS, modulates glucose utilization and mitochondrial respiration. p53 reduces glycolysis by induction of the expression of TIGAR, an enzyme downregulating levels of the glycolytic activator fructose-2,6-bisphosphate (F-2,6BP)11 and a simultaneous inhibition of the expression of glucose transporters GLUT 1 and GLUT412 and glycolytic enzyme phosphoglycerate mutase (PGM)13. Consequently, activation of glycolysis by overexpression of PGM abrogates p53-dependent senescence13. At the same time, p53 promotes mitochondrial respiration by transcriptional activation of subunit I of cytochrome c oxidase14 and activation of expression of synthesis of cytochrome c oxidase 2 (SCO2)15. Along these lines, recently a connection has been made between tricarboxylic acid cycle (TCA) associated malic enzymes (MEs) and p53-dependent senescence16. While p53 accumulation repressed the expression of ME1 and ME2, and therefore cell metabolism and proliferation, depletion of ME1 and ME2 reciprocally activated p53 in a forward feeding manner, leading to a strong induction of senescence. Also, regulation of nucleotide synthesis has been linked to OIS17: ribonucleotide reductase subunit M2 (RRM2), a rate-limiting protein in dNTP synthesis, is downregulated in senescence, resulting in suppression of nucleotide metabolism and DNA synthesis. A decrease in nucleotide pools is crucial for OIS as either addition of exogenous nucleosides or restoration of RRM2 abrogates senescence response. Moreover, we have demonstrated that OIS is associated with increased TCA cycle activity. This is mediated by the activation of the mitochondrial gatekeeper pyruvate dehydrogenase (PDH) as an enzyme linking glycolysis and oxidative phosphorylation18. While simultaneous deregulation of PDH-inhibitory kinase PDK1 and its activating phosphatase PDP2 switches on PDH and the OIS program, restoration of the levels of these enzymes inactivates PDH and leads to OIS escape. Interestingly, therapy-induced senescence (TIS) is also associated with enhanced glucose utilization in the TCA cycle19. Similarly, analysis of changes in protein expression in OIS revealed an upregulation of proteins involved in oxidative phosphorylation and a downregulation of proteins involved in glycolysis, further supporting an important role for oxidative metabolism in the senescence program20. This together demonstrates that senescence and metabolism are tightly intertwined. However, in spite of these observations, a comprehensive study of cellular metabolism in OIS has been lacking. Here, we used mass balance analysis and metabolic flux profiling to screen in an unbiased fashion for metabolic changes accompanying OIS.

Chapter 3

RESULTS We compared the metabolism of human diploid fibroblasts (HDF) undÂergoing OIS to that of cycling cells. To evoke OIS, we used oncogenic BRAFV600E, a common and strong inducer of OIS in vitro and in vivo18,21-24. With the use of liquid-chromatography mass spectrometry (LCMS), we have quantified the metabolic exchange rate of key metabolites (glucose, lactate, glutamine, glutamate, pyruvate and alanine) as well as their intracellular concentrations. We utilized two independent models, namely mass-balance and isotopomer dynamics, to calculate and predict intracellular metabolic fluxes. Mass balance analysis predicts differential glutamine, pyruvate and fatty acid metabolism in OIS and cycling cells In order to understand the potential metabolic alterations taking place during OIS, we delineated a mass-balance model utilizing the measured exchange rates of glucose, lactate, glutamine and glutamate, as well as that of pyruvate itself and alanine, a direct transamination product of pyruvate. Furthermore, the biosynthetic constraint of proliferating cells was translated to the required rate of production and utilization of amino acids, fatty acids and nucleotides. The doubling time of the cycling cells studied is 24 hours and the measured protein concentration is 165 Âľg/106 cells. Based on these observations, and (for simplicityâ&#x20AC;&#x2122;s sake) assuming an equal distribution of amino acids in proteins and an average molecular weight of an amino acid of 146 g/mol, we calculated the required rate for protein synthesis of key amino acids (which are either absent from the medium or directly derived from reactions in the model (alanine, aspartate, asparagine, glutamine and glutamate)), to be 2.3 nmol/(106 cells/hour). Assuming that the dry mass of cells consists of approximately 60% proteins, 20% lipids and 15% nucleotides25 and that the molecular weight of palmitate (as a readout for fatty acids) is 256 g/mol and the average molecular weight of nucleotide (monophosphate) is 340 g/mol, the rates of lipid (as palmitate equivalents) and nucleotide accumulation as biomass in proliferating cells were calculated to be 8.8 nmol/(106 cells/ hour) and 5.0 nmol/(106 cells/hour), respectively. Finally, oxygen consumption rate (OCR) was measured and utilized in this model to account for the oxidation rate of NADH (2 mol NADH per mol O2) produced in all the studied reactions. We fitted these measured and estimated fluxes into a central carbon metabolism model depicted in Figure 1, and calculated the best possible way to balance those metabolic rates in one consistent model. Due to multiple ways to transfer electrons between mitochondrial and cytosolic NAD(P)H/FADH2, a single pool representing all of the latter metabolites was assumed. The production (reduction) rate of these reducing equivalents was fitted into all the known reactions in the model (glycolysis, pentose phosphate pathway, malic enzyme and the TCA cycle).

The mass balance analysis made several predictions: (1) Glutamine utilization in cycling cells is sufficient to account for the required amino acids, which are either derived directly from glutamine (glutamine and glutamate) or from the TCA cycle (aspartate and asparagine), as well as for pyrimidine biosynthesis. On the other hand, OIS cells produce and secrete more glutamate from glucose (no net glutaminolysis) and hence require an active pyruvate carboxylase (PC) to support anaplerosis. Indeed, when cells were incubated for 24 hours in uniformly labeled [U-13C6]-glucose, and heavy [13C] isotopes were traced in different metabolites, it was evident that more glutamate was derived from glucose in OIS cells as compared to cycling cells and that citrate was labeled with 3 and 5 [13C] carbons, indicative of more pyruvate carboxylase activity in OIS cells (Figure 1). (2) The mass balance analysis computed a large increase in the rate of pyruvate oxidation by PDH in the mitochondria (Figure 1). This calculation was supported by the fact that OIS cells consumed a similar amount of glucose as cycling cells, but showed less alanine or pyruvate secretion (Figure 1), which was not balanced by an increase in biomass production (as OIS cells do not proliferate). It was also consistent with an increase in OCR in OIS cells, suggesting an increased rate of oxidative phosphorylation. In addition to that, measurements of decreased glutamine uptake and increased glutamate secretion in OIS cells supported the prediction of increased oxidation of pyruvate in these cells (Figure 1). (3) Although an increase in OCR during OIS was measured, this was insufficient to account for a complete oxidation of the excess pyruvate in the mitochondria. Therefore, it was predicted that the citrate/malate shuttle would remove excess acetyl-CoA (AcCoA) to the cytosol to be used for fatty acids biosynthesis (calculated in palmitate equivalents). Interestingly, the mass balance predicted that the rate of de novo palmitate synthesis in cycling cells is insufficient to support the minimum lipid biosynthesis required to sustain the measured proliferation rate and hence cycling cells will require a net uptake of exogenous fatty acids. On the other hand, the higher rate of palmitate production in OIS would require fatty acid secretion. These predictions were experimentally confirmed by LC-MS analyses of extracellular palmitate, oleate and stearate (Figure 1). To compute the difference in PDH flux between the OIS and cycling cells, while accounting for experimental error in the measurement of metabolite uptake and secretion rates, we assumed a Gaussian noise model per each uptake and secretion measurement (considering the experimental mean and standard deviation in uptake and secretion flux measurements). The standard deviation and confidence interval of PDH flux were calculated based on a linear combination of normal distributions. The expected difference in PDH fluxes (presented as nmol/(106 cells x hour) between OIS and cycling cells are: OIS: PDH = 307.570 std = 79.937 95% CI = [176.085, 439.055] Cycling: PDH = 63.050 std = 43.106 95% CI = [-7.853, 133.953] PDH difference = 244.520 std = 90.819 95% CI = [95.136, 393.904]

Cyc

fatty acids (peak area)

1.5x106

Cyc: -425.9± 40.9 OIS: -489.9± 35.3

glucose

MEDIUM

CYTOSOL

OIS

oleate

1x106 5x106

5x106

palmitate

stearate

palmitate

oleate

stearate

fatty acids

1x106

glucose R5P

Cyc

nucleotides

OIS

G3P pyruvate

100

lactate %

lactate Cyc:723.0±12.7 OIS:646.9±73.1

pyruvate

Cyc OIS

alanine

fatty acids

OAA+AcCoA

Chapter 3

pyruvate %

100

Cyc:51.7±1.0 OIS:16.8±0.5

palmitate ( equivalents )

AcCoA

Cyc OIS

100

lipids

oxaloacetate

Cyc OIS

TCA cycle

aspartate asparagine

MITOCHONDRION

Cyc OIS 100

aKG %

Cyc:13.0±0.6 OIS: 8.5± 0.5

citrate %

alanine %

100

Cyc

pyrimidines

glutamine

Cyc:-45.9± 5.1 OIS:-34.0± 7.6

NAD

protein

glutamine

glutamate

Cyc OIS

malate %

100

OIS

Cyc OIS

NAD

H 2O

nucleotides

Rate of metabolic flux [nmol/(106cell x hr)] Arrow scale

≥ 500 ≥ 250 ≥ 100 ≥ 50 < 50

Change in metabolite flux (OIS/Cyc) ≥ 1.5× 0.5≤x≤1.5 ≤ 0.5× non-determined flux

Counter-flow OIS Cyc measured metabolite exchange rate ± s.d. Cyc= cycling cells OIS= senescent cells negative value= uptake

Isotopologue pattern 24h U-12C 13 C1 13 C2 13 C3 13 C4 13 C5 13 C6

protein

glutamate Cyc:27.9± 0.3 OIS:38.8± 3.7

glutamate (peak area)

Cyc:-103.2±12.3 OIS:-177.3±21.0

6x107

3x107

Cyc OIS

Figure 1. OIS is accompanied by changes in metabolism of glutamine, pyruvate and fatty acids Metabolic alterations in OIS versus cycling cells were studied using the mass-balance model. Measured metabolic fluxes used for balancing the delineated metabolic network are indicated in solid boxes. The biosynthetic constraint of proliferating cells was translated to the required rate of production and utilization of amino acids, fatty acids and nucleotides. Based on these rates, the remainder of the reactions was fitted to balance the rates of metabolite production with their consumption and/or secretion. Also the rates of reactions that produce reducing equivalents (NAD(P)H and FADH2), (i.e., glycolysis, pentose phosphate pathway, malic enzyme and TCA cycle) were balanced with those that oxidize them (oxygen consumption, lactate dehydrogenase and fatty acid biosynthesis). The calculated reaction rates are presented as colored arrows where the thickness of the arrows represents the rate of reaction and the color represents the ratio between OIS to cycling (cyc) cells. All reactions are calculated as nmol/(106 cells x hour). Additional experimental data that provide important support for the predictions made by this model are presented as bar graphs. These represent the levels of metabolites (intracellular and extracellular) and isotopomers derived from [U-13C6]-glucose after 24 hours incubation. AcCoA - acetyl-CoA; G3P - glyceraldehyde 3-phosphate; OAA - oxaloacetate; aKG - alpha-ketoglutarate; R5P - ribose-5-phosphate; TCA cycle - tricarboxylic acid cycle.

Isotopomer flux analysis confirms an increase in the flux of pyruvate into the TCA cycle To validate the predicted increased rate of pyruvate oxidation in the mitochondria during OIS, we employed a 13C-tracer-based metabolic flux analysis to study the fate of glucose in these cells. For labeling experiments, cycling and OIS cells were incubated with medium supplemented with uniformly labeled [U-13C6]-glucose and at the time points between 15 minutes and 24 hours the distribution of glucose-derived isotopomers of key metabolites was quantified (Supplementary Tables 1 and 2). These metabolic fluxes, which were not measured directly, were determined based on fitting the measured distribution of [13C] isotopic isomers (isotopomers) of intracellular and secreted metabolites (Supplementary Tables 1 and 2) by the isotopomer dynamics method26. This method is based on the simulation of dynamics of isotopomer distributions and does not require reaching isotopic steady state. Application of this method was justified by the measured slow dynamics of metabolite labeling. The model consists of a system of ordinary differential equations describing the time course of concentrations of all the measured isotopomers. The algorithms for data fitting and determination of confidence intervals for metabolic fluxes are described in reference 27. Examples of fitting such data for cycling and OIS cells are shown in Figure 2. The fluxes are determined from fitting the entire set of the experimental data shown in Supplementary Table 1. Since the fraction of secreted lactate, pyruvate and alanine with respect to consumed glucose is less in OIS compared to cycling cells, the former directs much more pyruvate into citrate, a combined reaction of PDH and citrate synthase. This is clearly demonstrated by the fact that OIS cells have a higher concentration of citrate, which is labeled faster than in cycling cells (Supplementary Table 1). Table 1 shows the 95% confidence intervals (CI) for the metabolic fluxes. The intervals calculated for cycling and OIS cells are based on fitting the measured fluxes and distribution of isotopomers using the χ2 (sum of normalized squared deviation between measured and computed data) criterion as described in reference 27. Oxygen consumption (VO2) is calculated as the combined metabolic rates of PDH + (aKG → Mal) + (Cit → aKG)/2 + MDH/2. This calculation takes into account that conversion of 1 mol of pyruvate into AcCoA results in a reduction of 2 mol of NAD+ into NADH (one by GAPDH in the cytosol and another by PDH); further conversion in combined reactions (aKG → Mal) result in 1 mol NADH and 1 mol reduced FAD; the reactions (Cit → aKG) and MDH result in reduction of 1 mol NAD+ each. Oxidation of 2 mol NADH or FADH2 results in a reduction of one mol of oxygen. Thus, as predicted by mass balance analysis, the isotopomer flux analysis identified a sharp increase in the flux into the TCA cycle via PDH in OIS cells when compared to cycling cells. DISCUSSION In recent years, the mechanism of cellular metabolism, and its deregulation during oncogenic transformation, has been receiving more attention. The development of sensitive analytic

Table 1. cycling Gluc in Lac&Pyr out ALA out Pyr → AcCoA (PDH) Pyr → OAA (PC) aKG → Mal aKG → OOA (MDH) GLU out GLN in VO2

min 411.43 811.10 3.78 3.35 0 49.16 49.16 26.18 47.14 98.28

OIS max 411.43 812.01 4.29 4.15 0 54.19 54.19 27.49 47.14 108.06

min 484.69 782.33 1.19 140.34 0.07 14.59 14.59 21.62 24.28 169.64

max 484.69 812.95 1.44 159.41 8.25 18.97 18.97 27.28 30.55 189.98

95% confidence intervals of the indicated reaction rates [nmol/(106 cells x hour)] Gluc - glucose; Lac - lactate; Pyr - pyruvate; AcCoA- Acetyl-CoA; OAA - oxaloacetate; aKG - alpha-ketoglutarate; Mal - malate; GLU - glutamate; GLN - glutamine; PDH- pyruvate dehydrogenase; PC - pyruvate carboxylase; MDH - malate dehydrogenase.

tools to monitor metabolism in living cells has deepened our understanding of metabolic regulation. However, in spite of the widely recognized importance of OIS, and particularly abrogation thereof, for oncogenic transformation, little is known about the regulation and role of cellular metabolism in this context. For example, it is unclear how metabolic fluxes change when cells undergo OIS, and whether they are functionally connected to the execution of the senescence program.

Chapter 3

Figure 2. A comparison between the experimental measurements to computational fitting of the rate of glucose-derived 13C incorporation into each indicated metabolite Time course of unlabeled fractions of metabolites produced by cycling and OIS cells incubated with [U-13C6]-glucose is shown. Measured values with their standard deviations are indicated by symbols and the respective calculated values (best fit) are connected with lines. GLU - glutamate; mal - malate; cit - citrate; pyr - pyruvate.

Here, we report an unbiased and comprehensive analysis of the central carbon metabolism in cells undergoing OIS. Our study reveals that OIS cells display several metabolic alterations when compared to cycling cells. Both mass balance and isotopomer flux analyses show that entry into OIS is accompanied by an increase in pyruvate flux into the TCA cycle. Specifically, the rate of conversion of pyruvate to citrate, a combined reaction of PDH and citrate synthase, is highly increased during OIS. Notably, we have previously identified PDH to be an important modulator of OIS18. We have shown that PDH is activated in OIS due to the deregulation of expression levels of PDH regulatory enzymes: PDK1 and PDP2. Also, we have demonstrated that this process is not merely an epiphenomenon accompanying OIS as enforced normalization of the PDK1 or PDP2 levels resulted in an efficient abrogation of OIS. The fact that we found PDH to be activated in OIS not only confirms our previous observations, but also illustrates the versatility of metabolic modeling approaches to identify novel candidate pathways central for senescence program. Another prediction that arose from our metabolic profiling is that OIS cells have altered their glutamine metabolism when compared to cycling cells. Interestingly, pharmacological inhibition of glutaminase (GLS), the first enzyme involved in glutaminolysis, was shown to induce premature senescence in endothelial cells28. Notably, activity of GLS isozymes has been correlated to growth rates and malignancy in tumors. Along these lines, silencing of GLS expression or inhibition of its activity delayed tumor growth29-31. As several hallmarks of malignancy depend on the presence of glutamine32, a downregulation of glutamine utilization in OIS is likely to represent a tumor-suppressive response. Hence, from a therapeutic point of view, it would be interesting to study the regulation of glutamine in senescence in more detail. Next to the higher rate of pyruvate oxidation and alteration of glutamine metabolism in OIS, our study predicts that OIS is accompanied by an alteration in fatty acid metabolism. Notably, a few studies have previously linked senescence to changes in lipogenesis. For example, senescence was demonstrated to associate with reduced formation of phospholipids as well as biosynthesis and desaturation of fatty acid33. Along these lines, senescent cells showed an increased ratio of glycerophosphocholine (GPC) to phosphocholine (PC), both important in phospholipid metabolism34. Interestingly, the observed changes in the choline metabolism counteract the well-known changes in choline metabolism of tumor cells, which show a â&#x20AC;&#x153;GPC to PC switchâ&#x20AC;?35-37. Moreover, OIS cells were shown to have an increased steady-state level of certain free fatty acids38. Higher fatty acid levels do not reflect an increase in de novo synthesis of fatty acids, as the rates of lipid synthesis are reduced in OIS cells, but correlate with an increased rate of fatty acid oxidation. Nevertheless, inhibiting the fatty acid oxidation did not prevent a senescence-associated cell cycle arrest38, arguing against a causal role of that process in senescence. These observations, together with an increasing role of lipid metabolism in tumorigenesis39,40, suggest that modulating this metabolic

MATERIALS AND METHODS Cell culture, viral transduction, and senescence induction The human diploid fibroblast (HDF) cell line TIG3 expressing the ectopic receptor, hTERT and sh-p16INK4A was maintained in DMEM, supplemented with 9% fetal bovine serum (PAA), 2 mM glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin (GIBCO). Retroviral infections were performed using Phoenix cells as producers of viral supernatants. For senescence induction, HDFs were infected with BRAFV600E-encoding or control virus and pharmacologically selected with blasticidin. Cells were analyzed at the day 9 after introduction of BRAFV600E, the day that senescence is fully established. Plasmids pMSCV-blast-BRAFV600E and pMSCV-blast were previously described24. Measurement of metabolites by LC-MS 2 × 106 HDF were plated onto 10-cm dishes and cultured in standard medium for 24 hours. For 13C-flux analysis, the medium was replaced with 4.5 mM [U-13C]-glucose (Cambridge Isotope, UK). After incubation for the indicated time, cells and media were collected. For extracellular metabolite analysis, 200 µL of growth media from cell culture were added to 600 µL of acetonitrile for deproteinization. Samples were vortexed for 10 minutes and centrifuged for 10 minutes at 16000 g at 4 0C. The supernatant was stored for subsequent LC-MS analysis. For intracellular metabolite analysis, cells were lysed with a solution composed of 50% methanol and 30% acetonitrile in water in dry iced methanol (-80 0C) and quickly scraped from the plate. The insoluble material was immediately pelleted in a cooled centrifuge (4 0C) for 10 minutes at 16000 g and the supernatant was collected for subsequent LC-MS analysis. LC-MS analysis was carried out as described in reference 41. MS data was analyzed by LCquan™ (Thermo Scientific, UK) and quantifications of intracellular and extracellular metabolites were performed by the standard-dilution method as describe in reference 42. REFERENCES 1. 2. 3.

4. 5.

Cantor, J. R. & Sabatini, D. M. Cancer cell metabolism: one hallmark, many faces. Cancer Discov 2, 881–898 (2012). DeBerardinis, R. J., Sayed, N., Ditsworth, D. & Thompson, C. B. Brick by brick: metabolism and tumor cell growth. Curr Opin Genet Dev 18, 54–61 (2008). Tennant, D. A., Durán, R. V. & Gottlieb, E. Targeting metabolic transformation for

Chapter 3

pathway in senescence is a mechanism to prevent malignant transformation. Studying the nature of that regulation will be of great interest as it might reveal attractive therapeutic metabolic cancer targets.

10. 11. 12.

13. 14.

15. 16.

17.

18.

19. 20.

cancer therapy. Nat Rev Cancer 10, 267–277 (2010). Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033 (2009). Wellen, K. E. & Thompson, C. B. Cellular metabolic stress: considering how cells respond to nutrient excess. Mol Cell 40, 323–332 (2010). Lunt, S. Y. & Vander Heiden, M. G. Aerobic glycolysis: meeting the metabolic requirements of cell proliferation. Annu. Rev. Cell Dev. Biol. 27, 441–464 (2011). Campisi, J. Replicative senescence: an old lives’ tale? Cell 84, 497–500 (1996). Bensaad, K. et al. TIGAR, a p53-inducible regulator of glycolysis and apoptosis. Cell 126, 107–120 (2006). Schwartzenberg-Bar-Yoseph, F., Armoni, M. & Karnieli, E. The tumor suppressor p53 down-regulates glucose transporters GLUT1 and GLUT4 gene expression. Cancer Res 64, 2627–2633 (2004). Kondoh, H. et al. Glycolytic enzymes can modulate cellular life span. Cancer Res 65, 177–185 (2005). Okamura, S. et al. Identification of seven genes regulated by wild-type p53 in a colon cancer cell line carrying a well-controlled wild-type p53 expression system. Oncol. Res. 11, 281–285 (1999). Matoba, S. et al. p53 regulates mitochondrial respiration. Science 312, 1650–1653 (2006). Jiang, P., Du, W., Mancuso, A., Wellen, K. E. & Yang, X. Reciprocal regulation of p53 and malic enzymes modulates metabolism and senescence. Nature 493, 689–693 (2013). Aird, K. M. et al. Suppression of nucleotide metabolism underlies the establishment and maintenance of oncogene-induced senescence. CellReports 3, 1252–1265 (2013). Kaplon, J. et al. A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112 (2013). Dörr, J. R. et al. Synthetic lethal metabolic targeting of cellular senescence in cancer therapy. Nature 501, 421–425 (2013). Li, M. et al. Oncogene-induced cellular senescence elicits an anti-Warburg effect. Proteomics 13, 2585–2596 (2013).

21. 22. 23. 24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 41, 544–552 (2009). Dhomen, N. et al. Oncogenic Braf induces melanocyte senescence and melanoma in mice. Cancer Cell 15, 294–303 (2009). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). Frezza, C. et al. Haem oxygenase is synthetically lethal with the tumour suppressor fumarate hydratase. Nature 477, 225–228 (2011). Selivanov, V. A., Marin, S., Lee, P. W. N. & Cascante, M. Software for dynamic analysis of tracer-based metabolomic data: estimation of metabolic fluxes and their statistical analysis. Bioinformatics 22, 2806– 2812 (2006). de Mas, I. M. et al. Compartmentation of glycogen metabolism revealed from 13C isotopologue distributions. BMC Syst Biol 5, 175 (2011). Unterluggauer, H. et al. Premature senescence of human endothelial cells induced by inhibition of glutaminase. Biogerontology 9, 247–259 (2008). Gao, P. et al. c-Myc suppression of miR23a/b enhances mitochondrial glutaminase expression and glutamine metabolism. Nature 458, 762–765 (2009). Lobo, C. et al. Inhibition of glutaminase expression by antisense mRNA decreases growth and tumourigenicity of tumour cells. Biochem J 348 Pt 2, 257–261 (2000). Wang, J.-B. et al. Targeting mitochondrial glutaminase activity inhibits oncogenic transformation. Cancer Cell 18, 207–219 (2010). Hensley, C. T., Wasti, A. T. & DeBerardinis, R. J. Glutamine and cancer: cell biology, physiology, and clinical opportunities. J. Clin. Invest. 123, 3678–3684 (2013). Maeda, M., Scaglia, N. & Igal, R. A. Regulation of fatty acid synthesis and Delta9-desaturation in senescence of human fibroblasts. Life Sci. 84, 119–124 (2009). Gey, C. & Seeger, K. Metabolic changes during cellular senescence investigated by proton NMR-spectroscopy. Mechanisms

36.

37.

38.

39. 40. 41. 42.

Chapter 3

35.

of Ageing and Development 134, 130–138 (2013). Aboagye, E. O. & Bhujwalla, Z. M. Malignant transformation alters membrane choline phospholipid metabolism of human mammary epithelial cells. Cancer Res 59, 80–84 (1999). Glunde, K., Bhujwalla, Z. M. & Ronen, S. M. Choline metabolism in malignant transformation. Nat Rev Cancer 11, 835–848 (2011). Iorio, E. et al. Alterations of choline phospholipid metabolism in ovarian tumor progression. Cancer Res 65, 9369–9376 (2005). Quijano, C. et al. Oncogene-induced senescence results in marked metabolic and bioenergetic alterations. cc 11, 1383–1392 (2012). Santos, C. R. & Schulze, A. Lipid metabolism in cancer. FEBS J 279, 2610–2623 (2012). Wymann, M. P. & Schneiter, R. Lipid signalling in disease. Nat Rev Mol Cell Biol 9, 162–176 (2008). Frezza, C. et al. Metabolic profiling of hypoxic cells revealed a catabolic signature required for cell survival. PLoS ONE 6, e24411 (2011). Chaneton, B. et al. Serine is a natural ligand and allosteric activator of pyruvate kinase M2. Nature (2012). doi:10.1038/ nature11540

Glucose +0 Glucose +1 Glucose +5 Glucose +6 Sum GA3P +0 GA3P +3 Sum Pyr +0 Pyr +1 Pyr +2 Pyr +3 Sum Lac +0 Lac +1 Lac +2 Lac +3 Sum ALA +0 ALA +1 ALA +2 ALA +3 Sum Citrate +0 Citrate +1 Citrate +2 Citrate +3 Citrate +4 Citrate +5 Citrate +6 Sum a-KG +0 a-KG +1 a-KG +2 a-KG +3 a-KG +4 a-KG +5 Sum Mal +0 Mal +1 Mal +2 Mal +3 Sum GLU +0 GLU +1 GLU +2 GLU +3 GLU +4 GLU +5 Sum GLN +0 GLN +1 Sum

Compounds

SD(±) 2,77 0,21 0,30 4,29 6,54 0,13 0,16 0,17 0,01

0,01 0,01 1,21 0,06 0,01 0,33 1,36 0,21 0,01 0,00 0,02 0,24 0,13 0,01 0,00

0,14 0,04 0,00

0,04 0,05 0,00

0,00 0,06 2,86 0,17

3,03 3,83 0,18 4,00

Mean 12,43 0,74 5,24 97,78 116,19 1,92 1,68 3,60 0,10

0,05 0,15 18,26 0,55 0,16 5,81 24,78 5,87 0,18 0,00 0,66 6,71 0,76 0,05 0,01

0,81 0,52 0,03

0,55 0,74 0,03

0,00 0,77 113,58 6,45

120,02 112,98 6,33 119,31

Cycling 15min

120,75 116,39 6,56 122,95

0,01 1,02 114,28 6,43 0,04

0,72 0,97 0,04

1,11 0,69 0,03

0,10 0,21 24,32 0,74 0,25 9,35 34,66 5,42 0,16 0,00 0,73 6,31 1,02 0,09 0,01

Mean 8,25 0,51 6,52 120,03 135,31 1,14 1,89 3,04 0,10

1,25 7,60 0,47 8,07

0,00 0,03 1,20 0,05 0,01

0,06 0,03 0,00

0,33 0,06 0,00

0,03 0,07 8,91 0,31 0,04 1,39 10,60 0,36 0,01 0,00 0,02 0,39 0,30 0,04 0,00

SD(±) 2,80 0,20 2,27 42,95 48,15 0,02 0,18 0,17 0,04

Cycling 30min

122,37 108,63 6,19 114,81

0,01 1,31 115,31 6,51 0,56

0,80 1,25 0,05

4,28 4,29 0,27 4,55

0,00 0,12 3,95 0,25 0,09

0,11 0,12 0,00

0,26 0,10 0,01

0,07 0,48 0,24 0,02 0,00

0,90 6,24 0,65 0,04 0,01

0,70 0,76 0,04

0,04 0,08 6,13 0,17 0,06 2,07 8,16 0,40 0,01

SD(±) 2,59 0,19 1,72 29,70 34,06 0,15 0,12 0,24 0,04

0,11 0,20 19,53 0,64 0,30 11,27 31,74 5,18 0,15

Mean 9,78 0,67 5,89 106,51 122,85 1,20 2,12 3,32 0,08

Cycling 60min

124,02 126,63 7,17 133,80

1,13 1,56 0,06 0,00 0,02 1,64 115,63 6,60 1,79

1,13 1,07 0,06

1,79 7,73 1,03 0,06 0,04

0,30 0,43 21,48 0,69 0,66 24,19 47,02 5,77 0,17

Mean 8,29 0,54 7,13 128,50 144,47 1,00 2,48 3,48 0,13

5,01 6,20 0,36 6,56

0,14 0,17 0,01 0,00 0,00 0,17 4,62 0,32 0,07

0,25 0,13 0,01

0,09 0,38 0,23 0,02 0,01

0,07 0,10 5,22 0,18 0,14 4,87 10,39 0,31 0,01

SD(±) 1,56 0,14 1,44 25,58 28,54 0,08 0,22 0,20 0,03

Cycling 3 hr

121,47 113,85 6,47 120,32

1,97 1,99 0,09 0,08 0,04 2,19 107,19 6,29 7,38 0,42 0,20

1,58 1,75 0,10 0,11 0,01 0,00

3,93 9,35 1,12 0,08 0,34 0,02 0,01 0,01

Mean 7,17 0,45 5,10 92,42 105,14 1,05 3,63 4,68 0,17 0,01 0,01 0,89 1,07 21,95 0,65 1,44 54,51 78,55 5,25 0,16

3,67 5,77 0,31 6,08

0,28 0,07 0,00 0,00 0,00 0,07 3,17 0,22 0,32 0,05 0,02

0,35 0,25 0,01 0,01 0,00 0,00

0,43 0,95 0,29 0,02 0,05 0,00 0,00 0,00

SD(±) 1,28 0,08 1,33 24,68 27,26 0,04 0,25 0,29 0,05 0,00 0,00 0,11 0,16 5,56 0,21 0,40 13,30 18,96 0,50 0,01

Cycling 12 hr

2,84 2,45 0,14 0,30 0,03 0,02 0,00 2,94 1,96 0,10 0,14 0,07 2,26 92,59 5,65 11,52 0,99 0,73 0,01 111,49 122,03 6,93 128,96

6,91 12,53 1,93 0,15 0,64 0,04 0,06 0,02

Mean 9,64 0,57 6,29 112,90 129,41 1,01 5,37 6,38 0,28 0,01 0,05 2,17 2,51 33,53 1,07 3,61 134,46 172,66 5,44 0,17

0,77 0,17 0,01 0,02 0,00 0,00 0,00 0,20 0,16 0,01 0,01 0,00 0,18 1,49 0,13 0,02 0,03 0,03 0,00 1,64 6,86 0,39 7,25

0,30 0,56 0,57 0,04 0,13 0,01 0,01 0,00

SD(±) 2,49 0,17 1,75 31,54 35,85 0,13 0,56 0,68 0,06 0,00 0,01 0,62 0,69 9,01 0,31 0,95 33,55 43,46 0,26 0,01

Cycling 24 hr

63,08 80,39 4,55 84,94

0,34 58,99 3,33 0,76

0,20 0,33 0,01

2,42 0,19 0,01

0,10 2,54 2,15 0,15 0,11 0,01

0,04 0,09 29,21 0,80 0,21 8,59 38,81 2,38 0,06

Mean 5,92 0,37 6,57 90,65 103,51 2,09 6,46 8,54 0,05

OIS 15min

4,37 16,27 0,91 17,17

0,07 4,13 0,21 0,03

0,02 0,07 0,00

0,90 0,02 0,00

0,01 0,41 0,82 0,06 0,03 0,00

0,01 0,03 10,42 0,27 0,02 1,80 12,48 0,40 0,01

SD(±) 1,29 0,12 2,24 26,24 29,27 0,08 0,84 0,82 0,02

67,92 89,49 4,99 94,47

0,26 0,50 0,02 0,00 0,00 0,52 62,76 3,54 1,62

3,12 0,24 0,01 0,00

0,12 2,80 2,77 0,17 0,17 0,01

0,07 0,13 43,59 1,31 0,34 12,01 57,24 2,61 0,07

Mean 8,43 0,60 9,49 169,94 188,46 1,46 4,62 6,08 0,07

OIS 30min

1,78 9,74 0,47 10,21

0,01 0,03 0,00 0,00 0,00 0,03 1,67 0,10 0,02

0,72 0,01 0,00 0,00

0,01 0,16 0,67 0,04 0,01 0,00

0,02 0,02 6,46 0,17 0,05 1,07 7,63 0,15 0,01

SD(±) 1,03 0,08 2,10 38,83 41,95 0,05 0,33 0,37 0,02

Supplementary Table 1. Intracellular isotopomer quantities [nmol/106cells] at the indicated time point (n=3)

75,86 79,48 4,46 83,94

0,29 0,53 0,02 0,01 0,00 0,57 68,90 3,71 3,11 0,11 0,03

2,28 0,27 0,01 0,01

0,16 2,78 1,91 0,12 0,24 0,01

0,08 0,13 29,83 0,92 0,37 14,21 45,33 2,56 0,07

Mean 6,57 0,44 5,83 103,72 116,56 1,56 4,60 6,16 0,05

OIS 60min

2,42 9,42 0,57 9,99

0,04 0,09 0,00 0,00 0,00 0,10 1,78 0,65 0,17 0,02 0,01

0,48 0,03 0,00 0,00

0,00 0,21 0,47 0,03 0,02 0,00

0,02 0,04 10,71 0,30 0,08 3,22 14,16 0,21 0,00

SD(±) 1,25 0,11 1,27 24,14 26,72 0,26 0,38 0,61 0,01

75,42 100,26 5,72 105,98

0,37 0,67 0,03 0,04 0,01 0,76 63,43 4,04 6,90 0,66 0,40

3,33 0,32 0,02 0,03 0,00 0,00

0,40 3,45 2,52 0,18 0,52 0,05 0,04 0,01

0,23 0,33 41,41 1,28 0,93 34,86 78,48 2,98 0,08

Mean 8,11 0,55 8,12 143,14 159,92 1,37 4,67 6,04 0,10

OIS 3 hr

3,61 3,31 0,18 3,49

0,01 0,04 0,00 0,00 0,00 0,05 3,37 0,23 0,27 0,02 0,05

0,55 0,01 0,00 0,00 0,00 0,00

0,01 0,16 0,41 0,03 0,09 0,01 0,01 0,00

0,03 0,03 2,16 0,07 0,06 1,46 1,15 0,16 0,00

SD(±) 0,23 0,02 0,54 10,09 10,47 0,19 0,47 0,65 0,01

Mean 8,85 0,57 6,92 122,67 139,01 1,21 4,73 5,94 0,18 0,00 0,01 0,77 0,96 38,76 1,24 2,45 101,13 143,58 3,39 0,10 0,00 0,88 4,38 2,19 0,17 1,00 0,12 0,15 0,07 0,00 3,69 0,42 0,03 0,08 0,01 0,01 0,00 0,56 0,57 0,04 0,08 0,04 0,72 51,22 3,70 10,98 1,68 1,37 0,27 69,23 99,31 5,63 104,94

OIS 12 hr SD(±) 1,48 0,16 2,28 37,68 41,58 0,09 0,59 0,67 0,04 0,00 0,00 0,21 0,25 10,83 0,27 0,56 24,96 36,59 0,26 0,01 0,00 0,12 0,39 0,72 0,05 0,16 0,01 0,02 0,01 0,00 0,95 0,06 0,00 0,01 0,00 0,00 0,00 0,08 0,10 0,01 0,01 0,00 0,13 1,71 0,12 0,27 0,05 0,05 0,02 2,17 8,75 0,48 9,23

SD(±) 0,83 0,06 0,78 12,16 13,75 0,05 0,30 0,32 0,01 0,00 0,11 0,13 4,99 0,16 0,51 21,42 26,83 0,16 0,01 0,00 0,13 0,29 0,25 0,02 0,11 0,02 0,02 0,02 0,00 0,43 0,03 0,00 0,01 0,00 0,00 0,00 0,05 0,03 0,00 0,01 0,01 0,04 0,77 0,06 0,32 0,11 0,14 0,05 1,43 2,01 0,12 2,13

Mean 5,39 0,33 4,55 80,61 90,88 1,30 7,54 8,84 0,07 0,01 0,58 0,66 30,59 1,06 3,66 132,47 167,78 3,67 0,11 0,01 2,04 5,82 1,98 0,19 1,60 0,23 0,37 0,19 0,03 4,59 0,58 0,04 0,17 0,03 0,03 0,01 0,87 0,60 0,05 0,14 0,07 0,85 48,89 3,88 15,68 3,22 3,17 1,02 75,86 96,27 5,45 101,72

OIS 24 hr

1,1 0,4

1,1

6,3 4,9 88,8 100

84,1 2,7

13,2 100

Glucose +0 Glucose +5 Glucose +6 Sum

Pyr +0 Pyr +1 Pyr +2 Pyr +3 Sum

96,7 2,8 0,4 100

94,6 5,4 100

95,5 4,5

ALA +0 ALA +1 ALA +3 Sum

GLN +0 GLN +1 Sum

GLU +0 GLU +1 GLU +2 GLU +3 GLU +4 GLU +5 Sum

100

0,3

4,0 100

0,2 0,2

0,0 0,0

0,1 0,1 0,0

0,2 0,1

93,2 2,8

Lac+ 0 Lac +1 Lac +2 Lac +3 Sum

0,1 0,0 0,1

Cycling 15min Mean ±SD

Compounds

100

95,4 4,6

94,6 5,4 100

96,1 2,8 1,1 100

10,4 100

86,8 2,7

27,2 100

70,5 2,3

4,9 4,9 90,2 100

0,1 0,1

0,0 0,0

0,0 0,1 0,1

0,7

0,7 0,0

0,5

0,4 0,3

0,5 0,0 0,5

Cycling 30min Mean ±SD

100

95,5 4,5

94,6 5,4 100

94,3 2,8 3,0 100

21,4 100

76,2 2,3

52,9 1,8 1,0 44,4 100

5,3 4,9 89,8 100

0,0 0,0

0,4 0,0 0,3

0,9

0,9 0,0

0,9 0,1 0,1 0,8

1,2 0,1 1,1

Cycling 60min Mean ±SD

100

95,3 4,7

94,6 5,4 100

86,1 2,6 11,3 100

52,7 1,6 1,2 44,5 100

35,6 1,1 1,4 61,8 100

4,9 4,9 90,2 100

0,2 0,2

0,0 0,0

0,1 0,0 0,0

1,5 0,0 0,0 1,5

1,7 0,1 0,1 1,9

0,6 0,0 0,6

Cycling 3hr Mean ±SD

100

93,6 4,9 1,5

94,6 5,4 100

59,1 1,8 39,1 100

28,1 0,9 1,9 69,1 100

21,7 0,7 1,9 75,7 100

5,9 4,9 89,2 100

0,1 0,0 0,1

0,0 0,0

0,8 0,0 0,8

0,7 0,0 0,0 0,7

0,7 0,0 0,0 0,8

0,3 0,1 0,4

Cycling 12hr Mean ±SD

100

91,3 5,0 3,7

94,6 5,4 100

40,9 1,3 57,8 100

19,2 0,7 2,2 78,0 100

15,3 0,5 2,1 82,1 100

5,6 4,9 89,5 100

0,2 0,1 0,1

0,0 0,0

1,1 0,0 1,1

1,0 0,0 0,1 1,0

0,3 0,0 0,2 0,2

0,6 0,0 0,5

Cycling 24hr Mean ±SD

100

95,5 4,5

94,6 5,4 100

96,9 2,9 0,2 100

4,8 100

92,3 2,8

19,7 100

78,0 2,2

4,3 5,0 90,7 100

0,1 0,1

0,0 0,0

0,0 0,0 0,0

0,5

0,6 0,0

1,2

1,8 0,6

0,7 0,0 0,7

OIS 15min Mean ±SD

Supplementary Table 2. Extracellular isotopomer distribution [%] at the indicated time (n=3)

100

95,3 4,7

94,6 5,4 100

96,7 2,8 0,5 100

10,5 100

86,8 2,7

31,0 100

67,1 1,9

4,3 5,0 90,7 100

100

95,2 4,8

94,6 5,4 100

96,1 2,8 1,1 100

17,3 100

80,3 2,4

39,9 100

58,3 1,8

5,5 4,9 89,6 100

0,1 0,1

0,0 0,0

0,1 0,1 0,1

1,5

1,5 0,1

2,3

1,9 0,4

0,7 0,0 0,7

OIS 60min Mean ±SD

Chapter 3

0,1 0,1

0,0 0,0

0,0 0,0 0,0

0,7

0,6 0,1

1,2

1,4 0,2

0,3 0,1 0,3

OIS 30min Mean ±SD

100

93,8 4,9 1,2

94,6 5,4 100

92,5 2,7 4,8 100

59,0 1,8 1,0 38,1 100

41,3 1,4 1,3 56,0 100

5,0 4,9 90,0 100

OIS 3hr Mean

0,0 0,1 0,0

0,0 0,0

0,3 0,1 0,3

2,7 0,1 0,1 2,8

3,3 0,1 0,2 3,1

0,6 0,0 0,5

±SD

100

87,0 5,1 6,8 0,7 0,5

94,6 5,4 100

75,6 2,3 22,0 100

30,8 1,0 1,8 66,3 100

23,7 0,8 1,9 73,6 100

6,1 4,9 89,1 100

0,5 0,1 0,4 0,1 0,0

0,0 0,0

1,1 0,0 1,2

1,4 0,0 0,0 1,4

1,2 0,0 0,1 1,1

0,4 0,0 0,4

OIS 12hr Mean ±SD

79,6 5,1 11,9 1,7 1,4 0,2 100

94,6 5,4 100

54,8 1,7 43,5 100

18,9 0,7 2,3 78,2 100

15,2 0,6 2,2 82,0 100

5,7 4,9 89,4 100

0,4 0,1 0,2 0,1 0,0 0,1

0,0 0,0

0,3 0,0 0,3

0,4 0,0 0,1 0,5

0,8 0,0 0,0 0,8

0,3 0,1 0,3

OIS 24hr Mean ±SD

CHAPTER 4 A CRITICAL ROLE FOR THE MITOCHONDRIAL GATEKEEPER PYRUVATE DEHYDROGENASE IN ONCOGENE-INDUCED SENESCENCE

Nature 498, 109–112 (2013)

A CRITICAL ROLE FOR THE MITOCHONDRIAL GATEKEEPER PYRUVATE DEHYDROGENASE IN ONCOGENE-INDUCED SENESCENCE Joanna Kaplon1, Liang Zheng2*, Katrin Meissl1*, Barbara Chaneton2, Vitaly A. Selivanov3,4, Gillian MacKay2, Sjoerd H. van der Burg5, Elizabeth M. E. Verdegaal5, Marta Cascante3,4, Tomer Shlomi6,7, Eyal Gottlieb2& Daniel S. Peeper1 Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam,

The Netherlands. 2Cancer Research UK, Beatson Institute for Cancer Research, Switchback Road, Glasgow G61 1BD, Scotland, UK. 3Department of Biochemistry and Molecular Biology, Faculty of Biology, Universitat de Barcelona, Av Diagonal 643, 08028 Barcelona, Spain. 4Institute of Biomedicine of Universitat de Barcelona (IBUB) and CSIC-Associated Unit, Spain. 5Laboratory of Clinical Oncology, Leiden University Medical Center, of Technology, Haifa, 32000, Israel. 7Lewis-Sigler Institute for Integrative Genomics, Princeton University, Princeton, NJ, USA. *These authors contributed equally to this work. doi: 10.1038/nature12154

SUMMARY In response to tenacious stress signals, such as the unscheduled activation of oncogenes, cells can mobilize tumor suppressor networks to avert the hazard of malignant transformation. A large body of evidence indicates that oncogene-induced senescence (OIS) acts as such a break, withdrawing cells from the proliferative pool virtually irreversibly, thus crafting a vital pathophysiologic mechanism protecting against cancer1-5. In spite of the widespread contribution of OIS to the cessation of tumorigenic expansion in animal models and humans, we have only begun to define the underlying mechanism and identify the key players6. While deregulation of metabolism is intimately linked to the proliferative capacity of cells7-10, and in spite of the wide belief that senescent cells remain metabolically active11, little has been investigated in detail about the role of cellular metabolism in OIS. Here we show by metabolic profiling and functional perturbations that the mitochondrial gatekeeper pyruvate dehydrogenase (PDH) is a crucial mediator of senescence induced by BRAFV600E, an oncogene commonly mutated in melanoma and other cancers. BRAFV600E-induced senescence was accompanied by simultaneous suppression of the PDH-inhibitory enzyme pyruvate dehydrogenase kinase 1 (PDK1) and induction of the PDH-activating enzyme pyruvate dehydrogenase phosphatase 2 (PDP2). The resulting concerted activation of PDH enhanced utilization of pyruvate in the tricarboxylic acid (TCA) cycle, causing increased respiration and redox stress. Abrogation of OIS, a rate-limiting step towards oncogenic transformation, coincided with reversion of these processes. Further supporting a critical role of PDH in OIS, enforced normalization

Chapter 4

Albinusdreef 2, 2333 ZA Leiden, The Netherlands. 6Computer Science Department, Technion, Israel Institute

of either PDK1 or PDP2 expression levels inhibited PDH and abrogated OIS, thereby licensing BRAFV600E-driven melanoma development. Finally, depletion of PDK1 eradicated melanoma subpopulations resistant to targeted BRAF inhibition, and caused regression of established melanomas. These results reveal a mechanistic relationship between OIS and a key metabolic signaling axis, which may be exploited therapeutically. We compared the metabolism of human diploid fibroblasts (HDF) undergoing OIS to that of cycling cells. To evoke OIS, we used oncogenic BRAFV600E, a common oncogene and strong inducer of OIS in vitro and in vivo12-16. We first analyzed the oxygen consumption rate (OCR). Compared to cycling cells, “OIS cells” showed a significant increase in OCR, indicative of increased mitochondrial oxidative metabolism (Figure 1a). This was further supported by exchange rate (uptake or secretion) measurements of key metabolites: OIS cells secreted less than half the amount of pyruvate compared to cycling cells, which was not balanced by a significant change in glucose consumption nor an increase in lactate or alanine secretion (Figure 1b, c, Supplementary Figure 2a, b). While secretion of glutamate was increased in OIS (Figure 1d), glutamine consumption was decreased (Supplementary Figure 2c). In agreement with increased TCA cycle activity in OIS cells, stable isotope labeling with uniformly labeled [U-13C6]-glucose tracer showed faster accumulation of glucose-derived 2-carbon-labeled metabolic isotopomers of the TCA cycle (citrate, alpha-ketoglutarate and malate) as well as TCA cycle-derived 2-carbon-labeled glutamate (Supplementary Figure 2d). Together, these results indicate that OIS is accompanied by increased pyruvate oxidation, which was corroborated by a simultaneous rise in redox stress, as measured by increased reactive oxygen species (ROS) production, a decrease in the reduced:oxidized glutathione (GSH/GSSG) ratio and induction of several ROS-responsive genes (Supplementary Figure 3ai). The gatekeeping enzyme linking glycolysis to the TCA cycle is pyruvate dehydrogenase (PDH)17,18, which is regulated by reversible phosphorylation19-22: phosphorylation by pyruvate dehydrogenase kinases (PDK1 through 4) inhibits its action and halts pyruvate utilization in TCA cycle, whereas dephosphorylation by pyruvate dehydrogenase phosphatases (PDP1 and 2) stimulates PDH activity. Our metabolic profiles predicted PDH to be activated in OIS. Furthermore, if increased PDH activity has a causal role in mediating OIS, we expected this to be reversed upon OIS escape (by C/EBPb depletion13 or SV40 small t expression; Figure 1e, Supplementary Figure 4a). Indeed, PDH phosphorylation (on serine 293, one of three phosphorylation sites inactivating PDH) was strongly reduced in OIS cells, but this was restored upon senescence abrogation (Figure 1f, Supplementary Figure 4b). These changes in PDH phosphorylation translated into corresponding alterations in its activity (Figure 1g). Strikingly, in OIS, the two key PDH-modifying enzymes were regulated in opposite directions: whereas PDK1 was downregulated, PDP2 was induced (Figure 1f, h). Upon OIS

or BR

sh-C/EBPβ

g in

PDK1 PDP2

3.5 3 2.5 2 1.5 1 0.5 0

00 V6

cl in incon g hi ta bi c tio t n

PDP2 β-actin

PDK1

PDHE1α

Relative amount of transcript

o BR r AF V6

β-actin

pPDHE1α

cl in c g in on hi ta bi ct tio n

PDK1 PDP2

cl in RA g SG

PDHE1α

cli cy E

E 00 V6

BRAF pPDHE1α

ng OI S

cli cy cy

sh-C/EBPβ

% BrdU-positive cells

ng OI S E

g BR A ve F V60 0E ct o BR r AF

sh-C/EBPβ

cli

ng OI S

200

cli

400

***

600

100

800

glutamate secretion [nmol/(106cells x h)]

***

150

glucose uptake [nmol/(106cells x h)]

200

***

pyruvate secretion [nmol/(106cells x h)]

250

OCR [nmol/(106cells x h)]

Chapter 4

abrogation, their levels were normalized. Other PDK and PDP isozymes were not regulated in this manner (Supplementary Figure 4c, d). Similarly to cells undergoing BRAFV600E-induced senescence, RASG12V-senescent cells showed increased accumulation of glucose-derived TCA cycle metabolites (Supplementary Figure 5a), induction of OCR (Supplementary Figure 5b), a drop in PDH phosphorylation (albeit moderately, Supplementary Figure 5c) and increased PDH enzymatic activity (Figure 1i), but this was not accompanied by effects on PDK1 or PDP2 levels (Supplementary Figure 5c). In quiescent cells, labeling of pyruvate, lactate, alanine and citrate from glucose-derived carbons was synchronously increased over time at a higher rate than in proliferating cells (Supplementary Figure 6a, b), as previously reported23. However, the alterations in PDH activity were specific for OIS rather than a consequence of cell cycle arrest, as neither PDH phosphorylation nor PDH enzymatic activity was altered in quiescent cells (Figure 1 j, k). These results indicate that OIS, and escape thereof, is accompanied by antagonistic regulation of two key enzymes controlling PDH activity, PDP2 and PDK1.

sh-C/EBPβ

Figure 1. The PDK1-PDP2-PDH axis is deregulated in OIS a-d. Analysis of OCR (a), pyruvate secretion (b), glucose uptake (c) and glutamate secretion (d) in cycling and OIS HDF (upon expression of BRAFV600E); n=6. e-h. Cycling cells and cells undergoing OIS or abrogating OIS (upon C/EBPb depletion) were analyzed for BrdU incorporation (e), by immunoblotting (f), for PDH activity (g) and for regulation of PDK1 and PDP2 transcripts, as determined by qRT-PCR (h); n=3. i. PDH activity of HDF undergoing RASG12V-induced senescence. j-k. Immunoblotting analysis (j) or PDH activity (k) of cycling and quiescent HDF. All data are represented as mean ±s.d.

To determine whether deregulation of the PDK1-PDP2-PDH axis drives OIS-associated metabolic rewiring, we depleted PDP2. This reversed the decrease in PDH phosphorylation in OIS, leading to suppressed PDH activity (Figure 2a, b). Consistently, [U-13C6]-glucose and [13C3]-pyruvate labeling revealed that PDP2-depleted cells had less labeling of citrate (Supplementary Figure 7, Supplementary Figure 8a). In agreement, the ratio of labeled [13C2]-

citrate (emulating PDH product) to [13C3]-pyruvate (PDH substrate), indicating intracellular PDH activity, was neutralized by PDP2 depletion (Figure 2c). These cells decreased also OCR and redox stress (Figure 2d, Supplementary Figure 3b-i).

BRAF

V600E

vector

91.1±1.5

18.7±3.5

150 100

vector

shPDP2 #4

cycling

*** *

shPDP2

vector

BRAFV600E shPDP2 #4 #1

shPDP2

BRAFV600E

shPDP2

200

12.5±3.2

Relative amount of transcript

β-actin

cycling 1

MEK

250

BRAFV600E

f vector

BRAF pMEK

BRAFV600E

cycling

p15

60 60 50 50 40 40 30 30 20 20 10 10 00

% BrdU-positive cells

PDHE1α

0.0

cycling

0.4 0.2

pPDHE1α

4.7±1.8

*** ***

0.8 0.6

PDP2

cycling

1.0

OCR [nmol/(106 cells x h)]

shPDP2

%CIT+2/ %PYR+3

BRAFV600E #4

shPDP2

vector

cycling

BRAFV600E vector

cycling

1.2 1 0.8 0.6 0.4 0.2 0

1.2

IL6 IL8 C/EBP β

0.6 0.4 0.2

#1 #4 shPDP2 V600E BRAF

cycling vector

Figure 2. PDP2 regulates metabolic rewiring and OIS a. Immunoblotting analysis of HDF expressing empty vector or sh-PDP2 in the presence or absence of BRAFV600E. b-g. Cells from a were analyzed for PDH activity in cell extracts (b) or intracellularly, as denoted by [13C2]-citrate/[13C3]-pyruvate ratio (c), OCR (d), BrdU incorporation (e), cell proliferation (f) and SA-b-Gal activity (g); n=3. h. Regulation of IL8, IL6 and C/EBPb transcripts of the samples described in a, as determined by qRT-PCR; n=3. All data are represented as mean ±s.d.

We next investigated whether PDP2 regulates OIS. Indeed, its depletion abrogated BRAFV600Einduced arrest, which was not explained by loss of BRAFV600E signaling (Figure 2a, e, f). Senescence bypass was accompanied by a marked reduction in the levels of the senescenceassociated tumor suppressor p15INK4B (Figure 2a) and decreased activity of the SA-b-Gal senescence biomarker (Figure 2g). Furthermore, the induction of the OIS-associated and C/EBPb-dependent interleukins (ILs) 6 and 813 was curtailed by PDP2 depletion (Figure 2h). Because PDK1 (whose expression is suppressed by BRAFV600E) antagonizes PDP2 in regulating PDH activity, we addressed also its role in OIS. Ectopic restoration of PDK1 rescued the decrease in PDH phosphorylation and suppressed the increase in PDH activity in OIS (Figure 3a, b). Consistently, PDK1 expression reversed the increase in TCA cycle activity in OIS and blocked the rise in PDH activity as judged from the [13C2]-citrate : [13C3]-pyruvate ratio (Figure 3c, Supplementary Figure 7, Supplementary Figure 8b). These effects were mirrored by significant decreases in OCR and redox stress (Figure 3d, Supplementary Figure 3b-i). Restoration of PDK1 expression abrogated the induction of cell cycle arrest upon BRAFV600E

expression (Figure 3e, f), which was paralleled by suppression of several senescenceassociated biomarkers (Figure 3a, g, h). The OIS-associated metabolic rewiring was specific to the PDK1-PDP2-PDH axis: depletion of lactate dehydrogenase A (LDHA), which stimulates mitochondrial respiration in tumor cells24, did not change OCR nor promoted senescence, while BRAFV600E did not affect LDHA protein levels (Supplementary Figure 9a-f). Collectively, these results show that PDP2 and PDK1 are critically required for the metabolic wiring associated with, and the execution of, OIS.

BRAF

cycling

PDK1

vector PDK1

***

600

PDK1

cl in ve g ct o PD r K1

BRAFV600E

BRAFV600E PDK1 vector

1.2 1 0.8 0.6 0.4 0.2 0

IL6 IL8 C/EBP β

cycling PDK1 vector PDK1 BRAFV600E PDK1

PDK1

vector

10 20 30 days after injection

vector

PDK1

pPDHE1α

PDHE1α

cycling

***

300

PDK1

2.2±0.5 88.6±2.6 34.6±6.2

900

Tumor volume (mm3)

BRAFV600E

BRAFV600E vector PDK1

4.1±0.3

BRAFV600E

Relative amount of transcript

PDK1

Hsp90

vector

MEK

cycling

***

Chapter 4

pMEK

ve g ct o PD r K1

p15

200 100

PDHE1α

0.0

250

PDK1

vector

PDK1

0.2

150

0.1

% BrdU-positive cells

*** ***

0.4 0.3

PDK1 pPDHE1α

0.5

OCR [nmol/(106 cells x h)]

BRAFV600E

%CIT+2/ %PYR+3

PDK1

vector

PDK1

cycling

BRAFV600E

cycling

Figure 3. PDK1 regulates metabolic rewiring and OIS, and acts tumorigenically a. Immunoblotting analysis of HDF expressing empty vector or PDK1 in the presence or absence of BRAFV600E. b-h. Cells from a were analyzed for PDH activity in cell extracts (b), or intracellularly, as denoted by [13C2]citrate/[13C3]-pyruvate ratio (c), OCR (d), BrdU incorporation (e), cell proliferation (f), SA-b-Gal activity (g) and IL8, IL6 and C/EBPb transcripts (h); n=3. i. Growth curve of tumors formed by sh-p53/BRAFV600E melanocytes expressing empty vector or PDK1; n=8. j-k. Representative pictures (j) and immunohistochemical staining (k) of tumors described in i. Data are represented as mean ±s.d. (c-e, g, h) or ±s.e.m (i).

As OIS represents a rate-limiting step in oncogenic transformation25, we next investigated whether PDK1 can act oncogenically. Melanocytes from the skin of neonatal Tyr::CreER;BrafCA mice14 were depleted from p53, treated with tamoxifen to induce BRAFV600E expression and

infected with PDK1-encoding or control virus. Whereas transplanted sh-p53/BRAFV600E control melanocytes failed to form tumors, PDK1-expressing cells produced large tumors (Figure 3i, j, Supplementary Figure 10a). This was associated with robust PDH phosphorylation (Figure 3k, Supplementary Figure 10b). These results raised the possibility that, conversely, PDK1 depletion acts cytostatically. Indeed, its silencing from non-transformed cells induced proliferative arrest (Figure 4a, b Supplementary Figure 11a, b). This was accompanied by suppression of the DNA replication-associated protein PCNA and induction of several senescence biomarkers and tumor suppressors (Figure 4a, c, Supplementary Figure 11b, c), thereby underscoring the importance of PDK1 for cellular senescence. PDK1-depleted cells also showed decreased PDH phosphorylation, enhanced PDH activity and increased OCR (Figure 4a, d, e). Thus, PDK1 depletion from non-transformed cells causes senescence. Unexpectedly, in a panel of human BRAF mutant melanoma cell lines, this cell cycle arrest was followed by cell death a few days later (Figure 4f, Supplementary Figure 11d, e). This was not caused by differences in the extent of PDK1 knockdown, PDH phosphorylation or activity, nor OCR (Supplementary Figure 11f-h, Supplementary Figure 12a, b). This raised the intriguing possibility that PDK1 depletion negatively affects melanoma growth. Indeed, inoculation of (viable) melanoma cell lines into immunocompromised mice prevented tumor outgrowth almost completely (Figure 4g, h, Supplementary Figure 13a-d). A few small lesions that did develop invariably had lost PDK1 knockdown (Supplementary Figure 13b, e), indicating that PDK1 is essential for melanoma outgrowth in vivo. Because clinically, it would be more relevant to assess the role of PDK1 in tumor maintenance and progression than in initiation, we generated a doxocycline (DOX)-inducible shRNA system. DOX administration suppressed PDH phosphorylation and concomitantly caused melanoma cell death in vitro (Supplementary Figure 14a-c). In mice, uninduced sh-PDK1 cells produced tumors indistinguishably from control cells (Supplementary Figure 14d). In contrast, when DOX was administered starting from the time of injection, PDK1-depleted cells failed to produce tumors (Supplementary Figure 14e). Most importantly, when DOX was administered after melanomas had established, PDK1 depletion led to near-complete tumor regression, thereby greatly extending animal survival (Figure 4i-k). These results demonstrate that PDK1 is not only required for tumor initiation, but also for tumor maintenance and progression, implying that it may be beneficial to target this metabolic kinase for therapeutic intervention of BRAF mutant melanoma. Finally, we examined whether PDK1 depletion sensitizes BRAF mutant melanoma cells to treatment with PLX4720 (a preclinical analogue of vemurafenib, a specific clinical BRAFV600E inhibitor26,27). Dose-response curves of four BRAFV600E melanoma cell lines that are sensitive to PLX4720 (>90% cell death after treatment with 40 ÎźM PLX4720) and another four that are partially resistant (>20% cells surviving treatment with 40 ÎźM PLX4720) revealed that

PDK1

% BrdU-positive cells

shPDK1 #5

2.5±1.7

OCR [nmol/ (106 cells /h)]

shPDK1

mut BRAF melanoma nontransformed

300 200 0

10 20 30 40 days after injection

600

450

300

DOX *

150 0

80 60 40 20 0

shluc shPDK1

30 60 90 days after injection

shluc shPDK1 shluc DOX shPDK1 DOX

20 40 60 80 100 days after injection

vector

shPDK1 #3

shPDK1

100

Chapter 4

vector shPDK1 #3 shPDK1 #5

400

100

Percent survival

500

***

100

Residual cell fraction (% of total)

Tumor volume (mm3)

Fold cell death induction upon PDK1 depletion

***

200 150

shPDK1

β-actin

***

72.1±5.9 60.0±4.4

vector

p27

vector

p15

p16

vector

PCNA

250

50 40 30

shluc

shPDK1

Day 20 DOX D0

PDHE1α

Day 48 DOX D28

pPDHE1α

shPDK1 #3

vector

PDK1 depletion strongly sensitized all cell lines to BRAF inhibition (Supplementary Figure 15a, b). Remarkably, PDK1 depletion specifically eliminated melanoma subpopulations that resisted high PLX4720 concentrations (Figure 4l, Supplementary Figure 15b). Together, these observations indicate that PDK1 depletion synergizes with targeted BRAF inhibition to kill melanoma cells.

40μM PLX4720 40μM PLX4720 + shPDK1

20 10 0

93.15.2 00.08 93.03 01.14

Figure 4. PDK1 depletion causes melanoma regression and eradicates subpopulations resistant to targeted BRAFV600E inhibition a-e. HDF expressing empty vector or sh-PDK1 were analyzed by immunoblotting (a) and for BrdU incorporation (b), SA-b-Gal activity (c), PDH activity (d) and OCR (e); n=3. f. Cell death induction in melanoma (n=8) and non-transformed cells (n=4) upon PDK1 depletion. g-j. Growth curves (g, i) and representative images (h, j) of tumors upon constitutive (g, h) or DOX-inducible (i, j) PDK1 depletion; n=6. k. Kaplan-Meier survival curve of mice from i. l. Residual cell fraction of control or PDK1-depleted melanoma cells at a high dose of PLX4720. Data are represented as mean ±s.d. (c,e,f,l) or ±s.e.m (g, i).

In conclusion, by metabolic profiling and subsequent functional perturbations of a key metabolic axis we unveil that PDH, a gatekeeper linking glycolysis to oxidative metabolism, acts as a key regulator of OIS (Supplementary Figure 1). The observation that PDH activity is induced during OIS and normalizes upon OIS abrogation highlights this enzyme as potential barrier against malignant transformation. In agreement, high PDK1 expression drives PDH phosphorylation and promotes melanoma growth. Conversely, PDK1 depletion is highly toxic to melanoma cells. Indeed, the regression of established mutant BRAF melanomas, as well as the synergistic toxicity with targeted BRAF inhibition upon PDK1 depletion, raises the possibility that this metabolic kinase represents an attractive combinatorial target for therapeutic intervention of BRAF mutant metastatic melanoma. METHOD SUMMARY The human diploid fibroblast (HDF) cell line TIG3 expressing the ectopic receptor and hTERT (or its derivative expressing sh-p16INK4A), human retinal pigment epithelial cell line RPE1, human prostate cell line PNT1A and all BRAF mutant melanoma cell lines (A0, mel:00.08, 01.14, 04.01, 04.07, 06.04, 07.16, 93.03, 93.15.2, 634, SK-MEL-23, SK-MEL-28) were maintained in DMEM, supplemented with 9% fetal bovine serum (PAA), 2 mM glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin (GIBCO). Melanocytes were maintained as described previously12. Lentiviral and retroviral infections were performed using HEK293T cells and Phoenix cells, respectively, as producers of viral supernatants. For senescence experiments, HDF were infected with shRNA-encoding or protein-coding retro- or lentivirus, selected pharmacologically (puromycin or blasticidin) and subsequently infected with BRAFV600E-encoding or control virus. After selection, cells were seeded for cell proliferation assay, BrdU incorporation assay, SA-b-Galactosidase activity and analyzed. OCR was measured using a XF24 extracellular flux analyzer (Seahorse Bioscience). Exchange rate (uptake or secretion) measurements of key metabolites and stable isotope labelling were performed as described previously28-30. ROS production was measured with CellROX Deep Red Reagent from Invitrogen. GSH/GSSG ratio was determined by GSH/GSSG-Glo Assay from Promega. PDH activity was measured using the DipStick assay kit from MitoSciences (MSP90). Cell death induction was measured by trypan blue exclusion assay. Transcripts and/ or protein levels of the indicated genes were determined by qRT-PCR and immunoblotting or immunohistochemistry, respectively. Details are described in the Supplementary Methods. ACKNOWLEDGMENTS We thank J.-Y. Song for pathological analysis, M. McMahon for providing BrafCA mice, C. Vogel for sharing cell lines, R. van Amerongen for critical reading of the manuscript and all members of the Gottlieb and Peeper labs for their valuable input. This work was supported by Cancer Research UK, Spanish Government-EU-FEDER grants (SAF2011-25726 and ISCIII-

RTICC-RD6/0020/0046) and ICREA-Academia to M.C., Israel Cancer Research Foundation and Israel Science Foundation to T.S., a Vici grant from the Netherlands Organization for Scientific Research (NWO) and a Queen Wilhelmina Award grant from the Dutch Cancer Society (KWF Kankerbestrijding) to D.S.P.

COMPETING FINANTIAL INTERESTS A patent application for combined PDK and MAPK/ERK pathway inhibition in neoplasia has been filed with J.K. and D.S.P. as inventors. REFERENCES 1. 2. 3.

4. 5. 6.

Campisi, J. Suppressing cancer: the importance of being senescent. Science 309, 886–887 (2005). Collado, M. & Serrano, M. Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51–57 (2010). Vredeveld, L. C. W. et al. Abrogation of BRAFV600E-induced senescence by PI3K pathway activation contributes to melanomagenesis. Genes Dev 26, 1055–1069 (2012). Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). Lowe, S. W., Cepero, E. & Evan, G. Intrinsic tumour suppression. Nature 432, 307–315 (2004). Adams, P. D. Healing and hurting: molecular mechanisms, functions, and pathologies of cellular senescence. Mol Cell 36, 2–14 (2009). DeBerardinis, R. J., Sayed, N., Ditsworth, D. & Thompson, C. B. Brick by brick: metabolism and tumor cell growth. Curr Opin Genet Dev 18, 54–61 (2008). Tennant, D. A., Durán, R. V. & Gottlieb, E. Targeting metabolic transformation for

10.

11. 12. 13.

14. 15.

cancer therapy. Nat Rev Cancer 10, 267–277 (2010). Wellen, K. E. & Thompson, C. B. Cellular metabolic stress: considering how cells respond to nutrient excess. Mol Cell 40, 323–332 (2010). Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science 324, 1029–1033 (2009). Campisi, J. Replicative senescence: an old lives’ tale? Cell 84, 497–500 (1996). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 41, 544–552 (2009). Dhomen, N. et al. Oncogenic Braf induces melanocyte senescence and melanoma in mice. Cancer Cell 15, 294–303 (2009).

Chapter 4

CONTRIBUTIONS J.K., E.G. and D.S.P. conceived the project, analyzed the data and wrote the manuscript. J.K. performed all in vitro experiments and carried out the in vivo experiments together with K.M.. J.K., K.M., and B.C. performed metabolic experiments. L.Z. and G.M. performed LC-MS analyses. S.B. and E.V. provided low passage melanoma cell lines. V.S., M.C. and T.S. helped with metabolic analyses. All authors discussed the results and commented on the manuscript.

16. 17.

18. 19.

20.

21.

22. 23. 24.

25.

Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417, 949–954 (2002). Wieland, O. H. The mammalian pyruvate dehydrogenase complex: structure and regulation. Rev. Physiol. Biochem. Pharmacol. 96, 123–170 (1983). Patel, M. S. & Roche, T. E. Molecular biology and biochemistry of pyruvate dehydrogenase complexes. FASEB J 4, 3224–3233 (1990). Kolobova, E., Tuganova, A., Boulatnikov, I. & Popov, K. M. Regulation of pyruvate dehydrogenase activity through phosphorylation at multiple sites. Biochem J 358, 69–77 (2001). Roche, T. E. et al. Distinct regulatory properties of pyruvate dehydrogenase kinase and phosphatase isoforms. Prog. Nucleic Acid Res. Mol. Biol. 70, 33–75 (2001). Holness, M. J. & Sugden, M. C. Regulation of pyruvate dehydrogenase complex activity by reversible phosphorylation. Biochem. Soc. Trans. 31, 1143–1151 (2003). Patel, M. S. & Korotchkina, L. G. Regulation of the pyruvate dehydrogenase complex. Biochem. Soc. Trans. 34, 217–222 (2006). Lemons, J. M. S. et al. Quiescent fibroblasts exhibit high metabolic activity. PLoS Biol. 8, e1000514 (2010). Fantin, V. R., St-Pierre, J. & Leder, P. Attenuation of LDH-A expression uncovers a link between glycolysis, mitochondrial physiology, and tumor maintenance. Cancer Cell 9, 425–434 (2006). Mooi, W. J. & Peeper, D. S. Oncogeneinduced cell senescence--halting on the road

26.

27. 28.

29. 30.

31.

32.

33.

to cancer. N Engl J Med 355, 1037–1046 (2006). Tsai, J. et al. Discovery of a selective inhibitor of oncogenic B-Raf kinase with potent antimelanoma activity. Proc Natl Acad Sci USA 105, 3041–3046 (2008). Flaherty, K. T., Yasothan, U. & Kirkpatrick, P. Vemurafenib. Nature reviews Drug discovery 10, 811–812 (2011). Frezza, C. et al. Haem oxygenase is synthetically lethal with the tumour suppressor fumarate hydratase. Nature 477, 225–228 (2011). Frezza, C. et al. Metabolic profiling of hypoxic cells revealed a catabolic signature required for cell survival. PLoS ONE 6, e24411 (2011). Chaneton, B. et al. Serine is a natural ligand and allosteric activator of pyruvate kinase M2. Nature (2012). doi:10.1038/ nature11540 Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88, 593–602 (1997). Sviderskaya, E. V. et al. Complementation of hypopigmentation in p-mutant (pinkeyed dilution) mouse melanocytes by normal human P cDNA, and defective complementation by OCA2 mutant sequences. J Invest Dermatol 108, 30–34 (1997). Dankort, D. et al. A new mouse model to explore the initiation, progression, and therapy of BRAFV600E-induced lung tumors. Genes Dev 21, 379–384 (2007).

SUPPLEMENTARY METHODS Cell proliferation assay Cells were seeded into a six-well plate (at different densities: 2 × 105, 4 × 105 or 6 × 105 cells, resp.) and selected pharmacologically. Fixation and staining with crystal violet was performed 9 to 13 days after the BRAFV600E-encoding or control virus infection, or 6 to 9 days after shPDK1-encoding or control virus infection. Images of cell proliferation assays reflect representative results of at least three independent experiments. BrdU incorporation assay BrdU labelling was carried out for 3 h followed by fixation. Incorporated BrdU was detected by immunostaining as described in reference 31 and FACS analysis. Results are represented as mean with s.d. of at least three independent experiments.

Chapter 4

Analysis of SA-β-Galactosidase activity Senescence-associated (SA) β-Galactosidase was stained using the ‘Senescence Associated β-Galactosidase Staining Kit’ from Cell Signaling at pH 6 according to the manufacturer’s protocol. Images reflect representative results of at least three independent experiments. Trypan blue exclusion assay Cells were brought into suspension using trypsin, centrifuged and re-suspended in a small volume of culture medium. Trypan blue (Sigma) was added to the cell suspension (dilution factor=2) and stained cells were counted as dead. The number of dead cells was quantitated, and the values were expressed as the fold change over control. Results are represented as mean with s.d. of at least three independent experiments. Measurement of redox stress ROS production was measured with CellROX Deep Red Reagent from Invitrogen. Cells were incubated at 37oC for 30 min in DMEM medium supplemented with 9% fetal bovine serum (PAA) and containing 2.5 mM CellROX Deep Red Reagent. Afterwards cells were washed twice in PBS, treated with trypsin and re-suspended in PBS supplemented with 50% FCS. Fluorescence was immediately measured using FACS analysis. GSH/GSSG ratio was determined by GSH/GSSG-Glo Assay from Promega according to the manufacturer’s protocol. Results are represented as mean with s.d. of at least three independent experiments. Plasmids pMSCV-blast-BRAFV600E and pMSCV-blast were previously described13. For the overexpression of PDK1, the ORF of PDK1 derived from cDNA of a normal human skin sample was PCRamplified and cloned into pLZRS-IRES-puro or FG12eGFP. Empty pLZRS-IRES-puro or FG12eGFP was used as control. shRNAs Retroviral knockdown constructs were described previously12,13. Lentiviral constitutive knockdown constructs were purchased from Sigma-Aldrich in pLKO.1 backbone or cloned (shp53) into KH1eGFP backbone (a kind gift from M. Soengas): shPDP2#1–TRCN0000036739 – 5′- CCTTGAAAGCAGAGTCCCAAA-3′ shPDP2#4–TRCN0000036742–5′- GCTGAAGTGGAGTAAAGAGTT-3′ shPDK1#3–TRCN0000006261–5′- GCTCTGTCAACAGACTCAATA-3′ shPDK1#5–TRCN0000006263–5′- CCAGGGTGTGATTGAATACAA-3′ shp53 mouse–5′- GTACATGTGTAATAGCTCC-3′ As negative controls pRS-puro, pLKO.1-puro or KH1eGFP without insert were used. Lentiviral inducible knockdown plasmid pLKO.1-Tet-On was purchased from Addgene. shRNA targeting PDK1 were re-cloned from constitutive shPDK1#3 knockdown construct. As negative control pLKO.1-Tet-On with shRNA targeting luciferase was used. shluc–5′- CGCTGAGTACTTCGAAATGTC-3′

Quantitative real-time PCR Total RNA was DNase-treated with RQ1 RNase-Free DNase (Promega). Reverse transcription was performed with SuperScript II First Strand Kit (Invitrogen). qRT–PCR was performed with the SYBR Green PCR Master Mix (Applied Biosystems) on an ABI PRISM 7700 Sequence detection System. Primer sets used were as follows: IL6, IL8, C/EBPβ, and RPL13 (standard) primer sequence were described previously13. PDK1: 5′- CCAAGACCTCGTGTTGAGACC-3′ and 5′- AATACAGCTTCAGGTCTCCTTGG-3′; PDK2: 5′- GAGCCTCCTGGACATCATGG-3′ and 5′- TACTCAAGCACGCCTTGTGC-3′ PDK3: 5′- ACTGTATTCCATGGAAGGAGTGG-3′ and 5′- CTCCAATCATCGGCTTCAGG-3′ PDK4: 5′- AACTGTGATGTGGTAGCAGTGG-3′ and 5′- GATGTGAATTGGTTGGTCTGG-3′ PDP2: 5′- ACCACCTCCGTGTCTATTGG-3′ and 5′- CCAGCGAGATGTCAGAATCC-3′ NQO1: 5′- CAGCTCACCGAGAGCCTAGT-3′ and 5′- GAGTGAGCCAGTACGATCAGTG-3′ GCLC: 5′- ATGCCATGGGATTTGGAAT-3′ and 5′- AGATATACTGCAGGCTTGGAATG-3′ GSTa4: 5′- AGTTGTACAAGTTGCAGGATGG-3′ and 5′- CAATTTCAACCATGGGCACT-3′ GSTm4: 5′- TCATCTCCCGCTTTGAGG-3′ and 5′- CAGACAGCCACCCTTGTGTA-3′ HMOX1:5′- GGGTGATAGAAGAGGCCAAGA-3′ and 5′- AGCTCCTGCAACTCCTCAAA -3′ SOD1: 5′- TCATCAATTTCGAGCAGAAGG -3′ and 5′- CAGGCCTTCAGTCAGTCCTTT -3′ SOD2:5′- CTGGACAAACCTCAGCCCTA -3′ and 5′- TGATGGCTTCCAGCAACTC -3′ Except for C/EBPβ, all primer pairs span exon-exon borders. RPL13 was used as control. For analysis, the ΔCT method was applied. Data are represented as mean ±s.d. of three or more independent experiments. Antibodies Antibodies used for immunoblotting/immunohistochemistry were β-actin (AC-74; A5316; Sigma), BRAF (sc-5284; Santa Cruz), Hsp90 (4874; Cell Signaling), LDHA (2012; Cell Signaling), MEK1/2 (L38C12; 4694; Cell Signaling), phospho-MEK1/2 [Ser217/221] ((41G9); 9154; Cell Signaling), PCNA ((PC10); sc-56; Santa Cruz), PDHE1a ((9H9AF5); 459400; Invitrogen), phospho-PDHE1a [Ser293] (AP1062; Calbiochem), PDK1 (KAP-PK112; Stressgene), PDP2 (HPA01995; Sigma), p16INK4A ((JC3); sc-56330; Santa Cruz), p27Kip1 (610241; BD Transduction Laboratories); p15INK4B (sc-612; Santa Cruz). Immunohistochemistry Formalin-fixed paraffin-embedded tissue samples were stained according to common procedures. For antigen retrieval samples were incubated in 20mg/ml proteinase K (Z0622, Sigma for PDK1), in citrate buffer (Biogenex, for PDHE1a) or in Tris-EDTA pH 9.0 (for phosphor-PDHE1a [Ser293]). Sections were counterstained with hematoxylin. Drug treatment For dose-response curves melanoma cells were treated with the indicated concentrations of PLX4720 (Selleck) for 3 days. Cell viability was determined with the Cell Titer Blue assay

Chapter 4

(Promega) and fluorescence was measured with a TECAN infinite scanner. Results represent at least three independent experiments. In vivo assays All mouse experiments were performed according to a protocol approved by the Institutional Animal Experiment Ethics Committee. Experiments were repeated at least twice. Murine melanocyte derivation, culture and tumor growth in vivo Melanoytes were derived from neonatal skin of Tyr::CreER;BrafCA mice14 as described previously32. Primary melanocyte cultures were prepared on a mitomycin-treated XB2 (immortal murine keratinocytes) feeder cell layer for one passage only. Cells were grown in RPMI medium supplemented with 5% FCS, 200nM 12-O-tetradecanoyl phorbol 13-acetate, 200pM cholera toxin, 10ng/ml recombinant stem cell factor (SCF; R&D Systems) and 100nM endothelin 3 (Bachem) at 37°C and under low oxygen conditions (5% CO2 and 3% O2). Primary melanocytes were transduced with retro- or lentivirus in the presence of 2µg/ml Polybrene overnight. The transduction with lentivirus delivering shRNA targeting p53 and the PDK1 expression vector were done on consecutive nights. The transduction of virus allowing PDK1 expression was repeated two to four times. In addition melanocytes were treated with 0.2µM 4-hydroxytamoxifen for at least 9 days to induce the expression of BrafV600E by switching from the conditional to the mutated allele. One week after the last transduction 1-1.5 x 106 melanocytes (dependent on the experiment) were subcutaneously injected with 50% basement membrane MatrigelTM (growth factors reduced) into both flanks of NSG mice (NOD scid IL2 receptor gamma chain knockout mice). Tumor volume was determined by measurement of two dimensions and calculation with the following formula: V=4/3*p*a2*b (a is the shorter and b is the longer dimension). BRAF allele rearrangements in tumors were detected by PCR as previously described33. Tumors were analysed by immunohistochemistry and immunoblot. Xenograft experiments NOD scid mice were subcutaneously injected with 0.5 x 106 cells into both flanks. For doxycycline (DOX)-inducible shRNA xenograft experiments mice were exposed to 2mg/ml DOX administered in 5% sucrose-containing drinking water either directly after injection or when tumors reached a volume of 100 mm3. Mice were inspected twice a week and euthanized by CO2 when tumors reached the volume of 500 mm3 (1000mm3 per mouse). Tumor volume was determined by measurement of two dimensions and calculation with the following formula: V=4/3*p*a2*b (a is the shorter and b is the longer dimension). Metabolic profiling Measurement of OCR Basal OCR was measured using the XF24 extracellular flux analyzer (Seahorse Bioscience). At the end of the experiment 1 μmol l−1 antimycin A was added to measure mitochondriaindependent oxygen consumption. Each cycle of measurement consisted of 3 min mixing,

3 min waiting and 4 min measuring. OCR was normalized to the cell number calculated at the end of the experiments. To obtain the mitochondrial-dependent OCR, only the antimycinsensitive respiration was used. Homogeneous plating of the cells and cell count were assessed by fixing the cells with trichloroacetic acid 10% for 1 h at 4 °C and then staining the fixed cells with a 0.47% solution of sulphorhodamine B (SRB) (Sigma). Measurement of metabolites by LC-MS 2 × 106 HDF were plated onto 10cm dishes and cultured in standard medium for 24h. For stable isotope labeling analysis, the medium was replaced with 4.5 mM [U-13C]-glucose (Cambridge Isotope, UK). After incubation for the indicated time cells and media were collected. For extracellular metabolite analysis, 200 µl of growth medium from cell culture were added to 600 µl of acetonitrile for deproteinization. Samples were vortexed for 10 minutes and centrifuged for 10 minutes at 16000 g at 4oC. The supernatant was stored for subsequent LC-MS analysis. For intracellular metabolite analysis, cells were lysed with a solution composed of 50% methanol and 30% acetonitrile in water in dry ice methanol (-80oC) and quickly scraped from the plate. The insoluble material was immediately pelleted in a cooled centrifuge (4oC) for 10 minutes at 16000 g and the supernatant was collected for subsequent LC-MS analysis. For [U-13C]-pyruvate labeling analysis, samples were processed as for the 13C-labelled glucose labeling analysis, except that the medium was not replaced, but spiked with 0.11 mg/ml [13C-3]-sodium pyruvate (Cambridge Isotope, UK). LC-MS analysis was carried out as described in reference 29. Mass-spec data was analyzed by LCquan™ (Thermo Scientific, UK) and quantifications of intracellular and extracellular metabolites were performed by the standard-dilution method as described in reference 30. Measurement of PDH activity PDH activity in cell lysates was measured using the DipStick assay kit from MitoSciences (MSP90). Cells were lysed in the 10x sample buffer provided by the manufacturer, followed by centrifugation and measurement of the protein concentration with the BioRad Protein Assay. 140 mg of protein lysate was loaded and PDH activity was measured according to the manufacturer’s protocol. Statistical analysis Statistical analyses of metabolite exchange rates were done with t-tests. Analysis of all other data was done with a non-parametric two-tailed Mann-Whitney test with a 95% confidence interval (Prism; GraphPad Software, Inc.). P values of less than 0.05 were considered significant. *0.01 < P < 0.05; **0.001 < P < 0.01; ***P < 0.001.

SUPPLEMENTARY FIGURES

pyruvate

PDK1

malate BRAF

Supplementary Figure 1. Model for concerted activation of PDH necessary to drive OIS In response to an oncogenic trigger (BRAFV600E), non-transformed cells upregulate PDP2 and downregulate PDK1, causing the concerted activation of PDH. PDH is the gatekeeper enzyme linking glycolysis and the TCA cycle. Its activation during OIS promotes the utilization of glucosederived pyruvate in the TCA cycle, increasing cellular respiration and representing an essential element of the OIS program. Enforced normalization of PDP2 or PDK1 expression levels abrogates OIS, causing continued cell proliferation.

800

150

600

glucose

cycling

200

100

50 0

cycling

OIS

pyruvate

150 100

lactate

150

100

0 0 15 m 30 m 60 m 18 0m 12 h 24 h

cycling

0 15 m 30 m 60 m 18 0m 12 h 24 h

OIS

malate

100

15 m 30 m 60 m 18 0m 12 h 24 h

cycling

***

0 15 m 30 m 60 m 18 0m 12 h 24 h

S OI

ng cli cy

OIS

citrate

150

TCA cycle

100 50 0

0 15 m 30 m 60 m 18 0m 12 h 24 h

150

cycling

15 m 30 m 60 m 18 0m 12 h 24 h

cycling

0 15 m 30 m 60 m 18 0m 12 h 24 h

OIS

aKG

150 100

0 15 m 30 m 60 m 18 0m 12 h 24 h

cycling

150

100

glutamine

cycling

0 15 m 30 m 60 m 18 0m 12 h 24 h

OIS

0 15 m 30 m 60 m 18 0m 12 h 24 h

cycling

0 15 m 30 m 60 m 18 0m 12 h 24 h

S OI

cli ng cy

OIS

glutamate

150

0 15 m 30 m 60 m 18 0m 12 h 24 h

15 m 30 m 60 m 18 0m 12 h 24 h

glutamine uptake [nmol/(106 cells x h)]

0 15 m 30 m 60 m 18 0m 12 h 24 h

cycling

0 15 m 30 m 60 m 18 0m 12 h 24 h

ng cy

time

100

alanine

150

cli alanine secretion [nmol/(106 cells x h)]

G3P

150

OIS

U-12C 13 C1 13 C2 13 C3 13 C4 13 C5 13 C6

Chapter 4

0 15 m 30 m 60 m 18 0m 12 h 24 h

400

% Intracellular labeling

100

0 15 m 30 m 60 m 18 0m 12 h 24 h

lactate secretion [nmol/(106 cells x h)]

αKG

V600E

TCA OFF OIS bypass 1000

TCA

PDH

lactate

citrate

PDP2

0 15 m 30 m 60 m 18 0m 12 h 24 h

GLYCOLYSIS

glucose

OXIDATIVE PHOSPHORYLATION

TCA ON OIS induction

BRAFV600E

OIS

Supplementary Figure 2. Metabolic profiling of OIS vs cycling cells a-c. Analysis of lactate secretion (a), alanine secretion (b) and glutamine uptake (c) in cycling and OIS HDF (upon expression of BRAFV600E); n=6. d. Comparison of glucose metabolism in cycling and OIS HDF (upon expression of BRAFV600E). Cells were incubated with uniformly labeled [U-13C6]-glucose, extracted and analyzed at the indicated time points between 15 minutes and 24 hours. The distributions of the different isotopomers (U-12C, 13C1-6) of each metabolite are presented over time. Citrate, aKG and malate are indicated as part of the TCA cycle in the mitochondria. Though the exchange of these TCA cycle metabolites with the cytosol is possible, their labeling from glucose-derived carbon can take place only in the mitochondria; n = 3. All results are presented as mean ± s.d. G3P – glyceraldehyde 3-phosphate; aKG – alpha-ketoglutarate; TCA – tricarboxylic acid cycle.

relative amount of transcript

HMOX1

BRAFV600E

sh P

Pβ

sh P

EB C

or ct

in g

cy cl

***

#4 P2

#1 P2

***

SOD2

1.5 1.0 0.5

#4 P2

#1 P2

Pβ

or ct

C sh

Pβ

EB C

2.0

SOD1

relative amount of transcript

P2 #4

r Pβ

g in

***

BRAFV600E

GSTm4

BRAFV600E

***

BRAFV600E 2

***

BRAFV600E

Pβ

EB C

or ct

***

GCLC

***

BRAFV600E

relative amount of transcript

Pβ

g in cl

relative amount of transcript

***

ve ct o

cy cl in g

GSTa4

BRAFV600E

***

Pβ

or ct

BRAFV600E

10 4

NQO1

relative amount of transcript

GSH/GSSG ratio

100

relative amount of transcript

10 3

relative amount of transcript

10 2

Fluorescence intensity

***

200

20 0

10 1

***

10 0

***

300

Counts 60 80 100 120

sh PD

vector OIS ( BRAFV600E) vector + CellROX vector + menadione + CellROX OIS ( BRAFV600E) + CellROX OIS ( BRAFV600E) + menadione + CellROX

Supplementary Figure 3. OIS is accompanied by increased redox stress a. ROS production in cycling and OIS HDF (upon expression of BRAFV600E) measured by CellROX. Menadione-treated cells serve as a positive control. b. GSH/GSSG ratio in cycling, OIS and OIS escape cells (either upon PDK1 overexpression or PDP2 depletion); n=3. c-i. Regulation of transcripts of ROS-responsive genes in cycling, OIS and OIS escape cells (either upon PDK1 overexpression, or depletion of either PDP2 or C/EBPb), as determined by qRT-PCR; n=3. All data are represented as mean ± s.d.

cl in g

SV 40

SV 4 by (OI 0 s pa S t ss )

ve (O cto IS r )

st SV 40

c (v ycli ec ng to r)

BRAFV600E

SV 40 ve ct or

BRAF pPDHE1α PDHE1α

d PDK1 PDK2 PDK3 PDK4

OIS OIS bypass

-1 PDP1 PDP2

2.5

PDK1 PDK2 PDK3 PDK4

2 1.5 1 0.5 0

-2

cycling

SV40

vector

SV40

80 60 40 20

alanine

120 100

100 80

time (min)

***

200

100

cli

18 0

RAS

cycling

18 0

RAS

18 0

cycling

300

RAS

G3P

120

18 0

cycling

18 0

U-12C C1 13 C2 13 C3 13 C4 13 C5 13 C6

100

se RA ne S G sc 12V en ce

glucose

120

OCR [nmol/(106 cells x h)]

pyruvate

120

Chapter 4

BRAFV600E

% Intracellular labeling

log ratio (M)

relative amount of transcript

β-actin

Supplementary Figure 4. PDH regulation during senescence abrogation a. Cell proliferation assay of HDF expressing empty vector or SV40 st in the presence or absence of oncogenic BRAFV600E. b. Samples from a were analyzed by immunoblotting as indicated. c. Log2 ratio values for PDP and PDK family member transcripts in a gene expression data set comparing cycling, OIS and OIS escape HDF (ArrayExpress accession number E-NCMF-121). d. Regulation of PDK transcripts in cells from a, as determined by qRT-PCR; n=3. Data are represented as mean ± s.d.

80 60

RAS

citrate

120

PDHE1α

100

PDK1

80 60

18 0

TCA cycle

PDP2

cycling

RAS

18 0

RAS

18 0

in pPDHE1α

18 0

cycling

18 0

100

cycling

120

lactate

se RA ne S G sc 12V en ce

100

RAS

β-actin aKG

120

glutamate

120

100

18 0

cycling

18 0

RAS

18 0

cycling

RAS

Supplementary Figure 5. Analysis of glucose metabolism and PDK1-PDP2-PDH axis regulation in RASG12Vinduced senescence a. Comparison of the glucose metabolism in cycling and HDF rendered senescent by the expression of RASG12V, analyzed after 0, 15 and 180 minutes of labeling with [U-13C]-glucose. Results show the distributions of the different isotopomers (U-12C, 13C1-6) of each metabolite presented over time. b. OCR of cells from a. c. Immunoblotting analysis of samples from a. All data are represented as mean ± s.d. ; n=3. G3P glyceraldehyde 3-phosphate; aKG - alpha-ketoglutarate; TCA - tricarboxylic acid cycle.

b glucose

120 100

80 60 40

18 0

cycling

18 0

U-12C 13 C1 13 C2 13 C3 13 C4 13 C5 13 C6

quiescent

time (min)

alanine

120 100

100

60 40

18 0

cycling

18 0

quiescent

cycling

18 0

in con hi t bi ac tio t n

cy cl in g

18 0

G3P

120

18 0

% BrdU-positive cells

% Intracellular labeling

quiescent

pyruvate

120 100

Supplementary Figure 6. Analysis of glucose metabolism in quiescence a. BrdU incorporation of cycling and quiescent HDF. b. Comparison of the glucose metabolism in samples described in a, analyzed after 0, 15 and 180 minutes of labeling with [U-13C]-glucose. Results show the distributions of the different isotopomers 13 C1-6) of each (U-12C, metabolite presented over time. Data are represented as mean ± s.d. ; n=3. G3P – glyceraldehyde 3-phosphate; aKG – alpha-ketoglutarate; TCA – tricarboxylic acid cycle.

80 60 40 20

cycling

quiescent

80 60

citrate

120

100

80 60

18 0

TCA cycle

aKG

120

quiescent

glutamate

120

100

quiescent

% Intracellular labeling

citrate

120 100

100

[ C3] pyruvate 13

60 40

60 40 20

cycling

quiescent

cycling

pyruvate

120 100

PDH

60 40

cycling

15 18 0

20 0

cycling

15 18 0

PDK1/ BRAFV600E shPDP2#4 /BRAFV600E BRAFV600E

15 18 0

cycling

100 80

cis-aconitate

120

15 18 0

TCA cycle

15 18 0

100

LDH

120

U-12C C1 13 C2 13 C3

time (min)

PDK1/ BRAFV600E shPDP2#4 /BRAFV600E BRAFV600E

lactate

15 18 0

ALT

15 18 0

shPDP2#4 PDK1/ /BRAFV600E BRAFV600E

15 18 0

BRAF

15 18 0

cycling

V600E

15 18 0

18 0

cycling

18 0

quiescent

18 0

cycling

18 0

alanine

120

18 0

0 100

lactate

120

PDK1/ BRAFV600E shPDP2#4 /BRAFV600E BRAFV600E

Supplementary Figure 7. Analysis of pyruvate metabolism in OIS HDF expressing control vector, PDK1 or sh-PDP2 in the presence or absence of oncogenic BRAFV600E were incubated with fully labeled [13C3]-pyruvate and the abundance of the indicated metabolites was analyzed upon 15 and 180 minutes of labeling. The distributions of the different isotopomers (U-12C, 13C1-6) of each metabolite are presented over time. All data are represented as mean ± s.d. ; n=3. ALT – alanine aminotransferase; LDH – lactate dehydrogenase; PDH – pyruvate dehydrogenase; TCA – tricarboxylic acid cycle.

time (min)

G3P

40 20

BRAFV600E PDK1/ BRAFV600E

TCA cycle

BRAFV600E PDK1/ BRAFV600E

TCA cycle citrate

120

PDH

Chapter 4

cycling

PDH

BRAFV600E shPDP2#4 /BRAFV600E

0 0

citrate

120

100

cycling

BRAFV600E shPDP2#4 /BRAFV600E

cycling

BRAFV600E shPDP2#4 /BRAFV600E

100

LDH

pyruvate

120

ALT

100

BRAFV600E PDK1/ BRAFV600E

cycling

BRAFV600E PDK1/ BRAFV600E

lactate

120

LDH

cycling

pyruvate

100

PK 120

100

cycling

0 0

ALT

lactate

120

BRAFV600E shPDP2#4 /BRAFV600E

cycling

shPDP2#4 /BRAFV600E

100

G3P

120

100

alanine

120

BRAF

BRAFV600E PDK1/ BRAFV600E

cycling

100

cycling

120

100

V600E

time (min)

alanine

120

BRAFV600E shPDP2#4 /BRAFV600E

cycling

glucose

100

120

U-12C 13 C1 13 C2 13 C3 13 C4 13 C5 13 C6

% Intracellular labeling

glucose

120

U-12C 13 C1 13 C2 13 C3 13 C4 13 C5 13 C6

% Intracellular labeling

BRAFV600E PDK1/ BRAFV600E

vector

LDHA

β-actin

$!" #"10

BRAF

LDHA

'""

()*+),-$"

shLDHA

3.89 ± 1.02 E 00

ct or

cl in g

p27

#!"

shLDHA

4.54 ± 1.36

sh-C/EBPβ

$!"

p15

vector

PCNA

2.54 ± 0.52

$#"

PDHE1α

%#" %!"

vector

pPDHE1α

%BrdU-positive cells

100

shLDHA

150

vector

OCR [nmol/(106 cells x h)]

Supplementary Figure 8. Analysis of glucose metabolism in OIS Glucose metabolism analysis of HDF expressing vector control or either sh-PDP2 (a) or PDK1 (b) in the presence or absence of oncogenic BRAFV600E, upon 45 min (a) or 15 min of labeling with uniformly labeled [U-13C]-glucose. The results show distributions of the different isotopomers (U-12C, 13C1-6) of each individual metabolite. All data are represented as mean ±s.d. ; n=3. G3P - glyceraldehyde 3-phosphate; PK - pyruvate kinase; ALT - alanine aminotransferase; LDH - lactate dehydrohenase; PDH - pyruvate dehydrgenase; TCA tricarboxylic acid cycle. shLDHA b c e a shLDHA

β-actin

Supplementary Figure 9. LDHA regulates neither OCR nor senescence in HDF a. OCR of HDF expressing empty vector or sh-LDHA. b-e. Samples from a were analyzed by immunoblotting (b), and for cell proliferation (c), BrdU incorporation (d) and SA-b-Gal activity (e); f. Immunoblotting analysis of cycling, undergoing OIS or OIS escape (upon C/EBPb depletion) HDF. All data are represented as mean ± s.d with n=3. ()*+),-."

PDK1 #3

vector #3

PDK1 #2

specimens vector #2

***

vector #1 PDK1

0.6

Supplementary Figure 10. PDK1 acts tumorigenically a. Weight of tumors formed by sh-p53/BRAFV600E melanocytes expressing empty vector or PDK1 (n=8 tumors). Data are represented as mean ± s.e.m. b. Immunoblotting analysis of samples described in a.

pPDHE1α PDHE1α

0.0

K1 PD

vector

16 14

shPDK1

8.41±3.87

85.84±2.56

80.04±5.45

p27

β-actin

4 2 sh#5

vector

shPDK1

β-actin

shPDK1

800 600

***

*** ***

93.15.2 #3 #5

00.08 #3 #5

***

400 200 0

shPDK1 A0

shPDK1

93.15.2

TIG

00.08

PDHE1α

OCR [nmol/ (106 cells x h)]

FM186ae

PDK1

nontransformed

SK23

pPDHE1 α

mut BRAF melanoma

93.15.2

shPDK1

0.5

93.15.2

vector

00.08

shPDK1

vector

shPDK1

sh#3

vector

mut BRAF

sh#5

p15

non-transformed

sh#3

PDK1

shPDK1 vector

sh#5

Normalized BrdU incorporation

% BrdU-positive cells

sh#3

hsp90

vector

Weight (g)

1.2

PDK1 #1

shPDK1 00.08

Supplementary Figure 11. PDK1 depletion induces senescence in non-transformed cells and cell death in BRAF mutant melanoma a. BrdU analysis of human melanocytes expressing empty vector or shPDK1; n=3. b-c Samples from a were analyzed by immunoblotting (b) and for SA-b-Galactosidase activity (c); n=3. d. Cell proliferation assay of mutant BRAF melanoma cell lines and non-transformed cells expressing either empty vector (control) or shRNA targeting PDK1. e. BrdU incorporation assay of samples described in d. Data are normalized to the vector-expressing cells; n=3. f-h. Melanoma cell lines expressing either empty vector (control) or shRNA targeting PDK1 were analyzed by immunoblotting (f), for PDH activity (g) and for OCR (h); n=3. All data are represented as mean ± s.d.

PDK1 shPDK1#3

04.01

04.07

SK23

06.04

Hsp90 PDK1

shPDK1#3

RPE1

PNT1A

TIG

FM186ae

mut BRAF non-transformed

0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

Hsp90 PDK1

RPE1 TIG PNT1a FM186ae

Hsp90

1 0.9

SK23

06.04

04.07

00.08

93.15.2

00.08

93.15.2

SK28

shPDK1#3

Relative amount of PDK1 protein normalized to vector

shPDK1#3

Supplementary Figure 12. The extent of PDK1 knockdown between non-transformed and mutant BRAF melanoma cells a. Immunoblotting analysis of melanoma cell lines and non-transformed cells expressing either empty vector (control) or shRNA targeting PDK1. b. Quantification of the blot described in a.

shPDK1 #3

vector

tumors

PDK1

cells shPDK1 #3

vector

shPDK1 #3

PDK1

pPDHE1α

β-actin

93.15.2 cell line

β-actin

A0 cell line

10 20 days after injection

shPDK1 #3

shPDK1 #5

PDK1

0.2 0.1

vector

pPDHE1α

0.3

vector

100

Weight (g)

Tumor volume (mm3)

shPDK1 #5

200

93.15.2 cell line

0.4

shPDK1 #3

93.15.2 cell line vector

c 300

e H&E

PDHE1α

shPDK1

PDHE1α

vector

A0 cell line

Supplementary Figure 13. PDK1 depletion prevents melanoma outgrowth a. Immunoblotting analysis of BRAF mutant melanoma cell line A0 expressing either empty vector (control) or shRNA targeting PDK1. The cell batch was analyzed at the time that the remainder of the cells was inoculated into mice. b. Immunoblotting analysis of BRAF mutant melanoma cell line 93.15.2, expressing either empty vector (control) or shRNA targeting PDK1, as well as of the tumors these cells generated in mice. Cells were collected on the day of the injection into mice. c-d. Growth curve (c) and weight (d) of tumors formed by cells described in c; (n=4 tumors). e. Immunohistochemical staining of tumors formed from cells described in a. All data are represented as mean ± s.e.m. (c) or ± s.d. (d).

Chapter 4

non-transformed

mut BRAF melanoma cell lines

pLKOTet shluc

DOX

pLKOTet shPDK1 -

pLKOTet shluc

DOX

PDK1

pLKOTet shPDK1 +

pPDHE1α

100

% cell death

PDHE1α

Tumor volume (mm3)

shluc

40 20

shPDK1

600

400

200

20 40 days after injection

pLKOTet pLKOTet shluc shPDK1

600

Tumor volume (mm3)

800

DOX

β-actin

shluc

400

shPDK1 200 DOX 0

100

days after injection

Supplementary Figure 14. PDK1 is crucial for melanoma viability a-c. Immunoblotting analysis (a), cell proliferation assay (b) and cell death induction (c) of mutant BRAF melanoma cell line A0, expressing inducible shRNA targeting luciferase (control) or targeting PDK1. Cells were kept either under non-induced (DOX -) or induced (DOX +) condition. d. Growth curves of tumors formed by BRAF mutant melanoma cell line A0 carrying inducible sh-luciferase (sh-luc) or sh-PDK1. Cells were injected into immunocompromised mice under non-induced conditions (n=6 tumors). e. Growth curves of tumors formed by BRAF mutant melanoma cell line A0 expressing sh-luc or sh-PDK1. Cells were injected into immunocompromised mice under induced conditions (n=6 tumors). All data are represented as mean ± s.e.m. a

PLX4720 SENSITIVE 04.01

viability

100

SK28

100

-4

-2

-4

-2

07.16

100 50

634

100

-4

-2

-4

-2

vector shPDK1

PLX4720 dilution (log) PLX4720 RESISTANT

viability

93.15.2

100

-2

-4

93.03

100

-2

01.14 vector shPDK1

100 50

-4

00.08

100

-4

-2

-4

-2

PLX4720 dilution (log)

Supplementary Figure 15. PDK1 depletion sensitizes BRAF mutant melanoma cells to PLX4720 a-b. PLX4720 dose-response curves of BRAF mutant melanoma cells, which are sensitive (a) or partially resistant (b) to the inhibitor. Cells expressed either empty vector (control) or sh-PDK1; n=3.

CHAPTER 5 NEAR-GENOMEWIDE RNAi SCREENING FOR REGULATORS OF BRAFV600E-INDUCED SENESCENCE IDENTIFIES RASEF, A GENE EPIGENETICALLY SILENCED IN MELANOMA

Pigment Cell Melanoma Res (2014) 27, 640–652

NEAR-GENOMEWIDE RNAi SCREENING FOR REGULATORS OF BRAFV600E-INDUCED SENESCENCE IDENTIFIES RASEF, A GENE EPIGENETICALLY SILENCED IN MELANOMA Joanna Kaplon1, Cornelia Hömig-Hölzel1,3*, Linda Gao2*, Katrin Meissl1,4, Els M.E. Verdegaal5, Sjoerd H. van der Burg5, Remco van Doorn2 and Daniel S. Peeper1 Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam,

The Netherlands. 2Department of Dermatology, Leiden University Medical Center, Albinusdreef 2, 2300 RC Leiden, The Netherlands. 3 Current address: Department of Clinical Chemistry and Clinical Pharmacology, University of Bonn, Siegmund-Freud-Str., 2553105 Bonn, Germany. 4 Current address: Institute of Animal Breeding and Genetics, University of Veterinary Medicine, Veterinärplatz 1, A-1210 Vienna, Austria. 5 Clinical Oncology, Leiden University Medical Center, Albinusdreef 2, 2333 ZA Leiden, The Netherlands. *These authors contributed equally to this work.

SUMMARY The activation of oncogenes in primary cells blocks proliferation by inducing oncogeneinduced senescence (OIS), a highly potent in vivo tumor-suppressing program. A prime example is mutant BRAF, which drives OIS in melanocytic nevi. Progression to melanoma occurs only in the context of additional alteration(s) like the suppression of PTEN, which abrogates OIS. Here, we performed a near-genomewide short hairpin (sh)RNA screen for novel OIS regulators and identified by next generation sequencing and functional validation seven genes. While all but one were upregulated in OIS, their depletion abrogated BRAFV600E-induced arrest. With genome-wide DNA methylation analysis we found one of these genes, RASEF, to be hypermethylated in primary cutaneous melanomas compared to nevi. Bypass of OIS by depletion of RASEF was associated with suppression of several senescence biomarkers including senescence-associated (SA)-β-galactosidase activity, interleukins and tumor suppressor p15INK4B. Restoration of RASEF expression inhibited proliferation. These results illustrate the power of shRNA OIS bypass screens and identify a potential novel melanoma suppressor gene. SIGNIFICANCE This study describes an unbiased screening approach to identify novel melanoma susceptibility genes. We found seven novel genes contributing to oncogene-induced senescence, a critical cellular program preventing malignant transformation. By genomewide promoter methylation analysis we find that one of the genes, RASEF, is hypermethylated in melanomas when compared to senescent nevi. In agreement with a tumor suppressive role of RASEF, its restored expression in BRAFV600E-mutant melanoma acted cytostatically.

Chapter 5

doi: 10.1111/pcmr.12248

These results illustrate the versatility of oncogenomic screening approaches to identify novel candidate tumor suppressors in melanoma. INTRODUCTION Several tumor-suppressive mechanisms have evolved to prevent malignant transformation including self-destructive programs such as apoptosis1 and autophagy2. Work in recent years has uncovered that cells at risk for oncogenic transformation can embark on another strategy: in response to obstinate stress signals they activate signaling networks comprising several potent tumor suppressor proteins. In this way, a system is established enacting a dominant break that actively halts cell proliferation. This type of cell cycle arrest, termed Oncogene-Induced Senescence (OIS), can be maintained even in the context of persistent oncogenic cues3. In recent years a large body of compelling evidence has shown that OIS acts as a pathophysiologic mechanism suppressing cancer in model systems and humans alike. Indeed, senescence biomarkers have been reported for a plethora of precancerous lesions including pulmonary adenomas, prostate intraepithelial neoplasia, and mammary tumors4. Cellular senescence denotes a state of virtually irreversible withdrawal of cells from the proliferative pool. While we have come to realize that its onset can be triggered by a plethora of stress-related signals, it was first recognized in cultured human diploid fibroblasts following explantation and in vitro culturing. With every division telomeres get progressively shorter, causing cells to hit their “Hayflick limit” at the end of their replicative lifespan5. Such replicative senescence has correlates in the context of other stress signals, most notably the activation of oncogenes or loss of tumor suppressor alleles. As the latter occurs independently of telomere malfunction, it is referred to as premature cellular senescence3. Next to long-term cell cycle arrest, which is the central characteristic of senescent cells, OIS is characterized by the activation of tumor suppressor pathways including -but not restricted to- those constituted by p16INK4A-RB and ARF-p536-8. Conversely, cell-cycle progression factors such as cyclins A and B and PCNA are inhibited3. OIS is accompanied also by morphological changes, induction of senescence-associated β-galactosidase activity (SA-β-Gal)9 and chromatin condensation into senescence-associated heterochromatin foci (SAHF)10,11. Another characteristic of cells undergoing OIS is the activation of an inflammationassociated secretory program called Senescence-Messaging Secretome (SMS) as we and others have shown12-14 or Senescence-Associated Secretory Phenotype (SASP), the latter of which is associated with a DNA damage response15,16. The physiologic relevance of the inflammatory phenotype is reflected by the production of a number of interleukins and chemokines in senescent cells in benign murine and human neoplasms17. But secreted senescence-associated factors also limit tissue damage18 and correlate with premature

Chapter 5

aging19. Yet another function of SMS factors is to activate the innate immune system, setting up incipient cancer cells for their own eventual demise20. Further supporting a cell nonautonomous role for senescence, it was recently shown that this program can be brought about also in a paracrine fashion21. Lately, we and others have discovered functional connections between the cellular metabolism and OIS. Aird et al., have demonstrated that stable growth arrest in OIS is established and maintained through suppression of nucleotide metabolism22. Others have highlighted the importance of malic enzyme activity in OIS23. Recently, we identified mitochondrial gatekeeper pyruvate dehydrogenase (PDH) as an enzyme critically contributing to OIS24. In OIS, PDH is activated upon downregulation of its inhibitory kinase PDK1 and simultaneous upregulation of PDH-activating phosphatase PDP2. Normalization of the levels of these enzymes inactivates PDH resulting in abrogation of OIS. Translating these findings to a clinical context, we demonstrated that PDK1 depletion eradicates established melanomas, highlighting this metabolic enzyme as a potentially attractive therapeutic metabolic cancer target. The benign melanocytic nevus was the first human lesion for which evidence was shown in favor of the idea that OIS prevents malignant progression25,26. Three years prior to that, BRAFV600E was identified as a common mutation in melanoma and other cancer types27. Remarkably, in spite of the presence of oncogenic mutations in the BRAF (or, in some cases, NRAS) gene28, nevi are typically associated with an exceedingly low proliferative activity. However, while this proliferative arrest can be maintained for decades29, several studies provide genetic evidence that a fraction of nevi still progresses to melanoma30-32. As a BRAFV600E mutation alone is insufficient to drive melanomagenesis, nevi must acquire additional genetic and/or epigenetic alterations to evade growth restraints and become malignant. Several genetic events frequently occurring in melanoma have been described, including the loss of CDKN2A and ARF, amplification of CCND1 or CDK4 33,34, alterations in MMP8, GRM3, ERBB4, GRIN2A, MITF 35-40 and activation of the PI3K pathway41. We have demonstrated previously that the latter event, for example by reduction of the expression levels of PTEN, reflects a rate-limiting step in OIS abrogation on the path to oncogenic transformation42. Moreover, recent exome and whole genome sequencing studies reported melanoma frequently mutated genes including RAC1 and PREX2 43-45. In addition to genetic alterations tumor suppressor genes can be inactivated by epigenetic means. Promoter hypermethylation has emerged as an important epigenetic mechanism responsible for transcriptional repression of a multitude of genes in human cancer cells. In melanoma several established tumor suppressor genes have been shown to be inactivated secondary to promoter hypermethylation, including CDH1, RASSF1A and SERPINB546. Although the role of some of the above-mentioned genes affected by mutation or promoter hypermethylation in melanoma has been demonstrated, the mechanistic relationship with OIS is unclear for

most. Hence, the molecular mechanism and the identity of the factors underlying malignant transformation in relation to OIS and OIS bypass are largely unknown. Therefore, we set out to perform a near-genomewide (sh)RNA screen for novel OIS factors. Such genes would be predicted to function as tumor suppressor genes in tumor cells expressing an activated oncogene, with PTEN serving as a prime example47,48. We have recently performed a genome-wide analysis of promoter methylation in primary melanoma and benign nevus and identified a large number of genes, including several tumor suppressor genes, to be hypermethylated in melanoma49. Here, we have integrated the data from both RNAi screening and methylation analyses. RESULTS Near-genomewide shRNA screen identifies genes required for BRAFV600E-induced senescence In order to identify factors required for BRAFV600E to induce senescence, we performed a function-based knockdown screen for genes that can rescue this type of cell cycle arrest (Figure 1a). Human diploid fibroblasts (HDFs) were transduced with a near-genomewide lentiviral shRNA library targeting 15.000 human genes, organized in 10 pools with some five shRNAs per gene target. Transduction was performed at a multiplicity of infection (MOI) of 0.5 to ensure that only a single shRNA integrated per cell. The number of cells used for transduction was calculated to obtain minimum coverage of 10. After pharmacological selection, cells were transduced with BRAFV600E-encoding retrovirus. While cells containing empty vector control underwent cell cycle arrest as expected, cells that contained shRNAs abrogating BRAFV600E-induced senescence kept growing and formed colonies. These colonies were subsequently picked individually, expanded and their genomic DNA was isolated. To identify hits, gDNAs isolated from individual colonies derived from independent shRNA pools were grouped and received a unique index. shRNAs were re-amplified and analyzed by deep sequencing. If a particular shRNA gave rise to more than one colony, it would be present in more than one group. Candidates for validation were selected based on two criteria: 1) the presence of at least 2 shRNAs targeting the same gene to exclude off-target effects, 2) the extent of senescence abrogation: only shRNAs identified in independent colonies were further analyzed (Table S1.) On the basis of these criteria, shRNAs targeting 40 genes were re-evaluated, one by one, for their ability to abrogate BRAFV600E-induced cell cycle arrest. Seven genes (UBE2V1, NMRAL1, PCDHGA10, SLC1A4, WT1, GEMIN6, RASEF) could be validated in multiple rounds with two shRNAs in cell proliferation assays (Figure 1b) and BrdU incorporation assays (Figure 1c). We next tested whether the identified shRNAs were on-target, i.e., whether they deplete the gene of interest. Quantitative real-time (qRT) PCR analysis confirmed that in all cases the expression of the target gene was decreased (Figure 1 d-j). We previously demonstrated

100

EM IN 6 sh RA SE F

sh G

sh W

1A 4

A1 0

sh PC

sh SL C

RA L1

2V 1

sh N

sh U

*** *

***

0.5

cycling

* **

shSLC1A4 BRAFV600E

2 1.5 **

0.5 0

***

shWT1 BRAFV600E

4 1A

0 A1

WT1

2.5

BRAFV600E

Chapter 5

shPCDHGA10 BRAFV600E

GEMIN6

2.5 2 1.5 1

0.5 0

***

shGEMIN6

j 10

RASEF

8 6 4 2 0

***

BRAFV600E

***

0.5

cycling

1.5

Relative amount of transcript

SLC1A4

2.5

Relative amount of transcript

***

shNMRAL1

BRAFV600E

shUBE2V1

cycling

***

PCDHGA10

cycling

1 **

cycling

***

2.5

1.5

0.5

BRAFV600E

cycling

1 ***

1.5

Relative amount of transcript

NMRAL1

2.5

1.5

Relative amount of transcript

UBE2V1

2.5

g Relative amount of transcript

Relative amount of transcript

cycling

Relative amount of transcript

2V U sh

5 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2 #1 #2

***

cycling

isolate DNA, pool and identify positive shRNAs by deep sequencing

isolate and expand BRAFV600E- senescence bypassing colonies

BrdU positive cells normalised to OIS cells

shIGFR2

infect with BRAFV600E retrovirus & select

HDFs

transduce with human genome-wide lentiviral shRNA library & select

cycling

BRAFV600E

***

shRASEF BRAFV600E

Figure 1. Near-genomewide shRNA screen identifies genes required for BRAF -induced senescence a. Schematic summary of near-genomewide shRNA screen for genes required for BRAFV600Eâ&#x20AC;&#x201C;induced senescence. b. Cell proliferation assay on HDFs transduced with vector control or shRNA targeting candidate genes identified in the screen. Cells were fixed and stained 11 days after exposure to BRAFV600E. shRNA targeting IGFR2 was used as a positive control (Kuilman et al., unpublished data). c. BrdU incorporation of samples described in b., measured 9 days after exposure to BRAFV600E. Levels are represented as mean of three independent validation rounds. Error bars represent SD. Significance was determined by oneway ANOVA comparing OIS cells with cells depleted of the indicated target gene. d-j. Regulation of gene transcripts of the samples described in b. as determined by qRT-PCR. Measurements are standardized to the vector-expressing cells. Error bars represent SD from triplicate qRT-PCRs. V600E

101

that OIS genes are commonly induced by the oncogenic stressor, for example interleukins13. Consistent with this notion, the expression of all genes but SLC1A4 was induced in senescent cells. Taken together, this near-genomewide function-based shRNA screen identified seven genes crucial for BRAFV600E-induced cell cycle arrest. RASEF is a new OIS gene Out of the seven genes identified in the screen, the expression of RAS and EF-hand domain containing (RASEF) transcript, also known as RAB45, was induced by BRAFV600E to the highest extent (Figure 1j). Additionally, the RASEF locus was previously shown to be associated with predisposition to hereditary melanoma50 and epigenetically regulated in uveal melanoma51. In the light of those observations, and because BRAFV600E-induced senescence prevents nevi from progressing to melanoma26, we studied the role of RASEF in OIS and OIS abrogation in more detail. For this, with the use of additional shRNA, we confirmed that RASEF depletion (Figure 2a) abrogates BRAFV600E–induced cell cycle arrest as determined by cell proliferation (Figure 2b) and BrdU incorporation assays (Figure 2c). The abrogation of cell cycle arrest was seen already at day 10 after introduction of BRAFV600E (around the day when senescence is fully established) and was maintained up to at least 18 days (Figure 2b). Importantly, this could not be explained by loss of BRAFV600E expression or loss of activation of downstream MEK/ ERK signaling (Figure 2d). The rescue of proliferation by RASEF knockdown was associated with restoration of expression of the DNA replication-associated protein PCNA (Figure 2d). Moreover, RASEF depletion resulted in a profound suppression of other common senescenceassociated biomarkers SA-β-Gal activity (Figure 2e), expression of the senescence-associated p15INK4B tumor suppressor (Figure 2d) and the number of SAHF-positive cells (Figure 2f). Whereas RB as expected accumulated in its hypophosphorylated form in BRAFV600E senescent cells, RASEF depletion reversed its phosphorylation status (Figure S1), suggesting cross talk between RASEF and RB signaling. In contrast, p53 expression levels remained unaffected. Notably RASEF depletion also led to a marked reduction in IL6 and IL8 transcript levels, typical of OIS cells (Figure 2g). We have previously identified C/EBPβ to be a critical mediator in the interleukin pathway leading to OIS13. Indeed, while the C/EBPβ transcript was highly induced in OIS, this was significantly reduced upon RASEF silencing (Figure 2h). From these findings, we conclude that RASEF is a new and essential component of OIS. RASEF is hypermethylated in melanoma Recently, we reported a genome-wide analysis of promoter hypermethylation in primary cutaneous melanomas and benign nevi, interrogating the methylation status of 14.495 genes49. The identification in the present study of seven genes required for BRAFV600E-induced senescence prompted us to examine whether any of these genes is epigenetically silenced in melanoma harboring a mutant BRAF gene. We reanalyzed the epigenomic profiling data taking into account the BRAF mutation status of the melanoma and nevus samples. From

102

vector

94.18±1.43

BRAFV600E shRASEF#2

24.45±2.56

BRAFV600E vector

94.60±4.07

shRASEF#2

23.95±1.97

shRASEF#3

33.01±4.91

cycling PCNA p15

BRAFV600E

Hsp90

shRASEF#3

34.40±6.29

tMEK

#2 #3 shRASEF

1.2

IL6

IL8

0.8 0.6 0.4

*** ***

0.2

0 *** ***

***

shRASEF BRAFV600E

Chapter 5

BRAFV600E

1.59±0.76

pMEK

1.2 1 0.8 0.6

0.4

0.2 *** 0

5.44±2.41

#2 #3 d18 shRASEF

BRAF **

RASEF

***

shRASEF v

cycling

***

d14

BRAFV600E

Relative amount of C/EBPβ transcript

***

d10

cycling

Relative amount of transcript

shRASEF

cycling

6 4

BRAFV600E

% of BrdU positive cells

cycling

Relative amount of RASEF transcript

the samples subjected to genome-wide methylation analyses, all five nevi and five out of 24 melanomas harbored a BRAFV600E mutation, while two had a BRAFV600K mutation. It appeared that, out of the seven genes identified in the shRNA screen, only the RASEF gene was found to be differentially methylated in nevi and melanomas. In fact, RASEF was among the 16 most frequently hypermethylated genes in BRAF-mutant melanomas, whereas its methylation was absent from nevi. Comparative analysis pointed to hypermethylation in four of seven BRAF-mutant melanomas with an average β-value difference (a measure of differential methylation) of 0.32 (Figure 3a).

shRASEF BRAFV600E

Figure 2. RASEF is a new OIS gene a. Regulation of RASEF transcript in HDFs expressing vector control or nonoverlapping shRNAs targeting RASEF in the presence and absence of oncogenic BRAFV600E. Level of expression is determined by qRT-PCR. Measurements are based on three independent experiments and standardized to the vector-expressing senescent cells. Error bars represent SD. b. Cell proliferation assay on samples from a. Cells were fixed and stained 13 days after exposure to BRAFV600E. c. BrdU incorporation of samples described in a., measured 9 days after infection with BRAFV600E-encoding retrovirus. Levels are represented as mean of at least three independent experiments. Error bars represent SD. d. Samples from a. were analyzed by immunoblotting with antibodies as indicated. Hsp90 serves as loading control. e. Representative images of SA-β-galactosidase staining for the cells described in a. Quantification of SA-β-galactosidase positive cells was performed on three independent experiments, with SD. f. Representative images of DAPI staining for the cells described in a. Quantification of SAHF positive cells was performed on two independent experiments, with SD from triplicate. g-h. Regulation of IL6 and IL8 (g) and C/EBPβ (h) transcripts of the samples described in a., as determined by qRT-PCR. Measurements are based on three independent experiments and standardized to the BRAFV600E-expressing senescent cells. Error bars represent SD.

We proceeded with the validation of RASEF promoter hypermethylation in a panel of 76 primary melanoma and 15 nevus biopsy specimens using bisulphite melting curve analysis

103

(BMCA). In 16 of 76 melanoma samples (21%) the RASEF promoter was hypermethylated, whereas none of the 15 nevi showed methylation (Figure 3b). Bisulphite sequencing analysis (BSA) of five nevi, three primary melanomas and two early-passage melanoma cell lines further confirmed dense RASEF hypermethylation in the primary tumors and melanoma cell lines, against a general absence of RASEF methylation in the nevi (Figure 3c). To further assess whether an association exists between RASEF promoter methylation and BRAF mutation in melanoma, we performed mutation analysis of BRAF and NRAS on a subset of 24 primary melanomas. In four of these, RASEF was methylated; this included one melanoma with a BRAF mutation, no melanoma with NRAS mutation and three not harboring mutations in these melanoma oncogenes (Table S2). This result shows that RASEF methylation is not confined to melanomas that carry a BRAF mutation. Next we assessed the correlation between RASEF promoter methylation and transcript abundance in a panel of melanoma cell lines. In addition we analyzed whether a relationship exists between RASEF silencing, MAP kinase pathway activation and the RB pathway. RASEF transcript expression was very low to undetectable in the six melanoma cell lines with promoter hypermethylation, whereas significantly higher RASEF expression was present in several cell lines with absent RASEF promoter methylation (Figure 3d, Figure S2). RASEF expression levels neither correlated with MAP kinase pathway activation (as judged by lack of differences in ERK and MEK phosphorylation; Figure S3), nor with RB phosphorylation (Figure S4). Treatment with 5-aza-2â&#x20AC;&#x2122;-deoxycytidine of two different cell lines derived from melanoma metastases, both positive for RASEF promoter methylation, resulted in re-activation of RASEF gene expression with an induction of 7-fold and 3.5-fold in the cell lines 634 and 06.24 respectively (Figure 3e), suggesting that promoter methylation of RASEF is correlated with transcriptional silencing. Collectively, these results demonstrate that hypermethylation of the RASEF gene affects 21% of primary cutaneous melanomas. RASEF acts as potential tumor suppressor The essential role of RASEF in OIS, together with data on its differential methylation in nevi and melanomas, would be consistent with a tumor-suppressive role for RASEF. This hypothesis would predict that restoration of RASEF expression in cells in which the gene has been silenced by methylation acts cytostatically. Therefore, we re-expressed RASEF in two BRAFV600E-mutant melanoma cell lines (04.04 and A875) with a hypermethylated RASEF promoter and very low RASEF mRNA expression (Figure 3d). We ectopically expressed V5tagged RASEF (Figure 4a) and analyzed the effect on cell proliferation and survival. In both cell lines, this resulted in induction of cell cycle arrest as measured by cell proliferation and BrdU-incorporation assay (Figure 4b, c). This was associated with a decrease in the steady state levels of PCNA and induction in the levels of the senescence-associated protein p21Cip1 (Figure 4a). We could not detect p15INK4B protein in these cell lines, possibly owing to

104

a deletion of the locus, which is commonly seen in melanoma. The RASEF-induced cell cycle arrest was not associated with cell death (Figure 4d, e). Melanoma cell lines in which RASEF was methylated were more prone to halt cell cycle progression upon its re-introduction than were the RASEF-proficient cells (Figure S5). These observations together are compatible with a model in which RASEF has a tumor suppressor function.

exon 1 +383

amplicon

RASEF

Normalized expression

- RASEF

promoter methylation

1.2

+ RASEF

promoter methylation

0.8

BLM

A875

634

04.04

06.24

94.07

94.13

93.08

607B

Mz7.4

94.03C

0.4 05.06

+258 c c c cc c c c ccc cc c c c c 1 11 bisulphite seq.

TSS 0

9.0 8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0 0.0

mel. cell line 634

***

- AZA + AZA

mel. cell line 06.24

4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0

0 TSS

not methylated partially methylated fully methylated 1

1.4

amplicon

+383

primary melanoma

518A2

exon 1

cg10319505

pr. melanoma n=76 21%

nevus n=15 0%

04.07

CpG island

+258 ATG +261

nevus

0.0

RASEF CpG island

***

CpG dinucleotides

2 3 4 5

6 7 8 9 10 11

nevus #1 nevus #2 nevus #3 nevus #4 nevus #5 pr. melanoma #1 pr. melanoma #2 pr. melanoma #3 mel. cell line #1 mel. cell line #2

- AZA + AZA

Figure 3. RASEF is hypermethylated in melanoma a. Hypermethylation of the RASEF promoter in BRAF-mutant primary melanoma samples as measured by Infinium’s 27K genome-wide beadchip assay. Normalized average β-values represent extent of methylation of probe cg10319505 for 5 BRAF-mutant nevi and 7 BRAF-mutant primary melanoma samples. b. RASEF promoter methylation status by bisulphite melting curve analysis of the indicated RASEF CpG island region performed in 15 benign nevi and 76 primary melanoma biopsy samples. c. RASEF promoter methylation analysis by bisulphite sequencing of the indicated CpG dinucleotides in 5 benign nevi, 3 primary melanomas and 2 early-passage melanoma cell lines. d. RASEF mRNA expression levels in early-passage melanoma cell lines with known RASEF promoter methylation status as determined by qRT- PCR. Error bars represent SD from triplicate qRT-PCRs e. Re-activation of RASEF mRNA expression upon 5-aza-2’-deoxytidine treatment in early-passage melanoma cell lines 634 and 06.24 with pre-existent RASEF methylation. Measurements are representative for treatment experiments performed in duplicate, SD from triplicate qRT-PCRs.

RASEF contributes to BRAFV600E-induced senescence in melanocytes We identified and subsequently validated RASEF in a function-based screen for genes that are required for OIS in a model system, cultured human diploid fibroblasts (Figure 1, 2). The results above indicate that this gene is frequently silenced in melanoma (Figure 3). Together, these data predict that RASEF plays an important role also in melanocytes, in the context of OIS. We therefore validated the key results obtained in fibroblasts in cultured human melanocytes. Indeed, RASEF depletion in BRAFV600E-expressing melanocytes caused an almost five-fold induction of cell proliferation, as measured by BrdU incorporation assay (Figure 5ab), similar to what we previously reported for PTEN 42. This was associated with a drop in

105

Chapter 5

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0

Normalized expression

average β-value

several senescence biomarkers, particularly IL6 and IL8, as well as C/EBPβ transcripts (Figure 5c-e). We conclude from these results that RASEF is a candidate melanoma susceptibility gene that acts, at least in part, by contributing to OIS.

A875

0.4

p21

0.2

Hsp90

A875

Fold cell death induction

04.04

eGFP

RASEF

0 1.6 1.4

eGFP

RASEF

1.2 1 0.8 0.6 0.4 0.2

BRAFV600E

b RASEF

4 3

***

A875

1.2

IL6

0.8

***

0.6 0.2 ***

BRAFV600E

shRASEF

IL8

0.8

***

0.6

BRAFV600E

1.2

0.8 0.6

0.2

***

BRAFV600E

C/EBP β

0.4

0.2

shRASEF

***

1.2

0.4

cycling

04.04

Relative amount of transcript

cycling

***

shRASEF

shPTEN

***

Relative amount of transcript

45 *** 40 35 30 25 20 15 10 5 0

cycling

% of BrdU positive cells

***

04.04

0.6

A875

shRASEF

cycling

Figure 4. RASEF acts as potential tumor suppressor a. Immunoblotting analysis of two indicated melanoma cell lines expressing eGFP as control or RASEF. Hsp90 serves as loading control. b. Cell proliferation assay on samples from a. Cells were fixed and stained 12 days after exposure to RASEF. c. BrdU incorporation of samples described in a., measured 16 days after infection with RASEF-encoding lentivirus. Levels are represented as mean of at least three independent experiments. Error bars represent SD. d. Representative phase contrast images for the cells described in a. e. Cell death induction of cells described in a. as measured by trypan blue exclusion assay. Results represent mean of at least three independent experiments. Error bars represent SD.

Rasef

RASEF

Relative amount of transcript

eGFP

0.8

04.04

PCNA

1.2

shRASEF

V5 tag

c eGFP RASEF

cycling

Relative amount of transcript

A875

Normalized % of BrdU positive cells

04.04

eG FP RA SE F eG FP RA SE F

BRAFV600E

Figure 5. RASEF contributes to BRAFV600E-induced senescence in melanocytes a. BrdU incorporation in cultured human melanocytes expressing vector control or shRNA targeting RASEF in the presence and absence of oncogenic BRAFV600E. shRNA targeting PTEN was used as a positive control. Measurement was performed 15 days after infection with BRAFV600E-encoding lentivirus. Levels are represented as mean of two independent experiments with SD from triplicate. b. Regulation of RASEF transcript in samples described in a., as determined by qRT-PCR. Measurements are based on two independent experiments and standardized to the vector-expressing senescent cells. Error bars represent SD from triplicate qRT-PCRs. c-e. Regulation of IL6 (c), IL8 (d) and C/EBPβ (e) transcripts in samples described in a., as determined by qRT-PCR. Measurements are based on two independent experiments and standardized to the BRAFV600E-expressing senescent cells. Error bars represent SD from triplicate qRT-PCRs.

DISCUSSION Over the last decade, OIS has been established as a safeguard program protecting against cancer. However, we have only begun to dissect the underlying mechanism. There is ample evidence to argue that genes critically required for OIS are endowed with tumor suppressor

106

107

Chapter 5

functions, with p16INKA, p15INK4B, p53 and PTEN serving as a paradigms6,42. In this study, we performed a near-genomewide shRNA screen to uncover novel factors essential for BRAFV600E to induce senescence. We identified seven genes, one of which we characterized in more detail, because we found it was frequently methylated in melanomas. RASEF depletion prevented the induction of senescence by BRAFV600E both in fibroblasts and melanocytes. This was accompanied by the suppression of several biomarkers of senescence including SA-β-Gal activity, expression of p15 tumor suppressor protein and SAHF. Furthermore, RASEF depletion reversed the increase in components of the SMS/SASP, which represent key factors for the induction of senescence-associated cell-cycle arrest13. Function-based screens for mediators of OIS have been performed previously, too, in several different experimental settings and with gain-of-function libraries as well as loss-of-function libraries14,52-55. The latter study was done in the context of BRAFV600E-induced senescence; we found an overlap of two genes, BUB1 and NF2. Moreover, we identified two genes with an established function in OIS, that is, TSC22 and IL6R (albeit with single shRNAs in the primary screen)13,56. These results show that the current screens are not saturating, and that multiple approaches ought to be used to draw a complete map of critical OIS regulators. None of the seven genes identified in our screen has been previously studied in the context of BRAFV600E-induced senescence. Depletion of RASEF suppressed the expression of IL6, IL8 transcripts via downregulating C/EBPβ, a key controller of the SMS/SASP contributing to OIS13. It will be interesting to determine the relationship between C/EBPβ and RASEF. We have previously identified additional regulators of senescence, including the putative tumor suppressor TSC2256 and the metabolic enzymes PDK1 and PDP224. Interestingly, all these genes are essential to drive the expression of IL6 and IL8 transcripts in the face of an oncogenic signal. These results not only support model in which RASEF controls OIS by regulating components of the senescence secretome, but also highlight the central role of specific interleukins in OIS. Among the genes identified here, only Wilms Tumor 1 (WT1) has been previously associated with tumorigenesis57. Moreover, none were found to be among the genes that were recently reported to be frequently mutated in melanoma43-45. We find that, out of seven genes identified in the screen, only the RASEF gene shows promoter hypermethylation in a substantial fraction of primary cutaneous melanomas (21%). Benign melanocytic nevi did not show RASEF promoter methylation. Although results from our initial genomewide methylation analysis suggested selective RASEF hypermethylation in mutant BRAF melanomas, we observed in a subsequent assessment of a larger group of tumors that RASEF hypermethylation occurred equally in BRAF-mutant and -wild-type tumor samples. RASEF hypermethylation was associated with transcriptional downregulation and chemical demethylation resulted in transcriptional reactivation of this gene. RASEF promoter hypermethylation has previously been reported to occur in uveal melanoma, a melanoma

subtype with genetic and clinical features that are distinct from cutaneous melanoma51. In that study, methylation of RASEF was found in primary uveal melanomas and more prominently in uveal melanoma cell lines lacking RASEF expression. Only cell lines lacking RASEF methylation were found to express the gene. This pattern of correlation between promoter methylation status and gene expression is similar to our findings for RASEF in cutaneous melanoma. Interestingly, in another study of three Danish families with multiple cutaneous and uveal melanoma cases, a susceptibility locus was mapped to a 3 Mb chromosomal region on 9q21.32 harboring the RASEF gene50. Although the presence of germ line mutations in the RASEF gene was not analyzed in this study, reduced expression of this gene was found in melanoma patients, further supporting RASEF as a candidate melanoma susceptibility gene, but this should be explored further. To begin delineating such a role for RASEF, we show that its restored expression in a limited number of melanoma cells in which this gene is methylated causes cell cycle arrest. Clearly, a more definitive characterization of RASEF as a tumor suppressor will require in vivo models. Notably, RASEF has been previously reported to have differential functions depending on the cancer type; it has been attributed with a tumor-suppressive function in myeloid leukemia58 while it has an oncogenic function in lung cancer59. Our study is consistent with a model in which RASEF acts as a tumor suppressor in BRAFV600E -mutant melanomas. MATERIALS AND METHODS Cell culture, viral transduction, and senescence induction The human diploid fibroblast (HDF) cell line TIG3 expressing the ectopic receptor, hTERT and sh-p16INK4A as well as melanoma cell lines (04.04, 04.07, 05.06, 518A2, 06.24, 607B, 634, 93.08, 94.03C, 94.07, 94.13, A875, BLM, Mz7.4) were maintained in DMEM, supplemented with 9% fetal bovine serum (Sigma) and 2 mM glutamine (GIBCO). Melanocytes were maintained as described previously26. Lentiviral and retroviral infections were performed using HEK293T cells and Phoenix cells, respectively, as producers of viral supernatants. For senescence experiments, HDF were infected with shRNA-encoding or protein-coding retro- or lentivirus, selected pharmacologically (puromycin or blasticidin) and subsequently infected with BRAFV600E-encoding or control virus. After selection, cells were seeded for cell proliferation assay, BrdU incorporation assay, SA-β-galactosidase activity and analyzed. Cell proliferation assay Cells were seeded into a six-well plate (at different densities: 2 × 105, 4 × 105 or 6 × 105 cells, resp.) and selected pharmacologically. Fixation and staining with crystal violet was performed 9 to 18 days after the BRAFV600E-encoding or control virus infection, or 12 to 16 days after RASEF-encoding or control eGFP-encoding virus infection. Images of cell proliferation assays reflect representative results of at least three independent experiments.

108

109

Chapter 5

BrdU incorporation assay BrdU labeling was carried out for 3 h followed by fixation. Incorporated BrdU was detected by immunostaining as described in reference 3 and FACS analysis. Results are represented as mean with SD of at least three independent experiments. Analysis of SA-β-galactosidase activity Senescence-associated (SA) β-galactosidase was stained using the ‘Senescence Associated β-Galactosidase Staining Kit’ from Cell Signaling at pH 6 according to the manufacturer’s protocol. Images reflect representative results of at least three independent experiments. Trypan blue exclusion assay Cells were brought into suspension using trypsin, centrifuged and re-suspended in a small volume of culture medium. Trypan blue (Sigma) was added to the cell suspension (dilution factor=2) and stained cells were counted as dead. The number of dead cells was quantitated, and the values were expressed as the fold change over control. Results are represented as mean with SD of at least three independent experiments. Plasmids The human TRC1 shRNA Library was purchased from Sigma-Aldrich. pMSCV-blast-BRAFV600E and pMSCV-blast, KH1-GFP-shPTEN#1, HIV-CSCG-blast-BRAFV600E and HIV-CSCG-blast were previously described13,42. For the re-expression of RASEF, the pLX304 RASEF - V5 as well as control pLX304 - eGFP - V5 plasmids, both from CCSB-Broad Lentiviral Expression Library, were purchased from Thermo Scientific. shRNAs The human TRC1 shRNA Library was purchased from Sigma-Aldrich (http://www. sigmaaldrich.com/life-science/functional-genomics-and-rnai/shrna/products/lentiplex. html). Lentiviral knockdown constructs were purchased from Sigma-Aldrich in pLKO.1 backbone: shUBE2V1#1–TRCN0000033708 – 5′- CGCCTAATGATGTCTAAAGAA-3′ shUBE2V1#2–TRCN0000033706–5′- CCAAGAGCCATATCAGTGCTA-3′ shNMRAL1#1–TRCN0000036912–5′- GGGACATTCAAGGTTCGAGT-3′ shNMRAL1#2–TRCN00000036913–5′- CAAGATGACTCCTGAGGACTA-3′ shPCDHGA10#1–TRCN0000053343 – 5′- TTTCTATTTCATAGAAACCGG -3′ shPCDHGA10#2– TRCN0000053344–5′- ATTCCTCAGGAATTGAGTAGG -3′ shSLC1A4#1– TRCN0000038641–5′- AGAGGATCAGCAGGTTTAT-3′ shSLC1A#2– TRCN0000038642–5′- CCACCTGAATCAGAAGGCAA-3′ shWT1#1– TRCN0000001114–5′- AAAGTTTACATTAGCAGACAC-3′ shWT1#2– TRCN0000001117–5′- AAGTCACACTGGTATGGTTTC-3′ shGEMIN6#1– TRCN0000147641–5′- TTCATAGTTTCAACAGTCTGC-3′ shGEMIN6#2– TRCN0000146917–5′- TATACTCATTCTTCTCACTGG-3′ shRASEF#1– TRCN0000055624–5′- TAATGGGAACAGTCTCATGGG-3′

shRASEF#2– TRCN0000055625–5′- TTTCCGTGTCTTATGTTCTGC-3′ shRASEF#3– TRCN0000055626–5′- ATTCTCGTATGTTAAGAAAGC-3 Quantitative real-time PCR Total RNA was DNase-treated with RQ1 RNase-Free DNase (Promega). Reverse transcription was performed with SuperScript II First Strand Kit (Invitrogen). qRT–PCR was performed with the SYBR Green PCR Master Mix (Applied Biosystems) on an ABI PRISM 7700 Sequence detection System. Primer sets used were as follows: UBE2V1: 5’- GGAGAGGTTCAAGCGTCTTACC -3’ and 5’- TTCGAGTTCTTCCAACAGTCG -3’ NMRAL1: 5’- GCTTACGCCACCTTCATCG -3’ and 5’- CAGATCAGCGAGCAGCTTCC -3’ PCDHGA10: 5’- CTGCAAGCCATGATCTTGG -3’ and 5’- AGACATTCTGGCGGTAGTCG-3’ SLC1A4: 5’- TCCGAAGGAGAAGACCTCATCC -3’ and 5’- CTTCCAACAAGGAACATGATGC -3’ WT1: 5’- TACAGCACGGTCACCTTCG -3’ and 5’- CACCGAGTACTGCTGCTCAC -3’ GEMIN6: 5’- CGAGTGACAGCCAGTGAGAAG -3’ and 5’- ATCTTCAAGGAAGTTCACAAGG-3’ RASEF: 5’- GCTGCTACAGAGGGACAAAAA -3’ and 5’- CAGAATAATGCCCCATACGTC -3’ or 5’- ATCAGACTTCAAAGCACAGAAATGG-3’ and 5’- TTCCTCTTCCAACTCACTCAACTG-3’ IL6, IL8, C/EBPβ and RPL13 (standard) primer sequence were described previously13. All primer pairs except C/EBPβ span exon-exon borders. RPL13 was used as control. For analysis, the ΔCT method was applied. Data are represented as mean ± SD of three or more independent experiments. Antibodies Antibodies used for immunoblotting were BRAF (sc-5284; Santa Cruz), Hsp90 (4874; Cell

110

111

Chapter 5

Signaling), MAP kinase p44/42 (9102; Cell Signaling); phospho-MAP kinase p44/42 (9106; Cell signaling), MEK1/2 (L38C12; 4694; Cell Signaling), phospho-MEK1/2 [Ser217/221] ((41G9); 9154; Cell Signaling), PCNA ((PC10); sc-56; Santa Cruz), p21Waf1/Cip1 ((C19); sc-397; Santa Cruz), p15INK4B (sc-612; Santa Cruz), p53 ((DO-1);sc-126; Santa Cruz); phospho-RB (Ser807/811) (9308; Cell Signaling), total RB (9309; Cell Signaling); RASEF (HPA021431; Sigma), V5-tag (R960-25; Invitrogen). Patient samples Fresh-frozen and paraffin-embedded biopsy specimens were obtained from 76 primary cutaneous melanomas and from 15 benign melanocytic nevi, all containing at least 70% melanocytic cells (for detailed clinical information see Table S2). The BRAF and NRAS mutation status of a subset of 5 nevi and 24 primary melanomas was determined using allele-specific PCRs for BRAFV600E, BRAFV600K, NRASG12D, NRASQ61K, NRASQ61L, NRASQ61R and NRASQ61H mutations on a CFX384 Real-Time Detection System (Bio-Rad, Hercules, CA). DNA was extracted with the Genomic-tip kit (Qiagen, Hilden, Germany) or RecoverAll Nucleic Acid kit (Ambion, Carlsbad, CA). Bisulphite melting curve analysis (BMCA) and bisulphite sequencing analysis (BSA) Bisulphite conversion was performed with the EZ DNA methylation kit (Zymo Research, Orange, CA). RASEF bisulphite primers for BMCA and BSA were designed to amplify a 126-basepair region at +259 from the TSS, located within the CpG island that covers the promoter region and first exon of the RASEF gene (5’-GGGATGGAGGCGGATGGG-3’ and 5’-GGTATTGTGTACGGAGTTGCGG-3’) (Figure 3b). Sensitivity of the primer set for BMCA was validated using mixtures, 1:1, 1:3, 1:9 and 1:9.5, of completely methylated CpGenome universal methylated DNA (Chemicon, Hampshire, United Kingdom) and unmethylated semen DNA respectively, showing that methylation could be accurately detected if 10% of the total analyzed DNA was methylated. Generated bisulphite melting curves of the RASEF CpG island amplicon were scored as previously described49. Sequences obtained from bisulphite sequencing analysis were considered methylated if the density of methylated CpG dinucleotides within the interrogated amplicon was 15% or more, a threshold value that has been previously used as a scoring standard60,61. 5-aza-2’-deoxycytidine treatment Melanoma cell lines 634 and 06.24 were seeded at 15% confluency and treated with 2µM 5-aza-2’-deoxycytidine (Decitabine, Sigma, St. Louis, MO) for 96 hrs. Culture medium was replaced with medium containing freshly prepared 5-aza-2’-deoxycytidine every 24 hours. Cells were harvested for RNA extraction with the RNeasy Mini kit (Qiagen) and expression of RASEF was analyzed by qRT-PCR. Stable expression of the reference genes TBP and CPSF6 was validated using geNorm analysis62.

REFERENCES 1. 2. 3.

4. 5. 6.

7. 8. 9.

10.

11.

12. 13.

14.

112

Lowe, S. W., Cepero, E. & Evan, G. Intrinsic tumour suppression. Nature 432, 307–315 (2004). Mathew, R., Karantza-Wadsworth, V. & White, E. Role of autophagy in cancer. Nat Rev Cancer 7, 961–967 (2007). Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88, 593–602 (1997). Collado, M. & Serrano, M. Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51–57 (2010). HAYFLICK, L. THE LIMITED IN VITRO LIFETIME OF HUMAN DIPLOID CELL STRAINS. Exp Cell Res 37, 614–636 (1965). Campisi, J. Senescent cells, tumor suppression, and organismal aging: good citizens, bad neighbors. Cell 120, 513–522 (2005). Sherr, C. J. Divorcing ARF and p53: an unsettled case. Nat Rev Cancer 6, 663–673 (2006). Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). Dimri, G. P. et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci USA 92, 9363–9367 (1995). Narita, M. et al. Rb-mediated heterochromatin formation and silencing of E2F target genes during cellular senescence. Cell 113, 703–716 (2003). Zhang, R., Chen, W. & Adams, P. D. Molecular dissection of formation of senescenceassociated heterochromatin foci. Mol Cell Biol 27, 2343–2358 (2007). Acosta, J. C. et al. Chemokine signaling via the CXCR2 receptor reinforces senescence. Cell 133, 1006–1018 (2008). Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). Wajapeyee, N., Serra, R. W., Zhu, X., Mahalingam, M. & Green, M. R. Oncogenic BRAF Induces Senescence and Apoptosis through Pathways Mediated by the Secreted Protein IGFBP7. Cell 132, 363–374 (2008).

15.

16.

17. 18. 19.

20. 21.

22.

23.

24.

25. 26. 27. 28.

Coppé, J.-P. et al. Senescence-associated secretory phenotypes reveal cellnonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol. 6, 2853–2868 (2008). Rodier, F. et al. Persistent DNA damage signalling triggers senescence-associated inflammatory cytokine secretion. Nat Cell Biol 11, 973–979 (2009). Kuilman, T. & Peeper, D. S. Senescencemessaging secretome: SMS-ing cellular stress. Nat Rev Cancer 9, 81–94 (2009). Krizhanovsky, V. et al. Senescence of activated stellate cells limits liver fibrosis. Cell 134, 657–667 (2008). Baker, D. J. et al. Clearance of p16Ink4apositive senescent cells delays ageingassociated disorders. Nature 479, 232–236 (2011). Kang, T.-W. et al. Senescence surveillance of pre-malignant hepatocytes limits liver cancer development. Nature 479, 547–551 (2011). Acosta, J. C. et al. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat Cell Biol 15, 978–990 (2013). Aird, K. M. et al. Suppression of nucleotide metabolism underlies the establishment and maintenance of oncogene-induced senescence. CellReports 3, 1252–1265 (2013). Jiang, P., Du, W., Mancuso, A., Wellen, K. E. & Yang, X. Reciprocal regulation of p53 and malic enzymes modulates metabolism and senescence. Nature 493, 689–693 (2013). Kaplon, J. et al. A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112 (2013). Gray-Schopfer, V. C. et al. Cellular senescence in naevi and immortalisation in melanoma: a role for p16? Br J Cancer 95, 496–505 (2006). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417, 949–954 (2002). Pollock, P. M. et al. High frequency of BRAF mutations in nevi. Nat Genet 33, 19–20 (2003).

30.

31.

32. 33. 34. 35.

36.

37.

38.

39. 40.

41.

42.

Mooi, W. J. & Peeper, D. S. Oncogeneinduced cell senescence--halting on the road to cancer. N Engl J Med 355, 1037–1046 (2006). Bogdan, I., Smolle, J., Kerl, H., Burg, G. & Böni, R. Melanoma ex naevo: a study of the associated naevus. Melanoma Research 13, 213–217 (2003). Dadzie, O. E. et al. RAS and RAF mutations in banal melanocytic aggregates contiguous with primary cutaneous melanoma: clues to melanomagenesis. Br. J. Dermatol. 160, 368–375 (2009). Yazdi, A. S. et al. Mutations of the BRAF gene in benign and malignant melanocytic lesions. J Invest Dermatol 121, 1160–1162 (2003). Curtin, J. A. et al. Distinct sets of genetic alterations in melanoma. N Engl J Med 353, 2135–2147 (2005). Kamb, A. et al. A cell cycle regulator potentially involved in genesis of many tumor types. Science 264, 436–440 (1994). Garraway, L. A. et al. Integrative genomic analyses identify MITF as a lineage survival oncogene amplified in malignant melanoma. Nature 436, 117–122 (2005). Palavalli, L. H. et al. Analysis of the matrix metalloproteinase family reveals that MMP8 is often mutated in melanoma. Nat Genet 41, 518–520 (2009). Prickett, T. D. et al. Analysis of the tyrosine kinome in melanoma reveals recurrent mutations in ERBB4. Nat Genet 41, 1127– 1132 (2009). Prickett, T. D. et al. Exon capture analysis of G protein-coupled receptors identifies activating mutations in GRM3 in melanoma. Nat Genet 43, 1119–1126 (2011). Wei, X. et al. Exome sequencing identifies GRIN2A as frequently mutated in melanoma. Nat Genet 43, 442–446 (2011). Yokoyama, S. et al. A novel recurrent mutation in MITF predisposes to familial and sporadic melanoma. Nature 480, 99–103 (2011). Stahl, J. M. et al. Deregulated Akt3 activity promotes development of malignant melanoma. Cancer Res 64, 7002–7010 (2004). Vredeveld, L. C. W. et al. Abrogation of BRAFV600E-induced senescence by PI3K pathway activation contributes to melanomagenesis. Genes Dev 26, 1055–1069 (2012).

43. 44. 45.

46. 47. 48. 49.

50.

51.

52.

53.

54.

55.

Berger, M. F. et al. Melanoma genome sequencing reveals frequent PREX2 mutations. Nature 485, 502–506 (2012). Hodis, E. et al. A landscape of driver mutations in melanoma. Cell 150, 251–263 (2012). Krauthammer, M. et al. Exome sequencing identifies recurrent somatic RAC1 mutations in melanoma. Nat Genet 44, 1006–1014 (2012). Rothhammer, T. & Bosserhoff, A.-K. Epigenetic events in malignant melanoma. Pigment Cell Res. 20, 92–111 (2007). Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 41, 544–552 (2009). Lin, W. M. et al. Modeling genomic diversity and tumor dependency in malignant melanoma. Cancer Res 68, 664–673 (2008). Gao, L. et al. Genome-wide promoter methylation analysis identifies epigenetic silencing of MAPK13 in primary cutaneous melanoma. Pigment Cell Melanoma Res 26, 542–554 (2013). Jönsson, G. et al. Mapping of a novel ocular and cutaneous malignant melanoma susceptibility locus to chromosome 9q21.32. J Natl Cancer Inst 97, 1377–1382 (2005). Maat, W. et al. Epigenetic Regulation Identifies RASEF as a Tumor-Suppressor Gene in Uveal Melanoma. Investigative Ophthalmology & Visual Science 49, 1291– 1298 (2008). Berns, K. et al. A large-scale RNAi screen in human cells identifies new components of the p53 pathway. Nature 428, 431–437 (2004). Peeper, D. S. et al. A functional screen identifies hDRIL1 as an oncogene that rescues RAS-induced senescence. Nat Cell Biol 4, 148–153 (2002). Rowland, B. D. B., Bernards, R. R. & Peeper, D. S. D. The KLF4 tumour suppressor is a transcriptional repressor of p53 that acts as a context-dependent oncogene. Nat Cell Biol 7, 1074–1082 (2005). Vredeveld, L. C. W., Rowland, B. D., Douma, S., Bernards, R. & Peeper, D. S. Functional identification of LRF as an oncogene that bypasses RASV12-induced senescence via upregulation of CYCLIN E. Carcinogenesis 31, 201–207 (2010).

113

Chapter 5

29.

56. 57.

58.

59.

Hömig-Hölzel, C. et al. Antagonistic TSC22D1 variants control BRAF(E600)-induced senescence. EMBO J 30, 1753–1765 (2011). Huff, V. Wilms’ tumours: about tumour suppressor genes, an oncogene and a chameleon gene. Nat Rev Cancer 11, 111– 121 (2011). Nakamura, S. et al. Small GTPase RAB45mediated p38 activation in apoptosis of chronic myeloid leukemia progenitor cells. Carcinogenesis 32, 1758–1772 (2011). Oshita, H. et al. RASEF is a novel diagnostic biomarker and a therapeutic target for lung cancer. Molecular Cancer Research 11, 937–

60.

61. 62.

951 (2013). Garcia-Manero, G. et al. DNA methylation of multiple promoter-associated CpG islands in adult acute lymphocytic leukemia. Clin. Cancer Res. 8, 2217–2224 (2002). Toyota, M. et al. Methylation profiling in acute myeloid leukemia. Blood 97, 2823– 2829 (2001). Vandesompele, J. et al. Accurate normalization of real-time quantitative RTPCR data by geometric averaging of multiple internal control genes. Genome Biol 3, RESEARCH0034 (2002).

SUPPLEMENTARY TABLES Supplementary Tables S1 and S2 can be found with this article online at http://onlinelibrary. wiley.com/doi/10.1111/pcmr.12248/suppinfo.

114

SUPPORTING INFORMATION BRAFV600E

cycling

shRASEF #2

Figure S1. Regulation of RB and p53 proteins in cycling, OIS and cells bypassing OIS upon depletion of RASEF HDFs expressing vector control or nonoverlapping shRNAs targeting RASEF in the presence and absence of oncogenic BRAFV600E were analyzed by immunoblotting with antibodies as indicated. Hsp90 serves as loading control.

Ser807/811 RB RB p53 Hsp90

- RASEF promoter methylation

Figure S2. RASEF expression in early-passage melanoma cell lines RASEF mRNA expression levels in early-passage melanoma cell lines with known RASEF promoter methylation status as determined by qRTPCR. Bars represent mean of normalized RASEF expression in cell lines with hypermethylated RASEF promoter versus cells with absent RASEF methylation. Error bars represent S.E.M.

+ RASEF promoter methylation

0.4 0.2 **

0.0

pERK

Figure S3. Comparison of MAP kinase pathway activation in melanoma cell lines Melanoma cell lines with and without RASEF silencing were analyzed by immunoblotting with antibodies as indicated. Hsp90 serves as loading control.

tERK pMEK tMEK Hsp90

Ser807/811 RB tRB Hsp90

eGFP

RASEF silenced

518.A2

94.03C

05.06

Figure S4. Regulation of RB protein in melanoma cell lines Melanoma cell lines with and without RASEF silencing were analyzed by immunoblotting with antibodies as indicated. Hsp90 serves as loading control.

RASEF

A875

94.07

RASEF not silenced

94.13

93.08

RASEF not silenced A875

04.04

94.07

RASEF silenced

05.06

93.08

Figure S5. Effect of ecotropic expression of RASEF on melanoma cell lines viability Cell proliferation assay on melanoma cell lines with and without RASEF silencing expressing eGFP as control or RASEF. Cells were fixed and stained 12 days after exposure to RASEF.

115

Chapter 5

94.13

93.08

RASEF not silenced A875

04.04

94.07

RASEF silenced

518.A2

0.6

94.03C

0.8

05.06

Normalized expression

1.0

CHAPTER 6 PHOSPHOPROTEOME DYNAMICS IN ONSET AND MAINTENANCE OF ONCOGENE-INDUCED SENESCENCE

Mol. Cell Proteomics 13, 2089–2100 (2014)

PHOSPHOPROTEOME DYNAMICS IN ONSET AND MAINTENANCE OF ONCOGENE-INDUCED SENESCENCE E.L. de Graaf1,2,3, J. Kaplon4, Houjiang Zhou1,2, A.J.R. Heck1,2,3, D.S. Peeper4 and A.F.M. Altelaar1,2 Biomolecular Mass Spectrometry and Proteomics Group, Utrecht Institute for Pharmaceutical Sciences

and Bijvoet Center for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands. 2Netherlands Proteomics Centre, Padualaan 8, 3584 CH Utrecht, The Netherlands. 3Center for Biomedical Genetics, Padualaan 8, 3584 CH Utrecht, The Netherlands. 4 Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, the Netherlands.

SUMMARY Expression of the BRAFV600E oncoprotein is known to cause benign lesions, for example melanocytic nevi (moles). In spite of the oncogenic function of mutant BRAF, these lesions are arrested by a cell-autonomous mechanism called oncogene-induced senescence (OIS). Infrequently, nevi can progress to malignant melanoma, through mechanisms that are incompletely understood. To gain more insight into this vital tumor suppression mechanism, we performed a mass spectrometry-based screening of the proteome and phosphoproteome in cycling and senescent cells as well as cells that have abrogated senescence. Proteome analysis of senescent cells revealed the upregulation of established senescence biomarkers, including specific cytokines, but also several proteins previously not associated with senescence, including extracellular matrix-interacting. Using both general and targeted phosphopeptide enrichment by Ti4+-IMAC and phosphotyrosine antibody enrichment, we identified over 15,000 phosphorylation sites. Among the regulated phosphorylation sites we encountered components of the interleukins, BRAF and CDK-retinoblastoma (RB) pathways and several other factors. The extensive proteome and phosphoproteome dataset of BRAFV600E-expressing senescent cells provides molecular clues as to how OIS is initiated, maintained or evaded, serving as a comprehensive proteomic basis for functional validation. INTRODUCTION In order to sustain their reproductive lifespan multicellular organisms require several safeguard mechanisms to maintain cell homeostasis. Growth and cell replication are essential processes; however, uncontrolled growth can be detrimental resulting in cancer and eventually death. Therefore, several tumor suppressive mechanisms have evolved including apoptosis1 and replicative senescence2,3 that can lead to cell self-destruction and cell proliferation arrest, respectively. Recently it has been shown both in vitro4-6 and in vivo7-9

119

Chapter 6

doi: 10.1074/mcp.M113.035436.

that oncogene activation (e.g. through BRAF, RAS and E2F) can induce an irreversible cell growth inhibition mechanism termed oncogene-induced senescence (OIS). Human skin cells are naturally exposed to multiple stress factors that can induce gene mutations, which potentially lead to constitutive protein activation and ultimately tumor formation. Benign skin tumors that may remain dormant for decades are manifested in the form of melanocytic nevi (moles) and rarely progress into a malignant state. Nevi show markers of senescence including growth arrest, increase in senescence associated-βgalactosidase (SA-β-Gal) activity and induction of tumor suppressor p16 7. Strikingly, the BRAFV600E activating mutation10 is found with very high frequency (~50%) in both nevi as well as primary melanomas11. This suggests that this BRAF mutation alone is insufficient for melanoma development and additional mutations or other post-transcriptional alterations are needed for transformation. This idea is supported by BRAFV600E knock-in mouse models, which develop nevi that infrequently progress to melanomas12,13. Transcriptomic analysis has previously shown that the maintenance of BRAFV600E-induced senescence is dependent on an inflammatory network governed by the transcription factor CCAAT-enhancer-binding protein β (C/EBPβ)14. Senescence mechanisms defying tumor outgrowth are currently heavily investigated to understand endogenous tumor suppressive pathways involved, and to provide alternative drug solutions to cancer treatment. To better understand, at the molecular level, the mechanisms underlying the onset and maintenance of OIS in human fibroblasts, we used multiple complementary proteomics techniques to achieve a high coverage of both the proteome and the phosphoproteome15. Each protocol has been optimized previously16-18 to maximize the number of proteins and phosphorylation events quantified. Strong cation exchange (SCX) peptide fractionation together with both global phosphopeptide enrichment as well as phosphotyrosine site specific enrichment techniques were applied to allow for a deep coverage of the senescence phosphoproteome. Using stable isotope dimethyl labeling 5,997 proteins, 12,547 phosphoserine, 2,361 phosphothreonine and 590 phosphotyrosine sites could be quantified. EXPERIMENTAL PROCEDURES Cell culture and cell assays The human diploid fibroblast (HDF) cell line Tig3 expressing the ectopic receptor, hTERT and sh-p16INK4A (Tig3 (et)-16i) was maintained in DMEM with 4.5 mg/ml glucose and 0.11 mg/ ml sodium pyruvate, supplemented with 9% fetal bovine serum (PAA), 2 mM glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin (GIBCO). The Phoenix packaging cell line was used for the generation of ecotropic retroviruses. The plasmids pMSCV-blast and pMSCV-blast-BRAFV600E as well as pRS-puro and pRS-puro-C/EBPβ#1 were previously described14. For infections, filtered (pore size 0.45 mm) viral supernatant, supplemented

120

121

Chapter 6

with 4–8 μg/ml polybrene was used. In general, a single infection round of 6 h was sufficient to infect at least 90% of the population. Cells infected with shRNA-encoding retrovirus were selected pharmacologically (puromycin) and subsequently infected with BRAFV600Eencoding or control virus. After blasticidin selection, HDF were seeded for cell proliferation assays into a six-well plate or 6 cm plate (2 × 105, 4 × 105 or 6 × 105 cells, respectively) and maintained in selection medium. Fixation and staining with crystal violet was performed at day 3 and 9 after the last infection. Images of cell proliferation assays reflect representative results of several independent experiments. SA-β-Gal was stained using the ‘Senescence β-Galactosidase Staining Kit’ from Cell Signaling at pH 6 according to the manufacturer’s protocol. Images reflect representative results of several independent experiments. Proteomics sample preparation Frozen cell pellets were lysed by sonication in lysis buffer (8M Urea in 50 mM ammonium bicarbonate, 1 tablet Complete mini EDTA-free Cocktail (Roche) and 1 tablet PhosSTOP phosphatase inhibitor Cocktail (Roche) per 50ml of lysis buffer), samples for phophotyrosine peptide pulldowns contained an additional 1mM sodium orthovanadate. After centrifugation (20 000x g 30min at 4 °C), the supernatant was assayed for protein content using the BCA kit standard procedure (Pierce). Protein reduction and alkylation were performed using final concentrations of 5 mM dithiothreitol and 10 mM iodoacetamide, respectively. A first enzymatic digestion step was performed in 8 M urea lysis buffer using Lys-C at 37 °C for 4 h (enzyme:substrate 1:75). The second digestion was performed overnight (37 °C) with trypsin (enzyme:substrate 1:100) in 2 M Urea. Resulting peptides were chemically labeled and washed on Sep-Pak C18 columns (Waters, USA, Massachusetts) using stable isotope dimethyl labeling as described before19. Cycling HDF cells were labeled with light (L), OIS cells with medium (M) and OIS bypassing cells with heavy (H) dimethyl isotopes. In the replicate experiment the medium and heavy labels were swapped. The labeling efficiency for all labels was higher than 95%. An aliquot of each label was measured on a regular liquid chromatography-mass spectrometry (LC-MS/MS) run and samples were mixed 1:1:1 (L:M:H) based on their peptide intensities. This was found to result in a more precise ratio than using the total protein amounts as determined by a BCA assay. After mixing, peptides were dried in vacuum. Phosphotyrosine peptide immunoprecipitation Peptides were re-suspended in 800 μL of cold immuno-affinity purification buffer consisting of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, protease inhibitors (Roche Diagnostics, Germany) and 1% n-octyl-β-D-glucopyranoside (NOG) (Sigma, Germany). The peptide mixture was agitated on a shaker for 30 min to dissolve the peptides thoroughly and the pH was adjusted, if necessary, to pH 7.4. The immuno-affinity purification was performed as previously described20,21. In summary, the peptides were added to 50 μL slurry of PY99 antibody beads (Santa Cruz biotechnology, CA USA) and the peptide-antibody-bead mixture was incubated

overnight at 4°C on a rotator. The beads were spinned down and supernatant was used for a second iteration of immunoprecipitation using fresh PY99 beads. The pelleted beads were washed and bound peptides were eluted twice using 0.15% trifluoroacetic acid (TFA) (Sigma, Germany). Sample desalting was performed using homemade tips with C18 material (AquaTM C18, 5 μm, Phenomenex, Torrance, CA) as described elsewhere22. Finally, peptides were dried in vacuum and reconstituted in 40 μL of 10% formic acid (FA) prior to LC-MS/MS analysis. SCX chromatography for peptide fractionation SCX was performed on two separate systems optimized for sample amount and type. For the protein identification experiment, 150-200 μg of sample was injected on SCX system 1 and for the Ti4+-IMAC phosphopeptide identification experiment, 3 mg was loaded on SCX system 2. For SCX system 1, peptides were fractionated as described elsewhere23. Briefly, the SCX system consisted of an Agilent 1100 HPLC system (Agilent Technologies) with a Strata X 33u (Phenomenex, The Netherlands; 50 x 4.6 mm) trapping cartridge and a polysulfoethyl A SCX column (PolyLC, Columbia, MD; 200 mm x 2.1 mm, 5 µm, 200-Å). Labeled peptides were reconstituted in 10% FA and loaded onto the trap column at 100 µl/min and subsequently eluted onto the SCX column with 80% acetonitrile (ACN) (Biosolve, The Netherlands) and 0.05% FA. SCX buffer A was made out of 5 mM KH2PO4 (Merck, Germany), 30% ACN and 0.05% FA, pH 2.7; SCX buffer B consisted of 350 mM KCl (Merck, Germany), 5 mM KH2PO4, 30% ACN and 0.05% FA, pH 2.7. The gradient was as follows: 0% B for 10 min, 0–85% B in 35 min, 85–100% B in 6 min and 100% B for 4 min. A total of 45 fractions were collected for each set and dried in a vacuum centrifuge. The second SCX system24 was performed using an Opti-Lynx (Optimized Technologies, Oregon OR) trapping cartridge and a Zorbax BioSCXSeries II column (0.8-mm inner diameter × 50-mm length, 3.5 μm). SCX Solvent A consists of 0.05% FA in 20% ACN while solvent B was 0.05% FA, 0.5 M NaCl in 20% ACN. The SCX salt gradient as follows: 0-0.01 min (0-2% B); 0.01-8.01 min (2-3% B); 8.01-14.01 min (3-8% B); 14.01-28 min (8-20% B); 28-38 min (20-40% B); 38-48 min (40-90% B); 48-54 min (90% B); 54-60 min (0% B). A total of 50 SCX fractions (1 min each, i.e. 50-μl elution volume) were collected and dried in a vacuum centrifuge. Ti4+-IMAC phosphopeptide enrichment Prior to phosphopeptide enrichment, SCX fractions were desalted using Sep-Pak C18 columns and dried to completion using a speed vacuum. Ti4+-IMAC columns were prepared and used as described previously18,25. Briefly, microcolumns were created by loading Ti4+IMAC beads onto GELoader tips (Eppendorf) with a C8 plug to approximately 1–2 cm length. The enrichment procedure for all SCX fractions was as follows: Ti4+-IMAC columns were pre-equilibrated two times with 30 μL of loading buffer (80% ACN, 6% TFA). Next, each SCX fraction was re-suspended in 30 μL of loading buffer and loaded onto the equilibrated GELoader tips. Ti4+-IMAC columns were washed with 40 μL washing buffer A (50% ACN, 0.5%

122

123

Chapter 6

TFA, 200 mM NaCl) and subsequently with 40 μL washing buffer B (50% ACN, 0.1% TFA). Bound peptides were eluted by 30 μL of 10% ammonia into 30 μL of 10% FA. Finally, the remainder of the peptides was eluted with 4 μL of (80% ACN, 2% FA). The collected eluate was further acidified by adding 6 μL of 100% FA and subsequently stored at -20 °C for LCMS/MS analysis. LC-MS/MS For protein identification and quantification the SCX fractions containing doubly and triply charged peptides (approx. 20 fractions each SCX) were reconstituted in 10% FA and analyzed directly using nano flow reverse phase LC using an Agilent 1200 coupled to a LTQ-Orbitrap Velos mass spectrometer (Thermo, San Jose, CA). Depending on the SCX UV-trace, 1-10% of each fraction was injected. Densely populated 2+ fractions were injected twice to minimize undersampling of the mass spectrometer. About half of each phosphopeptide sample was injected to allow for erroneous events. Peptides were trapped on a trap column (ReproSil-Pur C18-AQ, 3μm, Dr. Maisch GmbH, Ammerbuch, Germany; 20 mm x 100 μm inner diameter, packed in house) at 5μl/min in 100% solvent A (0.1 M acetic acid in water). Next, peptides were eluted from the trap column onto the analytical column (ReproSil-Pur C18-AQ, 3μm, Dr. Maisch GmbH, Ammerbuch, Germany; 40 cm x 50-μm inner diameter, packed in house) at ±100 nl/min in 1h, 2h or 3 h linear gradients from 10 to 50% solvent B (0.1 M acetic acid in 8:2 (v/v) ACN:water). Nanospray was achieved with an in-house pulled and gold-coated fused silica capillary (360 μm OD; 20 μm inner diameter; 10 μm tip inner diameter) and an applied voltage of 1.7 kV. In all experiments the mass spectrometer was configured to perform a Fourier transform survey scan from 350 to 1500 m/z at a resolution of 30,000. For the protein identification/quantification experiment the top 10 most intense peaks were fragmented by higher energy collision fragmentation (35% normalized collision energy at a target value of 50,000 ions; resolution=7500). The phosphopeptide samples were analyzed using the higher energy collision-induced dissociation/electron transfer collisionally activated dissociation decision tree as described previously26, fragmenting the top 10 or top 5 most abundant peaks for the Ti4+-IMAC and phosphotyrosine immunoprecipitation samples, respectively. Data analysis All MS data were processed with Proteome Discover 1.3 (Thermo Scientific) using a standardized workflow. Peak lists, generated in Proteome Discover, were searched against a concatenated forward-decoy(reverse) Uniprot (v2010-12, taxonomy Homo Sapiens, 41 008 protein entries) database, supplemented with frequently observed contaminants, using Mascot (Matrix Science, UK). The following search parameters were used: 50 ppm precursor mass tolerance, 0.05 Da fragment ion tolerance for Fourier transform analyzed spectra, 0.6 Da fragment ion tolerance for ion tolerance analyzed spectra, trypsin cleavage with maximum of two miscleavages, cysteine carbamidomethyl static modification and methionine oxidation

and dimethyl labeling (L,M,H) of lysine residues and the peptide N-termini as dynamic modification. For the phosphopeptide experiments, Phosphorylation of serine (S), threonine (T) or tyrosine (Y) was used as an additional dynamic modification and the phosphoRS27 node was used in the workflow to calculate site occupation probabilities. Triplex dimethyl labeling was used as a MS1 XIC quantification method, with a mass precision of 2 ppm for consecutive precursor mass scans and normalization on the median peptide ratio. A retention time tolerance of 0.5 min was used to account for the effect of deuterium on the retention time in long gradients. To filter for high quality protein data and to control the false discovery rate on identifications, only the peptide spectrum matches adhering to the following criteria were kept for analysis: minimal Mascot score of 20, minimal peptide length of 7, no inconsistently labeled peptides (e.g. no light N-terminus and heavy lysine on a single peptide) only unique rank 1 peptides and a peptide mass deviance of 20 ppm or 15 ppm. As a result we obtained peptide false discovery rates of 1% and lower for all experiments. To distinguish between proteins with high similarities, only unique peptides were considered for protein identification and quantification. Proteins with a minimum average 2-fold change were considered regulated. The criteria used for phosphopeptide analysis were a 1% false discovery rate calculated by Percolator28 and an additional minimum Mascot score of 20. Phosphopeptide quantification was performed using an in-house developed script that calculated the average ratio, intensity and localization probability for every non-redundant phosphopeptide. The script mapped all the identified phosphopeptides on their protein keeping doubly/triply phosphorylated peptides together to potentially map differential crosstalk of close-by sites. Motif enrichment analysis was performed by submitting regulated sequences to Motif-x29 using the IPI Human database as the background database and a p value of 0.000001. The set of enriched peptides was submitted subsequently into IceLogo30, using the entire human proteome (Swiss-Prot human) as background and a p value of 0.01, to generate a sequence logo. Protein extraction and immunoblot analysis Cycling, OIS or OIS bypassing cells were harvested 9 days after BRAFV600E infection. Cells were scraped with 1X PBS, centrifuged @ 4000rpm for 4 minutes at 4°C and the pellets were frozen or used immediately. Fresh or frozen pellets were lysed on ice in RIPA buffer supplemented with protease inhibitor cocktail (Roche) and phosphatase inhibitors (10 mM β-glycerophosphate, 2 mM sodium fluoride, 0.2 mM sodium orthovanadate, 1 mM sodium pyrophosphate). Lysates were sonicated for 1 min (5 sec on/off interval), centrifuged at 4°C and 1200rpm for 10 min and supernatants were transferred to fresh Eppendorf tubes. Protein concentrations were determined using Bradford assay (Bio-Rad). Protein samples were prepared in 4X sample buffer (Invitrogen) supplemented with 2.5% β-mercaptoethanol. Proteins were separated on 4-12% polyacrylamide gels (Invitrogen), transferred onto a nitrocellulose membrane (Whatman) and blocked in blocking buffer (4% milk in 1X TBS-

RESULTS Model system To study the mechanisms underlying OIS we employed a previously described cell system14 comprising in vitro cultured HDFs in three differential states: cycling, OIS and OIS bypass (OISb). HDF transduced with retrovirus harboring an empty vector were used as normal cycling control cells. OIS was induced by infecting HDF with a retrovirus carrying the constitutively active BRAF gene. OIS was abrogated in the presence of mutated BRAF by depletion of C/EBPβ, a crucial component of the inflammatory pathway in OIS14. HDF cells were chosen as the model system because they are easily grown in the quantities required for proteomics experiments, and previous studies have shown that all observed factors crucial for OIS in HDF validate in primary melanocytes or in vivo models14,31. Cells were additionally modified for ectopic hTERT expression and p16 silencing to prevent the stochastic induction of replicative senescence producing heterogeneous cell populations. Cells expressing the BRAFV600E oncoprotein typically showed an initial burst of growth followed by the induction of senescence. Three days after infection, all cells showed equal proliferative kinetics (pre-OIS; Figure 1a). However, after 9 days, cells expressing BRAFV600E displayed a strong growth arrest accompanied by the induction of the senescence marker SA-β-Gal, whereas the cycling control cells continued proliferation without displaying this senescence marker (Figure 1b; cycling and BRAFV600E/vector). As expected, the C/EBPβ-depleted OISb cells, in contrast to OIS cells but similarly to cycling cells, failed to undergo cell cycle arrest or induce senescence markers by 9 days following BRAFV600E introduction (Figures 1a, b; BRAFV600E/ sh-C/EBPβ). Therefore, we hypothesized that the comparison of cycling, OIS, and OISb samples in the early onset and end-point phenotypes would allow us to find proteins and phosphosites regulated specifically in the onset and maintenance of senescence. Moreover,

125

Chapter 6

Tween) for 1h at room temperature. The membrane was probed with the indicated primary antibodies (overnight at 4°C in 4% milk in 1X TBS-Tween), followed by 1 h incubation with the corresponding secondary antibodies conjugated with horseradish peroxidase (HRP) enzyme. For detection of the signal the membrane was incubated 1 minute with ECL reagent (Amersham Biosciences) and visualized on films (GE Healthcare). Antibodies The primary antibodies used for immunoblot analysis were C2CD2 (D01P; H00025966; MaxPab), phospho-cdc2 (known as CDK1; Tyr15; #9111; Cell Signaling), cdc2 (known as CDK1; #9112; Cell Signaling), FGF-2 (C-18; sc-1360; Santa Cruz), Hsp90 (#4874; Cell Signaling), phospho-STAT-3 (Tyr705; 3E2; #9138; Cell Signaling), STAT-3 (sc-482; Santa Cruz). Secondary antibodies used were goat anti-mouse IgG (H+L) HRP conjugated (G21040; Invitrogen), goat anti-rabbit IgG (H+L) HRP conjugated (G21234; Invitrogen) and rabbit anti-goat IgG (H+L) HRP conjugated (R-21459; Invitrogen).

shvector C/EBPβ

vector

shC/EBPβ

3 days

cycling

Con

OIS

OISb

III

15E3

12E3

9E3

6E3

16E3 12E3 8E3 4E3

3E3 0

Multi- Loc. plicity Prob.

Amino Acid

90-180 min nLC-MS/MS Orbitrap Velos HCD

x 22

Lys-C & Tryp

18E3

Peptide SCX

Lysis red. alk.

Phosphopeptide SCX Dimethyl Labeling

Ti4+-IMAC

Ti4+

x 22

Ti4+

Ti4+ Ti4+ Ti

Ti4+

Ti4+ Ti4+

Ti4+ Ti4+ Ti4+

Ti4+

pT Pulldown 180 min nLC-MS/MS Orbitrap Velos HCD/ETD

pY, pS, pT Site Quan

BRAFV600E

9 days

3 days

cycling

Protein Quan

BRAFV600E

Σ # Unique phosphosites quantified

Σ # Unique phosphopeptides quantified

the comparison between OISb and OIS cells could potentially reveal processes involved in the bypass of senescence into tumor progression/malignancy.

Figure 1. Model system and approach Cell proliferation assay (a) and SA-β-Gal activity (senescence marker) assay (b) of HDF cells transduced with empty vector, BRAFV600E or both BRAFV600E and a shRNA targeting C/EBPβ that were subsequently fixed and stained 3 and 9 days after induction of BRAFV600E. Senescence is evident only 9 days after BRAFV600E induction (OIS), but not in cells with a C/EBPβ depletion (OISb) c. Proteomics workflow optimized for the quantification of proteins and phosphopeptides. d. Number of unique phosphopeptides quantified. Distributions of phosphopeptide multiplicity, site localisation probability (Class I:≥95%, II:≥ 50% and III:<50%) and type of residue phosphorylated.

Approach To uncover protein expression specific for cycling, OIS or OISb, related lysates were digested and resulting peptides were desalted and labeled with different dimethyl isotopes. After mixing the three labeled cell lysates, peptides were processed through three distinct pipelines. For full proteome analysis, peptides were fractionated by SCX and for unbiased global phosphopeptide analysis an additional SCX was ran followed by Ti4+-IMAC enrichment as described before18. For the specific analysis of tyrosine phosphorylation dynamics the protein digest was directly treated with phosphotyrosine (pY) antibody coupled beads (Figure 1c). The workflow was applied to cells harvested 3 days and 9 days post-transduction and the replicate experiment was performed using a label swap. The first experiment was focused on the analysis of unmodified peptides to assess protein identity and protein expression differences and resulted, following stringent filtering as described in “Experimental Procedures,” in the robust quantification of 5,997 proteins. The

126

extensive fractionation before the highly selective Ti4+-IMAC phosphopeptide enrichment together with the pY immunoprecipitation resulted in the quantification of 18,320 unique phosphopeptides (15,498 unique phosphosites) with a localization certainty of on average 90% (Figure 1d). Because of the different types of data distributions for different sample comparisons, we used a threshold of 2-fold up- or down-regulated in both replicates to pinpoint differentially regulated proteins/phosphopeptides. The full lists of all quantified proteins and phosphopeptides can be found in Supplementary Tables S1 and S2, and the raw files have been made available as Supplementary Data. The ratios determined in our three parallel experiments were used to define two sets of biologically distinct changes. The first set of regulated events was so-called BRAF specific and was observed when comparing both OIS cells and OISb cells to cycling cells. The second, and most interesting, scenario was termed OIS specific and consisted of factors up- or downregulated in comparisons of OIS versus cycling and OIS versus OISb. In order to find consistent regulations, we plotted the logarithmic base 2 ratio for both replicates in each sample comparison in Figure 2. 3210

3827

3 days - Presenescence

Proteins

6 ↑43 BRAF

CDKN1A

Phosphosites

-6

-4

-2

-6

-2

-6

ATM_T1885

NF1_S2543

CD44_S697

SEC22B_S137

-4

CD44_S697

ICAM1_T530

↓34 -6

-2

BAZ1B_S1342

RB1_T356

EPHB2_S776 CD44_S697 SEC22B_S137

EPHB2_S776

ICAM1_T530 SEC22B_S137

BRAF_S729

CDK1_Y15 CDK1_Y15

-4

↑77

BRAF_S729

CD44_S697 BAZ1B_S1342 ATM_T1885

SEC22B_S137

ATM_T1885

EPHB2_S776

BRAF

C2CD2

↓228 -2

↑328

ICAM1_T530

EPHB2_T590

CDK1_Y15

EPHB2_T590

CDK1_Y15

ATM_T1885

COIL_T303 RB1_T356

COIL_T303

RB1_S249,T252

-4 -6

-4

TFPI2

IFI16

TIMP3 USP36

COL1A2 COL1A1

↓2 ↑25

↑164

COL1A2 CDK1

USP36

↓50

FGF2

CDKN1A

RB1

CDK1

-6

RNF13

COL1A1

CDKN1A

RB1

COL1A2 COL1A1

-4

TIMP3

CDKN1A

IFI16

FGF2

-2

COL1A1

IL1B

ICAM1

ICAM1 BRAF

C2CD2

FGF2

COL1A2 BRAF

ICAM1

↑68

TFPI2

FGF2 RNF13

ICAM1

IFI16

-2

IFI16

TFPI2

↑159

↑14

3905

9 days - Senescence

Chapter 6

3180

RB1_S249,T252

-6

-4

-2

↓303 0

Log2 preOIS/Cycl

↓16 -6

-4

-2

RB1_T356

COIL_T303

Log2 preOIS/preOISb

RB1_T356

COIL_T303

-6

-4

-2

↓576 2

Log2 OIS/Cycl

↓111 -6

-4

-2

Log2 OIS/OISb

Figure 2. Proteins and phosphosites comparisons between different conditions a. Protein ratios determined in both replicates highlighting consistently regulated proteins (replicate one plotted versus replicate two); in total 5,997 proteins were quantified. b. Ratio plots of unique phosphopeptides quantified in both Ti4+-IMAC phosphopeptide enrichments and pY immunoprecipitation replicates. Down- or upregulated proteins/phosphopeptides (>2-fold) are highlighted with red and green diamonds, respectively. Proteins and phosphosites regulated specifically in OIS (and not in OISb) are highlighted with blue squares.

Proteome changes upon BRAF expression As displayed in the OIS versus OISb protein ratio plots, minor changes were observed after 3 days that became more apparent after 9 days when the full phenotypes were displayed (16 and 102 substantially regulated proteins after 3 and 9 days, respectively; Figure 2, Proteins; Supplementary Figure S1). Interestingly, larger numbers of differentially regulated proteins

127

were observed for the OIS versus cycling and OISb versus cycling cells, and those numbers increased over time (93 and 387 proteins regulated at 3 and 9 days, respectively; Figure 2, Proteins; Supplementary Figure S1). Although the OIS cells were cell cycle arrested after 9 days and both the OISb and cycling cells were still growing, substantially greater similarity in protein expression was observed between OISb and OIS cells than between OISb and cycling cells. This illustrates that oncogene activation has clearly a dominant effect on global protein expression levels. In line with expectations, BRAF-expressing cells displayed similarly elevated levels of BRAF protein expression after 3 and 9 days of growth, suggesting equal transduction levels (Figure 2, Proteins OIS/OISb; Supplementary Figure S1). Ontology enrichment analysis revealed that proteins concurrently downregulated in BRAFexpressing cells (OIS and OISb, 3 days and 9 days) are primarily involved in extracellular matrix (ECM) interactions (Figure 3a). ECM constituents such as collagens (I, III, VI, XII), fibronectin, cytoskeletal proteins filamin A and filamin C were among the most strongly downregulated proteins in both OIS and OISb (Figure 2; Supplementary Table S1). Interestingly, there is a difference between the early and late stages of oncogenic insult for some downregulated protein categories. DNA replication and cell cycle regulatory proteins were differentially enriched in BRAF-expressing cells at day 3, reflecting the early growth stimulatory response, whereas cell contact gene ontologies were specifically enriched in the late stage. The only significantly enriched protein group upregulated in both OIS and OISb belonged to the lysosome compartment. When looking more specifically into the data, it is interesting to note that IFI16 was upregulated in all BRAF-expressing cells after 3 and 9 days (Figure 2; Supplementary Figure S1). It has been shown previously that IFI16 is able to induce senescence via up-regulation of p21 (CDKN1A) and that subsequent p21 knockdown can bypass senescence32. In our model we saw p21 upregulation in both OIS and OISb, but only after 9 days, suggesting a delayed tumor-suppressive response to oncogene activation. Moreover, in line with the upregulation of p21, CDK1 protein expression levels showed downregulation after 9 days in OIS that seemed to be less strong in OISb (Figure 2; Figure 4b; Supplementary Figure S1). Phosphoproteome changes upon BRAF expression Similar to the proteome changes, the phosphoproteome changes were found to be the largest when we compared either OIS or OISb cells to cycling cells (Figure 2; Supplementary Figure S2). In line with an increase of BRAF protein expression, we also observed a strong increase in BRAF S729 phosphorylation in OIS and OISb cells. At the same time, phosphorylation on NF1 (S2543, T2565, S2802) and sprouty (related) proteins (Spry2 S167, Spry4 S125, Spred1 S238, and Spred2 S168) that are known to negatively affect BRAF activation were similarly upregulated after BRAF oncogene introduction. This is in line with a previous study by Courtois-Cox et al. demonstrating a negative feedback signaling loop involved in OIS33.

128

Log2 Enrichment factor

-4

-2

7 13 5 10 9 5

15 23

9 5

22 16 11 5

14 25

4 8

IL1B IL8

OIS OISb

6 7

Senescence Specific

FGF2 MMP1 SERPINB2 DPP4 TFPI2 TIMP1 DKK3 AGA CTSS RNF13 ACP2 PSAP GAPDH GLB1 GAA PGM2L1 GK

11 8 8 7 7 6

SOD2 ICAM1 THBD CEBPB PTGS2

Cycl OIS OISb

TSC22D1 CDR2L AKR1B1 CD55 ACO1 SELH AUH NAMPT SQRDL GPX1 GFPT2 HPCAL1 KYNU

PLIN2 NMES1 UPP1 ASNS TRHDE FZD7 CNTNAP1 FAM129A CNN2 CNN3

UGCG SPATS2L ANKRD13A USP36 TRIM16 HMOX1 DNAJB4 DOCK10 TJP2

Cycl

3 Days Cycl

OIS specific

Cycl

Extracellular region part Cell migration Secreted Lysosome Resp. to lipopolysaccharide Carbohydrate catabolic process

11 8

BRAF specific: (OIS & OISb)

Regulation of cell cycle DNA replication MCM complex Extracellular region part ECM organization ECM-receptor interaction ECM-structural constituent Collagen fibril organization Heparin binding VWC-domain Cell adhesion Cell motility Actin-binding LamininG2-domain Secreted Prot. digestion and absorption Resp. to amino acid stimulus Lysosome

OIS OISb

PTGS2 DPP4 ACO1 SQRDL SOD2 AUH ACP2 PLIN2 GK TIMP1 TFPI2 SERPINB2 THBD HPCAL1 TSC22D1 PGM2L1 SERPINB8 CD55 AGA AKR1B1 FGF2 GLB1

pre OIS

preOISb

9 Days Cycl

OIS

OISb

PSAP

GAA GAPDH NAMPT ICAM1 TNC TNFRSF10D MMP1 GFPT2 DOCK10 UGCG DKK3 TRHDE TJP2 DNAJB4 SPATS2L CNN3 FZD7 ASNS HMOX1 CNN2 ANKRD13A DSP

Z-Score -1.5

-0.5

0.5

1.5

FGF2 C2CD2

25kD

yc l

IS O IS b

Figure 3. Senescence signature proteome a. Functional classification of the proteins in the OIS and OISb proteome by GO categories (BenjaminHochberg corrected FDR<0.02). The number of proteins in each category is depicted next to the red and green bars indicating down- and upregulated categories, respectively. b. Hierarchical clustering of normalized average intensities displayed as a heat map for proteins specifically regulated in pre-OIS and/or OIS (3 and/ or 9 days post BRAFV600E expression); only proteins quantified in all replicates of all samples were plotted. c. Normalized average intensities displayed as a heat map of proteins specifically regulated after 9 days of BRAFV600E expression. Functional categories were colored as followed: secreted proteins (red), lysosomal (purple), LPS response (green), carbohydrate processing (cyan). Proteins present in multiple categories were colored only according to their first category.

20kD

CDK1 pY15

15kD

CDK1 STAT3 pY705

Hsp90

STAT3 BRAF : CDK concensus motif

56-fold enrichment (n=38)

Figure 4. Immunoblot validation of MS data and motif analysis a. Immunoblot probed with antibodies for FGF2 and C2CD2 confirm their regulation observed in the proteomics data (Hsp90 was used as a loading control). b. CDK1 and STAT3 phosphorylation and total protein levels confirm a strong reduction of phosphorylation in OIS and a smaller reduction in the OIS b cells. c. Motif-X sequence motif analysis of phosphopeptides downregulated in OIS and OISb, indicating a specific inhibition of CDK kinase activity in BRAF expressing cells.

129

Chapter 6

-6

In addition, up-regulation of ATM T1885 phosphorylation was observed, although ATM protein levels remained unchanged (Figure 2; Supplementary Figure S2). Although biological information about this specific site is absent, it could be potentially interesting, as the ATM stress kinase is known to phosphorylate important components of the stress response such as p5334. As described above, the increase in protein levels of the CDK inhibitor p21 was observed in both OIS and OISb only after 9 days. Therefore, we searched for the effects of p21 upregulation on phosphorylation levels of CDK targets. Our global phosphorylation screen revealed a strong reduction in CDK-dependent phosphorylation sites, as observed in the sequence motif analysis (Figure 4c). Relative to background, the CDK phosphorylation motif (SPxK) was strongly and solely enriched in the set of phosphosites downregulated in both BRAF-expressing cells after 9 days (56-fold enrichment). When looking at individual CDK targets, RB1 is one of the major Cyclin/CDK substrate proteins controlling cell cycle progression; RB1 hypophosphorylation is associated with transcriptional repression, whereas hyperphosphorylation relieves its tumor-suppressive role. In line with this, strong down-regulation of RB1 phosphorylation sites in senescence was observed, and the proteome screening indicated unchanged RB1 protein levels. After 3 days of growth, phosphorylated T356 and the dual phosphorylated sites S249, T252 and T821, T826 were downregulated in the pre-OIS cells only (Figure 2; Supplementary Figure S2). After 9 days, phosphorylation of both dual phosphorylation sites was nearly absent in both OIS and OISb but still present in cycling cells (Supplementary Figure S3). Furthermore, RB1 T356 was found to be downregulated in OIS and, to a lesser extent, in OISb (Figure 2B; Supplementary Figure S2). These changes in phosphorylation suggest that RB1 may play a partial role in the mutant BRAF-induced proliferative arrest. Another CDK1 substrate is histone acetyltransferase Myst2 T88, which was strongly downregulated in the OIS and OISb cells, whereas protein levels remained unregulated. Phosphorylation of T88 on Myst2 by CDK1 can be crucial for cell cycle progression35. Similar, another CDK1 substrate, S38 on Stathmin1 (Stmn1)36, showed downregulation of phosphorylation in both OIS and OISb. In summary, these findings indicate a strong tumor-suppressive response to mutant BRAF expression in this model system, involving CDK inhibition, leading to a potent cell cycle arrest. However, although all BRAF-expressing cells display greater sensitivity for cell cycle arrest, only the cells with intact C/EBPÎ˛ levels have the ability to enter a full cell cycle arrest. These senescence-specific events were studied next. Proteome changes specific for Senescence Even though the proteomes of (pre-) OIS and (pre-) OISb are very much alike at a global scale, the former cells are cell cycle arrested whereas OISb keep proliferating. This implies there are additional molecular mechanisms that distinguish OIS cells from OISb cells. Therefore,

130

131

Chapter 6

we focused on a small set of proteins that also displayed differential expression between OIS and OISb in the early and full OIS stages (i.e. 3 and 9 days), thereby discriminating between markers of senescence onset and maintenance. In Figure 3 an overview is given of the senescence-specific proteins and their are clustering into gene ontology (GO) groups. Upregulated enriched gene ontologies include mainly lysosomal, inflammatory, and ECM proteins (Figure 3a). To visualize the dynamics of senescence-associated proteins, we clustered the intensities for proteins regulated in OIS cells after 3 days and/or 9 days in Figure 3b. Interestingly, some ECM processing proteins such as SerpinB2/PAI-2 and TFPI2 were regulated after both 3 and 9 days, whereas other proteins belonging to the same category were exclusively upregulated after 3 days (SerpinB8) or 9 days (MMP1, TIMP1). This temporal regulation could indicate that different phases of extracellular modifications are required for the cell to progress into a senescence-like state. Consistent with our previous results on mRNA profiling14, strong upregulation was found for the cytokines IL-1b and IL-8 in senescent cell lines at day 9. At the pre-OIS stage, IL-1b was already clearly visible in both BRAF-expressing cells, whereas IL-8 was identified in one replicate only. Additional proteins specifically regulated in the final senescent phenotype (Figure 3c) included CEBP/Î˛ and TSC22D1, which we have shown previously to stimulate the production of cytokines such as IL-1b and IL-8 and to be crucial in maintaining the senescence phenotype in OIS14,37,38. Similarly, PTGS2 (COX2), the main enzyme in inflammation-linked prostaglandin production, was found to be upregulated in OIS in our data and was shown previously to be critical for senescence onset and maintenance39. Other upregulated proteins included glucose-linked biomolecule-processing enzymes (i.e. AGA, GFPT2) and lysosomal (i.e. GAA, GLB1/Î˛- galactosidase) proteins. Another factor strongly upregulated specifically in OIS was the fibroblast growth factor FGF2. The finding of this growth-promoting factor was at first sight counterintuitive, but immunoblot analysis validated a strong presence of both the short excreted para/autocrine FGF2 and the longer NLS-containing intracrine FGF2 variant of 18 kDa and 24 kDa, respectively (Figure 4a). Interestingly, FGF2 was shown previously in oncogenic RAS expressing cells to induce an irreversible cell cycle arrest40 One of the most strongly downregulated proteins in senescence was the deubiquitinating enzyme USP36 (Figure 2; Figure 3c). Interestingly, this enzyme was previously shown to positively control cell growth by stabilizing rRNA production and ribosome biogenesis41. Moreover, knockdown of both yeast and fly USP36 homologs was shown to result in an inhibition of cell growth41,42. Although further biological validation is required, in light of our model system, USP36 can potentially provide a novel link between deubiquitination and growth arrest in OIS. Conversely, we found the ubiquitin E3 ligase (RNF13) to be upregulated in OIS, suggesting that both USP36 and RNF13 proteins play an important role in ubiquitinmediated proteasomal degradation in OIS.

In summary, the protein expression analysis not only revealed a lot of different processes regulated in OIS previously shown to be directly or indirectly linked to growth arrest, but also gave a hint of novel mechanisms yet to be further explored in the context of senescence. Phosphoproteome changes specific for Senescence Among the most interesting phosphorylation events specifically regulated in OIS are phosphosites on Coilin/p80 (COIL), STAT3, p53, WSTF/BAZ1B, ephrin type-B receptor 2, and ECMR-III/CD44. For COIL, total protein levels remained unchanged, but its phosphorylation at T303 was downregulated in pre-OIS after only 3 days, and this was even more strongly and specifically pronounced in OIS after 9 days. Previously, COIL phosphorylation levels were shown to increase in mitosis43, suggesting a role of COIL phosphorylation in growth inhibition in OIS. Signal transducer and transcription activator 3 (STAT3) is a well-known mediator of cellular responses activated by several growth factors and cytokines. Y-phosphorylated STAT3 dimerizes to form an activated signal transducer that has been shown to be an important mediator in a variety of human cancers44,45. Moreover, it is interesting to note that inhibition of STAT3 activity also leads to the induction of pro-inflammatory cytokine expression, including IL-646. In our study, we indeed observed downregulation of STAT3 Y705 phosphorylation specifically in OIS, whereas its protein levels remained unchanged (Figure 4b; Supplementary Figure S4). This suggests that STAT3 may be an important mediator of OIS that acts by reducing growth and inducing the inflammatory response needed to maintain the OIS signature, which needs to be investigated further. Also phosphorylation of p53 was found specifically regulated in senescence (Supplementary Figure S5). At day 9, S315 of p53 was downregulated in OIS cells when compared to cycling cells, but not in OISb cells. Notably, p53 is a major tumor suppressor with established role in senescence47. Other phosphorylated proteins with sites specifically upregulated in OIS included WSTF/ BAZ1B, which is a component of the chromatin remodeling complex, the transmembrane proteins extracellular matrix receptor III (ECMR-III/CD44) and ephrin type-B receptor 2 that mediate the communication between the ECM and adjacent cells (Figure 2; Supplementary Figure S6). Especially for CD44, we identified many phosphorylation sites that were unchanged (S686, S706, and T720), downregulated in both OIS and OISb (S718), or specifically upregulated in OIS (S697). Ephrin type-B receptor 2 showed enhanced phosphorylation in OIS on T590 and S776, whereas T585 was downregulated in both OIS and OISb. DISCUSSION In this work we report, to date the most comprehensive dataset of protein and phosphosite regulations associated with the onset and maintenance of OIS. Using Ti4+IMAC phosphopeptide enrichment and pY peptide immunoprecipitation, we were able to

132

133

Chapter 6

monitor a vast amount of signaling events. The high specificity and selectivity of Ti4+-IMAC beads and pT antibodies allowed for the detection of important tumor suppressor (i.e. RB1, p53) and signaling proteins (i.e. STAT3, CDK1). This comprehensive high-quality dataset of proteins regulated in OIS and its bypass contains many interesting candidates that warrant further investigation in primary melanocytes and, perhaps, in vivo models to characterize the biological mechanisms involved. We instigated OIS by ectopic expression of the BRAFV600E oncoprotein in human primary fibroblasts. The expression of mutated BRAF led to substantial reprogramming of the proteome and phosphoproteome relative to cycling cells, both in the final OIS stage (9 days post-transfection) and in the early stage in which senescence was not yet apparent (3 days post-transfection) (Figure 2). However, not all the observed changes can be attributed to the onset of senescence, as became evident from our analysis of OISb cells, in which we introduced the mutant BRAF oncogene but rescued the cells from senescence with shRNA-mediated depletion of the proinflammatory transcription factor C/EBPÎ˛. The (phospho)proteome data also reveal that substantially less reprogramming of the proteome took place when comparing OIS to OISb cells, after both 3 and 9 days (Figure 2). Thus, BRAFV600E expression per se leads to substantial reprogramming of the proteome and phosphoproteome, and senescence-specific changes are less numerous, indicating that senescence is a tightly regulated mechanism that can be altered by a small number of alterations. That senescence is indeed abrogated in OISb cells is confirmed by several known senescence biomarkers that are differentially expressed between OIS and OISb, such as the relatively lower levels of interleukins (IL6, IL8) and SA-Î˛-Gal activity. These observations give credibility to our data, in that proteins and phosphosites found differentially regulated between OIS and OISb cells are potentially critical for onset and maintenance of OIS. It has been shown previously that senescence can be induced by inhibition of the cell cycle via the p53-p21 and p16INK4A-RB1 pathway48-50. We also found that BRAFV600E expression resulted in RB1 hypophosphorylation, upregulation of p21, and repression of CDK signaling. When CDKs are activated by cyclins, their t-loop is repositioned, exposing the active site of CDK. An essential step in this activation is the phosphorylation of Y15 by Wee1 and Myt151,52. Interestingly, CDK1/2/3 Y15 phosphorylation was downregulated strongly in OIS and to a lesser extent in OISb (Figure 2b; Figure 4b). Thus, the downregulation of CDK phosphorylation in OIS and OISb correlates well with the differentially elevated p21 levels, inhibiting CDKcyclin complex formation. However, we found that this mechanism was not specific for OIS, as p21 levels were enhanced and CDK protein and CDK target phosphorylation levels were repressed in both OIS and OISb cells (albeit to lesser extent in OISb). This suggests that these tumor-suppressive signals were not strong enough in the OISb cells to evoke a full cell cycle arrest. However, we did note other differentially regulated proteins and sites that can be either directly or indirectly linked to OIS onset and maintenance. A summary of these OISspecific events is displayed in Figure 5.

A factor strongly and specifically present in OIS was FGF2, which was shown previously in oncogenic RAS-expressing cells to induce an irreversible cell cycle arrest that could be negated by a tyrosine kinase inhibiton40. Despite the potentially different partners of RAS and RAF in OIS and malignant transformation, a similar regulation of FGF2 in our and this study calls for a more thorough investigation of a potential common mechanism. Interestingly, the phosphorylation of the tumor suppressor protein p53 at S315 was specifically downregulated in OIS cells. S315 site was previously shown to be phosphorylated by two different kinases. CDK inhibition reduces S315 phosphorylation levels and the p53 S315A mutant shows reduced transcriptional activity, resulting in lowered expression levels of p21 in cancer cell lines53. However, in our data similar levels of p21 were observed in OIS and OISb, and in contrast to the observation in OIS, S315 phosphorylation was not downregulated in OISb. This suggests that in our model p21 levels are independent of p53 S315 phosphorylation. The second kinase reported to phosphorylate p53 S315 is Aurora kinase A. Aurora A-mediated phosphorylation of S315 leads to p53 ubiquitination and subsequent protein degradation54. Therefore, reduced levels of S315 would suggest a greater stability of p53 in OIS cells than in cycling and OISb cells, potentially leading to a prolonged p53 effect in OIS. Cytokines / GF

ECM

ECM regulators

Collagen I, III, V, VI, XII Fibronectin

FGF2

TIMP3 DKK3, PAI-1 TFPI2, MMP1, TIMP1 PAI-2

IL8 IL1β

E P

IL6

H B

pT590 pS776 2

pY705

p53 pS315

P P

Lysosome

β-Gal,GAA, AGA, CTSS, ACP2, PSAP, RNF13 PLAU, CTSB, CTSL1

Cyclin CDK pY15 RB1

STMN1 pS38

pT821, T826 pT373 pT356 pS249, T252

MYST2 pT88

...

COIL pT303 pY

CEBPβ

BAZ1B pS1342

TSC22D1

OIS

I C A M pT530 1

B D

p16

p21

PTGS2

T H

pS106 IFI16

STAT3

D 4

pS686 pS697 4 pS706 pS718 4 pT720

OIS

H2AX

BRAF BRAF

Figure 5. Phosphosites and proteins regulated in senescence Up- and downregulated features, specific for OIS, are depicted in light green and light red, respectively. Proteins and phosphosites regulated by BRAF transformation (regulated in OIS and OISb) are depicted in dark green and dark red, respectively. Unregulated and unquantified proteins are indicated in black and gray, respectively. Gray lines correspond to interactions described in literature.

A potentially interesting phosphosite found upregulated in OIS was S1342 of the tyrosine protein kinase WSTF (or BAZ1B), whose total protein levels remained similar in all conditions (Figure 2; Supplementary Figure S6). WSTF is a part of the chromatin remodeling complex

134

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository65 with the dataset identifier PXD000522 and PXD000523. ACKNOWLEDGEMENTS This work was in part funded by the European Union 7th Framework Programme, PRIME-XS project Grant Agreement Number 262067, a VICI grant from the Netherlands Organization for Scientific Research (NWO) and a Queen Wilhelmina Award grant from the Dutch Cancer

135

Chapter 6

WICH55 and was shown to tyrosine phosphorylate histone H2AX, thereby regulating the formation of Îł-H2AX foci56. The appearance of Îł-H2AX foci is one of the hallmarks of senescent cells, and therefore specific upregulation of phosphorylated WSTF could provide a novel link between foci formation and OIS. Our study revealed that many proteins and sites specifically regulated in OIS are involved in the ECM compartment that is known to play an important role in autocrine and paracrine cell signaling. The ECM can function as a sequestering binding site or a pool of growth factors that can be released upon ECM degradation57. This is also the case for the ECM components heparin and heparan sulfate proteoglycans that can bind FGF2, leading to enhanced receptor tyrosine kinase signaling58,59. Interestingly, proteins specifically regulated in the senescence phenotype contain ECM interacting (i.e. FGF2, CD44), ECM protein cleaving (i.e. MMP1, DPP4, THBD), ECM cleavage inhibitors (i.e. TIMP3, TFPI2, SerpinB2/PAI-2), glycan processing (i.e. AGA, GLB1/b-galactosidase, PGM2L1), and membrane lipid processing proteins (i.e. GAA, PSAP). Although CD44, ephrin type-B receptor 2, and ICAM were extensively and differentially phosphorylated in OIS, not much is known about the effect of the specific phosphorylation sites on these ECM proteins. CD44 is known to bind to the glycosaminoglycan and ECM component hyaluronic acid and interact with matrix metalloproteinases60 that we found to be regulated in OIS as well. Moreover, CD44 was identified as an important cancer stem cell marker in several different cancers61,62. Therefore, the role of CD44 in cell-ECM interaction and its differential phosphorylation status in senescence and cancer might be interesting to functionally validate. It is clear that a substantial amount of OIS-regulated proteins are acting on the extracellular space and matrix. Consistent with our data, other studies have shown that extracellular communication through protein secretion is the driving force behind the OIS phenotype63,64. Therefore, functional studies on ECM regulation, organization, and ECMreceptor and ECM-growth factor interactions could further increase our understanding of OIS and cancer-related cell signaling in general. In summary, this large dataset describes differential protein and phosphorylation changes upon oncogene transduction and OIS, which sets the stage for function-based analysis of potentially novel tumor-suppressive mechanisms linked to senescence and cancer.

Society (KWF Kankerbestrijding) to D.S.P. The Netherlands Proteomics Center, embedded in The Netherlands Genomics Initiative, as well as the Netherlands Organization for Scientific Research (NWO) with the VIDI grant for A.F.M.A (723.012.102) is acknowledged for funding. REFERENCES 1. 2. 3.

7. 8. 9. 10. 11. 12.

13.

136

Cotter, T. G. Apoptosis and cancer: the genesis of a research field. Nat Rev Cancer 9, 501–507 (2009). HAYFLICK, L. THE LIMITED IN VITRO LIFETIME OF HUMAN DIPLOID CELL STRAINS. Exp Cell Res 37, 614–636 (1965). Deng, Y., Chan, S. S. & Chang, S. Telomere dysfunction and tumour suppression: the senescence connection. Nat Rev Cancer 8, 450–458 (2008). Serrano, M., Lin, A. W., McCurrach, M. E., Beach, D. & Lowe, S. W. Oncogenic ras provokes premature cell senescence associated with accumulation of p53 and p16INK4a. Cell 88, 593–602 (1997). Zhu, J., Woods, D., McMahon, M. & Bishop, J. M. Senescence of human fibroblasts induced by oncogenic Raf. Genes Dev 12, 2997–3007 (1998). Dimri, G. P., Itahana, K., Acosta, M. & Campisi, J. Regulation of a senescence checkpoint response by the E2F1 transcription factor and p14(ARF) tumor suppressor. Mol Cell Biol 20, 273–285 (2000). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). Collado, M. & Serrano, M. Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51–57 (2010). Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417, 949–954 (2002). Pollock, P. M. et al. High frequency of BRAF mutations in nevi. Nat Genet 33, 19–20 (2003). Dankort, D. et al. A new mouse model to explore the initiation, progression, and therapy of BRAFV600E-induced lung tumors. Genes Dev 21, 379–384 (2007). Dhomen, N. et al. Oncogenic Braf induces melanocyte senescence and melanoma in mice. Cancer Cell 15, 294–303 (2009).

14.

15.

16.

17.

18.

19.

20.

21.

22.

Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). Altelaar, A. F. M., Munoz, J. & Heck, A. J. R. Next-generation proteomics: towards an integrative view of proteome dynamics. 14, 35–48 (2013). de Graaf, E. L. et al. Spatio-temporal analysis of molecular determinants of neuronal degeneration in the aging mouse cerebellum. Mol. Cell Proteomics 12, 1350–1362 (2013). Gauci, S. et al. Lys-N and trypsin cover complementary parts of the phosphoproteome in a refined SCX-based approach. Anal Chem 81, 4493–4501 (2009). Zhou, H. et al. Enhancing the identification of phosphopeptides from putative basophilic kinase substrates using Ti (IV) based IMAC enrichment. Mol. Cell Proteomics 10, M110.006452–M110.006452 (2011). Boersema, P. J., Raijmakers, R., Lemeer, S., Mohammed, S. & Heck, A. J. R. Multiplex peptide stable isotope dimethyl labeling for quantitative proteomics. Nature protocols 4, 484–494 (2009). Zoumaro-Djayoon, A. D., Heck, A. J. R. & Munoz, J. Targeted analysis of tyrosine phosphorylation by immuno-affinity enrichment of tyrosine phosphorylated peptides prior to mass spectrometric analysis. Methods 56, 268–274 (2012). Di Palma, S. et al. Finding the same needles in the haystack? A comparison of phosphotyrosine peptides enriched by immuno-affinity precipitation and metalbased affinity chromatography. J Proteomics 91, 331–337 (2013). Gobom, J., Nordhoff, E., Mirgorodskaya, E., Ekman, R. & Roepstorff, P. Sample purification and preparation technique based on nano-scale reversed-phase columns for the sensitive analysis of complex peptide mixtures by matrix-assisted laser desorption/ ionization mass spectrometry. J Mass Spectrom 34, 105–116 (1999).

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Helbig, A. O. et al. Profiling of N-acetylated protein termini provides in-depth insights into the N-terminal nature of the proteome. Mol. Cell Proteomics 9, 928–939 (2010). Pinkse, M. W. H. et al. Highly robust, automated, and sensitive online TiO2based phosphoproteomics applied to study endogenous phosphorylation in Drosophila melanogaster. J. Proteome Res. 7, 687–697 (2008). Zhou, H. et al. Specific phosphopeptide enrichment with immobilized titanium ion affinity chromatography adsorbent for phosphoproteome analysis. J. Proteome Res. 7, 3957–3967 (2008). Frese, C. K. et al. Improved peptide identification by targeted fragmentation using CID, HCD and ETD on an LTQ-Orbitrap Velos. J. Proteome Res. 10, 2377–2388 (2011). Taus, T. et al. Universal and confident phosphorylation site localization using phosphoRS. J. Proteome Res. 10, 5354–5362 (2011). Käll, L., Canterbury, J. D., Weston, J., Noble, W. S. & MacCoss, M. J. Semi-supervised learning for peptide identification from shotgun proteomics datasets. Nat Methods 4, 923–925 (2007). Schwartz, D. & Gygi, S. P. An iterative statistical approach to the identification of protein phosphorylation motifs from largescale data sets. Nat. Biotechnol. 23, 1391– 1398 (2005). Colaert, N., Helsens, K., Martens, L., Vandekerckhove, J. & Gevaert, K. Improved visualization of protein consensus sequences by iceLogo. Nat Methods 6, 786–787 (2009). Kaplon, J. et al. A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112 (2013). Xin, H., Pereira-Smith, O. M. & Choubey, D. Role of IFI 16 in cellular senescence of human fibroblasts. Oncogene 23, 6209–6217 (2004). Courtois-Cox, S. et al. A negative feedback signaling network underlies oncogeneinduced senescence. Cancer Cell 10, 459– 472 (2006). Shiloh, Y. ATM and related protein kinases: safeguarding genome integrity. Nat Rev Cancer 3, 155–168 (2003).

35.

36.

37.

38. 39.

40.

41.

42.

43.

44.

45.

46.

Wu, Z.-Q. & Liu, X. Role for Plk1 phosphorylation of Hbo1 in regulation of replication licensing. PNAS 105, 1919–1924 (2008). Marklund, U., Brattsand, G., Shingler, V. & Gullberg, M. Serine 25 of oncoprotein 18 is a major cytosolic target for the mitogenactivated protein kinase. J Biol Chem 268, 15039–15047 (1993). Sebastian, T., Malik, R., Thomas, S., Sage, J. & Johnson, P. F. C/EBPbeta cooperates with RB:E2F to implement Ras(V12)-induced cellular senescence. EMBO J 24, 3301–3312 (2005). Hömig-Hölzel, C. et al. Antagonistic TSC22D1 variants control BRAF(E600)-induced senescence. EMBO J 30, 1753–1765 (2011). Martien, S. et al. Cellular senescence involves an intracrine prostaglandin E2 pathway in human fibroblasts. Biochim Biophys Acta 1831, 1217–1227 (2013). Costa, E. T. et al. Fibroblast growth factor 2 restrains Ras-driven proliferation of malignant cells by triggering RhoA-mediated senescence. Cancer Res 68, 6215–6223 (2008). Richardson, L. A. et al. A conserved deubiquitinating enzyme controls cell growth by regulating RNA polymerase I stability. CellReports 2, 372–385 (2012). Taillebourg, E. et al. The deubiquitinating enzyme USP36 controls selective autophagy activation by ubiquitinated proteins. Autophagy 8, 767–779 (2012). Carmo-Fonseca, M., Ferreira, J. & Lamond, A. I. Assembly of snRNP-containing coiled bodies is regulated in interphase and mitosis--evidence that the coiled body is a kinetic nuclear structure. J Cell Biol 120, 841–852 (1993). Alvarez, J. V. et al. Identification of a genetic signature of activated signal transducer and activator of transcription 3 in human tumors. Cancer Res 65, 5054–5062 (2005). Garcia, R. et al. Constitutive activation of Stat3 by the Src and JAK tyrosine kinases participates in growth regulation of human breast carcinoma cells. Oncogene 20, 2499– 2513 (2001). Wang, T. et al. Regulation of the innate and adaptive immune responses by Stat-3 signaling in tumor cells. Nat Med 10, 48–54 (2004).

137

Chapter 6

23.

47. 48.

49.

50.

51. 52.

53.

54.

55.

56. 57.

58.

59.

60.

138

Vousden, K. H. & Prives, C. Blinded by the Light: The Growing Complexity of p53. Cell 137, 413–431 (2009). Hara, E. et al. Regulation of p16CDKN2 expression and its implications for cell immortalization and senescence. Mol Cell Biol 16, 859–867 (1996). Shay, J. W., Pereira-Smith, O. M. & Wright, W. E. A role for both RB and p53 in the regulation of human cellular senescence. Exp Cell Res 196, 33–39 (1991). Campisi, J. & d’Adda di Fagagna, F. Cellular senescence: when bad things happen to good cells. Nat Rev Mol Cell Biol 8, 729–740 (2007). Morgan, D. O. Principles of CDK regulation. Nature 374, 131–134 (1995). Mueller, P. R., Coleman, T. R., Kumagai, A. & Dunphy, W. G. Myt1: a membrane-associated inhibitory kinase that phosphorylates Cdc2 on both threonine-14 and tyrosine-15. Science 270, 86–90 (1995). Blaydes, J. P. et al. Stoichiometric phosphorylation of human p53 at Ser315 stimulates p53-dependent transcription. J Biol Chem 276, 4699–4708 (2001). Katayama, H. et al. Phosphorylation by aurora kinase A induces Mdm2-mediated destabilization and inhibition of p53. Nat Genet 36, 55–62 (2004). Bozhenok, L., Wade, P. A. & Varga-Weisz, P. WSTF-ISWI chromatin remodeling complex targets heterochromatic replication foci. EMBO J 21, 2231–2241 (2002). Xiao, A. et al. WSTF regulates the H2A.X DNA damage response via a novel tyrosine kinase activity. Nature 457, 57–62 (2009). Flaumenhaft, R., Moscatelli, D. & Rifkin, D. B. Heparin and heparan sulfate increase the radius of diffusion and action of basic fibroblast growth factor. J Cell Biol 111, 1651–1659 (1990). Moy, F. J. et al. Properly oriented heparindecasaccharide-induced dimers are the biologically active form of basic fibroblast growth factor. Biochemistry 36, 4782–4791 (1997). Rapraeger, A. C., Krufka, A. & Olwin, B. B. Requirement of heparan sulfate for bFGFmediated fibroblast growth and myoblast differentiation. Science 252, 1705–1708 (1991). Marrero-Diaz, R. et al. Polarized MT1-MMPCD44 interaction and CD44 cleavage during

61.

62. 63.

64. 65.

cell retraction reveal an essential role for MT1-MMP in CD44-mediated invasion. Cell Motil. Cytoskeleton 66, 48–61 (2009). Prince, M. E. et al. Identification of a subpopulation of cells with cancer stem cell properties in head and neck squamous cell carcinoma. Proc Natl Acad Sci USA 104, 973–978 (2007). Li, C. et al. Identification of pancreatic cancer stem cells. Cancer Res 67, 1030–1037 (2007). Krtolica, A. & Campisi, J. Cancer and aging: a model for the cancer promoting effects of the aging stroma. Int. J. Biochem. Cell Biol. 34, 1401–1414 (2002). Kuilman, T. & Peeper, D. S. Senescencemessaging secretome: SMS-ing cellular stress. Nat Rev Cancer 9, 81–94 (2009). Vizcaíno, J. A. et al. The PRoteomics IDEntifications (PRIDE) database and associated tools: status in 2013. Nucleic Acids Res. 41, D1063–9 (2013).

SUPPLEMENTARY DATA a

Supplementary Figure S1. Protein expression comparisons for cycling (Cycl), BRAF transduced (OIS) and BRAF/sh-C/EBPÎ˛ transduced (OISb) human diploid fibroblasts after 3 days (a) and 9 days (b) of cell growth. The logarithmic base 2 ratio for two conditions of replicate 1 was plotted against the logarithmic base 2 ratio for the same two conditions of replicate 2 to visualize consistent changes in replicates.

Chapter 6

Supplementary Figure S2. Phosphosites regulated specifically in BRAF expressing cells (OIS and OISb). Phosphosite abundance comparisons for cycling (Cycl), BRAF transduced (OIS) and BRAF/sh-C/EBPÎ˛ transduced (OISb) human diploid fibroblasts after 3 days (a) and 9 days (b) of cell growth. The logarithmic base 2 ratio for two conditions of replicate 1 was plotted against the logarithmic base 2 ratio for the same two conditions of replicate 2 to visualize consistent changes in replicates.

139

RB1_S249,T252: TAVIPINGpSPRpTPR

RB1_T821,T826: ISEGLPpTPTKMpTPR

Ion score 54, pRS loc. Prob. S(9):100%,T(12):100%

Ion score 54, pRS loc. Prob. T(7):96%,T(12):100%

RB1_S249,T252: TAVIPINGpSPRpTPR

RB1_T821,T826: ISEGLPpTPTKMpTPR

9d Rep1 (L:CON,M:OIS,H:OISb):

9d Rep2 (L:CON,M:OISb,H:OIS):

Supplementary Figure S3. RB1 phosphorylation on S249, T252 & T821,T826 in BRAF expressing cell lines a. ETD peptide fragmentation spectra showing confident phosphopeptide identification and phosphosite localization. b. MS1 Quantification spectra with phosphopeptide levels clearly present in cycling samples but absent or near noise levels in BRAF expressing cell lines resulting in phosphopeptides without quantification ratios. a

STAT3_Y705: YCRPESQEHPEADPGSAAPpYLK Ion score 43, pRS loc. Prob. Y(20):100%

STAT3_Y705: YCRPESQEHPEADPGSAAPpY LK 9d pY IP Rep2 (L:CON,M:OISb,H:OIS):

Supplementary Figure S4. STAT3 Y705 phosphorylation downregulated in senescence a. CID peptide fragmentation spectra showing confident phosphopeptide identification and phosphosite localization. b. MS1 Quantification spectra with phosphopeptide levels clearly present in control sample and strongly downreguled in OIS. The phosphopeptide was only sequenced and quantified in a single biological replicate.

140

p53_S315: RALPNNTSS pS PQPK Ion score 71, pRS loc. Prob. S(10):100%

p53_S315: RALPNNTSSpS PQPK 9d Rep1 (L:CON,M:OIS,H:OISb):

9d Rep2 (L:CON,M:OISb,H:OIS):

Supplementary Figure S5. p53 S315 phosphorylation downregulated in senescence a. ETD peptide fragmentation spectra showing confident phosphopeptide identification and phosphosite localization. b. MS1 Quantification spectra with phosphopeptide levels clearly present in all samples but due to interference in the light channel of the second replicate no ratios were calculated for the control comparisons. When quantifying only the first isotopes, clearly a downregulation is observed for the senescence samples in both replicates.

Chapter 6

Supplementary Figure S6. Phosphosites regulated specifically in OIS Phosphosite abundance comparisons for cycling (Cycl), BRAF transduced (OIS) and BRAF/sh-C/EBPÎ˛ transduced (OISb) HDF after 3 days (a) and 9 days (b) of cell growth. The logarithmic base 2 ratio for two conditions of replicate 1 was plotted against the logarithmic base 2 ratio for the same two conditions of replicate 2 to visualize consistent changes in replicates.

141

Supplementary Tables S1 and S2 can be found with this article online at http://www. mcponline.org/content/13/8/2089.long.

142

足足足

CHAPTER 7 SIGNAL TRANSDUCTION REACTION MONITORING DECIPHERS SITE-SPECIFIC PI3K-mTOR/MAPK PATHWAY DYNAMICS

Manuscript submitted

SIGNAL TRANSDUCTION REACTION MONITORING DECIPHERS SITE-SPECIFIC PI3K-mTOR/MAPK PATHWAY DYNAMICS Erik L. de Graaf1,2,*, Joanna Kaplon3,*, Shabaz Mohammed1,2,4, Lisette A.M. Vereijken1,2, Daniel P. Duarte3, Laura Redondo Gallego1,2, Albert J.R. Heck1,2, Daniel S. Peeper3, A.F. Maarten Altelaar1,2 Biomolecular Mass Spectrometry and Proteomics, Utrecht Institute for Pharmaceutical Sciences

and Bijvoet Center for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, The Netherlands. 2Netherlands Proteomics Centre, Padualaan 8, 3584 CH Utrecht, The Netherlands. 3Division of Molecular Oncology, The Netherlands Cancer Institute, Plesmanlaan 121, 1066 CX Amsterdam, the Netherlands. 4Present address: Chemistry Research Laboratory, Department of Chemistry and Department of Biochemistry, University of Oxford, South Parks Road, OX1 3QU, Oxford, United Kingdom. * These authors contributed equally to this work.

Alterations in cellular signaling networks are the cause of many diseases and determine highly variable and microenvironment-dependent drug target potency1. Therefore, understanding perturbations in cellular signaling networks is of great interest from a clinical point of view. Signal transduction is still mostly studied by immunoblot analysis using phosphosite-directed antibodies. However, the limited availability of these reagents and the low throughput of immunoblot analysis severely hinder the analysis of complete signaling pathways. Other limiting factors of antibody-based protein phosphorylation assays are the semi-quantitative and often ambiguous nature of the results obtained. Phosphosites resident in highly conserved sequences are frequently undistinguishable between closely related protein isoforms, with ERK1 Y204 and ERK2 Y187 serving as prime examples. Moreover, close-proximity sites such as T202/Y204 on ERK1, T185/Y187 on ERK2 or S235/ S236 and S240/S244 on RPS6, are indistinguishable by antibody binding. In an attempt to address these problems, dual site-specific antibodies have been developed that recognize multiple isoforms and multiply phosphorylated domains. However, this approach does not allow for monitoring of possible mutually exclusive phosphorylation patterns and other

145

Chapter 7

SUMMARY Combining targeted quantitative mass spectrometry (MS) with highly selective phosphopeptide enrichment, we have monitored phosphorylation dynamics in the PI3KmTOR and MAPK signaling networks, in the context of oncogene-induced senescence (OIS). Our analysis reveals site-specific phosphorylation of ERK 1/2, p70S6K and RPS6, which were previously undetectable by phospho-antibodies. This study illustrates the applicability of our MS-based approach for high-resolution phosphodynamics screening and provides novel insights in signaling network perturbations associated with OIS.

cross-talk that may occur2. The current lack of antibodies that are specific for close-proximity or protein isoform phosphosites, has left the dynamics of these potentially important phosphosite differences largely unexplored, demanding the development of methods with increased resolving power. Global shotgun mass spectrometry (MS)-based phosphoproteomics can address many of these issues, allowing for the analysis of thousands of phosphorylation events with high specificity for protein isoform and phosphosite localization3. Currently, the main challenge of low phosphoprotein stoichiometry can be partially solved by enriching the phosphorylated peptides from the more abundant unphosphorylated peptides, using affinity-based techniques such as TiO24 or Fe3+/Ti4+-IMAC5,6. Nonetheless, the large dynamic range in the remaining phosphopeptide population greatly hampers global shotgun proteomics approaches. Detection reproducibility and sensitivity can be optimized using targeted MS, referred to as selected or multiple reaction monitoring (SRM/MRM). Moreover, when combined with stable isotope standard (SIS) phosphopeptides, technical variation can be reduced, resulting in increased accuracy in quantification7. Although SRM has been established for monitoring protein abundance8,9, its applicability to monitor selected protein phosphorylation events is still in its infancy. Up till now, only low-throughput targeted phosphosite experiments have been described using SRM, analyzing a small number of phosphosites from high abundant proteins7,10 or protein complexes11. However, no monitoring strategy has been reported using SRM phosphorylation dynamics of all components, at differing expression levels and in entire signal transduction pathways. Here, we have substantially improved phospho-SRM methods to accurately, reproducibly and comprehensively quantify phosphorylation events in the MAPK and PI3K-mTOR pathways. The success of this approach is determined by the combination of the sensitive and quantitatively reproducible Ti4+-IMAC enrichment strategy6,12 with SRM, using low-cost crude synthetic SIS phosphopeptides at two different concentrations (Figure 1). To measure the essential phosphosites selected, robust interference-free SRM assays were constructed using reproducible retention times, co-eluting SIS reference peptides for relative fragment ion intensity comparison together with SRM-triggered MS/MS for identification and phosphosite localization (Suppl. Figure 1-4. and Suppl. Data). Our method, termed signal transduction reaction monitoring (STREM), was applied to characterize the dynamics of the MAPK and PI3K-mTOR signal transduction pathways in oncogene-induced senescence (OIS). OIS is a largely irreversible state of cell cycle arrest, which can be triggered by the unscheduled activation of oncogenes13. It has been shown to act alongside death programs to suppress tumorigenesis14,15. For this purpose, â&#x20AC;&#x153;controlâ&#x20AC;? cycling cells (Cycl), OIS cells (upon introduction of BRAFV600E, a common oncogene and strong inducer of OIS) and OIS-bypassing cells (OISb)16,17, were analyzed in biological triplicates.

146

Additionally, cycling, OIS and OISb cells treated with BEZ235 (a dual PI3K/mTOR inhibitor), were included to screen for mTOR-sensitive perturbations. In total, 51 phospho-patterns from 27 different signaling proteins could be reproducibly quantified in 18 samples. + / - BEZ235

100X SIS

1X SIS

Ti4+-IMAC

Phosphopeptide enrichment

Signal transduction monitoring by SRM 100X SIS

1X SIS

Cycl

OIS

Lysis Red. & Alk

Sep-Pak C18

StageTip

Lys-C/Tryp digestion

Desalting

OISb

Several nodes in the MAPK pathway were expected to show changes in phosphorylation, as both OIS and OISb cells overexpress the constitutively active BRAFV600E mutant. Indeed, under each condition we observed a specific upregulation of the downstream ERK1/2 TEY dual phosphorylation motif, as seen in both immunoblot and STREM analyses. We also detected a strong BRAF S729 phosphorylation in the STREM measurements (Figure 2a, b). Comparing the two ERK isoforms, similar trends in phosphorylation were observed in both immunoblot and STREM analyses, indicating a level of redundancy between ERK1/2. However, the STREM data showed a stronger induction of ERK1 dual phosphorylation. Moreover, STREM allowed for the analysis of the single tyrosine sites (Y204/Y187) of which phosphorylation is known to precede that of the threonine residues (T202/T185). Surprisingly, single tyrosine site phosphorylation on both Y204 and Y187 was downregulated in OIS cells compared to cycling cells for both ERK1 and 2, which contrasts with the increase observed for the dual phosphorylation in OIS (Suppl. Figure 3). These results highlight the superior resolving power of our method, revealing differences in TEY-motif phosphorylation on different ERK1/2 isoforms that would remain undetected with classical approaches. Another major proliferation-controlling pathway is the one in which PI3K and mTOR are the main actors. Activation of PI3K-mTOR triggers protein synthesis via the phosphorylation of 4E-binding protein 1 (4EBP1) and the ribosomal protein S6 kinase (p70S6K). Both immunoblot and STREM assays confirmed a strong mTOR-dependent regulation of 4EBP1 S65, as evidenced by the total loss of S65 phosphorylation upon treatment with the dual PI3K/mTOR inhibitor BEZ235 (Figure 2a,b). Interestingly, a strong reduction of

147

Chapter 7

Figure 1. Experimental workflow scheme of signal transduction reaction monitoring (STREM) Biological triplicates of cycling (Cycl), undergoing oncogene-induced senescence (OIS) and OIS bypassing (OISb) cells were collected for mass-spectrometry (MS) analysis. After lysis and digestion, two concentrations (high and low) of stable isotope standard (SIS) phosphopeptides per each phosphosite to be analyzed were added, to achieve similar levels of endogenous and SIS peptides. Subsequently phosphopeptides were enriched using single stage Ti4+-IMAC, followed by a scheduled two hour LC-SRM run for each sample.

growth-stimulating 4EBP1 phosphorylation was observed in OIS and OISb, albeit less pronounced in the latter. A similar trend in signaling perturbation was observed for p70S6K phosphorylation on T412 (often referred to as T389) and S452, however not for S441 and S447 (Figure 2, 3a). Phosphorylation of p70S6K T412 is often used as a read-out for mTOR activity in immunoassays. For p70S6K S452, no site-specific antibody is available and no responsible kinase or biological role has been reported. By STREM analysis, we show an identical regulation of S452 and T412, including a total loss of phosphorylation upon BEZ235 treatment, indicating S452 as a novel putative mTOR phosphorylation site.

0.75

2.5

0.25

0.0

0.00

0.75

RPS6 S236/S240 *

1.50 1.25 1.00

+BEZ

yc l

IS IS b C yc l O IS O IS b

0.00

IS IS b C yc l O IS O IS b

0.50 0.2 0.1 0.0

0.25

p70S6K

+BEZ

0.00 1.25 1.00 0.75 0.50 0.25 0.04 0.02 0.00

p-4EBP1 S65

0.25

0.75

0.50

1.00 0.50

1.00

T412 (T389)

0.75

p70S6K S447 * *

p-p70S6K

RPS6 S240 1.25

4EBP1

RPS6 S235/S236/S240 * * *

p-RPS6 S235/236

p-RPS6 S240/244

RPS6

yc l

RPS6 S236

1.25

T202,Y204 T185,Y187

ERK1/2

p70S6K S452 *

* *

+BEZ

p-ERK1/2

0.50

1.50 1.25 1.00 0.75 0.50 0.25 0.00

b yc l

1.00

5.0

4EBP1 S65 *

IS IS b C yc l O IS O IS b

7.5

1.25

10.0

ERK1 T202/Y204 & ERK2 T185/Y187 * *

1.50

Fold change (relative to Cycl)

12.5

IS IS b C yc l O IS O IS b

BRAF S729 *

Fold change (relative to Cycl)

+BEZ

Figure 2. Phosphosite dynamics of several nodes in the MAPK and PI3K-mTOR signaling Using STREM (a) protein isoform phosphorylation domains as well as close-proximity phosphosites could be dissected, untraceable by immunoblot analysis (b).

Another well-studied downstream target of both the PI3K-mTOR and MAPK pathways is the ribosomal protein S6 (RPS6). RPS6 functions to integrate proliferation promoting signals from both pathways; its phosphorylation is increased in many types of cancer and is proposed to be a marker for drug resistance18. RPS6 is primarily known as a downstream mTOR target and is phosphorylated on S235, S236, S240 and S244 by p70S6K. However, S235 and S236 can be phosphorylated also by RSKs (RSK1/2) from the MAPK pathway19. In all studies on RPS6 signaling antibodies were used detecting only dual phosphorylation on S235/S236 or S240/S244. As this highly conserved and confined domain is functionally important and phosphorylated by two different kinases from two pathways, phosphorylation patterns are most likely more complex than merely two separate dual phosphorylations. The immunoblot data on RPS6 S235/S236 and S240/S244 phosphorylation showed no substantial differences in the level of phosphorylation between cycling, OIS and OISb cells.

148

149

Chapter 7

Treatment with BEZ235 revealed a strong dependence on mTOR activity in cycling cells (Figure 2b). Interestingly, the immunoblot data already reveal an altered response of these sites to BEZ235 treatment, showing only partial loss of phosphorylation in OIS and to a lesser extent in OISb. Exploiting the specificity of the STREM approach, we were able to dissect different patterns for S235, S236 and S240 phosphorylation, associated with different biological conditions (Figure 2a). For example, a strong mTOR-independent induction of S236 phosphorylation was observed in OIS. When interpreting the results obtained by the S235/S236 antibody one might falsely conclude that the phosphorylation of both sites is mainly dependent on mTOR. However, STREM revealed that S236 is strictly phosphorylated by RSK1/2, specifically in OIS. STREM analysis on single S240 phosphorylation showed comparable results to dual S240/S244 phosphoantibody analysis. Similar to immunoblot analysis, STREM showed no difference in the level of S240 phosphorylation between control, OIS and OISb cells. In addition, both analyses revealed only partial loss of phosphorylation in BEZ235-treated OIS cells when compared to BEZ235-treated control or OISb cells. Interestingly, the phosphorylation of the dual (S236/S240) and the triply phosphorylated domain (S235/S236/S240) were strongly (albeit not exclusively) dependent on mTOR. In samples without BEZ235 treatment, both dual and triple phosphorylation showed a significant downregulation in OIS compared to cycling and OISb, a pattern that was reversed when mTOR was inhibited (Suppl. Figure 3). The complex differential phosphorylation patterns on RPS6, which was found by STREM analysis only, indicate an important role of multiple signaling pathways in regulation of a confined domain on a single protein. Taken together, STREM analysis revealed a strong and specific regulation of different mTOR substrates in cells that undergo OIS, which is associated with, not reported before, remarkably high phosphosite pattern differentiation. Besides obtaining specific phosphosite pattern quantifications, STREM also enables comprehensive monitoring of pathway dynamics in a single run per sample, illustrated here for the full PI3K-mTOR and MAPK pathways (Figure 3a). It has been reported previously that phosphorylation of PRAS40 at S183 and T246 by mTOR and AKT, respectively, represses its inhibitory function on mTORC1 signaling20. Interestingly, our pathway analysis revealed that both S183 and T246 phosphorylation sites of PRAS40 are reduced in OIS, hinting at an important role of PRAS40 in the reduced mTORC1 activity associated with OIS. In line with this, PDK1, an upstream AKT-PRAS40 kinase, showed lower S241 phosphorylation in OIS that has been previously associated with decreased PDK1 activity21. Next, we systematically searched for pathway components specifically regulated by the PI3K-mTOR pathway (Figure 3b). Analysis of BEZ235-treated cells confirmed that phosphorylation of specific sites on p70S6K and 4EBP1 is mTOR-dependent whereas all measured sites on RSK1/2 are not. Other potential direct or indirect mTOR substrates showing regulation upon BEZ235 treatment include TSC2 S1388 and S1411, MEK2 T394 and eIF4B S445.

In conclusion, we report here a new strategy to comprehensively monitor pathway signal transduction dynamics in a site-specific manner. Precise and phosphosite-specific quantification with increased throughput is demonstrated, allowing the characterization of complex protein phosphorylation patterns associated with a particular biological phenotype. The need for such a comprehensive full pathway analysis method is illustrated by highlighting previously unknown and undetectable changes in protein phosphorylation in OIS before and after pharmacological inhibition of mTOR. a PI3K-mTOR mTORC2

mTOR

S813 S259 S338 S244

PIK3R4

MAPK

PIP2

RAS GDP

PIK3C2A PIK3C3

Rictor S21 S1385 S1591 S1777

S241 AKT T451

PIP3 PDK-1

MEK2

AMP

S183 T246

S1346 S1388 S1411 S1449

PRAS40 AKT1S1 S183

T202

TSC2

S510

SIN1

S863 S859

Raptor S722

S65

S65 4EBP1

4EBP1

S78 S951

eIF4E-T EIF4ENIF1

eIF4E

eIF4E Translation “off”

Kinase

PIK3C3

Rictor S21 S1385 S1591 S1777

S241 AKT T451

T246

S183 T246

S183

GDP

PIP3

AMP

SIN1

S863 S859

Raptor S722

RSK2 p90RSK3

S415 S715

S221 S380 S235 S240 S236

S93 S445 eIF4B

eIF3 eIF4A

RPS6

40S

S236 S240 S235,S240 S236,S240 S235,S236, S240

S65

S65 4EBP1

4EBP1

eIF4G2

eIF4E “Scanning” & Protein Synthesis

OIS vs. OISb OIS vs. Cycl OISb vs. Cycl

up regulated not regulated down regulated Senescence specific up Senescence specific down

Translation “off”

p70S6K

S78 S951

eIF4E-T EIF4ENIF1

eIF4E

MEK2

T394

T202

Y204

ERK2

Y204 T202,Y204 Y187 T185,Y187

RSK2 p90RSK3

S415 S715

ERK1

S722 S863 S859,S863

S380

RSK1 p90RSK1

T389/T412

S441 S447 S452

S446 S729

LKB1

TSC2

S221

S380

BRAF

S31

S377

S505

mTORC1

S186

RRAS2

GTP

RAS

TSC1

S510

RSK1 p90RSK1

eIF4E

RAS

PDK-1

S1346 S1388 S1411 S1449

PRAS40 AKT1S1

mTOR

T508 mRNA 5' cap-binding first step in translation initiation mRNA 5' cap-unwinding

inhibiting phosphorylation activating phosphorylation

Y204 T202,Y204 Y187 T185,Y187

S722 S863 S859,S863

S221

p70S6K

Y204

ERK1 ERK2

T389/T412

S441 S447 S452

MAPK

PIP2

AMPK

S505 mTOR

T394

PIK3R4 PIK3C2A

LKB1

TSC1 mTORC1

mTOR

S446 S729

S813 S259 S338 S244

S31

S377 AMPK

T246

BRAF

PI3K-mTOR

mTORC2

S186

RRAS2

GTP

RAS

S221 S380 S235 S240 S236

S93 S445 eIF4B eIF4E

eIF3 eIF4A

RPS6

40S

S236 S240 S235,S240 S236,S240 S235,S236, S240

eIF4G2

T508 mRNA 5' cap-binding first step in translation initiation mRNA 5' cap-unwinding

“Scanning” & Protein Synthesis

inhibiting phosphorylation Cycl+BEZ vs. Cycl activating phosphorylation OIS+BEZ vs. OIS OISb+BEZ vs. OISb Kinase

up regulated not regulated down regulated / PI3K-mTOR regulated

Figure 3. Regulations in the PI3K-mTOR and MAPK pathway measured by STREM a. Phosphosites specifically regulated in OIS. b. Phosphosites sensitive to PI3K-mTOR inhibition by BEZ235 treatment. Significantly up- or downregulated phosphosites are color-coded in green and red circles, respectively. Protein interaction and kinase substrate relationships were extracted from the UniprotKB and phosphositeplus databases and manually curated using literature.

ACKNOWLEDGEMENTS This work was supported by the Netherlands Proteomics Center, and the PRIME-XS project (grant agreement number 262067) funded by the European Union 7th Framework Programme. A.F.M.A. was supported by the Netherlands Organization for Scientific Research

150

(NWO) with a VIDI grant (723.012.102) and D.S.P. with a NWO VICI grant and a Queen Wilhelmina Award grant from the Dutch Cancer Society (KWF Kankerbestrijding). This work is part of the project Proteins At Work, financed by the Netherlands Organization for Scientific Research (NWO) as part of the National Roadmap Large-scale Research Facilities of the Netherlands (project number 184.032.201). AUTHOR CONTRIBUTIONS E.d.G., S.M., J.K., D.S.P., A.F.M.A., designed the study. E.d.G., J.K., L.A.M.V, D.P.D., L.R.G. performed the experiments. E.d.G. analyzed the data. A.J.R.H, D.S.P, A.F.M.A. supervised the study E.d.G., J.K., D.S.P, A.J.R.H, A.F.M.A. wrote the manuscript.

2. 3.

Ras/Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR Inhibitors: Rationale and Importance to Inhibiting These Pathways in Human Health. 2, 135–164–164 (2011). Hunter, T. The age of crosstalk: phosphorylation, ubiquitination, and beyond. Mol Cell 28, 730–738 (2007). Altelaar, A. F. M., Munoz, J. & Heck, A. J. R. Next-generation proteomics: towards an integrative view of proteome dynamics. Nat Rev Genet 14, 35–48 (2013). Pinkse, M. W. H., Uitto, P. M., Hilhorst, M. J., Ooms, B. & Heck, A. J. R. Selective isolation at the femtomole level of phosphopeptides from proteolytic digests using 2D-NanoLCESI-MS/MS and titanium oxide precolumns. Anal Chem 76, 3935–3943 (2004). Villén, J. & Gygi, S. P. The SCX/IMAC enrichment approach for global phosphorylation analysis by mass spectrometry. Nature protocols 3, 1630– 1638 (2008). Zhou, H. et al. Robust phosphoproteome enrichment using monodisperse microsphere-based immobilized titanium (IV) ion affinity chromatography. Nature protocols 8, 461–480 (2013). Gerber, S. A., Rush, J., Stemman, O., Kirschner, M. W. & Gygi, S. P. Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS. Proc Natl Acad Sci USA 100, 6940–6945 (2003). Picotti, P., Bodenmiller, B., Mueller, L. N., Domon, B. & Aebersold, R. Full dynamic range proteome analysis of S. cerevisiae

10.

11.

12.

13. 14. 15. 16.

by targeted proteomics. Cell 138, 795–806 (2009). Kennedy, J. J. et al. Demonstrating the feasibility of large-scale development of standardized assays to quantify human proteins. Nat Methods 11, 149–155 (2014). Narumi, R. et al. A strategy for large-scale phosphoproteomics and SRM-based validation of human breast cancer tissue samples. J. Proteome Res. 11, 5311–5322 (2012). Bisson, N. et al. Selected reaction monitoring mass spectrometry reveals the dynamics of signaling through the GRB2 adaptor. Nat. Biotechnol. 29, 653–658 (2011). de Graaf, E. L., Giansanti, P., Altelaar, A. F. M. & Heck, A. J. R. Single step enrichment by Ti4+-IMAC and label free quantitation enables in-depth monitoring of phosphorylation dynamics with high reproducibility and temporal resolution. Mol. Cell Proteomics (2014). doi:10.1074/mcp. O113.036608 Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). Michaloglou, C. et al. BRAFE600-associated senescence-like cell cycle arrest of human naevi. Nature 436, 720–724 (2005). Collado, M. & Serrano, M. Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51–57 (2010). Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008).

151

Chapter 7

REFERENCES

17.

18.

19.

20.

21.

de Graaf, E. L. et al. Phosphoproteome Dynamics in Onset and Maintenance of Oncogene-induced Senescence. Mol. Cell Proteomics 13, 2089–2100 (2014). Qian, Z. R. et al. Prognostic significance of MTOR pathway component expression in neuroendocrine tumors. J Clin Oncol 31, 3418–3425 (2013). Roux, P. P. et al. RAS/ERK signaling promotes site-specific ribosomal protein S6 phosphorylation via RSK and stimulates cap-dependent translation. J Biol Chem 282, 14056–14064 (2007). Vander Haar, E., Lee, S.-I., Bandhakavi, S., Griffin, T. J. & Kim, D.-H. Insulin signalling to mTOR mediated by the Akt/PKB substrate PRAS40. Nat Cell Biol 9, 316–323 (2007). Casamayor, A., Morrice, N. A. & Alessi, D. R. Phosphorylation of Ser-241 is essential for the activity of 3-phosphoinositidedependent protein kinase-1: identification of

22.

23.

24.

25.

five sites of phosphorylation in vivo. Biochem J 342 ( Pt 2), 287–292 (1999). de Graaf, E. L., Altelaar, A. F. M., van Breukelen, B., Mohammed, S. & Heck, A. J. R. Improving SRM assay development: a global comparison between triple quadrupole, ion trap, and higher energy CID peptide fragmentation spectra. J. Proteome Res. 10, 4334–4341 (2011). MacLean, B. et al. Skyline: an open source document editor for creating and analyzing targeted proteomics experiments. Bioinformatics 26, 966–968 (2010). Dickhut, C., Feldmann, I., Lambert, J. & Zahedi, R. P. Impact of digestion conditions on phosphoproteomics. J. Proteome Res. 13, 2761–2770 (2014). Taus, T. et al. Universal and confident phosphorylation site localization using phosphoRS. J. Proteome Res. 10, 5354–5362 (2011).

ONLINE METHODS Cell culture The human diploid fibroblast (HDF) cell line Tig3 expressing the ecotropic receptor, hTERT and sh-p16INK4A (Tig3 (et)-16i) was maintained in DMEM with 4.5 mg/ml glucose and 0.11 mg/ml sodium pyruvate, supplemented with 9% fetal bovine serum (PAA), 2 mM glutamine, 100 units/ml penicillin and 0.1 mg/ml streptomycin (GIBCO). The Phoenix packaging cell line was used for the generation of ecotropic retroviruses. For infections, filtered (pore size 0.45 mm) viral supernatant, supplemented with 4–8 μg/ml polybrene was used. In general, a single infection round of 6 h was sufficient to infect at least 90% of the population. For senescence experiments, cells were infected with shCEBP/β-encoding or control retrovirus, selected pharmacologically (puromycin) and subsequently infected with BRAFV600E-encoding or control virus. After 9 days of selection (blastomycin), cells were collected for mass-spectrometry (MS) and immunoblot assays. For screening for mTORsensitive perturbations, cells were pre-treated with 500 nM mTOR inhibitor BEZ235 for 24 h. Results represent data from three independent experiments. Plasmids The plasmids pMSCV-blast and pMSCV-blast-BRAFV600E as well as pRS-puro and pRS-puro-C/ EBPβ#1 were previously described14,16 Antibodies Antibodies used for immunoblotting were: phospho-p44/42 MAPK (ERK1/2) (T202,Y204/ T185,Y187; E10, #9106; Cell Signaling); total p44/42 MAPK (ERK1/2) (#9102; Cell Signaling); phospho-p70S6 (T389/T412; #9208; Cell Signaling); total p70S6 (#9202; Cell Signaling);

152

153

Chapter 7

phospho-4EBP1 (S65; #9456; Cell Signaling); total 4EBP1 (#9453; Cell Signaling); phosphoRPS6 (S235/S236; #2211; Cell Signaling); phospho-RPS6 (S240/S244; #2215; Cell Signaling); total RPS6 (#2217; Cell Signaling). Sample preparation Frozen HDF cell pellets were lysed by sonication in lysis buffer (8M Urea in 50 mM ammonium bicarbonate, supplemented with 1 tablet Complete mini EDTA-free Cocktail (Roche) and 1 tablet PhosSTOP phosphatase inhibitor Cocktail (Roche) per 50 mL of lysis buffer). After centrifugation (20 000x g 30min at 4°C), the supernatant was collected and assayed for protein content using the bicinchoninic acid assay (BCA) kit following manufacturer instructions (Pierce, IL USA). Protein reduction and alkylation were performed using final concentrations of 5 mM dithiothreitol and 10 mM iodoacetamide, respectively. The first enzymatic digestion step was performed with Lys-C at 37 °C for 4 h in lysis buffer (enzyme:substrate 1:75). For the second digestion, samples were diluted to 2 M Urea using 50 mM ammonium bicarbonate and incubated overnight with trypsin (Promega, USA) at 37 °C (enzyme:substrate 1:100). The resulting endogenous peptides were acidified with 10% formic acid (FA) and split into two aliquots of 500 μg peptides for each sample. Next, aliquots were spiked with either 10 pmol or 0.1 pmol of heavy stable isotope standard (SIS) peptides. Crude synthetic SIS peptides containing one C-terminal heavy Lysine (6C13,2N15: +8Da) or heavy Arginine (6C13, 4N15: +10Da) amino acid, were purchased from JPT technologies (Berlin, Germany). The SIS stock was created by mixing equimolar quantities of all peptides in 50% acetonitrile (ACN), 1% FA and divided into 100 pmol aliquots stored at -20°C. Prior to phosphopeptide enrichment, peptide mixtures were desalted on Sep-Pak C18 columns (Waters, USA, Massachusetts), dried to completion in vacuum and stored at -80°C. Conventionally, highly purified synthetic peptides are used for high precision absolute quantification. In our novel approach we used less pure crude synthetic phosphopeptides for each site and in two different concentrations per sample, to allow for high-throughput, sensitive and reproducible relative quantification. Phosphopeptide enrichment Two Ti4+-IMAC columns for each sample were prepared and processed in parallel using a micro centrifuge as described previously6. Briefly, microcolumns were created by loading 500 μg Ti4+-IMAC beads onto GELoader tips (Eppendorf) with a C8 plug. Ti4+-IMAC columns were pre-equilibrated two times with 30 μL of loading buffer (80% ACN, 6% trifluoroacetic acid (TFA)). Next, all samples were reconstituted in 250 μL loading buffer and 100 μl was loaded onto two equilibrated Ti4+-IMAC columns (corresponding to 200μg natural peptides (NAT) and 4 or 0.04 pmol SIS per column). After loading of samples, Ti4+-IMAC columns were washed with 60 μL washing buffer A (50% ACN, 0.5% TFA, 200 mM NaCl) and subsequently with 40 μL washing buffer B (50% ACN, 0.1% TFA). Bound peptides were eluted by 30 μL of 10% ammonia into 30 μL of 10% FA. Finally, the remainder of the peptides was eluted with 5 μL of (80% ACN, 2% FA). The

collected eluate was further acidified by adding 20 μL of 10% TFA and subsequently desalted using homemade reversed phase tips loaded with 10 μL of Aqua C18 5μm beads. Tips were washed twice with 30 μL 0.1% TFA, followed by elution of peptides with 20 μL 80% ACN, 1%FA. Prior to liquid chromatography (LC)-MS analysis, peptide mixtures were dried to completion in vacuum and stored at -80°C. For each sample only one Ti4+-IMAC column was analysed. The second loaded column served as a backup and was only used when primary column failed. Signal transduction reaction monitoring (STREM) assay development To confirm SIS peptide sequence and phosphosite localization and consistent LC-MS observability, all ordered crude SIS peptides were run separately on different LC-MS setups. Crude peptides were characterized by data-dependent acquisition (DDA) LC-MS runs on an Orbitrap Q-Executive (Thermo Scientific) because we have demonstrated previously that this type of MS/MS spectra are more beneficial for selected reaction monitoring (SRM) assay development than conventional ion trap CID spectra22. Peptides that failed to be identified or for which phosphosites were not localized by DDA LC-MS runs, were reanalyzed by targeted ETD fragmentation on an Orbitrap Elite (Thermo Scientific) for better phosphopeptide identification and localization, or by SRM triggered MS/MS on a TSQ Vantage (Thermo Scientific) for increased sensitivity. Subsequently, all beam-type MS/MS spectra (i.e. HCD and TSQ spectra) were used for SRM assay development. Assay development was performed using Skyline23 to pick the 8 most abundant fragment ions for SRM-mode validation and collision energy optimization. Collision energy optimization was achieved by ramping the collision energy (CE) from 7.5 eV lower to 7.5 eV higher than the calculated CE (standard skyline equation) in steps of 1.5 eV. For the final assay the 5 most abundant peptide fragment ions were manually picked that showed no (auto-)interference in standard (SIS) or endogenous (NAT) peptides measured in HeLa or HDF cell lysates enriched by Ti4+-IMAC. Peptide ID was confirmed by matching retention time (RT), relative fragment ion intensities and SRM triggered MS/MS runs that were scheduled in-between sample batches. Exemplary phosphopeptide RT reproducibility and XICs are shown in Supplementary Figure 1. Out of 49 phosphosites (on 51 phosphopeptides), 48 could be pinpointed with a localization certainty higher than 95% resulting in a total average localization probability of 98.6%. In some cases tryptic phosphopeptides inevitably contained miscleavages (MC) or methionine residues. MC in phosphopeptides are frequently observed when negatively charged phosphomoieties are in close proximity to positively charged arginine or lysine residues thereby hampering cleavage by trypsin. Our data and other studies24 show that quantification standard deviations are similar between MC peptides and non MC peptides, indicating MC peptides are reproducibly formed under our sample preparation conditions and therefore can be used for reproducible phosphosite quantification. In the case of methionine containing peptides, SRM assays for both the normal and oxidized forms were generated and measured

154

155

Chapter 7

in the final sample. However, when one of the two oxidation states was not present in a natural or SIS peptide, the peptide isoform was not taken into account for further analysis. In only two cases both oxidation states were present on both the SIS and NAT peptide. SRM analysis of these peptides showed identical regulation of both isomers with small deviations between biological replicate samples (Suppl. Fig. 2), indicating oxidation was similar on both SIS and NAT peptides in all samples. STREM LC-MS data acquisition All samples were reconstituted in 10 μL 10% FA and 30% was analysed on a TSQ Vantage coupled to an Easy nLC-1000 LC system configured with a single easy spray analytical column (ES803; 50 cm x 75 μm ID, 2 μm particles, 100 Å pore size) (Thermo Scientific) using 2 hour runs. Briefly, samples were loaded with 6 μL buffer A (0.1% FA) at 800 bar limited to a maximum of 300 nl/min, column equilibration was performed with 3 μL buffer A at 800 bar, peptides were separated on a gradient of 0-65 min 3-25% buffer B (0.1% FA, 99,9% ACN) and 65-75 min 25-40% buffer B at 200 nl/min, followed by a column wash for 10 min at 200 nl/min 100% buffer B. The TSQ Vantage spray voltage was set to 2.3 kV and was further configured to select peptides in Q1 at 0.7 FWHM and fragment them at 1.5 mTorr Argon in the second quadrupole. In total 89 phosphopeptides were measured in light and heavy resulting in 884 transitions. All transitions were measured using polytyrosine tuned S-lens values, transition optimized CE values, scheduled peptide RT window of 1.75 - 3 minutes and a cycle time of 1.8 seconds, resulting in a maximum of ~60 concurrent transitions. DDA LC-MS and STREM Data analysis Peptide MS/MS spectra were identified using Proteome Discoverer 1.4 (Thermo Scientific, Bremen) and Mascot 2.4 (Matrix Science, London) and phosphosite localization was performed using the phosphoRS25 PD node and SRM data were analyzed using Skyline23. For the final SRM data set, transitions with remaining interference, signal intensities above 13E6 or low signal-to-noise ratios were discarded for further analysis. To find samples or peptides with interfering irreproducible transitions, a t-test was performed comparing the average relative fragment intensity ratio between light and heavy peptides from biological triplicate runs. All transitions with a p-value below 0.1 indicated inequality between NAT and SIS relative fragment intensities and were therefore discarded. Additionally, all samples were inspected manually for transition interference and consistency in transition peak shape, peak integration and RT, using Skyline. In the end, only peptides with a minimum of 3 transitions for each SIS and NAT were kept for further analysis resulting in an average of 4,2 transitions per peptide isotopologue. To achieve relative quantification, crude synthetic peptides were used for determining relative ratios. To prevent quantification problems regarding upper and lower limits of detection and signal response linearity, preferably samples with equal intensity for SIS and NAT versions are used (SIS/NAT ratio close to 1). Therefore, as described above in sample preparation section,

each sample was split in two aliquots and each was spiked with a different SIS concentration allowing for each peptide analyte to select one of the two (or both) spike-in concentrations, ideally obtaining a SIS-NAT fold-difference less than 10. However, due to strong regulations in NAT levels between different phenotypes the ratio NAT/SIS could inherently not always be kept less than 10-fold. To obtain phosphosite abundances relative to control, each SISnormalized phosphopeptide abundance was divided by the average intensity of the control biological triplicate. Finally, two sided student t-tests with equal variances were used to calculate significant (p<0.05) changes between biological triplicate samples. Supplementary Data. All annotated spectra and STREM quantification results are available for download at the PASSEL data repository (acc. nr.: PASS00552).

156

SUPPLEMENTARY INFORMATION a

LNQPGTPTR - 532.2502++

Retention Time

LNQPGTPTR - 537.2544++ (heavy)

29 28 27

II.OISb_4_BEZ

III.OISb_4_BEZ

III.OIS_4_BEZ

I.OISb_4_BEZ

I.OIS_4_BEZ

II.OIS_4_BEZ

III.Cycl_4_BEZ

I.Cycl_4_BEZ

II.Cycl_4_BEZ

II.OISb_4

III.OISb_4

I.OISb_4

II.OIS_4

III.OIS_4

I.OIS_4

III.Cycl_4

I.Cycl_4

II.Cycl_4

Replicate IDSTEVIYQPR - 700.8265++

Retention Time

IDSTEVIYQPR - 705.8306++ (heavy)

56 55 54 53 III.OISb_4_BEZ

II.OISb_4_BEZ

I.OISb_4_BEZ

III.OIS_4_BEZ

II.OIS_4_BEZ

I.OIS_4_BEZ

II.Cycl_4_BEZ

III.Cycl_4_BEZ

I.Cycl_4_BEZ

III.OISb_4

III.OIS_4

I.OISb_4

II.OIS_4

I.OIS_4

II.OISb_4

III.Cycl_4

II.Cycl_4

I.Cycl_4

Replicate

IDS[+80]TEVIYQPR (LKB1 S31)

3000000

Intensity

I.Cycl_4_BEZ II.Cycl_4_BEZ III.Cycl_4_BEZ

I.Cycl_4 II.Cycl_4 III.Cycl_4

2500000

I.OIS_4 II.OIS_4 III.OIS_4

2000000

I.OIS_4_BEZ II.OIS_4_BEZ III.OIS_4_BEZ

I.OISb_4 II.OISb_4 III.OISb_4

1500000 1000000

Chapter 7

I.OISb_4_BEZ II.OISb_4_BEZ III.OISb_4_BEZ

500000 0

52.5

53.5

54 54.5 RT (min)

55.5

56.5

Supplementary Figure 1. Typical RT reproducibility and extracted ion chromatograms of two representative phosphopeptides a. RT reproducibility over all measured samples. LNQPGT[+80]PTR (MEK2 T394) (top) and IDS[+80]TEVIYQPR (LKB1 S31) (bottom) measured in high and low SIS spiked samples, respectively. b. Phosphopeptide chromatograms of endogenous (Red) and SIS (Blue) LNQPGT[+80]PTR from all 18 samples. c. Overlaid phosphopeptide chromatograms for all 18 samples of the SIS peptide IDS[+80]TEVIYQPR.

157

4EBP1 S65: FLMECRNSPVTK

1.5

p90S6K/RSK1 S221: AYSFCGTVEYMAPEVVNR

4 3

1.0

0.5

0.0

FLMECRNSPVTK

1.5

yc l+ BE IS Z + O BE IS Z b+ BE Z

IS b

yc l 4

AYSFCGTVEYMAPEVVNR

1.0

0.5

0.0

yc l+ BE IS Z + O BE IS Z b+ BE Z O

IS b C

yc l

yc l+ BE IS Z +B O IS EZ b+ BE Z O

IS b C

yc l

yc l+ BE IS Z +B O IS EZ b+ BE Z O

IS b

yc l

Supplementary Figure 2. Peptide methionine oxidation Two peptides (a and b) observed to be oxidized show identical regulation for their unoxidized and oxidized state (top and bottom panel). No significant differences were observed between both states in any sample (P-value below 0.05), indicating similar oxidative conditions for both SIS and endogenous peptide versions in all samples.

158

PDK1

Chapter 7

AKT

Supplementary Figure 3. All STREM acquired phosphosite measurements in the PI3K-mTOR and MAPK pathway

159

TSQ_140401_ES803_S90327_35C_HDF+SIS_MIDAS4 #1013 RT: 28.69 AV: 1 NL: 2.46E5 F: + c NSI d Full ms2 458.552 [180.000-1100.000]

200

250

300

350

400

450

500

550 m/z

616.28 600

650

y8-3H2O 707.34 700

802.98 751.21

840.03 885.52

750

800

850

TSQ_140401_ES803_S90327_35C_HDF+SIS_MIDAS6 #2686 RT: 39.44 AV: 1 NL: 2.75E3 F: + c NSI d Full ms2 727.308 [180.000-1400.000]

678.36

339.37

b2281.02 a 2+

10 0

270.56

395.79 433.56

200

300

400

579.18

548.39

499.05

249.20

726.42

562.18

b92+-2P-H2O

b3-P, b62+-2P

b5-2P b92+-P-H2O

500

737.89 750.15 600

m/z

700

872.96

929.96

1006.35 1069.55 990.48

854.37

800

b10-3P

MH2+-3P-2H2O

MH2+-3P

Relative Abundance

RPS6_S235,S236,S240: RLSSLRASTSK2+

MH2+-P

MH2+-2P 629.52 638.92 y6 -P

SRM-MS/MS RT: 39.44

100

666.91

552.07

MH2+-P-2H2O

210.55

608.86

y8-P

347.53

304.51

256.28

521.01

a6-P

387.51

y102+, a5

458.42

y92+, a92+

425.99

y82+, b92+-2P a102+-2P b102+-2P b5-P-H2O

RPS6_S235,S240: RLSSLRASTSK3+

MH3+-P, y72+

MH3+-2P

MH3+

y42+, a63+ b2, b52+-P y52+ a52+ y3-NH3 b62+-P

Relative Abundance

b62+-H2O, MH3+-2P-NH3

393.34

SRM-MS/MS RT: 28.69

b93+

b32+-NH3 a63+-P

900

1000

1100

TSQ_140401_ES803_S90327_35C_HDF+SIS_MIDAS6 #2723 RT: 39.56 AV: 1 NL: 6.07E3 F: + c NSI d Full ms2 485.208 [180.000-1400.000] 100 90

MH3+-3P

387.03

SRM-MS/MS RT: 39.56

RPS6_S235,S236,S240: RLSSLRASTSK3+

200

250

-2P-H2O y9 -P, b

-3P

523.11 543.18

639.17

310.59 300

551.77 350

y8-2P

468.24

y6-P

451.74

339.19

2+ 10 2+

503.25

246.61 235.01

y93+-2P y83+

270.62

425.41

2+ 10

MH3+-P, y82+-P-H2O

b8 -3P, b4-2P MH3+-2P y72+

y42+-H2O

Relative Abundance

400

450

500

550 m/z

620.37 600

660.78 709.96 650

700

782.77 821.16 853.96 768.68 750

800

850

Supplementary Figure 4. Manually annotated spectra of the doubly and triply phosphorylated RPS6 peptides Spectra of RLS[+80]SLRAS[+80]TSK2+ (a), RLS[+80]S[+80]LRAS[+80]TSK2+ (b) and RLS[+80]S[+80]LRAS[+80] TSK3+ (c) were recorded in SRM triggered MS/MS runs and the most abundant fragment ions were annotated using the Protein Prospector software package (Baker, P.R. and Clauser, K.R. http://prospector.ucsf.edu).

160

足足足

CHAPTER 8 GENERAL DISCUSSION AND PERSPECTIVE

163

Chapter 8

GENERAL DISCUSSION AND PERSPECTIVE In recent years, substantial evidence has emerged that cellular senescence acts as a potent antitumor mechanism. Replicative senescence stops the expansion of aged cells that have exhausted their proliferative potential. Similarly, premature senescence in response to harmful stresses, such as oncogenic signaling, actively halts the proliferation of cells at risk of oncogenic transformation. In fact, various precancerous lesions including pulmonary adenomas, prostate intraepithelial neoplasia, lymphomas and mammary tumors, show senescence biomarkers1,2. In these settings, a â&#x20AC;&#x2DC;driverâ&#x20AC;&#x2122; mutation triggers the activation of an oncogene or the loss of a tumor suppressor gene, setting in motion a program that contributes to the formation of a benign lesion. The senescence response manifests itself after an initial phase of cell proliferation, halting further expansion. Progression towards malignancy in this case can happen only in the event of additional tumorigenic alterations. In the work described in this thesis, by studying the mechanisms crucial for maintenance of oncogeneinduced senescence (OIS), we aimed to identify genes and signaling pathways involved in senescence escape and thereby find factors contributing to oncogenic transformation. Senescence associates with a distinct metabolic profile In Chapter 1 of this thesis, we reviewed the current knowledge on metabolic changes in malignant cells. Recent studies demonstrate that the deregulation of cellular metabolism is a key factor in driving oncogenic transformation. Cancer cells upregulate aerobic glycolysis and shift their metabolism towards biosynthesis, thereby providing the energy and building blocks that are necessary for the creation of progeny and tumor expansion (reviewed in Chapter 1). As described in Chapter 2, the molecular mechanisms underlying metabolic reprogramming are complex and require an alteration(s) in multiple growth pathways with PI3K/AKT/mTOR signaling as a prime example. Direct communication between the components of the cell cycle machinery and metabolic enzymes also plays a role in metabolic rewiring to meet proliferative needs. In spite of the widely recognized importance of OIS and OIS escape in development of cancer, only few studies have explored the role of cellular metabolism in this setting (described in Chapter 1). Through an unbiased and comprehensive analysis of cellular metabolism we found that OIS cells show a distinct metabolic profile (described in Chapter 3). Entry into OIS associates with an increased rate of glucose oxidation in mitochondria, lower uptake of glutamine and a higher rate of fatty acid (FA) secretion. As described below, in this thesis we have revealed a pivotal role of oxidative metabolism in OIS and tumorigenesis (described in Chapter 4). However, the regulation of glutamine and lipid metabolism in senescence and escape thereof requires further studies. Notably, changes in glutamine and lipid metabolism in senescence directly oppose those in cancer. In contrast to lower glutamine utilization found in senescent cells, cancer cells increase glutamine uptake and utilization by upregulating ASCT2 and SN2 glutamine transporters and glutaminase 1 (GLS1), a first

enzyme of glutaminolysis3,4. In fact, cancer cells are dependent on glutamine for survival5. Along these lines, the silencing of GLS1 delays tumor growth6-8 and glutamine withdrawal decreases the viability of cancer cells9. The observation of lower glutamine utilization in OIS raises the question as to how components of glutaminolysis are regulated in senescent cells. Interestingly, the pharmacological inhibition of GLS1 in endothelial cells leads to a growth arrest that is associated with the induction of several senescence markers including senescence-associated Î˛-galactosidase (SA-Î˛-Gal) activity and p16INK4A and p21Cip1 protein levels10. This suggests that GLS1 levels and/or its activity are downregulated in OIS. Alternatively, lower levels of glutamine transporters might prevent the uptake of glutamine in senescence. Studying the regulation of the particular components of glutaminolysis in OIS should be accompanied by stable isotope labeling with C13 labeled glutamine tracer. While several of such studies have been performed in cancer cells, a global overview of glutamine metabolism in OIS has not been reported yet (and conceivably several mechanisms exist depending on the oncogenic insult). Finally, an important question that needs to be answered is whether glutamine metabolism has a causal role in OIS and is not merely a consequence of cell cycle arrest. This is of particular interest since lower utilization of glutamine has been linked to quiescence11. Although a cancerâ&#x20AC;&#x2122;s dependence on glutamine in combination with the antitumor function of OIS suggests an instrumental role of glutamine in the regulation of malignant transformation, functional studies are yet to be performed. By means of metabolic profiling we found that OIS cells have a higher rate of FA secretion. Similarly, others have recently reported an increase in the steady-state levels of free FA in senescence12. Detailed analysis of FA metabolism in that study revealed that the upregulation of FA levels was not due to increased FA synthesis, which in fact was lower in senescent cells, but rather caused by a higher rate of FA oxidation. Taking into account the antitumor function of OIS, a decrease in FA biosynthesis is in line with the observation that cancer cells commonly upregulate de novo FA synthesis by increasing activity and expression of several lipogenic enzymes13. Yet, cancer cells similarly to OIS cells have also been shown to upregulate FA oxidation. The expression of carnitine palmitoyltransferase-1 isoform C (CPT1), an enzyme activating FA oxidation, is increased in cancer cells which stimulates FA oxidation-derived ATP production and resistance to glucose deprivation14. This suggests that increased FA oxidation in OIS is just a consequence of the activation of an oncogene rather than an important mechanism regulating OIS. In agreement with that, the inhibition of FA oxidation upon CPT1 depletion did not prevent a senescence-associated cell cycle arrest12. In light of these findings, a more detailed analysis of FA metabolism is required to understand its exact role in senescence and malignant transformation. Oxidative metabolism underlies OIS program In Chapter 4 of this thesis, by integrating metabolic profiling with functional analyses, we provide evidence that increased oxidative metabolism is a direct mediator, rather than

164

165

Chapter 8

only a phenomenon associated with OIS. We demonstrate that the rate of conversion of pyruvate to citrate, a combined reaction of pyruvate dehydrogenase (PDH) and citrate synthase (CS) is highly increased in OIS triggered by the mutant BRAF oncogene. This causes higher tricarboxylic acid (TCA) cycle activity and increased respiration. While PDH has been shown to be a central metabolic regulator in diabetes15,16, heart disease17,18 and more recently has been suggested to contribute to cancer19-21, its regulation and function in the context of cellular senescence had yet to be characterized and understood. We show that PDH is activated in OIS due to the deregulation of PDH regulatory enzymes: PDK1 and PDP2. Conversely, the abrogation of OIS concurs with PDH inhibition. Finally, normalization of PDK1 or PDP2 levels inhibits PDH and leads to OIS escape. These data, taken together, strongly support a critical role of PDH in OIS and establish a functional link between OIS and metabolic (de)regulation. Our study not only provides novel insights into the role and regulation of metabolic rewiring in senescence, but also points to the existence of close communication between OIS machinery and a key mitochondrial signaling axis. Yet, several aspects of this interaction remain to be resolved. For example, it is still uncertain how BRAFV600E deregulates PDP2 and PDK1 transcription. Previously, PDK1 has been shown to be a HIF-1α -responsive gene22,23. Among many genes regulated by the HIF-1α protein, a large subset is associated with cellular senescence, including cell-cycle regulators24,25 and several members of the senescencemessaging secretome (SMS)26. However, our previous study did not show a clear HIF1αtranscriptome signature in cells bypassing OIS27. This suggests that in OIS oncogenic BRAF acts on PDK1 via another pathway. Interestingly, PDK1 can be regulated also on the level of post-translational modification. Specifically, oncogenic tyrosine kinases such as FGFR1, were reported to localize in the mitochondria, where they phosphorylate and activate PDK1, thereby promoting cancer cell metabolism and tumor progression21. Although changes in PDK1 phosphorylation in OIS were not analyzed in our study, it is possible that, next to being regulated at the transcriptional level, PDK1 is also controlled by phosphorylation. One of the potential mediators of PDK1 phosphorylation is MEK, a tyrosine/threonine kinase acting downstream of BRAF. Such regulation of PDK1 could explain our finding that PDH is activated in RAS-induced senescence, even in the absence of changes in PDK1 and PDP2 expression levels. In addition to this, the precise nature and the mechanism of the connection between the higher mitochondrial respiration, redox stress and senescence program await better understanding. An increased production of reactive oxygen species (ROS) has been reported to induce and mediate RAS-induced senescence28 and has been associated with BRAFinduced senescence29. Similar to this, our study demonstrated that an increase in respiration associates with a rise in redox stress in OIS cells. Although only correlative, it is consistent with the idea that ROS induction is an important mediator of senescence. Even so, functional

studies, in particular on how antioxidants affect OIS, are required to demonstrate that ROS is a cause rather than a consequence of senescence. Our observation of increased TCA cycle activity in OIS raises the question as to whether any of the TCA cycle metabolites have a direct function in regulating senescence program. In fact, acetyl-CoA is not only an important intermediate for macrosyntheses, but it also a precursor for the acetylation of proteins. For example, acetylation of histones has been shown to be dependent on the acetyl-CoA levels30,31 and is an essential process in the release of DNA for replication and therefore cell cycle progression32,33. Although in our analyses we did not directly measure acetyl-CoA, higher PDH activity predicts an increase in acetyl-CoA levels. Thus, by regulating histone acetylation, acetyl-CoA might control the cell cycle and thereby the senescence program. Our study on oxidative metabolism and PDH regulation was focused on only one senescence type, OIS. Nevertheless, also therapy-induced senescence (TIS) has been recently associated with enhanced glucose utilization in the TCA cycle34. Along these lines, growth arrest upon inhibition of the melanoma-driver BRAF is accompanied by the induction of several senescence features35 and an increase in oxidative metabolism36,37. These studies show that enhanced oxidative metabolism is not restricted to OIS, but in fact represents a general senescence feature. Yet, whether PDK1-PDP2-PDH axis (de)regulation plays role in different types of senescence, remains to be answered. PDK1 as a potential therapeutic target Increased tumorigenicity correlates with a shift from oxidative phosphorylation towards glycolysis38. Similarly, fully transformed cancer cells are known to shut down oxidative metabolism in favor of lactate production, a phenomenon called aerobic glycolysis or â&#x20AC;&#x153;the Warburg effectâ&#x20AC;? (also described in Chapter 1)39,40. In this context, the increased oxidative metabolism upon PDH activation in OIS counteracts the tumor-supporting metabolic profile. Considering that for malignant transformation cells need to evade the senescence program, pathways crucial for senescence likely represent factors (de)regulated in tumors. In Chapter 4, we show that (de)regulation of PDH not only controls OIS, but also represents an important factor in malignant transformation. We find that ectopic expression of the PDH inactivating kinase PDK1 promotes melanoma growth in mice. Conversely, PDK1 depletion leads to cell death of melanoma cells, both in vitro and in vivo. Remarkably, depletion of PDK1 both prevents outgrowth of tumors and, perhaps more importantly from a clinical point of view, causes regression of established human melanomas. These results not only indicate a pro-oncogenic capacity of PDK1, but also reveal that PDK1 may serve as a potential novel therapeutic target in melanoma. The observation that PDK1 depletion induces senescence in primary cells but cell death in melanoma cells calls for more clarity about the mechanism responsible for these different sensitivities. One explanation might be a distinct level of dependence on glycolysis between cancer and non-transformed cells (reviewed in Chapter

166

167

Chapter 8

1). PDH activation upon PDK1 depletion diverts pyruvate into the mitochondrial TCA cycle. This indirectly suppresses glycolysis and thereby selectively kills cancer cells highly addicted to glycolysis. Alternatively, the activation of mitochondrial respiration might cause much higher redox stress in cancer cells than that in non-transformed cells. In fact, melanoma cells were reported to have elevated levels of mitochondrial respiration compared to melanocytes and to be largely dependent on mitochondrial activity for energy production41,42. Hence, further stimulation of oxidative metabolism upon PDH reactivation likely overloads the capacity of the oxidative chain, thereby boosting ROS production and melanoma cell death. The finding that pharmacological inhibition of PDKs with dichloroacetic acid (DCA) in melanoma potentiate the antitumor effects of a pro-oxidative drug elesclomol further supports such a scenario43. Having demonstrated the potential therapeutic function of PDK1 inhibition, we also studied how this result can aid current melanoma therapy, especially in the context of targeted BRAFV600E inhibition. This is of particular interest as, although targeted inhibition of melanoma-driving BRAF pathway initially causes a substantial tumor regression, most melanomas eventually become resistant and patients surrender to recurrent disease44-46. Remarkably, PDK1 depletion causes synergistic toxicity with targeted BRAF inhibition, even eliminating cultured melanoma cell populations resistant to the BRAF inhibitor. This observation, together with the fact that PDK1 is an important metabolic regulator, suggests a role for metabolic reprogramming in the emerging of resistance. In fact, several studies have recently indicated that targeted inhibition of BRAF pathway acts on the metabolic level. For example, BRAF inhibition in melanoma downregulates glucose transporters GLUT1, GLUT3 and the first enzyme of glycolysis, hexokinase 2 (HK2)47, thereby potently suppressing the uptake of glucose48-50. At the same time, BRAF inhibition activates the oxidative metabolism, leading to increased mitochondrial respiration and ROS production36,37,51. While a decrease in glucose metabolism is restored upon development of BRAF inhibitor resistance, drugresistant melanomas show high rates of mitochondrial respiration and oxidative stress, regardless of the presence of BRAF inhibitor. Importantly, high oxidative metabolism renders resistant melanomas prone to cell death, induced by pro-oxidants including the clinical trial drug elesclomol47,51. All of this together indicates that PDH is activated in response to BRAF pathway inhibition in melanoma. In this regard, PDK1 depletion in presence of BRAF inhibitor would potentiate the activation of PDH, oxidative metabolism and thereby redox stress. Although this is likely to be a mechanism of synergy between PDK1 and BRAF inhibition, it requires validation, especially as information on PDK-PDP-PDH axis regulation in response to BRAF inhibition is lacking. Our study on the role of PDK1 in response to BRAF inhibition does not go beyond cell culture. Clearly, an in vitro situation does not recapitulate the metabolic tumor microenvironment52 nor the tumor cell heterogeneity that exists in vivo53. Therefore, a validation of synergy

between PDK1 and BRAF inhibition in melanoma cells killing in in vivo models is a critical point to be addressed. Particularly important from a therapeutic point of view, it should be recapitulated with the use of specific PDK1 inhibitors instead of shRNA-mediated PDK1 depletion. Several PDKs inhibitors that could be used for that purpose are commercially available54. Although these inhibitors are not fully selective for PDK1 alone but act also on other PDK isoforms, the fact that they all activate PDH justifies their use, at least in first proof-of-concept experiments. Finally, an important question to be answered is how broad the effectiveness of the inhibition of PDK1 as a therapeutic target exactly is. Since we found that PDK1 depletion synergizes with BRAF inhibition in melanoma and because BRAF mutations are also found in thyroid, colonic and ovarian carcinoma55-57, these tumor types would be the first choice to test the effect of PDK1 inhibition. But any therapeutic value of PDK1 regulation might go beyond BRAF-driven tumors. The fact that PDK1 inhibition has been reported to decrease viability of small cell lung cancer, breast cancer and glioblastoma cancer cells19,58,59 implies that this approach should be explored for several cancer settings. Other senescence mediators with a possible role in malignant transformation In Chapter 5, Chapter 6 and Chapter 7 we describe other approaches we undertook, next to metabolic profiling, to identify novel senescence mediators. With a function-based short hairpin (sh)RNA screen we identified seven genes contributing to OIS (Chapter 5). One of them, RASEF, was characterized in more detail, because we found it was frequently methylated in melanomas. Although we focused on RASEF, other genes also represent interesting potential tumor suppressors for further characterization. While RASEF depletion abrogated OIS, causing continued cell proliferation, restoration of its expression in melanoma acted cytostatically (in some melanoma cell lines). This is in agreement with a potential tumor suppressive role of RASEF. An important limitation of our study however, is that it does not go beyond cell culture. Clearly, a more definitive characterization of RASEF will require in vivo studies. To confirm a tumor suppressive function of RASEF in an in vivo setting, a knock-in BRAFV600E mouse (Tyr::CreER; BRaf CA) would be a suitable model60. This mouse model was previously used to validate the importance of PTEN loss for development of BRAF-driven melanoma61. Also unclear remains the mechanism of action of RASEF in OIS and malignant transformation. Interestingly, RASEF is required for mutant BRAF to drive the increase of IL6, IL8 and C/EBPÎ˛ transcripts, components of SMS contributing to OIS27. This observation implies communication between senescence-associated and C/EBPÎ˛dependent inflammatory cytokines, and signaling involving RASEF. Together, the results support a model in which RASEF controls OIS by regulating components of the senescence secretome. Additionally, we performed mass spectrometry-based screening of the proteome and phosphoproteome in cycling, senescent, and senescence-escaping cells. By this approach

168

In conclusion, our study provides important insight into the mechanism underlying OIS, a vital program preventing malignant transformation. By three independent approaches: metabolic profiling, shRNA screens and mass spectrometry analysis, we identify novel factors having a pivotal role in senescence. We find several seemingly unrelated elements including the metabolic regulators PDK-PDP-PDH and the signalling molecule RASEF to act together with components of the SMS and the cell cycle machinery to regulate the senescence program in a coordinated fashion. Moreover, we provide pre-clinical evidence that these senescence mediators play a role in malignant transformation. The major challenge is to further explore the feasibility of their targeting for the clinical intervention of cancer.

169

Chapter 8

we identified many factors involved in the extracellular matrix (ECM) remodeling to be specifically regulated in senescence (Chapter 6). Furthermore, we revealed novel sitespecific phosphorylation of components of PI3K/AKT/mTOR and RAS/MEK/ERK signaling network, previously undetectable by phospho-antibodies (Chapter 7). Interestingly, several studies have indicated important roles for components of ECM remodeling both in senescence and in cancer62,63. Therefore, functional studies on ECM organization and regulation could further increase our understanding of the mechanism underlying OIS and malignant transformation. Also, as PI3K/AKT/mTOR and RAS/MEK/ERK pathways are wellestablished regulators of both senescence and malignant transformation, efforts are being made to efficiently target their components for clinical use64. In this regard, by providing novel mechanistic insight into the signaling pathways involved, our study contributes to finding specific inhibitors for anticancer therapy. A feedback loop comprising cytokines, metabolic regulators and transcription factors controls OIS We and others have demonstrated that senescent cells adapt a robust inflammatory transcriptome signature (called SMS or SASP)27,62,65-69. The C/EBPβ transcription factor is a major player for the establishment of the resulting secretory phenotype27. Previously, we have shown that TSC22, a transcription factor and critical OIS mediator, acts downstream of C/EBPβ on the inflammatory secretome70. Here, we demonstrate that also components of PDK-PDP-PDH axis influence the SMS (Chapter 4). Inhibition of PDH activity leads to a decrease in the expression levels of IL6, IL8 and C/EBPβ transcripts. Conversely, C/EBPβ depletion averts PDH activation by preventing BRAFV600E-induced changes in the expression of both PDP2 and PDK1. Furthermore, we find that yet another mediator of OIS identified in this thesis, RASEF, also regulates senescence-associated and C/EBPβ-dependent inflammatory cytokines (Chapter 5). Taken together, these findings not only highlight the central role of specific interleukins in OIS but also suggest the existence of a complex autostimulatory feedback mechanism, in which cytokines, metabolic regulators and transcription factors operate to coordinate the control of OIS.

REFERENCES 1. 2. 3.

4. 5.

9. 10.

11. 12.

13. 14.

170

Collado, M. & Serrano, M. Senescence in tumours: evidence from mice and humans. Nat Rev Cancer 10, 51–57 (2010). Kuilman, T., Michaloglou, C., Mooi, W. J. & Peeper, D. S. The essence of senescence. Genes Dev 24, 2463–2479 (2010). Wise, D. R. et al. Myc regulates a transcriptional program that stimulates mitochondrial glutaminolysis and leads to glutamine addiction. PNAS 105, 18782– 18787 (2008). Nicklin, P. et al. Bidirectional Transport of Amino Acids Regulates mTOR and Autophagy. Cell 136, 521–534 (2009). DeBerardinis, R. J. & Cheng, T. Q’s next: the diverse functions of glutamine in metabolism, cell biology and cancer. Oncogene 29, 313–324 (2010). Lobo, C. et al. Inhibition of glutaminase expression by antisense mRNA decreases growth and tumourigenicity of tumour cells. Biochem J 348 Pt 2, 257–261 (2000). Gao, P. et al. c-Myc suppression of miR23a/b enhances mitochondrial glutaminase expression and glutamine metabolism. Nature 458, 762–765 (2009). Wang, J.-B. et al. Targeting mitochondrial glutaminase activity inhibits oncogenic transformation. Cancer Cell 18, 207–219 (2010). Zamboni, N. Deficiency in glutamine but not glucose induces MYC-dependent apoptosis in human cells. 178, 93–105 (2007). Unterluggauer, H. et al. Premature senescence of human endothelial cells induced by inhibition of glutaminase. Biogerontology 9, 247–259 (2008). Lemons, J. M. S. et al. Quiescent fibroblasts exhibit high metabolic activity. PLoS Biol. 8, e1000514 (2010). Quijano, C. et al. Oncogene-induced senescence results in marked metabolic and bioenergetic alterations. cc 11, 1383–1392 (2012). Santos, C. R. & Schulze, A. Lipid metabolism in cancer. FEBS J 279, 2610–2623 (2012). Faubert, B., Berger, S. L., Jones, R. G., Thompson, C. B. & Mak, T. W. Carnitine palmitoyltransferase 1C promotes cell survival and tumor growth under conditions of metabolic stress. 25, 1041–1051 (2011).

15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

Wu, P., Inskeep, K., Bowker-Kinley, M. M., Popov, K. M. & Harris, R. A. Mechanism responsible for inactivation of skeletal muscle pyruvate dehydrogenase complex in starvation and diabetes. Diabetes 48, 1593– 1599 (1999). Sugden, M. C. & Holness, M. J. Therapeutic potential of the mammalian pyruvate dehydrogenase kinases in the prevention of hyperglycaemia. Curr. Drug Targets Immune Endocr. Metabol. Disord. 2, 151–165 (2002). Lewandowski, E. D. & White, L. T. Pyruvate dehydrogenase influences postischemic heart function. Circulation 91, 2071–2079 (1995). Terrand, J., Papageorgiou, I., RosenblattVelin, N. & Lerch, R. Calcium-mediated activation of pyruvate dehydrogenase in severely injured postischemic myocardium. Am. J. Physiol. Heart Circ. Physiol. 281, H722–30 (2001). Bonnet, S. et al. A mitochondria-K+ channel axis is suppressed in cancer and its normalization promotes apoptosis and inhibits cancer growth. Cancer Cell 11, 37–51 (2007). Mcfate, T. et al. Pyruvate dehydrogenase complex activity controls metabolic and malignant phenotype in cancer cells. J Biol Chem 283, 22700–22708 (2008). Hitosugi, T. et al. Tyrosine phosphorylation of mitochondrial pyruvate dehydrogenase kinase 1 is important for cancer metabolism. Mol Cell 44, 864–877 (2011). Papandreou, I., Cairns, R. A., Fontana, L., Lim, A. L. & Denko, N. C. HIF-1 mediates adaptation to hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab 3, 187–197 (2006). Kim, J.-W., Gao, P., Liu, Y.-C., Semenza, G. L. & Dang, C. V. Hypoxia-inducible factor 1 and dysregulated c-Myc cooperatively induce vascular endothelial growth factor and metabolic switches hexokinase 2 and pyruvate dehydrogenase kinase 1. Mol Cell Biol 27, 7381–7393 (2007). Koshiji, M. et al. HIF-1alpha induces cell cycle arrest by functionally counteracting Myc. EMBO J 23, 1949–1956 (2004). Goda, N. et al. Hypoxia-inducible factor 1alpha is essential for cell cycle arrest during hypoxia. Mol Cell Biol 23, 359–369 (2003).

27.

28.

29.

30. 31.

32. 33. 34. 35. 36. 37.

38.

39.

Welford, S. M. & Giaccia, A. J. Hypoxia and Senescence: The Impact of Oxygenation on Tumor Suppression. Molecular Cancer Research 9, 538–544 (2011). Kuilman, T. et al. Oncogene-Induced Senescence Relayed by an InterleukinDependent Inflammatory Network. Cell 133, 1019–1031 (2008). Lee, A. C. et al. Ras proteins induce senescence by altering the intracellular levels of reactive oxygen species. J Biol Chem 274, 7936–7940 (1999). de Keizer, P. L. J. et al. Activation of forkhead box O transcription factors by oncogenic BRAF promotes p21cip1-dependent senescence. Cancer Res 70, 8526–8536 (2010). Wellen, K. E. et al. ATP-citrate lyase links cellular metabolism to histone acetylation. Science 324, 1076–1080 (2009). Galdieri, L. & Vancura, A. Acetyl-CoA carboxylase regulates global histone acetylation. Journal of Biological Chemistry 287, 23865–23876 (2012). Berger, S. L. The complex language of chromatin regulation during transcription. Nature 447, 407–412 (2007). Li, B., Carey, M. & Workman, J. L. The role of chromatin during transcription. Cell 128, 707–719 (2007). Dörr, J. R. et al. Synthetic lethal metabolic targeting of cellular senescence in cancer therapy. Nature 501, 421–425 (2013). Haferkamp, S. et al. Vemurafenib Induces Senescence Features in Melanoma Cells. J Investig Dermatol 133, 1601–1609 (2013). Haq, R. et al. Oncogenic BRAF Regulates Oxidative Metabolism via PGC1&alpha; and MITF. Cancer Cell 23, 302–315 (2013). Vazquez, F. et al. PGC1α expression defines a subset of human melanoma tumors with increased mitochondrial capacity and resistance to oxidative stress. Cancer Cell 23, 287–301 (2013). Ramanathan, A., Wang, C. & Schreiber, S. L. Perturbational profiling of a cell-line model of tumorigenesis by using metabolic measurements. Proc Natl Acad Sci USA 102, 5992–5997 (2005). Warburg, O., Wind, F. & Negelein, E. The metabolism of tumors in the body. The Journal of general physiology 8, 519–530 (1927).

40. 41. 42. 43.

44. 45.

46. 47.

48.

49.

50.

51.

52.

WARBURG, O. On the origin of cancer cells. Science 123, 309–314 (1956). Barbi de Moura, M. et al. Mitochondrial respiration--an important therapeutic target in melanoma. PLoS ONE 7, e40690 (2012). Ho, J. et al. Importance of glycolysis and oxidative phosphorylation in advanced melanoma. Mol Cancer 11, 76 (2012). Kluza, J. et al. Inactivation of the HIF-1α/ PDK3 signaling axis drives melanoma toward mitochondrial oxidative metabolism and potentiates the therapeutic activity of prooxidants. Cancer Res 72, 5035–5047 (2012). Solit, D. B. & Rosen, N. Resistance to BRAF inhibition in melanomas. N Engl J Med 364, 772–774 (2011). Chapman, P. B. Mechanisms of resistance to RAF inhibition in melanomas harboring a BRAF mutation. Am Soc Clin Oncol Educ Book (2013). doi:10.1200/EdBook_AM.2013.33. e80 Jarkowski, A. & Khushalani, N. I. BRAF and beyond: Tailoring strategies for the individual melanoma patient. J Carcinog 13, 1 (2014). Parmenter, T. J. et al. Response of BRAFMutant Melanoma to BRAF Inhibition Is Mediated by a Network of Transcriptional Regulators of Glycolysis. Cancer Discov 4, 423–433 (2014). McArthur, G. A. et al. Marked, homogeneous, and early [18F] fluorodeoxyglucose-positron emission tomography responses to vemurafenib in BRAF-mutant advanced melanoma. J Clin Oncol 30, 1628–1634 (2012). Baudy, A. R. et al. FDG-PET is a good biomarker of both early response and acquired resistance in BRAFV600 mutant melanomas treated with vemurafenib and the MEK inhibitor GDC-0973. EJNMMI Res 2, 22 (2012). Carlino, M. S. et al. (18)F-labelled fluorodeoxyglucose-positron emission tomography (FDG-PET) heterogeneity of response is prognostic in dabrafenib treated BRAF mutant metastatic melanoma. Eur. J. Cancer 49, 395–402 (2013). Corazao Rozas, P., Guerreschi, P., Jendoubi, M. & André, F. Mitochondrial oxidative stress is the achille’s heel of melanoma cells resistant to Braf-mutant inhibitor. Oncotarget (2013). Vaupel, P., Kallinowski, F. & Okunieff, P. Blood flow, oxygen and nutrient supply,

171

Chapter 8

26.

53.

54.

55. 56. 57.

58. 59.

60. 61.

62. 63.

64.

65. 66.

172

and metabolic microenvironment of human tumors: a review. Cancer Res 49, 6449–6465 (1989). Salk, J. J., Fox, E. J. & Loeb, L. A. Mutational heterogeneity in human cancers: origin and consequences. Annu Rev Pathol 5, 51–75 (2010). Kato, M., Li, J., Chuang, J. L. & Chuang, D. T. Distinct Structural Mechanisms for Inhibition of Pyruvate Dehydrogenase Kinase Isoforms by AZD7545, Dichloroacetate, and Radicicol. Structure 15, 992–1004 (2007). Davies, H. et al. Mutations of the BRAF gene in human cancer. Nature 417, 949–954 (2002). Xing, M. BRAF mutation in thyroid cancer. Endocr. Relat. Cancer 12, 245–262 (2005). Di Nicolantonio, F. et al. Wild-type BRAF is required for response to panitumumab or cetuximab in metastatic colorectal cancer. J Clin Oncol 26, 5705–5712 (2008). Michelakis, E. D. et al. Metabolic modulation of glioblastoma with dichloroacetate. Science Translational Medicine 2, 31ra34 (2010). Sun, R. C. et al. Reversal of the glycolytic phenotype by dichloroacetate inhibits metastatic breast cancer cell growth in vitro and in vivo. Breast Cancer Res. Treat. 120, 253–260 (2010). Dankort, D. et al. Braf(V600E) cooperates with Pten loss to induce metastatic melanoma. Nat Genet 41, 544–552 (2009). Vredeveld, L. C. W. et al. Abrogation of BRAFV600E-induced senescence by PI3K pathway activation contributes to melanomagenesis. Genes Dev 26, 1055–1069 (2012). Kuilman, T. & Peeper, D. S. Senescencemessaging secretome: SMS-ing cellular stress. Nat Rev Cancer 9, 81–94 (2009). Gilkes, D. M., Semenza, G. L. & Wirtz, D. Hypoxia and the extracellular matrix: drivers of tumour metastasis. Nat Rev Cancer 14, 430–439 (2014). Ras/Raf/MEK/ERK and PI3K/PTEN/Akt/mTOR Inhibitors: Rationale and Importance to Inhibiting These Pathways in Human Health. 2, 135–164–164 (2011). Acosta, J. C. et al. Chemokine signaling via the CXCR2 receptor reinforces senescence. Cell 133, 1006–1018 (2008). Wajapeyee, N., Serra, R. W., Zhu, X., Mahalingam, M. & Green, M. R. Oncogenic BRAF Induces Senescence and Apoptosis

67.

68.

69.

70.

through Pathways Mediated by the Secreted Protein IGFBP7. Cell 132, 363–374 (2008). Coppé, J.-P. et al. Senescence-associated secretory phenotypes reveal cellnonautonomous functions of oncogenic RAS and the p53 tumor suppressor. PLoS Biol. 6, 2853–2868 (2008). Rodier, F. et al. Persistent DNA damage signalling triggers senescence-associated inflammatory cytokine secretion. Nat Cell Biol 11, 973–979 (2009). Acosta, J. C. et al. A complex secretory program orchestrated by the inflammasome controls paracrine senescence. Nat Cell Biol 15, 978–990 (2013). Hömig-Hölzel, C. et al. Antagonistic TSC22D1 variants control BRAF(E600)-induced senescence. EMBO J 30, 1753–1765 (2011).

足足足

ADDENDUM SUMMARY SAMENVATTING

CURRICULUM VITAE

LIST OF PUBLICATIONS

PHD PORTFOLIO

ACKNOWLEDGEMENTS

175

Addendum

SUMMARY Cancer is the result of a multistep process involving the gradual change of hea足足足lthy cells into immortal and often invasive cells. This process is called oncogenic transformation and requires the evasion of cell-intrinsic failsafe mechanisms that normally limit uncontrolled proliferation. Induction of a permanent cells cycle arrest in response to oncogenic stress, a process termed oncogene-induced senescence (OIS), represents such a built-in tumorsuppressive mechanism. Indeed, work of recent years clearly shows that OIS acts as a powerful pathophysiologic mechanism suppressing cancer, both in animal models and in humans. Also, it has become clear that factors controlling OIS are endowed with tumorsuppressive functions and are often absent or mutated in cancer. In this thesis, we have studied genes and signaling pathways critical for OIS in order to find factors important for oncogenic transformation. We have also made an attempt, wherever possible, to translate our findings to a clinically relevant setting. In Chapter 1, we introduce major cellular metabolic pathways and give an overview of metabolic changes occurring in malignant cells. Besides that, we review the current knowledge on metabolic regulation in senescence. To date, only few studies on that subject have been reported, showing that senescent cells rewire their metabolism in a way that counteracts changes seen in the cancer cells. Hence, the antitumor function of senescence is manifested also on the level of metabolic regulation. In Chapter 2, we give an account of how proliferation pathways, the cell cycle machinery and metabolism are interconnected. We illustrate that the molecular mechanisms underlying metabolic reprogramming involve not only alterations in multiple proliferation pathways, with PI3K/AKT/mTOR signaling as a prime example, but also direct communication between the components of the cell cycle machinery and metabolic enzymes. Chapter 3 presents our analysis of metabolic changes in OIS. Through mass balance analysis and metabolic flux profiling we show that OIS cells have a distinct metabolic profile when compared to cycling cells. Specifically, entry into OIS is associated with an increased rate of pyruvate oxidation in mitochondria, lower utilization of glutamine and a higher rate of fatty acid secretion. Notably, these changes in metabolism oppose those in cancer. In Chapter 4, we demonstrate that increased glucose oxidation in mitochondria in OIS is regulated by the mitochondrial gatekeeper pyruvate dehydrogenase (PDH). In OIS, PDH is activated upon downregulation of its inhibitory kinase PDK1 and simultaneous upregulation of PDH-activating phosphatase PDP2. Importantly, the abrogation of OIS, a rate-limiting step towards oncogenic transformation, coincides with reversion of these processes. This shows that PDH (de)regulation is a direct mediator of, rather than a phenomenon only associated with OIS. Further supporting a critical role of PDH in OIS, enforced normalization of either PDK1 or PDP2 expression levels inhibits PDH and abrogates OIS, thereby licensing melanoma development. Finally, depletion of PDK1 causes regression of established BRAF-driven

melanoma and eradicates melanoma subpopulations resistant to targeted BRAF inhibition. These results reveal a mechanistic relationship between OIS and a key metabolic signaling axis, which may be exploited therapeutically. Chapter 5, Chapter 6 and Chapter 7 describe unbiased screening approaches aimed at the identification of novel OIS regulators with a potential tumor suppressive function. In Chapter 5 we have used a function-based short hairpin (sh)RNA screen and identified seven genes, depletion of which abrogates senescence-associated cell cycle arrest. By genome-wide promoter methylation analysis we found that one of the genes, RASEF, is hypermethylated in melanomas when compared to senescent nevi. While RASEF depletion abrogates OIS, causing continued cell proliferation, restoration of its expression in melanoma acts cytostatically. This is in agreement with a potential tumor suppressive role of RASEF. In Chapter 6 and Chapter 7, we describe our mass spectrometry-based analysis of the proteome and phosphoproteome in cycling, senescent, and senescence-escaping cells. Among senescence-regulated proteins, we found several previously established senescence mediators, including cell cycle regulators and cytokines, but also multiple proteins previously not associated with the senescence program. In Chapter 8, we discuss our findings in more detail and speculate on their potential future implications. In addition to that, we formulate remaining and newly arisen questions concerning the mechanism controlling OIS and oncogenic transformation that need to be answered to better understand the complexity underlying tumorigenesis and to uncover new targets for therapeutic intervention.

176

177

Addendum

SAMENVATTING Kanker is het resultaat van een meerstaps proces van geleidelijke verandering van gezonde cellen in onsterfelijke cellen. Dit wordt ook wel oncogene transformatie genoemd en vereist het ontduiken van cel-eigen beschermingsmechanismen die ongecontroleerde proliferatie (groei) beperken. Het stoppen van de celcyclus in reactie op oncogene stress, ‘oncogengeïnduceerde senescence (OIS)’, is een voorbeeld van zo’n ingebouwd tumor-onderdrukkend mechanisme. Recent onderzoek, onder andere door onze onderzoeksgroep, heeft aangetoond dat OIS een belangrijke pathofysiologische kanker-onderdrukkende rol heeft, zowel in diermodellen als bij mensen. Daarnaast blijkt dat de factoren die OIS controleren ook tumor-onderdrukkende activiteit laten zien en vaak afwezig of gemuteerd zijn bij kanker. Om factoren te kunnen vinden die van belang zijn voor oncogene transformatie, hebben wij voor dit proefschrift OIS genen en ‘signaalpaden’ bestudeerd. In Hoofdstuk 1 introduceren we enkele belangrijke cellulaire stofwisselingsroutes en geven een overzicht van metabole veranderingen in tumorcellen. Ook bespreken we de huidige kennis over het metabolisme in cellen die in een staat van senescence verkeren. Enkele studies naar metabole regulatie in senescence geven aan dat het metabolisme van cellen in staat van senescence dusdanig wordt aangepast dat oncogene transformatie onmogelijk wordt gemaakt. Vandaar dat de antitumorfunctie van OIS zich ook manifesteert op het niveau van stofwisselingregulering. In Hoofdstuk 2 geven we een verslag van wat de relatie is tussen cellulaire groeipaden, de werking van de celcyclus en de samenhang daarvan met de regulering van metabolisme in cellen. We laten zien dat moleculaire mechanismen van metabolische herprogrammering niet alleen veranderingen in meerdere groeiwegen omvatten, met PI3K/AKT/mTOR signalen als een eerste voorbeeld, maar ook de directe communicatie tussen de componenten van het celcyclus-apparaat en metabole enzymen beïnvloeden. Hoofdstuk 3 bevat onze analyse van metabole veranderingen in OIS. Door middel van ‘massabalans-analyse’ en ‘metabolische flux profilering’ laten we zien dat OIS cellen een ander metabool profiel hebben dan groeiende cellen. Om precies te zijn, OIS gaat gepaard met een verhoogde oxidatie van glucose in de mitochondriën, minder opname van glutamine en een hoger percentage van vetzuuruitscheiding. Opmerkelijk is dat deze veranderingen in senescence tegenovergesteld zijn aan die van kanker. In Hoofdstuk 4 laten we zien dat de verhoogde glucose-oxidatie in mitochondriën in OIS wordt aangestuurd door de mitochondriale ‘poortwachter’ pyruvate dehydrogenase (PDH). In OIS is PDH geactiveerd wanneer gelijktijdig zijn remmende kinase PDK1 wordt afgeschakeld en de PDH activerende fosfatase PDP2 gestimuleerd wordt. Belangrijk is dat het ontsnappen aan OIS gepaard gaat met een omkering van deze processen. Dit toont aan dat PDH (de)regulering een directe schakel is in deze signaleringsroute, in plaats van slechts een met OIS geassocieerd fenomeen. Deze belangrijke rol voor PDH in OIS blijkt ook uit

het feit dat als we de niveaus van PDK1 of PDP2 normaliseren, PDH wordt geïnactiveerd. Zo wordt het OIS programma verstoord, en kunnen cellen uitgroeien tot een melanoom. Tenslotte, het onderdrukken van het niveau van PDK1 veroorzaakt regressie van melanomen met een BRAF mutatie en zelfs melanomen die resistent zijn geworden tegen remmers. Deze resultaten tonen een mechanistische relatie aan tussen OIS en een belangrijke metabole signalerings-as, die therapeutisch zou kunnen worden benut. Hoofdstukken 5, 6 en 7 beschrijven krachtige screeningsmethoden, gericht op het identificeren van onbekende OIS mechanismen met een potentiële tumor-onderdrukkende functie. In Hoofdstuk 5 hebben we met een zogenaamde short hairpin (sh) RNA screen zeven genen geïdentificeerd die, wanneer ze geïnactiveerd worden, de senescente staat van cellen opheffen. Met behulp van een genoom-brede promoter methylatie analyse hebben we ontdekt dat één van die genen, RASEF, gehypermethyleerd is in melanomen in vergelijking tot senescente naevi (moedervlekken). Terwijl het inactiveren van RASEF leidt tot opheffing van OIS, en daarmee tot ongebreidelde celproliferatie, werkt het herstellen ervan in melanoma cytostatisch (groeiremmend). Dit is in overeenstemming met een potentiële tumor-onderdrukkende rol van RASEF. In Hoofdstukken 6 en 7 beschrijven we onze op massa-spectrometrie gebaseerde analyse van proteoom en phosphoproteoom in delende, senescente en aan senescence ontsnapte cellen. Hiermee vonden we meerdere eiwitten met een tot nu toe onbekende functie in OIS. In Hoofdstuk 8 bespreken we de bevindingen uit dit proefschrift in meer detail en speculeren we over mogelijke toekomstige implicaties. Daarnaast formuleren we resterende en nieuw ontstane vragen over het mechanisme dat OIS en oncogene transformatie aanstuurt. Die vragen zullen beantwoord moeten worden om beter inzicht in de complexiteit van tumorvorming te krijgen en daarmee nieuwe aanknopingspunten voor therapeutische interventie te ontdekken.

178

Addendum

CURRICULUM VITAE Joanna Kaplon was born on July 20th, 1982 in Pulawy (Poland). After obtaining her high school diploma in 2001, she studied Biotechnology at the Warsaw University (Poland). In 2004, she received her Bachelor’s degree (cum laude) and followed the Master’s program Molecular Biology at Warsaw University. As part of the EU exchange program Erasmus, she subsequently did a six-month internship at the Department of Biology at the University of Groningen (The Netherlands), entitled “Role of RecQl4A in homologous recombination in Arabidopsis thaliana”. In 2005, she temporarily left Warsaw University to join the Topmaster Medical and Pharmaceutical Drug Innovation at the University of Groningen. During this two-year program, she performed three internships. First, she studied mechanisms that specify hematopoietic stem cells function, under the supervision of Dr. Leonid Bystrykh in the group of Prof. Dr. Gerald de Haan at the University Medical Center Groningen. For the second internship, she moved to Melbourne (Australia), where she worked in the group of Prof. Dr. David Huang at the Walter and Eliza Hall Institute of Medical Research on the role of Bax and Bak in mediating apoptosis. For this internship, she was awarded a Dutch Cancer Society (KWF) scholarship. Joanna performed her final Master’s internship at the Department of Medical Biology at the University Medical Center Groningen. Under the supervision of Dr. Bart Jan Kroesen, she studied the involvement of CD20-induced lipid raft dynamics in B-cell signaling. During that time, she was awarded a Keystone Symposia scholarship for participation in the meeting “Cell Signaling and Proteomics” (Colorado, USA). In 2007, she resumed her study at Warsaw University and received her Master’s degree (cum laude). In the same year, she obtained also her Master’s degree (cum laude) from the University of Groningen. After her graduation, Joanna joined the research group of Prof. Dr. Daniel Peeper at the Division of Molecular Oncology at the Netherlands Cancer Institute (NKI), where she investigated the relationship between cell metabolism, senescence and cancer. The results of her PhD work are presented in this thesis.

179

LIST OF PUBLICATIONS de Graaf, E.L.*, Kaplon, J.*, Shabaz, M., Vereijken, L.A.M., Duarte D.P., Gallego, L.R., Heck, A.J.R., Peeper, D.S., Altelaar, A.F.M. Signal transduction reaction monitoring deciphers sitespecific PI3K-mTOR/MAPK pathway dynamics. Manuscript submitted de Graaf, E.L., Kaplon, J., Zhou, H., Heck, A.J.R., Peeper, D.S., and Altelaar, A.F.M. (2014). Phosphoproteome Dynamics in Onset and Maintenance of Oncogene-induced Senescence. Mol. Cell Proteomics 13, 2089–2100 Kaplon, J., Hömig-Hölzel, C.*, Gao, L.*, Meissl, K., Verdegaal, E.M.E., van der Burg, S.H., van Doorn, R., and Peeper, D.S. (2014). Near-genomewide RNAi screening for regulators of BRAFV600E -induced senescence identifies RASEF, a gene epigenetically silenced in melanoma. Pigment Cell Melanoma Res 27, 640–652 Kaplon, J., Peeper, D.S., Inventor patent application no. WO2013109142 A1 named “Combined PDK and MAPK/ERK pathway inhibition in neoplasia”, Jul 25, 2013. Kaplon, J., Zheng, L.*, Meissl, K.*, Chaneton, B., Selivanov, V.A., Mackay, G., van der Burg, S.H., Verdegaal, E.M.E., Cascante, M., Shlomi, T., Gottlieb, E., and Peeper, D.S. (2013). A key role for mitochondrial gatekeeper pyruvate dehydrogenase in oncogene-induced senescence. Nature 498, 109–112

* these authors contributed equally to this work

180

Joanna Kaplon 2007-2014 Prof. dr. Daniel S. Peeper

I. PhD training Courses Year Workload (hrs) • Laboratory Animals Sciences course (Article 9) 2013 80 • “Klinische stage” for researchers 2012 80 • Writing and presenting in biomedicine 2012 32 • Statistics for Microarrays & High Throughput Sequencing 2010 10 • Introduction to R 2010 40 • Basic Medical Statistics 2010 24 • Gene Regulation Network 2010 80 • Drug Development 2008 80 • Replication stress and genome maintenance 2008 80 • Technology transfer: how to protect and utilize your 2007 8 scientific discovery Meetings and conferences Year • “Cancer and Metabolism”, De Rode Hoed, Amsterdam, 2013 The Netherlands (poster) • Keystone Symposia devoted to “Tumor Metabolism”, 2013 Keystone Resort, Keystone, Colorado, USA (poster) • Cold Spring Harbor meeting “Mechanisms and Models of Cancer”, 2012 Cold Spring Harbor, USA (oral presentation) • Prime Oncology Melanoma Conference, Paris, France 2012 (oral presentation) • CGC/ CBG meeting “Molecular mechanisms in cancer” 2010 Amsterdam, The Netherlands • MKFZ Conference “Metabolism in cancer”, 2010 Berlin, Germany (poster) • CNIO Cancer Conference “The energy of cancer”, Madrid, Spain 2009 • EMBO meeting, Amsterdam, The Netherlands 2009 • CGC/ CBG meeting “Molecular mechanisms and mouse models 2008 in cancer”, Amsterdam, The Netherlands • Beatson International Cancer Conference “Cell growth, 2008 metabolism and cancer”, Glasgow, Scotland 181

Addendum

PHD PORTFOLIO Name PhD student: PhD period: Name PhD supervisor:

• Annual graduate student retreat OOA, Texel, The Netherlands (poster and oral presentation) Other Visiting researcher at The Beatson Institute for Cancer Research, Group of Prof. dr. Eyal Gottlieb, Glasgow, UK

2007, 2008, 2009 Year Sept-Dec 2011

II. Teaching Supervising Year • Master student Loes van Dam 2013 Writing of Master Thesis on “Reciprocal regulation between metabolism and the cell cycle machinery: the molecular basis” • Bachelor student: Daniel Pereira Duarte 2012-2013 Project: Inhibition of pyruvate dehydrogenase kinase 1 (PDK1) in melanoma • Master student: Poornima Nair 2012 Project: Effects of NMRAL1, WT1 and GEMIN6 depletion on senescence biomarkers in the presence of an oncogenic stress

182

183

Addendum

ACKNOWLEDGMENTS It is done!!! Another chapter of my life has come to an end. Years of ups and downs, inspiring discussions, testing challenging hypotheses and satisfaction once experiments worked out well, but also endless weekends in the lab and frustration when a few-months-time effort ended up in the bin. Yes, motivation, hard work and good luck were certainly essential elements in my PhD journey. Yet, the successful completion of it would not be possible without help and support of many people inside and outside the NKI. Here, I would like to express my gratitude to all of you who were there for me when I needed you the most. First of all, I would like to acknowledge my promotor and supervisor. Daniel, thank you for an opportunity to work in your lab. I greatly appreciate the freedom that you gave me, whether it concerned the choice to work on difficult metabolic subjects, or the way to design my experiments. You have helped me to develop as a researcher and have shown that next to knowledge and critical attitude also diplomatic way of handling conflicts is important to be successful in science. I will never forget your faith in me and your support, especially during our ‘Nature battle’. Secondly, I am grateful to my copromotor Eyal. By opening Beatson’s doors for me, you let my metabolic dream to come true. Thank you for showing me the way in the jungle of isotope labeling and for teaching me how to transform metabolic nonsense to sense. I deeply admire your drive, knowledge, enjoyment in doing science and unlimited patience, virtues so crucial for dealing with research-embedded irritations (like “the pyruvate effect”) that made our metabolic adventure such a great success! Furthermore, I would like to acknowledge the members of my PhD committee: René Bernards, Maarten van Lohuizen and Huib Ovaa for keeping an eye on my progress during all these years, and the members of the reading committee: Anton Berns, Piet Borst, Albert Heck and Wolter Mooi for your revision of this thesis manuscript. Of course, this all would never have happened without my lab mates. Dear Katrin, working together with you was perhaps the happiest, the most educative and inspiring period of my PhD. Although we have quite different ways of performing experiments, I think we managed to complement each other perfectly. I cherish your support during metabolic-samplescollection weekends and stimulating discussions in the mouse house. I always felt that I could count on you, no matter what kind of issue it concerned. Your endless enthusiasm and constructive approach to science are truly impressive. Keep it up! I am proud to have you as my friend and paranymph and I wish you nothing but happiness in the future! Dear Sedef, although our projects did not have much in common, we shared a lot over the past few years. Thank you for all the fun we had during various NKI parties, Turkish gang’s New Year’s eves, Glasgow shopping experience, Q-day boat rides, Istanbul sight-seeing and our skiing trips, but also for sharing countless nights in the lab and all PhD-related stress. And of course for the best Hen Party ever! I am jealous (in a positive way) of your optimism and

your ability to say the suitable words in all situations. Thank you for being my friend and third paranymph. I wish you good luck with finishing your thesis. You are almost there! I would like to thank also other present and past members of the Peeper Lab for being great colleagues and friends. Patricia, before meeting you I did not believe that two such tempered people, like you and me, could ever interact. Still, all these years together in the lab, dinner dates, nights out, sailing weekends (actually “the one sailing weekend”) and even the “orange hair trip” have changed my mind. Although our relation has had ups and downs, I am really happy I have met you. Your advice and support meant a lot to me, especially the time when I wanted to give up. I wish you all the best and I hope our travelling in Brazil comes true! Cornelia, I bet that after leaving NKI you did not expect that this side-project screen of us would ever end up in a publication. After all, the enormous amount of both tissue cultures plates and hours in the crappy P1 lentilab, paid off. I miss these tissue culture sessions and even more cookie-baking parties and swimming lessons. Daniel D., it was nice to work with such a hardworking and accurate student like you. Good luck in choosing your career path! Liesbeth, thank you for taking me under your wings when I just started to fly in the lab. I learnt from you a lot, about science and non-science matters. I am sure your patience and honesty make you an excellent teacher! Judith, I truly value your critical attitude and advice, whether it concerns science, future career or daily life issues. It’s a pity we stopped running after one season, I really enjoyed having someone better to follow. Celia, thank you for your dedication in organizing the lab and for your sharp eye, which could find all spelling mistakes that my eye had missed. Sirith, thank you for your endless willingness to help and for your advices on must-see theater performances. Christelle, thank you for not leaving me alone in the senescence topic and for your critical comments. Thos, Thomas, Chrysiis and Marjon, your graduation made me to believe that after all ups and downs there is a positive end to a PhD journey. Marjon, I really enjoyed your handcrafted gifts and homemade apple-pies. Peeper Lab post-doc “(ex-)boys”: Alex, Christophe, Gireesh, Oscar, Naoki, Tristan and Xiangjun, thank you for your efforts to enlighten me with your wisdom and for bringing some testosterone to our female-dominated lab. Oscar, by providing all these references unavailable for NKI, you certainly had a big contribution to this thesis. Nils, thank you for helping me with the mice for these several months I did not have a license. Kristel and Paulien, thank you for the effort you put in the PDX project. Aida, Aranxa, Juliana, Kylie, Mirjam, Renee and Xinyao, thank you for fun times in the office and all the best for your future. Limitless appreciation for the people from Beatson Institute, especially for all the members of R12, which was my ‘second’ home for a few months. Leon, the lesson on how to interpret mass spectra, given by you in Beatson’s dark corridors at 2am on a Sunday night, was a remarkable experience! Thank you for measuring the hundreds of my samples and the introduction to the real Chinese cuisine! Barbara thanks for welcoming me at my arrival

184

185

Addendum

and you invaluable help during and after my stay, whether it was sharing excel macros, running seahorse experiments, organizing teleconferences or “pyruvate effect” discussions. Christian, you impressed me with your balanced personality as well as with the time at which you start lab work. Thank you for knowledgeable explanations to my countless metabolic questions. Laura, I am grateful for providing me a place to stay, personal talks and the highland hiking trip with the Torquay girls! All the best for you, George and Derek! Also many thanks to Marta and Vitaly from Barcelona, Tomer from Israel, Utrecht proteomic group, especially Maarten and Erik, as well as the Leiden team, Remco and Linda G., for a successful collaboration that resulted in nice publications. Linda, this coincidental common work discussion during your guest period at the NKI happened to be a turning point in my PhD. The NKI has some great facilities that I made regular use of. Sido, Sjaak and Henk thank you for looking after the mice. Ji-Ying thank you for evaluating all the slides I gave you. Ron and other members of sequencing facility, thanks for running my samples. Annegien, Ingrid and Bart thank you for help with IHC stainings. Lauran and Lenny thank you for your help with microscopy. Anita and Frank thank you for valuable FACS assistance. Minze and Erwin, thank you for taking good care of my cells. I had the honor of sharing the experience of being PhD student with a bunch of fantastic young people. Together, we exchanged not only scientific advices but also had great fun on Texel-retreats, summerparties, paper celebrations, Friday borrels and much more occasions. Johan, it didn’t take long that our screen collaboration turned into a friendship. Although I hate your habit of taking pictures in the late-night-and-too-much-drank situations and I am not chic enough for cocktails on the top of the Okura hotel, I deeply enjoyed our pancake dinners and philosophical discussions. Thanks for listening to my frustrations, for tolerating my lack of understanding for your perfectionism (no, I don’t care about different shades of black in thesis book) and for finally admitting that studying metabolism isn’t boring! Yme, your exceptional laugh instantly made me to forget all my worries. Thank you for first agreeing on, and then for dragging me to the ice-skating lessons. Marieke, although I do not share your passion for football, I am truly inspired with (and jealous about) yours (and Vicencio’s) amazing travels. Thanks for letting me know that there are more women above 30th that don’t cry for children. Fra, we made it! After all not all experiments fail. Unlike my attempts to properly cook pasta or morning swimming since you are gone. Izhar, thank you for your good heart and… for more good heart. I am grateful for pointing out and appreciating the beautiful things in life. I cross fingers for you and wish you all happiness with Maaike and your lovely kids! Jorma, even if I have never knew how to react to all these bizarre dilemmas of you, I always enjoyed your honesty. Andrej, thank you for all hugs and your never-ending enthusiasm about science. I still trust we will go sailing once! Joep, although the red wine stain on the wall got sold with my old house, the nice memories

of lovely dinners and scientific discussion remained. Thanks for all good times! Jaco, you proved me that research and family can go together. All the best for you and your ladies in Delft! Martijn, thanks for all the fun and for bringing all that heavy wooden floor up to our apartment. Success with your book! Guus, thanks for late afternoon chats in cafeteria and for the pub quiz. We should do it again! Ewald, thanks for sharing your opinions on any issue (like virgins theory), which always made me laugh. Jasper and Amida, you were the social hart of PhDs life at the NKI: work hard, play hard. Thank you for all legendary B7 parties. Big thanks to all people that I had the pleasure of sharing division with over the years. Andor, it was clear from the beginning that photography is your specialty, not science. Thanks for useful tips on the former and not on the latter. Gaetano, thank you for your advice on inducible shRNA system and for sharing the revision stress. Now I also know the nightmare of adjusting figures to fit requirements for publication. Paul, I am grateful for the nice BBQ in your beach house in Zandvoort and for the Zuiver tickets. Robert and Koen, thank you for answering all my silly questions about mass spectrometry and for borrowing me all sort of chemicals. Selvi, I appreciate our friendly chats and your advices about the thesis preparation process. Tom, thank you for great help in organizing things, especially sending all the packages and scheduling various meetings. You really saved me by catching me on the other side of the sloot, during boerensport uitje. Nathalie, thank you for pointing out the blunder in my thesis title, just in time to fix it. Inge, Zehlia, Vera thank you for sharing excitements and complaints on the PhD fate. It always feels better to know that everybody has ups and downs and you are not alone in the time of misery. Marco S. thanks for breaking up lab silence with pleasant music. The Neapolitan mozzarella is indeed delicious. Voest’s people (especially Julia, Marijn and Sahar) thank you for invading my bench, now it is not so obvious anymore that I am not frequent visitor there. Thanks to Alfred, Andre, Anita, Anna-Paulina, Asli, Ariena, Brigitte, Cesare, Charlotte, Danielle, Denise, Dilek, Elisabetta, Ellen, Els, Ewa, Gerjon, Guotai, Henri, Hilda, Huub, Ivo, Jacqueline, Jan Paul, Janneke, John Hilkens, John Z., Lorenzo, Marco B., Mandy, Margriet, Michela, Min Chul, Natalie, Nienke, Paul-André, Santiago, Seng, Stephanie, Sunny, Sven, Waseem, Wendy and all other past and present members of old P1 and H5 for the nice atmosphere you provided. But there is of course more than just P1 and H5 at the NKI. Bas, Ben, Christian, Dalila, Daniel V., Eitan, Erik V., Fred, Gözde, Guistina, Hein, Jacobien, Jacques N., Jelle, Jeroen, John Haanen, Jos, Kitty, Klaas, Linda C., Lodewyk, Marieke van K., Michael, Michiel B., Michiel vd H, Pasi, Philip, René M., Reuven, Rik, Roderick, Rodrigo, Rui and Ton thank you for scientific (or not) advice on various occasions or just casual chat during the Friday borrel. Without you NKI would not be the same! There are many other people that work or have worked at the institute, who I did not mention, and who one way or another contributed to this thesis. Thank you all!

186

187

Addendum

Finally I would like to thank people who supported me outside working hours. Maybe you did not directly contributed to this thesis but energy that you gave me certainly did. Annette although we met through NKI, we actually shared much more outside the lab. I enjoyed a lot famous glühwein parties, dinners, nights out and boat trips. I am amazed by your skill on getting alert of everything what is happening around and I admire how you always are there for your friends. Christine thanks for your warmth, enthusiasm and of course the nice Vienna trip! Erica, bedankt voor de keratine behandelingen, spa bezoeken (wanneer gaan we weer?), lekkere etentjes en natuurlijk voor het verzorgen van mijn kapsel en de foto’s bij onze bruiloft. Ewout, bedankt voor onze zeilweekenden die me hielpen mijn mislukte experimenten te vergeten. Met jouw positieve benadering heb je me geleerd dat zelfs een grote ramp een klein obstakel kan worden als je hem vanuit de juiste hoek bekijkt. Ik kijk uit naar meer gemeenschappelijke boottochtjes, ofwel met MaJo in de Amsterdamse grachten, ofwel met skûtsjes of jachtjes in Friesland. Al mijn zeilvrienden en -vriendinnen van Orionis en Euros, ik heb erg genoten van de team spirit maar ook de bierestafettes. Jasper, Jose, Sander, Bob, Nicole en Tjebbe bedankt voor alle spelletjesavonden en strandweekenden. Asia K. i Asia M. dziękuję za babskie kawki, lunche czy też obiady na mieście, podczas których, nawet po kilu miesięcznej przerwie, potrafimy szczerze rozmawiać bez owijania w bawełnę. Ubiegłe lata jasno pokazały że łączą nas nie tylko te same imiona, doświadczenie doktoratu czy holenderscy mężowie, ale przede wszystkim przyjaźń. Dorota, choć nie dzielimy imienia, doktoratu, czy holenderskiego męża (sorki ale chłopak się nie liczy…), urządzanie domów jest niewątpliwie łączącym wątkiem. Cieszę się że w końcu ktoś może znieść moje paplanie o dwuletnim remoncie. Magda dzięki za dzielenie ze mną uczucia jednoczesnej nienawiści i miłości do Holandii, aktywnego stylu życia i zapału do sportowych przygód. Chociaż na kajaki nie dam się namówić, chętnie powtórzę nurkowanie, narty czy żaglówki. Monika, kto by pomyślał piętnaście lat temu na pływalni UW, że nasza znajomość przekształci się w wieloletnią przyjaźń. Tak naprawdę to Twój wyjazd na Work&Travel dodał mi odwagi i rozpoczął moją zagraniczna przygodę. Dziękuję Ci za nieustannie płynąca pozytywną energie i pasję do odkrywania świata. Choć musieliśmy przerwać nasz on-line kurs degustacji wina, miło wspominam nasze wspólne sesje fotografii i dzielenie doświadczeń z podroży i z niecierpliwością czekam na nowe wyzwania. Ogromne podziękowania dla przyjaciół z Polski: Asi U., Luizy, Maćka, Marzeny, Mirka, Eli, Renaty i Tomka. Dziękuje, że zawsze jesteście dla mnie nie zależnie od odległości która nas dzieli. In the end I would like to express my gratitude for my family. Kochana Mamo, dziękuję Ci za wszystko a przede wszystkim za Twoje wsparcie, zrozumienie i obecność. Nie ma lepiej relaksujących wakacji jak w domu, u Ciebie na tarasie, zwłaszcza po okresie intensywnego remontu! Mirek, choć geograficzny dystans i natłok spraw życia dorosłego nie ma pozytywnego wpływu na częstotliwość naszych kontaktów, zawsze mogę na Ciebie liczyć. Mam nadzieje, że wkrótce znowu wybierzemy się na łódkę w Amsterdamie, tym razem z

Agą, Zuzią i Anielką. Betty, Paul, Luuk, Rinke, Eefje, Janneke en Guido, dank voor jullie hulp tijdens onze verbouwing en voor het verzorgen van afleiding waar nodig. Betty en Paul, ik zal de stoofvlees maaltijd tussen het stof en gereedschap, maar ook mijn onverwachte verjaardagsdiner nooit vergeten. Maarten, dear husband, my love, you are the one who held my hand through all phases of my PhD. The only one who actually knows how high were my ups and how deep were my downs. There are no words that can express my gratitude and appreciation for all you’ve done and been for me. Without your patience, support, believing in me and your unconditional love this book would have never happened. Without you I would not be me. I am so fortunate to have you in my life and proud to have you as my paranymph. Wherever you are, that is where I belong. Thank you for being my home!

188

SENTENCED TO SENESCENCE

SENTENCED TO SENESCENCE BRINGING REBELLIOUS CELLS TO JUSTICE

INVITATION to attend the public defence of the thesis

SENTENCED TO SENESCENCE BRINGING REBELLIOUS CELLS TO JUSTICE

by Joanna Kaplon Monday 26th January 2015 at 15:45 h in the aula of VU University Amsterdam The ceremony will be followed by a reception

Joanna Kaplon

Paranymphs:

Joanna Kaplon

Maarten Hoekstra mhhoekstra@gmail.com Katrin Meissl katrinmeissl@gmail.com