Regulation of oncogene-induced cell cycle exit and senescence by chromatin modifiers

Abstract

Oncogene activation leads to dramatic changes in numerous biological pathways controlling cellular division, and results in the initiation of a transcriptional program that promotes transformation. Conversely, it also triggers an irreversible cell cycle exit called cellular senescence, which allows the organism to counteract the potentially detrimental uncontrolled proliferation of damaged cells. Therefore, a tight transcriptional control is required at the onset of oncogenic signal, coordinating both positive and negative regulation of gene expression. Not surprisingly, numerous chromatin modifiers contribute to the cellular response to oncogenic stress. While these chromatin modifiers were initially thought of as mere mediators of the cellular response to oncogenic stress, recent studies have uncovered a direct and specific regulation of chromatin modifiers by oncogenic signals. We review here the diverse functions of chromatin modifiers in the cellular response to oncogenic stress, and discuss the implications of these findings on the regulation of cell cycle progression and proliferation by activated oncogenes.

Introduction

Paradoxically, oncogene activation leads to an irreversible cell cycle exit in primary cells, called oncogene-induced senescence (OIS).^1,2 Cellular senescence was first defined as a functional cellular state: a senescent cell is a cell that has exited the cell cycle permanently and is incapable of resuming proliferation, even upon mitogenic cues.^3,4 While this definition of cellular senescence has been helpful to functionally classify cell cycle exit situations, phenotypic characterization of cellular senescence requires the identification of specific markers. Among these potential markers, Senescence Associated β-Galactosidase activity in acidic conditions (pH 6.0) (SA-β-Gal) has been shown to consistently identify senescent cells in vitro and in vivo.^5-8 Additionally, several proteins that promote cell cycle exit are upregulated at the onset of senescence in vitro and in vivo, including p19^ARF in mouse cells, and p16^INK4A and p15^INK4B in mouse and human cells.^9-12 One caveat regarding the use of these markers is that their presence does not systematically discriminate between senescence and other forms of growth arrest. As discussed at length below, nuclei of senescent cells are subjected to a dramatic reorganization, as evidenced by the emergence of dense foci called SAHF (Senescence Associated Heterochromatic Foci)¹³ that are brightly stained with 4’-6-Diamidino-2-phenylindole (DAPI) and enriched in markers of constitutive heterochromatin.^14-16 While such foci were easily visualized in human cells, their detection in senescent murine cells is hindered by the fact that rodent pericentric loci are readily decorated by the aforementioned chromatin marks, regardless of the senescence status.

Functionally, pathways engaged by pro-senescence signals to induce permanent cell cycle withdrawal fall into three major pathways. The first pathway is the transcriptional silencing of pro-proliferative genes.^7,13,17,18 A second pathway is the DNA damage response pathway, triggered by aberrant DNA replication events, including replication fork collapse.^19-21 The third pathway is the Senescence-Associated Secretory Phenotype (SASP, also known as SMS), which corresponds to a specific set of secreted proteins and receptors expressed at the onset of senescence.^22-26 Engagement of each of these three programs appears to at least reinforce cellular senescence, suggesting they are likely to be coordinated at the molecular level. For example, SASP is detected upon persistent activation of the DNA damage response.²⁷ While these observations have immensely helped our understanding of the signal transduction leading to cellular senescence, how these pathways are being activated at the molecular level is only partially understood. Strikingly, the recent results that are discussed in this review have revealed that chromatin modifiers lie at the nexus of the cell cycle exit response triggered by oncogenic stress in primary cells.

Chromatin modifiers, cellular senescence and histone modifications

Numerous chromatin associated proteins and chromatin modifiers have now been implicated in the cell cycle exit triggered by oncogenic stress at the onset of senescence. An important and early study suggested that chromatin modifiers could modulate cellular senescence: the chromodomain (CHD) containing protein MORF4, was identified through its ability to induce senescence in immortalized human cells.²⁸ MORF4 is the founding member of the MRG family of proteins that are integral components of several transcriptional coregulator complexes.^29-31 Importantly, Eaf3, the yeast homolog of MRG proteins, binds directly histone 3 lysine 36 when methylated.^32-34 How the ability of MORF4 to induce senescence relates to the ability of MRG proteins to associate with modified histones and to modulate gene expression is still unclear. It is however intriguing to note that the ING family of proteins, which, like MRG proteins, are integral components of the NuA4 and the Sin3 co-regulator complexes, also modulate cellular senescence.³⁵ Whether the ability of these proteins/complexes to modulate cellular senescence depends on their capacity to modify chromatin is subject to debate. For example, ING2 promotes p300-mediated acetylation of p53, which is sufficient to trigger cellular senescence in primary fibroblasts.³⁶ As the number of non-histone proteins whose function is found to is regulated by acetylation or methylation increases exponentially,³⁷ the involvement of a chromatin-associated protein in cellular senescence may not always directly involve chromatin or histone modifications. Genome-wide mapping of these modifiers at the onset of senescence is likely to yield some important insight into their precise contribution, as chromatin modifiers, to oncogene-induced cell cycle exit in the future.

Nuclear reorganization at the onset of oncogene-induced senescence: Senescence associated heterochromatin foci

The most compelling evidence that chromatin modifiers contribute to the cell cycle exit triggered by oncogenic stress may have stemmed for the observation that discrete dense chromatin structures can be detected in the nuclei of human cells that have undergone oncogene-induced senescence.¹³ Such structures, named SAHF (for Senescence Associated Heterochromatin Foci) are comprised of condensed chromatin and are characterized by the presence of several hallmarks of heterochromatin, including trimethylation of histone 3 lysine 9 (H3K9me3), Heterochromatin Proteins 1β and 1γ (HP1β and HP1γ), the histone variant macroH2A and the chromatin-associated HMGA proteins.^14-16 The genesis of these SAHF domains is a complex mechanism that involves preexisting nuclear structures, the PML bodies, as well as the histone chaperones HIRA and Asf1.^16,38 The precise molecular mechanisms underlying the formation of SAHF at the onset of senescence have been extensively reviewed in the past.^39,40 Initially, it has been proposed that the formation of SAHF allows the heterochromatinization of pro-proliferative loci, including E2F response loci, and thus provides a molecular basis for their permanent transcriptional silencing. Indeed, H3K9me3 and HP1γ were detected on the promoters of PCNA and Cyclin A, only in senescent cells. Similarly, the locus encoding for Cyclin A was found embedded within SAHF domains in senescent cells in situ hybridization.¹³ Altogether, these data suggested that heterochromatinization of pro-proliferative genes through molecular mechanisms reminiscent of pericentric heterochromatinization, could mediate the observed cell cycle exit at the onset of senescence.⁴⁰ A genetic approach was undertaken to establish the relationship between heterochromatin, oncogene-induced senescence and cancer progression in vivo. The mammalian histone methyltransferases Suv39h1 and Suv39h2 (KMT1A and KMT1B) are responsible for most H3K9 trimethylation, and Suv39h knockout mice are viable.⁴¹ Using a mouse strain that expresses oncogenic N-Ras in lymphocyte and develops lymphoma with a high penetrance, Braig et al. showed that genetic inactivation of Suv39h1 significantly accelerates the incidence of cancer.⁴² In addition, Suv39h1-inactivated lymphocytes appear refractory to oncogene-induced senescence. From these observations, the authors concluded that Suv39h1 participates to a chromatin pathway that establishes senescence in response to oncogenic stress.⁴² While this study convincingly reinforced the molecular connection between cellular senescence and heterochromatin formation, several questions remain to be addressed. First, although Suv39h KMTs are essential for oncogene-induced senescence in lymphoma, they surprisingly appear dispensable for replicative senescence in primary fibroblasts.¹³ Second, the identity of the genomic loci that are subjected to Suv39-mediated heterochromatinization at the onset of senescence is unknown. They could be pro-proliferative loci, or more likely sites of DNA damage (see below). The current unavailability of Suv39h antibodies suitable for ChIP represents a major obstacle in expanding our understanding of this specific molecular pathway.

Despite this functional connection between heterochromatin and oncogene-induced senescence, the notion that global nuclear reorganization occurs at the onset of senescence to permanently silence pro-proliferative genes has recently been challenged. First, SAHF do not correspond specifically to pro-proliferative loci, but rather to whole chromosomes, as each chromosome becomes condensed upon oncogene activation to give rise to one senescence-associated focus.¹⁶ Second, it appears that mouse primary cells do not form SAHF in response to ectopic expression of oncogenic Ras.⁴³ Finally, two studies have recently shed a new light on the relationship between SAHF and the cellular senescence phenotype: SAHF are formed preferentially at the onset of oncogene-induced senescence but not during replicative senescence.^44,45 This surprising observation implies that in human cells, senescence can occur independently of SAHF formation, a phenomenon already suggested by the study of murine cells. Consistently, oncogene activation in primary cells devoid of p53 or ATM leads to the generation of SAHF, despite continuous cellular proliferation. On the other hand, disruption of the SAHF formation pathway, achieved though ATR knock-down, does not prevent senescence.⁴⁴ Together, these observations suggest that oncogenic stress triggers the activation of two independent pathways, an ATM-dependent repression of proliferation, and an ATR-dependent large-scale nuclear reorganization resulting in the formation of SAHF.⁴⁶ These conclusions are further supported by the inability to detect SAHF in human lung adenoma samples, despite the presence of other senescence markers.^44,45 So, if SAHF are not essential for the cell cycle exit triggered by oncogene activation, what is their contribution to the response to oncogenic stress? It has been shown in the past that ectopic expression of activated oncogenes, such as Ras or mos, promotes uncontrolled replication resulting in the fork collapse and accumulation of DNA double strand breaks (DSB).^19-21 Using HDAC inhibitors, Di Micco et al. demonstrated that preventing SAHF formation, through chromatin relaxation, results in the amplification of DNA damage response (DDR) throughout the genome, culminating in DNA-damage-induced apoptosis.⁴⁴ Thus, they hypothesized that SAHF enable cells to prevent the spreading of DNA damage induced by oncogenic stress. This exciting hypothesis underlines the complexity of the events that are triggered by oncogenic stress in mammalian cells. However, it does not fully explain the exquisite specificity for SAHF in senescent cells. If the formation of SAHF is a defense mechanism against DNA damage, why is there not more DNA damage in immortalized or transformed cells, which constitutively express activated oncogenes? This paradox raises the possibility that the molecular events that lie at the origin of SAHF formation may be disrupted in immortalized or transformed cells.

Lastly, and as highlighted above, the contribution of chromatin modifiers to the cellular events triggered by oncogenic stress needs to incorporate their function in the response to DNA damage. Indeed, most studies that have focused on the molecular mechanisms underlying cellular senescence appear to converge toward the DNA damage/repair pathways.² As presented earlier, oncogene activation results in the formation of DSB.^19-21 Recent results have demonstrated that chromatin modifiers, and in particular transcription repressors, are directly involved in the response to DNA damage.⁴⁷ While this participation may not be specific to the entry in a senescent state, it is however likely to correspond to an integral component of the response to oncogenic stress. Future studies focused on cell cycle exit driven by oncogene activation are expected to shed new light on the precise contribution of chromatin modifiers to this process.

Transcriptional repression of pro-proliferative genes

The ultimate effectors of oncogene-induced cell cycle exit comprise the pro-proliferative E2F target loci mentioned above. E2F factors control the progression from G₀/G₁ to S phase and are therefore central to the oncogene-induced senescence process.⁴⁸ Now that it has become clear that SAHF do not represent exclusively, if at all, E2F target loci that have been embedded in heterochromatin, the repression pathway leading to the permanent silencing of E2F targets remains to be explored. In the recent past, numerous studies have highlighted the molecular complexity underlying the repression of pro-proliferative genes in cells undergoing quiescence or differentiation.⁴⁹ Specifically, E2F factors associate with several chromatin modifiers comprised within large complexes to modulate transcription, including the dREAM complex,^50,51 the ARID1B Swi/Snf complex,⁵² the HCF-1-associated Sin3 complex and MLL-1 (KMT2A) complex,^53,54 and possibly the PRC1 complex.^55,56 Whether these observations can be extrapolated to the cell cycle exit driven by oncogenic stress at the onset of senescence is only beginning to be assessed. Notably, recent observations suggest that an integral dREAM complex is required for Ras^V12-induced senescence.^57,58 The central function of the retinoblastoma (Rb) family of proteins-E2F axis in the G₀/S transition has allowed the identification of a chromatin pathway driving cell cycle exit at the onset of senescence. While all three retinoblastoma-related proteins (Rb, p107, p130, also known as pocket proteins) share somewhat redundant functions in cellular senescence in murine cells,^59,60 a recent study highlighted the specific contribution of Rb in the regulation of Ras-induced cell cycle exit in human cells.⁶¹ Specifically, using ChIP-seq analysis, the authors demonstrate that Rb exquisitely binds to loci encoding proteins required for DNA replication, thus triggering their permanent silencing. This study undoubtedly provides a framework to investigate the chromatin modifying activities that are involved in the seemingly irreversible cell cycle exit upon expression of activated oncogenes. Identifying the activities associated with Rb at these loci specifically upon oncogenic stress should enable us to better understand the molecular basis for the irreversible repression of pro-proliferative loci. Concurrently, the knowledge gained upon studying cell cycle exit in other settings has provided initial clues as to how Rb tethering to chromatin contributes to gene silencing. As previously demonstrated for entry into quiescence, murine cells utilize specific histone deacetylases (HDAC)-containing complexes for active repression of E2F targets at the onset of senescence, targeted by their association with the Rb family of proteins. For example, the Sin3B-HDAC1/2 complex, previously shown to be recruited in a pocket protein-dependent manner on E2F target promoters at the onset of quiescence,^62-64 is also tethered to E2F target genes following oncogenic Ras expression.⁶⁵ Sin3B is also required for the senescence-dependent heterochromatinization of these loci, which includes trimethylation on H3K9 and HP1 recruitment, consistent with the notion that the recruitment of chromatin modifiers to particular loci is highly coordinated. Remarkably, Sin3B null cells are refractory to oncogene-induced senescence, and Sin3B protein appears to accumulate in preneoplastic lesions, reinforcing its functional relationship to cellular senescence.⁶⁵ In addition to HDAC-containing complexes, Rb associates with several chromatin modifiers at the onset of senescence, the H3K9 histone methyltransferases Suv39h1 and Suv39h2 (KMT1A and KMT1B), HP1γ, and the chromatin remodeler Brm1, to promote cellular senescence.⁶⁶ The identity of the loci targeted by such Rb-containing mega-repressor complex remains unclear. Although it is tempting to speculate that they stably repress E2F target loci, the reported participation of Rb in the heterochromatinization of large genomic regions, such as telomeres or centromeres, implies a potential non-sequence specific targeting and a remodeling of nuclear structures on a larger scale.⁶⁷ Notably, the relationship between these Rb-containing mega-complexes and heterochromatinization of SAHF needs to be explored further, especially in light of reports demonstrating the facultative nature of SAHF in cellular senescence, as discussed above.

Recently, an elegant genetic screen identified the H3K4 demethylase KDM5B as a direct transcriptional repressor of pro-proliferative genes at the onset of senescence.⁶⁸ The promoters of numerous pro-proliferative genes contain E2F binding sites that tether KDM5B specifically upon induction of senescence. KDM5B recruitment results in the demethylation of the H3K4 residues, correlating with the transcriptional repression of the corresponding loci. Consistent with its requirement for senescence-associated cell cycle exit, KDM5B depletion allows continuous proliferation and colony formation in primary mouse fibroblasts deficient for p107 and p130, reminiscent of Rb depletion in murine cells.^68,69 While KDM5B levels are upregulated in melanocytic nevi, its expression becomes restricted to a subpopulation of slow-growing cells in melanoma, which are essential to sustain the tumor.⁷⁰ The connection to oncogenic stress and cell cycle exit in this latter case remains unclear, but the possibility that a set of chromatin modifiers, including KDM5B and the related KDM5A, controls proliferation and resistance to genotoxic stress in a small portion of cells within a tumor is intriguing.^70,71 Future studies will likely explore the relationship between the resistance of these specific cells to chemotherapy and their response to oncogenic stress.

In conclusion, the transcriptional repression of E2F target genes at the onset of senescence requires the coordination of several histone modifying activities, some of them likely unique to senescence (Fig. 1). As senescence is defined as a permanent cell cycle exit, one would assume that it shares molecular pathways with terminal differentiation, particularly as far as silencing of proliferative genes is concerned. A deeper understanding of the chromatin modifiers that are involved in each one of the repression pathways discussed above (E2F target repression or SAHF formation) should enable us to better understand the connection between irreversible cell cycle exit and oncogenic stress.

Figure 1. Schematic representation of the contribution of chromatin modifiers to the transcriptional repression of E2F target genes driven by oncogenic stress. The promoters of proproliferative genes, initially in an open chromatin conformation, are subjected to the activity of specific histone methyltransferases (KMT), demethylases (KDM) and deacetylases (HDACs) at the onset of oncogene-induced senescence, leading to a switch from active marks (H3/H4 Acetylation, H3K4me3) to repressive marks (H3K9me2/3). See text for details.

Transcriptional activation of cell cycle inhibitors

While transcriptional repression of pro-proliferative genes is central to oncogene-induced cell cycle exit, the accumulation of transcripts encoding pro-senescence proteins is equally important and has been molecularly dissected in depth for specific transcripts in the recent past. One of the best studies loci in cancer biology is the INK4A locus, which encodes two proteins, the CDK4/6-inhibitor p16^INK4A, and the MDM2-intertactor p19^ARF (p14^ARF in human cells).⁷² Both transcripts share two common exons, translated however through alternative reading frames. As each protein targets and potentiates a major tumor suppressor pathway (Rb and p53 for p16^INK4A and p19^ARF, respectively), it comes as no surprise that the chromatin environment of this locus is regulated in an extremely precise manner upon oncogenic stress. Transcription through the INK4A locus is actively repressed in proliferating cells, including early passage primary fibroblasts. However, upon oncogenic stress elicited for example by ectopic expression of Ras^V12 or BRAF^V600E, p16^INK4A and p19^ARF transcripts accumulate. Importantly, genetic inactivation of the INK4A locus, as found in a large proportion of human cancers, allows cells to bypass oncogenic Ras-induced senescence.^7,73 In the past few years, several studies have aimed at identifying the enzymes responsible for the transcription changes affecting the INK4A locus upon oncogenic stress, promoted by the recent development in chromatin analysis, as well as the identification of novel chromatin modifying activities. The first hint linking cellular senescence to chromatin–mediated modulation of the INK4A locus stemmed from the observation that Bmi-1 null primary MEFs undergo premature senescence due to the accumulation of p16^INK4A and p19^{ARF. 74} Bmi-1 is an integral component of the Polycomb group complex PRC1, which can be recruited to genomic loci through a direct interaction between the PRC1 components CBX proteins, H3K27me3 and long non-coding RNAs (lncRNA).^75,76 In proliferating cells, the INK4A locus is decorated with H3K27me3 which, is the presence of the lncRNA ANRIL, results in the tethering of the PRC1 complex through the binding of CBX proteins.^76-79 The strong enrichment of Bmi-1 and its associated PRC-1 complex on the INK4A locus correlates with the inability of DNA-dependent RNA Polymerase II (RNAPII) to drive transcription through this locus, through partially understood mechanisms.^80,81 Importantly, several studies clearly demonstrated that pro-oncogenic signals, such as ectopic expression of the activated Ras oncogene (Ras^V12) or BRAF^V600E, lead to dramatic changes in the chromatin structure of the INK4A locus, through the modulation of specific chromatin modifiers level and recruitment. First, EZH2 (KMT6), the histone methyltransferase responsible for the addition of methyl groups on H3K27 on the INK4A locus in proliferating cells, is robustly downregulated at the transcriptional level upon oncogenic stress.⁷⁷ Concurrently, the histone demethylase responsible for the active removal of the H3K27 trimethyl groups, JMJD3 (KDM6B), accumulates and associates with the INK4A locus.^77,82 These events result in the active demethylation of the H3K27 residues throughout the INK4A locus upon oncogenic stress. The PRC1 complex is then unable to bind this locus and to repress its transcription. Of note, it has been suggested that activation of the MAP kinase pathway, downstream of Ras activation, results in the dissociation of Bmi-1 from chromatin, through its phosphorylation by 3pK.⁸³ Whether this event also contributes to the release of Bmi-1 from the INK4A locus at the onset of senescence remains to be investigated. Importantly, the removal of repressive marks from the INK4A locus at the onset of senescence is accompanied by the addition of active marks, including H3K4me3. While the H3K4 histone methyltransferase MLL-1 (KMT2A) is found on the human p16^INK4A locus with or without oncogenic signal, its presence, mediated by the CUL4-DDB1 complex, is absolutely necessary for p16^INK4A induction in oncogenic Ras-stimulated cells.⁸⁴ Additional chromatin modifiers have been shown to regulate transcription throughout the INK4A locus, including the SWI/SNF remodeling complex and the DNA methyltransferase DNMT3B protein.⁸⁵ While they modulate p16^INK4A/p19^ARFexpression, their contribution to these changes at the onset of senescence has not been directly explored. Finally, an unbiased approach revealed that at onset of senescence, the transcript levels of KDM2A and KDM2B decrease significantly.⁸⁶ KDM2A and KDM2B are histone demethylases that target H3K26me1/2 and are tethered to the INK4A locus in proliferating cells. In addition, it has been suggested that they are able to remove H3K4me3. Therefore, the downregulation of KDM2B at the onset of senescence prevents the removal of these two active marks, ultimately allowing RNAPII progression throughout the INK4A.^86,87 Interestingly, a separate study also demonstrated that KDM2B represses cellular senescence in growing cells, but the results from this report demonstrated that KDM2B targets the neighboring INK4B locus, rather than the INK4A.⁸⁸ The accumulation of the p15^INK4B, the product of the INK4B locus, is a well-recognized hallmark of cellular senescence.⁹ The basis for these seemingly conflicting results is unknown, but together, these studies highlight the direct regulation of the chromatin modifier KDM2B as an essential prerequisite for cell cycle exit as a response to oncogenic stress. Overall, the regulation of the INK4A/INK4B loci represents a paradigm in understanding how the transcriptional machinery adapts in response to oncogenic stress to produce a specific outcome (Fig. 2).

Figure 2. Schematic representation of the contribution of chromatin modifiers to the transcriptional activation of the INK4A locus driven by oncogenic stress. The INK4A locus, initially in a closed chromatin conformation, is affected by several histone methyltransferases (KMT) and histone demethylases (KDM) at the onset of oncogene-induced senescence, leading to a switch between the presence of H3K27me3 to the presence of H3K4me3. See text for details.

Transcriptional activation of the senescence associated secretory phenotype

In addition to the INK4A/INK4B loci, several loci become transcriptionally active at the onset of senescence. A family of proteins that have been shown to accumulate as a response to oncogenic stress and reinforce cellular senescence, in both paracrine and autocrine manner, are the pro-inflammatory cytokines and the metalloprotease that collectively form the SASP (for secretory associated secretory phenotype) or SMS (senescence messaging secretome).^22,25,89 In primary cells, including fibroblasts or melanocytes, acute expression of oncogenic proteins such as Ras^V12 or BRaf^V600E, leads to the rapid and substantial accumulation of secreted proteins. These secreted proteins include the interleukin IL-6, IL-8 and the metalloprotease MMP7. The precise identity of the proteins secreted at the onset of senescence, and their contribution to the senescent phenotype have been extensively reviewed in the past.^90,91 By contrast, the molecular bases underlying the transcriptional events that dictate their expression are just being uncovered. At the transcription level, the SASP response appears to be mainly dictated by NFκB and CEBP/β, two factors that had previously been implicated in oncogenic Ras signal transduction and oncogene induced senescence, respectively.^22,25,92-94 Although it was anticipated that chromatin modifiers contribute to the activation of these transcription factors in response to oncogenic stress, whether oncogenic stress specifically and directly targeted chromatin modifiers to regulate the SASP response was initially unknown. A surprising finding recently uncovered the direct and crucial regulation of specific histone methyltransferases triggered by oncogenic stress. Specifically, ectopic expression of oncogenic Ras in primary cells leads to the proteosomal degradation of the histone methyltransferases G9a and GLP (KMT1C and KMT1D, respectively), which are responsible for mono and dimethylation of histone H3 lysine 9.⁹⁵ This degradation is driven by the ubiquitin ligase complex APC/C^Cdh1, which ubiquitinates G9a and GLP upon DNA damage. Ultimately, G9a and GLP degradation correlates with their removal from the IL-6 and IL-8 promoters, allowing active transcription of these genes (as well as many other SASP genes) at the onset of senescence (Fig. 3).⁹⁵ Several important questions stem from this study; For example, it would be interesting to investigate whether all DNA damage settings trigger G9a and GLP degradation, or if alternatively, activation of the senescence program provides a specific signal for this degradation. Additionally, how this degradation affects genome wide expression, and in particular the expression of cell cycle regulators is unclear. Additional chromatin modifiers are likely to participate to this transcriptional activation. For example, administration of sodium butyrate (NaB), a HDAC inhibitor that promotes cellular senescence, induces the rapid accumulation of SASP, suggesting that HDAC modulate the repression of the secretory response.⁹⁶ Whether HDAC themselves, or any other chromatin modifiers are targeted by oncogenic stress to modulate the SASP response is currently unknown. However, the important discovery that specific KMT are degraded upon oncogenic stress highlights the central role of histone modifiers, not only as mediators, but also as direct targets of pro-senescence signals.

Figure 3. Schematic representation of the contribution of chromatin modifiers to transcriptional activation of the SASP genes driven by oncogenic stress. The chromatin surrounding the promoters of the genes corresponding to the SASP response, such as IL-8, are initially in a closed chromatin conformation, with the presence of H3k9 methylation, a repressive mark. Oncogenic stress induces the degradation of the corresponding histone methyltransferases and the subsequent recruitment of histone acetyltransferase, which result in the addition of active marks (H3/H4 Ac). See text for details.

Concluding remark

As highlighted by the examples presented in this review, chromatin modifiers play diverse and crucial roles in the modulation of cell cycle exit driven by oncogenic stress. While this was expected based on the wide functions of chromatin modifiers in the regulation of most, if not all, DNA transactions, the recent findings that multiple chromatin modifiers are specifically regulated, at the transcriptional or post-translational levels, by pro-senescence signals opens a new field of investigation. Indeed, the notion that the restoration of cellular senescence represents a therapeutic approach against cancer has gained momentum in the recent past.⁹⁷ A better understanding of the druggable enzymes, such as specific chromatin modifiers, that are required involved in sensing oncogene-induced stress, should allow the delineation of new and unsuspected therapeutic targets for the treatment of cancer. As small molecule inhibitors targeting specific chromatin modifiers are currently being developed at an increasing pace, one could envision that it will be possible in the near future to counteract the pro-tumorigenic effects of oncogenic stress by targeting specific chromatin modifying activities.

Glossary

Abbreviations:

OIS

oncogene induced senescence

SA-β-Gal

senescence associated β-galactosidase

SAHF

senescence associated heterochromatic Foci

DAPI

4’-6-Diamidino-2-phenylindole

CHD

chromodomain

H3K9me3

trimethylation of histone H3 lysine 9

SASP

senescence-associated secretory phenotype

SMS

senescence messaging secretome

DDR

DNA damage response

HDAC

Histone deacetylase

DSB

double strand break

KMT

histone methyltransferase

KDM

histone demethylase

RNAPII

RNA polymerase II

PCR1

Polycomb Group Repressive complex 1

lncRNA

long non-coding RNA