Oncogene (2007) 26, 7765–7772 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc REVIEW structure and DNA double-strand break responses in cancer progression and therapy

JA Downs

MRC Genome Damage and Stability Centre, University of Sussex, Falmer, Brighton, UK

Defects in the detection and repair of DNA double-strand recentwork proposes thatchromosomal instabilityis breaks (DSBs) have been causatively linked to tumourigen- likely to be a driving force in the process of tumourigen- esis. Moreover, inhibition of DNA damage responses esis (Michor, 2005). This suggests that the ability of cells (DDR) can increase the efficacy of cancer therapies that to detect and, appropriately and efficiently, respond to rely on generation of damaged DNA. DDR must occur DNA lesions is critical in preventing the onset of within the context of chromatin, and there have been tumourigenesis. significant advances in recent years in understanding how One of the most dangerous DNA lesions a cell can the modulation and manipulation of chromatin contribute sustain is a DNA double-strand break (DSB). If left to this activity. One particular covalent modification of a unrepaired, a single DSB can be lethal, and can also variant—the of H2AX—has been result in chromosomal translocations, deletions and loss investigated in great detail and has been shown to have of genetic information. In eukaryotes, there are two important roles in DNA DSB responses and in preventing main pathways for repairing DSBs; homologous re- tumourigenesis. These studies are reviewed here in the combination (HR) and nonhomologous end joining. In context of their relevance to cancer therapy and diagnos- addition to the repair activities, cells have pathways tics. In addition, there is emerging evidence for contribu- dedicated to the detection and signalling of these tions by proteins involved in mediating higher order dangerous lesions in order to orchestrate the appro- structure to DNA DSB responses. The contributions of a priate response, including cell cycle checkpoints and subset of these proteins—linker and high-mobility transcriptional upregulation. group box (HMGB) proteins—to DDR and their potential Not surprisingly, it has been known for some time that significance in tumourigenesis are discussed. defects in the DNA damage response (DDR) pathway, Oncogene (2007) 26, 7765–7772; doi:10.1038/sj.onc.1210874 such as the Mre11, Rad50, Nbs1 (MRN) complex (for review, see Williams et al., 2007) and in the kinases Keywords: chromatin; DNA repair; genome stability; ataxia telangiectasia mutated (ATM) and ATM and linker histone; H2A; H2AX; HMGB RAD3-related (ATR) (for review, see Jackson, 2002; Lavin et al., 2006), are associated with cancer predis- position syndromes in humans, and disruption of DNA repair systems in mouse models leads to an increased risk of cancer (Jackson, 2002; Lavin et al., 2006; Williams Introduction et al., 2007). Obviously, one critical function of these pathways is to respond to exogenously induced DNA Genomes of tumour cells exhibit multiple alterations damage. However, programmed DSBs are created in from the parent cells, and it is now well established that lymphocytes undergoing class switch recombination or tumourigenesis is a multistep process of genetic altera- V(D)J recombination. Importantly, components of the tions. These alterations allow the cell to acquire DDR program are essential for protecting the genome characteristics that are universal among tumours and from translocations in lymphocytes (for have been referred to as the ‘hallmarks of cancer’ example, Vogel et al., 1999). Unless the breaks created (Hanahan and Weinberg, 2000). It is estimated that this during these processes are efficiently resolved, they can is achieved in 4–7 rate-limiting stochastic events (Hana- serve as substrates for chromosome translocation, and han and Weinberg, 2000). Consequently, genomic deficiency in ATM promotes these aberrant joining instability has been termed as an ‘enabling characteristic’ events (Reina-San-Martin et al., 2003; Bredemeyer et al., of cancer, since the loss of systems to ensure genomic 2006; Franco et al., 2006; Ramiro et al., 2006). integrity allow the cell to more easily acquire these genetic Interestingly, recent work from two labs has demon- changes (Hanahan and Weinberg, 2000). In fact, more strated that the DDR are activated early during tumour- igenesis (Bartkova et al., 2005; Gorgoulis et al., 2005). This was thought to be due to DNA damage caused by Correspondence: Dr JA Downs, MRC Genome Damage and Stability Centre, University of Sussex, Science Park Road, Falmer, Brighton increased replicative stress in rapidly dividing pre-neoplas- BN1 9RQ, UK. tic lesions. These findings demonstrate the central E-mail: [email protected] importance of the DDR pathway, which acts as an Chromatin structure and DSB responses JA Downs 7766 inducible barrier againsttumourprogression and genetic mediating higher order chromatin structure, the linker instability. Once overcome, the cells are subsequently histones and HMG proteins, in DNA DSB responses is vulnerable to accumulating additional genetic changes. emerging. Second, recent studies have examined the Clearly, therefore, understanding the mechanisms of DDR ability to exploit one particular DNA damage-depen- is of central importance to cancer biology. dent chromatin alteration—the phosphorylation of H2AX—for clinical applications.

Chromatin and the DNA damage response The linker histone and DNA damage responses Clearly, these DDR must occur within the context of chromatin. At its simplest level, chromatin is made up of Linker histones are found throughout eukaryotic DNA wrapped around an octamer of histone proteins evolution (Ausio, 2000). In lower eukaryotes, they do (two each of H2A, H2B, H3 and H4) to form a not appear to be essential for viability (Ausio, 2000; . The histones have histone fold domains, Harvey and Downs, 2004). For example, in budding which are important for the formation of the central yeast, there is a single gene encoding a linker histone, structure of the nucleosome, and flexible tail domains, termed HHO1, and cells lacking this gene show very which protrude from the nucleosome core particle (Luger little phenotypic alterations (Harvey and Downs, 2004). et al., 1997). The packaging of DNA in this manner is In higher eukaryotes, multiple isoforms of linker generally inhibitory to processes such as histones exist. In mice, disruption any one of the eight and replication, and cells have numerous mechanisms by linker histone genes does not result in lethality, and it which this structure can be manipulated to make it more appears that there is upregulation of the remaining amenable to these activities. Broadly, they can be placed genes in order to compensate for the loss (Alami et al., into two groups; the addition of covalent modifications to 2003). However, once three genes are disrupted and the the histone proteins, and the activity of ATP-dependent total amount of linker histone falls significantly, this is chromatin remodelling activities that use the energy an embryonic lethal event (Fan et al., 2005). This derived from the hydrolysis of ATP to alter the contacts suggests that for the function required in higher between the DNA and histone proteins. eukaryotes to sustain viability, the different linker Chromatin has the ability to form compact structures histone genes are functionally redundant and it is the well beyond the initial level of . There are a overall level of linker histone protein in the cell that is number of factors that influence the formation of these critical. Similarly, in chicken DT40 cells, while there are higher order structures, including linker histones and six genes encoding linker histones, the presence of a high-mobility group (HMG) proteins. These are abun- single linker histone gene is sufficient for viability dant architectural chromatin proteins found throughout (Hashimoto et al., 2007). eukaryotic evolution. Not surprisingly, these proteins In yeast, it has been demonstrated that the loss of the are also subjected to complex control by numerous linker histone results in hyper-resistance to methyl covalent modifications that affect their chromatin- methane sulfonate (MMS). By using a combination of binding activities and protein–protein associations. genetic approaches, a model in which the linker histone Chromatin is, therefore, a highly complex polymer acts to inhibit the HR pathway was proposed (Downs capable of extraordinary variability in both structure et al., 2003). While the behaviour of the linker histone and composition. after treatment of cells with MMS was not directly There has been a great deal of research into the examined, it would be reasonable to hypothesize that changes in chromatin structure that occur in response to there may be some regulation of linker histones in DNA damage. In particular, there are many covalent response to DNA damage. In particular, the dissocia- modifications of histones that are now known to be tion of Hho1 from chromatin would be predicted to important for DDR, and there is likely to be a pattern of facilitate the repair of the DNA lesions by HR. modifications on histones in the proximity of DNA Interestingly in this regard, in mammalian cells, a DSBs that contributes to DDR. It is also now apparent population of all five of the linker histone isoforms that that multiple ATP-dependent chromatin remodelling are ubiquitously expressed in somatic cells re-localize complexes contribute to cellular survival after DNA from the nucleus to the cytosol in response to ionizing damage and play a role directly at the sites of DNA radiation (Figure 1; Konishi et al., 2003). However, the damage. These events have been extensively reviewed effect of this re-localization on HR activity was not elsewhere (Peterson and Cote, 2004; Foster and Downs, examined. Rather, in this study, the authors had set out 2005; Morrison and Shen, 2005; van Attikum and to identify pro-apoptotic cellular factors that induce Gasser, 2005; Costelloe et al., 2006; Downs et al., 2007). cytochrome c release from mitochondria by using Moreover, the contribution of chromatin modifying cytosolic extracts prepared from cells treated with complexes and enzymes to tumourigenesis was the ionizing radiation (IR). Although all five linker histones subjectof a recentissue of Oncogene reviews (volume were detectable in the cytosol after cells were treated 26, issue 37). Therefore, in this review, the focus will be with IR, they found that only H1.2 was capable of limited to two topics that have received less attention. efficiently promoting cytochrome c release in this assay, First, an important role for proteins involved in and this activity was dependent on the presence of Bak.

Oncogene Chromatin structure and DSB responses JA Downs 7767

Inflammation signalling

stress (some cell types)

necrosis

IR

senescence

apoptosis

Figure 1 The role of chromatin proteins during senescence, apoptosis and necrosis. In response to IR, linker histones (circles) are re- localized to the cytosol and one isoform (H1.2, blue) triggers apoptosis via interaction with the mitochondria (gray), while HMGB proteins (blue diamonds) are actively retained in the chromatin. In contrast, during necrosis and in response to stress (for some cell types) HMGB1 is released and acts to signal an inflammatory response. Linker histones are lost when cells undergo senescence. Gray bars represent chromatin. See text for details. The release of H1.2 from the nucleus was dependent on defectin a specific HR assay using an inducible p53, and occurred specifically in response to agents that restriction enzyme and looking at recombination of caused DNA DSBs, such as IR and etoposide, but not in two tandem NeoR cassettes. response to ultraviolet-induced apoptosis. Consistent The reason for the opposing activities of chicken with a direct role in mediating IR-induced apoptosis, H1.R and yeastHho1 in HR are notclear. One inhibition of H1.2 with anti-sense RNA resulted in possibility is that when multiple isoforms exist, they reduced apoptosis and very little detectable mitochon- are able to become more specialized. In yeast, where drial release of cytochrome c. Therefore, while the only one linker histone exists, there may be pressure to significance of IR-induced re-localization of all linker maintain certain activities. It is possible that globally, histones to the cytosol is still unknown, H1.2 plays an inhibition of HR is more advantageous than facilitation important role in mediating apoptosis in response to of this pathway. It has been shown that inappropriate DNA DSBs. recombination can result in accelerated aging, and The contribution of linker histones to DDR was indeed, we found that the linker histone in yeast is recently examined in chickens using DT40 cells important for maintaining a full life span. In chickens (Hashimoto et al., 2007). Interestingly, the authors however, if one linker histone that makes up less than found that only one linker histone, H1.R, contributes to one-third of the total linker histone in cells promotes cellular survival after exposure to DNA damage. HR, it may not be globally deleterious to the cell. However, in contrast to what might be predicted based Moreover, one or more of the other linker histones on the yeast studies, the cells lacking H1.R were could inhibitHR. This would allow higher eukaryotes hypersensitive rather than hyper-resistant to MMS. the potential for much greater regulation over genomic Yet, as in yeast, this appeared to be specifically relevant recombination. to HR, as H1.RÀ/À/Rad54À/À double-mutant cell line was One important aspect of this potential ability to regulate no more sensitive than either single. In contrast, a recombination is the affect on regulation and maintenance H1.RÀ/À showed additive effects when combined with of telomeres. In higher eukaryotes, telomerase is inactive in Ku70À/À. The mechanism by which H1.R contributes to somatic cells, and the ‘end replication’ problem contributes cellular survival is notclear, since H2AX phosphoryla- to the mortality of these cells. In the majority of tumours, tion and Rad51 foci formation after MMS appear the cells get around this problem by reactivating normal in the absence of H1.R, and there is no obvious telomerase. However, in a significant proportion of cancer

Oncogene Chromatin structure and DSB responses JA Downs 7768 cells, the need for telomerase is circumvented by hyper- Table 1 Role of linker histones in DNA damage responses and recombination of the repetitive sequences at telomeres to senescence maintain the ends of their in what has been Organism Linker histone Isoform and involvement in dubbed the Alternative Lengthening of Telomeres, or complement DNA damage/tumourigenesis ALT, pathway (Neuman and Reddel, 2002). In yeast, Mus musculus Eightisotypes All isoforms are released from when telomerase is inactivated, most cells die, but a (five somatic) the nucleus in response to IR and minority of cells survive this catastrophic event by using a H1.2 induces apoptosis similar mechanism, which is also dependenton HR (Teng All isoforms disappear during and Zakian, 1999). We found that the ability of yeast cells senescence to use this pathway is negatively influenced by the linker Gallus gallus Six isotypes H1R contributes to HR Saccharomyces One isotype Hho1 inhibits HR histone (Downs et al., 2003). If conserved, this would cerevisiae Hho1 inhibits telomere imply that linker histones can act as tumour suppressor maintenance by the ALT pathway genes by inhibiting immortalization via the ALT pathway (Table 1). Abbreviations: ALT, alternative lengthening of telomeres; HR, While not directly relating to DNA DSB repair, it has ; IR, ionizing radiation. been found recently that linker histones are lost during the process of cellular senescence (Funayama et al., 2006). However, it has been proposed that the senes- cence pathway is important for preventing tumourigen- While there is evidence that all three families impinge esis and this is therefore potentially relevant to the either directly or indirectly on the maintenance of genomic development of cancer (Campisi, 2005). In this study, stability (Reeves and Adair, 2005), the HMGB family has the authors found that in a variety of cell types, there been implicated specifically in DNA DSB repair activities. was a global loss of linker histones during senescence While some proteins containing an HMG box are (Figure 1; Funayama et al., 2006). Interestingly, this sequence specific DNA binding proteins, the HMG correlates with the compaction of chromatin into visible boxes of the abundant chromatin architectural proteins heterochromatic structures. While the linker histone is HMGB1 and HMGB2 have been shown to bind to normally associated with compaction of chromatin, the DNA in a sequence-independentmanner (for review, see authors also find that a number of histone modifications Thomas and Travers, 2001; Travers, 2003). These normally associated with condensed chromatin are not proteins are characterized by two tandem HMG boxes present in these senescent heterochromatic structures, followed by a basic linker region and an acidic C- suggesting that this is an atypical form of condensed terminal tail. The binding of these proteins to DNA chromatin. In addition, the authors also found a induces a bend, and consequently, the proteins have a proportionate increase in this of the higher affinity for pre-bent DNA. In addition, HMGB high- HMGA2 protein. This would suggest that repla- proteins preferentially bind crossover structures and cing linker histones with HMGA2 might help to create a hemicatenated DNA. These activities may suggest a role heterochromatic structure required for cellular senes- for HMGB proteins during HR, where these types of cence. In this model, linker histones would potentially structures exist as intermediates. be oncogenic if they could inappropriately prevent While no evidence currently exists to support a role senescence. However, it is not clear at this stage whether for HMGB proteins in mediating HR, there is some these changes are causative. reason to speculate that HMGB1 may play a role in What is emerging from these studies is that unlike the non-homologous end joining (NHEJ). Specifically, a ability of linker histones to function redundantly in higher number of reports found that the mammalian HMGB eukaryotes with respect to their essential function, the protein HMGB1 is able to promote end joining between differentisoforms can play highly specialized roles in DNA molecules in vitro. These proteins were able to DDR with little if any overlap (Konishi et al., 2003; very efficiently promote intramolecular ligation events Hashimoto et al., 2007). It will therefore be very interesting (Nagaki et al., 1998; Yamanaka et al., 2002), and this is to begin to more thoroughly dissect the activities of these likely to be due to the ability of these proteins to bend individual proteins in DNA repair activities. the DNA such that juxtaposition of the ends is energetically more favourable and therefore promotes successful ligation events. Somewhat more interestingly, intermolecular ligations were also promoted by the High-mobility group box proteins and DNA double-strand presence of HMGB1 (Stros et al., 2000; Yamanaka break repair et al., 2002), which is more relevant to NHEJ activity in the context of the genome. Moreover, using pull-down In addition to linker histones, HMG proteins are assays with DNA molecules labelled with either biotin abundant cellular proteins that regulate chromatin struc- or radioactivity, HMGB1 appears to be able to promote ture (for review, see Bianchi and Agresti, 2005). HMG the noncovalent association of two DNA molecules proteins have been divided into three families based on (Stros et al., 2000). Notably, however, there is very little their domain structure; HMGA (which contain AT evidence to suggest that this activity has any biological hooks), HMGB (which contain HMG Box domains) relevance to NHEJ in vivo. One exception to this is a and HMGN (which contain Nucleosome-binding activity). study showing that cells overexpressing HMGB1 or

Oncogene Chromatin structure and DSB responses JA Downs 7769 HMGB2 have higher rates of integration of transfected cells to radiation as well as chemotherapeutics. A exogenous DNA than the parental cells (Ueda et al., number of these inhibitors are now in clinical trials 2002). However, it is not clear whether the ability to and the data appear very promising (Xu et al., 2007). promote the association of two DNA molecules is the However, it may be worthwhile investigating whether activity responsible for this effect. It may instead be an these have the unintended consequence of promoting effect of altering chromatin structure, affecting gene HMGB1 release and consequently inflammatory re- regulation or binding to integration intermediates, and sponses. The use of HDAC inhibitors in combination this will need to be investigated further. with agents that block the HMGB1–RAGE signalling As discussed above, V(D)J recombination is the pathway may prove to be useful. process in higher eukaryotes, which allows the genera- tion of diversity in immunoglobulin-encoding genes. During this process, the RAG recombinases create DNA DSBs in the genome, which are subsequently H2AX and cancer processed and repaired by NHEJ. As previously mentioned, defects in the ability to detect and repair In addition to the incorporation of the four core histones these programmed DSBs is therefore very deleterious, into nucleosomes, histone variants also exist and can affect and can lead to increased levels of tumourigenesis in the structure and behaviour of chromatin. The H2AX these cell types. It has been known for some time that variant of has an extended C-terminal tail mammalian HMGB proteins, HMGB1 and HMGB2, that contains a consensus site for phosphorylation by ATM facilitate V(D)J recombination. Rather than affecting and ATR (Redon et al., 2002; Foster and Downs, 2005). NHEJ, however, this is because HMGB proteins The phosphorylation of this site in response to DNA promote DNA binding and cleavage by the RAG damage is one of the earliest and most robust alterations to proteins (for example Bergeron et al., 2005; Dai et al., chromatin known to occur in response to DNA damage. 2005 and references therein). Given the behaviour of the While budding yeastdo nothave an H2AX variant,the HMGB proteins in end joining assays in vitro, itwill be phosphorylated motif exists on the core histone H2A interesting to see whether they have additional, sub- protein (Foster and Downs, 2005) and is phosphorylated sequent roles in V(D)J recombination in which they by the homologues of ATM and ATR (Mec1 and Tel1) in promote the repair of the DSBs by NHEJ. response to DNA damage, respectively. While notspecifically relatedtoDNA DSB repair, The potential exploitation of H2AX phosphorylation HMGB proteins show an additional behaviour which for clinical applications related to cancer biology has impacts significantly on tumourigenesis (Figure 1). recently been investigated in several ways. First, H2AX When stressed, some cell types are able to secrete phosphorylation has been shown to act as a readout of HMGB1 and this results in an inflammatory response cellular radiosensitivity, thus acting to predict the (for review, see Raucci et al., 2007). Moreover, HMGB1 efficacy of radiotherapy (Taneja et al., 2004; Munshi is also passively released by necrotic cells, which has the et al., 2006). This suggests that H2AX phosphorylation same inflammatory consequence (Scaffidi et al., 2002). could be used to monitor patient responses to chemo- This response is mediated by HMGB1 binding to the and radiotherapies (for review, see Kao et al., 2006). In receptor for advanced glycation end products (RAGE). support of this possibility, the presence of H2AX foci Crucially, blocking this signalling pathway suppresses was recently examined in skin biopsies of patients tumour growth and metastases (Taguchi et al., 2000), undergoing radiotherapy and found to reproducibly indicating that this inflammatory response is an correspond with the dose given in a linear manner important consequence in the development of cancer. (Qvarnstrom et al., 2004). On the surface, this nuclear release of HMGB1 in Because tumour cells are dividing much more rapidly response to stress appears to be similar to the re- than most other somatic cells, agents that cause DNA localization of the linker histone from the nucleus in damagecanbeusedtoselectivelykillthem.Inthiscase, response to apoptosis-inducing treatment with IR. Inter- residual DNA damage detection and repair activities in the estingly, however, it was demonstrated that HMGB1 is tumour cells are problematic as they can allow some of not released from apoptotic cells induced with etoposide or these cells to evade IR- or chemotherapy-induced cell death. H2O2 (Scaffidi et al., 2002). This suggests that the use of Inhibition of the DDR has been shown to increase the nuclear proteins as signals in response to different types of radiosensitivity of cells and can therefore function to stress is more specialized. When undergoing apoptosis, increase the efficacy of existing cancer therapies (for review, cells sequester HMGB1 in the nucleus perhaps in order to see Madhusudan and Middleton, 2005). Specifically, avoid an unnecessary inflammatory response. It appears numerous reports have demonstrated that when radio- that the ability of cells to sequester HMGB1 to chromatin therapy is combined with drugs inhibiting the activity of the depends on the status of chromatin, and pre- ATM and ATR family of kinases, there is greater cell death treatment with the (HDAC) inhibitor than radiotherapy alone (for example, Cowell et al., 2005 Trichostatin A (TSA) prevents retention of HMGB1 and references therein). It has been shown that the gene (Scaffidi et al., 2002). encoding H2AX acts as a tumour suppressor gene in mice As previously mentioned, HDAC inhibitors have (Bassing et al., 2003; Celeste et al., 2003) and repression of strong anti-proliferative effects on transformed cells the H2AX gene in mouse teratocarcinoma cells results in relative to normal cells and enhance the sensitivity of increased radiosensitivity (Yoshida and Morita, 2004),

Oncogene Chromatin structure and DSB responses JA Downs 7770 suggesting that phosphorylation of H2AX provides an tumourigenesis (Peters et al., 2001; Gaudet et al., 2003). additional target for sensitizing cells to IR. Indeed, First, mice carrying a hypomorphic allele of the DNA treatment of tumour cells with peptide mimics of the methyltransferase 1 gene developed aggressive T-cell phosphorylated H2AX tail had no effect on un-irradiated lymphomas (Gaudet et al., 2003). While the mechanism tumour cells but enhanced cell death in irradiated radio- by which this occurs was not definitively elucidated, the resistant tumour cells (Taneja et al., 2004). Together, these authors found an effect on chromosomal instability and data support H2AX phosphorylation as a potential postulated that this was the mechanism by which the therapeutic target for improving the efficacy of radio- mice developed tumours. In addition, mice deficient in therapy. Because H2AX is also phosphorylated in response the Suv39 h histone methyltransferase, which targets to DNA damage caused by many chemotherapeutic agents K9, display chromosomal instabilities that (described in more detailbelow), itis notunreasonable to are associated with increased rates of tumourigenesis assume that H2AX phosphorylation may provide a target (Peters et al., 2001). It seems reasonable to predict that for improving the efficacy of chemotherapy as well. more genes involved in chromatin modulation will be Many drugs are identified as chemotherapeutic agents found to impact on tumourigenesis, and at least some of based on their growth inhibitory activity towards cells these will do so via influencing DDR. without any mechanistic insights about the molecular basis behind it. Thus, H2AX phosphorylation has been used to investigate the cellular and molecular effect of Histone modifications and cancer therapy and diagnostics: known anti-cancer agents, such as aminoflavone, which beyond H2AX was consequently found to activate the DDR (Meng et al., 2005). Inhibitors of HDACs function as very As discussed above with regard to H2AX phosphoryla- promising anti-cancer agents, but the mechanism by tion, inhibition of activities that promote or facilitate which they function in this regard is not known (Xu DNA repair can be clinically useful. Novel changes in et al., 2007). It has been postulated that they function to chromatin structure, modification patterns and organi- alter the transcription of oncogenes or tumour suppres- zation in response to DNA damage therefore have sor genes that have been misregulated due to epigenetic relevance to cancer biology since the identification of alterations common in cancer. Another possible mechan- additional DNA damage-dependent activities can iden- ism is that they influence the DNA repair pathway, tify novel targets for therapeutic drugs. which is regulated by acetylation (Hasan and Hottiger, In addition to therapeutic value, identification of 2002; Sun et al., 2005; Murr et al., 2006). Recently, it has novel DNA damage-dependent chromatin alterations been found that treatment of cells with HDAC inhibitors can serve as new biomarkers. Again, as described above, enhances the sensitivity of cancer cells to ionizing H2AX has been exploited in this way. Importantly radiation and DNA damage inducing chemotherapies however, studies have also found that H2AX phosphor- (Dokmanovic and Marks, 2005). H2AX phosphoryla- ylation does not function reliably as a predictor of tion was used to investigate this phenomenon and the radiosensitivity in all cases (Mahrhofer et al., 2006; Yu results suggest that HDAC inhibitors function as anti- et al., 2006). Identification of other DNA damage- cancer agents, at least in part, by inhibiting the repair of dependentmodificationscould provide more accurate DNA damage (Munshi et al., 2006; Qian et al., 2006). biomarkers. More likely, these could be used in Finally, the use of H2AX phosphorylation has been combination with our existing knowledge to provide adapted to determine whether drug treatments, such as more robust biomarkers and drug targets. norethindrone used in hormone replacement therapy, are Clearly, in light of the studies described above, the genotoxic (Gallmeier et al., 2005). DNA patterns and levels of H3 K9 methylation in tumours may turn out to be useful in this regard. In addition, two studies examined changes in Regulation of chromatin and tumourigenesis: a causative histone modifications that take place during tumour- link? igenesis without examining causality (Seligson et al., 2005). In the first, was isolated and analysed As discussed above, the loss of DDR is an enabling by mass spectroscopy. The authors found that a decrease characteristic which allows the generation of alterations in acetylation of K16 and trimethylation of K20 was to the genome that can result in tumourigenesis commonly found in cancerous cells and tissues compared (Hanahan and Weinberg, 2000). Also described above, with their normal counterparts (Fraga et al., 2005). In the most well-studied chromatin modification in DDR is the second study, specific antibodies against five the phosphorylation of H2AX and this has now been modifications (acetylated h3 K9, K18 and H4 K12 and exploited for use in clinical applications, and elegant dimethylated H4 R3 and H3 K4) were used to examine studies in mouse systems have demonstrated that the prostate tissue. The pattern of modifications was found gene encoding H2AX functions as a tumour suppressor to be a strong predictor of outcome, indicating that gene indicating a direct role in preventing tumourigen- histone modification can be used to determine prognosis esis (Bassing et al., 2003; Celeste et al., 2003). (Seligson et al., 2005). One intriguing possibility is that Two recent reports identify two further enzymes that some of these changes may be a part of the DDR that is affect chromatin structure whose loss can lead to the activated during tumourigenesis (Bartkova et al., 2005; induction of tumours suggesting a causal link with Gorgoulis et al., 2005), and these may therefore be useful

Oncogene Chromatin structure and DSB responses JA Downs 7771 not only as biomarkers, but as potential therapeutic critical to examine more complex chromatin changes. targets, as with H2AX phosphorylation. While not addressed here, the importance of ATP- dependentremodelling activitiesin DDR has also been recognized and is intimately related to the modification Summary and future perspectives pattern of chromatin (for review, see Downs et al., 2007). But additionally, proteins such as linker histones A picture is beginning to emerge about changes in and HMG proteins, which have received less attention, chromatin structure that occur during tumourigenesis. will need to be more thoroughly investigated with regard Obviously, only a subsetof thesechanges will be partof to their role in both DDR and tumourigenesis. Finally, the cellular response to DNA damage, whose loss may it will be important not only to more thoroughly identify contribute to ‘enabling’ tumourigenesis. At the moment, and investigate these changes in chromatin structure, one major focus has been on changes to the patterns of but also to dissect out changes that are part of the DDR covalent modifications on histone proteins. While this from those involved in other cellular activities that may information is fundamentally important, it will also be change during the course of tumourigenesis.

References

Alami R, Fan T, Pack S, Sonbuchner TM, Besse A, Lin Q Downs JA, Nussenzweig MC, Nussenzweig A. (2007). et al. (2003). Mammalian linker-histone subtypes differen- Chromatin dynamics and the preservation of genetic tially affect in vivo. Proc Natl Acad Sci 100: information. Nature 447: 951–958. 5920–5925. Fan Y, Nikitina T, Zhao J, Fleury TJ, Bhattacharyya R, Ausio J. (2000). Are linker histones (histone H1) dispensable Bouhassira EE et al. (2005). Histone h1 depletion in for survival? BioEssays 22: 873–877. mammals alters global chromatin structure but causes Bartkova J, Horejsi Z, Koed K, Kramer A, Tort F, Zieger K specific changes in gene regulation. Cell 123: 1199–1212. et al. (2005). DNA damage response as a candidate anti-cancer Foster ER, Downs JA. (2005). Histone H2A phosphorylation barrier in early human tumorigenesis. Nature 434: 864–870. in DNA double-strand break repair. FEBS Journal 272: Bassing CH, Suh H, Ferguson DO, Chua KF, Manis J, 3231–3240. Eckersdorf M et al. (2003). Histone H2AX: a dosage- Fraga MF, Ballestar E, Villar-Garea A, Boix-Chornet M, dependent suppressor of oncogenic translocations and Espada J, Schotta G et al. (2005). Loss of acetylation at tumors. Cell 114: 359–370. Lys16 and trimethylation at Lys20 of histone H4 is a Bergeron S, Madathiparambil T, Swanson PC. (2005). Both common hallmark of human cancer. Nat Genet 37: 391–400. high mobility group (HMG)-boxes and the acidic tail of Franco S, Gostissa M, Zha S, Lombard DB, Murphy MM, HMGB1 regulate recombination-activating gene (RAG)- Zarrin AA et al. (2006). H2AX prevents DNA breaks from mediated recombination signal synapsis and cleavage in progressing to chromosome breaks and translocations. Mol vitro. J Biol Chem 280: 31314–31324. Cell 21: 201–214. Bianchi ME, Agresti A. (2005). HMG proteins: dynamic Funayama R, Saito M, Tanobe H, Ishikawa F. (2006). Loss players in gene regulation and differentiation. Curr Opin of linker histone H1 in cellular senescence. J Cell Biol 175: Genet Dev 15: 496–506. 869–880. Bredemeyer AL, Sharma GG, Huang CY, Helmink BA, Gallmeier E, Winter JM, Cunningham SC, Kahn SR, Kern Walker LM, Khor KC et al. (2006). ATM stabilizes DNA SE. (2005). Novel genotoxicity assays identify norethindrone double-strand-break complexes during V(D)J recombina- to activate p53 and phosphorylate H2AX. tion. Nature 442: 466–470. 26: 1811–1820. Campisi J. (2005). Suppressing cancer: the importance of being GaudetF, Hodgson JG, Eden A, Jackson-Grusby L, Daus- senescent. Science 309: 886–887. man J, Gray JW et al. (2003). Induction of tumors in mice Celeste A, Difilppantonio S, Difilppantonio MJ, Fernandez- by genomic hypomethylation. Science 300: 489–492. Capetillo O, Pilch DR, Sedelnikova OA et al. (2003). H2AX Gorgoulis VG, Vassiliou LV, Karakaidos P, Zacharatos P, haploinsufficiency modifies genomic stability and tumor Kotsinas A, Liloglou T et al. (2005). Activation of the DNA susceptibility. Cell 114: 371–383. damage checkpoint and genomic instability in human Costelloe T, Fitzgerald J, Murphy NJ, Flaus A, Lowndes NF. precancerous lesions. Nature 434: 907–913. (2006). Chromatin modulation and the DNA damage Hanahan D, Weinberg RA. (2000). The hallmarks of cancer. response. Exp Cell Res 312: 2677–2686. Cell 100: 57–70. Cowell IG, Durkacz BW, Tilby MJ. (2005). Sensitization of Harvey AC, Downs JA. (2004). Whatfunctionsdo linker breast carcinoma cells to ionizing radiation by small histones provide? Mol Microbiol 53: 771–775. molecule inhibitors of DNA-dependent protein kinase and Hasan S, Hottiger MO. (2002). Histone acetyl transferases: a ataxia telangiectsia mutated. Biochem Pharmacol 71: 13–20. role in DNA repair and DNA replication. J Mol Med 80: Dai Y, Wong B, Yen YM, Oettinger MA, Kwon J, Johnson 463–474. RC. (2005). Determinants of HMGB proteins required to Hashimoto H, Sonoda E, Takami Y, Kimura H, Nakayama T, promote RAG1/2-recombination signal sequence complex Tachibana M et al. (2007). Histone H1 variant, H1R is assembly and catalysis during V(D)J recombination. Mol involved in DNA damage response. DNA Repair (in press). Cell Biol 25: 4413–4425. Jackson SP. (2002). Sensing and repairing DNA double-strand Dokmanovic M, Marks PA. (2005). Prospects: histone breaks. Carcinogenesis 23: 687–696. deacetylase inhibitors. J Cell Biochem 96: 293–304. Kao J, Milano MT, Javaheri A, Garofalo MC, Chmura SJ, Downs JA, Kosmidou E, Morgan A, Jackson SP. (2003). Weichselbaum RR et al. (2006). gamma-H2AX as a Suppression of homologous recombination by the Sacchar- therapeutic target for improving the efficacy of radiation omyces cerevisiae linker histone. Mol Cell 11: 1685–1692. therapy. Curr Cancer Drug Targets 6: 197–205.

Oncogene Chromatin structure and DSB responses JA Downs 7772 Konishi A, Shimizu S, Hirota J, Takao T, Fan Y, Matsuoka Y Reina-San-Martin B, Difilppantonio S, Hanitsch L, Masila- et al. (2003). Involvement of histone H1.2 in apoptosis mani RF, Nussenzweig A, Nussenzweig MC. (2003). induced by DNA double-strand breaks. Cell 114: 673–688. H2AX is required for recombination between immunoglo- Lavin MF, Delia D, Chessa L. (2006). ATM and the DNA bulin switch regions but not for intra-switch region damage response. EMBO Rep 7: 154–160. recombination or somatic hypermutation. J Exp Med 197: Luger K, Mader AW, Richmond RK, SargentDF, Richmond 1767–1778. TJ. (1997). Crystal structure of the nucleosome core particle Scaffidi P, Misteli T, Bianchi ME. (2002). Release of at2.8 A resolution. Nature 389: 251–260. chromatin protein HMGB1 by necrotic cells triggers Madhusudan S, Middleton MR. (2005). The emerging role inflammation. Nature 418: 191–195. of DNA repair proteins as predictive, prognostic and Seligson DB, Horvath S, Shi T, Yu H, Tze S, Grunstein M therapeutic targets in cancer. Cancer Treat Rev 31: 603–617. et al. (2005). Global histone modification patterns Mahrhofer H, Burger S, Oppitz U, Flentje M, Djuzenova CS. predict risk of prostate cancer recurrence. Nature 435: (2006). Radiation induced DNA damage and damage repair 1262–1266. in human tumor and fibroblast cell lines assessed by histone Stros M, Cherny D, Jovin TM. (2000). HMG1 protein H2AX phosphorylation. Int J Radiat Oncol Biol Phys 64: stimulates DNA end joining by promoting association 573–580. of DNA molecules via their ends. Eur J Biochem 267: Meng LH, Kohlhagen G, Liao ZY, Antony S, Sausville E, 4088–4097. Pommier Y. (2005). DNA-protein cross-links and replica- Sun Y, Jiang X, Chen S, Fernandes N, Price BD. (2005). A tion-dependent histone H2AX phosphorylation induced role for the Tip60 histone acetyltransferase in the acetylation by aminoflavone (NSC 686288), a novel anticancer agent and activation of ATM. Proc Natl Acad Sci USA 102: active against human breast cancer cells. Cancer Res 65: 13182–13187. 5337–5343. Taguchi A, Blood DC, del Toro G, CanetA, Lee DC, Qu W Michor F. (2005). Chromosomal instability and human cancer. et al. (2000). Blockade of RAGE-amphoterin signal- Philos Trans R Soc Lond B Biol Sci 360: 631–635. ling suppresses tumour growth and metastases. Nature Morrison AJ, Shen X. (2005). DNA repair in the context of 405: 354–360. chromatin. Cell Cycle 4: 568–571. Taneja N, Davis M, Choy JS, Beckett MA, Singh R, Kron SJ Munshi A, Tanaka T, Hobbs ML, Tucker SL, Richon VM, et al. (2004). Histone H2AX phosphorylation as a predictor Meyn RE. (2006). Vorinostat, a histone deacetylase of radiosensitivity and target for radiotherapy. J Biol Chem inhibitor, enhances the response of human tumor cells to 279: 2273–2280. ionizing radiation through prolongation of gamma-H2AX Teng S-C, Zakian VA. (1999). Telomere–telomere recombina- foci. Mol Cancer Ther 5: 1967–1974. tion is an efficient bypass pathway for telomere maintenance Murr R, Loizou JI, Yang YG, Cuenin C, Li H, Wang ZQ et al. in Saccharomyces cerevisiae. Mol Cell Biol 19: 8083–8093. (2006). Histone acetylation by Trrap-Tip60 modulates Thomas JO, Travers AA. (2001). HMG1 and 2, and related loading of repair proteins and repair of DNA double-strand ‘archetectural’ DNA binding proteins. Trends Biochem Sci breaks. Nat Cell Biol 8: 91–99. 26: 167–174. Nagaki S, Yamamoto M, Yumoto Y, Shirakawa H, Yoshida Travers AA. (2003). Priming the nucleosome: a role for M, Teraoka H. (1998). Non-histone chromosomal proteins HMGB proteins? EMBO Rep 4: 131–136. HMG1 and 2 enhance ligation reaction of DNA double- Ueda T, Shirakawa H, Yoshida M. (2002). Involvementof strand breaks. Biochem Biophys Res Commun 246: 137–141. HMGB1 and HMGB2 proteins in exogenous DNA Neuman AA, Reddel RR. (2002). Telomere maintenance and integration reaction into the genome of HeLa S3 cells. cancer- look, no telomerase. Nat Rev Cancer 2: 879–884. Biochim Biophys Acta 1593: 77–84. Peters AH, O’Carroll D, Scherthan H, Mechtler K, Sauer S, van Attikum H, Gasser SM. (2005). The histone code at Schofer C et al. (2001). Loss of the Suv39 h histone DNA breaks: a guide to repair? Nat Rev Mol Cell Biol 6: methyltransferases impairs mammalian heterochromatin 757–765. and genome stability. Cell 107: 323–337. Vogel H, Lim D-S, Karsenty G, Finegold M, Hasty P. (1999). Peterson CL, Cote J. (2004). Cellular machineries for Deletion of Ku86 causes early onset of senescence in mice. chromosomal DNA repair. Genes Dev 18: 602–616. Proc Natl Acad Sci 96: 10770–10775. Qian X, LaRochelle WJ, Ara G, Wu F, Petersen KD, Williams RS, Williams JS, Tainer JA. (2007). Mre11-Rad50- Thougaard A et al. (2006). Activity of PXD101, a histone Nbs1 is a keystone complex connecting DNA repair deacetylase inhibitor, in preclinical ovarian cancer studies. machinery, double-strand break signaling, and the chroma- Mol Cancer Ther 5: 2086–2095. tin template. Biochem Cell Biol 85: 509–520. Qvarnstrom OF, Simonsson M, Johansson KA, Nyman J, Xu WS, Parmigiani RB, Marks PA. (2007). Histone deacety- Turesson I. (2004). DNA double strand break quantification lase inhibitors: molecular mechanisms of action. Oncogene in skin biopsies. Radiother Oncol 72: 311–317. 26: 5541–5552. Ramiro AR, Jankovic M, Callen E, Difilippantonio S, Chen Yamanaka S, Katayama E, Yoshioka K, Nagaki S, Yoshida M, HT, McBride KM et al. (2006). Role of genomic instability Teraoka H. (2002). Nucleosome linker proteins HMGB1 and and p53 in AID-induced c-myc-Igh translocations. Nature histone H1 differentially enhance DNA ligation reactions. 440: 105–109. Biochem Biophys Res Commun 292: 268–273. Raucci A, Palumbo R, Bianchi ME. (2007). HMGB1: a signal Yoshida K, Morita T. (2004). Control of radiosensitivity of F9 of necrosis. Autoimmunity 40: 285–289. mouse teratocarcinoma cells by regulation of histone H2AX Redon C, Pilch DR, Rogakou E, Sedelnikova OA, Newrock gene expression using a tetracycline turn-off system. Cancer K, Bonner WM. (2002). Histone H2A variants H2AX and Res 64: 4131–4136. H2AZ. Curr Opin Genet Dev 12: 162–169. Yu T, MacPhail SH, Banath JP, Klokov D, Olive PL. (2006). Reeves R, Adair JE. (2005). Role of high mobility group Endogenous expression of phosphorylated histone H2AX in (HMG) chromatin proteins in DNA repair. DNA Repair tumors in relation to DNA double-strand breaks and (Amst) 4: 926–938. genomic instability. DNA Repair (Amst) 5: 935–946.

Oncogene