In vivo role of checkpoint kinase 2 in signaling telomere dysfunction

Checkpoint kinase 2 (CHK2) is a downstream effector of the DNA damage response (DDR). Dysfunctional telomeres, either owing to critical shortening or disruption of the shelterin complex, activate a DDR, which eventually results in cell cycle arrest, senescence and/or apoptosis. Successive generations of telomerase-deficient (Terc) mice show accelerated aging and shorter lifespan due to tissue atrophy and impaired organ regeneration associated to progressive telomere shortening. In contrast, mice deficient for the shelterin component TRF1 in stratified epithelia show a rapid and massive induction of DDR, leading to perinatal lethality and severe skin defects. In both mouse models, p53 deficiency can rescue survival. Here, we set to address the role of CHK2 in signaling telomere dysfunction in both mouse models. To this end, we generated mice doubly deficient for Chk2 and either Terc (Chk2−/− Terc−/−) or Trf1 (Trf1Δ/Δ K5Cre Chk2−/−). We show that Chk2 deletion improves Terc-associated phenotypes, including lifespan and age-associated pathologies. Similarly, Chk2 deficiency partially rescues perinatal mortality and attenuates degenerative pathologies of Trf1Δ/Δ K5Cre mice. In both cases, we show that the effects are mediated by a significant attenuation of p53/p21 signaling pathway. Our results represent the first demonstration of a role for CHK2 in the in vivo signaling of dysfunctional telomeres.


Introduction
Telomere dysfunction is caused by either critical telomere shortening or loss of the protein complex that protects telomeres, the so-called shelterin (de Lange, 2005a). Dysfunctional telomeres behave as double strand breaks (DSBs) and trigger a persistent DNA damage response (DDR) (d'Adda di Fagagna et al., 2003;de Lange, 2005a). DSBs are among the most deleterious lesions that challenge genomic integrity. Concomitant to the repair of DSBs, a rapid signaling cascade must be coordinated at the lesion site to prevent cell division by activation of cell cycle checkpoints. Ataxia telangiectasia mutated (ATM) and ATM and Rad-3-related (ATR) protein kinases are among the earliest signaling molecules known to initiate the transduction cascade at damage sites. This cascade ultimately results in the activation of p21 and p53, leading to senescence/apoptosis von Zglinicki et al., 2005). ATM has an essential role in signaling from DSBs arising from ionizing radiation (IR) through a CHK2-dependent pathway, while ATR is typically involved in signaling from replication-linked SSBs through the CHK1 kinase. Recent findings, however, have demonstrated an active cross talk between ATM and ATR signaling pathways in response to DNA damage Murga et al., 2009). Both CHK1 and CHK2 can activate p53 upon DNA damage (Shieh et al., 2000) and are important for cell cycle checkpoints, however, while CHK1 is crucial for intra-S and G2/S transition, CHK2 is more important for G1/S checkpoint (Liu et al., 2000;Takai et al., 2000Takai et al., , 2002Hirao et al., 2002).
Telomere shortening is one of the molecular pathways underlying organismal aging (Lopez-Otin et al., 2013). Telomerase-deficient mice show progressive telomere shortening, eventually impairing tissue regeneration and leading to tissue atrophy and premature aging phenotypes, as well as to cancer resistance (Greenberg et al., 1998;Lee et al., 1998;Rudolph et al., 1999;Gonzalez-Suarez et al., 2000). p53 deficiency can rescue both survival and tumorigenesis in Terc À/À p53 À/À mice, indicating that p53 is one of the main downstream mediators of the cellular response to DDR triggered by short telomeres (Artandi et al., 2000).
TRF1 is one of the six components of shelterin, which has an essential role in protecting telomeres from fusions and from telomere fragility resulting from telomere replication (De Lange, 2005b). Abrogation of Trf1 in mice results in early embryonic lethality (Karlseder et al., 2003). Conditional deletion of Trf1 in mouse stratified epithelia leads to perinatal lethality and severe skin morphogenesis defects, which are concomitant with rapid induction of telomere-originated DNA damage and activation of the p53/p21 pathway (Martinez et al., 2009). Similar to that shown for Terc-deficient mice, deletion of p53 in Trf1 D/D K5Cre mice rescues survival and leads to increased cancer incidence, indicating that p53 is also a main mediator of the cellular response to DDR induced by uncapped telomeres (Martinez et al., 2009).
Interestingly, it has been recently shown that deletion of Trf1 induces phosphorylation of both the ATR and ATM downstream kinases CHK1 and CHK2, respectively, suggesting that both checkpoint kinases may be important to mediate telomere dysfunction owing to telomere uncapping (Martinez et al., 2009;Sfeir et al., 2009). In line with this, inhibition of either ATM or ATR kinases in Trf1-deficient MEFs can significantly rescue DNA damage in vitro (Martinez et al., 2009). However, whether attenuation of these pathways rescues Trf1 deficiency-associated phenotypes in vivo is not known.
Several studies have previously addressed the role of CHK2 in the response to dysfunctional telomeres in vitro cultured human cells (Thanasoula et al., 2012;Cesare et al., 2013). Thus, Cesare et al. (2013) showed that telomere deprotection resuting from TRF2 downregulation to levels that do not induce telomere-telomere fusions, results in p53mediated G1 cell cycle arrest independently of CHK2 phosphorylation. Only when TRF2 was fully abrogated leading to chromosome fusions, the authors observed CHK2 phosphorylation (Cesare et al., 2013). The authors concluded that very few TRF2 molecules at telomeres are sufficient to prevent the nonhomologous end joining pathway (NHEJ) and to inhibit CHK2 activation, thereby allowing the cells to continue through mitosis avoiding the G2/M arrest (Cesare et al., 2013). In contrast, Thanasoula et al. (2012) showed that TRF2 depletion leads to CHK2 phosphorylation and to G2/M arrest, preventing fusions and genome instability.
In the case of the mouse, we have previously shown that dysfunctional telomeres induced by Trf1 deletion lead to Chk2 phosphorylation and G2/M arrest (Martinez et al., 2009). Furthermore, abolishment of end-to-end fusions owing to Trf1 deficiency by simultaneously deleting 53 bp1, and essential component of the NHEJ, does not inhibit Chk2 phosphorylation or 'in vivo' G2/M arrest (Martinez et al., 2009). Similarly, mouse cells depleted for Trf2 and lacking ligase 4, another essential component of the NHEJ, do not show telomere-telomere fusions but show Chk2 phosphorylation (Celli & de Lange, 2005). Thus, in contrast to human cells, mouse telomere dysfunction induced by either depleting TRF1 or TRF2 leads to Chk2 phosphorylation independently of chromosome fusions.
Here, we set to address the role of CHK2 activation in response to dysfunctional telomeres in vivo. In particular, we set to address whether Chk2 deficiency could rescue phenotypes associated to telomere dysfunction in Terc and Trf1 deficient mouse models previously generated by us (Blasco et al., 1997;Martinez et al., 2009).
Chk2-deficient mice are viable and do not show increased cancer incidence (Donehower et al., 1992;Barlow et al., 1996;Hirao et al., 2002), constituting an interesting model to test the importance of this checkpoint kinase in mediating telomere dysfunction in vivo. Thus, here, we generated Trf1 D/D K5Cre Chk2 À/À and Chk2 À/À Terc À/À compound mice. The findings described here indicate that Chk2 abrogation can partially rescue both survival and some of the phenotypes associated to both Terc and Trf1 deficiencies, revealing a potential for Chk2 inhibition in ameliorating phenotypes associated to both critically short telomeres and severe telomere dysfunction.

Results
Chk2 deficiency rescues mouse survival and degenerative pathologies in Trf1-deficient mice To address the in vivo role of CHK2 in signaling telomere dysfunction owing to Trf1 deficiency, we crossed Trf1 D/D K5Cre mice (Martinez et al., 2009) with mice with whole-body Chk2 deletion, Chk2 À/À (Hirao et al., 2002). Interestingly, Chk2 deficiency significantly rescued the survival of Trf1 D/D K5Cre mice, with one doubly deficient mice surviving for as long as 1 year of age compared to a maximum survival of 4 days for single Trf1-deficient littermates (Fig. 1A). On average, the median survival of the Trf1 D/D K5Cre Chk2 À/À mice was threefold higher than the Trf1 D/D K5Cre mice, 0.9 and 0.3 weeks, respectively (Fig. 1A). In agreement with the increased survival, we also observed a significant rescue of body weight in Trf1 D/D K5Cre Chk2 À/À newborns compared to the Trf1 D/D K5Cre newborns ( Fig. 1B-C).
In an analogous manner, the frequency and severity of epithelial pathologies associated to Trf1 deficiency such as hair follicle atrophy, skin interfollicular dysplasia and esophagus dysplasia were also significantly rescued in Trf1 D/D K5Cre Chk2 À/À newborns (1-2 day old mice) compared to the Trf1 D/D K5Cre Chk2 +/+ age-matched controls (Fig 1D,  E). Aberrant overexpression of cytokeratin 6 associated to Trf1 deficiency was also rescued in the Chk2-deficient background (Fig. 1D,E). Interestingly, at 9 days of age, at which point all single mutant Trf1 D/D K5Cre mice were dead, surviving Trf1 D/D K5Cre Chk2 À/À mice showed extensive dysplasia in the skin, suggesting an increased preneoplastic growth associated to massive Trf1-induced DNA damage in the absence of Chk2 Chk2 deficiency reduces p53 induction and increases proliferation of the skin of K5Cre Trf1 D/D Chk2 À/À newborns To further investigate how Chk2 deficiency could rescue both the severity of the skin pathologies and survival of Trf1 D/D K5Cre Chk2 À/À mice compared to Trf1 D/D K5Cre littermates, we first set to quantify DNA damage in the skin of these mice by determining the number of cells showing DNA damage foci as indicated by staining with cH2Ax and 53BP1. Both Trf1 D/D K5Cre Chk2 +/+ and Trf1 D/D K5Cre Chk2 À/À newborn skin showed approximately 80% of the basal cells positive for cH2Ax and 53BP1 staining (>5 foci per cell), indicating similarly high amounts of DNA damage in both cases, in agreement with similarly high levels of telomere uncapping owing to TRF1 deficiency ( Fig. 2A-B). Conditional Trf1deletion in MEFs by adenoviral infection with the Cre recombinase induces Chk2 phosphorylation (Fig. S1A,B) (Martinez et al., 2009). We also found similar amounts of DNA damage, as well as similar frequencies of chromosomal aberrations associated to Trf1 deficiency in in vitro grown MEFs deficient for both Trf1 and Chk2 (Fig. S1), thus confirming the in vivo findings.
Chk2 deficiency in Trf1 D/D K5Cre Chk2 À/À mice leads to epithelial abnormalities characteristic of telomere syndromes, including increased skin cancer Interestingly, cell division in the presence of persistent DNA damage in long-lived Trf1 D/D K5Cre Chk2 À/À mice resulted in development of epithelial pathologies, which recapitulate some of the skin abnormalities characteristic of the human telomere syndromes, such as occurrence of nail dystrophy and fibropapilomas (Fig. 2G,H). In addition, the back skin displayed preneoplasic lesions such as epithelial hyperplasia and dysplasia, orthokeratotic hyperkeratosis, and dermal infiltrates. In addition, long-lived Trf1 D/D K5Cre Chk2 À/À mice also showed spontaneous development of invasive squamous cell carcinomas (SCC) both in the tail and ear skin (Fig. 2G,H).
Altogether, these results demonstrate that Chk2 deletion attenuates induction of p53/p21 as well as rescues proliferative defects in the skin of Trf1 D/D K5Cre mice, which in turn leads to a partial rescue of survival but also to increased tumorigenesis.

Discussion
Here, we studied the role of the checkpoint kinase Chk2 in mediating the in vivo phenotypes induced by telomere dysfunction owing to either critical telomere shortening (Terc deficiency) or severe telomere uncapping (Trf1 deficiency). We found that Chk2 deficiency partially rescues both Terc deficiency in early generations and Trf1 deficiency-associated phenotypes, thus demonstrating a role for Chk2 in signaling telomere dysfunction in vivo. Furthermore, we demonstrate that this rescue is not mediated by a significant rescue of the amount of DNA damage but instead reflects an attenuation of the proliferative defects associated to p53/p21 induction. A previous work reported that Chk2 deletion in mice did not rescue survival or pathologies associated to Terc deficiency (Nalapareddy et al., 2010). Although our findings may seem to stand in contrast to those reported by Nalapareddy (Nalapareddy et al., 2010), it is relevant to note that these authors only focused their study in late generation iG4 telomerase-deficient mice which correspond to our G3 generation bearing very short telomeres and where we also do not see Chk2-mediated rescue. In contrast to our current study, Nalapareddy et al. did not address the effect of Chk2 deficiency in signaling critically short telomeres in G1 or G2 telomerase-deficient mice, which is where we observe a significant partial rescue in median survival by Chk2 deficiency. In addition, the different outcomes might also be due to the different ways by which both groups have generated the Terc-deficient mouse cohorts. Thus, Nalapareddy et al. crossed G3 Terc À/À Chk2 +/À with Terc +/À Chk2 +/À to generate their experimental cohorts; namely, Terc +/À Chk2 +/+ (iF1 Chk2 +/+ ), Terc +/À Chk2 À/À (iF1 Chk2 À/À ), Terc À/À Chk2 +/+ (iG4 Chk2 +/+ ), and Terc À/À Chk2 À/À (iG4 Chk2 +/+ ). The zygotes in all these cohorts harbor half of its chromosomes with long telomeres (those coming from the Terc +/À parental) and half with short telomeres (those coming from the G3 Terc À/À parental). In our work, we carried the Terc À allele in homozygosis from the first generation of G1 Terc À/À Chk2 +/À throughout the G2 and G3 generations, thus in a setting where all chromosomes present progressively shorten telomeres. It is conceivable that the Chk2-mediated response in G3 Terc-deficient mice is masked by a compensatory induction of a Chk2-independent mechanism that leads to p53/p21 activation. In fact, Nalapareddy et al. detected significantly higher levels of phosphorylated Chk1 in iG4 Chk2 À/À as compared to iG4 Chk2 +/+ (Nalapareddy et al., 2010). In agreement with Nalapareddy et al., we also detected a slight increase in Chk1 phosphorylation in the absence of Chk2 in Trf1-deficient keratinocytes, which explains the fact that p53 levels were still elevated in Chk2 cohorts compared to the wild-types. This fact also explains that Chk2 deficiency only partially rescued survival and phenotypes associated to Trf1 deficiency, while p53 deficiency was previously described by us to fully rescue Trf1 D/D K5-Cre perinatal lethality (Martinez et al., 2009). Thus, in the absence of Chk2, other components of the DDR, including Chk1 may be important to signal telomere dysfunction.
Finally, although single Chk2 deficiency has not been associated to increased cancer (Donehower et al., 1992;Barlow et al., 1996;Hirao et al., 2002), we found a slightly increased cancer incidence in the absence of Chk2 in the context of both Trf1 and Terc deficiencies. These findings are in line, with increased cancer incidence associated to telomere dysfunction in the absence of p53. Interestingly, abrogation of other components of the DDR, such as p21, Pms2 and Exo-1, can rescue survival of Terc-deficient mice without increasing cancer (Choudhury et al., 2007;Schaetzlein et al., 2007;Siegl-Cachedenier et al., 2007).
In summary, our findings represent the first demonstration in vivo for a role of Chk2 in signaling telomere dysfunction owing to both short telomeres and severe telomere uncapping. Furthermore, we show a role for Chk2 in blocking tumorigenesis associated to dysfunctional telomeres.

Supporting Information
Additional Supporting Information may be found in the online version of this article at the publisher's web-site.  Data S1 Experimental procedures. CHK2 role in signaling dysfunctional telomeres, M. García-Beccaria et al.