MutS homologue hMSH5: role in cisplatin-induced DNA damage response
© Tompkins et al; licensee BioMed Central Ltd. 2012
Received: 15 September 2011
Accepted: 8 March 2012
Published: 8 March 2012
Cisplatin (cis-diamminedichloroplatinum (II), CDDP) and its analogues constitute an important class of anticancer drugs in the treatment of various malignancies; however, its effectiveness is frequently affected by mutations in genes involved in the repair and signaling of cisplatin-induced DNA damage. These observations necessitate a need for a better understanding of the molecular events governing cellular sensitivity to cisplatin.
Here, we show that hMSH5 mediates sensitization to cisplatin-induced DNA damage in human cells. Our study indicates that hMSH5 undergoes cisplatin-elicited protein induction and tyrosine phosphorylation. Silencing of hMSH5 by RNAi or expression of hMSH5 phosphorylation-resistant mutant hMSH5Y742F elevates cisplatin-induced G2 arrest and renders cells susceptible to cisplatin toxicity at clinically relevant doses. In addition, our data show that cisplatin promotes hMSH5 chromatin association and hMSH5 deficiency increases cisplatin-triggered γ-H2AX foci. Consistent with a possible role for hMSH5 in recombinational repair of cisplatin-triggered double-strand breaks (DSBs), the formation of cisplatin-induced hMSH5 nuclear foci is hRad51-dependent.
Collectively, our current study has suggested a role for hMSH5 in the processing of cisplatin-induced DSBs, and silencing of hMSH5 may provide a new means to improve the therapeutic efficacy of cisplatin.
KeywordshMSH5 hMSH4 c-Abl Cisplatin Homologous recombination
Cisplatin (cis-diamminedichloroplatinum (II))
MutS homologue 5
Nucleotide excision repair
Despite being members of the MMR protein family, the MSH5 homologues have not been demonstrated to function in MMR. Instead, studies in mice, C. elegans, and S. cerevisiae have shown that MSH5 plays an array of diverse functions ranging from meiotic recombinational DSB repair, maintenance of chromosome integrity, to DNA damage response [1–6]. Purified hMSH4-hMSH5 protein complexes have been shown to possess binding activities towards recombination intermediate structures including the Holliday junction , and endogenous hMSH5 has been shown to interact with a Holliday junction binding protein . In addition, hMSH5 forms chromosomal foci in human fetal oocytes at different stages of meiotic prophase I .
Coherent with a conjectured role in recombinational DSB repair, hMSH5 has been reported to interact with several proteins related to DSB sensing and repair, including the c-Abl tyrosine kinase and HR protein hRad51 [10, 11]. It is observed that RAD51 silencing in MSH5-deficient C. elegans oocytes can result in chromosome fragmentation , suggesting that MSH5 and RAD51 may play a synergistic role in DSB processing at least during meiosis in C. elegans. In addition, interaction between endogenous hMSH5 and hMRE11 has been observed in human alveolar basal epithelial derived lung adenocarcinoma A549 cells . Studies performed with mouse models and human patient samples have also suggested a role for hMSH5 in class switch recombination during B and T cell development, whereas hMSH5 deficiency associates with long microhomologies at Ig switch joints . These observations have raised the possibility that, through interacting with various DSB repair proteins, hMSH5 could exert multiple roles in DNA damage surveillance and DSB repair. Although the link between hMSH5 mutation and diseases in humans has not been explored, a genome-wide association study has designated the hMSH5 locus at 6p21.33 as a high risk factor for lung cancer development . In addition to its potential role in DNA repair, hMSH5 interplays with c-Abl in mediating apoptotic response in cells treated with ionizing radiation--a process involved with the activation of p73 and caspase-3 .
In the present study, we have investigated the role of hMSH5 in cisplatin-induced DNA damage response. Cisplatin is the drug of choice for combination chemotherapy of testicular cancers , and hMSH5 is known to be expressed abundantly in the testis . The cytotoxicity of cisplatin is mainly caused by its ability to form adducts with DNA. The major types of biologically active cisplatin adducts are 1,2-intrastrand crosslinks between guanines or between guanine and adenine, and, to a lesser extent, interstrand DNA crosslinks . These DNA distortions can effectively block the progression of DNA replication and activate cell cycle checkpoint [18–20]. Cisplatin-induced DNA intrastrand crosslinks can be effectively removed by nucleotide excision repair (NER) [17, 21]; however, the removal of cisplatin-induced interstrand crosslinks requires the HR pathway . The importance of recombinational repair in resolving cisplatin-induced DNA damage has also been suggested by the observation that cisplatin increases the rate of recombination, presumably attributing to the formation of DSB [23, 24]. In addition, the single-strand breaks generated from the processing of cisplatin-induced DNA lesions by NER can also be converted into one-ended DSBs when single-strand breaks collide with the replication forks . Nevertheless, it becomes increasingly clear that HR plays an essential role in the repair of cisplatin-induced DSBs that may arise from stalled replication forks. In addition, HR-deficient cells are highly sensitive to DNA damaging agents that induce DNA cross-links, due to replication blockade.
Cisplatin leads to hMSH5 induction
c-Abl phosphorylates hMSH5 at Tyr742 in response to cisplatin-induced DNA lesions
Cisplatin stimulates hMSH5 chromatin association
hMSH5 facilitates the repair of cisplatin-induced DNA lesions
The appearance of γ-H2AX foci is an intrinsic biomarker for the extent of DSB formation and can be used to assess the combined effects of DNA damage and DNA lesion repair . It is known that the processing of cisplatin-induced DNA lesions coincides with DSB formation . Therefore, in order to analyze the effects of hMSH5 on the processing of cisplatin-induced DSBs, γ-H2AX foci positive cells were quantified at 24 hrs post-cisplatin treatment. Approximately 39.6% of 293T cells and 35.5% of 293T/f-hMSH5 cells were γ-H2AX foci positive (Figure 5C and Additional file 1: Figure S2). The small difference between these two cell lines indicates that the effects of exogenously over-expressed hMSH5 can be sufficiently masked by cisplatin-triggered induction of endogenous hMSH5 (Figure 5C). However, under identical experimental conditions, expression of hMSH5Y742F or silencing of hMSH5 significantly increased the γ-H2AX positive population to about 62.5% and 69.4%, respectively (Figure 5C), suggesting that hMSH5 deficiency can significantly delay the resolution of cisplatin-induced DSBs.
Since the key HR protein hRad51 coexists with hMSH5 and c-Abl in the same protein complex [10, 11], and Rad51 has been implicated in the repair of cisplatin-induced DSBs , we next investigated the possibility that hMSH5 functions together with hRad51 in the processing of cisplatin-induced DNA lesions. The results of immunostaining experiments demonstrated that, in response to cisplatin, both hMSH5 and hRad51 formed nuclear foci in both 293T and A549 cells (Additional file 1: Figure S3). To determine whether the formation of cisplatin-triggered hMSH5 foci was hRad51-dependent, RNAi-mediated hMSH5 and hRad51 knockdown were performed (Figure 5D). As shown in Figure 5E, the formation of cisplatin-induced hMSH5 foci was significantly diminished with the knockdown of either hMSH5 or hRad51 (left panel). However, the formation of cisplatin-induced hRad51 foci was not affected by the silencing of hMSH5 (Figure 5E, right panel), suggesting that hMSH5 acts downstream of hRad51 in recombinational processing of cisplatin-induced DSBs.
hMSH5 deficiency renders cells more sensitive to cisplatin toxicity
To validate whether cells harboring defective hMSH5 could confer a sensitive phenotype to cisplatin toxicity, MTT assays and clonogenic survival analyses were performed with 293T, 293T/f-hMSH5, and 293T/f-hMSH5Y742F cells treated with cisplatin. Evidently, 293T/f-hMSH5Y742F cells were very sensitive to the toxic effects of cisplatin in both the MTT and clonogenic survival assays, whereas over-expression of hMSH5 appeared to provide a moderate protection to cisplatin toxicity (Figure 6B and 6C). Consistent with the idea that hMSH5 functions together with hRad51 in the process of HR, Rad51-deficient cells have also been shown to exhibit a cisplatin-sensitive phenotype .
The dominant negative effects exerted by the phosphorylation-resistant mutant hMSH5Y742F may be attributed to competition between hMSH5Y742F and the endogenous hMSH5 for binding to partner proteins such as c-Abl, hRad51, hMRE11 and hMSH4. This view is supported by the observed normal interaction of hMSH5Y742F with c-Abl or hMSH4 (Additional file 1: Figure S4). Because c-Abl-mediated hMSH5 tyrosine phosphorylation is known to reduce hMSH5 interaction with c-Abl and hMSH4 , expression of hMSH5Y742F could also disrupt the dynamic composition of its associated protein complex. Together, this study suggests a role for hMSH5 in recombinational repair of cisplatin-induced DSBs.
Our current study has demonstrated an important role for hMSH5 in the processing of cisplatin-elicited S phase-dependent DSBs. Like other HR repair proteins, hMSH5 is predominantly expressed during S and G2/M phases of the cell cycle. Given the observation that silencing of hRad51 can significantly compromise the formation of cisplatin-induced hMSH5 foci, it is conceivable that hMSH5 acts downstream of hRad51 in the DSB repair process. Disruption of hMSH5 function either by RNAi or by the over-expression of a phosphorylation-deficient mutant significantly increased the number of cells displaying γ-H2AX foci at 24 hrs after cisplatin treatment, indicating an increase in the retention of cisplatin-induced DSBs. In addition, disruption of hMSH5 enhances a sustained cisplatin-triggered G2 arrest, thereby rendering cells more sensitive to cisplatin toxicity. It appears that the reduction of clonogenic survivability correlates well with the levels of γ-H2AX foci in 293T, 293T/f-hMSH5, and 293T/f-hMSH5Y742F cells at 24 hrs after cisplatin challenge. This observation is consistent with previous reports showing that the level of γ-H2AX foci retention 24 hrs after treatment is a useful indicator for cisplatin-mediated cell killing [28, 34]. Although the precise role of hMSH5 in this process remains to be delineated, the current evidence collectively supports a scenario by which hMSH5 functions downstream of hRad51 in recombinational repair of cisplatin-induced DSBs.
Due to its radio-sensitizing activity, cisplatin has been frequently used in combination chemoradiation therapy of human malignancies . Presently, however, the mechanisms underlying the effects of this radiosensitizer are still being studied. It is suggested that cisplatin adducts can block the repair of ionizing radiation-induced DSBs by the nonhomologous end-joining pathway [35, 36]. However, the results of our current study have raised another possibility that cisplatin-triggered hMSH5 induction may potentially contribute to the effectiveness of cisplatin combination chemoradiation therapy. It is known that higher levels of hMSH5 promote IR-induced apoptosis . Thus, in spite of the fact that cisplatin-triggered hMSH5 induction can facilitate the repair of cisplatin-induced DSBs, the higher levels of hMSH5 could promote a robust apoptotic response to IR during cisplatin combination chemoradiation therapy.
In addition to hMSH5, other MMR proteins are also involved in cisplatin-triggered DNA damage repair and response. In fact, the roles of several other MMR family members in mediating cellular responses to cisplatin-induced DNA lesions have been studied [37–46]. As a whole, these studies highlight two opposite effects of individual MMR proteins on cellular responses to cisplatin toxicity. By functioning in the recognition and signaling of cisplatin-induced DNA lesions, MMR proteins can promote cell killing. On the other hand, they can directly participate in the processing of cisplatin-induced DSBs, thereby exerting a protective effect. It has been shown that hMSH2 recognizes cisplatin-induced DNA lesions through direct binding of DNA-cisplatin adducts [41, 42]. Cells defective in the expression of hMSH2 or hMLH1 often exhibit a 2- to 4-fold increase in resistance to cisplatin in comparison to corresponding controls [37, 39, 40, 44–46], indicating that hMSH2 and hMLH1 are involved in mediating cisplatin-triggered DNA damage signaling. Accordingly, cisplatin-resistant cells derived from repetitive drug selection are frequently associated with defective hMSH2 or hMLH1 expression [47–49]. However, the effects of hMSH2 or hMLH1 in cellular sensitization to cisplatin have not been observed in a few other studies, reflecting the complex nature of cellular response to cisplatin-induced DNA damage (reviewed in ref. ). It is conceivable that the roles of hMSH2 and hMLH1 may be regulated differently in the processes of repair and DNA damage signaling in different cell types. In addition, difference in cell cycle regulation may be another important factor in controlling various levels of cellular sensitivity to cisplatin. In fact, it is demonstrated recently that disruption of RPA's role in cell cycle regulation synergistically enhances the cytotoxic effects of cisplatin .
In contrast to aforementioned observations, a recent study has revealed an important role for hMSH3 in the repair of cisplatin-induced DSBs . Using an isogenic HCT116-derived cell line in which the expression of hMSH3 can be controlled, Goel and colleagues  demonstrated that hMSH3 deficiency sensitizes cells to both cisplatin and oxaliplatin toxicity, and this effect of hMSH3 is not dependent on the canonical MMR pathway. In addition, in response to oxaliplatin treatment, hMSH3-deficient cells sustain a higher level of γ-H2AX, suggesting that hMSH3 plays an important role in DSB repair . Intriguingly, the role of hMSH5 in mediating cellular response to cisplatin-induced DSBs bears a resemblance to that reported for hMSH3. In spite of using different cell lines and different ways to disrupt gene expression, cells deficient in hMSH5 or hMSH3 show comparable levels of reduction in clonogenic survivability in response to the same doses of cisplatin (Figure 6C) . Although the timing for the elevation of treatment-induced γ-H2AX appears to be different in cells subjected to RNAi-mediated silencing of hMSH5 or hMSH3 (Figure 5C) , these observations warrant future studies to determine whether these two MutS homologues act in the same repair process of cisplatin-induced DSBs.
Intuitively, fertility preservation in male cancer patients undergoing chemotherapy is highly desired, and the relatively high levels of hMSH5 expression in the testis would be expected to provide a protection against cispaltin-induced DSBs. However, this effect of hMSH5 undoubtedly requires coordinated actions from a network of proteins involved in the repair process, and the efficiency of this pathway in various cell types in the testis is presently unknown. Since cisplatin represents a main treatment choice for testicular cancers , it would be interesting to investigate the relative expression levels of hMSH5 in testicular tumors and matched normal testicular tissues. This information will be useful for assessing the value of using hMSH5 as a prognosis biomarker. Finally, our study has implicated that combining hMSH5 disruption with cisplatin treatment might be an alternative strategy for enhancing the therapeutic effects of cisplatin.
In summary, our study has demonstrated a role for hMSH5 in protecting cells from cisplatin-induced DNA damage. Inactivation of hMSH5 by RNAi or by expressing a phosphorylation-deficient hMSH5 mutant elevates cisplatin-induced G2 arrest and renders cells susceptible to cisplatin toxicity.
Collectively, our data is compatible with the idea that hMSH5 is involved in HR repair of cisplatin-induced DSBs.
Cell lines and cell cultures
Stable cell line 293T/f-hMSH5Y742F was generated by a similar procedure that was previously described for 293T/f-hMSH5 . All human cell lines were maintained in DMEM (Invitrogen, Carlsbad, CA) containing 10% FBS (Biomeda, Foster City, CA) and 1x Penicillin-Streptomycin (Invitrogen). Cisplatin (Cis-diamminedichloroplatinum (II) or CDDP) (Sigma, St. Louis, MO) was used for the induction of DNA damage, and imatinib (Novartis, Basel, Switzerland) was used to inhibit c-Abl kinase activity. Cell synchronization at different cell cycle phases was performed by the use of standard hydroxyurea (Sigma), double thymidine (Sigma), and nocodazole (Sigma) based procedures [37, 51]. Cell cycle analysis was performed with the standard propidium iodide (Invitrogen) staining procedure as described previously , in which at least 10,000 cells were analyzed by the use of FlowJo V8 (Tree Star Inc., Ashland, OR). Clonogenic survival and MTT analysis were performed by the same procedures described previously [14, 52].
Antibodies, Western blot analysis, and immunoprecipitation (IP)
Antibodies used for performing IP and Western blot analysis included α-FLAG M2 (Sigma, St. Louis, MO), α-hMSH5 , α-hMSH4 , α-γ-H2AX (Upstate Laboratories, Inc., East Syracuse, NY), α-hRad51 (Calbiochem, Gibbstown, NJ), α-hMRE11 (Novus Biologicals Inc., Littleton, CO), α-c-Abl (BD Pharmingen, San Diego, CA), α-p-Tyr (Cell Signaling, Beverley, MA), α-α-tubulin (Sigma, St. Louis, MO), α-acetyl histone H3 (Upstate), α-histone H4 (Cell Signaling), and α-RNAPII (Upstate). IP and Western blotting were performed as described previously .
Immunofluorescence analysis of nuclear foci
Analysis of hMSH5, hRad51, and γ-H2AX foci formation was carried out by following a standard immunofluorescence protocol. Primary and secondary antibodies used for these experiments included α-hMSH5 , α-hRad51 (Ab-1) pAb (Calbiochem), α-hRad51 (14B4) mAb (Novus Biologicals Inc.), α-γ-H2AX (Upstate), Oregon Green goat α-mouse IgG and Texas Red goat α-rabbit IgG (Invitrogen, Carlsbad, CA). Cells were mounted with Vectashield mounting media containing DAPI (Vector Laboratories, Birminham, CA) and visualized by a Zeiss Axioplan fluorescence microscope. For nuclear foci quantification, cell counting was performed at least three times with a total of > 200 cells counted.
Bacterial Protein expression and purification
Expression of various hMSH5 cp-1 Tyr-to-Phe mutants and c-Abl in BL21(DE3)-RIL cells was performed using the same experimental procedure as described previously . PCR-based site-directed mutagenesis was adapted to generate DNA sequences encoding hMSH5 Tyr-to-Phe mutations. Recombinant hMSH5 cp-1 or hMSH5 cp-1 Tyr-to-Phe proteins were purified under native conditions by the use of TALON Metal Affinity Resins (Clontech, Mountainview, CA) as described previously .
Nuclear and cytoplasmic fractions were prepared by the use of NE-PER Nuclear and Cytoplasmic Extraction Reagents (Pierce, Thermo Fisher Scientific Inc., Rockford, IL) in accordance with the manufacturer's recommendation. Western blots with α-RNAPII (Upstate) and α-α-tubulin (Sigma) antibodies were performed to validate successful fractionation.
Chromatin association assays
Chromatin immunoprecipitation (ChIP) was performed according to the manufacturer's recommendation (EZ-ChIP kit, Millipore, Billerica, MA). Antibodies used in the experiments included α-acetyl histone H3 (Upstate), mouse IgG (Upstate), α-p-Tyr (Cell Signaling), α-hMSH5 , α-hMSH4 , and α-histone H4 (Cell Signaling). To analyze chromatin-associated proteins, bulk chromatin was immunoprecipitated with 10 μg of α-acetyl-histone H3. Immunoprecipitates were then subjected to Western blotting analysis.
The generation of pmH1P-based RNAi constructs was performed as previously described . The efficiency of RNAi-mediated gene silencing was validated by Western blotting, and those displayed greater than 50% knockdown were selected. The RNAi targets were hMSH5 sh-2  and hRad51 sh-1 (5'-AAGGAGAGTGCGGCGCTTC).
School of Molecular Biosciences, Mail Drop 64-7520, College of Veterinary Medicine, Washington State University, Pullman, WA 99164, USA
We thank Dr. Nianxi Zhao and Yen-Lin Chu for technical assistance. Imatinib was kindly provided by Novartis Pharma AG. Basel, Switzerland. This work was supported in part by NIH Grant GM084353 (C.H.).