MET inhibition in tumor cells by PHA665752 impairs homologous recombination repair of DNA double strand breaks

Authors

  • Michaela Medová,

    1. Department of Radiation Oncology, Inselspital, University of Berne, Berne, Switzerland
    2. Department of Clinical Research, Radiation Oncology, University of Berne, Berne, Switzerland
    Search for more papers by this author
  • Daniel M Aebersold,

    1. Department of Radiation Oncology, Inselspital, University of Berne, Berne, Switzerland
    2. Department of Clinical Research, Radiation Oncology, University of Berne, Berne, Switzerland
    Search for more papers by this author
  • Yitzhak Zimmer

    Corresponding author
    1. Department of Radiation Oncology, Inselspital, University of Berne, Berne, Switzerland
    2. Department of Clinical Research, Radiation Oncology, University of Berne, Berne, Switzerland
    • University of Berne, Department of Clinical Research, MEM-E807, Murtenstrasse 35, 3010 Switzerland
    Search for more papers by this author
    • Tel.: +41-31-632-2642, Fax: +41-31-632-3297


Abstract

Abnormal activation of cellular DNA repair pathways by deregulated signaling of receptor tyrosine kinase systems has broad implications for both cancer biology and treatment. Recent studies suggest a potential link between DNA repair and aberrant activation of the hepatocyte growth factor receptor Mesenchymal-Epithelial Transition (MET), an oncogene that is overexpressed in numerous types of human tumors and considered a prime target in clinical oncology. Using the homologous recombination (HR) direct-repeat direct-repeat green fluorescent protein ((DR)-GFP) system, we show that MET inhibition in tumor cells with deregulated MET activity by the small molecule PHA665752 significantly impairs in a dose-dependent manner HR. Using cells that express MET-mutated variants that respond differentially to PHA665752, we confirm that the observed HR inhibition is indeed MET-dependent. Furthermore, our data also suggest that decline in HR-dependent DNA repair activity is not a secondary effect due to cell cycle alterations caused by PHA665752. Mechanistically, we show that MET inhibition affects the formation of the RAD51-BRCA2 complex, which is crucial for error-free HR repair of double strand DNA lesions, presumably via downregulation and impaired translocation of RAD51 into the nucleus. Taken together, these findings assist to further support the role of MET in the cellular DNA damage response and highlight the potential future benefit of MET inhibitors for the sensitization of tumor cells to DNA damaging agents.

Introduction

Signaling crosstalk between receptor tyrosine kinase (RTK) systems and cellular DNA repair pathways is an emerging concept with broad implications for cancer biology and treatment. The prototype system in this category is that of the epidermal growth factor receptor (EGFR) that has been associated with tumor cell resistance to DNA damaging agents (DDAs)1, 2 and whose inhibition by the monoclonal antibody cetuximab both in preclinical and clinical setups demonstrated promising sensitizing capacity concerning tumor growth control when combined with ionizing radiation (IR).3 Although the mechanisms that underlie the molecular crosstalk between the EGFR and DNA repair remain yet largely unknown, Li et al. have reported in 2008 that blocking EGFR by the small-molecule erlotinib targets DNA double-strand breaks (DSBs) repair by homologous recombination (HR),4 providing, therefore, important evidence that RTK activity may influence cellular response to damage inflicted by therapeutic modalities that act primarily by exerting DNA damage. This notion is further supported by the observations by Choudhury et al., who reported that the small molecule Gleevec, an inhibitor of the breakpoint cluster region-V-abl Abelson murine leukemia viral oncogene homolog 1 (BCR-ABL), KIT, and platelet-derived growth factor receptor (PDFGR) kinases, enhances tumor sensitivity to DDAs by reducing HR efficiency.5

The MET RTK system regulates important biological processes, including cell scattering, invasion, and survival as well as epithelial remodeling and angiogenesis.6 MET is genetically altered through several mechanisms in multiple human cancers, and these aberrations are closely associated with cancer initiation and progression, identifying MET as an important therapeutic target.7 Recent evidence highlights additional roles for MET in cancer through a potential crosstalk with the DNA damage response (DDR) machinery. In that respect, we have recently reported a signaling axis, which consists of MET and the HR effectors V-abl Abelson murine leukemia viral oncogene homolog 1 (ABL) and RAD51.8 Treatment of cells expressing MET-mutated variants with the MET small molecule SU11274 diminished the signaling status of ABL and RAD51 and resulted in high and sustained levels of DSBs when combined with IR. To investigate further the potential link between MET and the DDR, we aimed in the current work to investigate the impact of MET inhibition on functional HR in the MET-overexpressing tumor cell lines GTL-16 and H1993. Our data demonstrate significant reduction of HR efficiency upon MET inhibition by the second-generation inhibitor PHA665752. Using MET mutants that respond differentially to PHA665752, we show a MET-dependent HR inhibition, which is not secondary to cell cycle changes. Mechanistically, the results demonstrate that PHA665752 considerably reduces IR-dependent increase of RAD51 nuclear levels and its physical association with BRCA2, two molecular events closely associated with HR repair of DSBs.

Material and Methods

Cells

Human gastric adenocarcinoma cells GTL-16 from Dr. Paolo Comoglio (Medical School University of Torino, Italy) and human lung carcinoma cells H1993 from Dr. Sunny Zachariah (Southwestern University, TX) were both grown in Roswell Park Memorial Institute medium (RPMI) medium (GIBCO, Invitrogen) supplemented with 5% fetal calf serum (FCS) (Sigma) and antibiotic-antimycotic (penicillin 100 U/ml, streptomycin sulfate 100 U/ml, amphotericin B as Fungizone 0,25 μg/ml; GIBCO). NIH3T3 cell lines that stably express the MET-mutated variants M1268T and Y1248H were provided by Dr. Laura Schmidt (National Cancer Institute, Frederick, MD) and maintained in Dulbecco's modified Eagle's medium (DMEM) (GIBCO) with 10% FCS, antibiotic-antimycotic and 0.5 mg/ml Geneticin/G-418 sulfate (GIBCO).

PHA665752 and siRNA transfection

PHA665752, (3Z)-5-[(2,6-dichlorobenzyl)sulfonyl]-3-[(3,5-dimethyl-4-{[(2R)-2-(pyrrolidin-1-ylmethyl)pyrrolidin-1-yl]carbonyl}-1H-pyrrol-2-yl)methylene]-1,3-dihydro-2H-indol-2-one (Pfizer, La Jolla, CA) was dissolved in DMSO, and working concentrations were prepared in corresponding medium.

To transfect nontargeting or anti-MET siRNAs (ON-TARGETplus siRNAs from Dharmacon, Thermo Fisher Scientific, Lafayette, Colorado), Lipofectamine™2000 (Invitrogen) was used according to manufacturer's instructions.

DR-GFP assay

HR repair of DSBs was monitored by a unique cellular method described previously.9–11 Briefly, cells stably expressing the direct-repeat (DR)-GFP construct (provided by R. Bristow, Canada) were plated into 12-well plates, 1 day later pretreated by MET inhibitor PHA665752 or siRNA transfected and subsequently transiently transfected with 5 μg of pGFP (transfection efficiency control), pCMV3xnlsI-SceI (functional endonuclease) or phCMV-1 I-SceI (carrying a mutated I-SceI sequence as a negative control) vector, using Lipofectamine™2000. Transient expression of I-SceI endonuclease generated a DNA DSB at the integrated GFP gene sequences and stimulated HR. GFP signal was assayed 72-hr post-transfection on a FACScan flow cytometer (BD Biosciences). For each experiment, 50,000 cells were scored per treatment group, and the frequency of recombination events was calculated from the number of GFP-positive cells divided by the number of cells analyzed after correction for transfection efficiency using the formula for recombination frequency (RF):

equation image

Cell cycle analysis

PHA665752-treated cells were trypsinized, washed with ice cold PBS, fixed in 70% EtOH and stored at 4°C. Before measurements, cells were washed with PBS, allowed to rehydrate for 1 min and incubated with RNaseA/propidium iodide (PI) cocktail (20 μg/ml PI + 40 μg/ml RNase A + 0.1% TRITON) for 30 min. PI fluorescence (20,000 cells per sample) was measured using FACScan. The single-cell-gated populations were analyzed using FlowJo to obtain the final cell-cycle distribution profiles.

Nuclear and cytoplasmic fraction separation

Cells treated with PHA665752 and irradiated with 10 Gy (Gammacell research irradiator GC40, MDS Nordion) were lysed 24 hr postirradiation and nuclear fractions were obtained using the Nuclear Extraction kit (Active Motif). Subsequently, Western blotting was performed as described previously.12

Coimmunoprecipitation and antibodies

Total protein lysates (0.5 mg) were incubated with 1 μg of specific anti-BRCA2 antibody (Cell Signaling Technology) and subsequently, μMACS protein G Microbeads (Miltenyi Biotec, Germany) were added. After calibration, columns were loaded with samples and washed with high- and low-salt buffers. Immunoprecipitated complexes were analyzed by Western blotting and probed for the presence of RAD51 using mouse monoclonal anti-RAD51 antibody (Upstate/Millipore).

MET phosphorylation (pMET) has been detected using an anti-p-MET(Tyr1234/1235) rabbit polyclonal antibody (Cell Signaling, Danvers, MA), mouse monoclonal anti-Actin was obtained from Sigma and rabbit polyclonal anti-SP-1 from Santa Cruz Biotechnology.

Statistics

Significance was calculated using Student's t test. A p value <0.05 was considered statistically significant.

Results and Discussion

When combined with IR, MET inhibition by the small molecule SU11274 leads to a delay of γH2AX resolution, suggesting a hindrance of DNA DSBs repair.8 As this observation is accompanied by a concomitant reduction of radiation-dependent tyrosine phosphorylation of RAD51 and ABL, two important HR effectors, a major goal of the current study was to investigate whether MET inhibition compromises HR efficiency.

To that end, we used GTL-16 human gastric adenocarcinoma and H1993 human lung carcinoma cells, which overexpress MET. GTL-16 and H1993 DR-GFP cells that carry a chromosomally integrated HR substrate were established similarly as previously described.13 Specific DSBs were introduced in GTL-16 and H1993 DR-GFP cells by transient expression of the I-SceI endonuclease. Using this system, restored GFP function may take place only via HR between the two mutated GFP repeats of the DR-GFP construct.

To test the impact of MET inhibition on restoration of GFP activity, RFs were calculated for GTL-16 and H1993 DR-GFP cells that were exposed to increasing doses of the small molecule PHA665752, whose IC50 toward MET is considerably lower than that of SU11274.14 As shown in Figure 1a, a consistent dose-dependent decrease in RF values following MET inhibition is detected starting already with 10 nM (10% reduction) for the GTL-16 line and reaches a nearly maximal HR inhibition with 40 nM of the drug (96% reduction). HR in H1993 cells was even more sensitive to the effect of the MET inhibitor, as a drop of 56% in GFP reconstitution was detected already with a low concentration of 10 nM of the small molecule.

Figure 1.

Effect of MET inhibition on HR efficiency in MET-overexpressing cells. (a) Left: Upper panel: Frequency of HR in DR-GFP GTL-16 cells pre-treated with PHA665752 (2 hr; 0–300 nM) assayed by DR-GFP assay (described in Material and Methods). The data represent the mean ± standard deviation (SD) of five independent experiments, Student's t-test (vs. untreated cells): p(25 nM) <0.0005 and p(40, 100, 300 nM) <0.5 × 10−6. Right: Frequency of HR in DR-GFP H1993 cells pretreated with PHA665752 (2 hr; 0–300 nM) assayed by DR-GFP assay. The data represent the mean ± SD of three independent experiments, Student's t-test (vs. untreated cells): p(10, 25 nM) = 0.05, p(40 nM) <0.05, p(100, 300 nM) <0.005. Lower panels: Representative diagrams of GTL-16 and H1993 cells expressing reconstituted GFP protein (population located left-up) and cells negative for GFP (main population). Cells were treated by 0, 25 or 300 nM PHA665752 and subsequently transiently transfected with pCMV3xnlsI-SceI plasmid. (b) Frequency of HR in DR-GFP GTL-16 cells (left) and in H1993 cells (right) transfected with nontargeting siRNA (siControl) or MET siRNA (siMET) assayed by DR-GFP assay. The data represent the mean ± SD of three experiments, p < 0.05 (GTL-16) and p < 0.005 (H1993). Representative Western blots are showing MET downregulation by siRNA approach, β-actin serves as a loading control.

To verify if targeting MET by another approach apart of kinase inhibition would also affect HR efficiency, we have transfected GTL-16 and H1993 cells with a pool of anti-MET siRNAs and analyzed the impact of MET knockdown on GFP reconstitution. As shown in Figure 1b, anti-MET siRNAs reduced HR efficiency in both cell lines by approximately 50% when compared with nontargeting siRNAs, which served as negative controls.

Furthermore, to exclude a possible off-target MET-independent activity of PHA665752 on HR, we next evaluated the RF values for NIH3T3 cells that ectopically express two distinct MET-mutated variants, M1268T and Y1248H, which display a respective sensitivity or resistance toward SU11274.15 Since the response pattern of these MET variants to PHA665752 has not been previously reported, we first determined pMET status following exposure to the inhibitor. To that end, we have used a 25-nM concentration of PHA665752, which was sufficient to reduce HR efficiency in GTL-16 by nearly 60% (Fig. 1). As Figure 2a shows, PHA665752 reduced the constitutive MET phosphorylation (pMET) of the M1268T variant, without affecting the activation status of Y1248H. A quantitative assessment of three independent experiments confirms a 30% average reduction of pMET in M1268T with no impact on Y1248H (Fig. 2b). Furthermore, the results in Figure 2c suggest that the PHA665752-sensitive variant M1268T displayed a dramatic 80% decrease in HR efficiency when treated by 25 nM of PHA665752 when compared with untreated cells. In contrary, Y1248H cells show a minor increase of HR-dependent repair when compared with untreated cells (Fig. 2d). However, this observation in PHA665752-treated Y1248H cells is not statistically significant.

Figure 2.

Impact of PHA665752 on MET phosphorylation status and HR efficiency in M1268T and Y1248H MET-mutated variants. (a) pMET levels in untreated (C) or PHA665752-treated (P; 25 nM for 48 hrs) NIH3T3 MET-mutated variants M1268T and Y1248H. β-Actin was used as a loading control. (b) Quantification of pMET status from three independent Western blot experiments described in (A). Bands were quantified using QuantityOne software (BioRad) and normalized to β-Actin levels. (c and d) Left: RFs in DR-GFP NIH3T3 cells expressing M1268T (C) and Y1248H (D) MET mutations (untreated or treated by 25 nM of PHA665752 for 16 hr). The data represent the mean ± SD of five independent experiments, P(M1268T; 25 nM) = 0,018, p(Y1248H; 25 nM) >0.15. Right: Representative diagrams of M1268T and Y1248H cells treated by 0 or 25 nM PHA665752 and transiently transfected with pCMV3xnlsI-SceI.

HR is one of two major distinct mechanisms in which eukaryotic cells use to repair DSBs and whose utilization is not evenly distributed throughout the cell cycle.16, 17 Thus, HR occurs only when a homologous template, e.g., sister chromatid, is available, and HR-related proteins are expressed. Consequently, HR activity is taking place predominantly during S and G2 phases of the cell cycle. As MET inhibition by PHA665752 results under certain conditions in a G1 arrest,18 we next questioned if the reduced HR efficiency following PHA665752 administration may be associated with an indirect effect such as MET inhibition-dependent cell cycle alterations and in particular, a significant G1 arrest.

Although the data in Figure 3a suggest a drop in percentage of S-phase GTL-16 cells treated with 100 and 300 nM of PHA665752 compared to controls (9.2–1.7% and 0.6%, respectively) with a concomitant increase in the sub-G1 apoptotic fraction (both 11.2% when compared with 3.9% in controls), cell cycle distribution of untreated versus PHA665752-treated GTL-16 cells at the time-point of GFP measurement did not substantially differ at 10–40 nM, which are the most relevant concentrations for the dose-dependent HR decrease. Similar data are seen also for the H1993 cell line. Thus, the observed attenuation of HR by MET inhibition could be primarily attributed to a direct impairment of the HR repair machinery. Importantly, these data are reminiscent of findings that were recently described by Li et al. for the EGFR system, which showed attenuation of HR by the small molecule erlotinib that are as well independent of cell cycle alterations,4 emphasizing, therefore, the increasing evidence for the involvement of RTK systems, which are strongly involved in cancer biology, in DDR processes.

Figure 3.

Cell cycle distribution and RAD51 and BRCA2 proteins status upon MET inhibition by PHA665752. (a) Cell cycle distribution of GTL-16 (left) and H1993 (right) cells measured 72 hr after the start of treatment with PHA665752. Columns—means of three independent experiments, bars—SD (±). (b and c) Subcellular localization and the interaction of RAD51 and BRCA2 proteins in GTL-16 (B, left) and H1993 (B, right) cells and (C) in NIH3T3 cells stably expressing MET mutated variants M1268T and Y1248H post PHA665752 treatment and irradiation. Upper panels: Nuclear fractions of cells treated with PHA665752 (300 nM; 16 hr) and subsequently irradiated by 10 Gy. The expression of SP-1 was used as a loading control. Lower panel: RAD51-BRCA2 interactions as detected by coimmunoprecipitation experiments. The results shown are representative data of at least three independent experiments.

Changes in the RAD51 recombinase levels, phosphorylation, cellular distribution and interactions with other effectors are all events, which are associated with DNA damage-induced HR.19, 20 In the search for potential HR-related molecular mechanisms that may be affected by the MET inhibitor, we focused, therefore, on the impact of PHA665752 on the liaison between RAD51 and the scaffold protein BRCA2, which plays a mandatory role in HR.21 A crucial step in this process is the formation of the RAD51-BRCA2 complex, which is required for accurate RAD51-mediated search for the homology and strand pairing stages.22 The data from this part of the study are summarized in Figures 3b and 3c. Although MET inhibition did not affect the expression and the nuclear localization of BRCA2 in control as well as in irradiated GTL-16 and H1993 cells, the induction of RAD51 expression and its nuclear translocation upon irradiation were significantly impaired when cells were treated by PHA665752 before irradiation (Fig. 3b). Moreover, even basal RAD51 levels in cells unchallenged by IR seem to be considerably diminishing by PHA665752 in both cell lines. Using coimmunoprecipitation, we next showed that MET inhibition disrupted the RAD51-BRCA2 interaction observed in irradiated cells, suggesting at least one potential mechanism responsible for the compromised HR in cells with MET inhibition. Importantly, in MET-mutated NIH3T3 cells, PHA665752 affected RAD51 nuclear translocation and the formation of RAD51-BRCA2 complex only in the drug-sensitive M1268T cell line, whereas no such effects were observed in the isogenic Y1248H variant (Fig. 3c), supporting, therefore, again that the decrease in HR via RAD51 nuclear depletion is mediated by blocking of MET.

In this study, we show that MET inhibition in tumor cells decreases DNA repair via HR. The data assist to further support a crosstalk between MET and HR-dependent DSBs repair that was previously reported, which demonstrated that MET targeting impedes the repair of IR-induced DSBs as monitored by persistent high cellular levels of γH2AX.8 Moreover, the observation that two different MET small molecules, SU11274 and PHA665752, can compromise HR effectors status and efficiency in a mutant-dependent manner substantiates an emerging MET-DDR link. Importantly, this study also highlights that at the low-concentrations range of the MET inhibitor used in this study, the mode of MET-dependent decrease in response to DNA damage is primarily through an impact on the DNA repair arm of the DDR rather than on its checkpoint branch, a conclusion, which emerges due to lack of major alterations in cell cycle phases relevant to HR. In summary, our results indicate that MET inhibition in tumor cells with deregulated activity strongly interferes with the ability of the cells to repair DNA damage via HR effectively, most probably by direct effect of MET signaling on HR proteins' functions and interactions. These findings support the role of MET in the cellular DDR, and the rationale for this RTK targeting for tumor cells sensitization to DNA-damaging agents. Finally, our work joins and supports recent studies, which showed that small molecules such as erlotinib and imatinib have the ability to compromise cellular response to DSBs by affecting the HR machinery.4, 5 These observations may be further exploited for combination protocols for the increase of tumor cytotoxicity to DDAs such as IR or other radiomimetic agents.23, 24

Acknowledgements

The authors thank Robert Bristow and Simon Powell for the HR DR-GFP system and Bruno Streit for technical assistance. This study was supported by the Oncosuisse grant (OCS 01681-02-2005) to D.M.A. and Y.Z. and by the Swiss National Science Foundation grant (SNF 31003A-125394) to Y.Z.

Ancillary