VAV2 is required for DNA repair and implicated in cancer radiotherapy resistance

Abstract

Radiotherapy remains the mainstay for treatment of various types of human cancer; however, the clinical efficacy is often limited by radioresistance, in which the underlying mechanism is largely unknown. Here, using esophageal squamous cell carcinoma (ESCC) as a model, we demonstrate that guanine nucleotide exchange factor 2 (VAV2), which is overexpressed in most human cancers, plays an important role in primary and secondary radioresistance. We have discovered for the first time that VAV2 is required for the Ku70/Ku80 complex formation and participates in non-homologous end joining repair of DNA damages caused by ionizing radiation. We show that VAV2 overexpression substantially upregulates signal transducer and activator of transcription 1 (STAT1) and the STAT1 inhibitor Fludarabine can significantly promote the sensitivity of radioresistant patient-derived ESCC xenografts in vivo in mice to radiotherapy. These results shed new light on the mechanism of cancer radioresistance, which may be important for improving clinical radiotherapy.

Fig. 1

VAV2 functions as a radioresistant oncogene in ESCC. a Schematic diagram for the constructions of patient-derived xenograft (PDX) of esophageal squamous-cell carcinoma (ESCC), PDX-derived cells (PDC) and irradiation (IR) treatment. b Growth curves of PDXs from 6 individuals with ESCC, showing different sensitivity to IR. Data at each time point are mean ± SD from 3 mice. c Hematoxylin-eosin (H&E) and immunohistochemical (IHC) staining of Ki67 in PDX tumors with or without IR. Scale bars, 100 µm. d Proliferation curves of PDC-2 and PDC-5 cells with or without IR (4 Gy), showing different sensitivity to IR. Data are mean ± SEM from 3 independent experiments and most error bars are with the symbols. **P < 0.01 and ****P < 0.0001 of Student’s t-test. e Heat map showing the differentially expressed genes detected by RNA-sequencing in radiosensitive or radioresistant PDXs. f Venn diagram displaying the overlapped genes among 1,660 apparently overexpressed genes in 3 radioresistant PDXs and 149 amplified and overexpressed genes in 94 clinical ESCC specimens as described previously. g Heat map displaying the inhibitory effects of silencing 8 genes on the malignant phenotypes of two ESCC cell lines. The number presents the inhibitory efficiency and each cell line had 3 independent replications. h H&E and IHC analysis of VAV2 protein level in radiosensitive or radioresistant PDXs. Scale bars, 100 µm. i Box and bar plots comparing the VAV2 protein levels between ESCC and non-tumor tissue samples. P of Mann–Whitney test. j Kaplan–Meier curves of patient survival according to the VAV2 IHC score in tumors. High, IHC score > 6 and low, IHC score ≤ 6. Also present with Kaplan–Meier plot is the hazard ratio (HR) and 95% confidence interval (CI) from multivariate Cox proportional hazard models, including age, sex, tumor stage as covariates

Fig. 2

Irradiation induces ESCC cell radioresistance by evoking aberrant VAV2 overexpression. a, b Forced VAV2 overexpression in KYSE150 cells (a) or radiosensitive PDC-2 cells (b) caused resistance of cells to IR treatment. Left panels show proliferation curves of cells and right panels shows fractions of cell survival by limiting dilution assays. IR, irradiation (4 Gy). c Silencing VAV2 expression by siRNA in radioresistant PDC-5 cells significantly increased sensitivity of cells to IR treatment. Left panels show proliferation curves of cells with siRNA#1 (the results of siRNA#2 are shown in Supplementary Fig. S2g); middle and right panels show fractions of cell survival by limiting dilution assays. IR, irradiation (4 Gy). d, e Western blot analysis showed overexpression of VAV2 and γ-H2AX in ESCC cells treated with radiation (10 Gy), which is time- (d) and dose-dependent (e). f Comparison of VAV2 mRNA (left panel) and protein (right panel) levels in IR-induced radioresistant KYSE150 cells (KYSE150R) and its parental KYSE150 cells. Results are mean ± SEM from three independent determinations and each had three triplicates. P of Mann–Whitney test. g Tumor spheres of KYSE150 and KYSE150R cells treated with or without radiation (4 Gy). Left panel shows representative images of tumor spheres and right panel shows statistics. Scales bars, 100 μm. Data are mean ± SEM from at least three independent experiments and five fields were randomly selected for each experiment. *P < 0.05; **P < 0.01; ***P < 0.001 and ****P < 0.0001 of Student’s t-test

Fig. 3

VAV2 overexpression promotes DNA repair in ESCC cells. a, b Volcano plots displaying upregulated (fold-change > 1.3; P < 0.05) or downregulated genes (fold-change < 0.7; P < 0.05) in KYSE150 cells with VAV2 overexpression (a) or knockout (b). Data are from three experiments. c Volcano plot of gene expression changes in both cell lines with VAV2 overexpression (OE) or knockout (KO). Orange, genes upregulated in cells with VAV2 OE and downregulated in cells with VAV2 KO. Purple, genes downregulated in cells with VAV2 OE and upregulated and in cells with VAV2 KO. Blue, genes upregulated in cells with VAV2 OE or KO. Red, genes downregulated in cells with VAV2 OE or KO. Gray, genes unchanged in cells with VAV2 OE or KO. All P < 0.05 except for genes in gray color. d Gene set enrichment analysis (GSEA) of genes in orange or purple in c showed several pathways including DNA repair-related pathway positively (left panel) or negatively (right panel) correlated with VAV2 expression. e Verification of the expression changes of genes in DNA repair pathway identified by GSEA in (b) in cells with VAV2 OE or KO. The level of mRNA was determined by real-time-quantitative PCR and data are mean ± SEM from three independent determinations. **P < 0.01; ***P < 0.001 and ****P < 0.0001 of Student’s t-test. f–h DNA double-strand breaks expressed by γ-H2AX level in VAV2-overexpressing KYSE150 (f), KYSE450 (g), and KYSE30 (h) cells treated with or without irradiation (IR, 4 Gy). Left panels show γ-H2AX and VAV2 levels by western blot analysis in cells 2 h after IR. Middle panels show images of γ-H2AX foci in cells at various time points of IR as indicated. Scale bars, 20 µm. Right panels represent the statistics. Data are means ± SEM from three experiments. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 and ns, not significant of Student’s t-test. i DNA double-strand breaks detected by comet assays in PDC-5 cells with or without VAV2 silenced by siRNA and treated with or without IR (4 Gy). Left panel shows fluorescence images of comet assays. Scale bars, 100 µm. Right panel shows the statistics. Data are mean ± SEM from three replicates and ten fields were randomly selected for each experiment. ***P < 0.001 of Student’s t-test

Fig. 4

VAV2 is necessary in Ku70 and Ku80-mediated DNA NHEJ repair. a Fourteen top proteins potentially associate with VAV2 identified by mass spectrometry in ESCC cells ectopically overexpressing FLAG-tagged VAV2. Cell lysates were immunoprecipitated with antibody against FLAG and antibody against IgG was used as negative control. b Proteins potentially associate with VAV2 in KYSE30 and KYSE450 cells. Left panel is a Venn diagram showing 68 proteins identified in both cell lines. Right panel shows score of each protein. c–e Western blot detection of VAV2, Ku70 and Ku80 proteins by reciprocal immunoprecipitation with antibody against VAV2 (c), Ku70 (d) or Ku80 (e) in KYSE150 and KYSE450 cells. IgG was used as control. f Immunofluorescence analysis of VAV2 and Ku70 co-staining in KYSE150, KYSE450, PDC-4 and PDC-5 cells with or without IR (4 Gy), showing colocalization of VAV2 and Ku70. DAPI was used to label the nucleus. Scale bars, 20 µm. g Multiple immunofluorescences analysis of VAV2, Ku70, and Ku80 in clinical ESCC tumor tissues. All tumors (T) show strong colocalization signal of the 3 proteins while the corresponding non-tumor tissues (N) from patient 3, which had low VAV2, show very weak colocalization signal. Merge 1 represents the merge of VAV2 and Ku70, Merge 2 represents the merge of VAV2 and Ku80 and Merge 3 represents the merge of VAV2, Ku70, and Ku80. Scale bars, 100 µm. h Western blot analysis of FLAG-VAV2, Ku70 and Ku80 in ESCC cells ectopically expressing FLAG-tagged VAV2 and treated with siVAV2. Cell lysates were immunoprecipitated with antibody against FLAG or IgG. i Western blot analysis of VAV2, Ku70, and Ku80 in ESCC cells overexpressing VAV2. Cell lysates were immunoprecipitated with antibody against Ku70 or IgG. j Effects of increasing VAV2 expression and decreasing Ku70 expression on DNA damage in ESCC cell lines KYSE150 (upper panels) and KYSE450 (lower panels). Left panels show Ku70, Ku80, VAV2 and γ-H2AX levels detected by Western blot analysis in cells with different treatment. Middle panels are image of comet assays showing overexpression of VAV2 significantly diminished DNA damage, which could be reversed by knockdown of Ku70 expression. Right panel are the statistics of tail moments in comet assays. OE, overexpression. Data are mean ± SEM from 3 replicate experiments and 10 fields were randomly selected for each experiment. **P < 0.01; ****P < 0.0001 and ns, not significant of Student’s t-test

Fig. 5

VAV2 overexpression excessively activates STAT1 signaling. a Scatter diagram shows the overlapped proteins in KYSE150 cells with VAV2 overexpression (OE) or knockout (KO). Orange, proteins upregulated in cells with VAV2 OE and downregulated in cells with VAV2 KO; Blue, proteins upregulated in cells with VAV2 OE or KO; Purple, proteins downregulated in cells VAV2 OE and upregulated in cells with VAV2 KO and Red, proteins downregulated in cells with VAV2 OE or KO. P < 0.05 of Student’s t-test compared with that in cells without VAV2 expression change (Control). The level changes of proteins in gray were not statistically significant between treated and control cells (P > 0.05). b Western blot analysis shows upregulation of STAT1, ATF7 and GTF2E1 in KYSE450 cells with VAV2 OE. p-STAT1, phosph-STAT1. c Western blot analysis of total STAT1, phosph-STAT1 (p-STAT1) and γ-H2AX levels in KYSE30 and KYSE150 cells with VAV2 OE or KO. d Comparison of VAV2, STAT1, p-STAT1, and γ-H2AX levels in KYSE150, KYSE450, irritation (IR)-induced radioresistant KYSE450R and KYSE450R cells. e Spearman correlation of STAT1 and VAV2 mRNA levels in clinical ESCC samples. f–i STAT1 inhibitor Fludarabine (Flud) significantly enhanced the radiosensitivity of radioresistant ESCC cells to IR in vitro. Shown are inhibitory effects of IR (2 Gy), Fludarabine (0.1 and 0.05 μM for cell growth or colony formation assays, respectively) or combination of IR and Fludarabine on cell growth detected by CCK-8 assays (f, h) and colony formation (g, i) of PDC-5 and KYSE150R. Data are mean ± SEM from 3 experiments and each had 3 replications. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 and ns, not significant of Student’s t-test. j STAT1 inhibitor Fludarabine (40 mg/kg) significantly enhanced the radiosensitivity of mouse xenografts derived from radioresistant PDC-5 to IR (10 Gy). Shown are curves of tumor growth overtime and tumors from each mouse in different groups are shown in Supplementary Fig. S5m. Data are mean ± SD from five mice. *P < 0.05; **P < 0.01 and ns, not significant of Student’s t-test. See methods for Fludarabine and IR treatment

Fig. 6

VAV2 predicts radiotherapy vulnerability of ESCC. a Chemoradiotherapy efficacy in patients with ESCC (n = 31) as function of the VAV2 level in tumor samples detected by immunohistochemical (IHC) staining, showing ESCC having IHC score >6 were significantly more resistant to the therapy compared with ESCC having the score ≤ 6. b–d Correlation between the VAV2 and γ-H2AX levels in ESCC tumors before and after chemoradiotherapy. Images show computed tomography of tumors before and after treatment (left) and H&E, VAV2 and γ-H2AX IHC staining in tumors before (middle) and after (right) treatment in ESCC sensitive (b) or resistant (c) to chemoradiotherapy. Scale bars, 100 µm. d Shown are the Spearman correlations between VAV2 and γ-H2AX levels in tumors before and after treatment. e Kaplan–Meier curves of patient survival according to VAV2 level in ESCC tumor. VAV2 high, IHC score > 6; VAV2 low, IHC score ≤ 6. Also present with the Kaplan–Meier curves is the hazard ratio (HR) and 95% confidence interval (CI) from multivariate Cox proportional hazard models, including age, sex, tumor stage as covariates. Four of 31 patients lost follow-up and thus were excluded in the analysis

Fig. 7

The schematic illustration for the possible mechanisms of VAV2-mediated radioresistance of ESCC cells and Fludarabine as a radiosensitizer. Overexpressed VAV2 in cancer cells increases VAV2-Ku70/Ku80 complex formation and activates signal transducer and activator of transcription 1 (STAT1) signaling, which enhances the ability of cells to repair DNA damage caused by ionizing radiation (IR) and thus promotes cancer growth. However, Fludarabine can inhibit STAT1 activity and reverses the vulnerability of VAV2-overexpressing cancer cells to IR. On the other hand, cancer cells without VAV2 overexpression are relatively sensitive to IR