FGFR3-induced Y158 PARP1 phosphorylation promotes PARP inhibitor resistance via BRG1/MRE11-mediated DNA repair in breast cancer models

Abstract

Poly(ADP-ribose) polymerase (PARP) inhibitors (PARPis) are used to treat BRCA-mutated (BRCAm) cancer patients; however, resistance has been observed. Therefore, biomarkers to indicate PARPi resistance and combination therapy to overcome that are urgently needed. We identified a high prevalence of activated FGF receptor 3 (FGFR3) in BRCAm triple-negative breast cancer (TNBC) cells with intrinsic and acquired PARPi resistance. FGFR3 phosphorylated PARP1 at tyrosine 158 (Y158) to recruit BRG1 and prolong chromatin-loaded MRE11, thus promoting homologous recombination (HR) to enhance PARPi resistance. FGFR inhibition prolonged PARP trapping and synergized with PARPi in vitro and in vivo. High-level PARP1 Y158 phosphorylation (p-Y158) positively correlated with PARPi resistance in TNBC patient-derived xenograft models, and in PARPi-resistant TNBC patient tumors. These findings reveal that PARP1 p-Y158 facilitates BRG1-mediated HR to resolve the PARP-DNA complex, and PARP1 p-Y158 may indicate PARPi resistance that can be relieved by combining FGFR inhibitors (FGFRis) with PARPis. In summary, we show that FGFRi restores PARP trapping and PARPi antitumor efficacy in PARPi-resistant breast cancer by decreasing HR through the PARP1 p-Y158/BRG1/MER11 axis, suggesting that PARP1 p-Y158 is a biomarker for PARPi resistance that can be overcome by combining FGFRis with PARPis.

Keywords: Breast cancer; Drug therapy; Oncology; Protein kinases; Therapeutics.

Figure 1. FGFR3 is activated in talazoparib-resistant cells.

(A) Colony formation of SUM149 parental and BR cells in response to talazoparib. (B) Half-maximal inhibitory concentration (IC₅₀) of TNBC cells in response to talazoparib. Cells were treated with talazoparib for 4 days before cell survival was analyzed by MTT assay. IC₅₀ was calculated using GraphPad Prism 8.0. Histogram shows the mean ± SEM. (Biological repeats: SUM149 n = 5, HCC1806BR n = 4, all other cell lines n = 3.) (C) Talazoparib and olaparib IC₅₀ of SUM149-BR cells according to MTT assay. Fold change (×) of IC₅₀ was compared with that of SUM149 parental cells (SUM). Histogram shows the mean ± SEM (n ≥3). The purple bars represent the cells used in the antibody array analysis in D and E, while the white bars represent the others. (D and E) Antibody arrays of RTK activation in SUM149 parental and SUM149-BR cells. Cells were treated with DMSO or 100 nM talazoparib overnight and harvested for RTK antibody array analysis. (D) The images of RTK antibody arrays in SUM149 parental, BR#09, and BR#17. (E) The signal intensities from all the arrays are shown as heatmaps.

Figure 2. Synergy between PARPis and FGFRis is independent of BRCA1 expression.

(A) BR#09 and BR#17 cells were treated with talazoparib (Tala) and PD173074 (PD), either alone or in combination (Combo), at the concentrations indicated for 10–12 days, and then cells were fixed for the colony formation assay. The number of colonies formed was normalized to that in the control group (not treated with talazoparib and PD173074), and the mean ± SD from 3 independent experiments is shown in the histogram. *P < 0.05, **P < 0.01, and ***P < 0.001. ANOVA was used for statistical comparisons. Representative images of colony formation are shown in Supplemental Figure 4, C and D. (B and C) Combination index (CI) of the talazoparib and PD173074 combination or the olaparib and AZD4547 combination in SUM149-BR (B), BT549 (C), and MDA-MB-157 (C) cells. Cells were treated with various concentrations of talazoparib and PD173074 or olaparib and AZD4547 for 4 days before cell survival was measured by MTT assay and the CI was calculated by CompuSyn. Fa, fraction affected. (D) Immunofluorescence of SUM149 parental, BR#09, and BR#17 cells staining for DAPI (DNA), RAD51 foci (homologous repair), and γ-H2AX foci (double-strand breaks) after 24 hours of 50 nM talazoparib treatment. Scale bars: 20 μm. (E) BRCA1 was knocked down with 2 different shRNAs (shBRCA1-1 and shBRCA1-3) in BR#09 and BR#17 cells. Moreover, BRCA1 was re-expressed in BR#09 and BR#17 shBRCA1-3 cells (WT-BRCA1). These cells, including the control cells (LKO.1), were treated with various concentrations of talazoparib and PD173074 combination, and the CI values were determined. The expression of BRCA1 was determined by Western blot, and the results are shown in Supplemental Figure 4, G and H.

Figure 3. Combination of talazoparib and PD173074 attenuates DNA repair.

(A–C) SUM149 parental (A), BR#09 (B), and BR#17 (C) cells were treated with 5 μM FGFRi (PD173074, JNJ-42756493, AZD4547) for 4 hours, then further exposed for 1 hour to 100 nM talazoparib (Tala) and 0.01% MMS along with the indicated FGFRis before Western blot analysis. (D and E) BR#17 cells were treated with MMS and the indicated inhibitors for 1 hour, followed by inhibitor treatment after MMS removal. Immunofluorescence images (D) display γH2AX (green) and DNA (blue). Scale bars: 20 μm. Scatterplot (E) shows mean ± SD from 3 independent experiments; scatterplot represents all counted cells. One-way ANOVA with Tukey’s test: *P < 0.05 and **P < 0.001. (F) The indicated cells were treated with 100 nM talazoparib and 0.01% MMS for 1 hour (+MMS), then recovered in fresh medium for 3 hours before comet assay. Scatterplot displays the mean ± SD from 3 experiments; scatterplot includes all cells counted. One-way ANOVA with Tukey’s test: *P < 0.05 and **P < 0.01. Representative comet assay images are shown in Supplemental Figure 5B. (G) BR#09 cells received 0.01% MMS, 100 nM talazoparib, and/or 5 μM PD173074 (alone or combined) for 1 hour before alkaline comet assay. DNA damage (olive moment) was normalized to the talazoparib-treated group. Scatterplot shows mean ± SD from 3 experiments; scatterplot represents all counted cells. One-way ANOVA with Tukey’s test: ***P < 0.001. Representative comet assay images are shown in Supplemental Figure 5C.

Figure 4. FGFR3 phosphorylates PARP1 at Y158, promoting PARPi resistance.

(A) BR#09 and BR#17 cells treated with 0.01% MMS and indicated inhibitors (Tala, 100 nM talazoparib; PD, 10 μM PD173074; Combo, Tala+PD) were subjected to PLA using FGFR3 and PARP1 antibodies. Nuclear PLA signals were quantified as mean ± SD from 3 independent experiments; scatterplots show all cells counted. Dunnett’s test: *P < 0.05 and **P < 0.01. Representative images are shown in Supplemental Figure 6, B and C. (B and C) BR#09 (B) and BR#17 (C) cells with endogenous PARP1 knockdown (shPARP1) were rescued by exogenous expression of PARP1^WT, PARP1^Y158D, or PARP1^Y158F. PARP1 expression was validated by Western blot (left). Cell survival with talazoparib treatment was assessed by MTT assay (right); mean ± SD from at least 3 independent experiments. (D) BR#09 and BR#17 cells expressing PARP1^WT, PARP1^Y158D, or PARP1^Y158F were treated with MMS (0.01%) and talazoparib (200 nM) for 30 minutes, followed by incubation with 100 nM talazoparib after MMS removal for indicated durations. γH2AX foci were quantified by immunofluorescence (BlobFinder). Scatterplots represent mean ± SD from 3 independent experiments; scatterplots show all cells counted. One-way ANOVA with Tukey’s test was performed to compare time points within each mutant cell line: *P < 0.05; **P < 0.01; ***P < 0.001. Representative images are shown in Supplemental Figure 7C. (E) BR#09 and BR#17 cells expressing PARP1^WT, PARP1^Y158D, or PARP1^Y158F were treated with talazoparib and PD173074 (constant ratio) for 6 days. Cell survival was determined by MTT assay, and combination index (CI) was calculated using CompuSyn. Fa, fraction affected. Results (mean from at least 3 experiments) are shown for BR#09 (left) and BR#17 (right). (F) BR#17 cells treated with 0.01% MMS plus 100 nM talazoparib (Tala), 10 μM PD173074 (FGFRi), or their combination were subjected to immunoprecipitation with anti–p-PARP (p-Y158) antibody, followed by Western blotting.

Figure 5. Inhibition of FGFR-mediated PARP1 Y158 phosphorylation prolongs PARP trapping.

(A and B) BR#17 cells expressing PARP1^WT, PARP1^Y158D, or PARP1^Y158F were harvested at different time points after 40 minutes of treatment with 100 nM talazoparib and 0.01% MMS, and subjected to chromatin fractionation, followed by Western blot (A). Chromatin-bound PARP1 signal intensities were normalized to histone H3 and compared with that of the cells at the beginning of DNA repair (0 minutes after releasing from talazoparib and MMS) (B). Means ± SD from 5 individual repeats are shown in the histograms. One-way ANOVA was used for statistical comparisons: *P < 0.05. (C–F) BR#09 and BR#17 cells were pretreated with 10 μM PD173074 for 2 hours, followed by a 40-minute incubation with either PD173074 (10 μM) plus talazoparib (100 nM) plus MMS (0.01%), or PD173074 (10 μM) plus talazoparib (100 nM). Also, these cells were treated with only talazoparib (100 nM) plus MMS (0.01%) for 40 minutes. After drug removal, these cells were harvested at 0, 30, or 60 minutes and subjected to chromatin fractionation. The chromatin-bound PARP1 levels in BR#09 (C) and BR#17 (E) were then determined by Western blot analysis. PARP1 signal intensities of BR#09 (D) and BR#17 (F) cells were normalized to histone H3 and compared with that of cells treated with talazoparib and MMS (MMS +, PARPi +, FGFRi +, 0 minutes). Mean ± SD from 4 individual repeats are shown in the histogram. One-way ANOVA with Tukey’s test was used for statistical comparisons: *P < 0.05 and **P < 0.01.

Figure 6. PARP1 Y158 phosphorylation enhances the recruitment of BRG1 and MRE11 on chromatin.

(A) BR#09 cells expressing PARP1^WT, PARP1^Y158D, or PARP1^Y158F were treated as described in Figure 5A and harvested at different time points as indicated. Then, chromatin fractions were isolated and subjected to Western blot analysis. Chromatin-bound BRG1 and MRE11 signal intensities were normalized to histone H3 and compared with those of the cells at the beginning of DNA repair (0 minutes after releasing from talazoparib and MMS). Graphs of quantification of BRG1 and MRE11 compared with H3 are shown below. (B) BR#09 cells expressing HA-PARP1 wild type or mutants were treated with 100 nM talazoparib and 0.01% MMS for 1 hour and subjected to immunoprecipitation with anti-HA or control IgG, followed by Western blot with the indicated antibodies. (C) BR#17 cells were cultured in the presence of 100 nM talazoparib, 2 μM PD173074, 100 nM BRGi, or their combination for 8 days, and cell viability was assessed by colony formation assay. Statistical significance was determined by 1-way ANOVA with Tukey’s test (a vs. b or c, b vs. c: P < 0.001). (D) The proposed model of FGFR3-mediated PARPi resistance via HR repair regulated by the FGFR3/PARP1/BRG1 complex.

Figure 7. FGFR and PARP inhibitors synergize in breast cancer xenograft models.

(A) Tumor-bearing mice were treated with olaparib and/or AZD4547. The top graph shows BR#09 xenografts (n = 5), and the bottom graph shows BR#17 xenografts (n = 4). Tumor volumes were measured over time and analyzed by 1-way ANOVA with Tukey’s test (mean ± SD; **P < 0.01 and ***P < 0.001). (B) Survival curves corresponding to the mice in A; BR#09 (top) and BR#17 (bottom). (C) Tumor growth in BR#09 (n = 5) and BR#17 (n = 6) xenograft models treated with talazoparib and PD173074, analyzed by 1-way ANOVA with Tukey’s test (mean ± SD; ***P < 0.001). (D) The tumor chunks of TNBC PDX model (BCX.070) were inoculated into the fourth mammary fat pad of female nude mice. Once tumors reached 80–100 mm³, mice were treated orally with vehicle (n = 3), olaparib (n = 4), AZD4547 (n = 4), or the combination (n = 4). Data are presented as the mean ± SEM; 1-way ANOVA with Tukey’s test (***P < 0.001). (E) Blood chemical test from 6 BALB/c mice treated with talazoparib and PD173074. Mean values and individual data points are shown. Dotted lines represent the reference levels of alanine aminotransferase (ALT), aspartate aminotransferase (AST), and blood urea nitrogen (BUN) in BALB/c mice.

Figure 8. p-Y158 PARP1 is detectable in TNBC patient tumors and PDX models with talazoparib resistance.

(A and B) p-Y158 PARP1 was detected in tissues of TNBC PDX tumors. IHC staining with 3,3′-diaminobenzidine chromogen for monoclonal antibody against p-Y158 PARP1. Representative images and the H-scores of p-Y158 PARP1 are shown in A and B, respectively (n = 9 for sensitive model, n = 13 for resistant model). Scale bars: 100 μm. (C and D) p-Y158 PARP1 was detected in tumor tissues from patients with BRCA1-mutated TNBC or HER2^– breast cancer with known talazoparib resistance (n = 2 each). The images and the H-scores of p-Y158 PARP1 are shown in C and D, respectively. Scale bars: 1 mm. (E and F) Baseline (pretreatment) core needle biopsies were collected from treatment-naive patients with BRCA-associated breast cancers who consented to receive neoadjuvant talazoparib monotherapy on clinical trial. FGFR3 expression levels (log TPM[FGFR3]) were determined using RNA-seq data from paired biopsies — collected at baseline and at week 8 (W8) of treatment — from a cohort of 7 HER2^–, BRCA1-mutated patients (E). Patient treatment outcome was determined by the end of the clinical trial and grouped into pathological complete response (pCR; n = 4) and residual tumor burden (RCB; n = 3). The FGFR3 expression level (log TPM[FGFR3]) was determined by analysis of RNA-Seq data (E). Detailed tumor characterization and RNA-Seq data can be found in a previous publication (32). Mann-Whitney U test was used to compare the change of FGFR3 expression level between pCR and RCB patients (F).