Fig. 2: High HRD scores by mutation signature analysis is associated with platinum and PARP inhibitor sensitivity in gastric cancer.

a Gastric cancer patients cell lines from CCLE stratified by HRD Score. Cut-off value of ≥42 for HR deficiency was previously defined (dashed-line); cell-lines with low (gray) and high (blue) HRD scores indicated. Starred cell-lines used in experimental validation studies. b Representative immunofluorescence images (top) and quantification (bottom) for RPE1 and AGS cells; Number of RAD51 (green) positive foci indicating DNA repair after treatment with 5 Gy gamma radiation; Data expressed as the mean number of RAD51 positive foci per cell ±SEM; Scale bar = 10 µm. c Representative immunofluorescence images (top) and quantification (bottom) from KE39 and HGC27 cell lines; Number of RAD51 (green) positive foci indicating DNA repair after treatment with 5 Gy gamma radiation; Data expressed as the mean number of RAD51 positive foci per cell ±SEM; Scale bar = 10 µm. d Cisplatin sensitivity between low (gray) and high (blue) HRD gastric cancer cell lines available in BROAD institute PRISM repurposing drug screen dataset. Difference between the cisplatin sensitivity (log2 fold change) is represented as the median with interquartile range. e Dose-response curve of non-neoplastic RPE1, low (gray) and high (blue) HRD (blue) gastric cancer cell lines to oxaliplatin. Error bars indicate SD. Best-fit IC50 scores are displayed. f Talazoparib sensitivity between low (gray) and high (blue) HRD gastric cancer cell lines available in BROAD institute PRISM repurposing drug screen dataset. Difference between the cisplatin sensitivity (log2 fold change) is represented as the median with interquartile range. g Dose-response curve of non-neoplastic cell line RPE1, low (gray) and high (blue) HRD gastric cancer cell lines to indicated concentrations of Talazoparib. Error bars indicate SD. Best-fit IC50 scores are displayed. h Colony formation assay (left) and the quantification shown as mean intensity from each well using ImageJ (right) of gastric cancer cell with high HRD score (blue) and non-neoplastic RPE1 (gray). i Platinum-specific survival in patients treated at Dana-Farber Cancer Institute with metastatic gastroesophageal adenocarcinoma who have Sigma 3 signature >0 or Cosmic ID6 signature >0 versus Sigma 3 signature = 0 and Cosmic ID6 signature = 0 (median platinum-specific survival 1015 days versus 496 days, HR = 2.43 [95% CI: 1.16−5.08], log-rank p = 0.0152). X axis, time from first dose of platinum-based chemotherapy in days; Y axis, proportion surviving. j Platinum-specific survival in patients from the TCGA-STAD cohort with localized gastric adenocarcinoma, who have HRD score >=42 and HRDetect score >= 0.7 (HR-D group) versus HRD score < 42 and HRDetect score <0.7 (HR-P group). Median platinum-specific survival 1015 days versus 496 days, HR = 3.39 [95% CI: 1.05−11.01], log-rank p = 0.0308. X axis, time from first dose of platinum-based chemotherapy in days; Y axis, proportion surviving, progression-free survival. k Overall survival in a multicenter cohort of patients with microsatellite stable gastroesophageal cancer who have a pathogenic/likely pathogenic mutation in BRCA1, BRCA2, or PALB2 versus those with Loss of Heterozygosity-Low (LOH-low) score (score < 16%) (median overall survival: 459 days versus 354 days, HR = 0.809 [95% CI: 0.677−0.968], log-rank p = 0.02). X axis, time from diagnosis in days; Y axis, proportion surviving.