Loss of the DNA repair gene RNase H2 and a unique subset of DDR-deficient leiomyosarcomas.

Abstract

11549 Background: Targeting the DNA damage response (DDR)/Homologous Recombination (HR) pathway is an emerging therapeutic approach for leiomyosarcoma (LMS). Loss of RNase H2 decreases DNA repair via the Non-Homologous End-Joining (NHEJ) pathway, leading to increased double stranded breaks, replication stress, and increased cell death. Therefore, we developed an assay to screen for RNase H2 loss in STS patient samples to determine its prevalence and prognostic significance, particularly in LMS patients. Methods: RNASEH2B homozygous deletion (HomDel) calls from TCGA samples were based on the Allele-Specific Copy number Analysis of Tumors (ASCAT2) algorithm. Immunohistochemistry (IHC) of RNase H2 was performed on tissue microarrays (TMAs) of uterine (U-LMS) and soft tissue (ST-LMS) leiomyosarcoma samples from MD Anderson Cancer Center (MDACC) using a selective antibody developed by Repare Therapeutics. RNase H2 loss was defined as < = 10% cells without nuclear IHC staining and scored by experienced pathologists in 2 separate cohorts. Genomic analysis was performed using SNiPDx™, a targeted-NGS assay designed to detect biallelic loss of function in select DDR related genes and whole genome sequencing (WGS). MDACC TMA samples were clinically annotated by retrospective review. Results: Using TCGA data, RNASEH2B HomDels were seen in 6% (5/80) of all LMS cases, with a higher proportion in U-LMS (15%; 4/27) as compared to ST-LMS (2%; 1/53). In a pan-tumor TMA, RNase H2 loss by IHC was found in 3.8% (32/843) of samples overall, and in 10% (9/88) of LMS samples. This proportion of RNase H2 loss was higher in a larger MDACC LMS cohort, where negative staining of RNase H2 was found in 30% (33/110) of U-LMS and 38% (39/102) of ST-LMS cases. 30 MDACC LMS cases (15 U-LMS and 15 ST-LMS) were analyzed for RNASEH2B HomDels by SNiPDx. In U-LMS, RNASEH2B HomDels were detected in 64% (7/11) of RNase H2 IHC loss cases vs. 0% (0/3) in RNase H2 IHC intact cases. No RNASEH2B HomDels were detected in ST-LMS cases (10 IHC loss, 5 intact). The median overall survival (mOS) of MDACC U-LMS patients (n = 109) was 4.3 years. No significant mOS difference was seen in RNase H2 IHC intact cases versus loss (mOS 4.4 v 3.3 years, p = 0.54). In a separate cohort, 45 U-LMS samples were screened by IHC; 22% (10/45) had RNase H2 loss. RNASEH2B HomDels were detected in 70% (7/10) of RNase H2 IHC loss cases using SNiPDx and confirmed using WGS. In contrast, RNASEH2B HomDels were detected in 3% (1/33) of RNase H2 IHC intact cases. In the combined U-LMS cohort with IHC and SNiPDx results (n = 71), the diagnostic accuracy, sensitivity and specificity of RNase H2 IHC for detecting RNASEH2B HomDels was 76%, 93% and 71% respectively. Conclusions: RNase H2 loss via RNASEH 2B HomDels is prevalent in U-LMS and is unique from other DDR pathway alterations. RNase H2 IHC is an effective screening tool and is being developed for future clinical trials targeting DDR in LMS.