(P 260) TARGETING HOMOLOGOUS RECOMBINATION DEFICIENCY (HRD) IN LEIOMYOSARCOMA WITH COMBINATION OF PARP INHIBITORS (PARPI) AND CHEMOTHERAPY (CTX).

Objective: Leiomyosarcomas (LMS) are genetically heterogeneous tumors that arise from smooth muscle. Currently, the mainstay of systemic treatment for patients with advanced/metastatic disease is doxorubicin (Dox) based CTx. Dox causes double strand (DSB) DNA breaks in exposed cells. Several genomic analyses of LMS reveal defects in the homologous recombination (HR) DNA repair pathway in about half of patients, consistent with a druggable “BRCAness” phenotype. In HR deficient cells, DSB are restored by non-homologous end joining (NHEJ), which can be impaired by PARPi. Therefore, the combination of chemotherapy and PARPi may represent a promising therapeutic option for LMS. To validate this hypothesis, we performed a series of experiments in vitro on LMS cell lines. To confirm that HR deficiency mechanisms were present in the experimental cell lines, we performed a comparative transcriptome analysis of tumor/cell line paired samples.

Methods: Dox, Docetaxel (Doc) and Temozolomide (Tmz) were evaluated in combination with several PARPi (Olaparib [Ola], Niraparib [Nira] and Talazoparib [Tala]) at 12 different drug concentrations. All drug concentrations were chosen according to EC50, identified in single drug experiments on 12 LMS cell lines. Drug combinations were tested in a high throughput manner on four LMS cell lines of either gynecological (GY), abdominal (A) or extremity (E) origin. Cells were incubated with each combination for 7 days. Evaluation of drug combination effect was performed using a Bliss synergy score. This system quantifies the degree of synergy as multiplicative effect of single drugs as if they acted independently. Specifically, with a synergy score of -5 to 5, the interaction between two drugs is considered as additive: <-5 antagonistic and > 5 synergistic, and therefore a promising combination.
Tumor-cell line pairs were then selected for transcriptome analysis. All analyses were performed in R (v4.2.3). Cell line and tumour data were combined and then normalized using edgeR (v3.40.2) calcNormalFactors function. Protein coding genes were used (n = 18471).
The normalized expression data were then run through GSVA (v1.46.0), using all available genesets from the Molecular Signature Database. GSVA scores were then analyzed using limma (v3.54.2) with cell lines vs. tumour as the coefficient of interest while accounting for cell line/tumour pairing. Benjamin Hochberg adjusted p values were reported.

Results: Anticancer activity, ranging from additive to synergistic was seen with all combinations. Results were consistent among all cell lines, independent of site of cell line origin (Table 1). The most synergistic combination in the majority of LMS cell lines was Dox or Tmz when combined with Tala, reaching up to 15 % and 27% above Bliss respectively. In contrast, Doc only had an additive effect with all analyzed PARPi. Comprehensive transcriptome analysis revealed similar expression of gene sets involved in HR between cell lines and corresponding tumors (Fig1, Table 2).

Conclusion: These data suggest that combining Dox or Tmz with PARPi may be a promising treatment option for LMS patients. Transcriptome analysis confirmed the presence of similar HRD mechanisms between tumors and corresponding cell lines. Recent clinical studies support this notion in uterine LMS. Importantly, our data suggest that such an approach may be extended to all sites of LMS. Preclinical in vivo studies are underway to identify the most promising combinations for future clinical trial design