Identification and treatment of patients with BRCA1 or BRCA2-defective breast and ovarian cancer

Abstract

Inactivating mutations in BRCA1 or BRCA2 confer a large lifetime risk of breast and ovarian cancer. These genes are involved in high-fidelity repair of DNA double strand breaks. Although defects in BRCA1 and BRCA2 are contributing to tumorigenesis, they may also form therapeutic targets. We investigated two treatment options, one being overloading the tumor with DNA double strand breaks using high dose alkylating chemotherapy. The other employing synthetic lethality between Poly(ADP)Polymerase (PARP) inhibiton and BRCA1 and BRCA2 deficiencies. To identify which patients could benefit from these treatments, tests to identify patients with defective BRCA1 and BRCA2 need to be identified and validated. In this thesis we studied several of these markers. We established that a previously identified tumor DNA profile with amplifications and deletions similar to that of BRCA1 mutated breast cancer can be robustly obtained using several (SNP) array based technologies and next generation sequencing. We then identified BRCA1-like and BRCA2-like DNA signatures in ovarian cancer. We further established that tumor type specific and mutation type specific signatures corresponded best with identifying tumors with a mutation in BRCA1 or BRCA2, but that certain combinations of genes and tumor types might be present. Subsequently, we used the BRCA1-like breast cancer test on tumor DNA of one balanced cohort of patients and one randomized trial in which patients were treated with high dose alkylating chemotherapy. We found in both studies that patients with a BRCA1-like tumor benefitted substantially from the high dose alkylating chemotherapy. Corrected for prognostic factors, patients with BRCA1-like tumors had 5-6-fold decreased risk of recurrence and death. The prediction of patients that benefit could be improved by measuring the expression of two other genes. Xist and 53BP1 were identified as resistance markers in BRCA1 deficient preclinical model systems. When used in patients with a BRCA1-like tumor, these two genes identified patients that were not benefitting from high dose alkylating chemotherapy treatment. Next evaluating these high dose alkylating regimens in patients with tumors with a BRCA-like DNA profile, we set up a trial to test PARP inhibition by olaparib in combination with carboplatin (a double strand break inducing chemotherapeutic). In a phase 1/randomized phase 2 trial we will first establish the dose of olaparib and carboplatin that can be administered in combination. In phase 2 we will randomize patients with a BRCA1 or BRCA2-mutated breast cancer to first line treatment with olaparib-carboplatin followed by olaparib monotherapy versus capecitabine, the currently established clinical standard for these patients. In conclusion, we performed technical validation of breast cancer BRCA-like profiles, established ovarian cancer BRCA-like profiles, validated the BRCA1-like breast cancer profile as predictor for high dose alkylating chemotherapy, performed early validation of the putative resistance markers Xist and 53BP1 and set up a randomized study to test olaparib-carboplatin as first line treatment of BRCA1 or BRCA2 mutated patients.