Abstract

Abstract Background: Breast cancer treatment is guided by baseline tumor tissue which does not take into account tumor interaction with therapy. We hypothesized that mutational change may occur early after treatment initiation and that “on-treatment” tumor profiling may be more informative of tumor behavior and, ultimately, patient outcome. Methods: We conducted whole exome sequencing on 49 tumor samples from 2 clinical trials of combined endocrine and targeted therapy (NCT00570921: with fulvestrant/everolimus, and NCT00525161: with tamoxifen/sorafenib). Tumor DNA was subjected to exome capture using Agilent V4+UTR 71Mb and Illumina's HiSeq2000 sequencer was used to generate 2×100bp paired-end reads. After mapping reads to the reference human genome (GRCh37), point mutations and insertions/deletions were detected using the Genome Analysis Tool Kit (GATK 2.6-5). We then retained breast somatic mutations from TCGA and COSMIC databases and excluded germline mutations from dbSNP. Further filtration was then done by removing modifier impact and synonymous mutations based on SnpEff, as well as removing mutations with <15X coverage depth. Results: Across 8 paired tumors on day 28 vs. day 1, 4-pairs on fulvestrant/everolimus and 4 on tamoxifen/sorafenib, there were a total of 50 mutations lost and 77 gained for an average of 6.25 and 9.6 per tumor pair, respectively. This represented, on average, <5% of the total mutations detected, with >95% similarity in the mutational profile of paired tumors. One consistent mutational change in response to treatment was the gain of MUC2 single nucleotide polymorphisms (SNPs) in all 8 paired day-28 vs. day-1 biopsies. This MUC2 mutation pattern persisted even with removal of the coverage depth filter. Other MUC genes had SNPs gained or lost on day 28 but these were infrequent events. Additional notable changes were the gain of HLA-DRB1 SNP's in 2/4 everolimus-treated patients, loss of NCOA3 in one, and loss of TP53 in another. Interestingly, across 5-paired primary and metastatic tumors treated with endocrine therapy alone prior to enrollment, there were more mutations lost than gained (68 vs. 34) for an average of 13.6 and 6.8 per tumor pair, respectively. This represented only slightly over 5% of the mutations detected across these tumors with no consistent genomic event observed. One patient with lobular carcinoma on tamoxifen developed a contralateral lobular associated with metastasis and gain of PIK3CA and NCOR2 SNPs in the new tumor. Conclusion: ER-positive breast tumors undergo early mutational change in response to therapy but the bulk of their mutational profile remains stable, even over long periods of time. MUC2 mutations emerge as a consistent early event which may reflect functional tumor differentiation in response to treatment pressure. Early emergence of mutational events suggests the need for a preoperative treatment window in future trials of this disease in order to interrogate tumor interaction with therapy prior to prolonged adjuvant therapy, which is currently based solely on pretreatment tumor assessment. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr PD3-8.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call