Brain metastasis is a major unmet clinical need associated with morbidity, neurological decline, virtually 100% mortality and high public health care cost. We need expand our repertoire of therapeutic options but the molecular mechanisms underpinning brain metastasis are not well understood. Metastases have not been extensively studied at the genomic level; particularly brain metastases, which are infrequency resected and rarely available in fresh frozen format suitable for high resolution molecular profiling. We hypothesize that metastatic cells with different evolutionary origins (e.g. primary tumour type, genetic background, clonal selection pressure) engage common mechanisms to exploit the unique brain microenvironment.
In this study we performed whole exome sequencing (Illumina HiSeq) and copy number analysis (Illumina HumanOmni 2.5M SNP arrays) on DNA extracted from a cohort of 36 fresh frozen human brain metastases originating from breast, lung, melanoma and esophageal cancers. RNA was extracted simultaneously with DNA (i.e. from single tissue samples) and also analyzed by HiSeq sequencing. Samples were donated by patients who underwent neurosurgical resection at the Royal Brisbane and Women’s and Gold Coast Hospitals in the last 5 years. A total of 22754 somatic single nucleotide variants and small insertion-deletion events (33-3282/patient) were identified across the cohort. The relative mutation loads and signatures of mutational process were consistent with the primary tumour types. Initial analysis of these data has focused on identifying genes and pathways that are recurrently altered across the cohort. We used a combination of systematic, knowledge- and pathways-based approaches. Key findings will be presented.
We hypothesize that analysis and further integration of these data in the future will lead to a better understanding of heterogeneity amongst brain metastases, and ultimately reveal common or recurrent features that could be therapeutically targeted with new approaches, since existing modalities are not broadly effective.