J. G. Grossman1, C. Maher2,4, B. S. White2,4, A. C. Lockhart3,4, T. Fleming1,3, K. Lim3,4, B. Goetz3, E. Pittman1, S. M. Strasberg1,3, D. C. Linehan1,3, W. Hawkins1,3, S. P. Goedegebuure1,3, E. Mardis2,5, R. K. Wilson2,3,5, T. Ley2,4,5, R. C. Fields1,3 1Washington University,Department Of General Surgery,St. Louis, MO, USA 2Washington University,The Genome Institute,Saint Louis, MO, USA 3Alvin J. Siteman Cancer Center,Saint Louis, MO, USA 4Washington University,Department Of Medicine, Oncology Division,Saint Louis, MO, USA 5Washington University,Department Of Genetics,Saint Louis, MO, USA
Introduction:
Colorectal cancer is the second leading cause of cancer mortality in the United States, and death from CRC occurs via sequelae of metastases. Our lack of understanding of mechanisms of metastasis formation has prevented the identification and direct targeting of pathways necessary for growth and survival of metastaseis. Next-generation sequencing gives us the capability to better study the evolutionary biology of metastasis, however a comprehensive analysis comparing matched primary and metastatic colorectal tumors has not been performed to date. Our group in collaboration with the Washington University Genome Institute is currently analyzing 10 patients’ primary and metastatic tumors by whole genome and transcriptome sequencing. We present the evolution of the clonal relationships of primary and metastatic tumors from three cases completed to date.
Methods:
Patients with metastatic colorectal cancer were consented, and primary tumor and liver metastases were procured during operative resection. Additionally, uninvolved colon, uninvolved liver, and peripheral blood were collected (germline controls). If necessary, tissue from prior resections was obtained from paraffin blocks. Deep exome and WGS were used to calculate variant allele frequency (VAFs) for the somatic single nucleotide variants. We chose to incorporate deep exome sequencing thereby enabling us to more accurately calculate VAFs, which in turn improves our ability to reconstruct the clonal architecture. To accomplish this, we used SciClone, a tool developed at Washington University Genome Institute for identifying VAF clusters via variational Bayesian Beta mixture modeling.
Results:
Among the primary tumors prevalent somatically altered genes were APC, TP53, KRAS, PIK3CA, and SMAD4. Upon evaluating the clonal evolution from primary to metastases, phylogenetic trees were generated for each patient illustrating tumor similarities and differences to each other and in relation to normal tissue. Additionally, the clonal evolution from primary to metastases was mapped and clearly shows every tumors’ subclonal makeup and how they relate to one another. Every patient had a dominant clone originating from the primary tumor that was present in all of its corresponding metastases. However, unique subclones appear to arise in all metastatic samples. In some cases, these subclones are shared among various metastases of the same patient, and these similarities may be due to the spatial or temporal proximity of the tumors.
Conclusion:
Exploring the clonal evolution from primary tumor to metastasis provides a greater understanding of cancer biology, and further elucidates the importance of tumor heterogeneity. Continued investigation is necessary to evaluate which subclones may be biologically relevant to disease progression and treatment resistance. In addition, future identification of altered genes associated with metastatic clones may lead to targeted cancer therapies.