Genomic heterogeneity in primary colorectal carcinomas and their metastases: born bad or brought up a villain?

Progression of solid cancers, colorectal carcinomas among them, from their primaries to metastatic lesions traditionally is thought to proceed by a stepwise acquisition of and selection for genomic aberrations. To test if patterns of genomic aberrations would be consistent with this model, we studied 10 colorectal carcinoma primary-metastasis pairs, 9 with 1 liver metastasis each and 1 with 2 metastases. Next-generation targeted sequencing (50-gene panel) with samples obtained from different regions of the primaries and their metastases demonstrated 1-11 gene mutations per lesion. But only in 2 tumors were there seen mutations in all samples from the metastasis and not any of the primaries (BRAFD594N and SMARCB1R377C mutation, respectively). However, allelotyping the multiregional samples with polymorphous microsatellite markers (17p13.1, D9S942, D9S1748, D5S346, D5S1385) and DNA methylation studies with a marker panel (MLH1, CDNK2A, NEUROG1, CRABP1, CACNA1G, IGF2, RUNX3, SOCS1) showed remarkably "insular" genomic aberrations in all cases for at least some of the analyses. The marked preponderance of mutations shared by the primaries and their metastases throughout the lesions over mutations private to metastases suggests that, at least in many cases, colorectal carcinomas might be endowed with a mutational load sufficient for fully fledged metastases even at a very early stage ("born bad"). But the very focal allelic imbalances and methylations observed here, hypothetically, could play a role in clinically metastasizing disease, a process of years rather than months and very much a matter of tumor-host interactions when tumor cells adapt to the host microenvironment.