Genome-wide case study of clonal evolution over 16 years in Barrett's Esophagus.

B71

Clonal evolution drives the process of neoplastic progression, yet little is known about the dynamics of that evolution. Recent evidence suggests that genetic diversity accumulates over time in a neoplasm, yet there is also evidence of clonal expansions which should reduce diversity. Only rarely are we able to track clonal evolution over long timespans in a single neoplasm from a single patient, since, once detected, benign tumors are usually removed. Thus, Barrett's Esophagus (BE) is an excellent model system to study neoplastic clonal evolution because endoscopic surveillance is the standard of care. We used Illumina's 317k SNP chip platform to detect loss of heterozygocity (LOH) and deletion events in longitudinal BE samples. We analyzed twelve biopsies from a single patient's neoplasm that were collected at four time points: 1989, 1993, 2001 and 2006 (16 years timespan). From each time point, three samples from different biopsy levels (distance from the incisors) were genotyped. An additional lymphocyte sample, collected in 2006, was used as control in paired comparisons against each of the 12 BE samples to detect LOH and deletion events. Illumina BeadStudio software was used to identify statistically significant regions of LOH and deletions. We estimated a phylogenetic tree that relates the samples based on the presence or absence of the events using parsimony. The DOLLOP program from the PHYLIP phylogenetics package was used to estimate the most parsimonious tree relating a set of 388 distinct events across the 12 BE samples. The data suggest a selective sweep between 1989 and 1993 mediated by loss of p16 (CDKN2A). In the 12 years following 1993, no other clone expanded to fixation in the neoplasm, although multiple clones persisted. Clones were characterized by LOH and/or deletions on chromosomes 1 (at ~49Mb, ~206Mb, ~226Mb), 2 (at ~116Mb, ~205Mb, ~219Mb), 3 (at ~60Mb, ~84Mb) , 4 (at ~92Mb), 6 (at ~148Mb), 7 (at ~110Mb, ~117Mb, ~156Mb), 9 (at ~22Mb, whole 9p arm), 11 (at ~102Mb), 14 (at ~35Mb), 16 (at ~77Mb), 21 (at ~36Mb) and X (at ~314Mb). Losses at fragile sites FRA3B, FRA7G and FRA16D evolved early in progression. This is the first study to track clonal evolution on a genome-wide scale in a neoplasm over 16 years. Evidence of an early clonal expansion followed by genetic divergence suggests a potential solution to the conflicting observations of increasing genetic diversity and homogenizing clonal expansions. The power of such a study can be seen through the identification of early and late alterations as well as those that drive clonal expansion. However, there is a need to devise better evolutionary models for clonal biomarkers (such as LOH) to estimate clonal phylogenies from longitudinal cancer data.