NEW YORK (GenomeWeb News) – In a paper scheduled to appear online this week in the Proceedings of the National Academy of Sciences, scientists used DNA methylation patterns to gauge what happens during cancer-related cellular expansions, as a lone cancerous cell progresses to become a tumor containing billions of cells.
Researchers from the University of Southern California and the University of Cambridge used bisulfite sequencing to measure methylation in colorectal cancers taken from a dozen men, comparing methylation patterns at two genomic sites in samples taken from different regions of each patient's tumor. They saw similar methylation patterns from opposite sides of the same tumors, suggesting they arose via so-called isotropic or "flat" clonal expansion — in a rapid burst — rather than by sequential expansion.
"We saw that we couldn't really tell the left side [of the tumor] from the right, which means that it was one clonal expansion," senior author Darryl Shibata, a pathologist at the University of Southern California Keck School of Medicine, told GenomeWeb Daily News.
Because it's difficult or impossible to actually observe human cancer growth over time, researchers were split on whether tumors result from flat clonal expansions of cells, in which different parts of the same tumor aren't easily distinguished by their location, or whether they grow through a series of expansions, leaving distinct cells in different areas.
For their part, Shibata and his team decided to track passenger methylation as a way to get at how colorectal tumors expand, comparing methylation patterns in different parts of the same tumor. After getting colorectal cancer samples from 12 men, the researchers sampled five to seven cancer gland fragments from opposite sides of each tumor. They then used bisulfite sequencing to look at methylation patterns at two CpG-rich tags, LOC and BGN, on the X-chromosome in cells taken from different parts of the same tumor.
The researchers selected these tags because they likely represent neutral passenger methylation, since LOC is found in an intergenic region and BGN is expressed in connective but not epithelial tissue. In addition, using these tags simplified the analysis, Shibata explained, since the regions were on the X-chromosome and all of the individuals sampled were men.
The overall approach is based on the idea that methylation and demethylation errors arise during cell division, similar to DNA point mutations. "Although this assumption has not been directly tested, DNA methylation appears to require cell division," the researchers explained, "and methylation patterns in the normal human colon are consistent with such replication errors."
The researchers found a range of polymorphic methylation patterns in the tumors. But while these patterns varied from one sample to the next, the general features were similar on both sides of the tumor. Rather than gaining mutations over time, which would give them different methylation patterns, the tumor cells on either side of the tumor appear to be "homogenously heterogeneous," exhibiting a similar range of methylation patterns, Shibata explained.
Based on these and other results, the researchers concluded that the tumors had undergone flat expansion, likely involving a fast expansion of the initial population of cells.
"It kind of tells us what we knew all along — that cancers are clonal expansions," Shibata said. "In many ways, this isn't a surprise ... but it's nice to actually have data."
The researchers also did computer simulations to try to figure out how many long-lived cancer stem cell lineages there are in the tumors. Their results suggest that the tumors tested likely didn't spring from rare cancer stem cells, but rather from several more common stem cell lineages.
"What we found was that the diversity was too high for very rare cancer stem cells," Shibata said. "It kind of says that the number of long-lived lineages were not very rare."
At the moment, Dr. Shibata and his co-workers are looking at metastatic colorectal cancers to see whether there are similarities and/or differences in methylation patterns compared to those observed in tumors in this study.
And, Shibata said, such work could provide new insights into tumor characteristics and diversity. And because chemotherapy often seems to fail due to the presence of variance in the tumor itself, he added, such information could also improve researchers' understanding of what makes chemotherapy effective in some cases and ineffective in others.