NEW YORK – Seemingly normal tissue falling next to cancer cells may contain large somatic copy number alterations (sCNAs) that may or may not line up with those found in tumors from the same individual, according to new research from the University of Texas MD Anderson Cancer Center.
As they reported yesterday in Nature Biotechnology, the researchers relied on an algorithm known as hapLOH to search for signs of sCNAs in array-based SNP data from thousands of tumor-adjacent tissue or blood samples from participants in the Cancer Genome Atlas project. That analysis, which hinged on maternal and paternal haplotype-resolved allelic data, suggested that somatic copy number shifts can vary significantly in their frequency and location from one tissue to the next, at times arising in parallel with alterations in nearby tumor tissues.
More generally, the work "lays out a framework for detection, characterization, and comparison of sCNAs across human tissues and within tissues from the same donor," first and corresponding author Yasminka Jakubek, a postdoctoral researcher in senior author Paul Scheet's epidemiology lab at MD Anderson Cancer Center, and her colleagues wrote.
Past studies suggest that somatic mosaicism can be quite pronounced in the blood and other tissues, particularly in samples from aging individuals. But although somatic changes arising in specific tissues have been linked to increased blood cancer risk, the alterations have also been implicated in other types of conditions and are sometimes found in healthy tissue, the team explained, making it tricky to know their role in cancer risk.
Using the hapLOH algorithm and Affymetrix SNP array data from TCGA, the researchers searched for subtle allelic imbalances stemming from sCNAs in 1,708 samples — classified as "normal-appearing, adjacent-to-tumor," or NAT, for simplicity — from more than two dozen cancer sites, and in 7,149 TCGA blood samples. A subset of 420 TCGA participants were profiled using both NAT and blood samples.
The team's analyses uncovered hundreds of sCNAs in 1.8 percent of the blood samples and 4.6 percent of the NAT samples spanning 27 sites in the body. This included alterations that the researchers verified with exome sequence data for a subset of the samples.
The normal tissues neighboring head and neck squamous cell carcinoma, breast invasive carcinoma, or kidney renal clear cell carcinoma, in particular, appeared to have sCNA profiles that sharply differed from those in NATs from other tissue types. sCNAs also turned up in NAT samples near bladder urothelial carcinomas, sarcomas, stomach adenocarcinomas, and ovarian cancers, but were not found in colon cancer-adjacent samples.
Roughly one-quarter of NATs from individuals with head and neck squamous cell carcinomas were marked by sCNAs, including recurrent chromosome 9q alterations — a change that also turned up in NAT near bladder urothelial carcinomas. Normal blood samples contained losses involving parts of chromosome 13 or chromosome 20, while chromosome 20 gains appeared in stomach adenocarcinoma-neighboring tissue.
When they focused on X chromosome alterations in nearly 4,100 blood samples and more than 800 NAT samples from females, the researchers tracked down more than three dozen blood samples containing 67 sCNAs involving the X chromosome. They also saw another 32 X chromosome sCNAs in 18 of the NATs.
In some cases, the same copy number clones found in the tumors seemed to carry over to nearby tissues, consistent with a clonal connection. But that was not always the case, the authors explained, noting that the analysis unearthed "instances in which both NAT tissue and tumor tissue harbor a gain of the same oncogene arising in parallel from distinct parental haplotypes."