NEW YORK – Tumor metastases may get their start years before a cancer is even diagnosed, a new study has found.
A team led by researchers at Stanford University examined exome sequencing data from paired primary and metastatic tumor samples to study how metastatic spread occurs, a process they noted is currently not well understood.
Through their analysis of samples from more than 100 patients with breast, colorectal, or lung cancer, the researchers found a low number of clonal mutations that were private to metastases, suggesting metastatic seeding occured early in disease development. They noted, though, that treatment influences clonal evolution, with treated metastases harboring more private mutations, as they reported Monday in Nature Genetics.
"[T]hese data suggest that the natural course of metastasis is selectively relaxed relative to early tumorigenesis and that metastasis-private mutations are not drivers of cancer spread but instead associated with drug resistance," Stanford's Christina Curtis and her colleagues wrote in their paper.
Her team amassed genome sequencing data from the literature from matched normal, primary tumor, and metastasis samples from 136 patients with breast, colorectal, or lung cancer, a total of 457 tumor samples. This set included 99 treated and 100 untreated metastatic samples.
Using three different computational approaches — PolyPhen2, FATHMM-XF, and CHASMplus — the researchers examined the number of functional driver gene mutations in the samples. In all, they uncovered more than 1,000 functional driver somatic single nucleotide variants or indels across the three cancer types, slightly more than 700 of which were clonal and about 300 of which were subclonal. Most of the clonal drivers were shared between primary tumors and metastases.
There was a decreased ratio of nonsynonymous and synonymous somatic SNVs between putative driver genes in metastases, which the researchers said indicated a relaxed selective pressure relative to early cancer development in breast and colorectal cancer, though not in lung cancer. This suggested to them that untreated metastases typically arise from the major or dominant clone of the primary tumor.
Treatment, though, exerts a selective pressure, as indicated by the enrichment of functional driver mutations and private mutations among treated metastases. This suggests that the genomic drivers needed for both invasion and metastasis arise early in the primary tumor.
When they examined whether tumors were typically formed through monoclonal or polyclonal seeding, the researchers found a mix. Untreated, distant metastases, for instance, were more likely to have been formed through polyclonal seeding, with 54 percent of lymph node, 26 percent of liver, 17 percent of brain, and 8 percent of lung metastases reflecting polyclonal seeding. Polyclonal seeding also appeared to be linked to worse prognosis, though they noted their sample size was limited and the finding needs to be further explored.
Biopsies taken after treatment, meanwhile, were more likely to exhibit monoclonal seeding.
Using a computational approach they adapted, the researchers also estimated the time from when the metastasis seeding event occured to when a biopsy was taken. On average, they found that the seeding events took place between about two and four years before diagnosis, indicating systemic spread can take place soon after a malignant transformation event and years before that event is detected.
"Our findings highlight the importance of studying the natural course of metastasis and the impact of therapy on this process," Curtis and her colleagues wrote. "Future studies of paired primary tumors and metastases with comprehensive treatment information subject to dense multi-region sampling and single-cell sequencing may provide additional resolution on these processes."