NEW YORK (GenomeWeb) – Two large-scale exome sequencing studies have independently discovered that somatic mutations in a small number of genes in a fraction of an individual's blood DNA increase their risk to develop blood cancer. In addition, one of the studies found that the mutations, which occur more frequently with age, also increase the risk for overall mortality, heart disease, and stroke.
The results are unlikely to have practical clinical applications in the near term, as no interventions exist at the moment for those that harbor the premalignant mutations in their blood, according to the authors. But knowledge about the premalignant state may allow researchers to develop drugs that could prevent its progression to full-blown cancer.
The two research teams, which used cohorts that were originally collected for projects unrelated to cancer – one for type 2 diabetes research, the other for research into schizophrenia and bipolar disorder – published their studies online in the New England Journal of Medicine today. Despite using very different approaches to analyze the exome data, both teams discovered mutations in essentially the same genes and came to very similar conclusions with regard to the increased cancer risk.
One study, led by Ben Ebert, an associate professor at Harvard Medical School and Brigham and Women's Hospital in Boston, analyzed whole-exome sequencing data from peripheral blood samples of more than 17,000 individuals. The samples came from 22 case-control cohorts for type 2 diabetes research and about 1,400 participants in the Jackson Heart Study; almost 8,000 of the patients had type 2 diabetes.
According to Ebert, whose research focuses on hematological cancer, the exome data had already been generated by the Broad Institute when he and his team decided to embark on the study and look for early signs of blood cancer in the blood-derived DNA. "We realized this was an ideal opportunity to look for a pre-malignant lesion in an extremely large sample set," he told GenomeWeb.
The mutations they were looking for were somatic in nature, meaning they only affected a fraction of the cells in the blood. Cancer is thought to arise because mutated hematopoietic stem cells clonally expand, a process called clonal hematopoiesis, and acquire additional mutations that make them divide uncontrollably.
The researchers focused their analysis on variants in 160 genes that had been previously described in the literature in myeloid and lymphoid cancers. They found them in a fraction of the blood DNA of about 750 people, most of whom harbored a single mutation. In total, they discovered about 800 variants in 73 different genes. Three genes – DNMT3A, TET2, and ASXL1 – were most commonly mutated; mutations in DNMT3A alone accounted for about half the variants found.
The frequency of the mutations increased with age: while they were very rare in patients younger than 40 years, they occurred in 10 percent of participants at least 70 years of age. Overall, "it is much more common than we expected for individuals to have a mutation in the blood," Ebert said.
Variants also persisted over time – those participants who had their blood reanalyzed after four to eight years still had the mutations.
For part of the study cohort, follow-up data on their health was available and the researchers looked at how mutations in their blood DNA correlated with disease and mortality. They found that the risk to develop a hematologic cancer was increased about 11-fold, which was somewhat expected, Ebert said, because the mutations represent "a first step on the way towards blood cancer."
Unexpectedly, the researchers also found that a blood mutation increased the risk for coronary heart disease 2-fold, the risk increase for ischemic stroke 2.6-fold, and the risk for overall mortality 1.4-fold. "That was completely surprising," he said.
While it is not known yet how cancer mutations might contribute to cardiovascular disease, it is known that many types of blood cells play a role in heart disease, and it may be that blood cells carrying such mutations "do not behave exactly as they should," Ebert said. His group plans to study the link between cancer mutations and cardiovascular disease risk further, in particular, how different genes contribute to that risk. "In the future, we might be able to home in on individual mutations conferring the greatest risk," he said.
In addition, they plan to analyze other exome sequencing datasets from peripheral blood in order to confirm and extend their findings.
The other study, led by Steven McCarroll, an assistant professor of genetics at Harvard Medical School and director of genetics at the Broad Institute's Stanley Center for Psychatric Research, analyzed whole-exome sequencing data from peripheral blood cells of about 12,500 individuals from Sweden. This cohort was originally assembled for schizophrenia research and included about 5,000 persons with schizophrenia, and 1,200 with biopolar disorder.
His team originally wanted to understand whether somatic mutations contributed to schizophrenia risk, McCarroll told GenomeWeb. To that end, they developed a computational method to discover somatic mutations across the exome that were present in unusual allelic fractions, rather than searching in a limited set of genes, as the other study did.
Overall, they found about 3,100 somatic mutations, the majority of them in four genes – DNMT3A, TET2, ASXL1, and PPM1D – three of which were the same genes the other study found. Mutations in these genes occurred in both the schizophrenia patients and the healthy controls. "As soon as we saw they were all four cancer genes, and in particular tumor suppressor genes, we realized this was potentially an early stage in the development of cancer," McCarroll said.
Again, the frequency of somatic mutations increased with age: only 1 percent of participants under age 50 had such variants compared to 10 percent of those over age 65.
Follow-up data for the Swedish cohort was available through their electronic medical records, and it turned out that those with somatic blood mutations developed blood cancer at about 13 times the normal rate. Also, 42 percent of those who later developed blood cancer had somatic mutations in their exome data at the time of the blood draw, more than six months before they were diagnosed with the cancer.
In two cases, the researchers were able to obtain and sequence a bone marrow biopsy from patients at the time of their cancer diagnosis. The analysis showed that both their cancers developed from the earlier pre-malignant clonally expanded blood cells.
Both groups essentially came up with the same results. "We looked at different sample sets, using different computational approaches, and had extraordinarily concordant findings," Ebert said. "We were extremely gratified to see that we had discovered the same thing in very different ways, and that gave us both great confidence that the results were correct."
But neither group thinks there is value right now in identifying individuals with the pre-malignant lesion in their blood DNA. "There is no reason to get this test done [now] because we don't have an intervention," Ebert explained. "The test would work, but at the moment, we would not have anything to do, other than to follow that individual more closely," he said.
And despite the elevated risk of those with premalignant mutations to develop blood cancer, their overall risk is still fairly small – an estimated 1 percent per year, he said.
"There is no clear benefit to having the information, other than worry," MacCarroll said, adding that "having one of these [mutated] clones is not a genetic destiny," though it is "a real risk factor."
"What needs to happen now is to develop therapies that reduce the likelihood that a clone will progress to full-blown cancer," he said. "And it will become much more possible to develop preventive therapies now that one can identify high-risk cohorts in which to do clinical trials."
The idea is to develop a drug that would inhibit the premalignant clone, Ebert said, and to prevent them from acquiring additional mutations that would turn them into cancer. "I don't think we have to necessarily think about removing it entirely, because the pace of developing cancer from this is quite slow, and if we can just decrease the likelihood a little bit, then it might decrease the likelihood that that individual might get cancer during their lifetime," he said.
This would represent a departure from most current cancer research, which focuses on developing treatments for existing cancers, rather than to prevent it. "But it's an attractive idea to start thinking about how we could develop therapies that would delay or prevent the development of blood cancers in those most at risk," Ebert said.