Researchers identify DNA changes, biological pathways associated with inherited cancer risk

Stanford researchers have conducted the first large-scale screen of these inherited changes, called single nucleotide variants, and homed in on fewer than 400 that are essential to initiate and drive cancer growth. These variants control several common biological pathways, including those governing whether and how well a cell can repair damage to its DNA, how it produces energy, and how it interacts with and moves through its microenvironment.

These common themes hint at new therapeutic targets aimed at preventing cancer or stopping its growth, the researchers believe. Understanding which variants contribute significantly to cancer risk may also enhance genetic screening meant to assess a person's lifetime risk of cancer.

"We distilled large compendia of information from millions of people diagnosed with any of the 13 most common cancer types, which constitute over 90% of all human malignancies," said Paul Khavari, MD, PhD, chair of dermatology. "This enormous funnel of data allowed us to identify 380 variants that control the expression of one or more cancer-associated genes. Certain variants, if you are unlucky enough to inherit them from your parents, can increase your risk of developing many types of cancer."

Khavari, who is the Carl J. Herzog Professor in Dermatology in the School of Medicine, is the senior author of the research, which was published on Feb. 17 in Nature Genetics. Former graduate student Laura Kellman, PhD, is the lead author of the study.

The risks we inherit

The study focused on DNA sequences inherited at conception, known as a person's germline genome, rather than on mutations that can accumulate during a person's lifetime as cells divide during development or to repair injury. Examples of well-known inherited cancer-associated mutations are the BRCA1 and BRCA2 genes that confer a significantly increased risk of breast and ovarian cancers. But only a few of these high-profile mutations are currently used to predict cancer risk.

The variants Kellman and Khavari identified are not in so-called "coding" genes, which encode the instructions to make proteins that do much of the work of the body. Instead, they are in regulatory regions that control whether, when and how much these genes are expressed. Often these regulatory regions influence the expression of nearby genes; sometimes they influence distant genes.

In 2020, Khavari launched a research project funded by the National Human Genome Research Institute to develop the Atlas of Regulatory Variants in Disease to pinpoint variants linked to the risk of developing 42 common complex diseases including cancers, develop individualized risk scores for each disease to aid in screening and prevention, and suggest new treatment strategies.

Previous studies, known as genome-wide association studies, identified variants found more often in people with specific kinds of cancers than in their cancer-free peers -- a kind of guilt-by-association metric. But these studies, of which there have been many, fall short of proving that the variations change the activity of the regulatory region in ways that pump up or tamp down the expression of the genes they regulate; they also do not identify which genes are affected.

Kellman, Khavari and their colleagues took a different approach. They amassed over 4,000 suspect variants identified by genome-wide association studies, or GWAS, in 13 types of cancer and tacked those regulatory regions -- along with control sequences -- to DNA sequences, each with a unique bar code. They then conducted what are known as massively parallel reporter assays to determine which variants changed the expression of the bar-tagged sequence in the relevant cell type, testing variants associated with lung cancer in human lung cells, for example.

Winnowing thousands of potential variants down to a few hundred functional regulatory regions allowed the researchers to combine information from pre-existing databases about DNA folding, tissue-specific gene expression profiles and others to identify about 1,100 target genes likely to play a role in cancer development. Some are specific to a certain type of cancer while others appear to increase the risk of several cancers.

"A lot of these genes make sense in the context of what we know about cancer development," Khavari said. "Some are involved in cell death pathways, and others affect how cells interact with the extracellular environment, for example. One of the most prominent pathways is involved in the function of cellular mitochondria -- tiny cellular energy factories that support cell growth and division."

The immune system's role

But the researchers also found things that surprised them.

"One pathway that really popped out includes a number of genes closely associated with inflammation," Khavari said. "While a connection has been established between inflammation and cancer, it's not been clear what was driving this process -- the cancer cells or the immune system. This finding suggests there may be cross talk between cells and the immune system that drives chronic inflammation and increases cancer risk."

Finally, the researchers used gene editing techniques in laboratory-grown cancer cells to show that as many as half of the variants are required to support ongoing cancer growth. They expect the study's findings will be a springboard for researchers around the world seeking to understand inherited cancer risk and develop new therapies.

"Now we have a first-generation cartographic map of functional single nucleotide variants that determine a person's lifetime cancer risk," Khavari said. "We expect that this information will be incorporated into increasingly informative genetic screening tests that will become available over the next decade to help determine who is most at risk for many types of genetically complex diseases, including cancer. This general approach may help provide an individualized risk assessment for common diseases to guide interventions, such as lifestyle changes, pharmacologic preventatives and diagnostic screening."

The study was funded by the U.S. Veterans Affairs Office of Research and Development and the National Institutes of Health (grants AR076965, AR43799, CA142635 and U24HG010856).