NEW YORK – New research suggests that analyzing cell-free DNA in blood from Li-Fraumeni syndrome patients can pick up early signs of cancer development. The condition, characterized by pathogenic germline variants in the tumor suppressor gene TP53, predisposes individuals to various types of cancer.
"Our method presents a novel approach to the management of patients with LFS that is both comprehensive, noninvasive, and sensitive," Derek Wong, a postdoctoral fellow at the Princess Margaret Cancer Center, and his colleagues wrote in an abstract outlining work Wong presented at the American College of Medical Genetics and Genomics annual meeting last week.
Individuals with hereditary cancer predisposition syndromes such as Li-Fraumeni are typically born with a germline variant affecting a cancer-related gene, such as TP53, Wong explained. In the "two hit" model for cancer development, additional somatic changes spur processes that lead to tumor formation.
Although heightened screening is recommended for such individuals, surveillance efforts are not always comprehensive or sensitive enough to catch cancer cases at an early stage, prompting interest in circulating tumor DNA-based methods for finding early signs of cancer in Li-Fraumeni syndrome patients.
"Nearly 50 percent of female [Li-Fraumeni syndrome] patients have developed a cancer by age 30," Wong noted, "which makes early detection vital for this population."
For their study, investigators with the "cfDNA in hereditary and high-risk malignancies" (CHARM) consortium and the Terry Fox Li-Fraumeni Syndrome New Frontiers Program Project used shallow whole-genome sequencing, cell-free methylated DNA immunoprecipitation (cfMeDIP), and targeted sequencing on a panel of CHARM project genes to assess copy number variants, DNA fragmentation features, methylation, and gene mutation profiles in cell-free DNA from more than 160 blood plasma samples from 55 adults and 36 children with Li-Fraumeni syndrome.
"Li-Fraumeni is a good proving ground for CHARM, due to the overall high risk of developing a variety of cancer types and multiple cancers over [the patients'] lifetime," Wong noted. "The most common cancers include breast, brain, bone, adrenal, and soft tissue sarcoma, and despite our advances in treatment and understanding of these cancers, early detection is often the best prognostic indicator for survival."
When the team searched for suspicious somatic mutations in a subset of 68 participants, for example, it found somatic changes to TP53 or other cancer genes in 24 percent of the Li-Fraumeni patients tested, including individuals with or without cancer diagnoses. Somatic changes were not found in all of the cancer-positive patients, though, indicating that the sensitivity of the approach depends on the input DNA available in the samples.
In their follow-up analyses of DNA fragment sizes in three of the individuals, the researchers found that somatic mutations in TP53 tended to show up on scraps of DNA that were smaller than the DNA fragments harboring germline TP53 mutations, helping them to distinguish between tumor and germline sources of these alterations.
Based on an analysis of sequential samples, the team also described a metastatic bladder cancer case that appeared to involve multiple tumor clones with distinct TP53 mutations, where two different somatic mutations were detected in the gene at multiple time points.
In other cases, blood plasma profiles from cancer patients pointed to the presence of somatic mutations in non-TP53 genes following treatment, Wong reported, while samples collected over time from seemingly cancer-free patients revealed somatic mutations that preceded cancer diagnoses or recurrence.
Because informative mutations identified by panel gene sequencing were often found at low levels, near the cutoff for variant allele frequencies in the clinic, the investigators searched for additional blood-based clues to cancer development, including information based on DNA fragment size and other DNA fragment features, which can differ between fragments released from tumor or normal cells.
"One emerging technology is fragmentomics, which is the analysis of fragment length, position, and coverage in cell-free DNA to determine the contributing sources," Wong explained, noting that "the DNA fragments released into the plasma [are] dependent on the cell of origin, and tumors in general release shorter fragments compared to healthy cells."
The team's results suggested that DNA fragments in Li-Fraumeni syndrome patients are somewhat shorter compared to Li-Fraumeni-free controls, and shorter still in Li-Fraumeni syndrome patients with cancer.
Those patterns appeared to be particularly pronounced in parts of the genome affected by somatic copy number changes in analyses of a prostate cancer case, Wong reported, though enhanced DNA fragmentation was also detected at some sites lacking copy number changes.
While machine learning analyses on DNA fragment data alone offered some cancer clues in the Li-Fraumeni syndrome cohort, the combination of somatic gene mutations, fragment features, and DNA methylation is expected to complement existing screening strategies.
"One challenge that we face in early detection is that we're often limited by the low tumor burden or signal," Wong said. "However, integration of independent biological analyses, such as the genome, fragmentome, and methylome can overcome these challenges."
To that end, he explained, the investigators are developing an integrated classifier tool that takes cell-free DNA mutation, copy number, fragment, and methylation features as well as germline variant profiles, clinical history, and family history into account. They are also working to establish a related clinical trial that would involve returning results to participating patients.