Researchers at the Johns Hopkins Kimmel Cancer Center, three other cancer centers and the Johns Hopkins Bloomberg School of Public Health compiled a comprehensive genetic architecture atlas for mutant RAS genes in human cancers. Their four-year study of the RAS family — including the KRAS, NRAS and HRAS genes that are mutated in approximately one-third of all human cancers — found that the frequency of mutant RAS genes differs among various tumor types, age, sex and racial groups, and co-mutation patterns among RAS genes and other genes potentially can result in different clinical outcomes or identify new areas for therapeutic intervention.
The work, published Sept. 8 in Cancer Research, focused on analyses of targeted next-generation sequence data of more than 600,000 mutations from more than 66,000 tumors in 51 cancer types from the AACR (American Association for Cancer Research) Project GENIE (Genomics Evidence Neoplasia Information Exchange) registry, which pools next-generation sequence data from multiple academic institutions. Investigators from the Dana-Farber Cancer Institute in Boston, the Vanderbilt University Medical Center and the Vanderbilt-Ingram Cancer Center in Nashville, and Memorial Sloan Kettering Cancer Center in New York, also contributed to this work. The study results are publicly available online at ras-hallmarks.jhmi.edu/ and through the AACR Project GENIE registry.
This work generated a comprehensive atlas of both co-occurring and mutually exclusive mutations among RAS and other genes at an unprecedented resolution. The results have immediate implications in how clinicians might be able to select patients for combination targeted therapies, such as KRAS inhibitor drugs, and understand why certain patients may not respond to certain therapies, says senior study author Valsamo Anagnostou, M.D., Ph.D., director of the molecular oncology laboratory and the thoracic oncology biorepository at the Johns Hopkins Kimmel Cancer Center and an associate professor of oncology at the Johns Hopkins University School of Medicine.
When treating patients who have RAS-mutant tumors, clinicians must consider clinical outcomes and the tumor’s aggressiveness as well as co-mutations and patient characteristics such as gender, racial background and age, Anagnostou says. “Context matters,” she says. “Our study shows that you need to consider who the host is and what the genetic makeup is of the tumor, because RAS-mutant tumors with different co-mutations have completely different profiles and clinical behavior.”
During the study, researchers examined several characteristics of the RAS genes. They first looked at the distribution and heterogeneity, or variations, of mutant RAS across cancer types and co-occurring mutations. Led by Robert Scharpf, Ph.D., associate professor of oncology at the Johns Hopkins Kimmel Cancer Center, the team developed novel analytical frameworks to assess the prevalence and co-mutation patterns of RAS genes in the AACR Project GENIE registry. They studied the cancer type-specific prevalence of KRAS, NRAS and HRAS mutant alleles (alternative forms of a gene) at codons (units of genetic code) 12, 13 and 61 in the overall population and stratified their results by patient age, race and gender.
Prevalence of the RAS mutations varied among cancer types — 74% in pancreatic cancers, 43.5% in colorectal cancers, 29.7% in non-small cell lung cancers, 25.3% in melanoma, 20.9% in cancer of unknown primary origin, 5.9% in precancerous blood and bone marrow diseases (myelodysplastic/myeloproliferative syndrome), and 1.5% in central nervous system tumors. Mutations were less common in prostate, breast and kidney cancer and mesothelioma, with mutation rates affecting about 1% of individuals.
KRAS mutations occurred at higher frequency in gastrointestinal tumors, lung cancers and gynecologic malignancies, while NRAS was more frequently mutated in melanoma, thyroid cancer and hematologic malignancies. HRAS was overall less frequently mutated. Diving deeper, the investigators found that non-small cell lung cancers primarily harbored KRAS G12C mutations, whereas these mutations were harbored in about 10% of colorectal cancers and 1% of pancreatic cancers.
Mutations were found at different frequencies depending on patient age, sex and gender. For example, RAS mutations were less prevalent among younger patients with melanoma (11.8% lower), cancer of unknown primary origin (12.5% lower), non-small cell lung cancer (11% lower) and pancreatic cancer (14.6% lower), but more prevalent in younger patients with ovarian cancer (15.8% higher) and B-lymphoblastic leukemia/lymphoma (8.2% higher) compared to the overall population. Sex-based differences also were seen among RAS mutations, with colorectal cancers (3.6% higher) and non-small cell lung cancer tumors (2.6% higher) in women more frequently harboring RAS mutations compared to men. By contrast, melanoma tumors in women had fewer RAS mutations (3.3% lower) than in men.
Race differences, too, were seen among the mutations. For example, Black individuals with colorectal cancers harbored a higher number of KRAS codon 12 mutations (6.5% higher) such as G12V, and KRAS codon 13 mutations (4.4% higher) such as G13D, than people in other racial groups. Codons are a sequence of DNA that corresponds to a specific amino acid during protein synthesis. In non-small cell lung cancers, RAS mutations occurred less frequently among Asian patients (18.6% lower). RAS mutations also were less common among Black individuals with uterine cancers (9.6% lower).
To assess co-occurring mutations, the team modeled dependencies among RAS genes and other genes across cancer types. RAS hotspot mutations co-occurred with mutations in the genes ATM, KEAP1, MAX, NKX2-1, RBM10, STK11 and USF1 in non-small cell lung cancers, for example. The team also discovered some instances in which RAS mutations were less likely to co-occur with other gene mutations, such as with the BRAF or RNF43 genes in colorectal cancers.
We confirmed and validated co-occurring mutation that we knew about from previous studies, such as KRAS and STK11 in non-small cell lung cancers. However, we also found new co-mutations, such as KRAS and NTRK3, which is very important, because these can represent potential therapeutic targets and can lead us to combination treatments.”
Robert Scharpf, Ph.D., associate professor of oncology, Johns Hopkins Kimmel Cancer Center
The team also investigated the genomic landscapes of RAS-mutant tumors, looking at all KRAS, NRAS and HRAS mutant alleles and considering co-occurrence with non-RAS mutations. The researchers found some distinct patterns depending on cancer type and patient age, sex and race.
Additionally, they looked at how RAS co-mutations are linked with cancer hallmarks and expression profiles of the cancer cells and the tumor microenvironment (the cells in and around tumors), evaluating sequence data from 9,258 tumors from the Cancer Genome Atlas, a National Cancer Institute genomics database of over 20,000 primary cancers spanning 33 cancer types. These analyses showed distinct gene expression programs in RAS-mutant tumors with different co-mutations.
Finally, the team evaluated the association between the distinct genomic tracks of RAS mutant tumors and clinical outcomes. They assessed 10,217 tumors from the Cancer Genome Atlas and evaluated differences in overall survival of individuals with tumors harboring RAS co-mutations, with some notable findings. Patients with lung cancers harboring KRAS G12C mutations and KEAP1, NTRK3, PIK3CA or TP53 co-mutations, for example, had a significantly shorter overall survival compared to those without KRAS co-mutations. Notable differences in outcomes with immunotherapy were also found in tumors with different KRAS mutations. For example, patients with non-small cell lung cancer harboring KRAS G12C/PIK3CA co-mutations had shorter overall survival, but patients with non-small cell lung cancer harboring KRAS G12C and TP53 co-mutations attained longer overall survival.
The team plans to expand this research into a platform that could be used to generate a blueprint of co-mutations in all driver genes in human cancers, Anagnostou says.
Scientists who contributed to the work include Archana Balan, Jacob Fiksel, Christopher Cherry, Chenguang Wang, James White, Alexander Baras, Jordan Anaya, Blair Landon, Marta Majcherska-Agrawal and Paola Ghanem of Johns Hopkins. Additional authors are from AACR Project Genie, Amgen, the Dana-Farber Cancer Institute, the Vanderbilt University Medical Center, the Vanderbilt-Ingram Cancer Center and Memorial Sloan Kettering Cancer Center.
