NEW YORK – By interrogating thousands of colorectal cancer (CRC) samples, members of two independent research teams have identified new and known genetic contributors to the disease along with molecular subgroups based on tumor alterations or gene expression patterns.
For the first of the studies, published in Nature on Wednesday, researchers at the University of Oxford, the Institute of Cancer Research, the University of Manchester, and other centers in Europe sequenced the genomes of 2,023 CRC tumor samples from 100,000 Genomes Project participants.
Their analyses uncovered a wide range of small insertions and deletions (indels), structural variants (SVs), and copy number alterations (CNAs), highlighting 250 apparent driver genes as well as recurrent alterations falling outside of protein-coding genes.
"[W]e provided … large and comprehensive analyses of the genomic landscape of more than 2,000 patients with CRC," corresponding and co-senior author Ian Tomlinson, a cancer researcher at the University of Oxford, and his colleagues explained. "In addition to providing a comprehensive set of mutations of all types, a principal strength of our study is the ability to detect uncommon features, as evidenced by the discovery of many new driver genes, including SNVs, small indels, SVs, and CNAs."
Along with clues to the pathways and processes that appeared to be altered during CRC development, the researchers explained, their findings pointed to molecular subgroups of CRC within the known microsatellite stable, microsatellite instability/mismatch repair-deficient, and DNA polymerase-deficient subtypes.
The collection included subgroups that were individually rare, as well as those expected to offer insights into potential treatment strategies. Some drivers may be targetable in the near term, Tomlinson explained in an email, while others may become targets in the future.
Findings from the study also pointed to potential environmental or lifestyle contributors to CRC — from the presence of a mutational signature linked to smoking or diet in some tumors to an apparent relationship between rectal cancer and a metabolite called colibactin that is produced by certain Escherichia coli strains.
The team also saw ties between genomic features in a tumor and its location or the age of the patient, Tomlinson noted, as well as a mutational signature that may offer insights into CRC etiology in individuals who develop the disease at a relatively young age.
"We anticipate that our work will fuel future studies," the study authors wrote, "including efforts to characterize putative driver genes, translational analyses, and multidisciplinary experiments to address specific questions in a focused fashion."
For the second study, also published in Nature, investigators at Uppsala University, the Chinese Academy of Sciences, BGI Research Shenzhen, and other centers in Sweden, China, and Norway focused on 1,063 CRC samples, including 120 tumor biopsies and 943 primary tumors obtained during surgery. They also profiled adjacent tissue samples for 541 participants and matched normal blood samples for 522 cases.
"This extensive dataset, coupled with a [five]-year complete clinical follow-up, provides a robust foundation for integrative analyses that link molecular findings with clinical outcomes at an unparalleled scale, which sets this study apart from most discovery genomics studies in cancer," co-senior and co-co-corresponding author Cong Lin, a specialized scientist affiliated with the BGI Genomics Institute, BGI Research, and the Chinese Academy of Sciences, said in an email.
Using whole-genome sequencing and RNA sequencing data for the tumors, along with clinical outcome information, the team was able to pick up mutations and gene expression features that coincide with CRC subgroups and treatment response patterns.
In the process, they tracked down 96 suspected driver genes, including two dozen genes that had not been linked to any cancer type in the past and nine genes that are newly implicated in CRC.
"Mutations in APC, KRAS, and TP53 are well-established contributors to CRC oncogenesis," Lin explained. "Our study supports these findings while adding insights into other, less studied genes."
The team's analyses also identified two pathway co-mutation patterns across the CRC samples, while highlighting driver mutations that tended to turn up earlier or later in the CRC development process.
The authors noted that their timing analyses "revealed that the vast majority of chromosomal losses are early events, whereas amplifications occur late." Mutations in the PIK3CA gene or loss of a chromosome 10 region containing PTEN tended to occur after indicated TP53 mutation events, they wrote.
Lin noted that such findings "offer clinical research strategies for the early detection of CRC and the development of targeted therapies, revealing important molecular changes associated with tumor invasion and metastasis in the later stages."
On the CRC subtype side, the investigators distinguished between MSI-instability tumors showing distinct hypoxia and immune cell infiltration features and focused in on gene expression signals that tracked with five prognostically informative molecular subgroups, dubbed "Colorectal Cancer Prognostic Subtypes" (CRPS).
Finally, the study provided its own peek at the mutational processes that can lead to CRC, including mutational signatures associated with colibactin exposure, hypoxia, mismatch repair deficiency, reactive oxygen species, ID18, and other processes, as well as signatures that appeared to be CRC-specific.
Together, Lin explained, the results suggested that "integrating genomic and transcriptomic data for molecular classification can enhance precision of prognostic stratification for patients, which is crucial for guiding personalized treatment for CRC."