NEW YORK (GenomeWeb) – An integrated molecular analysis of the stomach cancer gastric adenocarcinoma has prompted members of the Cancer Genome Atlas Research Network to put forth four molecular sub-types for the disease.
As they reported in Nature today, the researchers used a combination of exome sequencing, transcriptome sequencing, microRNA sequencing, and array-based copy number, methylation, and protein testing to assess samples from almost 300 individuals with gastric adenocarcinoma.
An analysis of the available sample data pointed to gastric cancer clusters characterized by Epstein-Barr virus positivity, microsatellite instability, chromosomal instability, or genomic stability.
The results are expected to offer a new framework not only for understanding the underlying biology of each gastric cancer sub-type, but also for placing patients in appropriate treatment trials and/or prognostic groups.
"Putting a tumor into a particular group definitely has value," the study's corresponding author Adam Bass, a medical oncology researcher at the Dana-Farber Cancer Institute, told GenomeWeb Daily News.
"That doesn't mean that every tumor within a particular class or group is identical," Bass added. "There's still variation between those and there are candidate targets that differ within that group."
Gastric cancers are widespread and often deadly. But while adenocarcinoma is regarded as the most common form of gastric cancer, efforts to further define the disease have proven difficult and competing classification schemes have yet to yield clinically informative subtypes.
"Even though there's been this understanding that there are differences between groups of gastric cancer, when it came time to think of how to actually develop clinical trials and approaches to treat the disease, it's very much been a one-size-fits-all model, where everyone has been lumped together," Bass explained.
For the current effort, the TCGA team focused on primary tumor samples collected from 295 individuals with gastric adenocarcinoma prior to treatment, along with matched normal blood or non-cancerous gastric mucosa tissue.
In addition to somatic copy number features, cytosine methylation profiles, and protein patterns gleaned from microarrays and reverse-phase protein arrays, respectively, the researchers generated whole-exome, messenger RNA, and miRNA sequences for tumor-normal gastric adenocarcinoma pairs.
Around 227 of the tumor-normal pairs were subjected to all six types of molecular profiling, Bass noted. Using low-coverage whole-genome sequencing, the team also got a glimpse at microsatellite stability traits in 107 of the matched tumor and normal samples.
Together, these molecular profiles delineated four forms of gastric adenocarcinoma. In the tumors showing high levels of Epstein-Barr virus, a known risk factor for gastric cancer, the team detected higher-than-usual promoter methylation levels, amplifications affecting a chromosome 9 region containing JAK2, PD-L1 (also called CD274), and PD-L2 (PDCD1LG2) genes, and an over-representation of PIK3CA gene mutations.
Alterations in PIK3CA are already being targeted for treatment in other cancer types, Bass said. In addition, he noted that PD-L1 and PD-L2 code for immune system down-regulators that are drawing interest from researchers focused on immune-based cancer therapies.
"These drugs really haven't been widely applied to gastric cancer yet," Bass explained. "But this points to a group of gastric cancers where these new, very promising agents should be prioritized for clinical trials."
Another gastric adenocarcinoma subtype — which often included tumors with a "diffuse" histological classification — was marked by genome stability and relatively low genome-wide mutation rates. When mutations and gene fusions did turn up in those genomically stable tumors, researchers reported, they tended to involve genes coding for proteins related to the function of the RHO-family GTPase signaling pathway.
"This opens up a new avenue of research into the basic biology of these tumors that we hope will ultimately be therapeutically relevant," Bass said. "But that new area needs a lot of basic investigation first."
In tumors with marked microsatellite instability, meanwhile, researchers saw a jump in overall mutation rates, particularly involving genes coding for components of oncogenic signaling pathways, and a rise in methylation in certain regions of the genome.
Aneuploidy was rampant in a fourth subtype that they classified based on its chromosomal instability, as was the focal amplification of genes coding for receptor tyrosine kinase enzymes.
Bass noted that several therapeutic agents targeting the receptor tyrosine kinases are already approved or in development, suggesting they may be useful against some tumors in that subtype.
Potential therapies based on such molecular targets still need to be tested in appropriate gastric cancer patients. But with the advent of better-defined gastric cancer subtypes, the study's authors argue that it should be possible to perform the sorts of clinical trials needed to evaluate such treatment possibilities.
"We have to think about the trials we're doing and how to move away from doing one-size-fits-all trials to doing more targeted testing of different drugs based on the profiles of the individual tumors," Bass said.
With an eye to streamlining tumor classification and treatment in the clinic, the team went on to establish a "decision tree" for slotting the available gastric adenocarcinomas into each of the four newly described sub-types based on key molecular features.
"What we tried to do was use this massive [set of] comprehensive data to identify key groups within gastric cancer," Bass said, "and then to find more simple features that marked those groups, so that it will be easier to try to classify patients in a real-world setting."