Skip to main content
Premium Trial:

Request an Annual Quote

Researchers ID Molecular Subtypes of Gastric Cancer Predictive of Survival and Drug Response


By Molika Ashford

Gastric cancer can be divided into two molecular subtypes that are associated with differing levels of patient survival and response to standard chemotherapy treatments, researchers from several research institutes in Singapore reported this week.

In the study, published this Monday in Gastroenterology, researchers from the National University of Singapore and elsewhere claimed to be the first to show that a molecular classification of gastric cancer can identify subtypes that respond differently different therapies, a crucial step in the push to personalize cancer treatment.

The researchers, led by Patrick Tan of the Genome Institute of Singapore, identified two intrinsic genomic subtypes of gastric cancer. They found that cell lines belonging to one subtype were significantly more sensitive to the drugs 5-fluorouracil and oxaliplatin, while cell lines belonging to the other subtype were more sensitive to cisplatin.

Tan wrote in an e-mail to PGx Reporter that this is not the first genomic analysis of gastric cancer. It has long been known that gastric cancers are histologically heterogeneous, and several cellular, genomic, and proteomic classifications have been made previously in this setting. However, according to Tan, no previous genomic classification systems have been able to provide reliable independent prognostic prediction or an association with specific treatment options.

"Most previous studies have been limited by analyses of small data sets and focused on analyzing primary tumors that are very complex entities," Tan wrote. "Here, we hypothesized that adopting a different approach, starting our analysis with a panel of gastric cancer cell lines … we might be able to derive cleaner molecular signatures that can then be extrapolated to primary cancers."

Tan wrote that the researchers took a three-step approach. "First, we analyzed the expression profiles of a large panel of gastric cancer cell lines… to identify highly distinct patterns of gene expression," he said. "This … revealed the existence of two very distinct subgroups of gastric cancer cell lines, which could be clearly distinguished by a gene expression signature of about 170 genes."

This initial discovery step was performed on 37 cell lines obtained from both commercial sources and various research centers. Profiling the lines using the Affymetrix Human Genome U133 plus 2.0 GeneChip, the group found 171 genes differentially expressed in the two subgroups, labeled genomic intestinal (G-INT) and genomic diffuse (G-DIF).

Then the group applied the signature to primary tumor samples from four independent patient cohorts representing 521 total cancer patients. The first two cohorts, totaling 270 patients, were used to confirm the ability of the gene signature to classify patients into one of the two subgroups.

Next the researchers investigated the concordance between their molecular subtypes and clinical and pathological features. The molecular classification lined up "significantly," though not entirely with Lauren's classification, a microscopic classification used to divide tumors into two types, diffuse vs. intestinal, based on cellular features. The group named their molecular subtypes with reference to these clinical features, though the overall association between the two classifications was only 64 percent, they reported.

The molecular subtypes were then examined for association with survival outcomes. One of the intrinsic subtypes, G-INT, was associated with significantly improved patient survival, Tan said, which marked the signature as a possible prognostic indicator. Classification according to Lauren's subgroups, meanwhile, was not prognostic, the authors wrote.

In multivariate analysis, the molecular subtypes remained independently prognostic, and the results were replicated in a third cohort of 65 patients using a different platform, the Illumina Human-6 v2 Expression BeadChip.

The group tested two genes from their discovered signature, LGALS4 and CDH17, for which immunohistochemical markers were available, as a possible immunohistocehmical signature for the G-INT subtype. This approach was also able to divide the cohort into different prognostic groups. However, the authors wrote, it is too early to conclude that these two genes represent the best markers for distinguishing the two subtypes.

Lastly, the researchers assessed the responses of the two subtypes to three chemotherapies currently used in adjuvant and first-line palliative treatment of gastric cancer: 5-fluorouracil, cisplatin, and oxaliplatin. Cisplatin is marketed by Bristol Myers Squibb as Platinol, oxaliplatin by Sanofi-Aventis as Eloxatin, and 5-FU is sold as Adrucil in the US by Teva.

The researchers treated 28 of the 37 cell lines in in vitro cell proliferation assays with increasing concentrations of the three chemotherapeutics.

"We found that cell lines belonging to the G-INT subtype were significantly more sensitive to 5-FU and oxaliplatin [in vitro], while cell lines belonging to the other subtype … were more sensitive to cisplatin," Tan said. "In the case of 5-FU, we were also able to clinically validate this association by examining the outcomes of gastric cancer patients specifically treated with 5-FU adjuvant chemoradiation."

The authors pursued their cell line-based discovery approach because they hypothesized that the simplified system, lacking vasculature and immune cells, would be more likely to reflect intrinsic molecular differences. However, Tan enumerated several downsides of this approach.

The researchers write in their paper that because the analysis was focused on only 37 cell lines, some of the rarer subtypes of gastric cancer may have been missed. Also, the study could not assess the full tumor microenvironment, which can also influence prognosis and response. "No approach is perfect," Tan said. "Nevertheless, we believe that the intrinsic subtype classification provided in this paper is a good foundation upon which other levels of complexity can be subsequently added."

According to Tan, the researchers didn't expect to discover two very distinct subtypes in the cell line panel. "We certainly expected to see subtypes, and were prepared to uncover these subtypes using more complex analyses, but this was not needed," he said, stressing that this ease of observation supports the conclusion that the group's molecular classification reflects two "very different versions of gastric cancer."

Reflecting on the varying responses seen in cell lines to chemotherapies, Tan said that this finding "was important as it is the first time that a proposed molecular classification of gastric cancer could be used to guide chemotherapy selection."

The authors stress in their paper that these findings should be considered early discovery results and require further validation.

The group is currently conducting retrospective validation of the recently published findings in larger patient cohorts and the researchers have launched a prospective phase II clinical trial where treatments will be allocated to either S1/oxaliplatin or S1/cisplatin treatment on the basis of tumor types as characterized by the group's genomic classifier.

Tan said no pharmaceutical companies were involved with this initial study, but he said the researchers are "in active discussions with several companies who are interested in applying their drugs to the cell lines to see if their drugs have subtype specificity" in advanced gastric cancer.

He said the group is interested in the possibility of developing a diagnostic classifier based on the findings that would help guide treatment choices. "However," he said, "these are still early days, and we need to generate more supporting data."

Have topics you'd like to see covered in Pharmacogenomics Reporter? Contact the editor at mashford [at] genomeweb [.] com.

The Scan

Mosquitos Genetically Modified to Prevent Malaria Spread

A gene drive approach could be used to render mosquitos unable to spread malaria, researchers report in Science Advances.

Gut Microbiomes Allow Bears to Grow to Similar Sizes Despite Differing Diets

Researchers in Scientific Reports find that the makeup of brown bears' gut microbiomes allows them to reach similar sizes even when feasting on different foods.

Finding Safe Harbor in the Human Genome

In Genome Biology, researchers present a new approach to identify genomic safe harbors where transgenes can be expressed without affecting host cell function.

New Data Point to Nuanced Relationship Between Major Depression, Bipolar Disorder

Lund University researchers in JAMA Psychiatry uncover overlapping genetic liabilities for major depression and bipolar disorder.