Skip to main content
Premium Trial:

Request an Annual Quote

Researchers Identify Five-Gene Metabolic Signature to Predict Pancreatic Cancer Patients' Survival

NEW YORK – Researchers from Zhongshan Hospital in Shanghai have developed a five-gene signature that may better characterize pancreatic cancer patients' survival and benefit from chemotherapy.

The research, published in Experimental Biology and Medicine earlier this month, analyzed pancreatic cancer patients' genetic data from several databases including the National Cancer Institute's The Cancer Genome Atlas (TCGA) and the National Center for Biotechnology Information's Gene Expression Omnibus (GEO).

The researchers analyzed genomic and clinical data from 168 pancreatic cancer samples from TCGA and 65 pancreatic cancer samples from the GEO. They then compared this data against normal pancreatic tissue samples from the Broad Institute's Genotype-Tissue Expression project and identified 109 differentially expressed metabolic genes associated with pancreatic tumors.

From those 109 genes, the researchers used statistical analysis to home in on five genes — B3GNT3, BCAT1, KYNU, LDHA, and TYMS — associated with survival. To validate the differential expression levels of these genes, the researchers used Thermo Fisher Scientific's Oncomine to sequence 25 Chinese pancreatic cancer patients at Zhongshan Hospital. While those data did validate the association between the genes and pancreatic cancer, the researchers noted that further validation is needed in patients from other ethnicities.

They also tested the signature's ability to distinguish high-risk and low-risk patients using an external dataset, called GSE62452, which included microarray gene-expression profiles of 69 pancreatic tumors and 61 normal pancreatic tissue samples. In this dataset, median overall survival for high-risk patients was 9.6 months, compared to 21.2 months among low-risk patients.

From there, they used data from 168 patients with clinicopathologic data in the TCGA cohort to identify other independent clinical features associated with survival and developed a prognostic nomogram that included the five-gene signature, patients' age, whether their cancer has spread to nearby lymph nodes, and chemotherapy. Similarly, the nomogram could separate patients into high- and low-risk prognostic groups, where patients deemed at higher risk "had markedly worse overall survival," and the researchers noted that the nomogram was better at predicting overall survival than any one feature in the algorithm.

To better understand the biological role of the genes in the risk signature, the researchers analyzed the tumor microenvironment of TCGA samples and found that pancreatic cancer patients with high risk scores also had weaker immune responses. The authors, led by first author Qiangda Chen, a researcher at Fudan University's Zhongshan Hospital, found that the expression levels of immunomodulatory molecules were also lower in the high-risk group, "further supporting the desertification of tumor immune microenvironment in patients with high risk scores," they wrote. "These results suggest that the patients in the high-risk group might be resistant to the immune checkpoint blockade."

Patients classified as high-risk by the five-gene signature also had greater tumor mutation burden and higher rates of oncogenic mutations in KRAS and TP53.

To determine if the signature could predict chemo response, the researchers measured drug potency using half-maximal inhibitory concentration levels, or IC50, of four chemotherapies: paclitaxel, cisplatin, erlotinib, and gemcitabine. They found that paclitaxel, cisplatin, and erlotinib had greater potency in high-risk patients, suggesting the high-risk group was more sensitive to these drugs. They also found that the prognosis of high-risk patients who did not receive chemo were poorer than those who did receive chemo.

"Taken together, this metabolic signature could assess the prognosis and response to chemotherapy for pancreatic cancer patients," Chen and colleagues concluded, further noting that the five genes in this signature should be further studied to explore their potential as molecular targets for treatment.