PerspectiveIn the Journals

Branched-chain amino acids linked to future pancreatic cancer diagnosis

Show Citation

October 28, 2014

Adults who had elevated plasma levels of branched-chain amino acids were at greater than twice the risk for pancreatic cancer, according to study results.

“The study demonstrates that increasing circulating branched-chain amino acids are a marker for future pancreatic cancer diagnosis,” Brian M. Wolpin, MD, MPH, assistant professor of medical oncology at Dana-Farber Cancer Institute, told HemOnc Today. “The association was particularly strong in the several years before diagnosis, leading us to hypothesize that early pancreatic tumors were impacting host metabolism to lead to elevated branched-chain amino acids.”

Brian Wolpin

Brian M. Wolpin

Wolpin and colleagues compared plasma metabolite samples from 454 patients with pancreatic cancer to samples from 908 matched controls who were enrolled on one of four prospective cohort studies. A median of 8.7 years passed between blood collection and pancreatic ductal adenocarcinoma diagnosis.

Overall, researchers identified 15 metabolites that were associated with a future pancreatic cancer diagnosis (P˂.05). Three circulating branched-chain amino acids —isoleucine, leucine and valine — met the researchers’ predefined significance threshold for an association with pancreatic cancer (P≤.0006).

Patients who were in the top quintile of branched-chain amino acid expression demonstrated more than twice the risk for pancreatic cancer compared with patients in the bottom quintile of expression (OR=2.13; 95% CI, 1.43-3.15). Researchers also noted the expression of isoleucine, leucine and valine were highly correlated.

Elevated branched-chain amino acid levels 2 to 5 years before diagnosis were associated with the strongest risk for pancreatic cancer, suggesting the presence of occult disease (OR=4.34; 95% CI, 1.82-10.35).

Stratified analyses indicated the association between branched-chain amino acid expression and pancreatic cancer persisted regardless of cohort, sex, smoking status, BMI and fasting status at blood collection (P≥.14).

Wolpin and colleagues sought to replicate this association in mice. They found plasma branched-chain amino acid expression was associated with KRAS-mutated early-stage pancreatic cancer, but not KRAS-mutated tumors in other tissues.

“[Our findings were] confirmed using two mouse models of pancreatic cancer, in which elevated branched-chain amino acids were seen with early tumors,” Wolpin said. “In further studies in mice, we discovered that branched-chain amino acids were coming for muscle wasting, which had developed before loss of weight or other outward signs of cancer.”

The tissue breakdown associated with increased branched-chain amino acid levels suggests that whole-body protein breakdown is an early sign of pancreatic cancer, researchers said.

“The primary implications are twofold,” Wolpin said. “One, better understanding the biology of early pancreatic ductal adenocarcinoma and the identified subclinical muscle wasting will enable further early detection markers to be developed. Circulating branched-amino acids on their own are not sufficiently sensitive or specific for population screening at this time. Two, if tumors are using the products liberated by muscle wasting to support their growth, interrupting this process could identify new treatments.”

Disclosure: The researchers report no relevant financial disclosures.

itj+ Infographic

itj+ Perspective

Jeanine Genkinger

Jeanine Genkinger

Pancreatic cancer is one of the most fatal cancers of all major cancers due to the lack of effective screening methods, few identifiable early symptoms and, thus, diagnosis at late stages. Worldwide, an estimated 232,306 cases of pancreatic tumors occur and 227,023 people die from the disease each year, according to data provided by GLOBOCAN. Data from the American Cancer Society indicate half of all individuals diagnosed with pancreatic cancer die within 6 months of diagnosis, and nearly 95% die within 5 years. Due to the significant morbidity and mortality of pancreatic cancer, it is critical to identify effective markers for early detection, and improvements in risk prediction for targeted screening are needed. 

Mayers and colleagues used four large prospective cohort studies in which blood samples were collected prior to cancer diagnosis to examine the association between circulating branched-chain amino acids and risk for pancreatic cancer in 454 pancreatic cancer cases and 908 controls.

They observed a higher risk (OR=2.13, 95% CI, 1.43-3.15) for pancreatic cancer in those individuals who had high plasma levels of isoleucine, leucine and valine combined compared with individuals who had lower circulating levels. 

Plasma levels of isoleucine, leucine and valine are modified by dietary intake, as well as by tissue use and breakdown of protein stores. These markers also have been associated with insulin resistance, diabetes and obesity, all of which are risk factors for pancreatic cancer. However, it appears that these markers work through other pathways besides obesity and diabetes, as the risk estimates were similar when they accounted for such risk factors (eg, BMI, diabetes). Their experiments on mouse models supported their findings in humans. In addition, they also were able to show that protein breakdown in pancreatic cancer is a process that occurs earlier than previously hypothesized.

This study, unlike prior studies, used blood samples collected prior to diagnosis, had a large number of cases, and examined this association both in humans and mouse models. This study has identified novel pathways that may contribute to pancreatic carcinogenesis. These markers may ultimately prove to be useful for risk prediction and stratification, as well as for preventive methods (eg, changes in dietary intake). However, due to the rarity of the disease and the high case fatality rate, we still need to improve treatment effectiveness and develop effective screening tests that can be applied in high-risk groups identified by better risk stratification models (Lennon AM. Cancer Res. 2014;74:3381-3389).

Jeanine Genkinger, PhD
Assistant professor of epidemiology
Mailman School of Public Health
Member, Pancreas Center
Columbia University Medical Center

Disclosure: Genkinger reports no relevant financial disclosures.