NEW YORK (360Dx) – Plasma levels of the metabolite methylthioadenosine might be able to distinguish sepsis patients who will survive the condition from those who will not.
Researchers from the US and UK used a combination of genetic and metabolomics analyses to find that methylthioadenosine (MTA) — part of the methionine salvage pathway — acts as a prognostic biomarker for sepsis. As they reported yesterday in Science Advances, a machine-learning model that combined MTA with other variables had an 80 percent accuracy in predicting death due to sepsis.
Sepsis, the dysregulation of host immune system in response to an infection, can be triggered by any sort of pathogen, and is the cause of death for a quarter of a million people in the US every year. It's tricky to both diagnose sepsis and gauge the severity of patients' outcomes, said Dennis Ko, an assistant professor of molecular genetics and microbiology at Duke University School of Medicine, because it can have different causes and varying effects on patients.
"So things are needed to make a more rapid diagnosis, and things are needed in terms of parsing out the complexity of all the different people who have sepsis," Ko said in an interview.
In a 2012 paper appearing in the Proceedings of the National Academy of Sciences, Ko and his colleagues conducted a cell-based genome-wide association study that searched for SNPs associated with increased pyrotosis, the pro-inflammatory cell death pathway, in response to a pathogen. While pyrotosis can destroy pathogens, it can also cause host damage if unconstrained.
Through this, they linked a SNP in the gene APIP to increased Salmonella-induced pyrotosis. The researchers also genotyped a cohort of systemic inflammatory response syndrome patients from Seattle to find that this SNP was associated with improved survival.
This led Ko and his colleagues in the new study to conduct a GWAS in Kenyan children with and without nontyphoidal Salmonella bacteremia. Again, they found a signal near the APIP gene that was associated with nontyphoidal Salmonella bacteremia. The SNP associated with improved survival in the Seattle patients was also linked to reduced odds of nontyphoidal Salmonella bacteremia in the Kenyan children. This, the researchers noted, confirms the role of APIP in the regulation of sepsis and bacteremia.
It also implicated the methionine salvage pathway, of which APIP is a part. A pathway analysis uncovered hits not just in APIP, but also in other components of the network, indicating that the whole pathway is involved in sepsis regulation.
This led Ko and his colleagues to shift gears.
"So we thought, OK, there's several of these SNPs here that appear to be regulating bacteremia. Maybe instead of looking at the genetics, now is when we should look at what they regulate," Ko said. "What's the metabolite that goes into this pathway?"
They then focused on MTA, which the methionine salvage pathway scoops up to convert to methionine. In a cohort of sepsis patients that were already undergoing metabolomics profiling at Duke and elsewhere, the researchers examined MTA levels to find that patients who survived had lower MTA levels than those who did not. In the Duke cohort, this association was strongest directly after the apparent onset of infection.
In the Duke cohort, MTA had a prognostic capability similar to that of the APACHE II score, which is calculated from a number of physiological parameters, they added. Within a day of suspected infection, MTA alone had a predictive power of 0.74 and, within two to three days, it had a predictive power of 0.79. A model combining APACHE II, MTA, and age boosted that predictive power slightly to 0.81.
"The potential would be that, OK, we can measure just a single metabolite and it's elevated in people who have sepsis who are more likely to die," Ko said. "So it can help clinicians figure out who is in the high-risk group here."
Sepsis patients with elevated MTA levels also had elevated levels of inflammation markers, the researchers found. Ko noted that there are two extremes of sepsis: one in which there is hyper-inflammation that leads to organ damage and, possibly, death, and one in which there is an insufficient immune response that leads to secondary infections and then to bad outcomes.
Because of this link to inflammation markers, he said that he and his colleagues think that MTA is identifying that subset of sepsis patients who have a hyper-inflammatory response.
This also suggested to Ko that MTA itself could be a possible therapeutic. In an early-stage mouse study, he and his colleagues dosed mice with MTA half an hour before infecting them or six hours after infecting them with Salmonella. Mice given MTA prior to infection had prolonged survival, they reported. Their levels of the inflammatory cytokines TNF-alpha and IL-6 were also reduced.
Ko noted that he felt there was stronger evidence to try this mouse study because he and his colleagues had collected both genetic and metabolite data. "If you just have the metabolite, you don't know if that's a cause of the disease or a consequence of the disease," he added. "But because we had both the genetics and the metabolite, I thought it was worth trying this in mice."
In addition to its possible use to identify patients at increased risk of death from sepsis, Ko said that being able to stratify patients could also inform drug development efforts. He noted that a number of sepsis drugs that failed clinical trials could be re-evaluated in different sepsis patient groups.
"Maybe that would tell us that, OK, for example, this therapeutic didn't work in all comers, but it does have an effect in those with high levels who have this pro-inflammatory response going on," Ko said.
He added that his lab is hoping to partner with pharmaceutical companies to do just that.