NEW YORK – A team led by researchers at Stanford University has developed a method for profiling metabolism at the single-cell level.
Described in a paper published this week in Nature Biotechnology, the approach measures proteins involved in various metabolic pathways as opposed to directly measuring the metabolites themselves, which are less amenable to single-cell analyses.
In the study, the researchers used the method to profile the metabolic state of single cytotoxic T cells in both healthy tissue and in samples from patients with colorectal cancer, finding that they could link metabolism to immune function and dysregulation.
In addition to providing insight into the metabolic status of immune cells under various conditions, the work also provides a more feasible route for single-cell metabolomics generally, said Sean Bendall, an assistant professor of pathology at Stanford and senior author on the study.
"In general, this is a whole new level of a kind of single-cell omics where we can now ask a whole set of questions around metabolism and general biology that we haven't been able to ask before," he said. "I think there will be a lot of applications within the immune system, but also there is a big opportunity to go outside the immune system and look at other tissues in the body."
Typically, metabolomics researchers use tools like spectroscopy or mass spectrometry to measure metabolite levels. However, these approaches are often not sensitive enough to detect changes in metabolite levels in single cells, Bendall said.
Additionally, the act of isolating and purifying single cells alters the metabolite levels of those cells, he said. "It's one of these Schrödinger's cat kind of problems where you are actually perturbing the system just by measuring it."
These factors led Bendall and his colleague to take a different approach to single-cell metabolomic profiling. Instead of looking at the metabolites, they looked at the expression of proteins involved in regulating different aspects of cell metabolism. Using Fluidigm's CyTOF technology, the researchers measured 110 proteins involved in eight areas of metabolism in a number of sample and tissue types. They benchmarked these measurements against traditional bulk-cell metabolomic profiling data, establishing that the protein measurements correlated well with the metabolite data.
"We had this hypothesis that we could get around a lot of the technical limitations of measuring the metabolites themselves," Bendall said. "And we did these experiments to show that if we measured the proteins, or you might call it the regulome, that it correlate with the cells' actual metabolic activity."
He said that the idea that the levels of proteins involved in metabolic regulation would reflect metabolite levels was not especially novel, but that it had not been concretely demonstrated experimentally.
"This wasn't like a 'eureka' moment, but it wasn't something that had been really well established and so we had to establish that first before we could convince people that they should even care about these measurements and how they related to cell identity and behavior," he said.
Having established the validity of the approach, which they named single-cell metabolic regulome profiling (scMEP), the researchers used it to examine how metabolism changes in single human T cells according to factors like cell activation state and tissue type. They also used the MIBI single-cell imaging technology developed by spatial proteomics firm IonPath, of which Bendall is a co-founder and board member, to look at how the location of T cells within a tissue correlated with their metabolism.
Bendall highlighted a number of findings he and his colleagues found interesting, among them the fact that different immune cell types could be consistently distinguished from one another by their metabolic profiles, including across different individuals and tissues.
"Not only do we know that different immune cells look different and have different activities, but the identities of those immune cells are also associated with really unique metabolic wiring," he said.
They also observed clear metabolic dysregulation of immune cells in cancerous tissue, Bendall said.
While cellular metabolic profiles were highly consistent across different healthy individuals, when the researchers looked at the colorectal cancer samples "we saw unique kinds of metabolic behaviors pop up," he said.
Specifically, he noted, they saw overrepresentation of metabolic pathways dependent on protein utilization along with expression of molecules like PD1 and CD39 that are associated with immune cell dysfunction.
"Just like in the healthy subject scenario where metabolism is linked to immune cell identity, there was a unique metabolic fingerprint also linked to a unique scenario of immune cell dysfunction," Bendall said.
Meanwhile, in their MIBI work examining the spatial distribution of immune cells within colorectal cancer tissue, the researchers found that different cell types in close proximity within that tissue showed very similar metabolic activity, suggesting, Bendall said, the existence of metabolic niches or microenvironments.
They also identified a population of cytotoxic T cells that only appeared in the tumor tissue that appeared to be highly dysfunctional from a metabolic standpoint. "They just had everything turned off,” he said. “Glycolysis was turned way down. The Krebs cycle was turned way down. They had a weird overabundance of molecules that are targeted mainly to protein metabolism."
These dysfunctional T cells also appeared disengaged from the tumor, Bendall noted. "Instead of being at the tumor-immune interface, which is typically where you would see activated cytotoxic T cells, they were kind of out in the [tissue] lost."
These cells also had high levels of expression of immuno-oncology targets like PD1 and CD39.
"It is strongly suggestive that these cells that companies are trying to target with things like checkpoint inhibitors… are also metabolically distinct and that metabolism is another avenue by which they could be targeted," Bendall said.
He said that he and his colleagues plan to continue to add targets to deepen the metabolic analysis enabled by the approach. He added that IonPath has received customer interest in the assays and that the company is working to put together assay panels that could be used for the kind of spatial metabolic profiling work demonstrated in the study.