How Does the Gut Immune System Distinguish Between Friends and Foes?

From cholera to norovirus, Giardia to hookworms, the gut is vulnerable to an impressive array of pathogens. To keep us alive in the face of this onslaught, the intestinal immune system must be constantly on the lookout for antigens that signal danger. However, the gut is also chock-full of other antigens that originate from the foods that we consume, posing a dilemma for these surveillance systems.

“It’s a simple but intriguing question: How do we survive eating? How do we not create an inflammatory response to food?” asked Maria Canesso, a postdoctoral researcher who studies mucosal immunology at the Rockefeller University.

In the intestinal immune system, specialized antigen-presenting cells (APCs) gather antigens from both food and pathogens in the gut contents and present them to T cells, causing them to differentiate into regulatory T cells (Tregs) or inflammatory T cells. In a study published in Science, Canesso and her colleagues used a recently-developed labeling technique to identify APC subtypes that present food antigens and aid in Treg induction.¹ They also showed that certain kinds of parasitic worm infections alter the balance of APCs that promote food tolerance and those that trigger inflammation, disrupting the development of tolerance to food antigens.

“The intestine is always seeing many different things, and the same thing can be seen in a different way depending on the context,” said Canesso. “It’s important to take this into account, especially when we’re thinking about the development of food allergies.”

Before Canesso and her team explored how these immune processes became dysfunctional, they wanted to understand how the system functioned under normal conditions. To accomplish this, they analyzed the APC-T cell interactions that occurred during food antigen exposure—in this case, the egg white protein ovalbumin (OVA)—in healthy mice using a technique called LIPSTIC (Labeling Immune Partnerships by SorTagging Intercellular Contacts). This technique involves engineering certain populations of T cells to express an enzyme that places a peptide marker on any cells they interact with.

In this experiment, the researchers expressed this enzyme in OVA-specific naïve T cells—T cells with receptors capable of recognizing the OVA antigen that had not yet “decided” whether they would become regulatory or inflammatory. These T cells would then mark any OVA-presenting APCs they connected with, enabling the scientists to isolate these APCs for further study.

Twenty-four hours after the mice received the OVA protein, the researchers collected the labeled APCs and analyzed them using single-cell mRNA sequencing, identifying two conventional dendritic cell subsets, cDC1s and cDC2s, in approximately equal numbers. When they cocultured both types of APCs with naïve T cells, they found that only the cDC1s increased the number of Tregs present in the culture, suggesting that this cell type is important for promoting tolerance to food antigens.

To explore the role of cDC1s in vivo, the researchers utilized a mouse model in which cDCs were unable to present antigens. They found that Treg numbers dropped by about 50 percent compared to mice with functional APCs. However, the fact that there were still Tregs present suggests that other cell types are also involved in this tolerization process.

“There’s been a lot of debate about, [the identity of] the tolerance-inducing antigen presenting cells,” noted Dan Littman, a molecular immunologist at New York University, who was not involved in the present study. While some research has suggested these cells are cDCs, more recent studies have called this into question.²

Littman said these experiments clearly show that cDC1s do ultimately influence the final proportion of OVA-specific Tregs. However, he added, “The function that they demonstrate is not necessarily an inducing function. It may be a role in expansion or maintenance of the Tregs. It could be some kind of positioning of the Tregs so that they can carry out their appropriate function.” Indeed, preprints from Littman’s laboratory and other research groups suggest that another subset of APCs that express the transcription factor RORγt are necessary for the induction of food antigen-specific Tregs, whereas cDC1s may be dispensible.^3,4

However, Littman is not discounting the relevance of this research, or of cDC1s themselves. “I think each of these [studies] adds a new layer of understanding of what’s happening,” he said.

Indeed, this study by Canesso and her colleagues supports the involvement of multiple APC types in the cascade of events leading to long-term tolerance of food antigens. At 24 hours after OVA administration, most of the LIPSTIC-labelled APCs—the APCs interacting with the OVA-specific T cells—were cDCs. However, at four hours, RORγt-expressing APCs made up a substantial portion of the cells marked with LIPSTIC. The authors proposed that these RORγt-expressing cells played an important part in tolerance induction immediately following food antigen exposure, while cDCs took on a larger role at slightly later time points.

After mapping out how this cellular crosstalk occurred during the development of tolerance to a novel food antigen, Canesso and her team set out to investigate what happens when this process breaks down. They had previously shown that mice infected with a specific type of intestinal parasite—the helminth Strongyloides venezuelensis—did not develop tolerance when they were exposed to novel food antigens.⁵

In the present study, they found that helminth-infected mice did not develop OVA-specific Tregs when they were exposed to this food protein. As a potential mechanism to explain this difference, they found that infection nearly abolished food antigen presentation by RORγt-expressing cells and cDC1s.

The researchers also analyzed gene expression in the unlabelled APCs—those that didn’t interact with the OVA-specific T cells—in the healthy mice and the infected mice. When they grouped the cells using these transcriptional profiles, they identified different subtypes of cDC2s present in the different groups of mice: Cells from infected mice contained a subset of inflammation-promoting cDC2s that were virtually absent in the healthy mice.

“This was surprising for us: You have tons of these other cells that induce an inflammatory response to the [parasite], but [they’re] not presenting dietary antigens,” said Canesso. “So, it's a compartmentalization of dietary versus pathogen-derived antigens by different antigen-presenting cells.”

Canesso said this line of research will help scientists understand why food allergies do or do not develop in different individuals and different contexts. Once researchers have characterized the cell types involved in tolerance versus inflammatory responses—the molecules they express, the signalling factors they release—they can begin to develop therapies to treat food allergies or even prevent them from emerging altogether.