The Molecules That Stop Our Immune System from Attacking Food

Lunch arrives: a bread roll, a bowl of edamame, a corn tortilla on the side. None of it is you. Every mouthful is structurally foreign, a procession of proteins the immune system has never made and cannot claim as self. And yet it lets them through. No alarm, no inflammation, no rejection: just digestion, quiet and unremarkable. How the body makes that particular peace with the outside world has puzzled immunologists for decades. Now, for the first time, researchers have identified the specific molecular fragments that broker that truce.

The study, published in Science Immunology on 6 March, is the work of Jamie Blum at Stanford University (now at the Salk Institute) and colleagues including Elizabeth Sattely, also at Stanford. What they found, buried inside ordinary mouse chow, are three small protein segments from corn, soy, and wheat that the immune system actively learns to ignore, or rather, learns to welcome.

Oral tolerance is the formal name for this process. It’s the reason most people can eat peanut butter or scrambled eggs without incident, and it operates through a class of immune cells called regulatory T cells. These cells are, in a sense, the peacekeepers: when they recognise a protein, they don’t trigger inflammation but actively suppress it, keeping the immune system from treating food as a threat. What nobody had managed to work out, until now, was which specific proteins were prompting that response in real food eaten under real conditions.

Not an abstract puzzle. Food allergies affect roughly 6 per cent of young children and 3 to 4 per cent of adults, and the therapeutic goal (reprogramming an overreactive immune system to stand down) has been well-articulated for years. The problem was molecular. You cannot instruct regulatory T cells to tolerate a food if you don’t know what they’re already recognising when tolerance is working normally.

Blum and colleagues started not with a targeted hypothesis but with a bowl of standard mouse chow. They generated 128 separate hybridoma cell lines, each carrying a different T cell receptor harvested from intestinal regulatory T cells, and screened them against each of the seven protein-containing ingredients in the feed: wheat, corn, oat, fish meal, soybean, alfalfa, and yeast. Seven of those 128 receptors responded to food proteins. Five lit up to corn, one to soy, one to wheat. All three proteins turn out to belong to the same biochemical family: seed storage proteins, the dense, water-insoluble reserves that plants pack into their seeds to fuel germination.

“As someone interested in foundational science, there’s value in understanding a normal immune process along with pathology,” says Blum, who completed the work at Stanford before joining Salk as an assistant professor. “Understanding how the immune system can normally see a protein as safe may lead to new therapies to promote tolerance in individuals with allergy.”

The corn epitope proved the most striking. The five corn-responsive receptors all converged on a single eleven-amino acid sequence at the tail end of a protein called alpha-zein, one of the prolamin family, water-insoluble, highly abundant in maize seeds. Using a specially built molecular probe called an MHC tetramer, the team traced these T cells through the gut and found that, in a healthy adult mouse eating normal chow, zein-specific regulatory T cells account for up to 2 per cent of the total regulatory T cell pool in the small intestine. That’s a substantial proportion to be dedicated to a single food protein, and it probably isn’t coincidental that corn is not a common allergen. The soy epitope, by contrast, came from a protein called glycinin G1, which is itself a known allergen. Crucially, though, the tolerogenic epitope the team identified sits in a completely different part of the protein from the antibody-binding sites already linked to soy allergy: an orthogonal target, the paper notes, that may explain why some people tolerate soy while others don’t.

The cross-reactivity picture is, if anything, more striking. The glycinin receptor doesn’t just respond to soy; when the team tested 26 different plant seed lysates against it, they found it activated by sesame, quinoa, pecan, and a range of others that share a closely related protein family called the 11S globulins. This may partly explain a poorly understood phenomenon: why eating one food can seem to protect against reactions to another. Some of that cross-tolerance, it turns out, may be molecular in origin, not merely dietary coincidence.

Timing matters too. The zein-specific T cells don’t appear at birth. They emerge around weaning, when solid food first enters the gut, suggesting the critical window for establishing these tolerogenic populations opens with the first encounter with protein in the context of a normal food matrix. Germ-free mice (animals raised without gut bacteria) had markedly fewer of these cells, revealing a dependency on the microbiome that adds another layer to an already complicated picture.

What this research does not do is explain why tolerance fails. The team found the peacekeepers; the conditions that prevent them from forming, or overwhelm them, remain an open question. Still, the tools they built in the process are now in others’ hands. The tetramer reagents used to track zein-specific T cells are available through the NIH Tetramer Core Facility, and the workflow used to map these epitopes is translatable to human tissue.

“Diet is our most intimate interaction with our environment,” says Blum. “Correctly recognizing foods as safe creates an anti-inflammatory environment to support nutrient acquisition and prevent allergy. Our research advances scientific understanding of the major dietary allergens, and points us toward future therapeutic interventions that could redirect allergic and autoimmune states.”

Whether any of those interventions eventually make it to a clinic depends on how well the mouse data maps onto human immunology. The team has already found evidence in human serum samples that the immune system does recognise these grain proteins, suggesting the underlying biology may well translate. The next step is doing for humans what they did for the mouse chow bowl: working systematically through the proteins, matching T cell receptors to their targets, and building a more complete molecular atlas of what it means for the body to decide that lunch is safe.

DOI / Source: https://doi.org/10.1126/sciimmunol.aeb4684

Frequently Asked Questions

Why do some people develop food allergies when most of us don’t? The short answer is that tolerance and allergy use some of the same proteins but in opposite ways. The new research found that the tolerogenic epitopes in soy, for instance, sit in a completely different part of the protein from the sites that trigger allergic antibody responses, meaning it may be possible to tip the same protein toward tolerance or toward allergy depending on which immune cells encounter which fragments, and under what conditions. What determines that fork is still not well understood, but identifying the tolerogenic side of the equation is a necessary first step.

Could these findings lead to new treatments for food allergies? Possibly, though not soon. Regulatory T cells have been a target for allergy immunotherapy for years, but one obstacle has been not knowing which specific protein fragments to use to programme them. Identifying the naturally occurring tolerogenic epitopes from corn, soy, and wheat gives researchers a more precise starting point. The team also built molecular tools, tetramer reagents, now available to other labs, which should accelerate the work. Translation to human therapies still requires demonstrating the same epitope targets exist and function similarly in human gut tissue.

Why are seed storage proteins specifically the ones involved in tolerance? The paper offers a suggestive explanation: the tolerogenic proteins identified are all water-insoluble, whereas proteins known to drive food allergy tend to be water-soluble. Insoluble proteins may be processed and presented differently by gut immune cells, potentially favouring the regulatory response over the inflammatory one. It’s a hypothesis at this stage, not a confirmed mechanism, but it has direct practical implications for understanding why some foods cause disproportionately high allergy rates while others almost never do.

Does eating one food actually protect against reactions to a related one? It might, in some cases, and there’s now a candidate molecular explanation. The soy-responsive T cell receptor the team characterised also responds to sesame, quinoa, pecan, and other plants sharing a related protein family. If those cross-reactive tolerance signals are established during early food introduction, they could plausibly reduce sensitivity to related proteins encountered later. Whether that holds in humans at clinically meaningful levels is not yet known, but the cross-reactivity pattern is real and wider than expected.