How the immune system learns which foods are safe, one corn protein at a time

Stanford University

Why can most people eat peanuts, wheat, and shellfish without incident while others face life-threatening reactions to the same foods? The standard explanation has long been that tolerance is simply the absence of allergy, the immune system ignoring what it eats. That explanation, it turns out, is wrong.

Tolerance is not passive. It is an active, targeted immune training program. And a new study from Stanford University, published in Science Immunology, has begun to map exactly how it works.

Peacekeepers with a specific target list

Elizabeth Sattely, associate professor of chemical engineering at Stanford, and her team discovered that a class of immune cells called regulatory T cells (Tregs) actively survey the foods we eat, searching for specific protein fragments called epitopes. When Tregs find these particular epitopes, they signal the immune system that the food is safe, dampening any inflammatory response that might otherwise produce an allergic reaction.

The surprise was how focused the mechanism is. Not all food proteins are treated equally. Tregs appear biased toward a small number of standout epitopes that preferentially trigger a calming regulatory response rather than an inflammatory one.

In the case of corn, the Tregs zeroed in on a single epitope that is part of zein, a protein in the starchy interior of the corn kernel. Co-first author Ryan Kong noted that given the enormous number of potential intestinal antigens a person encounters daily, it was striking to see such a targeted response.

Identifying the epitopes that matter

Co-first authors Jamie Blum and Kong, working with colleague Kazuki Nagashima, examined mouse chow for ingredients that overlap with human diets, focusing on corn, wheat, and soy. They used a combination of experiments and analyses to pinpoint which protein fragments were being presented to Treg cells in the intestines and which fragments preferentially stimulated the regulatory response rather than an inflammatory one.

The results suggest that the immune system learns oral tolerance from a limited set of molecular cues. This is not a broad survey of everything we eat. It is a targeted recognition program focused on specific peptides within specific food proteins.

Format and microbes both matter

One of the study's most interesting findings is that the development of zein-specific T cells depends on two factors: the physical format of the protein in the food and the intestinal microbial community. The same protein, presented in a different physical form or in the context of a different gut microbiome, might not trigger the same tolerogenic response.

Blum, who now leads her own lab at The Salk Institute, noted that the team is working to determine the exact biological mechanisms involved. Understanding why the gut microbiome matters for tolerance could explain why disruptions to gut bacteria, from antibiotics, diet changes, or illness, sometimes precede the development of food allergies.

Toward a tolerance map and maybe a vaccine

Sattely can envision several applications. A molecular map of tolerance-biased epitopes could guide treatment strategies that reduce existing food allergies by introducing the specific peptides that activate regulatory T cells. In patients who already have allergies, delivering the right epitopes might reprogram the immune system toward tolerance.

There is also the possibility of prevention. If the key epitopes can be identified for major food allergens, early-childhood exposures could be designed to guide allergy-prone children toward tolerance before allergies develop. Sattely described this as a potential tolerance "vaccine" for those at high risk.

The team plans to explore specific plant seed proteins, which make up a large portion of human dietary protein, and to synthesize versions with key epitopes disabled or removed to test how the immune system responds differently.

Still in mice, and the caveats that follow

All of this work has been demonstrated in laboratory mice. Mouse and human immune systems share fundamental architecture but differ in significant details, particularly in gut immune responses and microbiome composition. The specific epitopes that drive tolerance in mice may not be the same ones that matter in humans.

The study examined three food sources: corn, wheat, and soy. The major allergens that cause the most severe human reactions, including peanuts, tree nuts, shellfish, and eggs, were not studied. Whether the same epitope-focused tolerance mechanism applies to these foods remains an open question.

Translating mouse findings into human therapies for food allergies has historically been difficult. The immune environment of the human gut is influenced by factors including diet diversity, medication use, infection history, and genetic background that are hard to replicate in animal models.

Still, the conceptual advance is significant. If tolerance is an active process driven by specific molecular cues, then it is potentially something we can learn to control. That reframing, from tolerance as passive absence to tolerance as active program, opens a fundamentally different approach to food allergy treatment.