A Milk-Allergy Mystery

The forecast for March 14, 2017 warned of a late-season nor’easter, but pediatric gastroenterologist Victoria Martin and her undergraduate student Hannah Seay had to hit the roads. Their mission: transporting a giant Styrofoam cooler of frozen stool samples from Massachusetts General Hospital (MGH) across the Charles River, so they could reach the Broad Institute before they thawed. Afraid to trust their carefully assembled samples from babies seeing pediatricians at Newton-Wellesley Hospital to the vagaries of public transport or ride-sharing services, they took Seay’s car and arrived safely at the Broad, where computational biologist Moran Yassour shepherded the cooler’s contents into the freezers, and on to further processing and analysis. “I felt like my life was in that cooler,” Martin recalled.

The samples were crucial to a scientific collaboration whose mission is making sense of a medical mystery: why do some infants develop a cow’s milk protein allergy called allergic proctocolitis (AP) in their first year of life? Under the auspices of the Broad’s Food Allergy Science Initiative (FASI), Martin, Yassour, and their colleagues are trying to find out.

Even though food allergy impacts more than 220 million people worldwide, scientists still don’t understand the life-threatening condition’s basic underlying biology. FASI was founded to take on this challenge, bringing together experts from diverse fields to crack open the food allergy field. The initiative is a philanthropic partnership between parents of children with food allergy and scientists committed to making discoveries to improve diagnosis and treatment options and save sufferers’ lives.

In the case of the AP study, FASI has both built on an existing collaboration and brought new groups of scientists together. Five years ago, Wayne Shreffler, director of the Food Allergy Center at MGH and now a member of FASI’s scientific steering committee, and Qian Yuan, a pediatrician and gastroenterologist, came up with the idea of building a cohort to study why AP is so prevalent, and whether the seemingly benign treatment—abstaining from milk—might actually be harmful down the road. Soon thereafter, a landmark study revealed that lack of early peanut exposure could actually trigger severe peanut allergy later in childhood.

Martin, then a first-year fellow, joined Shreffler’s lab to oversee the day-to-day operations of the cohort: an effort to collect 10,000 samples from 1,000 healthy infants from birth to age three. The project became known as the Gastrointestinal Microbiome and Allergic Proctocolitis (GMAP) study.

Although the cohort is ongoing, the data GMAP has already collected point to some intriguing findings. About 17 percent of the children in the cohort develop AP—a much higher percentage than reported in the literature. And the best protection against AP seems to be diet that includes both breast milk and formula. But the researchers don’t yet understand why.

One factor on their minds from the outset was the role of the microbes in the infants’ guts. They had collected stool samples containing these microbes, but weren’t sure how to analyze them. That’s where Yassour came in. Separately, she had been contemplating the clinical applications of her research, which examined how babies acquire their gut microbiome. A chat with her former graduate advisor, FASI steering committee member Aviv Regev, convinced her to focus on allergy and led to a conversation with Shreffler—and their eventual collaboration.

Martin, Yassour, and their colleagues worked together to prioritize which samples to probe for their microbial contents. The combination of their different backgrounds was essential to determining what questions to ask and how best to answer them. “Not being a clinician, I’m not biased by what people have already shown—I came in with a very naïve perspective,” Yassour said.

After the team formed a plan, it was up to Martin and Seay to collect the samples from freezers at MGH for Yassour and her colleagues to study. Once the samples arrived safely, Yassour worked with a group at the Broad called the Microbiome Omics Core (MOC).

One of their first tasks was figuring out how to use as little of the precious material as possible in characterizing each sample’s microbial species diversity. That involved chipping more than 1,000 of the samples on a sterile, cooled metal surface.

Then the MOC began processing the batch: 11 plates of 96 samples each, the largest collection it had handled to date. They extracted the DNA to profile the microbial community composition and how it changes over time. Martin, Yassour, and colleagues are in the midst of analyzing the sequencing data, which should provide clues about the role microbes play in governing AP.

Though their work is in progress, the researchers are already thinking about their next steps—and topics they could explore in the future. One intriguing possibility is keeping track of the children with AP to see if they develop other food allergies—a link that is currently unclear. The planning for GMAP 2.0 is already underway, and with it, the chance to discover even more about babies, their microbes, and how we process foods in our earliest stages of life.