PHILADELPHIA— Type 1 diabetes (T1D) impacts nearly two million Americans, and by the time most people learn they have it, most of the body's insulin-producing cells are destroyed. Now, pinpointing a hidden group of immune "attack" cells in the pancreatic lymph nodes that appear earlier in the disease could offer the first real chance to detect – and even stop- T1D, according to new research from the Perelman School of Medicine at the University of Pennsylvania detailed in Science Immunology .
"For the first time, this research has caught the attack cells in the act, while the disease is still unfolding; we're not just seeing the wreckage after the immune system destroys insulin producing cells in the pancreas," Golnaz Vahedi, PhD , a professor of Genetics and co-corresponding author of the study.
Two proteins act like master switches in the pancreas
Analyzing nearly one million immune cells -- one cell at a time -- from the pancreatic lymph nodes and spleens of 43 organ donors, some with T1D, some showing early warning signs, and some healthy, researchers identified a unique subset of CD4 T cells: a type of "helper" immune cells in the pancreatic lymph nodes of people with active T1D. These cells ramp up two proteins, NFKB1 and BACH2, that act like master switches, turning genes on and off in ways that rev up the immune attack on insulin-making cells.
"The study showed the same cell pattern occurred in pre-type 1-diabetic people, who don't yet show the symptoms. This suggests that the immune misfire begins early, potentially while plenty of insulin-producing beta cells are still healthy," said Vahedi.
Blood-based clues in the spleen
In the spleen, certain white blood cells or B cells also showed T1D-specific molecular changes, and these signals could be spotted in simple blood samples. This hints at a future blood test to flag T1D risk years before symptoms like high blood sugar appear. "The spleen changes detectable in blood mean we could monitor at-risk children, like family members of type 1 diabetes patients, without invasive procedures," said Vahedi. "If we can block the pathways that fuel these rogue CD4 T cells, we might be able to delay or even prevent Type 1 diabetes."
The greatest gift fuels research
The research team studied pancreas and lymph node tissues generously donated by deceased organ donors and their families. "Gift of Life Donor Program personnel, Penn transplant surgeons, procurement teams, lab staff, and most of all organ donors made this research possible," explained Vahedi. The team has studied pancreas and lymph node tissues of over 200 organ donors so far, generating data other researchers can access free via PANC-DB , a public database run with partners at Vanderbilt University, University of Florida, and Stanford University. "Every dataset represents countless late nights and a donor's gift," said Robert Faryabi, PhD , an associate professor of Pathology and Laboratory Medicine and co-corresponding author of the study. "It's collaborative science at its best—surgeons, scientists, families, and funders making the impossible possible."
Using AI to map TD1
The research is conducted as part of the Human Pancreas Analysis Program (HPAP), co-directed by Ali Naji, MD, PhD , the Jonathan E. Rhoads Professor of Surgical Science II and Klaus Kaestner, PhD , the Thomas and Evelyn Suor Butterworth Professor in Genetics. The program launched in 2016 with funding from the National Institutes of Health (NIH) to conduct careful study of the pancreas to figure out exactly what goes wrong inside the organ before and during type 1 and type 2 diabetes. The program recently received funding for another four years . Vahedi and researchers now hope to develop AI models that can not only help detect T1D sooner, but map the disease at a molecular level "Our goal is to teach AI the molecular language of T1D—training it on the pathogenic cells in pancreatic lymph nodes so it can detect their faint traces in blood, even when they're like a needle in a haystack," Vahedi said.
The study was funded by the NIH (U01 DK112217, U01DK123594, R01DK133453, R01CA248041, R01CA230800, NIHR01AI168240, R01AI168136, U01DK127768, U01DA052715).
