Maybe the culprit in type 1 diabetes isn’t T cells gone bad.

Follow diabetes research and you start obsessing about beta cells—maybe a gram of cells buried across the pancreas that produce the insulin we need to live. Or stop producing it, in the case of type 1 diabetes.

These cells are heroic microbeasts. “The beta cell is a wonder of nature,” Bart Roep of the City of Hope National Medical Center told me during an interview for a Knowable story. “It’s the hardest-working cell in our body. Every second, each beta cell can make two thousand molecules of insulin; that’s daunting. It also has to be able to release insulin when it’s needed and only when it’s needed.”

In type 1 diabetes, some mix of fairly well understood genetics and not very well understood environmental factors goes wrong. T cells go haywire and begin to wipe out beta cells. So type 1 is described as an autoimmune disease in tens of thousands of research papers.

But maybe things work the other way around: Beta cells stress out and misbehave, and the immune system is just doing its job.

“I actually think that type 1 diabetes is not an immune problem, it is a beta cell problem,” said Roep.

He and his colleagues laid out some evidence in a 2017 Nature Medicine article. “We showed that if beta cells get stressed, which they do very quickly, they produce new antigens like those that expose cancers and infections to the immune system,” he noted.

“I now contend that the immune system is not making a mistake,” Roep said. “It’s the beta cell, and the immune system is actually responding with the best intentions, namely to target stressed tissue… The immune system is not interested in happy tissue.”

Roep is not the only prominent scientist who questions the T-cells-gone-wrong framework for type 1. At the Joslin International Symposium last month in Boston, Olle Korsgren of Uppsala University made another case, skimming through decades of studies on human pancreatic tissue samples analyzed by many researchers.

Among his points, Korsgren cited data suggesting that the T cell attack is surprisingly weak, this attack goes after the whole pancreas rather than just beta cells, and there are frequent signs of beta cell stress such as bleeding. “Could bleeding cells attract the immune system?” he asked.

His hypothesis: Type 1 is not an autoimmune disease that targets beta cells. Rather, it’s an inflammatory disease affecting the entire pancreas. Moreover, the inflammation might be driven by gut microbes invading the pancreas next door.

And Korsgren’s theory just might dovetail very nicely with recent research on the role of the gut microbiome in type 1, now well documented in large epidemiological studies and explored in many labs.

12/20 review from Roep and colleagues. Image: Pamela Itkin-Ansari lab