For decades, a piece of the puzzle has been missing in the study of Crohn’s disease and how it develops in the body.
But a recent discovery from researchers at the UC San Diego School of Medicine in La Jolla has unlocked that mystery with the help of artificial intelligence technology, paving the way for future studies and possible new therapies, according to the university.
Crohn’s is a form of inflammatory bowel disease that can cause pain and other symptoms and result in damage to the intestinal wall.
The human gut contains two types of macrophages, or specialized white blood cells, that have very different but equally important roles in maintaining balance in the digestive system. Inflammatory macrophages fight microbial infections, while non-inflammatory macrophages repair damaged tissue. In Crohn’s disease, an imbalance between the two types of macrophages can result in chronic inflammation.
“Crohn’s disease is one of those diseases of the immune system in which your own immune system attacks the gut and impacts every part from the mouth to rectum,” said Dr. Pradipta Ghosh, a senior author of the UCSD study. “Typically, the sad part of this disease is we do not not have a great grasp on how to treat it and there is no cure, so it can impact everyone from children to adults. It could be a lifelong disease with relapses and surgeries, so it’s a very challenging disease. The burden of the disease is so high economically because the drugs, surgeries and hospitalization are expensive, yet there is relentless progression.”
While it was already known that genetic mutations were the “bedrock” of how Crohn’s disease comes to be in the body, the specifics were unknown.
So the UCSD team applied AI, or machine learning, to the problem.
“Machine learning models look at the complexity of biology through mathematical rules,” Ghosh said. “In every field — wireless communication to modern transportation — if you look, you will see that AI unlocked the potential that was known to us after we grounded those methods in universal principles. Math is precision, so if you apply that precision and rules to biology, we could move mountains.”
In 2001, a gene known as NOD2 was linked to a heightened risk of Crohn’s disease, but the how and why remained a mystery. Ghosh said it was known that “NOD2 functions as the body’s infection surveillance system” and that inefficiencies in the gene had connections to Crohn’s disease.
Using a machine learning algorithm, researchers dug deeper.
Over the course of the study, the team learned that NOD2 binds to a protein called girdin that was deemed “unique and central to NOD2 function,” which was not previously known.
“When bound to girdin, it detects invading pathogens and maintains gut immune balance by swiftly neutralizing them,” Ghosh said. “Without this partnership, the NOD2 surveillance system collapses.”
Going further, the team found that NOD2 must be fully intact to bind to the protein. But occasionally, the tip or tail of the gene can be underdeveloped or missing, causing it to not function properly.
“If you knock off the NOD2 tip, it is the same thing if you knock off girdin,” Ghosh said.
Calling the discovery “foundational,” Ghosh said the same machine learning could be applied to other diseases to see if they share the same process.
“We wouldn’t have gone down that pathway [of study] unless we used a new way of looking at large amounts of data and being unbiased,” she said. “Machine learning makes it so much fun to look for things in places we wouldn’t know to look. [This study] opened up a slew of infectious diseases. We can study so many bacteria that use the same route.”
As for its implications with Crohn’s disease, Ghosh said the discovery opens the door to gene editing that targets NOD2.
“For the first time we have a way of turning on something that was turned off,” she said. “The process is faster now.”
Though any potential therapy is years away, she said, the findings in the study could contribute to the development of treatments aimed at restoring the relationship between girdin and NOD2. ♦
Source link
