Highlights from DDW
Crohn’s disease, FODMAP, SUN1 gene and more: Research updates from Michigan Medicine clinicians.
Digestive Disease Week, or DDW, is the largest meeting in the world involving physicians, researchers and other professionals dedicated to the fields of gastroenterology, hepatology, endoscopy and related surgery.
At this year’s gathering, gastroenterology and hepatology faculty, fellows and research scientists from Michigan Medicine presented on a variety of topics, ranging from the low FODMAP diet and its effects on irritable bowel syndrome, or IBS, to lipodystrophy and its connection to liver disease.
Here, three Michigan Medicine physicians talk about their featured research.
Protein-coding variant in SUN1 gene associates with hepatic steatosis and metabolic disease
Liver specialist and assistant professor Graham Brady, M.D., Ph.D., teamed up with gastroenterologist and professor of internal medicine, Elizabeth Speliotes, M.D., Ph.D., M.P.H, to examine nuclear envelope-related genes and their connection to hepatic steatosis and nonalcoholic fatty liver disease.
“NAFLD is the most common chronic liver disease in the United States, and it affects anywhere between 25 to 30% of the adult population,” said Brady. “While it’s a very common disease, and most patients with this condition will not advance to cirrhosis, a significant portion does – probably somewhere in the range of 5-10%.”
Brady notes that while researchers know quite a bit about how NAFLD manifests, as well as its associated conditions, there are likely to be more genetic associations that have yet to be identified.
“Nearly 50% of the risk for developing this condition is heritable,” he said. “And things like cardiovascular disease and obesity, for example, add to this statistic. But the genetic components are still incompletely understood.”
According to Brady, there are rare mutations in genes that encode parts of the nuclear envelope, which is the multi-protein structure that houses DNA within the cell, that lead to early onset NAFLD, diabetes and severe insulin resistance.
However, Brady said that “it has never been determined whether variants in some of these genes might contribute to NAFLD on a population level, or if there are more common variants that contribute to that 50% risk.”
The team of researchers restricted their study to 17 genes based on their biology, biochemistry and known properties from animal models and human genetics.
“We specifically focused on variants that were predicted to change the amino acids of the proteins, as they would most likely affect protein function,” said Brady. “Our datasets included the Michigan Genomics Initiative, which is a U-M cohort, the UK Biobank, which has publicly available data and the Genetics of Liver Disease Consortium, which Dr. Speliotes is a part of.”
Brady adds that they limited their data analysis to genetic variants that were at least tested in two of the three cohorts, if not all three.
“We didn’t want to chase down data for a variant that was in only one cohort, for example,” said Brady. “So, this ended up being 70 variants in 17 genes. And of those, by far, the top variant was a coding variant in a gene called SUN1, which encodes a protein of the inner nuclear membrane.”
The team was “intrigued” by this variant because it is present within the part of the gene that encodes the portion of the protein that is within the nucleus and therefore potentially associated with DNA.
“We analyzed this variant through a few validated algorithms, which predicted that it may be a potentially damaging variant that might affect the protein’s function,” said Brady. “Using HUH-7 cells, which are derived from liver cancer, we then further examined the effect of this SUN1 variant on the protein’s function.”
The team found that when a tagged version of the wild-type SUN1 protein was expressed in parallel with the histamine to tyrosine variant, there was lower expression of the latter.
“Initially we were skeptical, but we repeated this several times with fresh preparations of plasmid DNA to be extra certain,” said Brady. “We very consistently found lower expression, and increased degradation, of the SUN1 variant protein.”
Brady says that their work raises the possibility that degradation of the SUN1 protein could contribute to the cell’s nutrient-sensing process and therefore, this variant, which increases SUN1 degradation, may contribute to human disease.
“However, whether this is the mechanism that explains its association with disease in the population, we still do not know. Only time will tell.”
Low FODMAP diet improves colonic barrier function and mast cell activation in IBS-D patients
Irritable bowel syndrome, or IBS, that causes diarrhea is often referred to as IBS-D. While it is a relatively common condition, it has a significant impact on the quality of life for individuals with the disease.
“The current understanding of pathophysiology for IBS-D is fairly limited and due to this, sometimes treatments fail, which can lead to poor patient satisfaction,” said gastroenterologist and assistant professor, Prashant Singh, MBBS. “However, a low FODMAP diet, where you restrict certain groups of carbohydrates, has been shown to improve symptoms for nearly 60% of individuals with IBS-D. But the mechanism for the FODMAP-induced IBS pathophysiology is not clear.”
This inspired Singh and a team of experts to examine the connection between this diet and how it helps patients with IBS-D.
“We know that a subset of individuals with this condition have a ‘leaky gut’ and/or an increased activation of mast cells. These cells are in the lining of their colon and when activated can lead to barrier dysfunction,” said Singh. “In our study, we provided patients with IBS-D with a low FODMAP diet for four weeks and then assessed their barrier function and mast cell activation. This considered how leaky their guts were, as well as how activated their mast cells were before and after the diet.”
Singh notes that mast cells traditionally live in the mucosa of the colon and are not very activated. But sometimes, with changes to the microbiome in the gut or due to things like stress, they release compounds that cause increased pain and a “leaky gut.”
“We found that of the 28 subjects in our study, 24 responded to the diet,” said Singh. “And in those 24, their barrier function significantly improved after the diet. We measured this by having them consume a solution that allowed us to assess the lactulose in their urine, which is a marker for their intestinal permeability. The more leaky your gut is, the more lactulose will be present in your urine and vice versa. Surprisingly, we found that levels were significantly higher before the diet and later came down after the diet.”
The team also measured proteins that regulate the gap, or barrier, between two cells. If the proteins are expressed in higher amounts, for example, there would be less of a gap between the cells and therefore, less passing through the lining of the gut. For the patients who responded, the levels of these proteins significantly increased with the low FODMAP diet.
“The gut-barrier function for the four patients that did not respond to the low FODMAP diet did not change,” said Singh. “Therefore, it was the same before and after consuming this diet.”
Overall, Singh notes that this research is novel because it is the first time that a study has shown how a low FODMAP diet can improve barrier function in patients with IBS-D.
“Mast cells are generally dormant, but we ultimately found that they were more active at baseline and less active after consuming a low FODMAP diet.”
AI analysis for improved measurement of Crohn’s disease
According to gastroenterologist and associate professor, Ryan Stidham, M.D., M.S., imaging and computed tomography scans can often provide additional information unseen on colonoscopy that are vital for measuring Crohn’s disease.
“The challenge here is that CT and MRI scans can be even more difficult to consistently and accurately interpret than colonoscopy,” he said. “It is very hard to quantify how much Crohn’s disease is present, the total severity in a patient and how much the disease has changed over time. As a result of these limitations, we have not used imaging to its full potential when it comes to caring for patients with GI conditions.”
Stidham goes on to add that interpretations of CT and MRI scans can differ between radiologists, leaving patients dependent on who is assessing their imaging studies.
“We have been working on developing effective artificial intelligence, or AI, systems that can interpret CT scans and MRIs that provide consistent interpretations. These AI systems can also make measurements that, while important, are often too tedious or impossible for humans to collect.”
Stidham notes that research studies are just beginning to use imaging to help guide therapy. However, the extended time and expertise needed by radiologists to make very detailed and standardized measurements are simply unfeasible in routine patient care.
“But with emerging AI and computer vision, we believe we can train machines to do this type of interpretation almost instantaneously,” he said. “Then, the radiologists and clinicians can spend more of their cognitive efforts on interpreting what the results are for their patients moving forward.”
In a study funded by the National Institutes of Health and the United States Department of Defense, Stidham and a team of radiologists, computer scientists and image analysis engineers explored three different caveats.
“The first thing we did was teach machines how to find intestine on a CT scan automatically,” he said. “And then we ensured that the machine could take detailed measurements of the disease that were comparable to the accuracy of multiple expert radiologists. Once the radiologists did this, we were able to train the AI machines to replicate the accuracy of their work. We found that even the machine’s very precise measurements were made with 90% accuracy. What took our radiologists approximately nine months to do, only took our programmed machines a matter of hours. Perhaps more importantly, we were able to use AI to combine multiple measurements of Crohn’s related bowel injury into a single simplified metric that is easy for caregivers and patients to understand.”
Finally, the team applied their new AI scoring methods to a test group of 240 patients who were using a biologic treatment for Crohn’s disease.
“We wanted to figure out if this AI approach to measuring total bowel disease on imaging was helpful for predicting patient treatment outcomes,” said Stidham. “While our methods are far from perfect, we had a significantly improved accuracy for predicting if treatment would be successful or alternatively, if surgery would be needed, just from the CT scan. Our simple cumulative disease severity score from CT scans outperformed predictions that just used symptoms, labs and medical history, alone. We are very excited about the possibility to use these image analysis technologies to better personalize care for everyone with IBD.”
Live your healthiest life: Get tips from top experts weekly. Subscribe to the Michigan Health blog newsletter
Headlines from the frontlines: The power of scientific discovery harnessed and delivered to your inbox every week. Subscribe to the Michigan Health Lab blog newsletter