Computer Modeling Helps Uncover Possible Alternative for Diabetes Treatment

^{Jeremy Smith, PhD. Photo courtesy of University of Tennessee, Knoxville.}

Patients with type 2 diabetes could one day benefit from a new treatment option, and they owe the opportunity to advanced computer modeling, according to a new study.

Type 2 diabetes and metabolic syndrome, a cluster of cardiovascular and insulin-related conditions, are on the rise both in the US and worldwide. About 400 million people suffer from type 2 diabetes, at an annual global cost of roughly $827 billion. In 2012, 1.5 million people died because of the disease. And although it’s traditionally been a malady associated with older age, physicians report diagnosing more and more young people.

>> READ: How AI Is Shaking Up Healthcare, Beyond Diagnostics

But the good news in the face of those disheartening statistics is that people with type 2 diabetes can be treated, or at least have their disease progression slowed. In addition to encouraging patients to adopt diet and lifestyle changes, providers often prescribe the drug metformin as a first-line therapy. Metformin works by reducing the liver’s output of sugar, helping to normalize blood glucose levels.

However, people with kidney, liver, or heart issues may be unable to take metformin. The medication also can cause a rare but deadly buildup of lactic acid in the blood. And for some patients with type 2 diabetes, metformin simply may not be enough to control the disease as it progresses. Given these constraints, scientists have recognized the need for new therapies that help regulate insulin function.

So, a team of researchers from the University of Tennessee, Knoxville, the University of Tennessee Health Science Center, and the University of Alabama, Huntsville, set about trying to come up with alternatives to metformin.

Darryl Quarles, MD, a top-tier physician and scientist, had long been researching a particular protein—a receptor—that he thought could help treat type 2 diabetes, Jeremy C. Smith, PhD, a molecular biologist at the University of Tennessee, Knoxville, and a member of the study team, told Healthcare Analytics News™. “But he needed to find a chemical, a drug candidate, that would activate this receptor. [We] used our computers to find chemical ‘keys’ that would fit the receptor ‘lock.’”

The receptor was the protein GPRC6A, a key regulator of metabolic function. Using advanced computer modeling, the team, which also included Jerome Baudry, PhD, of the University of Alabama, came up with simulations that allowed them to identify chemicals that could bind to and switch on GPRC6A in order to reduce blood sugar levels.

Based on the results of these simulations, the scientists then created a variety of chemical compounds. One compound in particular, known as DJ-V-159, was found to be the most effective when it came to stimulating insulin production and lowering blood glucose levels after being tested in mice. It could also do the job at much lower doses than ones at which metformin typically is prescribed.

“[This] could be important because it suggests high potency,” Smith said. “If a drug is effective in small doses, it might be even more effective in higher doses. Moreover, smaller doses of drugs can sometimes reduce toxic side effects. However, until more testing is performed, we don’t know whether any of these potential advantages [are] indeed the case.”

Get the best insights in healthcare analytics directly to your inbox.

Related

The Dystopian Concerns of AI for Healthcare

How Sanofi and IBM Watson Unearthed a New Diabetes Finding

AI Can Help Patients with Diabetes Lose Weight