Rutgers researchers create enzyme with artificial intelligence that can aid in treating spinal cord injuries

Rutgers researchers recently conducted a study where they utilized artificial intelligence to create a therapeutic enzyme that can be used for treating spinal cord injuries and enhancing tissue regeneration.

Adam Gormley, an assistant professor of biomedical engineering at Rutgers School of Engineering and principal investigator of the study, said the enzyme that was created in the study is called Chondroitinase ABC (ChABC).

He said that while the enzyme is valuable in its ability to treat spinal cord injuries, it also loses all activity quickly when its temperature reaches 37 degrees Celsius. As a result, multiple infusions of the enzyme are required for it to work, which becomes expensive.

“Our lab utilized an artificial intelligence-driven approach coupled with liquid handling robotics to rapidly synthesize copolymers with different characteristics and tested their ability to stabilize ChABC and retain its activity at 37 degrees Celsius,” Gormley said.

Additionally, he said that synthetic copolymers have recently gained attention due to their ability to wrap around enzymes and protect them from surrounding harsh microenvironments, the surrounding areas of the injury.

Results from one copolymer combination showed that it retained 30 percent activity of the enzyme even at the end of seven days while a native enzyme lost all activity within 3 hours, Gormley said.

Shashank Kosuri, a graduate student at the Rutgers School of Engineering and lead author of the study, said that there is an immediate need to stabilize ChABC to be used as a common treatment in order to exploit its therapeutic potential.  

“Spinal cord injuries have enormous consequences on the physical, psychological and socio-economic well-being of patients as well as the families involved,” Kosuri said. “Soon after a spinal cord injury, a secondary cascade of inflammation results in a dense scar tissue that is highly detrimental toward nervous tissue regeneration.”

ChABC derives from bacteria and is known to degrade certain areas of the scar tissue around injuries and promote neuronal regeneration, Kosuri said.

Though, it also can be time-consuming and labor-intensive to identify each unique polymer combination on a case-by-case basis of the injury, Gormley said.

Gormley also said the researchers designed a seed library that took 504 copolymers and mixed them with ChABC while being thermally stressed at 37 degrees Celsius in an artificial cerebrospinal fluid (aCSF), which is a liquid found in your brain and spinal cord. 

The study used a “Design-Build-Test-Learn” approach that was repeated three times with 24 new polymer candidates generated per iteration. 

“By the end of the three iterations, we were able to identify few polymer compositions that stabilized ChABC remarkably well at body temperature,” Gormley said.

In order to improve the therapeutic effects of ChABC, the team set out to identify copolymer formulations that could stabilize the enzyme and retain its activity for longer durations, Kosuri said.

“Utilizing a machine learning guided approach, we were able to quickly test (more than) 500 unique polymer combinations for their ability to retain ChABC activity,” Kosuri said. “Using an iterative Design-Build-Test-Learn approach we were able to identify few polymer combinations that stabilized ChABC and retained (more than) 140 percent activity at the end of 24 hours while unstabilized enzyme lost all activity within the same time period.” 

One enzyme displayed positive results, which showed 30 percent activity after a period of seven days, he said. These results have motivated future preclinical studies to test their efficacy in degrading scar tissue and promoting tissue regeneration through animal models that are relevant to test with. 

Kosuri said researchers are currently evaluating the efficacy of their best performing tests in a mice contusion spinal cord injury model.

“If we don’t see any therapeutic benefits we might look into other approaches for delivering our (tests) into injured spinal cord tissue,” he said.