Poster #005, Sanford Burnham Prebys Medical Discovery Institute
Engineering Spinal Cord Repair: hiPSC-Derived dI4/V3 Interneurons and GaitFusion, a Markerless AI Gait Architecture for Assessing Recovery
Mentors: Su-Chun Zhang, MD, PhD and Adam Hall, PhD
Traumatic spinal cord injury (SCI) is a devastating condition, often leaving patients withserious motor and sensory paralysis that currently has no FDA-approved treatments. SCI destroys the neuronal tissue necessary to process motor and sensory function, while disrupting both ascending sensory and descending motor information. To address this damage, our lab has devised a cellular engineering method to craft specific spinal interneurons (SpINs). These SpINs aim to repair this damage while attending directly to the depleted sensory and motor capabilities of the spinal cord. This project focuses on both the necessary steps to engineer these pure populations of dI4 (sensory) and V3 (motor) interneurons, while simultaneously developing a novel gait kinematics system to interrogate the degree of recovery within rodent models upon later transplantation. The culture procedure utilized to make pure populations of dI4/V3 SpINs relies on fundamental biological processes to sculpt human induced pluripotent stem cells (hIPSCs) towards the neuronal fate of interest. This stepwise culture ensures that developing neural progenitors align with a spinal fate before conforming to narrow progenitor domains. Early results demonstrate the need for extended culture timelines to recapitulate mature SpIN identities. Anticipating the need for accurate motor testing following eventual transplantation of our SpINs, we developed GaitFusion. This novel machine learning 3-dimensional gait architecture enables analysis of otherwise missed aspects of gait function. This system utilizes transfer learning, multi-view 3D skeletal reconstruction, and anatomical constraint modeling (ACM) to recover accurate limb kinematics without physical markers. The ACM enhanced
accuracy and removed noise compared to previous machine learning procedures. Both the cellular engineering protocols and the GaitFusion system offer significant advancements towards the understanding and treatment of spinal cord injuries.