43.09 Quantification of Enteric Nervous System in Pediatric Small Intestine and Human Intestinal Organoids

J. Shin1, M. Herath1, J.E. Lewis1, A.L. Speer1  1McGovern Medical School at UTHealth Houston, Program In Children’s Regenerative Medicine, Department Of Pediatric Surgery, Houston, TX, USA

Introduction: Short bowel syndrome is prevalent in children and lacking curative treatment. Human intestinal organoid (HIO)-derived tissue-engineered intestine is a potential solution; however, integration of the enteric nervous system (ENS) requires further development. Traditional 2D sectioning is inadequate for analyzing the ENS, leading to recent development of a novel 3D imaging technique for reliable quantification of human colonic ENS. However, ENS quantification in human small intestine (hSI) and HIOs remains unreported. We aim to establish feasibility and reliability of this method in hSI and HIOs.

Methods: Pediatric hSI was obtained from an IRB-approved biobank. HIOs and enteric neural crest cells (ENCCs) were generated from human embryonic stem cells, co-cultured for 28-40 days, transplanted into mice (tHIOs), and harvested after 4-8 weeks. Tissues were fixed, immuno-stained, cleared and imaged using confocal microscopy. Volumes of ganglionated areas (Fig A,B) and ganglia (Fig C,D) were determined using Imaris software. Neurons (HuC/D) and glia (S100B) were manually counted by two blinded observers using Image J and automated neuron counts were obtained using Imaris. Neurons (Fig E), glia (Fig F), and neuron/glia ratio (Fig G) per ganglion were calculated (mean+/-SEM). Intraclass correlation coefficients (ICC) were computed.

Results: There was excellent reliability (ICC=0.93) between manual observer neuron counts and good reliability (ICC=0.83) between manual and automated neuron counts in pediatric hSI. There was moderate reliability (ICC=0.65) between manual and automated neuron counts in tHIOs. Pediatric jejunal ganglia contained 2,809,098+/-483,172 neurons/mm3, 8,271,046+/-1,173,641 glia/mm3, and a neuron/glia ratio of 0.34+/-0.04. Pediatric ileal ganglia contained 4,174,534+/-55,003 neurons/mm3, 9,453,455+/-759,670 glia/mm3, and a neuron/glia ratio of 0.45+/-0.04. Ganglion-like structures in 4-week tHIOs contained 10,090,145+/-4,021,523 neurons/mm3, 12,404,958+/-5,867,843 glia/mm3, and a neuron/glia ratio of 0.85+/-0.08, whereas 6-week tHIOs contained 5,572,456+/-4,084,134 neurons/mm3, 14,188,917+/-10,020,171 glia/mm3, and a neuron/glia ratio of 0.38+/-0.02.

Conclusion: This study demonstrates the feasibility and reliability of a novel 3D imaging method for ENS quantification in pediatric hSI and HIOs. While reliability decreased with automated counting, it was still amply reliable, indicating the potential for automated counting to replace manual counting in pediatric hSI. However, automated counting may not yet be sufficiently reliable in HIOs. Future studies will use machine learning to improve automated counting in HIOs.