(30b) Advancing clinical translation: An automated platform for massively scalable preclinical human testing

Over the past three decades, the engineering and characterization of extracellular matrix (ECM)-based hydrogels has enabled immense progress in the fields of tissue engineering and in vitro disease modeling. The implementation of these hydrogels in human organs-on-chips has ushered in an era of new alternative methods (NAMs) for drug development, overcoming many of the translational shortcomings and ethical issues associated with preclinical animal testing. However, current organs-on-chips are limited by low throughput, lack of reproducibility and the need for intensive manual labor.

To address these issues, we developed a high-throughput platform with end-to-end robotic automation for the seeding, cultivation, dosing, imaging, and multi-omic analysis of thousands of independent, functional, self-organizing, vascularized 3D human tissue models in parallel. By integrating realistic human tissue models, automated high-throughput screening and AI/ML with hydrogel design, we aim to better understand the cellular and molecular mechanisms whereby drugs modulate normal and disease tissue phenotypes. For example, we developed a vascularized, 3D human bone marrow (BM) model by leveraging the ability of adult stem cells to self-organize into a complex, specialized hematopoietic niche when encapsulated in ECM-hydrogels. The fabricated BM niche recapitulates key features of native marrow, including the self-renewal of hematopoietic stem and progenitor cells (HSPCs), multilineage hematopoiesis, and complex ligand-receptor signaling pathways.

Using our automated platform, we treated bone marrow tissues with 27 FDA-approved chemotherapeutic compounds in a blinded screening that included various drug modalities. Multi-modal AI algorithms analyzed the effects on cytopenia, revealing a concentration-dependent induction of neutropenia, consistent with clinical predictions. The compounds also showed varying toxicity levels on the BM vasculature. In separate experiments, we used robotic automation to deliver rapid, serial dosing, mimicking human clinical pharmacokinetic profiles. This approach aims to identify new interventions to mitigate chemotherapy-induced BM toxicity, potentially reducing the long-term, life-threatening effects of cancer therapy on the hematopoietic system.

In summary, the developed platform supports the high-throughput interrogation of complex in vitro models that faithfully recapitulate complex human phenotype and function, while also providing robotic reproducibility and walk-away automation to a formerly artisanal, effort-intensive process.

Author Biography: Jonathan H. Galarraga is the Director of Tissue Model Development at Vivodyne, Inc., a TechBio startup accelerating the discovery and development of human therapeutics through the convergence of novel biology, robotics, and AI. He earned his Ph.D. in Bioengineering from the University of Pennsylvania, where he specialized in biofabrication of cell-laden hydrogels. Jonathan has received numerous research awards, including the NSF Graduate Research Fellowship, and holds a Bachelor of Chemical Engineering from the University of Delaware.