Persona Adopted: Senior Research Director in Stem Cell Biology and Computational Genomics

Abstract

This compendium of research details the construction and validation of the Human Endoderm-derived Organoid Cell Atlas (HEOCA), an integrated transcriptomic framework comprising nearly one million single-cell profiles across nine tissue types. By harmonizing data from 218 samples and 55 publications, the HEOCA establishes a high-fidelity reference for assessing organoid maturation and "on-target" cellular composition against primary fetal and adult tissues. Building upon this atlas, the research introduces two advanced bioengineered platforms: Intestinal Immuno-Organoids (IIOs) and "Mini-Colons."

The IIO model incorporates autologous tissue-resident memory T (TRM) cells into epithelial structures, enabling the study of interlineage immune interactions and drug-induced inflammation, specifically regarding T-cell-bispecific (TCB) antibody toxicities. Simultaneously, the "Mini-Colon" platform utilizes hydrogel scaffolds and asymmetric growth factor stimulation to achieve in vivo-like cellular patterning, homeostatic cell turnover, and mature colonocyte differentiation. Collectively, these studies demonstrate that integrating large-scale transcriptomic references with bioengineered niche environments significantly improves the predictive power of in vitro models for drug safety, disease modeling, and regenerative medicine.

Key Takeaways and Technical Synthesis

• [HEOCA Integration] Large-Scale Transcriptomic Mapping:

  • Integrated approximately 1,000,000 cells from 218 experiments covering lung, liver, biliary system, stomach, pancreas, prostate, salivary glands, and small/large intestines.
  • Utilized scPoli for data-aware integration to mitigate batch effects across 12 different sequencing protocols (e.g., Smart-seq, 10x Genomics).
  • Identified 48 distinct level-2 cell types, establishing consistent markers such as OLFM4 for stem cells and TP63 for basal cells across all protocols.

• [Fidelity Assessment] Mapping Organoids to Primary Tissue:

  • PSC-derived organoids predominantly exhibit fetal transcriptomic signatures, whereas ASC-derived models align more closely with adult primary tissue.
  • Benchmarking revealed that pluripotent stem cell (PSC) models often produce "off-target" cells due to incomplete specification, requiring HEOCA-based "on-target" percentages to validate protocol accuracy.

• [IIO Development] Autologous Immuno-Organoid Architecture:

  • Successfully integrated tissue-resident memory T (TRM) cells (CD103+, CD69+) into epithelial organoids using an enzyme-free "crawl-out" isolation method.
  • Observed "flossing" behavior in intraepithelial lymphocytes (IELs), where T cells integrate into basolateral junctions at a physiological ratio of 1:16 (immune to epithelial cells).

• [Immune Toxicity] Recapitulating TCB-induced Inflammation:

  • IIOs effectively modeled clinical toxicities of Solitomab (EpCAM-targeting TCB), showing rapid caspase 3/7 induction at concentrations as low as 40 pg/ml.
  • Transcriptomic analysis identified a TH1-like CD4+ population as the primary driver of early inflammation, preceding the recruitment of cytotoxic CD8+ IELs.
  • Identified the Rho pathway (via ROCK inhibition) and TNF-alpha as viable targets to mitigate immunotherapy-associated intestinal damage.

• [Mini-Colon Bioengineering] Scaffold-Guided Morphogenesis:

  • Integrated organ-on-a-chip technology with laser-ablated hydrogel scaffolds to replicate colonic crypt-villus architecture.
  • Employed asymmetric growth factor stimulation (basal Wnt/NRG1 vs. apical differentiation media) to enable long-term (1 month+) homeostatic culture without passaging.
  • Achieved functional zonation: FABP1+ colonocytes at the luminal surface and OLFM4+ stem cells restricted to crypt bases.

• [Functional Maturity] Mucus Barrier and Nutrient Uptake:

  • Mini-colons produced a thick, physiological mucus layer composed of MUC2 and MUC5B, visible via Alcian Blue/PAS staining.
  • Demonstrated nutrient uptake capabilities using propargyl-choline click-chemistry, outperforming traditional Caco-2 models in metabolic fidelity (e.g., high CYP3A4 and CES2 expression).

• [Preclinical Validation] Predicting Gastrointestinal Toxicity (GIT):

  • Differential toxicological profiling of AML drugs: Cytarabine targeted S-phase proliferative cells in the crypt, while Idasanutlin (MDM2 inhibitor) triggered rapid, widespread p53-mediated apoptosis and epithelial shedding.
  • The platform successfully predicted rapid-onset diarrhea (Idasanutlin) vs. delayed mucositis (Cytarabine), correlating with observed clinical patient outcomes.