Transcriptomic Profile of Directed Differentiation of iPSCs into Hepatocyte-like Cells

TL;DR

This study compares 2D and 3D methods for turning stem cells into liver-like cells, finding that 3D cultures better mimic liver genes but still lack full adult liver function.

Contribution

The study provides a reproducible transcriptomic benchmark for iPSC-derived hepatocyte-like cells under 2D and 3D conditions.

Findings

3D organoids showed enhanced hepatic gene expression compared to 2D cultures.

3D cultures retained fetal characteristics and had significantly reduced CYP3A4 expression.

Transcriptomic profiles were highly reproducible across multiple iPSC lines.

Abstract

The liver is the central organ in metabolism; however, modeling hepatic diseases remains limited by current experimental models. Animal models frequently fail to predict human liver physiology, while primary hepatocytes rapidly dedifferentiate in culture. We performed comprehensive transcriptomic profiling of induced pluripotent stem cells (iPSCs) differentiation into hepatocyte-like cells (HLCs) under two-dimensional (2D) and three-dimensional (3D) culture conditions. RNA sequencing analysis revealed the sequential activation of lineage-specific markers across major developmental stages: definitive endoderm (FOXA2, SOX17, CXCR4, CER1, GATA4), posterior foregut (PROX1, GATA6), and hepatoblasts (HNF4A, AFP). Comparative analysis demonstrated a markedly enhanced hepatic gene expression of 3D organoids, as demonstrated by a 33-fold increase in HNF4A expression and elevated levels of mature hepatocyte markers, including ALB, SERPINA1, and UGT2B15. However, the 3D cultures retained fetal characteristics (290-fold higher AFP expression) and exhibited significantly impaired metabolic function, with CYP3A4 expression levels reduced by 2000-fold compared to the adult human liver. This partial maturation was further supported by a moderate correlation with adult liver tissue (ρ = 0.57). We demonstrated high reproducibility across five biologically distinct iPSCs lines, including those derived from patients with rare monogenic disorders. The establishment of quantitative benchmarks provides a crucial tool for standardizing in vitro liver models. Furthermore, we delineate the specific limitations of the current model, highlighting the need for further protocol optimization to enhance metabolic maturation and P450 enzyme activity. Functional validation of metabolic activity (CYP enzyme assays, albumin secretion) was not performed; therefore, conclusions regarding hepatocyte functionality are based on transcriptomic evidence.

Figures