Interleukin-4-mediated Pro-Regenerative Cellular Reprogramming in 3-dimensional Liver Culture

Abstract

Background & aims: Interleukin-4 (IL-4) is a key contributor to liver regeneration, but its effects remain poorly understood due to a lack of models that preserve the complex cellular interactions of the liver. Here, we use murine precision-cut liver slices (PCLS), a 3-dimensional tissue culture system that maintains both parenchymal and non-parenchymal cells, to investigate the role of IL-4 in hepatic cell reprogramming. Through longitudinal single-cell transcriptomics and protein-level validation, we demonstrate the proregenerative potential of IL-4.

Methods: We performed longitudinal single nucleus RNA sequencing on PCLS from 8- to 10-week-old C57BL/6 mice over 5 days of culture in the presence and absence of IL-4. We assessed intracellular ATP output to demonstrate slice viability. We further performed orthogonal evaluations of the impact of IL-4 treatment via immunhistochemical staining to confirm proliferation and cell identity within the slices. We then assessed the impact of IL-4 exposure in slices generated from the diseased livers (hepatonecrosis/fibrosis) of mice treated with thioacetamide.

Results: IL-4 induced transcriptional changes, including increased expression of tissue repair-associated markers in myeloid cells, expansion of hepatocyte and cholangiocyte progenitors, and inhibition of fibroblast activation. Additionally, IL-4 treatment significantly increased Ki67 protein expression and intracellular ATP production, indicating enhanced proliferation and viability. Notably, IL-4 also improved cellular viability in slices from thioacetamide-treated mice, highlighting its potential proregenerative effects in injured liver tissue.

Conclusions: Our study highlights the potential of IL-4-driven modulation of the liver microenvironment, paving the way for cytokine-based therapeutic strategies to enhance immune-mediated hepatic regeneration.

Keywords: Hepatocytes; Liver 3D Culture; Regeneration; Single-cell Transcriptomics.

Graphical abstract

Figure 1

Murine PCLS maintain viability in culture for 5 days. (A) The workflow for generating tissue slices from the whole liver and establishing them in culture. (B) Intracellular ATP production of PCLS based on slice weight (n = 4 for 3 mm, n = 5 for 6 mm at days 1, 3, & 5 of culture). (C) Intracellular ATP production of PCLS over 5 days of culture (n = 6 mice (3 slices per mouse); P = .002). Statistical significance was evaluated using a nonparametric Mann-Whitney test, ∗P < .05, ∗∗P < .01, ns P > .05; error bars represent SEM. (D) Morphology of PCLS slices with immunohistochemistry over 5 days of culture. H&E, TUNEL stains performed on 7-μm serial sections from the same FFPE block of a 250-μm PCLS. Viable nuclei are visualized in purple in the H&E stain and in blue in the TUNEL stain. Apoptotic nuclei are shown in brown in the TUNEL stain. Scale bars represent 20 μm. (E) Representation of apoptotic regions in brown in the TUNEL stains and viable regions in the purple in the H&E stains for 2 separate animals over 5 days of culture. Scale bars represent 250 μm. Images were captured with a microscope at 20× magnification. Files were visualized in QuPath (v0.4.4).

Figure 2

snRNA-seq reveals the maintenance of parenchymal and nonparenchymal populations in slice cultures and the presence of proliferative hepatic populations at day 5 of culture. (A) FACS viable cell sorting strategy with debris visualized in black is gated out. (B) Doublets represented in red are excluded. (C) DAPI-positive cells in blue are sorted. (D) Featureplot of feature RNA (left), mitochondrial RNA (middle), and Alb expression (right) indicating snRNA-seq sample quality. (E) Integrated UMAP containing 6 coarse populations and 10 distinct populations and split by day of culture (Day 1 and Day 3 and 5). (F) Marker genes for distinct cell types in the integrated map by day of culture (1, 3, 5). (G) Proportion of cells from different days of culture (Day 1, 3 and 5) contributing to each cell-type of the integrated data. (H) Heatmap depicting top differentially expressed genes (MAST algorithm) pre-culture (day Zero) and on days of culture (1, 3, 5) (∗P < .001).

Figure 3

Immune cell staining of healthy PCLS as protein-level validation of populations observed in single-nucleus RNA sequencing over the 5 day culture period.(A) T cells positive for the marker CD3, (B) B cells positive for marker B220, (C) myeloid cells positive for marker F4/80, and (D) cholangiocytes positive for marker CK7, are represented in brown in untreated and IL-4 treated PCLS for days 0, 1, 3, 5 of culture. Scale bars represent 50 μm. Images were captured with a microscope at 20× magnification. Files were visualized in QuPath (v0.4.4).

Figure 4

Changes in cell proportions, cell-cell interactions, and cell cycle state over culture time. (A) Cell count of each population over culture time (y axis: cell count, x axis: day of culture). (B) Top significantly enriched CellChat cell-cell interactions (∗P < .05; Wilcoxon test) on each day of culture. (C) Significantly depleted (NES <0) and enriched (NES >0) pathways in the hepatocyte (left) and mesenchyme (right) clusters across days of culture (1, 3, 5) compared with preculture (Day 0). (D) Heatmap of proliferation gene markers Mki67, Smc4, Smc2, and Top2a across days of culture (1, 3, 5) for each cell population (∗∗P <.01, ∗P < .05, avg_log2DC > .5 and minimum perfect expression of 0.1) to preculture (Day 0). (E) UMAP of murine PCLS split by day of culture representing the expression of cell cycle stages, where 0.5 π represents the S phase (DNA synthesis), π indicates the G2M phase (proliferative), 1.5 π indicates middle of M phase, and 1.75–0.25 π indicates G1/G0 state (growth/resting).

Figure 5

Single-nucleus RNA-sequencing reveals proliferation-related gene expression of IL-4-treated murine PCLS. (A) i. UMAP of murine PCLS split by day of culture and IL-4 treatment with annotated population, and ii. barplot of cell proportions (below). (B) UMAP of Tricycle cell cycle analysis, where 0.5 π represents the S phase (DNA synthesis), π indicates the G2M phase (proliferative), 1.5 π indicates middle of M phase, and 1.75–0.25 π indicates G1/G0 state (growth/resting). (C) Heatmap of proliferation gene markers Mki67, Smc4, Smc2, and Top2a across days of culture and treatment status for each cell population (∗∗P < .01, ∗P < .05, avg_log2DC > .5 and minimum perfect expression of 0.1, t-test of pseudobulk average).

Figure 6

Changes in cell proportions, cell-cell interactions and cell cycle state following IL-4 treatment. (A) Nuclei count of each annotated population based on culture time and treatment status (y axis: absolute nuclei count, x axis: day of culture). (B) QuPath quantification (v0.4.4) of B220 (i), CD3 (C) and CK7. Stains using positive cell detection (n = 4 animals). (D) Summary of key significant (P < .05) cell-cell interactions between the mesenchymal population and parenchymal populations, as measured by CellChat receptor ligand analysis.

Figure 7

Exogenous IL4 treatment shifts parenchymal and stromal cell trajectories. (A) i. UMAP of annotated hepatocyte and cholangiocyte subcluster (Hepatocyte1, Hepatocyte2, HepatoCholangiocyte, Cholangiocyte, and Proliferating Cholangiocyte clusters from integrated map in Figure 4A) (nuclei = 8454) (top) and ii. barplot of cell proportions (below). (B) Dotplot of key hepatocyte and cholangiocyte subpopulation markers. (C) UMAP of hepatocyte and cholangiocyte subcluster colored by sample ID with overlaid Slingshot pseudotime analysis showing predicted lineages within the hepatocyte clusters. (D) i. UMAP of annotated mesenchyme cluster from integrated map in Figure 4A. (nuclei = 1824), and ii. barplot of cell proportions (below). (E) Dotplot of key mesenchymal state and subpopulation markers. (F) UMAP of mesenchyme subcluster colored by sample ID with overlaid Slingshot pseudotime analysis showing predicted lineages within the hepatocyte clusters. (G) GSEA enrichment analysis on each day of culture comparing IL-4-treated slices to untreated, in hepatocytes and cholangiocytes, and (H) mesenchymal cells.

Figure 8

Distinct Arg1⁺ myeloid population in IL-4-treated slices express tissue repair associated genes. (A) i. Integrated UMAP of annotated myeloid populations (Marco+ Mac, Arg1+ Mac, Arg1+Marco+ Mac) (top), and ii. barplot of cell proportions (below). (B) Dotplot of key markers for myeloid subpopulations. (C) QuPath quantification (v0.4.4) of F4/80 stains with positive cell detection (n = 4 animals, 1 PCLS per animal per timepoint). (D) Violin plot depicting expression of classically activated macrophage, and (E) tissue repair macrophage genes across sample IDs. (F) GSEA on each day of culture comparing IL-4-treated slices with untreated in myeloid cells.

Figure 9

Exogenous IL-4 promotes intracellular ATP production, expression of proliferation-associated gene modules, and expression of nuclear Ki-67 in PCLS. (A) Intracellular ATP production (per slice) of IL-4-treated and untreated slices over 5 days of culture (n = 5 mice [3 slices per mouse]; P = .007). (B) Comparison of the number of Ki-67⁺ nuclei per mm² tissue in the IL-4-treated and untreated slices on days 1, 3, and 5 of culture (n = 4 [1 slice per mouse]; P = .0286). (C) H&E, TUNEL, and Ki67 stains performed on 3-μm serial sections from the same FFPE block of a 250-μm PCLS. Visualization of nuclei in purple and blue in the H&E and TUNEL stains, respectively, in the untreated and IL-4-treated slices. Apoptotic nuclei are represented in brown in the TUNEL stain. Proliferating nuclei are represented in brown in the Ki-67 stain. Arrows indicate viable, proliferating cells. Scale bars represent 50 μm. Statistical significance was evaluated using a nonparametric Mann-Whitney test, ∗P < .05, ∗∗P < .01, ns P > .05; error bars represent SEM. Images were captured with a microscope at 20× magnification. Files were visualized in QuPath (v0.4.4).

Figure 10

Exogenous IL-4 increases intracellular ATP production and reduces collagen deposition around portal tracts of PCLS generated from a TAA-treated mouse model. (A) Intracellular ATP production of TAA vs TAA + IL-4-treated PCLS generated from TAA-induced diseased mouse model over 5 days of culture (n = 4 mice [3 slices per mouse]; P = .02). Visualization of viable nuclei in the TAA and TAA + IL-4-treated slices with collapsed liver parenchyma in purple and blue in the H&E (B) and TUNEL (C) stains, respectively. Apoptotic nuclei are shown in brown in the TUNEL stains. Proliferating cells are marked brown in the Ki-67 stains (D). Collagen fibers are visualized around the portal vein in green in the trichrome stains (E). (F) Quantification of Ki-67⁺ nuclei in the TAA and TAA + IL-4 mice on day 3 of culture (n = 4 animals; n = 9 vessels). (G) Quantification of collagen deposition in the area surrounding each vessel in trichrome stains for TAA and TAA + IL-4 PCLS for day 3 of culture (n = 4 animals: 11 vessels from untreated, 9 vessels from IL-4-treated animals). Statistical significance was evaluated with a 2-way ANOVA test. (H) Representative images of vacuolization of hepatocytes from 2 animals are shown as pointed by the arrows in the H&E stains of TAA-mice on Day 1 of culture. Scale bars represent 50 μm. Statistical significance was evaluated using a nonparametric Mann-Whitney test, ∗P < .05, ∗∗P < .01, ns P >.05; error bars represent SEM. Images were captured with a microscope at 20× magnification. Files were visualized in QuPath (v0.4.4).