Engineering Modular 3D Liver Culture Microenvironments In Vitro to Parse the Interplay between Biophysical and Biochemical Microenvironment Cues on Hepatic Phenotypes

Abstract

In vitro models of human liver functions are used across a diverse range of applications in preclinical drug development and disease modeling, with particular increasing interest in models that capture facets of liver inflammatory status. This study investigates how the interplay between biophysical and biochemical microenvironment cues influence phenotypic responses, including inflammation signatures, of primary human hepatocytes (PHH) cultured in a commercially available perfused bioreactor. A 3D printing-based alginate microwell system was designed to form thousands of hepatic spheroids in a scalable manner as a comparator 3D culture modality to the bioreactor. Soft, synthetic extracellular matrix (ECM) hydrogel scaffolds with biophysical properties mimicking features of liver were engineered to replace polystyrene scaffolds, and the biochemical microenvironment was modulated with a defined set of growth factors and signaling modulators. The supplemented media significantly increased tissue density, albumin secretion, and CYP3A4 activity but also upregulated inflammatory markers. Basal inflammatory markers were lower for cells maintained in ECM hydrogel scaffolds or spheroid formats than polystyrene scaffolds, while hydrogel scaffolds exhibited the most sensitive response to inflammation as assessed by multiplexed cytokine and RNA-seq analyses. Together, these engineered 3D liver microenvironments provide insights for probing human liver functions and inflammatory response in vitro.

Keywords: Liver; biomaterials; cell-matrix interactions; hepatocytes; hydrogel; spheroids; tissue engineering.

Figure 1.. Engineering 3D Hepatic Liver Models Using a Perfused Bioreactor and 3D-Printing Based Alginate Microwells

A) Overview of hepatocyte seeding in the perfused LiverChip bioreactor. After seeding cells into wells in a polystyrene scaffold, flow is activated in the downward direction, which pushes cells to the bottom and sides of the flow channel. After 8 hours of attachment, flow is reversed in the upward direction to remove unattached cells and to even out the flow profile. B) Workflow for forming alginate microwells: a reverse-mold PDMS well is filled with alginate precursor, followed by the addition of a filter and dialysis membrane on top. Calcium chloride solution is added to initiate crosslinking, and after 24 hours the alginate can be manipulated and placed in a tissue culture plate. C) Dimensions and schematic of a single microwell patterned into an array that is 3D-printed to the size of a 24-well or 6-well plate. Multiple fitted well arrays are patterned into an aluminum base for PDMS casting. D) Hepatocytes form aggregated spheroids within 3 days. Scale bar = 500μm. E) Rat hepatocyte spheroids cultured in alginate microwells with variable spacing size showed higher albumin production with wider spacing. *p<0.05, N=3.

Figure 2.. Design of PEG Hydrogel Scaffolds Tuned for Perfused Liver Culture.

A) Schematic of PEG hydrogel components: an 8-arm PEG norbornene-functionalized (PEG-Nb) macromer is combined with a sortase-cleavable dithiol crosslinker and cell-adhesive peptide (PSHRN-K-RGD) and polymerized into a network using UV light and photoinitiator lithium phenyl-2,4,6-trimethylbenzoylphosphinate (LAP) B) Schematic of micromolding method used to create perfusable channels. The geometric structure is controlled using gold-coated posts, which block UV light during hydrogel polymerization. C) Scaffold images from left to right: computer-designed scaffold rendering, hydrogel without cells stained with trypan blue, hydrogel with cells after 11 days, F-actin (phalloidin, green) staining of cells in hydrogel. Scale bar = 1mm. D) Gel stiffness and E) swelling values in micromolded PEG-Nb hydrogel scaffolds with increasing PEG-Nb macromer wt%. Stiffness exhibits an expected linear trend, while swelling is minimized and attachment is maximized at 3.5wt%. *p<0.05, ***p<0.001, N=3. F) Hepatocyte attachment and G) hepatocyte spreading metrics in micromolded PEG-Nb hydrogel scaffolds with increasing PEG-Nb macromer wt%. Hepatocyte spreading does not significantly differ between gel formulations. Attachment and spreading are normalized to 3.5wt% PEG formulation. *p<0.05, N=4.

Figure 3:. Characterization of Microwell and PEG Systems for Long Term 3D Liver Culture

A) Rat and B) primary human hepatocyte albumin secretion time-courses over 15 days of culture. Three different culture conditions were compared: LiverChip polystyrene (LC-Polystyrene, black), LiverChip PEG hydrogel (LC-PEG, red), and microwell aggregated spheroids (Spheroids, green). C) CYP3A4 activity on day 11 in human hepatocytes compared across 3D cultures as well as 2D culture. D) Morphological comparison between polystyrene, PEG, and spheroid cultures using immunofluorescence (IF) staining. Top row: arginase (green) and MRP2 (magenta) markers, maximum projected 200µm z-stacks. Scale bar = 100μm. Bottom row: nuclei (blue) and F-actin (green) markers in wider view. Scale bar = 200μm for Polystyrene and PEG, Scale bar = 100μm for spheroids. E) Morphological IF comparison between human hepatocytes with standard and growth-factor supplemented medias on the Liverchip polystyrene scaffold, stained for nuclei (DAPI, blue) and F-actin (phalloidin, green). Scale bar = 100μm. F) CYP3A4 activity in human hepatocytes on Liverchip polystyrene (day 15) with standard hepatocyte culture media and growth-factor supplemented media. **p<0.01, N=3 all graphs.

Figure 4.. Growth Factors and Small Molecules Improve Liver Culture Performance While Increasing Inflammatory Response

A) Z-stack projected confocal immunofluorescence (IF) images of LiverChip polystyrene (left) and PEG (right) hydrogel scaffolds in the absence or presence of growth factors and small molecules (hepatocyte culture medium (HCM) and enhanced liver culture medium (eLCM), respectively) stained for nuclei (DAPI, blue), F-actin (phalloidin, green), and arginase-1 (magenta). Scale bar = 100μm. B) Albumin timecourse comparison in Liverchip polystyrene (PS), PEG hydrogel scaffolds, and spheroids (Sph) in the absence (HCM) or presence of growth factors and small molecules (eLCM). C) CYP3A4 comparison of PS, PEG, Sph conditions with HCM or eLCM. **p<0.01, ***p<0.001, N=3 for all graphs. D) Heatmap of hepatocyte-secreted cytokines/chemokines/growth factors (C/C/GF) with different scaffolds and media culture conditions (rows), grouped together by hierarchical clustering. C/C/GF concentrations are z-scored across conditions (columns). N=3 for each condition, with colored vertical bars right of the heatmap showing the groupings.

Figure 5.. C/C/GF Reveals Inflammation Differences Correlating to the Degree of YAP Nuclear Localization in Platform Responses to eLCM Media

A) Heatmap of cytokine/chemokine/growth factor (C/C/GF) fold changes (FCs) based on polystyrene, PEG, and spheroid cultured with enhanced liver culture medium (eLCM) normalized to the average of their regular hepatocyte culture medium (HCM) condition. N=3 replicates per condition. Vertical bars to the right of the heatmap show culture condition groups. Log2(Fold Change (FC)) values were z-scored across conditions to better visualize contrasts between conditions. B) Selected individual log2(FC) plots showing differential responses of C/C/GF secretion as a function of the culture condition. *p<0.05, **p<0.01, ***p<0.001, N=3 all graphs. C) Z-stack projected immunofluorescence (IF) imaging of YAP (magenta), nuclei (DAPI, blue), and cytoplasm (Arginase-1, green) in HCM media. Left: hepatocytes in polystyrene scaffolds in LiverChip both inside perfused channels (top) and the interstitial space between channels on top of the scaffold (bottom). Right: hepatocytes in PEG scaffolds (top) and as spheroids (bottom). YAP nuclear localization is most apparent in cells on top of the polystyrene scaffold.

Figure 6.. RNA-Seq Reveals Transcriptional Pathway Differences in Platform Responses to eLCM Media

A) RNA-Seq of culture platforms cultured with HCM or eLCM plotted in principle component space (PC1 and PC2). Each colored circle contains 3 biological replicates. B) Volcano plots of polystyrene (PS), PEG, and spheroid (Sph) cultures showing differentially expressed genes (DEGs) upregulated in eLCM (red) or HCM (blue). Top 15 most significant genes are labeled. C) Gene Set Enrichment Analysis (GSEA) dot plots of PS, PEG, and Sph culture platforms upregulated in eLCM normalized to HCM, run using the curated MSigDB KEGG database. The size of each dot corresponds to the size of the gene set, and the color corresponds to the false discovery rate (FDR). D-E) Heat map of DEGs for PS, PEG, and Sph derived from gene sets using D) KEGG-GSEA and E) genes associated with acute phase response and YAP/TAZ. All DEGs are for conditions cultured in eLCM normalized to HCM. Top bar plot above each heatmap shows the fraction of significantly upregulated (red), downregulated (blue), or nonsignificant (grey) genes in each column. F) Venn diagram of unique and overlapping differentially expressed genes, with each culture platform in eLCM culture normalized to HCM culture. Boxed terms show corresponding unique and overlapping terms as determined by GSEA/MSigDB-Reactome database.