Impact of the Physical Cellular Microenvironment on the Structure and Function of a Model Hepatocyte Cell Line for Drug Toxicity Applications

Abstract

It is widely recognised that cells respond to their microenvironment, which has implications for cell culture practices. Growth cues provided by 2D cell culture substrates are far removed from native 3D tissue structure in vivo. Geometry is one of many factors that differs between in vitro culture and in vivo cellular environments. Cultured cells are far removed from their native counterparts and lose some of their predictive capability and reliability. In this study, we examine the cellular processes that occur when a cell is cultured on 2D or 3D surfaces for a short period of 8 days prior to its use in functional assays, which we term: "priming". We follow the process of mechanotransduction from cytoskeletal alterations, to changes to nuclear structure, leading to alterations in gene expression, protein expression and improved functional capabilities. In this study, we utilise HepG2 cells as a hepatocyte model cell line, due to their robustness for drug toxicity screening. Here, we demonstrate enhanced functionality and improved drug toxicity profiles that better reflect the in vivo clinical response. However, findings more broadly reflect in vitro cell culture practises across many areas of cell biology, demonstrating the fundamental impact of mechanotransduction in bioengineering and cell biology.

Keywords: 3D cell culture; bioengineering; cytoskeleton; cytotoxicity; functionality; hepatocyte; mechanotransduction; microenvironment.

Figure 1

An innovative culture strategy that improves cellular function prior to use in an assay system. HepG2 cells, a well-characterised hepatocarcinoma cell line widely used in drug toxicity testing, was selected as a model cell line. Cells were propagated in traditional 2D culture then seeded onto either a 2D or 3D culture substrate for 8 days in a novel priming step. The 3D culture substrate was an inert porous polystyrene scaffold (Alvetex^® Strata), whereas 2D culture substrate was standard culture-ware. Following 8 days of growth on each priming substrate, cells were liberated through trypsinisation and either harvested for analysis and characterisation or re-seeded into a secondary culture system: hanging drop assay, which is a widely used pharmacological assay system.

Figure 2

Three-dimensional-primed cells exhibit enhanced morphology at both a cellular and population levels. Schematic representation of HepG2 cells primed on either 2D/3D substrates, liberated and allowed to adhere onto glass coverslips for 24 h prior to morphology observations (Aa). Gross population morphology highlighted by neutral red staining (Ab–g) demonstrates the differential interactions of 2D and 3D-primed cells and their ability to form colonies and close contact with one another. At a single-cell level observed through phalloidin staining of the actin cytoskeleton (green, Ah,i) and nuclei highlighted in blue, 3D-primed cells appear much more circular with a significantly reduced cell area (Al) (data represent mean ± SEM, n = 40). Scanning electron microscopy (SEM, Aj,k) supports this observation and also demonstrates a greater number of surface protrusions in 3D-primed cells. Schematic representation of cells primed on either 2D/3D substrates, liberated and seeded onto 3D culture scaffolds (Ba–c). Gross population morphology as visualised by neutral red staining (Ab–g) highlights enhanced colony formation in 3D-primed subpopulations. Scale bars: （Af,g, Bd,e） = 200 µm, （Ah,I） = 5 µm, （Aj,k） = 10 µm. ** = p < 0.01.

Figure 3

Nuclear structural markers are downregulated in 3D-cultured cells, indicating changes to nuclear stiffness. Schematic representation of HepG2 cells cultured on either 2D or 3D substrates and liberated prior to protein expression analysis (A). Expression of nuclear lamina proteins (SUN1 and SUN2) measured via Western blot analysis (Ba) reveals marked downregulation in the 3D-primed population. Relative expression of both markers determined by densitometry (Bb,c) confirms this reduction in the 3D-primed population, normalised to β-actin (data represent mean ± SEM, n = 6). Representative immunofluorescence images visualising expression of SUN1 (Ca,b) and SUN2 (Cc,d) in 2D and 3D culture cells supports this finding as reduced staining is observable in 3D-cultured samples (SUN1/2 stained red, Hoechst stain highlights nuclei (blue)). Scale bars 20 µm. *** = p < 0.001.

Figure 4

Changes in expression of hepatic and mechanotransduction-related genes in 3D hepatocyte cultures. Schematic representation of HepG2 cells cultured in 2D or 3D conditions and then harvested for gene expression analysis (A). Heatmaps of mechanotransduction (B,C) liver-related gene expression created with Heatmap.2 in R, from average DESeq2-normalised expression values in each. Colour key equates to log₁₀ of the expression value; darker colours indicating higher absolute expression (data represent n = 4, HL = Human Liver, PHH = Primary Human Hepatocytes). RT-qPCR analysis (D) confirms fold change of selected genes from RNAseq dataset, validating findings and reproducibility of dataset (data represent mean ± SEM, n = 8). * = p ≤ 0.05.

Figure 5

Enhanced metabolism and biosynthesis gene expression in 3D hepatocyte culture through gene enrichment analysis. Schematic representation of HepG2 cells cultured in 2D or 3D conditions and harvested for gene expression analysis (A). Gene set enrichment analysis of significantly differentially expressed genes in 2D/3D cultures calculated using log 2-fold change values of the genes (data represent n = 4, adjusted p value ≤ 0.05). WebGestalt was used to calculate non-redundant enriched biological processes using gene ontology annotation from www.geneontology.org (accessed on 1 February 2022) (B), and Wiki pathway data from www.wikipathway.org (accessed on 1 February 2022) (C). Length of bars indicates normalised enrichment score with metabolism- and biosynthesis-related genes upregulated in 3D culture, and mechanotransduction-related genes upregulated in 2D culture.

Figure 6

Enhanced protein expression of liver function biomarkers in 3D culture. Schematic representation of HepG2 cells culture in 2D/3D conditions and either harvested for protein analysis or biosynthesis and secretion of organic substances measured through ELISA assay (A). Hepatic functional proteins albumin and fibrinogen α-chain are upregulated in 3D-primed cells, as detected via Western blot (B) analysis. Densitometry measurements confirm enhancement of relative expression of fibrinogen α-chain (C) and albumin (D) in 3D cultures, normalised to the loading control β-actin (data represent mean ± SEM, n = 6). Increased production of metabolites albumin (E) and urea (F) in medium of 3D-cultured cells, as measured via ELISA and normalised to total protein (data represent mean ± SEM, n = 24). * = p ≤ 0.05, *** = p < 0.001.

Figure 7

Differential drug toxicity profiles in 3D-primed hepatocytes demonstrate functional-level changes as a result of culture substrate interactions. Schematic representation of HepG2 cells cultured on 2D/3D growth substrates for 8 days before the addition of drugs to the culture medium for 24 h (A). Drug toxicity curves were determined from an LDH assay of the culture medium. Cultures were exposed to increasing concentrations of Amiodarone (B), Gemfibrozil (C), Tamoxifen (D), Ibuprofen (E), Methotrexate (F) and Isoniazid (G) (data represent mean ± SEM, n = 8). The 3D cultures were generally less susceptible to drug-induced toxicity, suggesting a more reliable cell source for pharmaceutical modelling.

Figure 8

Enhanced functionality of 3D-primed cells in “gold standard” hanging drop assay. Schematic representation of 2D/3D-primed HepG2 cells, re-seeded into a secondary hanging drop culture system to form spheroid structures (A). Hanging drops were either analysed or dosed with drugs to develop drug toxicity profiles. Representative images of spheroids formed from 2D (Ba)- or 3D (Bb)-primed cells. F-actin is stained by phalloidin in green, propidium iodide stains dead cells in red and Hoechst highlights nuclei in blue. Increased production of metabolites, albumin (C) and urea (D) in the medium of 3D-cultured cells measured via ELISA, remained a consistent finding even when cells were in spheroid form. Measured via ELISA of culture medium and normalised to total protein (data represent mean ± SEM, n = 24). Spheroids were exposed to increasing concentrations of Amiodarone (E), Gemfibrozil (F), Tamoxifen (G) and Ibuprofen (H) (data represent mean ± SEM, n = 8). Drug toxicity curves were determined from an LDH assay of the culture medium. Scale bars: 100 µm. ** = p < 0.001.