Importantly, integrated omics analyses revealed that physiological phenotypes of 3D spheroid cultures were stable on transcriptomic, metabolomic and functional levels and superior alternative hepatic cell culture models.
Pronounced gene expression differences were detected between HepaRG cells, stem cell-derived hepatocytes and 3D spheroid cultures with 11,305 out of 17,462 genes (64.7%) being differentially expressed over the course of three weeks in culture (Figure 1a, FDR<0.01). Importantly, genes involved in the metabolism of endogenous as well as xenobiotic compounds were expressed at significantly elevated levels in spheroids compared to the other two models (padj=5*10-34), indicating drastically elevated metabolic competence (Bell et al., 2017).
Furthermore, orbitrap high-resolution mass spectrometry revealed that spheroids were highly stable on metabolomic level (Fig. 1b,c). When analyzing all identified endogenous compounds (nextra=1,132 and nintra=565 distinct chemical entities), it was found that relative concentrations were very similar with intra- and extracellular correlation coefficients of 0.93 and 0.96, respectively, between freshly isolated cells and spheroids after three weeks in culture.
For analysis of the functional stability of xenobiotic metabolism in PHH between 2D and 3D cultures a cocktail of five non-interacting CYP probe substrates (midazolam, dextromethorphan, phenacetin, amodiaquine and tolbutamide) was used and their metabolites were quantified by mass spectrometry (Vorrink et al., 2017). Importantly, functional activities of CYP1A2, CYP2C8, CYP2C9, CYP2D6 and CYP3A4 exponentially declined and were reduced by 90 % after 24 hours and >95% after 7 days in 2D culture compared to freshly isolated cells (Fig. 1d-h). In contrast, the metabolic activity of PHH in 3D culture was significantly higher for all five CYPs analyzed (p<0.0001; extra sum-of-squares F-test). When quantitatively comparing the amount of metabolites formed per time between 2D monolayer and 3D spheroid culture, the functional activity in 3D culture was found to be elevated between 10- and 1000-fold across all CYP enzymes studied (Fig. 1i).
Figure 1: The 3D spheroid cultures are stable on transcriptomic, metabolomic and functional levels for multiple weeks in culture. a, Transcriptomic profiling of three hepatic in vitro models reveals widescale differences in global gene expression. Heat map depicting differentially expressed genes in 3D spheroids (blue), HepaRG cells (red) and stem cell-derived hepatocytes (scd-Heps; green) at 48 hours, 7 days and 14 days. Overall, 11,305 out of 17,462 genes analyzed (65%) were found to be differentially expressed after multiple testing correction (Benjamini-Hochberg FDR<0.01). PHH spheroid cultures showed highly elevated expression of genes involved in endogenous and xenobiotic metabolism. Data is presented as mean-centred and sigma-normalized. b-c, Scatter log plots of intracellular (b ) and extracellular (c) metabolite abundances at day 21 in 3D culture and in freshly isolated cells. For each metabolite the average abundance of n=6 biological replicates is plotted. Dashed line depicts bisectrix corresponding to perfect correlation. Pearson correlation coefficients of r=0.93 and 0.96 for intra- and extracellular metabolite signatures, respectively, indicate that metabolic profiles are overall stable over the course of at least three weeks in culture. d-h, Column plots showing the levels of the metabolic activities of CYP1A2 ( d), CYP3A4 (e), CYP2C8 ( f), CYP2D6 (g) and CYP2C9 ( h) from three donors cultured in 2D monolayer (blue) and 3D spheroid culture (red) as measured by quantitative mass spectrometry. Dashed line indicates metabolite levels of freshly isolated cells (FIC). Regression lines and their 95% confidence intervals are shown. *** p<0.001, extra sum-of-squares F-test comparing the two fitted regression curves. i , Line plot of fold changes between 2D and 3D cultures of the same donors (n=3) demonstrate that metabolic activities are highly elevated in 3D HepaPredict spheroids. Figure modified from Bell et al., Drug Metab Disp, 2017 and from Vorrink et al., FASEB J, 2017.