Induction of in vitro Metabolic Zonation in Primary Hepatocytes Requires Both Near-Physiological Oxygen Concentration and Flux

Abstract

Pre-clinical drug screening is an important step in assessing the metabolic effects and hepatic toxicity of new pharmaceutical compounds. However, due to the complexity of the liver microarchitecture, simplified in vitro models do not adequately reflect in vivo situations. Especially spatial heterogeneity, known as metabolic zonation, is often lost due to limitations introduced by typical culture conditions. By culturing primary rat hepatocytes in varied ambient oxygen levels on either gas-permeable or non-permeable culture plates, we highlight the importance of biomimetic oxygen supply for the targeted induction of zonation-like phenotypes. Resulting cellular profiles illustrate the effect of pericellular oxygen concentration and consumption rates on hepatic functionality in terms of zone-specific metabolism and β-catenin signaling. We show that modulation of ambient oxygen tension can partially induce metabolic zonation in vitro when considering high supply rates, leading to in vivo-like drug metabolism. However, when oxygen supply is limited, similar modulation instead triggers an ischemic reprogramming, resembling metabolic profiles of hepatocellular carcinoma and increasing susceptibility toward drug-induced injury. Application of this knowledge will allow for the development of more accurate drug screening models to better identify adverse effects in hepatic drug metabolism.

Keywords: ADME-Tox; PDMS; hepatocyte; metabolic reprogramming; oxygen supply; zonation.

Figure 1

Microenvironmental effect on cellular energy balance. (A) Experimental setup for differential oxygenation of primary rat hepatocytes. Gas-permeable culture plates are used for high [+] oxygen flux. Selective blocking with polyester seals [–] limits gas diffusion to the medium side only. Further variation of oxygen tension is achieved by multi-gas incubators with ambient pO₂ set to 20, 10, 5, and 2.5%. (B) Cellular oxygen consumption rates (OCR) calculated from pericellular oxygen measurement. High flux conditions allow OCR in the range of previously reported in vivo values (Place et al., 2017) while blocked cultures are only viable in 20 and 10%. (C) Purine triphosphate balance and (D) pyrimidine triphosphate balance in conditions subjected to differential oxygenation. (E) Supernatant glucose and lactate concentration change per day on D5.

Figure 2

Quantification of intracellular carbon energy metabolism. (A) Immunofluorescent staining of SLC2A1 expression in different oxygen conditions. Translocation to the plasma membrane indicates active glucose transporter function (scale bar = 10 μm). (B) Residual supernatant insulin concentration after 24 h of culture at day 5. (C) Simplified carbon energy metabolism showing phenotype-specific metabolic patterning. High flux periportal-like cultures exhibit increased gluconeogenetic capabilities, whereas pericentral and hypoxic cultures accumulate citric acid cycle intermediates. Low flux cultures ambiguously exhibit cytosolic metabolism with PKM2-dependent pentose phosphate pathway activation.

Figure 3

Ischemic effect on nitrogen metabolism due to limited oxygen availability. (A) Intracellular amino acid profile of individual cultures as fold accumulation compared to control conditions (B) sPLS-DA principal component analysis based on cellular amino acid profiles. (C) Simplified pathway of cellular ammonia detoxification showing transcript level-independent urea cycle activity in low flux conditions, resulting in increased nitrogen efflux via urea and nitric oxide.

Figure 4

Oxygen-dependent signaling induces zonation-like phenotypes. Immunofluorescent imaging of (A) CDH2, (B) active β-catenin (CTNNB), and (C) glutamine synthetase (GLUL) observed in different conditions (scale bar = 10 μm). (D) Intracellular glutamine levels and (E) secreted albumin levels after 5 days of differential oxygenation. (F) Comparative metabolomic overrepresentation analysis of untreated cultures after 5 days of culture.

Figure 5

Drug metabolic capabilities of induced phenotypes. (A) Differential expression of drug metabolism-related genes in induced phenotypes as fold change versus control groups. (B) sPLS-DA principal component analysis of drug metabolism-related phenotypes. (C) Summary of relevant gene expression for APAP drug metabolism and transport. Quantification of related co-factors shown in terms of intracellular concentration of UDP-glucuronate for UGT1A1-mediated glucuronidation. ATP-availability required for ATP-dependent drug efflux and as a precursor for SULT1A1-related drug detoxification. Intracellular availability of reduced glutathione required for detoxification of monooxygenase-derived drug intermediates in (D) untreated and (E) 20 mM APAP treated groups. (F) Quantification of cytotoxicity after drug exposure compared to control groups. (G) Cellular morphology of hepatocyte sandwich cultures after exposure to 20 mM Acetaminophen (APAP) for 5 h on day 5 (scale bar = 100 μm). Arrowheads show pathological expansion of bile canaliculi.