Send this article to a friend Print the page

Publication reviews

From 2002, almost 400 publications incorporated the use of HepaRG™, mostly in the fields of hepatotoxicology, drug biotransformation and hepatic functions.
To give you a snap shot of the latest research using HepaRG™, Biopredic International, the HepaRG™ company, has summarized a selection of significant publications from 2014.

Integration of metabolic activation with a predictive toxicogenomics signature to classify genotoxic versus nongenotoxic chemicals in human TK6 cells. Buick et al, 2015

Buick JK, Moffat I, Williams A, Swartz CD, Recio L, Hyduke DR, Li HH, Fornace AJ Jr, Aubrecht J, Yauk CL. 
Environ Mol Mutagen. 2015 

Background
In order to better interpret the outcome of genotoxicity assays, more emphasis is being placed on the mode of action. Mechanism-based genomic investigations have identified the p53 pathway as being key to identifying chemicals that damage DNA. This work is an extension of previous work that identified a genomic gene signifier in TK6 cells that can classify chemicals as either genotoxic or non-genotoxic. The “TGx-28.65 classifier” included various genes associated with a DNA damage response, especially the significant overrepresentation of the P53-signaling pathway.

TK6 cells are a human lymphoblastoid cell line routinely used in standard genotoxicity assays such as the micronucleus assay. However, these cells lack metabolic competence and therefore need supplementation of S9 to detect bioactivated promutagens. Therefore, the effect of the addition of S9 on the gene expression profile should be determined to avoid spurious results when testing genotoxins that require metabolic activation. HepaRG™ cells do not need S9 supplementation because they are metabolically competent and were therefore also tested in these studies. Known genotoxic (benzo[a]pyrene (BaP) and aflatoxin B1 (AFB1)) and non-genotoxic (dexamethasone (DEX) and phenobarbital (PB)) were tested in both cells types to confirm whether the TGx-28.65 gene signature could be used for chemicals requiring metabolic activation in cell incubations with S9 or with HepaRG™ cells. Pervious published gene expression data for 28 chemicals incubated at one concentration and one time point (4h). This work extends the information on 4 chemicals treated at 3 dose levels and 3 time points up to 24h.

Investigations
Low, medium and high concentrations were selected for each chemical in dose range finder studies. Gene expression profiles (using Agilent) were measured after a 4h exposure, as well as 4h and 20h post-exposure. The TGx-28.65 classifier was used to assign a genotoxicity score for each chemical concentration and time point.

The genotoxic effects of BaP correlated with the induction of genetic damage (micronuclei (MN)), apoptosis and cytotoxicity at 24h, and with a time-dependent increase in the expression of three key stress response genes, gATF3, CDKN1A, and GADD45A. A comparison of the TGx-28.65 gene set in BaP-treated TK6 cells against a database resulted in the classification of BaP as genotoxic at the mid and high concentrations in the presence of S9 metabolic activation for all time points. The classifier was considered to be quantitative since the low concentration of BaP was classified as non-genotoxic at 4h and unclassified at 8 and 24h, under which conditions, this chemical did not cause overt signs of cytotoxicity or genotoxicity. In the absence of S9, the high concentration of BaP was not markedly toxic or genotoxic and the TGx-28.65 biomarker classified it as non-genotoxic at all three time points. These results confirm that BaP needs to be metabolised to the ultimate genotoxic metabolite and that the parent chemical does not cause the genotoxic signature. S9 itself did cause some changes in the gene profile but these did not mask the ability to detect non-genotoxic signatures or cause a genotoxic classification.

AFB1 + S9 caused a dose-dependent decrease in the relative survival (RS) and an increase in MN formation. CDKN1A was induced at all time points by all concentrations of AFB1, with the exception of the low concentration at 4h. ATF3 and GADD45A were not induced after 4h but were induced by the mid- and top-concentrations after 8 and 24h. Using the TGx-28.65 gene set, AFB1 classified as genotoxic at the mid and high concentrations at all time points. As with BaP, incubations without S9 did not cause genotoxicity or a decrease in RS and the top concentration was classified as non-genotoxic at all timepoints.

DEX was tested to a maximum concentration of 7.5mM (as recommended by the current ICH guidelines), which decreased the RS to 66%. MN formation was minimally affected (

PB (tested at 8 and 24h only) caused a dose-dependent decrease in RS but no change in MN formation. ATF3, CDKN1A, and GADD45A were minimally affected with the exception of ATF3 and CDKN1A, which were increased at the mid and high concentrations at 8h only. PB was classified as non-genotoxic at all concentrations and time points. The only exception to this was the high concentration with S9 metabolic activation at 24h, which was attributed to confounding effects of the high toxicity on the gene expression profiles.

Cisplatin, a direct-acting DNA-alkylating agent, was one of the 28 training chemicals used in the development of the TGx-28.65 genotoxicity classifier. This chemical was used in these studies as a positive control and to show that the effect of adding S9 to these incubations did not alter the classification of this chemical.

To confirm and validate the use of the TGx-28.65 genotoxicity classifier in cells other than TK6, the published gene profiles of HepaRG™ cells after incubation with 15 chemicals (including genotoxic carcinogens, non-genotoxic carcinogens and non-carcinogens) were analysed. All 15 chemicals were correctly classified using 53 of the 65 transcripts that were common to both studies (using either Affymetrix (HepaRG™) and Agilent (used in these studies).

Conclusions
This work supports the use of the TGx-28.65 classifier to predict the genotoxicity of bioactivated promutagens in TK6 cells supplemented with S9. These data also suggest that the classifier reflects quantitative effects, such that genotoxins were either equivocal or non-genotoxic at low concentrations. The effects were time-dependent and the optimal timepoint for classification of bioactivated chemicals was shown to be 8h. S9 does not alter the classification of vehicle control cells but is clearly required to bioactivate BaP and AFB1 in order to cause DNA damage and induction of genotoxicity-related genes. The classifier was also shown to be applicable to HepaRG™ cells, in which it was 100% accurate in classifying genotoxic and non-genotoxic chemicals. This is very encouraging since this indicates that the TGx-28.65 classifier can be used for different cell models. This methodology is medium throughput and relatively cost-effective and can be used to complement the current test battery as part of the safety assessment of chemicals.

Comparative localization and functional activity of the main hepatobiliary transporters in HepaRG™ cells and primary human hepatocytes. Bachour-El Azzi P et al, 2015

Bachour-El Azzi P, Sharanek A, Burban A, Li R, Le Guével R, Abdel-Razzak Z, Stieger B, Guguen-Guillouzo C, Guillouzo A. 
Toxicol Sci. 2015

Background
This group has characterised HepaRG™ cells and conventional (CCHH) and sandwich (SCHH) culture primary human hepatocytes with respect to their uptake and efflux transports in order to qualify them as suitable in-vitro models for choleastasis. Although the expression of transporters has been reported in these cells, there is little information about transporter activities and their cellular location. This work provides supporting evidence to show the polarised and selective location of different transporters. In addition, the contribution of sinusoidal and bile canalicular efflux transporters and passive diffusion of taurocholic acid (TCA) out of HepaRG™ cultures and primary cultures was compared. Time-lapse imaging was also used to show active clearing of the contents by contracting and relaxing of the canaliculi.

Investigations
Gross morphology of HepaRG™ and 4-5 day old conventional (CCHH) and sandwich (SCHH) cultures of fresh primary human hepatocytes indicated the presence of bile canaliculi. The bile canaliculi networks were more elongated in SCHH with the highest mean surface than in CCHH cultures or HepaRG™ cells. The latter were surrounded by large tight junctions (visualised by electron microscopy). Time-lapse imaging showed that some bile canaliculi contracted and relaxed in HepaRG™ cells, albeit in a non-synchronised fashion, indicating that they were active and periodically clearing accumulated products. These movements were slower in CCHH (it was not possible to see this effect in SCHH).

The distribution of uptake and efflux transporters was in line with their known in-vivo location. The uptake transporter, NTCP, and the efflux transporter, MRP3, were located on the sinusoidal membrane in all three hepatocyte formats, with HepaRG™ showing some diffuse intracytoplasmic staining. The efflux transporters, BSEP, MRP2, MDR1 and MDR3 were all located on the canalicular membrane (co-localized with cytoskeletal F-actin), again with HepaRG™ showing some diffuse intracytoplasmic staining of BSEP. The intracellular location of BSEP has been reported to occur in rat hepatocytes and is a function whereby BSEP can traffic within minutes to the canalicular membrane when the demand for BSEP is increased. This is the first time BSEP has been identified using immunostaining in HepaRG™ cells and primary human hepatocyte cultures.

Sodium-dependent uptake of TCA was demonstrated in HepaRG™ hepatocytes (but not primitive biliary cells), CCHH and SCHH incubations (uptake was 42-, 15- and 33-fold higher in the presence of sodium than in its absence). The uptake was lower in HepaRG™ cells than in primary hepatocytes (2.2- and 1.4-fold lower than SCHH and CCHH, respectively), reflecting the lower mRNA levels of NTPC reported in these cells compared to freshly isolated hepatocytes.    

Efflux activities were measured using CDF (MRP2) and TCA (BSEP), which accumulated in the bile canaliculi after 30min. Accumulation of CDF in all three culture models was decreased by using Ca2+/Mg2+-free buffer or by MK571 (an inhibitor of MRP2, MDR1 and BCRP). The biliary excretion index (BEI) values (%TCA excreted into the bile canaliculi) were 26.6%, 70.9% and 74.2% in HepaRG™ cells, CCHH and SCHH respectively and are in line with reported values for these cells. The lower BEI could be due to a number of factors, including a lower expression of NTCP and BSEP in HepaRG™ cells; an over-expression of MRP3 (and higher contribution of this transporter to excretion of TCA); more effective removal of TCA from the canaliculus by mechanical extrusion (by relaxation/contraction); and incomplete opening of the canaliculi in Ca2+-free buffer, causing an underestimation of the efflux.

The relative contribution of diffusion, and basolateral and canalicular efflux, was investigated by comparing the amount of 14C-TCA in cell lysates and buffer after incubation in Ca2+/Mg2+-free buffers at 4°C and 37°C. These conditions indicated that basolateral efflux and diffusion contributed to 60% to 76% of the efflux in all three culture models, with diffusion accounting for more of the efflux in CCHH than in HepaRG™ or SCHH incubations. The authors noted a large inter-assay variability using primary hepatocytes but not HepaRG™ and that the method used to estimate the contribution of diffusion may result in its overestimation.

MDR activity was assessed using JC-1, a dye that is translocated across the canalicular membrane via MDRs, particularly MDR1. JC-1 forms dimers above a critical intracellular concentration and these emit a red colour; whereas, at low concentrations, the dye is monomeric which emits green light. HepaRG™ cultures incubated with JC-1 dye showed areas of red (hepatocytes) and green (primitive biliary) cells. Green colour indicates a low concentration of JC-1 and thus active efflux of the dye. Pre-incubating cultures with verapamil (an inhibitor of MDRs) before JC-1 exposure resulted in inhibition of JC-1 efflux and the appearance of only red fluorescence in both cell types.

Conclusions
These studies contribute to the knowledge of the main hepatobiliary uptake and efflux transporters present in HepaRG™ and primary human hepatocytes cultured in conventional and sandwich cultures. In addition to confirming the correct location of NTCP, MRP3, BSEP, MRP2, MDR1 and MDR3 in each model, the function of NTCP, BSEP, MDRs were shown to be comparable in HepaRG™ and primary hepatocytes. Small differences in activities reflected the known expression levels in these cells, as well as the differences in the size of the bile canalicular network and the extent of mechanical extrusion of canalicular contents. These studies support the use of both primary hepatocytes and HepaRG™ cells for investigating hepatobiliary transport and canalicular efflux and its alteration by drugs. HepaRG™ cells offer the advantage of consistency in results; whereas, the availability of fresh human hepatocytes and inter-donor variability may limit their use.

Human HepaRG™ Cells can be Cultured in Hanging-Drop Plates for Cytochrome P450 Induction and Function Assays. Murayama et al, 2015

Murayama N, Usui T, Slawny N, Chesné C, Yamazaki H. 
Drug Metab Lett. 2015

Background
Hanging-drop technology allows for the incubation of compounds with cells cultured in 3D in a multi-well format. The 3D structure is more relevant to the in-vivo environment than conventional 2D cultures. The miniaturisation of cell spheres with a reproducible diameter and the ease of the cell handling allows for higher throughput screening. The application of primary human hepatocytes (PHH) to any screening assay is limited by the supply of quality cells; therefore, alternative hepatic models are needed to fill this gap. This group have optimised the application of HepaRG™ cells to the hanging drop format and investigated the CYP induction responses of these cells.

Investigations
All HepaRG™ hanging-drop cultures were pre-cultured for 3 days before adding inducers (omeprazole, phenobarbital, and rifampicin for CYP1A2, CYP2B6, and CYP3A4 induction, respectively). The fold-induction of CYP1A2 mRNA was higher after 24 than after 48h and the response decreased with increasing cell densities (between 72,000 – 18,000 per well). By contrast, CYP2B6 induction was higher after 48h than after 24h and was unaffected by the cell density. CYP3A4 mRNA induction was equivalent at both time-points and was also unaffected by cell density. There was a large variation in the extent of CYP mRNA induction depending on the CYP, time-point and cell density (CYP1A2: 2- to 220-fold, CYP2B6: 5- to 16-fold, and CYP3A4 5- to 10-fold).

Metabolism of midazolam to 1’- and 4’-hydroxymidazolam was compared at different densities (2300 – 72,000 cells per well) of HepaRG™ cells cultured in 2D and hanging drop 96-well plates for 2, 3 and 4 days after culture establishment (2-4 days). The formation of both metabolites was the same in 3D and 2D cultures regardless of the cell density. The major metabolite, 1’-hydroxymidazolam, was formed at similar rates in cultures of 2300 and 72,000 cells per well. Conjugation of 1’-hydroxymidazolam was also equivalent in hanging drop and 24-well cultures of HepaRG™ cells (the latter containing 6-fold more cells).

Conclusions
These results show that HepaRG™ cells cultured in hanging drop format retain CYP3A4 and phase 2 glucuronidation activities for up to 4 days. The 3D spheres were responsive to prototypical CYP1A2, CYP2B6 and CYP3A4 inducers and the extent of induction of CYP mRNA induction was comparable to that seen in 2D cultures of these cells. Metabolism and induction studies are both optimally conducted using with fewer starting cells per well than that recommended for monolayer culture, thus saving on cell usage and media consumption.

Cellular impact of combinations of endosulfan, atrazine, and chlorpyrifos on human primary hepatocytes and HepaRG™ cells after short and chronic exposures. Nawaz et al, 2014

Nawaz A, Razpotnik A, Rouimi P, de Sousa G, Cravedi JP, Rahmani R. 
Cell Biol Toxicol. 2014

Background
Typically, the toxicity of a single chemical is investigated despite that, in real-life many occur as mixtures at low non-acutely toxic levels e.g. pollutants such as pesticides. To address this, this group has compared the acute toxicity of three pesticides in primary human hepatocytes (PHH) and HepaRG™ cells exposed either as a single chemical or as a mixture. Chronic toxicity was modelled in long-term cultures of HepaRG™ cells only. The three chemicals selected belonged to different classes (a herbicide (atrazine(A)), an organophosphorous insecticide (chlorpyrifos (C)), and an organochlorine insecticide (endosulfan (E))), with different modes of action. The development of toxicity over time was monitored using a relatively new technology using a Real-Time Cell Impedance Analyzer. This method generates a cell index (CI) value to continuously quantify cell status based on the measured cell-electrode impedance – if cells are compromised (due to a change in morphology or toxicity) or detach, the impedance decreases. The effect of the pesticides on drug metabolising enzyme expression was also monitored over 2 weeks, since they have been shown previously to modulate both phase 1 and 2 metabolism in mouse primary hepatocytes.

Investigations
After a stabilization period (20h), the alterations in CI were expressed with respect to a normalised CI value of one. Initial studies were conducted in PHH and the results were comparable between fresh and cryopreserved donors. Depending of the type of culture the NC tended to decrease continuously over time. The effect of the test compounds over 24h was then compared to DMSO control cultures. Marked effects on the NC (i.e. decreasing to below 0.5) were only apparent at and above 75mM, either in combination or as a single chemical. Further tests analysed various permutations of mixtures and concentrations and classified them into different groups of toxicity potency. This resulted in an order of toxicity of CE>CE, AE>AC>E >>C, A, with E being the most toxic individual pesticide. The toxicity of E was exacerbated by the co-addition of the other 2 chemicals, even though they were non-toxic at the same concentration. Interestingly, contrary to previous reports of A being nontoxic to rat hepatocytes at similar doses, these data show that the combination of A with C causes significant hepatotoxic to primary hepatocytes regardless of the concentration tested.

In HepaRG™ cells, the stabilization period depended on the seeding density (which was 1 week when 1.3x105 cells/cm2 was used) and was stable for up to 3 weeks after this time. Compounds were tested for up to 2 weeks, during which time the medium was changed every 2-3 days. There were no acute effects (after 24 h) of the pesticides on HepaRG™ cells either alone or in combination at any concentration, indicating a lack of sensitivity of these cells to hepatotoxic effects of chemicals under these conditions. In chronic testing with HepaRG™ cells, the potency of each chemical was classified according to the length of time taken to reach a NCI of 0.5. In keeping with other reports, A was not toxic at any concentration tested (50, 75 and 100 mM) and C (±A) increased the NCI. E was the only compound found to be markedly toxic on its own and caused patches of cells to detach after 6 days which were not replenished by fresh dividing cells. The effect of E was exacerbated by co-addition of A and especially C, regardless of the concentration tested. The combination of A and C did not cause toxicity unless co-incubated with E; and the toxicity of the mix of all three pesticides was no more toxic than E+A or C.

The effect of different non-toxic (up to 25 mM) concentrations of pesticides either individually or in combination on CYP3A3, GSTA1 and UGT1A1 expression was measured qPCR. CYP3A4 expression was dose-dependently increased by E, either incubated individually or in combination with A or C in PHH and HepaRG™ after 24h and 2 weeks in HepaRG™ (by > 20% of the fold-induction by rifampicin). A and C have been reported to equally increase CYP3A4 mRNA levels in HepaRG™ cells; however, in these studies C increased CYP3A4 mRNA in HepaRG™ cells but not PHH after 24h and A induced CYP3A4 only in PHH.

UGT1A1 was unaffected by the pesticides in PHH and HepaRG™ cells after 24h but was significantly increased after 2 weeks in HepaRG™ cells. The induction was concentration-dependant, which was up to 4-fold for E alone and 6-fold of control values at 25 μM for AE, CE, and ACE. Unlike UGT1A1, GSTA1 was unaffected by any of the pesticides either alone or in combination.

Conclusions
The combination of the Real-Time Cell Impedance Analyzer and hepatic models allows the kinetics of acute and chronic repeat-dose effects of test chemicals to be measured. The inclusion of measurements of phase 1 and 2 enzyme expression levels allows for the link between underlying effects due to metabolism and changes in morphology or cell death to be evaluated.

PHH are ideal for detecting the acute effects of chemicals and HepaRG™ are more suited to long-term incubations, since the latter exhibited a lower acute sensitivity than PHH to the toxicity of the chemicals tested. Likewise, HepaRG™ cells may be less suited to studying effects on GSTs since this may be a GSTA1/A2 non-responsive donor. However, chronic effects of pesticides were demonstrated in HepaRG™ cells, as well as a synergistic effect of the pesticides. These studies show that the toxicity of chemicals that are normally present in mixtures should also be tested as mixtures, since the combination of two or more chemicals may be much more toxic than when tested individually.

Dual-color fluorescence imaging to monitor CYP3A4 and CYP3A7 expression in human hepatic carcinoma HepG2 and HepaRG cells. Tsuji S. et al, 2014

Tsuji S, Kawamura F, Kubiura M, Hayashi A, Ohbayashi T, Kazuki Y, Chesné C, Oshimura M, Tada M.
PLoS One. 2014 August.

Background
This paper describes how levels of CYP3A4 and its foetal form, CYP3A7, can be monitored in HepaRG™ and HepG2 cells.
Although these CYPs are very similar, their regulation, kinetics and substrate specificity are different; therefore, adult hepatocytes expressing CYP3A4 are needed to model interactions with this CYP. The high homology of the CYP proteins means that they cannot be differentiated between using immunostaining techniques. Unlike primary human hepatocytes, HepaRG™ and HepG2 cells represent a huge cell source and, in the case of HepaRG™ cells, the CYP expression is similar to that in primary hepatocytes and is sustained over weeks in culture.
The technology described involved the use of transgenic cells which contain reporter genes for both CYPs which are linked to either green or red fluorescence. For example, hepatoblasts from HepaRG™ cells express mainly CYP3A7 and thus exhibit a red fluorescence; conversely, as they differentiate and exhibit hepatic markers, they lose the red fluorescence and express only green marker, indicating the presence of CYP3A4 and an adult phenotype. The level of green fluorescence is proportional to the mRNA CYP3A4 expression and can therefore be used to monitor in real-time the changes in expression of this CYP e.g. in the presence of inducers.  Ultimately, this technology allows for the cost-effective and early high throughput screening of potential CYP3A4 inducers, as well as the selection of CYP3A4 expressing cells for cell sorting and enrichment.

Investigations
Fluorescent reporters were used whose expression is under the control of the enhancer and promoter regions of CYP3A4 and CYP3A7. The open reading frames (ORFs) of CYP3A4 and CYP3A7 were replaced with EGFP and DsRed, respectively, using a bacterial artificial chromosome (BAC) vector (4G/7R BAC). The optimal HepaRG™ clone in which the 4G/7R BAC was inserted into a loxP site on human chromosome 16 had a similar karyotype to the parental HepaRG™ cells and was thus mainly used. The 4G/7R BAC was inserted into the loxP site created on human chromosome 14 in HepG2 cells, and of the clones produced, clone 87 (C87) was mainly used.

The basal and rifampicin-induced expression of CYP3A4/7 in 17 hepG2 clones, were evaluated. The clone (C87) with a high signal to noise ratio for EGFP when treated with 100µM rifampicin and/or 200µM dexamethasone was selected for further investigations. HepG2 and HepaRG™ clones were treated with DMSO, rifampicin or dexamethasone and the induction response measured using fluorescence microscopy, qRT-PCR, Western blotting, or FACS. The relative induction of CYP3A4 and CYP3A7 by rifampicin and dexamethasone was reflected in all endpoints measured. The caveat to using transgenic cells is that there is a threshold for the reporter gene below which the fluorescence may not be detected; therefore, a lack of fluorescence does not mean the reporter gene is absent. Using transgenic HepaRG™ cells, the principle of screening for CYP3A4 induction was tested in 96-well format using a panel of known non-, weak and strong CYP3A4 inducers. The potency of the inducers was confirmed, together with the known species-specific induction effect of pregnenolone 16a-carbonitrile (non-inducer of CYP3A in human hepatocytes and transgenic HepaRG™ cells but positive in rat hepatocytes).

Using real-time imaging, the progression of the differentiation of transgenic HepaRG™ cells was monitored. As they differentiated from Day 7 to 12, the intensity of EGFP fluorescence increased; whereas, the intensity of DsRed fluorescence gradually disappeared and this was mirrored in the mRNA expression of these reporter genes. The expression levels of EGFP and CYP3A4 but not DsRed and CYP3A7 correlated, indicating high levels of DsRed mRNA transcription may be required to visualize DsRed fluorescence in each cell (which only carries a single copy of the DsRed reporter gene). Cells were separated by FACS into mainly EGFP-positive cells and mainly EGFP-negative cells. Some method optimisation was required but the resulting protocol allowed for the enrichment of EGFP-positive differentiated HepaRG™ cells. The use of these transgenic cells allowed the differentiation process of HepaRG™ cells to be investigated further. HepaRG™ cells are bipotent cells and can differentiate into CYP3A4-positive hepatocytes, CYP3A4-negative hepatic cells, or biliary-like cells. Undifferentiated cells were mostly DsRed-positive but the marker was undetectable after differentiation, at which point, EGFP-positive cells appeared. By monitoring the change from red to green over time, it was shown that the number of DsRed-positive cells decreased from differentiation Day 6 to 8, and the non-fluorescent cells (i.e. no CYP3A4 or CYP3A7) gave rise to EGFP-positive cells. Therefore, most EGFP-positive cells were derived from proliferating DsRed-negative cells and that most CYP3A4-positive hepatocytes indirectly appear from CYP3A7-positive hepatoblastic cells.

Conclusions
In this work, transgenic HepG2 and HepaRG™ cells were developed to monitor CYP3A4 and CYP3A7 expression in real-time. Changes in EGFP fluorescence intensity correlated to changes in the CYP3A4 mRNA and therefore, many of the time-consuming steps required to measure mRNA and protein levels normally carried out in CYP induction assays are eliminated.

This technology provided further insight into the changes in CYP3A4/7 expression during the differentiation from HepaRG™ hepatoblasts, via non-CYP3A-expressing proliferating cells, to CYP3A4-expressing, CYP3A7-negative differentiated hepatocytes. This aspect may be used to explore further the processes and factors involved in differentiating cells towards CYP3A4 phenotype.

The transgenic HepaRG™ and HepG2 cells produced were shown to be suitable for use in high throughput CYP3A induction screening. The ability to measure CYP changes in real-time enables the detection of the peak induction response, the time at which might be different from one compound to another. Transgenic HepG2 cells can be used directly in the CYP induction assay; whereas, HepaRG™ cells require 4 weeks of differentiation before use. However by preparing frozen stocks of EGFP-positive HepaRG™ cells, the differentiation process can be circumvented and the cells used in the same timeframe as transgenic HepG2 cells. Since HepaRG™ cells are used in assays along the drug development process they were regarded as more appropriate than HepaG2 cells for the early CYP3A4 induction screening. Importantly, EGFP-positive HepaRG™ cells can also be used instead of human hepatocytes in a panel of other preclinical tests, such CYP3A4-dependent drug metabolism, hepatotoxicity and carcinogenicity assays.

Generation of functional cholangiocyte-like cells from human pluripotent stem cells and HepaRG cells. Dianat N et al, 2014

Dianat N, Dubois-Pot-Schneider H, Steichen C, Desterke C, Leclerc P, Raveux A, Combettes L, Weber A, Corlu A, Dubart-Kupperschmitt A.
Hepatology. 2014 Aug

Background
Cholangiocytes are biliary epithelial cells that line the intra and extrahepatic ducts of the biliary tree. These cells regulate bile composition by modification of active secretion and resorption in addition to participating in the detoxification of xenobiotics. These cells are also primary targets of injury in a variety of cholestatic liver diseases. Despite their importance, cholangiocytes represent only 3% of the total liver mass and this, together with their intraheptic location, makes them difficult to isolate. Therefore there are very few in-vitro cholangiocyte models and these are mainly immortalised cell lines or rodent cells. To address this, alternative sources of cholangiocytes are needed.

This paper describes how the differentiation processes involved in cholangiocyte formation can be recapitulated in-vitro using human embryonic stem cells (hESCs) and induced pluripotent stem cells (hiPSC). The resulting cholangiocytes were characterised using gene expression analysis, morphological features and functional characteristics. The same differentiation process was also successfully applied to pluripotent HepaRG™ cells. These studies support the use of cholangiocytes derived from hESCs and HepaRG™ to further elucidate the molecular mechanisms of bile duct development. They also represent a suitable model for bile duct targeted therapeutic as well as use in bioengineered liver models.

Investigations
Initial studies used hESCs to derive hepatoblasts which were subsequently differentiated into cholangiocytes using a step-wise protocol. The potential of the hESC-HB to commit to the cholangiocyte pathway was confirmed by the presence in these cells of the biliary marker genes, FOXM1B, NOTCH2, and SALL4. The hESC-HB were then differentiated into cholangiocytes using a combination of epidermal growth factor (EGF) and interleukin 6 (IL-6) and growth hormone (GH), as well as sodium taurocholate. These factors cause the activation of appropriate signalling pathways known to be involved in biliary differentiation. In addition to sodium taurocholate stimulating the proliferation and differentiation of cholangiocytes, it also has anti-apoptotic activity.

The incubation of hESC-HB with EGF and IL-6 allowed the cells to proliferate and reach confluence. Once confluent, the cells were incubated further for 3 days with IL-6 and then with sodium taurocholate for 2 days. During this time, the gene expression and morphology of the cells was monitored and confirmed a significant increase in gene expression levels of cholangiocyte markers, CFTR, TGR5, aquaporin-1 (AQP1), SOX9, SCTR, and JAG1, (whereas the stemness marker, NANOG, was lost) and cholangiocyte-like morphology. The biliary receptors were also present; moreover, 90% of the cells showed the presence of CFTR.

The same protocol used for hESCs was applied to HepaRG™ cells but the differentiation included an additional 2 days of incubation with sodium butyrate. This was included to prevent spontaneous differentiation of HepaRG-HB into hepatocytes. The resulting HepaRG-derived cholangiocytes (HepaRG-Chol) expressed osteopontin (OPN), CK19, SCTR, and tetraspanin-8 (C0-029), as well as high levels of GGT1, CK19, JAGGED1, and TGR5 compared to hepatocyte-committed HepaRG™ cells (HepaRG-cHep) and HepaRG-HB. By contrast, hepatic markers e.g. ALB and HNF-4a were repressed in these cells. Hierarchical clustering analysis confirmed the commitment of progenitors into cholangiocytes such that hESC-Chol, and HepaRG-Chol, as well as the immortalised cholangiocyte line, H69, and normal human biliary epithelial cells fell into a single cluster; whereas a second cluster contained HepaRG-Hep and HepaRG-cHep. Within Cluster 1, HepaRG-Chol clustered with both H69 cell line and human normal biliary epithelial cell, thus reflecting it hepatocellular carcinoma origin.

Final investigations were based on hiPS cells, which were successfully reprogrammed into hepatoblasts in the same way as that used for hESCs. hiPSC-HB were then differentiated into cholangiocytes (hiPSC-Chol). hiPSC-Chol expressed a similar panel of cholangiocyte marker genes but not the hepatocyte marker, HNF-4a.

A number of morphological and functional characteristics of cholangiocytes were also demonstrated to be present in hESC-Chol, hiPSC-Chol and HepaRG-Chol. These cells have primary cilia (detected using immunostaining of acetylated a-tubulin) and formed round cysts with luminal spaces after 7 days in 3D culture. The cyts were polarised, as verified by the basolateral and apical localization of b‑catenin and F-actin, respectively. The secretion function of cholangiocytes was demonstrated using MDRI substrates and the fluorescent bile salt, cholyl-lysyl-fluorescein.

Conclusion
These authors demonstrated for the first time that cholangiocytes can be generated from human ESC and iPS cells, as well as HepaRG™ cells.
The differentiation methods were the same for all three cells, with the exception that HepaRG cells required an additional step with sodium butyrate to prevent spontaneous differentiation of HepaRG-HB into hepatocytes.

The resulting cholangiocytes derived from all three originating cells expressed multiple genes, morphology and functions characteristic of cholangiocytes, confirming their commitment to the cholangiocyte pathway. When placed into 3D culture, the cholangiocytes form polarized cysts and tubules with functional efflux transporters around a central lumen. The presence of cilia, the sensory organelle present on cholangiocyte apical surface is of importance since some primary cholangiopathies (e.g., polycystic disease), can be caused by defects in these cilia.

The development of a cholangiocyte in-vitro model represents an important advance in the area of bile duct development, liver fibrosis and other pathologies and the molecular mechanisms involved. The potential for a sufficient source of cholangiocytes may also provide a human relevant source of cells for bioengineering.

Publications reviews

 

 

  See all peer-reviewed publications about HepaRG™ 

Differentiation of liver progenitor cell line to functional organotypic cultures in 3D nanofibrillar cellulose and hyaluronan-gelatin hydrogels. Malinen MM et al, 2014

Malinen MM, Kanninen LK, Corlu A, Isoniemi HM, Lou YR, Yliperttula ML, Urtti AO.
Biomaterials. 2014 Jun.

Background
Standard culture conditions for primary human hepatocytes and other cell types used in drug development involve simple 2D monolayer formats. However, in order to better mimic the in-vivo environment and to extend the longevity of PHH, they should be cultured in 3D e.g. sandwich culture etc. This group explored the use of hydrogels of native wood-derived nanofibrillar cellulose (NFC) and hyaluronan-gelatin (HG) for the 3D culture of undifferentiated and differentiated HepaRG™ cells.

Investigations
Multicellular spheroids formed from undifferentiated cells in both 3D formats. Low density undifferentiated HepaRG™ cells grew to a maximum size of 40 µm in both NFC and HG after 1 week. The undifferentiated cells had a more organised structure than high density differentiated HepaRG™ cells, which exhibited lower mitochondrial activity and did not proliferate over the same time and retained approximately the same size throughout the 7 day incubation (40 µm).

The expression levels of hepatic markers, HNF-α, albumin, CYP3A4 and the transporter genes, MDR1 and MRP2, and the cholangiocyte marker, keratin 19 (K19), were monitored in HepaRG™ cells during the course of the 14 day incubation. The authors were able to show that different culture conditions i.e. low and high density in either 2D monolayers or 3D cultures in NFC and HG, affected the phenotype of the cells. For example, low density undifferentiated HepaRG™ cells cultured in all matrices differentiated towards hepatocytes over time, by contrast, the hepatic gene levels in high density differentiated cells in 3D culture decreased markedly over time. Notably, differentiation towards hepatocytes was accelerated by culturing undifferentiated cells in NFC or HG, which was attributed to the gels facilitating cell-cell contacts.

Basal CYP3A4 activities were markedly affected by the different formats. The most beneficial matrix was NFC, in which CYP3A4 activity in low density cultures increased by 4-fold over 2 weeks, in comparison to 2D, in which CYP3A4 activities remained constant, and HG, in which activities were very low. The density at which cells were cultured also affected CYP3A4 levels, such that they were highest in high density differentiated cells cultured in 2D than in 3D NFC and HG hydrogels. Despite the different basal activities, CYP3A4 was induced by rifampicin and phenobarbital to the same extent in HepaRG™ cells cultured in all formats.

Functional and polarised efflux transporters were demonstrated in 3D spheroids in 2D, NFC and HG cultures using the MRP2 and BSEP substrate, fluorescein, and the MRP2-selective substrate, calcein, and located in the canalicular structures by staining cells for F-actin and MRP2 and MDR1. The culture format also affected the level of transporter expression and number of canaliluar-like structures, which were highest in low density HG than in NFC cultures.

Conclusions
Wood-derived NFC is a promising new material for 3D cell culture. The 3D format combined with low density undifferentiated HepaRG™ cells improved hepatic functions and expression markers, NFC and HG being equally effective in this respect. The resulting spheroid cultures offer the potential for use in drug and chemical testing, especially with respect to transporter functions, which are not polarised in 2D formats.

A dynamic multi-organ-chip for long-term cultivation and substance testing proven by 3D human liver and skin tissue co-culture. Wagner I. et al, 2013

Wagner I, Materne EM, Brincker S, Süssbier U, Frädrich C, Busek M, Sonntag F, Sakharov DA, Trushkin EV, Tonevitsky AG, Lauster R, Marx U.
Lab Chip. 2013 Sep.

Background
One reason or failing to accurately predict drug toxicity is the lack of complexity of the models for human organs. The “human-on-a-chip” concept is recognised as a step towards solving discrepancies between different species and in-vitro versus in-vivo, and the restrictions for in-silico modelling. This paper reports on a multi-organ-chip in which co-cultures of different cell types from the liver and intact skin are cultured in 3D. This miniaturised organ model allows for adjustable fluid flow and a controllable local tissue-to-fluid ratio and thus, inter-organ cross-talk (shown to be an important physiological factor for functional tissue behaviour in almost all organs). Other fluidic systems have been developed but they do not focus on combining two or more organ compartments in one circuit and they use much larger volumes of medium (>3ml compared to only 0.6ml used here).

Investigations
The multi-organ-chip reported here consisted of a combination of a co-culture of different cell types from the liver (HepaRG™ and primary human hepatic stellate cells (HSCs)) and fresh skin biopsies in a perfused 3D 96-well format. Liver microtissue aggregates were formed using hanging drop technology using hepatocytes and HSC aggregates (50000 cells per aggregate and a total of 1 million cells per well). The cells within the aggregates exhibited polarisation and the existence of bile canaliculi-like networks (confirmed by positive MRP-2 staining in the bile canaliculi). Skin biopsies were trimmed to 2mm thickness and 4mm discs were mounted onto the well compartment. The ratio of number of skin cells to liver cells was the same as that in-vivo - each 1/100000 of the biomass of the in-vivo human organ. Tissues were perfused either submerged or at the air-liquid interface for the skin, using transwell inserts to protect the liver aggregates below and allow the culture of skin on the transwell membrane at air-liquid interface. Skin biopsies cultured on transwells were better maintained than those submerged and incubated directly in the flow of medium. Cultures were maintained for 14 -28 days without contamination. The functionality of the co-cultures was determined by measuring glucose consumption, lactate production and albumin secretion, all of which dropped over the first 4-5 days (the stabilisation period) and then remained constant thereafter. By day 28, CYP3A4 and CYP7A1 (involved in bile acid synthesis) were still clearly present in the liver aggregates. Likewise, in skin sections, the presence of cytokeratin 15 and 10 indicated that the epidermis still exhibited both basal and stratified keratinocytes and thus, an intact barrier function. There was only a low level of apoptosis in both liver and skin tissues, suggesting an adequate supply of nutrients throughout the whole culture period.

The toxicity of troglitazone was investigated by incubating the model with 5 and 50 M troglitazone, for 7 days, which was applied at 12 h intervals along with the medium changes. This known hepatotoxic compound caused a dose- and time-dependent increase in the release of LDH and a marked increase in CYP3A4 expression (measured using rt qPCR and immunohistochemistry) and mirrored effects reported by others.

Conclusions
This micro-fluidic model can maintain human liver and skin tissues for at least 28 days without loss of basic functions such as glucose consumption and albumin secretion, or structural integrity e.g. polarisation of liver aggregates or retaining the barrier function in skin. Future goals aim to more precisely monitor the oxygen consumption of the individual tissues in order to characterise their energy balance in various designs over extended culture periods.

Other human tissue organ models within a common fluid flow, such as liver, skin (epidermal tissues) and intestine (caco-2 cells) can also be applied to this model. Investigations into the cross-talk between organs can be made and how they may be altered by exposure to chemicals. This paper is of relevance to the cosmetics industry since topically applied chemicals such as those in cosmetics can be tested using this model to mimic skin absorption and subsequent hepatic metabolism and toxicity once it has entered the systemic circulation. Moreover, repeated dose studies for up to one month can be employed to better reflect the use conditions and regimens relevant to respective guidelines.

Impact of inflammation on chlorpromazine-induced cytotoxicity and cholestatic features in HepaRG cells. Bachour-El Azzi P et al, 2014

Bachour-El Azzi P, Sharanek A, Abdel-Razzak Z, Antherieu S, Al-Attrache H, Savary CC, Lepage S, Morel I, Labbe G, Guguen-Guillouzo C, Guillouzo A.
Drug Metab Dispos. 2014 Sep.

Background
This paper describes how the toxic effects of the idiosyncratic drug, chlorpromazine (CPZ), are affected by inflammatory mediators. The idiosyncratic nature of this and other drugs is thought to be linked to inflammation, as well as other factors such as polymorphisms. Cytokines down-regulate certain phase 1 and 2 drug metabolising enzymes and transporters, and for CPZ this could exacerbate the cholestatic effects of this drug. 

HepaRG™ cells were treated with non-toxic concentrations of the cytokines, IL-6 and IL-1β for 24 h, followed by treatment with a non-cytotoxic concentration of CPZ (20µM) and/or cytokine for 24 h or daily for 5 days. Endpoints measured were viability, ROS levels, CRP and IL-8 gene expression and protein production, F-actin distribution, NTCP activity and TA and CDF efflux, as well as quantification of the concentration of CPZ in the medium and cell lysates (biokinetics).

Investigations
IL-6 and IL-1β increased the toxicity of 20µM CPZ but only after 5 additions. At the concentrations used, both cytokines but not CPZ itself caused in inflammatory response, evident as a significant increase of CRP mRNA and protein levels at both time points. Levels of IL-8 mRNA and protein were unaffected by 20µM CPZ in the presence or absence of IL-6; whereas they were increased by IL-1β in the presence or absence of CPZ after 1 addition (less so after 5 additions).

The effect of a higher concentration of CPZ (50µM - at which it is known to cause cholestatic effects) on cytokine levels was also tested. At this concentration, CPZ was shown to increase IL-8 mRNA and protein levels, the latter being observed after only 2h. Moreover, CPZ also increased levels of IL-1β and IL-6 transcripts and secreted IL-6 protein was detectable after only 4h.

At the concentrations tested both cytokines markedly inhibited CYP3A4 and CYP1A2 expression by 58-90%. By contrast, CPZ itself induced these enzymes – an effect that was counteracted by co-incubation with cytokine, such that, over the course of 5 days, the inhibitory and induction effects of cytokines and CPZ cancelled each other out. The inhibitory effects of the cytokines was reflected in the relative amounts of CPZ metabolites in the culture medium and lysates, such that cytokines decreased the metabolite formation and increased the concentration of parent compound by approximately 2-fold.

CPZ (20µM) was shown to cause phospholipidosis, evident as the formation of lamellar bodies after 3 days and induction of genes related to this process. Cytokines affected CPZ induced genes involved in lipid metabolism but abolished the expression of a lipogenic gene (THRSP).

The cytokines did not affect ROS generation in HepaRG™ cells nor did they affect the small increase in ROS generation caused by CPZ. However, at the gene level, both cytokines and CPZ caused increases in the oxidative stress-related genes, heme oxygenase 1(HO1), manganese superoxide dismutase MnSOD), and NF-E2-related factor (Nrf2), which was increased further by co-treatment.

CPZ was shown to affect both efflux and uptake transporters. CPZ (but not cytokines) caused the accumulation of taurocholic acid (a BSEP substrate) and inhibited canalicular excretion of CDF (MRP2 substrate). The co-exposure of cytokines did not increase the inhibitory effects of CPZ on CDF excretion. As shown by others, CPZ altered the distribution of cytoskeletal pericanalicular F-actin, as well as the shape of the bile canaliculi within a few hours. These alterations were not aggravated by cytokine pre-treatment. NTCP-dependent activity was decreased by IL-6 and IL-1β but not by CPZ alone. Co-incubation of cytokines and CPZ increased the inhibition of NTCP activity further. Expression of the main genes involved in bile acids transport and synthesis revealed that several transporters were modulated by cytokines but not CPZ alone. The major effect was a down-regulation of influx transporters, especially NTCP, which was decreased further by co-incubation with CPZ.

Expression of CYP genes involved in bile acid synthesis was also analyzed. CYP8B1 mRNA levels were reduced by co-incubation of CPZ and IL-1β but not by other treatment combinations. Other CYPs were either only slightly over-expressed by co-treatment of CPZ and IL-1β (CYP7A1) or not affected by any treatment (CYP27A1). FXR expression was only minimally affected by cytokine and/or CPZ treatment but PXR and expression was down-regulated by both cytokines alone or in presence of CPZ.

Conclusion
This paper provides a comprehensive overview of multiple endpoints relating to the toxic and cholestatic effects of CPZ, as well as determining how the cellular modification affects the metabolism and accumulation of CPZ in hepatocytes. Short term (a few hours to 1 day) and longer-term (5 days) effects of CPZ were also considered which allowed for modifications at the gene level to be detected and more chronic consequences to take effect (e.g. CYP induction, phospholipidosis etc). More importantly, the effects of proinflammatory cytokines on the endpoints were determined. Of major importance to the in-vivo clinical situation was the confirmation that inflammatory cytokines cause a decrease in uptake transporter (NTCP) activity and a repression of CYP3A4 and CYP1A2 expression, which results in the inhibition of CPZ metabolism and its accumulation in hepatocytes. This effect could have major consequences on CPZ toxicity itself, as well as other drugs that are metabolised by these CYPs, should the patient have a heightened inflammatory state.

Table

 Parameter Single Co-treatment
Short-term
(1 day)
Longer-term
(5 days)
Short-term
(1 day)
Longer-term
(5 days)
CRP

CPZ = NC
IL-6, IL-1b = ↑

CPZ = NC
IL-6, IL-1b = ↑
Enhanced increase No statistical enhancement
IL-8 mRNA

CPZ = NC
IL-6 = NC
IL-1b = ↑

CPZ = NC
IL-6 = NC
IL-1b = ↑ (<D1)

No effect with IL-6
IL-1b = ↑

No effect with IL-6
IL-1b = ↑ (>D1)

CYP1A2

CPZ = ↑ ↑
IL-6, IL-1b = ↓↓

IL-6, IL-1b abolished induction by CPZ
CYP3A4

CPZ = ↑ ↑
IL-6, IL-1b = ↓↓

IL-6, IL-1b abolished induction by CPZ
Parent compound and metabolites Metabolites present in lysates and media 

CPZ = ↑
Metabolites = ↓

CPZ = ↑
Metabolites = ↓

NC = no change, ↑ = increased, ↓ = decreased, D1 = day 1, D5 = day 5

High Content Imaging and Analysis Enable Quantitative In Situ Assessment of CYP3A4 Using Cryopreserved Differentiated HepaRG Cells. Ranade AR et al, 2014

Ranade AR, Wilson MS, McClanahan AM, Ball AJ.
J Toxicol. 2014

Background
High content analysis (HCA) allows for multiparametric detection of events in individual cells in situ and is well suited for high-throughput assessment of hepatotoxicity. HepG2 cells are commonly used for HCA assays but these cells lack many metabolising enzymes and transporter functions. This study characterized cryopreserved differentiated HepaRG™ cells for use in HCA applications and compared them with results using HepG2 cells. HCA requires that cells are not tightly packed, a feature that is essential for HepaRG™ cells to retain their polarised morphology and high CYP levels. Therefore, these studies investigated optimal culture conditions for HepaRG™ cells that allowed HCA of differentiated and metabolically competent cell

Investigations
Initial studies showed that HepaRG cells exhibited phenotypic characteristics that were more comparable than HepG2 cells with primary hepatocytes. Characteristics measured included glycogen storage, glutathione content, α-1-antitrypsin and albumin secretion, as well as basal and induced CYP1A2, CYP2C9 and CYP3A4 activities. The authors also measured the effects of different seeding densities and culture times on basal and induced CYP1A2 and CYP3A4 activities.CYP2C9 was present at measureable levels but was not inducible.

After thawing and seeding in 96-well plates, CYP3A4 activities dipped before increasing to 200% of initial values by Day 12 in culture – this indicates that these cells should not be used until Day 3. A cocktail of Hoechst 33342 dye, an antibody against Cytokeratin 19 and an antibody against CYP3A4 enabled visualization of nuclei, epithelial cells and hepatocytes, respectively. HCA image analysis software was used to determine its ability to accurately quantify CYP3A4 induction by rifampicin. High seeding densities (75,000 cells/well) were difficult to analyse but densities of 50,000/well and 25,000/well were both suitable. The percentage of CYP3A4-expressing cells was closely linked to cell density and was increased by rifampicin.

An important observation was that HepaRG cultures have different-sized nuclei, with the hepatocyte population having smaller nuclei than biliary epithelial cells. The authors were able to determine that the hepatocyte and biliary populations within HepaRG culture may be separated from one another during analysis by filtering the data based upon nuclear area. Between 85–90% of the culture consisted of small nuclei cells. This suggests that analyzing only small nuclei cells within HepaRG™ cultures provides a greatly enriched population of hepatocytes for analysis and excludes the biliary epithelial cells. This saves both time and resources and frees up fluorescent imaging channels to examine other features of interest within the cells.

Changes in expression levels of CYP3A4 in HepaRG™ were detected with a good sensitively by HCA analysis, evident as changes in the cellular fluorescence intensity. Optimal culture conditions for HCA assessment were achieved at a seeding density of 50,000 HepaRG cells/well in 96-well plates. A proof-of-principle study was performed in HepaRG using a variety of cell function and cell tracing reagents relevant to hepatotoxicity.

Conclusions
Many multiplexed imaging-based assays can be performed using HepaRG cells, in which the test cells are metabolically competent and the hepatocyte subpopulation can be analyzed.
There is potential for more dynamic live cell analysis-based applications for hepatotoxicity using microfluidic devices.
Other HCA-compatible markers for cell function could be used to investigate the kinetics of drug metabolism and the cellular events underpinning hepatotoxicity.

 

Hepatitis B and D viruses exploit sodium taurocholate co-transporting polypeptide for species-specific entry into hepatocytes. Ni Y. et al, 2014

Ni Y, Lempp FA, Mehrle S, Nkongolo S, Kaufman C, Fälth M, Stindt J, Königer C, Nassal M, Kubitz R, Sültmann H, Urban S.
Gastroenterology. 2014 Apr.

Background
The receptor-mediated entry of the hepatitis B (HVB) and D (HVB) viruses is hepatocyte-specific, as well as species-specific; however, the identity of the receptor by which these viruses enter the cells was unknown until recently. Using a biochemical approach, others reported the entry receptor to be the uptake transporter, sodium-taurocholate co-transporter polypeptide (NTCP). Therefore these researchers set out to provide conclusive evidence to identify the receptor, using HepaRG™ cells that express the receptor and naive cells (HepG2) that do not.

Investigations
HBV/HDV are enveloped by the L-(large), M-(middle), and S-(small) HBV surface proteins and the N myristoylation and 75 aa of the N-terminus of L-protein are required for infectivity. By using this part of the L-protein, lipopeptides such as Myrcludex B can block infection in vitro and in vivo. The binding of Myrcludex B and its visualisation by immunoflourescent staining - and co-localisation with NTCP - was thus used to identify cells that contained the entry receptor, its correlation with the presence of NTCP, and under which conditions they were maximally expressed (e.g. differentiated or non-differentiated conditions for HepaRG™ cells).

Microarray expression screens of HepaRG™ cells and corresponding upregulated genes after full differentiation (against relevant selection criteria) identified a number of possible receptor genes. Serpin C1 and apolipoprotein A were ruled out because their small hairpin RNA mediated knockout in susceptible cells did not alter the infectivity. Only NTCP fulfilled all selection criteria.

Transduction of HuH7, HepG2, HepaRG™, and the 2 mouse hepatoma cells, Hepa1-6 and Hep56.1D, with a hNTCP-encoding lentivirus caused them to bind Myrcludex B, indicating hNTCP to be an HBV preS-specific receptor. Conversely, when hNTCP was silenced with shRNA-encoding lentiviruses targeting hNTCP-mRNA, infection by HVB and HVD was abolished.

The level of susceptibility to HBV infection was measured according to HBeAg secretion. The infection was high in hNTCP-transduced HepaRG™ cells and was inhibited by MyrB and two monoclonal antibodies against the preS or S-domain of the virus. Infection was dose-dependent and the time course was similar to that in primary human hepatocytes (PHH). The known influence of DMSO on HVB infection in PHH was also demonstrated in hNTCP transduced HepG2 and HuH7 cells such that 70% of transduced HepG2 cells cultured with 2.5% DMSO were infected by HBV (which was NTCP-specific because infectivity was decreased by Myrcludex B).

Inhibition assays measuring the uptake of 3H-labeled taurocholate in Flag-rNtcpeGFP expressing HepG2 cell lines showed that the IC50s for Myrcludex B inhibition of HBV infection (~ 100 pM) and bile salt transport (~ 5 nM) differed substantially, and supported observations that infection inhibition does not require binding saturation of NTCP. The natural substrates, taurocholate, taurodeoxycholate, and taurochenodeoxycholate, also inhibited HVB infection but at non-physiological concentrations. Therefore, under physiological conditions (

Mouse Ntcp binds Myrcludex B; however, primary mouse hepatocytes are not susceptible to HBV/HDV infection. Therefore the infectivity (measured according to HBeAg secretion) of HepG2 and HuH7 cells transduced with mouse Ntcp with HBV were compared with hNTCP-expressing equivalents. The mouse Ntcp transduced cells were resistant to HVB inoculation, indicating that this form of transporter lacks an essential determinant required for infection (in addition to receptor binding). Chimeric forms of hNTCP (from monkey) expressing HuH7 cells failed to bind Myrcludex B and were resistant to infection; whereas, humanized mNtcp-chimera supported HBV infection.

Conclusions
These studies provide robust supporting evidence that human NTCP is a specific receptor for HBV and HDV. The known species-specific entry of HVB and HVD was demonstrated, and the corresponding two short-sequence motifs in human NTCP required for this species-specific binding and infection were identified. This research paves the way for further in-vitro and in-vivo investigations into the mechanisms by which viruses enter cells, for the development of new antivirals.

 

BIOPREDIC International   |   customersinquiries@biopredic.com | Phone: +33 (0)2 99 14 36 14 | Parc d'Affaires de la Bretèche 35760 Saint Grégoire FRANCE

Home  |  Search  |  Site Map  |  Legal disclaimer