The liver is known to be the primary target for adverse drug reactions (ADRs) presumably due to its extensive involvement in drug metabolism and elimination (1). Idiosyncratic adverse drug reactions (IADRs) are considered as an important subset of ADRs, accounting for ∼13% of all acute liver failure cases (2). IADRs represent one of the leading causes for post-marketing drug withdrawal (3), predominantly due to being host-dependent, poorly predicted by standard preclinical or early clinical trials and unrelated to the pharmacologic target of the drug (2). However, a recently developed drug-cytokine co-treatment approach proved to be highly efficient in the early prediction of idiosyncratic drug-induced hepatotoxicity, especially in animal models (4). This approach is consistent with the inflammatory stress hypothesis, which states that “the co-occurrence of an acute episode of inflammation during drug therapy results in sensitization of the liver causing liver injury from an agent that typically would not cause hepatotoxicity” (5). Nevertheless, the underlying mechanisms through which idiosyncratic drugs synergize with pro-inflammatory cytokines to precipitate serious liver injury need further investigation and understanding for a better prediction or even prevention of IADRs. In the present work, we investigated the effect of idiosyncratic drugs on the activity of important hepatic transporters in the absence and presence of an inflammatory context, in the aim of demonstrating that impaired drug efflux and elimination may be an important underlying mechanism of idiosyncratic hepatotoxicity. Hepatic membrane transporters play an essential role in the absorption, distribution, and elimination of both endogenous substrates and xenobiotics (6), therefore possessing the ability to significantly modulate the efficacy and toxicity of pharmacological agents (7). From the several transport proteins found on the canalicular membrane of hepatocytes, we will be focusing on the implication of MDR1 and MRP2 in idiosyncratic drug-induced liver injury. While MDR1 (ABCB1, P-glycoprotein, P-gp) is involved in the transport of a wide spectrum of structurally varying substances (many drugs, but also natural compounds) preferentially large hydrophobic and positively charged molecules, MRP2 is more involved in the efflux of both hydrophobic uncharged molecules and water-soluble anionic compounds (8). It is noteworthy, that drug interactions, multi-drug resistance, and inter-individual variations in drug response have been attributed to altered transporter expression or activity mainly during diseases associated with an inflammatory response, such as hypoxia and cancer (6). The close correlation between altered transporters functionality and inflammatory diseases is probably due to the fact that pro-inflammatory cytokines possess a modulatory effect on the expression and activity of drug transporters (9), thus rendering the liver more susceptible to drug adverse reactions.
In order to elucidate the importance of hepatic efflux transporters in protecting or sensitizing the liver to drug-induced hepatotoxicity, we have investigated in this work the implication of MDR1 and MRP2 in inflammation-associated idiosyncratic drug-induced liver injury by incubating HepG2 cells with four referent idiosyncratic drugs (trovafloxacin, nimesulide, telithromycin, and nefazodone) in the presence and absence of an inflammatory environment. These drugs were specifically chosen because they are known to synergistically induce death when co-administered with a mix containing LPS and TNF-α (4). The effects of these idiosyncratic drugs on the efflux activity of MDR1 and MRP2 were analyzed by capillary flow cytometry using standard fluorescent transport assays.
Material and Methods
Chemicals, Drugs, and Cytokines
Rhodamine 123 (RH 123) and 5 (and 6)-carboxy-2',7'-dichlorofluorescein (CDF) and its diacetate promoiety (CDFDA) was purchased from Invitrogen. Verapamil, Trovafloxcacin, Nefazodone, Nimesulide, and Benzbromarone were purchased from Sigma-Aldrich. Telithromycin was obtained from Tebu-Bio. Unless otherwise noted, the following drug concentrations were used: 450 μM trovafloxacin, 70 μM nefazodone, 450 μM nimesulide, and 175 μM telithromycin. These drug concentrations were selected from initial dosing studies based on the criteria that the drug concentration (i) elicit minimal drug-only hepatotoxicity, (ii) induce robust supra-additive hepatotoxicity synergy with a representative cytokine mix, and (iii) be within a physiologically relevant dosing limit of 100-fold its Cmax value. Tumor necrosis factor-α was obtained from BD Pharmingen and was used at a final concentration of 100 ng/ml. Lipopolysaccharide (LPS) from E. coli serotype 055:B5 was purchased from Sigma-Aldrich and was used at 20 μg/ml.
Cell Staining and Flow Cytometry
The human hepatocellular carcinoma cell line HepG2 was obtained from the American Type Culture Collection (Maryland) and was maintained in DMEM supplemented with 10% FBS, P/S (100 unit/ml and 100 μg/ml) and glutamine (2 mM). Cells were grown at 37°C in a humidified incubator equilibrated with 5% CO2. Cells were trypsinized and subcultured twice a week.
To discriminate between negative and positive events in the analysis, a non-stained control sample from each culture condition always accompanied acquisition of the stained cells to define their cut off. Gates were drawn around the appropriate cell populations using a forward scatter (FSC) versus side scatter (SSC) acquisition dot plot after excluding dead cells based on propidium iodide staining so that only viable cells were taken into consideration. Cytometers performances are checked weekly using the Guava easyCheck Kit 4500-0025 (Merck/Millipore/Guava Hayward, CA). Flow cytometric measurement of MDR1 functional activity using Rhodamine 123 efflux assay. The fluorescent dye rhodamine 123 is a substrate for P-glycoprotein and its transport out of the cell has been demonstrated to reflect P-glycoprotein function. Analysis of variation in rhodamine 123 intracellular fluorescence together with the effect of P-glycoprotein modulating agents (i.e., verapamil) investigates the role played by multidrug resistance protein in idiosyncratic drug-induced hepatotoxicity.
Briefly, HepG2 cells were cultured in 24-well plates at a density of 105 cells for 48 h until they were 80% confluent. Then cells were loaded with 0.5 μM rhodamine-123 (RH123) for 30 min at 37°C in 5% CO2 in the presence or absence of 100 μM Verapamil, a standard P-gp inhibitor and four selected drugs (Trovafloxacin, Nimesulide, Telithromycin, and Nefazodone) (accumulation phase). Cells were then immediately transferred on ice, washed once with ice-cold phosphate buffered saline (PBS), and re-suspended in RH123-free medium for 120 min at 37°C to allow maximum efflux of fluorescent compounds (efflux phase). To be analyzed by capillary cytometry, cells were trypsinized and re-suspended in culture medium. MDR1-mediated efflux of rhodamine 123 was monitored on a Guava EasyCyte Plus capillary flow cytometer (Merck Millipore, Life Science division, Merck KgaA, Darmstadt, Germany) equipped with a 488 nm excitation laser and four emission band pass filters at 530/40, 585/42, 675/30, and 780/30 nm. The accumulated intracellular fluorescence intensity of rhodamine 123 at 530/40 nm was computed on the Guava ExpressPro software (Merck/Millipore/Guava Tech) in terms of x-geometric mean arbitrary units (AU). Dead cells were excluded based on propidium iodide staining. Final concentration of DMSO applied to cells during incubation with tested drugs was 0.5%. In the tested setup, these concentrations had no adverse effects on cell viability, cell morphology, nor on rhodamine-123 efflux. The inhibitory potential of tested compounds on rhodamine-123 efflux was expressed relative to maximum inhibition obtained with 100 μM verapamil in the same experiment.
Flow cytometric measurement of MRP2 functional activity using the CDF efflux assay. MRP2 transport activity was investigated using the 5 (and 6)-carboxy-2',7'-dichlorofluorescein (CDF) and its diacetate promoiety (CDFDA) efflux assay. Briefly, HepG2 cells were cultured in 24-well plates at a density of 105 cells for 48 h until they were 80% confluent; then they were incubated with a medium containing 1 μM CDFDA for 20 min at 37°C in 5% CO2 in the presence or absence of 250 μM benzbromarone, a known inhibitor of MRP2, and four selected drugs (Trovafloxacin, Nimesulide, Telithromycin, and Nefazodone). CDFDA is a nonfluorescent esterified form of CDF that freely diffuses into cells where it is cleaved by esterases to give CDF, a fluorescent dye effluxed by MRP2. The loaded cells were then washed three times with ice-cold medium and incubated in PBS for 120 min at 37°C to allow maximum efflux of CDF. To be analyzed by capillary cytometry, cells were trypsinized gently and re-suspended in culture media. MRP2-mediated efflux of CDF was monitored as described before on a Guava EasyCyte Plus System capillary cytometer. The accumulated intracellular fluorescence of CDF (530/40 nm) was computed on the Guava ExpressPro software in terms of percent of fluorescent cells. Dead cells were excluded based on propidium iodide uptake. Final concentration of DMSO applied to cells during incubation with tested drugs was 0.5%. In the tested setup, these concentrations had no adverse effects on cell viability, cell morphology, nor on CDF efflux results. The inhibitory potential of tested compounds on CDF efflux was expressed relative to maximum inhibition obtained with 250 μM benzbromarone in the same experiment.
The data are expressed as mean ± standard deviation (SD). Statistical analysis was performed using the unpaired t-test. Statistical significance was considered when P < 0.05.
Idiosyncratic Drugs Effects on MDR1 Activity
To evaluate the implication of MDR1 in drug-induced idiosyncratic hepatotoxicity, we have studied the effect of four idiosyncratic drugs on the efflux activity of MDR1 in the presence and absence of LPS and TNF-α. In the absence of an inflammatory context our results demonstrated, when compared to the specific MDR1 inhibitor verapamil, that telithromycin and nefazodone elucidated an inhibitory potential on the efflux activity of MDR1; such inhibition is represented by an increase in the intracellular fluorescence of rhodamine 123-loaded cells (Fig. 1 and Fig. 3). However, co-treatment of hepG2 cells with a pro-inflammatory mix containing TNF-α and LPS along with idiosyncratic drugs for 24 h has noticeably reduced their inhibitory potential on the efflux activity of MDR1 as demonstrated by the decrease in fluorescence of rhodamine 123-loaded cells (Fig. 1B). It should be noted that even verapamil, which is considered as the standard MDR1 inhibitor, revealed a decreased inhibitory potential within an inflammatory context.
Idiosyncratic Drugs Effects on MRP2 Activity
Based on the fact that MRP2 is extensively involved in protecting the liver of potentially toxic xenobiotics, we have investigated its role in idiosyncratic drug-induced hepatotoxicity in the presence and absence of our pro-inflammatory mix. Figure 2 demonstrates that besides trovafloxacin that exhibited a mild inhibitory potential on the efflux of MRP2, none of the studied idiosyncratic drugs inhibited the efflux of CDF when compared with benzbromarone in the absence of an inflammatory context. Contrary to MDR1, MRP2 was strongly implicated in the drug-cytokine-induced hepatotoxicity. Our results demonstrate that trovafloxacin, nimesulide, and to a lesser extent nefazodone noticeably inhibited the efflux activity of MRP2 in an inflammatory context as represented by an increase in the fluorescence of CDF-loaded cells (Fig. 2B). The co-treatment of HepG2 cells with both TNF-α and LPS along with idiosyncratic drugs for 24 h revealed that these drugs possess a potent inhibitory potential that was suppressed in the absence of an inflammatory context, suggesting that inflammation associated idiosyncratic drug synergy may be an effective tool in revealing the roles played by this hepatic transporter in drug-induced liver injury.
Idiosyncratic adverse drug reactions (IADRs) account for the majority of post-marketing drug withdrawal and “black box warnings” (4); nevertheless, the lack of adequately predictive pre-clinical and clinical assays (3) complicated the revelation and understanding of their underlying mechanisms. One of the hypotheses that have emerged to explain IADRs is that inflammatory stress induced by exogenous or endogenous inflammatory agents is a susceptibility factor for the precipitation of idiosyncratic drug-induced liver injury (2). Recently established animal models co-administering bacterial LPS to induce an inflammatory background during drug therapy have succeeded to predict the potential hepatotoxicity of certain drugs (4). However, the low throughput nature of these models necessitates the development of high throughput in vitro predictive models of idiosyncratic drug-induced hepatotoxicity to better understand its underlying mechanisms. Accordingly, we have developed in the present work a drug-cytokine cellular model in which we co-treated HepG2 cells with a mix of pro-inflammatory mediators (LPS and TNF-α) along with several idiosyncratic drugs, aiming at investigating the implication of two important efflux transporters, MDR1 and MRP2, in inflammation-associated idiosyncratic drug hepatotoxicity. We have chosen specifically LPS and TNF-α to induce inflammation for two main reasons: first, to mimic the in vivo situation of previously validated animal models in which the co-administration of minimal doses of LPS with idiosyncratic drugs better revealed their potential hepatotoxicity (3); second, to elucidate any implication of MDR1 and MRP2 in inflammation-associated drug-induced liver toxicity since the LPS-stimulated release of TNF-α is known to modulate hepatic drug transporters expression and activity (9).
Our results revealed that the co-occurrence of an episodic inflammatory reaction during drug therapy modulated in an opposing manner the efflux activity of MDR1 and MRPP2. Telithromycin and nefazodone proved to be potent inhibitors of the MDR1-mediated efflux of rhodamine 123 when compared to verapamil. Nefazodone is known to induce hepatotoxicity by inhibiting BSEP, thus leading to the accumulation of drug and bile acid in the liver (10). To our knowledge, no previous studies demonstrated a link between nefazodone and MDR1; however, our results showed that this drug possesses a potent inhibitory potential on MDR1 probably through a structure-specific interaction with this transporter as is the case with BSEP (10). Telithromycin is known to be a substrate of both MDR1 and MRP2; however, in our results it elucidated an inhibitory potential solely on MDR1. Being both a substrate and an inhibitor of MDR1 suggests that telithromycin blocks the efflux activity of this transporter by competitively inhibiting the extracellular translocation of rhodamine 123 (8). It is noteworthy that the inhibitory potential of nefazodone, telithromycin, and even verapamil proved to be reversible as it was noticeably lost within an inflammatory context, probably due to a TNF-α-stimulated induction of MDR1 protein expression and functionality (11). Concerning MRP2, besides trovafloxacin which exhibited a mild inhibitory effect on the efflux of CDF, none of the tested idiosyncratic drugs modulated the efflux activity of this transporter indicating a poor implication of MRP2 in idiosyncratic drug-induced liver injury. Conversely, the co-presence of pro-inflammatory mediators along with idiosyncratic drugs has markedly potentiated their inhibitory potential on the efflux activity of MRP2 (mainly trovafloxacin and nimesulide) probably due to a TNF-α-induced downregulation of MRP2 protein expression and activity (12). The fact that the hepatotoxicity of nimesulide is often correlated with a remarkable increase in conjugated bilirubin and cholestatic injury (13) explains the observed nimesulide-induced inhibition of MRP2, since the latter is responsible for the canalicular excretion of conjugates, including bilirubin, glutathione, and bile salts (10).
Trovafloxacin proved to cause severe liver injury in animals only after synergizing its administration with an LPS-induced inflammatory stress (2). Consistently, our results demonstrated that trovafloxacin exhibited an inhibitory effect on the efflux activity of MRP2 only within an inflammatory context. The fact that this inhibitory potential was completely hidden in the absence of TNF-α and LPS confirms that the presence of pro-inflammatory mediators is necessary to reveal the toxicity of this idiosyncratic drug. This hepatotoxicity might be closely correlated to the presence of TNF, especially that trovafloxacin pre-treatment in vivo proved to prolong the LPS-induced increase in plasma concentration of TNF (2).
The regulatory pathways by which pro-inflammatory cytokines synergize with idiosyncratic drugs to alter human hepatic drug transporter expression and activity remain to be technically proven. However, important consideration must be given to MAPKs, since these kinases are known to be implicated in the phenotypic effect of pro-inflammatory cytokines (12). Moreover, recent studies proved that type II nuclear receptors, such as pregnane X receptor (PXR), constitutive androstane receptor (CAR), farnesoid X receptor (FXR), PPARα (peroxisome proliferator-activated receptor), and retinoic acid receptor (RAR) play important roles in the regulation of human drug transporters expression and activity in response to both xeno- and endobiotics during inflammation, cholestasis, and cancer (14). Much evidence proved that the regulation of both MDR1 and MRP2 is PXR-mediated especially that several of their inducers such as rifampicin, ritonavir, and saquinavir happen to be also PXR-activating ligands (Teng and Miller 2008). Accordingly, the observed drug-induced inhibition of MDR1 and/or MRP2 efflux activity may be attributed to drug-mediated suppression of PXR. Furthermore, PXR activators are known to minimize inflammation-mediated downregulation of transporters and attenuate cholestatic liver injury; thus, it is perfectly logical that PXR inhibitors may be strongly correlated to altered drug disposition and cholestasis during an inflammatory reaction (6).
In the present work, we showed that micro-volume cytometry is an efficient technique to demonstrate the altered activity of both MDR1 and MRP2 during idiosyncratic drug-induced liver injury. Furthermore, it is noteworthy that the presence of pro-inflammatory mediators during idiosyncratic drug therapy can noticeably modulate such correlation. While MDR1 proved to be more implicated in idiosyncratic drug-induced hepatotoxicity than MRP2 in the absence of an inflammatory context, MRP2 exhibited a noticeable involvement in idiosyncratic drug-induced hepatotoxicity and this solely in the presence of an inflammatory context. Further research is needed to better elucidate the mechanisms through which the idiosyncratic drugs-inflammatory mediators synergy modulate hepatic transporters activity inducing liver toxicity.