Membrane Remodeling as a Key Player of the Hepatotoxicity Induced by Co-Exposure to Benzo[a]pyrene and Ethanol of Obese Zebrafish Larvae

The rise in prevalence of non-alcoholic fatty liver disease (NAFLD) constitutes an important public health concern worldwide. Including obesity, numerous risk factors of NAFLD such as benzo[a]pyrene (B[a]P) and ethanol have been identified as modifying the physicochemical properties of the plasma membrane in vitro thus causing membrane remodeling—changes in membrane fluidity and lipid-raft characteristics. In this study, the possible involvement of membrane remodeling in the in vivo progression of steatosis to a steatohepatitis-like state upon co-exposure to B[a]P and ethanol was tested in obese zebrafish larvae. Larvae bearing steatosis as the result of a high-fat diet were exposed to ethanol and/or B[a]P for seven days at low concentrations coherent with human exposure in order to elicit hepatotoxicity. In this condition, the toxicant co-exposure raised global membrane order with higher lipid-raft clustering in the plasma membrane of liver cells, as evaluated by staining with the fluoroprobe di-4-ANEPPDHQ. Involvement of this membrane’s remodeling was finally explored by using the lipid-raft disruptor pravastatin that counteracted the effects of toxicant co-exposure both on membrane remodeling and toxicity. Overall, it can be concluded that B[a]P/ethanol co-exposure can induce in vivo hepatotoxicity via membrane remodeling which could be considered as a good target mechanism for developing combination therapy to deal with steatohepatitis.


Introduction
The significant rise in obesity prevalence in recent decades constitutes an important public health concern worldwide. It exposes a person to several pathophysiological ailments including steatosis defined by an excessive lipid accumulation in hepatocytes [1]. Steatosis dominates liver diseases in countries consuming the western diet, i.e., containing an important amount of fat and/or carbohydrates [2,3]. It is viewed as a benign hepatic lesion but can sensitize hepatocytes towards subsequent aggressions, thereby leading to steatohepatitis, which is characterized by liver cell death, inflammation and recurrent involvement of oxidative stress [4][5][6][7][8][9][10][11]. Furthermore, steatohepatitis can manifest in more severe hepatic diseases like fibrosis, cirrhosis and ultimately hepatocellular carcinoma enzymes responsible for xenobiotic metabolism that resemble those expressed in humans like cytochrome P450 2E1 (CYP2E1) and alcohol dehydrogenase for alcohol or CYP1A for B[a]P [43,[46][47][48][49]. Beside the numerous advantages of zebrafish larva, notably for assessing hepatotoxicant effects [50][51][52][53], this model has already been demonstrated as very useful for studying fatty liver diseases [54][55][56][57][58][59]. In addition, the suitability of this model for studying the involvement of membrane remodeling in chemical-induced liver toxicity has already been reported [33]. Finally, we have recently described that co-exposure of HFD-fed larva to B[a]P and alcohol leads to a steatohepatitis-like state of the liver [27].
In the present study, the objective was, thus, to test the possible involvement of membrane remodeling in the disease progression of steatosis in zebrafish HFD-fed larvae upon co-exposure to two second hits, that is B[a]P and ethanol, at low doses. Therefore, at first, we evaluated steatosis with a Nile red staining; then its evolution towards steatohepatitis-like state after seven days of co-exposure was studied by examination of a histological liver injury and an assessment of the expression of several genes involved in the characteristic features of steatohepatitis-inflammation, cell death, markers of hepatotoxicity and general cellular stress response. In a second part, we evaluated membrane remodeling by staining with the fluoroprobe di-4-ANEPPDHQ, which is sensitive to membrane order. Finally, the involvement of membrane remodeling in this pathological progression of steatosis was explored by using pravastatin-a drug described for its capacity to disturb membrane properties, especially at the lipid-raft level through the inhibition of endogenous cholesterol synthesis.

Progression of High-Fat Diet Induced Steatosis to a Steatohepatitis-Like State in Zebrafish Larvae upon Co-Exposure to B[a]P and Ethanol
We previously developed an in vivo zebrafish larva model with or without steatosis for studying the effects of various toxicants [27,33]. We found that liver steatosis could be induced in zebrafish larvae at 5 dpf, with only one day of HFD that increased oil red o staining, liver size with respect to whole body, triglyceride content, and mRNA level of apolipoprotein A-II (apoa2 and cyp2y3 gene-homologous of the human CYP2E1 gene-in comparison to a standard diet (SD) [27]. In the present study, steatosis was further confirmed by using Nile red staining. The fluorescence ratio of the stained liver of HFD-fed larvae was, indeed, significantly higher compared to the SD-fed larvae liver ( Figure 1A,B), thus indicative of an accumulation of neutral lipids in the liver of HFD-fed larvae.
Following the onset of steatosis, larvae were then exposed to ethanol or B[a]P alone or in co-exposure at sub-lethal concentrations for seven days in order to elicit pathological progression of this disease. For each toxicant, a dose of exposure was chosen in respect to the human level of exposure. Thus, the dose used for ethanol was 43 mM that reached 10 mM (0.46 g/L) inside larvae (data not shown). This concentration is less than drinking guidelines for general populations issued by the International Alliance for Responsible Drinking in 2017 [60]. Considering B[a]P, a concentration range of 0.5-40 nM was obtained in serum from military personnel [61]. Thus, the dose of B[a]P, selected for the present study, was 25 nM. Liver cell damage-a prime characteristic of a progression towards steatohepatitis [6,11,62]-was analyzed by producing histological liver sections of zebrafish larva ( Figure 1C), and damaged hepatocytes-ballooning cells, vacuolated cells and hepatocyte dropouts-were counted in each experimental condition. As shown in the histogram ( Figure 1D), ethanol and B[a]P alone enhanced liver toxicity as visualized by an increased number of damaged cells in comparison to the control, with a further significant effect of co-exposure compared to all other conditions.
The second main characteristic of steatohepatitis-i.e., inflammation [6][7][8][9][10][11]62]-was tested by analyzing the mRNA expressions of various inflammatory gene markers, such as crp, nfkb, il1b and il6, in whole larvae (Figure 2A). mRNA expressions of all four inflammatory markers were significantly higher in larvae co-exposed with B[a]P and ethanol in comparison to control larvae with a more marked effect regarding crp. Ethanol or B[a]P alone were also able to induce these expressions but a further significant induction was observed on crp expression when larvae were exposed to both toxicants. Progression of high-fat diet (HFD) induced steatosis in zebrafish larvae to a steatohepatitis-like state upon co-exposure to ethanol and benzo[a]pyrene. Zebrafish larvae were fed with a HFD from 4 days post-fertilization (dpf) until 5 dpf and compared to larvae fed with a standard diet (SD) in order to observe the development of steatosis at 5 dpf (A,B). Lipid accumulation was analyzed after Nile red staining in HFD larvae as well as in SD larvae using confocal microscopy (excitation/emission (ex/em) wavelength: 488/500-560 nm, magnification ×400). (A) Representative images of larva staining are presented in which the liver has been outlined in white. (B) In order to estimate the relative amount of neutral lipids in the liver, the ratio of fluorescence intensity was calculated from images of more than 15 larvae per diet as follows: Fluorescence ratio = (intensity of neutral lipid staining with Nile red (ex/em wavelength: 488/500-560 nm)/(intensity of unspecific staining (autofluorescence; ex/em wavelength: 405/450-480 nm))). Values are the mean ± standard error of the mean (SEM) of at least 12 larvae per diet. Zebrafish larvae fed with HFD from 4 dpf and exposed to ethanol and/or B[a]P for seven days from 5 to 12-dpf to achieve four conditions-untreated (C) or treated with 25  The second main characteristic of steatohepatitis-i.e., inflammation [6][7][8][9][10][11]62]-was tested by analyzing the mRNA expressions of various inflammatory gene markers, such as crp, nfkb, il1b and Progression of high-fat diet (HFD) induced steatosis in zebrafish larvae to a steatohepatitis-like state upon co-exposure to ethanol and benzo[a]pyrene. Zebrafish larvae were fed with a HFD from 4 days post-fertilization (dpf) until 5 dpf and compared to larvae fed with a standard diet (SD) in order to observe the development of steatosis at 5 dpf (A,B). Lipid accumulation was analyzed after Nile red staining in HFD larvae as well as in SD larvae using confocal microscopy (excitation/emission (ex/em) wavelength: 488/500-560 nm, magnification ×400). (A) Representative images of larva staining are presented in which the liver has been outlined in white. (B) In order to estimate the relative amount of neutral lipids in the liver, the ratio of fluorescence intensity was calculated from images of more than 15 larvae per diet as follows: Fluorescence ratio = (intensity of neutral lipid staining with Nile red (ex/em wavelength: 488/500-560 nm)/(intensity of unspecific staining (autofluorescence; ex/em wavelength: 405/450-480 nm))). Values are the mean ± standard error of the mean (SEM) of at least 12 larvae per diet. Zebrafish larvae fed with HFD from 4 dpf and exposed to ethanol and/or B[a]P for seven days from 5 to 12-dpf to achieve four conditions-untreated (C) or treated with 25 nM B[a]P (B), 43 mM ethanol (E) or a combination of both toxicants (BE,C,D). (C) Liver damage was evaluated on zebrafish liver sections after HES staining (magnification ×400). Black dotted line outlines liver. Histological liver sections were magnified to show, surrounded by the white dotted line, a normal hepatocyte, a vacuolized hepatocyte, a cellular dropout and a ballooning hepatocyte (red arrow). Images are representative of at least five larvae. (D) From images obtained in (C), the histological count of damaged cells was realized. Values are the mean ± SEM of at least five larvae. ## Significantly different from SD larvae; * Significantly different from HFD control larvae; a Significantly different from larvae treated by ethanol only; b Significantly different from larvae treated by B[a]P only.
il6, in whole larvae (Figure 2A). mRNA expressions of all four inflammatory markers were significantly higher in larvae co-exposed with B[a]P and ethanol in comparison to control larvae with a more marked effect regarding crp. Ethanol or B[a]P alone were also able to induce these expressions but a further significant induction was observed on crp expression when larvae were exposed to both toxicants. . Zebrafish larvae were fed with HFD from 4 dpf and exposed to ethanol and/or B[a]P for seven days from 5 to 12 dpf to achieve four conditions-untreated (C) or treated with 25 nM B[a]P (B), 43 mM ethanol (E) or a combination of both toxicants (BE). For the experiments with paracetamol, 1 mM paracetamol was added to the incubation medium containing zebrafish larvae from 5 to 12 dpf. mRNA expression of genes characteristic of inflammation and cell death (A), hepatotoxicity (B,C) and general cellular stress response (D) are shown. Data are expressed relative to mRNA levels found in HFD control larvae, set at 0 (log 2 change). Values are the mean ± SEM. * Significantly different from HFD control The cell death marker, casp3a, was also found to increase significantly in toxicant co-exposed larvae, which is coherent with a classical increase of apoptosis in steatohepatitis [63,64] (Figure 2A). However, a similar increase as for co-exposure was also observed with each toxicant alone. Several studies have described markers characteristic of hepatotoxicity in zebrafish such as tgfb, tfa, zgc163022 in addition to nfkb, casp3a and some other [50,52]. For validation of these hepatotoxic markers in our conditions, we treated zebrafish larvae with paracetamol (1 mM), a well-known hepatotoxic agent, for the same long-term exposure as for B[a]P and ethanol and tested the expression of markers representative of hepatotoxicity ( Figure 2B). Our data clearly indicated that paracetamol significantly altered the expression of these hepatotoxic markers with a rise of tgfb and zgc163022 and an inhibition of tfa expression as already described for short-term exposure [50,56]. Therefore, we decided to quantify the mRNA expression of these hepatotoxic markers in larvae co-exposed with ethanol and B[a]P. As illustrated in Figure 2C, a significant change in expression of tgfb, tfa and zgc163022 was observed, in a similar way as with paracetamol, especially in co-exposed larvae. This thus confirmed the hepatotoxicity of the B[a]P/ethanol co-exposure. Finally, the expression of genes induced in response to general cellular stress, such as nrf2a, nqo1 and gstp1, was also tested since cellular stress is commonly associated with NAFLD and xenobiotic metabolism/toxicity [65,66]. We found that the expression of these genes was significantly augmented with either toxicant ( Figure 2D). In coherence with crp, co-exposure further enhanced, in a significant manner, the expression of all three genes induced in response to general cellular stress in comparison to toxicants alone ( Figure 2D). Together, these results confirm that co-exposure to both toxicants drives the progression of steatosis toward a steatohepatitis-like state even if further investigation will be required to fully confirm the inflammatory state, notably by looking for immune cell infiltration.

Involvement of Membrane Remodeling in the Hepatotoxicity Induced by B[a]P and Ethanol Co-Exposure in Zebrafish Larvae
Previously, our team described the reliability of the zebrafish model to study the effects of hepatotoxicants on plasma membranes [33]. Further, membrane remodeling was identified as a key mechanism of co-exposure to B[a]P and ethanol to induce hepatotoxicity in vitro [40] and for ethanol in vivo [33]. However, the involvement of membrane remodeling in vivo has never been investigated in the context of steatosis progression, notably upon co-exposure with B[a]P and ethanol. In the present study, the impact of such a co-exposure on membrane remodeling was thus determined by analysis of membrane order with the fluorescent hydrophobic probe-di-4-ANEPPDHQ. This allowed us to calculate a generalized polarization (GP) value representative of membrane order that depends on the chemical and physical properties of membranes-lipid composition and packing, fluidity and lipid bilayer thickness, and local hydration. In addition, lipid rafts-specialized membrane microdomains that can also be defined by their high membrane order-could be visualized through membrane areas with high GP values [33,[67][68][69]. After staining the whole zebrafish larvae with di-4-ANEPPDHQ, liver images-characteristic of membrane order-were acquired by computing the GP value obtained from fluorescence images of lipid bilayers with low-membrane lipid orderthe liquid disordered (L d ) phase-and with high-membrane lipid order-the liquid ordered (L o ) phase. It was observed that exposure of zebrafish larvae to ethanol or B[a]P alone has no significant effect on global membrane order in liver cells (data not shown). However, when tested in combination, they enhanced global membrane order in a significant manner in the liver of HFD-fed zebrafish larvae ( Figure 3A). Furthermore, numerous membrane domains with high GP value-defining lipid rafts, possibly reflecting their clustering, were observed at the level of the plasma membrane of liver cells in larvae treated with toxicant co-exposure in comparison to untreated larvae ( Figure 3B). This, therefore, indicated that B[a]P/ethanol co-exposure induced membrane remodeling in HFD-fed larvae.  On the left, some representative liver images of each treatment have been selected according to the respective mean of delta GP (magnification ×400). Pixels with higher GP values (which could be considered as lipid rafts) have been highlighted in yellow to pinpoint lipid raft spatial distribution. The liver area outlined in the white square on the left images are magnified on the right side to show lipid raft spatial distribution in the plasma membrane. Values are the mean ± standard error of the mean (SEM) of at least 25 larvae. * Significantly different from HFD control larvae.

Role for Membrane Remodeling in the Protective Effect of Pravastatin against Co-Exposure-Induced Hepatotoxicity in Zebrafish Larvae
Finally, with the aim of testing the involvement of membrane remodeling in hepatotoxicity produced by B[a]P/ethanol co-exposure, a lipid raft disrupter, pravastatin [70], was used as it was Membrane order in membranes of zebrafish liver was measured by computing the generalized polarization (GP) factor. (A) Changes in GP values were expressed as the difference between individual larva GP value and the mean of GP found in control larvae (∆GP). (B) On the left, some representative liver images of each treatment have been selected according to the respective mean of delta GP (magnification ×400). Pixels with higher GP values (which could be considered as lipid rafts) have been highlighted in yellow to pinpoint lipid raft spatial distribution. The liver area outlined in the white square on the left images are magnified on the right side to show lipid raft spatial distribution in the plasma membrane. Values are the mean ± standard error of the mean (SEM) of at least 25 larvae. * Significantly different from HFD control larvae.

Role for Membrane Remodeling in the Protective Effect of Pravastatin against Co-Exposure-Induced Hepatotoxicity in Zebrafish Larvae
Finally, with the aim of testing the involvement of membrane remodeling in hepatotoxicity produced by B[a]P/ethanol co-exposure, a lipid raft disrupter, pravastatin [70], was used as it was demonstrated to be effective in zebrafish larvae [33] (Figure S1). Pravastatin-a cholesterol synthesis inhibitor-prevented the effects of co-exposure to B[a]P and ethanol on the cell membrane in zebrafish larvae by significantly reducing the membrane order ( Figure 4A). Moreover, it also decreased the impact of co-exposure on the lipid raft spatial distribution in the plasma membrane of liver cells, thus pointing to a prevention of lipid-raft clustering ( Figure 4B). Regarding the consequences in terms of hepatotoxicity, the histological analysis of liver from steatotic zebrafish larvae co-exposed with pravastatin and toxicants showed less liver cell damage ( Figure 4C) compared to larvae unexposed to pravastatin ( Figure 1C). Indeed, this molecule significantly reduced the number of damaged cells ( Figure 4D). The last set of experiments was performed to test the impact of pravastatin on the mRNA expression of the genes altered by B[a]P/ethanol co-exposure. Our data showed that pravastatin prevented alterations in the expression of several genes involved in inflammation (crp, il6), cell death (casp3a) and also one hepatotoxic marker (zgc163022). However, no effect on tfa and on genes related to cellular stress response (nrf2a, nqo1 and gstp1) was detected ( Figure 5A-C). Note that pravastatin alone enhanced the expression of nfkb, il1b and, to a lesser extent, tgfb ( Figure 5A,B). Overall, our results indicated that pravastatin could protect liver from injury induced by toxicant co-exposure, thus indicating the involvement of membrane remodeling and especially lipid-raft clustering in the pathological progression of steatosis upon co-exposure to B[a]P and ethanol. demonstrated to be effective in zebrafish larvae [33] ( Figure S1). Pravastatin-a cholesterol synthesis inhibitor-prevented the effects of co-exposure to B[a]P and ethanol on the cell membrane in zebrafish larvae by significantly reducing the membrane order ( Figure 4A). Moreover, it also decreased the impact of co-exposure on the lipid raft spatial distribution in the plasma membrane of liver cells, thus pointing to a prevention of lipid-raft clustering ( Figure 4B). Regarding the consequences in terms of hepatotoxicity, the histological analysis of liver from steatotic zebrafish larvae co-exposed with pravastatin and toxicants showed less liver cell damage ( Figure 4C) compared to larvae unexposed to pravastatin ( Figure 1C). Indeed, this molecule significantly reduced the number of damaged cells ( Figure 4D). The last set of experiments was performed to test the impact of pravastatin on the mRNA expression of the genes altered by B[a]P/ethanol co-exposure. Our data showed that pravastatin prevented alterations in the expression of several genes involved in inflammation (crp, il6), cell death (casp3a) and also one hepatotoxic marker (zgc163022). However, no effect on tfa and on genes related to cellular stress response (nrf2a, nqo1 and gstp1) was detected ( Figure 5A-C). Note that pravastatin alone enhanced the expression of nfkb, il1b and, to a lesser extent, tgfb ( Figure 5A,B). Overall, our results indicated that pravastatin could protect liver from injury induced by toxicant co-exposure, thus indicating the involvement of membrane remodeling and especially lipid-raft clustering in the pathological progression of steatosis upon co-exposure to B[a]P and ethanol.  Values are the mean ± SEM of at least three larvae. * Significantly different from HFD control larvae; P Significant difference between larvae treated by pravastatin compared to untreated counterparts.  with 0.5 µM pravastatin as quoted as (C ± pravastatin) and (BE ± pravastatin), respectively. mRNA expressions of genes characteristic of inflammation and cell death (A), hepatotoxicity (B) and general cellular stress response (C) are shown. Data are expressed relative to mRNA level found in HFD control larvae, set at 0 (log 2 change). Values are the mean ± SEM. * Significantly different from HFD control larvae; P Significant difference between larvae treated by pravastatin compared to untreated counterparts.

Discussion
Several mechanisms are known to be involved in the toxicity of B[a]P and ethanol towards the liver-oxidative stress, cell death, inflammation and mitochondrial dysfunction [33,36,40,[71][72][73][74]. In addition, another process that has been highlighted in this context is membrane remodeling. In fact, B[a]P was suggested to repress HMGCR (3-hydroxy-3-methylglutaryl-CoA reductase) in vitro via AhR activation and ROS, thus hindering cholesterol synthesis and modulating the lipid content of lipid rafts, finally leading to hepatocyte cell death [36]. Ethanol is also described to alter cell membrane properties by modifying fluidity and lipid-raft clustering in plasma membrane in vitro and in vivo with consequences on cell death and liver injury [33,38]. More recently, co-exposure to B[a]P and ethanol was also shown in vitro in primary hepatocytes to induce membrane remodeling, with consequences in terms of hepatotoxicity [40]. Besides, these same compounds used in combination were demonstrated to induce the progression of steatosis to a steatohepatitis-like state, notably in a model of zebrafish larva. At the same time, HFD-the principal cause of steatosis-was also identified as modifying the physicochemical properties of the membrane by altering its lipid composition or lipid-raft protein activity; it was proposed that this process was involved in NAFLD progression in association or not with hepatotoxicants [41,75,76]. Therefore, in the present study, membrane remodeling was explored in vivo, using the steatotic zebrafish larva model co-exposed with B[a]P and ethanol in order to assess its implications for steatohepatitis development. We found that B[a]P and ethanol, when applied together, significantly altered zebrafish liver cell membrane properties by increasing the overall membrane order in comparison to the control. In parallel, a larger staining of high-ordered membrane domains-showing higher lipid-raft spatial distribution-was also seen in the plasma membranes of larvae liver cells when co-exposed with B[a]P and ethanol, thus emphasizing more lipid-raft clustering. This increase in membrane order and modification of lipid-raft spatial distribution-two indicators of membrane remodeling-were coherent with histological sections, showing more damaged hepatocytes in toxicant co-exposed larvae. However, exact characterization of membrane remodeling-global and local membrane fluidity and lipid-raft microdomain structures-still needs further investigation with special emphasis on B[a]P effects on cholesterol content and discrepancies over its effects in comparison to those observed in vitro [36,40]. Higher membrane remodeling and, notably, the higher level of plasma membrane lipid-raft clustering, suggest alteration of the lipid raft-associated signaling pathway. Several previous studies have proven that alterations in cell membrane properties can modulate several membrane receptors linked with lipid rafts such as toll-like receptors (TLR 2, 4 and 9), which induce, notably through the activation of NF-κB release, a variety of pro-inflammatory cytokines such as interleukins (IL-1β and IL-6) and TNFα [39,41,62,77,78]. Here, mRNA expression of several inflammatory and hepatotoxic markers was found to be increased including il1b, il6 and nfkb. The simultaneous membrane remodeling, hepatocellular damage, and increase in inflammatory markers associated with lipid rafts, therefore, suggested a link between hepatotoxicity of co-exposure to B[a]P and ethanol and membrane remodeling.
The participation of membrane remodeling in co-exposure-induced hepatotoxicity was thus tested in zebrafish larvae by assessing the impact of pravastatin; this molecule is indeed a known lipid-raft disrupter. The addition of pravastatin along with B[a]P and ethanol counteracted the effects of toxicant co-exposure on membrane order ( Figure 4A) and prevented changes in lipid raft spatial distribution ( Figure 4B). Histological analysis then showed less hepatocyte damage, likely due to the protective action of pravastatin on the hepatocyte cell membrane. To confirm the role of membrane remodeling, we tested pravastatin in vitro on the WIF-B9 hepatic cell line-an in vitro model of well differentiated hepatocytes [33,[79][80][81]. This cell line exhibited a similar type of results as in the zebrafish larvae model; indeed, pravastatin decreased the number of apoptotic cells induced by toxicant co-exposure in a steatotic state ( Figure S2). These protective effects towards cell death were further supported by the fact that pravastatin in zebrafish larvae prevented the effects of co-exposure on the mRNA expression of casp3a and the hepatotoxic marker zgc163022. Besides the protection afforded towards cell death, this molecule also inhibited the increase in mRNA expression of inflammatory markers, namely crp and il6, in line with the previously described role for lipid rafts in steatohepatitis-related inflammation [41,77]. Altogether, these results, therefore, indicate that pravastatin would decrease the observed hepatotoxicity by counteracting membrane remodeling, thereby further endorsing the contribution of membrane remodeling as a key player in the pathological progression of steatosis induced by a mixture of toxicants such as B[a]P and ethanol. One might have argued that the protective effect would have been through induction of cyp1a expression or through a decrease of ethanol metabolism via inhibition of cyp2y3 expression, the zebrafish homolog of CYP2E1, or through NRF2 pathway activation. Indeed, it has been previously reported that CYP1A1 can afford some protective action against NAFLD in dioxin- [82] or B[a]P-exposed mice [22]; besides, it is known that CYP2E1 is involved in ethanol toxicity [83], even in the zebrafish model [54]. Similarly, NRF2 pathway activation may also provide protection through xenobiotic metabolism or via an action against oxidative stress [84][85][86]. However, in our model, pravastatin had no significant effect on cyp1a or cyp2y3 expression (supplementary Figure S3A,B) nor on nrf2a and its regulated genes ( Figure 5C), thus further reinforcing a key role for membrane remodeling.
Although, pravastatin appeared to prevent hepatotoxicity, it seems that this protection would be only partial in our model; indeed the alterations observed in the mRNA expression of several genes were not all blocked. Henriksbo and Schertzer [87] have previously described the impact of pravastatin per se on inflammatory markers such as CRP, IL-1β and IL-6. Whereas they reported a decrease in CRP and IL-6 mRNA expression, which is clearly in favor of a protection afforded towards inflammation, they also found that pravastatin increased the expression of IL-1β. Quite a similar result was obtained in our study with an increased mRNA expression of il1b upon pravastatin; likewise, an increase in nfkb expression (another inflammatory marker) was detected. This increase in some inflammatory mediators/regulators might suggest exacerbation of inflammation in the liver and/or whole larvae by pravastatin. A relatively similar type of finding regarding the liver was previously observed with statins by others [88,89]. Such a proper effect of pravastatin might explain why no protective effect of this compound towards toxicant co-exposure impact on these genes could be observed ( Figure 5A). Regarding the other genes studied, that is tfa and tgfb, both previously shown as hepatotoxic markers [50,56], no effect of pravastatin towards co-exposure effects was detected. As already mentioned, B[a]P and ethanol could produce hepatotoxicity via several mechanisms. Based upon our results, pravastatin via cholesterol synthesis inhibition appeared to prevent toxicant effects on membrane remodeling, which thus pointed to membrane remodeling as being involved in steatosis progression. However, even though such a mechanism would be involved, we cannot yet exclude other mechanisms for the action of statins such as an effect on mitochondrial fatty acid oxidation [90], SREBP-2 (sterol regulatory element binding transcription factor 2) induced autophagy [91], and others [92][93][94], to be effective in NAFLD [95][96][97][98][99].
Overall, this study shows for the first time that toxicant co-exposure can favor the progression of liver steatosis towards a steatohepatitis-like state by inducing membrane remodeling, which is involved in both cell death and inflammation. This mechanism can be switched off by a lipid-raft disrupter. Therefore, this mechanism could be considered as a good target in addition to other mechanisms-oxidative stress, inflammation, apoptosis and fibrosis [63]-for developing combination therapy to deal with steatohepatitis.

Zebrafish Larvae Handling and Exposure
Animals were handled, treated and killed in agreement with the European Union regulations concerning the use and protection of experimental animals (Directive 2010/63/EU). All protocols were approved by local ethic committee CREEA (Comité Rennais d'Éthique en matière d'Expérimentation Animale, Rennes, France; approval number R-2012-NP-01). Fertilized zebrafish embryos-collected following natural spawning-were obtained from the Structure Fédérative de Recherche Biosit (INRA LPGP, Rennes, France). Embryos and larvae-sex unknown-were raised at 28 • C according to standard procedures and as previously described [33]. From 4 dpf until the last day of treatment renewal-at 9 dpf-larvae were fed daily with a SD, 10% of fat (Tetramin, Tetra, Blacksburg, VA, USA), or with a HFD made of chicken egg yolk,~53% of fat (Sigma-Aldrich, St. Louis, MO, USA), for 1 h before medium renewal. Both diets were also previously used in zebrafish [27,100,101]. At 5 dpf, larvae were exposed with 43 mM ethanol directly added to the incubation medium and/or by 25 nM B[a]P in dimethyl sulfoxide (DMSO)-DMSO final proportion: 0.001% v/v-or by this vehicle only until 12-dpf. For experiments with pravastatin, 0.5 µM pravastatin (Sigma-Aldrich), was added along with toxicants simultaneously; for experiments with paracetamol (1 mM; Acetaminophen; Sigma-Aldrich), this was added to the incubation medium.

Neutral Lipid Staining with Nile Red
At 5 dpf, after 24 h of feeding, zebrafish larvae were washed in phosphate buffered saline (PBS) and then fixed in 4% paraformaldehyde in PBS at 4 • C. A staining protocol of neutral lipids in liver with Nile red was adapted from previous works [59,102]. After washing in PBS, whole larvae were stained for 1 h with Nile red at 5 µg/mL (N3013, Sigma-Aldrich; stock solution was prepared at 100 µg/mL in acetone). Then, larvae were washed twice in PBS and mounted on slides with PBS. Images of zebrafish larvae were acquired with a confocal fluorescence microscope LEICA DMI 6000 CS (Leica Microsystems, Wetzlar, Germany). To evaluate neutral lipid content, a first image-characteristic of neutral lipid fluorescence-was taken under excitation at 488 nm using an argon ion laser with a photomultiplier tube (PMT) range of 500-560 nm (image A) whereas a second image-insensitive to neutral lipids-was taken under excitation at 405 nm with a diode laser with a PMT range of 450-480 nm (image B) (magnification ×400). Using Fiji imaging processing software(ImageJ, [103]), fluorescence intensity per liver area was calculated for both images; finally, the fluorescence ratio of image A to image B was determined.

Histological Analysis of Liver Toxicity in Zebrafish Larvae
Histological analysis was performed as previously described [27]. Briefly, after treatment, larvae at 12 dpf were washed in PBS and then fixed in 4% paraformaldehyde in PBS at 4 • C before being embedded in paraffin. Then, 5 µm sections were stained with hematoxylin, eosin and safran red (HES) and imaged on a Nanozoomer NDP (Hamamatsu Photonics K.K., Hamamatsu, Japan) (magnification ×400). A histological count of dead/damaged cells was performed from images (two or three sections) of at least three larvae per condition. Damaged/dead cells were counted as cellular dropouts [104], ballooning cells [105], and vacuolated hepatocytes [51].

Analysis of Gene mRNA Expression
Analysis of gene mRNA expression was performed as previously defined [27]. For mRNA extraction, 10-20 larvae were pooled and homogenized in 100 µL TRIzol reagent and total RNA was extracted according to the manufacturer's protocol with TRIzol reagent. RNA samples (1 µg) were then reverse-transcribed using the High-Capacity cDNA Reverse Transcription Kit (Life Technologies, Carlsbad, CA, USA). Quantitative reverse transcription polymerase chain reaction (RT-qPCR) (5 ng of cDNA per well) was performed using SYBR Green on the CFX384 Touch Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). mRNA expression was normalized by means of actb2, 18s and gapdh mRNA levels. The ∆∆Ct method was used to indicate the relative expression of each selected gene. Sequences of the tested zebrafish primers are provided in Table 1.

Membrane Order Determination by Fluorescence Staining
Plasma membrane order in zebrafish liver was assessed, as previously defined [33], by confocal fluorescence microscopy using the membrane order-sensitive fluorescent probe, di-4-ANEPPDHQ (Molecular Probes, Life Technologies). This probe displays a fluorescent spectral blue-shift from 620 nm when incorporated into lipid bilayers with a low-membrane lipid order (liquid disordered phase, L d ) to 560 nm when inserted into lipid bilayers with high-membrane lipid order (liquid-ordered phase, L o ). After acquisition using confocal fluorescence microscopy of both disordered and ordered-phase fluorescence images, a new image, indicative of membrane lipid order, was obtained by calculating the GP value-a ratiometric measurement of fluorescence intensities for each pixel which is associated to membrane lipid order [33,67]. Larva staining was realized as previously described [33]. After staining, they were mounted in 80% glycerol-PBS solution for the observation with confocal fluorescence microscope LEICA DMI 6000 CS (Leica Microsystems, Wetzlar, Germany). Under excitation at 488 nm with an argon ion laser, ordered membrane images were acquired with a PMT range of 500-580 nm, whereas for disordered membrane images the PMT range was 620-750 nm (magnification ×400). Using Fiji imaging processing software (ImageJ, [103]) and the macro published by Owen et al. [67], GP images were generated according to the following calculation: GP = (I500-580 − I620-750)/ (I500-580 + I620-750). In order to avoid potential variation due to the different batches of larva used or to different staining, for each experiment-one batch of zebrafish larvae/one staining procedure-GP values were expressed as the difference between individual larva GP value and the mean of GP found in control larvae (∆GP) within the same experiment.
Lipid-raft spatial distribution: lipid rafts are specialized membrane microdomains that can be defined by their high membrane order. Therefore, they were highlighted by selecting pixels with high GP values in comparison to the mean GP found in the control condition. Overall, the range of ∆GP values for lipid rafts was 0.142 to 0.381 in comparison to the mean value found for all membranes in the livers of the control larvae in the experiment (∆GP = 0). Visualization of the membrane area with high GP value reflects the membrane regions with a local high lipid-raft distribution suggesting a higher level of clustering. Using Fiji imaging processing software (ImageJ, [103]), pixels with a high density were selected in GP images, which highlighted high-ordered membrane domains in yellow through the membrane area of the liver cells. Images presented are pseudo-colored GP images in which ∆GP values are indicated on a colour scale [67].