Cerium oxide nanoparticles display antilipogenic effect in rats with non-alcoholic fatty liver disease

Non-alcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease worldwide, ranging from steatosis to non-alcoholic steatohepatitis (NASH). Recently, cerium oxide nanoparticles (CeO2NPs) have emerged as a new antioxidant agent with hepatoprotective properties in experimental liver disease. The aim of the current investigation was to elucidate whether CeO2NPs display beneficial effects in an experimental model of NAFLD.Therefore, fifteen Wistar rats were subjected to a methionine and choline deficient diet (MCDD) for 6 weeks and intravenously treated with CeO2NP or vehicle during the weeks three and four of the diet. The effect of CeO2NPs on serum biochemistry, hepatic steatosis, inflammation, fatty acid content and expression of reactive oxygen species (ROS) and lipid metabolism related genes was assessed. MCDD fed rats showed increased inflammation, enhanced hepatic lipid accumulation of both saturated and unsaturated fatty acids (FAs) and overexpression of genes related to fatty liver and ROS metabolism. Treatment with CeO2NPs was able to reduce the size and content of hepatocyte lipid droplets, the hepatic concentration of triglyceride- and cholesterol ester-derived FAs and the expression of several genes involved in cytokine, adipokine and chemokine signaling pathways. These findings suggest that CeO2NPs could be of beneficial value in NAFLD.

Body weight, liver to body weight ratio and serum biochemical parameters in control and MCDD rats treated with vehicle or CeO 2 NPs. In parallel to results previously obtained by other groups 38,40 , MCDD animals showed significantly decreased body weight and increased liver to body weight ratio than control rats (Table 1). Moreover, they also showed a remarkable alteration in plasma biomarkers of liver function. As shown in Table 1, rats fed with the MCDD displayed increased activity of transaminases, hypocholesterolemia, hyperbilirubinemia and significantly decreased levels of circulatory triglycerides. However, we were unable to detect any significant difference between rats treated and non-treated with CeO 2 NPs in any of the parameters assessed.
Histological examination of steatosis, inflammation and fibrosis in liver tissue. Figure 2A illustrates representative images of H&E, CD68 and Sirius red staining in liver biopsies of control and MCDD rats receiving vehicle or CeO 2 NPs. Macrovesicular steatosis was observed in both groups of MCDD rats as single large fat intra cytoplasmatic droplets displacing the nucleus. This alteration, consistent with a well-defined histological pattern of NAFLD, was significantly less pronounced in MCDD rats receiving CeO 2 NPs. Actually, the morphometric measurement of fat revealed a significant decrease of both, lipid content (48,91 ± 3,61 vs. 42,67 ± 5,75; %p < 0.001) and fat size (69 ± 6 µm 2 vs. 63 ± 9 µm 2 , p < 0.001) (data not shown) in rats receiving CeO 2 NPs compared to those receiving vehicle (Fig. 2B). In addition, the MCDD also resulted in a significant inflammatory Hepatic lipid peroxidation. In order to evaluate the oxidative stress-induced damage in the MCDD model of NAFLD and the antioxidant effects of CeO 2 NPs, lipid peroxidation was assessed by measuring malondialdehyde (MDA) content in the liver. A marked increment in the hepatic levels of MDA was found in MCDD rats treated with vehicle as compared to control rats. The level of MDA in the liver of the MCDD rats treated with CeO 2 NPs was significant lower than in those animals receiving vehicle (Fig. 3).  Hepatic lipid profiling. Further information on the metabolic alterations associated with the diet-induced experimental NALFD model was obtained by measuring the principal lipid components in the liver of control and MCDD rats. As shown in Table 2, total FAs in TG, CE, PC and PE showed marked differences between control and MCDD rats. As anticipated, the liver content of total TG-and CE-derived FAs was markedly increased in MCDD rats in comparison to control animals. However, these differences were not seen on analyzing PC-and PE-derived FAs. On the contrary, in these cases we observed significantly reduced content of total FAs in the liver of MCDD animals. This can be explained by the lack of methionine and choline in the diet of the MCDD group. The effect induced by CeO 2 NPs administration on PC-and PE-derived FAs are shown in Tables 3 and 4 respectively. Marked abnormalities were found in both chromatographic patterns of TG-and CE-derived FAs of MCDD rats (Fig. 4). As compared to control animals, the most remarkable differences were, in TG-derived FAs, the presence of high or very high hepatic content of C16:0, C17:0, C18:0, C18:1n9, C18:2n6, C18:3n6, C18:3n3, C20:0, C20:1n9, C20:2, C20:3n6, C20:4n6, C20:5n3, C22:1n9 and C22:6n3 FAs (Table 5). This was due to a significantly increase of both saturated (SFA) and unsaturated (UFA) FAs. Furthermore, in the latter case this augmentation was a consequence of higher levels of both mono (MUFA) and poly UFA (PUFA). Moreover, the peroxidisability index (PI), an indicator of PUFA peroxidation 41-43 that represents the degree of unsaturation of dietary lipids, was significantly higher in MCDD rats than in control animals (0.95 ± 0.05 vs 0.61 ± 0.06 nmol/mg tissue, p < 0.01). The pattern was quite similar in CE-derived FAs being C14:0, C15:0, C16:0, C16:1, C18:0, C18:1n9, C18:2n6, C18:3n6, C18:3n3, C20:0, C20:1n9, C20:2, C20:3n6, C20:4n6, C22:0 and C22:6n3 the FAs increased in this case. SFA and UFA were also found significantly increased in MCDD in comparison to control rats ( Table 6). The PI was higher too, although did not reach statistical significance (1.08 ± 0.14 vs 0.67 ± 0.01 nmol/mg tissue).
Administration of CeO 2 NPs markedly altered the lipogenic activity in MCDD animals as indicated by a 26% and 33% decrease in the liver content of total TG and CE, respectively. A significant decrease in TG-derived MUFA, almost exclusively due to a diminution in TG-derived oleic acid, was observed (Fig. 5A). The most remarkable effects, however, were noted on analyzing CE-derived FAs. CeO 2 NPs treatment decreased SFA, MUFA and PUFA by approximately 50.7%, 38.7% and 25.6%, respectively (Fig. 5B). In the former case, this diminution was due to a lesser abundance of myristic, pentadecylic and palmitic acids, whereas palmitoleic and oleic acids and linolelaidic and γ-linolenic acids were the principal contributors in MUFA and PUFA, respectively. CeO 2 NPs did not significantly modify the altered PI in MCDD rats. Interestingly, we also observed that CeO 2 NPs induced a significant increase in the CE-derived very long chain PUFA, C22:6n3 (docosahexaenoic acid) ( Table 6).  www.nature.com/scientificreports www.nature.com/scientificreports/ Effect of CeO 2 NPs on fatty liver metabolism related gene expression in liver tissue. Further insight on the effect of CeO 2 NPs in the liver of MCDD rats was obtained by assessing messenger expression of 86 genes involved in fatty liver metabolism using a commercially available PCR array. Table 7 depicts all the genes showing a 2-fold or greater change in expression between the liver of MCDD rats receiving vehicle and that of control rats. Nine genes were significantly upregulated, including Cd36, a gene encoding for an enzyme involved in the adipokine signaling pathway, genes related to metabolic pathways (Abcg1, Apoa1, Ctp1a, Gk and Lpl), the Il1β gene, related to inflammatory response, and apoptosis-related genes (Fas and Serpine1). By contrast, four genes were significantly down-regulated, including those encoding insulin signaling pathway enzymes (Igf1 and Pklr) or controlling other metabolic pathways (Scd1 and Slc27a5).
A 2-fold or greater change in expression with p < 0.05 was considered statistically significant on comparing rats treated with vehicle vs. the CeO 2 NPs treated MCDD rats. Volcano plots of the data are presented in Fig. 6. Interestingly, CeO 2 NPs exerted a significant inhibitory effect on the expression of two genes related to the  adipokine signaling pathway (Cd36 and Lepr); one gene related to the fatty acid oxidation (Cpt1a) and the inflammatory response-related genes Il1β, Il10 and Cebpb. A significant reduction in Igf1 expression was also observed, but the biological significance of this data was roughly less than 2-fold.
Effect of CeO 2 NPs on oxidative stress-associated gene expression in liver tissue. The relative expression of 86 genes from several pathways involved in oxidative stress and antioxidant defense was assessed in the liver of MCDD rats treated with vehicle or CeO 2 NPs using a commercially available PCR array.  www.nature.com/scientificreports www.nature.com/scientificreports/ in expression with p < 0.05 was considered statistically significant on comparing rats treated with vehicle vs. the CeO 2 NPs treated MCDD rats. Volcano plots of the data are presented in Fig. 7. CeO 2 NPs exerted a significant inhibitory effect on the expression of six genes related to antioxidant metabolism (Epx, Gpx7, Gstp1, Prdx2, Prdx4 and Vimp) and four genes related to ROS metabolism (Aox1, Ccl5, Hmox1 and Ncf1). However, an inhibitory effect greater than two fold was only observed on analyzing gene expression of Epx, Prdx4 and Ccl5. To verify the results obtained by PCR array, we used quantitative RT-PCR to assess the expression level of most genes showing differential expression in the presence of CeO 2 NPs. The quantitative gene expression analysis demonstrated paralleled the results previously found in the array profiler. In fact, administration of CeO 2 NPs significantly decreased mRNA abundance of all the assessed genes ( Fig. 8).

Discussion
In the current study we explored the effects of CeO 2 NPs on hepatic steatosis, inflammatory response, oxidative stress and liver fatty acid content in a MCDD-induced animal model of NAFLD. MCDD is among the most commonly used experimental methods to quickly induce liver steatosis and other hallmarks of NAFLD 44 . This diet, with high sucrose and 10% fat, but deficient in methionine and choline, results in macrovesicular steatosis within 3-4 weeks, progressing to inflammation and fibrosis 45 . As in previous investigations 38 , MCDD rats in the present study showed malnutrition, weight loss and a proportional increase in liver weight. Liver injury induced by MCDD was also associated with reduced serum levels of triglycerides and cholesterol, likely due to hepatic blockade of VLDL synthesis. Three weeks after starting the administration of the MCDD, when NAFLD was fully established but without reaching the most severe type of NASH, the iv administration of CeO 2 NPs was initiated. No noticeable effects on serum biochemistry parameters were observed following two weeks of CeO 2 NPs treatment. However, nanoparticles were able to reduce the hepatic fat content and the lipid droplet size in diet induced NAFLD animals. These apparently contradictory results at first may be explained by the fact that probably after www.nature.com/scientificreports www.nature.com/scientificreports/ 6 weeks of MCDD administration NAFLD intensity is already very high, and the significant steatosis reduction induced by CeO 2 NPs is not sufficient to represent a change in the analyzed serum parameters.
It is well known that feeding rats with MCDD increases lipid peroxidation and oxidative stress, which results in hepatocellular damage 45,46 . Actually, in the current investigation we observed a significant increment of MDA concentration in the liver of the MCDD rats, confirming thus the lipid peroxidation in these animals. Interestingly, CeO 2 NPs were able to reduce MDA concentration, indicating a marked reduction in the intensity of the lipid peroxidation. Moreover, this diet also results in remarkably increased hepatic FA accumulation 47 . Thus, we next sought to precisely assess whether CeO 2 NPs modify the FA pattern in animals with NAFLD. As expected, a dramatic difference in liver FA composition was observed between MCDD and control rats. MCDD animals showed between 2 to 30-fold higher FA content in hepatic TG and CE. Interestingly, regardless of which principal lipid component they were derived, the majority of these FAs were MUFA and PUFA. Interestingly, a marked reduction in SFAs and UFAs was observed in our MCDD animals treated with CeO 2 NPs. In view of the alterations induced by CeO 2 NPs, the goal of the next part of the study was to investigate the changes in expression of genes involved in hepatic lipid and ROS metabolism as a result of the MCDD, and to compare these changes to those obtained following CeO 2 NPs administration. The fatty liver and oxidative stress RT 2 Profiler ™ arrays used to determine hepatic gene expression have previously been successfully used in rat liver tissue 30,47,48 . In the present study we identified 14 MCDD-induced genes involved in β-oxidation pathways, adipokine signaling, inflammation and antioxidant and ROS metabolism that were significantly down-regulated or even normalized following   www.nature.com/scientificreports www.nature.com/scientificreports/ CeO 2 NPs administration. One additional gene related to antioxidant metabolism that was not induced by the MCDD was also markedly down-regulated by the administration of CeO 2 NPs. Finally, we also identified two genes involved in insulin signaling and ROS metabolism that displayed decreased mRNA expression in MCDD fed rats, an effect further accentuated when the animals also received CeO 2 NPs. The molecular function of most of these genes, including Gstp1, Hmox1, Igf1, Il1β, Il10, Ccl5, Cebpb, Cd36 and Cpt1a was related to protein binding interaction. However, a more stringent analysis of the alterations in gene expression induced by CeO 2 NPs in MCDD fed animals, considering only those genes showing both statistical and biological significance, revealed two major groups of genes involved in lipid oxidation and cytokine signaling. The former group comprises Epx, Prdx4 and Cpt1a whereas the second group is formed by Il1β, Il10, Ccl5 (RANTES), Cd36, Lepr and Cebpb.
Analysis of the tissue FA profile has become increasingly important in understanding the role of lipids in physiological or pathological processes 49,50 . FAs are the essential components of lipids. In the recent years, it has been suggested that FAs play important roles as intracellular signaling molecules involved for instance in turning on nuclear receptors, including the peroxisome proliferator activate receptors (Ppar) which regulate lipid and carbohydrate metabolism transport and cellular proliferation [51][52][53] . In addition, it has been reported that a high presence of unsaturated FAs could lead to an increased of ROS production and cause mitochondrial permeability leading to apoptosis or necrosis 54 . In this regard, oxidative stress has been proposed as a major mechanism involved in NAFLD pathology. The antioxidant N-acetylcysteine has been found to attenuate the progression of NASH 55 . In the present study, the analysis of FAs hepatic composition showed a dramatic increase of FA in MCDD compared   www.nature.com/scientificreports www.nature.com/scientificreports/ to control rats. Interestingly, the majority of these FAs were UFA (both MUFA and PUFA), which are more vulnerable to oxidation by free radicals and are pointed as a major source of lipid peroxidation 56 . Lipid peroxidation is a mechanism involved in the development and progression of NASH 57,58 triggered specifically by the presence of ROS 59 . Oxidizing agents such as ROS can attack double bonded FAs, especially PUFAs, producing lipids with peroxide and hydroxyperoxide radicals 56 . Lipid peroxides can have lipotoxic effects on mitochondrial DNA, RNA  www.nature.com/scientificreports www.nature.com/scientificreports/ and mitochondrial machinery proteins, thus contributing to mitochondrial dysfunction 60 . Among the secondary products of lipid peroxidation are MDA and 4-hydroxinonenal (HNE) 61 that contributes to liver fibrosis and inflammation 62,63 . In this study we observed a marked increment of lipid peroxidation in the liver of the MCDD rats, and, furthermore, we demostrated that CeO 2 NPs have the potential to significantly reduce this process, thus, attenuating the associated lipotoxic effects. N-acetyl cysteine (NAC) is a widely used antioxidant, often used to protect the liver from oxidative damage caused by ROS. Due to the short half-life of NAC, it is useful to treat acute oxidative stress, however, it is not practical for the treatment of chronic oxidative stress 64,65 . On the other side, CeO 2 NPs remain deposited in the liver for a period of at least 30 days 66 . This property, together with their regenerative nature, would make the use of nanoceria a more interesting antioxidant approach against chronic oxidative stress 66 . Previous studies have compared the biological effects exerted by CeO 2 NPs with that exerted by NAC, demonstrating that CeO 2 NPs have similar antioxidant effects compared to NAC, but a longer half-life. Specifically, CeO 2 NPs showed a trend for greater inhibition of lipoperoxidation and ROS production compared to the NAC-treated animals in mice with liver damage 66 , a similar ability to increase GSH compared to NAC in cells under H 2 O 2 -induced oxidative stress 67 , and a stronger protective effect than NAC reducing ROS production and the apoptosis due to the TNF and cycloheximide administration in U937 cells 68 . Taken together, these studies suggest that CeO 2 NPs antioxidant effects are similar to the exerted by NAC, however, given their long-time deposition in the tissue and their regenerative capacity, CeO 2 NPs could exert a sustained antioxidant effect that may be more useful for chronic oxidative stress treatment.
Previous studies have shown significantly augmented ROS production by liver mitochondria in MCDD rodents. In particular, they exhibit excessive production of H 2 O 2 69 . The marked reduction of both SFAs and UFAs observed in our MCDD animals treated with CeO 2 NPs is likely due to a diminished ROS production derived from the antioxidant properties of these nanoparticles. This would be also consistent with previous data from our laboratory showing that CeO 2 NPs reduce oxidative stress in H 2 O 2 -stimulated human derived cancer cells in culture 30 . In addition, CeO 2 NPs treatment also resulted in a significant increase in docosahexaenoic acid, a modulator of inflammatory response which has been considered as a potential treatment for NAFLD in children 70 . The current investigation provides for the first-time evidence that CeO 2 NPs induce changes in the hepatic gene expression of markers related to fatty liver and oxidative stress in the liver of MCDD fed rats, including Epx, Prdx4, Cpt1a, Il1β, Il10, Ccl5 (RANTES), Cd36, Lepr and Cebpb. Epx and Prdx4 are peroxidases whereas Cpt1a catalyzes the transfer of long chain FAs to carnitine for translocation across the mitochondrial inner membrane [71][72][73] . Moreover, increased Cpt1a expression has been previously observed in MCDD rats 74 . Our findings suggest that by decreasing Cpt1a expression CeO 2 NPs may improve the impairment in FA β-oxidation occurring in NAFLD animals. On the other hand, Il1β is one of the most potent proinflammatory cytokines and Cd36 and Lepr are involved in adipokine signaling. Cd36 mediates the cellular uptake of very long chain fatty acids and is known to be upregulated under hyperlipidemic conditions, contributing to the onset of hepatic steatosis 75 . Moreover, it has been shown that hepatocyte-specific disruption of Cd36 in high-fat diet mice reduces liver CE and TC (the largest being oleic acid) and improves inflammatory markers 76 . Lepr is the receptor for leptin. Ccl5 is a broader activator of several chemokine receptors, including Ccr1, Ccr3, Ccr4 and Ccr5 and it has been claimed as an important player in the pathophysiology of NAFLD-progression 77 . Finally, Cebpb is an important transcription factor that regulates the expression of genes involved in immune and inflammatory response which has been suggested to play a pivotal role in the pathogenesis of experimental NASH 78 . Taken together these results indicate that major beneficial effects of CeO 2 NPs are related to an anti-inflammatory effect secondary to general disruption of cytokine signaling.
In this study the experiments were performed using the MCDD experimental model of NAFLD. This model induces changes in the liver by nutritional deficiency and does not reflect the metabolic profile and the main etiopathogenic factors of human NAFLD [79][80][81] . However, in contrast to other NAFLD models, MCDD induces reproducible histological features of human NAFLD, with significant inflammation, fibrogenesis, and a liver redox balance similar to human patients 37,38 . Therefore, MCDD model is indicated for studying the consequences of fat accumulation, inflammation, oxidative stress and fibrogenesis in the liver 38 . Also, the MCDD model show sex differences regarding the degrees of steatosis and hepatic lipid content being higher in male rats than in female rats. The mechanisms underlying these differences are uncertain 82 . The current study has been performed in male Wistar rats, therefore, further studies would be necessary to confirm that the results are reproduced in females. Liver and spleen are the main target organs of CeO 2 NPs when administered intravenously 30,66,83 , thus, in this study, CeO 2 NPs were administered by iv route the rats to assure a high deposition in the liver. No toxicity or serious side effects were observed due to CeO 2 NPs iv administration, however, given the potential therapeutic value Gk, Glycerol kinase; Hmgcr, 3-hydroxy-3-methylglutaryl-Coenzyme A reductase; Igf1, Insulin-like growth factor 1; Igfbp1, Insulin-like growth factor binding protein 1; Il1B, Interleukin 1 beta; Lepr, leptin receptor; Lpl, Lipoprotein lipase; Mlxipl, MLX interacting protein-like; Nr1h4, Nuclear receptor subfamily 1, group H, member 4; Pck2, Phosphoenolpyruvate carboxykinase 2 (mitochondrial); Pdk4, Pyruvate dehydrogenase kinase, isozyme 4; Pklr, Pyruvate kinase, liver and RBC; Ppard, Peroxisome proliferator-activated receptor delta; Ppargc1a, Peroxisome proliferator-activated receptor gamma, coactivator 1 alpha; Scd1, Stearoyl-Coenzyme A desaturase 1; Serpine1, Serpin peptidase inhibitor, clade E (nexin, plasminogen activator inhibitor type 1), member 1; Slc27a5, Solute carrier family 27 (fatty acid transporter), member 5; Slc2a1, Solute carrier family 2 (facilitated glucose transporter), member 1; Slc2a4, Solute carrier family 2 (facilitated glucose transporter), member 4; Socs3, Suppressor of cytokine signaling 3; Srebf1, Sterol regulatory element binding transcription factor 1; Srebf2, Sterol regulatory element binding transcription factor 2; Tnf, Tumor necrosis factor (TNF superfamily, member 2). *p < 0.05, **p < 0.01, ***p < 0.001 vs. control rats. † p < 0.05 vs. MCDD rats receiving vehicle. Unpaired Student's t-test. (2019) 9:12848 | https://doi.org/10.1038/s41598-019-49262-2 www.nature.com/scientificreports www.nature.com/scientificreports/ of CeO 2 NPs in NAFLD it would be interesting to explore less invasive routes of administration. In this respect, some studies indicate that oral administration of CeO 2 NPs show higher excretion of the nanoceria and less accumulative nanodeposition than intravenous and intraperitoneal administration 66,83 . Even so, liver remains as the main target for CeO 2 NPs deposition after oral administration 84 . Furthermore, anti-inflammatory properties of CeO 2 NPs have been proved when administered intragastrically, preventing and liver injury in an obesity experimental model 33 . Still, further studies are required to enhance nanoceria administration options.
In conclusion, the results of the current investigation show that in the MCDD experimental model of NAFLD characterized by malnutrition, weight loss, reduced serum levels of triglycerides and cholesterol, activation of liver tissue proinflammatory pathways, enhanced liver concentration of FAs and significant overexpression of genes related to fatty liver and ROS metabolism, the administration of CeO 2 NPs reduced the size and content of hepatocyte lipid droplets, the hepatic concentration of TG-derived MUFA, CE-derived SFA and UFA and messenger expression of several genes involved in cytokine, adipokine and chemokine signaling pathways. These findings, therefore, suggest that CeO 2 NPs could be of beneficial value in NALFD.

Animal procedures and dietary induction of liver steatosis in rats.
The studies were performed in 20 male adult Wistar rats (Charles-River, Saint Aubin les Elseuf, France). Fifteen rats were fed with methionine and choline deficient diet (MCDD, TD 90262, Harlan Teklad). Five control rats were fed ad libitum with standard chow (Teklad global 14% protein rodent maintenance diet, Envigo). After 6 weeks of MCDD, rats were euthanized by isofluorane overdose. Livers obtained from each animal were immediately frozen in dry ice and stored at −80 °C for further analysis or fixed in 10% buffered formalin for hematoxylin and eosin (H&E) and immunostaining analysis. Serum samples were also obtained and kept at −20 °C until further analysis.   www.nature.com/scientificreports www.nature.com/scientificreports/ ceo 2 NPs administration in MCDD rats. CeO 2 NPs or vehicle were diluted in saline solution and given as a bolus (500 µl) through the tail vein. CeO 2 NPs (0.1 mg/kg bw) or vehicle (saline solution containing 0.8 mM TMAOH ammonium salts) were injected twice a week for 2 consecutive weeks starting at the third week after liver steatosis induction.
Morphometric measurement of steatosis and fibrosis. Liver sections (4 µm) were stained with H&E and digital images were obtained at a magnification of 100x with a microscope (Eclipse E600; Nikon, Tokio, Japan) and a digital camera (RT-Slider Spot; Diagnostic Instruments, Sterling Heights, MI, USA). For all the cases, the settings of the digital camera, microscope and software were the same. Ten digital images were taken for each slide. Image segmentation was made by selecting a few distinct fat droplets serving as reference. Area and roundness (R) were measured for each object. The formula (4 × π × area/perimeter 2 ) was used to calculate R, which is equal to 1 for perfectly round objects and decreases toward 0 for more irregular objects. Filters were set to exclude exceedingly large objects (~3800 µm 2 ) and objects with low roundness (R ≤ 0.35), which typically represented sinusoidal and vasuclar spaces or optically clear artifacts instead of fat droplets. All the indentified objects were manually inspected to ensure the quality. Fat content was defined as the percentage of total surface area occupied by fat droplets. The results were analyzed using imaging software (ImageJ, National Institutes of Health, Bethesda, MD, USA).
Fibrosis measurement was performed as described previously 85    area was assessed by analyzing 32 fields of Sirius red-stained liver sections per animal. Each field was acquired as described above and the results were analyzed using imaging software (ImageJ, National Institutes of Health, Bethesda, MD, USA). To evaluate the relative fibrosis area, the collagen area measured was divided by the net field area and then multiplied by 100. Subtraction of the vascular luminal area from the total field area yielded the final calculation of the net fibrosis area. The amount of fibrosis measured in each animal was analyzed, and the average value was presented as a percentage 85 . Immunodetection of CD68. Liver sections from fibrotic rats underwent microwave antigen retrieval to unmask antigens hidden by cross-linkage occurring during tissue fixation. Endogenous peroxidase activity was blocked by hydrogen peroxide pretreatment for 10 min and with 5% goat serum for 45 min. The sections were then stained with mouse anti-CD68 (1:150; AbD Serotec, Oxford, UK, RRID:AB_2291300) and incubated for 1.5 h at RT. For antigen detection, the LSAB 2 System-HRP (Dako Denmark A/S) was used, and antigen visualization was achieved with streptavidin peroxidase and counterstained with hematoxylin. Immunostaining was performed without the first antibody for the negative controls. Macrophages (CD68-positive cells) in the middle and margin of the septa were assessed by counting 20 random fields per each section. The mean cell count for each sample was calculated 85 . Hepatic lipid profiling by mass spectrometry analysis. Liver tissue (50 mg) was homogenized in chloroform: methanol (2:1, v/v, Scharlab, Barcelona, Spain, 372978 and PanReact AppliChem, Darmstadt, Germany, 1310911611, respectively) containing 0.005% butylated hydroxytoluene (Sigma Aldrich, Saint Louis, USA, W218405). Glyceryl trinonadecanoate (TG (19:0/19:0/19:0) Sigma Aldrich, Madrid, Spain), cholesteryl heptadecanoate (CE (17:0)), 1,2-dinonadecanoyl-sn-glycero-3-phosphocholine (PC (19:0/19:0)) and 1,2-dipentadecanoyl-sn-glycero-3-phosphoethanolamine (PE (15:0/15:0)) were used as internal standards. Separation of triglycerides (TG), cholesterol esters (CE), phosphatidylcholines (PC) and phosphatidylethanolamines (PE) from total lipid liver extracts dissolved in chloroform was performed by solid-phase extraction (SPE) using aminopropyl silica columns as described previously 86,87 . First, the CE and TG fractions were eluted with chloroform. Thereafter, PC were eluted with chloroform: methanol (3:2, v/v), and finally, PE were eluted with methanol. In order to isolate CE and TG, the first fraction was evaporated under nitrogen stream, dissolved in hexane (Merck Millipore, Darmstadt, Germany, 104374) and transferred to a fresh preconditioned aminopropyl silica column preconditioned with hexane. Then CE were eluted with hexane, and TG were eluted with hexane: chloroform: ethylacetate (100:5:5, v/v).
All solvent fractions containing isolated lipids were dried under nitrogen stream and transesterified (FAME) with 0.5 M NAOH and boron trifluoride (Sigma Aldrich, Saint Louis, USA, B1252) in methanol. GC-MS analyses of FAME were performed on a Shimadzu GCMS QP2010 Ultra instrument (Kyoto, Japan) as previously described 88 . Briefly, final extracts were injected in spitless mode (valve opened at 1 min) into the gas chromatograph interfaced with a mass selective detector. Chromatographic separation was achieved on a Sapines-5MS+ capillary column (30 m × 0.25 mm internal diameter × 0.25 μm film thickness) from Teknokroma (Barcelona, Spain) with helium as a carrier gas at a constant velocity of 50 cm/s. The temperature program was set to begin at 50 °C, maintained at this temperature for 3 min, elevated at 80 °C min −1 to 240 °C, then increased at 2 °C min −1 until 290 °C and finally maintained for 2 min at 290 °C. The ion source and transfer line temperatures were set at 250 °C and 280 °C, respectively. The mass detector was operated in scan mode. Identification of the FAME in the sample extracts was achieved by mass spectrum and GC retention time comparison with reference standards (Sigma). Results are expressed as nmol of FA/mg liver tissue.
Oxidative stress and fatty liver gene expression PCR array in the liver of MCDD rats. Total RNA was extracted using an RNA extraction column kit (RNAeasy, Qiagen, Venlo, The Netherlands). To remove residual DNA, RNA preparations were treated with RNase-Free DNAse set (Qiagen). First strand cDNA was synthesized from 500 ng total of RNA using an RT 2 first-strand kit (Qiagen), and PCR arrays were performed according to the manufacturer's protocols (SABiosciences, Frederick, MD). Real-time PCR arrays were performed using the rat Oxidative Stress RT 2 Profiler TM PCR array, (SABiosciences) and the rat Fatty Liver RT 2 Profiler PCR array according to the manufacturer's protocol. These PCR arrays combine the quantitative performance of SYBR Green-based real-time PCR with the multiple gene profiling capabilities of microarrays to profile the expression of 86 key genes involved in oxidative stress or NAFLD. PCR array plates were processed in a Light Cycler 480 (Roche Diagnostics) using automated baseline and threshold cycle detection. Normalization of gene expression was performed using internal controls to determine the fold change in gene expression between the test and the control samples. The relative quantity of product was expressed as fold-induction of the target gene compared with the reference gene according to the formula 2 −ΔΔCT . Data were interpreted using the SABiosciences web-based PCR array data analysis tool (http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php).