Alcoholic vs non-alcoholic fatty liver in rats: distinct differences in endocytosis and vesicle trafficking despite similar pathology

All animals received humane care in accordance with the guidelines established by the American Association for the Accreditation of Laboratory Animal Care (AAALAC) and Animal Research: Reporting of In Vivo Experiments (ARRIVE) [11] . All experimental procedures and ethical standards involving animals were reviewed and approved by the Institutional Animal Care and Use Committees at the Veteran’s Administration Nebraska-Western Iowa Health Care System Research Service (IACUC #11-067-07, approved on Aug. 24, 2014), site of the AFLD studies and University of Nebraska Life Sciences Annex (IACUC #956, approved on Oct. 23, 2013), site of the NAFLD studies. Male Wistar rats weighing 175 to 200 g were divided into two groups (AFLD and NAFLD) with 8 pairs in each group to get statistically relevant data (Student’s paired t Test). As described previously [12] for the AFLD model, rats were pair-fed with control or EtOH-containing Lieber–DeCarli diets [13] contained 18 % of total calories from protein, 35 % from fat, 11 % from carbohydrate, and 36 % from ethanol. In the control diet, ethanol was replaced isocalorically with maltodextrin. For the NAFLD model, rats were allowed ad libitum access to pellet diet and water. Rats in the HFD group (ResearchDiets #D08060104) were fed a diet with a caloric formulation of 60 % calories derived from fat (lard; a mixture of mono-, poly- and unsaturated fatty acids), 20 % from carbohydrates (corn starch, maltodextrin), and 20 % from protein (Casein) and rats fed a lean diet (Research Diets #D12450K with a caloric (kcal) composition of 10.0 % derived from fat (lard), 70 % from carbohydrates (corn starch, maltodextrin) and 20 % from protein (Casein). Both NAFLD and AFLD diets were similar for protein and carbohydrate content with a special care of avoiding sucrose and fructose. These diets have been proven to induce AFLD and NAFLD symptoms and disease that model human alcohol induced and non-alcohol induced fatty liver, respectively. Rats in both alcohol and high fat-diet groups were housed at AAALAC certified institutions (Omaha VAMC and UNL) in approved housing facilities and transported to the laboratories for terminal surgical procedures. Rats in the NAFLD groups were fed ad libitum throughout the study; rats in the AFLD group were fed the ethanol diet ad libitum, and the control rat received an equivalent amount of diet that its pair-fed ethanol consumed. On the day of sacrifice, control animals were meal-fed, to minimize variations in feeding patterns [15]. Rats were sacrificed in the morning hours eight weeks after the initiation of the diet regimen after being anaesthetized with 4 % isoflurane gas mixed with oxygen in a 0.9 m³ chamber. Blood samples were collected via the axillary artery, and serum used for analysis. A piece of whole liver was obtained for histological observation (after knotting a small portion of a small liver lobe with surgical polyester thread); the remaining liver was perfused with isotonic buffers containing collagenase (Type 1 V, Sigma #C5138) as described previously [14, 15].

The serum profile was measured by a Vetscan chemistry analyzer (Abaxis, Union City, CA). Serum samples were loaded on the Mammalian Liver Profile reagent rotor and read with VetScan VS2 Chemistry Analyzer. The Mammalian Liver Profile reagent rotor provided the quantitative measurements of alanine aminotransferase (ALT), alkaline phosphatase (ALP), albumin (ALB), bile acids, and total cholesterol.

Paraffin-embedded liver tissue sections were processed for hematoxylin/eosin staining and evaluated for steatosis and inflammation.

Immunohistochemical staining for lipid droplet protein (PLIN2) was performed as described previously [9]. Briefly, paraffin-embedded liver sections were deparaffinized in xylene and rehydrated in ethanol. Following deparaffinization, slides were subjected to antigen retrieval process with 10 mM sodium citrate buffer (pH 6) for 20 min. Sections were incubated overnight with PLIN2 antibody and followed by staining with appropriate Alexa Fluor secondary antibody. Sections were mounted with vectashield mounting medium containing DAPI. Confocal images were acquired using a Zeiss 510META laser scanning confocal microscope.

The extraction of lipids from hepatocytes was carried out according to the procedure of Folch and colleagues [16]. Aliquots of lipid extract were saponified to quantify the triglycerides (TGs) using the TG diagnostic kit (Thermo dimethyl adipimidate (DMA) kit; Thermo Electron Clinical Chemistry, Louisville, CO).

Liver homogenate (20 %) was prepared in 60 % sucrose in TE buffer (10 mM Tris–HCL, 1 mM EDTA, pH 7.4) containing a protease and phosphatase inhibitor cocktail (Sigma, St. Louis, MO). Liver post-nuclear supernatant (PNS) fractions were obtained by centrifugation (1000 × g) of the homogenate for 10 min. Liver PNS samples were separated by 12 % SDS-PAGE, blotted on nitrocellulose and proteins were detected with appropriate primary antibodies and then immunoreactive proteins were visualized and quantified using the Odyssey Infrared Imager and associated software.

Human orosomucoid (Sigma, St. Louis, MO) was desialylated by the neuraminidase procedure as described by Oka and Weigel [17]. ¹²⁵I-ASOR was prepared by the procedure described previously [15]. Briefly, 125 μg of ASOR in 200 μl of PBS was reacted with 38 μg 1,3,4,6-tetrachloro-3α,6α-diphenyl-glycoluril dried oxidizing reagent coated on the bottom of a glass tube and 0.3 mCi Na¹²⁵I at room temperature for 15 min. The labeled ASOR was then separated from unincorporated Na¹²⁵I with a PD-10 column and quantified by the Bradford assay.

Assessment of internalization of ASOR was determined by a modification of our previous studies [15]. Briefly, hepatocytes isolated from AFLD and NAFLD animals were plated in triplicate on collagen coated 24-well plates at a concentration of 10⁵ cells/well and incubated at 37 °C with 1.0 μg/mL ¹²⁵I-ASOR (25nM) for various periods of time (0, 30, 60, 90 & 120 min). Receptor specificity was measured in replicate wells containing excess unlabeled ASOR to assess background levels. After each incubation period, cells were washed with phosphate-buffered saline and the amount of radioactive ligand in cell lysates was determined and normalized to total cell lysate protein.

For quantification of mRNA, RNA was isolated from liver pieces using a PureLink RNA Mini Kit (Invitrogen, Carlsbad, CA) and was reverse transcribed from 1 μg of total RNA using oligo-dT primers and the Transkriptor kit (Roche Applied Science). To determine gene expression levels, real-time PCR reactions were performed using rat-specific primers from the TaqMan Gene Expression Assay System (Rab3D; catalog # rn00756153; Rab 7, catalog # rn00592246; Rab18, catalog # rn01526466; ASGPR, catalog # rn00560750) and samples were analyzed in the 7500 Real Time PCR System (Applied Biosystems, Carlsbad, CA). The ∆∆Ct method was used to determine the fold change using actin for normalization.

Metabolic hormones; leptin, amylin and insulin were measured using the Multiplex MAP Magnetic Bead-based immunoassay kits (Millipore Corp. Billerica, MA). The assay was conducted according to the manufacturer’s instructions using handheld magnetic separator block for 96-well flat bottom plates (Millipore, Millipore Corp) and analyzed using the Luminex 200 system (Luminex Corp., Austin, TX). All samples were run in duplicate and standards supplied by the manufacturer were run on each plate. Mean fluorescence intensity was analyzed using the BioPlex manager software version 5.0 (Bio-Rad, Hercules, CA) and compared to a standard curve to calculate the concentrations. Values below the range of the standard curve were set to the lower limit of detection.

Pro-oxidant formation was assessed by measuring protein carbonyl content as described [18] using 2, 4-dinitrophenylhydrazine and calculated using an extinction coefficient of 22 mM⁻¹ cm⁻¹. In addition, the extent of lipid peroxidation was assessed by quantitating thiobarbituric acid reactive substances (TBARS) following the procedure of Uchiyama & Mihara [19] using malondialdehyde (MDA) as a standard. Anti-oxidant defense was quantified by measuring total and oxidized glutathione (GSSG) using the enzymatic method [20].

At the outset of these experimental procedures, we confirmed those rats on the Lieber-DeCarli (EtOH) diet and their cohorts on the high fat diet (HFD) had fatty liver by H&E staining of the liver tissue (Fig. 1). The increased fat content was also quantitated by measuring triglyceride content in purified hepatocytes. The amount of triglyceride was similar between the control groups for both AFLD and NAFLD animals, and in the alcohol-fed and high-fat groups, indicating that induction of fatty liver by both diets were equivalent (Fig. 1e). PLIN 2 staining (a lipid droplet marker) also revealed increases in LD accumulation between controls and experimental groups (Fig. 2).

Serum from each animal was tested for markers of liver damage using the Liver Profile Rotorary disc (Abaxis). We measured alkaline phosphatase (ALP) is a marker for a number of disorders including blocked bile ducts [21] and some forms of cancer [22], but in our models, it is a sign of cholangiocyte stress and/or damage. ALP levels in the experimental animals (EtOH-fed or HFD) were significantly higher than the matched controls (Fig. 3a). Alanine aminotransferase (ALT) is an indicator of hepatocyte cell death [23] was also increased in both the high-fat groups (Fig. 4b). Bile acids (Fig. 3c) and cholesterol (Fig. 3d) were significantly higher in the EtOH-fed rats, but no significant differences were identified between high-fat and lean controls in NAFLD model. These results suggest that the synthesis of lipids from alcohol metabolic precursors has a greater impact on cholesterol biosynthesis than the absorption of fatty acids and esterification of triglycerides [24, 25]. Levels of albumin, an abundant serum protein produced by hepatocytes, were not affected in either model (Fig. 3e).

We also analyzed pancreatic hormones level in serum using the Rat Metabolic Hormone Panel provided by Milliplex systems. Insulin resistance has been shown to be associated with NAFLD/AFLD [26‐28] and, not surprisingly, our NAFLD model demonstrated hyperinsulinemia (Fig. 4a). In contrast, despite similar glucose levels observed in both control and ethanol rats in the fed condition (data not shown); insulin levels in the serum of EtOH-fed rats were lower than the controls. In NAFLD animals, along with increased insulin in the serum (Fig 4a), we also observed increased glucose (data not shown). These values were obtained in fed animals, and further studies on insulin resistance and determination of HOMA-IR will be important to examine in fasted animals. Like insulin, amylin is a peptide hormone produced by pancreatic beta cells and is co-secreted with insulin to decrease gastric emptying and increase satiety. Levels of amylin were unchanged in HF-fed rats suggesting that pancreatic beta cells are normal. In AFLD rats, the levels were significantly decreased in the EtOH-fed rats compared to the dietary control (Fig. 4b). The satiety hormone, leptin, was increased in NAFLD rats and unchanged in AFLD rats (Fig. 4c). These results may be reflective of the administration of the diet which is ad libitum for NAFLD and calorie restricted for the AFLD diet or the effects of ingested alcohol on pancreatic function.

Next, we focused on hepatocyte endocytosis in the isolated primary cells from the two models. Our previous work has shown that alcohol impairs multiple aspects of the process of receptor-mediated endocytosis (RME), using the ASGPR as a model (8–11, 15). In the present studies, we used a radiolabeled ligand (¹²⁵1-Acid glycoprotein or orosomucoid (ASOR)), we found that ASGPR internalization rates with cargo in AFLD are 50 % of the control (Fig. 5a), in contrast to NAFLD model in which endocytosis rates are unchanged between control and HF-fed rats. Follow-up binding studies at 4 °C, indicate that the level of surface receptors is lower in AFLD hepatocytes (Fig. 5b), but not in NAFLD hepatocytes. ASGPR protein and mRNA levels in the AFLD rats were decreased by about 30 and 45 %, respectively. In NAFLD rats, ASGPR levels were slightly increased as compared with the control rats, but these differences were not statistically distinguishable (Fig. 5c & d).

In previous studies, we have identified altered protein content for several Rab GTPases in the alcoholic fatty liver [9]. In particular, we identified decreased content for Rab3D (involved in exocytosis), Rab7 (involved in endocytosis/autophagy) and Rab18 (involved in Golgi-endoplasmic reticulum transfer). In the current study, we measured content of these proteins in livers of AFLD and NAFLD animals (Fig. 6a). Similar to what we have previously identified, ethanol treatment significantly decreased the content of Rab3D (75 %), Rab7 (25 %) and Rab18 (16 %) in liver (Fig. 6b). mRNA for Rab3D, Rab7 and Rab18 were also significantly impaired in the alcohol-fed, but not the high fat, non-alcoholic animals (Fig. 6c). In contrast, HFD animals in the NAFLD model did not show any significant change in Rab7 and Rab18, and increased Rab3D content. The mRNA profile for the NAFLD animals reflected the protein expression (Fig. 6c), and was unchanged.

Oxidative stress and lipotoxicity are known to occur in NAFLD [29] and AFLD [30]. In regards to the oxidative stress in both comparative model systems, we observed a significant increase in protein carbonyls content, TBARS (a by-product of lipid peroxidation) and oxidized glutathione (GSSG) in livers of both ethanol-fed and high-fed animals compared to their respective control-fed animals (Fig. 7a–d). In contrast, ~25 % lower content of reduced glutathione (GSH) in livers of both EtOH-fed and HFD rats compared to their controls was observed.