TRIzol^® reagent was obtained from Invitrogen; Thermo Fisher Scientific, Inc. (Waltham, MA, USA). RNase-free DNase was purchased from Promega Corporation (Madison, WI, USA). A Reverse Transcription System kit was purchased from Promega Corporation. The Light Cycler 480 SYBR Green^® I kit was obtained from Roche Diagnostics GmbH (Manheim, Germany). All other reagents were obtained from Sigma-Aldrich; Merck Millipore (Darmstadt, Germany), or as indicated in the specified methods.

Male Imprinting Control Region (ICR) mice (aged 8–10 weeks and weighing 28–30 g) were purchased from Beijing Vital River Laboratories Co., Ltd. (Beijing, China), whose foundation colonies were all introduced from Charles River Laboratories, Inc. (Yokohama, Japan). The animals were allowed free access to food and water at all times, and were maintained on a 12/12 h light/dark cycle with a controlled temperature (20–25°C) and humidity (50±5%) environment for 1 week prior to use. To identify the optimal reference gene in mouse models with ethanol-induced acute alcoholic liver injury, a total of 18 mice were divided into three groups (n=6 per group). The mice received ethanol (5 g/kg) by intragastric administration. The control group received saline (5 g/kg) by intragastric administration. At different time points (6 and 12 h) following intragastric ethanol administration, the mice were weighed and sacrificed. All the mice were sacrificed following fasting for 14 h, and liver and blood samples were collected. Liver tissue was collected and frozen immediately in liquid nitrogen for RT-qPCR analysis, or partially fixed in 4% paraformaldehyde for histological examination.

The present study was approved by the Association of Laboratory Animal Sciences and the Center for Laboratory Animal Sciences at Anhui Medical University (Hefei, China; permit no. 20150349). All procedures on animals conformed to the Guidelines for Humane Treatment set by the Association of Laboratory Animal Sciences and the Center for Laboratory Animal Sciences at Anhui Medical University.

Plasma was obtained from blood collected into tubes, after 2–6 h of storage at room temperature before centrifugation (5,000 × g, 10 min at 4°C). The plasma alanine aminotransferase level was measured using commercial available kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Histological evaluation was performed using hematoxylin and eosin-stained tissue sections (0.5×0.5 cm) and light microscopy. To quantify the extent of necrosis, the percentage of necrosis was estimated by measuring the necrotic area relative to the entire histological section. Analysis of the region was performed using NIH ImageJ software, version 1.44 (National Institutes of Health, Bethesda, MD, USA; http://rsb.info.nih.gov/ij/).

For the detection of nuclear DNA strand breaks, the paraffin-embedded sections were stained with the TUNEL technique using an in situ apoptosis detection kit (Promega Corporation), according to the manufacturer's protocol. The sections were counter-stained with hematoxylin. The TUNEL-positive cells were counted in 12 randomly selected fields from each slide at ×200 magnification with a light microscope. The percentage of TUNEL-positive hepatocytes was analyzed in six liver sections from the six mice in each group.

Total RNA was isolated from the liver tissues of the saline- and ethanol-treated mice using TRIzol reagent. The liver samples were homogenized using 1.2 ml TRIzol reagent per 50 mg liver tissue. DNase I was used to digest and remove genomic DNA contaminants. The RNA purity was determined by measuring the absorbance at 260 and 280 nm with a microplate reader (ELX800, Bio-Tek Instruments, Inc., Winooski, VT, USA). The purity was verified at OD260/OD280 nm, and the ratios of all samples ranged between 1.8 and 2.0. The RNA was stored at −80°C until further analysis.

The RNase-free DNase-treated total RNA (1.0 µg) was reverse-transcribed with AMV (Promega Corporation). In order to improve the reverse transcription, random primers were used. The reactions were incubated at 37°C for 30 min, 65°C for 10 min and 42°C for 60 min, and then diluted to a concentration of 0.5 µg/µl. All cDNA was stored at −20°C until required for the qPCR assay. The RT-qPCR analysis was performed with the Light Cycler 480 SYBR Green I kit using genetic-specific primers synthesized by Invitrogen; Thermo Fisher Scientific, Inc., as listed in Table I. The reaction mixture (20 µl) consisted of 7 µl of 10X buffer, 10 µl of Mix (2X Taq DNA Polymerase, 2X PCR Buffer, 2X dNTP), 2 µl of primer mix (forward and reverse primers) and 1 µl of diluted cDNA. The amplification reactions were performed on a Light Cycler 480 instrument (Roche Diagnostics GmbH), with an initial hold step (95°C for 5 min) and 50 cycles of a three-step PCR (95°C for 15 sec, 60°C for 15 sec and 72°C for 30 sec). For the quantification of primers, a dissociation curve was drawn at the end of the run.

Seven candidate reference genes (Actb, Gapdh, Gusb, Hprt1, 18S, Tbp and B2m) were analyzed. The quantification cycle (Cq) values were transformed into Raw Quantity (RQ) values via the ΔCq method [RQ=2^−(ΔCq)], ΔCq represents each corresponding Cq value - minimum Cq value (24). Two separate sets of independent samples from the control and treated mice were compared using an unpaired one-tailed t-test. Multiple-group comparisons were analyzed using one-way analysis of variance, followed by the Student-Newman-Keuls test. In all samples, P<0.05 was considered to indicate a statistically significant difference. The data are expressed as the mean ± standard deviation. In order to calculate the expression stability of the candidate reference genes, three validation mathematical algorithms were used, including BestKeeper (http://gene-quantification.com/bestkeeper.html), which identifies the appropriate reference gene by paired correlation analysis of all pairs of candidate genes; geNorm (https://genorm.cmgg.be/), which calculates a gene normalization factor based on a pairwise comparison analysis, without considering the experimental conditions; and NormFinder (http://www.mdl.dk/publicationsnormfinder.htm), which is based on a model selection method that enables estimation not only of the overall variation of the candidate normalization genes, but also of the variation between sample subgroups (20). The obtained RQ data were further analyzed with geNorm and NormFinder. BestKeeper analysis was based on the untransformed Cq values. For the rank of all candidate reference genes, the stability values from these three statistical algorithms were analyzed. The comparative Cq-method was used to determine the level of a target gene, normalized to a reference gene and relative to a calibrator (2^−∆∆Cq) using Lightcycler 480 software (Roche Diagnostics GmbH; version 1.5.0) (25,26).

The characteristics of the mice administered with saline or 5 g/kg ethanol intragastrically are listed in Table II. There was no statistically significant difference in mouse weights between the two groups. However, the ethanol-treated mice exhibited significantly elevated liver weights compared with those in the control group (P<0.05). In the ethanol group, the mice in the 12 h group exhibited a significant increase in hepatosomatic index (liver weight/body weight), compared with those in the 6 h group (P<0.05), whereas no significant difference in liver weight was found. The alanine aminotransferase activity was increased in the ethanol-treated group at different time points, and the activity in the ethanol 12 h group was significantly increased compared with that in the control group (P<0.05). Microscopic examination of the livers was performed to verify the damage caused by ethanol and to describe its typical histopathological characteristics. Characteristic hepatocyte necrosis was observed in the liver sections from the mice treated with ethanol (Fig. 1A). The area of necrosis was ~28% at 12 h post-ethanol administration (Fig. 1B). Ethanol-induced hepatocyte death was determined using a TUNEL assay, and numerous TUNEL-positive cells were observed in the livers of the ethanol-treated mice (Fig. 1C and D). The ethanol 6 h group exhibited mild inflammatory infiltration and mild liver cell degeneration, whereas the tissue in the 12 h ethanol group exhibited severe infiltration and hepatocyte necrosis.

Following data normalization, the Cq value was calculated for all PCR reactions. The Cq values for these seven reference genes were calculated for all mouse livers sampled. The median Cq values of the seven reference genes ranged between 11.2 cycles for 18S and 26.6 cycles for Tbp (Fig. 2A–G). Tbp and Gusb had the lowest expression levels, with median Cq values of 24–26 cycles. By contrast, 18S and B2m had high expression levels with median Cq values ranging between 11 and 18 cycles. Actb, Gapdh and Hprt1 distributed intermediate expression levels with median Cq values between 20 and 23 cycles. Among the seven genes, Actb had the maximum range of expression at 3.2 cycles (19.6–22.8 cycles), whereas the minimum range of Gapdh was 1.3 cycles., The range of Cq values within each gene are shown in Fig. 2A–G. The extended vertical bars show standard the minimum and maximum values deviation of the mean in each gene. To an extent, it reflects the expression stability of each reference gene in the different groups.

The present study first assessed the stability of expression for seven reference genes in the control, ethanol 6 h, ethanol 12 h and all groups (Table III). The expression stability of the reference genes was analyzed using geNorm software. When all groups were analyzed, the genes examined exhibited expression stability measures (M values) between 0.11 (Hprt1) and 0.68 (Gusb). Hprt1 and Gapdh had the lowest M values, representing the most stable reference genes/gene pair in the liver in all samples. Reference gene stability analysis revealed that the gene with the lowest variability in all groups was Hprt1; 18S had the highest M value in the other three groups, with the exception of the assessment of ʻall groupsʼ. The result revealed that the least stable gene was 18S. The expression stability of the seven reference genes is shown in Fig. 3A–D

As shown in Fig. 4, the pairwise variations V2/3, V3/4, V4/5, V5/6 and V6/7 were all lower than the limited value of 0.15, which indicated that the combination of the two reference genes with the lowest M values in the experiments was sufficient for normalization. As all pairwise variations were <0.15, the above-mentioned observations remain valid, whether the data were analyzed for each experimental group, or in a single set grouping all data.

NormFinder also analyzed the stability values of the seven candidate reference genes. NormFinder is an Excel-based mathematical tool, which analyzes each sample set separately and also estimates inter-group variations in expression across different sample sets. NormFinder ranks the control genes on the basis of their stability value, where the lower stability value represents higher gene expression stability and vice versa. In the 'all groups' group and in the ethanol 12 h group, Hprt1 was identified as the most stable reference gene (Fig. 5 and Table III).

Hprt1 was considered to be the top-ranked stable reference gene in all groups, whereas B2m and Gapdh were identified as the most stable internal reference genes in the control group, and in the ethanol 6 h and ethanol 12 h groups, respectively (Table III).

Ethanol can directly or indirectly lead to the occurrence of ER stress, and it may change the expression levels of glucose-regulated protein (GRP)78 and of other genes (27,28). It has been reported that the ER chaperone gene ER DNA J domain-containing protein 4 (ERdj4) is upregulated by ER stress (29). Protein disulfide isomerase (PDI) is a resident enzymatic chaperone and its expression is upregulated in ER stress (30). To demonstrate the importance of selecting appropriate reference genes as calibrators in ethanol treatment paradigms, ER stress-associated gene expression was normalized in the present study, with the most stable gene Hprt1 and the least stable gene 18S as the reference genes analyzed by the three mathematical algorithms, geNorm, NormFinder and BestKeeper. The ER stress-associated genes included GRP78, C/EBP homologous protein (CHOP), GRP94, spliced X-box binding protein 1 (XBP1s) total XBP1 (XBP1t), ERdj4 and PDI. Using Hprt1 as the reference gene, the mRNA levels of GRP78, CHOP and PDI were significantly increased in the livers of the ethanol 12 h group, compared with those in the control group, and the difference was statistically significant (P<0.05; Fig. 6). The relative expression levels of the remaining genes, ERdj4, GRP94 and XBP1t, exhibited a gradually increasing trend in the control to the ethanol groups. Using 18S as the reference gene, no statistically significant differences in gene expression were found between the ethanol and control groups. These results suggested that Hprt1 offers an advantage as a reference gene in mouse models of acute alcoholic liver injury.