The schedule of experimental study is depicted in Figs. 1 and 2. A total of 22 male New Zealand rabbits [KBT Oriental Co., Ltd., Saga, Japan; mean weight 2.5 kg (range 1.9–2.7 kg)] were used. The rabbits were housed at 23+/−1°C under a 12-h light/dark cycle (lights on at 7 a.m.) and 40–70% atmosphere, and were fed 100 g standard diet (LRC4: Oriental Yeast Co., Ltd., Kyoto, Japan) once per day and provided water ad libitum. The rabbits were divided into two groups (n=11/group): Normal diet and fat diet (NASH group). In the normal diet group, two subgroups were established as PVE (n=6) and Sham (n=5) (Fig. 1). In the fat diet group, PVE (n=5) and Sham (n=6) subgroups were established at 4 weeks after feeding a fat diet (Fig. 2).

All procedures were performed under general anesthesia. Animals were placed supine, and then general anesthesia was induced with intramuscular injection of a mixture of ketamine hydrochloride (25 mg/kg Ketalar Intramuscular 500 mg; Daiichi Sankyo Co., Ltd., Tokyo, Japan) and medetomidine chloride (0.5 mg/kg Domitor; Zenoaq, Fukushima, Japan), then maintained with administration of isoflurane (3% Isoflu; DS Pharma Animal Health Co., Ltd., Osaka, Japan) through a mask. Oxygen saturation was monitored using a pulse oximeter (BSM-5105; Nihon Koden, Tokyo, Japan). The study protocol was approved by the Institutional Animal Care and Use Committee of Kagoshima University (Kagoshima, Japan; approval no. MD11117).

Under deep anesthesia and using a fluoroscopic unit (TU-50; Hitachi Ltd., Tokyo, Japan), PVE was performed in the left lobe of each PVE group using 1 ml absolute ethanol (5). Following open laparotomy, the mesenteric vein was harvested and a microcatheter (Snyper 2; Terumo Corporation, Tokyo, Japan) was inserted and advanced to the left lateral portal vein. After verifying placement of the microcatheter in the lateral left lobe using a small amount (1 ml) of contrast medium (Omnipaque 300; Bayer Holding Ltd., Tokyo, Japan), 1 ml absolute ethanol was injected and a 3- or 4-mm-diameter Vortex microcoil was positioned in the lateral left portal vein (Boston Scientific Japan K.K., Tokyo, Japan). Following PVE, it was confirmed that the left portal vein branch had disappeared and the portal vein main trunk was dilated. Rabbits were sacrificed 2 weeks after PVE. For the Sham group, open laparotomy was performed and the suture was closed without any other interventions.

Rabbits were fed a high-fat diet (standard diet + 10% lard + 2% cholesterol) for 4 weeks (6). After feeding the fat diet for 4 weeks, a normal diet was resumed to avoid unintentional liver damage.

Measurements of d-ROM and BAP were performed as previously reported (1). Freshly drawn and heparinized peripheral blood (3 ml) was spun in a low-speed centrifuge to separate the plasma. Measurements of d-ROM and BAP in plasma were performed within 2 h using a Free Radical Analytical System 4 (FRAS4) designed by SEAC s.r.l. (Florence, Italy). For the measurement of d-ROM, 10 µl fresh plasma was mixed with 1 ml of a reagent containing FeCl₂ (Fe²⁺) and diethyl p-phenylene diamine chromogen, then incubated for 5 min at 37°C to develop the oxidized color. The change in optical density at 505 nm was measured over the 5-min reaction time. For the measurement of BAP, 1 ml of an oxyradical colorant comprising a thiocyanate compound and FeCl₃ (Fe³⁺) was reduced by the addition of 10 µl of the rabbit plasma, and the decrease in optical density at 505 nm over the 5-min reaction time at 37°C was measured in the FRAS4 according to the manufacturer's instructions. d-ROM and BAP readings were taken for each sample and used to assess the level of biological oxidative stress.

As rabbits produced a chylous plasma, for which the milky turbidity interfered with the optical densitometry in measurement of BAP and d-ROM, all milky plasma had to be clarified by skimming fatty components with petroleum ether (PE) prior to ordinary assay of BAP and d-ROM in the FRAS4. Prior to pathological assessment of NASH, it was clarified whether oxidative stress markers differed between PVE and Sham groups in normal and NASH rabbit models.

Laboratory data including serum aspartate transaminase (AST), alanine aminotransferase (ALT), triglyceride (TG) and total cholesterol (T-Chol) in the normal and NASH groups were also checked. After the experimental procedure, all rabbits were sacrificed and pathological specimens of the non-embolized liver lobe were investigated for the presence, grade and score of NASH.

Pathological diagnosis was assessed according to the classification of NASH reported by Brunt et al (7), with grade evaluated by dividing lesions into active and potentially reversible injury according to Nonalcoholic fatty liver disease Activity Score (NAS). The proposed NAS also separates three grades: Steatosis, lobular inflammation, and ballooning. Individual features were graded semiquantitatively according to the scale: 0, absent; 1, mild; 2, moderate; and 3, severe. Severity of steatosis was based on the proportion of hepatocytes involved: 1, <33%; 2, 33–66%; 3, >66%. Severity of lobular inflammation was based on the number of inflammatory foci per·x20 field: Score 1, 1–2 foci; score 2, 2–4 foci; score 3, >4 foci. Stage was evaluated as potentially less reversible and characterized by collagen deposition and architectural alterations that may evolve toward more permanent parenchymal remodeling.

Statistical analysis was performed using Student's t-test for continuous variables of paired data. The Mann-Whitney U test was used for continuous variables of unpaired data and the χ² test for categorical variables of unpaired data. MedCalc version 9.5.1.0 software (MedCalc Software bvba, Ostend, Belgium) was used for all analyses. A two-tailed value of P<0.05 was considered to indicate a statistically significant result.

Laboratory data are summarized in Tables I and II. A variable effect was observed on laboratory parameters when comparing the PVE and Sham treatment groups [PVE group: AST was marginally increased, ALT was relatively unchanged, TG was increased and total cholesterol was markedly increased (P=0.0206); Table I]. In the process of NASH creation (fat diet for 4 weeks), T-Chol was markedly increased in the NASH group (P<0.0001). In the NASH group, PVE induced increased AST (P=0.0318; Table II). In addition, significant differences were identified in NASH-related pathological findings including steatosis, ballooning, inflammation, NASH stage and NASH grade in selected samples (NASH, n=6; normal, n=7). The grading of NASH-related pathological findings was more severe in the NASH model than in normal liver (Table III).

Results of sequential changes in BAP and d-ROM values are summarized in Tables IV–VI.

BAP at 2 weeks after PVE was not significantly altered compared with that immediately prior to PVE (P=0.604). Likewise, the level of d-ROM at 2 weeks after PVE was not significantly changed relative to that prior to PVE (P=0.0501; Table IV).

BAP at 2 weeks after no treatment was not significantly altered compared with that immediately prior to the Sham procedure (P=0.6119). Interestingly, the level of d-ROM at 2 weeks after no treatment was significantly higher than that detected immediately prior to Sham operation (P=0.0152; Table IV).

In the process of NASH model creation, BAP was significantly decreased (P<0.0001), but d-ROM value was unchanged (P=0.3197; Table V).

BAP at 7 weeks (2 weeks after PVE) was not significantly altered compared with that at 5 weeks (P=0.5625). The level of d-ROM at 2 weeks after PVE was significantly elevated compared with that prior to PVE (P<0.0001; Table VI).

BAP at 2 weeks after no PVE did not differ significantly to that detected prior to initiation of the high-fat diet (P=0.249). The level of d-ROM at 2 weeks after no PVE was significantly increased compared with that prior to starting the high-fat diet (P=0.0189; Table VI).