A total of 48 male Sprague-Dawley rats (clean grade; 6 weeks old; average weight, 120±20 g) were purchased from the Laboratory Animal Center of the Hebei Medical University. The animals were maintained in the National Grade II Experimental Animal Center in a sterile environment (temperature, 22±1°C; relative humidity, 50±1% and a 12-h light/dark cycle). All animals had access to drinking water and forage. Additionally animals' health and behavior were monitored twice a day at 7:30 a.m. and 6 p.m. by a veterinary officer. In addition, all animal studies were approved and supported by Laboratory Animal Ethics Committee of Fourth Hospital Hebei Medical University (certificate no. 2017026).

The rats were randomly divided into four groups (n=12/group): Control, model, low-dose matrine and high-dose matrine groups. The model, low-dose matrine and high-dose matrine groups were used to generate preneoplastic lesions for the liver model. Briefly, rats were intraperitoneally injected with 200 mg/kg DEN (batch no. 1002279831; Sigma-Aldrich; Merck KGaA) just on the first day, then low-dose and high-dose matrine groups received intragastric administration of matrine (batch no. 0301010007160301; Ningxia bauhinia pharmaceutical Co., Ltd.) at a dosage of 1.25 and 12.5 mg/kg respectively twice per day at 8 a.m. and 5 p.m., which lasted from intraperitoneal injection of DEN to the killing of rats. The interval between two processes was no more than 10 min, so that this experiment can study matrine preventive treatment. While the rats in the model group were given an equal volume of saline solution via gavage. Two weeks after DEN injection, rats received daily intragastric administration of 2-AAF (batch no. 1002176257; Sigma-Aldrich; Merck KGaA) at a dosage of 15 mg/kg for 3 days. Following this, a partial (2/3) hepatectomy was performed. Preoperatively, rats were anesthetized by intraperitoneal injection of 10% chloral hydrate (batch no. 20161212; Tianjin Damao Chemical Reagent Factory) at a dosage of 300 mg/kg, which was much lower than the LD50 of 480 mg/kg. Three days after the surgery, rats continued to be administered 2-AAF for 1 week. Each rat was weighed before each intragastric administration and given the drug according to its weight to ensure a uniform standard for each administration. In the control group, equivalent abdominal opening/closing surgery and intragastric administration of an equal volume of saline were conducted. However, no DEN injection or 2-AAF supplementation was provided and no hepatectomy was performed in the control group. Partial hepatectomy and intraperitoneal injection of drugs always followed the principle of sterility, and 3 rats died due to postoperative intraperitoneal hemorrhage, which was caused by the surgical ligation suture falling off, but no obvious peritonitis was observed. In this experiment, there were 48 rats in total, 4 rats in each group at each time point. The 3 dead rats were distributed among different experimental groups. In each group, at least 3 rats were still survived at each time point, which met the needs of the experimental statistics (24,25). Each group of rats were sacrificed in 3 batches at 2, 4 and 7 weeks after hepatectomy. At the end of the administration period, animals ate and drank freely for 4 h. The rats were anesthetized by intraperitoneal injection of 10% chloral hydrate (300 mg/kg), then they were killed by extirpation of eyeballs and exsanguination (26–28). They were confirmed dead by observation of respiratory and cardiac arrest and the absence of any pain response to needle puncture of the extremities. The livers were then carefully dissected and immediately fixed in 10% formalin for 24 h at 4°C for subsequent experiments (Fig. 1).

Organs were collected, fixed with 10% formalin and dehydrated using a gradient alcohol series (60, 70, 80, 90, 95 and 100% ethanol) followed by two xylene treatments. The samples were embedded in paraffin, then tissues were cut into 5-µm-thick sections. Tissue sections were subsequently deparaffinized in xylene and rehydrated with a gradient alcohol series (100, 70 and 50% ethanol) and stained using H&E. The sections were stained with hematoxylin staining for 5 min at 37°C and 1% eosin for 3 min at 37°C. All liver sections were observed under a light microscope (magnification, ×400).

Immunohistochemistry was performed as described previously (29). Briefly, hepatic tissue was harvested, fixed in 10% formalin and permeated with paraffin wax. Subsequently, tissues were cut into 6–7-µm-thick sections, which was in line with international standards. The sections were then attached to the slides with polylysine membranes and incubated for 4.5 h at 65°C. Following this, sections were dewaxed, rehydrated and treated with hydrogen peroxide to block the endogenous peroxidase activity. Slides were incubated with rabbit anti-rat α-1-fetoprotein (AFP) polyclonal antibody (cat. no. ab46799; 1:80; Abcam), rabbit anti-rat albumin (ALB) monoclonal antibody (cat. no. ab207327; 1:75; Abcam), rabbit anti-rat Notch1 monoclonal antibody (cat. no. 3608; 1:400; Cell Signaling Technology, Inc.) and rabbit anti-rat Hes1 monoclonal antibody (cat. no. BM4488; 1:100; Wuhan Boster Biological Technology Co., Ltd.). Following this, slides were washed, incubated with a ready-to-use horseradish peroxidase-labeled goat anti-rabbit IgG polymer (cat. no. PV-6001; Beijing Zhongshan Golden Bridge Biotechnology Co., Ltd.) at 37°C for 20 min and then subjected to incubation with diaminobenzidine chromagen at 37°C for 20 min. Images were captured using an Olympus BX51T-PHD-J11 light microscope (Olympus Corporation). Immunohistochemical staining was quantified with Image-Pro Plus 5.1 for Windows software (Media Cybernetics, Inc.) using the measurement function. Randomly selected areas were used for analysis by electing 3 slides for each rat and 5 fields for each slide, and combined the number of positive cells with the strong expression of protein and showed the expression intensity of protein as accurately as possible in the form of quantitative data. The number of positive cells and the color intensity of positive cells were obtained, and the immunohistochemical score (IHS) was calculated as follows: A, positive cell number grading: 0–1%=0, 1–10%=1, 10–50%=2, 50–80%=3 and 80–100%=4; B, color intensity of positive cells: Grade 0 (negative), 1 (weak positive), 2 (positive), 3 (strong positive); IHS=A × B (30).

Statistical analyses were performed using SPSS 19.00 (IBM Corp). All data were expressed as mean ± standard deviation and experiments were repeated at least three times. Data were analyzed using one-way analysis of variance followed by Student-Newman-Keuls post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Normal gross morphology and hepatic lobule structures were indicated in the control group. Furthermore, no significant infiltration of inflammatory cells in the manifold area was observed in the control group (Fig. 2). By contrast, the hepatic lobule structures and the structure of liver cell cords were severely damaged, and the liver cell membrane was damaged 2 weeks after partial hepatectomy in model group. Moreover, edema degeneration was observed in a number of hepatocytes and a large number of inflammatory cells infiltrated the hepatic portal area in model group. A total of 4 weeks after surgery in model group, no significant alterations in liver tissue necrosis or inflammation were observed and injury of hepatic lobule structure was still visible, together with hepatocyte edema and necrosis. However, 7 weeks after surgery, a small proportion of cellular cord structures were detected (Fig. 2).

In the low-dose matrine group, a small proportion of the hepatocyte cord structures were restored 4 weeks after operation. 7 weeks after partial hepatectomy, the hepatocyte edema degeneration was reduced and hepatocyte nodule structures were recovered (Fig. 2). In the high-dose matrine group, a higher protective effect was observed in the week 4 and 7 compared with the model group and low-dose matrine group. Furthermore, discernible lobular structures and a large number of normal liver cell structures were indicated in the high-dose matrine group. In addition, edema degeneration of liver cells and inflammatory cell infiltration were reduced in the high-dose matrine group (Fig. 2).

AFP is a glycoprotein expressed in hepatoblast stem cells, hepatic cancer stem cells and hepatoma cells. When the liver is damaged, an increase of AFP often indicates the severity of damage. The continuous increase of AFP may indicate that part of liver tissue has become cancerous (31,32). According to immunohistochemistry results, AFP was distributed in the cytoplasm and nuclei of liver cells (Fig. 3). No AFP expression was detected in the control group. Notably, AFP expression was significantly upregulated at 2, 4 and 7 weeks after surgery in the model group (P<0.01; Fig. 3). In addition, the AFP expression levels in the model groups at 2 and 4 weeks were significantly increased compared with week 7 in the model group (P<0.05 at 2 and 4 weeks; Fig. 3). This indicated that HCC-like lesions occurred at earlier time points.

In the low-dose matrine and high-dose matrine groups, the expression of AFP was significantly decreased at week 2 and 4 compared with the model group; however, it was still increased compared with the control group (low-dose matrine group vs. model group: P<0.05 at 2 weeks and P<0.01 at 4 weeks; high-dose matrine group vs. model group: P<0.01 at 2 weeks and P<0.05 at 4 weeks; Fig. 3). In addition, there was no significant difference in AFP signal between the model group and low-dose or high-dose matrine groups at week 7 (Fig. 3).

ALB is secreted by normal hepatocytes and is often used to distinguish between hepatocytes and hepatoblasts. According to the immunohistochemistry results, ALB was mainly distributed in the cytoplasm of liver cells (Fig. 4). Strong ALB expression was detected in the control group. However, the expression of ALB in the model group was significantly suppressed at week 2, 4 and 7 compared with the control group (P<0.05). In the low-dose matrine and high-dose matrine groups, significantly increased expression was observed at week 2, 4 and 7 compared with the model group (P<0.05). In addition, significant differences were found between the low-dose matrine and high-dose matrine groups. Notably, the expression of ALB in the high-dose matrine group was increased compared with in the low-dose matrine group at 4 and 7 weeks (Fig. 4).

Notch1 is the primary receptor of the Notch signaling pathway and was used as the indicator for the activity of Notch pathway in the present study. According to immunohistochemistry results, Notch1 was primarily distributed in the cytoplasm of liver cells (Fig. 5). No Notch1 expression was detected in the control group. Notably, compared with the control group, Notch1 expression was significantly upregulated at week 2, 4 and 7 in the model group (P<0.01 at 2 weeks, P<0.01 at 4 weeks and P<0.001 at 7 weeks; Fig. 5). This suggested that the HCC-like lesions were accompanied with Notch pathway activation.

In low-dose matrine and high-dose matrine groups, significantly reduced Notch1 expression was observed at week 4 compared with the model group (low-dose matrine group vs. model group at 4 weeks: P<0.05; high-dose matrine vs. model group weeks: P<0.01; Fig. 5); however, no significant differences were found at week 4 between low-dose matrine and high-dose matrine groups. There was no significant difference in Notch1 staining between the model group and either of the matrine groups at 2 and 7 weeks (Fig. 5).

Hes1 is the predominant downstream target gene of the Notch signaling pathway and its expression level was used as the indicator for Notch signaling activity. The present data indicated that Hes1 was primarily distributed in the nuclei of liver cells (Fig. 6). No Hes1 expression was detected in the control group. Notably, compared with the control group, Hes1 expression was significantly upregulated at 2, 4 and 7 weeks after a partial hepatectomy in the model group (P<0.001 at 2 weeks, P<0.01 at 4 weeks and P<0.01 at 7 weeks; Fig. 6).

In the low-dose matrine group, Hes1 expression was not significantly altered compared with the model group. However, in the high-dose matrine group, significantly reduced Hes1 expression was observed at week 2 and 4 compared with the model group (P<0.05 at 2 weeks and P<0.05 at 4 weeks; Fig. 6). There was no significant difference in Hes1 staining between the model group and either of the matrine groups at 7 weeks (Fig. 6).