The study design was approved by the experimental ethics committee of the University of Kafkas (Turkey). All animals were maintained in accordance with university policies. In this study, female mice were used because females show greater sensitivity to FB₁-induced hepatotoxicity compared to males [9]. Ninety BALB/c female mice (University of 19 Mayis, Faculty of Medicine, Turkey) 10 weeks old and weighing an average of 30 ± 1.93 g were divided into six groups (15 females per group kept in separate cages). The mice were housed in stainless-steel wire-mesh cages in a well-ventilated, temperature-controlled room (23 ± 2℃) with 55% relative humidity and a 12-h light/dark cycle. The animals had free access to laboratory rodent chow (Toros Feed, Turkey) and tap water up to the time of sacrifice.

The mice were randomly divided into six groups (n = 15). Each group was further divided into three sub-groups (n = 5; A, B, and C). Only physiological saline (0.1 mL) was administered intraperitoneally (IP) every day to the control group (Group 1) for 14 days. The silymarin control group (Group 2) received only silymarin (CAS: 65666-01-1, S0292; Sigma-Aldrich, USA) by gavage (100 mg/kg) every day for 14 days. FB₁ (CAS: 116355-83-0; kindly provided by Dr. Ronald Riley; Toxicology and Mycotoxin Research Unit, United States Department of Agriculture-Agricultural Research Service, USA) was administered to Groups 3 (1.5 mg/kg FB₁, IP) and 4 (4.5 mg/kg FB₁, IP) every other day starting from Day 1 (for the details please see Table 1). For Groups 5 (1.5 mg/kg FB₁) and 6 (4.5 mg/kg FB₁), FB₁ was administered from day 1 for 14 days. Groups 5 and 6 also received silymarin (100 mg/kg, gavage) daily for 14 days starting from Day 0. Five mice from each group were sacrificed by cervical dislocation on Days 14, 17, and 21 (sub-groups A, B, and C, respectively; Table 1). The treatment protocol was based on previous studies showing that FB₁ induces liver injury in a dose-dependent manner [3,9].

Animals were euthanized by cervical dislocation. The liver was removed, weighed, cut into sections in 5 mm thickness with a surgical blade, and immediately fixed in 10% neutral buffered formalin. After fixation, the tissues were embedded in paraffin, cut into sections with a microtome (Leica RM2125RT; Leica, Germany) in 5 µm, and stained with hematoxylin and eosin (H&E; Hematoxylin crystals purchased from Merck, Germany; Eosin Y was purchased from Carlo Erba, Italy) for histological examination.

Sections from all tissue samples were cut into sections 5-µm thick and processed for immunohistochemical examination with a streptavidin-biotin-peroxidase method. The tissue sections were placed on slides (Isolab Laborgeräte, Germany) coated with 3-amino-propyltrieyhoxysilane (Sigma-Aldrich, USA), dewaxed, and hydrated by using decreasing dilutions of ethanol from 100% to 50%. Antigen retrieval was performed by heating the sections in citrate buffer (pH 6.0) for 10 min in a microwave oven (M1733N; Samsung, Malaysia) at 800 watts. The slides were then dipped in freshly prepared absolute methanol containing 3% (v/v) hydrogen peroxide for 15 min at room temperature to block endogenous peroxidase activity. The slides were then incubated with polyclonal or monoclonal antibodies (Table 2) for 1 h at room temperature. Next, the sections were exposed to biotinylated goat anti-rabbit IgG (Dako, USA) or rabbit anti-mouse IgG (Dako) diluted 1 : 300 in phosphate-buffered saline (PBS) for 30 min at room temperature. The slides were subsequently incubated with streptavidin peroxidase complex (ABC; Dako) diluted 1 : 300 in Tris-buffered solution for 30 min at room temperature. After repeated washing with PBS, the sections were incubated with 3,3'-diaminobenzidine tetrahydrocholoride (DAB; Dako) in distilled water (0.5 mg DAB/mL) containing 30% (v/v) hydrogen peroxide for 2 min. and the reaction was halted by immersing sections in tap water at room temperature. Finally, the sections were counterstained with Mayer's hematoxylin (Merck), dehydrated, and mounted (Entellan; Merck). Negative control tissue sections were incubated with undiluted normal goat or rabbit serum (Dako, Denmark).

Apoptotic cells were detected using a DeadEnd TUNEL system (Promega, USA) according to the manufacturer's protocol with some modifications. In brief, deparaffinized and rehydrated tissue sections were permeabilized with proteinase K (20 µg/mL; Promega) for 30 min at room temperature. After thoroughly washing with 3× PBS, the sections were incubated with equilibration buffer for 10 min. Next, 100 µL of terminal deoxynucleotidyl transferase reaction mixture was added to the sections except for the negative control and incubated at 37℃ for 1 h. The reaction was stopped by immersing the sections in 2× saline-sodium citrate buffer at room temperature for 15 min. Endogenous peroxidase activity was then quenched by incubation in 0.3% hydrogen peroxide at room temperature for 10 min. The sections were subsequently treated with horseradish peroxidase-conjugated streptavidin (1 : 500) for 30 min at room temperature. After repeated washing with PBS, the tissue samples were incubated with 3,3'-diaminobenzidine (Promega) in distilled water (0.5 mg DAB/mL) containing 30% (v/v) hydrogen peroxide and the reaction was halted by immersing sections in tap water at room temperature. Finally, the sections were counterstained with Harris hematoxylin (Merck), or methylene blue (Merck). The sections, then, were dehydrated by using increasing percentages of alcohol from 50% to 100% and cleared with xylene (Merck) and mounted (Entellan; Merck) before being examined at 400× magnification (Nikon Eclipse E600; Nikon, Japan) for TUNEL-positive cells.

Cells positive for Gal-3 and caspase-8 were semi-quantitatively assessed with a light microscope (Nikon Eclipse E600; Nikon). For immunostaining quantification, an area with the highest density of positive cells was chosen. To calculate the total number of positively stained cells, an ocular grid of 100 (10 × 10) squares was used at 400× magnification. Cells in a total of ten high-power fields (0.025 mm²) were counted. The percentages of positive versus negative cells were recorded in the ten fields and the mean value was the final score. The analyses of TNFα, FGF-2, and VEGF immunostaining were slightly different because staining intensities significantly differed between the sub-groups of every group. For this reason, percentages of positively stained cells were multiplied by cytoplasmic staining intensity scores: (1) weak, (2) moderate, and (3) intense.

Results of the Ki-67, PCNA, mitosis, and TUNEL assays were assessed using a grading system based on the total number of positive cells. The mitotic rate was calculated by counting the number of mitotic figures. All positive cells for the TUNEL assay as well as Ki-67 and PCNA staining or mitotic figures were counted among approximately 1,400 cells per sample from a total of 10 high-power fields (0.025 mm²). To calculate the total number of positively stained cells, an ocular grid of 100 (10 × 10) squares was used at 400× magnification. The intensity of Ki-67 and PCNA staining in the cytoplasm and nucleus was assessed semi-quantitatively based on the following scale: (0) none, (1) weak, (2) moderate, or (3) intense.

Differences between the groups were tested with a one-way analysis of variance (ANOVA) and Tukey's test using statistical package for the social sciences for Windows version 10.0 (SPSS, USA). Data are presented as the mean ± standard error (SE). A p values less than 0.05 were considered significant.

Histopathological examination revealed that livers from the control and silymarin groups were normal in appearance with typical histological structures. FB₁ treatment (Group 3; 1.5 mg/kg FB₁) caused degenerative changes in the hepatocytes including mild vacuolar degeneration, bile duct hyperplasia, megalocytosis (Fig. 1A), and mitosis (Fig. 1B). Severity of the lesions was similar among sub-groups 3A, 3B, and 3C. The degenerative changes were more severe in mice treated with 4.5 mg/kg FB₁. Silymarin administration alleviated lesion severity in both Groups 5 and 6. However, the occasional presence of apoptotic figures (Fig. 1C) and spherical, eosinophilic intranuclear inclusion bodies formed by invagination of the cytoplasm into nucleus persisted in all groups (Fig. 1D).

Apoptotic cells were identified in all the cases by the TUNEL assay. Three different patterns of TUNEL-positive reactions were observed either alone or in combination: brown products in both non-pyknotic nuclei and cytoplasm (Fig. 1E), cytoplasmic staining alone (Fig. 1F), and appearance of non-identifiable rounded or irregular structures (so-called "apoptotic bodies") either free among the hepatocytes or within the cytoplasm (Fig. 1G). TUNEL-positive signals were also observed among necrotic debris along with some Kupffer and other inflammatory cells. In the negative tissue section control, substitution of the TUNEL reaction mixture with labeling solution lacking terminal transferase produced negative results.

TUNEL-positive signals (Table 3) were occasionally observed in the control (Group 1) and mice treated with silymarin alone (Group 2). FB₁ administration caused a statistically significant increase in apoptotic bodies in Groups 3 and 4 (p < 0.0001). Silymarin treatment significantly ameliorated the deleterious effects of FB₁, particularly in Group 5 animals (p < 0.0001).

Mitotic figures (Table 3) were occasionally observed in the control (Group 1) and animals treated with silymarin alone (Group 2). FB₁ administration caused a significant increase of mitotic figures in Group 3 animals. Silymarin treatment significantly decreased mitotic figure formation in Group 5 (p < 0.05).

Immunohistochemical examinations revealed that FB₁ administration caused a statistically significant increase of caspase-8, Ki-67, PCNA, and TNFα levels in the liver compared to the control and mice treated with silymarin alone. Semi-quantitative assessment of the immunostaning results and statistical differences among all groups and sub-groups are summarized in Tables 3 and 4.

Immunohistochemistry revealed marked cytoplasmic immunoreactivity of the hepatocytes (Fig. 1H) from mice that received FB₁ alone (Groups 3 and 4). However, the staining intensity was significantly decreased in silymarin-treated Groups 5 and 6 (Fig. 2A) compared to that observed mice receiving FB₁ alone (p < 0.0001).

Three different patterns of Ki-67-specific staining were observed in the liver tissue either alone or in combination: (1) brown products in both the nucleus and cytoplasm (Fig. 2B), (2) brown product only in the cytoplasm with a negative nucleus (Fig. 2B), and (3) nuclear staining alone (Fig. 2C). The reactions were diffuse with periacinar preponderance (Fig. 2C). The groups treated with silymarin alone showed very limited reaction while staining specific for Ki-67 was not observed for the control group. In contrast, FB₁ administration caused a significant increase in Ki-67-positive staining, particularly in Group 4 (Fig. 2D). The differences between groups treated with FB₁ and silymarin were not statistically significant although a significant decrease was observed in Group 6 on day 21 (Group 6C, p < 0.005; Fig. 2E).

With the exception of some Kupffer cells, staining for the Gal-3 antibody was not observed in the control or silymarin-treated groups. FB₁ administration caused a mild elevation of Gal-3 staining in both Groups 3 and 4 (Fig. 2F). Prominent Gal-3-specific reactions were also detected in Kupffer cells, neutrophils, and macrophages (Fig. 2G). Silymarin treatment alleviated the effects of FB₁ in Groups 5 (Fig. 2H) and 6. The decrease was more pronounced on day 21 (Groups 5C and 6C, Table 4).

Three different patterns of PCNA-specific staining were observed in the liver tissue either alone or in combination: (1) both cytoplasmic and nuclear staining (Fig. 3A), (2) cytoplasmic staining alone with a negative nucleus (Fig. 3A), and (3) nuclear staining alone (Fig. 3A). Diffuse cytoplasmic or nuclear staining in the hepatocytes was mainly seen in the periacinar region (Fig. 3B). Expression levels for PCNA in the control and silymarin-treated groups were insignificant. On the other hand, FB₁ administration caused significant elevation of PCNA expression in Groups 3 and 4. The staining intensity was quite high on the 14th day in these two groups (Fig. 3C). PCNA levels decreased continuously and were at a minimal level on Day 21 in Groups 3 and 4. A combination of FB₁ and silymarin administration (Group 5) caused a significant decrease of PCNA expression on the 14th day (Group 5A; Fig. 3D). However, this decrease was not constant and PCNA levels started to increase on Day 17 (Group 5B) and returned to the base level on Day 21 (Group 5C).

TNFα-positive cytoplasmic staining was found in hepatocytes from all groups (Fig. 3E). Weak TNFα staining was observed in the control and silymarin-treated control groups. Liver sections from FB₁-exposed mice in Groups 3 and 4 exhibited a significant increase of TNFα expression (p < 0.0001). Silymarin application significantly alleviated the effects of FB₁ in only Group 6 (p < 0.0001). High TNFα levels were significantly decreased on Day 21 in both Groups 5C (Fig. 3F) and 6C (p < 0.0001).

Cytoplasmic FGF-2 reactivity was mainly seen in hepatocytes surrounding the central vein. Other hepatocytes were negative in the control (Fig. 3G) and silymarin-treated control groups. The expression levels did not change in the group treated with 1.5 mg/kg FB₁ (Group 3). Increased dosage of FB₁ (4.5 mg/kg) significantly elevated the FGF-2 expression level in Group 4 (Fig. 3H). Silymarin treatment resulted in a significant decrease (p < 0.0001) of FGF-positive hepatocytes in Group 6. FGF reaction was only seen in the intralobular cells in Group 6.

Hepatocytes from the control group and mice treated with silymarin alone showed weak, diffuse cytoplasmic staining limited to the periacinar and periportal areas. Immunostaining of the hepatocytes were elevated significantly upon FB₁ administration in Groups 3 and 4 (Fig. 4A). Silymarin treatment caused a significant decrease in VEGF expression, particularly in Group 6 (Fig. 4B; p < 0.0001).