Flutamide inhibits body weight growth, increases HFD-induced NAFLD score, and enhances preneoplastic hepatic lesions
During the study period, FL suppressed body weight gain after 6 weeks, regardless of the diet type (Fig. 1a). No discernible difference was observed in terms of food and water intake (Table 1, Supplemental Tables 3 and 4). In the final autopsy, FL significantly reduced the final body weight and intra-abdominal fat weight, independent of the HFD (Table 1). Importantly, HFD-induced obesity was not observed under these experimental conditions (Fig. 1a, Table 1).
Table 1
Final body weight, organ weight, food intake, and water intake in rats§
Group | BD | BD + FL | HFD | HFD + FL |
No. of animals | 6 | 6 | 6 | 6 |
Final body weight (g) | 301.5 ± 15.15a | 273.7 ± 6.96bc | 293.4 ± 15.13ab | 271.8 ± 12.05c |
Food intake (g/kg) | 47.1 ± 16.8 | 46.0 ± 15.5 | 34.1 ± 15.0 | 33.3 ± 13.9 |
Water intake (g/kg) | 66.6 ± 19.3 | 64.4 ± 17.3 | 53.8 ± 17.6 | 51.3 ± 16.5 |
Absolute liver weight (g) | 7.68 ± 0.72ab | 7.91 ± 0.32a | 6.89 ± 0.86b | 6.90 ± 0.38b |
Relative liver weight (%BW) | 2.54 ± 0.13a | 2.89 ± 0.07b | 2.34 ± 0.19a | 2.54 ± 0.16a |
Absolute intraperitoneal fat weight (g) | 6.27 ± 1.47a | 4.40 ± 0.40b | 6.78 ± 0.86a | 5.04 ± 1.33ab |
Relative intraperitoneal fat weight (%BW) | 2.07 ± 0.46ab | 1.61 ± 0.13b | 2.30 ± 0.18a | 1.85 ± 0.46ab |
Abbreviations: BD, basal diet; HFD, high fat diet; FL, flutamide, BW; body weight. | |
§: All animals were subjected to two-thirds partial hepatectomy at week 3. | | |
Data are shown as the mean ± standard deviation. | | | |
Different letters indicate significant differences between groups (p < 0.05, significantly different by Tukey's or Steel-Dwass test). |
Previous studies have shown that an HFD increases NAS and steatosis scores in the liver28–31. Similar results were obtained in the present study, where FL further increased these scores (Fig. 1b-1d). Notably, the scores associated with ballooning degeneration and inflammatory foci increased in response to an HFD but remained unaffected by FL (Fig. 1e, 1f). FL caused a significant increase in relative liver weight in the BD + FL group, which might have been caused by CYP induction27, as is evident from the Cyp1a gene expression data presented below (Table 2). However, no statistically significant change in relative liver weight was observed in the HFD + FL group (Table 1).
Table 2
Gene expression analysis of autophagy, mitochondria/mitophagy, lipid metabolism, drug metabolism, inflammation, and oxidative stress in liver samples.
Group | | BD | BD + FL | HFD | HFD + FL |
No. of animals | | 6 | 6 | 6 | 6 |
Autophagy-related genes | | | | | |
| Atg3 | 1.05 ± 0.34 | 1.69 ± 0.40 | 2.58 ± 1.23 | 1.60 ± 0.32 |
| Atg5 | 1.10 ± 0.51a | 1.77 ± 0.72ab | 2.61 ± 0.91b | 2.27 ± 0.83ab |
| Atg7 | 1.03 ± 0.26a | 1.10 ± 0.23a | 2.08 ± 0.64b | 1.98 ± 1.02ab |
| Lamp1 | 1.06 ± 0.40a | 1.50 ± 0.41ab | 2.08 ± 0.72b | 1.54 ± 0.59ab |
| Lamp2 | 1.05 ± 0.35a | 1.57 ± 0.36ab | 2.13 ± 0.93b | 1.19 ± 0.39a |
| Lc3 | 1.00 ± 0.08a | 1.24 ± 0.38ab | 2.06 ± 0.83b | 1.58 ± 1.04ab |
| p62 | 1.03 ± 0.26 | 1.30 ± 0.57 | 1.82 ± 0.72 | 1.67 ± 1.02 |
Mitochondria/Mitophagy-related genes | | | | | |
| Ambta1 | 1.14 ± 0.66 | 1.98 ± 0.75 | 3.95 ± 3.18 | 4.64 ± 5.32 |
| Parkin | 1.17 ± 0.64a | 2.54 ± 0.99ab | 3.27 ± 2.12ab | 3.57 ± 1.59b |
| Pink1 | 1.03 ± 0.24 | 1.09 ± 0.22 | 1.81 ± 0.65 | 1.12 ± 0.29 |
| Bnip3 | 1.01 ± 0.18 | 1.48 ± 0.41 | 1.43 ± 0.47 | 1.05 ± 0.34 |
| Nadh | 1.03 ± 0.26 | 1.32 ± 0.26 | 2.21 ± 0.97 | 1.17 ± 0.32 |
| Sdhd | 1.04 ± 0.30a | 1.67 ± 0.40ab | 2.17 ± 0.86b | 1.51 ± 0.51ab |
| ATP synthase | 1.03 ± 0.29 | 1.47 ± 0.45 | 1.79 ± 0.75 | 1.24 ± 0.30 |
Lipid metabolism-related genes | | | | | |
| Abca1 | 1.09 ± 0.40a | 1.89 ± 0.79ab | 3.26 ± 1.07c | 2.74 ± 0.89bc |
| Acox1 | 1.14 ± 0.59 | 1.80 ± 0.54 | 2.74 ± 1.36 | 1.89 ± 0.34 |
| Apob | 1.05 ± 0.30 | 0.82 ± 0.21 | 1.49 ± 0.76 | 1.31 ± 0.32 |
| Dgat2 | 1.06 ± 0.39a | 1.36 ± 0.41ab | 1.77 ± 0.51b | 1.70 ± 0.35ab |
| Fasn | 1.27 ± 0.87 | 3.04 ± 1.50 | 1.72 ± 0.96 | 2.43 ± 2.25 |
| Hmgcs1 | 1.09 ± 0.47 | 0.72 ± 0.36 | 1.51 ± 0.77 | 1.18 ± 0.36 |
| Hsd3b1 | 1.10 ± 0.48a | 0.38 ± 0.40a | 2.17 ± 0.90b | 0.83 ± 0.60a |
| Lpl | 1.06 ± 0.37 | 0.97 ± 0.27 | 1.50 ± 0.80 | 1.17 ± 0.46 |
| Lss | 1.11 ± 0.53 | 1.46 ± 0.72 | 1.69 ± 0.66 | 2.30 ± 1.19 |
| Ppara | 1.19 ± 0.69a | 1.81 ± 0.97a | 4.55 ± 1.44b | 2.08 ± 0.71a |
| Pparg | 1.05 ± 0.36 | 1.16 ± 0.98 | 0.61 ± 0.26 | 1.91 ± 1.47 |
| Scd1 | 1.03 ± 0.26 | 1.27 ± 0.38 | 1.69 ± 0.46 | 1.25 ± 0.62 |
| Srebf1 | 1.17 ± 0.73 | 0.84 ± 0.43 | 1.59 ± 1.26 | 1.58 ± 1.02 |
| Srebf2 | 1.08 ± 0.41a | 1.53 ± 0.44ab | 2.49 ± 0.95b | 2.33 ± 0.88b |
Drug metabolism-related genes | | | | | |
| Cyp1a1 | 1.27 ± 1.09a | 73.20 ± 45.16b | 1.90 ± 0.94a | 96.14 ± 93.51b |
| Cyp2b1 | 1.64 ± 1.91 | 1.97 ± 0.95 | 1.84 ± 0.91 | 2.51 ± 1.10 |
| Cyp3a1 | 1.05 ± 0.34 | 0.65 ± 0.25 | 1.40 ± 0.54 | 0.89 ± 0.79 |
Inflammation related genes | | | | | |
| Tnf-a | 1.11 ± 0.51 | 1.29 ± 0.64 | 1.81 ± 1.11 | 3.33 ± 3.87 |
Oxidative stress-related genes | | | | | |
| Catalase | 1.04 ± 0.33a | 1.75 ± 0.58ab | 2.70 ± 1.22b | 2.18 ± 0.98ab |
| Gpx1 | 1.16 ± 0.70 | 1.68 ± 0.67 | 1.92 ± 0.75 | 2.05 ± 0.95 |
| Gpx2 | 1.08 ± 0.47a | 5.36 ± 1.91b | 3.44 ± 2.86ab | 4.98 ± 2.16b |
| Mn-SOD | 1.04 ± 0.32 | 1.23 ± 0.38 | 1.65 ± 0.73 | 1.54 ± 0.56 |
| Sod1 | 1.11 ± 0.53 | 1.32 ± 0.54 | 1.18 ± 0.55 | 1.00 ± 0.40 |
| Sod2 | 1.13 ± 0.63 | 1.41 ± 0.23 | 2.07 ± 0.87 | 1.33 ± 0.35 |
Abbreviations: BD, basal diet; HFD, high-fat diet; FL, flutamide. | | |
§: All animals were subjected to two-thirds partial hepatectomy at week 3. | | |
Data are shown as the mean ± standard deviation. | | |
Different letters indicate significant differences between groups (p < 0.05, significantly different by Tukey's or Steel-Dwass test). |
The consumption of HFD induced a significant increase in the number and area of GST-P-positive foci (Fig. 1g, 1h, 1k), as previously reported28–31. A previous study showed that that higher FL doses (1000 and 2000 ppm) enhanced hepatocarcinogenesis in rats27; however, in the present study, a lower FL dose (500 ppm) caused a marginal increase in both the number and area of GST-P-positive foci (Fig. 1g-1j). The area of GST-P-positive foci was significantly increased in the HFD + FL group compared to the BD group (Fig. 1h-1j); however, the cumulative effects of HFD and FL on the area of GST-P-positive foci were marginal. The number of Ki-67-positive cells within hepatic precancerous lesions showed an increasing trend in the FL-treated groups (BD + FL and HFD + FL); however, no statistically significant change was detected (Supplemental Table 4). These findings suggest that FL exacerbated steatosis and marginally increased preneoplastic lesions in a rat model of non-obesity-related steatosis.
Expression analysis of mitophagy markers in background hepatocytes and precancerous lesions
FL exhibits mitochondrial toxicity35, while HFD inhibits mitophagy21, 36. In light of these insights, we assessed the expression of the autophagy marker LC3 and the mitophagy indicator AMBRA1 in both background hepatocytes and precancerous lesions of rats treated with HFD and/or FL. LC3 and AMBRA1 showed positive granular signals in hepatocytes (Supplemental Figs. 2 and 3). Quantitative analysis showed that LC3-positive granules exhibited an increasing trend in the FL-treated groups (BD + FL and HFD + FL groups) compared to the non-FL-treated groups (BD and HFD groups) (Supplemental Fig. 2b, d), both in background hepatocytes (non-foci) and precancerous lesions (foci). Similarly, AMBRA1-positive granules tended to increase in number in the HFD group compared to the BD group. However, when combined with FL, a decreasing trend was observed, particularly in the BD + FL group, in both background hepatocytes (non-foci) and precancerous lesions (foci) (Supplemental Fig. 3b, 3d). These results indicate a potential suppression of mitophagy in the HFD group, as indicated by the increased expression of AMBRA1. Conversely, in combination with the FL treatment, mitophagy might be induced by the suppression of increased AMBRA1 expression; however, these effects on mitophagy were found to be marginal in this standard analysis.
No statistically significant differences in Parkin expression were observed in either background hepatocytes or preneoplastic lesions in each group (Supplemental Fig. 4a-d).
Cluster analysis of LC3, AMBRA1, and Parkin expression in background hepatocytes
To further investigate the relationship between LC3 and AMBRA1 expression, we compared the expression ratios of both indicators in background hepatocytes using cluster analysis, as previously reported34, aiming to elucidate the dynamics of autophagy flux (Supplemental Fig. 1). Using the k-means, the clusters were divided into three groups (Fig. 2a-c) and examined the cluster composition of each group (Fig. 2d). These clusters were classified into three levels: C1, characterized by low expression of both LC3 and AMBRA1, indicative of a basal level; C2, characterized by low AMBRA1 expression and high LC3 expression, suggestive of mitophagy induction; and C3, characterized by low LC3 expression and high AMBRA1 expression, implying mitophagy inhibition (Fig. 2b). In the BD group, the major constitutive cluster was C1, followed by C3, and lastly C2 in a descending order (Fig. 2d). This pattern closely resembled that of the BD + FL group, where the primary constitutive cluster remained C1, similar to the BD group. However, the percentage of C2 clusters in the BD + FL group increased compared to the BD group, although this difference was not statistically significant. In contrast, in the HFD group, the prevalence of C3 composition significantly increased. In the HFD + FL group, the cluster composition was similar to that observed in the BD + FL group, with a significant increase in the rate of C2 composition.
We analyzed LC3 and Parkin expression to evaluate whether mitophagy induction was Parkin-dependent in the FL-treated groups (Fig. 2e-2h). The clusters were classified into four levels: C1, characterized by low LC3 expression and high Parkin expression, implying mitophagy inhibition; C2, with low Parkin expression and high LC3 expression, suggestive of Parkin-independent mitophagy induction; C3, a basal cluster with low expression of both LC3 and Parkin; and C4, featuring high expression of both LC3 and Parkin, indicative of Parkin-dependent mitophagy induction (Fig. 2f). In the BD + FL group, the primary cluster was C3, while in the HFD + FL group, the C4 composition was significantly augmented (Fig. 2h), suggesting that FL induced mitophagy in a Parkin-dependent manner in a population of preneoplastic lesions in HFD-fed rats.
Cluster analysis of the GST-P-positive foci area and LC3, AMBRA1, and Parkin expression in hepatic precancerous lesions
To determine whether AMBRA1 expression played a role in the development of hepatic preneoplastic lesions, we examined the relationship between the area of individual GST-P-positive foci and the expression rate of AMBRA1 in each lesion (Fig. 3a-d). Subsequently, we classified cluster into four levels, namely C1 to C4, and observed that C4 exhibited the highest AMBRA1 expression (mitophagy inhibition cluster) among all four levels (Fig. 3b). The percentage of C4 composition significantly increased in the HFD group, and this effect was attenuated when combined with FL (Fig. 3d), suggesting that FL might cancel mitophagy inhibition in HFD-fed rats.
We investigated the relationship between LC3 and AMBRA1 expression in hepatic precancerous lesions, which were divided into three clusters (Fig. 3e-3h), and examined the cluster composition of each group (Fig. 3h). These clusters were classified into three levels, namely C1–C3: C1 represented the basal level with low expression of both LC3 and AMBRA1; C2 indicated mitophagy induction with low expression of AMBRA1 and high expression of LC3; and C3 denoted mitophagy inhibition with high expression of AMBRA1 and low expression of LC3 (Fig. 3f), similar to the categorization in background hepatocytes (Fig. 2b). The percentage of C3 cells was significantly higher in the HFD group than in the BD group (Fig. 3h). The composition of clusters in the HFD + FL group was similar to that in the BD + FL group. The results suggested that FL might cancel mitophagy inhibition in HFD-fed rats as well.
Finally, we analyzed LC3 and Parkin expression in hepatic precancerous lesions to evaluate whether mitophagy induction was Parkin-dependent in the FL-treated groups (Fig. 3i-3l). The clusters were classified into three levels: C1 represented the basal level with low expression of both LC3 and Parkin; C2 indicated a cluster with low LC3 expression and high Parkin expression, indicative of mitophagy inhibition; and C3 represented a level with high LC3 expression and low Parkin expression, suggestive of Parkin-independent mitophagy induction (Fig. 3j). Unlike in background hepatocytes, we did not detect C4, a Parkin-dependent mitophagy inhibition cluster (Fig. 2f). C3 was minimally detected in the FL-treated groups BD + FL and HFD + FL (Fig. 3l).
HFD and HFD combined with FL alters hepatic gene expression
A comprehensive gene expression analysis was conducted on the liver samples to determine the contributions of autophagy, mitophagy, injured mitochondria, lipid metabolism, inflammation, drug-metabolizing enzymes, and oxidative stress-related genes, which might be related to NAFLD-related liver carcinogenesis.
We analyzed autophagy-related genes and showed that the expression of Atg5, Atg7, Lamp1, Lamp2, and Lc3 was significantly increased in the HFD group compared to the BD group. However, when HFD feeding was combined with FL administration, a significant decrease was observed in the expression of these genes (Table 2). These gene expression changes in the HFD group were considered a compensatory response to the suppression of mitophagy, as discussed earlier. Regarding mitophagy-related genes, Parkin expression was significantly elevated in the HFD + FL group compared to the BD group, along with higher levels of Ambra1 expression. Furthermore, a relatively higher increase in the expression of rkin and Ambra1 was observed in the BD + FL and/or HFD groups. This could potentially indicate the presence of Parkin-dependent mitophagy in background hepatocytes, aligning with the observations discussed earlier.
FL, a known inhibitor of the mitochondrial respiratory chain complex I, has been reported to impair mitochondrial functions 25. Moreover, 2-hydroxyflutamide, a major FL metabolite, impairs the functions of mitochondrial respiratory chain complexes II and V26. In this study, we analyzed the gene expression of complex I (NADH: quinone oxidoreductase), complex II (SDHD: succinate dehydrogenase), and complex V (ATP synthase). We found that the expression level of Sdhd tended to increase in the BD + FL group and was significantly elevated in the HFD group compared to the BD group. However, this expression trend reversed with the context of combined HFD feeding and FL administration, indicating a decreasing pattern (Table 2). Although no evident differences in the expression of Nadh and ATP synthase were observed between the groups, the expression of Nadh tended to increase in the HFD group; however, it remained comparable to that in the BD group. It is important to note that FL induces significant upregulation of Cyp1a1 expression in rats37. In our study, the expression of Cyp1a1 was significantly elevated upon FL administration, regardless of the presence of HFD.
Regarding the expression levels of genes related to lipid metabolism, Abca1, which is responsible for the production of high-density lipoproteins, and Srebf2, which is involved in cholesterol metabolism, were significantly increased in the HFD groups (HFD and HFD + FL) compared to the BD group (Table 2). In addition, the TG synthase gene Dgat2, steroid hormone synthase Hsd3b1, and transcription factor Ppara, which promotes β-oxidation, were significantly increased in the HFD group compared to the BD group, while they either showed a declining trend or a significant decrease in the HFD + FL group.
No obvious differences in the expression of Tnf-α, one of the most effective cytokines, between the groups (Table 2).
Regarding oxidative stress gene expression, the antioxidant enzyme Catalase exhibited a significant increase in the HFD group compared to the BD group. Furthermore, the expression level of Gpx2 significantly increased in the FL alone group and in the combination of HFD feeding and FL administration (Table 2). Thus, in the HFD group, antioxidant enzyme-related genes exhibited increased expression parallel to steatosis. However, the combined effect of HFD feeding and FL administration on gene expression was not detected.