Cholesterol (92.5%) and 2′,7′-dichlorofluorescein diacetate (DCFH-DA) were purchased from Sigma-Aldrich (St. Louis, United States). Pioglitazone (98%), Nile Red (95%), Oleic acid, and Palmitic acid were purchased from Aladdin (Shanghai, China). Dulbecco’s Modified Eagle Medium (DMEM) and Fetal Bovine Serum (FBS) were purchased from Gibco (Grand Island, NY, United States). Trypsin was purchased from Solarbio (Beijing, China). BSA-Fatty Acid Free was purchased from Yeasen (Shanghai, China). Penicillin-Streptomycin was purchased from Beyotime (Shanghai, China). HBA (analytical standard, purity >99%) was obtained from National Institutes for Food and Drug Control of China (Beijing, China). Anti-Nrf2 antibody (PA5-27882) was purchased from Invitrogen (Carlsbad, CA, United States). Goat Anti-Rabbit IgG H&L (FITC) secondary antibody (ab6717) was purchased from Abcam (Cambridge, MA, United States).

The primary food for Larval zebrafish (AP100) was purchased from Zeigler (PA, United States). The high-cholesterol diet (HCD) was prepared by mixing cholesterol with basic food. The final concentration of cholesterol in HCD was 5% (w/w). A total of 20.296 mg oil acid and 8.547 mg PA were dissolved with 2 ml 0.1 M NaOH in a 75°C water bath as a 50 mM FFA stock solution. We added 2 ml of 50 mM FFA stock solution to 8 ml 10% Fatty Acid Free BSA dropwise at a 55°C water bath to dissolve into 10 mM mother liquid, and this was percolated with 0.22 μM filter membrane. For drug solutions, due to the low solubility of Pioglitazone (PIG) in water, DMSO was used to dissolve the drugs first, followed by dilution in DMSO solution with water to achieve a final drug concentration of 1 μM/L (DMSO 0.001% v/v) for the administration of zebrafish. HBA was directly dissolved in water to achieve a final drug concentration of 20 mg/L, and each group was administrated with DMSO to the same concentration of 0.001% (v/v) in the zebrafish experiment and solved in PBS; this was then diluted by DMEM to a concentration of 20 μm to treat the BRL-3A cell.

The zebrafish embryos were generated by natural spawning from parent wild-type AB-line adult zebrafish. After three days of adaption from 5-days post-fertilization (dpf), larval zebrafish were randomly divided into eight groups in two processes (n = 100 for each group). The groups of the first process were as follows: 1) Control group, fed with a Normal diet (ND) for 7 days; 2) HCD group, fed with HCD (20 mg/tank per day) for 7 days; 3) PIG groups, fed with HCD (20 mg/tank per day) and PIG (1 μM) for 7 days; 4) HBA (20 mg/L) group, fed with HCD (20 mg/tank per day) and HBA (20 mg/L) for 7 days. In the second process, the groups were as follows: 1) Control group, fed with a Normal diet (ND) for 14 days; 2) HCD group, fed with HCD (20 mg/tank per day) for 14 days; 3) PIG groups, fed with HCD (20 mg/tank per day) and PIG (1 μM) for 14 days; 4) HBA (20 mg/L) group, fed with HCD (20 mg/tank per day) and HBA (20 mg/L) for 14 days. The dose of HBA treatment was administered according to the result of the survival rate, as seen in Supplementary Material Figure S1A . All the groups were maintained following the schedule showed in Figure 2A.

All the animal experiments were approved by the Science and Technology Department of Jiangsu Province and followed the Jiangsu Provincial standard ethical guidelines for the use of experimental animals under the ethical committees mentioned above (SYXK SU 2018–0,019, May 14, 2018).

Rattus norvegicus hepatocellular cell line BRL-3A (American Type Culture Collection, Manassas, VA, United States) was routinely maintained at 37°C in a humidified atmosphere of 5% CO2 in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin. For Nile red staining, ROS level measured, and immunofluorescent staining, BRL-3A cells (2 × 10⁵) were seeded in 60 mm plates until 70% confluent. Meanwhile, BAL-3A cells were treated with 1 mM FFA for 24 h for the model group, treated with 1 Mm FFA combination with 20 μM HBA for 24 h for the HBA group, and control cells were treated with BSA and NaOH in final concentrations of 1% and 0.2 μM, respectively. For q-PCR, BAL-3A cells (1 × 10⁶) were seeded in 100 mm plates until 70% confluent, and they were then treated as other experiments. The dose of HBA treatment on BRL-3A cell was in accordance with the result of MTT, and the maximum non-toxic dose was applied in Supplementary Material Figure S1B.

A total of 30 larval zebrafish of each group were collected; to measure the level of TG, TC, MDA, SOD, ROS, and HO-1, the whole Larval zebrafish were used to prepare the tissue homogenate with PBS in an ultrasonic cell homogenizer (SCIENT2-ⅡD), and we then obtained supernatant by centrifuge. The protein concentration of the tissue homogenate was measured by a BCA protein Quantitative Kit (BeyoTime, China). Triglyceride (TG) levels, total cholesterol (TC) levels, malondialdehyde (MDA) concentration, and superoxide dismutase (SOD) activity were measured by commercial assay kits (Jiancheng, Nanjing, China) following the manufacturer’s instructions. Quantitation of reactive oxygen species (ROS) was detected by Reactive Oxygen Species Assay Kit (BeyoTime, China) following the manufacturer’s instructions, and the fluorescence value read by a microplate reader, which represented the level of ROS. HO-1 level was measured by a Fish Heme oxygenase 1 (HO-1) Elisa Kit (Jiancheng, Nanjing, China). All quantitations of the above kits were read by a multifunctional microplate reader (BioTek, United States).

Nile red is a lipophilic fluorescence material that can stain the TG and fatty acid. It can be detected at 543 nm (excitation wavelength) and 598 nm (scattering light). DCFH-DA is an indicator of ROS. DCFH-DA do not have any fluorescence property; only when ROS is oxidizing, it will perform fluorescence property by its oxidized product. A total of 15 Larval zebrafish were incubated with 0.5 μg/ml Nile red or 10 μM/L DCFH-DA (diluted with embryo water) for 30 min at 28.5°C separately. The larval zebrafish were washed with water three times and kept into CMC-Na (4%) on a glass slide for fixation. After treated with FFA with or without HBA 24 h, the medium was removed, and BRL-3A cells were washed with PBS twice then incubated with 0.5 μg/ml Nile red. The fluorescence stereoscope (Olympus SZX16) was immediately used to observe and capture the lipid. All captures were taken with the same parameters (exposure time, ISO, and aperture) between different groups for comparison, and all the above procedures were carried out in the dark. The fluorescence intensity is measured by ImageJ: in brief, the images were converted to an 8-bit grey-scale map, and we adjusted the threshold under the same limited field and selected “limit to threshold” under analyze menu where the mean value represents the fluorescence intensity.

BRL-3A cells were treated with FFA with or without HBA 24 h, medium was removed, and BRL-3A cells were washed with ice PBS twice. Then cells were fixed in 4% paraform in ice for 15 min and incubated with a blocking solution containing 5% BSA and 0.3% Triton X-100 for 1 h at room temperature. After blocking, cells were incubated with Nrf2 (1:200, Invitrogen, Thermofisher) overnight in 4°C. After washing, cells were incubated for 20 min with FITC secondary antibody at 37°C and stained with DAPI to label cell nuclei. The signals were visualized, and digital images were obtained by the fluorescence stereoscope (Olympus SZX16).

A total of 30 liver tissues were separated from larval zebrafish in stereoscope and about 3 × 10⁶ cells of each group were sacrificed for the extraction of total RNA using Trizol reagent (Invitrogen, United States). Reverse transcription was performed by HiScript II qRT SuperMix (Vazyme, China) for the synthesis of cDNA. The qPCR was performed on the StepOnePlus Real-Time PCR System (Applied Biosystems, United States) by adding the ChamQTM universal SYBR qPCR Master Mix (Vazyme, China) and following the manufacturer’s protocol. The specific sequences of primers used in this study were synthesized by General Biotech Co., Ltd. (Shang Hai, China) and are shown in (Table 1). The 2^-∆∆Ct method was used to calculate the expression levels of each targeted mRNA normalized to GAPDH.

All the data are expressed as mean ± SD. Graph Pad PRISM (Graph Pad Software, United States) was used for comparing the treatment group and corresponding control by One-way ANOVA followed by Tukey’s test for the significant difference. The differences between groups were considered statistically significant at p-value <0.05.

To explore the role of HBA in isolated steatosis and mild nonalcoholic steatohepatitis, larval zebrafish models were studied at 7 and 14 days, respectively, by the inducement of HCD, as in a previous study (Ji et al., 2019) (Figure 2A). The effect of HAB in regulating lipid metabolism was measured by Nile Red staining, which is a red phenoxazine lipid fluorescent dye that can label the neutral lipid properties (Greenspan and Fowler, 1985). As the Nile Red results have shown, the intensity of Nile Red fluorescence within the model group increased gradually in the whole trunk and liver over time by the inducement of HCD compared with the Control. Whole-body fluorescent values counted by FIJI also saw extremely significant increases in model groups. However, the increase in Red fluorescence intensity and area were reversed by Pioglitazone and HBA, respectively, in two processes. Notably, the effect is more prominent in the second process (Figure 2B). Furthermore, the level of triglyceride (TG) and total cholesterol (TC) of larval zebrafish were also measured. In the first process, HCD increased the level of TG (0.033 ± 0.004 mMol/gprot, 2.358X to Control) and TC (0.150 ± 0.007mMol/gprot, 3.846X to Control) in the model group. However, both the level of TG and TC in the PIG group (TG: 0.017 ± 0.004 mMol/gprot, 0.515X to Model; TC: 0.054 ± 0.012 mMol/gprot, 0.360X to Model) and HBA group (TG: 0.026 ± 0.003 mMol/gprot, 0.788X to Model; TC: 0.081 ± 0.019 mMol/gprot, 0.540X to Model) were decreased. In the second processing, HCD increased the level of TG (0.042 ± 0.006 mMol/gprot, 2.100X to Control) and TC (0.350 ± 0.087 mMol/gprot, 4.160X to Control), and the level of TG and TC in both PIG group (TG: 0.020 ± 0.005 mMol/gprot, 0.476X to Model; TC: 0.130 ± 0.028 mMol/gprot, 0.377X to Model) and HBA group (TG: 0.030 ± 0.005 mMol/gprot, 0.714X to Model; TC: 0.177 ± 0.055 mMol/gprot, 0.506X to Model) were decreased. The results indicate that HBA has a lipid-regulating effect on HCD-induced larval zebrafish in a manner similar to PIG. To test the role of HBA in hepatic steatosis and steatohepatitis on HCD-induced NAFLD larval zebrafish, the HE staining was conducted in whole larval zebrafish (Figure 2D). In the first process, HCD induced 40–60% of the lipid area in the liver, and lipid accumulation in the liver was almost inhibited by HBA. In the second process, the macrovesicular steatosis almost occupied over 90% of the liver and presented rare inflammation. Notably, HBA also decreased the lipid area in the liver. In conclusion, our results suggested that HBA exhibits lipid regulation and improvement of NAFLD on HCD-induced isolated steatosis even though on a mild nonalcoholic steatohepatitis larval zebrafish model.

Lipid-modulating oxidative stress accelerates the development of NAFLD, and oxidative stress is also a prominent phenomenon in NAFLD. To reveal the anti-oxidative stress effect of HBA on an HCD-induced oxidant attack on larval zebrafish, 2′,7′-dichlorofluorescein diacetate (DCFH-DA), a fluorescent dye of ROS, was used to stain the oxidant stress product ROS on larval zebrafish. In the two processes, compared to the control group, HCD induced an increase in the fluorescence intensity, and it was distributed observably in the abdomen of larval zebrafish. PIG weakened the ROS-positive area and fluorescence intensity in both processes. However, the role in decreasing the ROS level of HBA is more remarkable in the second process compared with the first process (Figure 3A). ROS and MDA are all cytotoxicity products of lipid peroxidation. Meanwhile, HCD induced a higher level of ROS and MDA in all processes compared with control. Both PIG and HBA reversed the condition HCD induced (Figure 3B). In the first process, HCD increased the level of ROS (1706.0 ± 225.2, 1.7X to Control) and MDA (16.340 ± 1.932 mMol/gprot, 1.901X to Control). However, both the level of ROS and MDA in the PIG group (ROS: 1,157.0 ± 174.1 mMol/gprot, 0.7X to Model; MDA: 9.551 ± 1.080 mMol/gprot, 0.585X to Model) and HBA group (ROS: 1,410.0 ± 115.5, 0.8X to Model; MDA: 12.390 ± 1.445 mMol/gprot, 0.758X to Model) are decreased. Next, we measured the activity of antioxidase SOD and the level of antioxidase HO-1. The results (Figure 3C) showed that HCD did not disrupt the role of SOD (63.410 ± 2.963 U/mgprot, 1.268X to Control) and HO-1 (0.315 ± 0.057 ng/ml, 0.946X to Control) in the first processing. Intriguingly, the level of SOD in the PIG group (79.560 ± 8.41 0U/mgprot, 1.255X to Model) and HBA group (71.130 ± 8.434 ng/ml, 1.121X to Model) has increased, while the HO-1 level has not. In the second process, HCD also increased the activity of SOD (67.200 ± 5.000 U/mgprot, 1.347X to Control). Meanwhile, the level of SOD in the PIG group (88.46 0 ± 6.333 U/mgprot, 1.316X to Model) and HBA group (79.490 ± 7.269 U/mgprot, 1.183X to Model) has increased. Interesting, HCD decreased the level of HO-1 (0.250 ± 0.022 ng/ml, 0.749X to Control), while the level of HO-1 in PIG group (0.391 ± 0.024 ng/ml, 1.564X to Model) and HBA group (0.369 ± 0.021 ng/ml,1.476X to Model) has increased. These results indicate that HBA has the effect in regulating oxidative stress on HCD induced Larval Zebrafish by regulating multiple antioxidase.

HBA exhibits an effect on lipid regulating and anti-oxidative stress on HCD-induced NAFLD larval zebrafish. Thus, we further explored the underlying mechanism of HBA. The gene expression levels of srebf1, fasn, pparα, pparγ (related to lipid metabolism), tnfα, il-6, il-1β (related to inflammation), keap, nrf2, and ho-1 (related to oxidative stress) in the liver were measured. As the results suggest, ed lipogenesis-related genes srebf1 and fasn were significantly increased after HCD induction compared with control. Notably, the related gene expression noted above was reduced in the PIG and HBA groups in all processes (Figures 4A,B). However, the mRNA expression of the lipid-lowering related genes pparα and pparγ decreased in HCD-induced Larval Zebrafish at 14 days and was restored by PIG and HBA, while the change by HCD was induced within 7 days of processing has no significance (Figures 4A,B). Consistently, q-PCR results demonstrated that the inflammation-related gene expression was increased by HCD in after both 7 and 14 days of processing, while the inflammation-related gene expression decreased in both the PIG and HBA groups. Moreover, Nrf2 pathway gene expression did not differ between model and control in short-term HCD treatment but decreased following long-term HCD intervention. Notably, HBA restored the expression of Nrf2-pathway related gene expression as PIG did (Figures 4C,D).

Based on the results of the investigation into HCD-induced NAFLD Larval Zebrafish, we further explored the role of HBA in lipid regulation and oxidative stress by BRL-3A cells induced with FFA in vitro. FAs flow into the liver could induce lipid accumulation and lipotoxicity in hepatocytes. BRL-3A hepatocytes were induced by 1 mm FFA 24 h with or without HBA (20 μM), performed with Nile red staining to measure lipid level, and stained with DAPI to locate the cell. Fluorescence intensity in the cytoplasm increased by FFA, and HBA reduced the fluorescence intensity (Figure 5A). Treatment with 1 mM FFA (62.740 ± 2.986, 3.551X to Control) increased the fluorescence intensity, and HBA (45.370 ± 6.068, 0.723X to Model) decreased the fluorescence intensity (Supplementary Material Figure S2A). The ROS level in hepatocytes was measured, as rising lipid levels in hepatocytes could induce lipid peroxidation and produce ROS. As the results show in Figure 5B, FFA increased the ROS level, and HBA significantly reduced the ROS level. The results implied that HBA also plays a role in lipid-lowering and anti-oxidative stress in hepatocytes. Moreover, we investigate the intrinsic mechanism through immunofluorescence staining combined with q-PCR. The immunofluorescence staining result indicated that HBA facilitates Nrf2 nuclear translocation, unlike both the control and model groups (Figure 5C). Furthermore, the expression of the lipogenesis-related gene and Nrf2 pathway gene were estimated by way of q-PCR. The results showed that FFA could increase the expression of the lipogenesis-related gene and decrease the expression of the Nrf2 pathway gene. Notably, HBA decreased the expression of lipogenesis-related gene and increased the expression of Nrf2 pathway gene (Figure 5D). The results suggested Nrf2 may be the potential target of HBA in regulating oxidative stress on NAFLD.