Isoflurane (R510-22-10) was bought from RWD Life science Co., Ltd. (Shenzhen, China). Adezmapimod (SB203580), a potent p38-MAPK inhibitor, was purchased from Selleck chemicals, LLC, United States. An anti-PCSK9 monoclonal antibody, alirocumab, was obtained from Sanofi S.A. Penicillin-streptomycin liquid (P1400), 0.25% trypsin-EDTA solution (T1300), palmitic acid (N-16-A), and cis-9-octadecenoic acid (IC1350) were purchased from Solarbio Life Sciences (Beijing, China). Human hepatocellular carcinoma cell line HepG2 (SCSP-150) was bought from the Cell Bank of Chinese Academy of Sciences (Shanghai, China). Lipofectamine 3,000 reagent (L3000008) was purchased from Invitrogen Corporation, United States. Dulbecco′s modified Eagle′s medium (DMEM) (C11995500BT) and Opti-MEM I reduced serum medium (31,985,062) were the products of Gibco ThermoFisher Scientific Corporation (CA, United States). Fetal bovine serum (FBS) (BS1612-105) was obtained from Bioexplorer Life Sciences (C.A., China). High-fat diet (License No.: SCXK 2020–0004) was the product of Keao Xieli Feed Co., Ltd. (Tianjin, China). Heparinized capillary tubes (2501) were purchased from Kimble Chase Life Science and Research Products LLC. (NJ, United States). NEG-50™ frozen section medium (6502) was bought from Richard-Allan Scientific LLC. (Kalamazoo, MI, United States). The assay kits for total cholesterol (TC, 100,000,180) and TG (100,000,220) were provided by Biosino Bio-technology and Science Inc. (Beijing, China). The antibody information was provided in Supplementary Table S1. The rest reagents used in this study were of analytical grade.

Animal experiments were approved by the Laboratory Animal Ethical Committee of Shandong Second Medical University and carried out according to NIH guidelines for the Care and Use of Animals (2020SDL106). In this study, PCSK9 ^flox/flox mice and albumin (Alb)-Cre mice were purchased from GemPharmatech LLC. (Jiangsu, China). The mice were kept in a temperature-controlled room (23°C) with controlled light/dark cycles (12 h/12 h). PCSK9 ^flox/flox mice were bred with Alb-Cre mice to produce Alb-Cre^-/PCSK9 ^flox/flox (PCSK9^liver(+/+) or wild type) mice, Alb-Cre ⁺ /PCSK9 ^flox/flox (hepatic-specific PCSK9 knockout, PCSK9^{liver(−/−)} ) mice; and Alb-Cre ⁺ /PCSK9 ^flox/WT (PCSK9^liver(+/−) ) mice. The primers for genotyping are listed in Table 1.

Approximately 10-week-old male PCSK9^liver(+/+) (n = 10), PCSK9^{liver(−/−)} (n = 10), and PCSK9^liver(+/−) (n = 6) mice were fed an HFD containing 21% fat and 0.15% cholesterol for 9 weeks. On the last day of HFD feeding, mice were fasted for 6 h, weighted, and anaesthetized using isoflurane before sampling. Briefly, mouse blood was collected from the retro-orbital venous plexus by using heparinized capillary tubes, and plasma was obtained by centrifugation of the blood samples at 1,100 × g for 30 min. Next, mouse heart was exposed, and the auricle was cut using surgical scissors. The mouse was perfused with phosphate buffer saline (PBS, pH = 7.4) through the left ventricle to remove the residual blood before sampling. For comparative analyses, 6 age-matched PCSK9 ^flox/flox mice fed a chow diet were randomly selected for plasma levels of TC and TG determination and Western blotting experiments.

Plasma TC and TG levels were determined using appropriate assay kits in accordance with the manufacturers’ instructions. To determine TC and TG distributions in very low-density lipoprotein (VLDL), LDL, and high-density lipoprotein (HDL) particles, mixed plasma (100 μL) from each group was analyzed using a Superose™ 6 10/300 gel chromatography column linked to an ÄKTA Fast Protein Liquid Chromatography (FPLC) system. The column was eluted with 0.9% sodium chloride solution, and the eluate was collected using an automatic collector with 0.5 mL in each fraction. Hepatic lipids were extracted as reported previously with minor modifications (Tran et al., 2020; Li et al., 2022). Briefly, approximately 100 mg liver samples were homogenized in 1 mL PBS, and 4 mL chloroform and methanol mixture (2:1, v/v) was added and shaken well. This suspension was centrifuged at 1800 × g for 10 min to obtain the bottom chloroform phase containing lipid, which was subjected to blow-drying with nitrogen. The obtained lipids were resolved in 100 μL ethanol and used for determining TC and TG levels by using commercially available assay kits as described above.

Liver tissues were fixed with freshly prepared 4% paraformaldehyde for 24 h, embedded in paraffin by using a JB-P5 embedding machine (Wuhan, China), and cut into 7-µm-thick sections by using a RM2016 pathology slicer (Shanghai, China). Next, the sections were dewaxed in xylene, passed through a gradient of ethanol concentrations, and rinsed with tap water. Subsequently, the sections were stained with hematoxylin solution for 4 min, rinsed with tap water, and stained with eosin for 15 s. Finally, the sections were dehydrated and sealed with neutral gum (Zhang et al., 2022). For Oil Red O staining, fresh liver tissues were dissected, dried with a filter paper, trimmed into a cube shape (side length: 5 mm), and embedded in NEG-50™ frozen section medium. Next, the tissues were fixed on the slicer, and 8-µm-thick tissue sections were obtained using a Cryostar NX50 machine (Thermo, United States). The sections were subsequently stained with freshly prepared Oil Red O dye solution (Yin et al., 2019). The stained images were recorded on an Axiocam 506 color camera (Zeiss, Jena, Germany), and the staining areas were analyzed by using ImageJ software.

HepG2 cells were seeded in 6-well plates at a density of 2.0 × 10⁵/well. The cells were cultured in a humidified 5% CO₂ at 37°C. These cells were randomly divided into blank, model, alirocumab (Pm, 10 μg/mL), SB203580 (p38i, 30 µM), and combination (Pm + i, alirocumab 10 μg/mL and SB203580 30 µM) groups. Except for the blank group, the rest groups were treated with 0.5 mM cis-9-octadecenoic acid and 0.25 mM palmitic acid for 24 h to establish a lipid-loaded cell model (Li et al., 2020; Zulkapli et al., 2023; Zhang et al., 2024). Next, the cells were treated with the corresponding drugs for another 24 h prior to sampling.

For the siRNA transfection experiments, the HepG2 cells were initially incubated with 0.5 mM cis-9-octadecenoic acid and 0.25 mM palmitic acid for 24 h to establish a lipid-loaded cell model. Next, the cells were randomly divided into negative control (NC), PCSK9 siRNA1 (siR-1), and PCSK9 siRNA2 (siR-2) groups. The siRNAs against PCSK9 were synthesized by Guangzhou RiboBio Co., Ltd. (Guangzhou, China) and the sequences were listed in Table 2 (Tang et al., 2012; Xiao et al., 2019). Briefly, the cells were transfected with the corresponding siRNA by using Lipofectamine 3,000 as reported previously (Tang et al., 2012; Deng et al., 2020). The cells were cultivated for another 24 h prior to sampling. In this study, the normal HepG2 cells were used as blank control. Moreover, the expression levels of several inflammation-related proteins in the blank control, model, and NC groups were investigated to confirm lipid accumulation-induced inflammation and to rule out the potential effects caused by the control plasmid.

Approximately 50 mg liver tissues were homogenized in 500 µL RIPA lysis buffer and allowed to digest on ice for 30 min. The protein concentration was adjusted to approximately 3.0 g/mL according to the results of the bicinchoninic acid assay. The cell sample in each well was treated with 60 µL RIPA lysis buffer to obtain total proteins. This experiment was conducted by following previously reported methods (Li et al., 2020; Fu et al., 2021; Zhang et al., 2022; Zulkapli et al., 2023; Zhang et al., 2024). In brief, equal amounts of proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis on gels with different concentrations (6%–12%). The separated proteins were transferred onto 0.45 μm polyvinylidene fluoride (PVDF) membranes (or 0.22 μm PVDF membranes for proteins with a molecular weight of <30 kDa) by electroblotting. These PVDF membranes were blocked with 5% nonfat dry milk dissolved in Tris-buffered saline (TBS) at room temperature for 2–4 h. Next, these PVDF membranes were incubated with the corresponding primary antibodies at 4°C for 12 h. The PVDF membranes were then washed with TBS containing 0.1% Tween for 3–5 times and incubated with the corresponding secondary antibodies for 2 h at room temperature. Finally, immunoblots were visualized by enhanced chemiluminescence reaction, and the images were captured with an SH-Compact 523 ECL imaging system, a product of Shenhua Science Technology Co., Ltd. (Hangzhou, China). Densitometry analysis was conducted using ImageJ software. The protein levels were normalized by housekeeping protein glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or β-actin.

An appropriate amount of liver tissue (∼50 mg) was ground to powder under liquid nitrogen protection in a RNase free mortar. The obtained powder was extracted with Trizol (Spark Jade, Qingdao, China) in accordance with the manufacturer’s instructions. The concentration and purity of total RNA were determined by estimating the absorbance at 260 nm and 280 nm with a NanoDrop™ OneC UV spectrophotometer (Thermo Fisher, United States). cDNA was subsequently synthesized in an ABI Veriti™ 96-well thermal cycler (MA, United States) in accordance with the manufacturer’s instructions. RT-qPCR was performed in an ABI QuantStudio3 PCR system (Waltham, MA, United States) as reported previously (Yang et al., 2019; Li et al., 2020; Li et al., 2022). Gene-specific primers (Table 2) were designed and synthesized by Sangon Biotechnology Co., Ltd. (Shanghai, China). The relative expression of target genes was adjusted based on the C _t number of GAPDH and calculated according to the 2^−ΔΔCt method.

Statistical analysis was performed using one-way analysis of variance (ANOVA). The results were presented as the mean ± standard deviation (SD) for at least three independent experiments. Calculations were performed using GraphPad Prism sofeware, version 7.0 (San Diego, CA, United States). Differences were considered to be significant at a P < 0.05.

After several rounds of breeding, we obtained colonies of hepatic PCSK9^liver(+/+) , PCSK9^liver(+/−) , and PCSK9^{liver(−/−)} mice based on the genotyping results (Supplementary Figure S1). We used three and two pairs of primers to determine the gene phenotype of ALB-Cre⁺ and PCSK9 ^flox/flox , respectively. Table 1 shows the primers and target molecular weight for identifying PCSK9^liver(+/+) , PCSK9^liver(+/−) , and PCSK9^{liver(−/−)} mice. Taking the mice numbered as 1–9 for example: mice number 3, 4, and 6–9 were determined as PCSK9^{liver(−/−)} mice, while mice number 1, 2, and 5 were determined as PCSK9^liver(+/−) mice according to the molecular weight of the observed bands following agarose gel electrophoresis (Supplementary Figures S1A–C). In summary, hepatic PCSK9 conditional knockdown mice were successfully bred in our laboratory. Next, we investigated whether hepatic PCSK9 knockdown ameliorates hyperlipidemia-induced inflammation. The in vivo experimental design was shown in Supplementary Figure S1D.

First of all, we demonstrated that HFD successfully induced hyperlipidemia in mice compared to those fed a chow diet (Figures 1I, J, Supplementary Figures S2A, B). Importantly, the levels of PCSK9, TNF-α, IL-1β, TLR-2, and TLR-4, and the levels of phosphorylated NF-κB, AP-1, MyD88, phosphoinositide 3-kinase (PI3K), protein kinase B (PKB/AKT), mTOR, p38-MAPK, and ERK1/2 proteins were significantly upregulated in the liver of PCSK9^liver(+/+) mice fed a high-fat diet, as compared to the mice fed a chow diet (Supplementary Figure S2 and Supplementary Material S1). These results revealed the hyperlipidemia induces inflammation in liver by affecting multiple signaling pathways. In the following sections, we aimed to determine the effects of conditional knockdown of hepatic PCSK9 and the underlying mechanisms of action on modulating hyperlipidemia-induced chronic liver inflammation.

The gene and protein expression levels of hepatic PCSK9 and LDLR were determined in mice fed an HFD. As expected, compared to PCSK9^liver(+/+) mice, PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice exhibited a decrease in the mRNA expression level of PCSK9 by 66.4% and 60.9%, respectively (P < 0.01, Figure 1A). Conditional knockdown of hepatic PCSK9 did not affect the LDLR gene expression (Figure 1B). Consistent with this observation, PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice showed a significant decrease in the hepatic PCSK9 levels by approximately 65.5% and 47.0%, respectively, as compared to PCSK9^liver(+/+) mice (P < 0.05, Figure 1C). The hepatic LDLR levels were significantly increased by approximately 102% (P < 0.01) and 46.7% in PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice, respectively, as compared to those in PCSK9^liver(+/+) mice (Figure 1D). Notably, PCSK9^{liver(−/−)} mice exhibited significantly higher LDLR level than PCSK9^liver(+/−) mice. As shown in Figures 1E, F, mice body weight and liver indices showed no significant changes (Supplementary Material S2). We also did not find alterations in hepatic TC levels among the mice groups (Figure 1G); however, the hepatic TG levels were increased by 57.8% in PCSK9^{liver(−/−)} mice as compared to that in PCSK9^liver(+/+) mice (Figure 1H). A noteworthy finding is PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice showed a decrease in plasma TC levels by 29.7% (P < 0.001) and 23.4% (P < 0.05), respectively, as compared to PCSK9^liver(+/+) mice (Figure 1I). Consistent with this finding, ÄKTA-FPLC revealed a reduction in TC levels in LDL and HDL particles in hepatic PCSK9 knockdown mice, particularly in PCSK9^{liver(−/−)} mice (Figure 1K and Supplementary Material S2). However, the groups showed no significant differences in plasma TG levels (Figure 1J). Additionally, H & E staining revealed no apparent alterations in the liver among the groups (Figure 1L and Supplementary Material S3). The Oil Red O staining areas increased by 28.6% and 19.7%, respectively, in the liver sections of the PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice as compared to those of the PCSK9^liver(+/+) mice (Figures 1M, N, Supplementary Material S3). This finding was consistent with the higher levels of TG accumulation in the liver of hepatic PCSK9^{liver(−/−)} mice as compared with those of PCSK9^liver(+/+) mice (Figure 1H).

Next, we assessed whether hepatic PCSK9 affects hyperlipidemia-induced expression of proinflammatory cytokines in mice liver. Notably, conditional knockdown of hepatic PCSK9 remarkably decreased the mRNA expression levels of IL-6, IL-1β, and TNFα by 51.4% (P < 0.05), 65.6% (P < 0.001), and 66.1% (P < 0.001), respectively, in PCSK9^{liver(−/−)} mice as compared to those in PCSK9^liver(+/+) mice (Figures 2A–C). Consistent with this finding, PCSK9^{liver(−/−)} mice showed a decrease in IL-6, IL-1β, and TNFα protein levels by approximately 50.3% (P < 0.05), 49.7% (P < 0.01), and 58.3% (P < 0.01), respectively, as compared to PCSK9^liver(+/+) mice (Figures 2E–G). Moreover, IL-6, IL-1β, and TNFα expression levels were reduced by 54.4% (P < 0.05), 49.3% (P < 0.001), and 89.7% (P < 0.001), respectively, in PCSK9^liver(+/−) mice as compared to those in PCSK9^liver(+/+) mice (Figures 2A–C). In line with these changes, PCSK9^liver(+/−) mice exhibited a similar reduction in IL-6, IL-1β, and TNFα protein levels as compared to PCSK9^{liver(−/−)} mice (Figures 2E–G). Notably, PCSK9^{liver(−/−)} mice showed a further reduction in the IL-1β protein level as compared to PCSK9^liver(+/−) mice (P < 0.01, Figure 2F). However, PCSK9^liver(+/−) mice showed a significant reduction in TNFα mRNA (P < 0.001, Figure 2C) and protein expression levels (P < 0.05, Figure 2G) as compared to PCSK9^{liver(−/−)} mice.

Given the important roles of TLR in inflammation, we investigated whether this signaling pathway was involved in the abovementioned process. Notably, hyperlipidemia stimulation significantly decreased the mRNA expression levels of TLR-4 (Figure 2D) and TLR-2 (Figure 2I) by 69.1% (P < 0.001) and 68.1% (P < 0.05), and 52.1% (P < 0.01) and 38.6% (P < 0.05), respectively, in PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice as compared to that of PCSK9^liver(+/+) mice. Although both PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice groups exhibited a reduction in TLR-4 protein levels, only the former group showed a significant difference in the TLR-4 protein level as compared to the PCSK9^liver(+/+) group (P < 0.05, Figure 2H). Conditional knockdown of hepatic PCSK9 also significantly reduced the TLR-2 protein level by approximately 58.3% and 54.8%, respectively, in both PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice groups (P < 0.05, Figure 2M). Moreover, PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice groups showed a reduction in MyD88 expression levels by 85.8% (P < 0.001) and 47.8% (P < 0.01), respectively, as compared to PCSK9^liver(+/+) mice, with a significant difference between hepatic PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice groups (P < 0.001, Figure 2J). The phosphorylation levels of MyD88 were also decreased by approximately 59% in both PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice groups as compared to that in the PCSK9^liver(+/+) group (P < 0.01, Figure 2N). Importantly, conditional knockdown of hepatic PCSK9 significantly reduced the gene expression (P < 0.05) and phosphorylation levels of p65-NF-κB (P < 0.01) as compared to that in PCSK9^liver(+/+) mice (Figures 2K, O). Additionally, the AP-1 gene expression level was decreased by approximately 72.6% in PCSK9^{liver(−/−)} mice, but not in PCSK9^liver(+/−) mice, as compared to that in PCSK9^liver(+/+) mice (P < 0.001, Figure 2L). Consistent with this observation, PCSK9^{liver(−/−)} mice exhibited a significant reduction in the phosphorylation level of the AP-1 protein as compared to PCSK9^liver(+/+) or PCSK9^liver(+/−) mice (P < 0.001, Figure 2P).

Next, we investigated the alterations in hepatic MAPK and PI3K/AKT signaling pathways following hyperlipidemia challenge. As shown in Figures 3A–C, the PCSK9^{liver(−/−)} mice group showed significant reductions in the mRNA expression of p38-MAPK, ERK1/2, and JNK by 29.5% (P < 0.05), 59.6% (P < 0.01), and 49.2% (P < 0.001), respectively, as compared to that in the PCSK9^liver(+/+) mice group. There were similar reductions in the expression of p38-MAPK and ERK1/2, but not JNK, in the PCSK9^liver(+/−) mice group (P < 0.05). Moreover, in both PCSK9^liver(+/−) and PCSK9^{liver(−/−)} mice groups, the levels of phosphorylated p38-MAPK and ERK1/2 proteins were downregulated by approximately 60% (P < 0.01, Figure 3E) and 50% (P < 0.05, Figure 3F), respectively, as compared to those in the PCSK9^liver(+/+) mice group. However, the levels of phosphorylated JNK exhibited no significant alterations among the mice groups (Figure 3G). Moreover, in PCSK9^{liver(−/−)} mice, the gene expression of PI3K and AKT was decreased by 79.1% (P < 0.001) and 58.6% (P < 0.05), respectively, as compared to that in PCSK9^liver(+/+) mice (Figures 3D, I). Consistent with this finding, in PCSK9^{liver(−/−)} mice, the levels of phosphorylated PI3K and AKT proteins were downregulated by 60.8% (P < 0.01, Figure 3H) and 62.3% (P < 0.01, Figure 3J), respectively, as compared to those in PCSK9^liver(+/+) mice. Moreover, the mRNA expression of PI3K and AKT in PCSK9^liver(+/−) mice was downregulated by 54.4% (P < 0.001) and 38.3%, respectively, as compared to that in PCSK9^liver(+/+) mice. Consistent with this finding, the levels of phosphorylated PI3K and AKT proteins in PCSK9^liver(+/−) mice were decreased by 50.4% and 41.0%, respectively, as compared to those of PCSK9^liver(+/+) mice (P < 0.05, Figures 3H, J). Additionally, PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice showed a reduction in mTOR expression levels by 86.7% (P < 0.001, Figure 3K) and 42.7% (P < 0.01, Figure 3L), respectively, as compared to that in PCSK9^liver(+/+) mice. Consistent with this finding, the levels of phosphorylated mTOR protein were downregulated by approximately 39.2% (P < 0.001, Figure 3K) and 28.7% (P < 0.01, Figure 3L) in PCSK9^{liver(−/−)} and PCSK9^liver(+/−) mice, respectively.

Firstly, we demonstrated that the TC and TG levels in HepG2 cells were significantly increased after combined cis-9-octadecenoic acid and palmitic acid stimulation for 24 h, suggesting that a lipid-laden model was successfully established (Supplementary Figures S3A, B). Next, we demonstrated that lipid-loaded HepG2 cells exhibited higher levels of PCSK9, TNFα, IL-1β, IL-6, TLR-2, and TLR-4, and phosphorylated levels of NF-κB, MyD88, p38-MAPK, ERK1/2, JNK, PI3K, AKT, and mTOR proteins as compared to the blank control group. These findings suggested that lipid accumulation promoted inflammation in lipid-loaded HepG2 cells by regulating multiple signaling pathways as observed in mice. Importantly, there were no significant differences between the model and negative control groups, suggesting that the addition of control plasmid could not affect the levels of these proteins in lipid-loaded HepG2 cells (Supplementary Figure S3C–M and Supplementary Material S1). In the following, we investigated the effect of PCSK9 knockdown on lipid-loaded HepG2 cells and the underlying mechanisms of action. Lipid accumulation markedly increased the PCSK9 protein level by 136.2% (P < 0.001) in the NC group, while siR-1 and siR-2 intervention groups significantly decreased the PCSK9 protein level (P < 0.001, Figure 4A), as compared to that in the blank control group; this finding confirmed the effectiveness of these two PCSK9 siRNAs. Notably, TNFα, IL-6, and IL-1β protein levels were remarkably increased in the NC group by approximately 91.3% (P < 0.001), 70.4% (P < 0.01), and 92.3% (P < 0.01), respectively, as compared to those in the blank group, thus suggesting that lipid accumulation induced an inflammatory state in HepG2 cells. Compared to the NC group, the siR-1 intervention group showed a marked reduction in TNFα, IL-6, and IL-1β protein levels by 70.6% (P < 0.001), 40.6% (P < 0.001), and 26.5% (P < 0.05), respectively, while the siR-2 intervention group reduced their levels by 68.2% (P < 0.001), 57.9% (P < 0.001), and 21.8% (P < 0.05), respectively (Figures 4B–D). Lipid accumulation also elevated the TLR-2 and TLR-4 protein levels by approximately 40.0% and 53.5%, respectively, in the NC group (P < 0.05, Figures 4E, F). siR-2 alone significantly decreased TLR-2 and TLR-4 protein levels by 66.1% (P < 0.001) and 34.5% (P < 0.05), respectively. Although the NC group showed an increase by approximately 29.1% in the MyD88 protein phosphorylation level, this increase was not significant as compared to that in the blank group. In contrast, both siR-1 and siR-2 treatment significantly reduced MyD88 phosphorylation levels by approximately 60% as compared to that in the NC group (P < 0.01, Figure 4G). Notably, lipid accumulation in HepG2 cells significantly increased the p65 NF-κB protein phosphorylation level by 111.1% (P < 0.001, Figure 4H) without affecting AP-1 protein phosphorylation level as compared to that in the blank group (Figure 4I). However, transfection with siR-1 markedly decreased p65 NF-κB and AP-1 protein phosphorylation levels by approximately 77.6% (P < 0.001) and 54.3% (P < 0.001), respectively, while transfection with siR-2 decreased their levels by 80.9% (P < 0.001) and 52.4% (P < 0.01), respectively (Figures 4H, I), as compared to those in the NC group.

Additionally, as observed in the NC group, lipid accumulation significantly elevated p38-MAPK, ERK1/2, and JNK protein phosphorylation levels by 54.4% (P < 0.05), 208.1% (P < 0.001), and 55.5% (P < 0.01), respectively, as compared to the blank control group (Figures 4J–L). In contrast, as compared to the NC group, both siR-1 and siR-2 for PCSK9 significantly decreased p38-MAPK, ERK1/2, and JNK protein phosphorylation levels by more than 47% (P < 0.01), 23% (P < 0.01), and 65% (P < 0.001), respectively (Figures 4J–L). Lipid accumulation did not alter PI3K, AKT, and mTOR protein phosphorylation levels as compared to those in the blank control group (Figures 4M–O). However, as compared to the NC group, the siR-1 and siR-2 intervention groups showed a significant reduction in PI3K protein phosphorylation levels by 51.3% and 48.1%, respectively (P < 0.01, Figure 4M). siR-1 and siR-2 for PCSK9 also remarkably decreased the AKT protein phosphorylation levels by approximately 33.6% and 24.1%, respectively (P < 0.05, Figure 4N). Additionally, the mTOR protein phosphorylation levels in the siR-1 and siR-2 treatment groups were reduced by 64.1% (P < 0.01) and 87.3% (P < 0.001), respectively, as compared to those in the NC group (Figure 4O).

Finally, we investigated the role of PCSK9 in lipid-laden HepG2 cells and the underlying mechanisms of action. As shown in Figure 5A, lipid accumulation increased PCSK9 protein levels by 20.5% (P < 0.01) in HepG2 cells, while treatment of these cells with the anti-PCSK9 antibody alirocumab or the p38-MAPK inhibitor SB203580 alone or in combination significantly decreased PCSK9 protein levels by 65.3%, 69.6%, and 65.8%, as compared those in the model group (P < 0.001). Furthermore, compared to the blank control group, the model group showed an increase in the LDLR protein levels by 28.3%; however, neither alirocumab nor SB203580 significantly increased the LDLR protein level (Figure 5B). Notably, lipid accumulation significantly increased IL-6, IL-1β, and TNFα protein levels in the model group by approximately 52.7% (P < 0.01), 174.6% (P < 0.001), and 65.3% (P < 0.05), respectively, in the model group as compared to those in the blank control group (Figures 5C–E), while alirocumab or SB203580 alone or in combination attenuated these changes. A noteworthy finding is that the IL-6 protein levels in the combination group were significantly reduced as compared to that in the alirocumab alone intervention group (P < 0.05, Figure 5C). TLR-4 and TLR-2 protein levels were significantly increased by 36.1% and 137.6%, respectively, in lipid-loaded HepG2 cells as compared to those in the blank control group (P < 0.01, Figures 5F,G). Intervention with anti-PCSK9 antibody significantly decreased TLR-4 level (∼35.7%, P < 0.01) and particularly TLR-2 level (∼66.6%, P < 0.001) in lipid-laden HepG2 cells. SB203580 exhibited similar effects as those observed for alirocumab. However, alirocumab addition following SB203580 intervention did not further decrease TLR-4 or TLR-2 protein levels. Moreover, lipid accumulation significantly increased the phosphorylated MyD88 protein level by 70.2% (P < 0.01), while drug intervention significantly reversed this increase by approximately 65.9% (P < 0.001, Figure 5H). Importantly, alirocumab or SB203580 alone or in combination significantly decreased p65-NF-κB phosphorylation by approximately 36% (P < 0.05, Figure 5I) and reduced the phosphorylation levels of AP-1 by 61.0%, 75.5%, and 83.5%, respectively, as compared to those in the model group (P < 0.001, Figure 5J). The combination group showed a further decrease in the phosphorylated AP-1 protein level by 57.8% as compared to that in alirocumab alone intervention group (P < 0.01, Figure 5J).

Lipid accumulation also significantly increased the phosphorylation levels of p38-MAPK (∼30.6%, P < 0.05, Figure 5K), ERK1/2 (∼56.6%, P < 0.05, Figure 5L), and JNK (∼59.9%, P < 0.01, Figure 5M) proteins, as compared to those of the blank control group. Furthermore, the p38-MAPK and JNK proteins, but not ERK1/2 protein, showed a decrease in the phosphorylation levels by 23.5% (P < 0.05) and 26.6% (P < 0.05), respectively, in the alirocumab alone intervention group as compared to that in the model group. As a positive control, SB203580 significantly reduced p38-MAPK, ERK1/2, and JNK phosphorylation levels by approximately 46.1% (P < 0.001, Figure 5K), 45.4% (P < 0.05, Figure 5L), and 42.5% (P < 0.01, Figure 5M), respectively, as compared to those in the model group. Notably, the combination group also exhibited a decrease in the p38-MAPK phosphorylation level as compared to the alirocumab alone intervention group (P < 0.001, Figure 5K). Moreover, alirocumab further reduced the p38-MAPK protein phosphorylation level by 42.9% following SB203580 administration (P < 0.05, Figure 5K); this finding suggests that the anti-PCSK9 antibody alirocumab showed a synergistic effect with the p38-MAPK inhibitor SB203580. Notably, the combination group displayed the best effect in reducing the JNK phosphorylation level as compared to the model group (∼45.1%, P < 0.001, Figure 5M). Additionally, alirocumab or SB203580 alone or in combination remarkably decreased the phosphorylation levels of PI3K (∼30%, P < 0.05) and AKT (∼50%, P < 0.05) as compared to that in the model group (Figures 5N, O). Alirocumab, however, did not show a synergistic effect with SB203580 for reducing PI3K and AKT levels. Lipid accumulation in HepG2 cells also elevated the mTOR protein phosphorylation level by approximately 48.3% (P < 0.05) as compared to that in the blank control. In contrast, alirocumab, SB203580, or their combination significantly decreased the mTOR protein phosphorylation levels by 46.2% (P < 0.01), 55.3% (P < 0.001), and 74.6% (P < 0.001), respectively, as compared to those in the model group (Figure 5P). Furthermore, the combination intervention decreased the mTOR protein phosphorylation levels by 52.9% as compared to those in the alirocumab alone intervention group (P < 0.05, Figure 5P).