Gemilukast, 22-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl) amino)-23,24-bisnor-5-cholen-3β-ol (NBD-cholesterol), taurocholate and Filipin were purchased from J&K Scientific. All antibodies were purchased from Proteintech (Wuhan, Hubei, China). Commercial kits for total cholesterol (TC), triglycerides (TG), LDL-cholesterol (LDL-C), HDL-cholesterol (HDL-C), alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were purchased from Nanjing Jiancheng Bioengineering Institute. Enzyme-linked immunosorbent assay (ELISA) kits for TNF-α, IL-1β, and IL-6 were purchased from Boster Biological Technology Co., Ltd. The primers of IL-1β, IL-6, TNF-α, Arg-1, IL-10 and housekeeping gene of GAPDH were designed by Huayue Gene (Wuhan, China). The high-fat diet (HFD) feed, consisting of 77.6% maintenance feed, 10% lard, 10% yolk powder, 2% cholesterol, and 0.2% bile salt, was purchased from HuaFuKang Biotechnology (Beijing, China).

The cellular cholesterol-uptake assay was performed following the procedure of Jia et al. (23). NBD-cholesterol/taurocholate mixed micelles were prepared as described by Johnson et al. (27). Prior to uptake measurements, Caco-2 cells were cultured for 24 h in medium supplemented with lipoprotein-deficient serum. Cells were then rinsed three times with phosphate-buffered saline (PBS), incubated with gemilukast for 4 h, and subsequently exposed to NBD-cholesterol/taurocholate micelles for 1 h. After three additional PBS washes to remove non-internalized fluorescent cholesterol, fluorescence was recorded at 485 nm (excitation) and 535 nm (emission). Ezetimibe was included as a positive control. Vehicle control contained the same solvent concentration as the drug-treated groups. Fluorescence signals were background-subtracted and normalized to the vehicle control (set as 100%) for presentation.

The micellar cholesterol-solubility assay was adapted from Su et al. (28). Briefly, a bile salt micellar solution was prepared by combining 0.4 mM cholesterol, 10 mM sodium taurocholate, 132 mM NaCl, 1 mM oleic acid, and 15 mM sodium phosphate buffer (pH 6.8), followed by sonication until clarity. The suspension was equilibrated at 37°C for 24 h. Aliquots (1.0 mL each) of the micellar solution were then supplemented with gemilukast at the indicated concentrations (vehicle matched across groups), shaken in a water bath at 37°C for 2 h, and centrifuged at 15,000 rpm for 20 min. The supernatant was carefully collected without disturbing the pellet, and micellar cholesterol was quantified using an enzymatic assay according to the manufacturer’s instructions.

A 5 mg/mL Filipin stock solution was freshly prepared as described previously (23). The cholesterol-depleting medium was formulated according to Ge et al. (29). Caco-2 cells grown in confocal chambers were incubated in the depleting medium for 1 h, then treated with the indicated concentrations of gemilukast for an additional 1 h. Cells were rinsed three times with PBS and fixed in 4% formaldehyde. After two further PBS washes, cholesterol was visualized by staining with Filipin (50 μg/mL, 30 min, room temperature, protected from light), followed by three PBS washes. Images were acquired on a Nikon confocal microscope, with Filipin fluorescence displayed using red pseudo-color. For imaging, acquisition settings were kept constant across groups within each experiment. Fluorescence intensity was quantified using ImageJ by measuring mean fluorescence intensity per field (background-subtracted) from at least three randomly selected fields per condition.

Cytokines in culture supernatants were quantified using ELISA kits. For quantitative PCR, RAW 264.7 cells (5 × 10⁴ cells/mL) were stimulated with LPS (1.0 μg/mL) for 24 h. Total RNA was isolated with a commercial extraction kit and reverse-transcribed to cDNA. qPCR was performed with GAPDH as the endogenous control under the following cycling conditions: 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s, for 40 cycles. Primer sequences were designed using Primer 5.0 software (Supplementary Table 1). Each qPCR reaction was run in technical triplicates, and relative expression was calculated by the 2^-ΔΔCt method normalized to GAPDH.

Immunofluorescence staining was adapted from Yang et al. (30). Briefly, RAW 264.7 cells were seeded on coverslips in 24-well plates and stimulated with LPS (1.0 μg/mL) for 48 h. Cells were then treated with gemilukast (10 µM) for an additional 24 h and fixed in 4% paraformaldehyde. Primary antibodies against CD68 (1:100), iNOS (1:50), or CD206 (1:50) were applied at 4°C overnight, followed by incubation with the corresponding fluorescent secondary antibodies for 30 min. Nuclei were counterstained with DAPI and coverslips were mounted with antifade medium. Mean fluorescence intensity of iNOS and CD206 was quantified in ImageJ under identical acquisition settings for all groups.

Proteins were extracted and quantified by BCA assay. For each gel lane, 5-10 μg of total protein was loaded. After electrophoresis, proteins were transferred to membranes at a constant current of 250 mA, with transfer time adjusted to the molecular weight of the target. Membranes were then blocked for 15–25 min prior to antibody incubation. Membranes were incubated with primary antibodies at 4°C overnight, followed by incubation with HRP-conjugated secondary antibodies at room temperature. Protein bands were visualized using enhanced chemiluminescence. Signals were captured using an Amersham ImageQuant™ 800 imager (Cytiva). Band intensities were quantified by densitometry using ImageJ; phosphorylated proteins were normalized to the corresponding total proteins, and total proteins were normalized to GAPDH as indicated.

Animal experiments were performed under approval from the IACUC of Shandong First Medical University (W202403010164) in accordance with the NRC Guide for the Care and Use of Laboratory Animals. Male C57BL/6J mice and Sprague-Dawley (SD) rats (8 weeks) were purchased from Beijing Vital River (Beijing, China). After a 7-day acclimation at 25 ± 2°C with a 12:12 h light-dark cycle, animals were assigned to acute or long-term experimental protocols. At study endpoints, animals were humanely euthanized by cervical dislocation in accordance with institutional Standard Operating Procedures (SOPs). Trained personnel performed the procedure, and death was confirmed by cardiopulmonary arrest and fixed, dilated pupils.

In the acute lipid-absorption study, SD rats were randomly allocated to four groups (n = 5). A lipid emulsion was prepared from yolk powder and olive oil (1 g:10 mL, w/v) and sonicated for 10 min. Animals were fasted for 12 h before dosing. Experimental groups received the lipid emulsion by oral gavage (10 mL/kg); gemilukast was co-administered in the emulsion at 5 mg/kg or 10 mg/kg. Controls received an equal volume of water. Blood was collected at 1-h intervals for determination of plasma TC and TG.

In the long-term experiment, C57BL/6J mice were randomized into five groups (n = 5): control, high-fat diet (HFD), positive control (lovastatin), and two gemilukast groups (5 mg/kg/day and 10 mg/kg/day). Controls received standard chow, whereas the remaining groups were maintained on HFD. Lovastatin and gemilukast were administered once daily by oral gavage in the morning. Body weight was recorded weekly, and food intake was monitored every 2 days. At study termination, mice were euthanized and blood and tissues were collected for analysis; liver samples were processed for protein extraction and histopathology. Aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were measured using commercial assay kits following the manufacturer’s instructions. Fixed liver specimens were paraffin-embedded and sectioned at 3 µm. Sections were stained with hematoxylin and eosin (H&E) for histopathological evaluation of hepatic injury. Histopathological evaluation was performed in a blinded manner by an experienced investigator who was unaware of group allocation.

Data are presented as mean ± standard error of the mean. Group comparisons were performed by one-way analysis of variance (ANOVA). A two-sided P value < 0.05 was considered statistically significant. Analyses were conducted using GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA).

The chemical structure of gemilukast is shown in Figure 1A. Before evaluating the cholesterol-lowering and anti-inflammatory activities of gemilukast, its cytotoxicity was first assessed by cell viability assays in Caco-2 cells (cholesterol uptake model) and RAW 264.7 macrophages (inflammation model). As shown in Supplementary Figure 1, gemilukast exhibited no significant cytotoxicity at a concentration of 10 µM, with more than 90% of cells remaining viable. Therefore, a concentration of 10 µM was selected for subsequent experiments. Cholesterol uptake was evaluated in Caco-2 cells using the NBD-cholesterol assay, a commonly used method for assessing cholesterol absorption inhibitors (31). Gemilukast dose-dependently inhibited cholesterol uptake and achieved 51.6% inhibition at 10 μM versus vehicle group, exceeding the effect of ezetimibe (50 μM), an FDA-approved cholesterol absorption inhibitor (Figure 1B). Because micellar solubilization is required for cellular cholesterol uptake, we next examined whether gemilukast affects cholesterol solubilization in mixed micelles using a cholesterol-incorporation assay. The addition of gemilukast led to a distinct dose-dependent inhibition of cholesterol incorporation, as demonstrated in Figure 1C. Compared with micelles without gemilukast (control group), the cholesterol content in the micellar phase decreased by 41.5%, suggesting that gemilukast interfered with cholesterol incorporation into mixed micelles.

Filipin staining is widely used to visualize intracellular cholesterol (32). Here, the Filipin staining assay was also employed to further assess cellular cholesterol uptake (Figures 1D–F). The experimental sequence is illustrated in Figure 1D. Briefly, Caco-2 cells were first incubated for 1 h in a cholesterol-depleting medium to reduce intracellular sterol pools, followed by a 60 min exposure to gemilukast. Subsequently, a cholesterol-replenishing medium was applied. Confocal images were collected immediately before replenishment (-60 min) and after 60 min of replenishment. As shown in Figure 1E, 60 min in the depleting medium resulted in minimal Filipin fluorescence in Caco-2 cells. In contrast, 60 min of incubation with the replenishing medium markedly increased the cellular Filipin signal. Notably, gemilukast attenuated this fluorescence recovery, indicating effective inhibition of cholesterol uptake. Quantitative fluorescence analysis (Figure 1F) was concordant with these observations. Collectively, these findings support a mechanism whereby gemilukast limits cholesterol incorporation into mixed micelles, thereby restricting subsequent uptake by Caco-2 cells.

Gemilukast is a cysteinyl leukotriene receptor antagonist with antiasthmatic potential (21). To date, research on gemilukast has focused largely on target-level pharmacology, whereas studies addressing its anti-inflammatory activity and mechanisms at the cellular and in vivo levels remain limited. As shown in Figures 2A–C, cytokine analysis revealed that gemilukast significantly reduced the levels of pro-inflammatory cytokines TNF-α, IL-1β, and IL-6 at concentrations of 5 µM and 10 µM, indicating its anti-inflammatory activity. It has been demonstrated that macrophages play a crucial role in the initiation and progression of inflammation (33). M1 macrophages secrete pro-inflammatory cytokines that amplify inflammatory responses, whereas M2 macrophages release anti-inflammatory mediators that promote inflammation resolution (34). Therefore, promoting the polarization of macrophages from the M1 to the M2 phenotype represents an effective anti-inflammatory strategy. To further elucidate the anti-inflammatory mechanism of gemilukast, we investigated gemilukast’s effects on macrophage polarization. LPS-treated RAW 264.7 cells, exhibiting a typical M1 pro-inflammatory phenotype, were included as the positive control. Immunofluorescence staining was performed to analyze the expression of the M1 marker iNOS and the M2 marker CD206. Following gemilukast exposure, iNOS levels were reduced and CD206 staining was enhanced (Figure 2D), suggesting effective repolarization of macrophages from M1 to M2 type. In line with imaging data, fluorescence quantification indicated a decrease in M1 macrophages and an increase in M2 macrophages following gemilukast treatment (Figure 2E). To further confirm these findings, real-time polymerase chain reaction (RT-PCR) was conducted to evaluate the mRNA expression levels of macrophage markers, including IL-1β, IL-6, TNF-α, Arg-1, and IL-10. As shown in Supplementary Figure 2, compared with untreated cells, gemilukast significantly downregulated the expression of M1 markers (IL-1β, IL-6 and TNF-α) while upregulating M2 markers (Arg-1 and IL-10). These results indicate that gemilukast promotes the polarization of macrophages from the M1 to the M2 phenotype.

The PI3K/AKT signaling pathway is involved in the activation of M1 macrophages and represents a classic inflammatory signaling cascade that plays a crucial role in inflammatory diseases (35, 36). To investigate the regulatory effect of gemilukast on the PI3K/AKT signaling pathway, we examined the phosphorylation levels of PI3K and AKT (Figures 2F–H). Compared with the control group, LPS treatment significantly increased the ratios of p-PI3K/PI3K and p-AKT/AKT, whereas gemilukast markedly reduced these phosphorylation ratios relative to the LPS group, indicating that gemilukast inhibited the activation of the PI3K/AKT signaling pathway. Collectively, the current results indicate that gemilukast decreases M1 macrophage formation through modulation of PI3K/AKT signaling, leading to reduced inflammation.

To assess whether gemilukast impairs intestinal absorption of cholesterol and triglycerides in vivo, we measured plasma total cholesterol (TC) and triglycerides (TG) at serial time points in a rat model of acute hyperlipidemia (Figures 3A–C). Following oral lipid emulsion, TC in the model group rose markedly relative to the control group, whereas gemilukast administration significantly lowered TC at 2 h (Figure 3B). At 10 mg/kg, gemilukast reduced plasma TC by 14.8% at the 2 h time point compared with the model group, indicating that gemilukast may interfere with the intestinal uptake of cholesterol. For TG, there were no significant differences between the high-dose gemilukast and control groups throughout the 4-hour assessment (Figure 3C), which is plausibly explained by the requirement for luminal hydrolysis and subsequent re-esterification of TG prior to systemic appearance, a process that likely delays detectable changes.

The hypolipidemic activity of gemilukast was further assessed in a C57BL/6J hyperlipidemia model. The schedule of long-term protocol is presented in Figure 3D. The FDA-approved lipid-lowering drug lovastatin was used as the positive control. The body weight of each mouse was monitored and documented once per week throughout the study period. The concentrations of TG, TC, LDL-cholesterol (LDL-C) and HDL-cholesterol (HDL-C) were measured at the conclusion of the experiment. The overall body weight trend is shown in Supplementary Figures 3A, and the weight gain pattern is presented in Figure 3E. The HFD group exhibited a significantly greater body weight gain than the control group. However, with similar food consumption among groups (Supplementary Figure 3B), the body weight of gemilukast-treated mice did not differ significantly from HFD group (Figure 3E), suggesting that gemilukast did not affect body weight. In addition, as shown in Figures 3F–I, the effects of gemilukast on HFD-induced plasma lipid levels were investigated. HFD-induced robust increases in plasma TG and TC, while gemilukast significantly decreased TC compared with HFD groups (by 25.9%) but had no impact on TG (Figures 3F, G). Furthermore, although low-dose gemilukast had no effect on LDL-C and HDL-C, high-dose gemilukast significantly decreased LDL-C and markedly increased HDL-C (Figures 3H, I). Collectively, in vivo studies show that gemilukast not only significantly lowers TC in an acute hyperlipidemia model, but also improves the lipoprotein profile through reducing LDL-C and increasing HDL-C in long-term hyperlipidemia model, without affecting body weight.

Inflammation frequently accompanies hyperlipidemia induced by HFD (37). To evaluate the anti-inflammatory effects of gemilukast in vivo, the levels of inflammatory cytokines were measured using ELISA. The results show that HFD feeding resulted in higher serum concentrations of IL-1β, IL-6 and TNF-α. However, high-dose gemilukast treatment significantly reduced the levels of these pro-inflammatory factors (Figures 4A–C). Specifically, at 10 mg/kg, gemilukast reduced serum IL-1β by 22.0%, IL-6 by 31.5%, and TNF-α by 36.0% compared with the HFD group. In addition, lovastatin lowered plasma TNF-α compared with HFD, but had no significant effect on IL-1β or IL-6, suggesting that its anti-inflammatory efficacy is inferior to that of gemilukast. Extensive evidence shows that hyperlipidemia can cause liver damage by provoking chronic inflammatory responses (38). Considering alanine transaminase (ALT) and aspartate aminotransferase (AST) are the sensitive markers of liver injury, in this study, the serum levels of liver enzymes ALT and AST were measured to evaluate the hepatotoxicity after treatment (Figures 4D, E). The HFD group showed dramatic increases in serum ALT and AST activities compared with the control group, while treatment with high-dose gemilukast significantly decreased the activities of serum ALT and AST. In addition, the hepatic enzyme activities in the lovastatin-treated group showed no significant difference compared with the HFD group, which is consistent with previous reports that statins may induce hepatotoxicity (39). Moreover, H&E staining showed that high-dose gemilukast markedly reduced the inflammatory infiltrates associated with hyperlipidemia, while mild inflammation persisted after lovastatin treatment (Figure 4F). To assess gemilukast’s anti-inflammatory mechanism in vivo, hepatic PI3K and AKT phosphorylation were examined (Figures 4G–I), which has been demonstrated to play an important role in inflammatory diseases. The HFD group showed significant elevation in the ratios of p-PI3K/PI3K and p-AKT/AKT compared with the control group. Interestingly, compared with the HFD group, gemilukast (10 mg/kg) produced a statistically significant inhibition in the ratios of hepatic p-PI3K/PI3K and p-AKT/AKT. Taken together, these findings indicate that gemilukast exhibits significant anti-inflammatory activity with lower hepatotoxicity than lovastatin in vivo, and the anti-inflammatory activity of gemilukast is, at least in part, mediated by the PI3K/AKT signaling pathway.