A total of 60 male C57BL/6 mice weighed 18–22 g [used as wild-type (WT) mice] were purchased from the Experimental Animal Center of Nanjing Medical University (Nanjing, China). Thirty MARCKS-deficient mice were purchased from Jackson Laboratory (Bar Harbor, ME, USA). All mice were provided with drinking water, and housed in a controlled environment with a temperature of 22±2°C and relative humidity of 60±10% under a 12-h light/dark cycle. This study was approved by the Ethics Committee on Animal Research at the Department of Otolaryngology-Head and Neck Surgery, Huai'an Hospital Affiliated to Xuzhou Medical College, Huaian, China.

The 90 animals were divided into 6 groups (15 in each group) as follows: i) wild-type mice fed the normal chow diet (WT/Chow); ii) wild-type mice fed the HF diet (WT/HF); iii) MARCKS-knockout mice fed the normal chow diet (MARCKS^−/−/Chow); iv) MARCKS-knockout mice fed the HF diet (MARCKS^−/−/HF); v) wild-type mice fed the HF diet combined with 15 mg/kg CA (purity >98%; BioBioPha, yunnan, China); and vi) wild-type mice fed the HF diet combined with 30 mg/kg CA. CA was dissolved in distilled water at 0.3 mg/ml and administered to the mice with free access. Mice given distilled water were used as controls (WT/Chow and MARCKS^−/−/Chow). CA administration was initiated when the mice were fed with HF diets. Mice were administered a standard diet containing most essential nutrients, such as vitamins A (≥14,000 IU), D (≥1,500 IU), E (≥120 IU), K (≥5 mg), B1 (≥13 mg), B2 (≥12 mg), B6 (≥12 mg), B12 (≥0.022 mg), biotin (≥0.2 mg) and niacin (≥60 mg) per kg. During the period of study, the fodder was changed to a HF diet (60 kcal% fat, D12492; Research Diets, USA) until the mice were sacrificed for further analysis. At the end of week 8, body weight was monitored and all experimental mice were sacrificed after 12 h of fasting. Eyeball blood was collected, and serum was obtained by centrifugation at 13,000 rpm for 15 min at 4°C and stored at −80°C for analysis. The whole liver tissues were harvested and weighed on a 4°C glacial table, and were either frozen in liquid nitrogen and kept at −80°C for analysis, or fixed in 4% paraformaldehyde for histological analysis. All the methods were carried out in accordance with the approved guidelines.

At the end of the feeding period, OGTT and ITT were conducted. The mice from the WT/Chow, WT/HF, MARCKS^−/−/HF and MARCKS^−/−/Chow groups were treated with 20% glucose dissolved in saline orally or injected with insulin intraperitoneally (1 U/kg body weight). Tail-vein blood was collected at 0, 30, 60, 90 and 120 min after glucose or insulin treatment, and centrifuged at 4000 × g for 10 min at 4°C to obtain serum for glucose assay.

TG, total cholesterol (TC), aspartate transaminase (AST), alanine transaminase (ALT), uric acid and non-esterified fatty acid (NEFA) levels in serum or liver tissue obtained from the WT/Chow, WT/HF, MARCKS^−/−/HF and MARCKS^−/−/Chow groups of mice were determined using biochemical kits (Nanjing Jiancheng Biotechnology, Nanjing, China) following the manufaturer's instructions. TG, TC, AST, and ALT in serum or liver from the Con, HF, CAL and CAH groups were measured using biochemical kits.

Mice from the WT/Chow, WT/HF, MARCKS^−/−/HF, MARCKS^−/−/Chow, CAL, and CAH groups were fasted for 6 h, after which their blood was analyzed for glucose measurement with a glucose meter (Bayer, Mishawaka, IN, USA). For insulin analysis, the mice were fasted for 6 h and plasma insulin levels were measured with an Ultra Sensitive Mouse insulin enzyme-linked immunosorbent assay (ELISA) kit from Crystal Chem (Downers Grove, IL, USA).

The serum from 6 groups of mice was analyzed for the levels of major inflammatory cytokines, such as TNF-α (#MTA00B), IL-6 (#M6000B), IL-1β (#MLB00C), IL-18 (#7625), IL-2 (#M2000), IL-4 (#M4000B), IL-6 (#D6050), IL-12 (#M1270) and IFNγ (#MIF00), by ELISA, following the manufacturer's instructions (R&D Systems, Inc., Minneapolis, MN, USA). Color changes were determined at 450 nm.

The tissue isolated from WT/Chow, WT/HF, MARCKS^−/−/HF and MARCKS^−/−/Chow groups of mice were fixed with 10% buffered formalin, imbedded in paraffin and sliced into 4–5 µm thick sections. Following hematoxylin and eosin (H&E) staining, the pathological changes of the liver tissues were observed under a light microscope. Hepatic lipid content in mice from WT/Chow, WT/HF, MARCKS^−/−/HF, MARCKS^−/−/Chow, CAL, and CAH groups was determined using 5-µm-thick frozen sections stained with Oil Red O (Sigma-Aldrich). Masson's trichrome staining was performed according to the manufacturer's instructions (Diagnostic Biosystems Inc., Pleasanton, CA, USA). In addition, some tissues also were subjected to immunohistochemical (IHC) staining for the analysis of MARCKS (except for MARCKS^−/−/HF and MARCKS^−/−/Chow groups) and PPARα from each group of mice expression. The sections were stained with MARCKS (ab51100; Abcam, Shanghai, China) and PPARα (ab8934; Abcam). All histological examinations were carried out according to standard procedures reported previously (27). Furthermore, immunofluorescence assays for PPARα for liver tissue were performed according to the manufacturer's instructions.

The liver tumors were homogenized into 10% (wt/vol) hypotonic buffer (pH 8.0, 1 mM EDTA, 5 µg/ml leupeptin, 25 mM Tris-HCl, 1 mM Pefabloc SC, 5 µg/ml soybean trypsin inhibitor, 50 µg/ml aprotinin, 4 mM benzamidine) to yield a homogenate. Primary antibodies were shown in Table I. The final supernatants from cells and tumors were then obtained by centrifugation at 14,000 × g for 20 min at 4°C. The protein concentration was determined using a BCA protein assay kit (Thermo Fisher Scientific, Inc., Waltham, MA, USA) with bovine serum albumin as a standard. Sample-loading buffer was added, the mixture was boiled for 5 min. And the total protein extract was used for western blot analysis. Total protein (40 µg) was loaded and proteins were separated using 10% SDS-PAGE and electrophoretically transferred to the polyvinylidene difluoride membranes (Millipore, Billerica, MA, USA). The membranes were then blocked with 5% skim milk Tris-buffered saline with 0.1% Tween 20 (TBST), washed, and then incubated with primary antibodies overnight at 4°C. The membrane was then washed with TBST 3 times, followed by incubation with a horseradish peroxidase (HRP)-conjugated secondary antibody (1:2,500; KeyGen Biotech, Nanjing, China) at room temperature for 2 h. Following another round of washing with TBST, the membrane was then developed using ECL (Thermo Fisher Scientific, Inc.), and exposed to Kodak X-ray film (Eastman Kodak Company, Rochester, Ny, USA). Every protein expression level was defined as grey value using ImageJ 1.38 software (National Institutes of Health, Bethesda, MD, USA) and standardized to housekeeping gene of GAPDH and expressed as a fold of control. All experiments were performed in triplicate and done three times independently.

For analysis of qPCR, was performed as previously described (28). Total RNA was extracted from liver tissues using TRI-Reagent (Sigma-Aldrich) following the manufacturer's instructions and treated with deoxyribonuclease I. First-strand cDNA was then synthesized and amplified from 0.5 µg of total RNA using the ReverTra Ace qPCR RT kit (Toyobo, Tokyo, Japan). Real-time PCR was carried out for 35 cycles of 95°C for 20 sec, 54°C for 30 sec, and 72°C for 30 sec using SyBR Green Real-time PCR Master Mix (PE Applied Biosystems, Foster City, CA, USA) in a total volume of 20 µl. Fold changes in mRNA levels of target gene relative to the endogenous cyclophilin control were calculated. Briefly, the cycle threshold (=Ct) values of each target gene were subtracted from the Ct values of the housekeeping genecy clophilin (ΔCt). Target gene ΔΔCt was calculated as ΔCt of target gene minus ΔCt of control. The fold change in mRNA expression was calculated as 2^−ΔΔCt. The primers used in the study were showed in Table II.

qPCR analysis was performed as previously described (23). Fold induction values were calculated using the to 2^−ΔΔCq method, where ΔCq represents the differences in cycle threshold number between the target gene and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and ΔΔCq represents the relative change in the differences between the control and treatment groups. The primers used in this study are listed in Table II.

Data are expressed as the means ± SD. The treated cells, tissues and corresponding controls were compared using GraphPad PRI SM (version 6.0; GraphPad Software, Inc., La Jolla, CA, USA) by a one-way ANOVA with Dunn's least significant difference tests or Student's t-tests. Differences between groups were considered significant at p<0.05.

At the end of the experimental period, there was a striking difference in liver morphology in the Chow group compared to the high-fat diet feeding mice (Fig. 1A). As shown in Fig. 1A, the mice with MARCKS exhibited more severe liver injury. Significantly, the livers from the HF diet-fed MARCKS-deficient mice were pale in color, suggesting the evident accumulation of liver lipids in comparison to the HF diet-fed wild-type mice. Based on the liver morphology, we hypothesized that MARCKS might have a potential role in regulating lipid accumulation in the liver. Additionally, the results of H&E staining revealed that liver injury was induced in the HF diet-fed wild-type mice compared with the mice fed the normal chow. Importantly, the liver inflammation score was the highest in the HF diet-fed MARCKS-deficient mice (Fig. 1B). These data indicated MARCKS may be of great importance in preventing liver injury induced by HF. Similarly, Masson's Trichrome staining indicated that liver fibrosis was mildly increased in wild-type, but more robustly increased in MARCKS^−/− mice responding to HF diet (Fig. 1C). Indeed, the number of lipid droplets were also higher in the HF diet-fed wild-type mice compared with the normal chow-fed mice. However, an even higher number of lipid droplets was observed in the livers of the MARCKS-deficient mice fed the HF diet (Fig. 1D). These data indicated that MARCKS deficiency promoted liver injury and liver lipid accumulation in mice fed with HF diet.

We measured the body weight at the end of the experimental period. As shown in Fig. 2A, body weight was increased in the wild-type mice fed the HF diet compared with those fed the normal chow. MARCKS deficiency increased the body weight of the HF diet-fed mice even further (Fig. 2A). The brown adipose tissue weight was decreased in the HF diet-fed mice, and this was further decreased in the MARCKS-deficient mice fed with HF (Fig. 2B). In addition, liver weight was increased in the wild-type mice in response to the HF diet, and an even greater increase was observed in the MARCKS-deficient mice fed the HF diet (Fig. 2C). Similarly, serum uric acid was highly increased in wild-type mice induced by HF. In the MARCKS-deficient⁻ mice, the uric acid levels were even higher than those of the wild-type mice fed the HF diet (Fig. 2D). The HF diet-fed wild-type mice also exhibited significantly higher serum levels of TGs, TC and NEFA than the chow diet-fed wild-type mice. The mice with MARCKS deficiency fed the HF diet exhibited even higher levels of serum TGs, TC and NEFA (Fig. 2E–G). Consistent with the serum alterations, the hepatic TG and TC contents were significantly higher in the HF diet-fed wild-type mice compared with the chow diet-fed wild-type mice. The MARCKS-deficient mice fed the HF diet exhibited even higher TG and TC (Fig. 2H and I). These results indicated that MARCKS deficiency intensified the HF diet-induced accumulation of TGs and TC.

We then determined whether MARCKS deficiency is is involved in glucose and insulin resistance in HF diet-fed mice. The fasting blood glucose levels and serum insulin levels were increased in HF-fed mice, which was robustly upregulated in responding to MARCKS deficiency (Fig. 3A and B). We then performed a glucose tolerance test. Consistent with their elevated fasting glucose and insulin concentrations on the chow diet, MARCKS^−/− mice were more glucose intolerant than the mice on chow diet (Fig. 3C). The mice exhibited increased AST and ALT in serum as well as the increased concentrations of hepatic AST and ALT after HF diet (Fig. 3D–G). The MARCKS-deficient mice had even higher AST and ALT levels in serum and in the liver tissues, particularly those fed the HF diet (Fig. 3D–G).

HF diet-induced liver injury is always associated with inflammation (29). Thus, in this regard, we evaluated pro-inflammatory cytokine expression. The data indicated that the expression levels of IL-1β, IL-18, TNF-α, IL-2, IL-4, IL-6, IL-12 and IFN-γ in serum were markedly increased in mice fed the HF diet compared to those fed the normal chow diet. The MARCKS-deficient mice fed the HF diet exhibited even further upregulated levels of these pro-inflammatory cytokines (Fig. 4A–H). To further confirm these results, we also measured the pro-inflammatory cytokine levels in liver tissue samples by RT-qPCR. The results revealed that the hepatic mRNA levels of IL-1β, IL-18, TNF-α, IL-2, IL-4, IL-6, IL-12 and IFN-γ were elevated in the HF diet-fed mice, and MARCKS deficiency further stimulated these genes expressions (Fig. 4I). These data indicated that the presence of MARCKS greatly attenuated inflammation in liver of mice fed with HF diet.

As previously demonstrated, the PI3K/AKT signaling pathway is regulated by MARCKS (30). Hence, we attempted to explore whether MARCKS could regulate NAFLD in mice induced by high fat diet through regulating PI3K/AKT. Western blot analysis was performed to examine the levels of PI3K/AKT. As shown in Fig. 5A, we found that the PI3K/AKT signaling pathway was upregulated in the wild-type mice fed the HF diet. MARCKS deficiency further stimulated PI3K and phosphorylated AKT. AKT has a close association with the NLRP3-regulated NF-κB signaling pathway (31). As shown in Fig. 5B, NLRP3 expression was upregulated in the HF-fed wild-type mice and in the MARCKS-deficient mice. Subsequently, NF-κB was activated through caspase-1 stimulation in the HF diet-fed wild-type mice, and this was further enhanced by MARCKS deficiency (Fig. 5B and C). NF-κB activation is responsible for the release of pro-inflammatory cytokines, including IL-1β and IL-18 (32). In this study, both IL-1β and IL-18 were highly expressed in the HF diet-fed mice, particularly in the MARCKS-deficient mice (Fig. 5C). These data indicated that MARCKS, consistent with previous results (33), may play a potential role in suppressing the inflammatory response via the PI3K/AKT and NLRP3 inflammasome signaling pathways.

NAFLD development has a close association with lipogenesis-related proteins. SREBP-1c, ACCα, FAS and SCD1 have been well known to play a crucial role in lipid generation (34–36). We thus then attempted to investigate the mechanisms through which MARCKS regulates lipid metabolism via these proteins. As shown in Fig. 6A, SREBP-1c was highly expressed in HF diet fed mice, something that was consistent with a previous report (37). Similarly, the levels of ACCα, FAS and SCD1 were also elevated in the mice fed the HF diet. Of note, MARCKS deficiency markedly increased these proteins levels even further (Fig. 6A). PPARα is known as an important factor regulating lipid metabolism (38). In this study, we found that PPARα was expressed at low levels in the HF-fed mice. MARCKS deficiency even further downregulated PPARα levels in the HF diet-fed mice (Fig. 6B). Consistently, the results of western blot analysis further proved the PPARα alterations in the MARCKS-deficient mice fed the HF diet (Fig. 6C). These data suggested that MARCKS may be associated with lipid metabolism in HF diet-fed mice, which influenced NAFLD development.

We then attempted to reveal how CA affects NAFLD induced by a HF diet in mice. At the end of the treatment period, the H&E staining results showed that the higher inflammatory score in HF-treated mouse, the livers were much larger. By contrast, a lower inflammatory score was observed in the livers of mice treated with CA (Fig. 7A). Furthermore, Oil Red O staining results showed that the lipid droplets in livers of mice with HF diet were much larger. By contrast, fewer lipid droplets were observed in the livers of mice treated with CA in a dose-dependent manner (Fig. 5B). As mentioned above, MARCKS has a potential role in regulating HF diet-induced NAFLD. Thus, we performed immunohistochemical assays and the results indicated that a a high number of MARCKS-positive cells was observed in the control group; however, only a few MARCKS-positive cells were observed in the HF diet-fed mice. This number was increased by CA treatment (Fig. 7C). These data indicated that CA prevented liver injury and increased MARCKS expression in the livers of mice fed the HF diet.

The TG and TC levels in serum were markedly increased following feeding on the HF diet, consistent with the results mentioned above; however, these levels were reduced in the HF diet-fed mice treated with CA (Fig. 8A and B). In addition, TG and TC contents in liver induced by HF diet were significantly decreased for CA administration in a dose-dependent manner (Fig. 8C and D). We then examined whether treatment with CA improves glucose and insulin tolerance in the HF-fed mice. The fasting blood glucose levels and the plasma insulin levels were decreased following treatment with CA (Fig. 8E and F). Additionally, the serum AST and ALT levels were upregulated in the HF diet-fed mice, and were downregulated following treatment with CA (Fig. 8G and H). Similarly, the hepatic AST and ALT contents were also increased in HF groups compared to that in the Con group, which were decreased in HF diet-fed mice treated with CA (Fig. 8I and J). These results clearly indicated that CA treatment suppressed lipid accumulation by reducing TG and TC in serum and in liver, improving liver injury in mice administered with high fat diet.

Accelerated NAFLD is often associated with the increased production of inflammatory cytokines (39). Therefore, in this study, we examined whether CA represses the productions of inflammatory cytokines. To examine this, we evaluated the expression levels of the inflammatory cytokines, IL-1β, IL-18, TNF-α, IL-2, IL-4, IL-6, IL-12 and IFN-γ, in serum. The data indicated that the expression levels of IL-1β, IL-18, TNF-α, IL-2, IL-4, IL-6, IL-12 and IFN-γ, which had been increased by the HF diet, were markedly decreased by CA in the mice fed the HF diet (Fig. 9A). As can be seen in Fig. 9B, CA suppressed pro-inflammatory cytokines release through RT-qPCR analysis (Fig. 9B). These data indicated that treatment with CA markedly attenuated inflammation in mice fed a HF diet.

To elucidate the precise mechanisms responsible for the effects of CA in HF diet fed-mice and explore whether it can target MARCKS to attenuate NAFLD, we examined MARCKS expressed levels. The results revealed that the MARCKS levels were affected by the administration of CA in the HF diet-fed mice. MARCKS protein levels, which had been decreased by the HF diet, were markedly increased by CA (Fig. 10A). The mRNA level of MARCKS was also decreased in the HF diet-fed mice, and this reduction was reversed significantly following treatment with CA (Fig. 10B). To further examine this, we performed immunostaining using an antibody against MARCKS. Our results revealed that a great number of MARCKS-positive cells was observed in the control group; however, only a few MARCKS-positive cells were observed in the HF diet-fed mice, which were stimulated in mice with HF diet (Fig. 10C). These data indicated that CA attenuates NAFLD in mice fed a HF diet by regulating MARCKS expression.

PI3K/AKT is a potential target of MARCKS in attenuating NAFLD based on the results mentioned above. A recent study demonstrated that CA treatment attenuated the inflammatory response by regulating the NF-κB signaling pathway (40). To confirm this conclusion, we measured the levels of PI3K/AKT in the livers of mice fed a HF diet and treated with CA. We found that the upregulation of PI3K and p-AKT were reduced by carnosic acid in liver of mice with HF diet induction (Fig. 11A). Increased PI3K/AKT was also a pivotal factor for inducing NF-κB activation (41). Therefore, we examined whether NLRP3 and NF-κB expression in the liver was altered by CA. The overexpression of NLRP3 and the phophorylated levels of NF-κB in the HF diet-fed mice were reduced by CA treatment, whereas the levels of total NF-κB were not affected by CA (Fig. 11B). Caspase-1 was stimulated in HF diet fed mice, which was down-regulated due to carnosic acid administration, leading to NF-κB inactivity and pro-inflammatory cytokines suppression, including IL-1β and IL-18 (Fig. 11C). These data indicated that PI3K/AKT and NLRP3/NF-κB in the liver may be involved in the inhibitory effects of CA against the inflammatory response in mice with NAFLD induced by a HF diet.

In the end of the treatment period, we investigated whether lipogenesis is affected by CA. To examine this, we measured the expression levels of several key transcriptional factors in lipogenesis, including SREBP-1c, ACCα, FAS and SCD1. The overexpression of SREBP-1c, ACCα, FAS and SCD1, induced by the HF diet, was significantly attenuated by CA (Fig. 12A). Moreover, we also examined the PPARα levels by immunohistochemical assays, and the results revealed that PPARα expression was downregulated in the HF diet-fed mice, and that this effect was reversed by CA (Fig. 12B). The mRNA level of PPARα was also decreased in the mice fed the HF diet, and similar to the data mentioned above, CA increased PPARα mRNA expression (Fig. 12C). These results clearly indicated that CA treatment suppressed lipogenesis in HF diet induced animals with NAFLD.