Amitriptyline inhibits nonalcoholic steatohepatitis and atherosclerosis induced by high-fat diet and LPS through modulation of sphingolipid metabolism

Abstract

We reported previously that increased acid sphingomyelinase (ASMase)-catalyzed hydrolysis of sphingomyelin, which leads to increases in ceramide and sphingosine 1 phosphate (S1P), played a key role in the synergistic upregulation of proinflammatory cytokines by palmitic acid (PA), a major saturated fatty acid, and lipopolysaccharide (LPS) in macrophages. Since macrophages are vital players in nonalcoholic steatohepatitis (NASH) and atherosclerosis, we assessed the effect of ASMase inhibition on NASH and atherosclerosis cooperatively induced by high-PA-containing high-fat diet (HP-HFD) and LPS in LDL receptor-deficient (LDLR−/−) mice. LDLR−/− mice were fed HP-HFD, injected with low dose of LPS and treated with or without the ASMase inhibitor amitriptyline. The neutral sphingomyelinase inhibitor GW4869 was used as control. Metabolic study showed that both amitriptyline and GW4869 reduced glucose, lipids, and insulin resistance. Histological analysis and Oil Red O staining showed that amitriptyline robustly reduced hepatic steatosis while GW4869 had modest effects. Interestingly, immunohistochemical study showed that amitriptyline, but not GW4869, strongly reduced hepatic inflammation. Furthermore, results showed that both amitriptyline and GW4869 attenuated atherosclerosis. To elucidate the underlying mechanisms whereby amitriptyline inhibited both NASH and atherosclerosis, but GW4869 only inhibited atherosclerosis, we found that amitriptyline, but not GW4869, downregulated proinflammatory cytokines in macrophages. Finally, we found that inhibition of sphingosine 1 phosphate production is a potential mechanism whereby amitriptyline inhibited proinflammatory cytokines. Collectively, this study showed that amitriptyline inhibited NASH and atherosclerosis through modulation of sphingolipid metabolism in LDLR−/− mice, indicating that sphingolipid metabolism in macrophages plays a crucial role in the linkage of NASH and atherosclerosis.

INTRODUCTION

Epidemiological studies have well established that type 2 diabetes mellitus (T2DM) and metabolic syndrome (MetS), which is mainly characterized by obesity, insulin resistance, hypertension, and dyslipidemia (14, 20, 48), are associated with an increased risk of nonalcoholic fatty liver disease (NAFLD), a common cause of chronic liver disorder (36, 50, 54, 55). NAFLD ranges from triglyceride accumulation (steatosis) to nonalcoholic steatohepatitis (NASH), which is characterized by steatosis, hepatic inflammation, and hepatocellular ballooning and may progress to liver fibrosis, cirrhosis, and liver cancer (41).

It is known that overconsumption of a high-fat diet (HFD) that contains high amounts of saturated fatty acids (SFA) such as palmitic acid (PA) plays an important role in the pathogenesis of T2DM and MetS as well as their complications including NAFLD (9). In addition to HFD, innate immunological stimuli such as lipopolysaccharide (LPS) have been also implicated in T2DM and MetS as it was reported that circulating LPS was increased in patients with T2DM or MetS (10, 13) as a result of increased gut permeability and changes in composition and diversity of gut microbiome (17). Studies have shown that LPS activates Toll-like receptors (TLRs) in macrophages and hepatocytes, stimulates TLR-mediated inflammatory response, and promotes a systemic inflammation, which contribute to the development of NASH (43, 52). To explore the mechanisms by which T2DM promotes NASH, we have demonstrated that feeding high PA-containing high-fat diet (HP-HFD) and administration of a low dose of LPS cooperatively induce NASH in LDL receptor-deficient (LDLR−/−) mice (24), indicating the immunometabolic cross talk plays an important role in the pathogenesis of NASH.

It is well known that patients with NASH also have an elevated risk of cardiovascular diseases (CVDs) (27). In fact, CVDs, but not liver-related disorders, are the number one killer of patients with NASH (7, 40). However, the mechanisms by which NASH is associated with CVDs have not been fully elucidated. Our study showed that feeding HP-HFD and administration of a low dose of LPS in LDLR−/− mice cooperatively induced not only NASH (24) but also atherosclerosis, the major cause of CVDs (28), suggesting that the HP-HFD-fed LDLR−/− mice treated with a low dose of LPS are a good animal model to investigate the mechanisms whereby NASH is associated with atherosclerosis.

To further unveil the mechanisms involved in the cooperative promotion of NASH and atherosclerosis by HP-HFD and LPS, we have focused on sphingolipid metabolism in macrophages (19, 24, 29). Our study showed that PA and LPS synergistically increased ceramide (CER), a major sphingolipid (16), by stimulating acid sphingomyelinase (ASMase)-mediated sphingomyelin (SM) hydrolysis in macrophages, resulting in an increase in sphingosine 1 phosphate (S1P), a bioactive lipid involved in the upregulation of proinflammatory cytokines (18, 19). Our study also showed that targeting ASMase with pharmacological inhibitors or RNA interference effectively attenuated the stimulatory effect of PA and LPS on proinflammatory cytokine expression in macrophages in vitro (19).

ASMase is a lipid-metabolizing enzyme cleaving SM to CER and regulated by proinflammatory cytokines, UV radiation, LPS, cytotoxic agents, and others (58). ASMase is expressed by most tissues of human (44). Based on the above findings that ASMase plays a crucial role in the upregulation of inflammatory response by SFA and LPS, we conducted the current study to test our hypothesis that ASMase-mediated sphingolipid metabolism in macrophages plays a pivotal role in the association between NASH and atherosclerosis induced by HP-HFD and LPS in LDLR−/− mice and inhibition of ASMase with amitriptyline, a pharmacological ASMase inhibitor, attenuates both NASH and atherosclerosis. To show the specific involvement of ASMase-mediated sphingolipid metabolism in NASH and atherosclerosis, we also inhibited neutral sphingomyelinase (NSMase) with GW4869 in the study. Our previous studies showed that NSMase is responsible for the synergistic effect of LPS and PA on proinflammatory cytokine upregulation in vascular endothelial cells (29, 30) but not macrophages (19).

MATERIALS AND METHODS

Animals, diet, and treatments.

Forty-eight 8-wk-old male LDLR^−/− mice were purchased from Jackson Laboratory (Bar Harbor, ME) and housed at the animal facility of the Veterans Affairs Medical Center in Charleston, SC. The animal protocol was approved by the Institutional Animal Care and Use Committee. All mice were maintained on a 12-h light-dark cycle in a pathogen-free environment and had ad libitum access to water and food. The mice were randomly divided into six groups (n = 8 per group): 1) HFD-fed mice without LPS and amitriptyline or GW4869 treatments; 2) HFD-fed mice with amitriptyline treatment; 3) HFD-fed mice with GW4869 treatment; 4) HFD-fed mice with LPS treatment; 5) HFD-fed mice with both LPS and amitriptyline treatments; and 6) HFD-fed mice with both LPS and GW4869 treatments. All mice were fed TD.06414 (Envigo RMS, Inc., Indianapolis, IN), which is a lard-based HFD containing 34.3% of fat, 23.5% of protein, and 27.3% of carbohydrate and high PA content (8% of total fat), for 20 wk. During the last 10 wk, mice in groups 4, 5, and 6 received intraperitoneal injection of LPS (Escherichia coli serotype 055:B5, Sigma, St. Louis, MO) in 200 μL of phosphate-buffered saline (PBS), 25 μg per mouse, once a week while the other groups received equivalent amount of PBS, the vehicle for LPS. The dose of LPS was reported to be effective to induce atherosclerosis by the previous studies (32, 53). In addition, mice in groups 2 and 4 received intraperitoneal injection of amitriptyline (0.5 mg·mouse⁻¹·day⁻¹) while mice in groups 3 and 6 received intraperitoneal injection of GW4869 (0.1 mg·mouse⁻¹·day⁻¹). The above dose of amitriptyline and the route of amitriptyline administration were reported to be effective in inhibiting ASMase in mice (4, 37). The above dose of GW4869 and the route of GW4869 administration were also reported to be effective in inhibiting NSMase in mice (23, 35).

Metabolic assays.

Blood samples were obtained under the fasted condition, and glucose level was determined using a Precision QID glucometer (MediSense, Inc., Bedford, MA). Serum cholesterol and triglycerides were assayed using Cholestech LDX Lipid Monitoring System (Fisher Scientific, Pittsburgh PA). Serum fatty acids were determined using the EnzyChrom free fatty acid kit (BioAssay systems, Hayward, CA). Serum fasting insulin was assayed using the Ultra Sensitive Insulin ELISA Kit (Crystal Chem, Inc., Downers Grove, IL). Fasting whole body insulin sensitivity was estimated with the homeostasis model assessment of insulin resistance (HOMA-IR) according to the formula [fasting plasma glucose (mg/dL) × fasting plasma insulin (μU/mL)]/405.

Histological examination of liver tissue.

The liver tissue was embedded in Tissue-Tek OCT compound (EMS, Hatfield, PA), immediately frozen on dry ice, and stored at −80°C. The tissue with 6-μm thickness was sectioned and mounted on glass microscope slides. The sections were fixed in 10% of formalin and stained with Harris-modified hematoxylin and eosin (H&E) solution (Sigma, St. Louis, MO). Slides were dehydrated and mounted in Cytoseal-XYL mounting medium (Fisher Scientific, Waltham, MA). Photomicrographs of tissue sections were taken using an Olympus BX53 digital microscope with Cellsens digital image software (Olympus American, Inc., Center Valley, PA).

Oil Red O staining.

For Oil Red O staining, the frozen sections were fixed with 10% formalin for 10 min, placed in 60% isopropyl alcohol, and stained in 0.5% Oil Red O solution for 10 min. The slides were transferred to 60% isopropyl alcohol, rinsed in distilled water, and processed for hematoxylin counter staining. Photomicrographs of tissue sections were taken using an Olympus BX53 digital microscope, and the positively stained area was quantified with Image-Pro Plus 6 (Media Cybernetics, Rockville, MD).

Immunohistochemical analysis of expression of F4/80, ASMase, and NSMase.

Liver tissues were fixed in 4% paraformaldehyde for 10 min, and frozen sections were made using a cryostate. Immunohistochemical analysis with anti-F4/80 antibodies (cat. no. MCA497; Bio-Rad Laboratories, Inc., Hercules, CA) was performed as described previously (31). Immunohistochemical analysis with anti-ASMase antibodies (cat. no. ab83354; Abcam, Cambridge, MA) and NSMase antibodies (cat. no. abb85017; Abcam) was also performed. Counterstaining was performed with hematoxylin. Photomicrographs of tissue sections were taken using an Olympus BX53 digital microscope, and the positively immunostained area was quantified with Image-Pro Plus 6 (Media Cybernetics).

En face Sudan IV staining and histological analysis of atherosclerotic lesions.

Mice were euthanized, and aortas from heart to the iliac arteries were dissected out, soaked for 24 h in 4% paraformaldehyde for fixation, excised longitudinally, and then stained with 0.5% of Sudan IV as described previously (26). After staining, the aortas were laid onto the sponge block surface with the intimal surface up and pinned down using the Minutien (Fine Science Tools, Inc., San Francisco, CA). The images of the aortas were taken using an EPSON Perfection 2450 photo scanner and analyzed with Image-Pro Plus 6 (Media Cybernetics). For histological analysis, the tissues of aortic roots were embedded in Tissue-Tek OCT compound (EMS), immediately frozen on dry ice, and stored at −80°C. Starting from the aortic root, cryosetions with 6-μm thickness were cut and sections with a distance of 480 mm were collected and mounted on slides. Slides were fixed in 10% of formaline for 10 min, stained with Harris-modified hematoxylin (Fisher Scientific, Pittsburgh, PA) for 10 min, and then rinsed in deionized water. Histological analysis of atherosclerotic lesions was performed as described previously (28).

Cell culture and treatment.

The murine macrophage cell line RAW264.7 has been used extensively in the investigations of the role of macrophages in inflammation-related diseases (38). RAW264.7 cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA) and grown in DMEM (ATCC) supplemented with 10% heat-inactivated fetal calf serum. The cells were maintained in a 37°C, 90% relative humidity, 5% CO₂ environment. For cell treatment, LPS from Escherichia coli (serotype 055:B5) was used. The LPS was highly purified by phenol extraction and gel filtration chromatography and was cell culture tested. PA (Sigma) used in this study was bovine serum albumin-free as described previously (45). To prepare PA, PA was dissolved in 0.1 N NaOH and 70% ethanol at 70°C to make PA solution at concentration of 50 mM. The solution was kept at 55°C for 10 min, mixed, and brought to room temperature.

Culture of bone marrow-derived macrophages.

Bone marrow-derived macrophages were isolated and cultured as described previously (25). Briefly, bone marrow cells were obtained from tibiae and femora of 4- to 6-wk-old C57BL/6 mice and cultured with α-MEM containing 10% fetal bovine serum in a humidified incubator (5% CO₂) at 37°C. After 24 h, nonadherent cells were incubated for 4 days in the presence of 50 ng/mL of macrophage colony-stimulating factor (Sigma-Aldrich, Atlanta, GA). It has been shown that almost all the adherent cells expressed macrophage-specific antigens, such as Mac-1, Moma-2, and F4/80 (21).

Enzyme-linked immunosorbent assay.

IL-6 in medium was quantified using sandwich ELISA kits according to the protocol provided by the manufacturer (Biolegend, San Diego, CA).

RNA isolation from liver tissues.

Total RNA was isolated from mouse liver tissues using the RNeasy minikit (Qiagen, Santa Clarita, CA) by following the instructions provided by the company.

Real-time polymerase chain reaction.

Real-time PCR was performed as described previously (19). The Beacon designer software (PREMIER Biosoft International, Palo Alto, CA) was used for primer designing (mouse IL-6: 5′-primer sequence, TGGAGTCACAGAAGGAGTGGCTAAG and 3′-primer sequence, TCTGACCACAGTGAGGAATGTCCAC; mouse NSMase: 5′-primer sequence, AGAAACCCGGTCCTCGTACT and 3′-primer sequence, CCTGACCAGTGCCATTCTTT). Primers were synthesized by Integrated DNA Technologies, Inc. (Coralville, IA). Mouse glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a control (5′-primer sequence, CTGAGTACGTCGTGGAGTC and 3′-primer sequence, AAATGAGCCCCAGCCTTC). Data were analyzed with the iCycler iQ software. Mouse ASMase primers (PPM25140A) were purchased from Qiagen. The average starting quantity (SQ) of fluorescence units was used for analysis. Quantification was calculated using the SQ of targeted cDNA relative to that of GAPDH cDNA in the same sample.

Lipidomics.

Macrophages were collected, fortified with internal standards, extracted with ethyl acetate/isopropyl alcohol/water (60:30:10, vol/vol/vol), evaporated to dryness, and reconstituted in 100 μl of methanol. Simultaneous electrospray ionization tandem mass spectrometry (ESI/MS/MS) analyses of sphingoid bases, sphingoid base 1-phosphates, CERs, and SMs were performed on a Thermo Finnigan TSQ 7000 triple quadrupole mass spectrometer operating in a multiple reaction monitoring positive ionization mode. The phosphate contents of the lipid extracts were used to normalize the MS measurements of sphingolipids. The phosphate contents of the lipid extracts were measured with a standard curve analysis and a colorimetric assay of ashed phosphate (49).

Quantification of S1P.

Cellular S1P content was quantified using a S1P ELISA kit (ABcloneal, Woburn, MA) by following the instruction provided by the manufacturer.

Statistical analysis.

GraphPad Instat statistical software (Version 5.0) (GraphPad Software, Inc., La Jolla, CA) was used for statistical analysis. The two-way ANOVA was used to compare the data between groups with different treatments. To determine the statistical significance of differences between two experimental groups, parametric analysis using Student’s t test was performed for data with normal distribution, and nonparametric analysis using Mann-Whitney test was performed for data without normal distribution. No correction of multiple comparisons was made, and the chance of committing to a type I error could be very high due to having many experimental groups. In the animal experiments, n indicates the number of mice in each group. P < 0.05 was considered significant.

RESULTS

Both amitriptyline and GW4869 reduce glucose, cholesterol, triglycerides, and insulin resistance in HP-HFD-fed LDLR−/− mice with LPS treatment.

We first determined the effect of amitriptyline and GW4869 on metabolic parameters of HP-HFD-fed LDLR−/− mice with or without LPS treatment. As shown in Fig. 1, results showed that HP-HFD induced T2DM in LDLR−/− mice since glucose level reached 300 mg/dl, which is considered as the criteria for diabetes in mouse model (1). The addition of LPS to HFD did not significantly change body weight, glucose, total cholesterol, triglycerides, and HOMA-IR but reduced free fatty acid and insulin. Amitriptyline reduced body weight and glucose in HP-HFD-fed mice with LPS treatment and reduced total cholesterol, triglycerides, insulin, and HOMA-IR in HP-HFD-fed mice with or without LPS treatment. However, amitriptyline increased free fatty acids in HP-HFD-fed mice with LPS treatment. GW4869 inhibited body weight, free fatty acid, insulin, and insulin resistance in HP-HFD-fed mice without LPS treatment and inhibited glucose, total cholesterol, triglycerides, free fatty acid, insulin, and insulin resistance in HP-HFD-fed mice with LPS treatment. Taken together, these data indicate that inhibition of either ASMase or NSMase improved glucose, cholesterol, triglycerides, and insulin resistance in HP-HFD-fed mice treated with LPS.

Fig. 1.

A–G: the effects of amitriptyline (AMI) and GW4869 on the metabolic parameters including body weight (A), glucose (B), total cholesterol (C), triglycerides (D), free fatty acid (E), insulin (F), and insulin resistance [homeostasis model assessment of insulin resistance (HOMA-IR); G] in high-palmitic acid-containing high-fat diet (HP-HFD)-fed mice with or without LPS treatment. LDLR−/− mice were fed HP-HFD for 20 wk and treated with or without LPS in the last 10 wk. Part of the mice were also treated with amitriptyline or GW4869 during the last 10 wk. After the feeding and treatments, the metabolic parameters were quantified. The data are means ± SD (n = 8 per group). The Mann-Whitney test was performed for insulin and HOMA-IR analysis.

Amitriptyline markedly inhibits hepatic steatosis and inflammation induced by HP-HFD and LPS.

Our previous study has shown that HP-HFD and LPS cooperatively induced hepatic steatosis and inflammation in LDLR−/− mice (24). In this study, we assessed the effect of amitriptyline and GW4869 on hepatic steatosis and inflammation. H&E staining of liver tissues showed a large amount of fat droplets (hepatic steatosis) (Fig. 2) and enlarged cells with rarefied cytoplasm (hepatocellular ballooning) (Fig. 3A) in HP-HFD-fed mice with or without LPS treatment. Strikingly, amitriptyline treatment robustly attenuated the hepatic steatosis and hepatocellular ballooning in HP-HFD-fed mice with or without LPS treatment. In contrast, GW4869 did not change the hepatic steatosis and hepatocellular ballooning in HP-HFD-fed mice without LPS treatment and only modestly inhibited the hepatic steatosis and hepatocellular ballooning in HP-HFD-fed mice with LPS treatment. In addition, we determined the expression of ASMase and NSMase in hepatocytes and leukocytes in livers of HFD-fed mice using immunohistochemistry. Results showed that both hepatocytes and leukocytes in liver had similar level of ASMase and NSMase expression (Fig. 3B), which justifies our studies to compare the effects of ASMase and NSMase inhibition on hepatic steatosis and hepatocellular ballooning.

Fig. 2.

The effects of amitriptyline and GW4869 on steatosis in high-fat diet (HFD)-fed mice with or without LPS treatment. Representative images of liver sections stained with hematoxylin and eosin (magnification: ×100).

Fig. 3.

The effects of amitriptyline and GW4869 on steatosis and hepatocellular ballooning in high-fat diet (HFD)-fed mice with or without LPS treatment. A: representative images of liver sections stained with hematoxylin and eosin with high magnification (magnification: ×400). Insets: liver tissue images from HFD- or HFD plus LPS-treated mice were further enlarged (magnification: ×800) with the arrows indicating enlarged cells with rarefied cytoplasm (hepatocellular ballooning). B: immunostaining of acid sphingomyelinase (ASMase) and neutral sphingomyelinase (NSMase) in liver tissue (hepatocytes and leukocytes) of HFD-fed mice.

The above histological findings showing the effects of amitriptyline or GW4869 on hepatic steatosis were confirmed by the study using Oil Red O staining (Fig. 4). Results showed that while HP-HFD-fed mice with or without LPS treatment had marked fat accumulation, amitriptyline robustly inhibited the effect of HP-HFD or HP-HFD plus LPS on hepatic steatosis. In contrast, GW4869 had no effect on hepatic steatosis in HP-HFD-fed mice without LPS treatment but only moderately inhibited hepatic steatosis in HP-HFD-fed mice with LPS treatment.

Fig. 4.

The effects of amitriptyline (AMI) and GW4869 on steatosis in high-fat diet (HFD)-fed mice with or without LPS treatment. A: representative images of liver sections with Oil Red O staining (magnification: ×100). B: the positively stained areas of Oil Red O staining in the tissue sections were quantified and compared between all groups. The data presented are means ± SD (n = 8 per group).

To assess the effect of amitriptyline on hepatic inflammation, we performed immunohistochemical staining of F4/80, a marker for murine macrophages (2). Results (Fig. 5) showed that LPS significantly enhanced HP-HFD-induced F4/80 expression. Interestingly, amitriptyline strongly reduced F4/80 positively stained area (Fig. 5, A and B) or cells (Fig. 5C) in HP-HFD-fed mice with or without LPS treatment. In contrast, GW4869 failed to reduce F4/80 positively stained area or cells in HP-HFD-fed mice with or without LPS treatment.

Fig. 5.

The effects of amitriptyline (AMI) and GW4869 on hepatic inflammation in high-fat diet (HFD)-fed mice with or without LPS treatment. A: representative images of liver sections with F4/80 immunostaining (magnification: ×100). Insets: images were enlarged (magnification: ×400). B and C: the positively stained areas of F4/80 immunostaining (B) and F4/80 positively stained cells per high-power field (C) were quantified. The data presented are means ± SD (n = 8 per group).

Taken together, the above results demonstrated that amitriptyline robustly inhibited hepatic steatosis, hepatocellular ballooning, and hepatic inflammation while GW4869 only modestly inhibited hepatic steatosis but failed to inhibit hepatocellular ballooning and hepatic inflammation in HP-HFD-fed mice treated with LPS.

Both amitriptyline and GW4869 attenuate aortic atherosclerosis induced by HP-HFD and LPS.

Given that LDLR−/− mice are an animal model for atherosclerosis and we have shown that HP-HFD or HP-HFD plus LPS induces atherosclerosis in LDLR−/− mice (28), we also assessed the effect of amitriptyline and GW4869 on atherosclerosis induced by HFD and LPS. Results from Sudan IV staining on en face atherosclerotic lesions (Fig. 6, A and B) showed that while mice fed HP-HFD without LPS treatment had atherosclerotic lesions in aortic root and abdominal aorta, addition of LPS treatment to HP-HFD further increased atherosclerotic lesions. Interestingly, both amitriptyline and GW4869 had no effect on atherosclerosis in HP-HFD-fed mice without LPS treatment but attenuated atherosclerosis induced by HP-HFD-fed mice with LPS treatment. In addition, histological analysis of atherosclerotic lesions of aortic roots also showed that amitriptyline and GW4869 reduced intimal lesions of atherosclerosis induced by HP-HFD-fed mice with LPS treatment (Fig. 6, C and D).

Fig. 6.

The effects of amitriptyline (AMI) and GW4869 on atherosclerosis in high-palmitic acid-containing high-fat diet (HP-HFD)-fed mice with or without LPS treatment. LDLR−/− mice were fed HP-HFD for 20 wk and treated with or without LPS in the last 10 wk. Part of the mice were also treated with amitriptyline or GW4869 during the last 10 wk. A: after the feeding and treatments, aortas were dissected, opened longitudinally, and subjected to en face Sudan IV staining of atherosclerotic lesions. C: the histological analysis of atherosclerotic roots was also performed. B and D: the positively stained area of atherosclerotic lesions (B) and intimal lesion area (D) were quantified. The data presented are means ± SD (n = 8 per group).

Amitriptyline inhibits the upregulation of proinflammatory cytokines in macrophages.

Since macrophages play a pivotal role in both NASH and atherogenesis (5, 33) and ASMase is involved in the upregulation of proinflammatory cytokines in macrophages (19), we performed in vitro studies using murine RAW264.7 macrophages to elucidate the mechanisms by which amitriptyline inhibited NASH and atherosclerosis. Given that the HP-HFD used in this study contains a high content of PA, we tested our hypothesis that amitriptyline attenuates the synergy between PA and LPS on proinflammatory cytokine expression. Indeed, results showed that PA, a major SFA, and LPS had a synergistic stimulation on IL-6 expression, but amitriptyline inhibited IL-6 secretion (Fig. 7A) and mRNA expression (Fig. 7B) stimulated by LPS or LPS plus PA. To further demonstrate the effect of ASMase targeting on proinflammatory cytokine expression, we used bone marrow-derived macrophages prepared from wild-type and ASMase-deficient mice. Results showed that ASMase deficiency markedly reduced IL-6 secretion stimulated by LPS or LPS plus PA (Fig. 7C). In contrast to amitriptyline, GW4869 at the concentrations ranging from 1 to 50 μM had no effect on IL-6 secretion stimulated by LPS or LPS plus PA in RAW264.7 macrophages (Fig. 7D), which is consistent with our previous report (19). Although GW4869 had no effect on IL-6 secretion from macrophages, we have shown that GW4869 inhibited IL-6 secretion stimulated by LPS or LPS plus PA in vascular endothelial cells (29, 30). In addition, we determined the baseline expression of ASMase and NSMase mRNA in the untreated RAW264.7 macrophages. Results showed that RAW264.7 macrophages expressed similar levels of ASMase and NSMase mRNA (Fig. 7E), which justifies our studies to compare the effects of ASMase and NSMase inhibitors on IL-6 expression in RAW264.7 macrophages.

Fig. 7.

The effect of acid sphingomyelinase (ASMase) or neutral sphingomyelinase (NSMase) inhibition and ASMase deficiency on IL-6 expression in macrophages stimulated with LPS, palmitic acid (PA), or LPS plus PA. A and B: RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM of PA, or both in the absence or presence of 10 or 20 μM of AMI (A) or 20 μM of AMI (B) for 24 h. A and B: after the treatment, IL-6 in culture medium (A) and IL-6 mRNA (B) were quantified using ELISA and real-time PCR, respectively. C: bone marrow-derived macrophages were prepared using wild-type or ASMase knockout (KO) mice and treated with 1 ng/ml of LPS, 100 μM of PA, or both for 24 h. After the treatment, IL-6 in culture medium was quantified using ELISA. D: RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM of PA, or both in the absence or presence of different concentrations (0–50 μM) of GW4869 for 24 h. After the treatment, IL-6 in culture medium was quantified using ELISA. E: the expression of ASMase and NSMase mRNA in RAW264.7 macrophages. The expression of ASMase or NSMase mRNA was normalized to that of GAPDH. The presented data are means ± SD from 3 independent experiments.

Amitriptyline modulates sphingolipid metabolism and markedly reduces S1P content.

Since amitriptyline is an inhibitor of ASMase (15) and our previous studies have shown that sphingolipid metabolism plays a critical role in the synergistic upregulation of proinflammatory cytokines by LPS and PA (18, 19), we assessed the effect of amitriptyline on sphingolipid metabolism regulated by LPS and PA in RAW264.7 macrophages. Lipidomic analysis showed that PA or LPS plus PA significantly reduced SM, which is consistent with our previous report (19), but amitriptyline as an ASMase inhibitor increased SM as anticipated (Fig. 8A and Table 1). Given that amitriptyline inhibits ASMase-mediated SM hydrolysis, it is expected that CER production would be reduced. Surprisingly, results showed that amitriptyline increased CER production (Fig. 8B and Table 2). Lipidomic data showed that amitriptyline increased C16-CER and dhC16-CER, a sphingolipid produced during CER de novo synthesis (Fig. 8, C and D), suggesting that amitriptyline increased CER de novo synthesis. Furthermore, amitriptyline robustly diminished the production of sphingosine and S1P (Fig. 8, E and F), suggesting that amitriptyline suppressed CER hydrolysis by possible inhibiting ceramidase. This finding is in line with the previous reports that amitriptyline inhibited both ASMase and acid ceramidase (42, 46). To exclude the possibility that the inhibition of S1P production by amitriptyline is RAW264.7 cell specific, we used normal mouse bone marrow-derived macrophages. Results showed that amitriptyline also significantly inhibited S1P production stimulated by PA or LPS plus PA (Fig. 8G).

Fig. 8.

The effect of amitriptyline (AMI) on major sphingolipid metabolism in macrophages stimulated with LPS, palmitic acid (PA), or LPS plus PA. A–F: RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM of PA, or both in the absence or presence of 20 μM of AMI for 12 h; after the treatment, total sphingomyelin (A), total ceramide (B), C16-ceramide (CER; C), dihydro-C16-ceramide (dhC16-CER; D), sphingosine (E), and sphingosine 1 phosphate (S1P; F) were quantified by Lipidomics. G: the mouse bone marrow-derived macrophages were prepared from C57BL/6 mice and treated with 1 ng/ml of LPS, 100 μM of PA, or both in the absence or presence of 20 μM of AMI for 12 h. After the treatment, S1P was quantified. The presented data are means ± SD from 3 independent experiments.

Table 1.

The effect of amitriptyline on the cellular contents of major sphingomyelin species in RAW264.7 macrophages stimulated with LPS, PA, or LPS plus PA

Treatments	C16-SM	C18-SM	C20-SM	C22-SM	C24-SM	C24:1-SM	Total SM
Without AMI
Control	30.970 ± 3.465	2.122 ± 0.233	0.840 ± 0.087	4.379 ± 0.443	4.847 ± 0.534	8.692 ± 0.948	53.641 ± 4.470
LPS	27.689 ± 0.359	2.102 ± 0.269	0.709 ± 0.039	3.534 ± 0.021	4.244 ± 0.361	6.475 ± 0.169	46.409 ± 3.867
PA	25.054 ± 0.881	2.163 ± 0.045	0.790 ± 0.010	3.632 ± 0.149	2.711 ± 0.089	4.995 ± 0.167	40.667 ± 3.389
LPS + PA	28.360 ± 2.012	2.332 ± 0.112	0.815 ± 0.044	3.707 ± 0.275	2.684 ± 0.284	5.214 ± 0.341	44.533 ± 3.711
With AMI
Control	32.700 ± 0.045	2.649 ± 0.021	0.937 ± 0.001	4.771 ± 0.001	5.504 ± 0.002	9.505 ± 0.050	58.293 ± 4.858
LPS	34.397 ± 0.096	2.736 ± 0.047	0.917 ± 0.004	4.416 ± 0.003	5.094 ± 0.004	8.355 ± 0.084	58.319 ± 4.860
PA	33.960 ± 1.871	3.286 ± 0.134	1.056 ± 0.072	5.091 ± 0.374	3.860 ± 0.254	7.129 ± 0.319	56.272 ± 4.689
PA + LPS	39.299 ± 4.292	3.467 ± 0.400	1.133 ± 0.122	5.063 ± 0.539	4.011 ± 0.338	7.703 ± 0.844	63.045 ± 5.254

Data are presented as means ± SD of duplicates. The unit of the values of sphingomyelin (SM) is pmol/nmol phosphate. RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM palmitic acid (PA), or both LPS and PA for 12 h. After treatment, cells were harvested and subjected to the lipidomic analysis of SMs as described in materials and methods. AMI, amitriptyline.

Table 2.

The effect of amitriptyline on the cellular contents of major ceramide species in RAW264.7 macrophages stimulated with LPS, PA, or LPS plus PA

Treatments	C16-CER	C18-CER	C20-CER	C22-CER	C24-CER	C24:1-CER	dhC16-CER	Total CER
Without AMI
Control	0.558 ± 0.062	0.028 ± 0.001	0.030 ± 0.002	0.124 ± 0.002	0.685 ± 0.053	0.383 ± 0.036	0.024 ± 0.002	1.909 ± 0.243
LPS	0.653 ± 0.046	0.031 ± 0.002	0.028 ± 0.002	0.129 ± 0.006	0.644 ± 0.019	0.307 ± 0.012	0.035 ± 0.002	1.893 ± 0.245
PA	2.133 ± 0.028	0.182 ± 0.002	0.227 ± 0.016	0.783 ± 0.021	1.100 ± 0.022	0.544 ± 0.016	0.243 ± 0.001	5.114 ± 0.646
LPS + PA	2.705 ± 0.050	0.226 ± 0.008	0.313 ± 0.021	1.024 ± 0.042	1.329 ± 0.082	0.650 ± 0.040	0.161 ± 0.010	6.406 ± 0.815
With AMI
Control	1.207 ± 0.042	0.097 ± 0.002	0.079 ± 0.001	0.251 ± 0.007	0.708 ± 0.032	0.480 ± 0.002	0.100 ± 0.001	2.972 ± 0.375
LPS	1.234 ± 0.067	0.100 ± 0.001	0.088 ± 0.005	0.264 ± 0.004	0.771 ± 0.020	0.420 ± 0.010	0.111 ± 0.008	3.032 ± 0.385
PA	3.400 ± 0.015	0.338 ± 0.007	0.479 ± 0.019	1.396 ± 0.032	1.421 ± 0.054	0.593 ± 0.021	0.243 ± 0.002	7.807 ± 1.008
PA + LPS	5.500 ± 0.626	0.483 ± 0.048	0.608 ± 0.109	1.785 ± 0.311	1.822 ± 0.296	0.713 ± 0.114	0.700 ± 0.020	11.118 ± 1.584

Data are presented as means ± SD of duplicates. The unit of the values of ceramide (CER) is pmol/nmol phosphate. RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM palmitic acid (PA), or both LPS and PA for 12 h. After treatment, cells were harvested and subjected to the lipidomic analysis of CERs as described in materials and methods. AMI, amitriptyline.

To determine if the inhibition of CER hydrolysis by amitriptyline plays an essential role in the downregulation of proinflammatory cytokines, we inhibited acid ceramidase with pharmacological inhibitors D-NMAPPD and ceranib 1. Results showed that either D-NMAPPD (Fig. 9A) or ceranib 1 (Fig. 9B) effectively attenuated IL-6 secretion stimulated by LPS and PA in a concentration-dependent manner.

Fig. 9.

The effect of acid ceramidase inhibitors D-NMAPPD and ceranib-1 on major sphingolipid metabolism in RAW264.7 macrophages stimulated with LPS, palmitic acid (PA), or LPS plus PA. RAW264.7 macrophages were treated with 1 ng/ml of LPS, 100 μM of PA or both in the absence or presence of different doses of D-NMAPPD and ceranib-1 for 24 h. After the treatment, IL-6 in culture medium was quantified using ELISA. The presented data are means ± SD from 3 independent experiments.

DISCUSSION

HP-HFD-fed LDLR−/− mice are a commonly used animal model for T2DM, hyperlipidemia and atherosclerosis (3, 56). We have shown that HP-HFD and a low dose of LPS, which induces systemic inflammation, cooperatively induce T2DM-associated atherosclerosis in LDLR−/− mice (28). Since atherosclerosis is associated with NAFLD in patients with T2DM and MetS (27), we further used this animal model to determine if HP-HFD and LPS induced NAFLD. We found that HP-HFD induced NASH and the addition of LPS to HP-HFD further increased hepatic inflammation in LDLR−/− mice (24). These findings indicate that HP-HFD-fed LDLR−/− mice treated with a low dose of LPS are a good animal model to study the mechanisms involved in the linkage between NASH and atherosclerosis.

In the current study, we determine if ASMase is involved in both NASH and atherosclerosis induced by HP-HFD and LPS in LDLR−/− mice. Using amitriptyline, an inhibitor of ASMase, we showed that ASMase inhibition robustly inhibited not only NASH, but also atherosclerosis in LDLR−/− mice (Figs. 2–6). Since our previous studies have shown that ASMase plays a key role in PA-amplified inflammatory signaling triggered by LPS in macrophages (19) and macrophage-mediated inflammation is well known to involved in both NASH and atherosclerosis (5), these findings suggest that ASMase-related sphingolipid metabolism in macrophages may be a valid target for not only NASH but also atherosclerosis associated with T2DM.

To determine if ASMase is specifically involved in the pathogenesis of NASH, we also targeted NSMase using the NSMase inhibitor GW4869 as control. Interestingly, results showed that although inhibition of NSMase attenuated atherosclerosis, it had no significant inhibition on hepatic inflammation induced by HP-HFD or HP-HFD plus LPS (Figs. 2–6). These findings are consistent with our previous reports that NSMase was involved in inflammatory response of vascular endothelial cells but not macrophages (29, 30). It is known that while macrophages play a pivotal role in both NASH and atherosclerosis, endothelial cells are mainly involved in atherosclerosis (11). Thus these findings suggest that ASMase-related sphingolipid metabolism in macrophages specifically links NASH and atherosclerosis induced by HP-HFD and LPS in LDLR−/− mice.

Amitriptyline is a tricyclic antidepressant that has been used for depression and pain (34). Although it is not completely elucidated, the mechanism underlying the therapeutic effect of amitriptyline is to block transporters of the neurotransmitters norepinephrine and serotonin, leading to an increase in the levels of norepinephrine and serotonin at the nerve ending. Interestingly, it has been reported that amitriptyline also exerts its antidepressive effect via inhibition of ASMase-CER system (15). Thus it is likely that amitriptyline alleviates depression through multiple biological pathways.

With respect to metabolic parameters, it was reported that, in addition to its inhibition of SM hydrolysis, amitriptyline also reduced triglycerides and cholesterol in hepatic tissues of HFD-fed mice (12). Consistent with this report, our metabolic data also showed that amitriptyline reduced HP-HFD-increased plasma triglycerides and cholesterol. Furthermore, our metabolic data showed that amitriptyline reduced body weight, glucose, insulin and insulin resistance in animal model. Our results are consistent with the findings that amitriptyline improved glucose tolerance and insulin sensitivity reported by Boini et al. (6). However, our findings that LPS injection reduced insulin and insulin resistance are different from the previous report that the subcutaneous infusion of LPS increased insulin and insulin resistance by Cani et al. (8). It is possible that the different results were caused by different methods for LPS delivery as we injected LPS intraperitoneally once a week for 10 wk while Cani et al. delivered LPS by subcutaneous infusion for 4 wk. Collectively, all these metabolic data have demonstrated the positive effects of amitriptyline on hyperlipidemia, hyperglycemia, and insulin resistance in mice with T2DM. Interestingly, a clinical study showed that treatment with amitriptyline in patients with depression was associated with increased insulin sensitivity and improved parameters of cholesterol metabolism (51), which is consistent with the results from our current metabolic study in the animal model.

Our current study showed that amitriptyline strongly inhibited NASH induced by HP-HFD or HP-HFD plus LPS. Although it is not fully understood how amitriptyline inhibited NASH, our studies suggested two potential mechanisms that are likely to be involved. First, our study is consistent with the previous study (6) that showed the inhibition of insulin resistance and subsequent hyperinsulinemia by amitriptyline. It is well known that insulin resistance-related hyperinsulinemia increases triglycerides synthesis and accumulation in liver and impairs hepatic lipid metabolism (39). Second, our current study showed that amitriptyline inhibited the secretion and expression of proinflammatory cytokine IL-6 in macrophages. It is well known that proinflammatory mediators such as IL-6, TNF-α, IL-1α, and IL-1β released from immune cells play a key role in insulin resistance and dysregulation of lipid metabolism in liver (47). A recent study showed that amitriptyline mitigated sepsis-induced TNF-α expression in macrophages and blunted macrophage M1 polarization (57), which is in line with our findings that amitriptyline inhibited hepatic inflammation induced by HP-HFD and LPS.

To further understand how amitriptyline inhibited NASH and atherosclerosis through modulation of sphingolipid metabolism, we examined the modulation by amitriptyline of major sphingolipid metabolic pathways including SM hydrolysis, CER de novo synthesis, and production of sphingosine and S1P. Surprisingly, we found that amitriptyline regulates multiple metabolic pathways of sphingolipids (Fig. 10). First, amitriptyline increased cellular SM content as it inhibited ASMase-mediated SM hydrolysis as expected. Second, treatment with amitriptyline led to an increase in CER de novo synthesis as indicated by the significant increase of dhC16-CER, a CER species generated during CER de novo synthesis. Third, amitriptyline potently diminished the production of sphingosine and S1P via possible inhibition of CER hydrolysis. Taken together, treatment with amitriptyline resulted in increases in SM and CER and decreases in sphingosine and S1P (Fig. 10).

Fig. 10.

The schematic diagram showing the effect of amitriptyline treatment on sphingolipid metabolic pathways. Amitriptyline treatment of macrophages leads to increased sphingomyelin, increased ceramide via ceramide de novo synthesis, and decreased sphingosine and sphingosine 1 phosphate (S1P).

We have shown previously that S1P is critically involved in the inflammatory response of macrophages to LPS and PA (18). Although S1P alone did not stimulate proinflammatory cytokine expression, it augmented the stimulatory effect of LPS, PA, or LPS plus PA on cytokine expression (18). Our results also showed that the S1P receptor inhibitors VPC23019 and JTE013 inhibited proinflammatory cytokine expression stimulated by LPS and PA and knockdown of S1P receptor 1 or receptor 2 attenuated proinflammatory cytokine expression stimulated by LPS and PA. Furthermore, the sphingosine kinase inhibitors DMS and SK1-I inhibited proinflammatory cytokine expression stimulated by LPS and PA (18). All these findings clearly revealed a role of S1P in the stimulation of inflammatory response of macrophages by LPS and PA.

To determine if CER hydrolysis is involved in the inhibition of IL-6 expression by amitriptyline, we inhibited acid ceramidase with the inhibitor D-NMAPPD or ceranib 1. Results showed that ceramidase inhibitors D-NMAPPD and ceranib 1 effectively inhibited IL-6 expression in macrophages stimulated by LPS or LPS plus PA, suggesting that amitriptyline inhibited IL-6 expression in macrophages by inhibiting acid ceramidase and subsequent reducing S1P production.

It is noteworthy that amitriptyline and imipramine, both inhibitors of ASMase and tricyclic antidepressants (22), have different effects on sphingolipid metabolism in macrophages. Our previous study has shown that while both imipramine and amitriptyline inhibit SM hydrolysis and reduce S1P production, imipramine does not increase CER by stimulating CER de novo synthesis like amitriptyline (19). Since both amitriptyline and imipramine inhibit proinflammatory cytokine expression and amitriptyline increases but imipramine decreases CER production, it is indicated that the cellular content of CER is not crucial for proinflammatory gene expression in macrophages. It is likely, however, that S1P plays an essential role in the upregulation of proinflammatory gene expression stimulated by LPS and PA.

Taken together, we have demonstrated that amitriptyline is an effective inhibitor on both NASH and atherosclerosis in LDLR−/− mice with T2DM. We have also demonstrated that amitriptyline inhibits CER hydrolysis-dependent S1P production that is likely the mechanism involved in the inhibition of proinflammatory cytokine expression in response to LPS plus PA.

GRANTS

This work was supported by Merit Review Grant BX000855 from the Biomedical Laboratory Research and Development Program of the Department of Veterans Affairs and National Institute of Dental and Craniofacial Research Grant DE-027070 (to Y. Huang). The work on sphingolipid analysis was supported in part by the Lipidomics Shared Resource, Hollings Cancer Center, Medical University of South Carolina (National Cancer Institute Grant P30-CA-138313), the Lipidomics Core in the SC Lipidomics and Pathobiology Centers of Biomedical Research Excellence Grant P20-RR-017677, and the National Center for Research Resources and the Office of the Director of the National Institutes of Health through Grant C-06-RR-018823.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Y.H. conceived and designed research; Y.L. and Z.L. performed experiments; Y.L., Z.L., T.J.L. and Y.H. analyzed data; W.S., Z.W., M.F.L.-V. and Y.H. interpreted results of experiments; Y.L., Z.L., and Y.H. prepared figures; Y.H. drafted manuscript; W.S., Z.W., M.F.L.-V. and Y.H. edited and revised manuscript; Y.L., Z.L., W.S., Z.W., M.F.L.-V., T.J.L. and Y.H. approved final version of manuscript.