Simultaneous Inhibition of Peripheral CB1R and iNOS Mitigates Obesity-Related Dyslipidemia Through Distinct Mechanisms

Cardiovascular diseases are the major cause of morbidity and mortality in patients with type 2 diabetes (T2D). Increased plasma triglycerides (TGs) and decreased HDL levels observed in the context of T2D are among the major factors contributing to increased cardiovascular risk. Increased VLDL production is a major contributor to elevated TGs in subjects with insulin resistance/T2D (1), with increased de novo lipogenesis, increased microsomal triglyceride transfer protein (MTP) activity, and/or a reduction in apolipoprotein (apo)B degradation in hepatocytes (2). Diabetic dyslipidemia is strongly associated with the presence of nonalcoholic fatty liver disease (3).

The endocannabinoid system is composed of two receptors (cannabinoid-1 receptor [CB1R] and cannabinoid-2 receptor [CB2R]), endogenous ligands called endocannabinoids, and the enzymatic machinery of their biosynthesis/degradation (4). Accumulated evidence implicates the obesity-related overactivity of hepatic CB1R in the development of nonalcoholic fatty liver disease (5). Global CB1R blockade results in improved cardiometabolic parameters in obese and overweight patients (6,7). However, the therapeutic development of this class of compounds has been halted because of adverse neuropsychiatric side effects (5). In the meantime, numerous studies have implicated peripheral CB1R in the development of obesity and its metabolic sequelae (5). As such, second-generation peripherally restricted CB1R antagonists, which are devoid of central nervous system–mediated side effects, have shown promising therapeutic efficacy as antiobesity and/or antidiabetes drugs (5). A further approach to increase efficacy has involved the development of third-generation CB1R antagonists, which are non-brain-penetrant and have a secondary target, such as inducible nitric oxide synthase (iNOS). Nitric oxide (NO) is produced by NO synthases (NOS), from l-arginine, O₂, and NADPH through the l-arginine–NO pathway (8). The inducible isoform of NOS (iNOS) is expressed by many cells, including macrophages and hepatocytes (9) where its induction is associated with diet-induced steatohepatitis (10).

Here we report that (S)-MRI-1867, a peripherally restricted hybrid inhibitor of CB1R/iNOS with strong antifibrotic properties (11,12), improves hepatic lipid and lipoprotein metabolism in DIO mice by reducing liver lipoprotein production and stimulating its plasma clearance. Taking advantage of the CB1R-inactive stereoisomer (R)-MRI-1867, which retains the iNOS inhibitory activity, and JD-5037, a peripherally restricted CB1R antagonist, we further demonstrate that underlying mechanisms involve a CB1R-dependent reduction in VLDL assembly and an iNOS-dependent reduction of proprotein convertase subtilisin kexin 9 (PCSK9) activity associated with a restoration in hepatic LDLR levels.

Official French regulations for the use and care of laboratory animals were followed throughout the experiments. The experimental protocol was approved by the local ethics committee (CE2A, Dijon, France) for animal experimentation (APAFIS no. 16799-2018021914408858 v3).

Six-week-old male wild-type C57BL/6JRj mice (JANVIER LABS, Le Genest St Isle, France) were maintained under a 12:12-h light-dark cycle and fed ad libitum a high-fat diet (cat. no. E15742-34, 60% of calories from fat [lard], 20% from protein, and 20% from carbohydrates; ssniff, Soest, Germany) for 18 weeks. For the last 14 days, (S)-MRI-1867 (10 mg ⋅ kg⁻¹), (R)-MRI-1867 (10 mg ⋅ kg⁻¹), JD-5037 (3 mg ⋅ kg⁻¹), or vehicle was administered daily by oral gavage. Similar treatment was performed in aged-matched chow-fed mice. Body weight and individual food intake were monitored daily. After euthanasia, tissues were collected, weighed, and snap frozen in liquid nitrogen. Trunk blood was collected to determine endocrine and biochemical parameters.

(S)-MRI-1867 and (R)-MRI-1867 were synthesized as previously described (11,12), and JD-5037 was purchased from MCE (cat. no. HY-18697; Sollentuna, Sweden). Both compounds were administered by oral gavage once daily. Vehicle was used at a ratio of 1:1:18 DMSO:Tween 80:saline. 1400W dihydrochloride (1400W) (Bio-Techne, Lille, France) was purchased from Tocris Bioscience.

Glucose tolerance tests (GTT) and insulin sensitivity tests were performed as previously described (13).

iNOS activity was assessed as the rate of conversion of [¹⁴C(U)]l-arginine to [¹⁴C(U)]l-citrulline by use of an NOS Activity Assay Kit (Cayman Chemical, Ann Arbor, MI) as per the supplier’s instructions. This assay was used to measure the effects of the test substances on iNOS activity, with recombinant mouse iNOS (Cayman Chemical).

Intrahepatic TG content was determined as previously described (14), whereas glycogen content was determined based on the enzymatic reaction (15).

TGs, total cholesterol, HDL, LDL, ALT, and AST were assessed using the automated Vista 1500 and its dedicated reagents (Siemens, Erlangen, Germany). Leptin, insulin, and glucagon concentration were determined using the Luminex-based Bio-Plex Pro mouse assay (Bio-Rad, Marnes-La-Coquette, France). Adiponectin was measured using the Mouse Adiponectin ELISA Kit (Assaypro, St. Charles, MO). ApoB100 measurement was carried out using the Mouse ApoB100 ELISA kit (ELISA Genie, Dublin, Ireland). PCSK9 protein level was assessed using the Mouse Proprotein Convertase 9/PCSK9 Quantikine ELISA (R&D Systems, Minneapolis, MN). FGF21 circulating levels were assessed using the Mouse/Rat FGF-21 Quantikine ELISA Kit (R&D Systems).

Mice received an intraperitoneal injection of poloxamer-407 (1 g ⋅ kg⁻¹) (Sigma-Aldrich) as previously described (16). Blood was collected at 0, 0.5, 1, 2, 4, 8, and 24 h after the injection. Total cholesterol and TGs were assessed at all time points, whereas serum lipoprotein separation was performed on the 24-h time point collection using fast-performance liquid chromatography as described by Jourdan et al. (17).

Primary hepatocytes were isolated from DIO mice and cultured as previously described (18). Treatments were carried out as described in the figure legends. Human embryonic kidney (HEK)293 cells with a stable transfection of the murine CB1R (HEK293CB1Rm) with N-terminus hemagglutinin tag (cat. no. EIU005; Kerafast, Boston, MA) were cultured with DMEM, High Glucose, GlutaMAX (Thermo Fisher Scientific, Courtaboeuf, France), supplemented with 10% FBS at 37°C and 5% CO₂. HEK293CB1Rm cells were seeded in six-well plates (200,000 cells/well) for 48 h before switching culture medium to Opti-MEM (Thermo Fisher Scientific) for an additional 12 h before treatments. Cells were pretreated for 30 min with vehicle, JD-5037 (0.05 μmol/L), or (S)- or (R)-MRI-1867 (0.1 μmol/L) before an additional 5 min with vehicle or arachidonyl-2′-chloroethylamide (1 μmol/L) in the absence or presence of one of the above antagonists. Cells were then placed on ice, washed with ice-cold 1× PBS, and lysed in RIPA buffer before protein extraction and immunoblotting as described below.

Tissue or cell lysis, protein migration, transfer, immunoblotting, blot analysis, and quantification were performed as previously described (19). Primary antibodies are listed in Supplementary Table 1.

Total mRNA extraction, reverse transcription, and real-time PCR were performed and analyzed as previously described (20). Primer sequences used for amplification are available upon request.

Values are expressed as means ± SEM. Data were analyzed either by Student test or by ANOVA followed by the Tukey-Kramer post hoc test when multiple comparisons were required. Time-dependent variables were analyzed, and results in multiple groups were compared by ANOVA followed by Bonferroni test (version 7 for Windows; GraphPad Prism, San Diego, CA). Significance was set at P < 0.05.

All other data sets are available upon reasonable request.

Compared with mice given vehicle treatment, DIO mice treated with (S)-MRI-1867 displayed a transient decrease in food intake that normalized by day 7 and a progressive weight loss reaching 20% of initial body weight by the end of the 14-day treatment (−20.21% ± 1.44) as illustrated in Fig. 1A. The reduced body weight was mainly due to a reduction of adiposity (Supplementary Fig. 1A). Interestingly, (S)-MRI-1867 mice had no significant impact on food intake or body weight in lean mice and a slightly reduced body weight (−2.92% ± 0.86) (Fig. 1A).

(S)-MRI-1867 treatment also lowered resting blood glucose and decreased circulating levels of insulin and glucagon without affecting GLP-1 levels in DIO mice (Supplementary Fig. 1B). Consistently, glycemic control in DIO but not in lean mice was improved, as evidenced by the results of GTT and insulin sensitivity tests (Fig. 1B). Moreover, leptin levels were lowered by (S)-MRI-1867 treatment, while adiponectin concentration remained unchanged (Supplementary Fig. 1C).

The relative liver weight of mice treated with (S)-MRI-1867 was significantly higher than that of vehicle-treated DIO mice, with lower hepatic TGs and higher hepatic glycogen content, while total cholesterol content did not change (Fig. 1C). This was in agreement with earlier findings (21) indicating that HFD-induced increase in hepatic glucose production is due to a CB1R-mediated increase in glycogenolysis, resulting in reduced hepatic glycogen content. These differences were associated with lower levels of the hepatic injury markers ALT and AST (Supplementary Fig. 1D), suggesting that (S)-MRI-1867 reverses obesity-induced hepatocyte injury. Interestingly, (S)-MRI-1867 treatment did not influence these markers in lean mice, suggesting an absence of hepatotoxic side effects from this compound in healthy animals (Supplementary Fig. 1D). Intriguingly, we observed higher circulating levels of the ketone body β-hydroxybutyrate in (S)-MRI-1867–treated DIO mice, suggesting an enhancement of hepatic fatty acid β-oxidation activity (Fig. 1D). Furthermore, compared with vehicle-treated controls, (S)-MRI-1867–treated mice had higher levels of plasma free fatty acids and similar plasma TGs and lower total cholesterol concentration (Fig. 1E)—indicators of altered plasma lipoprotein profile. This was further associated with a marked decrease in the plasma levels of apoB100, the primary apolipoprotein present in lipoproteins, in (S)-MRI-1867–treated mice (Fig. 1F). In parallel, LDL was strongly reduced, while HDL was only slightly modified, leading to a much higher HDL–to–LDL cholesterol ratio in the (S)-MRI-1867–treated group compared with the vehicle-treated group (Fig. 1F).

We first tested whether (S)-MRI-1867 altered the expression of the genes encoding its pharmacological targets. The hepatic expression of the gene coding CB1R (Cnr1) was significantly reduced, similar to the effect of rimonabant (22), while the expression of Nos2 (iNOS) was unaffected by (S)-MRI-1867 (Fig. 2A). Interestingly, (S)-MRI-1867 compared with vehicle treatment strongly inhibited the expression of genes involved in the de novo lipogenesis pathways such as Srebf1, Acc1, Fas, and Scd1, predicting a marked decrease of fatty acid synthesis in the liver (Supplementary Fig. 2A). In the meantime, no change was observed for the two key regulatory genes of fatty acid oxidation, Cpt1a and Acox1 (Supplementary Fig. 2B). In addition, the expression of the genes encoding enzymes involved in glycolysis and gluconeogenesis was similarly unaffected by (S)-MRI-1867 (Supplementary Fig. 2C). Furthermore, (S)-MRI-1867 led to a decrease in the expression of the Tsukushi (Tsk) gene (Supplementary Fig. 2D) encoding the Tsukushi hepatokine, a deficiency of which has recently been associated with an improvement of nonalcoholic steatohepatitis–related features such as steatosis and circulating total cholesterol levels (23). A significant increase in the Ser⁵²⁶ phosphorylation of the transcription factor FoxO1 was observed in (S)-MRI-1867–treated livers, which should inhibit its activity (Fig. 2B). This may be related to an observed (S)-MRI-1867–induced decrease in both the gene and protein expression of the microsomal triglyceride transfer protein (MTP [Mttp]) in the liver (Fig. 2C), as FoxO1 is known to stimulate hepatic Mttp expression and activity (24). FoxO1 activity is under the control of insulin, and its inhibitory phosphorylation is the result of the activation of the phosphatidylinositol 3-kinase–AKT pathway. Interestingly, we found a strong activation of AKT in the liver of (S)-MRI-1867–treated mice, as illustrated by the marked increase in both Ser⁴⁷³ and Thr³⁰⁸ phosphorylation of AKT (Supplementary Fig. 2E). All together, these data further suggest that (S)-MRI-1867 reduces VLDL output at least in part through reducing the assembly and subsequent secretion of these lipoproteins. Because VLDL turnover is extremely rapid, we had to inhibit the LPL activity using a solution of poloxamer-407 in order to monitor VLDL output. After a single injection of poloxamer-407, TG accumulation in blood was much higher in vehicle-treated DIO mice than in (S)-MRI-1867–treated animals (Fig. 2D). Accordingly, 24 h post–injection of poloxamer-407, (S)-MRI-1867–treated mice displayed lower blood levels of TGs and total cholesterol compared with the vehicle-treated controls (Fig. 2D). This could be attributed to a much smaller VLDL fraction in (S)-MRI-1867–treated mice, as documented by the TG (Fig. 2E) and cholesterol (Fig. 2F) profile obtained after lipoprotein separation using fast-protein liquid chromatography. Interestingly, a similar treatment in lean mice showed no differences in circulating TGs and total cholesterol levels, as well as no changes in the HDL–to–LDL cholesterol ratio (Supplementary Fig. 3A). Similarly, TG accumulation in blood was unchanged after (S)-MRI-1867 treatment in lean mice (Supplementary Fig. 3B). Collectively, these data suggest that (S)-MRI-1867 can improve hepatic lipoprotein metabolism by reducing VLDL output in DIO mice.

The lower levels of circulating lipoproteins observed in (S)-MRI-1867–treated mice can also result from their higher uptake, mediated by the hepatic LDL receptor (LDLR). In line with this, (S)-MRI-1867 treatment of obese mice increased the gene and protein expression of LDLR (Ldlr) compared with vehicle treatment (Fig. 3A). Interestingly, the (S)-MRI-1867–induced increase in LDLR was paralleled by a marked reduction of circulating PCSK9 levels (Fig. 3B). PCSK9 binds to the extracellular domain of the LDLR and targets it for lysosomal degradation (25,26), thus lowering its cellular levels. PCSK9 expression is regulated through multiple pathways. We observed a profound reduction in the active form of SREBP-2 in livers of (S)-MRI-1867–treated mice compared with vehicle-treated mice (Fig. 3C). Additionally, (S)-MRI-1867 treatment resulted in increased gene expression of the chaperone heat-shock protein 90 kDa β member 1 (Gpr94) (Fig. 3D), which prevents PCSK9 from sequestering LDLR, thus allowing more LDLR to reach the cell surface (27). Interestingly, the gene expressions of histone deacetylase Sirtuin 6 (Sirt6) and the transcription factor FOXO3 (Foxo3) were both increased in livers of (S)-MRI-1867–treated mice compared with the vehicle-treated group (Fig. 3E). FOXO3 recruits SIRT6 to the PCSK9 promoter and represses PCSK9 transcription (28). Moreover, (S)-MRI-1867 led to a profound repression of the transcription factor E2F1 (E2f1), recently implicated in the regulation of PCSK9 transcription (29) (Fig. 3F).

To determine the relative role of CB1R blockade or iNOS inhibition in the observed effects of (S)-MRI-1867, we isolated hepatocytes from DIO mice and incubated them for 24 h with vehicle, the peripherally restricted CB1R inverse agonist JD-5037 (recently renamed as CRB-4001) (100 nmol/L), the specific iNOS inhibitor 1400W (10 μmol/L), or (S)-MRI-1867 (100 nmol/L). As expected from their target profiles, treatment with either JD-5037 or (S)-MRI-1867 but not 1400W resulted in reduced Cnr1 expression, whereas Nos2 expression was affected by 1400W and (S)-MRI-1867 but not by JD-5037 treatment (Supplementary Fig. 4A). Furthermore, treatment with either JD-5037 or (S)-MRI-1867 led to a marked decrease in intracellular TG (Supplementary Fig. 4B) and in TG secretion into the medium (Supplementary Fig. 4C), whereas 1400W failed to elicit similar changes but markedly increased intracellular cholesterol content (Supplementary Fig. 4D). All three compounds reduced Apob expression (Supplementary Fig. 4E), but only CB1R blockade led to a reduced Mttp expression (Supplementary Fig. 4F), suggesting a critical role for CB1R in the regulation of lipoproteins and TG output from hepatocytes. Interestingly, 1400W or (S)-MRI-1867 led to a strong repression of Pcsk9 gene expression, whereas JD-5037 treatment resulted in an opposite effect (Supplementary Fig. 4G), suggesting that the reduction in Pcsk9 expression elicited by (S)-MRI-1867 might be due to iNOS inhibition. Together, these data suggest that (S)-MRI-1867 effects on VLDL assembly/secretion are likely mediated by CB1R antagonism, while the improved LDL clearance may be due to iNOS inhibition.

To investigate which (S)-MRI-1867 target was responsible for the reduced VLDL assembly, we treated DIO mice for 14 days with either the single-target CB1R blocker JD-5037 or (R)-MRI-1867, the stereoisomer of (S)-MRI-1867 (chemical structure in Supplementary Fig. 5A). Because of its spatial conformation, (R)-MRI-1867 is no longer able to bind to and block CB1R effectively, but it retains the same iNOS inhibitory activity as the (S)-MRI-1867 isomer. Indeed, unlike JD-5037 and (S)-MRI-1867, the (R)-MRI-1867 compound was unable to reverse the arachidonyl-2′-chloroethylamide-mediated phosphorylation of p42/44 MAPK in HEK cells overexpressing mouse CB1R (Supplementary Fig. 5B). It did, however, exert inhibition of iNOS activity similar to that of (S)-MRI-1867, while JD-5037 had no impact on iNOS activity (Supplementary Fig. 5C). Besides, JD-5037–treated DIO mice displayed a 25% loss in body weight (−24.91% ± 1.63), similarly to what was observed with (S)-MRI-1867, likely due to the decrease in adiposity. CB1R blockade was also associated with an improvement of glucose homeostasis (Supplementary Fig. 6A). In contrast, neither body weight nor adiposity index was affected by (R)-MRI-1867. However, iNOS inhibition led to an improvement in glucose homeostasis (Supplementary Fig. 6B). JD-5037 reduced hepatic TG content, in agreement with previous reports (13,30), but it failed to impact hepatic cholesterol levels (Fig. 4A). Regarding plasma lipids, JD-5037 led to a reduction in total circulating cholesterol that was associated with a significant increase in HDL-to-LDL ratio without affecting triglyceridemia or circulating apoB100 levels (Fig. 4A). Conversely, iNOS inhibition had no impact on the steatosis (Fig. 4B), while it did, however, result in a small increase in hepatic cholesterol levels compared with vehicle control (Fig. 4B). Unlike CB1R blockade, iNOS deletion had no impact on total cholesterol but did increase the HDL-to-LDL ratio and resulted in a marked reduction in circulating apoB100 (Fig. 4B). Interestingly, both JD-5037 treatment and iNOS inhibition were associated with a decrease in Tsk expression (Supplementary Fig. 6C and D). As for VLDL assembly, CB1R blockade reduced MTP levels through an increase in Ser²⁵⁶ phosphorylation of FoxO1 (Fig. 4C), whereas iNOS inhibition had no impact on any of these parameters (Fig. 4D). Accordingly, we measured the hepatic lipoprotein production and profile in JD-5037–treated and vehicle-treated mice administered poloxamer-407. Similar to what we observed with (S)-MRI-1867, single-target CB1R blockade led to a much lower TG accumulation in the bloodstream and a much smaller VLDL fraction compared with vehicle-treated controls (Fig. 4E). These data confirm the prominent role of CB1R blockade in the reduction of VLDL assembly and secretion by the liver.

The stereoisomer (R)-MRI-1867 was also used to determine which target of (S)-MRI-1867 was responsible for the improved LDL clearance. CB1R blockade alone with JD-5037 did not influence the hepatic LDLR expression and resulted in a small but significant increase in PCSK9 plasma levels (Fig. 5A). Interestingly, iNOS inhibition by (R)-MRI-1867 was associated with an increase in LDLR expression along with a clear reduction of PCSK9 circulating levels (Fig. 5B). In agreement with these observations, CB1R blockade was unable to reduce SREBP-2 protein levels or Sirt6, Foxo3, or Gpr94 gene expression (Fig. 5C), although it downregulated E2f1 mRNA levels (Fig. 5C). On the other hand, iNOS inhibition led to a lower protein expression for SREBP-2; an increase in Sirt6, Foxo3, and Gpr94; and a decrease in E2f1 mRNA levels (Fig. 5D).

In further characterization of the mechanisms leading to PCSK9 reduction after iNOS inhibition, we hypothesized that mTOR complex 1 (mTORC1) signaling was involved. Indeed, mTORC1 inhibition with rapamycin led to an increase in PCSK9 and reduction in hepatic LDLR levels (31). We found that iNOS inhibition resulted in the activation of mTORC1, as evidenced by the increase in Ser²⁴⁴⁸ phosphorylation of mTOR and the increase in RAPTOR protein expression (Fig. 6A). On the other hand, iNOS inhibition had no effect on mTORC2 activation, as both RICTOR and the Ser²⁴⁸¹ phosphorylation of mTOR remained unchanged (Fig. 6B). The link between iNOS inhibition and mTORC1 activation is likely an increased bioavailability in l-arginine, iNOS substrate. Indeed, this amino acid can be sensed by mTORC1 and alleviate its inhibition (32). We then incubated primary hepatocytes with 0, 10, 25, 50, 75, or 100 μmol/L l-arginine for 24 h. Interestingly, high concentration of l-arginine was able to reduce Pcsk9 gene expression after 24 h (Fig. 6C). Strikingly, as little as 25 μmol/L l-arginine was enough to reduce PCSK9 secretion by 50%, with a maximum of 75% inhibition with 75 μmol/L l-arginine (Fig. 6D).

Dyslipidemia in patients with T2D significantly increases the risk of cardiovascular diseases, and T2D prevalence has increased by 31% worldwide from 2005 to 2015, rendering this disease and its metabolic sequelae an important public health issue (33). Current therapeutic approaches are mainly relying on lifestyle intervention and hypolipidemic drugs such as statins. While these interventions lead to a lowering of LDL, they do not reduce the residual cardiovascular risks (34). This unmet medical need highlights the urgency for new therapeutic options.

In the current study, we show that dual inhibition of hepatic CB1R and iNOS signaling using the hybrid inhibitor (S)-MRI-1867 improves the abnormalities of lipoprotein metabolism characteristic of diabetic dyslipidemia in DIO mice. (S)-MRI-1867 treatment improved the lipoprotein profile by 1) reducing VLDL synthesis and assembly, leading to reduced TG output from the liver via CB1R antagonism, and 2) promoting LDL clearance by increasing LDLR on hepatocytes through an inhibition of PCSK9 action via iNOS inhibition. Furthermore, through the use of selective, single-target inhibitors of peripheral CB1R and iNOS, we were able to attribute the reduced VLDL synthesis and secretion to CB1R blockade, while iNOS inhibition was responsible for the reduction in PCSK9 and increase in LDLR.

VLDL assembly depends on apoB100 and TG availability through MTP action, which is mainly under insulin control in normal conditions (35). Here, we found that (S)-MRI-1867 reduces MTP protein levels, apoB100, and TG availability in the liver. It is known that FoxO1 stimulates hepatic Mttp expression, and this effect is blocked by insulin under normal conditions (36,37). Interestingly, Kamagate et al. (24) nicely demonstrated that mice expressing a constitutively active FoxO1 transgene displayed an increased FoxO1 activity associated with higher MTP expression and increased VLDL production, whereas an RNA interference–mediated silencing of hepatic FoxO1 led to reduced MTP expression and VLDL production in adult mice. In addition, (S)-MRI-1867 improved both global insulin and hepatic sensitivity through the activation of the phosphatidylinositol 3-kinase–AKT pathway, which is directly involved in the control of FoxO1 (38). We also found an increase in FoxO1 phosphorylation induced by (S)-MRI-1867 leading to its translocation from the nucleus to cytoplasm, thus losing its transcriptional activity (39). By restoring proper insulin signaling in the liver, (S)-MRI-1867 treatment led to the activation of AKT, which then repressed FoxO1 activity, with a reduced MTP activity and VLDL assembly as a consequence. In addition, we found that (S)-MRI-1867 decreased circulating levels of apoB100 in DIO mice, which was likely due to an inhibition of apoB synthesis resulting from an increase in endoplasmic reticulum autophagy (40). This was further illustrated by the lower TG and VLDL secretion observed in (S)-MRI-1867–treated mice after poloxamer-407 injection. Strikingly, we observed a similar improvement of the lipoprotein profile with single CB1R blockade by JD-5037, while iNOS inhibition had no impact on this parameter, thus illustrating the CB1R-dependent regulation of these pathways.

Another important observation was the increase in LDLR expression and the decrease in serum PCSK9 levels after (S)-MRI-1867 treatment. Under normal circumstances, LDLR binds LDL and removes it from the circulation by endocytosis. While LDL is degraded by lysosomes, LDLR is recycled to the cell surface. However, when PCSK9 binds to the LDLR, the latter is no longer recycled to the cell surface and gets degraded in the lysosomes (41,42), leading to a lower amount of LDLR and an increase in circulating LDL. PCSK9 is usually upregulated by cholesterol depletion or by inhibition of intracellular cholesterol synthesis such as is achieved by statins, thus limiting the efficacy of statin treatment. This can be attributed to the presence of a sterol regulatory element negatively regulated by SREBP-2 (43). Here, we found that (S)-MRI-1867 did not affect hepatic cholesterol content and was able to strongly reduce SREBP-2 availability. Along this line, deficiency of the histone deacetylase sirtuin 6 (SIRT6) in the liver leads to elevated PCSK9 expression. Indeed, recruitment of SIRT6 to the PCSK9 promoter by the transcription factor FoxO3 represses PCSK9 transcription in part by suppressing hepatic SREBP-2 transcription (28,44). Accordingly, we found a strong upregulation of Sirt6 and Foxo3 in livers of (S)-MRI-1867–treated mice. Moreover, we also found a marked inhibition of E2f1 and a strong upregulation of Gpr94 expression in livers of (S)-MRI-1867–treated mice. The transcription factor E2f1 has been described as a major regulator of lipid and carbohydrate metabolism (45) and as a modulator of PCSK9 transcription by transactivation of PCSK9 promoter (29), whereas GPR94 can bind to PCSK9 and block its ability to sequester LDLR (27). As such, mice with GRP94 gene deletion display higher plasma LDL and markedly reduced hepatic LDLR levels compared with wild-type controls. Interestingly, single CB1R blockade with JD-5037 did not affect these parameters, while iNOS inhibition alone fully replicated the effects of (S)-MRI-1867 on PCSK9, LDLR, and SREBP-2. Finally, our data highlight for the first time the existence of a link between iNOS and PCSK9 metabolism and further indicate that iNOS inhibition reduces PCSK9 levels. PCSK9 inhibitors recently evaluated in clinical trials were reported to significantly improve the dyslipidemic state (46–48). As such, the lowering of PCSK9 levels by (S)-MRI-1867 is of particular importance, as it contributes to the improvement of dyslipidemia and, consequently, to a reduction of the residual cardiovascular risk. Although the exact mechanisms involved remain to be elucidated, we found that iNOS inhibition induced the activation of mTORC1, as evidenced by the increase in Ser²⁴⁴⁸ phosphorylation of mTOR and the increase in Raptor protein expression. Interestingly, Ai et al. (31) have elegantly demonstrated that mice treated with rapamycin displayed a major increase in PCSK9 and marked reduction in hepatic LDLR levels. The amino acid l-arginine can be sensed by mTORC1 through the specific GATOR2-interacting cellular arginine sensor for an mTORC1 (CASTOR1)-mediated mechanism (49,50) and can alleviate the inhibition of mTROC1, mediated by CASTOR1 or the lysosomal amino acid transporter SLC38A9 (32). We found that l-arginine reduced PCSK9 secretion in a dose-dependent manner, which strongly suggests that in the absence of iNOS activity, l-arginine availability is likely redirected toward mTORC1 activation, which then leads to the inhibition of PCSK9 secretion.

In conclusion, our findings demonstrate that in addition to the previously demonstrated antifibrotic efficacy (12), the concomitant inhibition of CB1R and iNOS by (S)-MRI-1867 could also regulate multiple critical pathways in lipoprotein homeostasis, as illustrated schematically in Fig. 7. Therefore, such a dual-targeting approach could have promising therapeutic potential in the treatment of dyslipidemia.

Acknowledgments. The authors thank Serge Monier from the FACS platform as well as Jean-Paul Pais de Barros and Hélène Choubley from the Lipidomic Analytical Platform, INSERM UMR1231 “Lipids, Nutrition, Cancer,” for excellent technical assistance in the fast-protein liquid chromatography and lipoprotein profile analysis. In addition, we are dedicating this work to the memory of Douglas Osei-Hyiaman, MD, PhD, who tragically died this year.

Funding. This work was supported by INSERM, University of Burgundy and Franche-Comté, by a French government grant managed by the French National Research Agency (ANR) under the program Investissements d’Avenir with the reference ANR-11-LABX-0021-01-LipSTIC LabEx and by a Conseil Régional Bourgogne grant (2019-Y-10658) to T.J.

Duality of Interest. R.C., M.R.I., and G.K. are listed as co-inventors on a U.S. patent covering (S)-MRI-1867 and related compounds (patent no. US 9765031 B2). No other potential conflicts of interest relevant to this article were reported.

Author Contributions. C.R., C.B., P.P.-D., and T.J. conceived and planned the experiments. C.R., C.B., T.M., J.L., R.C., and T.J. carried out the experiments. R.C., M.R.I., and G.K. provided the (S)- and (R)-MRI-1867 compounds. C.R., J.L., L.D., P.P.-D., R.C., M.R.I., G.K., B.V., P.D., and T.J. contributed to the interpretation of the results. C.R., G.K., P.D., and T.J. wrote the manuscript. All authors provided critical feedback and helped shape the research, analysis, and manuscript. T.J. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.