Ablation of Neuropilin 1 in Myeloid Cells Exacerbates High-Fat Diet–Induced Insulin Resistance Through Nlrp3 Inflammasome In Vivo

Obesity has become a global epidemic and a major cause of insulin resistance, which contributes to metabolic syndrome and type 2 diabetes (1). Chronic inflammation mediated by accumulated myeloid cells and lymphoid lineage in visceral adipose tissues is a major determinant of insulin resistance (2–4). Notably, macrophages infiltrated in visceral adipose tissues may constitute up to 40% of total obese visceral adipose tissue cells (3,4). Adipose tissue macrophages (ATMs) secrete proinflammatory cytokines, including tumor necrosis factor-α (TNF-α) (5) and chemokine (C-C motif) ligand 2 (Ccl2) (6). These cytokines inhibit insulin signaling in target tissues, impair glucose tolerance, and eventually instigate insulin resistance (5,7,8). Therefore, inhibition of macrophage-mediated chronic low-grade inflammation is an effective strategy to protect against obesity-induced insulin resistance and diabetes.

Emerging evidence supports an association between neuronal guidance cues and their receptors with insulin resistance. For example, axon-guiding class 3 semaphorin E (Sema3E) and its receptor plexin-D1 promote macrophage recruitment into adipose tissues, which triggers inflammation and systemic insulin resistance (9). Furthermore, another neuronal guidance cue, netrin-1 and its receptor Unc5, are highly expressed in visceral adipose tissues of obese human subjects and mice relative to lean ones (10). Netrin-1 has been shown to act as a local macrophage retention signal, leading to macrophage accumulation in adipose tissues (10). These studies suggest that the receptors of neuronal guidance cues may be required for inflammatory chemotaxis in adipose tissues. However, the direct links and molecular mechanisms of these receptors with respect to metabolic disorders remain largely unknown.

Neuropilin 1 (Nrp1), initially identified as a coreceptor for Sema3 and growth factors, including vascular endothelial growth factor, transforming growth factor-β, hepatocyte growth factor, and platelet-derived growth factor (11–14), has been shown to be strongly expressed in immune cells and to regulate immune response (15–17). In macrophages, ablation of Nrp1 results in decreased tumor growth and metastasis via enhanced infiltration of tumor-associated macrophages into normoxic tumor regions, which abolishes the proangiogenic and immunosuppressive functions of tumor-associated macrophages (18). Consistently, Miyauchi et al. (19) reported that Nrp1 ablation in glioma-associated microglia and macrophages suppresses glioma progression by promoting M1 macrophage polarization. Recent findings further establish the involvement of the Sema3A/Nrp1 axis, as well as M1/M2 macrophages, with respect to tumorigenic processes (20). However, specific functions of macrophage Nrp1 in the context of metabolic dysfunction, such as obesity-associated insulin resistance, have not been investigated. Here we report that a high-fat diet (HFD)–attenuated Nrp1 expression in macrophages correlates with accentuated insulin resistance, which is majorly caused by activated Nlrp3 inflammasome.

Nrp1^flox/flox mice (The Jackson Laboratory, Bar Harbor, ME) were backcrossed to C57BL/6J for at least 10 generations before crossing with lysozyme M (LysM)–Cre mice (The Jackson Laboratory) to generate myeloid cell–specific knockout (Nrp1^myel-KO) mice. Nrp1^flox/flox mice with the LysM-Cre transgene were Nrp1^myel-KO mice, and Nrp1^flox/flox littermates without the LysM-Cre transgene served as wild-type (WT) controls. To generate Nrp1^myel-KONlrp3^−/− mice, Nlrp3^−/− mice were crossed with Nrp1^myel-KO mice. Nrp1^myel-KONlrp3^+/+ littermates served as controls. All mice were C57BL/6J background, housed in temperature-controlled cages with a 12-h light-dark cycle, and given free access to water and food. Mice of different genotypes were housed in the same cage. Mice were fed an HFD (60% fat kcal, D12492; Research Diets Inc., New Brunswick, NJ) or a normal chow diet for 16 weeks. Mice were sacrificed, and white adipose tissue (WAT), liver, and skeletal muscle were isolated for biochemical analyses. The Georgia State University Institutional Animal Care and Use Committee reviewed and approved the animal protocol.

Peritoneal macrophages (PMs) were isolated as previously described with minor modification (21). Briefly, cells were collected by peritoneal lavage with 8 mL cold PBS with 10 mmol/L EDTA and 10% FBS. Macrophages were plated at 1.0 × 10⁶ per mL RPMI-1640 with 10% FBS. After incubation for 3 h at 37°C, the nonadherent cells were washed away, and adherent cells were collected for experiments.

Primary bone marrow–derived macrophages (BMDMs) were prepared by flushing the bone marrow of the femur and tibia of mice. Cells were cultured with complete RPMI-1640 medium supplemented with mouse macrophage colony-stimulating factor recombinant protein (50 ng/mL; eBioscience, San Diego, CA) for 7 days (22).

RAW264.7 cells (American Type Culture Collection, Manassas, VA) were cultured with DMEM containing 10% FBS and maintained at 37°C in 5% CO₂ and 95% air.

Nuclear factor (NF)-κB activation inhibitor (481406) was obtained from EMD Millipore (Billerica, MA). Ultrapure lipopolysaccharide (LPS) was bought from InvivoGen (San Diego, CA). ATP was purchased from Roche Applied Science (Basel, Switzerland). Nigericin was obtained from Sigma-Aldrich (St. Louis, MO). Nrp1 plasmid (#21934) was purchased from Addgene (Cambridge, MA). A Transcription Factor Kit for NF-κB p65 (#89859) was obtained from Thermo Fisher Scientific (Waltham, MA).

Glucose and insulin tolerance tests were performed as previously described (10). Briefly, for the glucose tolerance test (GTT), the mice were fasted for 16 h. After determination of fasted blood glucose levels, each animal received a glucose intraperitoneal injection of 1 g/kg body weight of glucose (25% d-(+)-glucose; G7528, Sigma-Aldrich). Blood glucose levels were determined after 15, 30, 60, and 120 min. The mice were fasted for 4 h, and insulin tolerance tests (ITTs) were performed by an intraperitoneal injection of 1.5 units/kg body weight of insulin (I9278, Sigma-Aldrich). Blood glucose levels were measured at 0, 15, 30, 60, and 120 min.

Glucose concentrations in whole venous blood were measured by an automatic glucose monitoring system (ReliOn Confirm). Serum insulin levels were measured by ELISA (ALPCO, Windham, NH). Serum free fatty acid (FFA) levels were measured using an enzyme kit (BioVision, Mountain View, CA). HOMA of insulin resistance (HOMA-IR) was calculated as (fasting insulin [mIU/L] × fasting glucose [mmol/L])/22.5 (23).

Quantitative data are expressed as mean ± SEM of at least three independent experiments. The difference between two groups was analyzed by the Student t test. The difference among more than two groups was analyzed by one-way ANOVA, followed by the Newman-Keuls multiple comparison test. Comparisons of different parameters between each group were made by two-way ANOVA, followed by the Bonferroni posttest. Statistical significance was evaluated with GraphPad Prism 5 software. P values of <0.05 were considered statistically significant.

To establish the association between Nrp1 and obesity, we adopted a diet-induced obesity model by feeding C57BL/6J mice with an HFD (60% kcal) for 16 weeks. Control mice were fed a normal diet (chow) in parallel with the HFD group. As expected, body weight and serum total cholesterol were significantly elevated in HFD-fed mice compared with chow-fed mice (Supplementary Fig. 1A and B). Furthermore, GTT and ITT results showed significantly impaired glucose tolerance and markedly blunted insulin responsiveness in HFD-fed mice relative to chow-fed mice (Supplementary Fig. 1C and D). These results indicated that HFD-fed mice successfully developed obesity-associated insulin resistance.

Next, we determined the expression of Nrp1 in visceral WAT, and as shown in Fig. 1A, Nrp1 mRNA levels were profoundly decreased in WAT of HFD-fed compared with chow-fed mice. Immunoblot analysis confirmed decreased Nrp1 protein in visceral WAT of HFD-fed mice relative to chow-fed mice (Fig. 1B). Furthermore, we assessed the expression of Nrp1 in subcutaneous WAT under the same dietary conditions but failed to detect differential Nrp1 expression (Fig. 1C and D), suggesting the reduction of Nrp1 by HFD is specific to visceral WAT. To identify the specific cellular compartment of WAT with decreased Nrp1 expression, we isolated the stromal vascular fraction (SVF) and adipocyte-rich fraction (ARF) from visceral WAT of chow- and HFD-fed mice. As shown in Fig. 1E and G, mRNA levels of Nrp1 were dramatically downregulated in the SVF but not the ARF of HFD-fed mice compared with samples from chow-fed mice. Attenuated Nrp1 protein was confirmed in SVF but not ARF by immunoblot analysis (Fig. 1F and H). Although macrophages represent the dominant cells in the SVF of adipose tissues, there are a variety of other cell types, including preadipocytes, mesenchymal stem cells, endothelial cells, T cells, and B cells (24,25). We therefore used immunofluorescence microscopy to visualize the costaining of Nrp1 with various established cell-specific markers, such as von Willebrand factor (endothelial cell), CD3 (T cell), or CD11c (dendritic cell) in visceral WAT of chow- or HFD-fed mice. The results showed that the HFD did not affect Nrp1 expression in these cells of visceral WAT (Supplementary Fig. 2A–C). We eventually isolated PMs from chow-fed and HFD-fed mice to ensure the purity of macrophages. As shown in Fig. 1I, Nrp1 mRNA was significantly downregulated by the HFD. Consistently, decreased Nrp1 protein was confirmed in PMs of HFD-fed mice relative to chow-fed mice (Fig. 1J). Together, these data suggest that the HFD suppresses Nrp1 expression in macrophages.

To investigate specific roles of macrophage Nrp1 in obesity-associated insulin resistance, Nrp1 was inactivated in myeloid cells by crossing Nrp1^flox/flox mice into mice expressing LysM-Cre recombinase. Quantitative RT-PCR data showed significantly reduced expression of Nrp1 mRNA in Nrp1^−/− BMDMs compared with WT BMDMs (Supplementary Fig. 3A). Decreased Nrp1 expression in Nrp1^−/− BMDMs was further confirmed by immunofluorescence staining and immunoblot analysis (Supplementary Fig. 3B and C. We also assessed tissue-specific deletion of Nrp1 in WAT, liver, and skeletal muscle of WT and Nrp1^myel-KO mice. Our data demonstrated that Nrp1 deletion was specific to macrophages (Supplementary Fig. 3D).

We next exposed WT and Nrp1^myel-KO mice to chow or the HFD for 16 weeks. As would be expected, the body weight of HFD-fed mice was markedly increased (Supplementary Fig. 4). However, no significant difference was observed between WT and Nrp1^myel-KO mice fed chow or the HFD (Supplementary Fig. 4). We used GTT and ITT to analyze possible alterations in glucose homeostasis in these mice. Although no significant difference was detected under chow-fed conditions in both mouse groups (Supplementary Fig. 5A and B), HFD-fed Nrp1^myel-KO mice were more intolerant to glucose and more resistant to insulin than their WT counterparts (Fig. 2A and B). Consistently, HFD-fed Nrp1^myel-KO mice displayed higher fasting serum insulin concentrations and elevated HOMA-IR index relative to their WT littermates (Fig. 2C and D). We also measured serum FFAs and observed no obvious difference between HFD-fed WT and Nrp1^myel-KO mice (Fig. 2E). Because the metabolic effects of insulin largely depend on phosphoinositol 3 kinase–protein kinase B (AKT) signaling in target tissues, we measured the phosphorylation (p) of AKT at Ser473 in WAT, liver, and skeletal muscle. Consistent with impaired insulin sensitivity in Nrp1^myel-KO mice fed the HFD, decreased p-AKT levels were evident in these tissues (Fig. 2F). Taken together, these data indicate that ablation of Nrp1 in myeloid cells leads to impaired maintenance of glucose homeostasis and blunted insulin sensitivity during obesity.

Obesity promotes the induction of a proinflammatory state and the accumulation of macrophages in adipose tissue (4,26). To investigate the potential inflammatory changes in WT and Nrp1^myel-KO mice fed chow or the HFD, we analyzed the gene expression profile of WAT, liver, and skeletal muscle by quantitative RT-PCR. The data demonstrated that upon chow feeding, Nrp1^myel-KO mice exhibited no significant difference with respect to inflammation profiles of WAT, liver, and skeletal muscle relative to these tissues of WT mice (Supplementary Fig. 5C–E). However, increased expression of genes encoding molecules typically linked to the development of insulin resistance (e.g., inducible nitric oxide synthase [iNOS], interleukin [IL]-13Rα1, Ccl2, and arginase II) were observed in HFD-fed Nrp1^myel-KO mice (Fig. 3A–C). In addition, data from immunohistochemistry analysis of CD68 revealed that the number of macrophages significantly increased in WAT and liver of HFD-fed Nrp1^myel-KO mice compared with the WT control group (Fig. 3D and E). In contrast, the number of T cells and dendritic cells was comparable in WAT of both mouse groups (Supplementary Fig. 6A and B). We also found that serum levels of TNF-α, IL-6, and IL-1β were dramatically increased in HFD-fed Nrp1^myel-KO mice (Fig. 3F–H). As a result of the association of liver steatosis with inflammation, we examined the accumulation fat in the liver. Comparable levels of fat accumulation were detected within both mouse groups (Supplementary Fig. 7). Together, these findings indicate enhanced systemic inflammation in Nrp1^myel-KO mice in response to the HFD.

Inflammasomes were assessed in tissues to explore potential involvement of inflammasomes in the heightened systemic inflammation of Nrp1^myel-KO mice. As shown in Fig. 4A, the expression of several inflammasomes, such as Nlrp1a, absent in melanoma 2 (AIM2), and Nlrc4, was similar in WAT of WT and Nrp1^myel-KO mice. Moreover, relative mRNA levels of inflammasome ASC adaptor and caspase 11 effector were comparable in WAT of both groups. However, relative mRNA levels of Nlrp3, caspase 1, IL-1β, and IL-18 were higher in WAT of Nrp1^myel-KO mice compared with WT littermates (Fig. 4A). In parallel, protein levels of Nlrp3, cleaved caspase 1 p20, pro–IL-1β, IL-1β p17, and IL-18 were profoundly elevated in WAT of Nrp1^myel-KO mice compared with WT littermates (Fig. 4B). Importantly, we also observed elevated priming and activation of Nlrp3 inflammasome in the liver and skeletal muscle of Nrp1^myel-KO mice (Fig. 4C–F). Taken together, myeloid cell–specific Nrp1 deficiency results in increased Nlrp3 inflammasome priming and activation in vivo.

Nlrp3 inflammasome function involves a twofold process: an initial priming phase and subsequent inflammasome activation (27). We investigated the effect of Nrp1 on LPS-induced Nlrp3 priming. As shown in Fig. 5A–D, Nrp1 deficiency markedly enhanced LPS-induced mRNA expression of Nlrp3, caspase 1, IL-1β, and IL-18. In line with these results, immunoblot analysis confirmed that pro–IL-1β, cleaved IL-1β p17, IL-18, Nlrp3, and cleaved caspase 1 p20 were substantially higher in LPS-primed Nrp1^−/− macrophages than those in WT cells, whereas ASC was not differentially modulated in the two genotypes (Fig. 5E and F). In contrast, LPS-induced mRNA expression of Nlrp3, IL-1β, and IL-18 was downregulated in RAW264.7 cells upon overexpression of Nrp1 (Supplementary Fig. 8A–C). Moreover, Nrp1 overexpression inhibited LPS-induced pro–IL-1β, cleaved IL-1β p17, and IL-18 protein in RAW264.7 macrophages (Supplementary Fig. 8D).

We also checked the effects of Nrp1 on Nlrp3 activation. WT and Nrp1^−/− BMDMs were primed, with or without LPS, and then treated with ATP or nigericin to trigger the activation of Nlrp3 inflammasome. As shown in Fig. 5G, pro–IL-1β, cleaved IL-1β p17, and IL-18 levels were markedly elevated in Nrp1^−/− macrophages compared with WT macrophages. Given the report of palmitate (PA)-mediated activation of the Nlrp3 inflammasome (28), WT and Nrp1^−/− BMDMs were primed with or without LPS to mimic elevated serum FFAs during obesity and then treated with BSA-PA. As shown in Fig. 5H, protein levels of pro–IL-1β and cleaved IL-1β were markedly increased in cell lysates and supernatants of Nrp1^−/− macrophages relative to WT macrophages. We further examined whether Nrp1 affected ASC cytoplasmic foci formation, an upstream event of caspase-1 cleavage. Consistent with previous study (29), ASC was diffusely localized in unstimulated WT and Nrp1^−/− BMDMs (Fig. 5I). However, the ASC cytoplasmic foci became obvious upon Nlrp3 activation with LPS and nigericin, (Fig. 5I). Importantly, Nrp1 deletion significantly promoted ASC foci formation (Fig. 5I). A similar pattern of Nlrp3-ASC complex formation was confirmed in WT BMDMs and markedly enhanced upon Nrp1 deletion (Fig. 5J). Taken together, these results suggest that Nrp1 deficiency promotes Nlrp3 inflammasome priming and activation in macrophages in vitro.

The essential role of the proinflammatory NF-κB pathway in priming the Nlrp3 inflammasome has been established (30). To determine whether increased Nlrp3 priming in Nrp1-deficient macrophages was dependent on the NF-κB pathway, we investigated potential functions of Nrp1 in NF-κB signaling. After LPS stimulation, Nrp1^−/− BMDMs exhibited enhanced phosphorylation of p65 and p105 compared with WT BMDMs (Fig. 6A). However, no significant difference in c-Jun N-terminal kinase or extracellular signal–regulated kinase 1/2 phosphorylation was observed between both groups (Fig. 6A). In line with these data, transcription factor NF-κB p65 activity detected by luminescence assay was significantly higher in Nrp1^−/− BMDMs than that in WT BMDMs (Fig. 6B). Next, we overexpressed Nrp1 in RAW264.7 cells to test direct roles of Nrp1 in the LPS-mediated signaling pathway. As shown in Fig. 6C, Nrp1 overexpression significantly inhibited LPS-induced p-IκBα (inhibitor of κBα) and iNOS levels in macrophages, suggesting that Nrp1 functions as an endogenous suppressor of LPS–NF-κB signaling. Finally, blockade of the NF-κB pathway could reverse Nrp1 knockdown-elevated Nlrp3, caspase 1 p20, pro–IL-1β, and cleaved IL-1β p17 levels in macrophages (Fig. 6D). Taken together, these data indicate that activated NF-κB signaling contributes to Nrp1 deficiency–induced Nlrp3 priming.

To test whether excessive activation of Nlrp3 inflammasome was a major contributor of insulin resistance in HFD-fed Nrp1^myel-KO mice, Nlrp3^−/− mice were crossed with Nrp1^myel-KO mice to generate Nrp1^myel-KONlrp3^−/− mice. Nrp1^myel-KONlrp3^+/+ (Nrp1^myel-KO) and Nrp1^myel-KONlrp3^−/− mice were fed the HFD for 16 weeks. Body weight was comparable in both groups (Fig. 7A). However, glucose tolerance and insulin sensitivity detected by GTT and ITT were improved in Nrp1^myel-KONlrp3^−/− mice compared with those in Nrp1^myel-KONlrp3^+/+ mice (Fig. 7B and C). Moreover, the improvement in glucose homeostasis of Nrp1^myel-KONlrp3^−/− mice fed the HFD was also validated by the lower fasting serum insulin concentrations and reduced HOMA-IR index in Nrp1^myel-KONlrp3^−/− mice compared with those in their Nrp1^myel-KONlrp3^+/+ littermates (Fig. 7D and E). However, the level of serum FFAs was not different between HFD-fed Nrp1^myel-KONlrp3^+/+ and Nrp1^myel-KONlrp3^−/− mice (Fig. 7F).

Next, we measured the p-AKT (Ser473) in WAT, liver, and skeletal muscle of HFD-fed Nrp1^myel-KONlrp3^+/+ and Nrp1^myel-KONlrp3^−/− mice. Consistent with improved insulin sensitivity of HFD-fed Nrp1^myel-KONlrp3^−/− mice, there was an increased p-AKT in WAT, liver, and skeletal muscle of HFD-fed Nrp1^myel-KONlrp3^−/− mice compared with that in HFD-fed Nrp1^myel-KONlrp3^+/+ controls (Fig. 7G).

Eventually, we detected the secretary proinflammatory cytokines in serum by ELISA. As shown in Fig. 7H–J, the levels of TNF-α, IL-6, and IL-1β were dramatically decreased in the serum of Nrp1^myel-KONlrp3^−/− mice compared with that in Nrp1^myel-KONlrp3^+/+ mice, respectively. Taken together, these data indicate that loss of Nlrp3 rescues dysregulated glucose homeostasis and insulin sensitivity during obesity in Nrp1^myel-KO mice, suggesting that the macrophage Nlrp3 pathway is at least partly responsible for HFD-induced insulin resistance in the mice.

In this study, we have uncovered a novel role for macrophage Nrp1 in repressing HFD-induced insulin resistance by suppressing Nlrp3 inflammasome priming and activation. Nrp1 expression in macrophages was markedly downregulated in HFD-induced mouse obesity. Importantly, the myeloid cell–specific Nrp1 deletion (Nrp1^myel-KO) in mice accentuated HFD-induced aberrant activation of Nlrp3, systemic inflammation, and insulin resistance. Finally, deletion of Nlrp3 was observed to markedly protect against both diet-induced inflammation and insulin resistance in Nrp1^myel-KO mice. Collectively, our data demonstrate that Nrp1 deficiency–accentuated Nlrp3 inflammasome in mouse models and macrophages is mediated by two interrelated mechanisms: NF-κB–mediated Nlrp3 priming and Nlrp3-ASC assembly. HFD-induced Nrp1 attenuation in macrophages promotes insulin resistance via Nlrp3-mediated chronic inflammation (Fig. 7K).

Nrp1 is a coreceptor for class 3 semaphorins, including Sema3A, B, C, D, E, and F, which possess axon guidance cue functions in neuronal cells (14). Nrp1 has been shown to promote endothelial tip rather than stalk cell function during sprouting angiogenesis (31). In this study, we have for the first time demonstrated a novel role for macrophage Nrp1 in attenuating insulin resistance in HFD-induced obesity. We found that Nrp1 levels were significantly decreased in macrophages of obese mice. Furthermore, in response to the HFD, Nrp1^myel-KO mice displayed impaired glucose homeostasis and blunted insulin sensitivity relative to WT mice. Unlike previous studies that majorly showed the involvement of macrophage Nrp1 in the regulation of tumor growth (19,20), our current study demonstrates the causative role of macrophage Nrp1 in HFD-induced insulin resistance. Our results are in line with other reports in which Sema3 family members and their cognate receptors have been linked to insulin resistance. For example, Shimizu et al. (9) described that Sema3E and its plexinD1 receptor are markedly increased in adipose tissue of obese mice. Sema3E was shown to function as a chemoattractant for plexinD1-expressing macrophages, resulting in the infiltration and accumulation of macrophages in adipose tissue and consequent adipose inflammation and insulin resistance (9). Another neuroimmune guidance cue, netrin-1 and its receptor Unc5b, have been shown to be upregulated in ATMs of obese human visceral adipose tissue (10). Netrin-1 promotes macrophage retention in adipose tissue via Unc5b and exaggerates metabolic dysfunction (10). Sema3A was previously identified as a chemoattractant for cortical apical dendrites (32) and regulates neuronal polarization by inhibiting axon formation and increasing dendrite growth (33). Recently, Casazza et al. (18) demonstrated that Sema3A is a chemoattractant for macrophages and that Nrp1 deletion significantly inhibits Sema3A-induced macrophage migration. Nrp1-dependent myeloid cell chemoattraction was also confirmed in neovascular retinal disease (34). Interestingly, we observed that myeloid cell–specific Nrp1 deficiency contributed to a higher accumulation of macrophages in WAT and liver of Nrp1^myel-KO mice.

One of the most important findings of this study is the novel function of Nrp1 as an important negative regulator of the NF-κB pathway. We found that Nrp1 deletion correlates with enhanced NF-κB activation, which in turn exerts proinflammatory effects through a Nlrp3 inflammasome-dependent and -independent manner. Nrp1 deletion–induced activation of NF-κB promotes Nlrp3 inflammasome via Nlrp3 priming; that is, NF-κB activation by Nrp1 loss increases the levels of inflammasome precursors. In addition, Nrp1 deletion promotes Nlrp3-ASC inflammasome assembly. Although the exact mechanisms are unclear, we have observed that Nrp1 deficiency results in attenuated LPS-induced 26S proteasome activation (data not shown). Because HOIL-1L is required for the assembly of Nlrp3-ASC inflammasome (29), we speculate that Nrp1 deletion induces proteasome inhibition, thereby delaying HOIL-1L degradation. Consequently, accumulated HOIL-1L promotes aberrant assembly of Nlrp3-ASC inflammasome. Further investigation is warranted to further delineate this complex and dynamic process.

Another key finding of this study is that Nlrp3 inflammasome is responsible for accentuated insulin resistance in HFD-fed Nrp1^myel-KO mice. The critical role of Nlrp3 within the condition was substantiated using Nlrp3-deficient/Nrp1^myel-KO mice. Marked protection against HFD-induced insulin resistance and decreased systemic inflammation was demonstrated in the double-KO mice. Together, these results reveal that Nrp1 may act as a critical endogenous inhibitor of Nlrp3 inflammasome. A published report by Wen et al. (28) has established a role of Nlrp3 inflammasome in HFD-induced insulin resistance. The authors showed that Nlrp3^−/− or Pycard^−/− mice exhibit improved insulin sensitivity compared with WT controls (28). Consistent with these observations, we found that Nlrp3 deficiency reversed HFD-induced insulin resistance in Nrp1^myel-KO mice, suggesting that the macrophage Nlrp3 pathway is at least partly responsible for HFD-induced insulin resistance in the mice. However, we cannot exclude the possibilities that Nlrp3-independent pathways may also contribute to insulin resistance in HFD-fed Nrp1^myel-KO mice.

Although this study provides insight on novel metabolic Nrp1 functions, a puzzling phenomenon remains unresolved about why Nrp1 levels become downregulated in macrophages but not in endothelial cells, T cells, or dendritic cells of HFD-fed obese mice. Recent studies have shown that serum FFAs are significantly increased during obesity and activate the proinflammatory NF-κB pathway (35) through fetuin-A, an endogenous ligand of Toll-like receptor 4 (36). Moreover, NF-κB p65 and p50 translocate to the nucleus upon activation and suppress the Nrp1 promoter in BMDM/macrophage precursor cells (37). We speculate that obesity may downregulate Nrp1 expression through activation of NF-κB signaling. However, future studies will be required to dissect whether obesity mediates Nrp1 regulation via other mechanisms. Potential consequences of Nrp1 with respect to additional obesity-related disease conditions, such as diabetes, should provide useful information on critical Nrp1 functions in metabolic disease.

In conclusion, our study identifies a novel function of Nrp1 as a negative regulator of Nlrp3 inflammasome in macrophages during obesity, suggesting potential therapeutic targeting of Nrp1 in chronic inflammation and insulin resistance.

Funding. This study was partly supported by grants from the National Heart, Lung, and Blood Institute (HL-079584, HL-080499, HL-089920, HL-110488, HL-128014, and HL-132500), National Cancer Institute (CA-213022), and National Institute on Aging (AG-047776). M.-H.Z. is an eminent scholar of Georgia Research Alliance.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. X.D., I.O., P.S., and M.-H.Z. conceived experiments and wrote the manuscript. X.D., Z.L., T.B., Q.W., T.R., and M.Z. performed experiments and analyzed data. All authors had final approval of the submitted and published versions. M.-H.Z. 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.