Specific NLRP3 Inhibition Protects Against Diabetes-Associated Atherosclerosis

Chronic sterile inflammation is a significant driver of diabetes-associated atherosclerosis, a macrovascular complication that is largely responsible for cardiovascular morbidity and mortality in the population with diabetes (1,2). Sterile inflammation, which occurs in the absence of viral or bacterial pathogens, and the downstream activation of large, cytosolic protein receptors known as inflammasomes have been implicated in all of the complications associated with diabetes (3,4). The Nod-like receptor family pyrin domain containing 3 (NLRP3) inflammasome has emerged as the most prominent member of the inflammasome family involved in diabetes and atherosclerotic disease pathogenesis (5,6).

In the diabetic milieu, hyperglycemia and increased oxidative stress, together with elevated oxidized LDLs, free fatty acids, and cholesterol crystals, act as endogenous host-derived triggers for the assembly and activation of the NLRP3 inflammasome (7,8). In response to these damage-associated molecular patterns, NLRP3 binds to its adaptor protein, apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC), which in turn recruits caspase-1. The fully formed NLRP3 inflammasome complex promotes caspase-1 cleavage and activation, which triggers the maturation of proinflammatory cytokines such as pro–interleukin-1β (IL-1β) and pro–IL-18 into their active secreted forms, triggering an inflammatory cascade (4,9). The importance of the NLRP3 inflammasome and its components in atherosclerotic disease is substantiated in several preclinical studies, which have demonstrated that deficiency in NLRP3, ASC, IL-1β, or caspase-1 protects against the development of diet-induced atherosclerosis (7,10,11). In diabetic mice, aortic protein expression of NLRP3, ASC, caspase-1, and IL-1β was upregulated compared with nondiabetic controls and was associated with enhanced lesion development and an elevation of reactive oxygen species levels (12), while knockdown of NLRP3 protected against diabetes-associated atherosclerosis (13). Moreover, in the clinical setting, increased expression of NLRP3, ASC, caspase-1, and IL-1β was observed in unstable carotid atherosclerotic plaques when compared with stable plaques and control patients who had no evidence of coronary artery stenosis (14).

As a potent inflammatory cytokine, IL-1β acts both locally on the vasculature, in particular on endothelial cells, smooth muscle cells, and lesional monocytes/macrophages, as well as systemically to contribute to atherogenesis. IL-1β increases expression of adhesion molecules such as vascular cell adhesion molecule 1 (VCAM-1), intracellular adhesion molecule 1 (ICAM-1), and MCP-1 in vascular endothelial cells that are important for monocyte recruitment (1,15,16). Additionally, IL-1β also acts on vascular smooth muscle cells and increases adhesion molecules and monocyte–smooth muscle cell interactions and promotes cell proliferation (1,17,18).

The recent success of the CANTOS trial, a large-scale multicenter clinical trial that tested the efficacy of a monoclonal antibody (canakinumab) targeted at IL-1β, has solidified the clinical significance of targeting IL-1β to reduce cardiovascular events (1,19). This pivotal trial provided the first proof-of-principle for the inflammation hypothesis in driving atherosclerosis in humans. Recent developments of specific NLRP3 small-molecule inhibitors have sparked great interest as anti-inflammatory therapeutics. One well-characterized NLRP3 small-molecule inhibitor is MCC950, which specifically binds to the NLRP3 ATP hydrolysis (or Walker B) motif, thereby preventing the second step in NLRP3 activation and IL-1β maturation (20,21). Indeed, MCC950 was shown to attenuate carotid artery plaque size in a short-term model of atherosclerosis (22). However, its efficacy has never been investigated in chronic inflammatory vascular disease induced by diabetes. In this study, we tested the hypothesis that specific and targeted inhibition of the NLRP3 inflammasome with MCC950 could protect against diabetes-associated atherosclerosis by reducing inflammation and improving vascular function.

To induce insulin-deficient type 1 diabetes, 8-week-old male apolipoprotein E–knockout (ApoE^−/−) mice were rendered diabetic by intraperitoneal injections of streptozotocin (100 mg/kg/day) or citrate vehicle (nondiabetic controls) on 2 consecutive days. After 9 weeks of diabetes, mice were randomized and injected intraperitoneally 3 days/week with vehicle (3% DMSO in PBS) or MCC950 (5 mg/kg in 3% DMSO) for 9 weeks (18-week cohort; flowchart of animal use and number for each experiment is shown in Supplementary Fig. 1). Throughout, mice were fed standard chow ad libitum and kept at room temperature. Blood glucose and body weights were recorded on a weekly basis to confirm diabetic status. Since changes in proinflammatory and oxidative stress-related gene expression (23) as well as changes in vascular function (24) occur earlier in the progression of atherosclerotic disease, additional cohorts of mice were administered MCC950 at the same concentration as above, after 5 weeks of diabetes for an additional 5 weeks (10-week cohort; flowchart of animal use and number for each experiment is shown in Supplementary Fig. 2), and assessed for endothelial function and gene expression.

All animal experiments were approved by the Alfred Medical Research and Education Precinct animal ethics committee (ethics number E1775/2017/B) (Melbourne, Victoria, Australia), and investigations conformed to National Health and Medical Research Council (NHMRC) of Australia guidelines.

Heparinized blood, collected after puncture of the right ventricle, was centrifuged to obtain plasma samples. Lipids were measured by standard assays using commercial enzymatic kits as described previously (25). For the 10-week cohort, aortae were cleaned in Krebs buffer to remove peripheral fat and cut into two sections. Thoracic aortae were mounted on a wire myograph (Danish Myo Technology) for vascular reactivity analysis, while arch and abdominal aortae were snap frozen for RNA extraction. For the 18-week cohort, atherosclerotic lesions were assessed using the en face method after Sudan IV staining to assess arch, thoracic, and abdominal lesions. Sections were then embedded in paraffin for immunohistochemistry. Heart sinuses were frozen in OCT, and sinus lesions were determined after staining with Oil Red O as described previously (26).

In order to investigate whether MCC950 improves vasodilation in diabetic ApoE^−/− mice, vascular reactivity studies were performed as published previously (27). Briefly, thoracic aortae were cut into 4-mm segments and mounted on two L-shaped metal prongs. Aortae were equilibrated for 30 min at a resting tension of 5 mN. All aortae were then exposed to high-potassium physiological salt solution to determine viability. Next, cumulative concentration responses to acetylcholine (ACh; 1 nmol/L to 100 µmol/L) were recorded in aortae preconstricted to ∼50% high-potassium physiological salt solution with phenylephrine (PE). All vasorelaxation responses are presented as percentage relaxation of the preconstriction response. Additionally, a concentration-response curve to PE (1 nmol/L to 100 µmol/L) was performed to assess vascular contractility. The variable slope sigmoidal concentration-response curves to all agonists for each mouse were calculated and plotted using GraphPad Prism (version 8.0).

Total RNA was extracted after homogenization of snap frozen aortae, and gene expression of VCAM-1, ICAM-1, nuclear factor-κB subunit p65 (p65 NF-κB), MCP-1, IL-1β, and tumor necrosis factor-α (TNF-α) was analyzed by quantitative RT-PCR (qRT-PCR) as described previously (26).

Plasma IL-6 and caspase-1 levels were measured using the Mouse IL-6 Quantikine ELISA Kit and mouse caspase-1 matched-pair detection set, respectively, according to the manufacturer’s instructions (Supplementary Table 1). White blood cells, neutrophils, and monocytes were calculated in whole blood using the Hemavet system. Urinary 8-isoprostanes were measured using an ELISA kit as per the manufacturer’s instructions.

Aortic IL-1β, NLRP3, nitrotyrosine, and CD68 localization and expression were determined by immunohistochemistry. In brief, aortic paraffin sections were dewaxed, and endogenous peroxidases were inactivated with 3% H₂O₂ in Tris-buffered saline. For CD68 staining, frozen aortic sections were thawed and fixed in ice-cold acetone. Sections were incubated with a serum blocking agent and a biotin-avidin blocking kit (Vector Laboratories). Sections were then incubated with the respective primary antibodies overnight at 4°C. Secondary antibody, biotinylated anti-rabbit Ig, or the biotinylated anti-rat secondary antibody was then added for 30 min followed by horseradish peroxidase–conjugated streptavidin (1:500), incubated for 3 min in 3,3′-diaminobenzidine tetrahydrochloride, and counterstained with hematoxylin. Images were visualized under light microscopy and quantitated using Image-Pro Plus.

IL-1β, α-smooth muscle actin (α-SMA), and Ly6C cells were examined in the aortic sinus using fluorescence microscopy as described previously (28). In brief, frozen aortic sinus sections were thawed and fixed in prechilled acetone and incubated in 3% H₂O₂ in distilled water. Sections were blocked with 5% normal goat serum for 60 min and stained with Ly6C, α-SMA, or IL-1β overnight at 4°C. Next, sections were washed and stained with anti-rat, anti-rabbit, or anti-goat fluorescent antibodies for 1 h at room temperature. Sections were then mounted with ProLong Gold Antifade mountant with DAPI. Lesions were imaged on a confocal microscope and analyzed using ImageJ. Data are expressed as the average number of Ly6C cells per section.

Frozen aortic sinus sections were thawed and fixed in prechilled acetone followed by staining with hematoxylin and eosin and 0.1% Picrosirius red.

Please refer to Supplementary Table 2 for detailed antibody information.

Human aortic smooth muscle cells (HAoSMCs) from subjects without and with diabetes (Lonza Clonetics) were grown in SmBM-2 media (Lonza) supplemented with 2% FBS at 37°C in 5% CO₂. Primary bone marrow–derived macrophages (BMDMs), isolated from C57/BL6 mice, were grown in l-cell conditioned media (29). BMDMs and AoSMCs were primed with lipopolysaccharide (LPS) (18 h; 0.1 μg/mL) and treated with MCC950 (1 h; 0.01–10 μmol/L) before ATP (4 h; 1 mmol/L) was added as the second stimulus. For cholesterol efflux assays, BMDMs were treated as per Mukhamedova et al. (30) in the presence of 1 µmol/L MCC950. For gene expression studies, cells were primed with LPS (18 h; 0.1 μg/mL), treated with MCC950 (5 h; 0.01–10 μmol/L), and activated with ATP (4 h; 1 mmol/L). IL-1β and caspase-1 secretion in the supernatant was determined by ELISA as per the manufacturer’s instructions. RNA extraction, cDNA synthesis, and qRT-PCR were performed as described previously (27). A separate set of BMDMs was activated with sodium palmitate (P9769; Sigma-Aldrich), which was added as a palmitate–fatty acid–free BSA complex (17 h; 200 µmol/L).

Monocyte–smooth muscle interactions were determined by in vitro static cell adhesion assays as described previously (27). Briefly, human monocytic THP-1 cells were labeled with a CellVue Burgundy Fluorescent Labeling Kit (Invitrogen) and incubated with primed and activated HAoSMCs (as above) in the presence or absence of 1 µmol/L MCC950 for 20 min at 37°C. Thereafter, HAoSMCs were washed twice with PBS and fixed with 10% neutral buffered formalin and plates scanned using an Odyssey infrared scanner. Fluorescence intensity (700 nm) of adherent THP-1 cells was quantified using Odyssey software. Representative images were taken using a cellSens microscope at ×10 magnification.

All data are expressed as mean ± SEM. For comparisons between groups, a one-way ANOVA with Tukey multiple-comparisons post hoc test was performed. All statistical analyses were performed using GraphPad Prism version 8.0. A P value <0.05 was considered statistically significant.

The data sets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Diabetic ApoE^−/− mice exhibited elevated blood glucose levels and lower body weight, which was assessed weekly, as compared with their nondiabetic counterparts (Fig. 1A and B) (P < 0.001). Glycated hemoglobin (HbA_1c) levels, a long-term indicator of diabetic status, were also increased in the diabetic cohorts (Fig. 1C). While MCC950 treatment had no effect on blood glucose (Fig. 1A) and HbA_1c levels (Fig. 1C [%] and Supplementary Table 3 [mmol/mol]), it did improve body weight in the diabetic cohort (Fig. 1B). MCC950 had no effect on the lipid profile (cholesterol, triglyceride, HDL, and LDL) or blood pressure under diabetic or nondiabetic conditions (Supplementary Table 3), suggesting that any effects of treatment are lipid and blood pressure independent.

As expected, diabetes increased atherosclerotic total plaque approximately fourfold (Fig. 1D and E) with significant increases observed in the arch, thoracic, and abdominal aortic regions (Fig. 1F–H). MCC950 treatment had no effect in the nondiabetic ApoE^−/− mice, but significantly attenuated atherosclerotic lesions in diabetic ApoE^−/− mice by 37% (Fig. 1D–H) (P < 0.001). Lesion deposition in the aortic sinus showed a tendency toward an increase in size after 18 weeks of diabetes; however, MCC950 had no effect on sinus lesion size (Fig. 1I and J).

Impairment of vascular function, triggered by oxidative stress and inflammation, two key characteristics in diabetes, is considered the initial step toward atherosclerotic lesion formation (31). Vascular endothelial cells and smooth muscle cells together play a role in the regulation of vascular function and the maintenance of vascular homeostasis (32). Therefore, we sought to investigate the effect of MCC950 on ACh-induced relaxation and PE-induced contraction ex vivo, which give an indication of vascular function. Aortic rings isolated from diabetic ApoE^−/− mice exhibited reduced maximal dilation in response to ACh (R_max: 85.2% [control] vs. 64.1% [diabetic]) and enhanced contraction in response to PE (R_max: 3.9 g [control] vs. 10.0 g [diabetic]) (Fig. 2A–C). MCC950 treatment showed a tendency toward improved maximal dilation (77.3%) in response to ACh in diabetic ApoE^−/− vessels; however, this did not reach significance (Fig. 2A and C). PE-induced contraction was significantly reduced with MCC950 treatment in diabetic ApoE^−/− vessels (Fig. 2B and C) (P < 0.05). Collectively, these data suggest that MCC950 enhances vascular tone and function. However, this is less likely to be dependent on the nitric oxide vasorelaxation pathway but rather on the adrenergic constriction pathway of the vascular smooth muscle cells.

Clinically, an inverse correlation between oxidative stress and endothelial function has been shown in patients with diabetes (33). In this study, tissue-specific oxidative stress in the aortae was assessed by nitrotyrosine staining, an indicator of peroxynitrite-induced nitrotyrosinated protein modification (26,27). Diabetic vessels exhibited increased nitrotyrosine immunostaining, which was attenuated with MCC950 treatment (Fig. 2D and E) (P < 0.01). Moreover, systemic oxidative stress, as assessed by urinary 8-isoprostane levels, was markedly increased in the diabetic ApoE^−/− cohort and significantly reduced with MCC950 treatment (Fig. 2F) (P < 0.05).

After 10 weeks of diabetes, aortic gene expression of proatherogenic mediators, in particular TNF-α, ICAM-1, and MCP-1, was significantly upregulated (Fig. 3A–C). This was significantly attenuated after 5 weeks of treatment with MCC950. Additionally, there was a tendency toward increased gene expression of IL-1β, VCAM-1, and p65 NF-κB (Fig. 3D–F). MCC950 treatment significantly reduced IL-1β gene expression and showed a trend toward reduction for VCAM-1 and p65 NF-κB (Fig. 3D–F).

Differential white blood cell counts from whole blood preparations were used as a measure of systemic inflammation. Total white blood cell count was significantly increased in the diabetic ApoE^−/− cohort, whereas specific monocyte and neutrophil counts were unaltered in the diabetic setting (Fig. 3G–I). MCC950 treatment had no effect on white blood cell count (Fig. 3G–I). In addition, plasma levels of IL-6 were measured as a surrogate marker for IL-1β, as IL-1β can lead to the downstream production of IL-6, which then acts to amplify inflammation (34,35). Diabetes caused an augmentation of plasma IL-6 levels, which were reduced with MCC950 treatment (Fig. 4F) (P < 0.01). Similar trends were observed when assessing secreted caspase-1 levels in plasma, where MCC950 treatment resulted in lower levels under diabetic and nondiabetic conditions (Fig. 4E).

Both NLRP3 and IL-1β protein levels were augmented in the diabetic aortae compared with nondiabetic controls, with localization demonstrated mainly in the endothelial and smooth muscle area of the vessel wall (Fig. 4A–D). NLRP3 and IL-1β levels were reduced in the diabetic vessels with MCC950 treatment (Fig. 4A–D).

In atherosclerosis, therapies that target monocyte recruitment or macrophage content stabilize atherosclerotic plaques (36). Therefore, we explored monocyte recruitment into atherosclerotic lesions by quantifying the abundance of Ly6C cells and macrophage content by staining for CD68, a protein highly expressed by circulating and resident macrophages, in the aortic sinus plaque region. There was a tendency toward increased Ly6C cells in diabetic plaque as compared with nondiabetic controls, with a trend toward a reduction with MCC950 treatment (Fig. 5A and B). Diabetic aortic sinus plaques demonstrated a significant increase in CD68 staining as compared with nondiabetic controls (Fig. 5C and D) (P < 0.05), with staining observed uniformly throughout the plaque (Supplementary Fig. 3). MCC950 treatment resulted in a decrease in macrophage content (Fig. 5C and D) (P < 0.05), with CD68 staining observed predominantly toward the periphery of the plaque (Supplementary Fig. 3). IL-1β levels were similarly lower in diabetic plaque after MCC950 treatment, although significance was not reached (Fig. 5E and F).

Importantly, MCC950 significantly reduced the necrotic core of diabetic plaque (P < 0.05) (Fig. 6A and B) without affecting α-SMA content (Fig. 6C and D) or collagen content as determined by Picrosirius red staining (Supplementary Fig. 4). Similarly, fibrous cap thickness, although significantly increased by diabetes (P < 0.05), was unaltered by MCC950 treatment in the diabetic plaque (Fig. 6E).

Since NLRP3 is highly expressed in macrophages and vascular smooth muscle cells, with both cell types playing a prominent role in atherogenesis, we determined the effect of NLRP3 inhibition in these mouse and human cell types in the context of an environment with diabetes to complement our in vivo findings.

Mouse BMDMs cultured in either low glucose (5 mmol/L) or high glucose (25 mmol/L) and primed with LPS revealed enhanced IL-1β and caspase-1 gene expression (Fig. 7A and B) (P < 0.001). The inclusion of l-glucose and mannitol confirmed that the observed changes were specific to d-glucose (Supplementary Fig. 5). Caspase-1 gene expression was mildly enhanced in BMDMs cultured in high glucose compared with low glucose (Fig. 7B) (P < 0.05); however, MCC950 treatment did not alter IL-1β or caspase-1 gene expression, which is consistent with other reports that MCC950 does not limit gene transcription (20). As expected, primed and NLRP3-activated BMDMs (LPS plus ATP) triggered the secretion of IL-1β and caspase-1 (Fig. 7C and D). Secretion of these proteins was further augmented by approximately twofold in BMDMs cultured in high glucose (Fig. 7C and D) (P < 0.01). Importantly, MCC950 treatment (0.01–1 μmol/L) resulted in a dose-dependent reduction of both secreted IL-1β and caspase-1, irrespective of glucose conditions (Fig. 7C and D). Similarly, human THP-1 cells and phorbol 12-myristate 13-acetate–differentiated macrophages secreted significantly increased amounts of IL-1β upon NLRP3 inflammasome activation with LPS plus ATP under high glucose conditions, which were significantly attenuated by MCC950 treatment (Fig. 8A and B, respectively). Palmitate, a saturated fatty acid known to increase under diabetogenic conditions, is reported to trigger IL-1β secretion in an NLRP3-dependent manner (37). We found that palmitate increased LPS-induced IL-1β transcription and secretion in BMDMs (Fig. 7E and F). Importantly, MCC950 reduced IL-1β secretion after palmitate activation but had no effect on transcription, in keeping with its known action in inhibiting NLRP3 at the posttranslational level (Fig. 7E and F). Since ABCA1 and ABCG1 transporters also contain ATP hydrolysis domains, we tested whether MCC950 affected cholesterol efflux from BMDMs to lipid-free apolipoprotein A-I, a substrate for ABCA1, or to isolated HDL, a substrate for ABCG1. MCC950 had no effect on cholesterol efflux mediated by either carrier, further demonstrating its specificity for NLRP3 (Supplementary Fig. 6).

HAoSMCs isolated from a patient without and with diabetes were primed with LPS and NLRP3 activated by ATP treatment. IL-1β gene expression was increased with LPS but was unaffected by MCC950 treatment (Fig. 8C), whereas significantly higher levels of secreted IL-1β were detected in the supernatants from HAoSMCs from patients with diabetes (Fig. 8D). While MCC950 treatment had no effect on HAoSMCs from control subjects without diabetes, there was a dose-dependent decrease in IL-1β secretion in HAoSMCs from subjects with diabetes (Fig. 8D) (P < 0.01).

Monocyte adhesion to vascular smooth muscle cells contributes to the retention of monocytes/macrophages within the subendothelial layer, thereby contributing to the atherogenic process (38). Since IL-1β has been implicated in monocyte–smooth muscle cell interactions and the increase in adhesion molecule gene expression, we performed a static adhesion assay to determine if MCC950 had an effect on this downstream inflammatory process. HAoSMCs from subjects with diabetes had an approximately twofold increase in the adherence of fluorescently labeled THP-1 monocytic cells under unstimulated conditions and when activated with LPS and ATP (Fig. 8E and Supplementary Fig. 7) (P < 0.001). This increase in adherence was closely associated with upregulated expression of VCAM-1 (Fig. 8F). MCC950 treatment concomitantly decreased THP-1 adhesion as well as VCAM-1 gene expression. Collectively, these data demonstrate that specific NLRP3 inhibition with MCC950 blocks IL-1β release, thereby effectively inhibiting known downstream mediators that induce inflammation in the diabetic setting.

Diabetes-associated atherosclerosis is considered a chronic inflammatory disorder in which the potent inflammatory cytokine IL-1β plays a causal role in disease pathogenesis. Indeed, in recent years, it has become increasingly apparent that clinically applicable interventions that interfere with IL-1β signaling or downstream pathways can improve cardiovascular outcomes, ushering in a new era of targeted anti-inflammatory therapies for atherosclerosis. In our interventional approach, we demonstrate that MCC950, a highly specific and selective inhibitor of the NLRP3 inflammasome, reduces atherosclerotic plaque by lowering inflammation and oxidative stress both systemically as well as locally within the plaque and vasculature. Specifically, in the diabetic macrovasculature, we demonstrate a reduction in NLRP3 and IL-1β protein with MCC950 treatment. This coincided with marked attenuation of several diabetes-induced proatherosclerotic inflammatory genes, such as TNF-α, ICAM-1, and MCP-1, most likely due to a feed-forward effect as a consequence of reduced IL-1β on downstream effector cells that in turn produce less effector-driven inflammatory responses. MCC950 exerted its antiatherogenic effects in a lipid-independent fashion, as reflected by the lack of an effect on the lipid profile after 9 weeks of MCC950 intervention. Our data concur with several studies that show increased activation of the NLRP3 inflammasome and downstream inflammation and oxidative stress induced by diabetic stimuli, which were reversed by genetic knockdown of the NLRP3 gene (13,39).

Patients with diabetes invariably exhibit clinical features of endothelial dysfunction, which correlates with enhanced oxidative stress and inflammatory parameters, all of which precede the development of atherosclerosis (40). Thus, to further clarify the impact of NLRP3 inhibition on endothelial dysfunction, we performed endothelial functional studies in the presence of short-term diabetes (10 weeks) with and without MCC950 treatment. Our endothelial functional studies reveal that while diabetes impairs ACh-induced endothelium-dependent relaxation and enhances PE-induced vasoconstriction, MCC950 had a predominant effect in reducing PE-induced contraction. These data indicate that MCC950 targets smooth muscle cell contraction in the aorta as opposed to a specific effect on endothelial cells. In support of this observation, acute IL-1β administration ex vivo was shown to directly enhance PE-induced contraction (41), while NLRP3 gene deletion protected against vascular smooth muscle cell remodeling in a hypertensive model (42), suggesting that MCC950 could improve contractile responses to PE in our model by inhibiting IL-1β activation. In addition, detailed electron microscopy and immunohistochemical analysis have revealed that smooth muscle cells are in direct contact with monocytes/macrophages, enhancing the retention of inflammatory cells within the early lesion (17,43). This is mainly achieved by secreting cytokines, including IL-1β, which induce the expression of cell-surface adhesion molecules such as VCAM-1 and ICAM-1 on smooth muscle cells (18,38,44). Our in vitro mechanistic data support this concept since we see static adherence of monocytes to vascular smooth muscle cells, which is further augmented in the diabetic setting. Importantly, MCC950 is able to reduce monocyte–smooth muscle cell interaction and limit the gene expression of these proinflammatory cellular adhesion molecules.

Currently, it is unknown what effect, if any, inhibition of the NLRP3 inflammasome might have on diabetes-associated inflammatory cytokines and the inflammatory cell types producing these cytokines. Based on previous patient studies, it appears that myeloid cells isolated from newly diagnosed patients with diabetes exhibit increased expression of the NLRP3 inflammasome and its components, including mature IL-1β, which were reversed with antidiabetic treatment (45). In line with this, we show two important findings. Firstly, high glucose treatment of BMDMs augments caspase-1 and IL-1β secretion, suggesting that under diabetic conditions, activation of the NLRP3 inflammasome is heightened. Secondly, NLRP3 inhibition by MCC950 is effective in dampening secretion of proinflammatory cytokines in response to diabetogenic mediators such as glucose and palmitate, highlighting its potential usefulness under diabetic conditions.

With respect to the involvement of monocytes and macrophages in diabetes-associated atherogenesis, tracking dyes have shown direct evidence of increased monocyte infiltration and macrophage retention in diabetic plaque (46). In agreement, we show that there are increased monocyte/macrophages within the plaque in a diabetic setting, which are reduced after MCC950 treatment. This has most likely arisen as a consequence of reduced monocyte infiltration rather than enhanced macrophage egress, in line with previous data (46) showing the importance of monocyte infiltration versus macrophage egression under diabetic conditions. Limitations of tissue availability and the size of the aortic sinus did not permit M1/M2 ratio analysis. Importantly, our data show reductions in necrotic core size after MCC950 treatment. In addition, MCC950 maintains fibrous cap thickness and collagen and α-SMA content of the plaque. Taken together, MCC950 appears to stabilize the plaque, making it less prone to rupture. These positive outcomes of MCC950 on plaque stability occurred in the absence of a reduction in lesion size in this region. This phenomenon is not uncommon in the mouse aortic sinus, where several lipid-lowering and antioxidant therapies have shown no reduction in lesion size (25,47,48) and where lesion development is driven mostly by hemodynamic parameters such as low shear stress (48). It is also important to note that unlike the recently reported brief study of van der Heijden et al. (22), atherosclerotic lesions were unaltered after MCC950 treatment in nondiabetic mice. This most likely reflects the less inflammatory nature of atherosclerosis observed under nondiabetic conditions in our model as reflected by the lower levels of NLRP3 and IL-1β in nondiabetic vessels (Fig. 4B and D) compared with diabetic counterparts. Lastly, a limitation of the study is that we did not investigate the role of proatherogenic NLRP3-mediated pyroptosis, and the implication of inhibiting this process in diabetes-associated plaque formation (49).

Interest in developing small molecules to target inflammation has been buoyed with the recent success of the IL-1 antibody canakinumab in the CANTOS trial, which demonstrated lower rates of recurrent cardiovascular events independent of lipid lowering (19). However, patients administered canakinumab had a higher incidence of fatal infections, suggesting that despite its positive effects on cardiovascular disease, long-term inhibition of this important cytokine may compromise host immune defenses (19). Additionally, targeting the resolution of inflammation, particularly via lipoxins, is yet another novel strategy to improve diabetes-associated atherosclerosis (50). As shown in this study, targeting the NLRP3 inflammasome with MCC950 might offer a unique way to circumvent fatal microbial infections since this small-molecule inhibitor is ineffective against pathogen-responsive inflammasomes such as NLRP1 and AIM2, highlighting its specificity (20). Our study provides a strong preclinical foundation that specific NLRP3-inflammasome inhibitors can reduce inflammation and lessen the complications of diabetes, thereby offering a therapeutic avenue to reduce cardiovascular events with potentially fewer side effects. In summary, in this era of transformative and targeted anti-inflammatory therapies, small-molecule NLRP3 inhibitors like MCC950 may offer a major advance in managing the unmet need of diabetes-associated cardiovascular complications.

Acknowledgments. The authors thank Professor Avril Robertson (The University of Queensland, St. Lucia, Queensland, Australia) for the MCC950 used in the study. The authors also thank Professor Owen Woodman (Monash Institute of Pharmaceutical Sciences, Parkville, Victoria, Australia) for the use of the vessel myograph and the Monash Micro Imaging facility for provision of instrumentation and training.

Funding. A.S. is supported by an early career fellowship from the National Health and Medical Research Council (NHMRC) and a Diabetes Australia Research Trust general grant. J.E.V. is supported by NHMRC Project grants (1145788 and 1101405), an Ideas grant (1183070), and a Fellowship (1141466). R.H.R. is supported by NHMRC Project Grants and a Senior Research Fellowship. J.B.d.H. is supported by a Baker Fellowship. This work was also supported by operational infrastructure grants through the Australian Government Independent Research Institute Infrastructure Support Scheme (9000220) and the Victorian State Government Operational Infrastructure Support, Australia.

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

Author Contributions. A.S. and J.B.d.H. conceived and designed the research and drafted the manuscript. A.S., J.S.Y.C., N.S., A.A.-S., N.M., and D.S.S. performed the experiments. A.S., J.S.Y.C, A.A.-S., D.S.S., N.M., K.J.-D., A.J.M., J.E.V., R.H.R., and J.B.d.H. provided intellectual input and analyzed the data. A.S., J.S.Y.C., and J.B.d.H. prepared the figures. A.S., J.S.Y.C., N.S., A.A.-S., D.S.S., N.M., K.J.-D., A.J.M., J.E.V., R.H.R., and J.B.d.H. approved the final version of the manuscript. J.B.d.H. and A.S. are the guarantors of this work and, as such, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. This study was presented in oral form at the 14th World Congress of Inflammation, Sydney, New South Wales, Australia, 15–19 September 2019.