Inflammation in the Pathophysiology and Therapy of Cardiometabolic Disease

Abstract

The role of chronic inflammation in the pathogenesis of type 2 diabetes mellitus and associated complications is now well established. Therapeutic interventions counteracting metabolic inflammation improve insulin secretion and action and glucose control and may prevent long-term complications. Thus, a number of anti-inflammatory drugs approved for the treatment of other inflammatory conditions are evaluated in patients with metabolic syndrome. Most advanced are clinical studies with IL-1 antagonists showing improved β-cell function and glycemia and prevention of cardiovascular diseases and heart failure. However, alternative anti-inflammatory treatments, alone or in combinations, may turn out to be more effective, depending on genetic predispositions, duration, and manifestation of the disease. Thus, there is a great need for comprehensive and well-designed clinical studies to implement anti-inflammatory drugs in the treatment of patients with metabolic syndrome and its associated conditions.

Essential Points

• In patients with metabolic syndrome, activation of the innate immune system contributes to the development of type 2 diabetes and associated complications

• Treatment of patients with type 2 diabetes with an IL-1 antagonist may improve insulin secretion and glycemia and prevent cardiovascular complications and heart failure and possibly other complications of diabetes

• Other anti-inflammatory treatments, alone or in combinations, may be more effective, depending on genetic predispositions, duration, and manifestation of the disease

• Anti-inflammatory treatments may improve simultaneously diabetes and associated inflammatory diseases such as rheumatoid arthritis, gout, or psoriasis

Over the past decades, major progress has been made in the understanding of the role of the immune system in the regulation of metabolism. It is now evident that inflammation contributes to the regulation of tissue adaptation to changes in metabolism in all situations, from physiology to pathology. Based on these pathogenetic findings, novel therapeutic opportunities arise, with the potential to palliate not only glycemia but also to prevent disease progression and beneficially target microvascular and macrovascular complications of diabetes. However, clinical translation is challenging, mainly due to the difficulties with implementing novel concepts. Therefore, the aim of this review is to generate interest in an emerging field and promote clinical translation. Accordingly, the main focus of this article is on clinically relevant aspects of metabolic inflammation.

Physiological Role of the Immune System in the Regulation of Metabolism

The endocrine system regulates the resorption, repartition, and metabolism of nutrients (commonly termed “metabolism”). The function of the immune system is to secure tissue integrity and repair. Although metabolism and immunity have distinct functions, both systems are required for maintaining and restoring homeostasis of an organism (1). Indeed, mounting an immune defense requires energy that needs to be directed to immune cells by endocrine signals. As an example, cortisol induces insulin resistance in muscle, liver, and fat, leading to increased circulating glucose concentrations that are consumed by immune cells during an infection. The availability of glucose will not only supply energy to immune cells, but also has a signaling function, delivering the message that sufficient energy and nutrients are available to mount a full immune response to a pathogen (2, 3). This intimacy between immunity and metabolism may explain why individuals are more prone to infection diseases in situations of a famine or food restriction. Indeed, it can be speculated that the limited calories available will then be used for basic vital functions, precluding the body to direct energy to the immune system.

Conversely, increasing evidence shows that the immune system contributes to the control of metabolism. Thereby, the NACHT, LRR, and PYD domain-containing protein 3 (NLRP3) inflammasome appears to be a sensor of metabolic changes, detecting increased glucose, fatty acid, and uric acid concentrations (4–9). The NLRP3 inflammasome is a multiprotein complex that consists of three subunits, the sensor molecules NLRP3 and apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) as well as the effector protease caspase-1 (also termed IL-1–converting enzyme). Upon stimulation, ASC self-associates and forms so-called ASC specks, which then allows the caspase-1 to become active in an autocatalytic manner. Active caspase-1 processes pro–IL-1β and pro–IL-18 into their mature active forms. In nonactivated myeloid cells, the cellular levels of ASC and caspase-1 are stable, but NLRP3 seems to be a limiting factor, suggesting a biphasic model, in which NLPR3 gene transcription needs to be induced first [(10), “inflammasome priming”] before an activating stimulus can trigger ASC and caspase-1 recruitment, which eventually leads to maturation of IL-1β and IL-18. Several etiological agents are known activators of the NLRP3 inflammasome, thereby driving metabolic disorders. As such, glucose, fatty acids, and islet amyloid polypeptide activate the inflammasome in type 2 diabetes, which leads to pancreatic islet inflammation and subsequent β-cell failure and destruction (11–13). Furthermore, depending on food intake and dietary fiber content, the microbiota may have either stimulating effects on NLRP3 via bacterial components or inhibitory effects via metabolites such as short-chain fatty acids (14–16). In gout, deposition of uric acid crystals in joints and periarticular tissues engages the NLRP3 inflammasome, leading to increased IL-1β and IL-18 (4). In atherosclerosis, cholesterol crystals in the artery wall activate the NLRP3 inflammasome, contributing to the inflammatory state in this disease (17). Rodent models that lack components of the inflammasome or downstream targets thereof were used to mimic human disease and identified the NLRP3 inflammasome as a required component in many of these entities. Clinical translation using NLRP3 inhibitors is currently in phase 1 to 2. Furthermore, there is also inflammasome-independent cytokine activation in metabolic diseases. For example, IL-1α, which does not require cleavage by the NLRP3 inflammasome to be active, mediates inflammasome-independent vascular inflammation and is released via free fatty acid–mediated uncoupling of mitochondrial respiration (18). Downstream of the inflammasome IL-1β, one of the first described cytokines (19), and the IL-1–dependent IL-6 and IL-33 emerge as important physiological modulators of insulin secretion and glucose disposal (20–22) (Fig. 1). Indeed, food ingestion increases the number of peritoneal macrophages, which are stimulated by the gut flora and glucose to increase the release of IL-1β in mice (20) and humans (23). Thereafter, IL-1β participates in the postprandial stimulation of insulin secretion (20, 24, 25) by activating the highly expressed IL-1 receptor type 1 on pancreatic β-cells (26). Further, insulin stimulates macrophage-derived pro–IL-1β maturation via the NLRP3 inflammasome (20) (Fig. 2). This occurs via enhanced phosphorylation of AKT downstream of the insulin receptor, upregulation of hexokinase 2, and increased glycolysis and reactive oxygen species. Elevated glucose metabolism promotes serine and glutathione production, which is essential for IL-1β mRNA expression (27). Finally, both insulin and IL-1β regulate glucose disposal, by which IL-1β promotes the major blood glucose regulatory hormone insulin and in addition directly stimulates glucose uptake into the immune cell compartment. These interactions between metabolism and innate immunity suggest that the postprandial increase of IL-1β contributes to the regulation of insulin secretion as well as glucose disposal. Thereby, IL-1β alerts and fuels the immune system, possibly to prevent the potential dissemination of microorganisms contained in the food (20) and to contribute to acute islet compensation (24).

Figure 1.

Immune regulation of insulin secretion in physiology. Various cytokines signal to the pancreatic β-cells and promote insulin secretion. Release of IL-6 from exercising muscle promotes the release of the incretin hormone GLP-1 from enteroendocrine l-cells and pancreatic α-cells to stimulate insulin secretion. IL-33 induces innate lymphoid cells 2 cells to release CSF2 and IL-13, which promote retinoic acid secretion from dendritic cells and macrophages. IL-22 derived from natural killer cells and innate lymphoid cells improve β-cell function. Acute IL-1β derived from macrophages upon feeding promote insulin secretion via IL-1 receptor 1 (IL-1R1) in the β-cells.

Figure 2.

Food intake regulation of IL-1β secretion. Food intake promotes IL-1β production in macrophages via stimulation of glucose metabolism and gut flora products. Glucose enters macrophages via Glut1 and is metabolized via the rate limiting enzyme hexokinase 2. Increased glycolysis and reactive oxygen species (ROS) production induce the processing of pro–IL-1β to mature IL-1β via the NLRP3 inflammasome. Enhanced glucose metabolism increases serine and glutathione production, which is required for IL-1β mRNA expression. The gut flora further promote IL-1β gene expression. Insulin reinforces the proinflammatory state of macrophages via the insulin receptor and stimulation of pAKT (protein kinase B). IL-1β induces the counterregulatory IL-1 receptor antagonist (IL-1Ra).

Another example of an adaptive cross talk between immunity and metabolism is the role of IL-6 in the regulation of GLP-1. Indeed, IL-6 can be seen as the first described myokine, because it is produced and released by the contracting skeletal muscle during exercise (28). Interestingly, IL-6 stimulates GLP-1 secretion from intestinal l-cells and reprograms pancreatic α-cells to process proglucagon to GLP-1 via upregulation of prohormone convertase 1/3 (21, 29). Hence, IL-6 mediates a metabolic cross talk among insulin-sensitive tissues, l-cells, and pancreatic islets to adapt to changes in insulin demand. This interaction between incretin hormones and cytokines is bidirectional, because the GLP-1 analog exenatide increases plasma concentration of the anti-inflammatory IL-1 receptor antagonist IL-1Ra in patients with type 2 diabetes (30). An additional aspect of the metabolic function of IL-6 is its role in the regulation of visceral adipose tissue. Hence, it has recently been shown that exercise-mediated loss of visceral adipose tissue mass requires IL-6 receptor signaling (31). Beyond these specific examples, IL-6 displays multiple complex metabolic effects that have been reviewed previously (32–34).

In line with the adaptive mission of the immune system, cytokines and immune cells often regulate metabolic adaptations in tissues in response to altered energy needs. Supporting this notion, IL-33, another member of the IL-1 family, also emerges as a positive regulator of insulin secretion (22). Indeed, IL-33 is expressed by islet mesenchymal cells and is enhanced by a diabetic milieu characterized by increased levels of glucose, IL-1β, and saturated fatty acids. IL-33 promotes β-cell function through islet-resident group 2 innate lymphoid cells by stimulating macrophages and dendritic cells via IL-13 and colony-stimulating factor 2 (22). These findings illustrate an additional cross talk between immune cells and endocrine cells that contributes to the maintenance of insulin secretion.

A particularly interesting cytokine is IL-22, a member of the IL-10 family, which plays a specific role in the regulation of metabolism. IL-22 contributes to the preservation of the gut mucosal barrier and thereby prevents endotoxemia with subsequent chronic inflammation and insulin resistance (35). Furthermore, IL-22 protects β-cells from oxidative and endoplasmic reticulum stress and improves glycemia in animal models of diabetes (36).

One of the newest members of the IL-1 family is IL-37 (37). Unlike other IL-1–related cytokines, IL-37 broadly suppresses innate immunity via binding to the IL-18 receptor and subsequent recruitment of the orphan decoy IL-1R8 to suppress proinflammatory cytokines. Circulating IL-37 concentrations are low in healthy humans but are induced locally and also systemically in several diseases such as rheumatoid arthritis (38), potentially as a counterregulatory mechanism to limit damage. However, other diseases are also associated with reduced levels of IL-37, suggesting that a failure to mount an IL-37 response could drive disease progression. Importantly for metabolic syndrome, IL-37 levels in adipose tissue positively correlated with insulin secretion and lower inflammatory status in humans and mice (39). Employing transgenic mice that express human IL-37 or therapy with recombinant IL-37 revealed a generally reduced innate immune cell response and a beneficial role for IL-37 in various disease models, such as systemic endotoxemia (40), myocardial infarction (41), colitis (42), and aging-induced damage (43). This supports a therapeutic potential for IL-37 in inflammatory and autoimmune diseases.

Finally, intra- and peri-islet macrophages may participate in the increase of β-cell mass during obesity by promoting β-cell proliferation (44). Thereby, M2 macrophages may promote β-cell proliferation by upregulation of SMAD7 (45, 46).

These selected examples and additional studies showing interactions between immunity and metabolic dysfunction (see later in text) initiated a new field of research called immunometabolism.

Pathological Role of Inflammation in the Development of the Metabolic Syndrome and Its Complication

Historically, preceding the description of the physiological and adaptive role of inflammation in metabolism, the focus has been on its pathological role. First evidence of an activation of the innate system in patients with type 2 diabetes came from the observation of elevated circulating levels of acute-phase proteins, including serum amyloid A, sialic acid, C-reactive protein (CRP), haptoglobin, fibrinogen, plasminogen activator inhibitor, TNF-α, IL-1β, IL-6, and IL-1Ra (47–50). Additionally, elevated levels of IL-1β, IL-1Ra, IL-6, and CRP are predictive markers for the development of type 2 diabetes (48, 50–54). Of note, increased CRP levels are also a strong predictor of cardiovascular disease (47–50). Importantly, these inflammatory proteins are regulated in an IL-1–dependent way, and IL-1 antagonism reduces the concentrations of CRP, IL-6, and leukocytosis (55–58).

The interpretation of levels of circulating inflammatory factor concentrations requires some caution. IL-1β is one of the most potent cytokines, and only a few molecules suffice to induce an answer in target cells (59). Therefore, even very low concentrations may have strong effects. Such low concentrations are often not detectable by standard protein assays, also because a majority of circulating IL-1β is bound to proteins. Furthermore, most cytokines are produced locally and act in a paracrine manner. Therefore, circulating levels of inflammatory biomarkers do not necessarily reflect the degree of inflammation in individual tissues. Small organs such as pancreatic islets are less likely to contribute to the circulating levels of inflammatory markers. In contrast, spillover of the relatively large adipose tissue and liver inflammation may disproportionately contribute to the circulating levels of inflammatory factors. Thus, to conclude on the real contribution of cytokines such as IL-1β, IL-6, or TNF-α to insulin secretion, insulin resistance, or secondary complications of diabetes, carefully designed clinical studies with specific antagonists and well-defined end points are warranted (56, 60).

The first report on the role of tissue inflammation in the pathogenesis of insulin resistance is the seminal observation by Hotamisligil et al. (61) of increased production of TNF-α by adipose tissues during obesity and improved insulin sensitivity upon TNF-α antagonism. The cellular source of TNF-α in adipose tissue was originally thought to be the adipocytes themselves. However, in the meantime, macrophages and other immune cells have been described in adipose tissues, which may account for the release of TNF-α and other cytokines, including IL-1β, IL-6, and IL-33 (62–64). As of today, it is well established that tissue inflammation plays a critical role in insulin resistance. Its pathogenesis has been extensively reviewed elsewhere (64–67).

Inflammation appears to play a considerable role not only in defective insulin action but also in insulin secretion. Indeed, pancreatic islets of patients with type 2 diabetes display increased expression of cytokines and immune cell infiltration mainly consisting of proinflammatory macrophages (68, 69). This chronic inflammatory process is associated with fibrosis and amyloid deposits, as observed in islets of the majority of patients with type 2 diabetes. Metabolic stress induces inflammation not only in insulin target tissues but also in islets, although the underlying mechanisms are not completely identical. Thereby, elevated glucose concentrations increase metabolic activity in islets, leading to the production of IL-1β (5, 68). This is enhanced by endoplasmic reticulum stress due to increased insulin production, bacterial products (endotoxins), and free fatty acids, which all contribute to inflammasome activation (70–72). Further, as insulin and the amyloid-precursor islet amyloid polypeptide are cosecreted from β-cells (73), increased insulin secretion may also increase amyloid deposition, which in turn also contributes to NLRP3 inflammasome activation in macrophages (11). Initially, this inflammatory process may be beneficial, as described above, promoting β-cell proliferation and insulin production to compensate for insulin resistance (24, 74, 75). IL-1β induces various cytokines and chemokines promoting macrophages and other immune cells, which eventually leads to deleterious inflammation (69, 76, 77). Indeed, prolonged metabolic stress will decrease the protective islet endogenous production of IL-1Ra (78, 79) and induce IL-1β autostimulation (26), and islet amyloid polypeptide will activate the inflammasome and thus promote the synthesis of IL-1β (11, 80, 81). Further, IL-1β, IL-6, and TNF-α were shown to promote β-cell dedifferentiation in cultured human and mouse islets, including IL-1β–mediated suppression of Foxo1, and TNF-α antagonism in vivo partly restored the loss of β-cell identity genes (82). This suggests that one deleterious mechanism of prolonged exposure to low-grade inflammation in islets could be due to β-cell dedifferentiation (83). The final pathway that will impair β-cell secretory function probably involves cytokines released by immune cells (84, 85), in which IL-1β appears to have a critical role in impairing β-cell proliferation (79).

Importantly, from a therapeutic point of view, long-term complications of diabetes also appear to be driven by an inflammatory process. This has been proven for IL-1β in cardiovascular complications, in which the initial preclinical hypothesis (86) has been confirmed with the CANTOS study [(87); see later in text for details]. Accumulating evidence also supports a role for inflammation in the manifestation of nonalcoholic steatohepatitis (88, 89), nephropathy (90), polyneuropathy (91–95), retinopathy, and macular edema (96, 97).

Finally, several inflammatory diseases are associated with diabetes, including rheumatoid arthritis, gout, psoriasis, and Crohn’s disease (98, 99). Moreover, other features of the metabolic syndrome are at least partly due to a pathological activation of the innate system and can be improved by IL-1 antagonism. These include low testosterone levels in men (99), increased cortisol levels (100), and fatigue (101, 102).

Reconciliation of Physiology and Pathology

The observed physiological role of immunity in glucose homeostasis appears to be at odds with its deleterious effect and the glucose-lowering effects of anti-inflammatory drugs (see later in text). This can be explained by the duration (acute vs chronic), the magnitude of the inflammatory effect, diminished anti-inflammatory mechanisms, and the concomitantly induced inflammatory diseases. One example for the dual role of proinflammatory cytokines in metabolism is IL-1β, which, on one hand, is an insulin secretagogue during feeding (20) and, on the other hand, contributes to β-cell failure (84, 85). Therapeutic inhibition of acute IL-1β may act as a brake for insulin secretion and allow β-cells to rest (103, 104) and recover from chronic overstimulation in states of metabolic stress. Indeed, when insulin secretion is reduced by genetic means, endoplasmic reticulum stress is alleviated and in turn, β-cell proliferation increases (105). Alternatively, excessive IL-1β stimulation could lead to resistance to IL-1β signaling, which may be corrected by blockage of IL-1. Indeed, islets from patients with type 2 diabetes fail to respond to IL-1β with insulin secretion, whereas islet from healthy donors are reactive to IL-1β (24). A further reason for the establishment of chronicity of inflammation may be a defective activation of counterregulatory pathways in affected organs. Islet β-cells, for example, locally express the protective IL-1Ra. Although there is an increase of IL-1Ra during obesity and type 2 diabetes in the circulation (50, 106), probably to counteract the systemic low-grade inflammation in vain, there is diminished IL-1Ra expression in human islets of patients with type 2 diabetes or in human islets exposed to toxic concentrations of glucose or human amyloid polypeptide toxicity in vitro (26, 78, 107). Further, mice lacking β-cell–specific IL-1Ra expression show impaired insulin secretion and altered islet morphology (79). Chronic exposure to elevated IL-1β along with simultaneous reduction of IL-1Ra may alter transcriptional programming, which regulate β-cell identity, proliferation, and apoptosis. In vivo constitutive deletion of the protective IL-1Ra in mouse islets indeed leads to reduced expression of proliferation genes along with diminished β-cell function (79).

Clinical Translation

Retrospectively, the first hint for a link between inflammation and metabolism was published in 1876, when it was reported that high doses of sodium salicylate improve glycosuria in patients with diabetes (108). This paper, published in the Berliner Klinische Wochenschrift, concludes that the aim of this publication was “to stimulate research testing whether salicylic acid should have a place in the treatment of diabetes.” Although it took >100 years, it is the merit of Yuan et al. (109), who have followed up on this message and described the underlying mechanisms, which involve inhibition of the nuclear factor κB pathway. The validity of this concept has been shown in animal studies and together with Goldfine et al. (110) in clinical trials with humans. Treatment with salsalate improves glycemia and reduces markers of systemic inflammation (110–112). This was confirmed in two multicenter, placebo-controlled studies (113, 114) and in smaller studies at early stages of the disease (115, 116). Salsalate is a widely used drug, for which safety has been confirmed in these studies. The only limitations are a small increase in low-density lipoprotein cholesterol levels and a reversible rise in urinary albumin. Thus, salsalate may be an effective and cheap drug to improve glycemic control. However, due to lack of a patent, salsalate is not a financially attractive drug to introduce into the market for the treatment of diabetes, and it is unclear whether a commercial company will invest the financial resources needed to make it available for patients. Despite these rather frustrating economic aspects, these important clinical studies should at least serve as supporting evidence for a role of inflammation in diabetes and encourage further studies in this direction.

Despite convincing preclinical studies showing that TNF-α induces insulin resistance in rodents (61), the clinical translation of this finding has not (yet) occurred, mainly due to superficial interpretation of the initial clinic studies. Indeed, as reviewed previously in detail (98), these studies have serious limitations that should not have led to firm conclusions. These include sample size and duration [10 and 7 patients for 4 weeks and 2 days, respectively, for the first two human trials (117, 118)]. Later studies were also statistically underpowered and of too short duration to detect meaningful effects (119, 120). However, other clinical trials using TNF-α antagonists in obese subjects without diabetes (121) and in patients with rheumatoid arthritis (122–126), psoriasis (127), and Crohn disease (128) showed improved glycemia. Importantly, treatment of patients with rheumatoid arthritis or psoriasis with TNF-α antagonists reduces the incidence of type 2 diabetes (129, 130). Altogether, this argues in favor of relevant beneficial metabolic effects of TNF-α antagonists in humans. However, well-designed clinical studies assessing its precise role are warranted.

Since the initial observations assigning a role for IL-1β in the pathogenesis of type 2 diabetes (68), numerous observations and clinical studies have demonstrated its involvement in glucose metabolism [reviewed in Donath (98)]. An initial proof-of-concept clinical study randomly assigned 70 patients with type 2 diabetes to receive either anakinra (recombinant human IL-1Ra) or placebo (56). Anakinra not only improved glycemia but also improved β-cell secretory function over the 3-month treatment period, thereby pointing to a disease-modifying potential. Beyond this direct metabolic benefit, IL-1 antagonism reduced CRP levels, providing a rational for the CANTOS study (see later in text). The ability to improve defective insulin secretion was confirmed in several follow-up studies using anakinra (131, 132). Although anakinra is well tolerated, it often causes injection site reactions and requires daily application. Much more attractive to treat metabolic diseases are anti–IL-1β antibodies, with a half-life that allows injection once every 3 months. Each of these antibodies demonstrated beneficial effects in patients with type 2 diabetes, albeit the magnitude of the effects varied depending on baseline HbA_1c levels and sample size (133–136).

As of today, the most informative study that investigates the role of inflammation in patients with a metabolic disease is the CANTOS study (87). This large cardiovascular outcome study involved >10,000 patients and demonstrated that treatment with canakinumab, an anti–IL-1β antibody, resulted in a lower number of cardiovascular events than placebo. A subanalyis focused on metabolic end points (137). However, it should be emphasized that the CANTOS study was not specifically devoted to diabetes; thus, diabetes was not a selection criterion, and in case of diabetes, lifestyle interventions and antidiabetic drugs were freely adapted. Furthermore, even patients with diabetes had an excellent glucose control with baseline HbA_1c levels within or even below treatment target (see later in the text). Keeping these limitations in mind, the following can be learned from this study: (1) by selecting patients solely on the basis of elevated CRP levels (≥2 mg/L) and a history of myocardial infarction, the resulting study population consisted of 90% individuals with impaired glucose metabolism (40% diabetes and 50% prediabetes) and other features of the metabolic syndrome. This strongly argues for a link between inflammation and diabetes. (2) IL-1β antagonism significantly decreased HbA_1c during the first 6 to 9 months of treatment with attenuation of the effect in the course of the study. This confirms findings of previous clinical studies mentioned previously and proves the role of IL-1β and inflammation in diabetes. The reason for the attenuation of the effect after 6 to 9 months is likely due to the design of the study, which allowed lifestyle interventions and adaptations of standard antidiabetic therapies. Indeed, in other cardiovascular outcome studies using classical antidiabetic drugs [dipeptidyl peptidase 4 (DPP4) or sodium-glucose–linked transporter 2 (SGLT2) inhibitors, the most widely used antidiabetic drugs today], similar patterns and magnitude of effects were observed (138). The magnitude of achievable effects depends on baseline HbA_1c. Baseline HbA_1c was 7.1%, which is far below the HbA_1c level targeted by diabetes treatment of this patient population (<8%, American Diabetes Association guidelines). However, during the first months, the pure (without changes in antidiabetic drugs) anti–IL-1 effect was apparent in subjects with diabetes who had a higher baseline HbA_1c of 7.5% (which is still quite low). Upon anti–IL-1β treatment, this group showed a statistically significant decreased HbA_1c. In patients without diabetes, canakinumab decreased HbA_1c for the whole duration of the study, albeit the effect was very mild, which is expected as patients were normoglycemic. (3) IL-1 antagonism prevented new onset of diabetes for 4 years. After this time, the number of patients, whom were further followed in the study, decreased by ∼50%, and the effect of canakinumab was not detected anymore. The reason for this loss of efficacy remains to be explained but could be related to the previously discussed patient selection or to the physiological role of IL-1β. Of note, none of the most widely used antidiabetic drugs (DPP4 inhibitor, SGLT2 inhibitor, insulin, and sulfonylurea) show diabetes prevention properties in contrast to the 4-year prevention by IL-1β blockade.

Following the CANTOS study, a meta-analysis of all 2921 reported cases with type 2 diabetes undergoing an IL-1 therapy (anakinra or anti–IL-1β antibody) has been performed (139). It demonstrated a highly substantial reduction in HbA_1c (P < 0.00001), confirming a previous meta-analysis involving fewer patients (140). As mentioned previously, the magnitude of achievable decrease in HbA_1c depends on baseline HbA_1c. Because most patients had a relatively low HbA_1c, no firm conclusion on the magnitude of the effect can be drawn. Interestingly, a meta-analysis also showed that anti–IL-1 therapy improves the level of C-peptide following meal intake (140).

The CANTOS study revealed several additional aspects of long-term IL-1β antagonism. Keeping in mind the multiple effects of IL-1β, the safety profile of canakinumab was surprisingly good. Nevertheless, canakinumab was associated with a higher incidence of fatal infections, which warrants caution in patients at risk. This has to be balanced with the additional favorable effects, including a decrease in cancer mortality and incident in lung cancer (141) and the confirmation of beneficial effects in arthritis, gout, and osteoarthritis.

Keeping in mind the unexpected protective effect of SGLT2 inhibitors on heart failure, a striking finding of the CANTOS study is that canakinumab reduced hospitalization for heart failure and the composite of hospitalization for heart failure or heart failure–related mortality (142). This was observed in particular in patients with higher body mass index, diabetes, hypertension, or prior coronary bypass surgery.

Concluding Remarks and Future Directions

Activation of the innate immune system is apparent at all stages of the development of diabetes and its complications. This includes impaired β-cell function (143), insulin resistance (64–67), cardiovascular diseases (87), heart failure (142), nonalcoholic steatohepatitis (88, 89), nephropathy (90), polyneuropathy (95–99), fatigue (105, 106), retinopathy, and macular edema (100, 101). Accordingly, IL-1β antagonism has been shown to prevent β-cell dysfunction and to improve glycemia (56, 78, 133–136), cardiovascular complications (87, 144), and heart failure (142), and it may counteract other complications of diabetes (Table 1) (4, 145, 146). Clearly, more studies are warranted to assess more precisely the magnitude of these effects. Furthermore, beyond IL-1β antagonism, other immunomodulatory drugs, either alone or in combination, may be more effective. However, the choice of the anti-inflammatory drugs to be tested should be based on precise pathophysiologic understanding of their role and not on an unspecific effect as with methotrexate, which does not decrease IL-1β and failed to prevent cardiovascular complications in patients with metabolic syndrome (147).

Table 1.

Benefits of Anti–IL-1β Treatment in Patients With Metabolic Syndrome

β-Cell function ↑	(56)
Glycemia ↓	(56, 131–136, 139)
Cardiovascular complications ↓	(87)
Heart failure ↓	(142)
Gout ↓	(4, 145, 146)
Arthritis ↓	(136)
Safe: no hypoglycemia (cave: severe infections)
Convenient (subcutaneous injection every 3 mo)
Possible additional effects: treatment of nephropathy, retinopathy, nonalcoholic steatohepatitis, and polyneuropathy

Until now, the only antidiabetic drugs demonstrating improvement in cardiovascular complications were GLP-1 analogs (148) and SGLT2 inhibitors in the EMPAREG study (149). With a similar design to the EMPAREG study (149), the CANTOS study showed that blocking IL-1β in patients with a metabolic syndrome has comparable effects, including prevention of cardiovascular diseases (87, 137) and heart failure (142). This is in contrast to the antidiabetic drugs insulin, sulfonylurea, and DPP4 inhibitors, which failed to show such preventive effects. This difference in outcomes can be explained by revisiting our understanding of the pathophysiologic role of insulin resistance, decreased insulin secretion, and glycosuria in the development of type 2 diabetes and its complication (Fig. 3). We suggest that once hyperglycemia due to overnutrition prevails, these three key features of type 2 diabetes are secondary protective mechanisms by preventing overloading of tissues with cell nutrients. If these mechanisms are overridden by increasing insulin production or action, damaging inflammation occurs, contributing to complications of diabetes. It follows that therapies of diabetes should prevent overfeeding (GLP-1 analogs), promote glycosuria (SGLT2 inhibitors), induce β-cell rest (IL-1β antagonists), and prevent inflammatory damages (IL-1β antagonists). In contrast, drugs increasing insulin action should be limited to patients with an absolute insulin deficiency. Beyond this glucocentric view, tissue inflammation in patients with metabolic syndrome is not only induced by glucose but also by cholesterol (17), fatty acids (71), uric acids (4), and bacterial products (150, 151).

Figure 3.

Development and treatment of complications of type 2 diabetes. Hyperglycemia and dyslipidemia due to overnutrition and genetic predisposition overload tissues with cell nutrients with subsequent tissue damage mediated by inflammatory processes. Overfeeding can be prevented with GLP-1 analogs, hyperglycemia can be reduced via glycosuria using SGLT2 inhibitors, and tissue damage can be alleviated via IL-1β antagonists. In contrast, increasing insulin signaling may enhance nutrient overload in tissues with subsequent induction of a damaging inflammation.

Pathological activation of the immune system plays a critical role in an increasing number of diseases (152), and some of them are associated with diabetes, such as rheumatoid arthritis, gout, and psoriasis. For all of these conditions, immunomodulatory treatments are approved and have substantially improved patient care. Somehow, the metabolic/endocrine field seems more reluctant to implement such anti-inflammatory treatments, despite increasing data supporting the beneficial role of these interventions. In particular, in cases of concomitant diseases, diabetes together with rheumatoid arthritis, gout, or psoriasis, a judicious choice of an anti-inflammatory treatment can improve both conditions with one drug. A convincing example is the recent study by Ruscitti et al. (136), showing that treatment of patients with diabetes and rheumatoid arthritis with an IL-1 antagonist decreased HbA_1c by >1% for 6 months and simultaneously reduced rheumatoid disease activity.

We hope that this review will encourage clinician scientists and industrial partners to translate the large knowledge accumulated with animal and proof-of-concept clinical studies into the treatment of patients with a metabolic syndrome. Thereby, anti-inflammatory treatments, alone or in combination, have the potential to improve diabetes, its progression, and complications, as well as associated inflammatory diseases.

Glossary

Abbreviations:

ASC

apoptosis-associated speck-like protein containing a caspase recruitment domain

CRP

C-reactive protein

DPP4

dipeptidyl peptidase 4

NLRP3

NACHT

LRR

and PYD domain-containing protein 3

SGLT2

sodium-glucose–linked transporter 2