Therapeutic Effects of Endogenous Incretin Hormones and Exogenous Incretin-Based Medications in Sepsis

Abstract

Background

Sepsis, a complex disorder characterized by a dysregulated immune response to an inciting infection, affects over one million Americans annually. Dysglycemia during sepsis hospitalization confers increased risk of organ dysfunction and death, and novel targets for the treatment of sepsis and maintenance of glucose homeostasis are needed. Incretin hormones are secreted by enteroendocrine cells in response to enteral nutrients and potentiate insulin release from pancreatic β cells in a glucose-dependent manner, thereby reducing the risk of insulin-induced hypoglycemia. Incretin hormones also reduce systemic inflammation in preclinical studies, but studies of incretins in the setting of sepsis are limited.

Methods

In this bench-to-bedside mini-review, we detail the evidence to support incretin hormones as a therapeutic target in patients with sepsis. We performed a PubMed search using the medical subject headings “incretins,” “glucagon-like peptide-1,” “gastric inhibitory peptide,” “inflammation,” and “sepsis.”

Results

Incretin-based therapies decrease immune cell activation, inhibit proinflammatory cytokine release, and reduce organ dysfunction and mortality in preclinical models of sepsis. Several small clinical trials in critically ill patients have suggested potential benefit in glycemic control using exogenous incretin infusions, but these studies had limited power and were performed in mixed populations. Further clinical studies examining incretins specifically in septic populations are needed.

Conclusions

Targeting the incretin hormone axis in sepsis may provide a means of not only promoting euglycemia in sepsis but also attenuating the proinflammatory response and improving clinical outcomes.

This manuscript provides a bench-to-bedside review of the evidence supporting a potential therapeutic role for endogenous incretin hormones and exogenous incretin-derived medications in sepsis.

Sepsis is a life-threatening illness characterized by overwhelming systemic inflammation in response to an infectious insult (1, 2). Organ dysfunction in sepsis results from a combination of direct injury to host tissues from the invading pathogen and subsequent release of proinflammatory cytokines, including IL-1β, IL-6, and TNF-α (3, 4). No sepsis-specific therapies have proven successful to date (5, 6); rather, the treatments demonstrating the greatest benefit have been timely administrative of appropriate antimicrobial agents and supportive care (7, 8). Novel approaches to minimize tissue injury and organ dysfunction in the acute phase of sepsis are needed.

Dysglycemia in sepsis

Hyperglycemia is common in patients with sepsis, both with and without diabetes, with several contributing factors (9–12). Proinflammatory cytokines contribute to the development of peripheral insulin resistance and hyperglycemia in sepsis by inhibiting the expression and membrane translocation of glucose transporter type 4 in skeletal muscle and adipose tissues and impairing signaling through the insulin-receptor substrate-1 and protein kinase B pathways (13–15). The acute effects of proinflammatory cytokines on pancreatic insulin release are unclear, with some studies demonstrating augmentation of insulin secretion and others demonstrating suppression (16–19). Regardless, prolonged exposure to cytokines is injurious and triggers apoptosis in pancreatic β cells (20–22). Endogenous glucocorticoids, released in the physiologic response to stress, increase blood glucose levels by stimulating gluconeogenesis and glycogenolysis (23). Additionally, vasopressors and exogenous glucocorticoids, typically initiated to treat sepsis-induced hypotension and dysregulation of the hypothalamic–pituitary adrenal axis, decrease insulin sensitivity and worsen glycemic control in septic animal models and human studies (24–26). Thus, the factors contributing to hyperglycemia in sepsis are complex.

Once present, hyperglycemia may further impair the host immune response and potentially increase the risk for further organ injury (27, 28). High glucose concentrations promote proinflammatory cytokine release and adversely affect several types of innate immune cells (29, 30). For example, hyperglycemia decreases granule release in neutrophils, reduces neutrophil extracellular trap formation, and prevents effective chemotaxis resulting in reduced capacity to combat invading pathogens (31). Additionally, blood glucose can complex with bacterial components and complement proteins, thereby preventing effective opsonization and phagocytosis (32, 33). Thus, the development of hyperglycemia has direct pathologic implications in sepsis.

Dysglycemia in sepsis does not comprise only hyperglycemia. Hypoglycemia is an independent predictor of increased in-hospital and 1-year mortality in critically ill patients with sepsis and often occurs secondary to exogenous insulin administration (34–36). Acute hypoglycemia in healthy individuals (both with and without diabetes mellitus) has been shown to increase oxidative stress, platelet aggregation, proinflammatory cytokine production, and vascular adhesion molecule expression (37–39). Additionally, glucose variability is of prognostic importance because increased glycemic lability in patients with sepsis is associated with higher in-hospital mortality (40). Maintenance of euglycemia remains the goal in hospitalized patients with sepsis but has proved difficult to achieve.

Clinical trials of glycemic control in patients with sepsis

The adverse impact of dysglycemia prompted several large, randomized, controlled clinical studies in critically ill patients. A landmark study published by van den Berghe et al. (41) in 2001 demonstrated a substantial mortality benefit with the use of exogenous intravenous intensive insulin therapy to maintain tight glycemic control (blood glucose maintained between 80 and 110 mg/dL) in a mixed population of mostly surgical critically ill patients. Subsequent studies of tight glycemic control, however, yielded variable and potentially contradictory results (42–45). A meta-analysis of 12 randomized controlled trials specifically in patients with sepsis determined that, compared with conventional glucose management, intensive insulin therapy did not reduce mortality or length of intensive care unit (ICU) stay in patients with sepsis but did increase incidence of hypoglycemia (46). Thus, an ideal strategy of glycemic control in sepsis that prevents hyperglycemia while minimizing the risk of hypoglycemia remains elusive (and has not been accomplished with approaches that rely on exogenous insulin alone) but can potentially be achieved through modulation of the incretin axis.

The Incretin Hormone Axis

Physiology of incretin hormones

Incretin hormones are intestine-derived peptides released primarily in response to enteral nutrients and promote insulin secretion from pancreatic β cells in a glucose-dependent manner. Glucose-dependent insulinotropic peptide (GIP) and glucagon-like peptide-1 (GLP-1) are the best characterized incretin hormones to date. GIP is released in response to enteral administration of nutrients (including carbohydrates and fatty acids) by enteroendocrine cells (K cells) located mostly in the proximal small intestine (47, 48). GIP is rapidly degraded by the protease dipeptidyl-peptidase 4 (DPP-4, also known as CD26) with a half-life of 7 to 8 minutes (49). The receptor for GIP is a seven-transmembrane G-protein coupled receptor that increases cAMP and intracellular calcium in response to ligand binding. In pancreatic β cells, GIP receptor signaling induces insulin exocytosis (47, 50, 51). Additionally, GIP promotes glucagon secretion under conditions of hypoglycemia (52). Thus, under physiologic conditions, GIP is important in the maintenance of euglycemia.

Although the incretin hormones GIP and GLP-1 are similar in their ability to promote insulin secretion in a glucose-dependent manner, there are some notable differences. GLP-1 is derived from the preproglucagon peptide encoded by the GCG gene (which also encodes glucagon) and is released from enteroendocrine cells (L cells) located predominantly in the distal small bowel and colon (48, 53). GLP-1 is similarly degraded rapidly by DPP-4, but with a shorter half-life compared with GIP of only 2 to 5 minutes (54, 55). Whereas GIP release is stimulated by direct interaction of enteroendocrine cells with enteral nutrients, release of GLP-1 following a meal is more complex. An initial peak of GLP-1 (approximately 10 to 15 minutes postprandial) is likely promoted by gastric distention or vagal stimulation, and a later peak (approximately 30 to 60 minutes postprandial) is likely promoted by direct interaction with enteral nutrients (47, 56). GLP-1 promotes glucose-dependent insulin secretion from pancreatic β cells through binding of a G-protein–coupled receptor that results in increased insulin gene expression, improved mRNA stability, and increased insulin exocytosis. Furthermore, GLP-1 suppresses secretion of glucagon from pancreatic α cells under conditions of euglycemia and hyperglycemia (47, 57, 58). Extrapancreatic effects demonstrated by GLP-1 agonists include the promotion of satiety and the promotion of weight loss (59–61). Uncovering roles of increased incretin activity in extrapancreatic tissues is an area of ongoing research.

Incretin release in response to inflammatory stimuli

Although GIP and GLP-1 are primarily secreted in response to enteral nutrients, recent studies have suggested that incretins are also released in response to inflammatory stimuli. Lebrun et al. (62) demonstrated that GLP-1 is released from L cells in response to the intestinal insults dextran sodium sulfate and ischemia reperfusion. Similarly, Ellingsgaard et al. (16) demonstrated that IL-6 stimulates GLP-1 production in immortalized enteroendocrine cells in a dose-dependent manner mediated by increases in Pcsk1 (encoding the prohormone convertase 1/3, which processes preproglucagon to GLP-1) and Gcg gene expression. Systemic IL-6 administration to mice in the Ellingsgaard study increased circulating GLP-1 (16). Subsequent studies in mice confirmed the ability of not only IL-6 but also of IL-1β and of the bacterial cell wall component lipopolysaccharide (LPS) to increase GLP-1 levels (17, 63). LPS has also been shown to increase GIP through an IL-1β–dependent mechanism (64). GLP-1 promotes local mucosal healing; thus, release of incretins following an insult may serve to limit intestinal injury under conditions of systemic inflammation (65).

Observational cohort studies have similarly demonstrated associations between circulating GLP-1 levels and systemic inflammation (17, 66, 67). In a cohort study examining critically ill patients (of whom 66% presented with sepsis), GLP-1 levels were positively correlated with inflammatory markers, including c-reactive protein and procalcitonin, as well as markers of organ dysfunction, including cystatin C and coagulation factors. Interestingly, increased endogenous GLP-1 levels in this study predicted ICU mortality (67). However, the elevated GLP-1 levels should not be interpreted to indicate that incretins are harmful in sepsis; rather, the findings suggest that GLP-1 may have a role in countering the adverse effects of systemic inflammation. The anti-inflammatory mediators IL-1 receptor antagonist and the soluble receptor for TNF-α are similarly released in response to overwhelming inflammation and are also positively correlated with organ failure and mortality in septic populations despite not being innately harmful (68, 69). Furthermore, findings from preclinical and clinical studies of incretin-based therapies in sepsis demonstrate benefit as opposed to harm.

Therapeutic Effects of Incretin-Based Medications

Current incretin-based medications

Two main categories of incretin-based therapies are currently in use for the management of diabetes mellitus: incretin mimetics and DPP-4 inhibitors. Incretin mimetics, including exenatide, liraglutide, and semaglutide, are structurally similar to GLP-1 but have been modified to resist degradation. DPP-4 inhibitors, including sitagliptin and linagliptin, increase incretin levels by blocking protease degradation by DPP-4, thus increasing endogenous levels of GIP and GLP-1 (59, 70, 71). Incretin mimetics and DPP-4 inhibitors improve glycemic control in patients with diabetes mellitus type 2, and recent clinical trials have highlighted a cardioprotective effect with incretin mimetic therapy (72–74).

Therapeutic effects of incretin mimetics on activated immune cells

Incretin mimetics demonstrate anti-inflammatory effects in preclinical studies, and several studies report the detection of incretin receptors in different immune cells, including macrophages, monocytes, and B- and T-cell lymphocytes (75–77). Yusta et al. (78) demonstrated that exendin-4, a GLP-1 receptor agonist, reduced proinflammatory cytokine expression in activated intestinal intraepithelial lymphocytes. In the same study, exendin-4 attenuated injury in a dextran sodium sulfate model of colitis, whereas GLP-1R knockout mice sustained more severe intestinal injury compared with wild-type controls. Restoration of GLP-1 signaling in lymphocytes in GLP-1R knockout chimera mice rescued intestinal integrity. In studies of cultured macrophages, treatment with exendin-4 attenuated the inflammatory response to LPS through a protein kinase A–cAMP-dependent reduction in nuclear translocation of NF-κB and consequent decrease in IL-6, IL-1β, and TNF-α production (79, 80). Exendin-4 also increases macrophage production of IL-10 and arginase (79), suggesting GLP-1R stimulation can influence macrophage polarization shifting from the M1 classically activated proinflammatory to the M2 anti-inflammatory phenotype (81). Thus, incretin mimetic therapies demonstrate the ability to directly modify the activity of immune cells.

Therapeutic effects of incretin-based therapies in nonimmune cells and tissues

Anti-inflammatory effects of incretin-based therapies have been demonstrated in several nonimmune cell and tissue types (Fig. 1). For example, in human aortic or vascular endothelial cells, treatment with liraglutide decreases expression of vascular adhesion molecules and reduces adhesion of monocytes induced by LPS (82) or TNF-α (83). Additional studies in human vascular endothelial cells exposed to LPS confirm that incretin mimetics decrease vascular permeability and reduce proinflammatory cytokine levels (84, 85). In cardiomyoblasts, treatment with liraglutide in a two-hit TNF-α and hypoxia model decreases NLRP3 inflammasome activation and is associated with improvements in cell viability (86). Incretin infusions with either GIP or GLP-1 in mouse models significantly reduce vascular inflammation and decrease macrophage infiltration into the endothelium, consistent with findings that chronic use of incretin-based therapies attenuate atherosclerotic plaque development (87–91).

Figure 1.

Extra-pancreatic effects of incretin hormones released in response to enteral nutrients or proinflammatory stimuli, endogenous incretin hormones, and incretin-based therapies demonstrate anti-inflammatory and protective effects across multiple organ systems.

Anti-inflammatory effects of incretin-based therapies are also evident in adipose tissues (92, 93), pancreatic islets (94–96), hepatocytes (97, 98), and renal glomerular cells (99–101) in preclinical studies. Some studies suggest GIP may worsen inflammation from adipose tissues in obese individuals (102, 103). However, recent preclinical studies with novel GIP analogs suggest an anti-inflammatory effect in adipose tissue (104), comparable to studies with DPP-4 inhibitors that raise levels of both GIP and GLP-1 (94, 105). Interestingly, GIP and GLP-1 analogs are being explored as therapeutic agents to reduce neuroinflammation in murine models of Alzheimer’s disease and Parkinson disease, with findings of decreased cytokine production, inhibition of microglial cell activation, improved neuronal insulin sensitivity, and improved cognitive performance (106–109).

Effects of incretin-based therapies on systemic inflammation

The anti-inflammatory effects of incretin-based therapies have been confirmed in clinical studies. An acute 3-hour infusion of exogenous GLP-1 reduced circulating IL-6 levels in a small study of obese individuals with diabetes mellitus type 2; a trend toward reduced inflammation with GIP infusion was also observed (110). Chronic incretin mimetic use in patients with diabetes decreases circulating proinflammatory cytokines, including IL-6, TNF-α, IL-1β, and monocyte chemoattractant protein-1 after 8 to 12 weeks on therapy, with persistent anti-inflammatory effects reported 6 to 12 months after initiating treatment (111–113). Chronic DPP-4 use similarly decreases proinflammatory cytokines and vascular adhesion modules and increases anti-inflammatory mediators, including IL-10 (114, 115). Given that sepsis is characterized by overwhelming systemic inflammation and multiorgan dysfunction, the consistent preclinical and clinical anti-inflammatory effects of incretins support further investigation of the incretin hormone axis in sepsis.

Incretin-based therapies in preclinical models of sepsis

Several studies have examined the use of incretin-based therapies in mouse models of sepsis. In an endotoxemic rat study by Yanay et al. (116), the GLP-1 analog exendin-4 decreased circulating IL-6, TNF-α, IL-1β, and interferon-γ levels and prevented LPS-induced hypoglycemia. A subsequent study by Steven et al. (117) demonstrated that both the incretin mimetic liraglutide and the DPP-4 inhibitor linagliptin decreased mortality, improved hypotension, suppressed inflammation, and reduced oxidative stress in a severe endotoxemic rat model. In a separate study in endotoxemic mice, linagliptin and liraglutide reduced oxidative stress, suppressed inflammatory mediator production, and prevented thrombocytopenia and microvascular thrombosis in the pulmonary vasculature (118). Protection against LPS-induced thrombocytopenia was lost in mice deficient for the GLP-1 receptor, highlighting the importance of intact incretin signaling. Interestingly, infusion of exogenous GLP-1 after intravenous LPS exposure reduces permeability in mesenteric vasculature, highlighting additional protective effects against intestinal injury in settings of systemic inflammation (119). Importantly, incretin-based therapies demonstrate favorable effects not only in sterile models of inflammation but also in preclinical bacteremic models.

Lee et al. (84) demonstrated benefit with incretin mimetics in a cecal-ligation and puncture model of sepsis, whereby high-dose exendin-4 administered twice daily starting 16 hours after septic insult improved survival, associated with decreases in the proinflammatory protein high mobility group box protein-1. In a separate study, a long-lasting polyethylene-glycol–conjugated exendin-4 molecule reduced lung injury; lowered IL-6, IL-1β, TNF-α, and monocyte chemoattractant protein-1 levels; decreased peritoneal leukocyte counts; and reduced mortality after cecal-ligation and puncture (85). Thus, exogenous incretin-based therapies consistently improve outcomes in preclinical septic models, and we hypothesize that physiologic activation of the incretin hormone pathway may induce similar beneficial effects.

Recently, our laboratory demonstrated that infusion of low-level enteral dextrose in a murine endotoxemia model increased glucose disposal, insulin secretion, and insulin sensitivity; improved mean arterial blood pressure; and decreased systemic inflammation in a GIP-dependent fashion (120). When GIP signaling was inhibited with pharmacologic blockade, the beneficial effects of enteral dextrose were lost, and infusion of exogenous GIP improved glucose disposal and reduced IL-6 levels. Our study demonstrated that activation of endogenous incretin hormones in the acute phase of sepsis may be achieved through provision of enteral nutrients and that therapeutic benefits of incretins may be realized in the absence of exogenous administration (120). Taken together, the improved outcomes with endogenous incretin stimulation and with exogenous incretin-based therapy support the notion that increased incretin activity is of potential clinical benefit to patients presenting with sepsis.

Clinical Studies of Incretin-Based Therapies in Critically Ill Patients

Incretin-based therapies have not been tested in clinical trials of sepsis. However, several small clinical trials have been conducted testing the effect of exogenous incretins in the broader context of critical illness (121, 122). In cardiac and in mixed medical and surgical ICU populations, exogenous GLP-1 is effective in reducing blood glucose and glycemic variability compared with placebo without an increased risk of hypoglycemia (123–129). Infusion of the GLP-1 analog exenatide may similarly improve glycemic control but has been associated with gastrointestinal side effects (130). Despite being well tolerated, infusions of exogenous GIP have not demonstrated benefit in reducing blood sugar in critically ill patients, but in these studies the effects of GIP were tested in the fasted state, where glucose-stimulated insulin secretion may be blunted, and in response to a liquid meal bolus, which stimulates endogenous incretin hormone production and may limit the host’s ability to respond to additional exogenous GIP (131, 132). Existing clinical trials were performed mostly in mixed ICU populations with few patients with sepsis enrolled (Table 1). Further clinical trials are needed to better characterize the incretin hormone axis during sepsis and to determine a potential therapeutic role. A prospective single-center, double-blind, placebo-controlled randomized clinical trial from our laboratory is currently recruiting patients to define the effects of a low-level enteral dextrose infusion on inflammation and on endogenous incretin hormone production in critically ill patients with sepsis (ClinicalTrials.gov no. NCT03454087).

Table 1.

Clinical Trials of Exogenous Incretion Therapies in Critically Ill Populations

Study	Year	Design	Population	Number of Participants	Participants With Sepsis, n (%)	Treatment	Primary Results	Adverse Effects
Meier et al. (133)	2004	Double-blind, placebo-controlled, randomized crossover study	Postsurgical patients with diabetes	8	Not reported	GLP-1 infusion	Reduced blood glucose, increased endogenous insulin secretion	None
Sokos et al. (124)	2007	Double-blind, placebo-controlled, randomized clinical trial	Cardiac patients undergoing coronary artery bypass grafting	20	Not reported	GLP-1 infusion	Reduced blood glucose, insulin requirements, and cardiac arrhythmias	Hypoglycemia in 1 patient in the GLP-1 group and 2 patients in the placebo group
Deane et al. (125)	2009	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients without diabetes	7	Not reported	GLP-1 infusion	Reduced blood glucose	None
Deane et al. (126)	2010	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients without diabetes	25	4 (16%)	GLP-1 infusion	Reduced blood glucose and small intestine glucose uptake	Hyperglycemia in 1 patient in the GLP-1 group and 2 in the placebo group
Deane et al. (127)	2011	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients with diabetes	11	5 (45%)	GLP-1 infusion	Reduced in blood glucose	None
Abuannadi et al. (130)	2012	Single-center feasibility study; all patients received intervention	Cardiac patients in the ICU	40	Not reported	Exenatide infusion	Blood glucose control similar to historical controls receiving insulin infusion protocols targeting blood glucose 100–140 mg/dL	Hyperglycemia in 3 patients, hypoglycemia in 4, nausea in 16, emesis in two
Lee et al. (132)	2013	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients without diabetes	20	9 (45%)	GIP infusion in patients already receiving GLP-1 infusion	No additional effect on blood glucose	None
Galiatsatos et al. (128)	2014	Double-blind, placebo-controlled, randomized clinical trial	Surgical and burn patients in the ICU	18	Not reported	GLP-1 infusion	Reduced blood glucose and glycemic variability	Hypoglycemia in 1 patient in the GLP-1 group and 3 in the placebo group
Kar et al. (131)	2015	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients without diabetes	20	2 (10%)	GIP infusion	No effect on blood glucose or gastric emptying	None
Miller et al. (129)	2017	Double-blind, placebo-controlled, randomized crossover study	Mixed medical-surgical patients without diabetes	12	1 (8%)	GLP-1 infusion	Reduced blood glucose and small intestine glucose uptake	None

Conclusion

Incretin-based therapies have revolutionized the chronic management of patients with diabetes by providing a means to improve glycemic control, mitigate the risk of hypoglycemia, increase weight loss, improve insulin sensitivity, and reduce the risk of cardiovascular disease. Therapies targeting increased incretin activity may provide similar benefit in the care of critically ill patients with sepsis by promoting the maintenance of euglycemia and modulating the host inflammatory response. Preclinical studies of incretin-based therapies demonstrate anti-inflammatory effects in various tissues, including vascular endothelium, pancreatic β cells, and hepatocytes, and improved mortality in sterile inflammation and infectious animal models of sepsis. Initial clinical studies of exogenous incretins in acutely ill patients demonstrate that GIP and GLP-1 infusions are well tolerated in mixed critically ill populations and may improve glycemic control without the risk of hypoglycemia. Additional clinical studies are needed to explore the role of incretin-based therapies as a novel target to optimize glycemic control, reduce systemic inflammation, and improve outcomes in patients with sepsis.

Glossary

Abbreviations:

DPP-4

dipeptidyl-peptidase 4

GIP

glucose-dependent insulinotropic peptide

GLP-1

glucagon-like peptide-1

ICU

intensive care unit

LPS

lipopolysaccharide