Incretin Agonism: Sustainable Efficacy or Surreptitious Hazard?

2. Role of Incretins in the Neurohormonal Axis of Appetite Control

Homeostatic feeding is a mechanism where energy for basic metabolic processes and survival will be obtained, while hedonic feeding is driven by sensory perception or pleasure. [58] Homeostatic feeding is tightly controlled by many molecules, hormones, and neuronal elements. These include sensing nutrients in the central nervous system (CNS), integrating afferent stimuli, reflecting the energy balance, and adjusting subsequent food intake. [59] The incretin effect is largely mediated by neuroendocrine actions and is correlated with the size of the meal. [38]

In the neuroendocrine control of food intake, the brainstem and hypothalamus are the core CNS areas because they receive, convey, and integrate peripheral signals. The area postrema (AP) and nucleus tractus solitarius (NTS) in the brainstem convey the peripheral signals, consisting of nutrients, hormones, and vagal afferent inputs, to the arcuate nucleus (ARC) of the hypothalamus. The ARC contains both orexigenic and anorexigenic neurons. The former expresses neuropeptide Y (NPY) and agouti-related peptides (AgRP) and the latter expresses pro-opiomelanocortin (POMC) and cocaine / amphetamine-related transcript (CART). They collectively process the received information and regulate eating and attain energy homeostasis. [60,61]

POMC cells activate melanocortin 4 receptor (MC4R) expressing neurons in the paraventricular nucleus of the hypothalamus (PVH) and other brain regions, thereby inhibiting food intake and increasing energy expenditure. The MC4R gene is involved in the brain's regulation of appetite and weight. Conversely, NPY/ AgRP neurons antagonize these effects. [62]

Also, leptin and serotonin are involved in regulating energy balance, appetite, and bone mass. Although both leptin and serotonin depolarize POMC neurons [63,64], there is a distinct selectivity in the responsive neurons. Namely, serotonin-responsive POMC neurons are not activated by leptin. Also, these two groups of neurons are anatomically segregated: leptin-activated POMC cells are located more laterally in the ARC than the serotonin-responsive cells. [64] Serotonin modulates the endogenous release of both agonists and antagonists of the melanocortin receptors, which are a core component of the CNS circuitry controlling body weight homeostasis. It should be noted that non-homeostatic or hedonic feeding can override this homeostatic pathway and result in overeating and obesity. Therefore, preventing hedonic feeding by food choices may be beneficial for weight homeostasis.

A sophisticated murine study reported the presence of GABAergic neurons in the dorsal vagal complex as a new player in the governing feeding behavior. [7] However, it was reported that GABAergic neurons, do not appear to express AgRP and reduce inhibitory tone to postsynaptic POMC neurons.[65] Vong’s study actually challenges the recent sophisticated studies in mice and insinuates that these studies might be flawed. The authors of murine models have shown that GABAergic neurons inhibit NPY [7] or food intake. [6] Moreover, the effects of GABAergic signals are not limited to NPY/ AgRP inhibition. GABAergic signals also inhibit (POMC) neurons during fasting.[61]

Both anorexigenic neurons as well as orexigenic neurons could be inhibited, or stimulated, depending on the energy status of the host.[61] Therefore, a multitude of feeding possibilities exists even without accounting for the brain region involved. The probability of non-exclusive event involving multiple molecules with 2 possible functions (inhibited or stimulated) is impossible to estimate in animal studies. The key point is that GLP-1 and GIP receptor agonists pharmacologically manipulate the neural signals and consequently, obese individuals lose appetite as well as body weight.

Peripheral injection of fluorescently labeled liraglutide revealed the presence of the drug in the circum-ventricular organs. [66] In this study, murine brain slices showed that GLP-1 directly stimulates POMC/CART neurons and indirectly inhibits neurons expressing neuropeptide Y (NPY) and agouti-related peptide (AgRP) via GABA-dependent signaling. [66] The labeled liraglutide was internalized in the neurons expressing proopiomelanocortin (POMC) and cocaine- and amphetamine-regulated transcript (CART).[66]

Beck also reported that the physiological activation of neuropeptides is dependent on energy availability of the host.[61,67] When energy is deprived or restricted, NPY is activated and when energy availability returns to normal, NPY synthesis returns to baseline level. Also, NPY metabolism is regulated by diet type, especially carbohydrate and fat content.[67] More recent findings support the view that the hypophagia was mediated by GLP-1 receptors in the brain.[68]

Toda summarized these facts as “arcuate melanocortin neurons consist of two distinct neuronal populations: (POMC)-expressing neurons and NPY/AgRP)-expressing neurons.”[62] However, it should be noted that POMC cells are activated by energy surfeits (excess) signals and inhibited by energy deficits.[61] In other words, the energy excess or deficits are the drivers of POMC cells. So, if the person’s energy state is the driver of POMC cell activation, then the energy state of humans should be controlled to activate or inhibit the POMC cells. We need to pause here to remind ourselves, that in a string of biologic actions, the earliest event is the cause. Therefore, food intake and resultant energy state of the host are the cause of POMC cell activation. Our view is supported by Beck and Rau who stated that “POMC cell stimulation or inhibition is the consequence of food intake and the resultant energy state.”[61,67]

The pharmacological incretin agonism manipulates these neural axes, and control the person’s appetite. However, once these neural inhibitions are lifted, weight regain can occur.[24,48] After discontinuation of tirzepatide, the study participants regained almost 2/3 of their lost weight within 4 weeks.[48] We will suggest a strategy to mitigate the weight regain after discontinuation of incretin agonists in the FUTURE DIRECTIONS section. These mechanisms and neurohormonal pathways are illustrated in Figure 1.

3. Challenges in Incretin Agonism

a. 

Involvement of β-Arrestins

As stated in section1-a, incretin and glucagon receptors (GCGR) belong to the class B GPCR and their efficacy is limited by β-arrestins. [69] β-arrestin-1 (β1arrs) and β-arrestin-2 (Β2arrs) are best known for their ability to mediate the desensitization and internalization of GPCRs. [70] The effect of B2arrs on insulin secretion was dominated by GLP-1R- and no clear GCGR-dependent effects are observed.[69]

Β2arrs are ubiquitously expressed and function as negative regulators of GPCR signaling by inhibiting GPCR and G protein coupling, via uncoupling cyclic AMP (cAMP)/protein kinase A (PKA) signaling at physiological levels of GLP-1.[71] This is a process called desensitization which dampens insulin secretion. Next, β2-arres scaffold enzymes, phosphodiesterase and diacylglycerol kinase, degrade second messengers generated by G protein activity. This will solidify the desensitization process.[72]

At pharmacological doses, the activation of extracellular signal-related kinase (ERK)/cAMP-responsive element-binding protein (CREB) requires Β2arrs. Also, GIP-dependent insulin secretion needs Β2arrs in human islets.[71]

β2arres involvement is a potential drawback of incretin agonism because it will cause plateau in the drug efficacies. Avoiding Β2arrs-dependent GPCR desensitization may alleviate the problem of tachyphylaxis of GLP-1 and GIP agonism.[73] The recent development of Tirzepatide is based on the previous study that revealed islets β-arrestin1 limits the insulin response of GLP-1, but not GIP, thus may not affect insulin secretion by GIP. [74]

b. 

β-Cell Exhaustion and Failure

Continuing stimulation of pancreatic β-cells may cause receptor downregulation and desensitization. A good example of this phenomenon is the insulin resistance in T2D patients.[75] Forcing the β-cells to secrete insulin when they are overwhelmed by insulin demand may accelerate β-cell failure as shown with sulfonylurea administration.[76] Simply put, chronic agonism eventually reduces the number of receptors leading to a condition similar to failure of function.[77] [78]

The mechanism of β-cell failure is presumed that overstimulation of β-cell increases cytoplasmic Ca⁺⁺ levels, and persistent elevation of cytoplasmic Ca⁺⁺ may trigger apoptosis of β-cells.[79] Under chronic agonism, some β-cells fail to differentiate properly and lose their identity.[80]

While these examples describe overstimulation of β-cells by hyperglycemia in T2D, chronic agonism by incretins may exert similar effects. [76,81,82] Indeed, animal studies reported increased pyknotic nuclei (a sign of apoptosis) of β-cells and acinar inflammation in rats given a high dose of exenatide.[83]

4. Potential Adverse Effects of Incretin Agonism

Although newer incretins appear to be much safer, increased risk of biliary disease,[12] pancreatitis,[84] bowel obstruction,[85] and gastroparesis have been noted. [86] Also an increased risk of pancreatic or thyroid cancers has been reported. [87,88] However, conflicting non-significant results on cancer risk were also observed. [89] But this study has very short follow-up (3.9 years) and compared GLP-1ra to dipeptidyl peptidase 4 inhibitors (DPP4i) which is on the causal pathway of GLP-1ra and thus, is an inappropriate comparator.

It is plausible that insulin secretion modulated by incretin agonism may affect pancreatitis and pancreatic cancer, because of their shared pathway. Indeed increased risks for multiple pathologies were reported.[90] Sodhi et al. [90] using a large health claims data set reported that the use of GLP-1ra compared with bupropion/naltrexone was associated with a 9-fold increased risk of pancreatitis, a 4.2-fold increase of bowel obstruction, and a 3.7- fold increase in gastroparesis. [90]

The cause of pancreatitis associated with GLP-1ra is hypothesized as incretin agonism may promote pancreatic hyperplasia, leading to increased pancreatic weight and exocrine duct occlusion. [91] However, the incidence of pancreatitis resulting from the incretin agonism is relatively rare and only large data sets such as pharmacovigilance data or insurance claims data will show significant results.

Using large historical population data, Dankner et al. reported no increase in GLP-1ra-associated pancreatic cancer incidence compared to insulin and metformin.[92] However, this study is fraught with selection bias.[92] Insulin is a well-known oncogene and this will generate selection bias. Choosing a worse comparator generate an aura of erroneously making GLP-1ras apper innocuous. [92] Further well-conducted studies are needed.

Other adverse effects such as retinopathy,[93] gastroparesis and bowel obstruction[85] were also reported. Lu and colleagues [94] explained the mechanism of intestinal obstruction as “long-term application of GLP-1ras may also elevate the release of endogenous GLP-2. GLP-2 is a cell-specific growth hormone regulating the growth of the small intestine, colonic villi and crypts, increasing villi height, and reducing antral motility.”. [94] This study provides the scientific basis for the intestinal blockage and raises concerns for the long-term use of incretin agonism. [94]

Recently several cases of pancreatitis and one fatality associated with Tirzepatide [95] have been reported. As more people use this class of medications, more serious adverse events may emerge. Especially, for tirzepatide with added GIP which masks the nausea and vomiting,[37] and allows the patients to tolerate the medication better and to take it for a longer duration. This may increase more serious adverse events.

Additionally, a case of leukocytoclastic vasculitis (LCV) induced by once-weekly subcutaneous semaglutide has also been reported. [96] Although 50% of LCV can be idiopathic, immune dysregulation is presumed cause. Especially, complete resolution of the lesions shortly after the discontinuation of semaglutide proves the causal role of semaglutide on LCV. [96]

Also, GLP-1R and GCGR agonism may increase heart rate (HR) [54,57,97,98] and sympathetic tone.[99,100] Both glucagon[101] and GLP-1[102] are known to increase sympathetic tone.[103,104] Although some studies reported no increase in HR , animal studies suggest that GLP-1 acts directly on the sinus node.[55,56]

The mechanism that glucagon receptor (GCGR) activation leads to increased HR is via adenylyl cyclase activation and increases in 3',5'-c- AMP production in the myocardium. [54] For this reason, glucagon is used as an antidote for β-blocker overdose.[101] It is conceivable that glucagon is associated with sympathetic tone because hypoglycemia activates stress response and the hypothalamic-pituitary-adrenal axis. [105,106,107,108]

The mechanism that GLP-1R contributes to increased HR is reasoned that the preproglucagon neurons that mediate GLP-1-associated anorexia are situated in the brainstem to receive signals of stress. This suggests a potential link with the sympathetic nervous system. [102,103] However, another study reported that the initial increase in sympathetic tone was reversed by GLP-1R activation via by acetylcholine and nitric oxide.[109] Making it more confusing, in a human trial, the addition of liraglutide to exercise appeared to blunt the beneficial effects of exercise on left ventricular diastolic function. [52] Many systematic reviews do suggest cardiac protective effects of incretins. [110,111] However, the absolute benefit size was only a 1.5% decrease. [112] Also, a meta-analysis of the CVD benefits of dulaglutide showed that all individual studies include relative risk of 1 (no effect). This means that the results are not statistically significant. [113]

5. Future Directions

How do we harness the observed beneficial effects of incretin agonism on glucose levels and weight loss minimizing potential harm? These are the wishes of experts and patients alike. They all raised the concern for potential adverse events from a long-term use of incretin agonism, sustainability of their efficacy, and the high cost. [114,115,116] We concur with the experts and suggest that the remarkable weight loss and glycemic control by incretin agonism should be juxtaposed to the high cost, it’s unsustainability of efficacy, and potential adverse effects. As we stated in the neuro-hormonal control of appetite section, manipulating neuronal axis will fail when the manipulation terminates. Thus, tapering GLP-1 agonism and enhancing lifestyle modification may be the key to healthy weight maintenance. It is pivotal to teach the patients so that they can make healthier food choices that maintain satiety [117], encouraging physical activities. A group in Denmark had done this by slowly tapering semaglutide and applying intense lifestyle modification in a trial.[118] More longer-term studies are needed emphasizing the lifestyle modifications in addition to tapering GLP-1ras.