Effect of exenatide use on cognitive and affective functioning in obese patients with type 2 diabetes mellitus: exenatide use mediates depressive scores through increased perceived stress levels

Abstract

Purpose/Background Glucagon-like peptide-1 (GLP-1) is a molecule used to treat type 2 diabetes mellitus (T2DM). Given their widespread expression in the nervous system, GLP-1 receptors also play a role in regulating mood and cognitive function. Here, we aimed to compare obese patients with T2DM, with or without exenatide (a GLP-1R agonist) use on cognitive and affective functioning.

Methods/Procedures A total of 43 patients with T2DM (23 on exenatide and 20 without exenatide) were evaluated with the Snaith-Hamilton Pleasure Scale, Cognitive Failures Questionnaire, Patient Health Questionnaire-9 (PHQ-9), Generalized Anxiety Disorder-7, Childhood Trauma Questionnaire, Perceived Stress Scale (PSS), and Chronic Stress Scale, in addition to laboratory-based measures of reward learning (the probabilistic reward task) and working memory (Letter-N-Back task).

Findings/Results Patients on exenatide had higher body mass index (BMI) (37.88 ± 5.44 vs 35.29 ± 6.30; P = 0.015), PHQ-9 (9.70 ± 4.92 vs 6.70 ± 4.66; P = 0.026), and PSS (29.39 ± 6.70 vs 23.35 ± 7.69; P = 0.015) scores. Other stress scales (Childhood Trauma Questionnaire and Chronic Stress Scale), Generalized Anxiety Disorder-7 scores, response bias, or discriminability as assessed by probabilistic reward task and self-report (Cognitive Failures Questionnaire) and laboratory-based (Letter-N-Back) cognitive measures were not significantly different between groups (both Ps > 0.05). Multivariate linear regression analyses adding BMI and PSS as covariates revealed that although BMI had no effect (P = 0.5), PSS significantly predicted PHQ-9 scores (P = 0.004). Mediation analysis showed that exenatide users reported higher PSS, with greater PSS associated with higher PHQ-9 levels (b = 0.236). There was no evidence on exenatide directly influencing PHQ-9 independent of PSS (c′ = 1.573; P = 0.305; 95% bootstrap confidence interval, −1.487 to 4.634).

Implications/Conclusions Based on previous research and our findings, exenatide use might be mediating depression scores through disrupting stress responses.

Glucagon-like peptide-1 (GLP-1) is a molecule used to treat type 2 diabetes mellitus (T2DM) due to its blood glucose–lowering effect, increasing satiety and weight loss. Originally, it is secreted from the distal ileum and colon cells, and it is also known to cross the blood barrier. It can bind to vagal afferents and enter the brain through the GLP-1 receptors (GLP-1R) on the nucleus of the solitary tract (NTS) and arcuate nucleus of the hypothalamus. It can also be secreted by neurons in the NTS1 and microglia and binds to GLP-1R that are widely expressed in regions including the hippocampus, neocortex, hypothalamus, and cerebellum.1

Based on a recent systematic review, GLP-1 analogs have a neuroprotective effect.2 On a cellular level, GLP-1 reduces apoptosis and restores neuronal growth and increases cell viability by enhancing cyclic adenosine monophosphate, protein kinase A (PKA), and cAMP response element binding protein (CREB) levels and altering glycogen synthase kinase 3 (GSK-3) Akt, Erk, and mammalian target of rapamycin (mTOR) phosphorylation levels, which results in changes of other downstream modulations important for cell survival. GLP-1 also promotes neurogenesis by stimulating neurotrophic factors such as glial cell line-derived neurotrophic factor (GDNF), vascular endothelial growth factor (VEGF), and brain-derived neurotrophic factor (BDNF) and reduces inflammation by decreasing tumor necrosis factor-a, interleukin-6 (IL-6), and IL-10 along with reducing microglia activation.2 GLP-1 administration also induces glutamate release and enhances hippocampal cell firing.2 The neuroprotective effect of GLP-1 results in improved cognitive assessments and avoidance of neuronal loss induced by neurotoxic compounds.2 Exendin-4 treatment improves spatial learning in the Morris water maze and associative learning in the radial maze paradigm; GLP-1R antagonist treatment reverses these effects.3 On the other hand, GLP-1R knockout mice have deficits in memory formation as they are unable to distinguish between familiar and novel objects in a novel object recognition task along with problems in spatial learning demonstrated in the Morris water maze.4 GLP-1R knockout mice also show impairments in contextual fear learning, and overexpression of GLP-1 in the hippocampus rescues these impairments and improves performance in the Morris water maze.5

In addition to its neuroprotective and cognitive restoring effects, GLP-1 is studied for its impact on reward pathways and stress-related systems. GLP-1R are abundant in the ventral tegmental area (VTA) and nucleus accumbens (NAc), regions of the mesolimbic reward pathway, and receive projections from the NTS.6 In addition, GLP-1's role in dopamine signaling has been implicated in the caudal and rostral lateral septum, where GLP-1R were located as opposed to dopamine terminals.7 Another study shows that GLP-1R antagonist has the ability to reduce lithium chloride–induced suppression of NAc dopamine neurons, indicating that this action of lithium chloride on phasic dopamine release is mediated by GLP-1R.8

GLP-1R are also located in areas crucial for emotion regulation, such as the amygdala, dorsal raphe, and the hippocampus.9 Research on rodents shows that acute administration of GLP-1 significantly induces anxiety-like behaviors, whereas chronic administration of GLP-1 causes an antidepressant effect.9 However, research on GLP-1 and mood are less clear, and the results are conflicting. A clinical study found that compared with healthy controls, the postmortem brains of patients with a diagnosis of major depressive disorder have reduced GLP-1R expression in the dorsolateral prefrontal cortex and hippocampus, even after adjusting for age, sex, ethnicity, treatments, substance use, and body mass index (BMI).10 Furthermore, 2 GLP-1 polymorphisms, A allele in rs10305492 and C allele in rs1042044, showed associations with response bias in a probabilistic reward task (PRT), a laboratory-based measurement of anhedonia, when controlled for age, sex, diagnosis, and discriminability.11 However, both polymorphisms failed to show associations with depressive disorder diagnosis. None of the GLP-1 polymorphisms were associated with the Snaith-Hamilton Pleasure Scale (SHAPS), a subjective measurement of anhedonia.

From a different perspective, GLP-1 also affects “satiety” in drugs of abuse. Systematic, intra-VTA, and NAc injections of GLP-1 significantly attenuate cocaine seeking in cocaine-experienced rats even at small doses.12 Besides cocaine, GLP-1R agonist use decreases reward-related abnormalities in multiple drugs of abuse such as alcohol,13,14 nicotine,15 and amphetamine.16,17 In the mesolimbic system, GLP-1R also specifically influences the control of palatable food intake.18

Despite the accumulating preclinical evidence of GLP-1R agonists' effects on cognitive and affective functioning, GLP-1R agonists' association with mood and cognition in patients with T2DM has not been thoroughly investigated, especially for the short term. Currently, clinical studies are limited, and it is hard to draw comprehensive conclusions on the subject. Therefore, new research focusing on the nature of GLP-1 on cognition, mood, and reward in humans is needed to estimate the psychiatric effects of GLP-1R agonists on obese patients with T2DM and answer if it can be an effective therapy for various psychiatric disorders. Based on this gap in the literature, we aimed to compare obese patients with T2DM diagnosis, with or without GLP-1R agonist treatment, in relation to psychiatric symptoms, as well as subjective and laboratory-based measurements of reward and anhedonia.