Possible Brain-Protecting Role of GLP-2 in Alzheimer's Disease Compared to GLP-1

Dementia syndromes are among the most significant public health concerns due to population aging. According to a study by Nichols et al. (2022), approximately 57.4 million people worldwide were living with dementia in 2019, with this number projected to rise to 152.8 million by 2050 [1]. Alzheimer’s disease (AD), an age-related neurodegenerative disorder, is estimated to account for 60–70% of all dementia cases and is the fifth leading cause of death among individuals over the age of 65 [2].

AD is characterized by two main pathological hallmarks: the accumulation of extracellular β-amyloid (Aβ) plaques and intracellular neurofibrillary tangles (NFTs), which are composed of hyperphosphorylated tau protein. These changes are strongly associated with synaptic and neuronal loss, dystrophic neurites, and gliosis, involving alterations in the morphology and function of microglia and astrocytes. Studies have shown that microglial overactivation in response to amyloid plaques increases the release of pro-inflammatory cytokines, which can damage neurons [3]. Oxidative stress, resulting from accumulating reactive oxygen species (ROS), is another key pathological feature of AD. Additionally, mitochondrial dysfunction leads to reduced ATP production, calcium dysregulation, and increased ROS generation. This dysfunction further exacerbates oxidative stress, creating a vicious cycle that accelerates neurodegeneration. Research has highlighted the role of oxidative stress and mitochondrial dysfunction in AD progression, as ROS contribute to Aβ plaque and tau tangle formation, ultimately leading to neuronal cell death [4,5]. Vascular changes also play an important role in AD pathogenesis. Studies have shown that cerebrovascular disease and AD often coexist, with vascular risk factors such as hypertension and diabetes accelerating disease progression [6].

All these pathological mechanisms contribute to synaptic degeneration, leading to progressive neurodegeneration and cognitive decline.

It is essential to highlight that hyperinsulinemia and insulin resistance (IR), key pathophysiological mechanisms of type 2 diabetes mellitus (T2DM), have been shown to impact cognitive function. IR, defined as reduced sensitivity to insulin, has been consistently identified as an essential risk factor for neurodegeneration and cognitive impairment [7,8]. AD is also linked to impaired insulin signaling in the brain, and T2DM patients have an increased risk of developing AD [9,10]. Therefore, ongoing research explores the influence of carbohydrate and insulin metabolism disorders on dementia-related processes.

Proglucagon is a precursor protein that gives rise to several biologically active peptides, some of which play crucial roles in regulating glucose homeostasis, insulin secretion, and appetite control. The proglucagon family comprises various peptides derived from the proglucagon protein, the most well-known of which are glucagon and glucagon-like peptides (GLP-1 and GLP-2). The proglucagon gene is expressed in multiple tissues, notably the pancreas, intestine, and brain, producing peptides with diverse physiological effects [11].

Notably, patients with type 2 diabetes treated with semaglutide had a lower risk of developing AD compared to those receiving other diabetes therapies [12]. Additionally, a recently published study comprising obese individuals revealed that GLP-1 medications not only aid in weight loss but also reduce the risk of AD and other neurological disorders [13]. These studies provide valuable insights into the potential neuroprotective benefits of GLP-1 medications in the context of AD. However, much less is known about GLP-2 and its possible role in dementia and AD.

GLP-2, secreted mainly through intestinal L cells and preproglucagonergic (PPG) neurons in the brainstem and hypothalamus, is primarily known for regulating energy homeostasis by influencing intestinal motility and nutrient absorption [14]. Its role in the central nervous system (CNS) remains less explored; however, emerging evidence suggests that GLP-2 may act as a neuroprotective agent and play a key role in neurophysiology [15].

The GLP-2 receptor (GLP-2R) has been localized not only in the gut but also in the CNS, including the hypothalamus, hippocampus, and cortex. Initial in vitro studies suggest that GLP-2R activation protects neurons from cytotoxic damage [16], reduces cell death, and functions as a trophic factor for astrocytes by promoting their proliferation [17]. Notably, GLP-2R activation in neurons stimulates de novo protein synthesis and is involved in critical processes, including cell growth, proliferation, survival, angiogenesis, and protein production [17]. Therefore, GLP-2-induced protein synthesis may be crucial for maintaining neuronal survival and facilitating the action of secondary mediators, including growth factors and neuropeptides.

Moreover, studies using knockout rat models have shown that the absence of GLP-2 receptor stimulation is associated with spatial cognitive dysfunction following chronic cerebral hypoperfusion (CCH), while increased GLP-2R expression prevents CCH-induced cognitive impairment. Furthermore, enhanced GLP-2R expression has been linked to improved long-term memory and reduced depressive symptoms, both of which are commonly observed in AD [16].

Despite significant advances in understanding AD mechanisms, current diagnostic capabilities remain insufficient. While precise methods to diagnose exist, including neuropsychological tests (e.g., the Mini-Mental State Examination [MMSE] and the Addenbrooke’s Cognitive Examination), neuroimaging techniques (positron emission tomography [PET] and magnetic resonance imaging [MRI]), and cerebrospinal fluid (CSF) protein analysis, their application is often limited due to invasiveness, cost, and time constraints [18,19]. PET imaging is expensive and available only at specialized centers, making it unlikely to be widely adopted for routine assessment of cognitive impairment [19]. Although CSF sampling is becoming increasingly common in clinical practice, lumbar punctures remain a highly invasive procedure [20].

Over recent decades, numerous biomarkers have been identified in body fluids, but only a few plasma biomarkers have demonstrated significant diagnostic relevance for AD [21]. Consequently, there is a growing need for an effective, simple, and affordable AD screening tool suitable for outpatient settings. Blood biomarkers, in particular, offer a relatively non-invasive diagnostic approach with minimal associated risks.

To our knowledge, data on GLP-2 levels in humans with AD are currently lacking. Therefore, we aimed to determine plasma GLP-2 concentrations and assess whether GLP-2 could serve as a novel, potential, non-invasive biomarker for sporadic Alzheimer’s disease.

Glucagon-like peptide-2 is primarily recognized for its role in promoting intestinal mucosal growth and enhancing nutrient absorption [22]. However, accumulating evidence suggests that GLP-2 may also exert significant neuroprotective effects. Several experimental studies have investigated its protective mechanisms in the context of neurodegenerative disorders [15,23]. To date, no published data address circulating GLP-2 concentrations in individuals with Alzheimer’s disease.

In the present study, plasma GLP-2 concentrations were significantly higher in patients with Alzheimer’s disease than in cognitively unimpaired controls (2.767 ± 1.488 vs. 1.941 ± 1.058 ng/mL; p = 0.0148). When stratified by disease severity, both mild and moderate-to-severe AD subgroups showed higher GLP-2 levels than controls (p = 0.0250 and p = 0.0489, respectively), with no significant difference between the two AD subgroups. The observed effect size was moderate (Cohen’s d ≈ 0.60), although the achieved post hoc power was approximately 68%, suggesting limited ability to detect effects smaller than moderate magnitude. Importantly, GLP-2 levels were positively associated with BMI, body weight, insulin concentrations, and HOMA-IR, indicating that circulating GLP-2 may be more closely linked to metabolic status than to clinical dementia severity.

The “type 3 diabetes” concept has been introduced to emphasize mechanistic convergence between type 2 diabetes mellitus (T2DM) and Alzheimer’s disease, in which brain insulin resistance and related metabolic stressors contribute to neurodegenerative signaling. Consistent with this framework, AD brains show impaired insulin/IGF-1 signaling accompanied by increased inhibitory IRS-1 (insulin receptor substrate 1) serine phosphorylation, changes reported even in the absence of overt diabetes and compatible with the notion of “brain insulin resistance” [24]. At the intracellular level, attenuated IRS-1/PI3K (phosphoinositide 3-kinase) signaling reduces Aktactivation, shifting the balance away from pro-survival and synaptotrophic programs [24,25]. Reduced Akt activity may, in turn, alter the inhibitory control of GSK-3β (glycogen synthase kinase 3 beta), thereby favoring tau phosphorylation and facilitating the emergence of neurofibrillary pathology, with secondary effects on microtubule stability, axonal transport, and synaptic integrity [24,25].

In parallel, insulin resistance may disrupt neuronal energy homeostasis by impairing GLUT-3/4 (glucose transporter type 3/4) trafficking and glucose utilization, thereby aggravating mitochondrial dysfunction and reactive oxygen species generation. These metabolic deficits can compromise ATP-dependent proteostasis and amyloid-β clearance, facilitating amyloid accumulation and synaptotoxicity. In addition, carbonyl stress and advanced glycation end-products (AGEs) may activate RAGE (receptor for advanced glycation end-products) and trigger NF-κB (nuclear factor kappa B)–dependent inflammatory signaling, reinforcing a feed-forward loop between oxidative stress and neuroinflammation. The AGE–RAGE axis also engages MAPK (mitogen-activated protein kinase) pathways (including ERK1/2 (extracellular signal-regulated kinase 1/2) and stress-activated signaling (JNK/STAT (c-Jun N-terminal kinase/signal transducer and activator of transcription), converging on transcriptional regulators such as NF-κB, AP-1 (activator protein 1), HIF-1α (hypoxia-inducible factor 1-alpha), and TGF-β (transforming growth factor beta). Through these mechanisms, inflammation can further impair IRS-1/PI3K/Akt signaling, creating a self-amplifying network linking insulin signaling failure with neuroinflammatory and amyloidogenic processes [26].

Collectively, this mechanistic architecture provides a biologically plausible route by which systemic and cerebral insulin resistance may lower the threshold for synaptic failure and accelerate neuronal loss in AD [24,25].

We failed to find any significant differences in insulin concentration between AD individuals and their non-demented counterparts. Based on the literature, plasma insulin findings in AD are contradictory, but several reports support a clinically relevant signal. In an ADNI-based study across the AD spectrum, higher plasma insulin levels were associated with worse cognitive performance and greater brain atrophy (particularly in AD-vulnerable regions), suggesting that hyperinsulinemia/IR may contribute to neurodegeneration [27]. Overall, available evidence supports the view that insulin abnormalities can be present in AD. Still, the direction and magnitude of peripheral changes in circulating insulin likely depend on disease stage, comorbidities (T2D/obesity), and other factors [28]. Genetic evidence may further support a role for metabolic factors: in a 2024 Mendelian randomization analysis, genetically predicted fasting insulin (but not diabetes status per se) was associated with AD dementia risk [29]. Beyond insulin levels alone, higher peripheral IR (e.g., HOMA-IR) has been associated with lower regional cerebral glucose metabolism on FDG-PET in late middle-aged adults at risk for AD, aligning systemic metabolic dysfunction with AD-vulnerable brain hypometabolism [30].

It should be noted that there is also a mechanistic bridge between the IR and incretin pathways in AD. In detail, GLP-1 receptor agonists improve systemic insulin sensitivity and have direct CNS effects in experimental models (affecting anti-inflammatory signaling, mitochondrial and synaptic support, and potentially impacting protein homeostasis) [25].

GLP-2 was not correlated with MMSE or CDR scores. Furthermore, in multivariable regression models that included age and clinical group, neither variable independently predicted GLP-2 concentrations (both p > 0.05), highlighting the potential influence of metabolic and other unmeasured factors. Overall, elevated GLP-2 levels in AD may reflect a compensatory or stress-related response; however, the clinical significance of this finding requires confirmation in larger and metabolically phenotyped cohorts.

Longitudinal assessments of cognitive performance, using the Mini-Mental State Examination and the Clinical Dementia Rating scale, confirmed disease progression across the study period. Importantly, no significant associations were observed between plasma GLP-2 levels and MMSE or CDR scores, suggesting that changes in GLP-2 concentrations were not associated with cognitive decline in our cohort.

The relatively small sample size, particularly at the study’s conclusion, represents a potential limitation that may have impacted the statistical power of our analyses. Nevertheless, given the exploratory nature of this investigation, these findings should be interpreted as preliminary. Future studies involving larger and more diverse populations are warranted. In addition, subsequent research should include individuals with mild cognitive impairment (MCI), an intermediate clinical stage between age-related cognitive changes and overt dementia, particularly AD. MCI is characterized by deficits in one or more cognitive domains, including but not limited to memory [31]. Approximately 50% of individuals with MCI progress to AD within five years, although MCI may also represent a prodromal phase of other dementias [32]. Therefore, we plan to continue the study with a larger cohort of patients with Alzheimer’s disease, as well as to expand it to include individuals with mild cognitive impairment.

Although clinical data regarding the role of GLP-2 in dementia, including AD, remain scarce, preclinical evidence suggests that GLP-2 and its analogues may confer neuroprotection through multiple mechanisms. These mechanisms encompass modulation of neurogenesis, oxidative stress, inflammation, mitochondrial function, and synaptic plasticity [33,34,35]. Experimental studies have demonstrated that GLP-2 receptors (GLP-2R) are expressed in several brain regions, albeit with a more restricted and less well-characterized distribution compared to GLP-1 receptors. GLP-2R has been identified in the hippocampus (CA1 (Cornu Ammonis 1), CA3 Cornu Ammonis 1), dentate gyrus), cerebral cortex, and brainstem (nucleus of the solitary tract) [34,36,37]. Notably, the hippocampus—particularly regions implicated in memory consolidation, learning, and adult neurogenesis—appears to be a significant site of GLP-2 activity. GLP-2 may facilitate neural progenitor proliferation (↑ Ki67, DCX (doublecortin)), upregulate neurotrophic factors such as BDNF (brain-derived neurotrophic factor), and support long-term potentiation (LTP) and long-term depression (LTD) [33]. GLP-2R expression has also been detected in neurons, astrocytes, microglia, and neural progenitor cells, suggesting a potential role in modulating inflammatory activation and promoting neurogenesis [15]. While GLP-2 demonstrates limited penetration across the blood–brain barrier (BBB), it may exert central effects indirectly via peripheral pathways, including vagal afferents and gut–brain axis signaling, particularly through the nucleus of the solitary tract [38].

Multiple mechanisms underlying the neuroprotective actions of GLP-2 have been proposed. GLP-2 appears to attenuate microglial and astrocytic overactivation, thereby reducing the release of pro-inflammatory cytokines, including IL-1β, TNF-α, and IL-6 [34,39]. Furthermore, GLP-2 enhances antioxidant defenses by upregulating enzymes such as superoxide dismutase (SOD) and catalase, thereby decreasing the accumulation of reactive oxygen species and mitigating oxidative damage [15,34,39]. In addition, GLP-2 has been shown to preserve mitochondrial function, counteracting mitochondrial dysfunction, autophagy impairments, and apoptotic processes, thereby preventing neuronal energy failure and cell death [35]. Experimental evidence further indicates that GLP-2 reduces neuronal apoptosis in both in vitro and in vivo models, suppressing caspase-3 activation and cytochrome c release while enhancing anti-apoptotic pathways [40]. GLP-2-mediated neuroprotection also involves the activation of signaling pathways critical for neuronal survival, including the PI3K/Akt and MAPK/ERK pathways, which are implicated in synaptic plasticity and memory processes [16,41,42]. Indeed, GLP-2R activation in the CNS stimulates the PI3K/Akt/mTOR pathway, facilitating protein synthesis, dendritic growth, synaptogenesis, and neurogenesis [43]. Upregulation of neurogenesis markers, such as Ki-67, DCX, and NeuN (neuronal nuclei), further supports this mechanism [33].

A summary of the potential neuroprotective effects of GLP-2 is presented in Figure 4.

Glucagon-like peptide-2 has been shown to have neuroprotective effects in several preclinical models of neurodegeneration. In rodents fed a high-fat diet, GLP-2 administration significantly reduced glial fibrillary acidic protein (GFAP) expression, a marker of reactive gliosis, and attenuated neurodegenerative changes (including reductions in NF-κB, pro-inflammatory cytokines, amyloid precursor protein, oxidative stress markers, and neuronal apoptosis) [34]. Additionally, GLP-2 has been shown to support blood–brain barrier integrity, partly of vascular endothelial growth factor (VEGF), which promotes angiogenesis and stabilizes the BBB [44,45].

Experimental studies using animal models of neurodegenerative diseases further corroborate the neuroprotective potential of GLP-2 analogues. In a murine model of sporadic Alzheimer’s disease induced by intracerebroventricular administration of streptozotocin, GLP-2 improved spatial learning in the Morris Water Maze test, normalized oxidative stress parameters (elevated glutathione, reduced thiobarbituric acid reactive substances), and promoted hippocampal neurogenesis, as evidenced by increased expression of Ki67 and doublecortin [33]. In a model of vascular dementia caused by chronic cerebral hypoperfusion, treatment with teduglutide—a dipeptidyl peptidase-4 (DPP-4)-resistant GLP-2 analogue—restored hippocampal GLP-2 receptor expression, improved long-term potentiation and long-term depression, and ameliorated cognitive performance [16]. In another study, long-term administration of Gly²-GLP-2 to mice on a high-fat diet attenuated oxidative stress, reduced markers of neuroinflammation (GFAP (glial fibrillary acidic protein), nuclear factor kappa B, amyloid precursor protein), and improved hippocampal function and cognition [34].

Further evidence comes from models of Parkinson’s disease (PD). In the MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine)-induced mouse model, Gly²-GLP-2 improved motor coordination, preserved mitochondrial function, reduced neuroinflammatory activity, and inhibited NLRP3 (NOD-, LRR- and pyrin domain-containing protein 3) inflammasome activation [23]. In vitro studies using SH-SY5Y and Neuro-2a neuronal cell lines exposed to mitochondrial toxins demonstrated that GLP-2 preserved mitochondrial membrane potential and ATP production, reduced apoptotic signaling (via caspase-3 suppression and Bcl-2 upregulation), and enhanced autophagic flux [35].

Despite these promising findings, clinical translation of GLP-2-based therapies remains limited. Currently, GLP-2 analogues are approved only for short bowel syndrome [46], and no clinical trials have evaluated their efficacy in neurodegenerative conditions. Challenges to clinical application include limited BBB penetration and a lack of safety and efficacy data in neurological populations. These factors may limit the translational potential of GLP-1 receptor agonists.

In contrast, GLP-1 and its receptor agonists have been extensively investigated for neuroprotective effects, particularly in Alzheimer’s disease. GLP-1 is produced not only in the gut but also in the brainstem (nucleus tractus solitarius), and its receptors are widely distributed in the hippocampus and cortex [47,48]. Activation of GLP-1 receptors counteracts multiple pathogenic mechanisms implicated in AD, including insulin resistance, mitochondrial dysfunction, oxidative stress, neuroinflammation, and neuronal apoptosis [49,50].

Preclinical studies using APP/PS1 (amyloid precursor protein/presenilin-1) and triple transgenic (3xTg)-AD mouse models have consistently demonstrated that GLP-1 RAs, such as liraglutide, semaglutide, and exendin-4, reduce β-amyloid accumulation, inhibit tau phosphorylation, restore synaptic plasticity, enhance neurogenesis, and improve cognitive outcomes [51,52,53]. These effects are independent of glycemic control, indicating a direct central mechanism of action. GLP-1 also restores insulin signaling via the IRS-1/PI3K/Akt pathways, enhances mitochondrial function by reducing reactive oxygen species production, and suppresses neuroinflammation by downregulating pro-inflammatory cytokines and microglial activity [50].

Clinical data, though still emerging, are encouraging. Currently, clinical trials are being conducted in individuals with Alzheimer’s disease to investigate the effects of various GLP-1 analogues on cognitive function, cerebral glucose metabolism, disease biomarkers, and treatment tolerance and safety [54]. Additionally, some observational studies on GLP-1 analogues have assessed their impact on cognitive deficits. However, observational findings are scarce due to the lack of granular information about cognitive scores in routinely collected data [54].

Below, we present the available data on studies investigating the roles of GLP-1 and GLP-2 in Alzheimer’s disease (Table 4).

Taken together, while both GLP-1 and GLP-2 share several mechanistic targets—such as modulation of neuroinflammation, oxidative stress, and mitochondrial protection—GLP-1 RAs benefit from superior blood–brain barrier penetration, broader receptor distribution in the CNS, and a more advanced stage of clinical development. Their potential as disease-modifying therapies is currently under investigation in large-scale randomized trials. Comparative analyses of GLP-1 and GLP-2 pathways may yield insights into complementary or synergistic therapeutic strategies for neurodegenerative diseases.

Despite promising preclinical data, clinical translation of GLP-2-based interventions remains limited. At present, GLP-2 analogues are approved for the treatment of short bowel syndrome, and no clinical trials have evaluated GLP-2-related therapies in neurodegenerative disorders. Key barriers include limited central nervous system exposure due to peptide-related constraints and insufficient safety and efficacy data in neurological populations. These factors may pose greater translational challenges than those posed by incretin-based therapies such as GLP-1 receptor agonists, for which clinical development in neurological indications is more advanced.

The present study is among the first to assess circulating GLP-2 concentrations in patients with Alzheimer’s disease. Plasma GLP-2 levels were higher in Alzheimer’s disease than in cognitively unimpaired controls; however, GLP-2 was not associated with cognitive performance or its change over 12 months. In contrast, GLP-2 showed positive associations with markers of adiposity and insulin resistance, suggesting that circulating GLP-2 may be more strongly linked to metabolic status than to clinical dementia severity. Future studies should replicate these findings in larger and age-matched cohorts with detailed metabolic phenotyping, evaluate longitudinal changes with adequate statistical power, and determine whether GLP-2 provides independent diagnostic or prognostic value beyond metabolic covariates.

Conceptualization, A.B.-B., M.C., A.D. and W.B.; methodology, M.C., A.L., M.K. and A.D.; validation, A.L., M.K. and A.D.; formal analysis, A.B.-B.; investigation, M.C., A.L., M.K., A.D. and A.K.; data curation, A.B.-B.; writing—original draft preparation, A.B.-B.; writing—review and editing, A.L., M.K. and W.B.; visualization, W.B.; supervision, W.B.; project administration, A.D.; funding acquisition, W.B. and A.D. All authors have read and agreed to the published version of the manuscript.

The study was approved by the Bioethics Committee of the Centre of Postgraduate Medical Education, Warsaw, Poland (31/PB/2020) on 11 March 2020. The study was conducted in accordance with the principles outlined in the Declaration of Helsinki.

Informed consent was obtained from all subjects or their legal representatives involved in the study.

The datasets presented in this article are not readily available because the data are part of an ongoing study. Requests to access the datasets should be directed to the corresponding author.

The authors declare that they have no conflicts of interest.

This research was funded by the Centre of Postgraduate Medical Education Grant no. 501-1-031-22-21.