Conquering Insulin Network Dysfunctions in Alzheimer’s Disease: Where Are We Today?

Insulin therapy-rationale

Conceptual breakthroughs that improved our understanding of the normal roles of insulin and IGF-1 signaling in the brain, together with the consequences of chronically dysregulated insulin signaling as mediators of cognitive-neurobehavioral impairment leading to dementia fueled consideration of insulin as a therapeutic or preventive strategy for neurodegeneration. Guidance was principally provided by the following observations: 1) Many normal homeostatic, metabolic, and cognitive-behavioral functions rely on intact insulin signaling mechanisms in the brain [67, 68]; 2) The core metabolic abnormalities (impaired glucose metabolism, insulin resistance, and impaired mitochondrial function) in T2DM and other systemic insulin resistance diseases such as obesity, metabolic syndrome, MAFLD, and PCOS, are shared with AD, PD, and VaD [8, 69]; 3) T2DM and other systemic insulin resistance diseases increase the risk of cognitive impairment [70], AD, and PD [69] due to insulin resistance [70]; 4) Insulin administration enhances working memory and cognition in people with MCI, AD, or PD, independent of co-existing T2DM [70]; and 5) The protean structural and functional pathologies in mild cognitive impairment, AD, PD, and VaD, including neuroinflammation, microvascular disease, mitochondrial dysfunction, increased Aβ deposition, hyperphosphorylation and ubiquitination of neuronal cytoskeletal proteins such as tau, white matter degeneration, and loss of synaptic plasticity can all be linked to impairments in insulin and IGF signaling in the brain [8, 69], and ameliorated with insulin therapy or insulin-sensitizing drugs [68].

Intranasal insulin: preclinical studies

Early preclinical experiments demonstrated significant effects of intranasal IGF-1 delivery for reducing cerebral ischemic injury and infarct volume [71–73]. Later studies showed that intranasal insulin remediated neuropathologic abnormalities including Aβ and pTau lesions, metabolic dysregulation [74], and neurocognitive deficits in experimental models of AD [75–79], and prevented dopaminergic neuronal loss in a PD model [80, 81]. Furthermore, repeated of intranasal insulin deliveries were found to have long-lasting effects on cognitive impairment and AD-type neurodegeneration in the ic-STZ rat model [82]. Treatment with long-acting insulin enhanced memory in aged rats, but without significant differences from short-acting insulin [78].

Intranasal insulin: clinical trials

Both preclinical and human studies provided proof of concept that cognitive impairment could be mediated by insulin deficiency and/or insulin resistance and that by providing exogenous insulin, cognitive and memory deficits of the types seen in diabetes mellitus and AD could be remediated [47, 84]. Given that older age groups have the greatest need for supplementing the function of insulin-related pathways, intranasal delivery was deemed to be the safest route of drug delivery. Intranasal insulin minimizes or by-passes potential adverse systemic effects including hypoglycemia, while optimizing the benefits of improved memory and cognition in MCI and AD, with or without T2DM [70, 85–87].

The initially positive results raised interest in the potential use of early intranasal insulin for preventing or significantly remediating incident MCI and AD [49, 88]. In a double-blind, placebo-controlled clinical trial to test safety, feasibility, and efficacy, intranasal insulin was used to treat participants with amnestic MCI or AD who had a Mini-Mental State Examination score of 20 or higher. The study, conducted between 2014 and 2018 over 27 sites found no significant cognitive or functional benefits for intranasal insulin versus placebo [89]. However, those results conflicted with the strong global evidence linking T2DM, dysregulated insulin metabolism, and AD via impairments in multiple insulin signaling mechanisms [90]. Correspondingly, in a meta-analysis that included 23 clinical trials, intranasal regular insulin enhanced verbal memory in most but not all studies [91]. In another randomized placebo-controlled trial that included 78 participants treated with intranasal regular insulin for 12 months, neuroimaging studies demonstrated significant reductions in cerebral white matter hyperintensity volumes [47]. Importantly, white matter hyperintensities progressively increased in correlation with worsening AD cerebrospinal fluid (CSF) biomarkers and cognitive function [47]. In their follow-up study, the authors showed that the intranasal insulin treatments that positively impacted white matter hyperintensities were associated with alterations in CSF immune, inflammatory, and vascular markers together with Aβ and pTau concentrations, suggesting that the interventions mediated their effects in part via compensatory shifts in the host immune responses [92]. In another systematic review of experimental and clinical trial manuscripts published between 2005 and 2022, intranasal insulin, thiazolidinediones, and incretins were shown to have significant positive effects on cognitive impairment in AD [93].

Although encouraging data from many studies support the use of intranasal insulin spray to treat or prevent cognitive impairment/declines and neurodegeneration linked to reduced insulin and IGF signaling in the brain, the results have been mixed [94]. Conceivably, the benefits of intranasal insulin in clinical trials may have varied due to differences in delivery methods/devices, type of insulin used (short- versus long-acting), duration of the study, or disease stage heterogeneity of the participants [94, 95]. Alternatively, additional measures are needed to address the insulin and IGF resistance components of neurodegeneration, as trophic factor replacement with impaired receptor responsiveness would not suffice for restoring the activation of critical signal transduction pathways. Additional limitations of intranasal insulin pertain to its limited delivery efficiency and, ultimately, its effectiveness in clearing Aβ [9]. These concerns inspired additional and alternative research approaches designed to enhance insulin signal transduction pathways. Those strategies included the use of metformin, PPAR agonists/insulin sensitizers, and incretins [70]. In addition, clinical trials are currently underway to investigate the efficacy of intranasal insulin with or without dulaglutide, a GLP-1RA with insulin sensitizer effects [96].

Metformin rationale

Metformin is the preferred first-line drug for T2DM due to its efficacy in lowering blood glucose, high level of safety, and affordability. Metformin (dimethylbiguanide) is a synthetic plant derivative that acts by reducing hepatic and peripheral glucose production. The few contraindications to its use in T2DM pertain to individuals with chronic kidney disease or severely impaired liver function. Repurposing Metformin to treat cognitive impairment in people with T2DM, MCI, or AD is justified because the drug readily crosses the BBB, and in the CNS, it reduces insulin resistance, activates neurons and glia, exerts neuroprotective, neurotrophic, and pro-neurogenesis effects, and protects against microvascular degeneration and disintegration, and BBB dysfunction [97]. Beyond T2DM, MCI, and AD, metformin treatments support the integrity of the neurovascular unit in cerebrovascular disease and stroke [98]. The significance of this effect is that intactness of the neurovascular unit is needed to support the centrally positioned astrocytes that modulate neuronal synaptic functions, oligodendrocytes, endothelial tight junctions, pericyte signaling, and microglial clearance of toxicants [98].

Metformin achieves its multifaceted neuroprotective effects via stimulation of AMP activated protein kinase (AMPK), and reducing inflammation, oxidative stress, mitochondrial dysfunction, and excitotoxicity, while supporting neurogenesis. In addition, some of metformin’s actions are linked to mechanistic target of Rapamycin (mTOR)-dependent pathways that are altered in AD [99]. Metformin’s anti-inflammatory effects are mediated by the inhibition of NFκB, one of the key pro-inflammatory drivers of aging. Metformin activation of AMPK supports bioenergetics by increasing ATP levels, inhibiting phosphorylation of acetyl-CoA carboxylase which increases oxidation and uptake of fatty acids, improving lipid metabolism and insulin sensitivity, and blocking gluconeogenesis through inhibition of mitochondrial glycerophosphate dehydrogenase. In addition, metformin activation of autophagy enhances neuronal bioenergetics, driving neuronal repair and reducing the accumulation of toxic protein aggregates [99]. Altogether, these findings suggest that metformin may be suitable for managing neurodegeneration in MCI, T2DM, and AD.

Metformin-AD studies: preclinical

Metformin was discovered as a glucose-lowering agent for potentially managing T2DM in 1918, but due to several clinical failures related to toxicity, it was not established as the drug of choice for T2DM until 1995. However, more recently, metformin was proposed as a potential therapeutic drug candidate for various CNS disorders, including neurodegeneration and stroke [98, 100], due to its neuroprotective effects and ability to enhance neuronal survival. For example, in an experimental rat model of dietary zinc deficiency which was associated with increased activation of GSK-3β and neurodegeneration, metformin treatment reduced neurodegeneration in concert with inhibition of GSK-3β activity, oxidative stress, and glutamate neurotoxicity [101].

Metformin-PD studies: preclinical

To evaluate the potential benefits of metformin in PD, studies were conducted in a standard mouse model generated by 6-hydroxydopamine (6-OHDA) induced lesions. Metformin co-treatment with L-DOPA did not significantly alter the pharmacotherapeutic effects of L-DOPA. However, it did attenuate L-DOPA-induced dystonia. Metformin’s therapeutic effects on abnormal involuntary movement scores were sustained for up to 20 days, along with persistent enhancement of the mTOR, dopamine D1 receptor, and extracellular signaling-regulated kinase 1/2 (ERK-1/2) signaling in dopamine-denervated striatum. Another aspect of metformin-induced neuroprotection was its normalization of 6-OHDA-mediated increases in GSK-3β activity along with Akt activation in the setting of chronic L-DOPA treatment. Altogether, these preclinical findings support the concept that metformin could be used to suppress or manage L-DOPA-induced motor complications in patients with PD [102].

Metformin-AD/MCI: clinical trials

In humans, the potential benefits of metformin for treating or preventing neurodegeneration initially stemmed from epidemiological studies identifying strong associations between T2DM and AD, and providing evidence that cognitive impairment and its progression to dementia in people with T2DM could be due to CNS insulin resistance. Those analytics inspired the intent to treat or prevent clinical trials using metformin, which penetrates the BBB and has both insulin-sensitizing and antioxidant properties [97]. In one of the early pilot randomized placebo control trials, metformin was evaluated for its effects on memory and AD Assessment Scale-cognitive subscale (ADAS-cog) score in participants who had amnestic MCI [103]. Within a 12-month interval, the investigators detected singularly significant improvements in recall on the selective reminding test [103].

A subsequent small pilot 8-week duration placebo-controlled crossover study that included 20 nondiabetic subjects with MCI or early AD, the effects of metformin on CSF, neuroimaging, and cognitive biomarker tests were evaluated [104]. In addition to showing safety, high degrees of tolerance, and measurable levels of metformin in CSF, the investigators observed significantly improved executive function and enhanced performance on learning/memory and attention tasks in the metformin-treated group, suggesting metformin could provide significant symptomatic disease remediation in people with MCI or early AD [104]. In addition, data analyzed in a subsequent review suggested that metformin could be repurposed to treat both aging-associated neurodegenerative diseases and ischemic stroke [98]. However, a meta-analysis that included 23 trials showed promising but inclusive results of metformin for treating AD or MCI [91]. Finally, in a prospective, hypothesis-driven clinical study, metformin was again found to have neuroprotective effects. Altogether, metformin’s abilities to enhance cognitive function and reduce incident dementia make it a strong candidate for use in individuals in the early stages of AD or who are at high risk for developing cognitive impairment, e.g., T2DM [105].

Metformin-PD: clinical trials

Metformin has been proposed as a therapeutically effective drug candidate in various CNS disorders, including PD. Although the dopamine (DA) precursor 3,4-dihydroxyphenyl-l-alanine (L-DOPA) is commonly used to treat PD-associated severe movement defects, prolonged therapy can lead to L-DOPA-induced dyskinesia. Contrary to the pilot MCI/AD and preclinical PD study results, a disappointing meta-analysis gave pause to the overall concept that metformin could be used to prevent or reduce neurodegeneration [106]. Among 23 comparisons in 19 studies that included 285,966 participants, metformin exposures were determined to have no significant effect on the incidence of neurodegenerative diseases, and worse yet, metformin monotherapy significantly increased the risk of PD incidence relative to non-metformin or glitazone users. These findings are consistent with those in an earlier 12-year follow-up retrospective study in Taiwan which detected increased risks of PD, AD, VaD, and all-cause dementias in patients with long-term metformin exposures [107]. Although not readily explained, one consideration is that long-term metformin use leads to vitamin B12 deficiencies which could abolish any neuroprotective actions of metformin [108]. Correspondingly, there are ample recent epidemiological or clinical studies showing reduced dementia risk with metformin treatment [109, 110] or clinical progression of neurodegeneration upon cessation of metformin therapy [111].

PPAR agonists-rationale

PPAR are nuclear receptors that function as ligand-activated transcription factors. The three PPAR subtypes, PPAR-α, PPAR-β/δ, and PPAR-γ are differentially expressed throughout the body, including the CNS [112, 113]. Lipids function as endogenous ligands for all PPAR subtypes, directly linking their interactions to metabolism. PPARs heterodimerize with retinoic X receptors to modulate downstream target gene expression which can be modulated by co-repressors or co-activators. PPAR activation can lead to changes in cell proliferation, survival differentiation, energy homeostasis, oxidative stress, inflammation, and mitochondrial fatty acid metabolism. PPARs are differentially expressed throughout the body. Numerous pharmaceutical compounds have been developed to activate PPARs and mediate selective effects on cellular functions and target diseases. Based on their modes of action, PPAR agonists hold promise for treating a broad range of diseases including metabolic syndrome, cardiovascular disorders, and neurodegeneration [114]. The potential use of PPAR agonists to treat or prevent neurodegenerative diseases addresses limitations of insulin-only treatment by supporting insulin receptor sensitivity, anti-inflammatory and antioxidant measures, and impairments in energy homeostasis, which in many respects parallels the purported CNS benefits of metformin.

Although all PPAR subtypes are expressed in the brain [112, 113], PPAR-β/δ is most abundant [3], yet most preclinical and clinical trial studies relevant to CNS diseases have been performed with PPAR-α or PPAR-γ agonists. Nonetheless, PPAR agonists share overlapping roles in the CNS, as their energy sensor functions, together with AMP-activated protein kinase-(AMPK), serve to regulate autophagy in relation to Aβ metabolism, tau phosphorylation, and neuroinflammation [19].

PPAR-α, a ligand-activated transcription factor and regulator of lipid metabolism, is activated by metabolites of fatty acids such as fibrates. PPAR-α agonists have been used to prevent complications of T2DM [115]. In the CNS, PPAR-α was shown to modulate neuronal functions related to synaptic plasticity and cognition in an AD mouse model [116], and regulate genes that encode proteins needed for glutamate homeostasis, cholinergic/dopaminergic signaling, and enzymatic processing of the amyloid-β protein precursor (AβPP), i.e., α- and β-secretases [112]. AβPP impairs lipid synthesis needed for cortical network activity and decreases PPAR-α expression in cultured cortical cells from AD transgenic mice. Correspondingly, in both late-onset AD and AβPP gene duplication associated early-onset AD, PPAR-α expression and activation are reduced, inversely proportional to the elevations in AβPP expression [116], which could be mechanistically relevant to Aβ buildup in AD brains. In addition, PPAR-α agonist modulation of lipid metabolism, fatty acid catabolism, and neuroprotection via the mitochondrial redox support are mediated by changes in lipid metabolism, [112]. Therefore, the abnormal lipid metabolism in AD could be addressed by targeted treatment with PPAR-α agonists [117].

PPAR-γ agonists together with PPAR-γ co-activator 1 alpha (PGC-1α) mediate anti-inflammatory and anti-oxidative stress effects that enable neuroprotection in the setting of neurodegeneration [118]. The expression levels of both PPAR-α and PGC-1α are reduced in AD brains [119], other neurodegenerative diseases, and psychiatric disorders [113]. Consequences may include failed anti-oxidative and anti-inflammatory mechanisms leading to altered fatty acid transport, lipid metabolism, and mitochondrial function [112]. However, over-expression of PGC-1α reduces Aβ by regulating BACE1 [119]. Therefore, therapeutic measures to coordinately upregulate PPAR-γ and PGC-1α could be used to enhance neuronal mitochondrial biogenesis for neuroprotection in AD, PD, Huntington’s disease, and amyotrophic lateral sclerosis [120].

PPAR agonists for AD and CNS vascular disease: preclinical

The two main PPAR-γ agonists used in preclinical models of neurodegeneration were pioglitazone and rosiglitazone. Rosiglitazone has insulin-sensitizing actions and has been used to treat T2DM and MAFLD. Efforts to assess the efficacy of PPAR-γ agonist treatments for AD-type neurodegeneration led to preclinical studies in established experimental models. In a Tg2576 AD mouse model study that included assessments of contextual fear, learning, and memory, spatial navigation, context discrimination, and object recognition learning and memory tasks, rosiglitazone enhanced selective cognitive domains that rely on the dorsal hippocampus [27]. In an earlier study, rosiglitazone reversed AD-associated memory loss in concert with downregulation of hippocampal glucocorticoid receptors [121]. However, the full effectiveness of rosiglitazone may not have been realized due to its aqueous nature and poor BBB penetration. To circumvent problems of drug delivery nano-formulated rosiglitazone was generated. In an ic-STZ sporadic mouse model of AD, nano-formulated rosiglitazone was found to be more neuroprotective than free drug treatment, attenuating behavioral and cognitive abnormalities, oxido-nitrosative stress, inflammatory cytokine expression, and neurotrophin expression [122]. In addition, nano-formulation of the PPAR-γ agonist provided strong neuroprotection, increased growth factor mRNA expression, and inhibited neuroinflammation [122].

Pioglitazone, like rosiglitazone, has insulin sensitizer, anti-inflammatory, and antioxidant effects. Pioglitazone mediates its anti-inflammatory actions by inhibiting several important pro-inflammatory cytokines (IL-1β, IL-6 and TNF-α). In addition, Pioglitazone modulates expression of mitochondrial proteins, increases signaling through NF-κB and p38 MAPK, inhibits Aβ, and negatively regulates the generation of reactive oxygen species and lipid peroxidation products [123]. Despite its positive effects as a therapeutic agent in MAFLD and T2DM, caution about its clinical use stemmed from negative/adverse effects including weight gain, edema, increased cardiovascular risk, and increased mortality. In the CNS, pioglitazone enhances synaptic plasticity, neurogenesis, and cognition, but like rosiglitazone, drug delivery across the BBB is poor [123], limiting its potential use for the treatment of neurodegeneration. However, pioglitazone may indirectly enhance cognitive function by increasing BBB transport of docosahexaenoic acid (DHA), essential fatty acid that is reduced in brains of AD mice. Experiments done with immortalized human brain endothelial cells demonstrated that pioglitazone was able to facilitate DHA transport in AD by regulating fatty acid-binding protein 5 at the BBB [124].

In an entirely unrelated series of experiments, taking advantage of the dominant CNS expression of PPAR-δ followed by PPAR-γ [3], the T3D-959 PPAR-δ/γ hybrid agonist was developed to simultaneously engage 80% PPAR-δ and 20% PPAR-γ [125–128]. T3D-959 has insulin sensitizer, antioxidant, and anti-inflammatory effects, is 100% BBB permeable, and reverses neurobehavioral dysfunctions including deficits in learning and memory, and AD neuropathological lesions in the ic-STZ rat model [33, 129]. In addition to preventing neurodegeneration, T3D-959 was shown to restore synaptic protein expression while reducing Aβ, pTau, and ubiquitin immunoreactivities [33].

In summary, preclinical studies suggest that PPAR agonists hold promise for treating human neurodegenerative diseases [130]. The functional and neuroprotective substrates of PPARs combined with their broad expression in the brain support the concept that pharmaceutical agonists could provide disease remediation in AD and other neurodegenerative diseases [131]. The most effective therapeutic targets in the brain appear to be PPAR-γ and PPAR-δ/γ. Pioglitazone, which activates PPAR-γ and inhibits long-chain acyl-CoA synthase family member 4 (ACSL4) has shown promise for preventing cognitive decline in AD and limiting damage from ischemic stroke [132]. In silico-designed hybrid PPAR-δ/γ agonists are expected to improve deficits in neurobehavioral function and impairments in synaptic plasticity while reducing neuroinflammation and Aβ deposition and toxicity [131]. Whether pan-PPAR agonists like bezafibrate can support mitochondrial dysfunction in neurodegenerative diseases remains to be tested [133].

PPAR agonists for AD: clinical studies

Initial small clinical trials with Rosiglitazone showed memory improvements in patients with early-stage AD [134, 135]. Similarly, Phase II rosiglitazone clinical trials demonstrated improved cognition and memory in mild to moderate AD [4, 136]. In another Phase II trial to assess the efficacy of an extended-release preparation of Rosiglitazone in mild to moderate AD, a treatment benefit in cognition was observed only in APOEɛ4-negative subjects [137]. However, in a 24-week, Phase III double-blind, randomized, placebo-controlled trial designed to prospectively stratify 693 participants by APOE genotype, there were no significant objective performance differences between the rosiglitazone and placebo groups, no differences from baseline. Instead, there was a significant effect on the clinician’s interview-based impression of change plus caregiver input [138]. Finally, the Reflect trials which included mild or moderate AD, assessed multiple effects of rosiglitazone in 360 participants between 50 and 90 years old and with Mini-Mental State Examination scores between 10 and 26 [139]. All participants had already been diagnosed and managed with standard therapy including donepezil or an acetylcholinesterase inhibitor. The study drugs were standard or extended-release rosiglitazone. The value of the study was that the investigators were able to identify a biomarker panel consisting of six analytes, interleukin-6 (IL6), tumor necrosis factor-alpha (TNF-α), fatty acid binding protein 3 (FABP-3), Peptide YY (PPY), Interleukin 10 (IL-10), and C-reactive protein (CRP), which accurately predicted positive responders to treatments in Phase II and Phase III trials [139]. Taken together, the outcomes of rosiglitazone clinical trials were mixed and distinctly less promising than the preclinical experimental results. In addition, the feasibility of extending the studies to larger populations was markedly curtailed by safety concerns related to cardiovascular problems including worsening of congestive heart failure and increased risk of cardiovascular events.

Pioglitazone was the second major PPAR-γ agonist used in clinical trials to treat mild to moderate AD. In contrast to rosiglitazone, pioglitazone has not been associated with safety or tolerability concerns. A meta-analysis of 9 studies that included 1314 patients and 1311 controls detected no statistically significant effects of pioglitazone treatment on cognition [140]. In a Phase III trial, pioglitazone reduced inflammatory markers, whereas acetylcholinesterase inhibitor treatment produced short-term benefits for mild to moderate AD and the NMDA receptor antagonist memantine transiently benefited severe AD [141]. A meta-analysis that included 23 clinical trials showed that pioglitazone improved cognition in participants with T2DM and MCI or AD in 3 of 5 trials, whereas detemir, a long-acting insulin, improved cognition after 2 months of treatment [91]. Thus far, the results of pioglitazone trials are not definitive. Additional studies are needed to better define the subject populations that would most benefit from early treatment. However, limitations pertain to poor BBB penetration [114], compromising CNS drug delivery, and possibly accounting for the disappointingly disparate results compared with preclinical model data. To circumvent this problem, a novel BBB-penetrant PPAR-γ agonist, Leriglitazone, is under development and evaluation [142].

Even with improved BBB penetration, PPAR- γ agonists would not engage PPAR-δ, the overwhelmingly dominant PPAR subtype expressed in the CNS [3]. both efficacy and specificity of PPAR engagement, given that PPAR-δ and PPAR-α responses are not included with PPAR-γ-specific targeting. To address the need to activate both PPAR-δ and PPAR-γ, a novel, 100% BBB-penetrant hybrid PPAR agonist, T3D-959 was designed to simultaneously engage PPAR-δ (80%) and PPAR-γ (20%) [125, 143]. In a Phase IIa exploratory trial clinical safety and tolerance were demonstrated. Importantly, in the 34 participants with mild to moderate AD, significant improvements in cerebral glucose and lipid metabolism and cognitive function (ADAS-cog11 and Digital symbol substitution test) were observed [144]. Additional studies are in progress to assess long-term responses for AD therapeutic remediation [143].

GLP-1 receptor agonists-rationale

The incretin family of hormones has critical roles in establishing and maintaining metabolic homeostasis via interaction with receptors expressed throughout the body, including the CNS. Glucagon-like peptide 1 (GLP-1), now one of the most studied of the incretins, is a peptide hormone produced by intestinal L-cells, and pre-proglucagon medullary neurons in the solitary nucleus [145]. GLP-1 is secreted in response to nutrient intake, and like other incretins, GLP-1 stimulates glucose-dependent insulin secretion and insulin biosynthesis, promotes healthy insulin signaling, regulates blood glucose levels, inhibits glucagon secretion and gastric emptying, and curtails food intake [146]. Consequently, as predicted, incretins can be used to control T2DM and obesity, and ameliorate their cardiovascular complications.

Within the past 15 years, multiple GLP-1 receptor agonists (GLP-1RAs) have been approved by the U.S. Food and Drug Administration (FDA) for the treatment of T2DM or obesity. Restrictions on the use of products geared for treating T2DM are based on the patient’s age, Hemoglobin A1c (should be 6.5% or higher), and failure to achieve satisfactory glycemic control with the maximum tolerated dose of metformin or one of the preferred SGLT2 inhibitors within 12-months prior treatment approval. The list of FDA-approved GLP-1RAs as of December 2023, includes: 1) Dulaglutide (Trulicity); 2) Exenatide (Byetta); 3) Exenatide Extended Release (Bydureon BCise); 4) Liraglutide (Victoza); 5) Liraglutide (Saxenda); 6) Lixisenatide (Adlyxin); 7) Semaglutide subcutaneous, tablet (Ozempic, Rybelsus); 8) Semaglutide (Wegovy); 9) Tirzepatide (Mounjaro); 10) Tirzepatide (Zepbound).

GLP-1RAs, and to some extent, glucose-dependent insulinotropic polypeptide (GIP) receptor agonists (GIP-RAs), are of particular interest because strong preclinical and clinical data have shown that beyond their insulin-stimulating effects to achieve glycemic control, these drugs have pleiotropic CNS actions. The dual expression of incretin receptors in the periphery and brain suggests links between systemic and brain metabolic functions. In the CNS, GLP-1 impacts neuronal functions such as blood pressure control, thermogenesis, energy homeostasis, neurogenesis, and retinal repair. In addition, GLP-1 and its analogs have potential therapeutic roles in neurodegeneration due to MCI, T2DM, AD, and PD, as evidenced by experimental and clinical data demonstrating neuroprotection and preventive actions against cognitive decline [60] [70, 147] [148] and dementia progression in AD and PD, independent of T2DM [70, 147] [149, 150]. GLP-1RA’s therapeutic effects on learning, memory, and cognition in T2DM or obesity have been linked to reduced hippocampal neurodegeneration and enhanced synaptic plasticity [151, 152]. In PD models, GLP-1RAs were shown to protect motor activity and dopaminergic neurons, whereas in AD models, they improved nearly all neuropathological features including cognition and they reduced Aβ peptide aggregation [145, 154]. Beyond MCI, AD, PD, and T2DM, GLP-1RAs have beneficial effects in reducing experimental cerebrovascular-mediated injury including cerebral infarct volumes and related neurological deficits caused by the inhibition of oxidative stress, inflammation, and apoptosis [145].

More recently, ‘shout-outs’ for repurposing GLP-1RAs to address cognitive impairments in T2DM and AD have gained steam. Similarly, the positive results of a 60-patient, 48-weeks, randomized, placebo-controlled washout trial for PD helped fuel the growing interest in using GLP-1RA drugs to retard PD progression [154, 155]. This subject of repurposing GLP-1RAs for neurodegenerative diseases and other neuropathological processes is now a major area of interest for multiple pharmaceutical companies hoping to bring GLP-1RAs forward as viable treatment options [155]. Extending the therapeutic application of GLP-1RAs to manage cognitive impairment is logically rooted in the known molecular and biochemical similarities between AD and T2DM, along with the higher rates of cognitive dysfunction and progression to AD in people with T2DM [51, 156]. Repurposing already-approved GLP-1RAs for ameliorating AD and PD pathology in humans is likely to be safe and cost-effective. Therapeutic efficacy in dementia and cognitive impairment associated with T2DM and PD is anticipated [32] due to GLP-1RA-mediated suppression of neuroinflammation, and support of neuroprotective and neurotrophic functions [146], energy metabolism, insulin signal transduction, and antioxidant effects [156]. On the other hand, differences in how brain energy metabolism changes in various states of insulin resistance, T2DM, and dementia highlight the therapeutic challenges. Furthermore, the heterogeneity of dementia syndromes will likely require a greater understanding of regional and mechanistic alterations in brain energy homeostasis for effective therapeutic targeting [157]. In the sections below, the results of pre-clinical studies and clinical trials with GLP-1RA therapies for neurodegeneration are discussed in relation to their mechanisms and the outcomes of treatment and prevention strategies.

Incretins target aging: preclinical

Although aging is the dominant risk factor for most neurodegenerative diseases, that trajectory is not the rule since many older adults have intact cognitive function [20]. The causes of “malignant brain aging” are not well understood, but many studies highlight impairments in brain insulin signaling and responsiveness as major culprits [20]. Correspondingly, in one study that focused on microvascular mediators of neurodegeneration, treatment with the exenatide GLP-1RA reversed brain aging-associated endothelial cell transcriptomic changes and BBB leakage and attenuated microglial priming [48]. In related work, exenatide treatment was demonstrated to reverse aging-associated genome-wide transcriptomic signatures in multiple brain cell types including glia and gliovascular units [158]. Importantly, many genes that mediate glucose uptake and utilization, lipid metabolism, and protein processing in both aging and neurodegeneration were impacted by exenatide treatment [158]. Furthermore, GLP-1RA treatment upregulated genes that encode metabolite receptors and transporters, neurotransmitter receptors, glial ion channels, synaptic plasticity [158]. These compound anti-aging effects of GLP-1RA bode well for potential slowing or pausing of brain aging and its progression to neurodegeneration.

Incretins target inflammation: preclinical

Inflammation is a consistent pathophysiological process in neurodegeneration. The key cellular mediators are activated microglia and astrocytes which secrete proinflammatory cytokines. Neuroinflammation promotes brain insulin-resistant states that contribute to neurodegeneration. In a mouse model, genetic depletion of GPR120, an omega-3 fatty acid receptor that has anti-inflammatory and insulin-sensitizing effects [159] increased prostaglandin D2 (PGD2)-mediated neuroinflammation in the hippocampus and reduced GLP-1 in plasma, both of which were abrogated by treatment with Liraglutide, a GLP-1RA [160]. In another study of aluminum-induced dementia in a rat model, the neuroprotective actions of Liraglutide prevented memory deficits associated with reductions in proinflammatory cytokine and oxidative stress indices [161]. In a third set of experiments that utilized the ic-STZ sporadic and 5xFAD genetic AD mouse models, Liraglutide’s attenuation of the brain insulin-resistant states was neuroprotective and had both anti-inflammatory and anti-Aβ effect [162]. These preclinical findings position Liraglutide and other GLP-1RAs at the intersections of neuroinflammation, brain insulin resistance, and neurodegeneration, supporting their potential use in human neurodegenerative diseases.

Incretins target metabolism: preclinical

AD and other neurodegenerative diseases exhibit dysregulated energy metabolism, mitochondrial dysfunction, and increased oxidative stress [43]. Although dysregulated signaling through insulin and IGF-1 networks are front and center in these pathologies, downstream pathways that crosstalk with mediators of cell survival, homeostasis, mitochondrial function, growth, and repair are also impacted. Accompanying increases in oxidative stress and neuroinflammation contribute to the neuropathologic processes in AD and PD. Preclinical models have aided in illuminating these points by demonstrating reduced pathology following treatments that support energy metabolism and insulin/IGF-1 signaling. For example, enhanced cognition measured in GLP-1RA treated 5xFAD mice was shown to be mediated by increased aerobic glycolysis, energy production, and neuroprotection, along with lower levels of oxidative stress in the brain [163]. In addition, in a DM rat model, exenatide inhibition of neuronal apoptosis-lined to cognitive impairment [164] illustrates the positive impact of GLP-1RAs for neuroprotection via support of insulin metabolic pathways.

Incretins target microvasculature: preclinical

In addition to supportive maintenance of neurons, GLP-1RAs support cerebrovascular functions required for metabolic homeostasis, neuroprotection, and lowering of pro-inflammatory and pro-oxidative stress responses. For example, in an L-methionine-induced rat vascular dementia model, the GLP-1RA sitagliptin, reduced cognitive deficits by improving oxidative stress biomarkers, inflammatory mediators, lipid profiles, and brain-derived neurotrophic factor expression in the hippocampus [165]. Those effects were associated with reduced levels of caspase-3 and glial fibrillary acidic protein, a marker of astrocyte activation, and increased Ki-67 immunoreactivity reflecting neurogenesis. Apart from reducing neuroinflammation associated with neurodegeneration in experimental animal models [166, 167], GLP-1RA treatment with exenatide reversed age-related brain endothelial cell transcriptomic changes and BBB leakage [48, 158]. Additional positive GLP-1RA effects on cerebrovascular function and structure pertain to autoregulation of blood flow and neuronal-glial metabolic coupling needed to deliver nutrients to the brain [168]. Therefore, the impairments of the neurovascular unit and uncoupling of neuronal-glial metabolic functions in neurodegenerative diseases could potentially be remediated by GLP-1RA treatment to normalize nutrient uptake, waste removal and integrity of the BBB [168].

Incretins target AD: preclinical

Operating under the premise that insulin deficiency is one of the key mediators of AD-type neurodegeneration, GLP-1RAs have been used to remediate related deficits in neurobehavioral function and structural neuropathological lesions. For example, in a db/db diabetic mouse model treated with siRNA to inhibit insulin synthesis-related gene expression in the brain, improvements in cognitive function accompanied by increased insulin synthesis and reductions in tau hyperphosphorylation were achieved by treatment with Exendin-4 [169]. Similarly, Liraglutide treatment of AD experimental models yielded significant benefits to cognitive function and AD pathology [170], and thwarted chronic unpredictable stress-induced depression linked to cognitive impairment [171]. Since depression can be an early feature of AD and other neurodegenerative diseases [172, 173], GLP-1RAs could potentially be used to halt or retard the development and progression of cognitive impairment.

PCOS is another insulin resistance disorder associated with metabolic disruption and cognitive impairment in women. Impairments in Notch pathway signaling have been identified as factors contributing to pathological processes in both the ovary and brain/cognitive function [174, 175]. However, Notch pathway activation is linked to insulin and IGF-1 signaling [176]. Correspondingly, treatment with the GLP-1RA liraglutide, enhanced cognitive function and Notch pathway signaling through targets that mediate neuroprotection and memory in a rat model of PCOS [176].

Incretins target AD and PD: clinical trials

Early clinical trials for MCI/AD showed that insulin administration improved memory, brain activity, neuronal energy utilization, and markers of inflammation [87]. However, limitations pertaining to delivery and long-term efficacy of intranasal insulin nudged the field to consider therapeutic targeting of brain insulin resistance for treating or preventing cognitive impairment in T2DM and AD. The basis for this modified approach stemmed from growth in the understanding of how insulin resistance drives many pathological, molecular, and pathophysiological processes in brains with neurodegeneration, together with strong evidence that the disease mechanisms in AD, T2DM, and other insulin resistance diseases are shared [20]. In essence, replacing insulin sources helps but does not address the problem of progressive brain insulin and IGF resistances. Metformin and PPAR agonists have already been evaluated in many preclinical studies and clinical trials, but enthusiasm for their widespread clinical use has for the most part peaked because the successes of preclinical studies were not robustly mimicked in human clinical trials, and PPAR agonist drugs optimized for brain delivery and treatment of neurodegeneration are still under investigation [131, 177–179]. Once again, attention has shifted due to rapidly growing interest in the re-purposing of incretin mimetics for treating cognitive impairment and dementias [60, 156].

Preclinical studies showed that incretin mimetics, in particular, GLP-1RAs, have multiple positive effects on processes that mediate neurodegeneration including neuroinflammation, tau hyperphosphorylation, Aβ accumulation, impaired synaptic plasticity, and loss of white matter integrity, which together contribute to deficits in attention, memory, and executive functions [180, 181]. The neuroprotective effects of incretin mimetics are likely mediated by the restoration of insulin signaling networks, together with the targeting of inflammatory responses, apoptosis, toxic protein aggregation, autophagy, and impairments in long-term potentiation. The growing list of GLP-1RAs is just beginning to address the relentlessly increasing rates of obesity, T2DM, insulin resistance diseases in general, and the associated cognitive impairment that could potentially lead to dementia. However, opportunities for incretin-mimetic therapy extend beyond AD and T2DM-related dementias since previous studies demonstrated deficits in insulin and IGF signaling mechanisms in brains afflicted with PD [32, 69], DLBD [182], FTLD [183], and multiple sclerosis [184]. Recent data suggest that incretin mimetics (GLP-1RAs) could provide neuroprotection in nearly all these conditions [168]. Furthermore, additional opportunities to deliver neuroprotection will likely emerge via creative combination targeting of incretin receptors, as suggested by the superior outcomes of early AD and PD clinical trials with co-administered GLP-1RA and GIP-RA [8], or dual GLP-1/GIP RAs that cross the BBB [185]. Although the fundamental actions of GLP-1RAs combined with preclinical data provide strong support that this class of drugs is effective in reducing the risk for AD in patients with T2DM, mechanistic details pertaining to the efficiency of drug penetration across the BBB and crosstalk between peripheral and CNS responses including inflammation are lacking [50].

Exenatide was one of the first GLP-1RAs used to examine the neuroprotective effects of incretins in early AD. In a small, 18-month double-blind randomized placebo-controlled trial, safety and tolerability were confirmed, and neuronal extracellular vesicle expression of Aβ was reduced, but without significant alterations in cognitive performance, cerebral cortex volume, or CSF AD biomarkers [186]. Although the finding of reduced Aβ in neuronal extracellular vesicles was encouraging, due to its relatively small size, the study design lacked sufficient statistical power to be conclusive.

In the ELAD double-blind placebo-controlled multicenter 12 months trial to examine the effects of liraglutide on cerebral cortical glucose metabolism in mild AD [187], the investigators detected significant improvements in ADAS-Cog performance and MRI cortical volume in the liraglutide-treatment group versus [188]. In a separate 12-week clinical trial to assess GLP-1RA’s therapeutic effects on cognitive impairment related to dysregulated metabolism in T2DM, liraglutide treatment was shown to improve cognitive scores, memory, and attention [189]. Functional gains were associated with selectively increased regional brain activation that was unassociated with changes in glycemic control, blood pressure, or body weight [189], dissociating the index responses from potential systemic metabolic effects of GLP-1RAs. In addition, in a small pilot study, Liraglutide was shown to enhance cerebral glucose metabolism and cortical connectivity in AD, whereas exenatide restored motor function and cognition in PD [154], perhaps suggesting differential targeting of specific GLP-1RA formulations or differential susceptibility of brain regions based on the presence, nature, and severity of neurodegeneration. In support of the concept that specific pharmaceutical formulations of GLP-1RAs can differentially target pro-metabolic effects, in the STEP clinical trials, the novel, long-acting GLP-1RA, Semaglutide, was demonstrated to have broad-spectrum application for treating insulin-resistance diseases [190].

In a randomized 16 weeks head-to-head clinical trial, the differential therapeutic effects of liraglutide, dapagliflozin, and acarbose on cognition and brain functional changes were examined in 36 people with T2DM in whom metformin monotherapy was deemed inadequate for achieving metabolic control [191]. Multipronged assessments included brain functional MRI (fMRI) scans and a battery of cognitive performance tests. Corresponding with earlier studies, only liraglutide significantly improved brain health and cognitive function as evidenced by enhancement of hippocampal activation and restoration of cognitive domains such as delayed memory [191]. However, in contrast to the 12-week study by Li et al. [189], this somewhat longer evaluation showed that liraglutide-associated cognitive improvements were associated with positive metabolic outcomes such as reduced waist circumference, body fat ratio, and fasting blood insulin [191]. Overall, the combined results of preclinical and clinical trial studies suggest that GLP-1RAs are neuroprotective against memory impairments associated with T2DM, AD, and PD [192]. However, further guidance is anticipated from ongoing Phase 2 and Phase 3 trials in AD and PD populations [180].

The main overall conclusion drawn from preclinical studies and clinical trials is that GLP-1RAs hold promise as effective disease-modifying therapeutic approaches for preventing and reducing the risk of T2DM, AD, PD, and probably cerebrovascular-related cognitive impairment and dementia. The objective measures of GLP-1RA’s neuroprotective effects including improvements in brain markers of hippocampal connectivity, cerebral glucose metabolism, and hippocampal activation by functional magnetic resonance imaging, even without strong correlations with cognitive scores, are hopeful. The most exciting outcome of this rapidly emerging data is that by targeting underlying metabolic pathologies linked to brain insulin resistance, a broad spectrum of neurodegenerative processes can be addressed in ways that are superior to the presently available approaches, e.g., anti-Aβ, anti-p-tau, and anti-inflammatory drugs. Furthermore, pharmaceuticals designed for treating systemic insulin-resistance diseases could be re-purposed in time-efficient manners to address the global pandemic crisis of neurodegeneration. Going forward, a major task will be to determine which GLP-1RA compounds have the greatest efficacy in normalizing systemic as well as CNS metabolic functions.

However, despite encouraging gains, two immediate concerns remain including: 1) Can long-term GLP-1RA treatments significantly ameliorate symptoms, neuropathology, and disease progression in humans with MCI or early to moderate AD; and 2) Will GLP-1RA mono-therapeutic approaches suffice for managing neurodegeneration. Regarding the former, epidemiological studies suggest GLP-1RAs mediate their neuroprotective actions by delaying progression of MCI to dementia in T2DM [193] rather than wholly aborting the process. Regarding the latter, it is not unreasonable to anticipate that additional and alternative measures will likely be required to support declining neurocognitive and behavioral functions in the setting of insulin resistance in light of the well-recognized eventual plateauing of weight loss in people treated with a mono-GLP-1RA [194] and greater efficacy of dual targeting such as with GIP and GLP-1 receptor agonists [195].

Beyond GLP-1RA monotherapy for neurodegeneration: rationale

GLP-1 is fundamentally an anorexigenic peptide. Related molecules include the glucose-dependent insulinotropic polypeptide, formerly gastric inhibitory peptide (GIP), peptide YY (PYY), and cholecystokinin (CCK). GIP-RAs, like GLP-1RAs, have pro-growth, pro-metabolic, and anti-inflammatory effects [153], and dampen the adverse effects of T2DM [30] and AD- as well as PD-related neuropathologic processes in experimental mouse models [153]. Although human data resulting from GIP-RA mono-therapy have not been generated, in a postmortem study, significant alterations in GIP and GLP-1 protein expression were detected in frontal cortex tissue from patients diagnosed with AD or PD [25]. These abnormalities suggest that brain incretin deficiencies are components of neurodegeneration contributing to cognitive-motor and behavioral deficits in these diseases, and that supplementation with exogenous GLP-1 and GIP (incretin) receptor agonists would provide therapeutic value. A number of preclinical and human studies have already demonstrated benefits of treating neurodegenerative diseases with novel dual -targeting GLP-1RA/GIP-RA compounds with promising results [151, 195–198]. Besides incretin receptor agonists, other approaches to enhance the functionality of incretin networks include the use of inhibitors that block the actions of DPP-4 (DPP-4i) or sodium-glucose co-transporter 2 (SGLT2) inhibitors [199]. DPP-4 promotes enzymatic degradation of incretins and DPP-4 inhibitors prevent or reduce incretin degradation [200]. SGLT2 inhibitors reduce blood glucose by blocking sodium-glucose transport proteins in renal nephrons and intestinal mucosa [199]. Both DPP-4 and SGLT2 inhibitors have been shown to have therapeutic effects in AD and PD. Examples of preclinical and clinical trial results with GIP-RAs, GLP/GIP dual receptor agonists, DPP-4 inhibitors, and SGLT2 inhibitors are discussed below.

GIP: preclinical

A long-acting GIP incretin, D-Ala(2)GIP was tested in preclinical experiments with genetic AD mouse models [30, 202]. Those studies demonstrated that the GIP agonist, which normalizes insulin signaling, exerted neuroprotective effects as demonstrated by reductions in Aβ plaque load, oxidative stress, neuroinflammation, and gliosis [30, 202]. In addition, D-Ala(2)GIP reversed or prevented neurocognitive deficits in association with enhanced hippocampal long-term potentiation, synaptic plasticity, and cAMP-/PKA/CREB signaling, inhibition of neuroinflammatory cellular reactions, and reduction in Aβ deposition [201].

GLP1/GIP dual analog

Research pertaining to the neuroprotective/anti-neurodegenerative effects of GIP-RAs evolved to focus on the therapeutic effects of a novel GLP-1RA/GIP-RA dual incretin analog, DA5-CH [202], that was shown to have substantial disease-modifying effects in AD and PD preclinical models [153] including the ic-STZ rat model of AD [197]. In this regard, DA5-CH reduced hippocampal tau phosphorylation, reversed working memory impairments, increased the energy of theta band activity in the hippocampal CA1 region, reduced mitochondrial stress, and normalized signal transduction with attendant enhanced transcription factor P-CREB(S133). In addition, DA4-JC, another GLP-1/GIP RA dual agonist, was demonstrated to significantly improve cognitive performance, synaptic number, dendritic spines, post-synaptic density protein 95, and mitochondrial integrity in brains of the APP/PS1/Tau triple AD mouse model [151]. Additional effects included improvements in the phosphatase and tensin homologue-induced putative kinase 1 (PINK1) – Parkin mitophagy signaling pathway, and downregulation of Aβ, p-tau, and the P62 autophagy marker [151]. To aid in prioritizing incretin receptor agonist-treatment approaches for human neurodegenerative diseases, a comprehensive comparison of multiple promising drugs was performed, focusing on their efficiencies in penetrating the BBB. The authors compared data obtained with single GLP-1RAs (exendin-4, liraglutide, lixisenatide, and semaglutide) and dual incretin receptor agonists (DA3-CH and DA-JC4) and found that DA5-CH and Exendin efficiency crossed the BBB, suggesting potential utility for treating humans with AD or PD [203]. Treatment with novel dual GLP-1/GIP receptor agonists that cross the BBB at enhanced rates [153], improved neuroprotection to greater degrees than a single GLP-1RA [192], suggesting a greater potential for providing disease-modifying care for patients with AD or PD.

Oxyntomodulin

This peptide hormone and growth factor released from the gut activates GLP-1 and glucagon (Gcg) receptors in hypothalamic neurons. Although oxyntomodulin binds to GLP-1 receptors with lower affinity than GLP-1 and other agonists, it nonetheless suppresses appetite, and like other GLP-1RAs, Oxm facilitates insulin signaling and has neuroprotective effects against Aβ-induced cytotoxicity [152]. Treatment of the APP/PS1 transgenic AD mice with a protease-resistant oxyntomodulin analogue (D-Ser2) prevented impairments in working and spatial memory and reduced Aβ plaques and deficits in synaptic long-term potentiation in the hippocampus. Mechanistically, these responses were associated with increased signaling through PI3K/Akt and inhibition of GSK-3β [152]. These findings are consistent with other reports that GLP-1RA treatments prevent AD symptoms and pathology by enhancing positive signaling through PI3K-Akt pathways that support cell survival, plasticity, metabolism, and homeostasis, and inhibit inflammatory and stress mediators, including Aβ and pTau accumulations in the brain [152].

DPP-4 inhibitors

Since DPP-4 is a negative regulator of incretins, DPP-4 inhibitors could potentially enhance the actions of incretins in the brain and prevent neurodegeneration. as observed with GLP-1RA drugs. DPP-4 inhibitors are used clinically and include sitagliptin, vildagliptin, saxagliptin, alogliptin, linagliptin, which can be administered alone or in combination with metformin. Treatment of an AD rat model with Linagliptin, a DPP-4i that crosses the BBB, enhanced GLP-1 and attenuated hippocampal Aβ, serine phosphorylation of IRS-1 (S307), GSK-3β activity, TNF-α, IL-1β, IL-6, acetylcholinesterase, and oxidative/nitrosative stress [42].

SGLT2 inhibitors

SGLT2 inhibitors have mainly been used as glucose lowering agents, but they also provide significant cardioprotective and renoprotective benefits [204]. In addition, there is growing evidence that SGLT2 inhibitors also have neuroprotective effects and therefore could potentially be used in AD or cognitive impairment caused by T2DM [65]. A limited number of human studies have investigated the potential use of glucose-lowering SGLT2 and DPP-4 inhibitors. In a retrospective meta-analysis of 10 studies that included a total of 819,511 subjects, treatment with an SGLT2 inhibitor, DPP-4 inhibitor, or with a GLP-1RA resulted in lower risks of all-cause dementia in subjects with T2DM [200]. Although specific benefits related to treatment of AD versus VaD could not be established.

Multi-pronged therapies for neurodegeneration: rationale

The field of AD has traveled a long and tortuous journey from its original clinical and neuropathological descriptions including extensively detailed analyses of structural and ultrastructural lesions that continue to serve as the gold standards for diagnosis. Molecular and biochemical breakthroughs that enabled identification of the nature and pathophysiology of Aβ and hyperphosphorylated tau lesions fueled over four decades of research including the generation of genetic models and non-invasive assays to detect the signature Aβ and hyperphosphorylated tau lesions in the brain for diagnostics, screening, monitoring, and enrollment into clinical trials. Those same avenues of research led to the development and testing of many anti-Aβ therapies that proved unsuccessful for remediating disease. Fortunately, parallel research efforts identifying dysregulated insulin signaling mechanisms as fundamental to AD from MCI to dementia, converge with studies that demonstrated strikingly similar abnormalities in AD and T2DM as well as most other insulin resistance diseases. Proof of concept experiments showed that in the brain, dysregulated insulin signaling leads to the multifaceted pathologies that characteristically affect neurons, oligodendrocytes, astrocytes, microglia, endothelial cells, and the BBB in AD, together with accumulations of Aβ and pTau lesions. At the same time, evidence was generated that Aβ and pTau lesions adversely impact the integrity of insulin signaling networks, and thereby contribute to a seemingly unrelated but actually highly interrelated spectrum of pathophysiological processes including increased neuronal damage and loss, impaired neuronal plasticity and synaptic function, oligodendrocyte-myelin degeneration with compromised neuronal conductivity, neuroinflammation, increased oxidative stress, microvascular disease (both Aβ+ and Aβ–), and loss of BBB integrity.

From the perspective of therapeutics, the identified deficits in insulin and insulin receptor responsiveness highlight the need to both supply insulin and support insulin receptor functions. The therapeutic strategies evaluated over time and the outcomes of both preclinical and clinical studies are summarized in Table 1. Certainly, the preclinical and human clinical studies showed benefits of supplying insulin to support cognition in the early stages of AD. However, technical problems related to intranasal insulin delivery and reproducibly assessing its effectiveness rendered this approach impractical on a broad scale. Efforts turned to the use of insulin sensitizers that mainly included metformin and PPAR agonists. Although metformin crosses the BBB, its supportive effects on cognitive function are strong in preclinical models but in human trials the outcomes have been mixed or inconclusive. PPAR agonists in theory should work given their actions in the nucleus, bypassing surface receptors and problems with receptor resistance, and robust anti-inflammatory, antioxidant, and pro-insulin signaling effects. Like metformin, the pre-clinical studies with PPAR agonists yielded promising results, but the human trial data are mixed. Unfortunately, study designs using suboptimum predominantly PPAR-γ rather than PPAR-δ agonists for correlation with the dominant PPAR subtype expressed in the brain, together with the low BBB permeability of most PPAR agonists compromised efforts to advance this therapeutic strategy. However, ongoing clinical studies with BBB permeable drugs that predominantly engage PPAR-δ followed by PPAR-γ, are in progress.

Rapid progress in the field of integrin research, inspired by the development of effective GLP-1RA drugs, opened opportunities to pursue another strategy for ameliorating brain insulin resistance in AD and T2DM. The results of those investigations are quite promising from both therapeutic and proof of principle perspectives. The enhancement in cognitive function observed thus far argue in favor of repurposing GLP-1RAs for the treatment of cognitive impairment and dementias. The fact that insulin sensitizer agents that enhance insulin signaling network functions can also reduce molecular, biochemical, and cognitive indices of AD-type neurodegeneration supports the concept that brain insulin resistance has causal roles in AD-type neurodegeneration and cognitive impairment in T2DM. Efforts are needed to optimize the drug selection to favor long-lasting effectiveness in the brain. Despite these gains, work is still needed to address the following issues: 1) besides insulin, insulin-like growth factor signaling networks are impaired in AD, PD, and other neurodegenerative diseases; 2) insulin sensitizers do not necessarily address problems due to trophic factor deficiencies; 3) emerging data suggest that more than one manner of targeting insulin/IGF receptor pathways will likely be needed to optimize and sustain therapeutic effects, i.e., therapeutic measures must be multi-pronged; 4) concerns about potential progressive declines or reaching plateaus in responsiveness GLP-1RAs or other therapeutics should be anticipated based on what is already known about the declines in weight loss efficacy in patients treated with GLP-1RAs. Going forward, these challenges must be addressed. However, the forecast is optimistic since evidence suggests multi-pronged, e.g., intranasal insulin plus a GLP-1RA [96], GLP-1RA/GIP-RA [192, 198], DPP-4i+metformin, GLP-1RA+SGLT2 inhibitor [65, 205], PPAR-δ+PPAR-γ [131, 177] are more effective than single-pronged therapeutic measures.