Did Your GLP-1 Drug Reach the Brain? Measuring CNS Target Engagement from a Blood Sample
GLP-1 receptor agonists are being tested for Alzheimer’s disease, Parkinson’s disease, depression and other brain disorders. But for most of these drugs, a basic question is still hard to answer in patients: did the drug actually reach the brain and change neuronal biology? This article explains why measuring GLP-1 CNS target engagement matters, why standard tools fall short, and how neuron-derived extracellular vesicles (NDEVs) in a routine blood sample can provide a direct readout of central drug activity.
Key Takeaways
- Most GLP-1 receptor agonists are large peptides with limited ability to cross the blood-brain barrier, so peripheral efficacy (lower glucose, weight loss) does not prove a central effect.
- Recent large trials, including EVOKE/EVOKE+ (oral semaglutide in early Alzheimer’s) and Exenatide-PD3 (exenatide in Parkinson’s), did not show clinical benefit. Without CNS target-engagement data, it is difficult to know whether the drugs failed biologically or never engaged the target sufficiently.
- Neuron-derived extracellular vesicles carry intracellular signaling proteins from neurons into the bloodstream. Measuring phosphorylated Akt, mTOR and related proteins inside NDEVs offers a minimally invasive window into neuronal insulin signaling.
- A 2026 study in The Journal of Clinical Endocrinology & Metabolism used NDEV biomarkers to detect central Akt–mTOR responses to liraglutide and pioglitazone, and found these responses were relatively independent of changes in peripheral insulin and glucose.
Why GLP-1 drugs are being tested in brain disorders
GLP-1 (glucagon-like peptide-1) receptor agonists such as exenatide, liraglutide and semaglutide were developed for type 2 diabetes and later approved for obesity. Over the past decade, researchers have become interested in their potential for neurological and psychiatric conditions for several reasons:
- Brain insulin resistance. Impaired insulin signaling in the brain has been linked to Alzheimer’s disease, Parkinson’s disease and depression. In 2015, Kapogiannis and colleagues reported abnormal insulin receptor substrate-1 (IRS-1) phosphorylation in neural-derived blood exosomes from people with Alzheimer’s disease, up to a decade before diagnosis [3].
- Neuroprotective signaling. In cell and animal models, GLP-1 receptor activation stimulates the PI3K–Akt pathway, a central regulator of neuronal survival, synaptic plasticity and autophagy. NeuroDex’s AKT panel page describes this pathway in more detail.
- Anti-inflammatory and metabolic effects. GLP-1 receptor agonists reduce systemic inflammation and improve metabolic health, both of which are risk factors for neurodegeneration.
These ideas have driven a wave of clinical trials. Yet the clinical results so far have been mixed, and that has made one gap in drug development very visible.
The problem: peripheral efficacy is not CNS target engagement
In a CNS drug program, target engagement means showing that a drug reaches its intended target in the brain and produces the expected biological change. It sits between pharmacokinetics (is the drug present?) and clinical efficacy (do patients improve?). Without it, a negative trial is very hard to interpret.
For GLP-1 receptor agonists this question is unusually difficult:
1. Limited and uneven brain access
GLP-1 receptor agonists are peptides, and most do not freely cross the blood-brain barrier. In mouse studies, fluorescently labeled liraglutide reached circumventricular organs and parts of the hypothalamus (regions with a leakier barrier), but showed little access to deeper brain structures [6]. A 2022 study in Cell Metabolism found that liraglutide did not cross brain endothelial cells in an in vitro barrier model; instead, it entered the hypothalamus through specialized cells called tanycytes [7]. These findings do not rule out central effects elsewhere in the brain, which could occur through indirect routes such as vagal signaling or reduced systemic inflammation. But they mean that a drug’s metabolic benefits tell you very little about what happens in the cortex, hippocampus or substantia nigra.
2. Recent large trials were negative, and hard to interpret
In the phase 3 Exenatide-PD3 trial, 194 people with Parkinson’s disease received weekly exenatide or placebo for 96 weeks. The drug showed no advantage over placebo on motor scores, patient-reported symptoms or brain imaging [5]. In late 2025, Novo Nordisk reported that the EVOKE and EVOKE+ phase 3 trials of oral semaglutide in 3,808 people with early Alzheimer’s disease did not slow progression on the primary endpoint (CDR-SB), even though some biomarkers moved in the expected direction [8].
These were well-run trials. But they leave sponsors with a difficult question: did the drugs engage neuronal signaling pathways and still fail to change the disease, or did they never reach the brain at a sufficient level in enough patients? Each answer points to a very different next step: a new mechanism, a different molecule with better brain penetration, a different dose, or a better-selected patient population.
3. Traditional tools are invasive, indirect or both
- Cerebrospinal fluid (CSF) sampling requires lumbar puncture, which limits how often it can be repeated and how many patients will agree to it. CSF drug levels also show exposure, not intracellular signaling.
- PET imaging is expensive, requires a suitable radiotracer and is rarely practical for repeated measures in large trials.
- Standard plasma biomarkers such as neurofilament light (NfL) reflect neuronal damage. They are valuable for tracking neurodegeneration but are not designed to show whether a specific signaling pathway inside neurons has been switched on.
- Peripheral metabolic markers such as glucose, insulin and HbA1c show that the drug is working in the body, but say nothing direct about the brain.
A window into neurons: neuron-derived extracellular vesicles
Extracellular vesicles (EVs), including exosomes, are small membrane-bound particles released by almost every cell type. They carry proteins, lipids and RNA from the cell that released them. Some EVs released by neurons cross into the bloodstream, carrying a sample of the neuron’s internal molecular state with them.
By isolating the subpopulation of blood EVs that come from neurons, researchers can measure intracellular proteins, including phosphorylated signaling proteins that never appear in plasma on their own. In 2017, researchers at the US National Institute on Aging described methods for enriching plasma EVs of neuronal origin and called them “a potential window into brain pathologic processes” [4].
The field has also had healthy debate. A 2021 paper in Nature Methods argued that L1CAM, the surface protein used in many early studies to capture neuronal EVs, is mostly present in a soluble form in plasma rather than on vesicles [9]. This prompted careful work on capture specificity and validation across the field. NeuroDex’s ExoSORT™ platform uses a proprietary immunoaffinity method developed to enrich neuron-derived vesicles from a standard blood sample, with assays run under CLIA and GLP standards.
Evidence that NDEV biomarkers can detect GLP-1 target engagement
Exenatide in Parkinson’s disease (2019)
The first human evidence came from a secondary analysis of the phase 2 Exenatide-PD trial, published in JAMA Neurology in 2019 [2]. Researchers measured insulin-signaling proteins in neuronal-derived exosomes from 60 patients. Compared with placebo, patients receiving exenatide showed:
- higher tyrosine phosphorylation of IRS-1 at 48 and 60 weeks,
- higher levels of total Akt and phosphorylated mTOR, and
- an association between mTOR levels and improvement in off-medication motor scores.
The authors concluded that neuronal exosome markers could help assess whether a drug engages its target in clinical trials, while noting that larger studies were needed.
Liraglutide and pioglitazone: the 2026 JCEM study
In 2026, a team including researchers from Stanford University and NeuroDex published a study in The Journal of Clinical Endocrinology & Metabolism that took this approach further [1]. The team analyzed stored samples from two placebo-controlled clinical trials in which drugs had already been shown to improve peripheral insulin resistance:
- Pioglitazone (a PPARγ agonist) in adults with depression that had not remitted after six months of usual treatment.
- Liraglutide (a GLP-1 receptor agonist) in middle-aged adults at genetic risk for Alzheimer’s disease.
Eleven proteins in the Akt–mTOR pathway were measured in NDEVs before and after 12 weeks of treatment. The key findings:
- Liraglutide significantly increased phosphorylated Akt (Ser473) and phosphorylated mTOR (Ser2448) compared with placebo. About 40% of participants showed an Akt response and 30% an mTOR response.
- Pioglitazone significantly increased phosphorylated GSK3β (Ser9), mTOR (Ser2448) and ribosomal protein S6, with 77% of participants showing an mTOR response.
- The NDEV responses were relatively independent of changes in fasting insulin and glucose. In other words, the central signal was not simply a reflection of improved peripheral metabolism.
The authors described these as CNS-specific biomarker responses to both drug classes. The cohorts were small (12 participants on pioglitazone and 15 on liraglutide), so the results should be viewed as a proof of concept that needs confirmation in larger studies.
What these findings mean for drug developers
Three points stand out:
- Central engagement can be detected from plasma, even for a drug with limited brain penetration. Liraglutide’s limited access to the brain did not prevent a measurable change in neuronal Akt–mTOR signaling.
- Not every patient responds. Response rates of 30–40% on liraglutide suggest that CNS target engagement varies between individuals. That variation may help explain why average clinical effects in large trials can be small or absent.
- Peripheral markers are not a reliable stand-in. Because NDEV changes were largely decoupled from glucose and insulin, a sponsor cannot assume that good metabolic control equals central drug activity.
How to use NDEV biomarkers in a GLP-1 CNS program
For teams developing GLP-1 receptor agonists, dual agonists or other metabolic drugs for brain indications, NDEV biomarkers can support decisions at several stages.
Phase 1 and early phase 2: dose selection
Peripheral pharmacodynamics usually guide dose selection for metabolic drugs. For a CNS indication, the dose that maximizes weight loss may not be the dose that best engages neuronal signaling. Serial NDEV measurements of phosphorylated Akt, mTOR, GSK3β and related proteins can help build a central dose-response curve from the same blood draws already collected for safety and PK.
Comparing molecules and formulations
Newer GLP-1-based drugs differ in size, half-life, receptor targets and likely brain access. NDEV pharmacodynamic markers offer a way to compare candidates on central activity, not just metabolic potency, before committing to a large efficacy trial.
Responder identification and enrichment
If only a subset of patients shows central engagement, early NDEV changes may help identify likely responders. The JCEM study also suggested that baseline Akt (Ser473) levels might predict response, an early finding that merits prospective testing. Biomarker-defined subgroups could support adaptive designs or enrichment strategies in later-phase trials.
Interpreting negative or ambiguous results
When a trial misses its clinical endpoint, NDEV data from banked samples can help separate “the drug did not engage the target” from “the target did not change the disease.” That distinction is critical for deciding whether to stop a program, change dose, or move to a different patient population.
Practical design considerations
- Sample collection: NDEV analysis uses standard plasma or serum. Collect baseline and on-treatment samples at consistent times relative to dosing and meals.
- Banked samples: Many completed trials have stored plasma. Retrospective NDEV analysis, as in the JCEM study, can add mechanistic insight without new enrollment.
- Analyte selection: For GLP-1 programs, a panel covering IRS-1, Akt, mTOR, GSK3β and downstream targets such as ribosomal protein S6 gives a fuller picture of pathway activity than a single marker.
- Combine with disease markers: Pair pathway markers with disease-relevant NDEV or plasma markers (for example tau, α-synuclein or NfL) to link central engagement to downstream biology.
- Regulatory-grade data: Choose a laboratory that runs assays under GLP and CLIA standards, with documented precision and batch-to-batch reproducibility.
Beyond GLP-1: other drugs with a blood-brain barrier question
The same approach applies to any therapy where brain exposure is limited or uncertain: insulin sensitizers, intranasal insulin, anti-inflammatory biologics, antibodies, and peptide or oligonucleotide drugs. Wherever a sponsor needs to show that a peripherally administered drug changes neuronal biology, NDEV pharmacodynamic biomarkers offer a minimally invasive option that can be repeated as often as blood can be drawn.
Frequently asked questions
What is CNS target engagement?
CNS target engagement is evidence that a drug reaches its intended target in the central nervous system and produces the expected biological effect. It is distinct from showing that the drug is present in the body (pharmacokinetics) or that patients improve (clinical efficacy).
Do GLP-1 receptor agonists cross the blood-brain barrier?
Most GLP-1 receptor agonists are peptides with limited ability to cross the blood-brain barrier. Animal studies show they reach circumventricular organs and parts of the hypothalamus, but have limited direct access to deeper brain regions. Central effects may also occur indirectly, which is why direct measurement of neuronal signaling is useful.
How can you measure GLP-1 effects on the brain from a blood test?
One approach is to isolate neuron-derived extracellular vesicles from plasma and measure insulin-signaling proteins inside them, such as phosphorylated Akt and mTOR. Changes in these proteins after treatment indicate that neuronal signaling has been affected by the drug.
What did the 2026 JCEM study find?
The study found that 12 weeks of liraglutide increased phosphorylated Akt and mTOR in neuron-derived extracellular vesicles compared with placebo, and that pioglitazone increased phosphorylated GSK3β, mTOR and ribosomal protein S6. These central changes were largely independent of changes in peripheral insulin and glucose.
Can NDEV biomarkers be measured in samples from completed trials?
Yes. NDEV analysis can be performed on properly stored plasma or serum, so sponsors can apply it retrospectively to banked samples to explore target engagement and responder subgroups.
Work with NeuroDex
NeuroDex provides CLIA-certified, GLP-compliant neuron-derived extracellular vesicle biomarker services for preclinical and clinical programs, including Akt–mTOR pathway panels for metabolic and CNS drugs. Learn more about our clinical trial services, browse our publications, or contact our team to discuss your program.
References
- Evers A, Watson K, Abbasi F, Haque K, Verma A, Eitan E, Rasgon N. Insights from changes in NDEV biomarkers of metabolism: effects of PPARγ and GLP1 receptor agonists on brain metabolism. J Clin Endocrinol Metab. 2026;111(10):2871–2882. doi:10.1210/clinem/dgag176
- Athauda D, Gulyani S, Karnati HK, et al. Utility of neuronal-derived exosomes to examine molecular mechanisms that affect motor function in patients with Parkinson disease: a secondary analysis of the Exenatide-PD trial. JAMA Neurol. 2019;76(4):420–429. doi:10.1001/jamaneurol.2018.4304
- Kapogiannis D, Boxer A, Schwartz JB, et al. Dysfunctionally phosphorylated type 1 insulin receptor substrate in neural-derived blood exosomes of preclinical Alzheimer’s disease. FASEB J. 2015;29(2):589–596. doi:10.1096/fj.14-262048
- Mustapic M, Eitan E, Werner JK Jr, et al. Plasma extracellular vesicles enriched for neuronal origin: a potential window into brain pathologic processes. Front Neurosci. 2017;11:278. doi:10.3389/fnins.2017.00278
- Vijiaratnam N, Girges C, Auld G, et al. Exenatide once a week versus placebo as a potential disease-modifying treatment for people with Parkinson’s disease in the UK: a phase 3, multicentre, double-blind, parallel-group, randomised, placebo-controlled trial. Lancet. 2025. doi:10.1016/S0140-6736(24)02808-3
- Secher A, Jelsing J, Baquero AF, et al. The arcuate nucleus mediates GLP-1 receptor agonist liraglutide-dependent weight loss. J Clin Invest. 2014;124(10):4473–4488. doi:10.1172/JCI75276
- Imbernon M, Saponaro C, Helms HCC, et al. Tanycytes control hypothalamic liraglutide uptake and its anti-obesity actions. Cell Metab. 2022;34(7):1054–1063.e7. doi:10.1016/j.cmet.2022.06.002
- Alzheimer Europe. Results of EVOKE and EVOKE+ trials show no effect of oral semaglutide on AD progression. December 2025. alzheimer-europe.org
- Norman M, Ter-Ovanesyan D, Trieu W, et al. L1CAM is not associated with extracellular vesicles in human cerebrospinal fluid or plasma. Nat Methods. 2021;18(6):631–634. doi:10.1038/s41592-021-01174-8
Leave a Reply