Delivering synaptic protein mRNAs via extracellular vesicles ameliorates cognitive impairment in a mouse model of Alzheimer's disease

Abstract

Background: Synaptic dysfunction with reduced synaptic protein levels is a core feature of Alzheimer's disease (AD). Synaptic proteins play a central role in memory processing, learning, and AD pathogenesis. Evidence suggests that synaptic proteins in plasma neuronal-derived extracellular vesicles (EVs) are reduced in patients with AD. However, it remains unclear whether levels of synaptic proteins in EVs are associated with hippocampal atrophy of AD and whether upregulating the expression of these synaptic proteins has a beneficial effect on AD.

Methods: In this study, we included 57 patients with AD and 56 healthy controls. We evaluated their brain atrophy through magnetic resonance imaging using the medial temporal lobe atrophy score. We measured the levels of four synaptic proteins, including synaptosome-associated protein 25 (SNAP25), growth-associated protein 43 (GAP43), neurogranin, and synaptotagmin 1 in both plasma neuronal-derived EVs and cerebrospinal fluid (CSF). We further examined the association of synaptic protein levels with brain atrophy. We also evaluated the levels of these synaptic proteins in the brains of 5×FAD mice. Then, we loaded rabies virus glycoprotein-engineered EVs with messenger RNAs (mRNAs) encoding GAP43 and SNAP25 and administered these EVs to 5×FAD mice. After treatment, synaptic proteins, dendritic density, and cognitive function were evaluated.

Results: The results showed that GAP43, SNAP25, neurogranin, and synaptotagmin 1 were decreased in neuronal-derived EVs but increased in CSF in patients with AD, and the changes corresponded to the severity of brain atrophy. GAP43 and SNAP25 were decreased in the brains of 5×FAD mice. The engineered EVs efficiently and stably delivered these synaptic proteins to the brain, where synaptic protein levels were markedly upregulated. Upregulation of synaptic protein expression could ameliorate cognitive impairment in AD by promoting dendritic density. This marks the first successful delivery of synaptic protein mRNAs via EVs in AD mice, yielding remarkable therapeutic effects.

Conclusions: Synaptic proteins are closely related to AD processes. Delivery of synaptic protein mRNAs via EVs stands as a promising effective precision treatment strategy for AD, which significantly advances the current understanding of therapeutic approaches for the disease.

Keywords: Alzheimer’s disease; Extracellular vesicles; Growth-associated protein 43; Messenger RNAs; Synaptic dysfunction; Synaptosome-associated protein 25.

Fig. 1

Representative MRI of MTA score rating. Representative T1-weighted images in the coronal plane depicting MTA scores of 2 (A), 3 (B), and 4 (C). MRI, magnetic resonance image; MTA, medial temporal lobe atrophy

Fig. 2

Neuronal-derived EV levels of synaptic proteins in patients with AD. Neuronal-derived EV levels of GAP43 (A), neurogranin (B), SNAP25 (C), and synaptotagmin 1 (D) were measured in patients with AD with MTA scores of 2, 3, and 4. A decreasing trend in the mean of synaptic proteins across MTA scores is depicted by the red lines. * Bonferroni-corrected P = 2.05 × 10⁻² (A), 4.27 × 10⁻² (B), 2.06 × 10⁻² (C), and 2.27 × 10⁻² (D) compared to MTA = 2 from one way ANOVA test. # Bonferroni-corrected P = 9.31 × 10⁻⁴ (A), 7.47 × 10⁻⁵ (B), 1.07 × 10⁻³ (C), and 1.22 × 10⁻³ (D) compared to MTA = 3 from one way ANOVA test. AD, Alzheimer’s disease; ANOVA, analysis of variance; EV, extracellular vesicle; GAP43, growth-associated protein 43; MTA, medial temporal lobe atrophy; SNAP25, synaptosome-associated protein 25

Fig. 3

CSF levels of synaptic proteins in patients with AD. CSF levels of GAP43 (A), neurogranin (B), SNAP25 (C), and synaptotagmin 1 (D) were measured in patients with AD with MTA scores of 2, 3, and 4. An increasing trend in the mean of synaptic proteins across MTA scores is depicted by the red lines. * Bonferroni-corrected P = 2.11 × 10⁻² (A), P = 1.79 × 10⁻² (B), 1.56 × 10⁻² (C), and 3.23 × 10⁻³ (D) compared to MTA=2 from one way ANOVA test. # Bonferroni-corrected P = 4.87 × 10⁻⁴ (A), 1.38 × 10⁻⁵ (B), 2.52 × 10⁻⁴ (C), and 1.65 × 10⁻⁵ (D) compared to MTA = 3 from one way ANOVA test. AD, Alzheimer’s disease; ANOVA, analysis of variance; CSF, cerebrospinal fluid; GAP43, growth-associated protein 43; MTA, medial temporal lobe atrophy; SNAP25, synaptosome-associated protein 25

Fig. 4

Reduction of GAP43 and SNAP25 levels in the hippocampus of 5×FAD mouse brain. A-D Gap43 (A), Snap25 (B), neurogranin (C), and synaptotagmin 1(D) mRNA levels quantified using RT-qPCR. E–I Representative Western blots (E) and quantitative assessment (F–I) of GAP43 (E and F), SNAP25 (E and G), neurogranin (E and H), and synaptotagmin 1 (E and I) protein levels. Group comparison was performed using t-test. GAP43, growth-associated protein 43; Ng, neurogranin; mRNA, messenger RNA; RT-qPCR, real-time quantitative reverse transcription-polymerase chain reaction; SNAP25, synaptosome-associated protein 25; Syn 1, synaptotagmin 1; WT, wild-type

Fig. 5

Construction of a neural-cell-targeting EV system loaded with mRNAs of Gap43 and Snap25. A Schematic representation of generation, collection, and administration of neural-cell-targeting EVs overexpressing Gap43 and Snap25 mRNAs. B HA pulldown assays and representative Western blots of LAMP-2B in HEK293T cells transfected with HA-RVG-LAMP-2B plasmid and their corresponding EVs. C RT-qPCR of HEK293T cells transfected with HA-RVG-LAMP-2B detects the mRNA of Rvg. D–G Quantitative assessment of Gap43 (D and F) and Snap25 (E and G) in HEK293T cells (D and E) and EVs (F and G) using RT-qPCR. H HA pulldown assays and representative Western blots of 5×FAD hippocampal tissues treated with EVs-TNGS. n = 3. I–K Representative Western blots (I) and their quantitative assessment (J and K) of protein levels of GAP43 (I and J) and SNAP25 (I and K) in treated 5×FAD mouse brains. EVs, extracellular vesicles; EVs-TN, extracellular vesicles targeting neural cells; EVs-TNGS, extracellular vesicles targeting neural cells overexpressing Gap43 and Snap25; GAP43, growth-associated protein 43; mRNAs, messenger RNAs; SNAP25, synaptosome-associated protein 25; RT-qPCR, real-time quantitative reverse transcription-polymerase chain reaction; RVG, rabies virus glycoprotein

Fig. 6

Characterization of EVs. A, B Size distribution of wild-type (A, 97.3 ± 42.8 nm) and transfected EVs (B, 99.5 ± 41.0 nm) measured using nanoparticle tracking analysis. C Protein levels of the EV markers Alix, TSG101, CD63, and APOB measured using Western blot analysis. D Transmission electron microscope image showing the distinctive spherical-shaped morphology of EVs. Scale bar = 100 nm. APOB, apolipoprotein B; EVs, extracellular vesicles; EVs-TNGS, extracellular vesicles targeting neural cells overexpressing Gap43 and Snap25; TSG101, tumor susceptibility gene 101

Fig. 7

Improvement of memory performance in 5×FAD mice following EV treatment. A Representative trajectories of each group in the probe test of MWM. B NOR test was performed 1 month after treatment. C NOL test was performed 1 month after treatment. D The parameters of escape latency were recorded for 5 days to assess the cognitive function of mice. E The time spent in the correct quadrant was recorded in the probe test of MWM. * P < 0.05, 5×FAD+EVs-TNGS compared to 5×FAD+EVs. EVs, extracellular vesicles; EVs-TNGS, extracellular vesicles targeting neural cells overexpressing growth-associated protein 43 and synaptosome-associated protein 25; MWM, Morris water maze; NOR, novel object recognition; NOL, novel object location; WT, wild-type

Fig. 8

Increase in levels of GAP43 and SNAP25 in the 5×FAD mouse brains treated with EVs-TNGS. A Upregulation of Gap43 mRNA was detected using RT-qPCR in EV-TNGS-treated hippocampal tissues. B Upregulation of Snap25 mRNA was detected using RT-qPCR in EV-TNGS-treated hippocampal tissues. C–E Representative Western blots (C) and their quantitative assessment (D and E) show protein levels of GAP43 (C and D) and SNAP25 (C and E) in EV-TNGS-treated hippocampal tissues. EVs, extracellular vesicles; EVs-TNGS, extracellular vesicles targeting neural cells overexpressing Gap43 and Snap25; GAP43, growth-associated protein 43; RT-qPCR, real-time quantitative reverse transcription-polymerase chain reaction; SNAP25, synaptosome-associated protein 25; WT, wild-type

Fig. 9

Increase in dendritic density in 5×FAD mouse brains treated with EVs-TNGS. Representative Golgi stains (A) and their quantitative data (B and C) show total dendritic length (B) and neuron branch number (C) in the hippocampus. Scale bar = 100 μm. EVs, extracellular vesicles; EVs-TNGS, extracellular vesicles targeting neural cells overexpressing growth-associated protein 43 and synaptosome-associated protein 25; WT, wild-type