SEARCHOTHERophthalmology, retina

other·ophthalmology, retina, neuro-ophthalmology, neurology, clinical trial·PMC10033723

New AAV tools fail to detect Neurod1-mediated neuronal conversion of Müller glia and astrocytes in vivo

eBioMedicine · 5 authors, 4 centres

AI SUMMARY

FIDELITY 100%

POPULATIONAdult mice (C57BL/6J, Glast-CreERT;Rosa-CAG-LSL-Sun1-GFP, Aldh1l1-CreERT2;R26R-YFP) at 4-6 weeks (retina) or 8+ weeks (brain) of age

INTERVENTIONAAV-based delivery of Neurod1, Math5, Ascl1, or Neurog2 via intravitreal, subretinal, or intracerebral injection, using various AAV serotypes (ShH10, PHP.eB, AAV5) and promoters (GFAP, CAG-FLEX)

COMPARISONAAV-GFAP-GFP-P2A-mCherry (reporter only), AAV-GFAP-mCherry-P2A-TF (reversed transgene position), AAV-CAG-FLEX-mCherry (dual system), and different AAV doses (e11-e13 vg/ml)

This summary was generated by AI from a single paper. It has not been reviewed by a clinician and is not clinical advice. Verify against the source before acting on it.

AGENT 01 · STRUCTURED SUMMARY

TRACE TO SOURCE

“

TL;DR

This study demonstrates that AAV-GFAP-mediated delivery of Neurod1 and other reprogramming factors causes severe leaky expression into endogenous neurons, which can be misinterpreted as glia-to-neuron conversion. The authors developed two new AAV-based tools that significantly reduce this neuronal leakage. Using these tools with stringent lineage tracing, they found that Neurod1 gene transfer fails to convert Müller glia or astrocytes into neurons in vivo, challenging previous reports of successful glia-to-neuron reprogramming.

Full summary

4,178 CHARS

**Background:** In vivo reprogramming of resident glial cells (Müller glia in retina, astrocytes in brain) into neurons is a promising therapeutic strategy for neurodegenerative diseases. AAV-based gene delivery of transcription factors like Neurod1 has been widely used, but recent studies suggest that AAV-GFAP-mediated expression can leak into endogenous neurons, leading to misinterpretation of glia-to-neuron conversion. This study aimed to systematically evaluate neuronal leakage for four reprogramming factors (Neurod1, Math5, Ascl1, Neurog2) in the mouse retina and brain, and to develop new AAV tools with improved specificity.

**Methods:** The authors used AAV vectors with a short GFAP promoter to deliver Neurod1, Math5, Ascl1, or Neurog2 (each fused to mCherry via a P2A site) into the retina of wild-type and Sun1-GFP lineage tracing mice (Glast-CreERT;Rosa-CAG-LSL-Sun1-GFP). Intravitreal injections used ShH10 serotype (1-5e13 vg/ml), and subretinal injections used PHP.eB serotype. For brain studies, AAV5 was injected into the cortex or striatum of Aldh1l1-CreERT2;R26R-YFP mice. Two new AAV tools were developed: (1) reversing the transgene order (mCherry-P2A-TF instead of TF-P2A-mCherry) to distance the reprogramming factor from the GFAP promoter, and (2) an AAV-CAG-FLEX dual system requiring Cre recombination for transgene expression. Retinas and brains were analyzed 3-6 weeks post-injection by immunohistochemistry for neuronal markers (HuC/D, Rbpms, NeuN), glial marker (Sox9), and lineage reporters (Sun1-GFP, YFP).

**Key Results:** - AAV-GFAP-mediated delivery of Neurod1-P2A-mCherry resulted in 80.2% of mCherry+ cells expressing HuC/D and 9.4% expressing Rbpms, but none of these cells were traced with Sun1-GFP, indicating leaky expression into endogenous neurons rather than MG-to-neuron conversion. Similarly, Math5-P2A-mCherry showed 80.3% HuC/D+ and 27.8% Rbpms+ cells, all non-MG derived. Ascl1 and Neurog2 showed milder leakage (31.0% and 9.9% HuC/D+, respectively), but again no lineage-traced MG were converted. - Lowering AAV titers 10-fold (to e12 vg/ml) did not correct leakage: Neurod1 still showed 75.0% HuC/D+ cells, Math5 71.1%, Ascl1 39.1%, Neurog2 6.8%. None were Sun1-GFP+. - Changing serotype to PHP.eB and injection route to subretinal shifted leakage patterns: Neurod1 showed 10.1% HuC/D+ cells (inner retina) but also leakage into photoreceptors; Math5 showed 27.8% HuC/D+ and leakage into horizontal cells. - Reversing transgene order (mCherry-P2A-TF) dramatically reduced leakage: Neurod1 dropped to 10.3% HuC/D+, Math5 to 0.7%, Ascl1 to 2.8%, Neurog2 to 1.8%. Sox9 positivity increased to 88.4-97.9%. - The AAV-CAG-FLEX dual system also effectively reduced leakage: Neurod1 showed 3.1% HuC/D+, Math5 5.4%, Ascl1 1.1%, Neurog2 1.2%. None of the HuC/D+ cells were Sun1-GFP+. - At 6 weeks post-injection using the new tools, Neurod1 and Math5 failed to convert MG: 87.2% and 98.4% of mCherry+ cells were Sox9+, and Sun1-GFP+ cells remained Sox9+ (93.9% and 96.7%). No mCherry+Sox9- cells were Sun1-GFP+. - In the brain, AAV5-GFAP-mCherry-P2A-Neurod1 showed 63.7-66.9% mCherry+NeuN+ cells in cortex (leakage), but none were YFP+ (non-astrocyte origin). In striatum, leakage was lower (9.7-14.7% NeuN+), but again no YFP+ cells.

**Clinical Implications:** This study challenges the validity of previous reports claiming successful glia-to-neuron conversion using AAV-GFAP-mediated Neurod1 or Math5 delivery. The findings emphasize that stringent fate mapping (e.g., using Cre-lox lineage tracing) is essential to distinguish true reprogramming from artifact due to leaky AAV expression. The two new AAV tools developed here—reversed transgene order and the CAG-FLEX dual system—provide improved specificity for glial targeting and will be valuable for future in vivo reprogramming studies. However, even with these tools, Neurod1 failed to convert glia to neurons, suggesting that additional factors or conditions (e.g., injury, combinatorial transcription factors) may be required. These results have significant implications for the development of gene therapies aimed at neuronal regeneration in the retina and brain.

PICO

PPOPULATION

Adult mice (C57BL/6J, Glast-CreERT;Rosa-CAG-LSL-Sun1-GFP, Aldh1l1-CreERT2;R26R-YFP) at 4-6 weeks (retina) or 8+ weeks (brain) of age

IINTERVENTION

AAV-based delivery of Neurod1, Math5, Ascl1, or Neurog2 via intravitreal, subretinal, or intracerebral injection, using various AAV serotypes (ShH10, PHP.eB, AAV5) and promoters (GFAP, CAG-FLEX)

OOUTCOME

Neuronal leakage (percentage of AAV-transduced cells expressing neuronal markers HuC/D, Rbpms, NeuN), glial identity (Sox9 expression), and lineage tracing (Sun1-GFP or YFP co-localization)

STUDY TYPE

other

SPECIALTY

ophthalmology

SUMMARISED BY

AI pipeline

FIDELITY CHECK

100% · A

ORIGINAL PAPER

PMC10033723

New AAV tools fail to detect Neurod1-mediated neuronal conversion of Müller glia and astrocytes in vivo

eBioMedicine

READ THE FULL PAPER

JUMP TO SECTION

ABSTRACT

Summary Background Reprogramming resident glial cells to convert them into neurons in vivo represents a potential therapeutic strategy that could replenish lost neurons, repair damaged neural circuits, and restore function. AAV (adeno-associated virus)-based expression systems are powerful tools for in vivo gene delivery in glia-to-neuron reprogramming, however, recent studies show that AAV-based gene delivery of Neurod1 into the mouse brain can cause severe leaky expression into endogenous neurons leading to misinterpretation of glia-to-neuron conversion. Methods AAV-based delivery systems were modified for improved in vivo delivery of Neurod1, Math5, Ascl1, and Neurog2 in the adult mouse retina and brain. To examine whether bona fide glia-to-neuron conversion occurs, stringent fate mapping experiments were performed to trace the lineage of glial cells. Findings The neuronal leakage is prevalent after AAV- GFAP -mediated delivery of Neurod1, Math5, Ascl1, and Neurog2. The transgene-dependent leakage cannot be corrected after lowering the AAV doses, using alterative AAV serotypes or injection routes. Importantly, we report the development of two new AAV-based tools that can significantly reduce neuronal leakage. Using the new AAV-based tools, we provide evidence that Neurod1 gene transfer fails to convert lineage traced glial cells into neurons. Interpretation Stringent fate mapping techniques independently of an AAV-based expression system are the golden standard for tracing the fate of glia cells during neuronal reprogramming. The newly developed AAV-based systems are invaluable tools for glia-to-neuron reprogramming in vivo . Funding The work in Chen lab was supported by National Institutes of Health (NIH) grants R01 EY024986 and R01 EY028921, an unrestricted challenge grant from Research to Prevent Blindness, the New York Eye and Ear Infirmary Foundation, and The Harold W. McGraw, Jr. Family Foundation for Vision Research. The work in Zhang lab was supported by NIH (R01 NS127375 and R01 NS117065) and The Decherd Foundation. Keywords: AAV, Gene delivery, Reprogramming factors, Glial cells, Glia-to-neuron reprogramming

CITATION EXPORT

Ye Xie, Jing Zhou, Lei-Lei Wang et al.. New AAV tools fail to detect Neurod1-mediated neuronal conversion of Müller glia and astrocytes in vivo. eBioMedicine. (2023). https://doi.org/10.1016/j.ebiom.2023.104531

AUTHORS

5 · 4 CENTRES

Ye Xie, Jing Zhou, Lei-Lei Wang, Chun-Li Zhang, Bo Chen

aDepartment of Ophthalmology, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
bDepartment of Molecular Biology and Hamon Center for Regenerative Science and Medicine, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA
cDepartment of Neuroscience, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA
dDepartment of Cell, Developmental and Regenerative Biology, Icahn School of Medicine at Mount Sinai, New York, NY 10029, USA