Comprehensive investigation of AAV tropism across human iPSC-derived neuronal subtypes

This study systematically evaluates the transduction efficiency and toxicity of 18 AAV serotypes across three distinct human iPSC-derived neuronal subtypes and cerebellar organoids, identifying AAV6, AAV6.2, and AAV2.7m8 as robust vectors while establishing a critical benchmark for optimizing gene delivery in human neural models.

The Gist

Imagine you are trying to deliver a very important package (a gene therapy) to a specific neighborhood in a giant, complex city (the human brain). For years, scientists have been testing different delivery trucks (viruses called AAVs) to see which ones work best. But here's the catch: they've mostly been testing these trucks on mouse cities, not human cities.

The problem is that human brain cells are like a different species of city entirely. What works perfectly in a mouse neighborhood might get lost, break down, or even cause a traffic accident in a human one.

This paper is like a massive, real-world test drive where scientists took 18 different delivery trucks and tried to deliver packages to three specific types of human "houses" (neurons) grown in a lab:

1. Cortical neurons (the city planners and thinkers).
2. NGN2 neurons (a specific type of forebrain cell, very delicate).
3. Dopaminergic neurons (the mood and movement specialists).

The Delivery Test

The scientists didn't just drop one package; they tried different amounts (doses) and watched closely with high-powered cameras to see:

• Did the truck get in? (Transduction efficiency)
• Did the package get delivered? (Gene expression)
• Did the truck crash the house? (Toxicity)

The Results: Finding the Best Trucks

After running this massive experiment, they found that not all trucks are created equal.

• The Winners: Three specific trucks—AAV6, AAV6.2, and AAV2.7m8—were the "Super Delivery Vans." They managed to get into all three types of human houses efficiently and reliably.
• The "Goldilocks" Truck: Among the winners, AAV2.7m8 was highlighted as a top contender, especially because it worked well even in a more complex, 3D "mini-city" (a brain organoid), proving it can navigate crowded, real-world environments.

The Safety Check: Don't Break the House

Just because a truck gets in doesn't mean it's safe. The scientists checked if the delivery caused damage to the houses.

• The Sensitive House: The NGN2 neurons were like a house made of glass; they were very fragile and got damaged easily if the truck was too big or the package too heavy.
• The Tough House: The Dopaminergic neurons were like a sturdy brick house; they could handle the delivery trucks much better without getting hurt.

Why This Matters

Think of this paper as a new "User Manual" for human brain research. Before this, scientists were guessing which truck to use, often relying on data from mice that didn't translate well to humans.

Now, researchers have a clear map. They know exactly which viral "truck" to pick to safely deliver genetic medicine to specific human brain cells without destroying them. This is a huge step forward for:

1. Drug Discovery: Testing new medicines on human cells that actually work.
2. Gene Therapy: Developing cures for brain diseases that are safe and effective for real humans, not just lab mice.

In short, this study took the guesswork out of delivering genetic packages to human brains, ensuring we use the right vehicle for the right job so we don't crash the house while trying to fix it.

Technical Summary

1. Problem Statement

Recombinant Adeno-associated viruses (AAVs) are the primary vectors for gene delivery in the central nervous system (CNS), serving as critical tools for both gene therapy and basic neuroscience. However, a significant knowledge gap exists regarding their efficacy in human neurons. While AAV serotypes have been extensively characterized in rodent models, their performance in human induced pluripotent stem cell (iPSC)-derived neurons remains poorly understood. This lack of characterization creates unpredictability in transduction efficiency and toxicity, hindering the reliability of human iPSC models for disease modeling and drug screening.

2. Methodology

The study employed a systematic, high-throughput approach to evaluate AAV performance across diverse human neuronal models:

• Viral Vectors: The researchers tested 18 distinct AAV serotypes, comprising both wild-type and engineered variants.
• Cellular Models: Three specific types of iPSC-derived neurons were utilized to represent different disease-relevant contexts:
  1. Cortical projection neurons.
  2. NGN2-induced forebrain-like neurons.
  3. Dopaminergic neurons.
• Experimental Conditions: Transduction was assessed across four viral doses (1×10³, 1×10⁴, 1×10⁵, and 2×10⁵ genome copies per cell).
• Data Acquisition: Automated high-throughput confocal imaging was used to quantify reporter gene expression for transduction efficiency.
• Safety Assessment: Toxicity was evaluated by quantifying cell counts and analyzing neurite morphology.
• Validation: To ensure physiological relevance, the top-performing serotype was further validated in a complex 3D human cerebellar organoid model (tested in both whole and dissociated states).

3. Key Contributions

• Systematic Benchmarking: This study provides the first comprehensive dataset comparing a wide array of AAV serotypes specifically within human iPSC-derived neuronal subtypes, moving beyond rodent-centric data.
• Dose-Response Characterization: It establishes how transduction efficiency and toxicity scale across a wide range of viral doses in human cells.
• Safety-Efficacy Correlation: The research explicitly links high transduction rates with potential cytotoxicity, offering a nuanced view of the safety profile for different neuronal subtypes.
• 3D Model Validation: By extending findings to 3D cerebellar organoids, the study bridges the gap between 2D monolayer cultures and complex tissue environments.

4. Key Results

• Top-Performing Serotypes: Among the 18 serotypes tested, AAV6, AAV6.2, and AAV2.7m8 demonstrated consistently robust and efficient transduction across all three neuronal subtypes.
• Subtype-Specific Sensitivity:
  • NGN2-induced neurons were identified as the most vulnerable to AAV-induced toxicity.
  • Dopaminergic neurons showed the highest resilience to viral transduction.
• Toxicity Correlation: A positive correlation was observed between high transduction efficiency/gene expression and cellular toxicity, though the magnitude of this effect varied significantly by cell type.
• Organoid Validation: The high performance of AAV2.7m8 was successfully replicated in human cerebellar organoids, confirming its utility in more complex, physiologically relevant 3D structures.

5. Significance

This work establishes a critical benchmark dataset for AAV tropism in human in vitro neural models. Its significance lies in:

• Guiding Preclinical Research: Providing practical guidelines for selecting the optimal AAV serotype and dose for specific human neuronal subtypes, thereby improving the reliability of disease modeling and drug discovery pipelines.
• Enhancing Gene Therapy Development: Offering essential safety and efficacy data that can inform the design of future gene therapies targeting human CNS disorders, reducing the risk of failure due to poor vector performance in human cells.
• Resource for the Community: Serving as a foundational resource for researchers aiming to manipulate gene expression in human neurons, facilitating more accurate translation from in vitro models to clinical applications.