AAV Tropism in Human iPSC-Derived Neurons: Selecting the Right Serotype for CNS Gene Delivery Research

Adeno-associated virus, or AAV, has become one of the most important gene delivery platforms in neuroscience, functional genomics, and gene therapy research. Its ability to deliver genetic payloads into neurons and other central nervous system, or CNS, cell types has made it a widely used tool for gene expression, gene silencing, CRISPR delivery, neuronal labeling, biosensor expression, optogenetics, chemogenetics, disease modeling, and preclinical therapeutic development.

However, AAV performance is not universal. Different AAV serotypes and engineered capsids can show strikingly different transduction profiles depending on species, tissue, cell type, delivery route, promoter, dose, and experimental model. This is especially important in CNS research, where many widely used AAV serotypes were first characterized in rodent systems. A capsid that performs well in mouse brain does not necessarily perform well in human neurons.

This translational gap has become increasingly important as human induced pluripotent stem cell, or iPSC-derived, neuronal models are used more widely in neurodegenerative disease research. Human iPSC-derived neurons can model disease-relevant genetic backgrounds and neuronal subtypes, including cortical neurons, dopaminergic neurons, motor neurons, and other specialized CNS cell types. They provide a human-cell-based platform for evaluating vector tropism, transgene expression, cytotoxicity, and therapeutic feasibility before moving into more complex preclinical systems.

A bioRxiv preprint, “Comprehensive investigation of AAV tropism across human iPSC-derived neuronal subtypes,” provides an important benchmark dataset for this question. The study systematically evaluated 18 wild-type and engineered AAV serotypes across three human iPSC-derived neuronal subtypes: cortical projection neurons, NGN2-induced forebrain-like neurons, and dopaminergic neurons. By combining transduction efficiency, fluorescence intensity, cell number, neurite morphology, and 3D organoid validation, the study offers practical guidance for AAV selection in human neuronal models.

Why AAV Serotype Selection Matters in Human Neuronal Models

AAV serotype selection is one of the most important variables in CNS gene delivery. The capsid influences which cells are transduced, how efficiently vector genomes enter cells, how strongly the transgene is expressed, and how much vector is needed to achieve the desired biological effect.

For human neuronal research, choosing the wrong AAV serotype can lead to several problems. Low-efficiency serotypes may require high doses, which can increase stress, toxicity, or cost. Strong but poorly controlled expression may cause artificial phenotypes unrelated to the biological question. A serotype selected from rodent data alone may fail in human iPSC-derived neurons, leading to weak expression and inconsistent experimental results.

AAV serotype screening can help researchers answer practical questions such as:

• Which AAV serotype gives the strongest expression in the target human neuronal subtype?
• Does the same AAV capsid work across cortical, dopaminergic, and induced neuronal models?
• Is the most efficient capsid also the safest for the cell type being studied?
• Can a serotype identified in 2D neurons also work in 3D brain organoids?
• Do rodent CNS-tropic capsids translate to human neuronal systems?
• Which capsid should be used for overexpression, knockdown, CRISPR editing, reporter delivery, or disease modeling?

These questions are central to both basic neuroscience and translational gene therapy research.

Human iPSC-Derived Neurons as a More Relevant AAV Screening Platform

Human iPSC-derived neurons offer a valuable bridge between simple cell culture and animal models. They retain human genetic context and can be differentiated into disease-relevant neuronal subtypes. This makes them especially useful for studying disorders such as Parkinson’s disease, Huntington’s disease, amyotrophic lateral sclerosis, spinocerebellar ataxia, Alzheimer’s disease models, and other neurological conditions.

Compared with immortalized cell lines, human iPSC-derived neurons are more biologically relevant. Compared with animal models, they provide human-cell-specific information that may be missed in rodent systems. This is particularly important for AAV capsid screening because capsid tropism depends on interactions between the viral capsid and host-cell surface factors, intracellular trafficking pathways, and transcriptional environments.

Human iPSC-derived neuronal systems are not perfect substitutes for in vivo studies. They may lack full tissue architecture, immune interactions, mature vascular barriers, and long-range circuit organization. Still, they provide a powerful model for evaluating whether an AAV capsid can efficiently transduce the target human cell type before moving into animal studies or more advanced organoid systems.

Key Findings from the 18-Serotype AAV Screen

The study screened 18 AAV serotypes across three human iPSC-derived neuronal subtypes and four vector doses. The tested neuronal models included cortical projection neurons, NGN2-induced forebrain-like neurons, and dopaminergic neurons. Automated high-throughput confocal imaging was used to quantify GFP-positive cells and fluorescence intensity, while cell number and neurite morphology were used to evaluate potential toxicity.

The screen identified three broadly efficient serotypes across the tested human neuronal subtypes:

• AAV6
• AAV6.2
• AAV2.7m8

AAV6.2 showed especially strong transduction in cortical neurons and dopaminergic neurons, while AAV2.7m8 also performed strongly across multiple neuronal models. AAV6 emerged as a strong wild-type comparator with broad transduction activity.

This result is important because several serotypes commonly viewed as strong CNS vectors in rodent systems, including AAV9, AAV-PHP.eB, and AAV-PHP.S, showed comparatively weak performance in these human iPSC-derived neuronal models. This finding reinforces a key principle: AAV tropism is species-dependent and must be validated in relevant human models whenever possible.

AAV2.7m8: A Strong Candidate for Human CNS Model Systems

AAV2.7m8 was one of the strongest performers in the human neuronal screen and was further validated in 3D human iPSC-derived cerebellar organoids. This is especially meaningful because 3D organoids introduce additional delivery challenges that are not present in 2D monolayer culture, including tissue penetration, cell-type heterogeneity, diffusion gradients, and more complex architecture.

In whole-mount cerebellar organoid imaging, AAV2.7m8 produced broad GFP expression, with signal extending from the organoid surface toward the core. In dissociated organoid analysis, GFP-positive cells included multiple CNS-related cell types, such as neural progenitor cells, mature neurons, and astrocyte-lineage cells. This suggests that AAV2.7m8 may have broad utility in human CNS in vitro models.

Independent peer-reviewed organoid research has also identified AAV2.7m8 as a strong performer in human brain organoid systems. In a comparative study using cortical and cerebral organoids, AAV2.7m8 showed strong performance at both vector entry and transgene expression levels. This independent evidence strengthens confidence that AAV2.7m8 is not merely a single-study outlier, but a meaningful candidate for human CNS model screening.

For researchers working with human neuronal cultures or brain organoids, AAV2.7m8 may therefore be a useful starting point for applications such as reporter delivery, neuronal labeling, transgene expression, CRISPR component delivery, or disease model perturbation studies. However, it should still be validated for each specific cell type, promoter, payload, dose, and readout.

AAV2-Retro and Dopaminergic Neuron-Biased Transduction

The study also identified an interesting subtype-specific pattern: AAV2-retro showed strong transduction in dopaminergic neurons compared with some other neuronal subtypes. This is notable because dopaminergic neurons are highly relevant to Parkinson’s disease modeling and therapeutic research.

AAV2-retro was originally engineered to support efficient retrograde access to projection neurons. It has been widely used in neural circuit studies because it can label neurons based on their projection patterns. Its strong performance in dopaminergic neurons in this human iPSC-derived screen may make it a useful candidate for DA-neuron-focused applications, although this should not be interpreted as universal dopaminergic specificity across all systems.

For Parkinson’s disease research, this finding is especially useful. Human iPSC-derived dopaminergic neurons are commonly used to study disease-associated mutations, mitochondrial dysfunction, alpha-synuclein biology, neuronal survival, and therapeutic rescue strategies. AAV2-retro may provide a valuable option for gene delivery in this context, particularly when paired with appropriate validation and controls.

Why Rodent CNS AAV Data May Not Translate Directly to Human Neurons

One of the most important lessons from this study is that rodent AAV tropism data cannot be assumed to predict human neuronal performance. This is not only an empirical observation; it is supported by mechanistic evidence from the AAV field.

AAV-PHP.B and AAV-PHP.eB are well-known examples. These engineered capsids show strong CNS transduction in certain mouse strains, but their performance depends on species- and strain-specific host factors. Studies have shown that LY6A is an important receptor or attachment factor for the enhanced CNS tropism of the AAV-PHP.B capsid family in mice. Because this biology is mouse-specific and does not translate directly to primates, AAV-PHP capsids have limited translational relevance for human systemic CNS delivery.

This helps explain why a capsid can look exceptional in one mouse model but perform poorly in human neuronal systems. AAV capsids interact with biological barriers, cell-surface receptors, co-receptors, trafficking pathways, and tissue environments that differ across species. Human iPSC-derived neurons and organoids therefore provide an important additional filter for evaluating capsids before advancing toward translational development.

Balancing AAV Efficiency and Toxicity

A highly efficient AAV serotype is not always the best choice. In the human neuronal screen, stronger GFP expression was associated with toxicity-related effects in some conditions, including reduced cell number and changes in neurite morphology. The degree of sensitivity differed across neuronal subtypes.

Dopaminergic neurons appeared relatively resilient. Cortical neurons showed intermediate sensitivity. NGN2-induced neurons were the most vulnerable, especially under high-dose, high-expression conditions. This suggests that AAV optimization should not focus only on maximizing expression. Instead, researchers should identify the lowest effective dose that provides sufficient expression while preserving neuronal health and morphology.

The study also suggests that toxicity may be related to high-level GFP overexpression rather than direct AAV capsid toxicity alone. This distinction is important because overexpression toxicity can occur even when the vector itself is well tolerated. The promoter, transgene, dose, expression duration, and protein burden all need to be considered.

For fragile neuronal models, researchers should consider:

• Reducing vector dose when possible.
• Using cell-type-appropriate promoters.
• Avoiding unnecessarily strong ubiquitous promoters.
• Evaluating expression over a time course.
• Including non-transduced and control AAV groups.
• Measuring cell viability, cell number, and neurite morphology.
• Confirming that observed phenotypes are not caused by reporter overexpression.
• Using orthogonal assays when possible, such as immunostaining, qPCR, Western blotting, functional readouts, or electrophysiology.

 

The Importance of AAV Vector Quality in Sensitive Neuronal Systems

In human neuronal cultures and organoids, vector quality can strongly influence experimental interpretation. Even when the capsid is appropriate, impurities or inconsistent vector quality can affect cell health, immune signaling, transduction efficiency, and reproducibility.

Important AAV quality attributes include:

• Accurate vector genome titer.
• Capsid titer and empty/full capsid ratio.
• Genome integrity.
• Residual host-cell DNA.
• Residual plasmid DNA.
• Residual host-cell protein.
• Endotoxin level.
• Aggregation status.
• Sterility, bioburden, and mycoplasma status where applicable.
• Batch-to-batch consistency.

For sensitive cell systems such as NGN2-induced neurons or 3D brain organoids, low endotoxin, low aggregation, high purity, and well-characterized capsid composition are especially important. If two vectors differ in purity or empty/full ratio, differences in apparent biological performance may reflect vector quality as much as capsid tropism.

Applications of AAV Screening in Human Neuronal Research

AAV serotype screening in human neuronal models can support a wide range of applications. In basic neuroscience, it can help researchers select vectors for studying gene function, neuronal morphology, synaptic biology, and pathway regulation. In disease modeling, it can improve the delivery of disease-associated genes, gene-silencing constructs, or rescue transgenes into patient-derived neuronal models.

In translational research, human neuronal AAV screening can support early evaluation of CNS gene therapy candidates. While in vitro models cannot fully predict in vivo performance, they can help eliminate poorly performing capsids, identify promising candidates, and reduce reliance on animal-only assumptions.

Common applications include:

• AAV-mediated overexpression in human neurons.
• AAV-shRNA or AAV-miRNA knockdown studies.
• AAV-CRISPR editing or gene perturbation.
• Fluorescent reporter and biosensor delivery.
• Optogenetic and chemogenetic tool expression.
• Neuronal morphology and neurite outgrowth studies.
• Disease model rescue experiments.
• Human brain organoid transduction.
• Early-stage capsid selection for CNS gene therapy programs.

Practical Guidance for Choosing an AAV Serotype

For human iPSC-derived neuronal work, AAV selection should be guided by the target cell type and experimental goal. A broadly efficient serotype may be useful for general transgene expression, while a subtype-biased vector may be better for disease-specific neuronal models.

A practical selection strategy may include:

• Start with human-cell-supported serotypes such as AAV6, AAV6.2, and AAV2.7m8 for broad neuronal transduction.
• Consider AAV2-retro for dopaminergic neuron-focused applications, while validating specificity in the exact model.
• Avoid assuming that AAV9, AAV-PHP.eB, or AAV-PHP.S will perform well in human iPSC-derived neurons based only on mouse data.
• Test multiple doses to identify the lowest effective dose.
• Measure both transduction efficiency and cell health.
• Validate findings in the relevant 2D or 3D system.
• Use matched control AAV vectors to separate transgene effects from vector effects.
• Confirm key findings with molecular and functional assays.

This approach can reduce trial-and-error and improve confidence in downstream biological conclusions.

Conclusion

AAV remains one of the most powerful platforms for CNS gene delivery research, but successful application depends on choosing the right serotype for the right biological system. Human iPSC-derived neurons and brain organoids provide valuable platforms for evaluating AAV performance in a human-cell context, helping researchers move beyond assumptions based only on rodent data.

The systematic screen of 18 AAV serotypes across human iPSC-derived neuronal subtypes identifies AAV6, AAV6.2, and AAV2.7m8 as broadly efficient candidates, while AAV2-retro shows strong potential for dopaminergic neuron-focused applications. The findings also emphasize the need to balance expression efficiency with neuronal health, especially in sensitive models such as NGN2-induced neurons.

For neuroscience and CNS gene therapy research, the key takeaway is clear: AAV serotype selection should be evidence-based, human-model-informed, and validated in the intended experimental system. By combining human iPSC-derived models, 3D organoid validation, careful dose optimization, and rigorous vector quality control, researchers can build more reliable AAV-based platforms for studying neurological disease and advancing translational gene delivery.

How PackGene Supports AAV-Based Neuroscience and CNS Research

PackGene provides customized AAV vector design, packaging, production, purification, serotype selection, and analytical testing services to support neuroscience, iPSC-derived neuronal models, brain organoid research, and CNS gene delivery development.

For researchers working with cortical neurons, dopaminergic neurons, NGN2-induced neurons, or 3D brain organoids, PackGene can support project-specific AAV strategies that consider capsid selection, promoter design, payload structure, vector dose, target cell type, delivery format, and downstream quality requirements.

PackGene offers a broad AAV serotype portfolio, including commonly used wild-type and engineered capsids for neuroscience applications. Combined with quality-focused production and analytical testing, PackGene supports AAV-based studies in gene expression, neuronal labeling, CRISPR delivery, disease modeling, organoid transduction, and preclinical CNS gene therapy research.

Accelerating Human CNS AAV Serotype Selection with PackGene AAV Capsid Discovery Kits

To help researchers move beyond one-serotype-at-a-time testing, PackGene offers AAV Capsid Discovery Kits designed for high-throughput, data-driven capsid screening. Developed in partnership with Children’s Medical Research Institute, these kits enable pooled screening of broad AAV capsid panels in a single experiment, helping researchers identify capsids with the best performance in their specific model system.

PackGene’s AAV Capsid Discovery Kits support:

• Broad capsid comparison using pre-validated wild-type, published, and engineered AAV serotypes.
• Barcoded pooled screening to reduce experimental variability and accelerate early capsid selection.
• Dual DNA and RNA readouts to distinguish physical vector entry from functional transgene expression.
• Screening in in vitro cultures, ex vivo organoids, and in vivo animal models.
• Tissue-focused options, including CNS/Brain add-on panels for neuroscience applications.
• Custom add-ons for researchers who want to include proprietary or project-specific capsids.

For CNS and neurodegenerative disease research, this approach is especially valuable. Instead of assuming that a rodent-validated AAV capsid will perform well in human neurons, researchers can directly test a diverse capsid panel in the model that matters most, whether that is a 2D iPSC-derived neuronal culture, a dopaminergic neuron model, or a 3D brain organoid system.

By combining human-relevant AAV screening with customized AAV production, quality-focused characterization, and downstream vector development support, PackGene helps researchers identify better-performing capsids earlier, reduce trial-and-error, and build stronger foundations for CNS gene delivery research and preclinical gene therapy development.

References

1. Comprehensive investigation of AAV tropism across human iPSC-derived neuronal subtypes. Linus Wiora, et al., bioRxiv.