Evaluating AAV therapies in human iPSC-derived retinal and neural models

Evaluating AAV therapies in human iPSC-derived retinal and neural models

Adeno-associated viral (AAV) vectors are a leading platform for gene therapy development, with approved therapies including Luxturna® (indicated for RPE65-mediated inherited retinal dystrophy) and Zolgensma® (indicated for spinal muscular atrophy) demonstrating the clinical utility of AAV-mediated gene delivery. Beyond approved gene replacement therapies, next-generation gene-silencing candidates are advancing rapidly. uniQure’s AMT-130 (ifezuntirgene inilparvovec), administered via a one-time stereotactic neurosurgical procedure directly into the striatum, reported a 75% slowing of disease progression at 36 months in Phase 1/2 trials compared to matched external controls1. Despite advances in vector engineering, translating preclinical findings into clinical success for complex disorders, such as inherited retinal dystrophies, spinal muscular atrophy, Huntington’s disease and hemophilia, remains challenging.

Traditional cell lines and animal models remain important components of AAV development, while physiologically relevant human models can provide complementary insights into disease biology and therapeutic activity. Human iPSC-derived systems are increasingly being used to evaluate AAV performance in disease-relevant cell types²,3.

Human iPSC-derived models support AAV evaluation

Successful AAV development requires evidence of efficient transduction, robust transgene expression, target engagement, therapeutic activity and safety profile in relevant human cell types. For neurodegenerative and retinal diseases, a key challenge is evaluating these endpoints in models that capture disease-specific phenotypes and clinically relevant cellular responses.

Human iPSC-derived models can complement existing preclinical workflows by providing access to disease-relevant cell types in a controlled experimental setting. Unaffected donor-derived models support vector characterization, transduction studies, assay optimization, and preliminary assessment of vector-related toxicity, while patient-derived disease models can enable assessment of therapeutic activity and disease-specific responses.

Advanced model systems, including patient-derived models, co-cultures, and organoids, are increasingly being incorporated into AAV workflows to generate more biologically relevant preclinical data. These approaches support evaluation of functional endpoints and patient-specific responses, helping developers assess both vector performance and biological activity before clinical development²,3.

Retinal organoids as a platform for AAV characterization

Inherited retinal diseases are a major focus of AAV-mediated gene therapy development. Human iPSC-derived retinal organoids provide a multicellular human retinal model that can be used to assess vector tropism and transgene expression in retinal cell populations.

In a study conducted by McClements ME and colleagues in collaboration with the University of Oxford and Axol Bioscience (former Newcells Biotech), human iPSC-derived retinal organoids were used to compare the performance of AAV5, AAV2 7m8, AAV2 quad mutant, AAV2 Y444F, and AAV8 Y733F vectors. Using green fluorescent protein (GFP) reporter constructs driven by the ubiquitous CAG promoter, the authors observed differences in transduction efficiency between vectors. The AAV2 quad mutant and AAV2 7m8 vectors achieved the highest proportion of GFP-positive cells within retinal organoids, transducing approximately 19.4% and 16.4% of cells, respectively. Reporter gene expression was detected as early as 5 days after transduction and increased over time, demonstrating sustained transgene expression within the retinal organoids.

Figure 1. Live cell imaging of reporter gene expression up to 27 days post-transduction. Arrows indicate the areas where the onset of reporter gene expression first appeared. Each retinal organoid was treated with 1E+10 genome copies, except for GRK1.GFP AAV8 Y733F, for which 5E+10 genome copies were used. Background fluorescence of live retinal organoids was unavoidable, but reporter expression was evident as higher intensity areas above this background. Adapted from McClements et al.4

The study also evaluated promoter-dependent expression patterns. While the ubiquitous CAG promoter supported broader retinal cell transduction, the photoreceptor-specific GRK1 promoter produced a more restricted expression profile associated with photoreceptor-like cells within the organoids. Immunohistochemical analysis demonstrated colocalization of GRK1-driven GFP expression with the photoreceptor marker recoverin, indicating successful transduction of photoreceptor-like cells within the organoids4.

Adapted from McClements et al.4

At Axol Bioscience, both iPSC-derived retinal organoids and RPE models are utilized to evaluate AAV vectors for transduction efficiency, transgene expression, target engagement, toxicity, and study-specific endpoints in a human retinal cell context.

Neural models for AAV evaluation

Neurodegenerative disorders represent a major focus of AAV therapeutic development, creating strong demand for models that capture disease-relevant responses in human neural tissue⁵,6. Human iPSC-derived neural models retain patient-specific genetic backgrounds and recapitulate key phenotypes for conditions such as Huntington’s disease (HD) and amyotrophic lateral sclerosis (ALS), enabling preclinical evaluation of gene delivery, target engagement, and functional outcomes in relevant human target cells5,7

At Axol Bioscience, we have utilized AAV extensively across multiple iPSC-derived cell types, including striatal neurons, sensory neurons, motor neurons, and microglia. We have examined the impact of promoter, serotype, and multiplicity of infection (MOI) to achieve optimal transduction efficiency with minimal cellular toxicity (Figures 3 and 4).

Figure 3. iPSC-derived microglia (ax0664) transduced with different AAV serotypes to evaluate transduction efficiency across serotypes. Three MOI were tested for each and fluorescence measured on an IncuCyte. A) demonstrates masking of fluorescence using IncuCyte software and B) representative image of cells showing no obvious cell death.

Figure 4. iPSC-derived sensory neurons (ax0055) transduced with different AAV-GFP serotypes to evaluate transduction efficiency across serotypes. A single MOI was tested for each and fluorescence measured on an IncuCyte. Different promoters were also tested, all AAV shown are CMV except for those labelled as CAG.  

The timepoint of vector addition is also an important consideration as transduction efficiency can vary with cellular maturation state, with some vectors achieving higher transduction in progenitor populations than in mature neurons. Using GFP-expressing AAVs provides a direct fluorescent measure of transduction efficiency, while monitoring cell morphology and neurite outgrowth allows us to assess short-term toxicity over several weeks and long-term toxicity over multi-month cultures. Finally, fluorophore-expressing AAVs are a useful tool for building co-culture systems, allowing neurons to be transduced prior to adding glial cells so individual cell types can be tracked independently.

Supporting preclinical AAV studies

At Axol Bioscience, we support preclinical AAV studies using physiologically relevant human iPSC-derived models. Depending on the study objectives, these systems can be used to assess vector tropism, transduction efficiency, promoter activity, transgene expression, target engagement and vector-associated toxicity in disease-relevant human cell types. Assays can be tailored to the biological question, incorporating live-cell imaging, qPCR, flow cytometry, electrophysiology and other functional readouts.

References

  1. Sung V, et al. AMT-130 Huntington’s disease gene therapy: clinical study outcomes at three years. Neurology. 2026;106(11 Suppl 1):4195.
  2. Li C, Samulski RJ. Adeno-associated virus as a delivery vector for gene therapy. Signal Transduct Target Ther. 2024;9:57.
  3. Wang D, Tai PWL, Gao G. Adeno-associated virus vector as a platform for gene therapy delivery. Nat Rev Drug Discov. 2019;18(5):358-378.
  4. McClements ME, et al. Tropism of AAV vectors in photoreceptor-like cells of human iPSC-derived retinal organoids. Transl Vis Sci Technol. 2022;11(8):16.
  5. Cappella M, Elouej S, Biferi MG. The potential of induced pluripotent stem cells to test gene therapy approaches for neuromuscular and motor neuron disorders. Front Cell Dev Biol. 2021;9:730329.
  6. García-González N, et al. Advances and challenges in gene therapy for neurodegenerative diseases: a systematic review. Int J Mol Sci. 2024;25(23):12485.
  7. Wang S, Elia L, Liu W, Finkbeiner S. Modeling human neurodegeneration with iPSCs: opportunities, limitations, and the challenge of aging. J Clin Invest. 2026;136(20):e209602.
Next
Modeling Friedreich’s ataxia cardiomyopathy with patient-derived iPSC cardiomyocytes