Steven Gray and the AAV9 platform for rare brain disease

The list of clinical trials Steven Gray's lab at UT Southwestern has fed is unusually long for a single academic group. By 2025 the published or active programs included gene therapies for giant axonal neuropathy, Rett syndrome, Tay-Sachs and Sandhoff disease, hereditary spastic paraplegia type 50, several Batten subtypes (CLN1, CLN5, CLN7), SLC6A1-related neurodevelopmental disorder, and additional programs at earlier stages. Most of these programs use the same AAV9 capsid, the same intrathecal route, and reuse manufacturing and toxicology templates the Gray Lab developed across the prior programs. The lab is the closest thing the rare-disease field has to an n-of-many AAV gene-therapy platform.

The Pirovolakis SPG50 trial, published in Nature Medicine in June 2024 as the first one-patient AAV gene therapy phase 1, was a Gray Lab vector. Michael Pirovolakis is the most public face of the Gray Lab's program. He is one of perhaps a dozen children whose treatments have come out of the same vector facility.

Training

Gray completed a B.S. with honors at Auburn University, then a Ph.D. in molecular biology at Vanderbilt University in 2006. His postdoctoral fellowship was in Jude Samulski's lab at the University of North Carolina at Chapel Hill, the lab that has been one of the foundational AAV gene-therapy training environments for two decades. Samulski's group developed many of the AAV vector engineering techniques the field now uses; Gray's training there gave him direct technical lineage from the early AAV development work.

He moved to UT Southwestern in the early 2010s, where he is now Professor in the Department of Pediatrics. He co-directs the UTSW Gene Therapy Program, which he and Berge Minassian founded in 2017, and he directs the UTSW Viral Vector Facility. He sits on the board of directors of the American Society of Gene and Cell Therapy and on the scientific advisory board of Sarepta Therapeutics.

The Gray Lab platform

The Gray Lab's core technical contribution is a generalizable AAV9 vector platform for the central nervous system. The platform consists of the AAV9 capsid (which crosses the blood-brain barrier), an intrathecal administration route (which delivers the vector to the cerebrospinal fluid and from there to the brain and spinal cord), a set of regulatory cassettes that the lab has validated across many target genes, and a manufacturing template that the lab's collaborators at academic and contract GMP facilities can adapt for new programs.

When a new rare-disease program approaches the Gray Lab with a candidate gene, the question is whether the gene will fit in the platform. The 4.7 kilobase cargo limit constrains which genes are eligible. The promoter and regulatory elements need to give the right level of expression in the right cell types. The disease has to be one where adding a functional copy of the gene addresses the molecular defect (loss-of-function recessive conditions are the cleanest fit; dominant-negative conditions can sometimes be addressed by knockdown-and-replace strategies the lab has also developed). When the answers line up, the lab can move from candidate gene to validated vector in approximately a year, faster than any commercial AAV company can.

A specific engineering refinement Gray's group has published is the miRARE element. miRARE, short for miRNA-Responsive Auto-Regulatory Element, is a feedback mechanism inserted into the AAV cassette that uses endogenous microRNA expression to limit transgene expression in cells that already produce sufficient endogenous protein. The element addresses an AAV-specific safety problem: overexpression toxicity in cells where the gene is supposed to be at a baseline level. A vector without auto-regulation can drive too much protein and produce cellular dysfunction. miRARE attempts to make the vector self-limiting. The element is now used in several Gray Lab programs, including the SLC6A1 program where overexpression of GAT-1 (the GABA transporter the gene encodes) would be cytotoxic.

Programs the lab has fed

The breadth of the Gray Lab's clinical pipeline is its distinguishing feature.

Giant axonal neuropathy (GAN). The first Gray Lab program to reach the clinic. GAN is caused by recessive mutations in GAN, encoding gigaxonin. The condition produces progressive peripheral and central nervous system degeneration. The Gray Lab's intrathecal AAV9-GAN vector was administered to a small case series of pediatric patients beginning in approximately 2015, with results published over the subsequent years showing safety and signs of clinical benefit.

Spastic paraplegia type 50 (SPG50). The Pirovolakis case. AAV9-AP4M1, intrathecal, single-patient phase 1, Nature Medicine June 2024. Twelve-month follow-up reported well-tolerated safety profile and disease stabilization.

Rett syndrome. Caused by mutations in MECP2. The Gray Lab developed AAV9-MECP2 with miRARE auto-regulation and partnered with Taysha Gene Therapies (which Gray co-founded as scientific advisor) to take the program to the clinic. Phase 1/2 trials are active.

Batten subtypes (CLN1, CLN5, CLN7). Several programs in development or active clinical evaluation, addressing the same NCL family disorders the Gray Lab has worked on for years. The CLN7 program is particularly relevant to the Mila Makovec case: an AAV9 gene-replacement approach is a different therapeutic class from the milasen ASO that targeted Mila's specific splicing variant, and the AAV approach could in principle treat children with other CLN7 mutations than Mila's.

SLC6A1-related neurodevelopmental disorder. SLC6A1 Connect, founded by Amber Freed after her son's diagnosis, partnered with the Gray Lab in October 2018. The program develops AAV9-SLC6A1 with miRARE for dose control. Preclinical efficacy has been demonstrated in mouse models, with rescue of EEG abnormalities and cognitive-behavioral deficits. Clinical trials are at advanced preclinical stages as of 2025.

Tay-Sachs and Sandhoff disease. AAV9-mediated gene therapy for the GM2 gangliosidoses. Programs were licensed to Sio Gene Therapies and have moved through clinical evaluation with mixed results.

The pattern is consistent. The lab develops the vector, validates it preclinically, and partners with foundations or companies for clinical execution. The lab does not run the clinical trials directly; the clinical work happens at the institutions that follow the patients (SickKids in Toronto for SPG50, sites running the Taysha and Sio trials for the other programs).

The institutional argument

The UTSW Gene Therapy Program model that Gray and Minassian built is the AAV-side equivalent of the n-Lorem Foundation model on the ASO side. The program aggregates demand for AAV gene therapies for ultra-rare conditions, partners with affected families and foundations for funding, runs the vector design and preclinical work in-house, and produces clinical-grade material through partnerships with GMP facilities.

The structural advantages are the same as n-Lorem's. The chemistry is shared (the AAV9 capsid, the intrathecal route, the manufacturing platform). The preclinical templates are shared. The regulatory experience compounds, so the seventh program submits a faster IND than the first one did. The cost per program comes down across the program slate.

The constraints are also similar. Manufacturing capacity is the throughput limit. AAV is harder than ASOs to produce, the small-batch GMP facilities are running at high utilization, and the cost per program is several times higher than the comparable ASO program. The Pirovolakis program raised approximately $3 million; published commentary on other Gray Lab programs reports total costs in the $3 million to $10 million range.

Gray has been explicit that the model works because of accumulated infrastructure, not because of an isolated technical breakthrough. The first AAV gene therapy his lab pushed to clinic took years of foundational work on capsid engineering, vector design, and manufacturing. The fifth program adapted the existing platform. The fifteenth program, when it arrives, will adjust the platform.

Recognition and what is next

Gray has published more than 90 peer-reviewed papers in journals including the New England Journal of Medicine, Nature Medicine, Nature Biotechnology, Brain, the Journal of Clinical Investigation, and Molecular Therapy. He holds more than 20 issued or pending patents on AAV vector engineering. He is a member of the ASGCT board of directors and a frequent speaker at the society's meetings.

The next decade of the Gray Lab's work will test whether the platform can scale to dozens of additional rare-disease programs without losing the speed and cost advantages that make single-patient AAV programs feasible. The Pirovolakis case proved the model works for one. The bottleneck has shifted to manufacturing throughput and to the question of whether the next ten and the next hundred programs can be pipelined through the same infrastructure without breaking it.