Pre-existing antibodies against adeno-associated virus, the most widely used vehicle for in vivo gene therapy, are present in roughly half or more of adults worldwide, and the immune reactions that follow vector administration range from silent neutralization of the therapy to fatal organ damage. AAV immunogenicity touches every stage of treatment: it determines who can receive a dose, whether the therapeutic protein lasts, and whether a second dose is even feasible. Despite AAV’s reputation as a mild, non-pathogenic virus, the immune system treats it with surprising aggression once it encounters the quantities used in clinical gene transfer.
Why So Many People Already Have Anti-AAV Antibodies
Wild-type AAV circulates naturally in the human population. Most people are exposed during childhood, and their immune systems form antibodies that persist for years. These neutralizing antibodies bind to the AAV capsid (the protein shell carrying the therapeutic gene) and block it from entering target cells. A large global seroprevalence study found that at a commonly used screening threshold, about half to three-quarters of adults tested positive for neutralizing antibodies against the most clinically relevant AAV serotypes. Antibodies against AAV1 were the most common, with roughly 75% of adults showing detectable activity at the lowest serum dilution, while AAV5 had the lowest prevalence at higher dilutions, dropping to about 27% at a 1:4 dilution. Antibodies to AAV6, AAV8, AAV9, and AAVRh74var fell in between.1Molecular Therapy. Global Seroprevalence of Neutralizing Antibodies against Adeno-Associated Virus Vectors
A separate multicenter study of adult males with hemophilia confirmed this pattern and added an important detail: antibody levels stayed stable over at least two years. About 47% had neutralizing antibodies to AAV8, roughly 53% to AAV2, and about 53% to AAV5. Nearly 40% had antibodies that cross-reacted with all three serotypes tested.2Gene Ther. Multicenter assessment and longitudinal study of the prevalence of antibodies and related adaptive immune responses to AAV in adult males with hemophilia That cross-reactivity is a real problem: switching from one AAV serotype to another does not always get around existing immunity, because the antibodies recognize structural features shared across serotypes.
For gene therapy programs that exclude seropositive patients, these numbers translate directly into shrinking the eligible population. Some clinical trials set strict cutoffs, requiring antibody titers below a 1:5 dilution, which can disqualify more than half of screened adults. Children tend to have lower seroprevalence than adults, which is one reason pediatric gene therapies like those for spinal muscular atrophy have had a somewhat easier path. But even in pediatric populations, maternal antibodies can cross the placenta. In a mouse model, maternal anti-AAV antibodies transferred efficiently to the fetus and impaired gene editing in a titer-dependent way, suggesting that very young patients are not always in the clear.3JCI Insight. Preexisting maternal immunity to AAV but not Cas9 impairs in utero gene editing in mice
How the Innate Immune System Detects AAV
Even in patients without pre-existing antibodies, the body’s first-line defenses respond to the vector within hours. One of the most important triggers is the presence of CpG motifs in the vector’s DNA. CpG sequences are stretches of unmethylated cytosine-guanine dinucleotides that mammalian cells recognize as foreign because bacterial and viral DNA tends to contain more of them. When AAV vector genomes carry these motifs, they activate a receptor called TLR9 inside a specialized group of immune cells. This triggers a cascade of inflammatory signaling that can suppress transgene expression and prime the adaptive immune system to attack transduced cells.
Research in mice showed that removing CpG motifs from the vector genome increased expression of the therapeutic protein by two- to threefold within the first 24 hours after dosing. The CpG-depleted vectors also reduced the appearance of a specific population of dendritic-like cells that directly bound the therapeutic protein.4PubMed. The presence of CpGs in AAV gene therapy vectors induces a plasmacytoid dendritic cell-like population very early after administration This innate immune flare is not just an inconvenience. It sets the stage for the more targeted adaptive response that follows, which can destroy the cells the therapy was designed to help.
Adaptive Immunity and the Loss of Transduced Cells
The adaptive arm of the immune response kicks in over days to weeks and has two main weapons: antibodies (humoral immunity) and killer T cells (cellular immunity). Both can undermine gene therapy.
Killer T cells are the more immediate threat to therapeutic durability. After AAV enters a target cell, the capsid proteins get broken down and displayed on the cell surface, effectively flagging that cell as infected. Research demonstrated that even at low antigen levels, this display is enough for capsid-specific killer T cells to destroy transduced liver cells.5Journal of Clinical Investigation. Capsid antigen presentation flags human hepatocytes for destruction after transduction by adeno-associated viral vectors The effect scales with dose: more vector means more capsid protein displayed, which means more aggressive T cell killing. This is one reason the field has struggled with a bitter paradox: higher doses that deliver more therapeutic gene also provoke fiercer immune attacks.
This capsid-directed T cell response has been directly linked to the liver enzyme elevations seen in hemophilia gene therapy trials. In trials of several AAV-based hemophilia B therapies, transient liver enzyme increases typically appeared two to six weeks after infusion. These elevations usually responded to corticosteroid-based immunosuppression, but the need for steroids raises its own concerns about long-term safety and about suppressing the very immune monitoring that protects against other threats.6PubMed. Liver dysfunction in AAV-mediated Hemophilia B gene therapy: Mechanisms and management strategies
When the Immune Response Turns Dangerous
Mild liver enzyme bumps are the most common immune-related side effect of AAV gene therapy, occurring in somewhere between 20% and 80% of patients depending on the therapy and dose. In rare cases, though, the immune response escalates to acute liver failure.7PubMed Central. Hepatic Manifestations Following Gene Therapy High-dose systemic administration carries the greatest risk. Patient deaths have been reported from liver, kidney, heart, and lung failure, with both innate and adaptive immune responses contributing to the toxicity.8PubMed Central. Lethal immunotoxicity in high-dose systemic AAV therapy
The timing of fatal outcomes offers clues about which arm of the immune system is responsible. Deaths within the first two weeks tend to present as a hyperinflammatory syndrome, likely driven by the innate response to massive vector quantities. Deaths occurring one to three months out appear to involve adaptive immunity, including T cell-mediated destruction of transduced cells. And some fatalities have emerged more than a year after treatment, through mechanisms still being worked out.9Molecular Therapy. AAV Immunogenicity: A Challenge for Gene Therapy
The mechanisms also differ by disease. In patients with spinal muscular atrophy, liver damage showed clear signs of T cell infiltration. In patients with X-linked myotubular myopathy, autopsies revealed relatively little inflammatory cell infiltration, suggesting that the sheer metabolic stress of processing enormous quantities of vector in an already compromised liver, rather than a T cell attack, was the primary driver.10Molecular Therapy. Challenges and Mechanisms of Lethality in High-Dose Systemic AAV Gene Therapy
Complement Activation and Blood Clotting Disorders
A particularly alarming immune complication is thrombotic microangiopathy, a condition where tiny blood clots form throughout small blood vessels, damaging red blood cells and consuming platelets. Clinical data showed that this reaction is antibody-dependent: patients who developed it had rising anti-capsid antibodies that activated both the classical and alternative complement pathways, visible as drops in complement component C4 and spikes in soluble C5b-9 and other markers. Patients who did not develop significant antibody responses showed minimal complement activation.11PubMed Central. Thrombotic microangiopathy following systemic AAV administration is dependent on anti-capsid antibodies
The interplay between AAV-specific factors, environmental variables, and individual patient biology makes complement-related outcomes difficult to predict. A convergence of all these factors likely explains why some patients tolerate the same vector and dose that causes serious harm in others.12PubMed. Complement System Response to Adeno-Associated Virus Vector Gene Therapy This unpredictability is one of the reasons regulators and trial sponsors are increasingly cautious about high systemic doses.
The Trouble with Screening
Given that pre-existing antibodies can neutralize the vector and potentially trigger dangerous complement reactions, screening patients before dosing seems like an obvious safeguard. In practice, it is frustratingly inconsistent. Antibody assays used across clinical programs measure different things: some detect neutralizing capacity, while others measure total antibodies that bind the capsid regardless of whether they block transduction. These assays are run with different protocols, different serum dilutions, and different thresholds for calling a result positive. The result is that two programs testing the same patient sample could reach opposite conclusions about eligibility.13Molecular Therapy. AAV Immunogenicity: A Challenge for Gene Therapy
Establishing the right cutoff is part of the challenge. Laboratories typically set a screening cut point that aims for a low false-positive rate using a panel of treatment-naive samples, but this process lacks standardization across the field.14Molecular Therapy. AAV Immunogenicity: A Challenge for Gene Therapy A titer that one sponsor considers safe for dosing might be above another sponsor’s exclusion threshold. Until the field converges on a harmonized approach, meaningful comparisons across programs remain nearly impossible.
Where the Vector Goes Shapes the Immune Response
Not all routes of administration provoke the same level of immune reaction. Intravenous delivery and intramuscular injection tend to be the most immunogenic. The liver, paradoxically, can be both a target organ and an immune-privileged site: directing transgene expression specifically to hepatocytes using liver-specific gene promoters has been shown to dampen immune responses compared to using a ubiquitous promoter that drives expression in many cell types.15Molecular Therapy. AAV Immunogenicity: A Challenge for Gene Therapy The liver’s natural tolerance mechanisms can help, but they are not foolproof: at high enough doses or in patients with pre-existing liver disease, those tolerance mechanisms get overwhelmed.
The therapeutic protein itself also matters. The immune system can mount responses against the transgene product, especially if the patient’s body has never made that protein before. In such cases, the therapeutic protein looks entirely foreign, and the immune system treats it accordingly. If the patient has some residual natural production, the immune system is more likely to tolerate the AAV-delivered version.16PubMed Central. Complexity of immune responses to AAV transgene products – Example of factor IX The choice of AAV serotype influences this, too. In autoimmune-prone mice, AAV1 delivering human alpha-1 antitrypsin provoked strong antibody and T cell responses, while AAV8 carrying the same gene did not.17PubMed Central. Distinct immune responses to transgene products from rAAV1 and rAAV8 vectors If the immune system clears the therapeutic protein or destroys the cells producing it, the treatment’s benefit fades.18PubMed. Assessment of Immune Responses Against AAV Encoded Transgene Products
Strategies to Outwit the Immune System
The field has pursued several parallel engineering approaches to reduce AAV immunogenicity, each targeting a different layer of the problem.
CpG Depletion of Vector Genomes
Stripping CpG motifs from the vector DNA tackles the innate immune trigger at its source. In mouse studies, CpG-depleted vectors maintained robust, stable transgene expression while dramatically reducing T cell infiltration and inflammatory signaling in transduced muscle and liver tissue. One study showed that CpG-depleted vectors evaded immune activation, prevented T cell infiltration, and avoided the upregulation of inflammatory markers that normally appear in transduced skeletal muscle.19JCI Insight. CpG-depleted adeno-associated virus vectors evade immune detection Another confirmed that CpG depletion markedly reduced capsid-specific killer T cell responses in hemophilia B mice, although it did not eliminate them entirely and had minimal effect on antibody formation.20PubMed Central. Effect of CpG Depletion of Vector Genome on CD8+ T Cell Responses in AAV Gene Therapy CpG depletion is promising for reducing T cell killing but does not solve the antibody side of the equation.
Capsid Engineering
If the capsid is the main target of neutralizing antibodies, then redesigning the capsid’s outer surface is a logical countermeasure. Three main strategies are being pursued: rational design based on structural knowledge of antibody-binding sites, directed evolution that tests enormous libraries of capsid variants, and machine learning methods that predict which variants will perform best. Together, these approaches have produced novel capsids with improved tissue targeting and reduced immunogenicity.21PubMed Central. Advances in AAV capsid engineering: Integrating rational design, directed evolution and machine learning
One especially creative approach used structure-guided evolution to create capsid variants with antigenic surfaces that do not exist in any natural AAV strain. A lead variant built on the AAV1 scaffold effectively evaded neutralizing sera from immunized mice and monkeys. It also resisted neutralization by non-human primate and human serum samples at dilutions as high as 1:5, a threshold used in several ongoing clinical trials.22PubMed Central. Structure-guided evolution of antigenically distinct adeno-associated virus variants for immune evasion This did not compromise the variant’s ability to enter cells or reach target tissues.
Exosome Shielding
Rather than redesigning the capsid, another approach hides it entirely. AAV particles can be enclosed within extracellular vesicles, tiny membrane-bound bubbles that cells naturally release. These “exosome-enveloped” AAV vectors cloak the capsid proteins inside a host-derived membrane, making the antibodies that recognize the naked capsid largely ineffective. In one study, exosome-enveloped AAV was up to 136-fold more resistant to neutralizing antibodies than standard AAV in cell culture. In mice with passively transferred human antibodies at levels that cut standard AAV brain transduction by 80%, the exosome-wrapped version maintained full transduction.23PubMed Central. Naturally enveloped AAV vectors for shielding neutralizing antibodies and robust gene delivery in vivo Purification of these vesicle-wrapped vectors has improved, with purified exo-AAV1 showing enhanced resistance to neutralization.24PubMed Central. Neutralizing Antibody Evasion and Transduction with Purified Extracellular Vesicle-Enveloped Adeno-Associated Virus Vectors The approach is still preclinical, but it offers a fundamentally different angle on the problem.
Clearing Antibodies to Enable Dosing and Redosing
For patients who already have anti-AAV antibodies, some groups are working on ways to temporarily remove those antibodies before vector infusion. Plasmapheresis, the physical filtering of blood plasma, achieved a 100-fold decrease in anti-AAV neutralizing antibody titers in a nonhuman primate, bringing levels down enough for successful readministration. The procedure required IgG supplementation afterward to avoid leaving the animal dangerously immunodeficient.25PubMed Central. Use of plasmapheresis to lower anti-AAV antibodies in nonhuman primates with pre-existing immunity to AAVrh74
An enzymatic approach offers a more targeted alternative. An IgG-degrading enzyme called IdeZ, administered intravenously, selectively cleaves circulating antibodies. In macaques, a single dose of IdeZ rescued AAV transduction by transiently reversing seropositivity. It also worked against a diverse panel of individual human donor sera in mice.26PubMed Central. Rescuing AAV gene transfer from neutralizing antibodies with an IgG-degrading enzyme Both approaches create only a temporary window for dosing, since the immune system quickly rebuilds its antibody stocks, but that window may be enough.
Redosing is one of the most pressing unresolved challenges. After a first dose of AAV, the immune system mounts a strong, durable antibody response that effectively blocks a second dose with the same serotype. Research in mice showed that after a high-dose brain injection of AAV5, neutralizing antibodies appeared in the blood within three days and reached extremely high titers within a week. In the brain itself, antibodies appeared by three weeks and were sufficient to block a second round of AAV transduction almost completely. Mice genetically engineered to lack B cells did not have this problem, confirming that the antibody response is the primary barrier.27Molecular Therapy Methods & Clinical Development. Brain-wide infiltration of B lymphocytes and neutralizing antibody production following intraparenchymal high-dose AAV administration impede gene therapy redosing For patients whose therapeutic protein expression fades over years, the inability to redose is a serious limitation.
Empty Capsids and Manufacturing Quality
An underappreciated contributor to immunogenicity comes from the manufacturing process itself. AAV production inevitably generates a mixture of “full” capsids carrying the therapeutic gene and “empty” capsids that contain no DNA payload. These empty shells still trigger immune responses, and in some ways they may be worse than full capsids. When tested against human and mouse immune cells, empty AAV8 capsids provoked stronger inflammatory cytokine production and greater dendritic cell migration than their gene-containing counterparts. And when empty and full capsids were mixed together, the combination activated natural killer cells and complement more aggressively than either type alone.28PubMed Central. Immunogenicity risk assessment of empty capsids present in adeno-associated viral vectors using predictive innate immune responses This finding has pushed the field to refine purification methods and consider whether acceptable limits for empty capsid contamination need to be tightened.
Why Animal Models Do Not Tell the Whole Story
Much of what we know about AAV immunogenicity comes from mice and nonhuman primates, but these models have real limitations. The translational gap between species arises from differences in how capsids interact with cell-surface receptors, how the immune system is configured, and even basic tissue architecture. Old World monkeys are the closest available model for human systemic delivery, yet they differ from humans in seroprevalence patterns, complement activity, and the biology of the blood-vessel lining.29PubMed Central. Species barriers in AAV tropism: mechanisms, models, and emerging solutions for clinical translation A capsid variant that escapes neutralization in a monkey may not do so in a human, and a dose that is safe in a smaller animal with a proportionally different liver-to-body ratio may overwhelm a human liver. These species barriers are a persistent source of unpleasant surprises in clinical trials and explain why the field continues to see safety signals that preclinical studies did not fully predict.