Pre-Existing Anti-Adeno-Associated Virus Immunity in Gene Therapy: Mechanisms,Challenges,and Potential Solutions

The immunological landscape of AAV exposure

The presence of pre-existing immunity to AAV is both common and variable in the human population, with seroprevalence estimates ranging from 30% to 80% in adult. ¹⁶ These antibodies, particularly NAbs, typically arise from exposure to wt-AAVs. However, the prevalence and profile of anti-AAV immunity vary widely across populations, influenced by environmental exposures, seasonal patterns, and geographical location. One potential driver of this variability is the environmental stability of AAV particles. Although direct data on the environmental stability of AAV are lacking, related parvoviruses such as canine parvovirus and human parvovirus B19 are known to be environmentally stable. They are highly resistant to degradation and remain infectious under a range of temperature and humidity conditions.^17,18 If AAV shares similar structural features, it may also persist in the environment, thereby increasing opportunities for human exposure. Similarly, insights from SARS-Cov-2, another environmentally stable virus, suggest that climate conditions can significantly influence viral persistence and transmission dynamics. ¹⁹ Such environmental variation in temperature and humidity could partly explain the geographical differences in AAV seroprevalence reported by Calcedo et al., ¹⁶ and other investigators, who observed that the prevalence of pre-existing anti-AAV antibodies differs across regions. ²⁰

Beyond geographical and environmental influences, host-related factors also shape the immunological landscape of AAV exposure. Seroprevalence generally increases with age, reflecting the cumulative likelihood of encountering AAV or related virus over time. This trend makes younger individuals, who are less likely to have developed high-titer NAbs, potentially more favorable candidates for AAV-based gene therapies.^21,22 However, maternal antibodies have been detected in infants under 6 months of age, indicating that passive immunity can temporarily elevate antibody levels even in those without direct exposure. This potentially complicates early-life gene therapy interventions. ²¹ These variations emphasize the need for population-specific immunity profiling when designing and implementing AAV gene therapy strategies.^16,19,21

Another factor adding complexity to the immune landscape is the inherent cross-reactivity of antibodies generated against AAV. Calcedo and Wilson demonstrated that a single natural AAV infection in Chimpanzees can elicit broad cross-NAbs capable of inhibiting multiple AAV serotypes. ²³ This breadth of response is thought to arise from the reactivation of latent AAVs during helper virus coinfections, which presents a diverse array of capsid antigens to the immune system. Such cross-reactivity within the AAV family suggests that antigenic overlap plays a central role in shaping seroprevalence patterns. Extending this concept, exposure to structurally related members of the parvoviridae family may also generate antibodies that cross-react with AAV capsids. In humans, prior exposure to parvovirus B19, a member of the parvoviridae family, could lead to the generation of antibodies that cross-react with AAV capsids due to shared structural motifs.^24,25 Similarly, in translational animal models such as pigs, routine vaccination against porcine parvovirus, another parvoviridae member, could produce antibodies with AAV-binding activity. These cross-reactive antibodies, whether generated by other AAV serotypes or by related viruses, have the potential to impair vector transduction efficiency and confound interpretations of preclinical efficacy data when pigs are used as translational or preclinical models. ²⁴

The magnitude and specificity of anti-AAV antibody responses are shaped by a combination of environmental influences, host factors such as age and maternal immunity, intrinsic cross-reactivity among AAV serotypes, and possibly, exposure to related viruses through natural infection and vaccination of animal models. Together, these variables define the baseline immune status of both human patients and animal models, with direct consequences for vector performance, transgene expression, and therapeutic outcomes. ²⁶

As AAV gene therapies advance toward widespread clinical application, understanding the prevalence, cross-reactivity, and functional impact of pre-existing anti-AAV immunity becomes essential. Addressing this need will require robust population-based seroprevalence studies, age-stratified analyses, and species-specific immune profiling.^16,21,22 By accurately characterizing these baseline immune landscapes, researchers and clinicians can improve patient selection criteria, vector design, and immunomodulatory strategies ultimately enhancing the accessibility and success of AAV-based therapies across diverse populations.

Mechanisms of AAV detection and clearance

AAV vectors are generally regarded as having low immunogenicity; however, they are still subject to detection and clearance by both innate and adaptive arms of the immune system. These responses can significantly impact the success of AAV-mediated gene therapy, especially in individuals with pre-existing immunity. Understanding these mechanisms is critical for designing more effective and durable therapeutic strategies.

Innate immune recognition of AAV

Upon initial exposure, AAV is recognized by the innate immune system through pattern recognition receptors (PRRs) that detect structural and nucleic acid components of the virus (Fig. 2). This recognition initiates a nonspecific but rapid immune response aimed at neutralizing the foreign vector and priming adaptive immunity.

OPEN IN VIEWER

Figure 2.

Innate immune recognition of adeno-associated virus (AAV) vectors. This illustration depicts the key pathways involved in the innate immune sensing of AAV vectors following administration. Viral capsid proteins and vector genomes are detected as pathogen-associated molecular patterns (PAMPs) by pattern recognition receptors (PRRs). On the cell surface, Toll-like receptor 2 (TLR2) recognizes AAV capsid components, initiating MyD88-dependent signaling through IRAK kinases and NF-κB activation, which induces proinflammatory cytokines such as TNFα, IL-6, and IFNγ. Within endosomes, degradation of the AAV capsid exposes viral DNA to toll-like receptor 9 (TLR9), which also activates the MyD88–IRAK–NF-κB axis. In the cytosol, leaked mitochondrial DNA (mtDNA) and AAV genomes are sensed by cyclic GMP-AMP synthase (cGAS), which catalyzes cGAMP production to activate the STING–IRF3 pathway, driving type I interferon responses. These innate sensing mechanisms contribute to inflammatory signaling and prime adaptive immunity, including the development of anti-AAV antibodies that limit transduction efficacy upon re-exposure.

Among the key PRRs involved, Toll-like receptor 2 (TLR2), located on the cell surface or within the endosomal membranes, recognizes structural components of the AAV capsid.^27–30 Upon activation, TLR2 recruits the adaptor protein myeloid differentiation primary response 88 (MyD88), leading to phosphorylation of interleukin-1 receptor-associated kinases (IRAKs) and activation of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB). This signaling cascade drives the production of proinflammatory cytokines such as tumor necrosis factor-alpha, interleukin-6 (IL-6), and interferon-gamma (IFN-γ). ³¹ Following endocytosis, AAV capsids may break down in the endosome, exposing the viral genome to TLR9, which similarly activates the MyD88/IRAK pathway, further amplifying inflammatory signaling. ³²

Additionally, both vector-derived genomes and stress-induced mitochondrial DNA (mtDNA) can leak into the cytosol where they serve as danger-associated molecular patterns that activate cytosolic DNA sensors such as cyclic GMP-AMP synthase (cGAS). ³³ Notably, AAV vectors have been shown to trigger DNA damage responses in transduced cells, which may further promote the release or mislocalization of nuclear and mtDNA into the cytosol. Upon sensing cytosolic DNA, cGAS catalyzes the formation of cyclic GMP-AMP (cGAMP), which activates the stimulator of interferon genes (STING) pathway and downstream interferon regulatory factor 3 (IRF3), ultimately leading to robust type I interferon production.^33,34 Other DNA sensors, including interferon-inducible factor 16 (IFI16) and Absent in Melanoma 2 (AIM2), have also been implicated in the recognition of foreign DNA.^35,36 IFI16 has been shown to act as a restriction factor for AAV2, interfering with sp1-dependent transactivation and thereby reducing transgene expression independently of traditional immune-modulatory pathways. ³⁷ AIM2, a well-characterized inflammasome sensor, can detect cytoplasmic double-stranded DNA and promote maturation of proinflammatory cytokines such as IL-1β and IL-18. While AIM2 activation reflects an innate response to foreign DNA, current evidence suggests that inflammasome activation may not be a dominant contributor to AAV-induced innate inflammation, especially in comparison with other DNA sensors such as TLR9 or STING. ³¹

Furthermore, cytosolic RNA sensors such as retinoic acid-inducible gene I (RIG-I) and melanoma differentiation-associated protein 5 (MDA5) recognize viral RNA and activate downstream signaling through mitochondrial antiviral-signaling protein (MAVS), leading to phosphorylation of TANK-binding kinase 1 (TBK1) and IRF3, ultimately triggering type I interferon responses in several RNA virus infections.^38–40 Following AAV vector transduction, bidirectional transcription from strong promoters and ITRs can produce complementary sense and antisense RNAs that anneal to form double-stranded RNA (dsRNA). This dsRNA can trigger innate immune response through MDA5. In this context, MDA5 activation leads to IFN-β production and reduced transgene expression, as demonstrated in human hepatocytes and xenografted mouse liver models. ⁴¹ Although MDA5 and RIG-I both signal through MAVS and can activate TBK1 and IRF3 in most infections by RNA viruses, their specific involvement in AAV-mediated responses remains to be established, and current data suggest a minimal role for RIG-I. Collectively, these sensing pathways lead to the expression of interferon-stimulated genes, inflammation, and the priming of adaptive immunity, including the generation of NAbs that can limit the efficiency of AAV-based gene therapy upon re-exposure. ³⁴

In parallel, the complement system serves as an additional innate immune defense mechanism against AAV. When antibodies bind to the AAV capsid, they can initiate the classical complement pathway, leading to cleavage of complement component 3 (C3), opsonization of the vector, and formation of the membrane attack complex (MAC), which may damage viral particles and amplify immune signaling.^42,43 Notably, complement activation has also been observed in the absence of antibodies, particularly at high AAV vector doses, through the alternative complement pathway, contributing to immune-mediated vector clearance and, in some cases, toxicity.^42,44 A recent study by Salabarria et al., ⁴⁵ however, provided compelling clinical evidence that complement activation following systemic AAV9 administration is predominantly antibody-driven. In pediatric patients treated with high-dose AAV9, the classical pathway was triggered by anti-capsid IgM and IgG and further amplified through alternative pathway, as indicated by elevated levels of Ba and Bb. The activation cascade led to increased production of C3a, C5a, and MAC (SC5b-9), which are key effectors of inflammation and endothelial injury. Notably, this immune activation was associated with the development of thrombotic microangiopathy (TMA), a complement-mediated complication characterized by thrombocytopenia and microvascular damage. In contrast, patients who received targeted immune modulation with rituximab and sirolimus exhibited minimal antibody production and complement activation and did not develop TMA. This study underscores the critical role of anticapsid antibodies and complement activation in driving AAV-toxicity and the importance of risk assessment in AAV gene therapy. ⁴⁵

Adaptive immune responses to AAV

The adaptive immune system represents a significant hurdle in AAV-based gene therapy, particularly due to the presence of pre-existing humoral immunity in a substantial portion of the human population. The key contributors in this response are antibodies directed against the AAV capsid. For clarity and consistency, we define total AAV antibodies as all immunoglobulins capable of binding to AAV capsid epitopes, regardless of their functional consequences. This category includes both NAbs and non-NAbs.

NAbs directly inhibit AAV infection by sterically blocking the interactions between the viral capsid and cellular receptors or disrupting intracellular trafficking, thus preventing transduction. Non-NAbs, in contrast, bind to the capsid without directly blocking entry but can impair vector performance through Fc-mediated effector functions. These include complement activation, antibody-dependent phagocytosis (ADCP) by macrophages, antibody-dependent cellular cytotoxicity (ADCC) by natural killer (NK) cells, and antibody-dependent complement deposition (ADCD).^43,46

NAbs are among the most well-characterized components of this immune barrier. These antibodies can persist for years following natural exposure to wt-AAV or related viruses and are often cross-reactive among different AAV serotypes, limiting not only initial vector efficacy but also the feasibility of repeat vector administration.^15,25,47 We will adhere to these definitions throughout the remainder of this article.

Although antibody responses are more commonly evaluated, pre-existing cellular immune memory can also influence AAV gene therapy. Capsid-specific cytotoxic CD8⁺ T lymphocytes have been observed to recognize AAV-derived peptides presented on major histocompatibility complex class I (MHC I) molecules by transduced cells. These cells may be eliminated by T cell-mediated lysis, reducing transgene expression and therapeutic durability.^48,49 While this response is more typically induced after vector exposure, pre-existing AAV-specific T cells have been detected in some individuals, suggesting that recall responses could contribute to reduced efficacy in subsequent dosing scenarios.

Together, these adaptive responses, especially pre-existing antibodies, form a well-recognized and significant barrier to the successful application of AAV gene therapies. Their role in neutralizing vectors, mediating clearance of transduced cells, and precluding redosing highlights the need for robust patient screening, novel capsid designs, and adjunct immunomodulatory strategies to ensure clinical efficacy.

Mechanisms of antibody-mediated AAV neutralization

As outlined in earlier sections, pre-existing anti-AAV antibodies are both common and immunologically diverse in both human population and preclinical animal models. Building on this foundation, we now turn to the functional consequences of these antibodies’ influence on AAV-based gene therapy. Neutralization in this context is driven predominantly by extracellular mechanisms, which either block essential steps in vector entry or promote clearance via Fc-dependent immune pathways. The following subsections examine these principal categories, detailing their molecular underpinnings, structural correlates, and clinical relevance.

Direct extracellular neutralization of AAV vectors

NAbs can block AAV infection at the earliest stages by binding to epitopes on the AAV capsid critical for receptor engagement or capsid stability, thus obstructing essential steps in the AAV entry process. ²⁵ Structural studies using high resolution cryo-electron microscopy and pseudoatomic modeling have shown that many potent NAbs target the conformational epitopes located on or near the protrusions surrounding the icosahedral threefold axes, with footprints extending toward the fivefold channel. Antibody binding to these regions frequently overlaps with known receptor-binding sites, such as the heparan sulfate proteoglycan site in AAV2/AAV6 and the sialic acid sites in AAV5. This indicates that steric hinderance at these conserved interfaces is a primary neutralization mechanism. ²⁵ In some serotypes, a single Fab molecule can block multiple receptor-binding sites due to its angle of approach and spatial arrangements of protrusion, enhancing neutralization efficiency. Structural mapping also reveals that certain capsid features, including regions adjacent to the fivefold axis, serve as cross-reactive epitopes. For example, Hsi et al. demonstrated that avian AAV, despite only ∼54–58% VP1 sequence identity with human/primate serotypes, is neutralized by ∼32% of human sera due to conserved antigenic surfaces in the fivefold region and can be recognized by a human anti-AAV9 monoclonal antibody. ⁵⁰

Complement activation is another major extracellular barrier, often initiated when pre-existing anti-AAV antibodies form immune complexes with AAV capsids. This triggers classical or alternative pathways, resulting in C3 generation, opsonization, recruitment of phagocytes, and potential lytic damage through MAC formation. These events impair vector entry and can induce systemic inflammation. Inhibiting complement activation at the level of C3 or C5 disrupts this cascade, reducing immune-mediated clearance and tissue damage. As a result, complement inhibition helps preserve vector integrity and target cell viability, thereby improving transduction efficiency and safety, especially in seropositive individuals.^42,51

These insights emphasize that the extracellular neutralization is a multifaceted process, and effective strategies must simultaneously address steric hinderance, cross-reactive epitopes, and complement activation to preserve vector efficacy in seropositive host.

Fc-mediated clearance pathways and immune cell activation

Antibodies also contribute to AAV vector clearance through Fc-mediated effector functions. These mechanisms include ADCD, ADCC, and ADCP and are particularly relevant to individuals with pre-existing immunity or those receiving repeat vector administration.

ADCD is the most characterized effector mechanism. It involves activation of the classical complement pathway through C1q binding to antigen–antibody complexes, leading to C3 cleavage and MAC formation. In studies involving AAV vectors, the classical activated pathway can be activated in vitro leading to the deposition of complement component on opsonized vectors, promoting their uptake by phagocytic cells. ⁵² In vivo, high-dose AAV administration has been linked to complement activation and clinical syndromes such as TMA, implicating both classical and alternative complement pathways in these events.^42,43,53,54

ADCC is a cytolytic response in which FcγRs on effector immune cells recognize IgG-coated transduced cells and mediate cell lysis. NK cells are primary mediators, using FcγRIIIA to trigger the release of cytotoxic granules and cytokines. ADCC has not been demonstrated in the context of AAV gene therapy. However, ADCC has been observed in other viral infections and in models using AAV to deliver monoclonal antibodies with enhanced Fc activity.^55–57 These findings suggest ADCC may contribute to reduced transgene persistence in certain seropositive individuals.

ADCP is mediated by FcγR-expressing phagocytes, such as macrophages and neutrophils, that internalize and degrade IgG-opsonized particles. This mechanism is well-characterized in viral infections such as influenza and HIV, where non-NAbs enhance clearance despite low neutralizing titers. ⁵⁸ However, its role in AAV vector clearance has not been directly demonstrated. While opsonized AAV particles and transduced cells could, in theory, become targets of Fc-mediated phagocytosis, no evidence for this mechanism has been shown to date. Further studies are needed to determine if ADCP contributes to AAV vector elimination or reduced transgene expression, particularly in the context of pre-existing immunity or vector re-administration.

Among the Fc-mediated effector functions, only ADCD has been clearly demonstrated in the context of AAV gene therapy, where it has been shown to contribute to opsonization, vector clearance, and inflammation. By contrast, ADCC and ADCP, though well-documented in other viral infections, remain largely unexplored in AAV vectors. This lack of direct evidence leaves a critical gap in our understanding of how Fc-mediated pathways may influence transgene persistence or contribute to loss of transduced cells, particularly in seropositive individuals or during re-administration. Defining the presence, magnitude, and triggers of this mechanism could reveal previously overlooked barriers to efficacy and identify novel intervention points, such as selective Fc-receptor blockade, that could mitigate the impact of preformed antibodies on AAV-based gene therapy.

Determinants of susceptibility to pre-existing anti-AAV immunity

The efficiency of AAV-mediated gene transfer is not solely dictated by vector dose or route of administration. It is also critically shaped by interplay between viral capsid properties and the host’s pre-existing immune repertoire. Susceptibility to antibody-mediated neutralization varies widely, influenced by multiple determinants that act at both host and vector levels. Host-related factors include species-specific differences in immune recognition, prior exposures, and the overall immunological status of the target population. Vector-related factors encompass capsid serotype, structural epitopes targeted by neutralizing or cross-reactive antibodies, and the ability to evade the immune-mediated restriction mechanisms. These interdependent determinants are summarized in Figure 3. Understanding these determinants is essential, as even low-titer or non-NAbs can alter vector distribution, reduce transgene expression, or accelerate the clearance of transduced cells in ways that in vitro systems may fail to predict. This section examines these variables in detail beginning with the influence of species and serotype on susceptibility to pre-existing humoral immunity.

OPEN IN VIEWER

Figure 3.

Key factors affecting susceptibility to anti-AAV immunity include species and serotype differences, age-related variability in antibody titers, and tissue or route-specific influences on immune exposure and vector clearance.

Influence of species and serotype on antibody-mediated neutralization

Species-specific immune responses introduce a critical variable in AAV gene therapy, as different animals exhibit unique immune profiles, which affect antibody-mediated neutralization. For example, AAV serotypes isolated from nonhuman primates (NHPs), such as AAV7 and AAV8 from rhesus monkeys, demonstrate reduced neutralization by human sera and significantly enhanced liver-directed gene transfer compared with human-derived serotypes. ⁵⁹

Serotype-specific differences also play a crucial role in susceptibility to neutralization by pre-existing antibodies. Among the commonly used serotypes, AAV5 has historically exhibited the lowest global seroprevalence of NAbs in human populations. In a large multicenter study by Klamroth et al., AAV5 NAbs were detected in only 34.8% of participants compared with 58.5% for AAV2 and 45.6% for AAV8, supporting the widespread assumption that AAV5 is less likely to encounter pre-existing humoral immunity. ²⁰ This feature has made AAV5 a favorable option for the delivery of gene therapy candidates that might otherwise have been excluded due to immunity against other AAV serotypes. However, more recent studies, such as that by Pabinger et al., have demonstrated that immunity screens remain an important consideration. That study found a higher than expected prevalence of AAV5 antibodies (53.4%) among adult males with hemophilia, underscoring that seroprevalence can vary significantly by patient cohort, region, and assay platforms. ⁶⁰

Importantly, even when pre-existing antibodies to AAV5 are present, their impact on gene transfer appears limited. In a study by Majowicz et al., ⁶¹ patients with hemophilia B who tested positive for anti-AAV5 NAbs using a highly sensitive luciferase-based assay still achieved therapeutic levels of factor IX expression following AAV5-hFIX administration. Notably, one patient with the highest recorded pretreatment NAB titer (1:340) exhibited the greatest transgene expression within the low-dose cohort. Consistent with these observations, non-NHPs with anti-AAV5 NAB titers as high as 1:1030 also showed robust liver transduction and circulating hFIX levels. These results contrast sharply when other serotypes, such as AAV2 and AAV8, are used, when low NAb titers can often abrogate transduction. ⁶¹ The apparent resilience of AAV5 may be attributed to its capsid architecture, which shares low sequence homology with other serotypes and may be less targeted by cross-reactive or low–affinity antibodies. ⁶¹

However, resistance to antibody interference may not solely depend on classical steric inhibition observable in vitro. Non-NAbs can also impair vector efficacy. In a seminal study by Li et al., ²⁶ mice actively immunized with adenoviral vectors encoding AAV8 capsid (Ad-AAV8) mounted an immune response that significantly reduced transgene expression following AAV2-mediated gene transfer, despite the absence of cross-NAbs in vitro or cytotoxic T cell responses. Passive transfer of serum from AAV8-immunized mice was sufficient to impair AAV2 transduction in vivo, indicating that cross-binding non-NAbs can still interfere with vector efficacy. Vector genomes were initially detectable in the liver but declined progressively over time, suggesting that vector uptake occurred and inhibition occurred postcellular entry. Although the precise mechanism was not fully elucidated, it could be inferred that these non-NAbs may facilitate intracellular degradation or misrouting of the vector. Importantly, this effect was entirely antibody-dependent: AAV2 transduction remained unaffected in B-cell deficient (RAG⁻/⁻) mice, despite the presence of AAV8-specific CD8⁺ T cells. In contrast, AAV5 has shown a consistent ability to transduce target tissues even in the presence of measurable antibody levels, potentially due to its limited cross-reactivity and structural features that help evade intracellular restriction. ²⁶

Collectively, these findings underscore that both the origin of the host species and the structural properties of the AAV capsid profoundly influence susceptibility to antibody-mediated neutralization. While serotypes such as AAV2 and AAV8 are highly vulnerable even to low-titer neutralizing or cross-reactive antibodies, AAV5 appears to retain functional transduction under conditions that would typically impair gene transfer. Importantly, the ability of non-NAbs to inhibit transduction via intracellular pathways adds a layer of complexity not captured by conventional seroprevalence or in vitro neutralization assays. These insights emphasize the need for a more refined approach to AAV vector selection—one that integrates serological profiling with mechanistic understanding of antibody-mediated interference across species and serotypes to maximize therapeutic success.

Age-dependent variability in antibody responses

The influence of age on antibody titers introduces another layer of complexity to AAV-based therapies, as antibody levels generally increase with age due to cumulative viral exposures.^47,62 This age-related escalation in antibody titers has been associated with a reduced AAV transduction efficiency, both in clinical and preclinical studies. Consequently, younger individuals with fewer viral exposures and lower baseline antibody levels exhibit improved gene delivery and expression. In contrast, older patients, with higher and more complex antibody repertoires, may face greater challenges in achieving efficient transduction, making immune modulation strategies potentially necessary to improve therapeutic outcomes. ⁶³ These insights emphasize the need to consider age-stratified antibody screening in clinical trial design and patient selection.

A recent study profiling anti-AAV9 antibodies across the Chinese population (0–90 years) found a significantly higher seroprevalence of NAbs in adults (75.0%) compared with children (34.3%), with the lowest rates observed in children aged 6 months to 3 years, an age range identified as optimal for gene therapy. Notably, children under 6 months of age had elevated antibodies, likely due to maternal transfer rather than their own infection or exposure. While average seropositivity increased with age, age by itself was not a reliable predictor of NAb or total binding antibody (TAb) titers, suggesting that individual antibody levels vary independently of age. ²¹

A similar trend was observed in the United States, where a validated AAV6 transduction inhibition assay demonstrated age-associated variation in seroprevalence. Among 120 healthy adults aged 18–62 years, 40% were seropositive for preexisting anti-AAV6 NAbs compared with 19% of pediatric donors under 12 years of age, with detectable responses observed as early as 2 years of age. ²² These findings underscore the importance of considering both age and developmental immune status when determining the optimal timing for AAV-based interventions.

Tissue- and route-specific factors modulating antibody-mediated neutralization

The impact of pre-existing immunity on AAV gene therapy is strongly influenced by both the target tissue and the route of vector administration. Immune-privileged sites such as the central nervous system (CNS) and the eye exhibit reduced susceptibility to antibody-mediated neutralization presumably due to limited lymphatic drainage, lower density of antigen-presenting cell (APC), and reduced expression of MHC molecules.^64,65 Studies targeting these regions have demonstrated minimal humoral and cellular immune responses, even in the absence of immunosuppression. For example, AAV vectors delivered via intrathecal or intravitreal injection elicited no significant T cell activation or antibody-mediated transgene clearance, highlighting the protective immune environment of these sites. Lower doses of AAV were sufficient to achieve therapeutic expression reinforcing the advantage of immune privilege for durable gene transfer. ⁴⁹

Localized delivery methods, such as intraparenchymal or intracerebroventricular injections, can reduce the vector’s exposure to circulating NAbs. In contrast, intravenous (IV) administration introduces the vector directly into the bloodstream, increasing susceptibility to neutralization in seropositive individuals. Highly vascularized organs such as the liver and spleen are particularly vulnerable, where rapid antibody binding and antigen presentation may lead to loss of transgene expression and immune-mediated toxicity. ⁶⁶

These findings emphasize that both tissue-specific immune characteristics and administration routes significantly influence susceptibility to neutralization. Tailoring administration strategies based on the patient’s immune profile and the intended target tissue is essential to maximize therapeutic efficacy and mitigate immune-related barriers.

Strategies to circumvent pre-existing immunity

Addressing the immune barriers outlined in the preceding section requires an integrated set of solutions that neutralize, bypass, or evade antibody- and cell-mediated interference. These strategies fall into three overlapping domains: immune-focused strategies that transiently reduce or modulate humoral and cellular responses; delivery-focused strategies that optimize vector administration to minimize immune exposure; and capsid-focused strategies that engineer or shield the vector from immune recognition. This framework summarized in Figure 4 serves as a roadmap for the following discussion, which examines each domain in turn, from targeted immunoglobulin depletion to precision delivery routes and next-generation vector design.

OPEN IN VIEWER

Figure 4.

Classification of strategies to overcome pre-existing anti-AAV immunity. (A) Immune-focused strategies include plasmapheresis/immunoadsorption, enzymatic IgG/IgM cleavage (Imlifidase, KJ103), corticosteroids, and depletion of B and plasma cells (e.g., CD20 mAb and Bortezomib). (B) Delivery-focused strategies involve tissue-targeted administration and targeting immune-privileged sites. (C) Capsid-focused strategies utilize decoy capsids and engineered capsids with altered antigenicity.

Immune-focused strategies

The first category targets the immune system directly, aiming to lower circulating antibody titers, impair antibody function, or modulate immune activation to create a permissive window for vector administration. These approaches range from extracorporeal techniques such as plasmapheresis and immunoadsorption to enzymatic degradation of immunoglobulins, pharmacologic immunosuppression, and depletion of antibody-producing B cells. While each method varies in specificity, duration of effect, and clinical feasibility, they share the goal of reducing the humoral barriers to levels compatible with efficient AAV transduction.

Immunoglobulin removal strategies

Plasmapheresis: Plasmapheresis and immunoadsorption are extracorporeal blood purification techniques that remove circulating antibodies, including anti-AAV NAbs, thereby providing a temporary window for effective vector administration. These methods have been explored in NHPs and human patients as potential strategies to overcome the barrier posed by pre-existing humoral immunity to AAV vectors.^67,68

Plasmapheresis, or therapeutic plasma exchange, nonselectively removes plasma proteins, including immunoglobulins, from circulation and has been shown to reduce anti-AAV NAb titers below inhibitory thresholds in seropositive subjects. In NHP studies, multiple rounds of plasmapheresis significantly lowered anti-AAV NAb levels, enabling successful systemic transduction that would otherwise be blocked. This has been comprehensively reviewed by Mingozzi et al. ⁶⁹ Importantly, Bertin et al. demonstrated in a pilot study that three cycles of plasmapheresis in cynomolgus macaques led to a ∼100-fold reduction in NAb titers, enabling successful AAV8 vector re-administration and robust liver transduction, comparable to that seen in naïve animals. ⁷⁰ To mitigate hypogammaglobulinemia, these animals received IgG-reconstituted plasma from AAV-seronegative donors.

In a more recent study, Potter et al. evaluated the safety and efficacy of plasmapheresis as a pretreatment to allow redosing of an AAVrh74-based gene therapy for Duchenne muscular dystrophy in NHPs. Two to three rounds of plasmapheresis significantly reduced circulating anti-AAVrh74 antibodies and enabled safe redosing. Animals re-dosed without plasmapheresis at high antibody titers (≥1:51,200) experienced acute hypersensitivity reactions, whereas those that underwent plasmapheresis exhibited no adverse events. Importantly, vector genome biodistribution, transgene expression, and immune responses (including T cell and complement activation) were not negatively affected by plasmapheresis. ⁷¹

Despite its therapeutic potential, plasmapheresis presents notable limitations in the context of AAV gene therapy. The procedure often requires multiple cycles to sufficiently reduce NAb levels, particularly in individuals with high baseline titers. Moreover, its lack of specificity leads to the broad removal of immunoglobulins, which can result in hypogammaglobulinemia and increased susceptibility to infection. ⁷² In autoimmune and transplant settings, this risk is mitigated by the coadministration of IV immunoglobulin (IVIg) to restore immunoglobulin levels. However, this approach is not feasible in AAV gene therapy, as IVIg preparations contain high-titer anti-AAV antibodies that could reintroduce or even exacerbate the humoral barrier to vector transduction. ⁷³ Thus, while plasmapheresis can offer a temporary solution for reducing anti-AAV antibody levels, its logistical complexity, limited durability, and risk of immunological disruption highlight the need for more selective and scalable approaches to safely enable gene therapy in seropositive patients.

Enzymatic cleavage of immunoglobulins and immune modulation

Targeted enzymatic cleavage of antibodies is an emerging strategy to transiently overcome humoral immunity and facilitate AAV gene transfer. Imlifidase (IdeS), a cysteine protease derived from Streptococcus pyogenes, specifically cleaves human IgG below the hinge region, abolishing Fc-mediated effector functions such as complement activation and ADCC. Originally developed to enable kidney transplantation across donor-specific HLA antibody barriers, ⁷⁶ IdeS has been employed to degrade anti-AAV IgG and restore AAV-mediated transduction in seropositive individuals. ⁷⁷ This approach has enabled successful AAV delivery and even vector re-administration in NHPs with pre-existing NAbs. ⁷⁷ However, its clinical utility is constrained by the high prevalence of pre-existing anti-IdeS antibodies, also known as anti-drug antibodies, which have been observed in over 90% of the general population and may neutralize IdeS or potentially provoke hypersensitivity reactions.^78,79 Moreover, IdeS exhibits minimal activity against murine IgG, thus limiting its use in conventional mouse models and necessitating alternative strategies such as passive immunization with human IgG or studies in humanized mice. This scenario not only complicates mechanistic investigations but also reduces the predictive power of murine models for human translation.

To address some of these limitations, KJ103, a next-generation recombinant IgG-degrading enzyme derived from Streptococcus equi, has been developed with enhanced substrate breadth. Unlike IdeS, KJ103 efficiently cleaves all human IgG subclasses at low doses. In two Phase I trials, it reduced circulating IgG by over 90% within 6 hours, with recovery toward baseline levels in over half of the subjects by 6 months, suggesting a favorable safety and immunogenicity profile. ⁸⁰ In vitro, KJ103 substantially lowered NAb titers against AAV2, even in highly seropositive individuals. However, KJ103 has not yet been fully evaluated in murine models, and its activity against mouse IgG appears minimal. As a result, further studies in preclinical models, particularly large animals with closer immunological homology to humans, are warranted to fully assess their translational utility and optimize dosing strategies.

Recent efforts have expanded antibody-cleaving approaches beyond IgG to address a broader spectrum of immune barriers that limit AAV-based gene therapy. While IgG has traditionally been the primary focus due to its neutralizing capacity and Fc-mediated effector functions, emerging evidence indicates that IgM plays a key role in initiating the classical complement cascade, which contributes to vector clearance and treatment-related toxicities. ⁸¹ To mitigate these effects, Smith et al. engineered IceM, an IgM-cleaving enzyme, and IceMG, a bifunctional enzyme combining IceM with the IgG protease IdeZ. ⁸² These enzymes target both soluble and membrane-bound forms of immunoglobulins, including B cell receptors, thereby impairing antigen recognition and downstream signaling. IceMG was shown to reduce both IgM- and IgG-mediated complement activation and significantly diminished C3a production in human serum exposed to AAV9. In NHPs, IceMG transiently suppressed NAb levels, reduced complement activity, and improved transduction efficiency. ⁸² These findings underscore the relevance of targeting both IgM and IgG in mitigating complement-related toxicities such as TMA as reported by Salabarria. ⁸³

Beyond enzymatic approaches, pharmacologic immunosuppression remains a widely used strategy to dampen innate immune responses and enhance vector efficacy. Corticosteroids such as prednisolone have been used prophylactically to reduce acute innate immune responses and improve vector uptake. In liver gene transfer models employing AAV5, daily prednisolone administration initiated before vector dosing enhanced transgene expression and vector genome persistence. ⁸⁴ Handyside et al. continued to show that, mechanistically, corticosteroids suppressed type I interferons, NF-κB activation, and inflammasome responses, while upregulating platelet-derived growth factor receptor alpha, which is a coreceptor for AAV5 on hepatocytes. This led to improved vector internalization and reduced intersubject variability. ⁸⁴ These findings are supported by a recent meta-analysis of 73 studies, which found that prophylactic corticosteroids were used in over 75% of AAV gene therapy trials to mitigate immune-related toxicities and sustain transgene expression. Despite variations in immunosuppressive regimens, corticosteroids consistently emerged as an effective and safe component of clinical protocols. ⁸⁵

B cell and plasma cell depletion strategies

Monoclonal antibodies targeting CD20, such as rituximab or its murine equivalent 18B12, deplete mature and memory B cells. In hemophilia A mice, anti-CD20 therapy reduced inhibitors against factor VIII and eliminated FVIII-specific memory B cells. When combined with rapamycin, this regimen suppressed anti-AAV8 and anti-FVIII antibody responses and prevented resurgence upon rechallenge. ⁸⁶ These findings support combined immunosuppression to facilitate vector redosing.

Bortezomib, a clinically approved proteasome inhibitor, targets terminally differentiated plasma cells through the inhibition of 26S proteasome, thereby reducing sustained antibody production. ⁸⁷ In a key study by Choi et al., a combination therapy with bortezomib and an anti-CD20 monoclonal antibody significantly reduced anti-AAV8 NAbs and restored vector transduction in mice. ⁸⁸ The most effective suppression of NAbs was observed with a 16-week combination therapy, which enhanced liver transgene expression and vector genome levels. ⁸⁸ However, consistent findings from Karman et al. showed that bortezomib monotherapy, despite lowering AAV-specific titers at 1 mg/kg over 20 weeks, failed to permit successful vector redosing. This limitation was attributed to persistence of memory B cells, which were not depleted by bortezomib alone. ⁸⁹

Supporting these limitations, Chaanine et al. demonstrated that in a rat pressure-overload heart failure model, bortezomib did not enhance the efficacy of AAV9.SERCA2a gene therapy. There was no improvement in SERCA2a expression, vector genome delivery, or cardiac function. Interestingly, in vitro assays revealed that while bortezomib enhanced AAV9 transduction in HeLa cells and neonatal cardiomyocytes, it inhibited transduction in adult cardiomyocytes. These results suggest that the impact of proteasome inhibition is strongly cell type dependent and may even be detrimental in certain adult tissues. ⁹⁰

In contrast, Monahan et al. demonstrated that bortezomib can significantly augment AAV transduction in both small and large animal models of hemophilia, particularly when using oversized transgenes. In mice, bortezomib increased factor VIII expression by ∼600% with AAV2 and ∼300% with AAV8 vectors. In hemophilia A dogs, coadministration of AAV8 and bortezomib normalized clotting time and reduced bleeding events by 90% over a 32-month period. These effects were attributed to enhanced nuclear import of AAV genomes and stabilization of vector particles against proteasomal degradation. ⁹¹

Together, these findings indicate that bortezomib’s utility depends on both the transgene payload and the tissue context. While it holds promise as an adjuvant for enhancing AAV-mediated expression, particularly for large transgenes, its ability to overcome pre-existing humoral immunity remains limited unless paired with agents that also deplete memory B cells. Therefore, combinatorial strategies that simultaneously target both plasma and memory B cells offer a more robust approach to immunomodulation for AAV redosing.

Emerging B cell-targeted therapies

In addition to monoclonal antibodies and proteasome inhibitors, emerging immunotherapies targeting B cell survival and function have shown promise in enabling AAV vector redosing. Anti-BAFF (B cell activating factor) therapies, which disrupt B cell maturation and survival, have demonstrated efficacy in suppressing anti-AAV immune responses in preclinical models. For example, combination regimens using anti-BAFF with anti-CD20 monoclonal antibodies facilitated successful vector readministration in mice, ⁸⁸ although this strategy has not yet been shown to eliminate pre-existing antibodies. A related approach was validated by Salabarria et al., ⁸³ where targeted immune modulation consisting of rituximab and sirolimus effectively suppressed the anti-capsid IgG and IgM responses, as discussed earlier in the “Innate Immune Recognition of AAV” section. Another innovative approach involved the use of CD19-targeted chimeric antigen receptor (CAR) T cells, which selectively eliminated CD19⁺ B cells, including antibody-producing precursors. A recent patent describes the use of CD19-targeted immunotherapies including CAR-T cells, to deplete B cells and lower NAb levels against AAV vectors, thereby enabling effective transduction in seropositive mice. ⁹² These B cell-targeted therapies, alone or in combination with established immunosuppressive agents, hold potential for overcoming pre-existing humoral immunity and expanding the applicability of AAV gene therapies.

In summary, these immune-focused strategies, ranging from antibody depletion through plasmapheresis or immunoadsorption, enzymatic cleavage of immunoglobulins, corticosteroid-based modulation of inflammatory responses to B and plasma cell targeting, represent useful tools to transiently overcome pre-existing humoral immunity and facilitate successful AAV transduction. While each strategy offers distinct advantages, their effectiveness often depends on patient-specific factors such as antibody titers, immune history, and vector serotype. However, overcoming immune barriers does not solely depend on systemic interventions. Optimizing the route and site of vector delivery also plays a critical role in minimizing immunological interference. The following section explores delivery-focused approaches that further enhance AAV gene therapy efficacy by leveraging tissue-specific immune environments and targeted administration techniques.

Immune evasion through targeted delivery

The route and location of AAV vector administration critically influence the interaction between the vector and the host immune system. By modulating delivery strategies, it is possible to minimize immune exposure, bypass systemic antibody barriers, and improve therapeutic efficacy, particularly in seropositive individuals.

Targeting immune-privileged sites

Immune-privileged tissues such as the CNS, retina, and joints provide a relatively protected microenvironment.^64,93 In diseases where these tissues are the direct therapeutic targets, such as spinal muscular atrophy (CNS), Leber congenital amaurosis (retina), or hemophilic arthropathy (joints), AAV delivery via intrathecal, subretinal, or intra-articular routes has shown reduced exposure to circulating NAbs and APCs.^94–97 This localized delivery can preserve transduction efficiency even in seropositive individuals. Clinical studies have demonstrated favorable immune profiles and long-term expression following gene transfer to these compartments, reinforcing the feasibility of targeting immune-privileged tissues when they are directly involved in the disease pathology. ⁴⁹

Localized administration

Localized delivery methods such as intramuscular (IM), intraparenchymal (e.g., intracerebral or hepatic), or intracerebroventricular injections restrict the biodistribution of AAV vectors and help limit their interaction with circulating antibodies and immune cells. These approaches facilitate direct tissue transduction, reduce vector dose requirements, and potentially minimize the induction of systemic immune responses. For example, IM administration has been associated with reduced vector leakage into systemic circulation, which has been proposed as a way to lower the risk of capsid-specific immune activation and complement-mediated toxicity. ⁴⁹

Other delivery techniques such as hydrodynamic limb vein injection ⁹⁸ or convection-enhanced delivery (CED) ⁹⁹ have been explored to improve AAV distribution while mitigating systemic immune detection. CED allows direct infusion of vector into the interstitial space under positive pressure, enabling broader distribution while circumventing the blood–brain barrier and limiting exposure to circulating NAbs. Studies confirm CED’s utility in safely and effectively distributing AAV vectors in CNS tissues.^100,101

Timing and dosing considerations

Vector dose and infusion rate also play critical roles in determining immune activation. Gradual or fractionated dosing may help modulate acute innate responses and reduce systemic capsid burden. In some settings, transient immunosuppressive regimens have been coadministered to enhance the safety and persistence of localized delivery. The timing of immunosuppression is critical, as peritransduction immunomodulation can affect immune priming versus tolerance. ⁴⁹

Together, these delivery-focused approaches provide important tools for improving transduction efficiency while reducing vector immunogenicity, especially in patients with pre-existing anti-AAV antibodies.

Capsid-focused strategies

Use of decoy particles and empty capsids

Decoy particles and empty capsids have historically been explored as a strategy to shield therapeutic AAV vectors from pre-existing NAbs. Because empty capsids are structurally identical to full AAV capsids but devoid of genetic material, these particles can competitively bind circulating antibodies, thereby preserving the functionality of genome-containing vectors and improving transduction efficiency. Studies have focused on optimizing the capsid composition and dosing ratio of empty to full particles to maximize decoy effectiveness while minimizing immunogenicity and vector competition. ¹⁰²

A study by Mingozzi et al. ¹⁰² provided strong experimental evidence that coadministering excess empty capsids with therapeutic AAV vectors can effectively titrate out NAbs and restore transgene expression. Their findings demonstrated that empty capsids act as functional antibody decoys in a dose-dependent manner in mice, humans, and NHPs, thus rescuing vector transduction even in the presence of moderate-to-high antibody titers. Furthermore, they introduced mutant empty capsids with disrupted receptor-binding domains to prevent cellular entry and minimize MHC class I antigen presentation, thereby reducing the risk of capsid-specific CD8⁺ T cell activation. These noninfectious capsids retained full decoy function while improving the safety profile, suggesting that, when precisely formulated based on individual NAb titers, decoy capsids can be both effective and safe.

However, this strategy is not without limitations. A pivotal study by Gao et al. ¹⁰³ demonstrated that the presence of empty capsids, particularly partially empty (PE) ones that contain fragmented genomes, can significantly reduce transduction efficiency and exacerbate immune-mediated side effects in vivo. In both C57BL/6 and BALB/c mouse models, increasing the proportion of empty virions in AAV8 vector preparations led to a marked decrease in reporter gene expression (up to 70%) and elevated serum ALT levels, a marker of liver damage. Notably, PE capsids derived from the production process were more immunogenic than completely empty (CE) capsids, likely due to enhanced capsid antigen presentation and residual genome content.

Together, these studies illustrate the dual nature of decoy capsids. On one hand, controlled use of high-purity, noninfectious empty capsids offers a promising strategy to overcome NAbs and enable gene delivery in seropositive individuals, especially if the capsids are engineered to reduce immunogenicity and minimize receptor competition.^102,104 On the other hand, impure capsid preparations, particularly those containing PE virions, can dilute vector potency, exacerbate immune responses, and compromise safety. ¹⁰⁵ These findings offer mechanistic insight into prior clinical concerns, where preparations with high ratios of PE capsids not only reduced efficacy but also triggered capsid-specific CD8⁺ T-cell responses and complement activation.

As a result, the field has shifted toward producing high-purity AAV vectors with minimal levels of empty or partially filled capsids. Current efforts are focused on developing next-generation decoy strategies, such as mutant CE capsids, capsid mimetics, or transient immunosuppression, to balance immune shielding with safety.^105,106

Prioritizing the most effective strategies to overcome pre-existing humoral immunity

Overcoming pre-existing humoral immunity to AAV vectors is most effective when combining immune-modulating and capsid-focused strategies. Dual IgG/IgM-cleaving enzymes such as IceMG offer potent neutralization and complement blockade, whereas B cell depletion regimens—especially anti-CD20 antibodies combined with proteasome inhibitors—reduce both memory and plasma cells. Immune-privileged delivery routes (e.g., CNS and retina) provide validated options to bypass circulating antibodies without systemic immunosuppression, although they are only situationally useful. Additionally, rational capsid engineering and the use of high-purity decoy capsids can minimize vector clearance, provided partially filled virions are excluded. Ultimately, a multipronged approach integrating antibody removal, targeted delivery, and capsid design offers the best chance for safe and repeatable AAV therapy in seropositive patients.

Implications and future directions for AAV gene therapy in the context of pre-existing immunity

The influence of pre-existing immunity on AAV-based gene therapies is shaped by species-specific immune responses, serotype specificity, patient age, tissue type, and administration route. This complexity presents a critical challenge for researchers and clinicians, who must anticipate immune barriers and develop strategies that overcome them. Optimizing administration methods, selecting AAV serotypes with low seroprevalence, and incorporating individualized immune profiling into patient selection are essential components of successful gene therapy design.

Looking ahead, several avenues hold promises for advancing our understanding and management of pre-existing anti-AAV immunity. First, it is important to have an improved understanding of the impact of natural parvovirus infections and vaccinations (when using animal models) on the generation of antibodies that cross-react with AAV capsids. Given that AAV belongs to the Parvoviridae family, antibodies generated against related viruses (e.g., porcine or human parvoviruses) may neutralize AAV vectors unintentionally. Future studies using animal models such as pigs could clarify how such immune cross-reactivity develops and how it affects AAV-based transduction. These findings could inform both vaccine development and gene therapy vector design to minimize unintended immune interference.

Another promising direction is the investigation of Fc-mediated mechanisms, such as ADCC and ADCP, in the context of AAV immunogenicity. While these effector functions are well characterized in other viral systems, their role in shaping the immune response to AAV vectors remains poorly understood. Understanding how Fc receptor engagement influences vector clearance, antigen presentation, and inflammation could reveal new opportunities for mitigating immune responses. For instance, strategies that reduce opsonization by IgG such as creating capsid modifications that minimize antibody binding may indirectly attenuate Fc-mediated effector functions such as ADCC and ADCP. Although Fc receptor engagement occurs via antibodies rather than direct capsid interaction, limiting antibody binding to the capsid could reduce downstream FcγR activation.

Future work should continue to refine and validate immunomodulatory tools such as IgG- and IgM-cleaving enzymes, B cell depletion strategies, and engineered immune regulatory cells (e.g., AAV-CAR Tregs). Emerging hypotheses, such as the potential role of TRIM21 in recognizing intracellular antibody-bound AAV particles, also merit investigation. Although TRIM21’s function has been demonstrated to act during the intracellular neutralization of other nonenveloped viruses, direct evidence in the context of AAV remains lacking. However, it is plausible that antibody-coated AAV vectors that evade extracellular neutralization may be targeted by TRIM21 once internalized. Exploring this pathway could reveal novel intracellular mechanisms of AAV restriction and inform capsid or payload design to improve vector durability. A better understanding of the timing, dosing, and specificity of these interventions will be essential for their safe and effective clinical translation.

Finally, the development of re-dosable AAV platforms remains an important goal. Approaches such as orthogonal capsid engineering, transient immunosuppression, or vector delivery to immune-privileged sites should be prioritized.^66,107 These efforts will help broaden access to gene therapies and reduce the exclusion of patients with pre-existing immunity.

Altogether, integrating immunological insights with vector design and clinical protocols will be key to realizing the full potential of AAV gene therapies in diverse patient populations.