Clinical immunogenicity in rAAV gene therapy: Insights and implications

Abstract

Recombinant adeno-associated viruses (rAAV) provide prolonged therapeutic protein expression in multiple target organs after single administration, delivering significant clinical benefits with a well-characterized safety profile, and can be further supported by active immune management strategies. Systemic rAAV gene therapy is associated primarily with transient hepatocellular injury, while complement-mediated thrombotic microangiopathy and immune-related cardiac events occur in specific contexts; DRG-associated neurotoxicity remains predominantly nonclinical with only isolated suspected human cases. The extent, kinetics, and clinical consequences of immune activation are influenced by vector-related factors, dose, route of administration, and patient-specific characteristics. Hence, the immune responses may vary significantly across products, treatment regimens, and patient populations. These responses depend on disease indication, dosing and immunosuppression regimens, route of administration, vector serotypes, and underlying immune status. Significant progress has been made in understanding mechanistic and clinical aspects of rAAV immunogenicity. Strategies for clinically managing immunogenicity are being developed to achieve desired efficacy, minimize immune-mediated adverse events, and potentially expand treatment options for patients with or without previous AAV exposure. This review focuses on clinical manifestations of immune responses to rAAV, temporal patterns of these safety events and underlying immunological mechanisms, and discusses clinical considerations, risk assessment, and potential emerging immunogenicity mitigation strategies.

Graphical abstract

Recombinant AAV gene therapies deliver durable clinical benefit but face immune-mediated challenges that vary by vector, dose, route, and patient. Gulve and colleagues synthesize the clinical manifestations, temporal patterns, and mechanisms of rAAV immunogenicity, highlighting risk assessment and emerging mitigation strategies to support safer, more effective gene therapy development.

Introduction

Adeno-associated viruses (AAVs) are preferred vectors for delivering therapeutic payloads in gene therapy primarily due to their low pathogenic nature, ability to enable long-term gene expression, and minimal risk of insertional mutagenesis. The recombinant AAV (rAAV) genome consists of single-stranded DNA of approximately 4.7 kb that includes a transgene expression unit, defined as a promoter, a transgene and optional regulatory element/s. Inverted terminal repeats (ITRs) flank both ends of the viral genome. ITRs are important signals necessary for packaging of this rAAV genome in an AAV capsid. Differences in the capsid proteins lead to different AAV serotypes which are identifiable by antibodies.¹ At least 13 naturally occurring AAV serotypes and numerous natural variants have been identified, while a large and continually expanding number of engineered capsid variants have been generated through directed evolution and rational engineering approaches.^2,3,4,5

Upon administration, the rAAV particles have the potential to elicit innate and adaptive immune responses. Innate immune responses are triggered through Toll-like receptors (TLRs) expressed by macrophages or dendritic cells (DCs) leading to elevated expression of pro-inflammatory cytokines and chemokines.^6,7,8 The complement system is another crucial component of the innate immune system, that can contribute to capsid opsonization and clearance while also initiating an inflammatory response. Complement proteins can promote neutralization of viruses through several mechanisms including opsonization of viral particles, virion aggregation, and in some contexts, complement-mediated lysis. Complement activation by rAAV can lead to capsid opsonization, immune complex formation, and downstream inflammatory signaling. Additionally, complement opsonization of the capsid makes the vector more susceptible to phagocytosis by immune cells.⁹ The adaptive immune response is antigen-specific and involves B cells, which are responsible for eliciting an antibody response, along with T cells, which orchestrate cell-mediated cytotoxicity and help antibody producing B cells.^6,10 The magnitude of the innate and adaptive immune responses in gene therapy depends upon multiple factors such as dose, route of administration, capsid serotype, pre-existing antibodies, CpG content of the vector, and homology of the transgenic protein to the endogenous proteins.

With more than 100 rAAV gene therapies at different stages of clinical development,¹¹ a considerable repository of data detailing clinical outcomes is available in the literature. rAAV gene therapies have demonstrated clinically meaningful efficacy following single-administration treatment in RPE65-mediated inherited retinal dystrophy, spinal muscular atrophy (SMA), hemophilia A, hemophilia B, and aromatic L-amino acid decarboxylase (AADC) deficiency, and OTOF-associated sensorineural hearing loss, with pivotal trials in these indications leading to regulatory approvals by the FDA (Table 1). Patterns of clinical outcomes observed in multiple clinical studies may help inform development of new interventions leading to successful efficacious outcomes and adverse event management. Because rAAV immunogenicity is shaped by multiple vector-, treatment-, and host-related factors, its effective management will likely depend on an integrated strategy incorporating vector design, patient selection, optimized dosing, targeted immunomodulation, and monitoring of innate and adaptive immune responses.

Table 1.

FDA-approved rAAV gene therapies, associated immunogenicity-related adverse events, and immunosuppression regimen as listed in the label

Trade name	Generic/proper name(s)	Initial US approval	Indication	Route of administration	Dose and AAV serotype	Potential immunogenicity-related adverse events	Immunosuppression regimen
LUXTURNA	VORETIGENE NEPARVOVEC-RZYL	2017	RPE65 mutation-associated retinal dystrophy	subretinal administration	1.5 × 1011 vg AAV2	N/A	corticosteroids (prednisone)
Zolgensma	ONASEMNOGENE ABEPARVOVEC-XIOI	2019	spinal muscular atrophy (SMA)	intravenous infusion	1.1 × 1014 vg/kg AAV9	acute serious liver injury, systemic immune response, thrombocytopenia, TMA	corticosteroids (prednisolone)
HEMGENIX	ETRANACOGENE DEZAPARVOVEC	2022	hemophilia B	intravenous infusion	2 × 1013 gc/kg AAV5	infusion related reactions, hepatotoxicity	corticosteroidsb (prednisolone)
ELEVIDYS	DELANDISTROGENE MOXEPARVOVEC-ROKL	2023	Duchenne muscular dystrophy (DMD)	Intravenous infusion	1.33 × 1014 vg/kga AAVrh74	infusion-related reactions, acute serious liver injury, immune-mediated myositis, myocarditis	corticosteroidsc (prednisone)
ROCTAVIAN	VALOCTOCOGENE ROXAPARVOVEC-RVOX	2023	hemophilia A	intravenous infusion	6 × 1013 vg/kg AAV5	infusion-related reactions, hepatotoxicity	corticosteroidsd (prednisone)
KEBILIDI	ELADOCAGENE EXUPARVOVEC-TNEQ	2024	aromatic L amino acid decarboxylase (AADC) deficiency	intraparenchymal infusion	1.8 × 1011 vg/kg AAV2	N/A	N/A
BEQVEZ	FIDANACOGENE ELAPARVOVEC-DZKT	2024	hemophilia B	intravenous infusion	5 × 1011 vg/kg AAVRh74 var	infusion-related reactions, hepatotoxicity	corticosteroids (prednisone)
ITVISMA	ONASEMNOGENE ABEPARVOVEC-BRVE	2025	spinal muscular atrophy (SMA)	intrathecal injection	1.2 × 1014 vector genomes (vg) AAV9	hepatotoxicity, thrombocytopenia, TMA	corticosteroids (prednisolone)
OTARMENI	LUNSOTOGENE PARVEC-CWHA	2026	OTOF-associated sensorineural hearing loss	intracochlear infusion	7.2 × 1012 vg AAV1	N/A	corticosteroids (prednisone)

^a9.31 × 10¹⁵ vg/kg for >70 kg body weight.

^bOther regimens to be considered if corticosteroids fail or contraindicated.

^cOther immunosuppressants e.g., calcineurin-inhibitor may be considered.

^dtacrolimus, mycophenolate mofetil if corticosteroids contraindicated.

In this review, we examine the landscape of important immune-associated and multifactorial toxicities relevant to rAAV gene therapy, including thrombotic microangiopathy (TMA), myocarditis, hepatotoxicity, and dorsal root ganglion (DRG) toxicity.⁶ Many reviews have addressed how the immune system responds to rAAV gene therapy, whereas others have offered extensive summaries that include topics such as bioanalytical methods, potential risks, and mitigation strategies. In contrast, this review focuses specifically on the clinical experience of immunogenicity-related events following rAAV gene therapy and their association with product- and patient-related risk factors. These events are discussed in the context of rAAV capsid type and relevant patient-specific risk factors. We further highlight unresolved challenges in the gene therapy field and discuss potential strategies for effective monitoring and mitigation of immunogenicity-related adverse events. The following sections address mechanisms of immune responses to rAAV vectors, clinical observations from gene therapy studies, and determinants of adverse events related to rAAV vector-specific and patient-specific factors.

Characteristics of immune responses to rAAV

Natural AAV infection is typically asymptomatic in humans and considered non-pathogenic, whereas rAAVs have been shown to elicit both the innate and adaptive immune responses in human clinical trials, in a dose- and route-of-administration-dependent manner.^10,12 The following section describes the current understanding of immune mechanisms involved in rAAV recognition and response, drawing on both preclinical studies that have elucidated underlying molecular pathways and clinical observations that have confirmed or suggested their relevance in humans. Where evidence derives primarily from animal models, this is explicitly noted.

Dose of rAAV plays a crucial role in determining the characteristics of the immune response. Even low dose levels of rAAV may elicit both an antibody response and a T cell response; however, T cell responses may not be always detected due to their limited number in circulation.^13,14 In addition, the route of administration also plays a significant role in determining the intensity of immune response. Administration in the immune privileged sites like the ocular compartment or brain is less likely to stimulate a robust immune response compared to a more systemic intravenous (i.v.) administration.^15,16 In addition to route of administration, both the target tissue or organ and the intrinsic tissue tropism of the rAAV vector can influence the immune response, as these factors govern the sites of vector distribution, transgene expression, and exposure to distinct local immune environments. The liver has a specialized immunoregulatory microenvironment that can promote peripheral tolerance under some conditions; however, induction of transgene-specific tolerance by liver-directed AAV has been demonstrated most clearly in murine models, including immunocompetent mice and factor IX-deficient hemophilia B mice.¹⁷

Beyond these determinants, the qualitative features of the immune response to rAAV can be understood within the classical framework of a protective immune response against an infectious agent, which generally involves three signals: an innate immune danger signal, an antigen-specific signal, and a co-stimulatory signal.¹⁸ Among the innate pathways implicated in rAAV sensing, TLR9-mediated recognition of unmethylated cytosine-phosphate-guanine (CpG) motifs within the vector genome is the most extensively characterized, primarily through in vitro and mechanistic studies in murine models, including TLR9- or MyD88-deficient mice.^7,19 Additional preclinical studies have suggested potential contributions from AAV capsid recognition by TLR2 in human nonparenchymal liver cells and from cytosolic sensing of double-stranded RNA (dsRNA), potentially generated through bidirectional transcription from AAV ITRs, in human cell lines, primary human hepatocytes, and human hepatocyte–xenografted mice; however, evidence for these pathways remains limited.^20,21 Another arm of the innate immune response, the complement system, can also be activated by immune complexes formed between pre-existing or treatment-induced anti-AAV antibodies and the rAAV vector.^22,23

Innate immune responses not only aid in the initiation of but may also potentiate adaptive immune responses.²⁴ In murine models, coordination between plasmacytoid dendritic cells (pDCs, sensing the unmethylated CpG motifs using TLR9) and the conventional DCs promotes generation and stimulation of capsid-specific cytotoxic CD8+ T cells, resulting in killing of rAAV-transduced cells.^12,19,25,26 In addition to this TLR9/pDC-dependent axis, an alternative pathway of CD8⁺ T cell activation has recently been identified in murine models. In comparative studies using wild-type mice and mice deficient in components of TLR9/type 1 interferon or IL-1R1-MyD88 signaling, following hepatic AAV gene transfer, transgene product-specific CD8+ T cell responses occurred independently of TLR9 and type 1 interferon signaling and instead depended on interleukin-1α and interleukin-1β signaling through interleukin-1 receptor 1 (IL-1R1)-MyD88.²⁷ In a separate murine model of AAV1-mediated muscle gene transfer, IL-1R1 signaling also contributed to transgene-product specific CD8+ T cell responses, where it functioned redundantly with TLR9-type 1 interferon signaling and potentially other nucleic acid-sensing pathways in a vector dose-dependent manner.²⁸

Building on these innate signals, the adaptive immune response to rAAV includes activation and proliferation of antibody-producing B cells and helper CD4+ T lymphocytes. Activated CD4⁺ T cells provide critical signals (e.g., CD40L, IL-2) that drive B cell differentiation into plasma cells producing both binding and neutralizing antibodies. The majority of the human population is known to have pre-existing antibodies to various serotypes of AAV.^29,30,31,32 These pre-existing as well as treatment-emergent anti-AAV antibodies can negatively impact safety and efficacy of the treatment.^33,34,35 Treatment-emergent anti-AAV antibodies generally emerge during the first two weeks post-administration.³⁶ Following i.v. rAAV administration, treatment-emergent anti-capsid IgM titers typically become detectable by day 5 and IgG by day 7–14, with peak anti-AAV antibody titers rising to levels substantially exceeding those observed in natural AAV infection. In clinical hemophilia studies, post-treatment NAb titers have been reported to reach levels that would preclude effective re-administration with the same serotype.^32,37 The reported magnitude varies markedly among programs because of differences in capsid, dose, assay format, dilution range, and sampling time; therefore, no single post-administration titer is generalizable across rAAV therapies.³⁸ Induction of antibody response against the transgene-product itself is rare, potentially because most rAAV gene therapies typically encode proteins that are identical in their amino acid sequence to their endogenous counterparts and are targeted to the same cellular location where the endogenous protein resides.

In parallel with the humoral response, cellular adaptive immunity contributes to the elimination of transduced cells. Antigen-presenting cells (APCs) cross-present capsid peptides on MHC class I via the cytosolic pathway to prime or reactivate CD8⁺ cytotoxic T lymphocytes, which then eliminate transduced cells.^12,39 Both central and effector memory T cell subsets specific for conserved capsid epitopes have been observed to persist long-term in humans and non-human primates. In the initial AAV2-based clinical trial for hemophilia, presence and persistence of these responses were shown to correlate with loss of transgene expression.^12,40 Clinical studies in hemophilia have identified capsid-specific T cell responses associated with transaminase elevations and, in some patients, reduced transgene expression, with these events reported more frequently at higher vector doses in certain studies.^{10,12,14,41,42} Reactivation of capsid-specific memory CD8⁺ T cells induced by prior natural AAV exposure has been proposed as a potential mechanism, although a causal role has not been definitively established.¹² These clinical observations have been most extensively studied in hemophilia, where AAV-mediated gene transfer has been extensively evaluated in hemophilia, where AAV-mediated gene transfer has been evaluated using diverse capsids, transgene cassettes, genome configurations, vector engineering strategies, and dose levels. In an AAV8-based hemophilia B trial, capsid-specific T cell responses and elevations in liver aminotransferases were observed at the higher dose of 2 × 10¹² vg/kg; glucocorticoid treatment normalized aminotransferase levels and maintained transgene expression. In the same study, patients treated at the lower dose of 6 × 10¹¹ vg/kg developed a capsid-specific T cell response without evidence of transgene loss or elevated transaminases.¹⁴ In another hemophilia A trial employing an AAV5-based vector, elevated liver enzymes were reported despite corticosteroid therapy, while no corresponding increase in capsid-specific circulating T cells was observed.⁴³ Several clinical studies of hemophilia gene therapies support the concept that CD8+ T cell-mediated loss of transduced cells and the accompanying decline in transgene expression may be influenced by vector dose and presence of immunosuppressive therapy.^10,42,44,45 Nevertheless, in three hemophilia clinical trials, loss of transgene expression was observed despite corticosteroid therapy.^46,47,48 In two studies, the lack of durable efficacy was attributed to the higher CpG content of the vector, which appeared to limit sustained expression of the FIX Padua or wild-type FIX transgene.^46,47,49,50 In the third study, complete loss of transgene expression in two patients was associated with anti-AAV capsid immune responses that were refractory to immunosuppression.⁴⁸ Overall, these studies highlight the heterogeneity in the relationship between capsid-specific immune responses, hepatotoxicity, and transgene durability across hemophilia rAAV gene therapy trials. Taken together, the available evidence suggests that clinical outcomes are shaped by a complex interplay among vector dose, capsid and genome design, CpG content, and the effectiveness of immunosuppressive intervention, underscoring the need for more precise strategies to predict and mitigate immune-mediated loss of expression.

In addition to these effector responses, the liver can also activate regulatory mechanisms that promote tolerance to rAAV-delivered transgene products. In some hemophilia gene therapies, where the encoded transgene expressed a protein that was not identical to the endogenous protein, it is possible that the hepatic route of administration may have promoted the activation of regulatory T (Treg) cells, possibly inducing peripheral tolerance, as observed pre-clinically in wild-type mice following hepatic AAV2-mediated human FIX gene transfer.⁵¹ The liver fosters immune tolerance through a specialized microenvironment that supports Treg cell development and function while suppressing effector responses. Hepatic antigen-presenting cells including Kupffer cells, DCs, liver sinusoidal endothelial cells (LSECs), and hepatic stellate cells produce anti-inflammatory mediators such as interleukin-10 (IL-10), transforming growth factor β (TGF-β) and can express inhibitory molecules like programmed death-ligand 1 (PD-L1), which collectively limit CD4⁺ and CD8⁺ T cell activation, promote T cell apoptosis, anergy, and exhaustion, and maintain peripheral tolerance.^52,53 Within this microenvironment, Treg induction is driven by liver-derived TGF-β and all-trans retinoic acid from hepatic stellate cells, while Kupffer cells enhance Treg proliferation and IL-10 secretion, and LSECs facilitate Treg differentiation and deletion of activated T cells.^54,55 Tregs then exert suppression via IL-10, TGF-β, CTLA-4-mediated downregulation of costimulatory signals, IL-2 consumption, and direct cytotoxicity toward effector T cells and antigen-presenting cells.⁵⁶ Clinically, hepatic Tregs are critical for resolving inflammatory liver injury, preventing autoimmunity, and promoting allograft tolerance in transplantation. Leveraging this pathway through adoptive Treg therapies has shown promise in preclinical and clinical liver transplantation settings.⁵⁷ In the gene therapy context, preclinical studies suggest that liver-directed transgene expression can induce antigen-specific Tregs and suppress immune responses against the transgene product, thereby supporting durable expression. However, direct demonstration of antigen-specific regulatory T cell induction has largely been limited to murine models. Preliminary clinical findings nevertheless provide observations consistent with liver-mediated immune tolerance in humans. In one participant with hemophilia A and an active factor VIII inhibitor, liver-directed AAV5-mediated factor VIII gene transfer was followed by a decline in the inhibitor to an unmeasurable level by week 32 and an increase in factor VIII activity to 13 IU/dL, consistent with partial restoration of hemostatic function. Mechanistic immunologic assessments were not available at the time of the interim report; therefore, the contribution of regulatory T cells to this clinical outcome remains undetermined.^58,59

The innate, adaptive, and regulatory immune mechanisms described in this section influence clinical outcomes by eliciting a spectrum of clinical safety events usually observed with rAAV therapies. The principal innate and adaptive immune responses to systemic rAAV administration are summarized in Figure 1. In the next section, each of these immune mechanisms is mapped to their associated major adverse events detailing their pathophysiology, clinical presentation, diagnostic biomarkers, and current clinical management strategies.

Figure 1.

Systemic rAAV-mediated immune responses

Following systemic administration, rAAV vectors distribute to the liver, spleen, vascular compartment, and target tissues, where the capsid and vector genome may engage multiple immune pathways. Cellular innate recognition includes endosomal TLR9-MyD88 signaling in dendritic cells and plasmacytoid dendritic cells, promoting type 1 interferon production, and reported TLR2-mediated capsid sensing by macrophages or Kupffer cells, leading to inflammatory cytokine and chemokine responses. Downstream cellular responses may include monocyte/macrophage-derived inflammatory mediators, reactive oxygen species, and proteases, as well as neutrophil activation and neutrophil extracellular trap formation. In parallel, anti-capsid antibody-vector complexes may initiate the classical complement pathway, followed by alternative-pathway amplification through formation of the C3bBb convertase. Complement activation leads to opsonization, generation of anaphylatoxins C3a and C5a, and the terminal complement complex C5b-9, which can promote phagocytic uptake, inflammation, and membrane injury. Adaptive immunity is initiated through antigen presentation and CD4⁺ T cell activation. CD4⁺ T cell help supports B cell differentiation and production of anti-capsid antibodies, which may present a barrier to vector redosing, and promote antibody-dependent complement activation. Capsid-specific CD8⁺ cytotoxic T cells may recognize capsid-derived peptides presented on MHC class I by transduced cells and mediate their clearance, potentially reducing the durability of transgene expression. Product- and patient-dependent transgene-directed immune responses may additionally include transgene-specific CD4⁺ and CD8⁺ T cells and anti-transgene antibodies. The magnitude and clinical relevance of these pathways vary according to the vector capsid, genome design, dose, route of administration, target tissue, transgene, pre-existing immunity, and patient population. Created in BioRender. Gulve, N. (2026) https://BioRender.com/f1fblyq.

Clinical manifestations of rAAV immunogenicity

rAAV gene therapies have been shown to have various immune-mediated adverse events and immune responses, such as thrombocytopenia, TMA, capsid-specific neutralizing antibodies and anaphylactic reactions. Most of these immune-mediated adverse events and immune responses can be effectively managed with different immunosuppressive regimens, which may include corticosteroids, with additional agents such as sirolimus, rituximab used in selected settings; complement inhibitors have been used primarily in suspected complement-mediated TMA following systemic rAAV gene therapy.⁶⁰ The immune-related clinical observations that have been typically observed as rAAV-associated adverse events are discussed in the next section with primary emphasis on the underlying pathology, the target therapeutic indication, underlying causative immune mechanisms, diagnostic considerations, and potential mitigation strategies.

TMA

TMA is a pathological condition characterized by aggregation of platelets in the microvasculature of kidneys or other vital organs. Microangiopathic hemolytic anemia (MAHA) is a defining feature of TMA. It results from fragmentation of erythrocytes within injured microvasculature containing platelet-rich microthrombi.^61,62 Primary TMA is caused by genetic defects, while secondary TMA may be caused by pathogens, drugs or other pathological conditions that affect the complement system.^61,62 It can manifest in several forms such as thrombotic thrombocytopenic purpura (TTP) and hemolytic uremic syndrome (HUS) or atypical hemolytic uremic syndrome (aHUS).⁶¹ aHUS commonly occurs due to dysregulation of alternate complement pathway and/or due to the overactivation of complement system.⁶³

TMA is recognized as a significant potential safety concern in the gene therapy field. 12 cases with available supportive clinical data were reported in the Zolgensma Risk Management Plan.⁶⁴ rAAV gene therapy-mediated TMA is primarily associated with activation of the complement system through classical, lectin, or alternate pathways. These three pathways function through a series of effector molecules (C1-C9) to neutralize the foreign entity.⁶⁵ Clinical TMA complications were observed at high doses of rAAV (equal or above 1 × 10¹³ vg/kg) and only in i.v. administration.^61,62,66 TMA is one of the most concerning immune-mediated complications of the rAAV gene therapy and it usually manifests within the first two weeks of treatment. If it is not clinically managed, it may quickly cascade into other serious sequalae such as kidney or other organ failure.⁶⁷ TMA has also been observed in rAAV gene therapy trials for the correction of SMA,⁶⁷ Duchenne muscular dystrophy (DMD),⁶⁰ Fabry disease cardiomyopathy,⁶⁰ Danon disease,⁶⁸ and methylmalonic acidemia (MMA).⁶⁹ In these studies, most patients were successfully treated with plasmapheresis, steroids, platelet transfusion, hemodialysis, and complement inhibitors such as eculizumab.⁷⁰ However, one patient in each of these indications i.e., SMA, DMD, and MMA succumbed to complications due to TMA.⁶⁷ In a Zolgensma clinical trial for SMA, it is believed that in one of the patients who developed TMA, either classical or lectin pathway may have played a role in the activation of the complement system due to C4 consumption along with sC5b9 increase.⁶⁷ Another patient in the same trial exhibited TMA symptoms such as acute renal failure, thrombocytopenia, hyponatremia, hemolytic anemia, and elevated transaminases 12 days post-administration. Levels of sC5b9 complex were elevated on day 12 with no change in C3 and C4, or factors H and I levels.⁶⁷

rAAV gene therapy-mediated complement activation may be dependent on treatment-emergent anti-AAV antibodies and silencing these B and T cells have shown to mitigate complement activation in in rAAV9 clinical studies.⁶⁰ Clinical studies using systemic rAAV9 infusion without targeted immunomodulation have noted robust complement activation via both classical (C1q-dependent) and alternative (Ba/Bb-mediated) pathways which correlated with TMA events. In one prospective study, participants who received prophylactic B cell depletion using rituximab and T cell suppression using sirolimus did not experience laboratory evidence of complement activation or TMA. These observations support a potential contribution of anti-capsid antibodies to complement activation in that study. However, the mechanisms underlying complement activation and TMA following rAAV gene therapy are likely multifactorial, and the relative contribution of anti-capsid antibodies compared with other proposed pathways remains to be elucidated.⁶⁰

As mentioned previously through different examples, multiple factors such as route of administration, dose, and immunosuppression regimen may play a role either independently or synergistically in rAAV gene therapy-mediated TMA abnormalities. Diagnosis of TMA and rapid clinical management can potentially mitigate the gene therapy-associated TMA risk. To facilitate early detection, clinicians must establish pre-treatment baselines for all hematologic and renal parameters. TMA is clinically characterized by microangiopathic hemolytic anemia and thrombocytopenia. Accordingly, close monitoring after systemic administration of gene therapy should include complete blood count with platelet counts, lactate dehydrogenase (LDH), hemoglobin, peripheral blood smear for schistocytes, blood pressure, sC5b9 levels, and random urine protein to creatinine ratio (rUPCR).⁷¹ Given the absence of validated diagnostic criteria for rAAV-associated TMA, monitoring may include assessment of clinical and laboratory features associated with TMA, including thrombocytopenia, anemia, elevated LDH, schistocytosis, hypertension, proteinuria, and evidence of complement activation (e.g., elevated sC5b-9). These findings are informed by established TMA frameworks and should be interpreted collectively and in the context of the overall clinical presentation rather than as validated diagnostic criteria for rAAV-associated TMA.⁷¹ As thrombocytopenia may be among the earliest laboratory manifestations of evolving TMA, a decline in platelet count should prompt careful evaluation of concurrently monitored hematologic, hemolytic, renal, and complement parameters for evidence of evolving TMA. Given the high susceptibility of the microvasculature of the kidneys to TMA, concurrent and frequent monitoring of renal function (e.g., serum creatinine, blood urea nitrogen, and urinalysis) is essential, and concurrent use of nephrotoxic medications should be avoided.⁷¹ Assessment of potential innate immunogenicity markers of complement including C3, C4, C3a, C5a, sC5b9, and Bb can guide decisions on C5 inhibitor use. C3 and C4 may have limited diagnostic potential due to the poor sensitivity and specificity of the assays, whereas sC5b9 and Bb are relatively more consistently associated with complement activation in TMA.⁷² While tissue histology of a kidney biopsy remains the gold standard for definitive TMA diagnosis,⁷³ it has limited practical value in clinical trial settings due to the severe bleeding risk associated with concurrent thrombocytopenia. Treatment for TMA is a medical emergency that cannot be delayed for biopsy confirmation, necessitating reliance on peripheral blood markers. At the first clinical or laboratory suspicion of TMA, prompt multidisciplinary consultation (hematology and nephrology) is strongly recommended to guide rapid intervention. Prophylactic use of complement inhibitors has been considered in certain rAAV gene therapies with i.v. administration and where symptoms potentially related to complement activation have been observed in clinical studies.⁷⁴

Complement-mediated endothelial injury in rAAV gene therapy may occur through multiple pathways. The classical pathway, activated by immune complexes between anti-capsid antibodies and rAAV particles, has been implicated in several reported cases. Emerging evidence also suggests that non-classical mechanisms, including possible amplification through the alternative complement pathway, may contribute to endothelial cell damage independently of antibody-mediated activation (Figure 2). In some clinical contexts, high levels of transgene expression in endothelial cells may directly contribute to endothelial injury and capillary leak syndrome, a mechanism distinct from complement-mediated damage.⁷⁵ Furthermore, certain disease backgrounds may confer greater susceptibility to alternative pathway-mediated endothelial damage. In Danon disease, a fatal case potentially involving complement-mediated endothelial injury has highlighted that disease-specific vascular vulnerability may amplify complement-mediated toxicity through pathways not fully addressed by classical pathway inhibition alone.⁷⁶ Overall, these observations underscore the challenge of controlling complement-related toxicities using antibodies against individual complement components when the underlying mechanism may be heterogeneous, incompletely understood, or influenced by disease-specific endothelial fragility.

Figure 2.

Non-immune contributions to recombinant adeno-associated virus toxicity

Systemic AAV administration can generate high intracellular capsid and vector-genome burden, broad systemic biodistribution with off-target transduction, and high or ectopic transgene expression. These exposure-related factors may converge on cell-intrinsic stress responses, including proteostasis and endoplasmic reticulum stress, activation of DNA-damage and p53-associated pathways, and context-dependent oxidative or mitochondrial stress. The resulting injury may affect susceptible cell populations, including hepatocytes, cardiomyocytes, dorsal root ganglion neurons, and endothelial cells, and contribute to hepatocellular stress and injury, cardiomyocyte injury or cardiotoxicity, DRG neuronal degeneration and axonopathy, and sinusoidal or endothelial injury with platelet sequestration. The pathways are presented as a simplified conceptual framework; their relative contribution varies according to vector design, dose, route of administration, biodistribution, promoter and transgene properties, and tissue-specific susceptibility. Immune-mediated mechanisms may overlap with and amplify several of these outcomes but are not depicted here. AAV, adeno-associated virus; DRG, dorsal root ganglion; vg, vector genome. Created in BioRender. Gulve, N. (2026) https://BioRender.com/f1fblyq.

Hepatotoxicity

Hepatotoxicity is one of the most frequently reported adverse clinical effects in patients treated with rAAV gene therapy products.⁷⁷ Although it is often mild in many patients, hepatotoxicity may lead to acute liver injury and death, which has resulted in the termination of some clinical trials.^14,36,78,79

Immune-mediated hepatotoxicity has been observed in both treatments that are currently undergoing clinical trials as well as those that have received FDA approval (Table 1). Immune-mediated events observed in context of rAAV gene therapies are usually associated with cytotoxic T cell responses directed against the rAAV capsid. However, high vector doses or high transgene expression resulting in cellular stress in the hepatocytes, and pre-existing disease-related hepatic abnormalities may also contribute to liver injury through non-immune or multifactorial mechanisms.⁸⁰ The hepatic dysfunction seen in patients receiving rAAV gene therapy is not solely due to anti-AAV immune responses. Some patients have pre-existing liver disease or other liver co-morbidities that are exacerbated by the anti-AAV immune response or rAAV treatment-linked hepatic stress. Selected examples of immune-mediated hepatotoxicity are discussed in the following sections, followed by examples illustrating non-immune and multifactorial contributions.

Zolgensma, which was approved in 2019 for treatment of pediatric patients less than 2 years of age with SMA with bi-allelic mutations in the survival motor neuron 1 (SMN1) gene. Clinical data revealed some degree of elevated concentrations of serum aminotransferase liver enzymes-alanine aminotransferase (ALT) and/or aspartate aminotransferase (AST) in up to 90% of the treated patients, with a few reports of serious acute liver injury.³⁶ The elevation in liver enzymes peaked a week after administration and again after a month when the steroid taper commenced.³⁶ In two instances, patients displayed increased liver inflammation with prominent CD8-positive T cell infiltration, supporting an immune-mediated component; however, the capsid specificity of the infiltrating T cells was not directly demonstrated.⁷⁸ Both patients responded to a high dose of corticosteroids. Although SMA is primarily a neurodegenerative disease, patients often also have an impaired hepatic function.⁸¹ Over 60% of patients enrolled in clinical trials already had elevated ALT/AST and/or bilirubin prior to receiving gene therapy, and more than 40% received potentially hepatotoxic medications in parallel.³⁶ Given the potential baseline hepatic vulnerability of treated patients, additional immune-mediated liver injury following rAAV administration may further compromise clinical stability. Accordingly, corticosteroids have been routinely used both prophylactically and reactively to mitigate rAAV-associated hepatocellular inflammation. The current prescribing information also recognizes that hepatotoxicity may be immune-mediated and recommends adjustment of the corticosteroid dose or duration when liver abnormalities persist.⁸²

In case of both hemophilia A (factor VIII deficiency) and hemophilia B (factor IX deficiency), hepatocytes were the cellular targets for rAAV transgene expression. Early clinical studies showed an initial dose-dependent expression of factor IX levels; however, a rise of liver transaminases in serum was noted in patients with higher rAAV doses. High levels of serum transaminases were also associated with a drop in factor IX protein levels, consistent with immune-mediated elimination of transduced hepatocytes by capsid-specific CD8-positive T cells.⁴¹ In these patients, immunosuppressive treatment with corticosteroids (prednisolone, tapered over several weeks) resulted in stabilization of liver enzymes and rescued long-term expression of factor IX.⁴¹ These observations provided early clinical evidence for capsid-directed cellular immunity as a mechanism of rAAV-associated hepatotoxicity, although the relationship among cellular immune responses, aminotransferase elevations, and loss of transgene expression has not been consistent across all vectors and clinical studies.

Human hemophilia A gene therapy trials with systemic AAV-FVIII vectors have also reported ALT elevations in most recipients within first 26 weeks post-infusion, though late ALT elevations after 52 weeks have also been reported.^43,83,84 For valoctocogene roxaparvovec, the median duration of ALT elevations was approximately 25 days and in a minority of cases elevated levels persisted up to 64 weeks.⁵² The majority of patients in the pivotal study required corticosteroid treatment, with a median corticosteroid-treatment duration of approximately 35 weeks, and some patients required treatment for more than 1 year. These transaminase elevations have been generally managed with standardized corticosteroid protocols, although the extended duration in some patients underscores the complexity of hepatic injury mechanisms in FVIII gene therapy. In the GENEr8-1 study, AAV5 capsid-specific cellular immune responses were detected in most patients but showed only a weak association with ALT elevations and were not associated with changes in factor VIII activity, indicating that these events are not uniformly explained by capsid-directed cytotoxicity. These findings support weekly ALT/AST monitoring for at least six months and comprehensive pre-infusion hepatic assessment.⁸⁵

Although adaptive immune responses directed against the AAV capsid represent a recognized mechanism of liver injury following rAAV gene transfer, hepatotoxicity may also arise through non-immune or mixed mechanisms (Figure 2). These non-immune contributions should be distinguished from adaptive immune responses, as they may require different management approaches and have distinct implications for vector dose, transgene design, and patient selection.

In hemophilia A gene therapy, ALT/AST elevations have been observed to persist for several weeks to months in many treated patients, including their occurrence during immunosuppressive treatment, suggesting that factors beyond a self-limited capsid-directed T cell response may contribute to hepatocellular stress.^84,86,87 Factor VIII is a large and structurally complex glycoprotein whose folding and secretion place substantial demands on the protein-processing machinery of the endoplasmic reticulum. In contrast to physiological factor VIII production, which occurs predominantly in LSECs, current liver-directed AAV-factor VIII vectors drive expression of B-domain-deleted factor VIII in hepatocytes.^88,89,90 Preclinical studies in hemophilia A Exon 16 knockout mice and RAG2^−/− mice have demonstrated that high levels of factor VIII expression following AAV administration can result in intracellular factor VIII accumulation, activation of the unfolded protein response, and endoplasmic reticulum stress in hepatocytes. These findings provide biological plausibility for transgene-specific proteostatic stress as a contributor to hepatocellular injury.^89,90 However, its contribution to clinical ALT and AST elevations has not been established. Liver biopsies collected from five patients 2.6–4.1 years after valoctocogene roxaparvovec administration showed no clinically relevant chronic inflammation or evidence of endoplasmic reticulum stress in factor VIII-expressing hepatocytes. Because these biopsies were obtained several years after treatment, they do not exclude the possibility of transient cellular stress during earlier periods of higher transgene expression. Factor VIII-associated cellular stress should therefore be considered a biologically plausible, but clinically unconfirmed, contributor to hepatotoxicity rather than an established mechanism.^89,90,91

In the ASPIRO study where a rAAV8 vector encoding the MTM1 transgene under a muscle-specific promoter was delivered i.v. to treat X-linked myotubular myopathy (XLMTM), fatal hepatotoxicity was observed in four patients and one patient demonstrated severe liver injury within 3–4 weeks post-administration. Three out of four patients who died received a dose of 3.5 × 10¹⁴ vg/kg and one patient received a dose of 1.3 × 10¹⁴ vg/kg. All patients had pre-existing baseline liver abnormalities. In rAAV gene therapies for hemophilia, SMA and DMD, hepatic injury was predominantly hepatocellular which was marked by serum AST/ALT elevations with normal alkaline phosphatase (ALP) and was resolved with corticosteroid-based immunosuppression.^43,83,84 In contrast, the ASPIRO study of AT132 exhibited cholestatic liver injury, evidenced by elevated ALP and histologic cholestasis, which did not respond to standard immunosuppressive regimens.⁹² Subsequent studies showed that XLMTM itself can affect the liver and bile-drainage system, with some patients developing episodes of cholestasis even without gene therapy. In a zebrafish model of XLMTM, loss of myotubularin disrupted the normal structure and functions of bile canaliculi and impaired the transport and flow of bile. These findings indicate that patients with XLMTM may already have an underlying susceptibility to cholestatic liver injury. More recent studies in a mouse model of XLMTM support a multihit model in which MTM1 deficiency creates this baseline vulnerability, while additional metabolic, environmental, or treatment-related stressors, including systemic AAV8 exposure, may trigger or worsen liver injury.^93,94,95 The ASPIRO experience therefore illustrates how underlying disease-specific organ pathology, high systemic vector dose, and additional patient-specific stressors may interact to produce severe treatment-related toxicity independently of, or in combination with, adaptive immune activation. These dose- and patient-dependent hepatotoxic events highlight the need for thorough pre-infusion hepatic assessment and conservative vector-dosing strategies in AAV gene therapy.^79,92

Collectively, the hemophilia A and XLMTM examples illustrate that hepatotoxicity following rAAV gene therapy is multifactorial and may involve capsid-directed cellular immunity, transgene-specific cellular stress, vector dose-related effects, and underlying disease-associated organ vulnerability, either independently or in combination. Accurate characterization of the predominant mechanism in each clinical context is essential because non-immune hepatotoxicity may respond incompletely or not at all to immunosuppressive treatment and may instead require dose optimization, alternative promoters or transgene-expression cassettes, and careful patient selection based on baseline hepatic and hepatobiliary function.

Detection of rAAV-mediated hepatotoxicity relies on serial monitoring of serum ALT and AST to detect the characteristic transaminitis and may be supplemented by noninvasive imaging to exclude alternative etiologies.^43,96 Cellular immune parameters, particularly capsid-specific CD8⁺ T cell frequencies measured by IFN-γ ELISpot may provide supportive evidence for a cellular immune response. However, correlations among capsid-specific T cell responses, transaminase elevations, and loss of transgene expression have been inconsistent across clinical studies; therefore, these assays should not be interpreted as standalone evidence that liver injury is immune-mediated.⁹⁷ Elevated proinflammatory cytokines like IL-6 and TNF-α can further inform on the intensity of immune activation.⁹⁶ Management of suspected immune-mediated hepatocellular injury generally begins with corticosteroid therapy. Many patients will already be on a prophylactic immunosuppression regimen that contains corticosteroids. Based on severity of hepatotoxicity, these patients may have a delay in their tapering schedule or receive pulse (high-dose i.v.) corticosteroids. Beyond corticosteroids, there is currently no consensus regarding the optimal second-line treatment for persistent or steroid-refractory hepatotoxicity. Limited clinical experience suggests that mammalian target of rapamycin inhibition, particularly with sirolimus, may be beneficial in selected patients. In a recent case series of four patients with DMD who developed liver inflammation following delandistrogene moxeparvovec treatment, the addition of sirolimus after an incomplete or transient response to high-dose corticosteroids was associated with improvement in liver enzymes and enabled corticosteroid tapering. Many gene therapy programs now include sirolimus in their prophylactic immunosuppression.^80,98 Clinical experience with alternative immunosuppressive regimens remains limited, and treatment is not standardized across rAAV products.³⁶ Although transaminase elevations accompanied by evidence of capsid-directed cytotoxic T cell activation may respond favorably to immunosuppression; cholestatic, transgene-related, vector dose-related, or disease-associated hepatic injury may respond incompletely or not at all to immunosuppression. Management should therefore be guided by the timing, biochemical and histopathologic pattern, associated immune findings, and likely mechanism of liver injury rather than by elevations in ALT and AST alone.

Myocarditis

Myocarditis is an inflammatory condition of the myocardium, the muscular layer of the heart wall. Inflammation can lead to impaired cardiac function and in certain severe cases result in fulminant myocarditis causing cardiac dysfunction and leading to end-stage cardiac failure.⁹⁹ In rAAV gene therapy, transgene-directed cytotoxic T cell responses have been implicated as a mechanism of myocardial injury in selected clinical settings and demonstrated in nonclinical models. In DMD, such responses have been associated with recognition of transgene epitopes absent from the endogenous protein because of underlying genomic deletions.¹⁰⁰ A study in rhesus macaques for Pompe disease using an rAAV vector encoding human acid-α-glucosidase (hGAA) demonstrated CD8+ T cell infiltration of the myocardium and elevated troponin I. Myocarditis was associated with elevated ELISpot responses to transgene peptide pools and the Mamu-A002.01 MHC I haplotype in three of five animals that developed cardiac toxicity.¹⁰¹ These findings highlight the potential for transgene-driven immune toxicity in specific settings and suggest that host MHC genotype and transgene epitope design may warrant careful consideration in clinical rAAV gene therapy.¹⁰¹ Chest pain, arrhythmia, or heart failure/cardiogenic shock are primary signs of potential rAAV-mediated myocarditis, with clinical evaluation mirroring that of other etiologies.¹⁰² In the context of rAAV therapy, symptoms typically manifest 3–6 weeks post-infusion, corresponding to the period of transgene expression and peak T cell activation.¹⁰¹ Myocarditis is a rare but serious adverse event associated with rAAV gene therapies.¹⁰³ Representative examples of myocarditis documented in clinical rAAV gene therapy trials are discussed in the next section.

Another related example of AAV-associated immune-mediated myocarditis was seen in AAV9-based micro-dystrophin gene therapy trial for DMD.⁷⁶ Multiple cases of immune-mediated myocarditis, characterized by elevated troponin levels, arrhythmias, and CD8⁺ T cell infiltration on biopsy were observed. These events were generally seen arising 4–8 weeks after high-dose vector infusion (≥1 × 10¹⁴ vg/kg).⁷⁶ These events included a fatal cardiac arrest in a phase 1b study of PF-06939926 and myocarditis-related muscle weaknesses in the CIFFREO phase 3 trial.^76,104 Patients with DMD exon-deletion mutations, especially those affecting exons 8–11, experienced higher rates of myocarditis after AAV-mediated micro-dystrophin delivery, implicating transgene-specific cytotoxic T cell responses against neo-epitopes rather than anti-capsid immune response.¹⁰⁰ Similarly, one patient in a DMD clinical trial developed myocarditis, which was eventually resolved by steroid treatment. Beyond myocarditis, these patients also developed severe myositis presenting as severe muscles weakness.¹⁰⁰ In the ASPIRO trial for XLMTM, a subject developed myocarditis seven weeks after rAAV8 infusion, marked by elevated troponin I and creatine kinase (CK) levels and requiring brief hospitalization, followed by atrial tachycardia at week 21 necessitating extended electrocardiogram (ECG) monitoring. Both events resolved under standard prednisolone prophylaxis without escalation to high-dose steroids or additional immunosuppressants without permanent dysfunction.⁹² These cardiac complications highlight the need for cardiac assessments, serial troponin and CK measurements, and ongoing electrocardiographic surveillance in rAAV gene therapy protocols. Additional rAAV clinical trials reporting myocarditis with limited published details are summarized in Table 2.

Table 2.

Immunogenicity-related adverse events observed in FDA-approved and on-going rAAV gene therapies clinical trials as of March 2026

Adverse event	Voretigene neparvovec launched 2018	Onasemnogene abeparvovec launched 2019	Valoctocogene roxaparvovec launched 2022	Delandistrogene moxeparvovec launched 2023	Etranacogene dezaparvovec launched 2023	Fidanacogene elaparvovec registered 2023	Rebisufligene etisparvovec pre-registered	Dirloctocogene samoparvovec phase 3	Resamirigene bilparvovec phase 3	RP-A501 phase 2	AT845 phase 1/2	FKzhAspA-Opt phase 1/2	FLT190 phase 1/2	LB-001 phase 1/2	NGN-401 phase 1/2	RGX-501 phase 1/2	SB-913 phase 1/2	SGT-001 phase 1/2	SGT-003 phase 1/2	CRD-TMH-001 phase 1	GNT-0004 phase 1	Fordadistrogene movaparvovec discontinued	scAAV9/JeT-GAN discontinued
mmune-mediated hepatic disorder	–	–	ROCTAVIAN105	Duan et al.76	HEMGENIX106	BEQVEZTM107	–	–	–	–	–	–	–	–	–	–	–	solid bbiosciences74	–	–	–	–	–
Immune-mediated cardiac disorder (myocarditis, myopericarditis)	–	Day et al.108	–	ELEVIDYS109; Proud et al.110; Duan et al.76	–	–	–	–	Shieh et al.92	–	–	–	therapeutics111	–	–	–	–	–	–	Lek et al.112	Laugel et al.113	Pfizer114	–
Immune-mediated adverse reaction	Luxturna115	Zolgensma116	–	–	–	BEQVEZTM107	–	–	–	Rocket Pharmaceuticals, Inc.117	Astellas118	–	–	–	–	–	–	–	–	–	–	Pfizer119	–
Thrombotic microangiopathy	Benemei et al.120	Zanoteli et al.121; Desguerre et al.122	–	Duan et al.76	–	–	–	–	–	B Greenberg et al.123; Greenberg et al.68	–	–	–	Payton et al.124	–	–	–	Benemei et al.120	–	–	–	–	–
Thrombocytosis	–	–	–	FDA125	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	Bharucha-Goebel et al.126
Thrombocytopenia	–	Gowda et al.127; Servais et al.128	–	ELEVIDYS109; Duan et al.76	–	–	Flanigan et al.129	–	Shieh et al.92	Greenberg et al.68	–	Eichler et al.130	–	–	–	–	a phase 1/2, multicenter131	Solid Biosciences74; Carl et al.132	Solid Biosciences133	–	–	Muntoni et al.134	–
Anaphylactic reaction	–	–	Ozelo et al.86; Mahlangu et al.87; ROCTAVIAN105	–	HEMGENIX106; Etranacogene Dezaparvovec135	BEQVEZTM107; BEQVEZTM136	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
Complement activation	–	–	–	–	–	–	–	–	–	–	–	–	–	–	Suter et al.137	–	–	Dreghici et al.138	–	–	–	–	–
Cytokine release syndrome	–	–	–	–	–	–	–	–	–	–	–	–	–	LogicBio Therapeutics, Inc139	–	–	–	Solid Biosciences74	–	–	–	–	–
Immune system disorder	–	Favia et al.140; Ruggiero et al.141	–	–	–	–	–	High et al.48	–	–	–	–	–	–	–	Marina Cuchel et al.142	–	–	–	–	–	–	–
Anaphylactic reaction	–	–	Ozelo et al.86; Mahlangu et al.87; ROCTAVIAN105	–	HEMGENIX106; Etranacogene Dezaparvovec135	BEQVEZTM107; BEQVEZTM136	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–
Infusion-related hypersensitivity reaction	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	–	∗∗∗solid Biosciences133	–	–	–	–

Assessments considered for detection of myocarditis in clinical trial settings have included baseline and serial ECG and echocardiography to detect arrhythmias and ventricular dysfunction, and measurement of cardiac biomarkers such as troponin I/T and CK-MB (myocardial isozyme of CK).¹⁰² Elevations in cardiac biomarkers above the 99th percentile may trigger further evaluation.¹⁴³ Cardiac magnetic resonance serves as the noninvasive gold standard for tissue characterization and endomyocardial biopsy can confirm diagnosis in cases of unknown etiology.¹⁰² Asymptomatic myocarditis in the initial weeks after rAAV infusion is generally managed with frequent monitoring and immunosuppressive regimen consisting of corticosteroid administration, often in combination with sirolimus.¹⁰³ Progression of symptoms may warrant escalation to i.v. methylprednisolone followed by a tapering oral regimen. In refractory cases, other immunomodulatory approaches may be considered. Rituximab has been used in selected forms of autoimmune giant-cell myocarditis; however, their relevance to rAAV gene therapy remains uncertain and warrants cautious evaluation.¹⁴⁴ Hemodynamically unstable patients may require further specialized heart failure treatment and arrhythmia management.¹⁰³

DRG toxicity

Dorsal root ganglion (DRG) is a cluster of sensory nerve cell bodies susceptible to off-target rAAV transduction and high local transgene expression, triggering neuronal stress, innate immune activation, and mononuclear infiltration.¹⁴⁵ Lesions are characterized by neuronal degeneration, axonopathy, and inflammatory infiltrates in the absence of clinical signs.¹⁴⁵

DRG toxicity has been observed primarily in nonclinical rAAV gene therapy studies, particularly in nonhuman primates (NHPs), with clinical evidence remaining limited to isolated and infrequent suspected cases. A meta-analysis of 33 NHP studies (n = 256) demonstrated that rAAV gene therapy-mediated DRG toxicity in NHPs was characterized by minimal to moderate lesions while the animals displayed no clinical signs.¹⁴⁵ DRG pathology was characterized by mononuclear cell infiltration and neuronal degeneration. It was most frequently observed when rAAV was administered in the cerebrospinal fluid (CSF) as compared to i.v. injection.¹⁴⁵ DRG toxicity was found to be unrelated to the types of capsids, promoters, or transgenes employed.¹⁴⁵ While some studies suggest overexpression of transgene and subsequent innate immune activation as a potential mechanism of DRG toxicity,¹⁴⁶ other studies indicate route of administration-dependent DRG toxicity in animals. Route of administration and dose critically influence both efficacy and immunogenicity. In addition to governing systemic antigen exposure and immune cell engagement, the route of delivery directly dictates the level and distribution of transgene expression in target and non-target tissues.¹⁴⁵ Preclinical evidence in rhesus macaques demonstrates that incorporating DRG restricted miR-183 target sequences into the vector genome attenuates DRG toxicity without compromising therapeutic transduction, suggesting a potential capsid-independent, transgene-mediated mechanism.¹⁴⁷ Although DRG toxicity is primarily a preclinical finding, select human rAAV clinical trials that have incorporated DRG assessments are discussed in the following section.

In a meta-analysis of seven i.v. onasemnogene abeparvovec (Zolgensma) studies (n = 102 SMA patients), no clinical sensory neuropathy or signs of DRG inflammation were observed over up to 50 months follow-up.¹⁰⁸ By contrast, in an AAVrh10 gene therapy trial for amyotrophic lateral sclerosis (ALS), one of two patients developed DRG neuron depletion with CD8⁺ T cell infiltration in proximal roots around 10 weeks post-infusion under corticosteroid-only immunosuppression; this coincided with peak ALT and capsid-specific ELISpot responses at week 10 and rising anti-capsid neutralizing antibodies through week 50. The second patient, who received corticosteroids plus rituximab and sirolimus, did not develop DRG lesions.¹⁴⁸

DRG toxicity can be assessed by pre- and post-gene therapy sensory nerve conduction studies to monitor sensory nerve action potentials (SNAP). Definitive DRG toxicity diagnosis requires histopathology from DRG or nerve-root biopsy, which is rarely, performed in rAAV gene therapy clinical trial settings. Noninvasive biomarkers such as serum and CSF neurofilament light chain (NfL) rise in a dose- and time-dependent manner in preclinical models where elevations are observed by day 14 in primates and strongly correlate with DRG lesion severity.¹⁴⁹ Non-clinical studies suggest blood NfL may have greater sensitivity than other in-life measures including neurophysiologic assessments.¹⁴⁹

Although not a confirmed case of DRG toxicity, a reported intrathecal AAV case in ALS suggests that intensified prophylactic immunosuppression may reduce inflammatory neurologic complications after gene therapy administration. In that report, after one patient developed post-treatment meningoradiculitis, a subsequent patient who received corticosteroids, rituximab, and sirolimus did not develop sensory symptoms. This observation may support further evaluation of intensified combination immunosuppression as a strategy to mitigate post-AAV inflammatory neurologic complications.¹⁴⁸ Although a protective effect cannot be concluded from this limited experience, emerging mitigation strategies under investigation include vector design approaches such as embedding DRG-restricted microRNA target sequences to reduce transgene expression in DRG neurons without affecting target tissue efficacy.¹⁴⁷ In addition, dose level and route of administration, and circulating NfL may represent important variables to better characterize, monitor, and potentially reduce DRG-related toxicity.

Patient selection and risk stratification in clinical trials

In rAAV gene therapy trials, risk stratification refers to the evaluation of factors that may influence vector transduction, immunogenicity, and the likelihood of treatment-related adverse events. This process is closely linked to patient selection, as it helps identify individuals who may be at increased risk of reduced efficacy or enhanced toxicity and informs trial eligibility criteria, monitoring strategies, and in some cases prophylactic immunomodulation. The following sections discuss key humoral clinical and genomic factors that may be considered during risk stratification and trial planning.

Preexisting humoral immunity against rAAV capsids may reduce vector bioavailability, tissue transduction, and transgene expression, particularly following systemic administration. Although preexisting anti-AAV antibodies have also been considered a potential safety risk through immune-complex formation, complement activation, or increased vector uptake by Fc receptor-expressing cells, the clinical relationship between baseline antibody titer and immune-mediated toxicity remains less clearly established than its effect on transduction and efficacy.^33,37,150 Thirteen naturally occurring AAV serotypes (AAV1–AAV13) have been identified, alongside a growing number of engineered capsid variants, with baseline anti-AAV seroprevalence varying according to capsid serotype, geographic region, age, and other demographic factors.^2,151 Route of administration may also be considered for risk stratification as systemic administration exposes the vector to circulating antibodies and typically requires higher vector doses than localized administration.¹⁵²

There is currently no standardized anti-AAV antibody assay or universally applicable antibody threshold for patient eligibility. The selection of a total antibody (TAb) or neutralizing antibody (NAb) assay and the corresponding clinical decision limit is product-specific and may depend on the capsid, route of administration, vector dose, target tissue, assay performance, nonclinical neutralization data, clinical exposure-response relationships, and the overall benefit-risk profile of the intended population.

In clinical gene-therapy settings, two principal antibody assays, TAb assay and NAb assay, are used to assess patient eligibility. TAb assays use ligand-binding formats, which detect antibodies that bind the AAV capsid. The immunoglobulin isotypes detected depend on the assay format and detection reagents. These assays are generally more sensitive even to low-affinity and non-neutralizing antibodies, which can form immune complexes, alter vector biodistribution, and trigger complement or Fc receptor-mediated uptake. NAb assays are cell-based transduction-inhibition (TI) assays that measure functional inhibition of vector transduction. NAb assays are technically complex and are affected by non-antibody inhibitors including soluble receptor decoys, serum components such as complement factors, lactoferrin, and other serum proteins that can reduce cellular transduction independently of capsid-specific antibodies, potentially leading to false-positive results.¹⁵³ NAb assay performance can vary with the reporter vector, vector multiplicity of infection, target cell line, serum concentration, transduction-enhancement conditions, inhibition threshold, and titer-calculation method.^33,37,154 Hence, TI assays pose greater challenges for in vitro diagnostic (IVD) assay validation than TAb assays. Because of their broad capture, TAb assays are often preferred for initial screening in patient selection, and cutoff titers are defined based on an upper reference limit derived from a seronegative reference population.

Consequently, titers generated using different capsids, assay formats, laboratories, or clinical-development programs are not directly interchangeable. A numerically higher eligibility threshold for one product should not be interpreted as evidence that the product permits a biologically higher level of preexisting immunity than another product with a lower numerical threshold. Furthermore, some programs use quantitative reciprocal titers, whereas others use qualitative “detected/not detected” or “positive/negative” companion-diagnostic results. Representative patient-selection criteria used for selected rAAV gene therapies are summarized in Table 3. The product-specific criteria are stated in FDA prescribing information and companion-diagnostic documents. ZOLGENSMA used an anti-AAV9 threshold of ≤1:50, whereas ELEVIDYS requires a titer <1:400; ROCTAVIAN and BEQVEZ use qualitative companion-diagnostic results rather than directly comparable titers.^82,105,109 HEMGENIX is the principal exception among the examples because the pivotal study enrolled participants irrespective of AAV5 NAb status. The current label reports 21 participants with baseline titers ranging from 1:8.5 to 1:3,212.^155,156

Table 3.

Representative baseline anti-AAV antibody assays and patient-selection criteria used in rAAV gene-therapy programs

Gene therapy	Vector and route	Baseline anti-AAV assay	Patient-selection criterion
ZOLGENSMA (onasemnogene abeparvovec-xioi)	AAV9; intravenous	anti-AAV9 binding-antibody enzyme-linked immunosorbent assay	clinical trial participants were required to have baseline titers ≤1:50; safety and efficacy in patients with titers >1:50 have not been evaluated
ELEVIDYS (delandistrogene moxeparvovec-rokl)	AAVrh74; intravenous	anti-AAVrh74 TAb enzyme-linked immunosorbent assay	patients should have anti-AAVrh74 TAb titers <1:400; administration is not recommended at titers ≥1:400
ROCTAVIAN (valoctocogene roxaparvovec-rvox)	AAV5; intravenous	qualitative AAV5 TAb bridging electrochemiluminescence companion diagnostic	patients with a “not detected” result are eligible; patients with a “detected” result are not eligible
BEQVEZ (fidanacogene elaparvovec-dzkt)	AAVRh74var; intravenous	qualitative cell-based NAb/transduction-inhibition companion diagnostic	result is positive when inhibition is >50% at the 1:1 cutoff dilution and negative when inhibition is ≤ 50% at 1:1; only patients with a negative result were eligible
HEMGENIX (etranacogene dezaparvovec-drlb)	AAV5; intravenous	cell-based AAV5 NAb assay used in the clinical-development program	participants were enrolled irrespective of baseline AAV5 NAb status; twenty-one of 54 participants had detectable baseline titers ranging from 1:8.5 to 1:321

Available data do not establish that products using higher numerical anti-AAV antibody eligibility thresholds are associated with a greater incidence of adverse outcomes. Direct comparisons across products are limited by differences in assay format, capsid, vector dose, route of administration, target tissue, and patient population. In addition, most pivotal trials excluded participants exceeding the prespecified antibody criterion, resulting in limited safety data at higher baseline titers. Although the Hemophilia B (HOPE-B) study enrolled participants irrespective of baseline AAV5 NAb status, the available data do not establish a relationship between baseline antibody titer and toxicity. Overall, preexisting anti-AAV antibodies are more consistently associated with reduced vector transduction and transgene expression than with a generalized increase in adverse outcomes.

Beyond preexisting anti-AAV antibodies, patient selection must also account for organ-specific comorbidities and genomic risk factors. Preexisting liver disease should be assessed especially in liver-targeting rAAV gene therapies as the immune reaction may exacerbate liver injury.⁸⁰ Thus, baseline liver-function tests, elastography, and other imaging are essential to identify high-risk individuals.⁴³ Genomic factors may also be relevant to risk stratification, particularly when the endogenous gene defect results in complete absence of the native protein or large deletions affecting major coding regions. In such settings, expression of the therapeutic transgene may introduce epitopes that are perceived as non-self, potentially increasing the risk of transgene-directed cellular immune responses. This consideration may be particularly relevant in disorders such as DMD, where genotype-dependent differences in endogenous dystrophin exposure may influence immunogenic risk by mounting cytotoxic T cell responses against novel epitopes from the transgenic protein.¹⁰⁰ Excluding such genotypes may help mitigate transgene-directed immunogenicity.¹⁵⁷ Underlying genotype may, therefore, represent an important factor in patient selection and immunogenicity risk assessment during trial planning.

Patient age may influence both the risk and clinical consequences of immune-mediated adverse events following recombinant rAAV gene therapy, although direct evidence establishing age as an independent risk factor remains limited. Older patients, particularly those with advanced disease or comorbidities, may have reduced physiological, hepatic, or renal reserve, and may therefore be less able to tolerate clinically significant toxicities such as hepatocellular injury or TMA.^158,159 Evidence specifically isolating age as an independent determinant of immunosuppression-related morbidity in rAAV gene therapy remains limited. Nevertheless, age may influence morbidity associated with the immunosuppressive regimens used to prevent or mitigate immune responses against the vector capsid, transgene product, or transduced cells. Although direct comparative data by age remain limited, available evidence suggests that older patients may be at greater risk of treatment-related morbidity because of immunosenescence, reduced physiological reserve, greater comorbidity burden, and altered hepatic, renal, and metabolic handling of immunosuppressive agents.^160,161 A more informative assessment should integrate disease stage, comorbidities, baseline anti-AAV antibody status, frailty, organ function, baseline infection risk, concomitant medications, vector dose, route of administration, and the intensity and duration of immunosuppression.

Finally, a risk-benefit assessment that weighs the potential risks due to AAV seropositivity versus severity and high unmet medical need, should drive the inclusion/exclusion decision.³³ For life-threatening indications with no alternatives, combining humoral, cellular, genetic, and clinical data into a structured flowchart or algorithm to guide inclusion/exclusion may be considered along with appropriate prophylactic immunomodulation.³³

Risk mitigation strategies

Effective mitigation of immunogenicity in rAAV gene therapy requires an integrated strategy combining vector engineering, targeted immunosuppression regimens, adjunctive and emerging approaches, optimized dosing protocols, and monitoring innate and adaptive immune markers. The following section explores each component of this integrated strategy in detail.

During preclinical development, vector engineering can effectively enhance transgene expression and minimize off-target tissue tropism, thereby reducing the minimum efficacious dose. rAAV vectors are engineered via directed evolution and rational surface-residue modification, which yield capsids with reduced antibody binding and innate sensing. Preclinical studies have shown that incorporation of tissue-specific microRNA target sequences can reduce off-target transgene expression; examples include miR-183-family target sites for DRG detargeting in rhesus macaques and miR-122 target sites to suppress expression in hepatocytes in Institute of Cancer Research (ICR) mice when liver expression is undesirable.^147,162 Codon optimization of the transgene open reading frame and balancing host-preferred codon usage improves translation efficiency. Minimizing immunostimulatory CpG motifs enhances protein yield and reduces innate immune activation.¹⁶³ In preclinical studies using wild-type and FIX-deficient hemophilia B mice, the strategic depletion of unmethylated CpG dinucleotides from vector genomes and regulatory elements have been shown to circumvent TLR9-mediated sensing, which correlated with prolonged transgene expression and reduced cytotoxic T cell responses.¹⁶⁴ In the BAX 335 clinical trial, high CpG content in the vector cassette was proposed as a potential contributor to immune activation and loss of transgene expression, supporting the potential clinical relevance of CpG-mediated innate sensing.⁴⁶

Short-course regimens of corticosteroids remain the backbone of immunosuppression in rAAV gene therapy and have been most widely used agents across clinical studies, including approved rAAV therapies, because of their broad anti-inflammatory effects on innate and adaptive immune responses. (Table 1).^165,166 Additional immunosuppressive approaches have also been used selectively in certain settings, including calcineurin inhibitors such as tacrolimus or cyclosporine to suppress T cell activation, anti-proliferative agents such as mycophenolate mofetil to block lymphocyte expansion and mTOR inhibitors such as sirolimus to attenuate cytokine-driven lymphocyte proliferation.^167,168,169 In some contexts, B cell-depleting therapy with rituximab has been used selectively to reduce humoral immune responses.^38,168,169 As part of the immunosuppression regimen, more recent programs have incorporated sirolimus and/or rituximab to help manage immune responses.^170,171

In contrast to corticosteroids, these other agents are generally used selectively rather than routinely, and their use must be balanced against associated safety limitations, including infection risk and antigen-specific toxicities such as nephrotoxicity with calcineurin inhibitors, hematologic toxicity with lymphodepleting regimens, and the rare risk of progressive multifocal leukoencephalopathy with rituximab.^169,172

Apart from conventional approaches of targeting lymphocytes, emerging approaches such as adoptive transfer of antigen-specific Tregs, and enzymatic antibody depletion may offer additional means to preserve transgene expression while minimizing both innate and adaptive immune responses against the viral vector and the transgene product. Antigen-specific regulatory T cells (CD4⁺CD25⁺FoxP3⁺) induce durable transgene tolerance in preclinical models.^173,174,175 Other approaches include enzymatic IgG cleavage via IgG-degrading enzymes such as Imlifidase (IdeS) and IdeZ prior to vector administration to transiently eliminate neutralizing antibodies. IdeZ has shown significant efficacy in enabling rAAV transduction in passively immunized mice without the need for broad immunosuppression.¹⁷⁶ In cynomolgus macaques, a single dose of IdeS transiently reverses seropositivity, rescuing AAV transduction.¹⁷⁶ IdeZ acts by cleaving the heavy chains of all human IgG subclasses and eliminating Fc-dependent effector functions, including both ADCC and CDC.¹⁷⁶ This approach may enable expanding the eligible patient population and enhance rAAV gene therapy efficacy. Enzymatic IgG cleavage may also enable AAV vector re-dosing in patients who develop high-titer anti-AAV antibodies after initial rAAV administration. By transiently removing circulating IgG before a second dose, this approach may theoretically permit re-transduction, although this approach has not yet been demonstrated in human rAAV gene therapy clinical trials and would require careful evaluation of the kinetics of antibody regeneration and potential immunological consequences of repeated enzyme administration.¹⁷⁷

Route of administration and dose critically influence efficacy as well as immunogenicity. Localized delivery such as subretinal injections of 1.5 × 10¹¹ vg for Luxturna limits systemic antigen exposure and reduces the need for immunosuppression. In contrast, high-dose i.v. infusions (e.g., 2 × 10¹³ vg/kg for Zolgensma) necessitate more intensive monitoring and prolonged immunosuppressive courses. Patients typically undergo a short-term immunosuppressive regimen involving corticosteroids like prednisone before and after vector administration. This regimen aims to dampen immune responses against the rAAV vector and the transgene product, thereby enhancing the persistence of transgene expression and reducing the risk of immune-mediated adverse events.¹⁶⁵ Specific immunosuppression protocols may vary depending on factors such as the target tissue, route of administration, dose, transgene, and the patient’s immune status. Detailed corticosteroid regimens listed in the label for approved rAAV gene therapies are summarized in Table 1.

Immune markers including both innate and adaptive immune indicators should be monitored in the clinical protocol to enable tailoring immunomodulation strategies in real time and guide patient management. Anti-capsid and anti-transgene total antibodies using validated TAb assays are necessary to assess humoral immunogenicity. ELISpot assays have been used to monitor capsid- or transgene-specific IFN-γ secreting T cell responses in rAAV gene therapy trials; however, these data are generally interpreted in conjunction with clinical and laboratory findings rather than being used as a standalone trigger for immunosuppression escalation. Contemporary risk-mitigation strategies typically rely on corticosteroid-based prophylaxis or treatment, with additional agents such as sirolimus or calcineurin inhibitors used specifically selectively in specific clinical contexts.^97,168,169 Multiplex cytokine panels including IFN-γ, IL-6, TNF-α, IL-10, and MCP-1 identify early innate activation within 24–48 h post-infusion and inform preemptive steroid administration. In cynomolgus macaques, NfL levels in serum or CSF rise by day 14 and correlate with DRG injury, guiding decisions on intensified immunosuppression.¹⁷⁸ Assessment of complement split products such as C3a, C5a, Ba, Bb, and sC5b-9 may provide supportive information regarding complement pathway involvement. However, treatment decisions often are driven primarily by clinical presentation and routine laboratory abnormalities due to long turnaround times of complement assays with results used to further inform management when available. These readouts should be incorporated into adaptive trial designs where predefined thresholds could trigger additional immunomodulation. Such monitoring strategies may enable striking a balance between patient safety and durable transgene expression.

The risk mitigation strategies presented are primarily based on current knowledge, published experience, and established practices within the gene therapy field, with limited discussion of additional or novel approaches that may warrant consideration. Despite substantial advances, the field continues to evolve, and several important areas remain the subject of active investigation. First, the use of immunosuppression in rAAV gene therapy remains an evolving area, with strategies varying across products and indications according to differences in vector design, target tissues, dosing paradigms, and patient populations. While clinical evidence and experience from approved and investigational gene therapy products have informed current practice, further research will be important to refine risk-based approaches to immune management. Second, vector engineering strategies such as CpG depletion and codon optimization have shown promise in reducing innate immune activation and enhancing transgene expression. However, uncertainties remain regarding the consistency and generalizability of these findings across different vectors, indications, dosing regimens, and patient populations, and their contribution to long-term clinical outcomes remains under investigation. Third, emerging approaches, including IdeS-mediated antibody reduction and regulatory T cell-based therapies, are being explored to address humoral and cellular immune responses. Although early findings are encouraging, their clinical utility in rAAV gene therapy remains to be established. Finally, the complex interplay among innate and adaptive immune responses, vector-related factors, and non-immune toxicities suggests that the effectiveness of mitigation strategies may vary across products, indications, and patient populations, requiring approaches tailored to product-specific characteristics and individual risk factors.

Taken together, continued refinement of these strategies guided by mechanistic insights and assessment of associated immunogenicity parameters will be essential to minimize immunogenicity risk while maximizing therapeutic benefit of rAAV based gene therapies.

Future directions and perspectives

rAAV vectors exhibit efficient transduction of dividing and nondividing cells, serotype-specific tropism, stable transgene expression, minimal inflammatory burden, low insertional mutagenesis risk, and scalable manufacturing.³ While these attributes underpin their clinical promise, ongoing advances in elucidating immune activation pathways, identifying patient- and treatment-specific risk factors, and optimizing mitigation strategies are actively converting immunogenicity from a critical challenge into an opportunity for further improving the safety and efficacy of rAAV-based gene therapies.

Preexisting anti-AAV antibodies are common due to natural AAV infections and must be clinically assessed before patient enrollment in clinical trials. TAb assays offer reproducible quantification of all binding including non-neutralizing antibodies, while cell-based NAb assays measure only functional inhibition by neutralizing antibodies.³³ TAb assays are plate-based immunoassays that offer high sensitivity, reproducibility, and the ability to capture complement-activating antibodies.¹⁵³ NAb assays use cell-based transduction inhibition readouts to quantify functional blockade of vector entry but are subject to greater inter-laboratory variability, lower sensitivity for low-titer responses, and longer development timelines.¹⁵³ Studies have shown significant correlation in terms of sensitivity between NAb and TAb titers.¹⁷⁹ Harmonized guidelines on assay selection, validation, and cutoff criteria are urgently needed to ensure consistent patient eligibility decisions.^33,153 This decision will have an enduring impact on sponsors as well as thousands of patients as it will determine their eligibility to receive these therapies. Beyond preexisting antibodies, patient-specific risks may also be characterized based on presence of genomic deletions in corresponding endogenous genes, which may lead to cellular immune response against the neo-epitopes introduced by transgenic proteins. Preexisting liver health should also be considered, and baseline characterization of transaminases and other hepatic markers must be performed to avoid immune-related exacerbation of liver injury.

With preexisting patient-related risk factors addressed, vector design becomes a critical determinant of rAAV immune-mediated safety. Vector engineering during early pre-clinical development may help minimize TLR-mediated innate immune system activation. Despite steroid intervention in a clinical study, rapid loss of transgene activity was observed in clinical studies with the overstimulation of innate immune responses.⁴⁶ It is uncertain whether the corticosteroid regimen used was inadequate to counteract an rAAV vector containing significant CpG content triggering innate and concurrent adaptive immune activation cascades.⁴⁶ Designing vectors to eliminate or minimize these innate immune danger signals is expected to significantly decrease immune responses to rAAV, thereby facilitating more effective steroid interventions. CpG motifs can be found throughout the rAAV vector genome and activate TLR9 due to their hypomethylated status. Depleting CpG motifs from coding regions leads to poor transgene expression, and in isolation this may not be sufficient to dampen the immune response if detrimental CpG motif remains elsewhere in the payload. Complete CpG depletion is achievable but often comes at a high cost, resulting in lower rAAV titers and potentially reduced promoter activity. Rational approaches may include blocking TLR9 activation strategies by administration of TLR9 antagonist or by adding TLR9-blocking sequences to rAAV payloads. Prediction tools have been developed to determine the relative TLR9 activation risk by rAAV payload sequences to inform on potential immunogenicity.¹⁸⁰

Manufacturing process optimizations such as minimizing capsid impurities, reducing empty-capsid ratios, and eliminating residual host-cell DNA may provide an additional avenue for mitigating rAAV immunogenicity at the manufacturing stage. Rational rAAV plasmid design can minimize packaging of highly immunogenic bacterial backbone DNA sequences outside of the ITR-flanked vector cassette, fragmented genomes, and other potentially immunogenic impurities. Maintaining a high full capsid to empty capsid ratio during manufacturing is critical, as studies in wild-type and TAP^−/− mice have shown that empty capsids and denatured rAAV proteins can be efficiently presented via MHC-I and trigger a T cell response.¹⁸¹ These empty/partial capsid products are relevant as they can contribute to the immune responses observed in patients. Therefore, while quantifying the vector dose genomes per kilogram (vg/kg) is straightforward, this metric should also account for packaged impurities, partially full, and empty capsids, which can constitute a significant fraction of the administered drug product.

As the first line of defense, innate immune system remains a key determinant of rAAV safety. High-dose i.v. rAAV administration can trigger complement-mediated TMA, which presents with microangiopathic hemolysis, thrombocytopenia, and organ injury. Integrating early and frequent measurement of complement split products such as C3a, C5a, Bb, and soluble C5b9 in the clinical protocol may help assess complement activation-associated patient risks and enable timely intervention. Use of complement inhibitors has a mechanistic rationale for mitigating rAAV-associated TMA. However, their use in systemic rAAV gene therapy remains insufficiently defined and has been potentially associated with safety concerns, warranting caution and further clinical evaluation.¹⁸²

Baseline patient characteristics may also potentially influence immune response to rAAV-based gene therapies. Subsequent to innate activation, the adaptive immune response poses a critical barrier to durable transgene expression. Preexisting anti-AAV antibodies can neutralize vector particles and predict transduction failure. Upon i.v. administration of rAAV, treatment-emergent IgM titers normally rise by day 5 and IgG by day 7, with higher peak titers correlating inversely with transgene persistence. Antigen-presenting cells process and present capsid and transgene peptides on MHC II to CD4⁺ helper T cells, which in turn support B cell maturation, and on MHC I to prime CD8⁺ cytotoxic T lymphocytes capable of lysing transduced cells. IFN-γ ELISpot assays may help refine monitoring of capsid- or transgene-specific cellular immune responses in rAAV gene therapy, although their role in guiding immunosuppressive management remains to be better defined and will likely require interpretation alongside clinical and other laboratory data. Cytotoxic T cell activation and associated cytokine profile is implicated in rAAV induced hepatotoxicity, myocarditis, and DRG toxicity.⁶ In rAAV-mediated hepatotoxicity, serum transaminase elevations correlate with capsid-specific T cell activation, although excessive transgene expression and pre-existing hepatic inflammation may exacerbate injury. Hepatoxicity has been successfully mitigated in rAAV clinical studies by suppressing T cell responses through immunosuppressive strategies including corticosteroids. Myocarditis occurs across different AAV serotypes and target organs, suggesting recognition of transgene-derived epitopes that may be recognized as non-self in certain patients. Myocarditis is not exclusively associated with gene therapies targeting cardiac or skeletal muscles either; it has also been observed in clinical trials of liver-directed gene therapies (Table 2). Reported cases used a truncated version of the transgene significantly different from its endogenous counterpart. The temporal association of myocarditis onset with the activation of transgene-specific CD8+ T cells indicates that myocarditis may occur due to T cell mediated cytotoxicity.¹⁰⁰ Taken together, these observations suggest that future risk assessment may need to consider baseline hepatic inflammatory conditions as well as truncations or mutations in the counterpart endogenous genes.

Tailored immunosuppression regimens are critical for modulating host innate and adaptive immune responses and preserving durable transgene expression in rAAV gene therapies. Standard immunosuppressive regimens include corticosteroids as the primary approach, with sirolimus, rituximab, calcineurin inhibitors, or mycophenolate mofetil used selectively in clinical context; collectively, these approaches attenuate immune responses but show variable efficacy and remain incompletely characterized mechanistically.⁷⁷ Further evaluation of immunosuppressive regimens, including dosing and timing, is required to better understand and optimize immunosuppression in the context of rAAV gene therapy. Early initiation of corticosteroids following liver enzyme elevation has been associated with improved preservation of transgene expression in some hemophilia studies.^43,44 However, relationships among liver enzyme elevations, capsid-specific T cell responses, corticosteroid responsiveness, and transgene expression have varied across clinical studies, and the underlying mechanisms of hepatotoxicity remain incompletely understood.¹⁸³

In summary, the immunogenicity of rAAV gene therapy requires a coordinated strategy comprising of standardized immunogenicity assays, tailored immunosuppression regimen, optimized vector and payload design, and rational clinical monitoring. Continued mechanistic studies and adaptive clinical protocols will enhance the safety and durability of rAAV gene therapies. Data from ongoing clinical studies will help to highlight the intricacies of rAAV immunogenicity and to refine the approaches necessary for the successful clinical application of future rAAV gene therapies.

Acknowledgments

This work was funded by Alexion, AstraZeneca Rare Disease. The authors thank Dr. Akhilesh Chamediya for his valuable input in reviewing the manuscript and suggesting edits.

Author contributions

N.G. and M.R. contributed to conceptualization; N.G. and M.B. contributed to writing the original draft; N.G., M.B., H.G., Z.K., O.T., S.K., and M.R. contributed to writing, review, and editing.

Declaration of interests

All authors are employees of Alexion, AstraZeneca Rare Disease.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used AI-assisted tools, including ChatGPT in order to assist with identifying potentially relevant literature sources and reviewing literature summaries. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.