Global seroprevalence of neutralizing antibodies against adeno-associated virus serotypes used for human gene therapies.

Adeno-associated virus (AAV) vectors are promising gene therapy candidates, but pre-existing anti-AAV neutralizing antibodies (NAbs) pose a significant challenge to successful gene delivery. Knowledge of NAb seroprevalence remains limited and inconsistent. We measured activity of NAbs against six clinically relevant AAV serotypes across 10 countries in adults (n = 502) and children (n = 50) using a highly sensitive transduction inhibition assay. NAb prevalence was generally highest for AAV1 and lowest for AAV5. There was considerable variability across countries and geographical regions. NAb prevalence increased with age and was higher in females, participants of Asian ethnicity, and participants in cancer trials. Co-prevalence was most frequently observed between AAV1 and AAV6 and less frequently between AAV5 and other AAVs. Machine learning analyses revealed a unique clustering of AAVs that differed from previous phylogenetic classifications. These results offer insights into the biological relationships between the immunogenicity of AAVs in humans beyond that observed previously using standard clades, which are based on linear capsid sequences. Our findings may inform improved vector design and facilitate the development of AAV vector-mediated clinical gene therapies.

Complement Activation by Adeno-Associated Virus-Neutralizing Antibody Complexes.

Treatment of monogenetic disorders using vectors based on adeno-associated virus (AAV) is an area of intense interest. AAV is non-pathogenic human virus, and preexisting capsid antibodies are prevalent in the population posing a challenge to the safety and efficacy of AAV-mediated gene therapies. In this study, we investigated the risk of AAV-mediated complement activation when sera from a cohort of human donors were exposed to AAV9 capsid. Seropositive donor sera carrying neutralizing antibodies from a previous environmental exposure activated complement when admixed with AAV9 capsids and complement activation was associated with donors who had higher levels of anti-AAV IgG1 antibodies. These findings were consistent with mass spectrometry analysis that identified increased binding of immunoglobulins and complement factors when AAV9 capsids were admixed with seropositive sera. Finally, complement activation was abrogated after IgG-depletion using affinity columns or serum pretreatment with an IgG degrading enzyme. Overall, these results demonstrate an important role of preexisting neutralizing antibodies in activating complement; a risk that can be mitigated by using adequate immunosuppression strategies when dosing seropositive patients with vector.

Binding and neutralizing anti-AAV antibodies: Detection and implications for rAAV-mediated gene therapy.

Assessment of anti-adeno-associated virus (AAV) antibodies in patients prior to systemic gene therapy administration is an important consideration regarding efficacy and safety of the therapy. Approximately 30%-60% of individuals have pre-existing anti-AAV antibodies. Seroprevalence is impacted by multiple factors, including geography, age, capsid serotype, and assay type. Anti-AAV antibody assays typically measure (1) transduction inhibition by detecting the neutralizing capacity of antibodies and non-antibody neutralizing factors, or (2) total anti-capsid binding antibodies, regardless of neutralizing activity. Presently, there is a paucity of head-to-head data and standardized approaches associating assay results with clinical outcomes. In addition, establishing clinically relevant screening titer cutoffs is complex. Thus, meaningful comparisons across assays are nearly impossible. Although complex, establishing screening assays in routine clinical practice to identify patients with antibody levels that may impact favorable treatment outcomes is achievable for both transduction inhibition and total antibody assays. Formal regulatory approval of such assays as companion diagnostic tests will confirm their suitability for specific recombinant AAV gene therapies. This review covers current approaches to measure anti-AAV antibodies in patient plasma or serum, their potential impact on therapeutic safety and efficacy, and investigative strategies to mitigate the effects of pre-existing anti-AAV antibodies in patients.

In vivo overexpression of frataxin causes toxicity mediated by iron-sulfur cluster deficiency.

Friedreich's ataxia is a rare disorder resulting from deficiency of frataxin, a mitochondrial protein implicated in the synthesis of iron-sulfur clusters. Preclinical studies in mice have shown that gene therapy is a promising approach to treat individuals with Friedreich's ataxia. However, a recent report provided evidence that AAVrh10-mediated overexpression of frataxin could lead to cardiotoxicity associated with mitochondrial dysfunction. While evaluating an AAV9-based frataxin gene therapy using a chicken ß-actin promoter, we showed that toxic overexpression of frataxin could be reached in mouse liver and heart with doses between 1 × 1013 and 1 × 1014 vg/kg. In a mouse model of cardiac disease, these doses only corrected cardiac dysfunction partially and transiently and led to adverse findings associated with iron-sulfur cluster deficiency in liver. We demonstrated that toxicity required frataxin's primary function by using a frataxin construct bearing the N146K mutation, which impairs binding to the iron-sulfur cluster core complex. At the lowest tested dose, we observed moderate liver toxicity that was accompanied by progressive loss of transgene expression and liver regeneration. Together, our data provide insights into the toxicity of frataxin overexpression that should be considered in the development of a gene therapy approach for Friedreich's ataxia.

Higher Seroprevalence of Anti-Adeno-Associated Viral Vector Neutralizing Antibodies Among Racial Minorities in the United States.

Anti-adeno-associated viral vector (AAV) neutralizing antibodies (NAbs) can ablate efficacy of transgene expression following intravenous vector administration. This observation in both preclinical and clinical trials has led to exclusion of NAb-positive patients from receiving AAV gene therapy. AAV drug development includes selection of capsids with lower NAb seroprevalence that also possess other favorable traits. Often a limited number of healthy volunteers are screened to gauge NAb seroprevalence. However, limited data sets can be biased leading to inaccurate estimates of NAb incidence. In this study, we evaluated AAV NAbs against a panel of vectors among healthy donors within the United States. While the overall seroprevalence against most AAVs was lower, we did observe increased NAb incidence among black and Hispanic donors. These findings of increased NAb seroprevalence among the minority races were confirmed in a second set of donors who also demonstrated higher seroprevalence among these races. Interracial- and intraracial differences within genders were also observed among donors. The increased incidence of AAV NAb among racial minorities was unexpected. Our findings underscore the need for removing bias in sample data sets and evaluating seroprevalence within the patient population while selecting capsids.

Developing a second-generation clinical candidate AAV vector for gene therapy of familial hypercholesterolemia.

Gene therapy for hypercholesterolemia offers the potential to sustainably ameliorate disease for life with a single dose. In this study, we demonstrate the combinatorial effects of codon and vector optimization, which significantly improve the efficacy of an adeno-associated virus (AAV) vector in the low-density lipoprotein receptor (LDLR)-deficient mouse model (Ldlr -/-, Apobec1 -/- double knockout [DKO]). This study investigated vector efficacy following the combination of intervening sequence 2 (IVS2) of the human beta-globin gene and codon optimization with the previously developed gain-of-function, human LDLR triple-mutant variant (hLDLR-L318D/K809R/C818A) in the treatment of homozygous familial hypercholesterolemia (HoFH). Vector doses as low as 3 × 1011 genome copies (GC)/kg achieved a robust reduction of serum low-density lipoprotein cholesterol (LDL-C) by 98% in male LDLR-deficient mice. Less efficient LDL-C reduction was observed in female mice, which was attributable to lower gene transfer efficiency in liver. We also observed persistent and stable transgene expression for 120 days, with LDL-C levels being undetectable in male DKO mice treated with the second-generation vector. In conclusion, codon and vector optimization enhanced transgene expression and reduced serum LDL-C levels effectively at a lower dose in LDLR-deficient mice. The second-generation clinical candidate vector we have developed has the potential to achieve therapeutic effects in HoFH patients.

Adenovirus-Antibody Complexes Contributed to Lethal Systemic Inflammation in a Gene Therapy Trial.

Intra-arterial administration of an adenovirus serotype 5 (Ad5) vector in a gene therapy trial caused lethal, systemic inflammation in subject 019 with ornithine transcarbamylase deficiency. This unanticipated inflammatory response was absent in another subject receiving the same vector dose and in 16 subjects receiving lower vector doses. We hypothesized that an immune memory to a previous natural adenovirus infection enhanced the immune response to high-dose systemic Ad5 vector, causing the exaggerated immune response in subject 019. To investigate this, we found that rabbit polyclonal sera to Ad5 and pooled human immunoglobulin (Ig) inhibited Ad5 vector transduction of non-immune cells in vitro, but enhanced transduction and activation of human dendritic cells (DCs). Sera from approximately 7% of normal human subjects and 50% of patients treated topically with Ad5 vectors enhanced Ad5 transduction and activation of DCs, apparently from formation of Ig-Ad5 immune complexes and binding to DCs through FcγR. Subject 019's blood substantially increased Ad5-vector activation of human DC primary cultures at levels exceeding those from normal subjects. Although this study is based on one event in a single subject, the results implicate a pre-existing humoral immune response to Ad5 in the lethal systemic inflammatory response that occurred in subject 019.

TLR9 signaling mediates adaptive immunity following systemic AAV gene therapy.

An ongoing concern of in vivo gene therapy is adaptive immune responses against the protein product of a transgene, particularly for recessive diseases in which antigens are not presented to lymphocytes during central tolerance induction. Here we show that Toll-like receptor 9 (TLR9) signaling activates T cells against an epitope tagged mitochondria-targeted ornithine transcarbamylase (OTC) following the administration of a systemic adeno-associated virus (AAV) vector. Using a transgenic mouse model system, we demonstrate that TLR9 signaling extrinsic to T cells induces a robust cytotoxic T-cell response against the transgene and results in transgene expression loss. Overall, our results suggest that inflammation mediated by TLR9 signaling and the presence of high affinity transgene-specific T cells is important for the development of adaptive immune responses to transgene products following AAV gene therapy.

Mapping an Adeno-associated Virus 9-Specific Neutralizing Epitope To Develop Next-Generation Gene Delivery Vectors.

Recent clinical trials have demonstrated the potential of adeno-associated virus (AAV)-based vectors for treating rare diseases. However, significant barriers remain for the translation of these vectors into widely available therapies. In particular, exposure to the AAV capsid can generate an immune response of neutralizing antibodies. One approach to overcome this response is to map the AAV-specific neutralizing epitopes and rationally design an AAV capsid able to evade neutralization. To accomplish this, we isolated a monoclonal antibody against AAV9 following immunization of BALB/c mice and hybridoma screening. This antibody, PAV9.1, is specific for intact AAV9 capsids and has a high neutralizing titer of >1:160,000. We used cryo-electron microscopy to reconstruct PAV9.1 in complex with AAV9. We then mapped its epitope to the 3-fold axis of symmetry on the capsid, specifically to residues 496-NNN-498 and 588-QAQAQT-592. Capsid mutagenesis demonstrated that even a single amino acid substitution within this epitope markedly reduced binding and neutralization by PAV9.1. In addition, in vivo studies showed that mutations in the PAV9.1 epitope conferred a "liver-detargeting" phenotype to the mutant vectors, unlike AAV9, indicating that the residues involved in PAV9.1 interactions are also responsible for AAV9 tropism. However, we observed minimal changes in binding and neutralizing titer when we tested these mutant vectors for evasion of polyclonal sera from mice, macaques, or humans previously exposed to AAV. Taken together, these studies demonstrate the complexity of incorporating mapped neutralizing epitopes and previously identified functional motifs into the design of novel capsids able to evade immune response.IMPORTANCE Gene therapy utilizing viral vectors has experienced recent success, culminating in U.S. Food and Drug Administration approval of the first adeno-associated virus vector gene therapy product in the United States: Luxturna for inherited retinal dystrophy. However, application of this approach to other tissues faces significant barriers. One challenge is the immune response to viral infection or vector administration, precluding patients from receiving an initial or readministered dose of vector, respectively. Here, we mapped the epitope of a novel neutralizing antibody generated in response to this viral vector to design a next-generation capsid to evade immune responses. Epitope-based mutations in the capsid interfered with the binding and neutralizing ability of the antibody but not when tested against polyclonal samples from various sources. Our results suggest that targeted mutation of a greater breadth of neutralizing epitopes will be required to evade the repertoire of neutralizing antibodies responsible for blocking viral vector transduction.

Alternative Start Sites Downstream of Non-Sense Mutations Drive Antigen Presentation and Tolerance Induction to C-Terminal Epitopes.

CTL responses to the transgene product remain an active area of concern for the gene therapy field. A patient's underlying genetic mutation may influence the qualitative nature of these potentially destructive T cell responses. Individuals with a mutation that introduces a premature termination codon (PTC) that prevents synthesis of the full-length peptide are considered more likely to mount a transgene-specific T cell response because of a lack of immune tolerance to C-terminal epitopes as a consequence of absent endogenous Ag presentation. In this article, we demonstrate that a human ornithine transcarbamylase gene containing various PTC-inducing non-sense mutations is able to generate and present epitopes downstream of the termination codon. Generation of these epitopes occurs primarily from alternative translation start sites downstream of the stop codon. Furthermore, we show that expression of these genes from adeno-associated virus vectors in C57BL/6 mice is able to induce peripheral tolerance to epitopes downstream of the PTC. These results suggest that, despite the lack of full-length endogenous protein, patients with PTC-inducing non-sense mutations may still present T cell epitopes downstream of the premature termination site that may render the subject tolerant to wild-type transgene products.

Class I-restricted T-cell responses to a polymorphic peptide in a gene therapy clinical trial for α-1-antitrypsin deficiency.

Adeno-associated virus (AAV)-mediated gene therapy is currently being pursued as a treatment for the monogenic disorder α-1-antitrypsin (AAT) deficiency. Results from phase I and II studies have shown relatively stable and dose-dependent increases in transgene-derived wild-type AAT after local intramuscular vector administration. In this report we describe the appearance of transgene-specific T-cell responses in two subjects that were part of the phase II trial. The patient with the more robust T-cell response, which was associated with a reduction in transgene expression, was characterized more thoroughly in this study. We learned that the AAT-specific T cells in this patient were cytolytic in phenotype, mapped to a peptide in the endogenous mutant AAT protein that contained a common polymorphism not incorporated into the transgene, and were restricted by a rare HLA class I C alleles present only in this patient. These human studies illustrate the genetic influence of the endogenous gene and HLA haplotype on the outcome of gene therapy.

Stable liver-specific expression of human IDOL in humanized mice raises plasma cholesterol.

AIMS: IDOL (inducible degrader of the low-density lipoprotein receptor, LDLR) is an E3 ubiquitin ligase that promotes the ubiquitination and degradation of the LDLR. IDOL is a potential therapeutic target for the development of a novel class of low-density lipoprotein cholesterol (LDL-C)-lowering therapies. In an attempt to develop a mouse model for testing IDOL inhibitors, we examined the effects of adeno-associated virus (AAV)-mediated stable expression of human IDOL in the livers of mice 'humanized' with regard to lipoprotein metabolism. METHODS AND RESULTS: Using a liver-specific AAV serotype 8 (AAV8)-mediated delivery, AAV-hIDOL produced a dose-dependent increase in LDL-C levels and a decrease in liver LDLR protein. Furthermore, we expressed hIDOL in a 'humanized' mouse model of heterozygous familial hypercholesterolaemia (LDLR(+/-)/Apobec1(-/-)/hApoB-Tg, LAhB). In this model, total cholesterol (TC) and LDL-C levels were increased by ∼60% starting from 1 week and were sustainable for at least 3 weeks post-injection. Finally, we demonstrate that the effects caused by hIDOL expression are LDLR- dependent given the unchanged plasma lipids in LAhB mice lacking LDLR. CONCLUSION: In conclusion, our study demonstrates a dose-dependent physiological effect of human IDOL on LDL metabolism in mice. This provides a potential model for preclinical testing of IDOL inhibitors for reduction of LDL-C levels.

Comparative Study of Liver Gene Transfer With AAV Vectors Based on Natural and Engineered AAV Capsids.

Vectors based on the clade E family member adeno-associated virus (AAV) serotype 8 have shown promise in patients with hemophilia B and have emerged as best in class for human liver gene therapies. We conducted a thorough evaluation of liver-directed gene therapy using vectors based on several natural and engineered capsids including the clade E AAVrh10 and the largely uncharacterized and phylogenically distinct AAV3B. Included in this study was a putatively superior hepatotropic capsid, AAVLK03, which is very similar to AAV3B. Vectors based on these capsids were benchmarked against AAV8 and AAV2 in a number of in vitro and in vivo model systems including C57BL/6 mice, immune-deficient mice that are partially repopulated with human hepatocytes, and nonhuman primates. Our studies in nonhuman primates and human hepatocytes demonstrated high level transduction of the clade E-derived vectors and equally high transduction with vectors based on AAV3B. In contrast to previous reports, AAVLK03 vectors are not superior to either AAV3B or AAV8. Vectors based on AAV3B should be considered for liver-directed gene therapy when administered following, or before, treatment with the serologically distinct clade E vectors.

In vivo evaluation of adeno-associated virus gene transfer in airways of mice with acute or chronic respiratory infection.

Patients with cystic fibrosis (CF) often suffer chronic lung infection with concomitant inflammation, a setting that may reduce the efficacy of gene transfer. While gene therapy development for CF often involves viral-based vectors, little is known about gene transfer in the context of an infected airway. In this study, three mouse models were established to evaluate adeno-associated virus (AAV) gene transfer in such an environment. Bordetella bronchiseptica RB50 was used in a chronic, nonlethal respiratory infection in C57BL/6 mice. An inoculum of ∼10(5) CFU allowed B. bronchiseptica RB50 to persist in the upper and lower respiratory tracts for at least 21 days. In this infection model, administration of an AAV vector on day 2 resulted in 2.8-fold reduction of reporter gene expression compared with that observed in uninfected controls. Postponement of AAV administration to day 14 resulted in an even greater (eightfold) reduction of reporter gene expression, when compared with uninfected controls. In another infection model, Pseudomonas aeruginosa PAO1 was used to infect surfactant protein D (SP-D) or surfactant protein A (SP-A) knockout (KO) mice. With an inoculum of ∼10(5) CFU, infection persisted for 2 days in the nasal cavity of either mouse model. Reporter gene expression was approximately ∼2.5-fold lower compared with uninfected mice. In the SP-D KO model, postponement of AAV administration to day 9 postinfection resulted in only a two fold reduction in reporter gene expression, when compared with expression seen in uninfected controls. These results confirm that respiratory infections, both ongoing and recently resolved, decrease the efficacy of AAV-mediated gene transfer.

AAV vectors expressing LDLR gain-of-function variants demonstrate increased efficacy in mouse models of familial hypercholesterolemia.

RATIONALE: Familial hypercholesterolemia is a genetic disorder that arises because of loss-of-function mutations in the low-density lipoprotein receptor (LDLR) and homozygous familial hypercholesterolemia is a candidate for gene therapy using adeno-associated viral vectors. Proprotein convertase subtilisin/kexin type 9 (PCSK9) and inducible degrader of LDLR (IDOL) negatively regulate LDLR protein and could dampen adeno-associated viral vector encoded LDLR expression. OBJECTIVE: We sought to create vectors expressing gain-of-function human LDLR variants that are resistant to degradation by human PCSK9 (hPCSK9) and IDOL and thereby enhance hepatic LDLR protein abundance and plasma LDL cholesterol reduction. METHODS AND RESULTS: Amino acid substitutions were introduced into the coding sequence of human LDLR cDNA to reduce interaction with hPCSK9 and human IDOL. A panel of mutant human LDLRs was initially screened in vitro for escape from PCSK9. The variant human LDLR-L318D was further evaluated using a mouse model of homozygous familial hypercholesterolemia lacking endogenous LDLR and apolipoprotein B mRNA editing enzyme catalytic, APOBEC-1 (double knockout). Administration of wild-type human LDLR to double knockout mice, expressing hPCSK9, led to diminished LDLR activity. However, LDLR-L318D was resistant to hPCSK9-mediated degradation and effectively reduced cholesterol levels. Similarly, the LDLR-K809R\C818A construct avoided human IDOL regulation and achieved stable reductions in serum cholesterol. An adeno-associated viral vector serotype 8.LDLR-L318D\K809R\C818A vector that carried all 3 amino acid substitutions conferred partial resistance to both hPCSK9- and human IDOL-mediated degradation. CONCLUSIONS: Amino acid substitutions in the human LDLR confer partial resistance to PCSK9 and IDOL regulatory pathways with improved reduction in cholesterol levels and improve on a potential gene therapeutic approach to treat homozygous familial hypercholesterolemia subjects.

Human Treg responses allow sustained recombinant adeno-associated virus-mediated transgene expression.

Recombinant adeno-associated virus (rAAV) vectors have shown promise for the treatment of several diseases; however, immune-mediated elimination of transduced cells has been suggested to limit and account for a loss of efficacy. To determine whether rAAV vector expression can persist long term, we administered rAAV vectors expressing normal, M-type α-1 antitrypsin (M-AAT) to AAT-deficient subjects at various doses by multiple i.m. injections. M-specific AAT expression was observed in all subjects in a dose-dependent manner and was sustained for more than 1 year in the absence of immune suppression. Muscle biopsies at 1 year had sustained AAT expression and a reduction of inflammatory cells compared with 3 month biopsies. Deep sequencing of the TCR Vß region from muscle biopsies demonstrated a limited number of T cell clones that emerged at 3 months after vector administration and persisted for 1 year. In situ immunophenotyping revealed a substantial Treg population in muscle biopsy samples containing AAT-expressing myofibers. Approximately 10% of all T cells in muscle were natural Tregs, which were activated in response to AAV capsid. These results suggest that i.m. delivery of rAAV type 1-AAT (rAAV1-AAT) induces a T regulatory response that allows ongoing transgene expression and indicates that immunomodulatory treatments may not be necessary for rAAV-mediated gene therapy.

Biodistribution of AAV8 vectors expressing human low-density lipoprotein receptor in a mouse model of homozygous familial hypercholesterolemia.

Recombinant adeno-associated viral vectors based on serotype 8 (AAV8) transduce liver with superior tropism following intravenous (IV) administration. Previous studies conducted by our lab demonstrated that AAV8-mediated transfer of the human low-density lipoprotein receptor (LDLR) gene driven by a strong liver-specific promoter (thyroxin-binding globulin [TBG]) leads to high level and persistent gene expression in the liver. The approach proved efficacious in reducing plasma cholesterol levels and resulted in the regression of atherosclerotic lesions in a murine model of homozygous familial hypercholesterolemia (hoFH). Prior to advancing this vector, called AAV8.TBG.hLDLR, to the clinic, we set out to investigate vector biodistribution in an hoFH mouse model following IV vector administration to assess the safety profile of this investigational agent. Although AAV genomes were present in all organs at all time points tested (up to 180 days), vector genomes were sequestered mainly in the liver, which contained levels of vector 3 logs higher than that found in other organs. In both sexes, the level of AAV genomes gradually declined and appeared to stabilize 90 days post vector administration in most organs although vector genomes remained high in liver. Vector loads in the circulating blood were high and close to those in liver at the early time point (day 3) but rapidly decreased to a level close to the limit of quantification of the assay. The results of this vector biodistribution study further support a proposed clinical trial to evaluate AAV8 gene therapy for hoFH patients.

Adeno-associated virus serotype 8 gene therapy leads to significant lowering of plasma cholesterol levels in humanized mouse models of homozygous and heterozygous familial hypercholesterolemia.

Familial hypercholesterolemia (FH) is a life-threatening genetic disease caused by mutations in the gene encoding low-density lipoprotein receptor (LDLR). As a bridge to clinical trials, we generated a "humanized" mouse model lacking LDLR and apolipoprotein B (ApoB) mRNA editing catalytic polypeptide-1 (APOBEC-1) expression and expressing a human ApoB100 transgene in order to permit more authentic simulation of in vivo interactions between the clinical transgene product, human LDLR (hLDLR), and its endogenous ligand, human ApoB100. On a chow diet, the humanized LDLR-deficient mice have substantial hypercholesterolemia and a lipoprotein phenotype more closely resembling human homozygous FH (hoFH) than in previous mouse models of FH. On injection of an adeno-associated virus serotype 8 (AAV8) vector encoding the human LDLR cDNA, significant correction of hypercholesterolemia was realized at doses as low as 1.5 × 10(11) genome copies (GC)/kg. Given that some patients with heterozygous FH (heFH) cannot be adequately treated with current therapy, we then extended our studies to similarly "humanized" mice that were heterozygous for LDLR deficiency, and that have a lipoprotein phenotype resembling heterozygous FH. Injection of AAV8-hLDLR brought about significant reduction in total and LDL cholesterol at doses as low as 5 × 10(11) GC/kg. Collectively, these data demonstrate the safety and efficacy of the liver-specific AAV8-hLDLR vector in the treatment of humanized mice modeling both hoFH and heFH.

Inflammation promotes the loss of adeno-associated virus-mediated transgene expression in mouse liver.

BACKGROUND & AIMS: Non-self transgenes delivered to mouse liver via adeno-associated virus (AAV) are expressed stably due to the induction of immune tolerance. However, such transgene expression has been reported to be lost in higher-order primates. We investigated whether inflammatory processes, which likely differ between species, impact the stability of transgene expression. METHODS: We developed a mouse model that mimics a scenario in which a subject that has received hepatic AAV-mediated gene transfer develops subsequent, vector-unrelated, systemic inflammation. RESULTS: Inflammation eliminated previously stable expression of transgenes delivered by AAV; the limited tissue destruction and persistence of AAV genomes implicated pathways besides the cytotoxic T-cell response. Tumor necrosis factor-a down-regulated expression of the transgene from the AAV, indicating a role for similar inflammatory cytokines in such loss of transgene expression. CONCLUSIONS: Inflammation can block AAV-mediated expression of transgenes in mouse liver. The presence of inflammation might therefore affect hepatic expression of transgenes from viral vectors in humans.