Immune Responses to AAV in Clinical Trials

Current Gene Therapy, 2011, 11, 321-330 321 Immune Responses to AAV in Clinical Trials

Federico Mingozzi1 and Katherine A. High1,2,*

1Children’s Hospital of Philadelphia, Philadelphia, PA, USA; 2Howard Hughes Medical Institute, Philadelphia, PA, USA

Abstract: Findings in the first clinical trial in which an adeno-associated virus (AAV) vector was introduced into the liver of human subjects highlighted an issue not previously identified in animal studies. Upon AAV gene transfer to liver, two subjects developed transient elevation of liver enzymes, likely as a consequence of immune rejection of transduced hepatocytes mediated by AAV capsid-specific CD8+ T cells. Studies in healthy donors showed that humans carry a population of antigen-specific memory CD8+ T cells probably arising from wild-type AAV infections. The hypothesis formulated at that time was that these cells expanded upon re-exposure to capsid, i.e. upon AAV-2 hepatic gene transfer, and cleared AAV epitope-bearing transduced hepatocytes. Other hypotheses have been formulated which include specific receptor- binding properties of AAV-2 capsid, presence of capsid-expressing DNA in AAV vector preparations, and expression of alternate open reading frames from the transgene; emerging data from clinical trials however fail to support these compet- ing hypotheses. Possible solutions to the problem are discussed, including the administration of a short-term immunosuppression regimen concomitant with gene transfer, or the development of more efficient vectors that can be administered at lower doses. While more studies will be necessary to define mechanisms and risks associated with capsid-specific immune responses in humans, monitoring of these responses in clinical trials will be essential to achieving the goal of long- term therapeutic gene transfer in humans. Keywords: Adeno-associated virus, AAV, capsid, CD8 T cells, clinical trial, gene therapy, immune response, liver.

INTRODUCTION nantly in a non-integrated form as episomes; a smaller proportion of genomes integrate within chromosomes, although Clinical gene therapy has made considerable progress in it is not clear at this point whether integration events are as- the past several years [1]. First-in-human studies have high- sociated with tumor formation [14, 15]. lighted some of the major challenges to be overcome in order to achieve safe, effective gene transfer and, as a result, al- A large number of studies in experimental animals have though to date there are still no licensed gene therapy prod- established the potential of AAV vectors as a therapeutic tool ucts in the U.S. or Europe, successful clinical trial results [4-8, 16-21]. However, translation of these results into clini- have recently been reported using both integrating [2] and cal studies revealed some of the limits of animal models in non-integrating viral vectors [3]. While for integrating vi- fully predicting outcomes in humans, a finding by no means ruses the biggest concern is insertional mutagenesis, for unique to gene transfer therapeutics [22]. adeno-associated virus (AAV) vectors, the most commonly Differences in mechanisms by which the immune system used vectors for in vivo gene transfer, host immune re- recognizes and reacts to the AAV capsid may account for sponses are still the main concern. some of these limitations. Early findings in a clinical trial in AAV vectors are vehicles of choice for in vivo gene which an AAV-2 vector was introduced into the liver of he- transfer, as they can transduce a wide variety of tissues, me- mophilia B subjects [23, 24] suggested that immune re- diating long-term expression of the donated gene after a sin- sponse directed to the AAV capsid represents one of the last gle in vivo administration [4-8]. Wild-type AAV is not asso- major roadblocks to the development of a successful thera- ciated with any pathology in humans, and is also naturally peutic platform based on AAV-mediated gene transfer. replication-defective, requiring a helper virus such as adeno- In recent years, in response to this hypothesis, a number virus to replicate [9]. The low efficiency in transducing pro- of investigators started to look prospectively at T cell re- fessional antigen presenting cells (e.g. macrophages or den- sponses to capsid after vector administration in humans. dritic cells) [10-12] perhaps contributes to the generally low While the emerging data lend credence to this original con- immunogenicity of AAV vectors. Recombinant AAV vectors cept, the debate has begun to shift from whether there is a are one of the simplest gene therapy vectors, containing only detectable T cell response to capsid, to what it means clini- the transgene expression cassette flanked by two non-coding cally. viral inverted terminal repeats (ITRs) enclosed in a capsid composed of three structural proteins, VP1, 2, and 3 [13]. In this article we shall review the data available on host Once within a cell, AAV vector genomes persist predomi- immune responses to the AAV capsid in humans, the current hypothesis formulated to explain differences between hu- *Address correspondence to this author at the Children’s Hospital of Phila- mans and experimental animals, and strategies proposed to delphia, Philadelphia, PA, USA; Tel: +1 2155904521; Fax: +1 2155903660; avoid capsid recognition by the human immune system. E-mail: [email protected]

IMMUNE RESPONSES TO AAV CAPSID IN LIVER tolerant to the transgene product [5, 38-41]. Finally, beyond GENE TRANSFER: THE HEMOPHILIA B TRIALS hemophilia, the establishment of a gene transfer technology platform for hepatic gene transfer would allow the treatment The lack of circulating functional coagulation factor VIII of a wide range of genetic disorders and infectious diseases. (F.VIII) or factor IX (F.IX) results in the X-linked inherited disease hemophilia A or B, respectively [25, 26]. Hemophilia The pre-clinical work supporting AAV-mediated gene is characterized by defective coagulation, resulting in in- transfer to liver for hemophilia B is compelling [4, 5, 40, 42, creased risk of bleeding into joint spaces, with consequent 43]. Long-term expression (>9 years) at therapeutic levels arthropathy, or in bleeding in other internal sites, exposing (6-8% of normal levels) in hemophilic dogs was achieved patients to life-threatening hemorrhagic episodes. Although a after infusing a dose of 1x1012 vector genomes (vg)/kg of conventional protein replacement therapy is available to AAV-canine F.IX into the portal vein without production of manage hemophilia, several limitations of this treatment neutralizing antibodies to the transgene product [5]. prompted the study of a gene replacement approach to treat With the expectation that immune responses to the trans- the disease [27]. Furthermore, hemophilia represents an ideal gene product would be avoided in a liver-directed approach, disease model for gene therapy as: a) clinical endpoints are the first dose escalation study of hepatic artery delivery of an well defined and b) these are easily measured; c) circulating AAV-2 vector expressing human F.IX under the control of a levels between 5 and 100% of normal result in significant liver-specific promoter was initiated [23]. Subjects affected amelioration of the disease phenotype and are not associated by severe hemophilia B (pre-treatment F.IX levels <1% of with side effects; d) functional clotting factors can be pro- normal) were enrolled into three dose cohorts, receiving duced by a variety of tissues, including liver, muscle, and 8x1010 to 2x1012 vg/kg. Vector infusion proceeded unevent- fibroblasts [27-30]. However, one severe complication of fully for all subjects. both conventional, protein replacement therapy, and (poten- tially) gene therapy is the development of inhibitory antibod- None of the subjects enrolled in the first two dose cohorts ies (termed inhibitors) to the coagulation factor, which se- showed evidence of vector-related toxicity or efficacy de- verely complicate clinical management of the disease [31, fined as levels of human F.IX expression >1% of normal. th 32]. The 5 subject, subject E, was the first enrolled at what was expected to be a therapeutic dose (2x1012 vg/kg) based on Initial efforts to develop an AAV gene transfer approach the canine studies. This individual did indeed show therapeu- to treat hemophilia B focused on skeletal muscle as the target tic levels of F.IX expression initially, in the range of 10- cell [33, 34]. Clinical translation of results to affected hu- 12%, sufficient to convert his disease phenotype from severe mans resulted in long-term detection of F.IX transgene prod- to mild. However, beginning 4 weeks after vector infusion, uct in muscle biopsies [35-37]; however, at the doses tested, F.IX levels began to fall, and gradually returned to baseline circulating F.IX levels did not reliably rise above 1%. No (<1%) at 10 weeks after vector infusion. Concurrently, liver acute or long-term toxicity was observed following vector enzymes, which had been normal for the first few weeks administration, documented by normal clinical laboratory after vector infusion, began to rise, then slowly returned to testing, including absence of inhibitory and non-inhibitory normal without medical intervention. No inhibitory or non- antibodies to the secreted transgene product, absence of inhibitory antibodies to the F.IX transgene product were muscle enzyme elevation, and absence of lymphocytic infil- measured in this subject. Possible toxic or infectious causes trates in repeated muscle biopsies analyzed over several of the increase in liver enzymes were sought and excluded. years. Exposure to HIV, HBV, and/or HCV as a conse- At the request of regulatory agencies, the next subject stud- quence of prior infusion with infected plasma-derived prod- ied, subject G, was infused at a 5-fold lower dose than E and ucts did not alter the outcome of gene transfer. Beyond test- also experienced an asymptomatic, self-limited liver enzyme ing for antibodies to AAV and to F.IX, no specific immu- elevation, with similar kinetics. nological assays (e.g. ELISpot) were performed in this study. ELISpot analysis of T cell responses in peripheral blood The low efficiency of protein secretion into the circula- mononuclear cells (PBMC) isolated from subject G demon- tion from a site of synthesis in the muscle, combined with a strated IFN- production in response to AAV capsid pep- relatively higher immunogenicity (vide infra) compared to tides, but not to F.IX peptides [23]. Identification of AAV other target tissues, shifted the attention to the liver. Com- capsid peptide epitopes to which subjects E and G reacted, pared to other tissues, liver presents several advantages as a allowed the synthesis of MHC class I (MHC I) pentamers for target for AAV-mediated gene transfer. First, it is the natural the direct quantitation of capsid-specific CD8+ T cell popula- site of biosynthesis of clotting factors, so the necessary post- tions in peripheral blood [24]. This population was shown to translational modifications take place accurately and effi- expand then contract after vector infusion, with a time course ciently; second, since it is a highly vascularized organ, thera- that closely matched the rise and fall of serum transaminases peutic proteins produced in the liver can easily and effi- [24]; moreover, even long (>2 years) after vector exposure, ciently gain access to the systemic circulation, conferring a although capsid-specific CD8+ T cells are no longer detect- dose advantage compared to muscle as documented by ani- able in peripheral blood by ELISpot, in vitro expansion of mal studies; third, expression of antigens in transduced hepa- lymphocytes with the relevant epitope in the presence of IL- tocytes is associated with induction of tolerance rather than 2 and IL-7 produces a robust expansion of a population of immunity. Hepatic administration of an AAV vector express- capsid-specific CD8+ T cells, indicating the presence of a ing human F.IX results in sustained expression without neu- pool of memory T cells, which through homeostatic prolif- tralizing antibody formation, even in animals not previously eration are maintained long after gene transfer. Immune Responses to AAV in Clinical Trials Current Gene Therapy, 2011, Vol. 11, No. 4 323

Differently from direct intramuscular AAV vector deliv- [50-52]. In addition, cardiac muscle has also been targeted ery [37], the presence of neutralizing antibodies (NAb) to the through an intravascular approach, for cardiac failure [53]. AAV-2 capsid prior to vector infusion had a profound effect Following the initial findings in the AAV2-F.IX liver on liver transduction. Subjects E and F both received the gene transfer trial, monitoring of capsid-directed T cell re- highest dose of vector, however only subject E, who had low sponses was implemented in several clinical studies, with the pretreatment NAb, achieved appreciable levels of F.IX ex- result that the largest dataset on capsid T cell responses in pression, while F, who had a titer of 1:17 did not. Subject F humans undergoing AAV gene transfer comes from muscle- did not experience an increase in liver enzymes. Similarly, directed AAV gene transfer trials (Table 1). One important among subject C, D, and G, who all received the same vector point that should be kept in mind when analyzing and com- dose, only subject G, with the lowest pretreatment NAb to paring results from different laboratories is that the assay AAV-2, had a rise in liver enzymes [23]. These data indicate used in all the studies (an IFN- ELISpot assay specific for that even low levels of NAb can prevent transduction when the AAV capsid) may be a source of variability itself [54]. the AAV vector is introduced through the systemic circulation, and that the presence of antibodies to AAV is not pre- Table 1 summarizes the results of monitoring capsid T dictive of T cell responses against AAV capsid. cell responses in muscle gene transfer studies. With few exceptions, the magnitude of T cell responses directed against Following the initial observations in the AAV2-F.IX the AAV capsid shows a rough correlation with dose of vec- trial, strategies for avoiding capsid immune response have been proposed (vide infra). Investigators at St. Jude Chil- tor administered, with an increase in frequency of circulating dren’s Research Hospital in Memphis, TN, and University reactive T cells in PBMC (expressed in the ELISpot assay as College London, UK, developed a more efficient expression spot forming units, SFU, per million cells plated in the as- cassette and vector, based on the hypothesis that such a vec- say) and frequency of reactive subjects at higher vector tor could achieve therapeutic levels of F.IX expression at doses. Studies performed by our laboratory in the context of vector doses too low to elicit an immune response. Features the AAV1-LPL study [47] showed that dose-dependence in that make this vector more efficient include use of an AAV-8 the activation of capsid-specific T cell responses also mani- serotype which has a strong tropism for liver and can be in- fested with a faster kinetics of detection of T cells in periph- fused intravenously [42, 43]; use of a self-complementary eral blood at higher vector doses [55]. Additionally, in this vector design, to overcome dependence on host cell machin- study higher vector doses seemed to be associated with a Th1 and Th2 response (with activation of both CD8+ and ery for synthesis of the necessary second strand [44]; and + codon optimization to improve translational efficiency [4]. A CD4 T cells, respectively), compared to a mostly Th2 re- clinical trial testing this newly developed vector has been sponse detected at lower vector doses. recently initiated. Hemophilia B subjects received the More recent results from a clinical study of AAV-1 in- AAV8-F.IX vector intravenously at doses of 2x1010vg/kg, 10 11 tramuscular gene transfer of alpha-sarcoglycan (-SG) in 6x10 vg/kg, and 2x10 vg/kg. IFN- ELISpot analysis of T children affected by limb-girdle muscular dystrophy type 2D cell responses directed against the AAV-8 capsid in the first [51, 52] seem to confirm the initial findings in the context of two subjects showed no activation of capsid T cell responses AAV1-LPL. In this study, -SG expression was detectable in peripheral blood. At the next vector dose, IFN- ELISpot in muscle biopsies collected at week 6 (n=2 subjects) and 3 on PBMC showed a robust activation of capsid specific T 6 months (n=1 subject) [52]. In these first 3 subjects, muscle cells (spot forming units/10 PBMC up to ~1,700). Expres- inflammation and T cell infiltrates were evident, together sion of the F.IX transgene product was observed both at the with upregulation of MHC I expression in injected muscles. low and mid vector doses. Follow up of the infused subjects, Tunel staining showed apoptosis of T cells locally in the as well as further dose escalation, is ongoing at this time. No muscle, a phenomenon previously described in mice receiv- loss of FIX transgene expression and no elevation of liver ing an AAV vector expressing GFP intramuscularly [56]. enzymes have been observed in these 4 subjects. Additional Low but detectable capsid-specific T cell responses were studies will be needed to determine the safety of further dose documented in one subject. escalation, and whether there are serotype-specific differences in capsid processing and presentation that affect bio- In the second part of the study, 3 subjects received the logical consequences of T cell activation [45]. same vector at the same dose and muscle biopsies were collected 6 months after gene transfer [51]. In 2/3 subjects, a- Importantly, these findings suggest that infusion of SG expression and increase in muscle fiber size was docu- AAV-8 vectors can trigger activation of T cells in a dose- dependent manner, a result consistent with the experience in mented, together with upregulation of MHC I. In one sub- AAV gene transfer to muscle in humans (vide infra), and at ject, the muscle biopsy showed no evidence of transduction. odds with an earlier report suggesting that AAV-8 infusion Interestingly, after vector administration this subject had a would not trigger T cell activation [46]. rapid rise in anti-AAV-1 neutralizing antibody titer to levels higher than 30-fold the other subjects; the subject also had a IMMUNE RESPONSES TO AAV VECTORS IN MUS- detectable T cell response to the AAV capsid at week 2 after CLE GENE TRANSFER vector administration detected by IFN- ELISpot on PBMC. These results further support the hypothesis that humans can Other groups have adopted protocols similar to the he- mount T cell responses against AAV transduced target cells. mophilia B muscle trial [36, 37] for studies of AAV gene Interestingly, lack of muscle transduction and brisk immune transfer for lipoprotein lipase (LPL) deficiency [47], 1- responses against the AAV capsid were documented in 0/3 antitrypsin deficiency [48, 49], and muscular dystrophies subjects aged 11 to 14 years and in 1/3 subjects aged 23 to 324 Current Gene Therapy, 2011, Vol. 11, No. 4 Mingozzi and High

Table 1. T Cell Responses to Capsid After Intramuscular Injection of AAV Vectors in Humans

IFN- Magnitude Time Point First Serotype Transgene (ref) Dose (vg)* ELISpot SFU/106 cells Detected

3.25 x 1011 1/3 60-70 2 wks, 6 wks -Sarcoglycan (51, AAV-1 52) 3.25 x 1011 2/3 150-250 Day 2, 2 wks

AAV-1 SERCA2a (53) 3 x 1012 1/3 90 4 wks, 6 wks

7.0 x 1012 2/4 ~300 12 wks AAV-1 LPL (47, 55) 2.1 x 1013 3/6 ~300 4-6 wks

2.2 x 1013 2/2 ~100 2 wks Alpha 1-Antitrypsin AAV-1 (48) 6.0 x 1013 3/3 <200 - 428 2-4 wks

*Assuming weight of a child = 30kg, adult = 70kg; SFU = spot forming unit.

43 years, suggesting a correlation between capsid immune of reactive T cells [56], and variability related to the sub- responses and timing of exposure to wild type AAV. jects’ HLA alleles. A third example of AAV muscle gene transfer trial worth Finally, it is important to point out that potential issues mentioning was also recently published [48]. In this study, related to transgene product immunogenicity in muscle- intramuscular administration of an AAV-1 vector encoding directed gene transfer were highlighted in a recent study of 1-antitrypsin resulted in detectable but subtherapeutic levels AAV gene transfer for micro-dystrophin in DMD-affected of transgene expression. Vector administration was associ- children [50]. In this study, an AAV vector encoding a func- ated with detection of capsid-specific CD8+ T cells in tional truncated version of the dystrophin protein was deliv- PBMC, which did not seem to affect the duration of trans- ered intramuscularly at doses of 2x1010 vg/kg and 1x1011 gene expression. While the lack of cytolytic activity of cap- vg/kg. Following gene transfer, transgene expression was sid-specific T cells in this particular case could be imputed to documented in 2/4 subjects tested at day 42; at day 90, 0/2 the immunomodulatory properties of the transgene product subjects had detectable dystrophin positive muscle fibers, itself [57], an open question is whether further dose escala- suggesting loss of transgene expression. Additional studies tion to achieve therapeutic efficacy will eventually trigger in the subjects’ PBMC revealed the development of both immune-mediated clearance of transduced muscle fibers. CD4+ and CD8+ T cell responses directed against the transgene product. In particular, one subject from the low-dose In summary, while it is now clear that AAV vector ad- cohort and three subjects from the high-dose cohort devel- ministration in humans results in antigen-specific T cell acti- oped an IFN- response to peptide pools derived from the vation, the open question is the significance of the detection mini-dystrophin transgene. Two of the subjects had a re- of these T cells. In at least two studies [51, 55], early detec- sponse to mini-dystrophin prior to gene transfer, an unex- tion of capsid-specific T cells has been associated with lack pected finding with potential implications for all gene addi- of evidence of target tissue transduction. tion studies, which further reinforce the importance of care- Further analysis of these results is complicated by several ful monitoring of both vector and transgene T cell responses factors, which may be relevant to the experience with AAV in all AAV gene transfer studies. Intravascular delivery of gene transfer to liver. First, reported results to date have been AAV vectors to achieve widespread muscle transduction [8, confined to PBMCs, while the cells of true interest are those 59-61] may offer a solution to the problem of transgene im- in the target tissue itself. Because in most clinical trials these munogenicity, as it seems less prone than intramuscular ad- cells are not easily accessible, analysis has been restricted to ministration to trigger immune responses directed against the PBMCs as surrogates, with all the shortcomings this implies. delivered transgene [62, 63]. Safety and feasibility of the Second, if as seems to be the case, the magnitude of the im- technique is currently being tested in Becker’s muscular dys- mune response is rather low, it may be difficult to define a trophy patients by infusing saline instead of vector [64]. “clinical correlate” in the case of some target tissues. Addi- tional factors may play a role in shaping the outcome of ex- GENE TRANSFER TO IMMUNOPRIVILEGED SITES pansion of capsid-specific T cells, including immunomodulatory properties of the transgene product [48, 57], levels of Delivery of relatively small doses of AAV vectors to the expression of MHC I in the target tissue [51, 52], baseline brain or the subretinal space is generally associated with levels of inflammation within the target tissue [58], apoptosis little to no immune response to the capsid detectable in serum and PBMC. A modest increase in anti-AAV antibody Immune Responses to AAV in Clinical Trials Current Gene Therapy, 2011, Vol. 11, No. 4 325 titer was documented following vector administration in a Alternate Hypotheses small proportion of subjects undergoing intracranial gene While the memory CD8+ T cell response hypothesis has transfer for Parkinson’s disease [65-68], Canavan disease been difficult to prove or disprove, a number of alternative [69], and late infantile neuronal ceroid lipofuscinosis [70]. hypotheses have been proposed, none entirely satisfactory, Similarly, subretinal administration of an AAV-2 vector en- and also difficult to prove or disprove. One proposed that the coding the RPE65 transgene resulted in minimal and tran- plasmid encoding the AAV capsid genes (rep/cap plasmid) sient anti-capsid humoral responses and no detectable capsid had been packaged as a low level contaminant in the manu- T cell responses in humans [71, 72]. facturing process. Although encapsidation of prokaryotic Safety of AAV vector readministration to the contralat- sequences during AAV vector production has been docu- eral eye in subjects who previously received AAV2-RPE65 mented [78], studies performed on clinical lots of vector util- gene transfer [71] is currently being tested (ClinicalTri- ized in AAV-infused subjects who developed transaminitis als.gov ID# NCT01208389). Preclinical studies in RPE65- revealed only very low levels (<1%) of DNA other than the deficient dogs and non-human primates suggest that the ap- AAV expression cassette in viral preparations. Additionally, proach is safe, and subretinal vector readministration to ani- in these experiments no evidence of capsid expression has mals previously exposed to AAV vectors does not seem to been found in vitro or in vivo using sensitive RT-PCR tech- trigger harmful immune responses against the AAV capsid niques after transducing cell lines at high multiplicity of in- [73]. fections (MOIs) or injecting mice with AAV vectors at high doses [79]. Another hypothesis proposed that the expression LEADING HYPOTHESES FOR IMMUNOTOXICITY of alternate open reading frames in the F.IX expression cas- IN AAV GENE TRANSFER sette had triggered a CD8+ T cell response to these neoepi- The “Prior-Exposure to Wild-Type AAV” Hypothesis topes, which resulted in destruction of the transduced cells [80, 81]. This seemed difficult to reconcile with long-term One open question is why human subjects manifest T cell expression of AAV vectors expressing human F.IX in ex- responses to AAV capsid after vector infusion, while animal perimental animals, who would presumably also mount a models do not [24]. Prior exposure to AAV capsid may un- response to these neoepitopes. Finally, another study in mice derlie the difference in response. A high percentage of hu- and NHP proposed that the AAV-8 capsid would not activate mans, the only natural hosts for wild-type AAV-2 infection, a CD8+ T cell response because it does not transduce den- are infected by AAV-2 in childhood [74]. Because AAV is dritic cells (DC) [46]. The study speculated that the RXXR naturally replication defective, this initial infection invaria- heparan sulfate proteoglycans (HSPG) binding motif found bly takes place together with a helper virus infection such as within the AAV-2 capsid, but not within AAV-8, would me- adenovirus. Although AAV-2 on its own may not induce the diate the entry of the virus into a DC and its consequent inflammatory reactions needed for stimulation of a maximal processing and presentation on MHC I. This hypothesis does adaptive immune response, in combination with the helper not seem to extend to human subjects, who were recently virus, which causes activation of the innate immune system, shown to exhibit a dose-dependent activation and prolifera- it is likely that CD8+ T cells directed to the antigens of both tion of capsid-specific T cells after intravenous infusion of the helper virus and of AAV are formed. Upon controlling AAV-8 vectors [82]. the infection, the frequency of AAV-specific CD8+ T cells would be expected to decline, leaving behind a small pool of MODELS OF CAPSID IMMUNOGENICITY memory T cells, which through homeostatic proliferation are maintained throughout the lifetime of an individual. On re- Independent findings published by three groups [83-85] exposure to capsid, these memory CD8+ T cells are activated indicate that animals fail to develop T cell responses to AAV and eliminate the AAV capsid-harboring cells (the trans- capsid. Mice immunized against the AAV capsid by either duced hepatocytes). Because memory T cells are more read- an intramuscular injection of an adenovirus expressing AAV ily triggered than naïve lymphocytes, human subjects, un- capsid or by repeated injection of dendritic cells pulsed with capsid do indeed develop AAV-specific cytotoxic T lympho- dergoing re-exposure, have an outcome different from ex- + perimental animals, undergoing what amounts to a primary cytes, however, these primed CD8 T cells fail to clear in infection with AAV. vivo AAV-transduced hepatocytes. More recently, two in vivo models of capsid T cell responses have been developed. The very high prevalence of anti-AAV antibodies in Both models used AAV vectors modified by inserting within healthy donors [75] and the fact that humans carry a pool of the AAV capsid one or multiple copies of the ovalbumin expandable AAV-specific memory CD8+ T cells able to pro- MHC I epitope SIINFEKL [86, 87]. Proliferation of ova- duce IFN- in response to AAV-derived peptide epitopes specific CD8+ T cells in response to AAV-SIINFEKL vector [24] and to kill AAV transduced target cells [76] support this administration was observed in both studies. Interestingly in model. the study from Dr. Ertl’s lab [87] a significant difference in the kinetics of T cell proliferation was evidenced between Additional factors may contribute to the difference in + outcome between humans and other species. In particular, AAV-2 or AAV-8 vectors. SIINFEKL-specific CD8 T cells differenced in reactivity of T cells (vide infra) [77] may also showed a delayed but prolonged kinetics of T cell prolifera- explain why non-human primates, which like humans are tion in animals injected with AAV-2 vectors compared with exposed to wild type AAV, do not mount cytotoxic re- AAV-8 vectors. This is in agreement with previously pub- sponses against the AAV capsid. lished data on differences in the vector genome uncoating rate between AAV-2 and AAV-8 [88]. 326 Current Gene Therapy, 2011, Vol. 11, No. 4 Mingozzi and High

The difference in T cell reactivity to AAV between ex- limitations of both in vitro and in vivo models of AAV cap- perimental animals and humans may be explained by the sid immunogenicity developed so far (vide supra) further greater responsiveness of human T cells to T cell receptor highlight the importance of careful immunomonitoring in stimulation due to the loss of sialic acid-recognizing Ig- AAV studies in humans. superfamily lectins (Siglecs) on human T cells [77]. Indeed, studies in a transgenic mouse model lacking Siglecs (cmah-/-) Reduce Antigenic Load + showed a higher degree of CD8 T cell activation in re- Combined data on the occurrence of capsid T cell response to various stimuli [89]. sponses in humans seem to argue strongly for this strategy to In addition to differences at the levels of T cell recogni- evade capsid-triggered immunotoxicity. Results from both tion, other possible explanations include better homing of liver [23, 45, 82] and muscle (Table 1) gene transfer trials, as primed human T cells to liver [90] or inefficient presentation well as results from studies on subretinal gene transfer [71, of capsid antigen in murine liver. 72], would argue that lower vector doses are less likely to be “seen” by the immune system. Additionally, data generated In vitro systems using human cells have been more suc- in vitro [76, 92] also suggest that there is a clear dose- cessful than experimental animals in modeling T cell immu- + response relationship between the MOI of infection of target nity against AAV vectors. Using capsid-specific CD8 T cells and the extent of capsid antigen presentation, which in cells expanded from normal donor PBMC, Pien and col- turn triggers immune-mediated clearance of transduced cells leagues developed an in vitro cytotoxic T lymphocyte (CTL) Fig. (1). assay in which an HLA-matched human hepatocyte cell line transduced with AAV vectors was used as target [76]. Effi- Thus, diminishing the amount of antigen (i.e. vector cap- cient in vitro killing of transduced cells correlated with the sid) introduced into the human organism may help to evade detection of capsid antigen presentation on MHC I, visual- immune responses, as will introduction of vector into closed, ized by confocal microscopy on transduced cells stained with immunoprivileged sites [65-68, 71, 72, 94-96]. Therapeutic a soluble T cell receptor (TCR) multimer [91] specific to the levels of transgene expression may be achieved at lower vec- AAV capsid. These results were further extended using a T tor doses using more efficient expression cassettes, AAV cell line engineered to recognize the AAV MHC I epitope serotypes with higher tropism for the target tissue [97], self- VPQYGYLTL (Jurma-VPQ) and express luciferase upon complementary AAVs [44], codon-optimized transgenes [4, TCR engagement [92]. When incubated with AAV- 42], or, when possible, transgene variants with higher spe- transduced human hepatocytes, Jurma-VPQ T cells ex- cific activity [98]. Some of these strategies are being tested pressed luciferase in a dose-dependent manner over a wide in a clinical study of AAV-8 gene transfer for hemophilia B range of MOIs, and treatment of transduced target cells with (vide supra). the proteasome inhibitor bortezomib blocked AAV antigen recognition (vide infra). Modulate T Cell Responses to Capsid Together, in vitro studies support a model of AAV capsid A short course of immunosuppression (IS), given around antigen presentation in which the AAV capsid enters the cell the time of gene transfer, has been proposed to create a win- through receptor-mediated endocytosis. Upon endosomal dow of time during which immune responses are blunted escape, AAV virions are ubiquitinated and become targets while capsid antigen is cleared from the transduced cells. for proteasomal degradation and loading onto MHC I. This Pre-clinical studies in non-human primates (NHP) [38, 99] results in flagging of transduced cells for recognition by cap- highlighted the importance of CD4+CD25+FoxP3+ regulatory sid-specific CTLs Fig. (1). T cells (Tregs) in establishing and maintaining tolerance to transgene product, confirming previous data developed in It should be noted that in vitro models of AAV capsid mice [39, 41]. In particular, IS regimens that interfere with antigen presentation suffer from two main limitations: the Treg induction or homeostasis should be avoided, as they first comes from the poor in vitro transduction efficiency of can trigger unwanted immune responses against the trans- most AAV serotypes other than AAV-2, which makes com- gene product. parison of alternate AAV serotypes in vitro almost impossi- ble; the second is related to the fast cell cycling in tissue cul- A short course of IS (10 weeks) consisting of mycophe- ture, which is not representative of the in vivo rate of cell nolate mofetil (MMF) and sirolimus, which was demon- replication and therefore complicates the extrapolation of in strated to be safe in NHP [38], was tested in a human subject vitro data on capsid half-life. undergoing AAV2-F.IX hepatic gene transfer. The study demonstrated that the approach was safe and IS was well Overall, the lack of a fully predictive animal model to tolerated (Mingozzi and High unpublished observation). study T cell responses to capsid still represents a major limi- Similarly, no adverse events were reported in studies for tation to understanding the mechanism(s) of this phenome- AAV-1 intramuscular gene transfer for lipoprotein lipase non. Development of murine models in which the liver can deficiency in which the vector was co-administered a 12 be repopulated with human hepatocytes [93] may afford a weeks course of IS with MMF and cyclosporine A [100]. solution, with the caveat that a humanized immune system may also be required to fully mimic the findings in the clinic. As a general principle, immunomodulation strategies should be carefully tested for each gene transfer application POSSIBLE SOLUTIONS TO THE PROBLEM in the relevant animal model to highlight possible paradoxi- cal effects of immunosuppressive drugs on the immune sys- Different strategies could be effective in escaping, or tem itself [38, 101], or other effects on vector persistence or limiting, T cell responses to AAV capsid in humans. The transduction efficiency [102]. Immune Responses to AAV in Clinical Trials Current Gene Therapy, 2011, Vol. 11, No. 4 327

Fig. (1). Model of AAV capsid antigen presentation. Upon entering the cell via receptor-mediated endocytosis, AAV vectors are released into the cytoplasm. This sequence of events is essential for the vector to reach the nucleus where it will eventually results in stable cell transduction. Within a cell, AAV capsid is also subject to ubiquitination and cleavage by the proteasome, a process that leads to the loading of capsid peptide epitopes on MHC I and, ultimately, to capsid antigen presentation to CD8+ T cells.

Alternate Serotypes and remains an important question to address. Possibilities include the capsid proteins being put “on hold”, until the Results accumulating from AAV clinical trials seem to proteasome inhibition has subsided and then all capsids be- support the idea that the high degree of conservation of cap- ing processed and degraded en masse, or more likely, the sid sequences among alternate AAV serotypes [97], and the capsids being eliminated by some other cellular degradation consequent high degree of cross-recognition by T cells [24], mechanism such as autophagy [106]. together with the complexity of the human HLA antigen recognition system, make it less likely that switching sero- Select Naïve Subjects types will succeed in avoiding T cell immunity against the AAV capsid. AAV-8 administration, in particular, seems to If indeed the memory T cell hypothesis is correct (vide result in capsid T cell activation in humans in a dose- supra), selecting subjects that have never encountered the dependent manner [82], in contrast to an earlier hypothesis AAV capsid antigen may be an effective way to escape im- (vide supra). Recently, Srivastava and colleagues described mune responses directed against transduced cells. However, AAV capsid mutants in which tyrosine residues that are tar- the approach may be challenging both because available as- geted for ubiquitination were mutated to phenylalanine says are not sensitive enough to determine definitively [103]. While several reports indicate that these capsid vari- whether a subject is naïve to the AAV capsid. With the ca- ants exhibit higher transduction efficiency [104, 105], it is veat that some studies are targeting immunoprivileged sites, not yet clear whether prevention of ubiquitination of surface- monitoring of capsid immune responses in children undergo- exposed tyrosine residues will impact capsid antigen presen- ing AAV gene transfer for both non-lethal disorders like tation. congenital blindness [71] or lethal conditions like lysosomal storage disorders (Clinicaltrials.gov ID# NCT01161576) and Pharmacological Alteration of Capsid Antigen Presenta- muscular dystrophies [51, 52] will help to test this hypothe- tion sis. Recent data indicate that treatment of AAV-transduced CONCLUSIONS cells with the proteasome inhibitor bortezumib, at doses comparable to those used in the clinic (serum levels 30–100 Clinical experience with AAV vectors continues to ac- nmol/l), results in decreased capsid antigen presentation [92]. cumulate. Translation of preclinical results into humans with This may be a feasible and relatively safe intervention to severe hemophilia highlighted a previously unrecognized reduce capsid immunogenicity. However the fate of capsid problem related to immunogenicity of AAV capsid. Absent proteins after proteasome inhibition is currently unknown an animal model that faithfully recapitulates the findings in 328 Current Gene Therapy, 2011, Vol. 11, No. 4 Mingozzi and High humans, further analysis and solution of this problem will [8] Arruda VR, Stedman HH, Haurigot V, et al. Peripheral need to rely on in vitro studies [76, 92] and on carefully transvenular delivery of adeno-associated viral vectors to skeletal muscle as a novel therapy for hemophilia B. Blood 2010; 115: monitored studies in human subjects. On the other hand, 4678-88. experience to date indicates that injection of small doses of [9] Muzyczka N, Berns KI. Parvoviridae: The viruses and their AAV vector into immunoprivileged sites [65-68, 71, 72, 94- replication. 4th ed. Knipe DM, Howley PM, editors. Philadelphia: 96] may be tolerated uneventfully, thus leading to long-term Lippincott, Williams and Wilkins; 2001. expression of transgene. [10] Jooss K, Yang Y, Fisher KJ, Wilson JM. Transduction of dendritic cells by DNA viral vectors directs the immune response to A difficulty in interpreting clinical data to date is the lack transgene products in muscle fibers. J Virol 1998; 72: 4212-23. of easily quantifiable endpoints; exceptions to this include [11] Vandendriessche T, Thorrez L, Acosta-Sanchez A, et al. Efficacy and safety of adeno-associated viral vectors based on serotype 8 plasma proteins such as F.IX and 1-antitrypsin. It is possi- and 9 vs. lentiviral vectors for hemophilia B gene therapy. J ble that reports suggesting efficacy at early time points, fol- Thromb Haemost 2007; 5: 16-24. lowed by loss of efficacy [107] are actually reflecting undi- [12] Zaiss AK, Liu Q, Bowen GP, Wong NC, Bartlett JS, Muruve DA. agnosed immune-mediated elimination of transduced cells. Differential activation of innate immune responses by adenovirus and adeno-associated virus vectors. J Virol 2002; 76: 4580-90. Again, for some target organs such as liver, or skeletal or [13] Samulski RJ, Berns KI, Tan M, Muzyczka N. Cloning of adeno- cardiac muscle, immune-mediated destruction of transduced associated virus into pBR322: rescue of intact virus from the cells is easily appreciated by measuring relevant serum en- recombinant plasmid in human cells. Proc Natl Acad Sci USA zyme levels, while for other cell types such as those lining 1982; 79: 2077-81. the respiratory tract, this is more difficult to detect. [14] Donsante A, Miller DG, Li Y, et al. AAV vector integration sites in mouse hepatocellular carcinoma. Science 2007; 317: 477. Future studies will need to address fundamental questions [15] Li H, Malani N, Hamilton SR, et al. Assessing the potential for such as: a) is there a reliable way to predict which subjects AAV vector genotoxicity in a murine model. Blood 2010. + [16] Song S, Morgan M, Ellis T, et al. Sustained secretion of human will mount a CD8 T cell response to AAV capsid; b) are alpha-1-antitrypsin from murine muscle transduced with adeno- some serotypes, delivery methods, and target tissues more or associated virus vectors. Proc Natl Acad Sci USA 1998; 95: 14384- less immunogenic than others; c) what are the biomarkers 8. that can predict immune rejection of transduced cells; finally, [17] Mas A, Montane J, Anguela XM, et al. Reversal of type 1 diabetes by engineering a glucose sensor in skeletal muscle. Diabetes and most importantly d) what is the fate of the AAV capsid 2006;55: 1546-53. inside the transduced cell and how long does the capsid per- [18] Goyenvalle A, Vulin A, Fougerousse F, et al. Rescue of dystrophic sist in an immunologically detectable form. muscle through U7 snRNA-mediated exon skipping. Science 2004; 306: 1796-9. Capsid-specific ELISpot screening for all subjects under- [19] Ghosh A, Yue Y, Long C, Bostick B, Duan D. Efficient whole- going AAV-mediated gene transfer is warranted [80] and, if body transduction with trans-splicing adeno-associated viral performed, data produced in the only relevant model avail- vectors. Mol Ther 2007;15:750-5. able to date, humans, will contribute to the full understand- [20] Flotte TR. Recent developments in recombinant AAV-mediated gene therapy for lung diseases. Curr Gene Ther 2005; 5: 361-6. ing of the mechanisms and risks associated with T cell re- [21] Adriaansen J, Tas SW, Klarenbeek PL, et al. Enhanced gene sponses to AAV capsid. transfer to arthritic joints using adeno-associated virus type 5: implications for intra-articular gene therapy. Ann Rheum Dis 2005; ACKNOWLEDGEMENTS 64: 1677-84. [22] Suntharalingam G, Perry MR, Ward S, et al. Cytokine storm in a This work was supported by NIH P01 HL078810, by phase 1 trial of the anti-CD28 monoclonal antibody TGN1412. N NIH PPG, 2P01 HL64190, by NIH R01 HL094396, the Engl J Med 2006; 355: 1018-28. [23] Manno CS, Pierce GF, Arruda VR, et al. Successful transduction of Howard Hughes Medical Institute, and the Children’s Hospi- liver in hemophilia by AAV-Factor IX and limitations imposed by tal of Philadelphia the host immune response. Nat Med 2006; 12: 342-7. [24] Mingozzi F, Maus MV, Hui DJ, et al. CD8(+) T-cell responses to REFERENCES adeno-associated virus capsid in humans. Nat Med 2007; 13: 419- 22. [1] Kay MA. State-of-the-art gene-based therapies: the road ahead. Nat [25] Pollak ES, High KA. Hemophilia A: Factor IX deficiency. In: Rev Genet 2011; 12: 316-28. Scriver CR, Beaudet AL, Valle D, Sly WS, editors. The Metabolic [2] Naldini L. Ex vivo gene transfer and correction for cell-based and Molecular Bases of Inherited Disease. 5th ed. New York: Mc- therapies. Nat Rev Genet 2011; 12: 301-15. Graw Hill; 2001. p. pp. 4393-413. [3] Mingozzi F, High KA. Therapeutic in vivo gene transfer for genetic [26] Kazazian HH, Tuddenham EGD, Antonarkakis SE. Hemophilia A: disease using AAV: progress and challenges. Nat Rev Genet 2011; deficiency of coagulation factor VIIII. In: Scriver CR, Beaudet AL, 12: 341-55. Sly WS, Valle D, editors. The Metabolic and Molecular Bases of [4] Nathwani AC, Gray JT, Ng CY, et al. Self-complementary adeno- Inherited Disease. New York: McGraw-Hill; 2001. p. pp. 4367-92. associated virus vectors containing a novel liver-specific human [27] High KA. AAV-mediated gene transfer for hemophilia. Ann N Y factor IX expression cassette enable highly efficient transduction of Acad Sci 2001 ; 953: 64-74. murine and nonhuman primate liver. Blood 2006; 107: 2653-61. [28] Snyder RO, Miao C, Meuse L, et al. Correction of hemophilia B in [5] Niemeyer GP, Herzog RW, Mount J, et al. Long-term correction of canine and murine models using recombinant adeno-associated inhibitor-prone hemophilia B dogs treated with liver-directed viral vectors. Nat Med 1999; 5: 64-70. AAV2-mediated factor IX gene therapy. Blood 2009; 113: 797- [29] Palmer TD, Thompson AR, Miller AD. Production of human factor 806. IX in animals by genetically modified skin fibroblasts: potential [6] Davidson BL, Stein CS, Heth JA, et al. Recombinant adeno- therapy for hemophilia B. Blood 1989; 73: 438-45. associated virus type 2, 4, and 5 vectors: transduction of variant [30] Arruda VR, Hagstrom JN, Deitch J, et al. Posttranslational cell types and regions in the mammalian central nervous system. modifications of recombinant myotube-synthesized human factor Proc Natl Acad Sci USA 2000;97: 3428-32. IX. Blood 2001; 97: 130-8. [7] Acland GM, Aguirre GD, Ray J, et al. Gene therapy restores vision [31] Hay CR, Brown S, Collins PW, Keeling DM, Liesner R. The in a canine model of childhood blindness. Nat Genet 2001; 28: 92- diagnosis and management of factor VIII and IX inhibitors: a 5. guideline from the United Kingdom Haemophilia Centre Doctors Organisation. Br J Haematol 2006; 133: 591-605. Immune Responses to AAV in Clinical Trials Current Gene Therapy, 2011, Vol. 11, No. 4 329

[32] DiMichele D. Inhibitor development in haemophilia B: an orphan (CUPID Trial), a first-in-human phase 1/2 clinical trial. J Card Fail disease in need of attention. Br J Haematol 2007; 138: 305-15. 2009; 15: 171-81. [33] Herzog RW, Hagstrom JN, Kung SH, et al. Stable gene transfer [54] Janetzki S, Panageas KS, Ben-Porat L, et al. Results and and expression of human blood coagulation factor IX after harmonization guidelines from two large-scale international Elispot intramuscular injection of recombinant adeno-associated virus. proficiency panels conducted by the Cancer Vaccine Consortium Proc Natl Acad Sci USA 1997; 94: 5804-9. (CVC/SVI). Cancer Immunol Immunother 2008; 57: 303-15. [34] Herzog RW, Yang EY, Couto LB, et al. Long-term correction of [55] Mingozzi F, Meulenberg JJ, Hui DJ, et al. AAV-1-mediated gene canine hemophilia B by gene transfer of blood coagulation factor transfer to skeletal muscle in humans results in dose-dependent IX mediated by adeno-associated viral vector. Nat Med 1999; 5: activation of capsid-specific T cells. Blood 2009; 114: 2077-86. 56-63. [56] Velazquez VM, Bowen DG, Walker CM. Silencing of T [35] Jiang H, Pierce GF, Ozelo MC, et al. Evidence of multiyear factor lymphocytes by antigen-driven programmed death in recombinant IX expression by AAV-mediated gene transfer to skeletal muscle in adeno-associated virus vector-mediated gene therapy. Blood 2009; an individual with severe hemophilia B. Mol Ther 2006; 14: 452-5. 113: 538-45. [36] Kay MA, Manno CS, Ragni MV, et al. Evidence for gene transfer [57] Hudig D, Haverty T, Fulcher C, Redelman D, Mendelsohn J. and expression of factor IX in haemophilia B patients treated with Inhibition of human natural cytotoxicity by macromolecular an AAV vector. Nat Genet 2000; 24: 257-61. antiproteases. J Immunol 1981; 126: 1569-74. [37] Manno CS, Chew AJ, Hutchison S, et al. AAV-mediated factor IX [58] Matzinger P. The danger model: a renewed sense of self. Science gene transfer to skeletal muscle in patients with severe hemophilia 2002; 296: 301-5. B. Blood 2003; 101: 2963-72. [59] Rodino-Klapac LR, Montgomery CL, Bremer WG, et al. Persistent [38] Mingozzi F, Hasbrouck NC, Basner-Tschakarjan E, et al. expression of FLAG-tagged micro dystrophin in nonhuman Modulation of tolerance to the transgene product in a nonhuman primates following intramuscular and vascular delivery. Mol Ther primate model of AAV-mediated gene transfer to liver. Blood 2010; 18: 109-17. 2007; 110: 2334-41. [60] Su LT, Gopal K, Wang Z, et al. Uniform scale-independent gene [39] Mingozzi F, Liu YL, Dobrzynski E, et al. Induction of immune transfer to striated muscle after transvenular extravasation of tolerance to coagulation factor IX antigen by in vivo hepatic gene vector. Circulation. 2005 ; 112(12):1780-8. transfer. J Clin Invest 2003; 111: 1347-56. [61] Toromanoff A, Cherel Y, Guilbaud M, et al. Safety and efficacy of [40] Mount JD, Herzog RW, Tillson DM, et al. Sustained phenotypic regional intravenous (r.i.) versus intramuscular (i.m.) delivery of correction of hemophilia B dogs with a factor IX null mutation by rAAV1 and rAAV8 to nonhuman primate skeletal muscle. Mol liver-directed gene therapy. Blood 2002; 99: 2670-6. Ther. 2008 Jul;16(7):1291-9. [41] Cao O, Dobrzynski E, Wang L, et al. Induction and role of [62] Haurigot V, Mingozzi F, Buchlis G, et al. Safety of AAV factor IX regulatory CD4+CD25+ T cells in tolerance to the transgene peripheral transvenular gene delivery to muscle in hemophilia B product following hepatic in vivo gene transfer. Blood 2007; 110: dogs. Mol Ther 2010;18: 1318-29. 1132-40. [63] Toromanoff A, Adjali O, Larcher T, et al. Lack of immunotoxicity [42] Nathwani AC, Gray JT, McIntosh J, et al. Safe and efficient after regional intravenous (RI) delivery of rAAV to nonhuman transduction of the liver after peripheral vein infusion of self- primate skeletal muscle. Mol Ther 2010; 18: 151-60. complementary AAV vector results in stable therapeutic expression [64] Powers W, Fan Z, Kocis K, et al. Safety and feasibility of of human FIX in nonhuman primates. Blood 2007; 109: 1414-21. transvenous limb perfusion with saline in human muscular [43] Nathwani AC, Rosales C, McIntosh J, et al. Long-term Safety and dystrophy. Mol Ther 2010; 18 (Suppl 1): S126. Efficacy Following Systemic Administration of a Self- [65] Christine CW, Starr PA, Larson PS, et al. Safety and tolerability of complementary AAV Vector Encoding Human FIX Pseudotyped putaminal AADC gene therapy for Parkinson disease. Neurology With Serotype 5 and 8 Capsid Proteins. Mol Ther 2011. 2009; 73: 1662-9. [44] McCarty DM, Fu H, Monahan PE, Toulson CE, Naik P, Samulski [66] Kaplitt MG, Feigin A, Tang C, et al. Safety and tolerability of gene RJ. Adeno-associated virus terminal repeat (TR) mutant generates therapy with an adeno-associated virus (AAV) borne GAD gene for self-complementary vectors to overcome the rate-limiting step to Parkinson's disease: an open label, phase I trial. Lancet 2007; 369: transduction in vivo. Gene Ther 2003; 10: 2112-8. 2097-105. [45] Nathwani AC, Tuddenham E, Rosales C, et al. Early clinical trial [67] Marks WJ, Jr., Bartus RT, Siffert J, et al. Gene delivery of AAV2- results following administration of a low dose of a novel self neurturin for Parkinson's disease: a double-blind, randomised, complementary adeno-associated viral vector encoding human controlled trial. Lancet Neurol 2010;9: 1164-72. Factor IX in two subjects with severe hemophilia B. Blood 2010; [68] Marks WJ, Jr., Ostrem JL, Verhagen L, et al. Safety and 116: 114 (Abstract #248). tolerability of intraputaminal delivery of CERE-120 (adeno- [46] Vandenberghe LH, Wang L, Somanathan S, et al. Heparin binding associated virus serotype 2-neurturin) to patients with idiopathic directs activation of T cells against adeno-associated virus serotype Parkinson's disease: an open-label, phase I trial. Lancet Neurol 2 capsid. Nat Med 2006; 12: 967-71. 2008; 7: 400-8. [47] Stroes ES, Nierman MC, Meulenberg JJ, et al. Intramuscular [69] McPhee SW, Janson CG, Li C, et al. Immune responses to AAV in administration of AAV1-lipoprotein lipase S447X lowers a phase I study for Canavan disease. J Gene Med 2006; 8: 577-88. triglycerides in lipoprotein lipase-deficient patients. Arterioscler [70] Worgall S, Sondhi D, Hackett NR, et al. Treatment of late infantile Thromb Vasc Biol 2008; 28: 2303-4. neuronal ceroid lipofuscinosis by CNS administration of a serotype [48] Brantly ML, Chulay JD, Wang L, et al. Sustained transgene 2 adeno-associated virus expressing CLN2 cDNA. Hum Gene Ther expression despite T lymphocyte responses in a clinical trial of 2008;19: 463-74. rAAV1-AAT gene therapy. Proc Natl Acad Sci USA 2009; 106: [71] Maguire AM, High KA, Auricchio A, et al. Age-dependent effects 16363-8. of RPE65 gene therapy for Leber's congenital amaurosis: a phase 1 [49] Brantly ML, Spencer LT, Humphries M, et al. Phase I trial of dose-escalation trial. Lancet 2009; 374: 1597-605. intramuscular injection of a recombinant adeno-associated virus [72] Maguire AM, Simonelli F, Pierce EA, et al. Safety and efficacy of serotype 2 alphal-antitrypsin (AAT) vector in AAT-deficient gene transfer for Leber's congenital amaurosis. N Engl J Med 2008; adults. Hum Gene Ther 2006;17: 1177-86. 358: 2240-8. [50] Mendell JR, Campbell K, Rodino-Klapac L, et al. Dystrophin [73] Amado D, Mingozzi F, Hui D, et al. Safety and efficacy of immunity in Duchenne's muscular dystrophy. N Engl J Med 2010b; subretinal readministration of a viral vector in large animals to treat 363: 1429-37. congenital blindness. Sci Transl Med 2010; 2: 21ra16. [51] Mendell JR, Rodino-Klapac LR, Rosales XQ, et al. Sustained [74] Erles K, Sebokova P, Schlehofer JR. Update on the prevalence of alpha-sarcoglycan gene expression after gene transfer in limb- serum antibodies (IgG and IgM) to adeno-associated virus (AAV). girdle muscular dystrophy, type 2D. Ann Neurol 2010; 68: 629-38. J Med Virol 1999; 59: 406-11. [52] Mendell JR, Rodino-Klapac LR, Rosales-Quintero X, et al. Limb- [75] Calcedo R, Vandenberghe LH, Gao G, Lin J, Wilson JM. girdle muscular dystrophy type 2D gene therapy restores alpha- Worldwide epidemiology of neutralizing antibodies to adeno- sarcoglycan and associated proteins. Ann Neurol 2009; 66: 290-7. associated viruses. J Infect Dis. 2009; 199: 381-90. [53] Jaski BE, Jessup ML, Mancini DM, et al. Calcium upregulation by [76] Pien GC, Basner-Tschakarjan E, Hui DJ, et al. Capsid antigen percutaneous administration of gene therapy in cardiac disease presentation flags human hepatocytes for destruction after 330 Current Gene Therapy, 2011, Vol. 11, No. 4 Mingozzi and High

transduction by adeno-associated viral vectors. J Clin Invest 2009; [92] Finn JD, Hui D, Downey HD, et al. Proteasome inhibitors decrease 119: 1688-95. AAV2 capsid derived peptide epitope presentation on MHC class I [77] Nguyen DH, Hurtado-Ziola N, Gagneux P, Varki A. Loss of Siglec following transduction. Mol Ther 2010; 18:135-42. expression on T lymphocytes during human evolution. Proc Natl [93] Azuma H, Paulk N, Ranade A, et al. Robust expansion of human Acad Sci USA 2006 May 16; 103: 7765-70. hepatocytes in Fah(-/-)/Rag2(-/-)/Il2rg(-/-) mice. Nat Biotechnol [78] Chadeuf G, Ciron C, Moullier P, Salvetti A. Evidence for 2007; 25: 903-10. encapsidation of prokaryotic sequences during recombinant adeno- [94] Eberling JL, Jagust WJ, Christine CW, et al. Results from a phase I associated virus production and their in vivo persistence after safety trial of hAADC gene therapy for Parkinson disease. vector delivery. Mol Ther 2005; 12: 744-53. Neurology 2008; 70: 1980-3. [79] Hauck B, Murphy SL, Smith PH, et al. Undetectable transcription [95] Bainbridge JW, Smith AJ, Barker SS, et al. Effect of gene therapy of cap in a clinical AAV vector: implications for preformed capsid on visual function in Leber's congenital amaurosis. N Engl J Med in immune responses. Mol Ther 2009; 17: 144-52. 2008; 358: 2231-9. [80] http://oba.od.nih.gov/rdna_rac/rac_meetings.html. Report on [96] Hauswirth WW, Aleman TS, Kaushal S, et al. Treatment of leber Immune Responses to Adeno-Associated Virus (AAV) Vectors. congenital amaurosis due to RPE65 mutations by ocular subretinal Office of Biotechnology Activities, NIH Recombinant DNA injection of adeno-associated virus gene vector: short-term results Advisory Committee Meeting, June 19, 2007. of a phase I trial. Hum Gene Ther 2008; 19: 979-90. [81] Li C, Goudy K, Hirsch M, et al. Cellular immune response to [97] Gao GP, Alvira MR, Wang L, Calcedo R, Johnston J, Wilson JM. cryptic epitopes during therapeutic gene transfer. Proc Natl Acad Novel adeno-associated viruses from rhesus monkeys as vectors for Sci USA 2009; 106: 10770-4. human gene therapy. Proc Natl Acad Sci USA 2002; 99: 11854-9. [82] Basner-Tschakarjan E, Mingozzi F, Chen Y, et al. Dose-dependent [98] Simioni P, Tormene D, Tognin G, et al. X-linked thrombophilia activation of capsid-specific T cells after AAV serotype 8 vector with a mutant factor IX (factor IX Padua). N Engl J Med 2009; administration in a clinical study for hemophilia. Mol Ther 2011; in 361: 1671-5. press. [99] Finn JD, Favaro P, Wright JF, Mingozzi F, High KA, Arruda VR. [83] Li C, Hirsch M, Asokan A, et al. Adeno-associated virus type 2 Rabbit Anti-Thymocyte Globulin (rATG) Administrated (AAV2) capsid-specific cytotoxic T lymphocytes eliminate only Concomitantly with Liver Delivery of AAV2-hFIX Can Promote vector-transduced cells coexpressing the AAV2 capsid in vivo. J Inhibitor Formation In Rhesus Macaques Blood 2010; 116. Virol 2007; 81: 7540-7. [100] Gaudet D, deWal J, K. T, et al. Review of the clinical development [84] Li H, Murphy SL, Giles-Davis W, et al. Pre-existing AAV capsid- of alipogene tiparvovec gene therapy for lipoprotein lipase specific CD8+ T cells are unable to eliminate AAV-transduced deficiency. Atheroscler Suppl. 2010; 11: 55-60. hepatocytes. Mol Ther 2007; 15: 792-800. [101] Araki K, Turner AP, Shaffer VO, et al. mTOR regulates memory [85] Wang L, Figueredo J, Calcedo R, Lin J, Wilson JM. Cross- CD8 T-cell differentiation. Nature 2009; 460: 108-12. presentation of adeno-associated virus serotype 2 capsids activates [102] Montenegro-Miranda PS, Bloemendaal LT, Kunne C, de Waart cytotoxic T cells but does not render hepatocytes effective cytolytic DR, Bosma PJ. Mycophenolate mofetil impairs transduction of targets. Hum Gene Ther 2007; 18: 185-94. single-stranded adeno-associated viral vectors. Hum Gene Ther [86] Li C, Hirsch M, DiPrimio N, et al. Cytotoxic-T-lymphocyte- 2011; 22: 605-12. mediated elimination of target cells transduced with engineered [103] Zhong L, Li B, Mah CS, et al. Next generation of adeno-associated adeno-associated virus type 2 vector in vivo. J Virol 2009; 83: virus 2 vectors: point mutations in tyrosines lead to high-efficiency 6817-24. transduction at lower doses. Proc Natl Acad Sci USA 2008; 105: [87] Li H, Tuyishime S, Wu TL, et al. Adeno-associated virus vectors 7827-32. serotype 2 induce prolonged proliferation of capsid-specific CD8+ [104] Li M, Jayandharan GR, Li B, et al. High-Efficiency Transduction T cells in mice. Mol Ther 2010; 19: 536-46. of Fibroblasts and Mesenchymal Stem Cells by Tyrosine-Mutant [88] Thomas CE, Storm TA, Huang Z, Kay MA. Rapid uncoating of AAV2 Vectors for Their Potential Use in Cellular Therapy. Hum vector genomes is the key to efficient liver transduction with Gene Ther 2010. pseudotyped adeno-associated virus vectors. J Virol 2004; 78: [105] Hoffman BE, Martino AT, Sack BK, et al. Nonredundant Roles of 3110-22. IL-10 and TGF-beta in Suppression of Immune Responses to [89] Buchlis G, Mingozzi F, Campos Soto P, et al. Intrinsically Hepatic AAV-Factor IX Gene Transfer. Mol Ther 2011. Hyperactive and Hyperproliferative CD8+ T Cells In Cmah-/- Mice [106] Ding WX, Ni HM, Gao W, et al. Linking of autophagy to as a Model of Human Gene Transfer Responses Blood 2010; 116. ubiquitin-proteasome system is important for the regulation of [90] Lang KS, Georgiev P, Recher M, et al. Immunoprivileged status of endoplasmic reticulum stress and cell viability. Am J Pathol 2007; the liver is controlled by Toll-like receptor 3 signaling. J Clin 171: 513-24. Invest 2006; 116: 2456-63. [107] Moss RB, Milla C, Colombo J, et al. Repeated aerosolized AAV- [91] Subbramanian RA, Moriya C, Martin KL, et al. Engineered T-cell CFTR for treatment of cystic fibrosis: a randomized placebo- receptor tetramers bind MHC-peptide complexes with high affinity. controlled phase 2B trial. Hum Gene Ther 2007; 18: 726-32. Nat Biotechnol 2004; 22: 1429-34.

Received: April 29, 2011 Revised: May 04, 2011 Accepted: May 10, 2011

PMID: 21557723