Methods of viral neutralizing antibody epitope mapping

The present application is a divisional of U.S. patent application Ser. No. 15/306,429 titled METHODS OF VIRAL NEUTRALIZING ANTIBODY EPITOPE MAPPING, filed on Oct. 24, 2016, which is a U.S. National Phase patent application under 35. U.S.C. § 371 of International Application No. PCT/US2015/027536, filed Apr. 24, 2015, which claims priority to U.S. Provisional Patent Application No. 61/984,553, filed Apr. 25, 2014, all of which are each expressly incorporated herein by reference.

This application was made with US Government support under grant number RO1DK078388 and NS088399, awarded by the National Institutes of Health. The US Government has certain rights in this application.

The disclosure generally relates to methods of high-throughput mapping of viral neutralizing antibody epitopes. More specifically, the disclosure relates to in vitro immunoprecipitation-based adeno-associated virus Barcode-Seq-based methods of mapping viral neutralizing antibody epitopes.

Viral neutralizing antibody (NtAb) epitope mapping can assist in the development of new vaccines and pharmaceuticals for the prevention and/or treatment of infectious diseases. Additionally, viral NtAb epitope mapping can assist in the development of gene delivery vectors. Identification of and knowledge regarding viral NtAb epitopes may help in the genetic engineering of components of viral vectors that may evade a host immune response, as the host immune response can be a significant obstacle to effective in vivo gene therapy.

Adeno-associated virus (AAV) is a promising in vivo gene delivery vector for gene therapy. Various issues remain to be overcome, however, in the use of AAV as an in vivo gene delivery vector, including the requirement of high vector dose for clinically beneficial outcomes, efficacy-limiting host immune response against viral proteins, promiscuous viral tropism, and the high prevalence of pre-existing anti-AAV neutralizing antibodies in humans. Despite these issues, interest in the use of AAV in gene therapy is growing. A number of naturally occurring serotypes and subtypes have been isolated from human and non-human primate tissues (Gao G et al., J Virol 78, 6381-6388 (2004) and Gao G et al., Proc Natl Acad Sci USA 99, 11854-11859 (2002), both of which are incorporated by reference herein). Among the newly-identified adeno-associated virus isolates, AAV serotype 8 (AAV8) and AAV serotype 9 (AAV9) have gained much attention because recombinant adeno-associated vectors (rAAVs) derived from these two serotypes can transduce various organs including the liver, heart, skeletal muscles, and central nervous system with high efficiency following systemic administration via the periphery (Foust K D et al., Nat Biotechnol 27, 59-65 (2009); Gao et al., 2004, supra; Ghosh A et al., Mol Ther 15, 750-755 (2007); Inagaki K et al., Mol Ther 14, 45-53 (2006); Nakai H et al., J Virol 79, 214-224 (2005); Pacak C A et al., Circ Res 99, e3-e9 (2006); Wang Z et al., Nat Biotechnol 23, 321-328 (2005); and Zhu T et al., Circulation 112, 2650-2659 (2005), all of which are incorporated by reference herein).

The robust transduction by rAAV8 and rAAV9 vectors has been presumed to be ascribed to strong tropism for these cell types, efficient cellular uptake of vectors, and/or rapid uncoating of virion shells in cells (Thomas C E et al., J Virol 78, 3110-3122 (2004), incorporated by reference herein). In addition, emergence of capsid-engineered rAAV with better performance has significantly broadened the utility of rAAV as a vector toolkit (Asokan A et al., Mol Ther 20, 699-708 (2012), incorporated by reference herein). Proof-of-concept for rAAV-mediated gene therapy has been shown in many preclinical animal models of human diseases. Phase I/II clinical studies have been initiated or completed for genetic diseases including hemophilia B (Manno C S et al., Nat Med 12, 342-347 (2006) and Nathwani A C et al., N Engl J Med 365, 2357-2365 (2011), both of which are incorporated by reference herein); muscular dystrophy (Mendell J R et al., N Engl J Med 363, 1429-1437 (2011), incorporated by reference herein); cardiac failure (Jessup M et al., Circulation 124, 304-313 (2011), incorporated by reference herein); blinding retinopathy (Maguire A M et al., Lancet 374, 1597-1605 (2009), incorporated by reference herein); and α1 anti-trypsin deficiency (Flotte T R et al., Hum Gene Ther 22, 1239-1247 (2011), incorporated by reference herein), among others.

Although rAAV vectors have widely been used in preclinical animal studies and have been tested in clinical safety studies, the current rAAV-mediated gene delivery systems remain suboptimal for broader clinical applications. The sequence of an AAV viral capsid protein defines numerous features of a particular AAV vector. For example, the capsid protein affects features such as capsid structure and assembly, interactions with AAV nonstructural proteins such as Rep and AAP proteins, interactions with host body fluids and extracellular matrix, clearance of the virus from the blood, vascular permeability, antigenicity, reactivity to NtAbs, tissue/organ/cell type tropism, efficiency of cell attachment and internalization, intracellular trafficking routes, and virion uncoating rates. Furthermore, the relationship between a given AAV capsid amino acid sequence and the characteristics of the rAAV vector are unpredictable.

High prevalence of pre-existing NtAbs against AAV capsids in humans poses a significant barrier to successful AAV vector-mediated gene therapy. There has been strong enthusiasm about developing “stealth” AAV vectors that can evade NtAbs; however, creation of such AAVs requires more comprehensive information about NtAb epitopes, which currently remains very limited.

DNA-barcoded AAV2R585E hexapeptide (HP) scanning capsid mutant libraries have been produced in which AAV2-derived HPs were replaced with those derived from other serotypes. These libraries have been injected intravenously into mice harboring anti-AAV1 or AAV9 capsid antibodies, which has led to the identification of 452-QSGSAQ-457 (SEQ ID NO:1) in the AAV1 capsid and 453-GSGQN-457 (SEQ ID NO:2) in the AAV9 capsid as epitopes for anti-AAV NtAbs in mouse sera (Adachi K et al., Nat Commun 5, 3075 (2014)). These epitopes correspond to the highest peak of the three-fold symmetry axis protrusion on the capsid. In addition, this region may also function as an epitope for mouse anti-AAV7 NtAbs using the same in vivo approach. A sequencing-based high-throughput approach, termed AAV Barcode-Seq, can allow characterization of phenotypes of hundreds of different AAV strains and can be applied to anti-AAV NtAb epitope mapping.

The embodiments disclosed herein will become more fully apparent from the following description and appended claims, taken in conjunction with the accompanying drawings. These drawings depict only typical embodiments, which will be described with additional specificity and detail through use of the accompanying drawings in which:

FIG. 1 A depicts a map of the DNA-barcoded AAV genome containing a pair of 12 nucleotide-long DNA barcodes (lt-VBC and rt-VBC) downstream of the AAV2 pA. Each virus barcode (VBC) can be PCR-amplified separately.

FIG. 1 B is a representation of double alanine (AA) scanning mutagenesis of the AAV9 capsid.

FIG. 1 C is a representation of hexapeptide (HP) scanning mutagenesis of the AAV2R585E capsid at a two amino acid interval.

FIG. 1 D is a representation of a procedure for AAV Barcode-Seq analysis. PCR products obtained from each sample are indexed with sample-specific barcodes attached to the PCR primers. This allows multiplexed ILLUMINA sequencing. Phenotypic Difference (PD) values provide information about a spectrum of phenotypes (receptor binding, transduction, tropism, blood clearance, reactivity to NtAbs, blood-cerebrospinal fluid barrier (BCSFB) penetrability, etc.) for each serotype or mutant.

FIG. 2 A is a graph showing pharmacokinetic profiles of 117 HP scanning AAV2R585E mutants following intravenous injection of AAV2R585E-HP-VBCLib in anti-AAV1 NtAb-positive mice.

FIG. 2 B is a graph showing pharmacokinetic profiles of 117 HP scanning AAV2R585E mutants following intravenous injection of AAV2R585E-HP-VBCLib in anti-AAV9 NtAb-positive mice.

FIG. 2 C is a graph showing pharmacokinetic profiles of 117 HP scanning AAV2R585E mutants following intravenous injection of AAV2R585E-HP-VBCLib in naïve mice. For each of FIGS. 2 A, 2 B, and 2 C , blood AAV concentrations of each AAV mutant relative to those of AAV2R585E were determined by AAV Barcode-Seq. Only the results of AAV2R585E, 451-16000, and 451-00009 are highlighted with the patterned lines, as indicated. The results of all the other 115 HP mutants are shown with gray lines. The 451-16000 and 451-00009 exhibited significantly accelerated blood clearance only in anti-AAV1 and anti-AAV9 NtAb-harboring animals, respectively. This was not observed in the naïve animals. In the 451-16000 and 451-00009 mutants, the native AAV2R585E sequence 451-PSGTTT-456 (SEQ ID NO:3) located at the 3-fold spike was replaced with QSGSAQ (AAV1) (SEQ ID NO:1) and GSGQN (AAV9) (SEQ ID NO:2), respectively. This indicates that QSGSAQ (SEQ ID NO:1) and GSGQN (SEQ ID NO: 2) are anti-AAV1 and anti-AAV9 capsid antibody epitopes, respectively.

FIG. 3 A is a representation of an RNA barcode-expressing recombinant AAV (rAAV). Clone-specific DNA barcodes (lt-VBC and rt-VBC) are transcribed into RNA under the control of the U6 promoter.

FIG. 3 B is a graph depicting a scatter plot showing a linear correlation between relative DNA and RNA quantities determined by AAV DNA/RNA Barcode-Seq. HEK293 cells were infected with two different AAV libraries containing 25 rAAV2 clones mixed at an equal amount or at approximately a 1:3:10:30:100 ratio, and harvested 48 hours post-infection. Each clone was tagged with a clone-specific barcode. Relative DNA and RNA quantifies of each clone in the same HEK293 cell sample were determined by ILLUMINA barcode sequencing read numbers and plotted.

FIG. 4 depicts AAV2R585E and AAV9 HP scanning mutants included in the DNA/RNA-barcoded dsAAV-U6-VBCLib-1. The amino acid sequences around the highest peak of the AAV capsids derived from AAV2R585E, devoid of HP mutations and AAV2R585E-HP mutants, are aligned to the left and those derived from wild-type AAV9 and AAV9-HP mutants are aligned to the right. Bold letters and hyphens indicate amino acid mutations and deletions compared to the parental sequences, respectively. The name of each mutant is given to the amino acid sequences based on the naming system as follows. The left three digits indicate the first amino acid position of the hexapeptide based on AAV2 VP1 (left panel) and AAV9 VP1 (right panel). The right five digits indicate AAV serotype from which each hexapeptide is derived: 10000, AAV1; 06000, AAV6; 00700, AAV7; 00080, AAV8; 00009, AAV9; and 00002, AAV2. When a hexapeptide amino acid sequence is shared with multiple serotypes, the right five digits have more than one positive integer.

FIG. 5 A is a graph plotting the averages of the relative quantities of two different clones carrying the same HP mutation in an IP precipitate.

FIG. 5 B is a graph plotting the averages of the relative quantities of two different clones carrying the same HP mutation in an IP supernatant. For each of FIGS. 5 A and 5 B , four 8-week-old C57BL/6 male mice (Mouse 1, 2, 3 and 4) were injected intravenously with AAV1-CMV-lacZ vector at a dose of 1×10 11 vector genomes (vg) per mouse. Serum samples containing anti-AAV1 NtAbs were collected 3 weeks post-injection. 20 μl of PROTEIN A/G PLUS-AGAROSE beads were first coated with sample immunoglobulins by incubating the beads with 25 μl of serum samples at 4° C. for one hour, and then reacted with 1×10 9 vg of DNA/RNA-barcoded dsAAV-U6-VBC-Lib-1 at 4° C. overnight. This library contained 72 AAV clones composed of 24 HP mutants and two reference controls (AAV2R585E and wild-type AAV9, 15 clones each). Viral genomic DNAs were extracted from agarose beads-bound and unbound AAV particles in the IP precipitates and the IP supernatants, respectively, and subjected to the AAV Barcode-Seq analysis (id.). The relative quantity of each clone (two clones per mutant) determined by ILLUMINA sequencing read numbers was normalized with the ILLUMINA sequencing read numbers of the reference control AAV2R585E. The Y-axis shows Phenotypic Difference (PD) values (id.) of each mutant relative to the control AAV2R585E in antibody-positive sera, normalized with PD values obtained with naïve mouse serum. Plotted are the averages of the relative quantities of two different clones carrying the same HP mutation. Arrows indicate mutants harboring the heterologous peptides that bind to anti-AAV1 antibody, and therefore, represent anti-AAV1 antibody epitopes.

FIG. 6 is a graph wherein the Y-axis shows Phenotypic Difference (PD) values (id.) of each mutant relative to the control wild-type AAV9. An arrow indicates a mutant harboring the heterologous peptide that binds to anti-AAV2 antibody, and therefore, represents an anti-AAV2 antibody epitope. Four 8-week-old C57BL/6 male mice (Mouse 1, 2, 3 and 4) were injected intravenously with AAV2-CMV-lacZ vector at a dose of 1×10 11 vg/mouse. Serum samples containing anti-AAV2 NtAbs were collected three weeks post-injection. The subsequent experimental procedure is the same as that for FIGS. 5 A and 5 B .

FIG. 7 is a graph wherein the Y-axis shows Phenotypic Difference (PD) values (id.) of each mutant relative to the control AAV2R585E. An arrow indicates a mutant harboring the heterologous peptide that binds to anti-AAV7 antibody, and therefore, represents an anti-AAV7 antibody epitope. Three 8-week-old C57BL/6 male mice (Mouse 1, 2, and 3) were injected intravenously with AAV7-CMV-lacZ vector at a dose of 1×10 11 vg/mouse. Serum samples containing anti-AAV7 NtAbs were collected three weeks post-injection. The subsequent experimental procedure is the same as that for FIGS. 5 A and 5 B .

FIG. 8 is a graph wherein the Y-axis shows Phenotypic Difference (PD) values (id.) of each mutant relative to the control AAV2R585E. An arrow indicates a mutant harboring the heterologous peptide that binds to anti-AAV9 antibody, and therefore, represents an anti-AAV9 antibody epitope. Four 8-week-old C57BL/6 male mice (Mouse 1, 2, 3 and 4) were injected intravenously with AAV9-CMV-lacZ vector at a dose of 1×10 11 vg/mouse. Serum samples containing anti-AAV9 NtAbs were collected three weeks post-injection. The subsequent experimental procedure is the same as that for FIGS. 5 A and 5 B .

FIG. 9 is a graph depicting anti-AAV2 capsid mouse monoclonal antibody (A20) epitope identification by the magnetic beads-based IP-Seq analysis in conjunction with the dsAAV9-HP-U6-VBCLib-2 library. 1×10 9 vg of the dsAAV9-HP-U6-VBCLib-2 library was reacted with Pierce Protein A/G Magnetic Beads coated with the A20 antibody, and subjected to immunoprecipitation. Subsequently, viral genomic DNAs were extracted from magnetic beads-bound and unbound AAV particles in the IP precipitates (A20_IP) and the IP supernatants (A20_Sup), respectively, and a demographic change of the library composition in the IP precipitates (A20_IP) and the IP supernatants (A20_Sup) was determined by the AAV Barcode-Seq analysis. The relative quantity of each clone (2 clones per mutant) determined by Illumina barcode sequencing read numbers was normalized with the Illumina sequencing read numbers of the reference control AAV9. The Y-axis shows Phenotypic Difference (PD) values of each mutant relative to the control AAV9. The X-axis shows different AAV strains. They are AAV9, AAV2R585E, AAV2, 009-00002, . . . , 718-00002 from the left to the right. Due to space limitations, not all the AAV2 strains are labeled. Plotted are the averages of the relative quantities of two different clones carrying the same HP mutation. An arrow indicates the mutant 261-00002 harboring a heterologous AAV2-derived peptide that binds to the A20 antibody (261-SSQSGA-266 (SEQ ID NO:50)). Another peak to the leftmost represents AAV2R585E and AAV2.

FIGS. 10 A, 10 B, 10 C, and 10 D are graphs depicting anti-AAV2 mouse polyclonal antibody epitope identification by the magnetic beads-based IP-Seq analysis in conjunction with the dsAAV9-HP-U6-VBCLib-2 library. Four 8-week-old C57BL/6 male mice (Mouse 1, 2, 3 and 4 in FIGS. 10 A, 10 B, 10 C, and 10 D , respectively) were injected intravenously with AAV2-CMV-lacZ vector at a dose of 1×10 11 vg/mouse. Serum samples containing anti-AAV2 neutralizing antibodies were collected 3 weeks post-injection and subjected to the IP-Seq analysis for epitope mapping. Demographic change of the library composition in the IP precipitates (Mousel, 2, 3 and 4_IP) and the IP supernatants (Mousel, 2, 3, and 4_Sup) were determined by the AAV Barcode-Seq analysis. The relative quantity of each clone (2 clones per mutant) determined by IIlumina barcode sequencing read numbers was normalized with the Illumina sequencing read numbers of the reference control AAV9. The Y-axis shows Phenotypic Difference (PD) values of each mutant relative to the control AAV9. The X-axis shows different AAV strains. They are AAV9, AAV2R585E, AAV2, 009-00002, . . ., 718-00002 from the left to the right. Due to space limitations, not all the AAV2 strains are labeled. Plotted are the averages of the relative quantities of two different clones carrying the same HP mutation. Gray arrows, in FIGS. 10 A- 10 D , indicate the two hexapeptide mutants containing the dominant epitope, 513-RDSLVNPG-520 (SEQ ID NO:52). The thick black arrow, in FIG. 10 D , indicates the same epitope identified for the A20 antibody. Thin black arrows, in FIGS. 10 A- 10 C , also indicate mutants that may contain epitopes. The peak to the leftmost represents AAV2R585E and AAV2.

It will be readily understood that the embodiments, as generally described herein, are exemplary. The following more detailed description of various embodiments is not intended to limit the scope of the present disclosure, but is merely representative of various embodiments. Moreover, the order of steps or actions of the methods disclosed herein may be changed by those skilled in the art without departing from the scope of the present disclosure. In other words, unless a specific order of steps or actions is required for proper operation of the embodiment, the order or use of specific steps or actions may be modified.

The term “viral vector” as used herein means any vector that comprises or derives from components of a given virus and is suitable to infect mammalian cells, including human cells, of any of a number of tissue types, such as brain, heart, lung, skeletal muscle, liver, kidney, spleen, or pancreas, whether in vitro or in vivo. The term “viral vector” may be used to refer to a viral particle (or virion) comprising at least a nucleic acid molecule encoding a protein of interest.

The term “AAV vector” as used herein means any vector that comprises or derives from components of AAV and is suitable to infect mammalian cells, including human cells, of any of a number of tissue types, such as brain, heart, lung, skeletal muscle, liver, kidney, spleen, or pancreas, whether in vitro or in vivo. The term “AAV vector” may be used to refer to an AAV type viral particle (or virion) comprising at least a nucleic acid molecule encoding a protein of interest.

Additionally, the AAVs disclosed herein may be derived from various serotypes, including combinations of serotypes (e.g., “pseudotyped” AAV) or from various genomes (e.g., single-stranded or self-complementary). In particular embodiments, the AAV vectors disclosed herein may comprise desired proteins or protein variants. A “variant” as used herein, refers to an amino acid sequence that is altered by one or more amino acids. The variant may have “conservative” changes, wherein a substituted amino acid has similar structural or chemical properties, e.g., replacement of leucine with isoleucine. More rarely, a variant may have “nonconservative” changes, e.g., replacement of a glycine with a tryptophan. Analogous minor variations may also include amino acid deletions or insertions, or both.

Methods of producing AAV vectors as disclosed herein are well known in the art, including methods, for example, using packaging cells, auxiliary viruses or plasm ids, and/or baculovirus systems (see, e.g., Samulski et al., J Virol 63, 3822 (1989); Xiao et al., J Virol 72, 2224 (1998); Inoue et al., J Virol 72, 7024 (1998); W01998/022607; and W02005/072364).

Methods of producing pseudotyped AAV vectors are also known (see, e.g., W000/28004), as well as various modifications or formulations of AAV vectors, to reduce their immunogenicity upon in vivo administration (see, e.g., W001/23001; W000/73316; W004/112727; W005/005610; and W099/06562). In some embodiments, AAV vectors may be prepared or derived from various serotypes of AAVs which may be mixed together or mixed with other types of viruses to produce chimeric (e.g., pseudotyped) AAV viruses.

In some embodiments, a method of high-throughput mapping of viral NtAb conformational epitopes can be utilized, which may comprise HP scanning of mutant viral libraries, immunoprecipitation (IP), and/or next-generation sequencing (NGS) technology.

As discussed, NtAb epitope mapping can be used in the development of new vaccines and drugs for the prevention and treatment of infectious diseases. NtAb epitope mapping can also be used for the development of novel gene delivery vectors. Identification of and knowledge regarding NtAb epitopes may help in the genetic engineering of viral components of novel vectors that can evade, or better evade, the host immune response, as the host immune response can be a significant obstacle in effective in vivo gene therapy. In various embodiments, the present disclosure can overcome the limitations of currently available methods for NtAb epitope mapping (e.g., X-ray co-crystallography, cryoelectron microscopy, synthetic peptide array, phage display, etc.) and may effectively identify conformational epitopes of viral antigens in a high-throughput manner.

NtAbs may recognize mostly conformational epitopes as opposed to linear epitopes. X-ray diffraction and cryoelectron microscopic analyses of co-crystallized antigen-antibody complexes can identify conformational NtAb epitopes. However, it may be technically challenging to apply this method to conformational epitope mapping of multiple samples due to its low-throughput, time-intensive, and cost-consuming nature. Synthetic peptide array-based methods can adopt a high-throughput format but they may primarily identify linear epitopes and may be inefficient in identifying conformational epitopes. Phage display approaches have some ability to identify conformational epitopes but may not necessarily be able to present antigens in their native three dimensional protein structures. In contrast, the methods disclosed herein can express epitopes in the context of native tertiary and quaternary structures of viral proteins and identify conformational epitopes of multiple samples at one time in a high-throughput manner.

In various embodiments, the methods disclosed herein can use DNA/RNA-barcoded HP scanning libraries in which platform viral proteins that are not neutralized by NtAbs of interest carry HPs derived from homologous viral proteins that are neutralized by the NtAbs of interest for which one may wish to identify conformational epitopes. HPs can be expressed in the platform viral proteins with appropriately juxtaposed amino acids in native-like tertiary and quaternary structures. Examples of HP scanning libraries include, but are not limited to, the AAV2R585E-HP and AAV9-HP libraries.

Conventional epitope mapping methods such as peptide array- or phage display-based systems can map epitopes of one antigen or one antibody at a time. The methods of the present disclosure can make it possible to map epitopes of multiple antibodies against multiple antigens at one time by an NGS-based method such as, but not limited to, multiplexed ILLUMINA sequencing technology.

In certain embodiments, the present method may comprise the following steps: 1) incubation of samples and a HP scanning library; 2) IP; 3) viral DNA extraction; 4) PCR amplification of viral DNA barcodes; 5) multiplexed ILLUMINA barcode sequencing; and/or 6) data analysis.

An in vitro IP-based AAV Barcode-Seq method can be used to identify anti-AAV antibody epitopes as opposed to the previously reported in vivo approach described above. Also, the highest peak around the amino acid positions 452-457 in the capsid may be a potential dominant epitope for antibodies against various AAV serotypes.

In some embodiments, IP conditions may be optimized using a recombinant AAV9 vector, anti-AAV9 mouse sera, and protein A/G agarose beads. Additionally, a DNA-barcoded AAV HP scanning capsid mutant library comprising 78 AAV clones can be produced, which comprises nineteen AAV2R585E-derived HP mutants and five AAV9-derived HP mutants that span a 14-amino-acid region in and flanking the vicinity of the highest peak of the AAV1, AAV2, AAV7, AAV8, and AAV9 capsids. Using this library, and anti-AAV1, anti-AAV2, anti-AAV7, anti-AAV8, anti-AAV9, and naïve mouse sera collected from four mice per serotype/native, IP can be performed and AAV library genomes from both immunoprecipitates and supernatants may be recovered, which may then be subjected to the AAV Barcode-Seq analysis. Mutants harboring AAV1-, AAV2-, AAV7-, AAV8-, and AAV9-derived peptides spanning eight amino acids within the vicinity of the highest peak, but not other mutants or the control AAV2R585E or AAV9, can be found to be captured by the corresponding anti-AAV serotype sera of some, if not all, of the immunized mice, indicating that the identified peptides constitute a dominant epitope. Thus, a combination of AAV Barcode-Seq with IP can map anti-AAV antibody epitopes in a high-throughput manner.

In some embodiments, a method of identifying one or more dominant epitopes in a viral vector may comprise contacting a mutant capsid of a virus with serum from a subject previously exposed to the virus and immunoprecipitating serum immunoglobulins from the serum. In various embodiments, the viral vector may be an AAV vector. In certain embodiments, the mutant capsid of the virus may be an AAV mutant capsid. In some embodiments, the mutant capsid of the virus of the disclosed method may be included in a mutant capsid library, wherein the capsids in the mutant capsid library are barcoded.

In some embodiments, an AAV1 viral vector may comprise an antibody neutralizing mutation in amino acids 452-457 in the AAV1 capsid (452-QSGSAQ-457) (SEQ ID NO:1). In other embodiments, an AAV9 vector may comprise an antibody neutralizing mutation in amino acids 453-457 in the AAV9 capsid (453-GSGQN-457) (SEQ ID NO: 2). In certain embodiments, the antibody neutralizing mutation of either the AAV1 viral vector or the AAV9 viral vector may comprise a mutation to an alanine.

A dsAAV9-HP-U6-VBCLib-2 HP-scanning mutant capsid library containing a total of 338 AAV clones has been created. These clones are composed of 153 AAV9-HP mutants, AAV2, AAV2R585E and AAV9. The IP-Seq (Immunoprecipitation followed by AAV Barcode-Seq) has been optimized using Protein A/G magnetic beads. An epitope in the AAV2 capsid that is recognized by the mouse monoclonal antibody against intact AAV2 particles (A20) has been mapped by IP-Seq. Epitopes in the AAV2 capsid have been mapped that are recognized by the mouse polyclonal antibodies developed in mice immunized by intravenous injection of an AAV2 vector. Strategies for the creation of anti-AAV neutralizing antibody-escaping AAV capsid mutants have been developed based on the new IP-Seq data.

AAV Barcode-Seq

AAV Barcode-Seq, an NGS-based method that allows the characterization of phenotypes of hundreds of different AAV strains (i.e., naturally occurring serotypes and laboratory-engineered mutants) in a high-throughput manner with significantly reduced time and effort and using only a small number of subjects (e.g., tissue cultures and experimental animals), has recently been established (Adachi K et al., Nat Commun 5, 3075 (2014)). Using this approach, biological aspects including, but not limited to, blood clearance rate, transduction efficiency, tissue tropism, and reactivity to anti-AAV NtAbs can be assessed. FIGS. 1 A- 1 D schematically depict the AAV Barcode-Seq approach. The principle of this approach is as follows. When a library stock comprising many different AAV strains is applied to certain types of samples (e.g., cells), the composition of the AAV population would in theory not change between the original input library and the library recovered from the samples if each of the AAV strains had exactly the same biological properties in a given context. However, if some strains show a different biological property (e.g., faster blood clearance or more efficient cellular internalization) compared to the others, there would be a change in the population composition between the input library (i.e., the library stock) and the output library (i.e., the library recovered from the samples). The basic method consists of a bioinformatic comparison between the input and output libraries using a similar principle as that employed in RNA-Seq (Wang Z et al., Nat Rev Genet 10, 57-63 (2009)). This method allows the quantification of phenotypic differences between different AAV strains as a function of strain demographics. Such an analysis becomes possible by tagging each AAV strain with a unique short DNA barcode and applying ILLUMINA barcode sequencing to the resulting population (Smith AM et al., Genome Res 19, 1836-1842 (2009)).

In Vivo-Based Anti-AAV NtAb Epitope Mapping Using AAV2R585E Hexapeptide Scanning Libraries

Construction has been completed of a total of 452 hexapeptide (HP) scanning AAV2R585E capsid mutants that carry all the AAV1-, AAV6-, AAV7-, AAV8-, and AAV9-specific HPs that are not present in the AAV2 capsid (see Table 1). In Adachi K et al., Nat Commun 5, 3075 (2014), AAV2R585E-HP-VBCLib-1 and 2 libraries were produced containing a total of 117 capsid-forming HP mutants, the libraries were injected intravenously into anti-AAV1 or AAV9 NtAb-harboring C57BL/6 mice (n=3) or naive mice (n=2) at 1×10 13 vg/kg, and relative blood concentrations of each mutant were determined at one, 10, 30 and 60 minutes post-injection by AAV Barcode-Seq. Because there is no or minimal serologic cross-reactivity between AAV2 and any of AAV1, AAV6, AAV7, AAV8, and AAV9 (Gao G et al., J Virol 78, 6381-6388 (2004)), only AAV2R585E mutants with a HP containing an antibody epitope would be neutralized, and therefore would be cleared faster than other mutants in the same immunized animal or faster than the same mutant in naïve animals. By taking this approach, 452-QSGSAQ-457 (SEQ ID NO:1) and 453-GSGQN-457 (SEQ ID NO:2) were identified as epitopes for mouse anti-AAV1 and AAV9 NtAbs developed by viral immunization (see FIGS. 2 A- 2 C ). Taking the same approach, 449-RTQSNPGGTAG-459 (SEQ ID NO:4) was identified as a mouse anti-AAV7 NtAb epitope. These observations establish that injection of AAV2R585E-HP scanning libraries into anti-AAV NtAb positive mice combined with AAV Barcode-Seq can identify anti-AAV NtAb epitopes effectively.

Establishment of a Universal AAV DNA/RNA Barcode-Seq System Expressing RNA Barcodes

A universal Barcode-Seq system expressing RNA barcodes, termed AAV DNA/RNA Barcode-Seq, has been devised. In this system, AAV libraries are produced in which each viral particle contains a DNA genome that is devoid of the rep and cap genes but is transcribed into an RNA barcode unique to its own capsid. To show proof-of-principle of this new method, two libraries of 25 recombinant AAV2 viral clones mixed at defined ratios were constructed, HEK293 cells were infected with each library in duplicate, and the cells were harvested at 48 hours post-infection. In these libraries, each viral clone carried the dsAAV-U6-VBCLib genome expressing RNA containing a pair of clone-specific 12 ribonucleotides transcribed from the corresponding DNA barcode sequences placed downstream of the human U6 snRNA promoter (see FIG. 3 A ). ILLUMINA sequencing of DNA-PCR and reverse-transcription (RT)-PCR barcode amplicons from total DNA and RNA extracted from the same library-infected cells showed that viral genome expression could be determined by Barcode-Seq in at least a 2-log dynamic range with a Pearson's correlation coefficient of 0.98 (see FIG. 3 B ). This RNA barcode system, AAV DNA/RNA Barcode-Seq, has been employed for anti-AAV NtAb epitope mapping.

AAV Libraries for Anti-AAV NtAb Epitope Mapping Created Based on the Universal AAV DNA/RNA Barcode-Seq System

In this new system, DNA/RNA-barcoded dsAAV-U6-VBCLib libraries packaged with HP scanning mutants can be produced. Such HP mutants can be AAV2R585E-HP scanning mutants for anti-AAVx NtAb epitope mapping (x=any strains other than AAV2 that do not cross-react with anti-AAV2 NtAb) and AAV9-HP scanning mutants for anti-AAV2 NtAb epitope mapping. The structure of AAV2R585E-HP mutants is shown in FIG. 1 C . AAV9-HP mutants are those in which AAV9 HPs are replaced with those derived from the AAV2 capsid. All of them are HP scanning libraries that contain the dsAAV-U6-VBCLib genome (see FIG. 3 A ). DNA/RNA-barcoded dsAAV2R585E-HP-U6-VBCLib libraries for anti-AAV1, anti-AAV6, anti-AAV7, anti-AAV8, and anti-AAV9 NtAb epitope mappings and DNA/RNA-barcoded dsAAV9-HP-U6-VBCLib libraries for anti-AAV2 NtAb epitope mapping can be created. The former set of libraries can comprise a total of 452 AAV2R585E-HP scanning mutants, and the latter set of libraries can comprise 153 AAV9-HP mutants. These mutants, in theory, can cover all the potential hexapeptide epitopes of AAV1, AAV2, AAV6, AAV7, AAV8, and AAV9. Each library will contain two clones per mutant plus 15 clones each of the reference controls, AAV9 and AAV2R585E.

Immunoprecipitation (IP)-Seq Based Anti-AAV Antibody Epitope Mapping

The IP-Seq based method does not require animals and is capable of mapping antibody epitopes of multiple samples at one time using multiplexed ILLUMINA sequencing. Differentiation between NtAb epitopes and non-NtAb epitopes may be achieved by integrating an AAV RNA Barcode-Seq-based neutralization antibody assay into a system as detailed below in the section “AAV RNA Barcode-Seq-based analysis of the neutralizing ability of anti-AAV antibodies with defined epitopes.”

The procedure for IP-Seq based anti-AAV antibody epitope mapping can be as follows. First, 25 μl of serum samples (containing anti-AAV NtAbs) and 20 μl of PROTEIN A/G PLUS-AGAROSE (SANTA CRUZ sc-2003) can be incubated in a total volume of 100 μl in PBS in 1.5 ml tubes at 4° C. for 1 hour on a rotation device. After washing with PBS, a DNA/RNA-barcoded dsAAV-U6-VBCLib library and the agarose beads coated with immunoglobulins can be mixed in a total volume of 100 μl PBS, and may then be incubated at 4° C. overnight on a rotation device. On the next day, a standard IP procedure may be followed, the supernatants and immunoprecipitates can be collected, and viral genome DNA can be extracted using a WAKO DNA Extraction Kit following Proteinase K treatment of the samples. The subsequent procedure may be similar to that used for AAV Barcode-Seq as described in Adachi K et al., Nat Commun 5, 3075 (2014). Briefly, left and right viral clone-specific barcodes (lt-VBC and rt-VBC in FIGS. 1 A- 1 D ) may be PCR-amplified using viral genome DNA recovered from the IP supernatants and precipitates. The PCR primers can be indexed with sample-specific DNA barcodes. All the PCR amplicons may then be mixed into a pool and the pool may be subjected to ILLUMINA sequencing. The ILLUMINA sequencing data may be bioinformatically analyzed to detect demographic changes of the AAV library in each sample. The principle of the method is that viral clones with higher avidity to sample immunoglobulins than others can be detected as clones that are decreased or depleted in the supernatants while enriched in the precipitates by ILLUMINA barcode sequencing. Such clones may likely carry epitopes for anti-AAV antibodies under investigation, and the epitopes targeted by the antibodies may likely be the heterologous peptides incorporated into the capsid of particular AAV clones showing a demographic change. 1×10 7 , 1×10 8 , and 1×10 9 vg per 1.5 ml tube have been used and it has been found that this range of the virus quantity may give clear results as described below.

To show proof-of-principle, a DNA/RNA-barcoded dsAAV-U6-VBCLib-1 library exhibiting low diversity was produced. This library was designed to identify anti-AAV1, AAV2, AAV6, AAV7, AAV8, and AAV9 antibody epitopes at the highest peak around the amino acid positions 452-457 in the capsid. Based on data obtained in the in vivo-based epitope mapping study as described above, it was hypothesized that this region is a potential dominant epitope for antibodies against various AAV serotypes. Therefore, it was assumed that targeting this region in the proof-of-principle experiments may have a higher success rate in finding anti-AAV antibody epitopes for various AAV serotypes. This library was composed of 78 AAV clones, which included 19 AAV2R585E-derived HP mutants and five AAV9-derived HP mutants that spanned a 14-amino-acid region in and flanking the vicinity of the highest peak of the AAV1, AAV2, AAV7, AAV8, and AAV9 capsids (see FIG. 4 ) in addition to 15 clones each of reference control AAV strains, AAV9, and AAV2R585E devoid of HP mutations. Using this library and anti-AAV1, anti-AAV2, anti-AAV7, anti-AAV8, anti-AAV9 mouse sera collected from 3-4 mice per serotype, the IP-Seq-based epitope mapping procedure described above was performed. The mice from which anti-AAV sera was collected had been immunized by intravenous injection of 1×10 11 vg of AAV-CMV-lacZ vector packaged with the corresponding serotype capsids. The data presented below were obtained when 1×10 9 vg per tube was used for IP-Seq. The IP-Seq procedure was also performed using native mouse sera to control nonspecific binding of the AAV library to the immunoglobulin-coated agarose beads.

FIGS. 5 A and 5 B show the results of IP-Seq using anti-AAV1 mouse sera collected from four mice. Mouse 1 exhibited >10 fold enrichment of 451-16000 and 453-16000 in the IP precipitations, and Mouse 3 also showed >10 fold enrichment of 453-16000 in the IP precipitation. Consistent with this observation, reduction of 451-16000 and 453-16000 in the supernatant, although not dramatic, was found in the Mouse 1 serum. Mouse 2 and Mouse 4 exhibited a lesser degree of enrichment of 453-16000 in the precipitation. The reduction of these mutants in the supernatants was not detected in the Mouse 2, 3, and 4 sera. These observations indicate that, in Mouse 1 and 3 and perhaps Mouse 2 and 4, an anti-AAV1 antibody epitope resides within the heterologous peptides contained in 451-16000 and 453-16000 (i.e., 452-QSGSAQNK-459 (SEQ ID NO:5)). The fact that the heterologous peptides in these mutants are derived from the AAV1 capsid may also lend support to this conclusion. In addition, this conforms to the result obtained by the in vivo-based epitope mapping by AAV Barcode-Seq (Adachi K et al., Nat Commun 5, 3075 (2014)). The sensitivity of IP-Seq may be increased by decreasing the amount of AAV library added to the IP reaction. For example, the use of 1×10 7 vg or 1×10 8 vg per tube may identify epitopes more effectively and clearly in both the IP supernatants and the IP precipitations. A preliminary experiment using 1×10 7 vg and 1×10 8 vg per tube of an AAV library has supported this prediction.

FIG. 6 shows the results of IP-Seq using anti-AAV2 mouse sera collected from four mice. All the AAV2R585E mutants were found enriched in the IP precipitations. In Mouse 2, 453-00002 was enriched by >10 fold in the precipitates, demonstrating that the AAV2-derived heterologous peptide in this mutant, 451-PSGTTT-456 (SEQ ID NO:3), may be an epitope for anti-AAV2 antibodies developed in Mouse 2.

FIG. 7 shows the results of IP-Seq using anti-AAV7 mouse sera collected from three mice. 451-00700 was significantly enriched in the IP precipitates in Mouse 1 and 2. This demonstrates that AAV7-derived heterologous peptide in the 451-00700 mutant, 453-NPGGTAG-459 (SEQ ID NO:6), may be an epitope for anti-AAV7 antibodies developed in Mouse 1 and 2.

FIG. 8 shows the results of IP-Seq using anti-AAV9 mouse sera collected from four mice. 451-00009 was significantly enriched in the IP precipitates in Mouse 1 and 3. This demonstrates that AAV9-derived heterologous peptide in the 451-00009 mutant, 453-GSGQN-457 (SEQ ID NO:2), may be an epitope for anti-AAV9 antibodies developed in Mouse 1 and 3. This also conforms to the result obtained by the in vivo-based epitope mapping by AAV Barcode-Seq (Adachi K et al., Nat Commun 5, 3075 (2014)). Regarding the IP-Seq analysis of anti-AAV8 mouse sera, no epitopes could be detected at the sensitivity of the assay used.

In summary, a series of proof-of-concept experiments demonstrates that the IP-Seq using AAV capsid hexapeptide scanning libraries is a means to map anti-AAV antibody epitopes, presumably including conformational epitopes, effectively and in a high-throughput manner. Although the AAV library used for this preliminary set of experiments contained only 24 hexapeptide (HP) scanning mutants, 452 AAV2R585E-HP mutants were created to look for anti-AAV1, anti-AAV6, anti-AAV7, anti-AAV8, and anti-AAV9 antibody epitopes. A total of 153 AAV9-HP mutants to cover the entire region of AAV2 VP1 capsid protein can also be created. Additionally, the same approach can be exploited for epitope mapping of antibodies against other AAV serotypes or capsid-engineered mutants. This method should also have a potential to be adapted to antiviral antibody epitope mapping for any viruses other than AAV.

AAV RNA Barcode-Seq-Based Analysis of the Neutralizing Ability of Anti-AAV Antibodies with Defined Epitopes

AAV DNA/RNA Barcode-Seq may be used to assess the neutralizing ability of anti-AAV antibodies that recognize defined epitopes. The principle of this new assay system is as follows. A DNA/RNA-barcoded dsAAV-U6-VBCLib library that has been prepared for IP-Seq can be pre-incubated with samples under investigation (serum samples, purified monoclonal/polyclonal antibodies, etc.) at 37° C. for one hour or pre-incubated with a naive animal serum devoid of anti-AAV NtAbs (e.g., naïve mouse serum) as a control. The mixture can then be applied to a reporter cell line in vitro in duplicate or in triplicate. Two to three days after AAV library infection, total RNA can be recovered from cells and reverse-transcribed using an AAV genome-specific RT primer. Then clone-specific viral RNA barcodes may be PCR-amplified and subjected to AAV Barcode-Seq (i.e., multiplexed ILLUMINA barcode sequencing followed by data analysis). When the ILLUMINA sequencing data are compared between the samples and the control, AAV clones that are neutralized by anti-AAV antibodies can be identified as a relative decrease of ILLUMINA sequencing reads among all the AAV clones in the library used for the analysis. By combining the heterologous peptide information of each AAV clone and the AAV Barcode-Seq results, it may be determinable whether or not an anti-AAV antibody, or anti-AAV antibodies, that recognizes a defined antibody epitope on the capsid can neutralize the virus and impair the virus infectivity. This assay complements IP-Seq because IP-Seq by itself may not be able to differentiate NtAb epitopes from non-neutralizing antibody epitopes, although the in vivo-based epitope mapping approach is capable of differentiation between these two types of anti-AAV antibodies. The reporter cells should be selected carefully because in vitro transduction efficiencies significantly vary depending on cell types and AAV strains. For example, HEK293 cells can be appropriate for anti-AAV2 antibody epitopes and Chinese Hamster Ovary (CHO) Lec2 cells can be appropriate for anti-AAV9 antibody epitopes.

The following examples are illustrative of disclosed methods. In light of this disclosure, those of skill in the art will recognize that variations of these examples and other examples of the disclosed method would be possible without undue experimentation.

Example 1—Development of an In Vitro IP-Based AAV Barcode-Seq Method to Identify Anti-AAV Antibody Epitopes

IP conditions were optimized using a recombinant AAV9 vector, anti-AAV9 mouse sera, and protein A/G agarose beads. Then, a DNA-barcoded AAV HP scanning capsid mutant library comprising 78 AAV clones was produced, which included nineteen AAV2R585E-derived HP mutants and five AAV9-derived HP mutants that spanned a 14-amino-acid region in and flanking the vicinity of the highest peak of the AAV1, AAV2, AAV7, AAV8, and AAV9 capsids. Using this library and anti-AAV1, anti-AAV2, anti-AAV7, anti-AAV8, anti-AAV9, and naive mouse sera collected from four mice per serotype/native, IP was performed and AAV library genomes were recovered from both immunoprecipitates and supernatants, which were then subjected to the AAV Barcode-Seq analysis. As a result, it was found that mutants harboring AAV1-, AAV2-, AAV7-, AAV8-, and AAV9-derived peptides spanning eight amino acids within the vicinity of the highest peak, but not other mutants or the control AAV2R585E or AAV9, were clearly captured by the corresponding anti-AAV serotype sera of some, if not all, of the immunized mice, indicating that the identified peptides constitute a dominant epitope. Thus, these results demonstrated that a combination of AAV Barcode-Seq with IP can be a substantially easy and effective approach to map anti-AAV antibody epitopes in a high-throughput manner.