Capsid Variants Research Articles

A Rationally Engineered Capsid Variant of AAV9 for Systemic CNS-Directed and Peripheral Tissue-Detargeted Gene Delivery in Neonates

Adeno-associated virus (AAV) has provided the gene therapy field with the most powerful in vivo gene delivery vector to realize safe, efficacious, and sustainable therapeutic gene expression. Because many clinically relevant properties of AAV-based vectors are governed by the capsid, much research effort has been devoted to the development of AAV capsids for desired features. Here, we combine AAV capsid discovery from nature and rational engineering to report an AAV9 capsid variant, designated as AAV9.HR, which retains AAV9’s capability to traverse the blood-brain barrier and transduce neurons. This variant shows reduced transduction in peripheral tissues when delivered through intravascular (IV) injection into neonatal mice. Therefore, when IV AAV delivery is used to treat CNS diseases, AAV9.HR has the advantage of mitigating potential off-target effects in peripheral tissues compared to AAV9. We also demonstrate that AAV9.HR is suitable for peripheral tissue-detargeted CNS-directed gene therapy in a mouse model of a fatal pediatric leukodystrophy. In light of recent success with profiling diversified natural AAV capsid repertoires and the understanding of AAV capsid sequence-structure-function relationship, such a combinatory approach to AAV capsid development is expected to further improve vector targeting and expand the vector toolbox for therapeutic gene delivery.

The Three-Fold Axis of the HIV-1 Capsid Lattice Is the Species-Specific Binding Interface for TRIM5α.

ABSTRACTRhesus TRIM5α (rhTRIM5α) potently restricts replication of human immunodeficiency virus type 1 (HIV-1). Restriction is mediated through direct binding of the C-terminal B30.2 domain of TRIM5α to the assembled HIV-1 capsid core. This host-pathogen interaction involves multiple capsid molecules within the hexagonal HIV-1 capsid lattice. However, the molecular details of this interaction and the precise site at which the B30.2 domain binds remain largely unknown. The human orthologue of TRIM5α (hsTRIM5α) fails to block infection by HIV-1 both in vivo and in vitro. This is thought to be due to differences in binding to the capsid lattice. To map the species-specific binding surface on the HIV-1 capsid lattice, we used microscale thermophoresis and dual-focus fluorescence correlation spectroscopy to measure binding affinity of rhesus and human TRIM5α B30.2 domains to a series of HIV-1 capsid variants that mimic distinct capsid arrangements at each of the symmetry axes of the HIV-1 capsid lattice. These surrogates include previously characterized capsid oligomers, as well as a novel chemically cross-linked capsid trimer that contains cysteine substitutions near the 3-fold axis of symmetry. The results demonstrate that TRIM5α binding involves multiple capsid molecules along the 2-fold and 3-fold interfaces between hexamers and indicate that the binding interface at the 3-fold axis contributes to the well-established differences in restriction potency between TRIM5α orthologues.IMPORTANCE TRIM5α is a cellular protein that fends off infection by retroviruses through binding to the viruses' protein shell surrounding its genetic material. This shell is composed of several hundred capsid proteins arranged in a honeycomb-like hexagonal pattern that is conserved across retroviruses. By binding to the complex lattice formed by multiple capsid proteins, rather than to a single capsid monomer, TRIM5α restriction activity persists despite the high mutation rate in retroviruses such as HIV-1. In rhesus monkeys, but not in humans, TRIM5α confers resistance to HIV-1. By measuring the binding of human and rhesus TRIM5α to a series of engineered HIV-1 capsid mimics of distinct capsid lattice interfaces, we reveal the HIV-1 capsid surface critical for species-specific binding by TRIM5α.

Intravenous administration of the adeno-associated virus-PHP.B capsid fails to upregulate transduction efficiency in the marmoset brain.

Intravenous administration of adeno-associated virus (AAV)-PHP.B, a capsid variant of AAV9 containing seven amino acid insertions, results in a greater permeability of the blood brain barrier (BBB) than standard AAV9 in mice, leading to highly efficient and global transduction of the central nervous system (CNS). The present study aimed to examine whether the enhanced BBB penetrance of AAV-PHP.B observed in mice also occurs in non-human primates. Thus, a young adult (age, 1.6 years) and an old adult (age, 7.2 years) marmoset received an intravenous injection of AAV-PHP.B expressing enhanced green fluorescent protein (EGFP) under the control of the constitutive CBh promoter (a hybrid of cytomegalovirus early enhancer and chicken β-actin promoter). Age-matched control marmosets were treated with standard AAV9-capsid vectors. The animals were sacrificed 6 weeks after the viral injection. Based on the results, only limited transduction of neurons (0-2%) and astrocytes (0.1-2.5%) was observed in both AAV-PHP.B- and AAV9-treated marmosets. One noticeable difference between AAV-PHP.B and AAV9 was the marked transduction of the peripheral dorsal root ganglia neurons. Indeed, the soma and axons in the projection from the spinal cord to the nucleus cuneatus in the medulla oblongata were strongly labeled with EGFP by AAV-PHP.B. Thus, except for the peripheral dorsal root ganglia neurons, the AAV-PHP.B transduction efficiency in the CNS of marmosets was comparable to that of AAV9 vectors.

Glial fibrillary acidic protein promoter determines transgene expression in satellite glial cells following intraganglionic adeno-associated virus delivery in adult rats.

Recombinant adeno-associated viral (AAV)-mediated therapeutic gene transfer to dorsal root ganglia (DRG) is an effective and safe tool for treating chronic pain. However, AAV with various constitutively active promoters leads to transgene expression predominantly to neurons, while glial cells are refractory to AAV transduction in the peripheral nervous system. The present study evaluated whether in vivo satellite glial cell (SGC) transduction in the DRG can be enhanced by the SGC-specific GFAP promoter and by using shH10 and shH19, which are engineered capsid variants with Müller glia-prone transduction. Titer-matched AAV6 (as control), AAVshH10, and AAVshH19, all encoding the EGFP driven by the constitutively active CMV promoter, as well as AAV6-EGFP and AAVshH10-EGFP driven by a GFAP promoter (AAV6-GFAP-EGFP and AAVshH10-GFAP-EGFP), were injected into DRG of adult male rats. Neurotropism of gene expression was determined and compared by immunohistochemistry. Results showed that injection of AAV6- and AAVshH10-GFAP-EGFP induces robust EGFP expression selectively in SGCs, whereas injection of either AAVshH10-CMV-EGFP or AAVshH19-CMV-EGFP into DRG resulted in a similar in vivo transduction profile to AAV6-CMV-EGFP, all showing efficient transduction of sensory neurons without significant transduction of glial cell populations. Coinjection of AAV6-CMV-mCherry and AAV6-GFAP-EGFP induces transgene expression in neurons and SGCs separately. This report, together with our prior studies, demonstrates that the GFAP promoter rather than capsid tropism determines selective gene expression in SGCs following intraganglionic AAV delivery in adult rats. A dual AAV system, one with GFAP promoter and the other with CMV promoter, can efficiently express transgenes selectively in neurons versus SGCs.

Spinal nociceptive circuit analysis with recombinant adeno-associated viruses: the impact of serotypes and promoters.

Recombinant adeno-associated virus (rAAV) vector-mediated gene transfer into genetically defined neuron subtypes has become a powerful tool to study the neuroanatomy of neuronal circuits in the brain and to unravel their functions. More recently, this methodology has also become popular for the analysis of spinal cord circuits. To date, a variety of naturally occurring AAV serotypes and genetically modified capsid variants are available but transduction efficiency in spinal neurons, target selectivity, and the ability for retrograde tracing are only incompletely characterized. Here, we have compared the transduction efficiency of seven commonly used AAV serotypes after intraspinal injection. We specifically analyzed local transduction of different types of dorsal horn neurons, and retrograde transduction of dorsal root ganglia (DRG) neurons and of neurons in the rostral ventromedial medulla (RVM) and the somatosensory cortex (S1). Our results show that most of the tested rAAV vectors have similar transduction efficiency in spinal neurons. All serotypes analyzed were also able to transduce DRG neurons and descending RVM and S1 neurons via their spinal axon terminals. When comparing the commonly used rAAV serotypes to the recently developed serotype 2 capsid variant rAAV2retro, a >20-fold increase in transduction efficiency of descending supraspinal neurons was observed. Conversely, transgene expression in retrogradely transduced neurons was strongly reduced when the human synapsin 1 (hSyn1) promoter was used instead of the strong ubiquitous hybrid cytomegalovirus enhancer/chicken β-actin promoter (CAG) or cytomegalovirus (CMV) promoter fragments. We conclude that the use of AAV2retro greatly increases transduction of neurons connected to the spinal cord via their axon terminals, while the hSyn1 promoter can be used to minimize transgene expression in retrogradely connected neurons of the DRG or brainstem. Cover Image for this issue: doi. 10.1111/jnc.13813.

Optimization of design and production strategies for novel adeno-associated viral display peptide libraries.

Libraries displaying random peptides on the surface of adeno-associated virus (AAV) are powerful tools for the generation of target-specific gene therapy vectors. However, for unknown reasons the success rate of AAV library screenings is variable and the influence of the production procedure has not been thoroughly evaluated. During library screenings, the capsid variants with the most favorable tropism are enriched over several selection rounds on a target of choice and identified by subsequent sequencing of the encapsidated viral genomes encoding the library capsids with targeting peptide insertions. Thus, a high capsid-genome correlation is crucial to obtain the correct information about the selected capsid variants. Producing AAV libraries by a two-step protocol with pseudotyped library transfer shuttles has been proposed as one way to ensure such a correlation. Here we show that AAV2 libraries produced by such a protocol via transfer shuttles display an unexpected additional bias in the amino-acid composition which confers increased heparin affinity and thus similarity to wildtype AAV2 tropism. This bias may fundamentally impair the intended use of AAV libraries, discouraging the use of transfer shuttles for the production of AAV libraries in the future.

Expansion of the assembly of cowpea chlorotic mottle virus towards non-native and physiological conditions

The cowpea chlorotic mottle virus (CCMV) is a nanoparticle that holds promise for diagnostic and therapeutic applications. The empty virus-like particle, however, is not stable under physiological conditions. Here, we describe a systematic study into the expansion of the assembly properties of a protein-based block copolymer of the CCMV capsid protein and an elastin-like polypeptide. By systematically changing the hydrophobicity of the stimulus-responsive elastin-like polypeptide block, assembly of the capsid proteins could be achieved at close to physiological conditions. This strategy may prove to be useful in the development of a physiologically stable CCMV capsid variant.

Present Molecular Limitations of ON-Bipolar Cell Targeted Gene Therapy.

Recent studies have demonstrated the safety and efficacy of ocular gene therapy based on adeno-associated viral vectors (AAVs). Accordingly, a surge in promising new gene therapies is entering clinical trials, including the first optogenetic therapy for vision restoration. To date, optogenetic therapies for vision restoration target either the retinal ganglion cells (GCs) or presynaptic ON-bipolar cells (OBCs). Initiating light responses at the level of the OBCs has significant advantages over optogenetic activation of GCs. For example, important neural circuitries in the inner retina, which shape the receptive fields of GCs, remain intact when activating the OBCs. Current drawbacks of AAV-mediated gene therapies targeting OBCs include (1) a low transduction efficiency, (2) off-target expression in unwanted cell populations, and (3) a poor performance in human tissue compared to the murine retina. Here, we examined side-by-side the performance of three state-of-the art AAV capsid variants, AAV7m8, AAVBP2, and AAV7m8(Y444F) in combination with the 4xGRM6-SV40 promoter construct in the healthy and degenerated mouse retina and in human post-mortem retinal explants. We find that (1) the 4xGRM6-SV40 promoter is not OBC specific, (2) that all AAV variants possess broad cellular transduction patterns, with differences between the transduction patterns of capsid variants AAVBP2 and AAV7m8 and, most importantly, (3) that all vectors target OBCs in healthy tissue but not in the degenerated rd1 mouse model, potentially limiting the possibilities for an OBC-targeted optogenetic therapy for vision restoration in the blind.

Cryo-electron Microscopy Reconstruction and Stability Studies of the Wild Type and the R432A Variant of Adeno-associated Virus Type 2 Reveal that Capsid Structural Stability Is a Major Factor in Genome Packaging.

The adeno-associated viruses (AAV) are promising therapeutic gene delivery vectors and better understanding of their capsid assembly and genome packaging mechanism is needed for improved vector production. Empty AAV capsids assemble in the nucleus prior to genome packaging by virally encoded Rep proteins. To elucidate the capsid determinants of this process, structural differences between wild-type (wt) AAV2 and a packaging deficient variant, AAV2-R432A, were examined using cryo-electron microscopy and three-dimensional image reconstruction both at an ∼5.0-Å resolution (medium) and also at 3.8- and 3.7-Å resolutions (high), respectively. The high resolution structures showed that removal of the arginine side chain in AAV2-R432A eliminated hydrogen bonding interactions, resulting in altered intramolecular and intermolecular interactions propagated from under the 3-fold axis toward the 5-fold channel. Consistent with these observations, differential scanning calorimetry showed an ∼10°C decrease in thermal stability for AAV2-R432A compared to wt-AAV2. In addition, the medium resolution structures revealed differences in the juxtaposition of the less ordered, N-terminal region of their capsid proteins, VP1/2/3. A structural rearrangement in AAV2-R432A repositioned the βA strand region under the icosahedral 2-fold axis rather than antiparallel to the βB strand, eliminating many intramolecular interactions. Thus, a single amino acid substitution can significantly alter the AAV capsid integrity to the extent of reducing its stability and possibly rendering it unable to tolerate the stress of genome packaging. Furthermore, the data show that the 2-, 3-, and 5-fold regions of the capsid contributed to producing the packaging defect and highlight a tight connection between the entire capsid in maintaining packaging efficiency. The mechanism of AAV genome packaging is still poorly understood, particularly with respect to the capsid determinants of the required capsid-Rep interaction. Understanding this mechanism may aid in the improvement of AAV packaging efficiency, which is currently ∼1:10 (10%) genome packaged to empty capsid in vector preparations. This report identifies regions of the AAV capsid that play roles in genome packaging and that may be important for Rep recognition. It also demonstrates the need to maintain capsid stability for the success of this process. This information is important for efforts to improve AAV genome packaging and will also inform the engineering of AAV capsid variants for improved tropism, specific tissue targeting, and host antibody escape by defining amino acids that cannot be altered without detriment to infectious vector production.

Insight into the mechanisms of enhanced retinal transduction by the engineered AAV2 capsid variant -7m8.

Recently, we described a modified AAV2 vector-AAV2-7m8-having a capsid-displayed peptide insertion of 10 amino acids with enhanced retinal transduction properties. The insertion of the peptide referred to as 7m8 is responsible for high-level gene delivery into deep layers of the retina when virus is delivered into the eye's vitreous. Here, we further characterize AAV2-7m8 mediated gene delivery to neural tissue and investigate the mechanisms by which the inserted peptide provides better transduction away from the injection site. First, in order to understand if the peptide exerts its effect on its own or in conjunction with the neighboring amino acids, we inserted the 7m8 peptide at equivalent positions on three other AAV capsids, AAV5, AAV8, and AAV9, and evaluated its effect on their infectivity. Intravitreal delivery of these peptide insertion vectors revealed that only AAV9 benefited from 7m8 insertion in the context of the retina. We then investigated AAV2-7m8 and AAV9-7m8 properties in the brain, to better evaluate the spread and efficacy of viral transduction in view of the peptide insertion. While 7m8 insertion led to higher intensity gene expression, the spread of gene expression remained unchanged compared to the parental serotypes. Our results indicate that the 7m8 peptide insertion acts by increasing efficacy of cellular entry, with little effect on the spread of viral particles in neural tissue. The effects of peptide insertion are capsid and tissue dependent, highlighting the importance of the microenvironment in gene delivery using AAV. Biotechnol. Bioeng. 2016;113: 2712-2724. © 2016 Wiley Periodicals, Inc.

Pulmonary Targeting of Adeno-associated Viral Vectors by Next-generation Sequencing-guided Screening of Random Capsid Displayed Peptide Libraries.

Vectors mediating strong, durable, and tissue-specific transgene expression are mandatory for safe and effective gene therapy. In settings requiring systemic vector administration, the availability of suited vectors is extremely limited. Here, we present a strategy to select vectors with true specificity for a target tissue from random peptide libraries displayed on adeno-associated virus (AAV) by screening the library under circulation conditions in a murine model. Guiding the in vivo screening by next-generation sequencing, we were able to monitor the selection kinetics and to determine the right time point to discontinue the screening process. The establishment of different rating scores enabled us to identify the most specifically enriched AAV capsid candidates. As proof of concept, a capsid variant was selected that specifically and very efficiently delivers genes to the endothelium of the pulmonary vasculature after intravenous administration. This technical approach of selecting target-specific vectors in vivo is applicable to any given tissue of interest and therefore has broad implications in translational research and medicine.

253. Expanded Packaging Capacity of AAV by Lumenal Charge Alteration

Adeno-associated virus (AAV) is widely believed to be the safest viral vector for gene therapy. Moreover, the array of AAV serotypes available allows transduction of different tissues in vivo. However, one major limitation of AAV is the relatively small DNA packaging size (4,700 nt). Existing serotypes have been over-packaged with limited success and varying reproducibility. The consensus appears to be that AAV can be over packaged by ~10%, but with a concomitant reduction in both viral titers and in vivo transduction.We have taken a novel approach to increase the packaging limitations of AAV by focusing on the lumen of AAV where vector DNA interacts with the capsid. We have created a series of AAV capsid variants that alter the capsid lumenal charge using AAV-DJ as the scaffold. The variants add positively charged lysine and arginine residues at lumenally exposed sites within the capsid that, we believe, may allow for interaction and condensation of vector DNA increasing the packaging capacity and/or stabilizing the capsid in an over-packaged state. We present in vitro characterization of the library of capsid variants and demonstrate that variants with intermediate charge (+4 to +7 over wildtype) can be produced at high titer when over-packaged up to 6.2 kb. In addition, these variants can transduce HEK293 cells with a 5.6 kb length vector better than wildtype. Finally, these variants can completely encapsidate at least a 5.4 kb vector by alkaline Southern blot. To validate that these new vectors are functional, we establish that a subset of these vectors are still able to transduce mouse liver in vivo. We are currently establishing the relative transduction efficiencies using standard length and expanded genome sizes in order to establish the maximal DNA length that can be functionally packaged by these new vectors. We believe our capsid alteration strategy can be extrapolated to other serotypes with distinct tissue tropism and applied in vivo to treat hemophilia A and other monogenic disorders suitable to replacement therapy.

133. Highly Efficient Homology-Driven Genome Editing in Human T Cells with Combined Zinc-Finger Nuclease mRNA and AAV6 Donor Delivery and Improved Efficiency Under Serum-Free Conditions

Immunotherapy using gene-modified T cells for adoptive cell transfer is a rapidly expanding field that is currently being tested in early- and late-stage clinical studies, with recent successes seen in the treatment of hematologic malignancies using T-cell receptor or chimeric antigen receptor (CAR)-retargeted T cells. Ex vivo T cell modification can potentially further enhance the activity and safety of these cells via the insertion of the desired transgenes at a specific pre-selected site such as a safe harbor locus. We describe the development of a highly efficient method to genome edit both primary human CD8 and CD4 T cells by homology-directed repair at a pre-defined site of the genome. We have identified AAV serotype 6 as a capsid variant with high tropism for both human CD8+ and CD4+ T cells. Two different homology donor templates were evaluated, representing both minor gene editing events (restriction site insertion) to mimic gene correction, or the more significant insertion of a larger gene cassette. By combining this AAV6 donor delivery method with the delivery of ZFNs as mRNA, we could ‘gene correct’ >40% of CCR5 or 55% of PPP1R12C (AAVS1) alleles in CD8+ T cells. We were also able to achieve targeted insertion of a GFP transgene cassette in >40% of CD8+ and CD4+ T cells at both the CCR5 and AAVS1 safe harbor loci, potentially providing a robust genome editing tool for T cell-based immunotherapy. In addition, we found that transducing T cells with AAV6 donor and ZFN encoding mRNA under serum-free conditions resulted in a marked improvement in HDR-mediated genome editing at significantly lower vector doses (2-log lower doses of AAV6 donor), representing an essential stride in the production of clinical-scale genome edited T cells for use in adoptive T cell therapy.

617. Identification of Novel AAV Capsid Variants with Enhanced Tropism for the Canine Outer Retina Following Intravitreal Delivery

The delivery of AAV vectors across physical barriers to target the outer retina is key to the success of retinal gene therapies. Subretinal injection of AAV vectors is the most common approach to deliver genes to photoreceptors and RPE; however, this route of administration presents surgical challenges and potential complications, as well as mediates transfer to only a limited region of the retinal surface. Intravitreal administration may represent a safer way of targeting larger areas of the outer retina; physical barriers such as the internal limiting membrane significantly limit efficiency of natural AAVs through this route. In a prior murine proof of concept study, we showed that directed evolution is promising method for engineering new AAV capsid variants with enhanced permissivity for the outer retina. Here, we show results from a screen performed in the canine retina, a large animal model with eye structure and size similar to the human eye. A library of >107 distinct capsid variant AAV vectors was injected intravitreally into the eye of a wildtype dog. A subset of the AAV vectors injected was recovered from samples of outer retina by PCR amplification. Isolated variants were then pooled and re-injected in additional wildtype canine eyes, and following 5 total rounds of selection the pool had converged to a relatively small number of variants. High throughput sequencing was used to characterize the convergence of variants over all rounds of selection. The final candidate variants were then evaluated by packaging a GFP reporter controlled by a ubiquitous CAG promoter. Retinal distribution and cellular tropism of the variants was examined by cSLO (AF mode) imaging, histology and immunocytochemistry. Nineteen variants were identified as candidate vectors for targeting the canine outer retina. Intravitreal injection of a mixture of equal amounts of these variants carrying GFP fused to a barcode resulted in efficient targeting of all retinal layers including photoreceptors and RPE in 6 eyes of 3 wildtype dogs. Thus, directed evolution performed in a large animal model, with an eye structure similar to that of humans, resulted in new AAV capsid variants with improved abilities to bypass significant structural barriers and efficiently transduce the outer retina. High throughput sequencing allowed for identification of promising variants and monitoring of convergence. These vectors show promise for gene therapy delivery via the intravitreal route of administration, potentially providing a safer approach for targeting a larger area of outer retina than currently achieved by the more traumatic and focal subretinal injection method.

541. Use of Endogenous Viral Elements to Deduce the Sequence of an Ancient Adeno-Associated Virus That Circulated Amongst Australian Marsupials ~30 Million Years Ago

Vectors based on adeno-associated viruses (AAVs) have been successfully applied in clinical trials and pre-clinical settings. AAV capsid proteins are the primary determinants of vector tropism and immunogenicity, and are continually being developed to impart desirable properties, such as immune evasion and specificity for selected target tissues in humans and those of model organisms. One approach that was recently adopted by at least two groups to generate novel capsid variants is the computational deduction of putative ancestral AAV sequences, based upon observable genetic diversity amongst closely related contemporary AAV isolates. Such inferred ancestral sequences potentially recapitulate the evolutionary history of AAVs for hundreds or possibly thousands of years. Germline endogenous viral elements (EVEs) can compensate for the lack of a viral fossil record by genetically preserving viral nucleotide sequences over a geologic time span. Here we describe an AAV-EVE (mAAV-EVE1) that is lineage-restricted to the germlines of members of the Australian marsupial suborder Macropodiformes (present-day kangaroos, wallabies, and related macropodoids), to the exclusion of other Diprotodontian marsupial lineages. Orthologous mAAV-EVE1 sequences from sixteen macropodoid species, representing a speciation history spanning an estimated 30 million years, facilitated in silico reconstruction of an inferred ancestral AAV sequence. This sequence represents the genome of an ancient marsupial AAV that circulated among Australian metatherian fauna sometime during the late Eocene to early Oligocene. The deduced ancient AAV bears remarkable resemblance to modern AAVs in its predicted structural and non-structural genes, underscoring the concept that despite their relatively rapid mutation rates, AAVs have evolved under tight constraints imposed by both form and function. Moreover, the evolutionary history of mAAV-EVE1 since its integration into the pro-macropodoid lineage reflects that of the respective host species. In addition to providing an inferred ancient AAV capsid sequence for future transduction studies, this sequence can serve as input for the generation of further novel capsids by approaches such as capsid shuffling and directed capsid engineering/evolution. Furthermore, by comparing it to contemporary AAV capsids, this sequence can inform rational capsid design approaches by providing insights into long-term capsid evolution.

722. AAV Capsid Evolution for Enhanced Antibody Delivery to Human Skeletal Muscle for Use in Next-Generation HIV Vaccines and Muscle Gene Therapies

Our goal is to develop an AAV vector that can transduce human skeletal muscle after intramuscular administration at levels sufficient to express therapeutic levels of antibodies with broad-spectrum protection against HIV. AAV has begun being used in passive vaccines for HIV and influenza. Yet for many indications, greater human skeletal muscle transduction is needed than has been achieved with existing AAV serotypes. To bioengineer novel capsids with unparalleled human skeletal muscle transduction, we utilized primary human skeletal muscle cells from surgical resections to screen libraries of replication-competent shuffled AAV capsids. Two screens were performed in pools of primary human skeletal muscle stem cells or myotubes from six patients. Six rounds of replicating selection were carried out and the five most highly selected variants from each screen were vectorized and validated against existing serotypes with muscle tropism (AAV1, 6 and 8). In primary human muscle stem cells, variants NP22, NP66 and NP94 had significantly increased transduction that ranged from a 65 to 284-fold improvement over AAV1, 27 to 118-fold over AAV6, and 10 to 45-fold over AAV8. In primary human myotubes, NP22, NP66 and NP94 again showed significantly increased transduction that ranged from a 13 to 464-fold improvement over AAV1, 25 to 871-fold over AAV8, and NP94 showed a 15-fold improvement over AAV6. To assess human skeletal muscle transduction capabilities in vivo, we injected xenografted humanized muscle NSG mice with shuffled or control capsids expressing luciferase by intramuscular injection (1E9/leg) and assessed transduction weekly by live imaging in time course studies over two months. AAV6 was the first to uncoat but NP66 produced the highest sustained transduction levels. To control for species-specific transduction, we performed the same time course experiment in strain- and gender-matched non-transplanted mice. Strikingly, AAV6 outperformed shuffled variants when no human muscle fibers were present, highlighting the importance of performing preclinical studies in human cells and xenograft systems whenever possible to prevent misleading results which are mouse-specific. To further demonstrate the superiority of our shuffled variants and more accurately predict eventual muscle transduction in humans, we transduced human skeletal muscle explants from surgical resections ex vivo. Four adult patients (two male and two female) had healthy skeletal muscle tissue removed for ex vivo transduction analyses with luciferase. In all 4 patients, NP22 and NP66 (and NP94 in 3/4 patients) had significantly increased transduction by live fiber luciferase imaging as well as luciferase assays on lysed fibers that ranged from a 4 to 116-fold improvement over all control serotypes. To support preclinical vaccine testing in non-human primates, we assessed transduction in rhesus macaque skeletal muscle explants ex vivo. Here again we showed that NP22 and NP66 showed a 30 to 57-fold improvement in transduction over control serotypes. Taken together, our results demonstrate that capsid variants NP22 and NP66 (and in some cases NP94), have highly significant increased human skeletal muscle fiber transduction when assessed in vitro, in vivo and ex vivo. These capsids represent powerful tools to express therapeutic quantities of human monoclonal antibodies for use in passive vaccines, or transgenes for various muscle disorders in gene therapy, specifically in humans.

257. Selection of Next Generation AAV Gene Therapy Vectors for Specific and Precise Gene Delivery

The ultimate goal in human gene therapy is the specific and exclusive modification of the desired target cells upon systemic vector administration. Especially vectors derived from adeno-associated virus (AAV) are among the most promising gene transfer systems for in vivo application and have received broad attention due to substantial clinical benefit. However, AAV specificity for a particular target cell or tissue has been hampered by the broad tropism of different AAV serotypes. Over the last several years, new approaches have been initiated to create and select for more effective and selective recombinant AAV vectors by genetically modifying the capsid protein. These methods include random and/or rationale amino acid substitutions, creation of chimeric capsid variant libraries and various selection screens, and/or peptide insertion. A different approach involves the incorporation of highly specific binding molecules (DARPins) into the AAV capsid. DARPins are derived from ankyrin-repeat proteins that have been developed as alternative to antibody-based scaffolds, which are selected by high-throughput screens from DARPin libraries. At present, the bottleneck remains the cumbersome selection of DARPin molecules and the low number of functional DARPin capsid chimeras, which are able to assemble into a functional vector and still bind their intended extracellular target. Thus after months of screening and analysis the majority of the pre-selected DARPins are found not to be suitable for AAV targeting approaches. Here we describe a novel reliable and faster selection approach. First we eliminate the cumbersome and often misleading prokaryotic selection steps. To do this, we incorporate an entire DARPin library into the AAV capsid as VP2-fusion protein, generating either replication deficient or competent AAV library systems allowing for direct screening on therapeutically relevant cell types in vitro or in vivo. Our studies show that replicating deficient and replicating AAV libraries can be generated with high diversity up to 5×107 and high functional titers retaining their infectivity. We have generated a DARPin-capsid library in AAV-DJ (serotype with diverse cellular tropism) and AAV-LK03 (human selective serotype) and after only two selection rounds, found up to 10% of selected DARPins sharing identical repeat motifs responsible for target receptor binding in the pancreas and the liver. We are currently making vectors from these chimeric capsids and will determine their target specificity both in vivo and in vitro. This new approach takes only six days per selection cycle, allowing discovery of novel DARPin molecules and selection for targeted vectors in substantially reduced time. This approach expands the potential diversity in creating and defining novel rAAV vectors with medically relevant transduction properties. Such viral vectors will not only open the door for an array of new approaches to treat acquired human diseases but also push the development of new AAV-based gene therapy vectors to the next level.

A brain microvasculature endothelial cell-specific viral vector with the potential to treat neurovascular and neurological diseases.

Gene therapy critically relies on vectors that combine high transduction efficiency with a high degree of target specificity and that can be administered through a safe intravenous route. The lack of suitable vectors, especially for gene therapy of brain disorders, represents a major obstacle. Therefore, we applied an in vivo screening system of random ligand libraries displayed on adeno‐associated viral capsids to select brain‐targeted vectors for the treatment of neurovascular diseases. We identified a capsid variant showing an unprecedented degree of specificity and long‐lasting transduction efficiency for brain microvasculature endothelial cells as the primary target of selection. A therapeutic vector based on this selected viral capsid was used to markedly attenuate the severe cerebrovascular pathology of mice with incontinentia pigmenti after a single intravenous injection. Furthermore, the versatility of this selection system will make it possible to select ligands for additional in vivo targets without requiring previous identification of potential target‐specific receptors.

Accumulation of MxB/Mx2-resistant HIV-1 Capsid Variants During Expansion of the HIV-1 Epidemic in Human Populations

Recent studies have identified human myxovirus resistance protein 2 (MxB or Mx2) as an interferon induced inhibitor of HIV-1 replication. However, whether HIV-1 can overcome MxB restriction without compromise of viral fitness has been undefined. Here, we have discovered that naturally occurring capsid (CA) variants can render HIV-1 resistant to the activity of MxB without losing viral infectivity or the ability to escape from interferon induction. Moreover, these MxB resistant HIV-1 variants do not lose MxB recognition. Surprisingly, MxB resistant CA variants are most commonly found in the Clade C HIV-1 that is the most rapidly expanding Clade throughout the world. Accumulation of MxB resistant mutations is also observed during HIV-1 spreading in human populations. These findings support a potential role for MxB as a selective force during HIV-1 transmission and evolution.

Microglia-specific targeting by novel capsid-modified AAV6 vectors

Recombinant adeno-associated viruses (rAAV) have been widely used in gene therapy applications for central nervous system diseases. Though rAAV can efficiently target neurons and astrocytes in mouse brains, microglia, the immune cells of the brain, are refractile to rAAV. To identify AAV capsids with microglia-specific transduction properties, we initially screened the most commonly used serotypes, AAV1–9 and rh10, on primary mouse microglia cultures. While these capsids were not permissive, we then tested the microglial targeting properties of a newly characterized set of modified rAAV6 capsid variants with high tropism for monocytes. Indeed, these newly characterized rAAV6 capsid variants, specially a triply mutated Y731F/Y705F/T492V form, carrying a self-complementary genome and microglia-specific promoters (F4/80 or CD68) could efficiently and selectively transduce microglia in vitro. Delivery of these constructs in mice brains resulted in microglia-specific expression of green fluorescent protein, albeit at modest levels. We further show that CD68 promoter–driven expression of the inflammatory cytokine, interleukin-6, using this capsid variant leads to increased astrogliosis in the brains of wild-type mice. Our study describes the first instance of AAV-targeted microglial gene expression leading to functional modulation of the innate immune system in mice brains. This provides the rationale for utilizing these unique capsid/promoter combinations for microglia-specific gene targeting for modeling or functional studies.