Correlation of antigen expression with epigenetic modifications after rAAV delivery of a human factor IX variant in mice and rhesus macaques.

We investigated long-term human coagulation factor IX (huFIX) expression of a novel variant when delivered into mice and rhesus macaques and compared transduction efficiencies using two different adeno-associated virus (AAV) capsids. In hemophilic mice injected with KP1-packaged recombinant AAV (rAAV) expressing the hyperactive FIX variant specific activity plasma levels were 10-fold or 2-fold enhanced when compared with wild-type or Padua huFIX injected mice, respectively. In rhesus macaques AAV-LK03 capsid outperformed AAV-KP1 in terms of antigen expression and liver transduction. Two animals from each group showed sustained low-level huFIX expression at 3 months after administration, while one animal from each group lost huFIX mRNA and protein expression over time, despite comparable vector copies. We investigated whether epigenetic differences in the vector episomes could explain this loss of transcription. Cut&Tag analysis revealed lower levels of activating histone marks in the two animals that lost expression. When comparing rAAV genome associated histone modifications in rhesus macaques with those in mice injected with the same vector, the activating histone marks were starkly decreased in macaque-derived episomes. Differential epigenetic marking of AAV genomes may explain different expression profiles in mice and rhesus macaques, as well as the wide dose response variation observed in primates in both preclinical and human clinical trials.

Improved Genome Editing through Inhibition of FANCM and Members of the BTR Dissolvase Complex.

Recombinant adeno-associated virus (rAAV) vectors have the unique property of being able to perform genomic targeted integration (TI) without inducing a double-strand break (DSB). In order to improve our understanding of the mechanism behind TI mediated by AAV and improve its efficiency, we performed an unbiased genetic screen in human cells using a promoterless AAV-homologous recombination (AAV-HR) vector system. We identified that the inhibition of the Fanconi anemia complementation group M (FANCM) protein enhanced AAV-HR-mediated TI efficiencies in different cultured human cells by ∼6- to 9-fold. The combined knockdown of the FANCM and two proteins also associated with the FANCM complex, RecQ-mediated genome instability 1 (RMI1) and Bloom DNA helicase (BLM) from the BLM-topoisomerase IIIα (TOP3A)-RMI (BTR) dissolvase complex (RMI1, having also been identified in our screen), led to the enhancement of AAV-HR-mediated TI up to ∼17 times. AAV-HR-mediated TI in the presence of a nuclease (CRISPR-Cas9) was also increased by ∼1.5- to 2-fold in FANCM and RMI1 knockout cells, respectively. Furthermore, knockdown of FANCM in human CD34+ hematopoietic stem and progenitor cells (HSPCs) increased AAV-HR-mediated TI by ∼3.5-fold. This study expands our knowledge on the mechanisms related to AAV-mediated TI, and it highlights new pathways that might be manipulated for future improvements in AAV-HR-mediated TI.

Bioengineered AAV Capsids with Combined High Human Liver Transduction In Vivo and Unique Humoral Seroreactivity.

Existing recombinant adeno-associated virus (rAAV) serotypes for delivering in vivo gene therapy treatments for human liver diseases have not yielded combined high-level human hepatocyte transduction and favorable humoral neutralization properties in diverse patient groups. Yet, these combined properties are important for therapeutic efficacy. To bioengineer capsids that exhibit both unique seroreactivity profiles and functionally transduce human hepatocytes at therapeutically relevant levels, we performed multiplexed sequential directed evolution screens using diverse capsid libraries in both primary human hepatocytes in vivo and with pooled human sera from thousands of patients. AAV libraries were subjected to five rounds of in vivo selection in xenografted mice with human livers to isolate an enriched human-hepatotropic library that was then used as input for a sequential on-bead screen against pooled human immunoglobulins. Evolved variants were vectorized and validated against existing hepatotropic serotypes. Two of the evolved AAV serotypes, NP40 and NP59, exhibited dramatically improved functional human hepatocyte transduction in vivo in xenografted mice with human livers, along with favorable human seroreactivity profiles, compared with existing serotypes. These novel capsids represent enhanced vector delivery systems for future human liver gene therapy applications.

AAV-mediated genome editing is influenced by the formation of R-loops.

Recombinant adeno-associated viral vectors (rAAV) hold an intrinsic ability to stimulate homologous recombination (AAV-HR) and are the most used in clinical settings for in vivo gene therapy. However, rAAVs also integrate throughout the genome. Here, we describe DNA-RNA immunoprecipitation sequencing (DRIP-seq) in murine HEPA1-6 hepatoma cells and whole murine liver to establish the similarities and differences in genomic R-loop formation in a transformed cell line and intact tissue. We show enhanced AAV-HR in mice upon genetic and pharmacological upregulation of R-loops. Selecting the highly expressed Albumin gene as a model locus for genome editing in both in vitro and in vivo experiments showed that the R-loop prone, 3' end of Albumin was efficiently edited by AAV-HR, whereas the upstream R-loop-deficient region did not result in detectable vector integration. In addition, we found a positive correlation between previously reported off-target rAAV integration sites and R-loop enriched genomic regions. Thus, we conclude that high levels of R-loops, present in highly transcribed genes, promote rAAV vector genome integration. These findings may shed light on potential mechanisms for improving the safety and efficacy of genome editing by modulating R-loops and may enhance our ability to predict regions most susceptible to off-target insertional mutagenesis by rAAV vectors.

Genome engineering with Cas9 and AAV repair templates generates frequent concatemeric insertions of viral vectors.

CRISPR-Cas9 paired with adeno-associated virus serotype 6 (AAV6) is among the most efficient tools for producing targeted gene knockins. Here, we report that this system can lead to frequent concatemeric insertions of the viral vector genome at the target site that are difficult to detect. Such errors can cause adverse and unreliable phenotypes that are antithetical to the goal of precision genome engineering. The concatemeric knockins occurred regardless of locus, vector concentration, cell line or cell type, including human pluripotent and hematopoietic stem cells. Although these highly abundant errors were found in more than half of the edited cells, they could not be readily detected by common analytical methods. We describe strategies to detect and thoroughly characterize the concatemeric viral vector insertions, and we highlight analytical pitfalls that mask their prevalence. We then describe strategies to prevent the concatemeric inserts by cutting the vector genome after transduction. This approach is compatible with established gene editing pipelines, enabling robust genetic knockins that are safer, more reliable and more reproducible.

The AAV capsid can influence the epigenetic marking of rAAV delivered episomal genomes in a species dependent manner.

Recombinant adeno-associated viral vectors (rAAVs) are among the most commonly used vehicles for in vivo based gene therapies. However, it is hard to predict which AAV capsid will provide the most robust expression in human subjects due to the observed discordance in vector-mediated transduction between species. In our study, we use a primate specific capsid, AAV-LK03, to demonstrate that the limitation of this capsid towards transduction of mouse cells is unrelated to cell entry and nuclear transport but rather due to depleted histone H3 chemical modifications related to active transcription, namely H3K4me3 and H3K27ac, on the vector DNA itself. A single-amino acid insertion into the AAV-LK03 capsid enables efficient transduction and the accumulation of active-related epigenetic marks on the vector chromatin in mouse without compromising transduction efficiency in human cells. Our study suggests that the capsid protein itself is involved in driving the epigenetic status of the vector genome, most likely during the process of uncoating. Programming viral chromatin states by capsid design may enable facile DNA transduction between vector and host species and ultimately lead to rational selection of AAV capsids for use in humans.

A strategy for high antibody expression with low anti-drug antibodies using AAV9 vectors.

Introduction: Use of adeno-associated virus (AAV) vectors is complicated by host immune responses that can limit transgene expression. Recent clinical trials using AAV vectors to deliver HIV broadly neutralizing antibodies (bNAbs) by intramuscular administration resulted in poor expression with anti-drug antibodies (ADA) responses against the bNAb. Methods: Here we compared the expression of, and ADA responses against, an anti-SIV antibody ITS01 when delivered by five different AAV capsids. We first evaluated ITS01 expression from AAV vectors three different 2A peptides. Rhesus macaques were selected for the study based on preexisiting neutralizing antibodies by evaluating serum samples in a neutralization assay against the five capsids used in the study. Macaques were intramuscularly administered AAV vectors at a 2.5x10^12 vg/kg over eight administration sites. ITS01 concentrations and anti-drug antibodies (ADA) were measured by ELISA and a neutralization assay was conducted to confirm ex vivo antibody potency. Results: We observed that ITS01 expressed three-fold more efficiently in mice from AAV vectors in which heavy and light-chain genes were separated by a P2A ribosomal skipping peptide, compared with those bearing F2A or T2A peptides. We then measured the preexisting neutralizing antibody responses against three traditional AAV capsids in 360 rhesus macaques and observed that 8%, 16%, and 42% were seronegative for AAV1, AAV8, and AAV9, respectively. Finally, we compared ITS01 expression in seronegative macaques intramuscularly transduced with AAV1, AAV8, or AAV9, or with the synthetic capsids AAV-NP22 or AAV-KP1. We observed at 30 weeks after administration that AAV9- and AAV1-delivered vectors expressed the highest concentrations of ITS01 (224 µg/mL, n=5, and 216 µg/mL, n=3, respectively). The remaining groups expressed an average of 35-73 µg/mL. Notably, ADA responses against ITS01 were observed in six of the 19 animals. Lastly, we demonstrated that the expressed ITS01 retained its neutralizing activity with nearly the same potency of purified recombinant protein. Discussion: Overall, these data suggest that the AAV9 capsid is a suitable choice for intramuscular expression of antibodies in nonhuman primates.

Selection of viral capsids and promoters affects the efficacy of rescue of Tmprss3-deficient cochlea.

Adeno-associated virus (AAV)-mediated gene transfer has shown promise in rescuing mouse models of genetic hearing loss, but how viral capsid and promoter selection affects efficacy is poorly characterized. Here, we tested combinations of AAVs and promoters to deliver Tmprss3, mutations in which are associated with hearing loss in humans. Tmprss3tm1/tm1 mice display severe cochlear hair cell degeneration, loss of auditory brainstem responses, and delayed loss of spiral ganglion neurons. Under the ubiquitous CAG promoter and AAV-KP1 capsid, Tmprss3 overexpression caused striking cytotoxicity in vitro and in vivo and failed to rescue degeneration or dysfunction of the Tmprss3tm1/tm1 cochlea. Reducing the dosage or using AAV-DJ-CAG-Tmprss3 diminished cytotoxicity without rescue of the Tmprss3tm1/tm1 cochlea. Finally, the combination of AAV-KP1 capsid and the EF1α promoter prevented cytotoxicity and reduced hair cell degeneration, loss of spiral ganglion neurons, and improved hearing thresholds in Tmprss3tm1/tm1 mice. Together, our study illustrates toxicity of exogenous genes and factors governing rescue efficiency, and suggests that cochlear gene therapy likely requires precisely targeted transgene expression.

Selection of rAAV vectors that cross the human blood-brain barrier and target the central nervous system using a transwell model.

A limitation for recombinant adeno-associated virus (rAAV)-mediated gene transfer into the central nervous system (CNS) is the low penetration of vectors across the human blood-brain barrier (BBB). High doses of intravenously delivered vector are required to reach the CNS, which has resulted in varying adverse effects. Moreover, selective transduction of various cell types might be important depending on the disorder being treated. To enhance BBB penetration and improve CNS cell selectivity, we screened an AAV capsid-shuffled library using an in vitro transwell BBB system with separate layers of human endothelial cells, primary astrocytes and/or human induced pluripotent stem cell-derived cortical neurons. After multiple passages through the transwell, we identified chimeric AAV capsids with enhanced penetration and improved transduction of astrocytes and/or neurons compared with wild-type capsids. We identified the amino acids (aa) from regions 451-470 of AAV2 associated with the capsids selected for neurons, and a combination of aa from regions 413-496 of AAV-rh10 and 538-598 of AAV3B/LK03 associated with capsids selected for astrocytes. A small interfering RNA screen identified several genes that affect transcytosis of AAV across the BBB. Our work supports the use of a human transwell system for selecting enhanced AAV capsids targeting the CNS and may allow for unraveling the underlying molecular mechanisms of BBB penetration.

Fludarabine increases nuclease-free AAV- and CRISPR/Cas9-mediated homologous recombination in mice.

Homologous recombination (HR)-based gene therapy using adeno-associated viruses (AAV-HR) without nucleases has several advantages over classic gene therapy, especially the potential for permanent transgene expression. However, the low efficiency of AAV-HR remains a major limitation. Here, we tested a series of small-molecule compounds and found that ribonucleotide reductase (RNR) inhibitors substantially enhance AAV-HR efficiency in mouse and human liver cell lines approximately threefold. Short-term administration of the RNR inhibitor fludarabine increased the in vivo efficiency of both non-nuclease- and CRISPR/Cas9-mediated AAV-HR two- to sevenfold in the murine liver, without causing overt toxicity. Fludarabine administration induced transient DNA damage signaling in both proliferating and quiescent hepatocytes. Notably, the majority of AAV-HR events occurred in non-proliferating hepatocytes in both fludarabine-treated and control mice, suggesting that the induction of transient DNA repair signaling in non-dividing hepatocytes was responsible for enhancing AAV-HR efficiency in mice. These results suggest that use of a clinically approved RNR inhibitor can potentiate AAV-HR-based genome-editing therapeutics.

Tracking Adeno-Associated Virus Capsid Evolution by High-Throughput Sequencing.

Despite early successes using recombinant adeno-associated virus (rAAV) vectors in clinical gene therapy trials, limitations remain making additional advancements a necessity. Some of the challenges include variable levels of pre-existing neutralizing antibodies and poor transduction in specific target tissues and/or diseases. In addition, readministration of an rAAV vector is in general not possible due to the immune response against the capsid. Recombinant adeno-associated virus (AAV) vectors with novel capsids can be isolated in nature or developed through different directed evolution strategies. However, in most cases, the process of AAV selection is not well understood and new strategies are required to define the best parameters to develop more efficient and functional rAAV capsids. Therefore, the use of barcoding for AAV capsid libraries, which can be screened by high-throughput sequencing, provides a powerful tool to track AAV capsid evolution and potentially improve AAV capsid library screens. In this study, we examined how different parameters affect the screen of two different AAV libraries in two human cell types. We uncovered new and unexpected insights in how to maximize the likelihood of obtaining AAV variants with the desired properties. The major findings of the study are the following. (1) Inclusion of helper-virus for AAV replication can selectively propagate variants that can replicate to higher titers, but are not necessarily better at transduction. (2) Competition between AAVs with specific capsids can take place in cells that have been infected with different AAVs. (3) The use of low multiplicity of infections for infection results in more variation between screens and is not optimal at selecting the most desired capsids. (4) Using multiple rounds of selection can be counterproductive. We conclude that each of these parameters should be taken into consideration when screening AAV libraries for enhanced properties of interest.

Using a barcoded AAV capsid library to select for clinically relevant gene therapy vectors.

While gene transfer using recombinant adeno-associated viral (rAAV) vectors has shown success in some clinical trials, there remain many tissues that are not well transduced. Because of the recent success in reprogramming islet-derived cells into functional ß cells in animal models, we constructed 2 highly complex barcoded replication competent capsid shuffled libraries and selected for high-transducing variants on primary human islets. We describe the generation of a chimeric AAV capsid (AAV-KP1) that facilitates transduction of primary human islet cells and human embryonic stem cell-derived ß cells with up to 10-fold higher efficiency compared with previously studied best-in-class AAV vectors. Remarkably, this chimeric capsid also enabled transduction of both mouse and human hepatocytes at very high levels in a humanized chimeric mouse model, thus providing a versatile vector that has the potential to be used in both preclinical testing and human clinical trials for liver-based diseases and diabetes.

Bioengineered Viral Platform for Intramuscular Passive Vaccine Delivery to Human Skeletal Muscle.

Skeletal muscle is ideal for passive vaccine administration as it is easily accessible by intramuscular injection. Recombinant adeno-associated virus (rAAV) vectors are in consideration for passive vaccination clinical trials for HIV and influenza. However, greater human skeletal muscle transduction is needed for therapeutic efficacy than is possible with existing serotypes. To bioengineer capsids with therapeutic levels of transduction, we utilized a directed evolution approach to screen libraries of shuffled AAV capsids in pools of surgically resected human skeletal muscle cells from five patients. Six rounds of evolution were performed in various muscle cell types, and evolved variants were validated against existing muscle-tropic serotypes rAAV1, 6, and 8. We found that evolved variants NP22 and NP66 had significantly increased primary human and rhesus skeletal muscle fiber transduction from surgical explants ex vivo and in various primary and immortalized myogenic lines in vitro. Importantly, we demonstrated reduced seroreactivity compared to existing serotypes against normal human serum from 50 adult donors. These capsids represent powerful tools for human skeletal muscle expression and secretion of antibodies from passive vaccines.

Semi-synthetic DNA shuffling of aveC leads to improved industrial scale production of doramectin by Streptomyces avermitilis.

The avermectin analog doramectin (CHC-B1), sold commercially as Dectomax, is biosynthesized by Streptomyces avermitilis. aveC, a gene encoding an unknown mechanistic function, plays an essential role in the production of doramectin (avermectin CHC-B1), modulating the production ratio of CHC-B1 to other avermectins, most notably the undesirable analog CHC-B2. To improve the production ratio for doramectin, the aveC gene was subjected to iterative rounds of semi-synthetic DNA shuffling. Libraries of shuffled aveC gene variants were transformed into S. avermitilis, screened using a miniaturized 96-well growth and production format, and analyzed by high throughput mass spectrometry to determine CHC-B2:CHC-B1 ratios. Several improved aveC variants were identified; the best shuffled gene encoded 10 amino acid mutations, and conferred a final CHC-B2:CHC-B1 ratio of 0.07:1, a 23-fold improvement over the starting gene (aveC wild type). Chromosomal insertion of an improved aveC shuffled gene into a high titer S. avermitilis strain yielded an improved doramectin production strain. This strain is under development to be used commercially, and is expected to provide considerable cost savings in large-scale manufacture, as well as significantly reducing by-product levels of CHC-B2 requiring disposal.

Chimeric human immunodeficiency virus type 1 containing murine leukemia virus matrix assembles in murine cells.

Murine cells do not support efficient assembly and release of human immunodeficiency virus type 1 (HIV-1) virions. HIV-1-infected mouse cells that express transfected human cyclin T1 synthesize abundant Gag precursor polyprotein, but inefficiently assemble and release virions. This assembly defect may result from a failure of the Gag polyprotein precursor to target to the cell membrane. Plasma membrane targeting of the precursor is mediated by the amino-terminal region of polyprotein. To compensate for the assembly block, we substituted the murine leukemia virus matrix coding sequences into an infectious HIV-1 clone. Transfection of murine fibroblasts expressing cyclin T1 with the chimeric proviruses resulted in viruses that were efficiently assembled and released. Chimeric viruses, in which the cytoplasmic tail of the transmembrane subunit, gp41, was truncated to prevent potential interference between the envelope glycoprotein and the heterologous matrix, could infect human and murine cells. They failed to further replicate in the murine cells, but replicated with delayed kinetics in human MT-4 cells. These findings may be useful for establishing a murine model for HIV-1 replication.

Evolution of a human immunodeficiency virus type 1 variant with enhanced replication in pig-tailed macaque cells by DNA shuffling.

DNA shuffling facilitated the evolution of a human immunodeficiency virus type 1 (HIV-1) variant with enhanced replication in pig-tailed macaque peripheral blood mononuclear cells (pt mPBMC). This variant consists exclusively of HIV-1-derived sequences with the exception of simian immunodeficiency virus (SIV) nef. Sequences spanning the gag-protease-reverse transcriptase (gag-pro-RT) region from several HIV-1 isolates were shuffled and cloned into a parental HIV-1 backbone containing SIV nef. Neither this full-length parent nor any of the unshuffled HIV-1 isolates replicated appreciably or sustainably in pt mPBMC. Upon selection of the shuffled viral libraries by serial passaging in pt mPBMC, a species emerged which replicated at substantially higher levels (50 to 100 ng/ml p24) than any of the HIV-1 parents and most importantly, could be continuously passaged in pt mPBMC. The parental HIV-1 isolates, when selected similarly, became extinct. Analyses of full-length improved proviral clones indicate that multiple recombination events in the shuffled region and adaptive changes in the rest of the genome contributed synergistically to the improved phenotype. This improved variant may prove useful in establishing a pig-tailed macaque model of HIV-1 infection.