ALBERT

Description: Messenger RNA (mRNA) therapeutics have been explored to treat various genetic disorders. Lipid-derived nanomaterials are currently one of the most promising biomaterials that mediate effective mRNA delivery. However, efficiency and safety of this nanomaterial-based mRNA delivery remains a challenge for clinical applications. Here, we constructed a series of lipid-like nanomaterials (LLNs), named functionalized TT derivatives (FTT), for mRNA-based therapeutic applications in vivo. After screenings on the materials, we identified FTT5 as a lead material for efficient delivery of long mRNAs, such as human factor VIII (hFVIII) mRNA (~4.5 kb) for expression of hFVIII protein in hemophilia A mice. Moreover, FTT5 LLNs demonstrated high percentage of base editing on PCSK9 in vivo at a low dose of base editor mRNA (~5.5 kb) and single guide RNA. Consequently, FTT nanomaterials merit further development for mRNA-based therapy.

Description: During intracellular processing, factor VIII (FVIII) undergoes proteolysis at multiple sites with the most predominant cleavage at a Paired basic Amino acid Cleaving Enzyme (PACE)-furin cleavage site at the carboxy-terminus of the B-domain at residue 1648 to give rise to two polypeptide chains, the heavy chain and the light chain. Through a metal ion dependent association, these chains form a heterodimer that is the secreted form of the protein. Previously, we observed that canine B-domain deleted FVIII (cFVIII-BDD) is secreted primarily as a single chain (SC) molecule (170kD) while human FVIII-BDD (hFVIII-BDD) is secreted predominantly as a heterodimer. In addition, cFVIII-BDD is more stable and has higher biological activity. Amino acid sequence analysis revealed a single amino acid difference between cFVIII (1645-HHQR-1648) and hFVIII (1645-RHQR-1648) at the PACE-furin cleavage recognition site. Characterization of hFVIII-R1645H demonstrated that this residue is responsible for the secretion of cFVIII predominantly as a single polypeptide chain and its inherent increased stability and specific activity. Furthermore, hFVIII-R1645H has enhanced hemostatic effects upon vascular injury and in the setting of AAV delivery was associated with increased circulating levels compared to wild type hFVIII-BDD. These data suggest that there is suboptimal cleavage of cFVIII and hFVIII-R1645H by PACE-furin. Here we tested whether deletion of part or all of the PACE-furin recognition sequence may further decrease the efficiency of cleavage at the site resulting in a larger portion of single SC molecules and thus may confer increased stability compared to hFVIII-R1645H. A series of hFVIII-BDD deletion variants (n=5) of the PACE-furin cleavage recognition site (1645-1648) were generated: del1645, del1645-46, del1645-47, del1645-48 and del1648. Stable BHK cell lines were established for each variant and protein was purified using ion exchange chromatography. On an SDS-PAGE gel under reducing conditions the deletion variants migrate in a similar manner, however, the proportion of the hFVIII in the single chain form compared to hFVIII-BDD (15% SC) was 3.5-fold higher for del1645-46 (57% SC), del1645-47 (53% SC) and del1645-48 (48% SC) which is similar to R1645H (52% SC). The single deletion variants on SDS-PAGE were 2-fold higher (del1645, 34% SC) or identical (del1648, 15% SC) to hFVIII-BDD. In a one-stage aPTT assay, all variants had activity similar to hFVIII-BDD. While in the two-stage aPTT assay, three variants (del1645-46, del1645-47, del1645-48) had 2-fold higher activity than hFVIII-BDD. Variants del1645 and del1648 had similar activity to hFVIII-BDD. Using an A2 domain dissociation clotting-based assay, we found that del1645-46, del1645-47, del1645-48 and del1648 were more stable following thrombin (IIa) activation compared to hFVIII-BDD suggesting that the A2 domain of these variants may dissociate more slowly than the wild type human A2 domain following IIa cleavage. These findings explain, at least in part, the superior clotting activity determined in multiple assays. The hFVIII PACE-furin variants were also introduced into an adeno-associated viral vector serotype 8 (AAV8-hAAT-hFVIII-BDD) for liver targeted delivery. Each AAV8-hAAT-hFVIII-P/F variant was delivered to hemophilia A mice (5x1011vector genomes/mouse)(n=6/group). At 4 weeks post vector administration, the hFVIII expression of del1645-47 was 3-fold higher than hFVIII-BDD followed by del1645, del1645-48, del1645-46 while del1648 was similar to hFVIII-BDD. In summary, these in vitro data demonstrate that deletion of residues 1645-46, 1645-47 or 1645-48 of the PACE-furin cleavage recognition site confers 2-3-fold higher biological activity than wild type hFVIII-BDD. In vivo these variants result in 2-3 fold higher levels of hFVIII expression after AAV delivery compared to wild type hFVIII-BDD. Together these data suggest that the hFVIII-P/F deletion variants have enhanced hemostatic function and are superior to wild type hFVIII. In the setting of gene-based therapeutics for hemophilia A, these novel variants provide a unique strategy for increasing factor VIII expression which will allow the use of a lower vector dose which is a critical improvement for translation of this approach. Disclosures Nguyen: Pfizer, Inc.: Research Funding. Camire:Pfizer: Patents & Royalties, Research Funding. Sabatino:Spark Therapeutics: Research Funding; Pfizer, Inc.: Research Funding.

Description: During intracellular processing, B-domain deleted human factor VIII (hFVIII-BDD) undergoes proteolysis at multiple sites with the most predominant cleavage at the furin recognition site (1645-RHQR-1648) at the carboxy-terminus of the B-domain. This cleavage gives rise to two polypeptide chains, the 90 kDa heavy chain (HC) and the 80 kDa light chain (LC), that form a heterodimer which becomes the major secreted form of the protein. In contrast, B-domain deleted canine factor VIII (cFVIII-BDD) is secreted primarily in the uncleaved single chain (SC) form, exhibits higher activity and shows increased expression in the gene therapy setting. To investigate the difference in the amino acid sequence at the furin recognition site between cFVIII-BDD (1645-HHQR-1648) and hFVIII-BDD, our previous work generated and characterized a series of hFVIII-BDD furin site deletion variants. Our studies demonstrated that hFVIII- BDD-del1645-1647 (Δ3) variant is secreted predominantly as SC, with a two-fold increase in activity and improved secretion over hFVIII-BDD. N-terminal sequencing of the hFVIII LC revealed that while hFVIII-BDD is cleaved after R1648 and S1657, cleavage of Δ3 occurs exclusively after S1657. Amino acid sequence analysis revealed differences at this site between hFVIII (S1657/D1658) and cFVIII (P1657/E1658). We hypothesized that modifications to this downstream cleavage site in hFVIII-BDD may further decrease cleavage efficiency, leading to a higher portion of SC, and further increase protein expression after gene therapy. We generated a new variant where this downstream cleavage sequence was modified to that of cFVIII, S1657P/D1658E, in combination with our previously described Δ3 variant. hFVIII-BDD, hFVIII-Δ3, and hFVIII-Δ3-S1657P/D1658E were purified using ion exchange chromatography. Interestingly, optical densitometry analysis of all purified proteins on a reducing SDS-PAGE gel revealed that hFVIII-Δ3-S1657P/D1658E was expressed almost entirely as a SC (91.7% ± 5.1%), similar to what was observed with cFVIII (86.8% ± 1.0%). hFVIII-Δ3 and hFVIII-BDD were 56.5% ± 5.7% and 24.3% ± 0.6% SC, respectively. Once activated by thrombin, hFVIII-Δ3-S1657P/D1658E yielded the same expected species as hFVIII-Δ3 and hFVIII-BDD. In a one-stage aPTT assay, hFVIII-Δ3-S1657P/D1658E had activity comparable to hFVIII-Δ3 and hFVIII-BDD. In the two-stage aPTT assay, hFVIII-Δ3 and hFVIII-Δ3-S1657P/D1658E exhibited a two-fold increase in activity over hFVIII-BDD. Thus, hFVIII-Δ3-S1657P/D1658E was secreted primarily in the SC form and had higher activity similar to the hFVIII-Δ3 variant in vitro. In the setting of gene-based therapeutics, strategies to increase hFVIII expression provide a platform for reducing the vector dose required to achieve therapeutically relevant FVIII levels. To investigate the efficacy of this variant in vivo, the hFVIII-Δ3-S1657P/D1658E variant was introduced into an adeno-associated viral vector serotype 8 expressing the wild type hFVIII-BDD cDNA sequence (AAV8-hAAT-hFVIII-BDD) for liver-targeted delivery. The vector was delivered to immune deficient hemophilia A mice (5x1011 vector genomes/mouse) (n=4/group). At 12 weeks post-vector administration, the hFVIII-Δ3-S1657P/D1658E had a 4.5-fold increase in protein expression over hFVIII-BDD, a significant improvement over hFVIII-Δ3, which was two-fold higher than hFVIII-BDD. Altogether, the data supports that the modifications at the downstream cleavage site at residue S1657 and D1658, in combination with a furin evading variant in hFVIII-BDD, further hinders cleavage during intracellular processing. This unexpectedly leads to additional enhancement of hFVIII-BDD expression in the gene therapy setting. Not only might the studies of these variants improve our understanding of the intracellular processing of hFVIII-BDD, but they also provide a novel approach to increase hFVIII expression that will allow the use of a lower vector dose, improving the safety of gene-based therapeutics. Disclosures Sabatino: Spark Therapeutics: Research Funding.

Description: The hemophilia A (HA) dogs have been a valuable model for evaluating the efficacy of novel hemophilia therapeutics. These dogs have a mutation that is analogous to the most common mutation in humans with severe disease, the intron 22 inversion. The advantages of the hemophilia dog model include: (1) it was predictive of the therapeutic adeno-associated viral (AAV) vector dose in human factor IX (hFIX) clinical trials, (2) it is an outbred immunocompetent species, (3) it demonstrates a clinical bleeding phenotype consistent with patients, including hemarthrosis, and (4) it permits long-term investigation of both efficacy and safety. Our previous studies delivering AAV-canine factor VIII (cFVIII) in the HA dogs demonstrated long-term (up to 10 years of follow-up) dose-dependent cFVIII expression without evidence of an immune response to the cFVIII protein. In hemophilia B preclinical studies, the comparable biological characteristics between canine FIX and hFIX allows preclinical results using cFIX to be translated to clinical studies. In contrast, for hemophilia A, our studies of recombinant cFVIII and human FVIII (hFVIII) proteins demonstrate that cFVIII is a more stable protein that has higher biological activity and is secreted better than hFVIII (Sabatino et al. 2009). Thus, expression of cFVIII following AAV delivery does not accurately predict the therapeutic dose of AAV-hFVIII which is relevant for translation to clinical trials. The challenge of administering AAV-hFVIII to the HA dogs is that this expressed xenoprotein (hFVIII) results in inhibitor formation that precludes the ability to measure transgene (hFVIII) expression. We hypothesized that tolerizing HA dogs to hFVIII will (1) permit accurate evaluation of hFVIII expression and thus predict the therapeutic vector dose and (2) allow the evaluation of the potential immune response of AAV8-hFVIII versus a novel hFVIII variant that has increased activity and secretion (Nguyen et al. 2017). In this study we used a neonatal retroviral (RV) delivery approach to tolerize the HA dogs (n=5) to B-domain deleted hFVIII (hFVIII-BDD). HA neonatal male dogs (S28, S29, V06, V26, V27) were treated with the retrovirus (RV-hAAT-hFVIII-BDD-WPRE) (3x10e9 TU/kg) on day 2 of life. The levels of hFVIII expression after RV delivery were 0.3-6% 4 weeks after RV delivery and plateaued after 6 months to 0.8% (S28), 0.3% (S29), 0% (V06), 1.5% (V26) and 1.7% (V27) based on Coatest assay. At 5-6 months of life the dogs (S28, S29, V06) were challenged with hFVIII-BDD (Xyntha; 25IU/kg; I.V.) weekly for 6 consecutive weeks. Samples were collected before and 15 minutes after each protein challenge to demonstrate the successful infusion of the protein. At 4 weeks after the final challenge, no anti-hFVIII IgG1 or IgG2 antibodies were detected consistent with no evidence of an inhibitor. At 4.5 years of age, S28 and S29 had stable expression of 0.5-1% of hFVIII and were rechallenged with hFVIII-BDD (Xyntha; 25IU/kg per week x 6 wks). No anti-hFVIII IgG1 or IgG2 antibodies were detected 4 weeks after the final protein challenge. Thus, these data demonstrate that all of the RV-hAAT-hFVIII-BDD-WPRE treated dogs that have been challenged (n=3) have been tolerant to hFVIII-BDD. Since the goal of this study is to generate a cohort of HA dogs tolerant to hFVIII-BDD to address the efficacy and safety of AAV8-hFVIII-BDD versus a hFVIII-BDD variant, we treated the first dog (S29) 12 weeks after the second series of protein challenges with an optimized AAV vector cassette containing a codon-optimized hFVIII sequence with a modified transthyretin (TTRm) promoter, AAV8-TTRm-hFVIII-BDD (2x10e12 vg/kg). Prior to AAV administration the hFVIII activity was 0.2%. At 8 weeks post AAV administration, the hFVIII activity increased to 3.5%. No anti-hFVIII-BDD IgG1 or IgG2 was detected after AAV-hFVIII administration. These data demonstrate that low levels of sustained hFVIII expression of 0.2-2% up to 4 years post-retroviral delivery were able to induce and maintain tolerance to hFVIII. Overall, these studies demonstrate that the neonatal HA dogs treated with a retrovirus targeting hFVIII expression to the liver are tolerant to hFVIII and provide a unique large animal model to evaluate both efficacy as well as potential immunogenicity of our novel FVIII variants. Disclosures Sabatino: Spark Therapeutics: Patents & Royalties.

Description: The hemophilia dog models are valuable for evaluating the efficacy of novel hemophilia therapeutics. The hemophilia B dog was predictive of the therapeutic dose of adeno-associated viral (AAV) vector delivery of human factor IX in clinical trials. In addition, it provides an opportunity to study the long-term efficacy and safety after AAV administration. However, there are several challenges in using the hemophilia A (HA) dog model for gene therapy studies. First, canine factor VIII (cFVIII) has higher specific activity and increased rate of secretion compared to human FVIII (hFVIII). This significant difference between cFVIII and hFVIII prevents the use of a species-specific transgene to predict the efficacy of AAV-hFVIII. Second, the HA dogs are immunocompetent and develop an immune response to the xenoprotein, hFVIII, that precludes the ability to measure transgene expression. Therefore, in order to employ this valuable model for gene therapy studies, we generated a unique cohort of HA dogs that are tolerant to B-domain deleted (BDD) hFVIII. We hypothesized that tolerizing dogs to hFVIII will (1) permit accurate evaluation of hFVIII expression and thus predict the therapeutic vector dose and (2) allow the evaluation of the potential immune response to a novel hFVIII variant. To tolerize the dogs to hFVIII, neonatal HA dogs were treated with a retrovirus (RV-hAAT-hFVIII-BDD-WPRE, 3x109 TU/kg) (n=5) on day 2 of life. The hFVIII expression was between 0.3%-6% at 4 weeks after RV delivery and plateaued after 6 months to 0.8% (S28), 0.3% (S29), 0% (V06), 1.5% (V26) and 1.7% (V27) based on Coatest assay. To determine if the dogs were tolerant to hFVIII-BDD, the dogs were challenged with hFVIII-BDD protein at 5-6 months post-RV administration (Xyntha, 25IU/kg per wk x 6 wks, I.V.). Anti-hFVIII antibodies were monitored closely throughout the challenge and up to 8 weeks after the last challenge. In 4 out of 5 dogs, no anti-hFVIII immune response was observed based on IgG1, IgG2, total IgG or Bethesda titer. In contrast, naïve HA dogs (n=2) developed high level anti-hFVIII IgG2 (1.2-3.2 μg/mL), total IgG (3.4-5.0 μg/mL), and Bethesda titer (4.1-67.8 BU/mL) after the same challenge regimen. Interestingly, the hFVIII activity in one RV-treated dog (V06) was undetectable at 6 months post-RV administration. After the challenge, V06 had anti-hFVIII IgG2 (1.7 μg/mL), total IgG (2.6 μg/mL), and a Bethesda titer (9.5 BU/mL), suggesting that FVIII must be maintained to achieve tolerance. These dogs were used to evaluate the efficacy of AAV serotype 8 (AAV8) delivery of a hFVIII-BDD codon-optimized sequence driven by a hepatocyte promoter, modified transthyretin promoter (TTRm). S29 was delivered AAV8-TTRm-hFVIII-CO (2x1012 vg/kg). Prior to AAV delivery, the levels of hFVIII activity were 0.5-1% from the tolerization with the RV. After AAV administration the hFVIII activity was 3.8% at d168 and 4.7% at d387, resulting in a 4% increase in hFVIII expression. No anti-hFVIII antibodies were detected. The annual bleeding rate (ABR) for S29 post-RV delivery was 5 and after AAV delivery was 0, showing an improvement in the bleeding phenotype in contrast to untreated HA dogs (ABR=13, n=11). A hFVIII-tolerized littermate, S28, was recently treated with a hFVIII variant, AAV8-TTRm-hFVIII-CO-Δ3-SP/DE (2x1012 vg/kg). The hFVIII-Δ3-SP/DE variant has a deletion of the furin site (1645-47) and replaces residues SD at 1657-58 with PE. This variant showed higher specific activity (2-fold) in vitro and increased secretion (4-fold) compared to wild type hFVIII-BDD in the setting of AAV delivery in HA mice. Based on the results in S28, we will determine the dosing of V26 and V27. These studies demonstrate that sustained low level hFVIII expression of 0.2-2% up to 4 years post-retroviral delivery were able to induce and maintain tolerance to hFVIII, while allowing for the subsequent assessment of AAV efficacy. A clinically relevant dose of AAV8-TTRm-hFVIII-CO resulted in therapeutic levels of hFVIII expression while ongoing studies will allow investigation of the efficacy of the hFVIII-BDD variant, Δ3-SP/DE, in the setting of AAV administration in a large animal model. Overall, these studies demonstrate that RV-targeting of hFVIII-BDD expression to the liver in neonatal HA dogs leads to tolerance to this xenoprotein and provide a unique large animal model to evaluate both efficacy as well as potential immunogenicity of novel FVIII variants. Disclosures Sabatino: Spark Therapeutics: Patents & Royalties.

Description: Hemophilia is an X-linked bleeding disorder caused by a deficiency in clotting factor VIII (FVIII)(hemophilia A, HA) or factor IX (FIX)(hemophilia B, HB). While early clinical trials of AAV delivery of FIX for HB have demonstrated stable FIX expression for 〉8 years, an ongoing clinical trial of AAV-FVIII delivery for HA achieved high levels of transgene expression that unexpectedly declined after 1 year. Here we describe preclinical studies of AAV-canine FVIII (cFVIII) delivery in nine HA dogs with sustained FVIII expression for the duration of the study, as long as 10 years. FVIII was delivered using two delivery approaches: (1) co-administration of two AAV vectors encoding separate cFVIII heavy and light chains driven by the thyroxine binding globulin (TBG) promoter (Two chain approach)(TC) (n=5) at two AAV doses (2.5 x 1013vg/kg; F24, Woodstock, J60) and (1.2 x 1013vg/kg; Linus, H19) or (2) delivery of cFVIII as a single chain driven by the human alpha-1 anti-trypsin (hAAT) promoter (Single chain approach)(SC)(n=4) at two AAV doses (4 x 1013 vg/kg; M50, M06) and (2 x 1013vg/kg; M66, L51) (Sabatino 2011). We demonstrated that both strategies were efficacious; preventing 〉95% of spontaneous bleeding episodes without toxicity. We now report the long-term follow-up of between 2.2 and 10.1 years for these treated dogs. Dose-dependent cFVIII:C (Coatest SP4 FVIII) was observed. At the final time point, the cFVIII:C was 2.7% (F24), 7.1% (Woodstock), 4.5% (J60), 11.3% (Linus) and 2.5% (H19) for TC dogs. For the SC dogs, the cFVIII:C was 9.4% (M06), 10.3% (M50), 1.9% (L51) and 3.7% (M66). Stable FVIII expression was maintained for seven of the dogs over the course of the study. Two dogs (Linus, M50) had a gradual increase in FVIII:C that began about three years after vector administration and continued for an additional seven years (Linus) and four years (M50), until the termination of the study. Liver function tests, serum alpha-fetoprotein concentrations, fibrinogen levels as well as liver pathology did not suggest altered liver function or tumor development in Linus and M50 compared to the other dogs. Clinically, there was no evidence of malignancy and no tumors were detected at the time of necropsy in any dog. One of the safety concerns for AAV-mediated gene therapy approaches is the potential for AAV integration events to be genotoxic and lead to tumorigenesis. While recombinant AAV primarily remains as an episome, integration events have been observed in mouse models and hepatocellular carcinoma has been observed after neonatal delivery of AAV vectors. In addition, the increase in FVIII expression in Linus and M50 prompted us to investigate integration and clonal expansion as a potential mechanism for these observations. Vector copy number (VCN) analysis was performed on liver samples (5-29 per dog, n=8 dogs) by Q-PCR and detected DNA copy numbers between 0.0 and 7.8 per diploid genome (Fig 1A). We performed integration target site analysis on liver samples (n=3/dog) from six of the AAV-treated HA dogs and naïve HA dogs (n=2) by ligation-mediated PCR, Ilumina paired-end sequencing and analysis using the custom software pipeline, AAVenger. Analysis of the 20 samples identified 〉2,000 unique AAV integration events (IE). There was a correlation between the DNA copy number and the number of integration events detected. Clonal abundances were estimated by counting the unique genome breaks associated with integration positions, which showed that the maximum clonal abundance ranged from 1 to 138. The integration events were distributed across the canine genome. Clonal expansions were observed with integration near genes previously associated with growth control and transformation in humans (Fig 1B) with the most abundant clones located in DLEU2L (Linus), PEBP4 (J60) and EGR3 (M50). Integration events in EGR3, EGR2, CCND1, LTO1 and ZNF365 were detected in multiple dogs. Validation of integration sites in the most abundant clones was performed using targeted PCR to isolate junction fragments followed by Sanger sequencing. While AAV integration and clonal expansion was observed, the dogs had no evidence for tumorigenesis and it is not clear if the increase in FVIII expression is associated with the clonal expansions detected. Overall, these studies demonstrate long-term sustained FVIII expression for up to 10 years with clonal expansion, but without clinical adverse events after AAV-mediated gene therapy for hemophilia. Disclosures Sabatino: Spark Therapeutics: Patents & Royalties.

Description: Quantum dots are both excellent single-photon sources and hosts for single spins. This combination enables the deterministic generation of Raman-photons—bandwidth-matched to an atomic quantum-memory—and the generation of photon cluster states, a resource in quantum communication and measurement-based quantum computing. GaAs quantum dots in AlGaAs can be matched in frequency to a rubidium-based photon memory, and have potentially improved electron spin coherence compared to the widely used InGaAs quantum dots. However, their charge stability and optical linewidths are typically much worse than for their InGaAs counterparts. Here, we embed GaAs quantum dots into an n-i-p-diode specially designed for low-temperature operation. We demonstrate ultra-low noise behaviour: charge control via Coulomb blockade, close-to lifetime-limited linewidths, and no blinking. We observe high-fidelity optical electron-spin initialisation and long electron-spin lifetimes for these quantum dots. Our work establishes a materials platform for low-noise quantum photonics close to the red part of the spectrum.