Depth filtration for clarification of intensified lentiviral vector suspension cell culture

Depth filtration significantly impacts efficiency of lentiviral (LV) vector purification process. However, it is often deprioritized in the overall scope of viral vector manufacturing process optimization. The demand for LV vectors has increased with the rise in disease indications, making it crucial to improve current manufacturing processes. Upstream bioreactor process intensification has enabled cell densities of over 107 viable cells/mL, creating challenges for harvest unit operations. The larger size of LV vectors and their physiochemical similarity to host cell-DNA (HC-DNA) and poor clarification performance causes significant challenges for the subsequent chromatography-based purifications. As a result, a robust and scalable harvest of LV process is needed, especially for LV in vivo therapeutic quality needs. In this study, we systematically evaluated the overlooked yet important issue of depth filtration systems to improve enveloped LV functional vector recovery. We found that an established depth filtration system in process A that provided 94% (n = 6) LV functional recovery could not be translated to intensified Process B cell culture. Hence, the depth filtration process became a bottleneck for the purification performance in an intensified process. We demonstrated an improvement in LV functional vector recovery from 34% to 82% via filter train optimization for an intensified suspension cell culture system (>107 cells/mL with higher titer), while still maintaining a loading throughput of ≥82 L/m2 and turbidity ≤20 NTU. It was demonstrated that the two or three-stage depth filtration scheme is scalable and more suitable for high cell density culture for large scale for LV manufacturing process.

Liver-directed lentiviral gene therapy corrects hemophilia A mice and achieves normal-range factor VIII activity in non-human primates

Liver gene therapy with adeno-associated viral (AAV) vectors delivering clotting factor transgenes into hepatocytes has shown multiyear therapeutic benefit in adults with hemophilia. However, the mostly episomal nature of AAV vectors challenges their application to young pediatric patients. We developed lentiviral vectors, which integrate in the host cell genome, that achieve efficient liver gene transfer in mice, dogs and non-human primates, by intravenous delivery. Here we first compare engineered coagulation factor VIII transgenes and show that codon-usage optimization improved expression 10-20-fold in hemophilia A mice and that inclusion of an unstructured XTEN peptide, known to increase the half-life of the payload protein, provided an additional >10-fold increase in overall factor VIII output in mice and non-human primates. Stable nearly life-long normal and above-normal factor VIII activity was achieved in hemophilia A mouse models. Overall, we show long-term factor VIII activity and restoration of hemostasis, by lentiviral gene therapy to hemophilia A mice and normal-range factor VIII activity in non-human primate, paving the way for potential clinical application.

“Lentiviral gene therapy to the liver establishes stable long-term normal to supra-normal coagulation factor VIII activity in mouse models of hemophilia A and in non-human primates, representing a potential new treatment option for people with hemophilia A.”.

Phagocytosis-shielded lentiviral vectors improve liver gene therapy in nonhuman primates

Liver-directed gene therapy for the coagulation disorder hemophilia showed safe and effective results in clinical trials using adeno-associated viral vectors to replace a functional coagulation factor, although some unmet needs remain. Lentiviral vectors (LVs) may address some of these hurdles because of their potential for stable expression and the low prevalence of preexisting viral immunity in humans. However, systemic LV administration to hemophilic dogs was associated to mild acute toxicity and low efficacy at the administered doses. Here, exploiting intravital microscopy and LV surface engineering, we report a major role of the human phagocytosis inhibitor CD47, incorporated into LV cell membrane, in protecting LVs from uptake by professional phagocytes and innate immune sensing, thus favoring biodistribution to hepatocytes after systemic administration. By enforcing high CD47 surface content, we generated phagocytosis-shielded LVs which, upon intravenous administration to nonhuman primates, showed selective liver and spleen targeting and enhanced hepatocyte gene transfer compared to parental LV, reaching supraphysiological activity of human coagulation factor IX, the protein encoded by the transgene, without signs of toxicity or clonal expansion of transduced cells.

Recombinant factor VIII Fc (rFVIIIFc) fusion protein reduces immunogenicity and induces tolerance in hemophilia A mice

Anti-factor VIII (FVIII) antibodies is a major complication of FVIII replacement therapy for hemophilia A. We investigated the immune response to recombinant human factor VIII Fc (rFVIIIFc) in comparison to BDD-rFVIII and full-length rFVIII (FL-rFVIII) in hemophilia A mice. Repeated administration of therapeutically relevant doses of rFVIIIFc in these mice resulted in significantly lower antibody responses to rFVIII compared to BDD-rFVIII and FL-rFVIII and reduced antibody production upon subsequent challenge with high doses of rFVIIIFc. The induction of a tolerogenic response by rFVIIIFc was associated with higher percentage of regulatory T-cells, a lower percentage of pro-inflammatory splenic T-cells, and up-regulation of tolerogenic cytokines and markers. Disruption of Fc interactions with either FcRn or Fcγ receptors diminished tolerance induction, suggesting the involvement of these pathways. These results indicate that rFVIIIFc reduces immunogenicity and imparts tolerance to rFVIII demonstrating that recombinant therapeutic proteins may be modified to influence immunogenicity and facilitate tolerance.

FcRn Rescues Recombinant Factor VIII Fc Fusion Protein from a VWF Independent FVIII Clearance Pathway in Mouse Hepatocytes

We recently developed a longer lasting recombinant factor VIII-Fc fusion protein, rFVIIIFc, to extend the half-life of replacement FVIII for the treatment of people with hemophilia A. In order to elucidate the biological mechanism for the elongated half-life of rFVIIIFc at a cellular level we delineated the roles of VWF and the tissue-specific expression of the neonatal Fc receptor (FcRn) in the biodistribution, clearance and cycling of rFVIIIFc. We find the tissue biodistribution is similar for rFVIIIFc and rFVIII and that liver is the major clearance organ for both molecules. VWF reduces the clearance and the initial liver uptake of rFVIIIFc. Pharmacokinetic studies in FcRn chimeric mice show that FcRn expressed in somatic cells (hepatocytes or liver sinusoidal endothelial cells) mediates the decreased clearance of rFVIIIFc, but FcRn in hematopoietic cells (Kupffer cells) does not affect clearance. Immunohistochemical studies show that when rFVIII or rFVIIIFc is in dynamic equilibrium binding with VWF, they mostly co localize with VWF in Kupffer cells and macrophages, confirming a major role for liver macrophages in the internalization and clearance of the VWF-FVIII complex. In the absence of VWF a clear difference in cellular localization of VWF-free rFVIII and rFVIIIFc is observed and neither molecule is detected in Kupffer cells. Instead, rFVIII is observed in hepatocytes, indicating that free rFVIII is cleared by hepatocytes, while rFVIIIFc is observed as a diffuse liver sinusoidal staining, suggesting recycling of free-rFVIIIFc out of hepatocytes. These studies reveal two parallel linked clearance pathways, with a dominant pathway in which both rFVIIIFc and rFVIII complexed with VWF are cleared mainly by Kupffer cells without FcRn cycling. In contrast, the free fraction of rFVIII or rFVIIIFc unbound by VWF enters hepatocytes, where FcRn reduces the degradation and clearance of rFVIIIFc relative to rFVIII by cycling rFVIIIFc back to the liver sinusoid and into circulation, enabling the elongated half-life of rFVIIIFc.

A single chain variant of factor VIII Fc fusion protein retains normal in vivo efficacy but exhibits altered in vitro activity

Recombinant factor VIII Fc (rFVIIIFc) is a fusion protein consisting of a single B-domain-deleted (BDD) FVIII linked recombinantly to the Fc domain of human IgG1 to extend half-life. To determine if rFVIIIFc could be further improved by maintaining the heavy and light chains within a contiguous single chain (SC), we evaluated the activity and function of SC rFVIIIFc, an isoform that is not processed at residue R1648. SC rFVIIIFc showed equivalent activity in a chromogenic assay compared to rFVIIIFc, but approximately 40% activity by the one-stage clotting assay in the presence of von Willebrand Factor (VWF), with full activity in the absence of VWF. Moreover, SC rFVIIIFc demonstrated markedly delayed thrombin-mediated release from VWF, but an activity similar to that of rFVIIIFc upon activation in FXa generation assays. Therefore, the apparent reduction in specific activity in the aPTT assay appears to be primarily due to delayed release of FVIII from VWF. To assess whether stability and activity of SC rFVIIIFc were affected in vivo, a tail vein transection model in Hemophilia A mice was utilized. The results demonstrated similar pharmacokinetic profiles and comparable efficacy for SC rFVIIIFc and rFVIIIFc. Thus, while the single chain configuration did not promote enhanced half-life, it reduced the rate of release of FVIII from VWF required for activation. This impaired release may underlie the observed reduction in the one-stage clotting assay, but does not appear to affect the physiological activity of SC rFVIIIFc.

Effects of transient immunosuppression on adenoassociated, virus-mediated, liver-directed gene transfer in rhesus macaques and implications for human gene therapy

In a clinical study of recombinant adeno-associated virus-2 expressing human factor IX (AAV2-FIX), we detected 2 impediments to long-term gene transfer. First, preexisting anti-AAV neutralizing antibodies (NABs) prevent vector from reaching the target tissue, and second, CD8(+) T-cell responses to hepatocyte-cell surface displayed AAV-capsid-terminated FIX expression after several weeks. Because the vector is incapable of synthesizing viral proteins, a short course of immunosuppression, until AAV capsid is cleared from the transduced cells, may mitigate the host T-cell response, allowing long-term expression of FIX. To evaluate coad-ministration of immunosuppression, we studied AAV8 vector infusion in rhesus macaques, natural hosts for AAV8. We administered AAV8-FIX in 16 macaques via the hepatic artery and assessed the effects of (1) preexisting anti-AAV8 NABs, (2) a standard T-cell immunosuppressive regimen, and (3) efficacy and safety of AAV8-FIX. We found that low titers (1:5) of preexisting NABs abrogate transduction, whereas animals with undetectable NABs are safely and effectively transduced by AAV8-FIX. Coadministration of mycophenolate mofetil and tacrolimus with vector does not induce toxicity and does not impair AAV transduction or FIX synthesis. These findings enable a clinical study to assess the effects of immunomodulation on long-term FIX expression in patients with hemophilia B.

Multiyear therapeutic benefit of AAV serotypes 2, 6, and 8 delivering factor VIII to hemophilia A mice and dogs

Hemophilia A, a deficiency of functional coagulation factor VIII (FVIII), is treated via protein replacement therapy. Restoring 1% to 5% of normal blood FVIII activity prevents spontaneous bleeding, making the disease an attractive gene therapy target. Previously, we have demonstrated short-term activity of a liver-specific AAV2 vector expressing canine B-domain-deleted FVIII (cFVIII) in a hemophilia canine model. Here, we report the long-term efficacy and safety of AAV-cFVIII vectors of serotypes 2, 5, 6, and 8 in both hemophilia A mice and dogs. AAV6-cFVIII and AAV8-cFVIII restored physiologic levels of plasma FVIII activity in hemophilia A mice. The improved efficacy is attributed to more efficient gene transfer in liver compared with AAV2 and AAV5. However, supraphysiologic cFVIII levels correlated with the formation of cFVIII-neutralizing antibodies in these mice. Of importance, hemophilia A dogs that received AAV2-cFVIII, AAV6-cFVIII, and AAV8-cFVIII have persistently expressed therapeutic levels of FVIII, without antibody formation or other toxicities, for more than 3 years. However, liver transduction efficiencies are similar between AAV2, AAV6, and AAV8 serotypes in hemophilia A dogs, in contrast to mice. In summary, this is the first report demonstrating multiyear therapeutic efficacy and safety of multiple AAV-cFVIII vectors in hemophilia A dogs and provides the basis for human clinical studies.

Novel caprine adeno-associated virus (AAV) capsid (AAV-Go.1) is closely related to the primate AAV-5 and has unique tropism and neutralization properties

Preexisting humoral immunity to adeno-associated virus (AAV) vectors may limit their clinical utility in gene delivery. We describe a novel caprine AAV (AAV-Go.1) capsid with unique biological properties. AAV-Go.1 capsid was cloned from goat-derived adenovirus preparations. Surprisingly, AAV-Go.1 capsid was 94% identical to the human AAV-5, with differences predicted to be largely on the surface and on or under the spike-like protrusions. In an in vitro neutralization assay using human immunoglobulin G (IgG) (intravenous immune globulin [IVIG]), AAV-Go.1 had higher resistance than AAV-5 (100-fold) and resistance similar to that of AAV-4 or AAV-8. In an in vivo model, SCID mice were pretreated with IVIG to generate normal human IgG plasma levels prior to the administration of AAV human factor IX vectors. Protein expression after intramuscular administration of AAV-Go.1 was unaffected in IVIG-pretreated mice, while it was reduced 5- and 10-fold after administration of AAV-1 and AAV-8, respectively. In contrast, protein expression after intravenous administration of AAV-Go.1 was reduced 7.1-fold, similar to the 3.8-fold reduction observed after AAV-8 administration in IVIG-pretreated mice, and protein expression was essentially extinguished after AAV-2 administration in mice pretreated with much less IVIG (15-fold). AAV-Go.1 vectors also demonstrated a marked tropism for lung when administered intravenously in SCID mice. The pulmonary tropism and high neutralization resistance to human preexisting antibodies suggest novel therapeutic uses for AAV-Go.1 vectors, including targeting diseases such as cystic fibrosis. Nonprimate sources of AAVs may be useful to identify additional capsids with distinct tropisms and high resistance to neutralization by human preexisting antibodies.

Human immunoglobulin inhibits liver transduction by AAV vectors at low AAV2 neutralizing titers in SCID mice

Long-term cures of hemophilia B have been achieved using AAV2 delivering the factor IX gene to the liver of adeno-associated virus (AAV)-naive hemophilic animals. However, the clinical success of this approach requires overcoming pre-existing AAV neutralizing antibodies prevalent in humans. To better define the inhibition of neutralizing antibodies on AAV2-mediated liver transduction, we developed an in vivo passive immunity model. SCID mice were first reconstituted to a defined neutralizing titer with pooled plasma-derived human immunoglobulin. AAV2-FIX vectors then were administered to the liver, and the transduction efficiency was measured by plasma FIX levels. Unexpectedly, AAV2 neutralizing titers lower than 1:10 were sufficient to neutralize 4 to 20 x 10(12) vg/kg of AAV2 vectors in vivo, a capacity that was underestimated by in vitro neutralizing assays. We also evaluated strategies to evade neutralization, including the use of alternative delivery routes, infusion parameters, empty capsids, and alternative AAV serotypes 6 and 8. The results indicate that low AAV2 neutralizing titers can be inhibitory to the tested human and primate AAV vectors delivered into the circulatory system. Therefore, novel nonprimate AAV vectors or compartmentalized delivery may offer more consistent therapeutic effects in the presence of pre-existing AAV neutralizing antibodies.

Regulatable liver expression of the rabbit apolipoprotein B mRNA-editing enzyme catalytic polypeptide 1 (APOBEC-1) in mice lacking endogenous APOBEC-1 leads to aberrant hyperediting

Apolipoprotein (apo) B mRNA editing is the deamination of C(6666) to uridine, which results in translation of the apoB-48 protein instead of the genomically encoded apoB-100. ApoB-48-containing lipoproteins are cleared more rapidly from plasma and are less atherogenic than apoB-100-containing low-density lipoproteins (LDLs). In humans, the intestine predominantly produces apoB-48 whereas the liver secretes apoB-100 only. To evaluate a potential therapeutic use for liver-induced apoB mRNA editing in humans, we investigated the efficiency and safety of transgenic expression of apoB mRNA-editing enzyme catalytic polypeptide 1 (APOBEC-1) in the absence of endogenous editing in the mouse model. Here we show that regulatable tetO-mediated APOBEC-1 expression in the livers of gene-targeted mice lacking endogenous APOBEC-1 results in 30% apoB mRNA editing. In a time-course experiment, the expression of tetO-APOBEC-1 mRNA was suppressed within 2 days after mice were fed doxycycline and apoB mRNA editing and apoB-48 formation were suppressed within 4 days. However, tetO-APOBEC-1 expression resulted in regulatable aberrant hyperediting of several cytidines downstream of C(6666) in apoB mRNA and in novel APOBEC-1 target 1 (NAT1) mRNA. Several of the cytidines in apoB mRNA were hyperedited to a level similar to that of C(6666), although editing at C(6666) was lower than that in wild-type mice. These results demonstrate that even moderate APOBEC-1 expression can lead to hyperediting, limiting the single-gene approach for gene therapy with APOBEC-1.

A dinucleotide deletion in amyloid precursor protein (APP) mRNA associated with sporadic Alzheimer's disease results in efficient secretion of truncated APP isoforms from neuroblastoma cell cultures

Recently, two dinucleotide deletions were detected in the mRNA of the amyloid precursor protein (APP) from cerebral cortex neurons of patients with sporadic Alzheimer's disease (AD) or Down's syndrome. These deletions resulted in truncation of APP, producing an APP isoform with a 38-kDa N-terminus and a novel carboxyl terminus (APP+1). We investigated the subcellular localization and the processing of APP+1 in the neuroblastoma cell line B103. cDNA constructs were generated encoding fusion proteins of APP+1 or full-length APP with the enhanced green fluorescent protein (eGFP). In transient transfection experiments using B103 cells, the APP+1-eGFP fusion protein showed a reticular localization with intense staining in the Golgi complex. Unlike full-length APP fused to eGFP, the APP+1-eGFP fusion protein did not localize to the perinuclear area or to the plasma membrane. Western blot analysis of cell extracts confirmed the translation of the expected fusion proteins. Analysis of the supernatant by western blot indicated that the APP+1-eGFP fusion protein was efficiently secreted from B103 cells, whereas the secreted form of full-length APP fusion protein (APPs) was hardly detectable. Thus, both dinucleotide deletions in the APP mRNA result in truncated APP+1 that is not membrane associated and is readily secreted from neurons.

Phylogenetic analysis of the apolipoprotein B mRNA-editing region. Evidence for a secondary structure between the mooring sequence and the 3' efficiency element

Apolipoprotein (apo) B mRNA editing is the deamination of C(6666) to uridine, which changes the codon at position 2153 from a genomically encoded glutamine (CAA) to an in-frame stop codon (UAA). The apoB mRNA-editing enzyme complex recognizes the editing region of the apoB pre-mRNA with exquisite precision. Four sequence elements spanning 139 nucleotides (nt) on the apoB mRNA have been identified that specify this precision. In cooperation with the indispensable mooring sequence and spacer element, a 5' efficiency element and a 3' efficiency element enhance editing in vitro. A phylogenetic comparison of 32 species showed minor differences in the apoB mRNA sequence, and the apoB mRNA from 31 species was robustly edited in vitro. However, guinea pig mRNA was poorly edited. Compared with the consensus sequences of these 31 species, guinea pig apoB mRNA has three variations in the 3' efficiency element, and the conversion of these to the consensus sequence increased editing to the levels in the other species. From this information, a model for the secondary structure was formulated in which the mooring sequence and the 3' efficiency element form a double-stranded stem. Thirty-one mammalian apoB mRNA sequences are predicted to form this stem positioning C(6666) two nucleotides upstream of the stem. However, the guinea pig apoB mRNA has a mutation in the 3' efficiency element (C(6743) to U) that predicts an extension of the stem and hence the lower editing efficiency. A test of this model demonstrated that a single substitution at 6743 (U to C) in the guinea pig apoB mRNA, that should reduce the stem, enhanced editing, and mutations in the 3' efficiency element that extended the stem for three base pairs dramatically reduced editing. Furthermore, the addition of a 20-nucleotide 3' efficiency element RNA, to a 58-nucleotide guinea pig apoB mRNA lacking the 3' efficiency element more than doubled the in vitro editing activity. Based on these results, a model is proposed in which the mooring sequence and the 3' efficiency element form a double-stranded stem, thus suggesting a mechanism of how the 3' efficiency element enhances editing.