Platelet-Targeted Hyperfunctional Factor IX (FIX) Gene Therapy for Hemophilia B Mice with Pre-Existing Anti-FIX Immunity

Hemophilia B (HB) is a prime model for gene therapy. While data from current clinical trials using AAV-mediated liver-targeted FIX gene therapy are very encouraging, this protocol can only be applied to adults without liver disease or anti-AAV antibodies. Thus, developing another gene therapy protocol is desired. Our previous studies have demonstrated that platelet-targeted FIX expression driven by the platelet-specific αIIb promoter (2bF9) restores hemostasis and induces immune tolerance in HB mice (Chen et al. Mol Ther 2014). To improve the efficacy, we used a codon-optimized hyperfunctional FIX Padua (2bCoF9R338L) to replace the normal FIX expression cassette. We showed a 5.8-fold higher platelet-FIX antigen (plt-F9:Ag) and a 28-fold activity (plt-F9:C) levels in HB mice (the non-inhibitor model), respectively, in the 2bCoF9R338L group compared to the 2bF9 group. Here we evaluate 2bCoF9R338L gene therapy in HB mice with preexisting anti-F9 immunity (the inhibitor model). Both donor and recipient HB mice were immunized with rhF9 in the presence of Incomplete Freund's Adjuvant to induce inhibitor development. Plt-F9 expression was introduced by 2bCoF9R338L lentivirus transduction of Sca-1+ cells followed by syngeneic transplantation into primed recipients preconditioned with either a lethal 11Gy or a sub-lethal 6.6Gy total body irradiation. Animals were analyzed starting at 4 weeks after transplantation. FACS analysis showed that there was 21.4 ± 11.4% (n = 6) transduced platelets in the 11Gy group, which was not significantly different compared to the 6.6Gy group (12.5 ± 10.9%, n = 7) or the 6.6Gy non-inhibitor model (17.7 ± 11.7%, n = 4). Plt-F9:C levels in the 11Gy and 6.6Gy groups were 18.68 ± 8.61 and 9.17 ± 12.22 mU/108 platelets, respectively. Plt-F9:Ag levels in the two groups were 3.96 ± 1.57 and 1.70 ± 1.64 mU/108 platelets, respectively, which were not significantly different compared to that obtained in the non-inhibitor model. To determine if plt-F9 is γ-carboxylated, BaSO4 precipitation was performed on platelet lysates from 2bCoR338L-transduced recipients. Diluted platelet lysate samples were treated with or without BaSO4 , and the remaining unprecipitated FIX (meaning un-γ-carboxylated) in supernatants was determined by FIX ELISA. No F9:Ag remained in the supernatants of BaSO4-treated platelet lysates, suggesting that FIX can fully undergo functional γ-carboxylation in platelets. Of note, no anaphylaxis occurred in primed HB mice after platelet gene therapy. The inhibitor titers declined to undetectable with time in recipients after treatment with a half-life (t1/2) of inhibitor reduction in the 11Gy group of 49.2 ± 3.9 days, which was significantly shorter than in the 6.6Gy group (82.6 ± 32.6 days) and in the untransduced 6.6Gy control group (77.7 ± 8.7 days), indicating that plt-F9 does not trigger a memory immune response in primed F9null mice. Since our previous studies have shown that the clinical efficacy of plt-F9 is limited in the presence of anti-F9 inhibitors (Zhang et al. Blood 2010), we did not assess the bleeding phenotype in recipients until inhibitor titers dropped to an undetectable level. When a 6-hour tail bleeding test was used to assess the clinical efficacy of 2bCoF9R338L gene therapy, the bleeding time in the 11Gy and 6.6Gy groups was 2.5 ± 1.3 and 2.5 ± 1.1 hours, respectively, which was similar to that obtained in the wild-type (WT) control (1.7 ± 0.9 hours). In contrast, none of the F9null mice clotted within 6 hours. The remaining hemoglobin levels after the test were 79.5 ± 15.5% and 64.7 ±23.4% in the 11Gy and 6.6Gy groups, respectively, which were not significantly different compared to the WT group (70.6 ± 13.9%), but significantly higher than in the F9null control (38.8 ± 6.7%). We used ROTEM to analyze the whole blood clotting time (WBCT). WBCT in the 2bCoF9R338L-transduced group was significantly shorter than in the F9null control group (1328 ± 236 vs. 2426 ± 476 sec, respectively), confirming that hemostasis was improved in 2bCoF9R338L transduced recipients. Together, our data demonstrated that 2bCoF9R338L gene therapy could efficiently introduce sustained levels of plt-F9 in HB mice even with pre-existing immunity, resulting in antibody eradication and phenotypic correction when inhibitors diminished, suggesting that 2bCoF9R338L gene therapy can be a promising protocol for all HB patients, including patients with inhibitors. Disclosures No relevant conflicts of interest to declare.

Platelet-Derived Codon-Optimized Hyperfunctional FIX Gene Therapy of Hemophilia B Mice

Our previous studies demonstrate that targeting factor IX (FIX) expression to platelets under control of the platelet-specific αIIb promoter (2bF9) can restore hemostasis and induce immune tolerance in hemophilia B (HB) (FIXnull) mice (Chen, et al. Mol Ther 2014). However, functional platelet-FIX activity levels in transduced mice were only around 3% in whole blood even when a lethal 11Gy total body irradiation (TBI) was employed. To reduce potential toxicities associated with this gene therapy protocol from preconditioning and insertion site-mediated mutagenesis, it is desirable to optimize our vector for better clinical efficacy and safety. Recent studies have demonstrated that the combined effect of codon optimization and hyperfunctional FIX Padua can significantly enhance the efficacy of hepatocyte-targeted FIX gene therapy in HB. Thus, we engineered a novel lentiviral vector, 2bCoF9R338L, in which codon-optimized FIX Padua was used to replace the normal FIX expression cassette in our 2bF9 construct. FIXnull mice that we used in this study were originally developed by Lin, et al. (Blood 1997). Platelet-FIX expression was introduced by 2bCoF9R338L lentivirus transduction and syngeneic transplantation under a clinically relevant non-myeloablative preconditioning regimen 6.6Gy TBI. The levels of FIX expression were determined by ELISA for FIX antigen (FIX:Ag) and chromogenic assay for functional FIX activity (FIX:C). Both antigen and activity levels of FIX in platelets from 2bCoF9R338L-transduced recipients were significantly higher than those from normal 2bF9LV-transduced animals. There are approximately a 5.8-fold higher antigen (10.9±3.9 vs. 1.9±1.3 mU/108 platelets) and 28-fold activity (29.1±9.8 vs. 1.1±0.3 mU/108 platelets) levels, respectively, in the 2bCoF9R338L group compared to the 2bF9 group. Flow cytometry analysis showed that 17.7±5.8% of platelets expressed hFIX, which was not significantly different from the 2bF9 group (14.8±10.7%), demonstrating that lentivirus harboring 2bCoF9R338L has similar transduction efficiency as the 2bF9 lentivirus. To assess whether the bleeding phenotype was rescued in FIXnull mice after receiving 2bCoF9R338L-transduced HSCs, we used a 6-hour tail bleeding test. All 2bCoF9R338L-transduced recipients' tail bleeding clotted within 6 hours with a clotting time of 2.5±0.6 hours and the remaining hemoglobin level of 69.3±8.8%, which were not significantly different from those of the wild type controls (1.9±0.3 hours and 67.2±4.2%). In contrast, none of the FIXnull control mice clotted within 6 hours and the remaining hemoglobin level (40.5±1.9%) was significantly lower than in the 2bCoF9R338L group. To investigate whether anti-FIX immune tolerance was induced in 2bCoF9R338L-transduced recipients, 6 months after HSCT, animals were immunized with recombinant human FIX (rhF9) at a dose of 200 U/kg via intraperitoneal injection two times with a 3-week interval, and the anti-FIX inhibitory antibodies (inhibitors) were determined by Bethesda assay. We found that none of the FIXnull mice that received 2bCoF9R338L-transduced HSCs developed anti-FIX inhibitors even after extensive rhF9 immunization in the presence of adjuvant. In contrast, all FIXnull control mice developed anti-FIX inhibitors when the same immunization protocol was employed. Of note, anaphylaxis can occur in these FIXnull mice with rhF9 infusion if the immune system was primed by FIX. To confirm that platelet-FIX expression is sustained in 2bCoF9R338L-transduced recipients, sequential transplantation was carried out using bone marrow from primary recipients that received 2bCoF9R338L-transduced HSCs. Platelet lysate FIX assays showed that hyperfunctional platelet-FIX was sustained in the secondary recipients resulting in phenotypic correction and immune tolerance in the secondary transplantation FIXnull recipients. Together, our data strongly suggest that immune tolerance is induced in FIXnull mice after 2bCoF9R338L gene therapy. In summary, we have demonstrated that we are able to significantly augment platelet-FIX expression utilizing codon-optimized FIX Padua for platelet-specific gene therapy of HB, resulting in phenotypic correction and immune tolerance induction in FIXnull mice. Our data suggest that platelet-targeted codon-optimized gain-of-function FIX gene therapy is a promising approach for gene therapy of HB. Disclosures No relevant conflicts of interest to declare.

Blockade of Inhibitor Formation by B-Cell Delivered Tolerance to Immunodominant A2 and C2 Domains.

A major impediment in the treatment of hemophilia is the formation of inhibitory antibodies, which occurs in approximately 25–30% of Hemophilia A patients treated with therapeutic Factor VIII (fVIII). We have focused on the development of a gene therapy protocol for tolerance induction, with an emphasis on the elimination of inhibitor production. Our lab has demonstrated that LPS-stimulated B-cell blasts, transfected with a retrovirus encoding an IgG-peptide fusion protein, such as fVIII domains, are tolerogenic in both normal and primed recipients. Last year, we reported (http://www.abstracts-on-line.com/abstracts/hemphiladelphia03; Scott and Lei 2003) that specific tolerance to the immunodominant epitopes in the C2 domain of fVIII (a major target of inhibitors) could be induced by our protocol. However, the immune response to full length fVIII was only modestly affected. Most inhibitory antibodies are reactive with conformational epitopes on the exposed surfaces of the A2, as well as the C2, domain of fVIII. Therefore, in this study, we inserted residues S2173-Y2332 of the C2 domain and S373-R740 of the A2 domain onto the IgG heavy chain backbone, respectively, to induce tolerance in hemophilic mice. Specific tolerance to each domain was induced by this protocol. Importantly, a combination of A2-IgG and C2-IgG expressing B cells induced tolerance to the full length fVIII molecule, a result which supports the dominance of these domains in the immune response to fVIII. Tolerance was manifested in terms of ELISA, T-cell proliferation and especially Bethesda Unit titers (95% reduction). Similar results were obtained even when treatment was initiated after priming injections of fVIII. In conclusion, this protocol offers great promise for prevention and potential reversal of this serious complication of fVIII replacement therapy. (Supported by HL61883 and a Laboratory Grant from the National Hemophilia Foundation).

Haploidentical Allogeneic Versus Gene Modified Autologous Stem Cell Transplantation for Severe Combined Immunodeficiencies (S.C.I.D). The Risk/Benefit Balance Is in Favor of Gene Therapy.

For the last 35 years or so, allogeneic hematopoietic stem cell transplantation (HSCT) has been the only curative approach for patients affected by severe combined immunodeficiencies (SCID), a medical emergency. According to the most recent European survey haploidentical transplants allow a survival rate at 3 years of 75% (taking only into account patients transplanted since 1995). The mortality rate is heavily influenced by age at transplantation (being very low within young infants), in correlation with the infection burden and also the incidence of acute graft-versus-host reaction (GVHD). Between 1998 and 2003, 6 patients with NK(−) B(+) SCIDs (either gc or JAK3 deficiency) received an haploidentical HSCT at Necker Hospital, 3 of them died from an infection, one in the context of a severe GVHD. In addition, despite the improvement in the survival rate, over the time a number of long-term concerns have been detected including a frequent persistence of a B cell deficiency, and a decline in T-cell functions related to the absence of donor stem cell engraftment and may be also to a premature decline in the thymus function. Therefore, several patients have received a second transplant with a low efficiency in an haploidentical setting. These significant limitations set the rationale for the development of an alternative strategy such as gene therapy. From March 1999 up to May 2002, ten children with gc deficiency under the age of one year were enrolled. All of them are alive today. The gc gene transfer into patients’ CD34(+) cells led in 9 out of 10 patients to the emergence of T and NK lymphocytes. It took 10 to 12 weeks to detect mature T cells in the periphery, a delay which is significantly faster than the one observed after haploidentical HSCT. In all but 2 patients, T cell counts normalized up to 5.3 years after gene therapy. The occurrence of two severe adverse events led to put transiently the trial on hold in order to understand the pathophysiology of these events and assess its overall risks. Following a thorough analysis of the retrospective data and a prospective analysis, it was concluded that the benefit/risk balance in favor of the gene therapy approach at least for patients older than 3 months, hence our gene therapy protocol has been reopened. An update of the clinical data will provided at the meeting.

Correction of a mouse model of sickle cell disease: lentiviral/antisickling β-globin gene transduction of unmobilized, purified hematopoietic stem cells

Although sickle cell anemia was the first hereditary disease to be understood at the molecular level, there is still no adequate long-term treatment. Allogeneic bone marrow transplantation is the only available cure, but this procedure is limited to a minority of patients with an available, histocompatible donor. Autologous transplantation of bone marrow stem cells that are transduced with a stably expressed, antisickling globin gene would benefit a majority of patients with sickle cell disease. Therefore, the development of a gene therapy protocol that corrects the disease in an animal model and is directly translatable to human patients is critical. A method is described in which unmobilized, highly purified bone marrow stem cells are transduced with a minimum amount of self-inactivating (SIN) lentiviral vector containing a potent antisickling β-globin gene. These cells, which were transduced in the absence of cytokine stimulation, fully reconstitute irradiated recipients and correct the hemolytic anemia and organ pathology that characterize the disease in humans. The mean increase of hemoglobin concentration was 46 g/L (4.6 g/dL) and the average lentiviral copy number was 2.2; therefore, a 21-g/L /vector copy increase (2.1-g/dL) was achieved. This transduction protocol may be directly translatable to patients with sickle cell disease who cannot tolerate current bone marrow mobilization procedures and may not safely be exposed to large viral loads. (Blood. 2003;102:4312-4319)