Gene therapy in an era of emerging treatment options for hemophilia B

Factor IX gene therapy and the landscape of treatment for hemophilia B

Hemophilia B, resulting from deficient activity of clotting factor IX, accounts for 20% of hemophilia, while 80% of hemophilia results from deficient factor VIII activity (hemophilia A). Throughout most of the last half century, hemophilia B therapies have developed more slowly and often have been less available than treatments for hemophilia A. The description in 1964 of cryoprecipitate as a plasma preparation rich in factor VIII was a transformative moment in the care of hemophilia A, but large volume plasma infusion remained the only source for replacement factor IX. Widespread availability of plasma-derived factor IX concentrates was delayed almost a decade after factor VIII concentrates became available in the 1970s. In response to the tragic contamination of plasma-derived factor concentrates with hepatitis and HIV, recombinant FVIII was generated and licensed by 1992; hemophilia B patients waited until 1997 in the U.S. (and much later in some countries) for the licensure of recombinant factor IX (rFIX), the last real innovation in hemophilia B care.

Then in a one month period between December, 2011–January, 2012, reports appeared that two goals had been achieved by new factor IX therapies: 1) the successful, persistent expression of factor IX by human gene therapy [1] and 2) the demonstration in a human clinical trial several-fold extension of factor IX circulating half-life via recombinant protein engineering [2]. Hemophilia A still awaits these successes. In the interim between January 2012 and now, the licensure of a second recombinant factor IX (biosimilar rFIX concentrate having the wild type FIX molecular sequence and structure) expands product choice for conventional hemophilia B replacement therapy. During the same period three different approaches to modify the wild type factor IX protein have yielded comparable, >3-fold prolongations in the factor IX circulating half-life in human trials: the recombinant expression of factor IX monomer protein genetically fused to the constant region of immunoglobulin(rFIXFc); the recombinant expression of factor IX monomer protein genetically fused to albumin; and the modification of recombinant factor IX via the addition of a 40k polyethylene glycol moiety to the N-glycan acceptor sites at FIXAsn157 or FIXAsn167 (as reviewed by Kaufman and Powell) [3]. One rFIX fusion protein now is licensed (rFIXFc) and a U.S. Food and Drug Administration Biologic License Application has been submitted for the second (rIX-FP). This review limits its scope to considering how insights gained from three pioneering factor IX gene therapy trials have informed the objectives and design of several ongoing and planned trials of hemophilia B gene therapy. Hemophilia gene therapy’s first clinical success is reviewed in the context of an evolving era of choice for men with severe hemophilia B, while also considering the unique characteristics of this population (see Table 1) that currently limit the broad clinical translation of liver-directed adeno-associated virus (AAV) vector and other gene delivery approaches.

The first success: Adeno-associated virus vectors for the expression of factor IX

The first and only successful approach for human gene transfer of a coagulation factor used an adeno-associated virus (AAV) as the vector to deliver a human factor IX cDNA to the liver, the natural site of factor IX production [1] [4]. The wild type (wt) AAV is a small single-stranded DNA virus that is non-pathogenic and replication defective. Most individuals experience an asymptomatic exposure to wt AAV during childhood and mount a humoral immune response, which often results in persistent titers of anti-AAV antibodies sufficient to neutralize subsequent infection by wt AAV and by recombinant AAV vectors [5] [6] [7]. All coding sequences of the 4680 nt wt AAV genome are removed during the generation of AAV gene therapy vectors, adding further to their safety. Following transduction of the target cell, the recombinant AAV (rAAV) therapeutic gene sequences are retained primarily as concatameric episomes and uncommonly integrate into the host genomic DNA.

Following the initial recognition that rAAV vectors transduce a variety of relatively end-differentiated, non-dividing cells and tissues, the earliest attempts to correct hemophilia B targeted muscle expression. The capacity for intramuscular (I.M.) injection of AAV.FIX vectors to direct human factor IX expression and improve coagulation was demonstrated in mouse and dog models of hemophilia B [8] [9] [10] [11]. A Phase I clinical trial of AAV.FIX delivered to skeletal muscle commenced, sponsored by Avigen, Inc. Eight subjects were treated with a single strand rAAV serotype 2 vector following a single-dose, dose-escalation design via multiple direct I.M. injections. There were no safety concerns but also no sustained circulating factor IX activity was achieved [12]. Biopsy of injected skeletal muscle at 2–10 months after administration revealed persistence of the therapeutic transgene and factor IX protein not only intracellularly but also extensively bound to extracellular matrix collagen IV surrounding myocytes[13] [14]. Skeletal muscle from one subject examined at 10 years after AAV gene delivery demonstrated local persistence of vector-derived FIX mRNA transcripts and FIX protein [15]. Recently, large animal studies have validated a strategy to achieve more efficient and uniform transduction of skeletal muscle by avoiding multiple local percutaneous I.M. injections and delivering rAAV.FIX to the entire vascular bed of limb muscles by using a peripheral transvenular approach [16].

AAV serotype 2 factor IX gene delivery to the natural site of factor IX expression in the liver was explored in mice and found to generate greater transgenic factor IX activity than equivalent vector doses delivered to the muscle; at least a part of the difference results from less efficient post-translational modification of factor IX in muscle when compared to the natural organ of synthesis of the protein. An unusual complication of severe hemophilia B, occurring in about 2–4% of patients treated with factor IX protein infusions, is the development of factor IX-specific IgG that neutralizes the procoagulant activity of factor IX (“factor IX inhibitors”). Not only is factor IX therapy rendered ineffective by this loss of “tolerance” of factor IX by the patient’s immune system (to which the factor IX appears to be a neoantigen) but potentially life-threatening anaphylactic reactions may be triggered by factor IX exposure. In the context of gene therapy, liver transduction in animals did not lead to loss of tolerance and factor IX inhibitor formation as readily as muscle gene delivery [10]. Multiple investigations have reproduced the observation that the tolerance of clotting factor following hepatic transduction by AAV does not result from anergy or immunologic ignorance of the potentially neoantigenic clotting factor, but results from the induction of tolerogenic regulatory T lymphocytes [11] [17] [18]. Convincing preclinical data for the liver-directed AAV approach was generated by showing that null mutation hemophilia B dogs having a high risk for factor IX inhibitor formation produced persistent circulating factor IX activity following hepatic delivery of AAV2.FIX, experienced decreased bleeding, and did not form factor IX inhibitors [12].

Avigen sponsored a human gene therapy trial delivering an AAV2.FIX construct with a liver specific promoter directly into the hepatic artery, reported by Manno [19]. At two lower doses no safety concerns or factor IX expression were observed, prompting escalation to the planned highest dose of 2 × 10¹² Vector Genomes (VG)/kg. Circulating factor IX activity peaked at 11.8% at 2 weeks after vector infusion, however factor IX then declined starting around 4 weeks after the AAV2 infusion, coincident with an ~ 10 times normal elevation of the liver transaminases ALT and AST. Peripheral blood T cells from a subsequent subject who received AAV2.FIX and experienced a less profound transaminitis suggested the most likely etiology was a cell-mediated adaptive immune response directed toward the liver[19] [20]. Following successful FIX gene delivery to hepatocytes, the capsid of the vector is degraded intracellularly and capsid-derived peptides are presented on the surface of the transduced hepatocyte in the context of MHC class I molecules. At relatively higher exposures of AAV capsid, memory T cell recognition of AAV capsid peptides then may trigger a cytotoxic lymphocyte (CTL)-mediated elimination of AAV-transduced hepatocytes.

Several strategies were subsequently incorporated into AAV.FIX vectors to increase the efficiency of liver transduction, so that the vector particle dose and potential immune recognition of AAV could be minimized. The rhesus macaque is the natural host for AAV8 and the seroprevalence of AAV8-neutralizing antibodies (NAb) in humans is lower than AAV2 NAb [6] [7] [5]. AAV8 transduces liver more efficiently than AAV2 and with less off-target distribution in organs outside the liver, allowing infusion by peripheral vein rather than the invasive hepatic artery route [21]. A rate-limiting step in AAV transduction, which is the need to convert single strand AAV into a double-stranded configuration to serve as a template for transcription, was overcome by using self-complementary (sc) AAV vectors, increasing transduction efficiency [22] [23] [24]. Conservative nucleotide changes leading to codon optimization also yielded robust increases in factor IX expression [17] [24].

In 2011 St Jude Children’s Research Hospital and the University College London (UCL/SJCRH) investigators reported the first hemophilia trial to achieve persistent clotting factor expression [1] [4]. Following the single scAAV8.FIX codon-optimized vector administration all ten subjects in the trial decreased markedly their use of exogenous factor IX protein infusions. Steady state expression of factor IX activity for the six subjects treated at the highest dose (2 × 10¹² VG/kg) maintained a mean of 5.1 ±1.7%, measured at least a year after vector administration. At this high dose, however, 4 of 6 subjects developed AAV capsid-specific T lymphocyte responses at 7 to 10 weeks after vector administration. The pro-inflammatory T effector cell responses were temporally associated with transient alanine aminotransferase (ALT) increases to 1.5–4 times the upper limit of normal in two subjects; two additional subjects demonstrated an increase above their individual baseline, although the ALT remained in the normal range. Corticosteroid therapy was initiated promptly in response to the transaminase changes and continued for 8–12 weeks. Despite a decline in factor IX expression in these four subjects some factor IX expression has been maintained in all subjects, which is an encouraging improvement over the outcomes in the ssAAV2 liver trial. The first subject treated has expressed 2% factor IX steadily for more than 4 years of follow up.

Ongoing and planned hemophilia B clinical trials using the adeno-associated virus vector

The UCL/SJCRH study investigators describe the hepatic inflammation seen in their high dose study subjects as a manageable toxicity. Nevertheless, the complication limits the ability to offer the therapy to adults with hemophilia who are active carriers of hepatitis or to increase the vector dose to achieve truly normal hemostasis (e.g. to achieve baseline factor activity of ≥15% that is expected to prevent all joint bleeding) [25]. Subsequent to the 2011 report, three AAV8.FIX trials have commenced with the goal of achieving greater factor IX expression without toxicity. The UCL/SJCRH clinical vector contained, in addition to the therapeutic scAAV8.FIX_co, empty AAV8 capsids carrying no FIX gene [26]. These contaminating empty AAV capsids comprised 80% of the vector preparation and were a potential trigger for the dose-dependent CTL response directed against AAV. The AAV vector manufacturing process at the University of North Carolina at Chapel Hill Vector Core Facility incorporates empty capsid removal steps. Vector made at this facility containing less than 10% empty capsids is being used in a trial sponsored by Baxter Healthcare, testing a scAAV8.FIX vector designated BAX335 [27, 28]. Purification steps that remove empty capsids are now employed by the UCL/SJCRH group, resulting in a 90% “full” capsid vector stock [29]. Each of these two groups is currently enrolling subjects to test whether scAAV8 vector having minimal empty capsid contamination will permit dose escalation without stimulating cell-mediated immunity.

Investigators at the Children’s Hospital of Philadelphia (CHOP) proposed an alternative approach to the empty capsids, theorizing that the defective particles may have aided AAV transduction in the UCL/SJCRH trial. Mingozzi and colleagues suggested that empty capsids titrated into therapeutic AAV vector preparations could act as “decoys” in the eventuality that circulating AAV NAb were present, thereby allowing therapeutic virus to evade neutralization and transduce liver more efficiently. This strategy was to be incorporated into a single-stranded AAV8.FIX trial [30]. A clinical trial initiated by these investigators (sponsored by Spark Therapeutics) has yet to be reported [31].

Given that a vector dose-dependent immune response is the only toxicity of AAV.FIX observed to date in humans, and the fact that capsid epitopes from empty capsids that enter cells may contribute to the CTL response, the CHOP investigators next introduced further mutations at amino acids 585/588 of their decoy capsids to disrupt the AAV2 binding for heparin sulfate (HS), the receptor for AAV on hepatocytes [30]. The investigators showed in mice that these mutations prevented receptor-mediated hepatocyte entry by the decoy empty capsids, presumably eliminating the ability of the hepatocyte to present capsid-derived peptides and trigger a hepatocyte-targeted CTL response. The use of excess HS-non-binding (585/588 mut) decoy capsids has not been tested in human application. One concern is that Kern and colleagues have previously shown that, although hepatocyte entry is lost in mutant rAAV585 capsids, entry into cardiac muscle is maintained or increased, and so the potential exists that the CTL could be redirected from liver to myocardium [32].

An alternative approach to achieve higher factor IX activity without increasing vector exposure, now validated in preclinical work, is to employ a gain-of-function variant of factor IX. Structure-function studies of the catalytic domain of factor IX in the Stafford laboratory revealed that substitution of alanine for arginine at amino acid 338 in the catalytic domain increased the specific activity of factor IX [33], and Simioni and colleagues subsequently characterized a family that expressed a hyperactive factor IX as a result of a leucine substitution at this same amino acid (FIXR338L) [34]. Multiple groups have now demonstrated in hemophilia B animal models that the incorporation of FIXR338L (Factor IX Padua) into AAV vectors for transduction of liver [35] or muscle [16], or into lentiviral vectors for transduction of liver [36] [37] leads to the expression of factor IX protein having at least 6–8 times normal specific activity without an increased incidence of thrombosis. Crudele and colleagues recently reported that AAV8-mediated hepatic expression of FIXR338L was associated with the development of factor IX tolerance in FIX inhibitor-prone hemophilia B dogs, as well as eradication of the pre-existing factor IX inhibitor in one dog [17]. Preclinical data has been reported recently that supported the initiation in 2013 of a human trial of the scAAV8FIXR338L codon-optimized BAX335 vector (the vector introduced above), with trial results pending [35].

Spark Therapeutics and Dimension Therapeutics each have promised Phase 1 trials for hemophilia B using proprietary AAV capsids; additional details are not publicly available. Uniqure has initiated a European Phase 1/2 trial using the identical factor IX expression cassette (LP1-hFIXco) used in the UCL/SJCRH trial, packaged within an AAV serotype 5 capsid instead of AAV serotype 8 and delivered via peripheral vein. Interestingly, the LP1-hFIXco expression construct pseudotyped with AAV5 and with AAV8 has previously been compared in non-human primates by Nathwani and colleagues [21]. Although the studies demonstrated comparable liver transduction and factor IX expression using AAV5- or AAV8-pseudotyped vectors, the two serotypes were not interchangeable. Following peripheral vein administration, when compared to scAAV8-LP1-hFIXco, the expression was slower from AAV5-pseudotyped sc-LP1-hFIXco, the clearance of vector genomes from plasma more delayed, and the off-target transduction of spleen, kidney and testis were greater for the serotype 5-pseudotyped vector. The investigators’ conclusion was that scAAV8 capsid proteins are preferable for liver transduction via peripheral vein but that scAAV5-pseudotyped vectors should be effective for treating individuals with pre-existing immunity to AAV8 or AAV2 [21]. Should pre-existing AAV8 NAbs limit enrollment in clinical trials, the Uniqure AAV5-pseudotyped vector might prove advantageous. A recent study of prevalence of naturally-occurring antibodies against AAV serotypes conducted in a healthy European population documented neutralizing antibody seroprevalence of 59% for AAV2, 19% for AAV8 and 3% for AAV5 [7]. Higher seroprevalence of AAV5 NAbs has been reported in other populations [38] [39], and the lack of concensus regarding a standardized methodology for NAb measurement complicates comparison between studies. The only AAV NAb study of a similar size evaluated in a hemophilia population is a pediatric study showing that by age 5–6 years boys with hemophilia had NAb seroprevalence of 27% for AAV2, 13% for AAV8 and 15% for AAV5 [6]. Uniqure has presented in abstract form a safety and dose-finding study in non-human primates to justify the trial design [40]. Five subjects will be treated with AAV5-hFIX 5.0 × 10¹² VG/kg, followed by an escalation to 2.0 × 10¹³ VG/kg to treat the final five subjects.

Sangamo BioSciences in collaboration with Shire AG has announced plans for a human trial of factor IX gene editing using AAV to deliver zinc finger nuclease (ZFN) technology. The planned trial is based on work from the research group of High, in collaboration with Sangamo, who used a hemophilia B mouse model, engineered to carry a human nonsense mutant Y155stop F9 minigene inserted in the ROSA26 locus (hF9mut mice), to examine ZFN gene correction in vivo. The delivery of AAV8 ZFN to target intron 1 of the F9 gene along with a second AAV8 donor template vector with arms of homology flanking F9 exons 2–8 to neonatal mice (i.e. mice undergoing rapid growth of the liver) as well as adult mice resulted in apparent ZFN-induced double strand breaks in host DNA and homology-directed repair of the F9 gene [41]. Co-delivery of AAV8 expressing the hF9-specific ZFN and AAV8 expressing the corrective partial F9 cDNA was able to direct sustained expression averaging 23% of normal human factor IX in the adult hF9mut mouse model.

Although gene editing is an exciting direction for the field, several caveats exist with the work described to date. Off-target cleavage by the ZFN vector remains a concern (for example, treatment of WT littermates lacking the ZFN target site nevertheless led to expression of 1% normal human factor IX) [42]. Rates of off-target cleavage differ between ZFN and more recently developed nuclease systems that systems that are finding application in hemophilia gene and cell therapy [43], such as transcriptional activator-like effector nucleases (TALENS) and the two-component CRISPR (clustered regularly interspaced short palindromic repeat)– Cas9 (CRISPR-associated nuclease 9). Off-target rates depend not only on the nuclease reagent but also how long and at what level the nuclease is expressed and how many likely off-target sites exist in the genome from the outset. The AAV8.ZFN F9-targeting dual vector strategy described above provided for constitutive, rather than transient, expression of the ZFN by AAV vector transduced cells. This creates the potential for generation of an immune response against the ZFN (as a foreign protein), as well as the possibility for lasting genomic instability due to long-term low-level ZFN-mediated genomic damage. Furthermore, although ZFN-mediated gene editing is expected to introduce permanent gene correction (when compared to an episomally located AAV-delivered factor IX transgene), the proposed approach did not circumvent the need for doses of AAV vector that have been immunogenic in the previous hemophilia trials; the equivalent of 2.4 × 10¹³ VG/kg of the combined AAV vectors was used to promote gene correction in the murine model (i.e. ten times greater vg/kg dose than the dose associated with transaminitis in the Avigen and UCL/SJCRH trials). For human clinical application, Sangamo has developed an alternative approach in which F9 gene insertion is targeted downstream of the highly active albumin promoter, potentially driving robust factor IX production while requiring fewer vector particles/integration events than were required in the mouse studies. Such an approach may not require a nuclease to promote site-specific recombination [44].

An unknown regarding the ultimate potential for AAV gene therapy to provide a cure for hemophilia B is how long transgene expression from this non-integrating vector will persist, particularly in tissues like the liver that are felt to undergo some degree of turnover (albeit at a slow rate). The F9 transgene integration status and the persistent, stable expression of AAV-delivered factor IX from the livers of the null-mutation dogs has been extensively characterized at 8 years of follow up and subsequent reports confirm phenotypic correction for >10 years [45] [11]. Likewise, ongoing hepatic factor IX expression for greater than 8 years continues to be followed in non-human primates [4]. Consistent with the large animal experience, the first subjects treated on the UCL/SJCRH trial continue to demonstrate stable factor IX expression for greater than four years and counting.

Preclinical Development: Lentivirus vectors for liver gene therapy

Lentiviruses (LVs) efficiently transduce the relatively quiescent hepatocytes along with multiple cells of the liver. These include antigen-presenting cells, and a major task for adapting lentiviruses for hemophilia gene therapy has been to avoid immune responses to the potentially neoantigenic factor IX and factor VIII. In regard to factor IX, restricting gene expression to the liver with the use of liver-specific promoters and opposing gene expression in APCs via the incorporation of hematopoietic-specific microRNA target sequence (miR142-3p) achieved factor IX expression in and secretion from the liver in both mouse and dog models of hemophilia B [46] [47]. Moreover, lentivirus-directed liver-restricted factor IX expression appears to actively promote tolerance induction via the induction of CD4+ CD25+ Foxp3+regulatory T cells (Tregs) [47]. Tolerance induction and eradication of pre-existing factor IX inhibitors has recently been demonstrated following liver-restricted LV.FIX transduction [18].

Given somewhat less robust factor IX expression from LV (when compared to several serotypes of AAV) much recent emphasis has been on LV development for the task of delivering the large F8 gene, which is very difficult to accommodate in tiny AAV vectors [47] [48]. Investigation of the concern that LV genomic integration could cause genotoxicity is the subject of active investigation, and appears to be considerably lower risk than has been observed with the use of gamma retroviral vectors [46]. Nevertheless, several groups have addressed the concern of potential genotoxic risk by engineering lentiviral vectors with inactivating mutations in the integrase to create integration-defective lentiviral vectors. Although the transgene expression is reduced by loss of integration, optimization of the vectors and incorporation of gain-of-function FIXR338L transgene has achieved disease correction in hemophilic mice, while maintaining the potential for factor IX tolerance induction by hepatic expression [46] [36] [37] [47].

Preclinical Development: Cell-based approaches

Lentivirus-mediated delivery of the F9 gene has also been employed in a cell-based approach that results in circulating platelet delivery of factor IX to maintain hemostasis [49]. Adapting an approach that has demonstrated phenotypic correction in large animal models of hemophilia A and Glanzmann Thrombasthenia [50], a LV construct expressing human factor IX under the control of the platelet-specific glycoprotein IIb gene (αIIb) promoter was used to transduce FIX^−/− mouse hematopoietic stem cells (HSC) ex vivo. Recipient FIX^−/− mice were conditioned for HSC transfer with either a myeloablative (1100 cGy) or a nonmyeloablative (660 cGy) dose of total body irradiation 24 hours prior to infusion of the LV-transduced HSCs. Although negligible factor IX activity was measurable in plasma, human factor IX expression was present in cells of megakaryocyte lineage, co-localized with Von Willebrand Factor, provided partial hemostatic protection during a bleeding challenge, induced tolerance to human factor IX in FIX^−/− mice, and could be transferred via bone marrow transplant to secondary recipient FIX^−/− mice. Moreover, the nonmyeloablative conditioning was adequate to achieve all of these outcomes. Interestingly, an LV strategy employing a platelet-specific factor VIII expression cassette has recently been demonstrated to transduce hematopoietic stem cells and lead to correction of the hemophilia A bleeding phenotype in mice following direct intraosseous LV delivery with no requirement for a bone marrow conditioning regimen [48]. The latter approach has not been explored for the correction of hemophilia B.

Severe hemophilia B: Goals and Options

It has been stated many times that hemophilia B is an ideal disease model for the investigation of gene therapy approaches [47] [1]. It is worthwhile to examine the target patient population of adult men with severe factor IX deficiency and consider whether gene delivery in its current application is the ideal therapeutic approach and what patient-related factors may determine whether and how gene therapy translates from a primarily academic effort into wider clinical application.

An unprecedented time of new treatment options for men with severe hemophilia B approaches. An examination of the self-reported outcomes of men with severe factor IX deficiency captured by the U.S. surveillance instrument (Table 1) reveals that a third of men report potentially damaging high rates of joint bleeding (adjusted joint bleeding rate ≥10/yr), while the adoption of regular factor IX prophylaxis is low. (Table 1) Adherence to prophylactic infusions is an unrelenting demand whether the goal is to reduce bleeding or prevent it altogether. Hemophilia patients and physicians have recently been presented with three studies taking different approaches to maintaining joint health. One study compared on-demand therapy with unmodified recombinant factor IX (rFIX, Benefix®, Pfizer) using either a conventional approach to factor IX prophylaxis (50 IU/kg) twice weekly or a higher dose once-weekly (100 IU/kg) regimen. There were no treatment-related safety concerns. The mean annualized bleeding rates (ABR) were 35.1, 2.6, and 4.6 bleeding events/year for the on-demand, the twice-weekly therapy and the once-weekly therapy. The rationale for conventional twice weekly therapy is to maintain measurable (>1% activity) trough factor IX concentrations throughout a week; the higher once-weekly dose would be expected to result in some portion of the week with unmeasurable circulating factor IX, Nevertheless, there was no statistical difference between the ABR for the two prophylaxis approaches (P=0.22) [51]. These study results and others justify investigation of the possibility that factor IX bound in extracirculatory sites may contribute to hemostasis [52] [53]. Understanding the amount and the physiologic relevance, if any, of the factor IX distributed in extracirculatory sites could be relevant for understanding the total endogenous factor IX expression required from gene therapy.

A separate study compared recombinant factor IX Fc fusion protein (rFIXFc, Alprolix^®, Biogen Idec), a newly licensed extended circulating half-life factor IX, infused on demand for bleeding events, compared to either 1) rFIXFc prophylaxis infused weekly and starting at a dose of 50 IU/kg or 2) prophylaxis infused at a dose of 100 IU/kg every ten days. Trough factor IX activity levels were obtained regularly and the dose of the lower prophylaxis dose was adjusted during the trial to maintain trough plasma factor IX activity 1–3%; the interval of the higher prophylaxis dose was adjusted to target the same trough levels. The median annualized bleeding rates were 17.7, 3.0, and 1.4 bleeding events/year for the on-demand, the lower dose, weekly therapy and the higher dose, less frequent therapy, with no statistical difference between the two prophylaxis approaches (P=0.22) [51]. There were no treatment-related safety concerns. The third recent study to consider alongside the rFIX and the rFIXFc studies is the UCL/SJCRH gene therapy trial, in which six subjects received the highest dose of 2 × 10¹² VG/kg of scAAV2/8-LP1-hFIXco vector. This dose resulted in mean steady state plasma factor IX activity of 2.9–7.2% for the six subjects, a median annual bleeding rate of 1.0 episodes, and a relative 94% reduction of bleeding compared to before gene therapy [4]. Laboratory monitoring was intensive for the first four months of the trial and then tapered. Four of six participants received 8 to 12 weeks of corticosteroids to treat vector-associated hepatic inflammation. These examples of choices and outcomes are offered as a reminder that progress in gene therapy occurs in the context of constant shifts in hemophilia care standards and community outlook.

To illustrate an example of a developed world country, Table 1 compiles available data for the population of men with severe hemophilia B in the United States (keeping in mind that 80% of men with hemophilia B live in developing countries and have no access to factor IX therapy). While there are approximately 1000 men in the U.S. who could be eligible for gene therapy trials based on factor IX activity ≤2 % of normal (≤2 U/dl), the relative and the absolute exclusionary criteria that apply to liver-directed AAV.FIX trials (ongoing or planned) considerably narrow the eligible population.

The hepatocyte-directed immune response that has been elicited by AAV capsid epitopes at the most therapeutic vector doses drastically limits the eligibility of men over the age of ~ 30 years due to the high hepatitis B virus (HBV) and hepatitis C virus (HCV) carrier prevalence in this population. Almost 80% of the adult severe hemophilia B population in the United States, for instance, is seropositive following exposure to HCV. Rates of active carrier state (HCV PCR +) are not available for the hemophilia population. Although it is likely that spontaneous clearance of HCV may occur in as many as 20% of HCV exposures, at the start of the current decade most HCV Ab+ hemophilic men in the U.S. had never received therapy to attempt to eliminate the HCV carrier state. The last two years have witnessed the advent of new and well tolerated antiviral agents that are effective against even the most historically resistant strains of HCV. It can be hoped that eliminating active HCV will be a major accomplishment in hemophilia care in the next decade. Men who have a sustained HCV clearance response following therapy and no evidence of liver failure are eligible for AAV.FIX therapy and one consequence of eliminating the HCV carrier state would be to broaden the eligible population for systemic gene delivery. Alternatively, continued development of gene therapy targeting non-hepatic tissues (e.g. targeting muscles, joints, hematopoietic stem cells including platelets) may present the safest approaches in this large subset of individuals with hepatitis. [16] [53] [48] [49].

Although it is uncommon for inhibitors to complicate hemophilia B, the history of a factor IX inhibitor antibody is exclusionary for all current gene therapy trials [54]. As discussed above, multiple small and large animal studies have demonstrated that steady state endogenous factor IX expression via gene therapy may promote tolerance via the induction of factor-specific regulatory T cells (e.g. platelet-targeted gene expression) or even eradicate pre-existing factor IX inhibitors (as has been demonstrated via hepatic expression following AAV or lentivirus factor IX gene delivery) [55] [56] [18] [49]. Clinically, factor IX inhibitors are often associated with hypersensitivity reactions, and there are no reliable predictors of the risk or timing of the reactions. A deliberate attempt to reverse a factor IX inhibitor by constitutively expressing factor IX using an expression cassette that cannot be silenced involves some risk of triggering hypersensitivity that cannot be controlled. It is intriguing to consider combining formal factor IX desensitization or anti-IgE monoclonal antibody therapy with liver-directed gene therapy as a potential therapy for the very difficult clinical complication of allergy-associated factor IX inhibitors.

An additional immunologic consideration already mentioned is the growing understanding that pre-existing NAbs against AAV may cross-neutralize multiple serotypes of AAV and abrogate vector transduction at very low titers (e.g. neutralizing titers of 1:5 to 1:10) [57]. Older prevalence studies of AAV NAbs often defined as a lower limit of detection titers of 1:20 or higher, and methods of detecting neutralizing antibodies are not uniform between laboratories, so that the scope of AAV NAb interference remains unclear. Standardization of methods will be critical as the nascent field of hemophilia gene therapy grows beyond isolated, primarily academic trials and into multiple concurrent industry-supported trials. Implementation of a recently reported reference standard for the titering of AAV serotype 8 viral vectors is an important first step in being able to rationally compare dose-response and dose-toxicity results between clinical trials [58]. Currently children are excluded from gene therapy trials for diseases that do not cause death or great morbidity in childhood. It has been shown that once infants lose maternally transferred AAV NAb, there is a window around one year of life when essentially 0% of infants have AAV NAb [5]. Should liver-directed AAV gene therapy prove to be safe, to support factor IX tolerance and concerns about NAbs (and perhaps by extension also memory lymphocyte-driven CTL response)in infants are minimal, then broadening the application of gene therapy to children could be advantageous.

Following decades of limited choices regarding hemophilia B therapy, patients and their doctors now may consider variety of protein and (in the context of clinical trials) gene therapies. While gene therapy involves the most uncertain risks, extended therapeutic benefit has now been demonstrated in a human clinical trial. With multiple additional trials ongoing or commencing, one can expect that additional knowledge and experience will facilitate the discussion between patients, families and their doctors about how gene therapy may fit into the therapeutic options for hemophilia B.