Another treatment for hemophilia is liver transplantation, which completely eliminates the symptoms of the disease
[9][7]. However, this intervention is radical and carries the risk of serious adverse reactions.
Currently, alternative therapeutic approaches are being developed: recombinant proteins with an extended half-life, monoclonal antibodies directed to tissue factor pathway inhibitor (TFPI)
[10][8], systemic delivery of
F9 mRNA
[11][9], antithrombin-specific small interfering RNAs
[12][10], and various gene therapy options with delivery of the
F9 gene or other factors (FVII or FV) of the coagulation pathway, and genome editing of the
F9 gene.
Gene therapy is a promising direction, as it can potentially provide long-term expression of factor IX after a single injection. For patients with severe disease, expression of factor IX as low as 5% will prevent spontaneous bleeding episodes and significantly improve quality of life. Among the possibilities for hemophilia B gene therapy, the following approaches can be distinguished: (1) correction of a defective gene copy using in vivo or ex vivo genome-editing tools, (2) control of protein translation without affecting the gene sequence using RNA interference, and (3) delivery of a functional gene copy of the
F9 gene with the help of viral vectors in vivo or ex vivo in autologous cells of the patient, with subsequent transplantation.
However, gene therapy has so far been used only for a small number of diseases. An in vivo editing approach is being studied in two clinical trials for the treatment of Leber’s amaurosis (EDIT-101, NCT03872479) and hereditary transthyretin amyloidosis (NTLA-2001, NCT04601051). The ex vivo approach is more common, including the possibility of CRISPR/Cas editing of B-lymphocytes for CAR-T therapy of various types of oncohematological diseases (NCT04037566, NCT04637763), hematopoietic stem cells in sickle cell anemia (CTX001, NCT05477563), beta-thalassemia (BRL-101, NCT05577312) and HIV infection, and pancreatic endodermal cell editing in type 1 diabetes (NCT05210530). In the case of hemophilia B, genome editing is used mainly at the stage of preclinical studies.
A drug based on RNA interference for the treatment of hemophilia B, Fitusiran (ALN-AT3), has been developed and is currently in phase 3 clinical trials. The FDA has approved four small interfering RNA drugs for the treatment of rare metabolic disorders (Patisiran, Givosiran, Lumasiran, and Inclisiran).
The vast majority of clinical studies on hemophilia B gene therapy are focused on the systemic delivery of a functional copy of the
F9 gene using AAV vectors. AAV vectors have low immunogenicity, generally do not integrate into the genome, and more often form an extrachromosomal structure, the episome, which remains in the cell for a long time. These advantages of AAV make this delivery system promising for the development of gene therapies, as evidenced by the growing number of approved drugs and drugs in clinical trials
[13,14][11][12].
Luxturna (Voretigene neparvovec, Spark Therapeutics, Philadelphia, PA, USA) and Zolgensma (Onasemnogene abeparvovec, Novartis Gene Therapies, Bannockburn, IL, USA), gene therapies based on AAV, have been approved by the FDA for the treatment of Leber’s congenital amaurosis and spinal muscular atrophy, respectively. In August 2022, the first gene therapy for hemophilia A (Roctavian, BioMarin Pharmaceutical, San Rafael, CA, USA) was approved in the European Union. Later, in November 2022 an AAV-based drug (Hemgenix, from UniQure, Amsterdam, The Netherlands, CSL Behring, King of Prussia, PA, USA) was approved by the FDA for the treatment of hemophilia B, which has become the most expensive drug in the world.
2. Clinical Gene Therapy Studies
Most of the gene therapy drugs being tested in clinical trials are AAV-based gene therapies. The general approach in these studies is to intravenously administer an AAV vector with liver affinity carrying a transgene containing a functional copy of the F9 gene under a liver-specific promoter. Codon optimization and elimination of immunogenic CpG motifs from expression cassettes are frequently utilized. Several clinical studies use F9-Padua, a variant of the F9 gene with one amino acid substitution (R338L), which is 5–10 times more active than wild-type F9 and was initially found in patients with thrombophilia [15][13]. Despite the recognition of F9-Padua as a game-changer in hemophilia B gene therapy, it is important to emphasize that there is a problem of assay discrepancies when evaluating post injection factor IX Padua levels.
AMT-060 (UniQure), an AAV5 vector carrying wild-type codon-optimized F9, was initially less effective, with a factor activity of 5.2–7.5% in phase 2 clinical trials after 3 years following administration of the drug to 10 patients. However, after replacing wild-type F9 with F9-Padua (AMT-061, UniQure), with the preservation of all other elements of the vector [18][14] in phase 3 clinical trials (HOPE-B, NCT03569891), the average factor IX activity 18 months after injection was 34.3%. A total of 52 of the 54 participants who received the injection stopped prophylactic substitution therapy; of the remaining two, one participant with a low response level had a high titer of neutralizing antibodies to AAV5, and the other received only a partial dose of the drug (10% of intended) due to an adverse event of hypersensitivity and continued prophylactic replacement therapy [19][15]. In November 2022, FDA approval was obtained for the use of this drug under the trade name Hemgenix.
The use of highly active F9-Padua has significantly influenced the development of gene therapy for hemophilia B, with many companies making use of it. In general, its use is considered safe, but a case with severe side effects has been reported. In phase ½ clinical trials of FLT180A (Verbrinacogene setparvovec, Freeline, Hertfordshire, UK), one patient experienced thrombosis of an arteriovenous fistula associated with an increase in the level of factor IX activity up of to 310% at the 4th week after administration [21][16].
Recombinant and modified AAV vectors continue to be developed and tested. In November and December 2022, studies of a new drug, ZS801, based on AAV with a synthetic capsid were registered in China (sponsored by the Institute of Hematology and Blood Diseases Hospital). In July 2022, the results of a phase 1 trial of BBM-H901 from the same sponsor based on AAV with a synthetic capsid carrying F9-Padua were published. The safety of the drug was shown in 12 patients with prophylactic use of glucocorticosteroids 1 year after administration. No serious side effects associated with BBM-H901 have been found [23][17].
Several studies of AAV-based drugs that deliver a functional F9 variant were terminated in 2021–2022, in particular, SHP648 (NCT04394286, Shire, Lexington, MA, USA) and DTX101 (NCT02618915, Ultragenyx Pharmaceutical, Novato, CA, USA), but not due to drug safety problems. Also, there is one study with lentivaral delivery of F9 gene to autologous hematopoietic and mesenchymal stem cells and their subsequent transplantation into a patient (YUVA-GT-F901, SGIMI, and NCT03961243), but the current status of this clinical trial is unknown.
3. Hemophilia B Models
In preclinical studies, the efficacy and safety of a gene therapy drug is tested in model objects: in cell cultures, which evaluate the efficiency of transduction, the level of transgene expression, cytotoxicity, and in animal models, which evaluate biodistribution, tissue specificity, transgene expression, toxicity, and immunogenicity and enable dose-administration-method selection.
To test the vectors used to deliver a functional copy of a gene or editing tools, cell lines derived from the liver are used, particularly Huh7, PLC/PRF/5, and Hep3B, with the necessary mutations introduced through editing
[29][18].
A more accurate model is iPSC-derived hepatocytes from hemophilia B patients. For example, for therapy testing, HB-iPSC (SXMUi001-A) with the c.223C>T mutation (p.R75X) was created
[30][19]. Mouse ESCs with the same nonsense mutation were also obtained
[31][20].
An even more accurate in vivo model that reproduces not only the cell type from the patient but also the effects at the level of intercellular interaction that exists in a real organ are 3D organoids. For example, there has been research into the differentiation of fibroblast-derived iPSCs derived from a patient with severe hemophilia B (mutation c.1297G>A) into hepatocytes with 3D organization, for which, compared with 2D culture, a higher level of expression of albumin, a marker of hepatocytes, was found
[32][21].
Hemophilia B has not been found in wild-type mice; therefore, different variants of transgenic mice are used to reproduce the pathology. For example,
F9-knockout mice are used to test therapies with the delivery of a healthy copy of the gene, and humanized mice are used to test editing tools on human regions of the genome.
A popular mouse model with undetectable levels of mRNA and factor IX in blood plasma has a knockout for the
F9 gene (B6.129P2-F9tm1Dws, Jackson Laboratory, USA)
[33][22]. Knock-in mice expressing various variants of factor IX are also frequently used, in particular, mice carrying human
F9 with a missense mutation found in patients with severe hemophilia B (R333Q-h
F9) under the mouse
F9 promoter, mice with wild-type human
F9-coding sequence, and mice with mutations in mouse
F9 (K5A in the Gla domain of
factor IX). Notably, in mice with R333Q-h
F9, the transcript and factor IX are expressed at a level of less than 1%, which is typical for patients with severe hemophilia B, and K5A mice have a mild disease phenotype.
All mouse models of hemophilia B are characterized by the absence of spontaneous bleeding. However, they die shortly due to blood loss after the tip of the tail is cut off. In knockout mice, when h
F9 is injected with AAV, antibodies are typically formed, while knock-in mice do not form them; therefore, the former are preferable for studying immune response to therapy and for testing therapies for the inhibitory form of the disease
[34][23]. Immunodeficient
F9-knockout mice (
Rag2-KO,
IL2-KO,
Fah-KO, and
F9-KO) were also created for transplantation of hepatocytes obtained from iPSCs of patients
[35][24].
As an alternative animal model of hemophilia B, dogs with spontaneously acquired mutations that were further established during breeding are used: a Cairn terrier with a mutation leading to the amino acid substitution E379G and the absence of detectable factor IX; a Lhasa Apso with a 5 nt deletion 772–776 and a g.777C > T substitution, also without detectable factor IX and without development of inhibitors upon administration of canine factor IX; and a Labrador retriever with a complete
F9 deletion that developed inhibitors
[36][25].
Using CRISPR/Cas9 and the somatic nuclear transfer method, a porcine model of hemophilia B with
F9 knockout (117 bp deletion in the 5’-UTR and exon 1) was also created, which is characterized by frequent episodes of spontaneous bleeding and joint damage
[37][26]. In this model, it was shown that the insertion of h
F9 facilitated bleeding, which indicates the possibility of in situ replacement of a defective gene with a functional one.
4. Genome-Editing Studies
Various groups of researchers are actively studying the possibility of genome editing in vivo or ex vivo by editing progenitor cells of the organ of interest to obtain a lifelong therapeutic effect in hereditary diseases of renewable organs. Viral vectors are used not only to deliver a healthy copy of the gene (which acts as a template for homologous replacement) but also editing tools (CRISPR/Cas or zinc finger nuclease—ZFN).
During the early stages of preclinical in vivo genome-editing research for hemophilia B, hemophilic mice were subjected to the administration of the AAV6-h
F9 and mAlb-targeted AAV8-ZFNs. The outcome of th
ise study revealed notably elevated levels of circulating human factor IX, reaching approximately 3000 ng/mL
[39][27].
CRISPR/Cas-based therapies for hemophilia B have only reached preclinical trials. The primary concern is potential off-target effects, especially with in vivo editing, so the use of the CRISPR/Cas system on patients will require constant monitoring for undesirable off-target effects. In this regard, ex vivo editing with subsequent transplantation of edited cells may be a safer and more controlled approach.
Viral vectors (based on AAV and adenoviruses) are typically used to deliver the editing system into the body. In 2019, a long-term restoration of the normal phenotype was shown after adenoviral delivery of the CRISPR/Cas9 system and a template for homologous recombination, with the aim of inserting a normal copy of the
F9 gene into a ROSA26 locus safe for insertion in hepatocytes in R333Q mice characterized by the expression of defective h
F9 and the absence of m
F9 [41][28]. The experimental group was injected with two AAV5 vectors (the first one with Cas and guides; the second one with the m
F9 matrix) at a concentration of 1:3; the control group received the same adenovirus with Cas but without guides. It was shown that the plasma concentration of mouse factor IX in the experimental group was significantly higher after 238 days than in the group without editing.
The efficiency of CRISPR/Cas editing using homologous recombination in vivo or in cell lines between different studies is about 5%
[43,44][29][30]. Therefore, to achieve effective knock-in, administration of high doses of AAV with Cas9 and donor DNA is usually required, which is associated with off-target risks and higher production costs. An alternative method is to edit through non-homologous end joining, which does not require the introduction of a vector with donor DNA. Insertion of
F9 into the 3’-UTR of the mouse albumin locus by non-homologous termination with the Cas9-delivered AAV2/8 vector has been shown to correct hemostasis in adult and neonatal m
F9-knockout mice for at least 48 weeks. Germ cell editing did not occur, and off-target effects were not detected.
As an alternative in vivo editing option, lipid nanoparticles have been used to deliver CRISPR/Cas with guides targeting the antithrombin gene. Affecting antithrombin expression, which is an endogenous negative regulator of thrombin generation, may improve blood clotting and relieve symptoms of hemophilia. When using lipid particles, no active off-targets, liver toxicity, and significant immune response to Cas9 were found. The use of lipid particles made it possible to reduce the time of CRISPR/Cas action as well as carry out repeated administration of the drug, unlike AAV
[48][31].
An ex vivo editing approach is also being explored, in particular, autologous transplantation of hepatocytes differentiated from patient iPSCs after CRISPR/Cas mutation correction. In iPSCs obtained from a patient with severe hemophilia (mutation g.31280G>A), a functional
F9 was inserted into a safe AAVS1 locus (AAV integration site in the first intron of the PPP1R12C gene on chromosome 19) with the help of CRISPR/Cas9
[32][21]. On day 11 of differentiation,
F9-KO mice were transplanted with corrected hepatocytes, and restoration of the normal phenotype was observed. The authors suggest that this approach may be relevant for children with severe hemophilia B, as cells in their bodies are actively dividing, which can lead to loss of AAV.
5. Current Challenges and Limitations
5.1. Immune Response to Factor IX
The formation of inhibitors (antibodies to factor IX) is the most severe complication in the treatment of hemophilia B, which can occur both with standard enzyme replacement therapy and with the use of gene therapy approaches. During standard prophylaxis, antibodies to factor IX in different studies were reported in only 3–5% of patients with severe hemophilia B. However, one of the latest studies with prospective follow-up showed a higher cumulative inhibitor incidence, namely 9.3% at 75 days after exposure and 10.2% at 500 days after exposure
[52][32]. While inhibitors do not affect the amount and location of bleeding, they significantly increase the risk of acute allergic reactions and death in such patients. There are currently no effective protocols of the induction of immune tolerance (IIT) for hemophilia B.
A high level of inhibitors in the blood of patients is a criterion for excluding patients from the sample of possible participants in clinical trials and future therapies. It has been shown that delivery of
F9 by AAV or a lentiviral vector can lead to the elimination of inhibitors and the subsequent establishment of factor IX expression at therapeutic levels
[53][33].
As an alternative approach to IIT, the possibility of using genetically edited B cells is being studied. Specifically, a lentiviral vector with IgG-h
F9 was developed that targeted CD20-expressing B cells and prevented the development of inhibitory antibodies to factor IX in a mouse model of hemophilia B
[54][34]. Resting human B-cell receptor specificity was achieved by mutating the envelope glycoproteins of the measles virus combined with lentivirus (MV-LV vector) and adding a single-chain variable fragment specific for hCD20.
To address the problem of antibodies to factor IX, oral immune therapy is also being developed to ameliorate IIT. The use of plant cells (lettuce or tobacco) producing CTB-hFIX by a patient may help prevent the formation of inhibitors and anaphylactic reactions during substitution therapy
[55][35]. The plant’s thick cell walls keep antigens from being destroyed by acid in the stomach until the cells are destroyed by intestinal bacteria, followed by release of the antigen combined with a transmucosal transporter (CTB) for passage through the intestinal epithelium.
Another method to get around the inhibition problem is the development of alternative gene therapy approaches that do not involve delivery or editing of the
F9 gene. Fitusiran (ALN-AT3, Sanofi, Paris, France), which is based on RNA interference, has shown its effectiveness in patients with hemophilia A and B regardless of the presence of inhibitors. Fitusiran is a double-stranded small interfering RNA, with one strand binding a 23 nt region in the SERPINC1 gene after insertion into the RISC complex. This gene encodes antithrombin, a serine protease produced in hepatocytes and inactivating thrombin, factor FXa, and, to a lesser extent, factors FIXa, FXIa, and FXIIa
[56][36].
Another option is using FVIIa in gene therapy for hemophilia B as a bypassing agent to promote blood clotting, as this factor can activate the coagulation cascade independently of factor IX
[57][37] or FVa, which has the capacity to enhance the rate of thrombin generation almost 10,000 fold
[58][38]. In hemophilia B mice, normal aPTT level was achieved over 28 weeks after AAV8/hFVa vector injection, and no risk of thrombosis has been shown.
5.2. AAV Immune Response
A significant limitation of all AAV therapies is the presence of AAV-neutralizing antibodies in some patients who have had a previous infection—such patients are usually excluded from clinical trials. Various approaches are being explored to overcome immunity to the delivery vector, including plasmapheresis, the use of immunosuppressants, IgG proteases, CpG reduction, induction of regulatory T cells, capsid variant switching, the addition of empty capsids, and the creation of synthetic capsids
[60][39].
The presence of neutralizing antibodies to AAV may not always lead to the absence of a therapeutic effect in gene therapy. In a study of AAV5-based AMT-060, a test with AAV5 reporter vectors with luciferase was retested for neutralizing antibodies more sensitive than the original test with AAV5 reporter vectors with GFP. The results showed the presence of antibodies to AAV5 in three of ten patients who were initially negative for anti-AAV5 neutralizing antibodies, and in two of them, an increase in the immune response was confirmed after administration of the AAV5-hFIX preparation. Despite this, no correlation was found between the level of
F9 expression and the level of antibodies since one patient had a low level of factor IX activity, and the second had the highest level of the low-dose cohort
[61][40].
5.3. Disadvantages of AAV as a Delivery System
AAVs are characterized by low capacity compared to other viral vectors. For example, the capacity of a lentiviral vector is up to 9 kb, which is twice the capacity of AAV (4.7 kb). The capacity of the AAV vector is not an obstacle in the case of the delivery of a functional copy of the
F9 gene, the coding part of which is 1386 bp; however, during genome editing, when AAVs are used to deliver the CRISPR/Cas system, the capacity of the vector becomes critical.
AAV vectors are stored as an episome in the cell, so they are advantageously used in non-dividing or infrequently dividing cells rather than in cells of rapidly growing organs, such as the liver in children, where loss of transgene expression can occur. In such cases, integration into the genome of a functional
F9 sequence is an attractive approach.