Gene therapy for haemophilia

New technologies for the treatment of haemophilia offer a future with better control of bleeding, fewer long term complications, a lower treatment burden and, in the case of gene therapy, the prospect of a life largely unaffected by the risk of spontaneous bleeds. They also bring challenges, with major change in the expectations of people with haemophilia, a transformation in the way haemophilia services are provided and increased pressure on resources.

Positive results of early stage clinical trials have focused the attention of the haemophilia community on gene therapy. Investigators and clinicians have been cautious in their interpretation of achievements to date. Nevertheless, talk of a ‘cure’ is not wholly unjustified and people with haemophilia and their families may include gene therapy among the future treatment options they are now considering. This review summarises the evidence available in September 2020.

Why haemophilia?

Haemophilia is an ideal candidate for gene therapy. It is caused by mutation of a single gene (F8 for haemophilia A, F9 for haemophilia B), meaning there is a single target for therapy. The mutation has a large effect, so that replacing the faulty gene with a functional version has the potential for a large benefit. Further, the level of clotting factor does not need to be very high to achieve a large reduction in the risk of spontaneous bleeds. Finally, conventional factor replacement therapy is relatively expensive and burdensome; the high cost of gene therapy is therefore offset by large savings from reduced clotting factor consumption and the potential for gains in quality of life.

How is gene therapy carried out?

At its simplest, gene therapy involves cloning the gene that encodes for factor VIII (FVIII) or factor IX (FIX), packaging it into a system that delivers it into a cell, and administering this to an individual to enable them to produce sufficient clotting factor to avoid bleeds. However, no step in this process is straightforward.

Gene therapy has developed more quickly for haemophilia B than haemophilia A. This is because the F9gene is four to five times smaller than the F8 gene, and it has taken longer to develop a vector that can accommodate the replacement F8. Progress in developing gene therapy for von Willebrand’s disease has been slowed by the challenge of incorporating the large gene encoding for von Willebrand factor into a viral vector than can deliver it to the target organ. To date, studies have not progressed beyond animal models.

Most trials to date have evaluated technology that creates a stable pool of modified genes within a cell. The pool (known as an episome) remains in the cell cytoplasm, is independent of the cell’s nucleus and produces clotting factor autonomously. This approach has the potential to abolish haemophilia or greatly reduce its severity for the person treated; however, it does not change the risk of passing the condition on to his children (all trial participants have been men so far). Most trials are targeting liver cells as the repository for the modified gene; one trial is underway to evaluate transplantation of autologous CD34+ hematopoietic stem cells containing the viral vector ( NCT04418414).

Trials are now being initiated to evaluate the effect of modifying the structure of DNA itself by gene editing – something that could conceivably be a ‘cure’ for haemophilia in that it might stop genetic transmission. One method of gene editing is to use a designer molecule (e.g. a zinc finger nuclease or CRISPR-cas9) to snip the DNA molecule and introduce a new functional gene. Zinc finger proteins, in which zinc ions stabilise the 3D protein structure, occur naturally in human genes; their role is to bind specific regions of DNA and RNA during cellular processes such as gene transcription. This process of gene integration has huge therapeutic potential but also raises uncertainties about the long-term risks of altering DNA.

First, a word about genes

The ‘gene’ in gene therapy is not identical to the gene that occurs in human DNA. Instead of the complete F8 and F9 genes, only the part of the gene required to code for the synthesis of the clotting factor is necessary. The terms used in publications for this section of DNA include cDNA (complementary DNA, which refers to the way it is synthesised) and transgene (meaning a gene transferred from one organism to another).

The natural structure of the transgene is altered to increase its activity and to assist its formulation as a medicine. It is modified by adding regions within the molecule (introns) that regulate or promote its activity (e.g. to enhance uptake or transport in a cell), and by switching the units (codons) that code for specific amino acids (codon optimisation). Alterations to the transgene affect its efficiency in inducing the expression of clotting factor, and possibly its immunogenicity and long-term safety.

The transgene for FVIII is further modified from F8 by removing a large portion that is not required for coagulation, known as the B domain. This substantially reduces the size of the molecule but also impairs how it is transported within the cell ‒ overcoming this means modifying the molecule further or restoring parts of the protein that were removed.

Small changes in the structure of the transgene can have a large effect on the expression of clotting factor. Such ‘gain of function point mutations’ occur naturally in humans, and one in particular has been used to improve the effectiveness of FIX gene therapy. In 2009, investigators described two brothers with thrombophilia whose FIX had a mutation at one point (position R338L), in which the amino acid leucine replaced arginine [1]. This small change rendered the mutant FIX five to ten times more active than ‘normal’ FIX and increased clotting activity eight-fold. The research was carried out in Padua, Italy, and the mutation became known as FIX-Padua. Current gene therapy for haemophilia B now uses a transgene with the Padua mutation (known as FIX-Padua or FIX-R338L), making it possible to achieve very high levels of FIX.

… and vectors

The transgene synthesises clotting factor only after it gains access to the cell cytoplasm. This is achieved by harnessing the mechanism by which viruses infect cells, using a virus as a vector to penetrate the barrier presented by the cell membrane. The virus currently used in most clinical trials is adeno-associated virus (AAV), a type of parvovirus. Many people naturally acquire AAV infection in infancy; the virus does not cause illness but it does provoke an immune response. There are several subtypes of AAV and they share a high proportion of antigens, meaning that immunity to one subtype is likely to affect the others. Clinical trials have used AAV subtypes 2, 5, 6 and 8 as vectors; AAV5 has fewest shared antigens and therefore the lowest number of potential patients with antibodies to it.

The transgene is inserted into the AAV genome, together with smaller sections that act as an enhancer, an intron and a promoter (a section that increases affinity for a particular cell type ‒ hepatocytes in the case of haemophilia gene therapy). The modified genome is then packaged within the AAV protein shell (capsid) to create the vector.

After intravenous infusion, the capsid binds to receptors on the surface of the hepatocyte (for gene therapy targeting the liver) and passes into the cell cytoplasm, where it is separated from the transgene. The transgene is not integrated into the cell’s DNA but remains in the cytoplasm, where it produces FVIII or FIX without regulation by the cell. The viral capsid proteins are transported to the cell membrane where they may be recognised by T cells, inducing an immune response. Recent gene therapy regimens have included a course of steroid treatment to suppress this response (see below). Gene therapy utilising stem cells involves inserting the transgene into stem cells before infusing them into the patient.

In early clinical trials, the response to gene therapy using an AAV vector was reduced or abolished by AAV antibodies; subsequently, potential trial participants have been screened to exclude those with antibodies. This suggests that gene therapy with the vectors currently in use may be unsuitable for many people. Alternative vectors that are now being evaluated include lentivirus and gamma-retrovirus, though both have safety concerns. Trials with adenovirus and non-viral vectors have shown them to be associated with a higher risk of adverse effects or less effective than AAV.

… and dose units

The dose of gene therapy is usually expressed as vector genomes per kg (vg/kg). The size of the dose is given as A multiplied by ten to the power of B – that is, A x 10B. In the case of valoctocogene roxaparvovec, the lower dose in clinical trials is 4 x 1013 vg/kg; this is usually abbreviated to 4e13, where e stands for ‘exponential’.

Less often, doses may be expressed as genome copies per kilogram (gc/kg), also as an exponential figure (e.g. 2 x 1013 gc/kg).

Clinical trials

The clinical trials registered with are summarised in the linked Table. Many are early stage (Phase I or II) studies involving small numbers of participants and designed to determine acute safety and the dose that should be tested in subsequent studies. Of these, three have been published in full [2,3,4]. An update of one of these trials, with follow up of 2–3 years, has now been published [5]. Two trials have started in patients with antibodies to the AAV5 vector (who have previously been excluded from these trials) and in patients with FVIII inhibitors. Most Phase I or II trials are too small to provide robust evidence of efficacy or safety; however, gene therapy is not a conventional medicine and the results from completed trials and the interim outcomes from ongoing trials have been interpreted very positively.

Seven Phase III trials designed to provide definitive evidence of efficacy and safety are currently underway in patients with haemophilia A (valoctocogene roxaparvovec, giroctocogene fitelparvovec) or haemophilia B (etranacogene dezaparvovec, FLT180a, fidanacogene elaparvovec).

Interpreting the trials

There are several caveats about the outcomes of trials of gene therapy for haemophilia:

  • Not all people with haemophilia were eligible to participate

The inclusion and exclusion criteria were carefully selected to optimise outcomes. For example, the eligibility criteria for Spark’s Phase I/II trial of SPK-9001 for haemophilia B were:

Inclusion criteria Exclusion criteria
Able to provide informed consent and comply with requirements of the study Evidence of active hepatitis B or C or currently on antiviral therapy for hepatitis B or C
Males ≥18 yrs with confirmed diagnosis of haemophilia B (≤2 IU/dL or ≤2% endogenous factor IX) Have significant underlying liver disease
Received ≥50 exposure days to factor IX products Have serological evidence of HIV-1 or HIV-2 with CD4 counts ≤200/mm3 (subjects who are HIV+ and stable with CD4 count >200/mm3 and undetectable viral load are eligible to enrol)
A minimum average of four bleeding events per year requiring episodic treatment of factor IX infusions or prophylactic factor IX infusions Have detectable antibodies reactive with AAV-Spark100
No measurable factor IX inhibitor as assessed by the central laboratory and have no prior history of inhibitors to factor IX protein Participated in a gene transfer trial within the last 52 weeks or in a clinical trial with an investigational drug within the last 12 weeks
Agree to use reliable barrier contraception until three consecutive samples are negative for vector sequences Unable or unwilling to comply with study assessments


These criteria exclude the individuals who have acquired hepatitis B or C or are experiencing the hepatic complications of past infection. Most people with severe haemophilia B would otherwise meet the criteria except for the requirement for no antibodies against the AAV vector: in one large study, the prevalence of antibodies to AAV in a general population in France varied with the subtype, being higher for AAV1 (67%) and AAV2 (72%) than AAV5 (40%), AAV6 (46%), AAV8 (38%) and AAV9 (47%) [6]. There was also cross-reactivity between AAV types but, more positively, titre levels were low in most people. Trials are now underway to determine whether high vector antibody titres can be overcome with high-dose gene therapy.

All trials reported so far have excluded children and it is not clear how effective gene therapy may be in younger people (because the transgenes are not replicated as the liver grows). Also excluded were people with mild or (in most trials) moderate haemophilia and women. No gene therapy for other inherited bleeding disorders has so far reached the clinical trial stage, so this development is confined to the haemophilia community.

Selecting a trial population in this way means that the results can reliably be applied only to people who share these characteristics, meaning that – at first – treatment will only be offered to a subgroup of the haemophilia population. In time, trials will be conducted with less restrictive criteria, provided no major safety concerns emerge.

  • Factor levels are not restored to 100% over time

The typical response appears to be a rapid increase in factor levels in the first weeks after treatment, followed by a decline and then a plateau or a very slow decline. The initial peak may be very close to 100%, but plateau levels range up to 80% in individuals or average 30% – 50%. Severe haemophilia is therefore being converted to mild haemophilia for most people.

  • Most outcomes have been reported for a few years at most

Gene therapy is a lifelong treatment and one for which there is no precedent. The long-term risks are small, as far as is known, and apparently greatly outweighed by the potential benefit. However, follow-up in most trials is only one to three years and the duration of response is far from clear. Studies in mice suggest the possibility of a small risk of hepatocellular carcinoma, perhaps due to low-level integration of the viral vector into host DNA; however, there is currently no evidence of this in humans, or of new safety concerns over time.

  • Results to date have come from small numbers of participants

Gene therapy is such a radical change in treatment that early trials have received unparalleled attention. It is not unusual for such trials to involve only tens of participants because they are intended to provide information about dosage and safety. Phase I and II trials of gene therapy have been greeted as if they were more definitive Phase III trials but, encouraging as the results have so far been, the numbers of participants have been too small to provide statistically robust evidence. Further, it is not possible to pool the results of several small trials because of differences between the vectors. Definitive evidence of outcomes will be available from a Phase III trial in 2020; two more will report in 2022.

Data from the Phase I/II trial of SPK-9001 provide a good example of how outcomes vary between individuals and how these differences make it difficult to extrapolate the results to a larger population in a way that is meaningful [7]. One year after infusion, FIX levels in eight participants ranged from about 10% to about 60%. After 34 weeks, levels were increasing in three people, had plateaued in one and were falling in three; one participant had stable levels at week 34 but no later data. Of two participants whose factor levels declined after developing an immune response within one month, one achieved a stable FIX level of about 80% at six months that was subsequently falling, and the second had a stable level of about 15% at one year, though with signs that the immune reaction may have persisted.

  • Immune response to gene therapy

In early trials of gene therapy, initially high factor levels rapidly declined in some individuals within two months of treatment. This was associated with an increase in blood levels of the liver enzymes alanine aminotransferase (ALT) and aspartate aminotransferase (AAT). Most cases appear to be due to a cytotoxic T cell response to the vector capsid protein expressed by hepatocytes, but additional possible mechanisms include an innate immune response associated with an inflammatory reaction. Reactive treatment with prednisone, initiated when liver enzymes start to increase, was usually sufficient to suppress the immune response and save the transduced cells. It is now routine to provide prophylaxis with a tapering dose of prednisone, starting at a dose of 60 mg/day.

  • Monitoring requirements for trial participants

Gene therapy is associated with intensive post-treatment surveillance, to the extent that the burden of regular infusions for factor replacement is substituted by a regimen of frequent clinic visits and blood monitoring. For many people this is possible only through shared care arrangements between the treatment centre and local healthcare services, and makes participation more difficult for individuals whose time is constrained by commitments such as employment or providing care. Whether and how this will impact on individuals and their families as gene therapy becomes more commonplace is not clear.

Latest outcomes

The most recent information about progress in current clinical trials comes from the innovator companies, usually as statements for investors. Different companies have reported different outcomes and it is not possible to compare products. Due to differences between the vectors and transgenes, it is not possible to pool the results of different clinical trials to improve analysis of outcomes.

Cost of gene therapy

It is currently unclear how much gene therapy is likely to cost in the UK or how its cost-effectiveness will compare with the best care now available (which might be prophylaxis with recombinant factor, extended half-life factor or newer agents such as emicizumab). Biomarin has indicated that, in 2020, it was considering a price of $2 – $3 million for  valoctocogene roxaparvovec, arguing that the lifetime cost of haemophilia is about $25 million [8]. The basic cost of emicizumab in the UK is about £470,000 per year; Spark Therapeutics has already marketed a gene therapy, Luxtarna, to treat a rare hereditary disorder of the eye, at a cost of $425,000. Novartis markets onasemnogene abeparvovec-xioi, a gene therapy for spinal muscular atrophy, which is priced at $2.1 million.

This cost of gene therapy will be offset by reduced spending on alternative treatments. First, factor replacement therapy and emicizumab are expensive (compared with drug therapies for more common long term disorders); trial evidence to date has shown that gene therapy is associated with a very large and sometimes total reduction in factor use. Second, gene therapy offers the prospect of higher factor levels, which would be expected to reduce the risk of long-term complications by even more than factor prophylaxis, offering large savings in the cost of future care. Third, gene therapy may offer improved quality of life for people with haemophilia and their carers.

US analysts have published a cost effectiveness analysis (funded by state and federal grants with no competing financial interests; available free online [9]). Assuming that gene therapy will cost $1 million over a period of ten years for a 30-year-old individual with uncomplicated severe haemophilia A, they concluded it would be less expensive than recombinant FVIII prophylaxis ($1.7 million) and would provide a greater gain in quality of life (using NICE’s favoured measure, the quality-adjusted life-year or QALY: 8.33 QALYs gained vs 6.62). The model assumed that gene therapy would be 90% effective and associated with a complication rate of 1% per month in the first year. Adjusting the various assumptions that went into the economic model showed that their conclusion was statistically robust, though it is unclear whether the impact of long term monitoring and the cost of prophylaxis to cover surgery were evaluated.

Innovative treatments are rarely both less expensive and more effective than the established options, and if this analysis is indicative of the outlook in other countries, it seems likely that demand for gene therapy will be high. This poses challenges for health budgets, unless risk-sharing arrangements are agreed to spread the high initial cost over a period of years so that the pattern of expenditure more closely resembles current spending on factor prophylaxis. However, gene therapy will be disruptive for health providers: it will transform the way in which haemophilia care is delivered for some people, while others who have already developed complications will need ongoing care. It is also unclear whether countries that presently have difficulty meeting the cost of conventional factor replacement therapy will be able to afford gene therapy.


Few innovations in healthcare offer the potential of gene therapy for haemophilia: the prospect of lifelong freedom from spontaneous bleeds and bleeding complications for individuals and their carers. However, gene therapy entails a long-term commitment from patients, and current technology does not remove the risk of transmitting haemophilia to their children. Early results from clinical trials are very promising for most people; however, they should be interpreted cautiously as they are based on relatively few participants and short-term outcomes. The cost of gene therapy will be high but, if early evidence of efficacy and safety is confirmed in the longer term, it is likely to be cost-saving over the lifetime of the patient compared with prophylaxis with a recombinant factor. How it will compare with emicizumab and extended half-life factors is not known.


  1. Simioni P, Tormene D, Tognin G, et al. X-linked thrombophilia with a mutant factor IX (factor IX Padua). N Engl J Med 2009;361:1671-5.
  2. Nathwani AC, Tuddenham EG, Rangarajan S, et al. Adenovirus-associated virus vector-mediated gene transfer in hemophilia B. N Engl J Med 2011;365:2357-65.
  3. Rangarajan S, Walsh L, Lester W, Perry D, et al. AAV5-Factor VIII Gene Transfer in Severe Hemophilia A. N Engl J Med 2017;377:2519-2530.
  4. George LA, Sullivan SK, Giermasz A, et al. Hemophilia B Gene Therapy with a High-Specific-Activity Factor IX Variant. N Engl J Med 2017;377:2215-2227.
  5. Pasi KJ, Rangarajan S, Mitchell N, et al. Multiyear Follow-up of AAV5-hFVIII-SQ Gene Therapy for Hemophilia A. N Engl J Med 2020;382:29-40.
  6. Boutin S, Monteilhet V, Veron P, Leborgne C, Benveniste O, Montus MF, Masurier C. Prevalence of serum IgG and neutralizing factors against adeno-associated virus (AAV) types 1, 2, 5, 6, 8, and 9 in the healthy population: implications for gene therapy using AAV vectors. Hum Gene Ther 2010 Jun;21(6):704-12.
  7. George LA, Sullivan SK, Giermasz A, et al. Hemophilia B gene therapy with a high-specific-activity factor IX variant. N Engl Med 2017; 377; 2215-27.
  8. BioMarin Sets High Price Tag for Hemophilia Gene Therapy Candidate. 1 March 2020.
  9. Machin N, Ragni MV, Smith KJ. Gene therapy in hemophilia A: a cost-effectiveness analysis. Blood Adv 2018; 2: 1972-8.