Summary. In the last three decades, hemophilia has moved from the status of a neglected and often fatal hereditary disorder to that of a fully defined group of molecular-pathological entities for which safe and effective treatment is available.
Hemophilia is likely to be the first widespread severe genetic condition to be cured by gene therapy in the third millennium.
In the socio-economic arena it remains a challenge to humanity to know that four-fifths of the world’s hemophiliacs still receive no treatment at all. Production of factor (F) VIII and IX in the milk of transgenic farmyard animals could provide a source of less expensive replacement therapy for developing countries.
Affordable gene transfer will be the ultimate solution for hemophilia in the third world as in the first. Thus it may be confidently predicted that the early new millennium will see an end to this ancient scourge.
Keywords: factor VIII, factor IX, hemophilia.

Introduction
The modern management of hemophilia started in the 1970s, when the increased availability of plasma-derived concentrates of coagulation factors and the widespread adoption of home replacement therapy led to the early control of hemorrhages and thereby to the reduction of the musculoskeletal damage typical of poorly treated patients. Prophylactic treatment was successfully initiated in Sweden, achieving the goal of preventing the
majority of bleeding episodes and further minimizing the impact of arthropathy. Hemophilia care became one of the most gratifying examples of successful secondary prevention of a chronic disease. No rose is without thorns, however, as concentrates manufactured from plasma pooled from thousands of donors were invariably contaminated with blood-borne viruses, that caused post-transfusion hepatitis B and non-A non-B (hepatitis C) in practically all treated hemophiliacs.
Chronic hepatitis was very frequent but appeared to be mild and non-progressive, so that benefits of concentrates seemed to outweigh risks at that time.
This optimistic perception of hemophilia changed dramatically in the early 1980s, when in Europe and in the USA 60–80% of persons with severe hemophilia became infected with the human immunodeficiency virus (HIV) that had contaminated concentrates. The scientific community reacted promptly to this tragedy. The last decade of the second millennium has witnessed the production of safer and safer plasma-derived concentrates of coagulation factors. Availability of recombinant factors was the outstanding result of dramatic progress in DNA
technology. Analysis of patients’ DNA has permitted identifi-cation of the gene lesions that cause hemophilia and allowed the disease to be controlled through carrier detection and antenatal diagnosis. New treatments have improved substantially the prognosis of patients infected with the hepatitis viruses and HIV and of those who develop alloantibodies (inhibitors) to factor VIII (FVIII) or factor IX (FIX). Finally, in the last few years the first experiments of somatic gene transfer have started in persons with hemophilia. This article will review the currently available options of treatment with plasma-derived and recombinant coagulation factors in patients with hemophilia, with and without inhibitors. The progress that can be expected in the near future, through the development of improved recombinant products and gene transfer, will also be outlined.
For other information on hemophilia, see the review of Mannucci and Tuddenham [1].
Plasma-derived factors The current safety of these products is based upon the adoption of measures meant both to decrease viral load in source plasma and to inactivate viruses that have escaped plasma screening. As
a consequence of these measures, progressively adopted in 1984–5 (moderate ‘dry’ heating, minimizing the risk of HIV infection, but not that of hepatitis C virus (HCV) infection), 1986–7 (stronger ‘dry’ and ‘wet’ heating or solvent/detergent, minimizing the risk of HCV transmission), 1996–7 (adoption of more than one virucidal method to inactivate non-enveloped viruses such as the hepatitis A virus) and 1999–2000 (adoption of PCR testing and quarantine of source plasma), the safety of plasma-derived factors has dramatically improved, so that no
significant transmission of the aforementioned blood-borne viruses has been unequivocally documented since the progressive adoption of these measures [2]. However, the highly thermoresistant B19 parvovirus is still transmitted by plasma concentrates [3]. Even though B19 infection is normally of little consequence in hemophiliacs, a few clinically significant events have been reported [3]. B19 infection must also be seen as
evidence that blood-borne viruses other than the hepatitis agents and HIV may still be transmitted.
Another perceived threat is that of new variant Creutzfeldt-Jakob disease [4], with the fear that the abnormal prion protein might be contained in, and transmitted by plasma coagulation factors and albumin [5]. Even though several studies, carried out also in multitransfused hemophiliacs, have shown that sporadic Creutzfeldt-Jakob disease is not transmitted by blood or its derivatives [6–9], these data cannot be necessarily extrapolated
to variant disease. The latter has a different incubation period and the number of blood donors potentially incubating the transmissible agent may be much higher than that of donors incubating sporadic disease. An objective cause for concern is the demonstration of transmission by whole blood transfusion of spongiform encephalopathies in sheep [10]. On the other hand, the fractionation processes used to purify plasma proteins,
including albumin and coagulation factors, contribute signifi-cantly to clear abnormal prions (more than six infectivity logs), making it unlikely that these agents even if present in plasma would be carried into the final products at concentrations capable of causing clinical disease [11]. Are other ‘new’ infectious agents a possible threat? The dramatic experience with HIV tells us that this possibility should not be overlooked.
For instance, the recently documented transmission through transfusion and organ transplantation of the West Nile virus prompts specific surveillance, even though this enveloped flavivirus is likely to be inactivated by the currently used virucidal methods [12].

Recombinant factors

Current status
Two preparations of full-length recombinant FVIII (rFVIII) formulated in human albumin were licensed in the early 1990s.
Several clinical studies have since demonstrated their excellent efficacy, approximately 80% of bleeding episodes being controlled by a single dose [13–15]. Very high levels of safety from viral transmission and immunological reactions against animal proteins used in cell culture and in the manufacturing process have been observed. The current view is that rFVIII products trigger no more inhibitors than plasma-derived factors. Second generation rFVIII products are now licensed. One is formulated in sucrose instead of human albumin [16], another lacks the
large B domain of the full-length protein but still retains coagulant activity [17]. Human albumin is employed in the culture medium but none is present in the final formulation of this product and the manufacturing process includes a virus inactivation step. A newer preparation of FVIII with no exposure to human or animal proteins during manufacturing or formulation (except for mouse and hamster proteins) is currently undergoing clinical trials [18]. On the whole, secondgeneration rFVIII products are perceived as an improvement because there is no other human protein in the final formulation, but are more expensive than albumin-formulated products (20–
30% more). No human or animal protein (except hamster protein) is used during the purification steps of rFIX, or is added to the final product for formulation. Pharmacokinetic and efficacy studies in previously treated patients gave satisfactory results [19], even though the in vivo recovery is substantially lower than that of plasma-derived FIX, probably because of minor posttranslational differences in the recombinant protein.
Which option?
The choice between plasma-derived and recombinant factors has to consider that plasma-derived factors are becoming safer and safer; that recombinant factors cost from 20–50% more; and, most importantly, that the production capacity of recombinant factors is improving but is still limited, as witnessed by a recent period of dramatic shortage. On the other hand, recombinant factors are inevitably perceived as safer than plasmaderived
factors, so that countries like Canada and Ireland chose to switch all hemophiliacs to recombinant products. In the USA, approximately 60–70%of severe hemophiliacs currently use recombinant products,andtheproportionisincreasing. InEurope the proportion of recombinant factor users is generally smaller, so that some countries like Italy had to develop priority guidelines [20]. We recommended choosing recombinant factors first for newly diagnosed, previously untreated hemophiliacs and then for those who have been spared from blood-borne infections despite previous exposure to plasma-derived factors [21]. These restrictions and selection policies may change soon, as factor production is increasing and cost should decrease. It can be
predicted that in the first decade of the twenty-first century replacement therapy will continue to evolve towards using more and more recombinant factors, at least in the richest countries.
On the other hand, one should make a point of reassuring the large number of persons with hemophilia that are using and will continue to use plasma-derived factors, the only foreseeable option for 80% of the persons with hemophilia worldwide who have at the moment no or limited access to any replacement material. The risks of blood-borne infections transmitted by plasma factors are more theoretical than real, and patients and
policymakers should be educated to distinguish real from perceived risks. One would expect increased availability and decreasing prices for these products but this is not occurring, perhaps because the very demanding precautionary measures currently enforced by regulatory authorities make source plasma more and more scarce and expensive. Progress is warranted in the quality of plasma fractionation technologies. The yield of
FVIII from source plasma is still only 510%, a loss that is difficult to accept in an era of high technology!
Finally, it should be reiterated that desmopressin (DDAVP) is the treatment of choice in responsive patients with mild hemophilia A (and vonWillebrand disease). Its early adoption in Italy in the late 1970s and early 1980s, at the time of the onset of the HIV epidemic, has minimized the proportion of patients with mild hemophilia Awho became infected. This is much smaller than that of a comparison group of Italian patients with mild hemophilia B, who being unresponsive to desmopressin could only be treated with unheated plasma-derived FIX [22].
Future developments The current limitations for rFVIII and rFIX are availability and high cost, which depend at least in part on the relatively poor expression of these factors in the mammalian cells systems used to produce them. An attempt to improve expression of FVIII has already been realized with the B-domainless FVIII, which is
expressed 2-fold more efficiently than wild-type FVIII. Other targets for the improvement of recombinant coagulation factors are an increase of the specific activity of FVIII and greater resistance to inactivation by activated protein C (reviewed by Saenko et al. [23]). Efforts are also directed to engineer recombinant forms of FVIII that preserve the functions in coagulation but may be less immunogenic, i.e. less likely to trigger the onset of inhibitors because the strongest and more frequent antigenic determinants are removed and replaced with
less immunogenic epitopes of porcine FVIII [24]. Such a molecular product could be considered the first line treatment in patients previously unexposed to any source of FVIII (previously untreated patients, PUPs), because it would have less ability to elicit an immune response. It can also be expected that this product may be successfully employed to treat bleeding episodes in patients who have already developed an inhibitor,
because multiple substituted human/porcine hybrid FVIII should be less inactivated than wild-type molecules [24].
Engineered molecules with a longer plasma half-life should help to increase the time intervals between doses, particularly in the setting of continuous prophylaxis. One might envisage, for instance, the possibility of disrupting the sites in the FVIII molecule involved in its binding to the low-density lipoprotein
receptor-related protein, the major hepatic mechanism for FVIII clearance from the circulation [25,26]. Other targets for slower FVIII clearance and longer plasma half-life are cell-surface heparin sulfate proteoglycans, which are among the major glycoprotein components of the extracellular matrix and cooperate with lipoprotein receptor-related protein in FVIII clearance [27]. To my knowledge, only the approach attempting to reduce immunogenicity and antigenicity by introducing porcine substitutions within the immunogenic epitopes located in
human FVIII A2, A3 and C2 domain, is being considered for clinical purposes at the moment [24].

Treatment of patients with inhibitors

Current status
Until the 1980s, the risk of death due to uncontrollable bleeding was high in hemophiliacs with inhibitors, particularly when emergency surgery was needed, and limb-threatening hemorrhages such as hemarthoses and muscle hematomas could not be treated optimally. Treatments that bypass the need for FVIII and FIX in intrinsic coagulation have improved this situation.
The principle underlying these treatments is to bypass the defect in intrinsic coagulation by activated forms of FVII, FIX and FX contained in the prothrombin complex concentrates used in the routine treatment of FIX deficiency; or in those purposely manufactured to contain Factor Eight Inhibitor Bypassing Activity (FEIBA) in ‘controlled’ amounts. Randomized, double-blind, placebo-controlled trials have shown that both types of products are efficacious in controlling 40%60% of spontaneous bleeding episodes [28–30]. This success rate is
definitely a great improvement over the past situation of no effective treatment, but is lower than the 90%95% success rate obtained with two doses of FVIII or FIX in hemophiliacs without inhibitors [13–15].
In the last few years a new product, recombinant activated FVII (rFVIIa), has been licensed. It is thought to ensure hemostasis by binding, directly or in complex with tissue factor, to negatively charged phospholipids exposed on the surface of activated platelets [31]. According to an alternative theory the therapeutic effect is due to increasing the ratio of FVIIa to FVII [32]. Cell localization of the enzymatic reactions that lead to the
generation of FXa and ultimately of thrombin should reduce the risk of systemic coagulation activation and of thrombotic complications.
However, myocardial infarction has been reported after administration [33,34]. Infused as bolus at the recommended doses of 90–120 mgkg1, to be repeated 2–4 times at 2–3 h intervals, rFVIIa is claimed to stop approximately 80%90% of spontaneous hemorrhages and to prevent excessive bleeding during major surgical procedures [35–39]. The best efficacy results were obtained in uncontrolled studies carried out in the frame of home therapy, that makes possible early intervention [35,36]. Recombinant DNA manufacturing means this product is perceived as safer than activated plasma-derived factors. A randomized controlled trial comparing the clinical efficacy and safety of rFVIIa with that of the plasma-derived activated prothrombin complex concentrate FEIBA is ongoing. There are attempts to improve the currently available bypassing agents. A recombinant preparation of activated prothrombin complex factors, and FVIIa analogs with higher specific activity
and affinity for platelets, is being developed [40].
The concept of suppressing the production of FVIII inhibitors by building tolerance in patients through repeated exposure to the antigen was first implemented by Brackmann [41] through the administration of huge daily doses of FVIII (200Ukg1).
Schedules of immune tolerance based on lower doses (25–50Ukg1 given every other day) are also apparently successful [42]. These treatments do not eliminate the production of FVIII inhibitors, but induce the production of neutralizing anti-idiotypic antibodies. About two-thirds of the treated hemophilia A patients respond to immune tolerance with a decrease of inhibitor levels below 1 mmL1, but only one-fourth also have normal FVIII recovery and plasma half-life [43]. Predictors of response are low levels of inhibitor before immune tolerance, Hemophilia: treatment options 1351 lower peak levels and lower historical peak levels. More than 20 years after the first study on immune tolerance, there is still little ground for choosing between high- and low-dose and highand low-purity regimens of induction, but an international
randomized study is now addressing these issues.
Future developments A few newer approaches to the treatment and prevention of inhibitors are emerging, mostly in experimental animals. Idiotypic regulation may form the basis of new methods for the induction of long-term immune tolerance in patients with anti- FVIII antibodies (reviewed by Lacroix-Desmazes et al. [44]).
Since anti-FVIII antibodies with inhibitory activity can be neutralized by anti-idiotypic antibodies, active immunization with idiotypic antibodies or with polypeptides that mimic idiotypes of anti-FVIII antibodies may generate anti-idiotypes capable of neutralizing the neutralizing activity of inhibitors.
Another approach is based upon the disruption of T celldependent B-cell activation by antigen-independent blockade of the interaction between B and T cells. Attempts to inhibit the production of FVIII antibodies by blocking the CD40-CD40L pathway has been started in three hemophilia A patients with high titer inhibitors who received monthly exposures to FVIII in the presence of a humanized mouse monoclonal antibody to human CD40L [45]. Blockade strategies targeting the B7/ CDC28 costimulatory pathway are other potential approaches
to the neutralization of secondary anti-FVIII immune responses [46,47]. The use of these approaches to prevent the primary onset of inhibitors is less promising, because our capacity to identify patients at high risk through the identification of the type of DNA lesion [48] or HLA genotyping [49] is limited.
Another potential approach may be the design of molecules that mimic the prevalent epitopes recognized by inhibitor in the FVIII molecule and that function as ‘inhibitor inhibitors’.

Gene transfer
Hemophiliacs provide excellent combinations of features for a favorable response to gene transfer [50–53]. Clinical manifestations are entirely attributable to lack of one or other single specific gene product, the gene product circulates in minute amounts in plasma, plasma levels of FVIII and FIX do not require strict control, a minor increase in plasma levels will markedly ameliorate the symptoms of severe cases, murine and
canine models of hemophilia are available, FVIII and FIX do not have to be expressed in their normal tissue of synthesis, and they can be produced by any cell type provided the proteins can gain access to blood.
Early efforts focused on retroviral vectors proved to havemany problems. These and other problems have been gradually overcome, partly by switching to adenovirus and adeno-associated virus, and partly by redesigning the inserts and promoters and using novel gene delivery systems in vivo and ex vivo. Using these approaches sustained correction of hemophilia A and B in mice and dogs has been achieved [50–54].
Three gene transfer trials in patients with hemophilia A and B are now completed and two are ongoing Table 1. The first Avigen study in patients with hemophilia B was a dose escalating safety trial based upon the intramuscular injection to eight adults of an adeno-associated virus-based vector. A favorable effect on plasma levels of FIX (up to 3.7%) and/or concentrate usage was evident in three of eight patients [55]. Muscle biopsy
demonstrated the presence of vector genome and expression of 1352 P. M. Mannucci