SEARCH

SEARCH BY CITATION

Keywords:

  • factor VIII;
  • gene therapy;
  • hemophilia A;
  • inhibitors;
  • platelet

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Summary. Background: Gene therapy strategies directed at expressing factor (F)VIII in megakaryocytes has potential advantages in the treatment of hemophilia A. Among these is that platelet (p) FVIII may be effective in the presence of circulating anti-FVIII inhibitors. Objective: We examined in a murine transgenic model whether pFVIII could correct the coagulation defect in FVIIInull mouse in the presence of circulating inhibitors. Methods: FVIIInull mice that were transgenic for pFVIII (pFVIII/FVIIInull) were compared with FVIIInull mice receiving infused FVIII in a FeCl3 carotid injury model in the presence of anti-FVIII inhibitors. Results: After injury, pFVIII/FVIIInull mice were significantly more resistant to circulating inhibitors than after plasma FVIII correction in both an acute and chronic models of inhibitor exposure even although in the chronic model, significant amounts of inhibitor were stored within the platelets. Furthermore, bleeding in the pFVIII mice in the presence of inhibitors was not as a result of the development of thrombocytopenia. Conclusion: In FVIIInull mice, pFVIII provides improved, but limited, protection in the presence of inhibitors of ∼6-fold greater Bethesda Units per mL relative to infused FVIII. Our findings differ from a recent report using a tail-clip exsanguination assay on the degree of efficacy of pFVIII in the presence of inhibitors. We propose that this difference in outcome is as a result of the sensitivity of the tail-vein exsanguination model to low levels of pFVIII.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Substantial effort has been directed at developing gene therapy for the care of patients with hemophilia A in whom there are significant qualitative or quantitative defects in coagulation factor (F)VIII [reviewed in 1]. While a number of different tissues have been targeted in these efforts, all aim at correcting plasma FVIII levels to a therapeutic range [reviewed in 2]. One exception has been gene therapy directed towards expression of FVIII within megakaryocytes [3]. In a murine model of platelet (p) FVIII, we had shown that the FVIII is localized within platelet α-granules without concurrent plasma correction [3,4]. Potential advantages of expression of FVIII within platelets include that pFVIII may have a longer half-life than plasma FVIII and be more effective than an equal plasma level in correcting the coagulation defect in FVIIInull mice [5]. Our data showed that these advantages did occur and that low levels of platelet human B-domainless (hbd) FVIII, equivalent to a plasma correction of ∼9% antigen, was sufficient to correct the clotting defect in a FeCl3 carotid artery injury model equivalent to giving a 25–50% human FVIII infusion [3].

Subsequently, using mice with a lower pFVIII level per platelet, Qizhen Shi et al., demonstrated that another advantage of pFVIII is to retain this correction even in the presence of high-titer anti-FVIII inhibitors [6]. The relative improvement of pFVIII over infused FVIII appeared to be greater than a two hundred and fiftyfold increase in the level of Bethesda Units (BU) per mL. Several editorials noted that this pFVIII gene therapy model may be of benefit in patients with hemophilia A who have problematic inhibitors [7,8]. As ∼30% of patients with severe hemophilia A develop anti-FVIII inhibitors [9] and because not all of these can be managed with present-day therapies [10], such a gene therapy model would be welcomed. However, the bleeding model used to demonstrate protection from inhibitors in this study involved overnight survival following tail-clipping. The tail-clip exsanguination model was based on the original hemostatic defect noted in FVIIInull mice and can be corrected by tail cauterization or by infusing FVIII [5]. However, we had described the characterization of a series of different pFVIII mice transgenic founder lines that included lines with low and even with no detectable pFVIII antigen or activity that still survived this exsanguination model [3]. Several lines that had no detectable hbdFVIII mRNA by reverse transcription-polymerase chain reaction (RT-PCR) survived (data not shown). These finding suggest that the tail-clip exsanguination model may be too sensitive to low levels of pFVIII and may overestimate efficacy in the presence of anti-FVIII inhibitors.

We therefore tested our transgenic line with the highest pFVIII level in the FeCl3 carotid artery injury model [11] to examine how effective pFVIII was in the presence of circulating anti-FVIII inhibitors. We had previously shown gradations of improvement that correlate with the amount of human FVIII infused into FVIIInull mice over a 5% to 50% antigen correction range. We also demonstrated that FVIIInull mice, which were also co-transgenic for pFVIII (pFVIII/FVIIInull), showed improvement in the FeCl3 carotid artery injury model correlated with pFVIII content. Using this model to study the efficacy of pFVIII in the presence of inhibitors, we confirm that pFVIII is protective against circulating inhibitors, but with only an ∼6-fold improvement in BU per mL level relative to a comparable FVIII infusion. The implication of these findings is that platelet-directed gene therapy may be of limited benefit in gene therapy for patients with hemophilia A who have high-titer inhibitors.

Materials and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Characterization of mice studied

The FVIIInull mice were previously described and had exon 16 deleted from the F8 gene [5]. The pFVIII transgenic mouse had also been described [3] and had the hbdFVIII cDNA (kindly provided by Randy Kaufman, University of Michigan [12]) driven by the glycoprotein Ibα promoter (kindly provided by Jerry Ware, University of Arkansas [13,14]). All FVIIInull and pFVIII/FVIIInull mice studied were littermates. Genomic phenotype was determined by PCR analysis of DNA isolated from tail clippings as described [3]. Experimental approval was obtained from the Children’s Hospital of Philadelphia Animal Care and Use Committee.

Antibodies studied, inhibitor levels and determination of antibody half-lives

The following murine monoclonal inhibitors were studied: ESH8 (American Diagnostica, Stamford, CT, USA) [15,16] and GMA-8021 (Green Mountain Antibodies, Burlington, VT, USA). Most studies shown were done using a 1:5 (μg/μg) of ESH8:GMA-8021.

Levels of anti-FVIII activity in these antibodies were determined using a modified Bethesda assay [17]. Briefly, sequential dilutions of individual antibodies or antibody mixtures were made in FVIII-deficient mouse plasma or buffer and then incubated with an equal volume of 1 U mL−1 recombinant human FVIII [Advate, kindly provided by Baxter Healthcare (Deerfield, IL, USA)] or hbdFVIII (Refacto, Wyeth, Andover, MA, USA) at 37 °C for 2 h, and residual FVIII coagulation activity was subsequently analyzed by chromogenic assay (Coamatic®Factor VIII kit, Chromogenix, MA, USA) according to the manufacturer’s instruction. The definition for Bethesda titer is the dilution of the inhibitor that causes a 50% inhibition of 1 U mL−1 of FVIII.

Half-life of the inhibitor mix antibodies was done after intravenous (i.v.) injection of 20 μg of inhibitors in 100 μL phosphate-buffered saline (PBS) into FVIIInull mice. Blood was collected at intervals for up to 10 days after antibody administration and isolated platelet poor plasma (PPP) was stored at −80 °C until analysis by ELISA as described below.

Inhibitor/FeCl3 carotid artery injury studies

FeCl3-induced carotid artery injury was performed as described previously after a 2-min exposure to 15% FeCl3 [3] using FVIIInull and pFVIII/FVIIInull littermate 20–25 g, 6–10-week-old mice of both sexes. In acute studies, FVIIInull mice were given a 12.5% FVIII (Advate) correction [3] delivered via an exposed jugular vein. Two minutes after this infusion, inhibitor was injected in doses up to 20 μg per 20-g mouse in 100 μL of PBS. After an additional 2 min, a sclerotic injury was induced in the carotid artery as described [3]. pFVIII/FVIIInull mice were studied in parallel and given the same FVIII inhibitor infusions as FVIIInull mice before they had a similar carotid injury induced.

For chronic inhibitor studies, pFVIII/FVIIInull animals were given 2.0 μg total per 20 g mouse of the 1:5 (μg/μg) FVIII inhibitor mixture on days 0, 3, 6 and 9 by retro-orbital injection. A control group was given control isotype IgG, a gift from G. Arepally (Duke University, NC), in the same amounts over the same time schedule. FeCl3 injury studies were performed at day 10, one day after the last inhibitor infusion.

After the vascular injury, mice were analyzed over a 30-min window [3]. Outcome was measured as ‘percent occluded flow’, representing the percent of the total anticipated blood flow over the 30 min of study occluded by the developing thrombus. ‘Percent complete occlusion’ outcome was measured as the percent of mice tested that developed complete, stable occlusions that lasted for ≥ 10 min.

Blood studies

Plasma levels of FVIII inhibitor were measured by ELISA in PPP and in platelet extract using blood obtained by retro-orbital puncture into acid citrate dextrose (ACD; pH 4.5, 10:1 vol/vol) and immediately diluted 1:1 (vol/vol) with modified Tyrode’s buffer (134 mm NaCl, 3 mm KCl, 0.3 mm NaH2PO4, 2 mm MgCl2, 5 mm HEPES, 5 mm glucose, 12 mm NaHCO3, 1 mm EGTA, 0.1% BSA; Sigma A7030, fatty acid–free, pH 5.5) to prevent platelet activation. In chronic studies, blood was taken 2 min before each injection of inhibitor and from the jugular vein immediately before FeCl3 carotid injury. Samples were spun 15 min at 170 × g to remove red cells, and then respun at 800 × g for 10 min to separate platelets and PPP. Platelets were washed twice with modified Tyrode’s buffer, adjusted to 1 × 107 in 100 μL PBS, stimulated with thrombin (1 U mL−1, Sigma, St. Louis, MO, USA) for 15 min at 37 °C, and stored at −80 °C along with the isolated PPP. Nunc 96-well flat-bottom immuno plates (Nalge Nunc international, Rochester, NY, USA) were coated overnight at 4 °C with 25 units human FVIII (Advate). Plasma (50 μL of 1:100 dilution) or platelet extract (50 μL of 1:10 dilution) were added per well and incubated overnight at 4 °C. The 1:5 (μg/μg) inhibitor mix was used to generate standard binding curves. Binding of FVIII inhibitor IgG to the antigen was detected using 1:104 dilution of peroxidase-conjugated Affinipure goat anti-mouse IgG (Jackson Immuno Research Laboratories, Westgrove, PA, USA) and developed using a TMB Peroxidase Substrate Kit (KPL, Gaithersburg, MA, USA). Optical density (OD) at 450 nm was measured in a plate reader (Spectramax 190, Molecular Devices, Sunnyvale, CA, USA).

After an infusion of FVIII inhibitors into pFVIII/FVIIInull mice, complete blood counts were determined on retro-orbital samples using a Hemavet 1500 FS set to mice (Drew scientific, Dallas, TX, USA). These studies were done prior to infusing the inhibitor, 30 min afterwards and 1, 3 and 7 days afterwards.

Statistical analysis

Statistical analysis was done using a two tailed t-test and was generated using Microsoft Excel 2004 for Mac (Version 11.3.7). P-values for percent occluded flow, and the percent complete occlusions in the chronic inhibitor studies were compared with the control arm receiving IgG control.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Characterization of monoclonal anti-FVIII inhibitors studied

The ability of inhibitors to interfere with clotting in the pFVIII/FVIIInull mice was studied using two monoclonal anti-human FVIII inhibitors: ESH8 binds to the light chain of the C2 domain of human FVIII [15,16], and requires that the FVIII be bound to von Willebrand factor (VWF) for its inhibitory activity [18]. The other murine monoclonal anti-human FVIII antibody, GMA-8021, binds to the A2 domain, but its VWF dependency has not been tested (F. Church, personal communication). We measured the inhibitory activity per μg of each monoclonal antibody alone and of a 1:5 (μg/μg of ESH8:GMA-8021) mixture using recombinant human FVIII (Fig. 1) in the presence and absence of FVIII-deficient murine plasma as a source of VWF (Fig. 1). In the absence of the FVIII-deficient plasma, ESH8 is not active, while both the GMA-8021 and the 1:5 (μg/μg) mix demonstrated 50% inhibition of activity of 1 unit of the FVIII per 0.04 μg of inhibitor (Fig. 1A). In the presence of the FVIII-deficient plasma, GMA-8021 and the mixture demonstrated the same level of inhibitor activity as in its absence, but ESH8 showed 50% inhibition of 1 unit of FVIII per 0.13 μg of inhibitor (Fig. 1B). These data are consistent with the known dependency of ESH8 on VWF and with ESH8 having an ∼3-fold lower inhibitory potency than GMA-8021. In addition, Fig. 1B shows that GMA-8021’s inhibitory activity is not VWF dependent, consistent with GMA-8021 binding to FVIII distal to the VWF-binding site [19]. As we had previously shown that functional pFVIII is largely stored in platelets independent of VWF, we believed that these two distinct inhibitors might help target the site of any observed anti-FVIII inhibitory activity against pFVIII.

image

Figure 1.  Characterization of inhibitors. Inhibition of factor (F)VIII activity as measured by Bethesda assay [17] for ESH8 (open diamonds), GMA-8021 (black circles), a 1:5 (μg/μg) mix of ESH8:GMA-8021 (gray squares) and an IgG control (x). (A) With no added plasma or (B) with FVIIInull murine plasma. Representative of two separate studies done in duplicate. The data shown were done using human FVIII (Advate) although similar results were seen using hbdFVIII (ReFacto, data not shown).

Download figure to PowerPoint

Ability of inhibitors to block pFVIII in an acute model

We have previously shown that unlike wild-type (WT) mice, which consistently develop stable, occlussive clots after FeCl3 carotid injury, FVIIInull mice never develop such clots [3]. We also found that pFVIII/FVIIInull mice, whose platelets contain the equivalent of 9% plasma human FVIII antigenic levels and of ∼3% activity correction in FVIIInull mice [3], develop such clots to virtually the same extent as WT mice. Furthermore, infused human FVIII into FVIIInull mice is corrective in this model in a dose-dependent fashion so that a 25–50% antigenic correction results in near-normal clotting. We now used this system to test whether pFVIII is protective in the face of infused inhibitors. We focused on infused inhibitors because it allowed us to closely regulate the antibody level that was present in the recipient animal, and because non-adjuvant exposure to human FVIII was non-immunogenic in the pFVIII/FVIIInull mice (data not shown). Figure 2 supports our prior studies showing that pFVIII/FVIIInull mice effectively correct the coagulation defect, while an infusion of FVIII to give a 12.5% antigen correction was almost equally effective as measured by percent overall occlusion of blood flow over the 30-min study and by the percent of animals studied that achieved stable complete occlusions of ≥ 10 min (Fig. 2A,B, respectively, at baseline with no inhibitor infused).

image

Figure 2.  Acute inhibitor infusion FeCl3 carotid artery studies. Shown is a comparison of the relative protection from infusion of the inhibitor mixture in pFVIII/FVIIInull mice (grey squares) vs. FVIIInull mice (black circles) after a 12.5% antigenic correction. The amount of inhibitor mixture infused per 20-g mouse is shown on the X-axis. In (A), the mean ± 1 standard error (SE) of the percent-occluded flow representing percent of total anticipated blood flow over the 30-min study occluded by the developing clot. Top gray bar is mean ± 1 SE for pFVIII/FVIIInull mice without inhibitor and bottom gray bar is the same for FVIIInull mice. In (B), the percent of studied animals under a specific set of conditions that had stable total occlusions lasting ≥ 10 min is shown. Top gray bar is mean for pFVIII/FVIIInull mice without inhibitor and bottom gray bar is same for FVIIInull mice, where none of the mice developed stable occlusions. The numbers next to the circles and squares are the number of animals evaluated at that point.

Download figure to PowerPoint

Infusion of the two different anti-human FVIII murine monoclonal inhibitors were individually unable to limit clot formation in the pFVIII/FVIIInull mice except at doses of >100 μg per 20-g mice, and at this amount, infused control IgG also began to interfere with thrombosis (data not shown). However, a combination of 1:5 (μg/μg) of the two inhibitors was able to effectively inhibit thrombosis in these mice at lower doses (Fig. 2A,B). Comparison of the inhibitory curve for the pFVIII/FVIIInull mice and for the corrected FVIIInull littermate mice show that pFVIII is protective to inhibitor concentrations ∼ seventyfold higher than after a similar plasma correction, which would be equivalent to an ∼6-fold higher efficacy in BU mL−1, whether one measures the percent of total decrease in blood flow (Fig. 2A) or the percent of studied animals that develop stable occlusions (Fig. 2B).

FVIII is known to bind to the platelet surface [20]. One possible mechanism that could explain the effect of the infused inhibitors limiting clotting in the pFVIII/FVIIInull mice is that these mice developed significant thrombocytopenia secondary to the inhibitors binding to surface FVIII. We therefore infused a single dose of 20 μg per 20-g mouse of the mixed inhibitor, a dose shown to completely prevent thrombosis in the pFVIII/FVIIInull mice (Fig. 2), and followed platelet counts over the following week. This infusion had no effect on platelet counts (Fig. 3A), hemoglobins or total white cell counts (data not shown).

image

Figure 3.  Change in platelet counts after infusion of inhibitor and inhibitor half-life. A single dose of the 20 μg per 20-g mice dose of the inhibitor mix was infused. (A) Platelets counts were measured after inhibitor infusion (black circles) or IgG control (gray squares) in pFVIII/FVIIInull mice as mean ± SE relative to baseline counts. (B) Measured level of anti-human FVIII inhibitor over a 10-day period of study. Initial and terminal half-lives are shown. Mean ± 1 SE is shown. In (A) and (B), studies were done in four mice, each measured in triplicate.

Download figure to PowerPoint

Ability of inhibitors to block pFVIII in a chronic model

Megakaryocytes and platelets absorb immunoglobulins from the surrounding milieu and store this immunoglobulin in their α-granules [21,22]. The study described in Fig. 2 involved a short exposure to anti-human FVIII antibodies and would not involve significant uptake of these antibodies into platelets (see below). As we had shown that pFVIII is mainly stored within circulating platelet α-granules [3,4], we wondered whether the simultaneous presence of stored inhibitors along with FVIII would decrease the efficacy of pFVIII. As the likely therapeutic application of a platelet-based gene therapy may involve patients with hemophilia A and chronically circulating inhibitors, it is important to see the effect of having inhibitors localized within the circulating platelets on pFVIII efficacy.

Before we established a model of chronically infused inhibitors in the mice, we determined the half-life for the murine inhibitor mix. We again infused a single dose of the mixed inhibitor (20 μg per 20-g mouse). The initial equilibration half-life of the mouse inhibitors was ∼24 h (Fig. 3B), and the later half-life was ∼10 days, consistent with half-lives reported for other murine antibodies in mice [23].

We, therefore, established a model of chronic inhibitors by repeatedly infusing inhibitors into pFVIII/FVIIInull mice over a 10-day period. Each dose of inhibitor was 5 μg per 20-g mice, a dose that had a threshold affect on the ability to clot in the acute FeCl3 carotid artery injury model (Fig. 2). The design of the protocol was to have a total plasma inhibitor level in the chronically infused mice similar to that in mice after a single dose (Fig. 4A). Intermediate plasma inhibitor levels were consistent with the determined half-life (data not shown). Final plasma inhibitor titers in these chronically infused mice were slightly higher than mice infused acutely with 10 μg per 20-g mice of inhibitors, but less than that in mice infused acutely with twice that dose (Fig. 4B).

image

Figure 4.  Comparison of acute to chronic inhibitor infusions. All studies were done in pFVIII/FVIIInull mice except when specifically noted to be FVIIInull mice. (A) Schematic of infusion and study schedule, including when inhibitor mix was infused (black thin arrows), blood was sampled (circle-head arrows), and the FeCl3 carotid artery injury studies done (open arrow). The single infusion, acute model is shown above the chronic model. In (B), the mean ± 1 SE of the plasma inhibitor levels are shown on the left for two different dose acute studies of 5 and 10 μg per 20-g mouse. On the right two chronic studies, one with the inhibitor mix and one with control IgG, each involved giving repeat doses of 5 μg per 20 g mouse. Gray bars are plasma inhibitor levels and the black bars are platelet-associated inhibitor levels. In (C), the mean ± 1 SE of the percent-occluded flow (open bars) and the percent of animals with complete occlusion (dark gray bars) are shown. Also FeCl3 injury studies for FVIIInull mice are shown for comparative purposes. *P < 0.01 compared with the chronic IgG control study. In (B) and (C), each study was done in ≥ 6 mice.

Download figure to PowerPoint

At the end of the 10-day study, inhibitors were detectable in the platelets of the pFVIII/FVIIInull mice, but were not present in the two acute studies done in parallel (Fig. 4B). In spite of the presence of inhibitors within the platelets, the decrease in the ability to form an occlusive thrombi in the FeCl3 injury model in the chronically treated mice was similar to that seen in concurrently studied acutely treated mice at the higher inhibitor dose (Fig. 4C). This was true whether we measured the ability to clot by percent decrease in blood flow over the 30-min study or by the percent of mice that developed stable occlusive thrombi.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

We had proposed that platelet delivery of an ectopically-expressed protein may be a strategy for targeted delivery of that protein to a site of vascular injury [24]. We have successfully presented two proof-of-principle examples of this: (i) the efficacy of pFVIII in the amelioration of hemophilia A in spite of no rise in plasma FVIII levels [3], and (ii) the ability of platelet-stored urokinase to lyse thrombi without untoward systemic fibrinolysis [24]. These unconventional delivery systems for plasma proteins may have certain advantages and disadvantages compared with plasma delivery, but require careful analysis to define their positive and negative features as these may not be intuitively obvious.

One potential advantage of pFVIII is that it may be protected from inhibitors and remain effective in providing FVIII at sites of vascular injury. Studies by Qizhen Shi et al., using pFVIII/FVIIInull mice, showed a remarkable ability of pFVIII to protect FVIIInull mice [6]. pFVIII may protect to >two hundred and fiftyfold higher BU mL−1 level more than infused FVIII although this study involved only a single FVIII infusion at its initiation, and given the short half-life of infused human FVIII in mice [25], little may have been present at the conclusion of this overnight injury. This protection occurred in the setting of mice with levels of pFVIII ∼3-fold lower than in the present study. However, their study utilized an overnight tail-clip exsanguination model. We had reported that in mice with different transgenic levels of pFVIII that while all survived tail clipping when on a FVIIInull background, they all developed significant anemia relative to WT mice in a related 8-h cuticular bleeding model [3], suggesting that the tail-clip study may underestimate the remaining bleeding diathesis. Consistent with this insensitivity to the risk of bleeding in the tail-clip model, pFVIII/FVIIInull mice lines with undetectable platelet FVIII message by RT-PCR or with no pFVIII antigen or activity detectable survived tail clipping. In a recent preliminary report by Qizhen Shi et al., bone marrow-transplanted FVIIInull mice, reconstituted with a 100:1 mixture of WT:pFVIII stem cells, still survived tail-clip exsanguination [26], suggesting that levels of pFVIII equivalent to a plasma activity level of ∼0.01% were enough to survive in this bleeding model. We believe that following tail-clipping, increasingly anemic mice shunt blood flow away from their tails. The small amounts of FVIII released from slowly flowing or even stationary platelets would then be effective to prevent death. When inhibitors are being studied, the apparent efficacy of pFVIII would be further biased by the absence of blood flowing bringing in additional anti-FVIII antibodies as well as washing out locally released pFVIII.

Thus, we believe that the events occurring in this low to no flow model may not closely reflect physiology at sites of bleeding in patients with hemophilia A. In our study, we focused instead on the FeCl3 carotid artery injury model in which the activated platelets were exposed to significant flow and in which we have demonstrated a clear dose response both for infused human FVIII and for the level of FVIII within the platelets needed to correct the clotting defect in pFVIII/FVIIInull mice [3]. We compared the effect of infused inhibitors into pFVIII/FVIIInull mice and into FVIIInull mice that had received an i.v. infusion of a slightly higher amount of FVIII replacement than the plasma equivalency of the pFVIII. As the study was completed in ∼30 min, issues related to infused FVIII half-life were of lesser consequence. Using this model, we found that pFVIII is ∼ seventyfold protective against the inhibitor mix relative to a similar plasma correction or has an ∼six-fold increased efficacy in BU mL−1 protection.

Accumulation of anti-FVIII inhibitor antibodies inside platelets did not significantly decrease the efficacy of the pFVIII. This last point may be as a result of three possibilities: (i) Insufficient amounts of inhibitors are accumulating in the α-granules relative to the molar amount of pFVIII. If one calculates the concentration of FVIII within the α-granules of the pFVIII/FVIIInull mice, it is very high even although when released from activated platelets it amounts to only a 9% antigenic correction in whole blood [3]. (ii) Inhibitors are accumulating in a subpool of α-granules distinct from that containing pFVIII as recently described for other proteins [27]. This segregation may be especially true for pFVIII that is synthesized within the megakaryocyte and inhibitors that are endocytosed. (iii) The environment within the α-granule prevents inhibitor activity. ESH8 is unable to inhibit FVIII in the absence of VWF, and we have previously shown that ∼70% of the stored FVIII within platelets is independent of VWF [4]. Perhaps pFVIII and VWF are stored in separate α-granules or pFVIII does not bind VWF within common granules so that the released pFVIII is not VWF bound?

The FeCl3 carotid artery model with the infused mixture of inhibitors studied in this paper has its limitations as well. In our hands, we were unable to use a single monoclonal antibody to inhibit thrombus development in pFVIII/FVIIInull mice, but required the described combination of inhibitors. The basis for this need for combinatorial inhibitors is not clear. Whether two compartments of pFVIII need to be inhibited, one VWF-dependent and one independent, is unclear. In addition, this model involves an arterial injury with no venous component. Coagulation factors are more critical in venous clots, while platelets are more critical in arterial clots [28]. In particular, the FeCl3 carotid artery injury model involves endothelial denudation, collagen exposure and adhesion of platelets [29]. However, this feature of thrombus development in the FeCl3 injury model should mean that pFVIII should be particularly effective in improving thrombosis relative to FVIII infusions and exaggerate observed pFVIII efficacy. Moreover, the carotid injury model is analyzed only over the first 30 min, while the bleeding diathesis associated with FVIII deficiency is often delayed or recurrent [30], more similar to the events seen with the tail-clip exsanguination model or even longer. Issues related to long-term stability of the resulting clot may not be addressed by our model.

Certainly developing a strategy of gene therapy with improved protection against inhibitors or preventing the development of inhibitors is of particular interest in the care of hemophilia A. Our studies concur with prior studies that there is a clear protective advantage of pFVIII over plasma infusions, but we propose that this protection is not of the degree previously suggested and may be of limited benefit in the treatment of hemophilia A patients with high-titer inhibitors. Future studies in larger animal models, which undergo injuries more consistent with that seen in hemophilia A patients, are needed to resolve this issue of efficacy in the face of inhibitors as well as to determine additional benefits and risks of delivery of FVIII via platelets.

Authors’ contributions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

J. Gewirtz was the primary individual to perform and evaluate the presented studies, and to help with manuscript preparation. L. Rauova assisted in the chronic inhibitor studies and in the evaluation of plasma and platelet immunoglobulin levels. She assisted in reviewing the manuscript. M.A. Thornton began these studies and taught J. Gewirtz the necessary technologies to carry out these studies. M. Poncz directed the proposed research, assisted in data analysis and in the preparation of the manuscript.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The authors thank W.R. Church of Green Mountain Antibodies for providing the Green Mountain Human FVIII monoclonal antibody, for a control murine IgG from G. Arepally at Duke University and for human FVIII from Baxter Healthcare. We thank S. Krishnaswamy for critical review of the manuscript. Support for these studies was provided by NHLBI PO1 HL64190.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Results
  6. Discussion
  7. Authors’ contributions
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  • 1
    Castaldo G, D’Argenio V, Nardiello P, Zarrilli F, Sanna V, Rocino A, Coppola A, Di Minno G, Salvatore F. Haemophilia A: molecular insights. Clin Chem Lab Med 2007; 45: 45061.
  • 2
    Ponder KP. Gene therapy for hemophilia. Curr Opin Hematol 2006; 13: 3017.
  • 3
    Yarovoi H, Kufrin D, Eslin DE, Zhu H, Wilcox DA, Camire R, Fakharzadeh SS, Kowalska MA, Montgomery RR, Poncz M. Factor VIII ectopically expressed in platelets. Efficacy in hemophilia A treatment. Blood 2003; 102: 400613.
  • 4
    Yarovoi H, Nurden AT, Montgomery RR, Nurden P, Poncz M. Intracellular interaction of Von Willebrand Factor and Factor VIII depends on cellular context: lessons from platelet expressed Factor VIII. Blood 2005; 105: 46746.
  • 5
    Bi L, Lawler AM, Antonarakis SE, High KA, Gearhart JD, Kazazian HH Jr. Targeted disruption of the mouse factor VIII gene produces a model of haemophilia A. Nat Genet 1995; 10: 11921.
  • 6
    Shi Q, Wilcox DA, Fahs SA, Weiler H, Wells CW, Cooley BC, Desai D, Morateck PA, Gorski J, Montgomery RR. Factor VIII ectopically targeted to platelets is therapeutic in hemophilia A with high-titer inhibitory antibodies. J Clin Invest 2006; 116: 197482.
  • 7
    High KA. The leak stops here: platelets as delivery vehicles for coagulation factors. J Clin Invest 2006; 116: 18402.
  • 8
    Andrews N. Putting platelets to work to circumvent hemophilia inhibitors. Hematologist 2006; 3: 8.
  • 9
    Dasgupta S, Navarrete AM, Delignat S, Wootla B, Andre S, Nagaraja V, Lacroix-Desmazes S, Kaveri SV. Immune response against therapeutic factor VIII in hemophilia A patients – a survey of probable risk factors. Immunol Lett 2007; 110: 238.
  • 10
    Mathew P. Current opinion on inhibitor treatment options. Semin Hematol 2006; 43: S813.
  • 11
    Fay WP, Parker AC, Ansari MN, Zheng X, Ginsburg D. Vitronectin inhibits the thrombotic response to arterial injury in mice. Blood 1999; 93: 182530.
  • 12
    Pittman DD, Alderman EM, Tomkinson KN, Wang JH, Giles AR, Kaufman RJ. Biochemical, immunological, and in vivo functional characterization of B-domain-deleted factor VIII. Blood 1993; 81: 292535.
  • 13
    Hashimoto Y, Ware J. Identification of essential GATA and Ets binding motifs within the promoter of the platelet glycoprotein Ib alpha gene. J Biol Chem 1995; 270: 245329.
  • 14
    Fujita H, Hashimoto Y, Russell S, Zieger B, Ware J. In vivo expression of murine platelet glycoprotein Ibalpha. Blood 1998; 92: 48895.
  • 15
    Scandella D, Gilbert GE, Shima M, Nakai H, Eagleson C, Felch M, Prescott R, Rajalakshmi KJ, Hoyer LW, Saenko E. Some factor VIII inhibitor antibodies recognize a common epitope corresponding to C2 domain amino acids 2248 through 2312, which overlap a phospholipid-binding site. Blood 1995; 86: 18119.
  • 16
    Villard S, Piquer D, Raut S, Leonetti JP, Saint-Remy JM, Granier C. Low molecular weight peptides restore the procoagulant activity of factor VIII in the presence of the potent inhibitor antibody ESH8. J Biol Chem 2002; 277: 272329.
  • 17
    Verbruggen B, Novakova I, Wessels H, Boezeman J, Van Den Berg M, Mauser-Bunschoten E. The Nijmegen modification of the Bethesda assay for factor VIII:C inhibitors: improved specificity and reliability. Thromb Haemost 1995; 73: 24751.
  • 18
    Saenko EL, Shima M, Gilbert GE, Scandella D. Slowed release of thrombin-cleaved factor VIII from von Willebrand factor by a monoclonal and a human antibody is a novel mechanism for factor VIII inhibition. J Biol Chem 1996; 271: 2742431.
  • 19
    Saenko EL, Shima M, Rajalakshmi KJ, Scandella D. A role for the C2 domain of factor VIII in binding to von Willebrand factor. J Biol Chem 1994; 269: 116015.
  • 20
    Hsu TC, Pratt KP, Thompson AR. The factor VIII C1 domain contributes to platelet binding. Blood 2008; 111: 2008.
  • 21
    Handagama P, Rappolee DA, Werb Z, Levin J, Bainton DF. Platelet alpha-granule fibrinogen, albumin, and immunoglobulin G are not synthesized by rat and mouse megakaryocytes. J Clin Invest 1990; 86: 13648.
  • 22
    George JN, Saucerman S, Levine SP, Knieriem LK, Bainton DF. Immunoglobulin G is a platelet alpha granule-secreted protein. J Clin Invest 1985; 76: 20205.
  • 23
    Millward TA, Heitzmann M, Bill K, Langle U, Schumacher P, Forrer K. Effect of constant and variable domain glycosylation on pharmacokinetics of therapeutic antibodies in mice. Biologicals 2008; 36: 417.
  • 24
    Kufrin D, Eslin DE, Bdeir K, Murciano J-C, Kuo A, Kowalska MA, Degen JL, Sachais BS, Cines DB, Poncz M. Anti-thrombotic thrombocytes. Ectopic expression of urokinase-type plasminogen activator in platelets. Blood 2003; 102: 92633.
  • 25
    Mordenti J, Osaka G, Garcia K, Thomsen K, Licko V, Meng G. Pharmacokinetics and interspecies scaling of recombinant human factor VIII. Toxicol Appl Pharmacol 1996; 136: 758.
  • 26
    Shi Q, Fahs SA, Wilcox DA, Kuether EL, Weiler H, Montgomery RR. In the presence of pre-existing factor VIII (FVIII) immunity, hematopoietic stem cells (HSC) that are genetically modified to express FVIII in platelets were successfully transplanted into hemophilic mice under myeloablative and various non-myeloablative conditions. Blood 2007; 110: 236a.
  • 27
    Italiano J, Richardson JL, Folkman J, Klement G. Blood platelets organize pro- and anti-angiogenic factors into separate, distinct alpha granules: implications for the regulation of angiogenesis. Blood 2006; 108: 120a.
  • 28
    Olson PS, Ljungqvist U, Bergentz SE. Analysis of platelet, red cell and fibrin content in experimental arterial and venous thrombi. Thromb Res 1974; 5: 119.
  • 29
    Furie BC, Furie B. Tissue factor pathway vs. collagen pathway for in vivo platelet activation. Blood Cells Mol Dis 2006; 36: 1358.
  • 30
    Spira J, Plyushch OP, Andreeva TA, Andreev Y. Prolonged bleeding-free period following prophylactic infusion of recombinant factor VIII reconstituted with pegylated liposomes. Blood 2006; 108: 366873.