Prophylactic efficacy of BeneFIX vs Alprolix in hemophilia B mice

Abstract

FIX binds tightly to collagen IV. Furthermore, a FIX mutant, FIXK5R, which binds better than wild-type FIX to collagen IV, provides better hemostasis than wild-type FIX, long after both are undetectable in the plasma. There is also credible evidence of extravascular FIX. Here, we use the saphenous vein bleeding model to compare the efficacy of recombinant FIXFc (Alprolix) and wild-type FIX (BeneFIX) in hemophilia B mice 7 days postinfusion. Although the terminal half-life of Alprolix is significantly longer than that of BeneFIX, at equal doses Alprolix is not better at controlling bleeding 7 days postinfusion, presumably because of the extravascular FIX. Both BeneFIX and Alprolix exhibit a linear response in clotting efficacy up to 150 IU/kg, where they appear to saturate an extravascular compartment, because there is no additional prophylactic benefit from higher doses. A robust pool of extravascular FIX is clearly observed surrounding blood vessels, localized to the same region as collagen IV, in 2 representative human tissues: liver and skeletal muscle. We see no increased risk for thrombosis at 250 IU/kg FIX at 6 hours postinfusion. In summary, 7 days postinfusion into hemophilia B mice, BeneFIX and Alprolix are hemostatically indistinguishable despite the latter's increased half-life. We predict that doses of FIX ∼3 times higher than the currently recommended 40 to 50 IU/kg will, because of FIX's large extravascular compartment, efficiently prolong prophylactic hemostasis without thrombotic risk.

Figure 1

Comparison of BeneFIX and Alprolix at different doses in Hemophilia B mice. Each point in (A) represents the number of times that clotting can recur in a hemophilic mouse with a saphenous vein injury 7 days postinfusion with either Alprolix or BeneFIX. The P values shown within the plot are from Mann-Whitney tests. (B) Plot of median values for the number of clots formed at different doses of BeneFIX or Alprolix. The medians were taken from the data shown in (A) together with additional similar data from higher doses. The line connecting the points has no theoretical meaning but serves to draw attention to the proportional increase. (C) Mice expressing FIX_K5A at 120% of WT levels are about 50% as active in the saphenous vein bleeding model as are WT mice. The in vitro activity of FIX_K5A is indistinguishable from FIX_WT in an activated partial thromboplastin time assay. HemB, hemophilia B.

Figure 2

Hemophilia B mice, that had been infused with 150 IU/kg BeneFIX 7 days previously, were subjected to the saphenous vein injury. After the vein ceased clotting, the contralateral saphenous vein was transected. The data show that there is no significant difference between the 2 veins in the number of clots formed, despite undetectable levels of circulating FIX at the time of the injury of the first vein. This indicates that it is not residual circulating FIX that is responsible for clotting, but instead extravascular FIX.

Figure 3

Histology of extravascular FIX distribution. Antibodies to FIX (A,D), Collagen IV (B), FX (C), and FVII (F) were used to stain human liver sections with 3,3′-diaminobenzidine. (E) No primary antibody; all no primary antibody controls lacked background noise. Notice that collagen IV in the sinusoids (B) is not stained with FIX, but that collagen IV and FIX staining generally coincide. The tissues stained with antibodies to FX and FVII differ noticeably from those stained for FIX. (G) A cluster of arteries sectioned in the human skeletal muscle stained for FIX. (H) An adjacent section stained for collagen IV. (I) No primary antibody control. Staining was visualized using a Vectastain ABC Elite Kit (Vector Labs PK-6100), followed by 3,3′-diaminobenzidine. The images were captured with a Nikon Optiphot-2 with a plan-apo lenses and an Olympus DP-70 camera.

Figure 4

Venous thrombosis model at 250 IU/kg. The intravital thrombosis assay: an electrolytic injury was delivered to the surface of a mouse femoral vein for 30 seconds (1.5 V anodal current) after the injection of fluorophore labels for fibrin (A) and platelets (B) captured with time-lapse fluorescence imaging from 1 to 60 minutes later. Data are quantitated and normalized every video frame showing comparative levels for WT (blue), FIX knockout mice without treatment (red), or FIX knockout mice with 250 IU/kg BeneFIX injected 6 hours prior to thrombus induction (green lines), using a group of 5 to 6 mice per genotype and treatment. Data are means; errors bars are standard error of the mean. There are no statistical differences between the WT and HemB mice with BeneFIX treatment, whereas the untreated HemB mice had lower levels of accumulation for both thrombotic parameters (P < .001 at 10 or more minutes; analysis of variance).