Recombinant factor VIIa (rFVIIa) is used widely as an effective hemostatic agent for treatment of severe hemophilia patients with inhibitors against FVIII or FIX. Recent reports have shown that rFVIIa, when administered in prophylaxis, can prevent the development of mild-moderate joint bleedings [1–7]. Given the short biological half-life of rFVIIa, it is unclear by what mechanism rFVIIa could be effective in prophylactic treatment. A recent study characterizing the bio-distribution of pharmacologically administered rFVIIa by immunohistochemistry revealed that rFVIIa (i) readily associates with the endothelium, (ii) enters into extravascular tissues, where it can bind to TF, and (iii) remains there for longer time periods than in the circulation . Although these immunohistochemistry data provide convincing evidence that rFVIIa from the blood stream enters into extravascular compartments, they provide no quantification of the amount of rFVIIa transferred from the blood stream to extravascular tissues, and no information regarding the functional status of rFVIIa associated with these tissues. Moreover, the recent studies of Hoffman et al.  indicated that perivascular TF is occupied by endogenous FVII/FVIIa in the absence of injury. In such a scenario, the exogenously administered rFVIIa may be simply supplanting the endogenous FVII/FVIIa bound to peri/extravascular tissues without increasing the net levels of FVIIa in tissues. Therefore, we thought it was important to evaluate FVIIa levels in tissues following its administration in order to assess the extent of rFVIIa transfer from the blood stream into tissues, and the functional status of rFVIIa associated with extravascular tissues.
First, to determine whether TF in perivascular tissue is saturated with endogenous FVII/FVIIa, we have examined the presence of mouse FVII in adventitia of blood vessels from mice receiving saline by immunohistochemistry using polyclonal antibodies specific to mouse FVII/FVIIa. The antibodies were raised in rabbits using purified recombinant mouse rFVIIa as the antigen. The antibodies detect both mouse FVII and FVIIa. Out of 25 or more tissue sections of skin, liver and heart that we have examined, positive immunostaining of FVII was found only once around a single blood vessel in skin (data not shown). None of the other blood vessels or tissues stained positive for FVII/FVIIa. It is unclear why our immunohistochemistry data differ from those of Hoffman et al. . It is possible that different mouse FVIIa antibodies used in these studies could have contributed to this difference. Although our present data suggest that it is unlikely that the peri or extravascular TF is saturated with endogenous FVII, we cannot rule out the possibility that mouse FVIIa antibodies used in this study may not have sufficient sensitivity to detect traces of FVII/FVIIa in tissues. Therefore, in additional studies we have examined whether exogenously added FVIIa binds to tissue sections. Skin and other tissue sections were incubated with mFVIIa (10 nm), CaCl2 (5 mm) or mFVIIa + CaCl2 for 1 h, washed and then stained with anti-mFVIIa. Only tissue sections incubated with mFVIIa in the presence of calcium ions stained positively, whereas other tissue sections stained negatively for FVIIa (data not shown). These data indicate that exogenously added FVIIa is capable of binding to extravascular TF, suggesting that TF sites are not saturated with endogenous FVII.
To determine the extent of rFVIIa transported to tissues following its administration into the blood stream, mice were injected intravenously with saline or mFVIIa (120 μg kg−1 body weight, n = 3 to 4) via the tail vein. Thirty minutes, 6 and 24 h post-administration of rFVIIa (and 30 min post-administration of saline), mice were exsanguinated by flushing 10 mL of ice-cold saline (supplemented with CaCl2, 5 mm) through the heart and draining the blood by severing the renal artery. Various tissues (e.g. lung, liver, kidney, brain, spleen, heart and skin) were collected and stored at −80°C until homogenization. The tissues were weighed and homogenized in ice-cold TBS (0.01 m Tris–HCl, 0.15 m NaCl, pH 7.4, 0.5 mL per 100 mg tissue). Initially, we measured FVII/FVIIa activity in total tissue homogenates as well as supernatants of tissue homogenates after the addition of EDTA (20 mm) in a FX activation assay using saturating concentrations of relipidated TF (100 ng mL−1). We obtained similar levels of FVII/FVIIa activity in both tissue homogenates and the supernatants. FVII/FVIIa activity levels were higher in tissues derived from mice receiving rFVIIa compared with mice receiving saline. These data suggest that rFVIIa administered to mice entered into tissues and remained functionally active. Based on this information, we have used the EDTA supernatants for all of our subsequent assays. We have chosen the supernatant over the tissue homogenates for measuring FVIIa levels because varying levels of TF present in different tissue extracts may introduce error in determining FVIIa levels accurately. More importantly, the presence of cell-derived full-length TF does not permit measurement of FVIIa activity in a FVIIa-specific clotting assay using soluble TF . Furthermore, it is not feasible to measure FVIIa antigen accurately in tissue homogenates if it is bound to TF or other cell-associated proteins.
Measurement of FVIIa activity in tissues from mice administered saline, using FVIIa-specific clotting assay , showed that all tissues contained traces, but measurable amounts, of FVIIa, varying from 0.1 to 1 ng mg−1 tissue protein. Following the administration of rFVIIa, FVIIa levels increased in most of the tissues (Fig. 1A). The extent of increase in FVIIa in different tissues varied. FVIIa levels increased by 5 to 10-fold in the heart, spleen, kidney, liver and lung. The increase in FVIIa in tissues was highest at 30 min following the administration of rFVIIa. In some tissues, FVIIa levels remained high (similar to those that were found at 30 min) even at 24 h following rFVIIa administration, at which time rFVIIa in plasma was undetectable. No detectable increase in FVIIa was found in the brain and skin at any time following the infusion of rFVIIa. The well-recognized property of the blood-brain barrier may be responsible for curtailing the transport of FVIIa from blood to brain. In addition to measuring rFVIIa activity, we also measured FVII/FVIIa antigen in tissue extracts using ELISA specific for mouse FVII/FVIIa. In mice administered saline, FVII/FVIIa antigen levels in various tissues varied from 1 to 6 ng mg−1 protein (Fig. 1B), which is approximately 6 to 10-fold higher than FVIIa levels determined in the FVIIa-specific clotting assay. This indicates that most of the endogenous FVII associated with the tissues remains as the zymogen and not activated. Similar to that observed in FVIIa activity measurements, FVII/FVIIa antigen in tissues of mice was significantly higher following rFVIIa administration. The increase in FVII antigen was evident at 30 min and 6 h after rFVIIa administration (Fig. 1B). However, the fold-increase of FVII/FVIIa measured by estimating the antigen levels in tissues following rFVIIa administration was much lower than that estimated using the FVIIa-specific clotting activity assay. A significant increase in FVII antigen was not found in tissues harvested at 24 h following rFVIIa administration. Here, it is important to note that, unlike the FVIIa-specific clotting activity assay, the antigen assay measures both FVII and FVIIa, thus a small yet potentially significant increase in FVIIa activity following rFVIIa administration may be difficult to detect relative to total FVII/FVIIa antigen present endogenously. While it is unlikely that endogenous FVII/FVIIa detected in tissues is plasma protein remaining following perfusion of the animal as we adequately perfused the animal and no breach in vascular endothelium was evident (visualized by EPCR immunohistochemistry), we can not exclude this possibility completely.
Although from the above data it may appear that only traces of rFVIIa administered to mice enter into the extravasculature and are retained in tissues, it is important to compare the levels of FVIIa in tissues in relation to FVIIa levels in plasma following rFVIIa administration to appreciate the significance of FVIIa levels in tissues. As shown in Fig. 1(C), 30 min following rFVIIa administration, most of the injected rFVIIa in mice was in plasma and only about 4% of rFVIIa was in tissues, mostly in the liver. Six hours following rFVIIa administration, the majority of rFVIIa found in mice was in tissues. At 24 h, all of the rFVIIa detected in mice was in tissues and none in plasma. This clearly illustrates that rFVIIa distributed to the extravasculature is retained much longer than rFVIIa in circulation.
Overall, the quantitative determination of FVIIa activity and antigen in various tissues in mice following rFVIIa administration is consistent with our recent findings of bio-distribution of rFVIIa by immunohistochemistry , in that rFVIIa administered i.v. enters into extravascular tissues and FVIIa is retained in tissues for a much longer time compared with plasma levels. However, the data obtained in the present study differ with the earlier data in two aspects. First, in the earlier study , we could not detect rFVIIa in lung tissue sections by immunohistochemistry whereas in the present study we found a significant increase in both FVIIa activity and antigen levels in lung tissue following rFVIIa administration. It is possible that diffused distribution of rFVIIa throughout the lung could have escaped detection by immunohistochemistry. Second, in the earlier study we found rFVIIa accumulation in skin by immunohistochemistry, particularly in cuboidal epithelial cells of the sebaceous glands adjacent to hair follicles and in the squamous epithelial cells of the epidermis. In the present study we did not detect any significant increase in FVIIa activity and antigen in skin homogenates. However, it may be pertinent to point out here that it was difficult to homogenize skin tissue as it is elastic and consistently resulted in a very viscous and glue-like mixture. It is possible that FVIIa associated with skin is readily internalized and degraded, yet this degraded FVIIa could still be detected by AF488 antibodies in immunohistochemistry. It is also possible that differences in tissue densities could have contributed to this anomaly.
At present, it is unclear whether or how FVIIa associated with extravascular tissue contributes to the hemostatic and prophylactic effects of rFVIIa. In this context, it is interesting to point out that a recent gene therapy study in hemophilia B dogs showed that one of the hemophilia B dogs, which had no significant increase in circulating levels of FVIIa, still did not develop any spontaneous bleedings during the 34 months of observation . These findings support the hypothesis that rFVIIa deposited perivascularly prevents joint bleeding in hemophilia by stopping early bleeding in the microvasculature induced by normal moving patterns.