The first human tumor derived protein with in vivo angiogenic activity to be obtained in pure form has been isolated from serum-free supernatants of an established human adenocarcinoma cell line (HT-29) and named angiogenin. It was purified by cation-exchange and reversed-phase high-performance liquid chromatography; the yield was approximately 0.5 microgram/L of medium. Biological activity of angiogenin was monitored throughout purification by using the chick embryo chorioallantoic membrane assay. Statistical evaluation demonstrates that it displays activity in this system with as little as 35 fmol per egg. Moreover, only 3.5 pmol is required to induce extensive blood vessel growth in the rabbit cornea. The amino acid composition of this basic (isoelectric point greater than 9.5), single-chain protein of molecular weight approximately 14 400 has been determined. The amino terminus is blocked, and the carboxyl-terminal residue is proline.

Coagulation factor VIII (FVIII) is a cofactor in the intrinsic pathway of blood coagulation for which deficiency results in the bleeding disorder hemophilia A. FVIII contains a domain structure of A1-A2-B-A3- C1-C2 of which the B domain is dispensable for procoagulant activity in vitro. In this report, we compare the properties of B-domain-deleted FVIII (residues 760 through 1639, designated LA-VIII) to wildtype recombinant FVIII. In transfected Chinese hamster ovary (CHO) cells, LA- VIII was expressed at a 10- to 20-fold greater level compared with wildtype FVIII. The specific activity of purified LA-VIII was indistinguishable from wild-type recombinant FVIII and both exhibited similar thrombin activation coefficients. Wildtype recombinant-derived FVIII and LA-VIII also displayed similar timecourses of thrombin activation and heavy chain cleavage. However, compared with wildtype recombinant-derived FVIII, the light chain of LA-VIII was cleaved fivefold more rapidly by thrombin. Addition of purified von Willebrand factor (vWF) did not alter the kinetics of thrombin cleavage or activation of either wildtype recombinant-derived FVIII or LA-VIII. The immunogenicity of LA-VIII was compared with wildtype FVIII in a novel model of neonatal tolerance induction in mice. The results did not detect any immunologic differences between wildtype FVIII and LA-VIII, suggesting that LA-VIII does not contain significant new epitopes that are absent in wildtype FVIII. LA-VIII was tolerated well on infusion into FVIII-deficient dogs and was able to correct the cuticle bleeding time similar to wildtype recombinant factor VIII. In vivo, LA-VIII was bound to canine vWF and exhibited a half-life similar to wildtype recombinant FVIII. These studies support that B-domain-deleted FVIII may be efficacious in treatment of hemophilia A in humans.

The angiogenic capacity of the class 1 heparin-binding growth factor from bovine brain, an anionic endothelial cell mitogen of Mr 16 000, has been evaluated. Its ability to induce the growth of new blood vessels has been assessed by means of two established assay systems. On the embryonic chick chorioallantoic membrane dose-response studies demonstrate that 160 ng (10 pmol) of mitogen is required to induce angiogenesis in greater than 50% of the eggs within 72 h. In the presence of 1 unit of exogenous heparin only 40 ng of mitogen (2.5 pmol) is needed to induce a similar response. Moreover, this occurs within 48 h, indicating that heparin also augments the angiogenic response by enhancing the rate of induction of angiogenesis. Eighty nanograms (5 pmol) of mitogen also induces the ingrowth of new blood vessels into the rabbit cornea, both in the presence and in the absence of heparin. These results establish that the class 1 heparin-binding growth factor from bovine brain is an angiogenesis factor. Importantly, the neovascularization induced by this angiogenesis factor is enhanced by heparin. The mechanistic implications for neovascularization under certain normal and pathological conditions are discussed.

Background. Bone morphogenetic proteins (BMPs) are potent inducers of bone formation. Functional and immunohistochemical studies have identified BMPs in a subset of osteosarcomas. In the present study, the authors extend the analysis of BMP expression to other bone and soft tissue sarcomas. Methods. Monoclonal antibody AbH3b2/17 against human BMP‐2 and BMP‐4 was used in avidin‐biotin‐immunoperoxidase assays with frozen sections of bone tumors (71 specimens), soft tissue sarcomas (69 specimens), and normal tissues. Results. Among bone tumors, BMP was detected in osteosarcomas (17 of 29 samples), malignant fibrous histiocytomas (MFHs) (6 of 6), and the spindle cell sarcomatous components of spindle cell (dedifferentiated) chondrosarcomas (4 of 4), but not in conventional chondrosarcomas (0 of 10) or Ewing's sarcomas (0 of 14). Histologic subtypes of osteosarcoma differed for BMP expression, with 8 of 9 fibrohistiocytic, 9 of 13 osteoblastic, and 0 of 5 chondroblastic lesions showing immunostaining. In all BMP‐positive bone tumors, immunostaining was localized in the cytoplasm of primitive mesenchymal cells, with little or no staining in tumor matrix and more mature osteoblastic/chondrocytic cells. Among soft tissue sarcomas, MFHs (11 of 12), liposarcomas (5 of 11), leiomyosarcomas (3 of 9), and malignant schwannomas (3 of 8) showed cytoplasmic BMP immunostaining. Synovial sarcomas (0 of 9), rhabdomyosarcomas (0 of 8), and fibrosarcomas (0 of 7) were BMP‐negative. All normal humantissues tested, including the tissues of a 16‐week‐old fetus, lacked BMP immunoreactivity. Conclusions. Bone morphogenetic protein is expressed in a subset of osteosarcomas, a high proportion of MFHs of bone and soft tissue, and in spindle cell chondrosarcomas. In these tumors, BMP is localized predominantly to the cytoplasm of malignant cells with primitive mesenchymal features; no or little BMP is detected in the more differentiated elements of bone and soft tissue sarcomas, Different histologic types of bone and soft tissue sarcomas may mimic discrete stages of mesenchymal differentiation as defined by BMP expression and histologic criteria.

Background. Bone morphogenetic proteins (BMP) induce ectopic bone formation in vivo and may play a role in normal bone development. In addition, bone morphogenetic activity, as measured in a bone‐forming assay in immunodeficient, athymic nu/nu mice, is present in a proportion of osteosarcomas; this activity, which may be mediated by BMP, is correlated with a poor prognosis. Methods. The development of a monoclonal antibody against recombinant human BMP‐2, AbH3b2/17, has allowed immunohistochemical localization of BMP in tumor tissues. Cryostat sections of osteosarcomas (21 tumor samples), chondrosarcomas (5 samples), and Ewing's sarcomas of bone (5 samples) were examined with AbH3b2/17 using the avidin‐biotin‐immunoperoxidase method. Results. The authors found AbH3b2/17 immunoreactivity in 12 of the 21 osteosarcoma samples (57% sensitivity) obtained from 20 patients. For one patient, samples of the primary lesion and a subsequent metastasis were tested, and only the latter showed AbH3b2/17 immunore‐activity. The chondrosarcomas and Ewing's sarcomas examined showed no immunoreactivity. In antigen‐positive osteosarcomas, AbH3b2/17 immunostaining was localized predominantly in the cytoplasm of tumor cells. Moreover, the proportion of AbH3b2/17‐reactive cells varied among osteosarcomas with disparate histologic features. Conclusions. The authors identified a rapid and widely applicable method for detecting BMP expression in intact tissues, which may complement and enhance the bone‐forming assay in nu/nu mice as a prognostic procedure in osteosarcomas.