OBJECTIVE: Hemangioma is an endothelial cell tumor that grows rapidly during infancy and regresses slowly during childhood. However, little is known about the natural history of this common tumor. To gain insight into the cellular mechanisms that underlie the switch from uncontrolled growth to involution of endothelium, we investigated the extent of cellular apoptosis versus proliferation in hemangioma specimens that spanned the natural life cycle of the tumor. METHODS: We analyzed apoptosis and cellular proliferation in frozen sections from 16 hemangioma specimens using the TUNEL assay to detect apoptotic cells and the Ki67 antigen to detect dividing cells. RESULTS: Apoptosis was low in proliferative phase hemangiomas but increased fivefold in involutive phase specimens obtained from children one to four years of age. Immunofluorescence double-labeling experiments showed that at least one third of the apoptotic cells were endothelial. As expected, cellular proliferation was high in specimens up to 2 years of age but decreased significantly thereafter. Apoptosis was consistently low in nine normal skin tissues (newborn to 4 years of age) obtained from discarded pathology specimens. CONCLUSIONS: These results suggest that increased apoptosis during the second year of life can offset cellular proliferation and may be involved in initiating regression of hemangioma.

Human dermal microvascular endothelial cells are used to analyze the functions of microvascular endothelium in vitro. However, the low yield and short lifespan of these cells in culture has limited the types of analysis that could be performed. Human microvascular endothelial cells are typically grown in media containing supplements such as dibutyryl cyclic AMP, hydrocortisone, bovine brain extract, and antifungal agents, each of which increase the complexity of experimental design and interpretation of results. In the present study, endothelial cells were transferred after Ulex europeus-I selection into a simplified medium consisting of Endothelial Basal Medium 131, 10% fetal bovine serum, and 2 ng/ml basic fibroblast growth factor and analyzed over 3 mo. The human microvascular endothelial cells expressed the endothelial markers von Willebrand factor, CD31, P-selectin, and E-selectin. In addition, the cells showed increased proliferation in the presence of basic fibroblast growth factor (0.5 ng/ml) or vascular endothelial cell growth factor (10 ng/ml). Tumor necrosis factor-alpha-induced expression of E-selectin was similar in cells at Passages 3, 6, and 12, indicating that the cells maintained responsiveness to cytokines over several weeks. Furthermore, the endothelial cells attained a typical cobblestone morphology with increased cellular density and also formed capillarylike tubes on Matrigel. In summary, the human dermal microvascular endothelial cells display the expected endothelial characteristics when grown for several passages and, therefore, provide a valuable in vitro model for human microvascular endothelium.

Angiostatin, a 38 kilodalton fragment of plasminogen, is a potent inhibitor of angiogenesis. However, little is known about how angiostatin affects endothelial gene expression. To learn more about its effect on endothelial-specific genes implicated in angiogenesis, we examined E-selectin expression and function in bovine capillary endothelial cells treated with recombinant angiostatin. Angiostatin caused a four to five-fold increase in E-selection polypeptide levels in proliferating endothelial cells but little or no increase in confluent cells. P-selection polypeptide levels were unaffected by angiostatin in either proliferating or confluent cells. E-selectin mRNA and adhesion activity in proliferating endothelial cells were also increased by angiostatin. Angiostatin had little effect on the distribution of endothelial cells in G0/G1, S, and G2/M, indicating angiostatin does not alter cell cycle progression significantly. These data demonstrate that angiostatin selectively upregulates E-selectin in proliferating endothelial cells in vitro. This selectivity may provide insights into the mechanism by which angiostatin inhibits tumor growth in vivo without apparent effects on quiescent endothelium.