Search

Human umbilical vein endothelial cells (HUVEC) propagated in co-culture with fibroblasts form capillary-like networks of tubes. Here we characterize the morphology and ultrastructure of HUVEC in such co-cultures and investigate the influence of different angiogenesis inhibitors on endothelial cell morphology. Addition of angiogenesis inhibitors to the co-culture disrupted endothelial network formation and influenced endothelial cell morphology in two distinct ways. Instead of characteristic capillary-like networks, the endothelial cell morphology appeared as either short cords or compact cell clusters of variable size. Electron microscopy (EM) showed that in co-culture untreated HUVEC formed capillary-like tubes with lumina and retained important ultrastructural and physiological properties of endothelial cells in functional vessels as they contained both Weibel-Palade bodies and transport vesicles. Immuno-EM showed that the endothelial cell marker CD 31 stained endothelial membranes at cell-cell contacts, and at the luminal and abluminal side of the capillary-like tubes, although most abundantly at the luminal membranes. No ultrastructural signs of apoptosis were seen in HUVEC in inhibitor-treated co-cultures. Our results demonstrate that treatment with levamisole or anti-VEGF inhibits endothelial cell differentiation into tubes or instead induces formation of compact endothelial cell clusters. Treatment with platelet factor 4, suramin and TNP-470 results in formation of short endothelial cell cords. We discuss the implications of these findings.

Since solid tumours and metastases depend on adequate blood supply, much research is focused on inhibition of angiogenesis. Unfortunately, most known angiogenesis inhibitors have serious side effects when used as therapeutic agents in man. It is therefore important to develop methods to identify well-tolerated and efficient angiogenesis inhibitors. As a method for identification of new angiogenesis inhibitors we have further developed the procedure described by Bishop et al. (Angiogenesis 1999;3:335-44) to a quantitative ELISA-based fibroblast and endothelial cell co-culture angiogenesis assay. In each well of a 96-microwell plate, human umbilical vein endothelial cells (HUVEC) are seeded onto normal human dermal fibroblasts (NHDF) and propagated in co-culture for 72 h with or without a potential angiogenesis inhibitor. The effect on total cell proliferation is evaluated by quantitative immunochemical measurement of DNA, and on endothelial tube formation by quantification of CD 31, von Willebrand factor, and collagen IV. After ELISA reading, the morphology of the tubular structures formed by HUVEC is visualised with BCIP/NBT, permitting a quantitative result and a qualitative evaluation of cell morphology from the same well. We have used the assay to demonstrate the effect of well-known angiogenesis inhibitors on HUVEC tube formation.

An experiment was set up to investigate the relationship, if any, between cell surface MHC class I expression and the growth rate for skin tumors induced by two different UV radiation regimens in hairless mice. Two groups of 20 hairless mice were each irradiated with either a UVA radiation source (2 SED per session) or broad-spectrum UV radiation (UVB) (8.1 SED per session) 5 days a week during the entire experiment. In the UVA group, 17 out of 20 animals developed tumors, and 10 of these grew to a diameter of > or = 5 mm. In the UVB group, 19 out of 20 animals developed tumors, and 15 of these grew to a diameter of > or = 5 mm. The tumor induction time, i.e. the time from the start of UV treatment to tumor appearance, was found to be significantly longer (p<0.01) in the UVA than in the UVB group. This is in accordance with previous findings. Of the 25 tumors growing to a diameter of > or = 5 mm, 11 were established as cultured cell lines (4 UVA and 7 UVB tumors). These uncloned cell lines were analyzed for surface expression of major histocompatibility complex class I by FACS analysis. There was a clear correlation between high MHC class I expression and slow growth of the individual tumors (p<0.05). This suggests a role for the MHC class I governed, i.e. cytotoxic T-cell-mediated, reactions in deciding the fate of UV-induced skin cancers. No correlation was found between MHC class I expression and tumor induction time.

In order to study the role of the T-cell-mediated immune defense in tumor development, a total of 93 sarcomas were induced using different doses (8 micrograms (0.1%), 40 micrograms (0.5%) and 400 micrograms (5%)) of 3-methylcholanthrene in athymic nude Balb/c mice and phenotypically normal immunocompetent Balb/c mice. A shorter tumor induction time and a higher tumor incidence after treatment with low doses of methylcholanthrene were seen in nude mice than in immunocompetent mice, indicating that they have a lower resistance to the carcinogen. Contrary to expectations we found that the MHC class I expression of tumors from nude mice was lower than that of tumors from normal mice. Higher surface expression of MHC class I was demonstrated on high dose tumors from normal mice than on low dose tumors from normal mice. The cellular composition of the individual tumors raised in nude mice was more heterogeneous with respect to MHC class I expression. Since the mice differ genetically only with respect to the nu gene, these results indicate that a lack of T-cell-mediated defense mechanisms may confer upon the bearer a lower resistance to 3-methylcholanthrene and a different MHC profile of the ensuing tumor.

Many human tumors express low amounts of HLA class I molecules relative to the normal cells from which they are derived. From experimental work it is clear that the malignant behavior of a tumor cell may depend on its MHC class I expression. Therefore, it is of obvious interest to study the HLA class I expression of human tumors in their various stages. We have studied the HLA class I expression by the cells in premalignant epithelial lesions and invasive carcinoma of the bladder and uterine cervix using immunoperoxidase staining for beta 2-microglobulin of paraffin-embedded tissue. We here assume that beta 2-microglobulin expression by malignant and premalignant cells equals HLA class I expression. Thirty-two of the 36 invasive tumors expressed less overall beta 2-microglobulin than cells from the normal epithelium. In contrast, approximately two-thirds of 34 premalignant bladder epithelia and 47 premalignant cervix epithelia displayed higher overall beta 2-microglobulin expression than the normal epithelium. Thus, a systematic large-scale elimination of HLA class I high-expressing tumor cell variants may take place only after the tumor penetrates the basement membrane.

Tumors were induced in athymic, T-cell-deficient nude mice and in syngeneic normal haired mice by treatment with low doses of 3-methylcholantrene (MCA). The tumors were studied for tumor cell expression of MHC class I molecules and for immunogenicity by transplantation to syngeneic haired recipients. Ten tumors were obtained by the MCA treatment, six from nude and four from haired mice. They were all fibrosarcomas as judged from their microscopic appearance. Five of the "nude" tumors expressed measurable amounts of MHC class I molecules and two of them expressed high amounts. Both were immunogenic in the sense that they evoked a cytotoxic T-cell response in transplanted haired recipients. Only one of the four "haired" tumors expressed measurable amounts of MHC class I, and none of them were immunogenic. These findings support the concept that some tumors are immunoselected at an early point of time in their existence in a host with a normal immune system and that this results in an elimination of tumor cell variants which are highly immunogenic for the T-cell system, leaving the low or non-immunogenic variants. These take over and grow and kill their host. The results suggest that tumor cell variants expressing high amounts of MHC class I are important targets in the immunoselection in hosts with a normal immune system.