The accumulation of blood leukocytes at sites of inflammation depends upon their localized adhesion to the vascular lining. We have investigated the hypothesis that this adhesive interaction involves inducible endothelial cell-surface structures that can bind leukocytes. Certain infammatory! immune cytokines, namely interleukin 1, tumor necrosis factor, and lymphotoxin, as well as bacterial endotoxin, act on cultured human endothelial cells (HEC) in a time-and proteinsynthesis-dependent fashion to increase leukocyte adhesion. (unpublished observations). In all cases, fusions of splenocytes to the NS-1 myeloma cell line and all subsequent procedures were done as described (15).

No abstract

Human recombinant tumor necrosis factor (rTNF) was found to act directly on cultured human vascular endothelium to induce a tissue factor-like procoagulant activity (PCA). After a 4-hr incubation in rTNF (100 units/ml), serially passaged endothelial cells isolated from umbilical veins, saphenous veins, iliac arteries, and thoracic aortae demonstrated a dramatic increase (4-to 15-fold, 21 experiments) in total cellular PCA as measured with a one-stage clotting assay. rTNF-induced PCA was also expressed at the surface of intact viable endothelial monolayers. Induction of PCA by rTNF was concentration dependent (maximum, 500 units/ml), time dependent, reversible, and blocked by cycloheximide and actinomycin D, and it occurred without detectable endothelial cell damage. Actions of rTNF were compared with those of natural human interleukin 1 (IL-1) derived from stimulated monocytes and two distinct species of recombinant IL-1, each of which also induced endothelial PCA. The use of recombinant polypeptides and specific neutralizing antisera established the distinct natures of the mediators. The minetics ofthe endothelial PCA responses to TNF and IL-1 werf similar, demonstrating a rapid rise to peak activity at -4 hr, apd a decline toward basal levels by 24 hr. This characteristic decline in PCA after prolonged incubation with TNF or IL-1 was accompanied by selective endothelial hyporesponsiveness to the initially stimulating monokine. Interestingly, the effects of TNF and IL-1 were found to be additive even at apparent maximal doses of the individual monokines. Endothelial-directed actions of TNF, alone or in combination with other monokines, may be important in the initiation ofcoagulation and inflammatory responses in vivo.

Dormant solid tumors were produced in vivo by prevention of neovascularization. When small fragments of anaplastic Brown-Pearce carcinoma were implanted directly on the iris in susceptible rabbits, they always vascularized. A characteristic growth pattern, consisting of prevascular, vascular, and late phases, was observed, which terminated with destruction of the eye within 2 wk. The beginning of exponential volume increase was shown to coincide with vascularization of the implant, as demonstrated by perfusion with intravenous fluorescein and by histologic sections. In contrast, implants placed in the anterior chamber, at a distance from the iris, did not become vascularized. After initial growth into spheroids, they remained arrested at a small size comparable to prevascular iris implants, for periods as long as 6 wk. Although dormant in terms of expansion, these avascular tumors contained a population of viable and mitotically active tumor cells. When reimplanted on the iris, vascularization was followed by rapid, invasive growth. These observations suggest that neovascularization is a necessary condition for malignant growth of a solid tumor. When a small mass of tumor cells is prevented from eliciting new vessel ingrowth from surrounding host tissues, population dormancy results. These data suggest that the specific blockade of tumor-induced angiogenesis may be an effective means of controlling neoplastic growth.