The cell-cell adhesion molecule E-cadherin has been shown to suppress invasive growth of epithelial cells in vitro, and loss of its expression is thought to be important in invasion and metastatic potential of epithelial tumors in vivo. We retrospectively studied the level of E-cadherin expression in 50 primary head and neck squamous-cell carcinomas (HNSCC) by immunohistochemical methods, on frozen sections, using anti-E-cadherin monoclonal antibody (MAb) 6F9. It concerned patients with different stages of carcinoma of larynx or oral cavity who had been treated with curative intention 30 months or more before. Percentages of membranous stained tumor cells were scored in 1 of 5 categories. Scores were generally low, as in 11/50 lesions < or = 5% cells were stained, and in 19/50 lesions only 6-25% cells showed membranous staining. In 9 lymph-node metastases evaluated, E-cadherin expression was in the same range as in the primary tumors. There was a significant correlation between the level of membranous E-cadherin expression in the primary tumor and the degree of differentiation. No relation was found with tumor size (pT) or regional lymph-node classification (pN). Nevertheless, 29 patients surviving> or = 30 months without evidence of disease had significantly higher levels of membranous E-cadherin expression in their primary tumors than 10 patients with unfavorable clinical course clearly related to recurrent and/or metastatic HNSCC. Moreover, this could only partially be explained by distinctions in differentiation grade between both groups. Our results suggest that membranous E-cadherin expression has prognostic importance in patients with HNSCC.

Summary We studied the relation between tumour vascular density and tumour growth rate, metastatic incidence and vascular permeability factor (VPF) mRNA levels in a human xenograft model described previously. Vascular density was determined by automated image analysis.Xenografts derived from cell lines BLM and MV3 showed the highest mean vascular density (MVD), the highest in vivo growth rate, high VPF mRNA levels and rapid development of lung metastases. Xenografts of cell lines M14, Mel57 and MV1 showed a significantly lower degree of vascularization, lower in vivo growth rates and lower levels of VPF mRNA, but formed lung metastases with a similar incidence as those of BLM and MV3. Xenografts from cell line 1 F6 did not form lung metastases, whereas tumours derived from a spontaneous mutant of 1 F6, designated 1 F6m, gave rise to lung metastases to the same extent as Mel57, M14 and MV1 tumours. MVD values in 1 F6 and 1 F6m xenografts, VPF mRNA levels and in vivo growth rates of 1 F6 and 1 F6m xenografts, however, were similar. In conclusion, in the melanoma xenograft model vascular density is correlated with in vivo growth rate and with in vitro VPF mRNA levels, but not with the ability to metastasize.

No abstract

Twenty-nine stage IIIA/B melanoma patients treated by isolated limb perfusion (ILP) with a high dose of recombinant human tumour necrosis factor alpha (rHuTNF alpha), interferon gamma (IFN gamma), and melphalan were histologically documented with emphasis on therapy-induced changes of the tumour vasculature. Sequential biopsies were taken at various intervals before and after the treatment to compare the morphological change. In order to visualize microvascular changes, immunostaining was performed for von Willebrand factor (VWF), type IV collagen, alpha-smooth muscle actin, endothelial antigen PAL-E, tissue factor, CD41 (thrombocyte marker), and fibrin. In biopsies prior to perfusion, necrosis, haemorrhage, and fibrin thrombi were not found. Within 3 h following triple combination therapy, a change in the distribution of VWF staining occurred, from a discrete endothelial pattern in the untreated lesions to a fuzzy perivascular and subepidermal pattern in the treated lesions. Within 24 h, this was accompanied by intravascular thrombocyte aggregation and erythrostasis, in the absence of tissue factor and fibrin deposits. These findings indicate that the thrombocyte aggregation observed is not caused by local procoagulant activity, but is rather the result of the therapy-associated vascular damage or haemostasis. Although it is difficult to derive the dynamics of this process from static images, we assume that TNF alpha induced endothelial cell damage, leading to VWF release. Release VWF may play a role in the adhesion between thrombocytes and the damaged endothelium or the denuded subendothelium. As a consequence, the blood flow is impaired, leading to congestion and oedema, compatible with an early stage of haemorrhagic infarction.

In uveal melanoma, macrophages accumulate at sites of EMAP-II expression. Based on the results, it may be hypothesized that this process of chemotaxis is facilitated by EMAP-II-dependent expression of ICAM-1 on vascular endothelial cells and concomitantly leads to localized vascular damage, as indicated by release of von Willebrand factor.

HLA class I and II antigen expression in routinely processed head and neck squamous cell carcinoma (HNSCC) primary lesions was evaluated. Paraffin embedded samples from 66 squamous cell carcinomas and 7 verrucous carcinomas were studied immunohistochemically using recently developed anti-HLA class I monoclonal antibodies (MAbs) HClO and HCA2, and anti-HLA-DR rabbit serum. Percent stained tumor cells were scored in I of 5 categories. The scores of 40 tumors were compared t o the staining results obtained on frozen sections of the corresponding lesions, including those of the anticlass I MAb W6/32. High percentage-matched scores for paraffin and frozen sections were obtained, with HClO vs. HCIO, HClO vs. W6/32, and anti-HLA-DR vs. anti-HLA-DR showing the best correlations. In the squamous cell carcinomas HLA class I expression was high (Le., in 49/66 lesions more than 50% cells were stained), and correlated with the degree of differentiation, and inversely with the modified Jakobsson score. HLA class II expression (more than 5% cells stained) was found in 21/66 tumors and correlated inversely with the degree of differentiation. All verrucous carcinomas exhibited very high HLA class I expression, whereas class II was locally expressed in 517 lesions. Comparison of 4 subsites of HNSCC showed that carcinomas of the oral cavity had the highest HLA class I expression. This suggests susceptibility to CD8+ T cells, and together with the well developed submucosal lymphoid tissue, makes the oral cavity carcinomas probably well suited for local immunotherapeutic approaches in HNSCC.HLA antigens are supposed to play a role in the host's immune response towards neoplastic cells. On a variety of cancer cells both defective HLA class I and de novo class 11 expression have been described, mostly using immunohistochemistry on frozen tissue samples . Low expression of HLA class I antigens correlates with poor differentiation in colorectal cancer (Momburg et al., 1986) and in larynx carcinoma , and with poor prognosis in patients with loco-regional melanoma metastases (van Duinen et al., 1988). HLA class I1 expression in primary melanoma is associated with shorter disease-free survival (Brocker et al., 1985). The significance of HLA expression on tumor cells in relation to immunotherapy is mainly speculative, and has seldom been investigated. High HLA class I1 expression in regressing melanoma lesions following recombinant interleukin-2 (rIL-2)-based immunotherapy has been described by Rubin et al. (1989).Immunohistochemical analysis of HLA antigen expression is largely restricted to fresh frozen tissue. Routine tissue processing, i.e., formaldehyde-fixation and paraffin-embedding , denatures the antigens, and as a result the reactivity of most antibodies is lost. HCA2 and HClO MAbs, however, were prepared against free denatured HLA-A and -B locus heavy chains, respectively, and found to be reactive in paraffin sections of normal tissues (Stam et al., 1990). For detecting HLA class I1 antigen expression in frozen and paraffin secti...

No abstract

Endothelial injury of the tumor microvasculature after isolated limb perfusion (ILP) with TNF-a and melphalan is considered to play an important role in the pathogenesis of tumor necrosis. It is thought to follow endothelial cell activation and subsequent attraction of polymorphonuclear cells (PMNs). The observed selectivity for the tumor could be due to preferential overexpression of cell-adhesion molecules by the tumor vasculature. We tested this proposition by analyzing sequential biopsies from both tumor and normal distant skin, taken from melanoma and sarcoma patients before ILP and at 30 min and 24 h after ILP. Histopathologically confirmed complete response was observed in six of seven melanoma patients, 1-8 months after ILP. By using immunohistochemistry on the light-and electron-microscopic level, the expression patterns of intercellular adhesion molecules-1 (ICAM-I), E-selectin (ELAM-I), VCAM-I, and PECAM-1 were examined. In addition, the results were compared with the effects on HUVECs (human umbilical vein endothelial cells) in vitro of transient exposure of the agents used during ILP. ICAM-1 and PECAM-1 were constitutively expressed on vascular endothelial cells, both in normal tissues and in the tumor lesions. In biopsies taken 30 min after termination of the perfusion, a moderate induction of E-selectin expression on the vascular endothelium in the tumors and a marked expression on the vasculature in the perfused normal skin were observed. It decreased within 24 h after perfusion in both normal skin and in the tumor. The upregulation of E-selectin was accompanied neither by an influx of neutrophils nor by hemorrhagic necrosis. There were no drastic changes in the expression of VCAM-1, ICAM-I, or PECAM-I. These findings imply that the upregulation of E-selectin after ILP is not restricted to the tumor microvasculature and that, therefore, these microvascular events seem not to be the decisive pathomechanism responsible for tumor regression. Key Words: Melanoma-Sarcoma-TNF-a-Melphalan-Isolated limb perfusion.Recombinant human tumor necrosis factor-alpha (TNF-a) has potent antitumor activity on human