Cervical lymph node metastasis: model for study of head/neck melanoma.

Tumour cells spread from primary tumours to form distant metastatic deposits by both lymphatic and blood routes. Melanomas occurring in the head and neck have an extremely poor prognosis largely in part due to late detection resulting in extensive dissemination by lymphatic metastasis. The purpose of this study was to develop an animal model for the study of head and neck melanoma metastasis. B16-F1 parental cells were injected into the subcutis of the ear mid-lobule of C57BL/6 mice. At selected time periods after inoculation, animals were killed by cervical dislocation and autopsied. In some animals tumours had spread to the cervical lymph nodes. Examination of organ systems revealed no evidence of distant metastases. Histological examination of the cervical lymph nodes revealed tumour invasion, beginning at a subcapsular sinus and progressing into the paracortical sinuses. Cells from these nodes were adapted to cell culture, expanded by passage and reinjected into new mice. Subsequent generations of lymph node-selected B16 cell lines were more metastatic than their parental cell line, as evidenced by a more rapid appearance of cervical lymph node and extensive node invasion. Morphologically, the lymph node-selected B16 cell lines were more dendritic than the original B16-F1 parent line and had a larger number of pseudopodial projections. Perhaps increased expression of pseudopods by the metastatic variants may allow for greater migratory potential and hence increased metastatic ability. These results indicate that highly mobile variant B16 sublines can be selected with an increased capacity for cervical lymphatic metastasis.

Metastatic melanoma cells interact with the reticular fibres of the lymph node.

Murine B16 melanoma sublines showing enhanced metastasis to lymph nodes were selected in vivo. Successive selections of tumours metastasizing from the footpad to para-aortic nodes yielded variant tumour cell lines, including an amelanotic line, with moderately increased potential for lymph node metastasis. The phenotype of the variant cells was distinct from that of the parental cells. The lymph node-selected cells had extensive dendritic-like pseudopodial projections and were more motile than the parental cells. In addition, the variant cells were more efficient than the parental cells in attaching to and spreading on preparations of lymph node extracellular matrix. This matrix is composed of an array of reticular fibres containing a core of collagen type III decorated with a basement membrane-like material rich in laminin and type IV collagen. In adhesion assays, the melanoma cells attached best to laminin, collagen, and fibronectin, and poorly to the interstitial matrix proteins collagen types I and III. This pattern of ligand preference was confirmed in adhesion assays to cryostat tissue sections of amnion, in which the tumour cells attached to the basement membrane aspect but not the interstitial stromal matrix. Experiments using specific antibodies established that cell attachment to lymph node reticular fibres was mediated by the beta 1 class of integrin receptor complexes. These results indicate that highly motile variant B16 sublines can be selected for distant lymphatic dissemination, and that interaction between invasive tumour cells and nodal reticular fibres may facilitate this metastatic process.

Analysis of integrin receptors for laminin and type IV collagen on metastatic B16 melanoma cells.

As tumor cells invade surrounding tissue, they adhere to various extracellular matrix components. Previously we reported that B16-BL6 melanoma cell adhesion to both basement membrane and purified protein substrates was blocked by antibody to beta 1-integrin adhesion receptors (R. H. Kramer et al., Cancer Res., 49: 393-402, 1989). In the present study we found, using immunofluorescent staining, that beta 1-integrin complexes were colocalized with vinculin in focal adhesion plaques on laminin, type IV collagen, and fibronectin substrates. To identify potential adhesion receptors on B16 cells, the cells were surface-labeled with 125I, solubilized with detergent, and chromatographed on laminin-, type IV collagen-, and fibronectin-Sepharose columns. On laminin-Sepharose, an integrin heterodimer complex was eluted with EDTA that contained a beta 1 chain at Mr 120,000 and an alpha subunit at Mr 140,000 (nonreduced). This complex was specific for laminin and failed to bind to collagen- or fibronectin-Sepharose columns. Immunoprecipitation with specific monoclonal antibody identified this complex as alpha 6 beta 1 (VLA-6). Furthermore, monoclonal antibody to the alpha 6 beta 1 complex effectively blocked the attachment of B16-BL6 cells to laminin but did not affect adhesion to fibronectin or type IV collagen. We recovered a different integrin complex from type IV collagen-Sepharose columns that was composed of a beta 1 chain and an alpha chain of Mr 180,000 (nonreduced). This same complex also exhibited a weak affinity for laminin-affinity chromatography. The laminin-binding complex and the type IV collagen-binding complex were clearly distinct from the fibronectin-binding receptor and were not eluted by arginyl-glycyl-aspartate-containing peptides. The results suggest that the B16 melanoma cells express multiple integrin-related receptors that appear to mediate cell adhesion to basement membrane matrices.