Expression of the gangliosides GM3, GD3 and GD2 was studied in tissue sections from 19 naevi, 29 primary and 83 metastatic melanoma using the ABC immunoperoxidase technique. GM3 was not detected in normal skin whereas GD2 was detected on the basal and stratum spinosum of the epidermis and on peripheral nerves in the dermis. GD3 was expressed on melanocytes but not on most other components of normal skin. However, GD3 was strongly expressed on epidermis adjacent to naevi and primary melanoma whereas GD2, in contrast to that in normal skin, was not expressed on the epidermis adjacent to 26/29 primary melanoma. All naevi were positive for GM3 and GD3 except that GM3 was not detected on junctional components of naevi. GD2 was not expressed on naevi except in areas showing neuroid differentiation. Studies on melanoma revealed that approximately 60% of primary and 75% of metastatic melanoma expressed GM3 to a varying extent. With 2 exceptions, all primary and metastatic melanomas expressed GD3 although there was variable expression within most of the individual tumours. GD2 was detected in only approximately 25% of primary and 50% of metastatic melanomas. Both GD2 and GD3 were detected on lymphocytes surrounding melanoma. The higher expression of GD2 on metastases compared to primary melanomas was consistent with the view that GD2 expression was associated with increased metastatic potential. However, the low proportion of metastases expressing GD2 and the absence of any correlation with thickness of the primary tumour suggested that GD2 expression was not a reliable marker of metastatic potential. No differences could be detected in ganglioside expression on metastases in skin or lymph nodes. These results appear to have implications for the use of MAbs against gangliosides in therapy of melanoma and in the study of melanocytic differentiation.

No abstract

No abstract

Patients with stage II melanoma were vaccinated with vaccinia virus-induced melanoma cell lysates (VMCL). The vaccine contained viable vaccinia virus, membranous fragments and no intact nuclei. A number of antigens defined by monoclonal antibodies were detected in the vaccine including the ganglioside GD3 and DR antigens. Administration of the vaccine was associated with depression of natural killer cell activity against melanoma and K562 target cells in the first 3-6 months of treatment. Leucocyte dependent antibody (LDA) activity against melanoma cells was induced or increased in titre in approximately half of the patients studied. Continued vaccination was associated in a number of patients with a decrease in LDA titres. Studies on a small sample of patients revealed that this was associated with the development of serum factors which inhibited LDA activity. LDA activity appeared directed to non-MHC antigens on melanoma cells which were of at least two specificities. One specificity which was shared with antigens on a number of non-melanoma carcinoma cells was removed by absorption on fetal brain and may be similar to oncofetal antigens described by other workers. Reactivity against melanocytes was induced in some patients and may underline the development of vitiligo in several patients. These results suggest that vaccines prepared from VMCL may be a favourable method for increasing immune responses against melanoma.