Significance, specificity, and ultrastructural localization of HMB-45 antigen in pigmented ocular tumors.

PURPOSE

To evaluate the expression of the monoclonal HMB-45 antibody in melanocytic and nonmelanocytic ocular tumors and seek "activated" cellular subpopulations in an attempt to distinguish between benign and malignant melanocytes, to compare HMB-45 and S100 activity, and to determine the specificity of this tumor marker for melanocytic ocular lesions.

METHODS

Immunohistologic investigations were performed with paraffin-embedded tissue of 10 acquired conjunctival melanoses, 19 conjunctival nevi, 34 conjunctival melanomas, 69 uveal melanomas, 20 basal cell carcinomas of the lid, 20 cystic dermoids, 15 hemangiomas of the lid, 20 conjunctival papillomas, 20 squamous cell carcinomas, 20 pterygia, 11 sebaceous gland carcinomas, 10 retinoblastomas, and 5 choroidal metastatic carcinomas. The avidin-biotin peroxidase technique and monoclonal HMB-45 antibody were used. The distribution of S100 protein was studied in the melanocytic tumors for comparison. To localize the HMB-45 antigen, lowicryl-embedded tissue of uveal melanomas was investigated immunoelectron microscopically.

RESULTS

More than 95% of the conjunctival and choroidal melanomas expressed the HMB-45 antigen, while S100 was found in all melanomas of the conjunctiva and in 91% of the uveal melanomas. In benign melanocytic lesions of the conjunctiva (nevi and melanocytes), especially the intraepithelial and junctional components stained with HMB-45, and at the site of tumor invasion, infiltrating cells showed increased HMB-45 reactivity. On the whole, HMB-45 antigen was less evenly distributed in the melanocytic tumors investigated than S100 antigen. All nonmelanocytic ocular tumors revealed no HMB-45 expression. Retinal pigment epithelium and tumor-free choroid were negative for HMB-45. The HMB-45 antigen was immunoelectron microscopically found in melanosomes at stages II and III.

CONCLUSION

HMB-45 immunohistology helps in distinguishing melanocytic from nonmelanocytic ocular tumors and often clarifies the front of tumor invasion. The stronger HMB-45 reactivity probably reflects melanocytic activation, but a sharp line between benign and malignant melanocytes cannot be drawn.