We used an immunoperoxidase procedure to examine the tissue distribution of the platelet a-granule membrane protein, GMP-140. In addition to its presence in megakaryocytes and platelets, GMP-140 antigen was found in vascular endothelial cells of diverse human organs, but it was not detected in other types of secretory cells. 35S]Cysteine-labeled human umbilical vein endothelial cells synthesized a GMP-140 molecule containing complex N-linked oligosaccharides similar to those previously demonstrated in platelets and the megakaryocytic HEL cell line. Using an immunogold procedure on frozen thin sections of endothelial cells, we found GMP-140 antigen to be localized to membranes of electron-dense storage granules. In double-label experiments there was colocalization of GMP-140 with vWf, indicating that these granules are Weibel-Palade bodies. When endothelial cells were stimulated with histamine, GMP-140 rapidly redistributed to the plasma membrane. Immunoassays of cell lysates indicated that, relative to total cell protein, less GMP-140 is present in human umbilical vein endothelial cells than in platelets. The restricted expression of GMP-140 in secretory granules of platelets and endothelium suggests that it has a specific function in the vascular system rather than a general role related to inducible secretion.

The majority of renal neoplasms can be distinguished on the basis of histologic examination alone; however, there are morphologic similarities between clear cell renal carcinoma and chromophobe cell carcinoma, as well as between the granular/eosinophilic variants of these tumors and renal oncocytoma. Only a limited number of histochemical markers are available to aid in the differential diagnosis of these neoplasms. Hale's colloidal iron usually yields strong, diffuse cytoplasmic staining of chromophobe cell carcinomas whereas clear cell carcinomas are generally negative; however, interpretation of this stain is not always straightforward. By immunohistochemistry, vimentin is detectable in most clear cell carcinomas and is absent from most chromophobe cell tumors and oncocytomas, but reliance on a single antibody can be misleading. In this report we examine the use of commercially available monoclonal antibodies to RCC and CD10 in the differential diagnosis of common renal tumors. Eighty-five percent of clear cell carcinomas (53 of 62) had detectable surface membrane staining for RCC, and 94% (58 of 62) were positive for CD10. Papillary carcinomas were likewise strongly positive for RCC and CD10 in nearly all cases (13 of 14 each). In contrast, all 19 chromophobe cell carcinomas examined were completely negative for surface membrane staining with both of these markers. Oncocytomas were also negative for RCC (0 of 9), but CD10 was detectable in some cases (3 of 9). These results suggest that the presence of surface membrane staining for RCC and CD10 may be used to confirm a diagnosis of suspected clear cell or papillary renal carcinoma. Chromophobe cell carcinomas should be negative for both markers. The absence of RCC staining may also be helpful in the diagnosis of renal oncocytoma.

Immunohistochemistry is a powerful tool for tissue diagnosis and research. Although the frozen section has remained the gold standard for this important approach to evaluation of antigens in tissues, there is widespread acknowledgment of many limitations. Routine paraffin-embedded sections are widely used for morphological examinaaon of tissues but are not optimal for antigen preservation. In this study, paraffin-embedded tissues fmed with a simple buffer con-tainingzincastheprimaryfuati~~comparedwithtissues frxed with routine formalin, zinc-formalin, paraformaldehyde, ethanol, a variety of commercial (non-formalin-mntaining) fuatives that have been recommended for r e d u d toxicity and improved antigen survival, and frozen sections. Human lymphoid tissues and a group of antibodies to ' Corrrspondence ta Dept. o f h t h o l~, L113, Oregon H e a l t h Sciences U,, Portland, OR 97201.

A patient with an unusual, distinctive cutaneous histiocytosis is described. Extensive morphologic, antigenic, and enzymatic studies indicate that this histiocytosis represents a proliferative disorder of cutaneous indeterminate cells. Features that distinguish this disorder from other histiocytoses are discussed.

Prion proteins (PrP) were localized in the brains of normal and scrapie-infected hamsters by immunohistochemistry and Western blotting. PrP monoclonal antibodies and monospecific anti-PrP peptide sera, which react with both the cellular (PrPC) and scrapie (PrPSc) isoforms of the prion protein, were used to locate PrP in tissue sections. In normal hamsters, PrPC was located primarily in nerve cell bodies throughout the CNS; whereas, in the terminal stages of scrapie, PrP immunoreactivity was shifted to the neuropil and was absent from most nerve cell bodies. Prion proteins were not uniformly dispersed throughout the gray matter of scrapie-infected hamster brains; rather, they were concentrated in those regions that exhibited spongiform degeneration and reactive astrogliosis. Since earlier studies showed that the level of PrPC remains constant during scrapie infection as measured in whole brain homogenates and no antibodies are presently available that can distinguish PrPC from PrPSc, we analyzed individual brain regions by Western blotting. Analysis of proteinase K-digested homogenates of dissected brain regions showed that most of the regional changes in PrP immunoreactivity that are seen during scrapie infection are due to the accumulation of PrPSc. These observations indicate that the tissue pathology of scrapie can be directly correlated with the accumulation of PrPSc in the neuropil, and they suggest that the synthesis and distribution of the prion protein has a central role in the pathogenesis of this disorder.