Susan Banks-Schlegel

Keratin proteins and carcinoembryonic antigen in lung carcinoma: An immunoperoxidase study of fifty-four cases, with ultrastructural correlations

This study evaluates the use of immunoperoxidase stains for keratin and carcinoembryonic antigen (CEA) in the diagnosis of lung cancer. Immunohistochemical staining for these antigens was performed on paraffin sections from 54 lung neoplasms, and the results were correlated with the histologic, histochemical, and ultrastructural appearances of the tumors. Strong staining for keratin proteins was evident in all squamous cell carcinomas. This reaction was particularly helpful in identifying squamous differentiation in poorly differentiated areas and in cases in which tonofilaments appeared to be sparse or absent on electron microscopy. Weak focal staining for keratin was evident in only two cases of adenocarcinoma, and the difference in frequency of staining between squamous cell carcinomas and adenocarcinomas was statistically significant ( P =0.01). Carcinoembryonic antigen was variably present in all types of lung carcinoma, but staining for it was strongly positive in cases of adenocarcinoma. Positive staining for keratin with negative or focal weak staining for CEA was characteristic of mesotheliomas, which was helpful in differentiating them from adenocarcinomas. When evaluated by multiple means including immunohistochemical analysis, tumors with mixed differentiation were frequently identified and most large cell carcinomas showed evidence of glandular or squamous differentiation. Electron microscopic demonstration of neurosecretory granules was the best single diagnostic criterion for small cell anaplastic carcinomas ( P =0.01, as determined by stepwise logistic regression analysis). Also helpful in differentiating between small and large cell anaplastic carcinomas were a significantly smaller mean cell diameter in the former, as measured on electron micrographs ( P =0.05), and positive staining for keratin, which occurred significantly more often in large cell carcinomas ( P =0.03). These findings suggest that immunoperoxidase stains for keratin and CEA supplement routine histochemical and electron microscopic studies in the diagnosis of lung tumors, and may provide additional specific objective criteria for their classification.

Keratin proteins and carcinoembryonic antigen in synovial sarcomas: An immunohistochemical study of 24 cases

Twenty-four synovial sarcomas were examined for the presence of keratin proteins by an indirect immunoperoxidase method with paraffin-embedded tissues. Keratin proteins were identified in 16 of 24 cases (67 per cent). Both the pseudoglandular and spindle cell areas of all eight of the biphasic synovial sarcomas and the spindle cells of eight of the 16 monophasic synovial sarcomas contained keratin proteins. In spindle cell areas, staining was observed in single cells and small cords and clusters of cells in the absence of cleft formation or other evidence of a pseudoglandular component. The predominant cytologic staining pattern in all cases was peripheral, with localization of staining to the cell membrane or adjacent areas, but diffuse and focal cytoplasmic staining patterns were also observed. No staining for keratin proteins was seen in 101 control cases, including 52 sarcomas of various types. Carcinoembryonic antigen was also identified in four of the 24 synovial sarcomas by an indirect immunoperoxidase technique. The identification of keratin proteins may be helpful in the pathologic diagnosis of synovial sarcoma, particularly the spindle cell monophasic variant.

Keratin protein domains within the human epidermis

Abstract Three species of human keratins are shown to have specific localizations within the epidermis. Using an immunohistochemical technique with rabbit antisera prepared against purified human keratins, two distinct epidermal domains were defined. The 45K and 46K MW keratins occur predominantly in the basal epidermal layer, whereas 55K keratin protein occurs chiefly in the suprabasal, differentiated squamous cells. Commitment of proliferating basal cells to terminal differentiation is accompanied by changes in the proportions of keratin species.

Localized fibrous mesothelioma: An immunohistochemical and electron microscopic study

The absence of keratin staining in tumor cells from localized fibrous mesotheliomas in both paraffin-embedded and frozen sections with sensitive peroxidase-antiperoxidase and avidin-biotin techniques is described. In addition ot the absence of staining for whole-stratum corneum keratin proteins, sections were negative for keratins of different molecular weights (45, 55, and 63 kilodaltons) that are characteristically present in mesothelial cells. Ultrastructurally, the cells most closely resembled mesenchymal cells of the fibroblastic type. These findings are in accordance with recent theories that relate the derivation of localized fibrous mesotheliomas to nonmesothelial cells, including subpleural connective tissue. Based on differences in immunohistochemical staining, the tumors appear to be unrelated to diffuse malignant mesotheliomas.

Tissue-specific expression of keratin proteins in human esophageal and epidermal epithelium and their cultured keratinocytes☆

In contrast to the simplified keratin content of bovine, rabbit, and rat esophageal epithelium (composed mainly of a 57 and 46 or 51 kD keratin, depending on the animal species), human esophageal epithelium contained a quantitatively different array of keratin proteins, ranging in molecular weight from 37 to 61 kD. The pattern of keratin proteins from human esophageal epithelium differed qualitatively and quantitatively from that of human epidermis. Human esophageal epithelium lacked the 63, 65, and 67 kD keratins characteristic of human epidermis, consistent with the absence of a granular layer and an anucleate stratum corneum. Moreover, human esophageal epithelium contained a distinctive 61 kD keratin protein which was either not present or present in only small amounts in human epidermis and variable amounts of a 37 kD keratin. Whereas the 56, 59, and 67 kD keratins were the most abundant keratins in human epidermis, the 52, 57, and 61 kD keratins predominated in human esophageal epithelium. During in vitro cultivation, both human epidermal and esophageal keratinocytes produce colonies which are stratified, but the morphologic appearance of these cultured epithelia differs. Only cultured human epidermal keratinocytes contain keratohyalin granules in the outermost layers and a prominent 67 kD keratin on immunoprecipitation. Otherwise the keratin contents appear similar. In conclusion, human esophageal epithelium exhibited intertissue and interspecies differences in the pattern of keratin proteins. During in vitro cultivation, human esophageal keratinocytes retained some aspects of their distinctive program of differentiation.

Elevated epidermal growth factor receptor gene copy number and expression in a squamous carcinoma cell line.

The human epidermal growth factor (EGF) receptor is known to be homologous to the v-erb B oncogene protein of the avian erythroblastosis virus. Overexpression of the EGF receptor gene in A431 epidermoid carcinoma cells is due to gene amplification. In this study, a variety of squamous cell carcinomas were examined and one, SCC-15, contained high levels of the EGF receptor as determined by immunoprecipitation via an EGF receptor-specific polyclonal antibody. Using a cloned EGF receptor complementary DNA as a probe, the level of EGF receptor RNA was found to be elevated four-fold in SCC-15 relative to normal cultured keratinocytes. When the same probe was used to identify EGF receptor gene fragments on a genomic DNA blot, the SCC-15 cell line was shown to possess an EGF receptor gene copy number amplified four to five times. Gene amplification results in the enhancement in the level of the EGF receptor in several carcinomas and could be responsible for the appearance of the transformed phenotype in these cells.

Keratoacanthoma and squamous cell carcinoma of the skin: Immunohistochemical localization of involucrin and keratin proteins

Fifteen keratoacanthomas and fifteen squamous cell carcinomas of the skin were examined by immunoperoxidase methods for involucrin and both 45- and 63-kilodalton keratins. Keratoacanthomas showed a relatively homogeneous staining pattern for involucrin; all cells except basal cells stained with mild to moderate intensity. Squamous cell carcinomas disclosed a highly irregular involucrin staining pattern with marked variation in staining intensity from cell to cell. Staining patterns for keratin proteins did not appear to distinguish between keratoacanthomas and squamous cell carcinomas. The 45-kilodalton keratin pattern showed diffuse staining within both keratoacanthomas and squamous cell carcinomas, and the 63-kilodalton keratin pattern consisted of focal staining, mostly of dyskeratotic cells. These results suggest that involucrin may serve as a diagnostic aid in differentiating between squamous cell carcinomas and keratoacanthomas. In addition, other lesions in the differential diagnosis of keratoacanthoma and squamous cell carcinoma were also examined for involucrin.

Involucrin in intraepithelial and invasive squamous cell carcinomas of the cervix: An immunohistochemical study

Immunohistochemical staining for involucrin, a cytoplasmic protein synthesized during squamous maturation, was assessed in histologic sections from hysterectomy and cone biopsy specimens from patients with cervical neoplasia. In normal and condylomatous squamous epithelium, diffuse cytoplasmic staining was seen in the suprabasal layers, with no staining of the basal cells. Staining was absent in two cases of cervical intraepithelial neoplasia (CIN), grade III, in which the lesions were composed entirely of undifferentiated cells and markedly decreased in cases involving large numbers of basal cells. In 19 of 23 cases (83 per cent) of CIN, however, focal staining for involucrin was seen in large differentiated cells in the more superficial layers, and in two cases of keratinized CIN diffuse suprabasal staining was observed. Similarly, strong staining for involucrin was present in differentiated areas in one case of microinvasive squamous cell carcinoma and in 93 per cent of cases of infiltrating squamous cell carcinoma. These findings suggest that involucrin is a marker for maturation in cervical squamous epithelial neoplasms. Patterns of immunohistochemical staining for involucrin in keratinized dysplasia and differentiated squamous carcinomas should be taken into consideration if loss of involucrin staining is used as a criterion for neoplastic transformation of cervical epithelium, as has been proposed.

Immunocytochemical localization in normal and transformed human cells in tissue culture using a monoclonal antibody to the src protein of the Harvey strain of Murine Sarcoma Virus

Using a rat monoclonal antibody directed against the p21 src protein of the Harvey strain of Murine Sarcoma Virus (MSV), we have examined the reactivity of human cells in tissue culture using immunofluorescence and electron microscopy. Qualitative results indicated that untransformed mouse and human fibroblastic cells have undetectable amounts of p21; these levels were greatly increased after transformation with Harvey MSV. A group of human tumor cell lines adapted to tissue culture were examined and almost all of the epithelial tumor lines showed significant localization with this antibody. Notable exceptions were two melanoma cell lines which were negative for p21 by immunofluorescence. When normal human epithelial cells derived from esophageal or foreskin epithelium were examined, the antibody showed significant reactivity with subconfluent growing cells. After the normal cells were allowed to become quiescent, the reactivity with this antibody decreased. All of the localization seen by fluorescence was in a distribution consistent with the previously demonstrated location of p21 scr on the inner aspect of the plasma membrane. Electron microscope localization showed labeling for this antigen on the inner surface of the plasma membrane in both transformed mouse cells and in the human tumor cell lines MCF-7 and HTB-2 (RT4). These results suggest that the amounts of p21-like proteins detectable in human epithelial tumor cells do not necessarily reflect their malignant potential, but may be related to their epithelial nature. The loss of detectable localization at quiescence suggests that p21 levels decrease when these epithelial cells stop growing, and raises the possibility that an analog of p21 may be used by these human epithelial cells to regulate cell growth.

Localization of keratin proteins in the human epidermis by a postembedding immunoperoxidase technique.

Individual keratin species were localized ultrastructurally to the tonofilaments and tonofibrils of the human epider-misusing a postembedding immunoperoxidase method. Low molecular weight keratin (45 kD) was localized primarily to the stratum basale and displayed a fine cytoplasmic staining of individual tonofilaments and some tonofibrillar staining. Higher molecular weight keratins (55 and 63 kD) were found predominantly in the suprabasal, differentiated squamous epithelial cells, staining tonofibrils with a web-like pattern. The technique utilized is suitable for studying the distribution of keratin proteins in normal and disease states.