Vitamin A in epithelial differentiation and skin carcinogenesis

Our concepts on epithelial phenotypes are illustrated in Figure 1, which summarizes the multiplicity of the principal epithelial phenotypes under normal nutritional intake of vitamin A. These phenotypes can be (1) simple columnar, as found in the endocervix; (2) pseudostratified, as for the trachea; (3) squamous stratified, as for the ectocervix; and (4) keratinizing squamous stratified, as for the epidermis. Nutritional deficiency of vitamin A permits transformation of the simple columnar endocervical epithelium, initially into a pseudostratified epithelium as found in trachea by the appearance of basal-like 'reserve' cells, then into the squamous stratified phenotype, and eventually into the keratinizing squamous stratified phenotype. Obviously, nutritional as well as hormonal factors are involved in the definition of morphological characteristics of the various lining epithelia of the body. We are studying the interplay between RA and steroid hormones in the development of these changes. In situ hybridization revealed a specific epithelial localization of the RARs. In particular, RARβ isoforms were found to be localized in the columnar epithelium, suprajunctionally, and RARγ isoforms in the squamous stratified epithelium, subjunctionally. RARα isoforms were found distributed throughout the stratified squamous and the simple columnar phenotypes. The interplay or differential expression of specific RARs, and possibly of RXRs and estrogen receptors in the cervical epithelium, may be an important determinant of the differentiated phenotype. Our present efforts are aimed at understanding the regulation of receptor expression in normal, preneoplastic, and neoplastic cervical skin epithelia. Preliminary results indicate a significant down-regulation of RARα, specifically in skin carcinomas. Whether this is a consequence or a cause of malignant conversion remains to be determined. A second item of interest is that dietary RA, in pharmacologic doses, inhibits conversion of the benign lesion papilloma to carcinoma without influencing papilloma formation to any significant extent. A likely hypothesis to explain this effect is that RA upregulates retinoid receptor expression, and that this receptor protein might form complexes with AP-1 proteins, such as c-fos, thereby preventing their action in malignant conversion. This hypothesis is being investigated in our laboratory and is particularly relevant, since c-fos has been shown to cause malignant conversion of papilloma to carcinoma cells in grafting experiments. Alternatively, RA may induce differentiation and cell loss in papilloma and/or carcinoma cells, thereby reducing the chance for the formation visible malignant lesions. This latter mechanism has already been demonstrated in the induction of differentiation of APL cells in vivo and in vitro. The retinoids represent a family of compounds that is uniquely capable of interacting with processes as diverse and yet interconnected as epithelial phenotypic interconversions and benign to malignant neoplastic transformation. This versatility is reflected in their use as differentiation agents in APL. The work of our laboratory is presently focused on understanding the steps involved in the neoplastic process and developing clinical strategies to control neoplastic progression.