Skin tumors were induced on the backs of Swiss white mice by application of a tumor initiator, dimethylbenz(a)anthracene, followed by chronic application of a tumor promoter, 12-O-tetradecanoylphorbol-13-acetate. Some mice received vehicle (acetone) 1 hr before each 12-O-tetradecanoylphorbol-13-acetate application, others received retinoic acid before each 12-O-tetradecanoylphorbol-13-acetate treatment (chronic retinoic acid treatment), and the remaining mice received four applications of retinoic acid at 12-hr intervals in the 48 hr before the final 12-O-tetradecanoylphorbol-13-acetate application (acute retinoic acid treatment). Mice promoted for 10 weeks formed papillomas. Control mice promoted for 22 weeks had carcinomas, while the chronic retinoic acid-treated mice formed papillomas. The tumors were excised and incubated in medium, and the tumors and media were radioactively labeled, either chemically by iodination or metabolically by incubation in [35S]methionine. The media were collected and the tumors were extracted with urea to give cell-associated material. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that the media and cell-associated fraction of the retinoic acid-treated mice contained more of a high-molecular-weight protein (Mr >400,000) than did the control tumors. We subjected the media and cell-associated protein to affinity chromatography on gelatin attached to Sepharose 4B. The chronic retinoic acid-treated tumors contained the most gelatin-binding protein, the control tumors the least, and the acutely retinoic acid-treated tumor an intermediate amount. Carcinomas contained less gelatin-binding protein than did papillomas. The protein retained on the gelatin column showed behavior on sodium dodecyl sulfate-polyacrylamide gel electrophoresis similar to that of authentic mouse fibronectin.

Papillomas were induced in another group of mice, by the two-step regimen described above. One half of the group received vehicle (acetone) before each of the 12-O-tetradecanoylphorbol-13-acetate applications during the 16-week promotion regimen, while the other half received a topical application of retinoic acid. The tumors were excised and incubated either in [3H]leucine or [3H]fucose for 4 hr, the media were collected, and the tumors were extracted with urea and deoxycholate to extract fibronectin. The media and cell-associated proteins were analyzed by gelatin affinity chromatography and by immunoprecipitation with anti-mouse fibronectin antibody. The media and tumor extracts of the retinoic acid-treated mice contained approximately 2 times as much fibronectin as did the control tumors. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the immunoprecipitated protein and the protein isolated by gelatin affinity chromatography gave protein bands characteristic of fibronectin. We conclude that, in vivo, retinoic acid, which inhibits tumor induction at the promotion step, stimulates the production of a high-molecular-weight glycoprotein identified as fibronectin.


Financial support for this work was provided by NIH Grant CA 13792.

This content is only available via PDF.