Retinoic acid resistance at late stages of human papillomavirus type 16-mediated transformation of human keratinocytes arises despite intact retinoid signaling and is due to a loss of sensitivity to transforming growth factor-beta.

In our in vitro model of human cell carcinogenesis, normal human foreskin keratinocytes (HKc) transfected with human papillomavirus type 16 DNA (HKc/HPV16) progress toward malignancy through several phenotypically defined and reproducible "steps" that include immortalization, growth factor independence (HKc/GFI), differentiation resistance (HKc/DR), and ultimately malignant conversion. While HKc/HPV16 are very sensitive to growth inhibition by all-trans-retinoic acid (RA) at early passages, they lose their sensitivity to RA during progression in culture. However, gel mobility shift assays using the retinoid response elements DR1 and DR5 showed no changes in binding activity of nuclear extracts obtained from HKc/HPV16 at different stages of in vitro progression. Similarly, Western blot analyses for retinoic acid receptor gamma-1 and the retinoid X receptors failed to reveal any decreases in the levels of these retinoid receptors throughout progression. In addition, luciferase activity driven by the SV40 promoter with a DR5 enhancer element was activated following RA treatment of HKc/DR that were resistant to growth inhibition by RA. Since RA induces transforming growth factor-beta2 (TGF-beta2) in normal HKc and HKc/HPV16, we investigated whether this response changed during progression. Again, RA induced TGF-beta2 mRNA in early and late passage HKc/HPV16, HKc/GFI, and HKc/DR approximately to the same extent, confirming that the RA signaling pathways remained intact during in vitro progression despite the fact that the cells become resistant to growth inhibition by RA. We then investigated the sensitivity of HKc/HPV16 to growth inhibition by TGF-beta. While early passage HKc/HPV16 were as sensitive as normal HKc to growth inhibition by TGF-beta1 and TGF-beta2, the cells became increasingly resistant to both TGF-beta isotypes during in vitro progression. In addition, while both RA and TGF-beta produced a decrease in the levels of mRNA for the HPV16 oncogenes E6 and E7 in early passage HKc/HPV16, this effect was also lost at later stages of progression. Finally, blocking anti-TGF-beta antibodies partially prevented RA inhibition of growth and E6/E7 expression in early passage HKc/HPV16. Taken together, these data strongly suggest that inhibition of growth and HPV16 early gene expression in HKc/HPV16 by RA is mediated by TGF-beta and that a loss of RA sensitivity is linked to TGF-beta resistance rather than alterations in RA signaling.