Transforming growth factor-beta1-mediated Slug and Snail transcription factor up-regulation reduces the density of Langerhans cells in epithelial metaplasia by affecting E-cadherin expression

Abstract

Epithelial metaplasia (EpM) is an acquired tissue abnormality resulting from the transformation of epithelium into another tissue with a different structure and function. This adaptative process is associated with an increased frequency of (pre)cancerous lesions. We propose that EpM is involved in cancer development by altering the expression of adhesion molecules important for cell-mediated antitumor immunity. Langerhans cells (LCs) are intraepithelial dendritic cells that initiate immune responses against viral or tumor antigens on both skin and mucosal surfaces. In the present study, we showed by immunohistology that the density of CD1a(+) LCs is reduced in EpM of the uterine cervix compared with native squamous epithelium and that the low number of LCs observed in EpM correlates with the down-regulation of cell-surface E-cadherin. We also demonstrated that transforming growth factor-beta1 is not only overexpressed in metaplastic tissues but also reduces E-cadherin expression in keratinocytes in vitro by inducing the promoter activity of Slug and Snail transcription factors. Finally, we showed that in vitro-generated LCs adhere poorly to keratinocytes transfected with either Slug or Snail DNA. These data suggest that transforming growth factor-beta1 indirectly reduces antigen-presenting cell density in EpM by affecting E-cadherin expression, which might explain the increased susceptibility of abnormal tissue differentiation to the development of cancer by the establishment of local immunodeficiency responsible for EpM tumorigenesis.

Figure 1

Correlation between the density of CD1a⁺ LCs and the E-cadherin expression in squamous exocervical epithelium and in EpM (mature and immature). A–C: Involucrin expression in normal squamous epithelium and in areas of mature and immature EpM. Compared with exocervical epithelium (A) and mature EpM (B), the involucrin immunoreactivity is strongly decreased in immature EpM (C). D–F: Density of CD1a⁺ cells in normal squamous epithelium and in EpM. D: The normal squamous epithelium shows a high density of CD1a⁺ cells in the basal and suprabasal cell layers. Mature (E) and immature (F) EpM are infiltrated by a low density of CD1a⁺ cells. G–I: E-cadherin expression in normal squamous epithelium and in EpM. G: The normal squamous epithelium shows typical cell-surface E-cadherin staining of basal and intermediate keratinocytes. The E-cadherin immunoreactivity was intermediate and low in mature (H) and immature (I) EpM, respectively. J and K: Semiquantitative evaluation of E-cadherin and CD1a expression, respectively, in normal exocervix (n = 51), mature (n = 23), and immature (n = 39) areas of EpM. Asterisks indicate statistically significant differences (***P < 0.001). Original magnifications: ×100 (A, D–G); ×200 (B, C, H, I).

Figure 2

The methylation status of CDH1 (E-cadherin) gene in exocervical and metaplastic tissues. A: Representative examples of PCR results for methylated and unmethylated CDH1 gene. After a treatment with the restriction enzyme HpaII, a PCR product is visualized under UV illumination for methylated tissue samples. Genomic DNA without enzyme digestion and DNA treated with methylation-unsensitive endonuclease MspI represent positive and negative control, respectively. B: Methylation analysis of CDH1 gene in cervical tissues with exocervical epithelium (n = 12), mature (n = 14), or immature EpM (n = 14).

Figure 3

TGF-β1 (A–C), Slug (D–F), and Snail (G–I) immunostaining in cervical biopsy specimens. The exocervical epithelium shows a low TGF-β1 staining (A) whereas mature (B) and immature (C) EpM demonstrate, respectively, an intermediate and high expression of TGF-β1. The normal squamous exocervical epithelium shows a medium expression of Slug (D) and a low expression of Snail (G) transcription factors only in upper cell layers. The dashed line delineates the epithelium from the stroma. In contrast, mature (E and H) and immature (F and I) EpM demonstrates a strong Slug (E and F) and Snail (H and I) immunoreactivity. J: Semiquantitative evaluation of TGF-β1 expression, respectively, in normal exocervix (n = 51), mature (n = 23), and immature (n = 39) areas of EpM. Semiquantitative evaluation of Slug (K) and Snail (L) expression, respectively, in normal exocervix (n = 12), mature (n = 14), and immature (n = 14) areas of EpM. Asterisks indicate statistically significant differences (***P < 0.001). Original magnifications: ×100 (A–C, D, G); ×200 (E, F, H, I).

Figure 4

TGF-β1 inhibits E-cadherin and induces Slug and Snail expression in keratinocytes. A: The immunofluorescence demonstrates a diminished expression of E-cadherin (green) at the cell membrane after incubation with TGF-β1 during 48 hours. The nuclei were stained with DAPI (blue). Western blot (B) and RT-PCR (C) confirm the down-regulation of E-cadherin and show the increased expression of Slug and, at a lower level, of Snail transcription factors after TGF-β1 treatment. Results are representative of three independent experiments performed in duplicates. The mean ± SD is shown. Asterisks indicate statistically significant differences (*P < 0.05; **P < 0.01; ***P < 0.001). Original magnifications, ×630.

Figure 5

The decrease in E-cadherin expression caused by TGF-β1 is partially attenuated by Slug and Snail silencing. A: Efficiency of Slug and Snail silencing was demonstrated by Western blot 48 hours after siRNA transfection. B: Immunofluorescence analysis and subcellular localization of E-cadherin (green) in HaCaT-siRNA control, in HaCaT-siRNA Slug and in HaCaT-siRNA Snail stimulated or not with TGF-β1 for 72 hours. The nuclei are stained with DAPI (blue). The images are representative of results obtained in three different experiments. C: Expression of Slug and Snail in TGF-β1-treated cells was controlled by RT-PCR for the indicated times after siRNA transfection. D: Quantification of RT-PCR data demonstrated that the decrease in E-cadherin expression is greater in HaCaT transfected with siRNA control than in HaCaT transfected with siRNA Slug or siRNA Snail. Results are representative of three independent transfection experiments performed in duplicates. The mean ± SD is shown. Asterisks indicate statistically significant differences (*P < 0.05; ***P < 0.001). Original magnifications, ×630.

Figure 6

The decrease in E-cadherin expression induced by Slug and Snail transcription factors affects the interactions between LCs and keratinocytes. A: Seventy-two hours after transfection with Slug and Snail transcription factors, the expression of E-cadherin, Slug and Snail was assessed by Western blot. B: Immunofluorescence analysis and subcellular localization of E-cadherin (green) in human Slug or Snail transfected keratinocytes. Transfections with corresponding empty expression vectors were used as controls. For each condition of transfection, a representative example of LC (red) density observed by field in the heterotypic cell adhesion assay is shown. The nuclei are stained with DAPI (blue). C: Graphic representation of the mean number ± SD of LCs observed by field (original magnification ×200) in the co-culture experiments. For each condition, three independent experiments were performed. The adhesion of LCs to human Slug- or Snail-transfected keratinocytes was significantly lower compared to cells transfected with empty vectors. Asterisks indicate statistically significant differences (**P < 0.01; ***P < 0.001). Original magnifications: ×630 (B, left); ×200 [B (right)].