The transcription factors Slug and Snail act as repressors of Claudin-1 expression in epithelial cells

Abstract

Claudin-1 is an integral membrane protein component of tight junctions. The Snail family of transcription factors are repressors that play a central role in the epithelial-mesenchymal transition, a process that occurs during cancer progression. Snail and Slug members are direct repressors of E-cadherin and act by binding to the specific E-boxes of its proximal promoter. In the present study, we demonstrate that overexpression of Slug or Snail causes a decrease in transepithelial electrical resistance. Overexpression of Slug and Snail in MDCK (Madin-Darby canine kidney) cells down-regulated Claudin-1 at protein and mRNA levels. In addition, Snail and Slug are able to effectively repress human Claudin-1-driven reporter gene constructs containing the wild-type promoter sequence, but not those with mutations in two proximal E-box elements. We also demonstrate by band-shift assay that Snail and Slug bind to the E-box motifs present in the human Claudin-1 promoter. Moreover, an inverse correlation in the levels of Claudin-1 and Slug transcripts were observed in breast cancer cell lines. E-box elements in the Claudin-1 promoter were found to play a critical negative regulatory role in breast cancer cell lines that expressed low levels of Claudin-1 transcript. Significantly, in invasive human breast tumours, high levels of Snail and Slug correlated with low levels of Claudin-1 expression. Taken together, these results support the hypothesis that Claudin-1 is a direct downstream target gene of Snail family factors in epithelial cells.

Figure 1. Snail and Slug induce a disruption of TJs in epithelial cells

Snail-, Slug- and pcDNA3- (control) stably transfected MDCK cells (1×10⁵) were plated on Transwell polycarbonate membrane inserts with an area of 1.1 cm². TEER values were determined after 48 h in culture in DMEM supplemented with 10% (v/v) foetal calf serum. Snail- and Slug-MDCK overexpressing clones (Snail-High, Snail-Low, Slug-High and Slug-Low) exhibited a complete abolition of TEER values when compared with control cells, independent of transcription factor expression levels. Results are means±S.D. for three independent experiments, each performed in triplicate.

Figure 2. Overexpression of Snail and Slug are associated with a full repression of Claudin-1 expression

(A) Control (pcDNA3), Snail- and Slug-transfected MDCK clones were analysed by immunofluorescence and confocal microscopy for the expression of Claudin-1, E-cadherin and ZO-1. In Snail- and Slug-transfected cells, Claudin-1 and E-cadherin became almost undetectable. (B) Western blot analysis of whole-cell extracts for the indicated protein was conducted. Snail and Slug transfection caused a reduction in the protein levels of Claudin-1. (C) The reduction in the protein levels of Claudin-1 was caused by a reduction in the transcript. The presence of cClaudin-1 transcripts in control, Slug- and Snail-transfected clones was analysed by RT–PCR using different amounts of RNA. The expression of GAPDH was analysed in the same sample as a control. (D) RT–PCR for mSnail, mSlug cSnail, cSlug and GAPDH expression in control, Snail and Slug clones. (E) Expression patterns of TJ and adherens junction proteins in control and transfected cells with different levels of Snail and Slug expression. (F) RT–PCR for cClaudin-1, mSnail, mSlug and GAPDH expression in control and different Snail and Slug clones. Representative data from three independent experiments are shown.

Figure 3. Snail- and Slug-induced repression of Claudin-1 promoter activity

Luciferase reporter constructs carrying the human Claudin-1 promoter (−748 to +252) were transfected in MDCK cells (300 ng) with empty vector (pcDNA3) (control), or together with Snail or Slug, or a combination of both factor expression constructs. (A, B, C) When Snail or Slug was co-expressed in MDCK cells, the activity of the Claudin-1 reporter construct was repressed in a dose-dependent manner, an additive effect that was observed when both factors were co-transfected (C). (D) Luciferase reporter constructs carrying a large or short fragment of the human Claudin-1 promoter (300 ng) were transfected in MDCK cells, together with (150 ng) Snail or Slug expression constructs, or with an empty vector (pcDNA3). A shorter fragment was also sensitive to Snail and Slug repression. Results are means±S.D. for three independent experiments, each performed in quadruplicate. *, P<0.05.

Figure 4. Impairment of Snail- and Slug-induced repression of Claudin-1 promoter activity by mutations in the E-boxes

(A) Schematic representation of the E-boxes in the promoter region of human Claudin-1 and E-cadherin: open box, E box; +1, putative transcription star point; ORF, open reading frame. (B) Comparison between the 5′-flanking region of human and canine Claudin-1. Alignment of the sequences revealed the presence of conserved E-boxes (underlined). (C) Mutational analyses. The core 5′-CA(G/C)(G/C)TG-3′ sequence of the E-boxes (E1, E2 and E1-2) was mutated to 5′-TG(G/C)(G/C)TG-3′, in various combinations (shadowed boxes). Luciferase reporter constructs carrying wild-type or mutated boxes were transfected in MDCK cells (300 ng), together with (150 ng) Snail or Slug expression constructs or the empty vector (pcDNA3). Luciferase activity in cells co-transfected with wild-type reporter constructs and the pcDNA3 empty vector was defined as 1. Results are means±S.D. for three independent experiments, each performed in quadruplicate. *, P<0.05.

Figure 5. Direct binding of Snail and Slug to the E-boxes of the Claudin-1 promoter

EMSA for the interaction of Snail and Slug with the E-box sequence. (A) Snail binding. Affinity-purified GST (lane 2) or GST–Snail (lanes 3–9) (10 ng) was incubated with double-stranded ³²P-labelled oligonucleotides (2.5 ng) (+165 to +201) corresponding to the sequence containing two E-boxes of the Claudin-1 promoter; in lane 1, no protein was added. GST–Snail (arrowheads in lane 3), but not GST (lane 2), formed DNA complexes. Competitions were performed with a 200-fold excess of E1 (+153 to +192) and E2 (+176 to +215) wild-type probes (unlabelled oligonucleotides) (lanes 4 and 5), or mutant E1 and E2 boxes (lanes 6 and 7). Complex formation was affected by unlabelled wild-type E1 and E2 boxes, but not by unlabelled mutated boxes. Supershift experiments were performed by adding 200 ng of anti-Snail (lane 9) or non-specific (lane 8) antibody. Arrows in lane 9 indicate the supershifted retarded complexes. (B) Slug binding. Affinity-purified GST–Slug (lanes 2–8) (10 ng) was incubated with double-stranded ³²P-labelled oligonucleotides (2.5 ng), containing E1 and E2 boxes (+165 to +201); in lane 1, no protein was added. GST–Slug formed two DNA complexes (arrowheads in lane 2). Competitions were performed with a 200-fold excess of wild-type E1 and E2 (lanes 3 and 4), or mutant E1 and E2 boxes (lanes 5 and 6). Results were very similar to those shown in (A). Supershift experiments were analysed by adding 200 ng of anti-Slug (lane 8) or non-specific (lane 7) antibody. Representative results from six independent experiments are shown. wt, wild-type; mut, mutant; Ab, antibody.

Figure 6. Expression of Snail, Slug and Claudin-1 in breast cancer cell lines and in human breast tumour samples

(A) Total RNA from a panel of breast cancer cell lines was reverse-transcribed into cDNA. Results show the cell line/fibroblast ratio of normalized mRNA levels of Claudin-1, Snail and Slug using the comparative threshold cycle method. Claudin-1 compared with Slug: Spearman r=−0.64; P=0.042. Claudin-1 compared with Snail: Spearman r=−0.15; P=0.6. (B) Luciferase reporter constructs carrying wild-type or mutated boxes (300 ng) were transfected in a panel of breast cancer cell lines. Luciferase activity in MDCK co-transfected with wild-type reporter constructs was defined as 1. Results are means±S.D. for three independent experiments. *, P<0.05. (C) Gene cluster assay of 295 human primary invasive breast carcinoma samples using three genes: Snail, Slug and Claudin-1. For each gene, the ratio was calculated with respect to the intensity of a reference pool made up of equal amounts of cRNA from all tumours [20].