Tissue factor pathway inhibitor-2 was repressed by CpG hypermethylation through inhibition of KLF6 binding in highly invasive breast cancer cells

Abstract

Background: Tissue factor pathway inhibitor-2 (TFPI-2) is a matrix-associated Kunitz inhibitor that inhibits plasmin and trypsin-mediated activation of zymogen matrix metalloproteinases involved in tumor progression, invasion and metastasis. Here, we have investigated the mechanism of DNA methylation on the repression of TFPI-2 in breast cancer cell lines.

Results: We found that both protein and mRNA of TFPI-2 could not be detected in highly invasive breast cancer cell line MDA-MB-435. To further investigate the mechanism of TFPI-2 repression in breast cancer cells, 1.5 Kb TFPI-2 promoter was cloned, and several genetic variations were detected, but the promoter luciferase activities were not affected by the point mutation in the promoter region and the phenomena was further supported by deleted mutation. Scan mutation and informatics analysis identified a potential KLF6 binding site in TFPI-2 promoter. It was revealed, by bisulfite modified sequence, that the CpG island in TFPI-2 promoter region was hypermethylated in MDA-MB-435. Finally, using EMSA and ChIP assay, we demonstrated that the CpG methylation in the binding site of KLF-6 diminished the binding of KLF6 to TFPI-2 promoter.

Conclusion: In this study, we found that the CpG islands in TFPI-2 promoter was hypermethylated in highly invasive breast cancer cell line, and DNA methylation in the entire promoter region caused TFPI-2 repression by inducing inactive chromatin structure and decreasing KLF6 binding to its DNA binding sequence.

Figure 1

Expression of TFPI-2 in human breast cancer cells with different metastasis potential. (A). Western blot analysis of TFPI-2 protein expression. The HUVEC cell line was used as positive control and the protein levels of GAPDH were determined as control. (B). Quantitative real-time PCR analysis of TFPI-2 mRNA levels. All expression levels of TFPI-2 in breast cancer cell lines were normalized to the level of its expression in HUVEC. Lane 1: MDA-MB-435. Lane 2: T47D. Lane 3: MCF-7. Lane 4: HUVEC.

Figure 2

Analysis of TFPI-2 promoter. (A) Breast cancer cell lines were transfected with promoter-luciferase plasmid cloned from different cell lines, bearing respective genetic mutation. The promoter from different cell line has nearly the same luciferase activity in both MDA-MB-435 and MCF-7 cell lines. Results are shown as mean ± SD. The apparent lower level of promoter activity in MDA-MB-435 cell was not statistically significant (p = 0.342>0.05). (B) Luciferase reporter gene constructs with 5'ends between nucleotides -1436 and -144 and a common 3'end at +75 were transiently transfected into breast cancer cells. The minimal construct P-144 has the same luciferase activity as P-1436. (C) Linker-scan mutation analysis of TFPI-2 promoter -144 to +1 region. The fragment from -144 to the transcriptional start point (+1) of TFPI-2 promoter were systematically replaced with an BamH I & Xba I polylinker (GGATCCTCTAGA), the resulting sequence were analyzed for not introducing new transcription factor binding site by MatInspector [21]. SM1 indicate the promoter sequence from -12 to +1, and in turn every 12 bases from -144 to -12 was indicated by SM12 to SM2. Then the luciferase promoter reporter plasmids were transfected into MDA-MB-435 cell line and the reported luciferase activity was measured. Results are shown as mean ± SD relative to the wild type (*p < 0.05; **p < 0.001).

Figure 3

KLF6 can transactivate and binds to the promoter of TFPI-2. (A) The p-144 promoter-luciferase activity induced by cotransfection with various amount of KLF6 expression plasmids in MDA-MB-435 cells. The KLF6 cDNA was a generous gift of Scott Friedman and was cloned into PCDNA 3.0 expression vector. (B) Both the predicted KLF6 binding sequence (K6) and the authentic KLF6 binding (Kc) [46] can form specific bands with nuclear extract (NE) from MCF-7 cell lines. And the specific bands were supershifted by the antibody of KLF6. (C) No difference of K6 binding activity was found between MCF-7 and MDA-MB-435 cell lines. Lane 1 to 6: NE from MCF-7; lane 7, 8: NE from MDA-MB-435. Lane 5, 7: 2.5 μg NE; lane 6, 8: 5 μg NE.

Figure 4

(A) The sequence of CpG islands in TFPI-2 promoter. The CpG dinucleoide detected by bisulfite modified sequence was indicated in gray shadow. (B) The TFPI-2 gene promoter is highly methylated in MDA-MB-435 cells, whereas it is largely unmethylated in MCF-7 cells. For each cell type, the methylation status of 8 individual clones as determined by bisulfite sequencing analyses is shown in rows 1 to 8. The filled or open circles represent the methylated or unmethylated CpG sites, respectively. (C, D) MDA-MB-435 cells were seeded into 10 cm tissue culture dishes at an initial density of 33% confluence, allowed to attach over night, and then replaced with medium containing 1.25, 2.5 and 5 μM of AZA for 72 hour, then expression of TFPI-2 was detected by western-blot (C) and real-time PCR (D).

Figure 5

CpG methylation blocked the binding of KLF6 to TFPI-2 promoter. (A) CpG methylation in core matrix decreased the binding ability of KLF6 in vitro. Competition were used to determine the affinity of K6, Kc, methylated K6 (mK6) and muted K6 (Kmute). A 50- fold, 100-fold and 200-fold (50×, 100×, 200×) molar excess of cold K6, Kc, mK6 and Kmute cold probe were chased with labeled Kc. All the sequences were listed in table 1 (B) ChIP reveals that hypermethylation blocks KLF6 binding to the TFPI-2 promoter in vivo. Both CpG Unmethylated MCF-7 and methylated MDA-MB-435 cells were tested for binding of KLF6 by ChIP in vivo. Templates for PCR corresponded to the input used in the ChIP assay (input) and DNA obtained from immunoprecipitations performed in the absence (No antibody) or presence of anti KLF6-specific antibody.