The tumor suppressor gene KCTD11 ^REN is regulated by Sp1 and methylation and its expression is reduced in tumors

TSGs often locate at chromosomal regions, which are frequently deleted and/or methylated in tumors. High levels of 17p13 somatic alterations have been showed in several tumors, distal and independent of the p53 locus [1‐4].

Our group has identified KCTD11 as an immediate-early gene induced by neurogenic signals [5] and encoding a novel adaptor of Cullin3 ubiquitin E3 ligase complex targeting Histone Deacetylase 1 [6]. Importantly, KCTD11 is a novel TSG that inhibits cell growth and is mapping on human chromosome 17p13.2, whose expression is frequently lost in human MB [4].

To analyze whether the down-regulation of KCTD11 represents a specific feature of MB, as well to other cancers, we performed a wide screening for KCTD11 expression, analyzing 177 human tumor samples and 177 normal matching samples, representing 18 different cancer types. Normal tissues, including larynx, esophagus, stomach, colon-rectum, urinary bladder, lung, breast, gallbladder and endometrium, exhibited a nuclear KCTD11 positive immunohistochemical staining between 40 to 78% (Fig. 1B), whereas the matching tumor samples showed a significant reduction of 0 to 18% of nuclear KCTD11 staining (Fig. 1A and 1B). Reduced KCTD11 expression was not observed in thyroid and kidney tumor tissues vs normal suggesting a tumorigenic specific role of KCTD11 for the above mentioned tissues (Fig. 1A and 1B and data not shown). Moreover KCTD11 was undetected both in normal and cancer tissues from liver, lymph-node and exocrine pancreas (data not shown). Together, these findings clearly indicated that selective tissues expressing KCTD11 have down-regulated this gene during tumorigenesis.

To understand the transcriptional regulation of KCTD11, we identified and analyzed the promoter region. Human KCTD11 proximal promoter is a 623 bp region (Fig. 2A). It turned out to be a TATA- and CAAT-less promoter. The transcription start site (TSS) was previously identified [4] (Fig. 2A). Using the TRANSFACT software, we identified six putative binding sites for Sp1 transcription factor (TF), two of them located closely to the TSS (see Fig. 2A, Sp1-E and Sp1-F). Sp1 is a well-characterized transcriptional activator [7, 8]. It is essential for proper expression of a large variety of genes involved in development, cell growth regulation and cancer [9, 10]. Moreover, Sp1 is responsible for recruiting TATA-binding protein and fixing the TSS at TATAless promoters [9, 11]. Thus, the Sp1-E and Sp1-F binding sites on KCTD11 promoter (Fig. 2A) are likely to be responsible for assembling of the basal transcription complex.

To understand the regulation of KCTD11 promoter by Sp1, we analyzed the transcriptional activity of full-length and various deleted forms of KCTD11 promoter (Fig. 2B-D). Surprisingly, full-length KCTD11 promoter showed a strong basal transcriptional activity (see Fig. 2C, -522/+99), as the luciferase activity was 2.5 folds higher than that of pGL3-promoter, containing the strong SV40 promoter. Although slightly decreased, KCTD11 promoter activity was not drastically dropped down by deletion of the Sp1-A, Sp1-B, Sp1-C and Sp1-D sites, whereas it resulted strongly dependent on Sp1-E and Sp1-F sites (Fig. 2B-C). Importantly, the interval between Sp1-B and Sp1-C sites is too short to allow Sp1 to bind both sites, and this might explain why -398/+99 (containing Sp1-B and Sp1-C) and -338/+99 (lacking Sp1-B) mutants showed the same transactivation ability. Similar results have been obtained by transfecting KCTD11 promoter constructs in other cell lines [additional file 1]. To verify the importance of Sp1-E and Sp1-F binding sites for KCTD11 promoter activity, a full-length KCTD11 promoter construct bearing a deletion -156/-38 - the region containing the Sp1-E and Sp1-F - was generated (Fig. 2B, -522/-157Δ-37/+99). The deletion of Sp1-E and Sp1-F sites strongly affected the ability of KCTD11 promoter to drive transcription (Fig. 2D, compare -522/-157Δ-37/+99 to -522/+99). Most likely, Sp1 binds only to one of these two sites, considering the short interval between Sp1-E and Sp1-F sites. To exclude that the Sp1-E upstream region -156/-82, which was also deleted in the -522/-157Δ-37/+99 construct, was relevant to the promoter activity, we further generated and tested an -81/+99 construct (Fig. 2B), demonstrating that this region did not have any effect on KCTD11 transcriptional activity (Fig. 2D, compare - 156/+99 and -81/+99).

Because of Sp1 protein is constitutively expressed TF [9], to assess specifically the role of Sp1 in the regulation of KCTD11 promoter, we co-transfected the KCTD11 promoter constructs (Fig. 2B) along with Sp1, Sp2 or Sp3 expression vectors into Drosophila Mel-2 cells (Fig. 3A), which provide a null background for Sp TFs activity [9]. In this cellular system, none of KCTD11 promoter constructs showed basal activity when co-transfected along with an empty vector, whereas Sp1 co-transfection strongly activated the promoter (Fig. 3A). Moreover, the functional significance of each Sp1 site in the KCTD11 promoter was evaluated. Sp1-A and Sp1-C resulted to drive KCTD11 promoter activity (Fig. 3A, compare -398/+99 + pPac-Sp1 vs -522/+99 + pPac-Sp1 and - 268/+99 + pPac-Sp1 vs -338/+99 + pPac-Sp1), whereas Sp1-B and Sp1-D did not (Fig. 3A, compare -338/+99 + pPac-Sp1 vs -398/+99 + pPac-Sp1 and -156/+99 + pPac-Sp1 vs -268/+99 + pPac-Sp1). As a control, deletion of all Sp-1 sites completely abrogated KCTD11 promoter activation by Sp1 (Fig. 3A, -37/+99 + pPac-Sp1 column).

As D. Mel-2 cells are null for Sp TFs, Sp1 should not be responsible for recruiting TATA-binding protein and fixing the TSS at TATA-less promoters in this cell context. In fact, full-length KCTD11 promoter lacking Sp1-E/Sp1-F was just slightly (30%) less active than the full-length promoter (Fig. 3A, -522/-157Δ-37/+99 + pPac-Sp1 vs -522/+99 + pPac-Sp1; see also Fig. 2D). Therefore we demonstrated that the activation of KCTD11 promoter was Sp1 specific; conversely the co-transfection with Sp2 or Sp3 expression vectors failed to induce KCTD11 promoter activity (Fig. 3A). In addition, Sp3 did not influence Sp1-driven KCTD11 transcription (Fig. 3A, left panel). Next, we demonstrated that Sp1 binds all the Sp1 binding sites of KCTD11 promoter, except for Sp1-D, by performing EMSAs (Fig. 3B).

Finally, we assessed the in vivo relevance of Sp1-mediated regulation of KCTD11 gene, performing ChIP-qPCR for Sp1, Sp3 and Acetyl-H3 on ARO cells (a thyroid cancer cell line, matching a tumor tissue sample in which KCTD11 is expressed; see Fig. 1) and HCT15 cells (a colon cancer cell line, matching a cancer type in which KCTD11 is inhibited; see Fig. 1). As shown in Fig. 3C, Sp1 binds to KCTD11 promoter much strongly in ARO cells than in HCT15 cells, confirming that Sp1 plays a pivotal role in regulating KCTD11 gene expression, in vivo. At contrary, ARO and HCT15 cell lines showed no differences in Sp3 binding to KCTD11 promoter. Importantly, Acetyl-H3 binds to KCTD11 promoter much strongly in ARO cells than in HCT15 cells, indicating that this promoter is more acetylated in ARO cells than in HCT15 cells.

The Sp-regulated promoters are often associated with GC-rich regions of the genome known as CpG islands, which are important for gene regulation [12]. The presence of six Sp1 binding sites on KCTD11 promoter and the mapping of human KCTD11 at the highly methylated chromosomal region 17p13 prompted us to investigate the involvement of DNA methylation in KCTD11 regulation.

We found a putative CpG island overlapping almost all KCTD11 promoter (-522 to +70), which contains 72 CpG dinucleotides (Fig. 2A; 4A). To evaluate whether methylation might regulate KCTD11 transcription, we investigated the effects of demethylation on KCTD11 expression, in two human colon cancer cell lines, HCT15 and HCT116. 5'-Aza-2'-deoxycytidine (Aza) treatment of these cells significantly increased KCTD11 mRNA level (Fig. 4B). It is worth to note that the induction of KCTD11 expression that we observed here with Aza is comparable with the induction by RA, NGF and EGF or during differentiation of cultured cerebellar GCPs observed in previous studies and there considered physiologically relevant [5, 13]. We therefore investigated the methylation profile of KCTD11 promoter in HCT15 and HCT116 cell lines. DNAs from sodium bisulfite-treated HCT116 and HCT15 were amplified by PCR, subcloned and sequenced. Sequences showed that KCTD11 promoter is methylated in both HCT-15 and HCT-116 cell lines with a similar pattern (Fig. 4C). As control, DNAs from Aza and sodium bisulfite-treated HCT116 and HCT15 cells showed, as expected, C-T conversion of all cytosines (not shown). Together, these data clearly demonstrate a direct role of the methylation on KCTD11 transcriptional activity. Interestingly the regions of KCTD11 promoter resulting methylated are located mainly in the 5' region (Fig. 4C), in particular the region upstream of Sp1-A binding site (-517/-501 meaning CpGs 2 to 4), the region in between of Sp1-A and Sp1-C binding sites (-443/-331 meaning CpGs 8 to 24) and the region upstream of Sp1-D binding site (-279/-278 meaning CpG 30) (Fig. 2A, 4C).

Thus, the majority of the methylated CpGs in KCTD11 promoter are not located at Sp1 binding sites, but in the surrounding regions. To this regard, published reports have implicated Sp1 binding sites involved in the maintenance of the methylation-free status of the CpG islands [12, 14, 15]. Thus, the balance between DNA methylation and Sp1 binding to the promoter may represent a mechanism to regulate the tissue-specific expression pattern of KCTD11. It is noteworthy that the chromosomal region 17p13, where KCTD11 localized, is frequently hypermethylated in cancer. Therefore, the strong down-regulation of KCTD11 in cancers (Fig. 1) may be due to promoter hypermethylation. Of course, we cannot exclude that other mechanisms might play a role in the physio-pathological regulation of KCTD11 expression, such as inhibition of Sp1 activity by other proteins. In summary, we reported a dramatic down-expression of KCTD11 protein in several types of cancer. Importantly, we identified a CpG island and several Sp1 binding sites in the KCTD11 promoter region and further demonstrated that Sp1 transcription factor and DNA methylation contributed, at least in part, to regulate KCTD11 expression. These findings provide a basis to implicate that decreased expression of KCTD11 tumor suppressor gene is associated with cancers of specific tissue types.