Transregulation of microRNA miR-21 promoter by AP-1 transcription factor in cervical cancer cells

In recent years, there has been great effort to elucidate the precise mechanisms involved in carcinogenesis, as well as to identify which genes are involved. Understanding of the complexity of tumor cells has been significantly enhanced with the discovery of genes producing small non-coding RNAs known as microRNAs. Moreover, it has become evident that abnormalities in microRNA expression can contribute to carcinogenesis [1]. MicroRNAs are non-coding RNAs that silence the expression of target genes at the posttranscriptional level by binding to 3′-UTRs of mRNAs and cleaving target mRNAs or repressing target mRNA translation [2, 3]. Although the effect of microRNAs on the initiation and progression of cancer has been well documented, the molecular regulatory mechanisms of microRNAs involved in cancer are poorly understood. MicroRNA biosynthesis is regulated at different levels, and transcriptional regulation at the promoter level is similar to cellular genes. Most DNA binding elements and transcription factor binding sites in microRNA promoter regions are the same as those that control protein-coding genes. However, transcription of primary microRNA transcripts can be dynamically regulated in response to stimulation by different transcription factors, and this process can be altered during carcinogenesis [4]. Thus, microRNAs and transcription factors can cause biological alterations in tumor cells leading to the event cascades of regulatory genetic networks and represent a challenge in the study of carcinogenesis.

Studies have uncovered several transcription factors as pioneer factors that can affect or even cause biological changes required in the initial steps of the transformation, invasion and metastasis cascades [5]. For instance, AP-1 is an early transcriptional regulator mainly composed of c-Fos and c-Jun family members, which bind to variations of the consensus DNA binding sequence TGAG/CTCA, usually located in promoter regions of target genes [6, 7]. An effect of the induction of both c-Fos and c-Jun oncogenes in response to upstream oncogenic signal transduction cascades is that AP-1 activity is increased in many tumor cells and transformed cell lines [8, 9]. By this molecular mechanism, AP-1 coordinates a complex program of gene expression involved in the processes of transformation, invasion and metastasis that constitute the malignancy phenotype [10]. Thus, enhanced AP-1 activity is associated with cellular oncogenic properties, and may be involved in the regulation of miR-21 transcriptional activity. The expression of microRNA miR-21 has been found to be altered in almost all types of cancers and it has been classified as an oncogenic microRNA or oncomir [11‐13]. Due to the critical functions of its target proteins in various signaling pathways, miR-21 is an attractive target for genetic regulation analysis in cancer. The ability of AP-1 and miR-21 to mediate the tumorigenesis process in a variety of cell contexts, their role in oncogene biological activity, and their relevance to cell invasion during metastasis, reflect the critical function that AP-1 and miR-21 play in cancer progression.

It is well established that persistent infection with high-risk oncogenic HPV is the main etiologic agent and initial step in cervical carcinogenesis; however, other factors are required for the development and progression of the malignant phenotype [14]. An event that occurs in HPV-associated carcinogenesis is the global perturbation of cellular gene expression [15]. Although the relationship between HPV infection and cervical cancer has been well documented [16, 17], the detailed regulatory genetic networks of events leading from HPV infection to tumor development have yet to be elucidated. Despite the fact that we have a partial understanding of the molecular mechanisms responsible for microRNA gene regulation, the complete process remains to be clarified. Hence, the enhancement of AP-1 and miR-21 reflects robust regulatory genetic networks, which can contribute to oncogenic potential.

HPV16 and HPV18 are present in 70% of squamous cervical carcinomas [14]. For this reason, cell lines of cervical cancer derived from patients infected with those viral types have been intensively studied. In present study, we used human cervical cancer cells transformed with HPV16 and HPV18 as a cervical cancer model to investigate the impact of the AP-1 complex on the transcriptional activity of the miR-21 promoter. This experimental design allowed the identification of the transcriptional regulation mechanism of the miR-21 promoter in the malignancy phenotype regardless of the presence of intrinsic variations, such as the cell transformation by HPV. The principal novel contribution of our data is to propose that AP-1 transcription factor specifically upregulates miR-21 promoter activity in human cervical cancer cells, which can explain in part miR-21’s increased expression during cervical cancer development.

Previously, we reported that miR-21 down-regulates the expression of target cellular genes in cervical intraepithelial neoplasia-derived cell lines [20]. In order to understanding the upstream regulatory genetic network of the miR-21 gene, we characterized the mechanism of miR-21 gene regulation in cervical cancer cells and described how this molecular pathway contributes to cervical carcinogenesis. In the present study, we demonstrated that members of the AP-1 transcription factor family, including c-Fos and c-Jun, have an essential role in miR-21 gene regulation in cervical cancer cells transformed with high-risk oncogenic HPV.

Several pieces of evidence support the idea that changes in AP-1 complex composition are involved in gene transcriptional regulation during cervical carcinogenesis. We demonstrated than miR-21 expression is induced by constitutive activation of AP-1 in cervical cancer cells (Figs. 1 and 2). AP-1 is a transcription factor involved in a wide variety of pathophysiological responses including cancer. MAPKs including ERKs, JNKs and p38 kinases are known to be the most general common signaling pathways modulating AP-1 activity in response to external stimuli. Therefore, the agents that stimulate or inhibit AP-1 activation such as PMA or SR11302, respectively, may act in the MAPKs signal pathways and in this manner regulate miR-21 expression. Interestingly, this regulatory mechanism seems to be specific to cervical cancer cell lines, since the miR-21 expression in non-tumor cell line (HaCaT cells) was not robustly affected by AP-1 transcriptional factors.

Furthermore, we found endogenous differences in each cell type, with levels of the c-Fos/c-Jun heterodimer increased in SiHa and HeLa cells (Fig. 3). These data suggest a heterogeneous behavior of AP-1 transcription factor specific to each cervical cancer cell line. Previously, De Wilde et al., reported that mRNA expression levels of c-Fos, Fra-1, Fra-2, c-Jun and JunB change at different stages during the malignant transformation of primary keratinocytes by high-risk HPV [21]. While the onset of deregulated expression varied among the AP-1 members, a shift in AP-1 complex composition toward c-Fos and c-Jun expression was favored in tumorigenic cells. They concluded that HPV-mediated transformation cell is associated with altered expression of AP-1 transcription factor among the different genes during the transition from early to late immortalization. It has been also reported that in cells that do not follow a neoplastic process, the AP-1 complex is composed mainly of c-Jun/Fra-1 heterodimers and c-Fos protein is almost absent. On the other hand, in malignant cell lines such as HeLa, SW756 and SiHa, the levels of c-Fos protein are increased and the levels of Fra-1 are decreased, resulting in a prevalent c-Jun/c-Fos dimerization pattern [22]. Our findings of increased expression of c-Fos and c-Jun in SiHa and HeLa cells provide insight into the increased expression of the c-Fos/c-Jun heterodimer in the AP-1 complex in immortal and tumorigenic cells, which closely resemble to malignant cervical lesions. Furthermore, our data suggest that constitutive activation of AP-1 results in high levels of c-Jun that could also activate the c-Fos gene promoter and induce formation of the oncogenic dimer c-Fos/c-Jun, which may be involved in the transcriptional regulation of the miR-21 promoter. In addition, we found that the c-Fos and c-Jun proteins co-localize into the nucleus of cervical cancer cells (Fig. 4). To strengthen our evidence that c-Fos and c-Jun proteins form heterodimers, and translocate into nuclei of cervical cancer cells, is necessary to carry out additional assays such as western blot with a nuclei-cytoplasm protein fractionation, or confocal microscopy assays. However, we do not have the necessary resource to perform this type of assays and this represents a limitation of our study. This finding confirms that the c-Fos and c-Jun proteins has ability to translocate into the nuclei in cervical cancer cells, and that miR-21 promoter activity can be dependent of AP-1 binding sites.

In order to enhance understanding of the AP-1/miR-21 regulatory genetic network, we analyzed the direct binding of AP-1 transcription factor to the miR-21 promoter region in human cervical cancer cells (Fig. 5). The data presented herein demonstrated that c-Fos protein is required for binding to the miR-21 promoter. Interestingly, we observed some DNA–protein signal in the autologous competition in HeLa cells. This DNA–protein signal can be explained by the different alignment of AP1D, AP1M and AP1P sequences, which can recruit different members of AP-1 transcription factor family, and/or other co-factors associated to the AP-1 complex. We found that the sequences identified as AP1D (GTTAATCAC from − 270 to − 262 nt), AP1M (GATGACG from 202 to − 196 nt) and AP1P (GATGACA from 95 to − 89 nt) possess divergent sequence motifs in comparison to the canonical AP-1 binding site (TGAG/CTCA). The AP-1 complex was originally identified as a dimeric transcription factor composed of either a homodimer or heterodimer of AP-1 family members that bind to variations of the canonical DNA binding sites TGAG/CTCA in the 5′-end promoter region of target genes [23]. Binding of AP-1 produces transcriptional synergy of DNA binding that could result in highly specific protein–protein interactions between AP-1 and other promoter-bound transcription complexes. Once the c-Fos/c-Jun heterodimer binds promoter regions, these non-consensus AP-1 sites synergistically enhance transcription, and this transcriptional machine is essential to regulation of several human genes involved in cellular homeostasis, as has been reported for other AP-1-driven promoter interactions [24]. For instance, using recombinant full-length human AP-1 dimers formed between c-Fos and c-Jun for DNA binding and transcriptional analysis, each AP-1 complex exhibits differential activity for distinct non-consensus AP-1 recognition sequences present in the long control region of HPV [25]. Our results are consistent with other research groups, with the novel finding that AP-1 binding sites in the miR-21 promoter can be selectively recognized by constitutive activation of AP-1 complex. In addition, our findings provide insight into the importance of AP-1 complex composition, the relevance of non-canonical binding sites in the miR-21 promoter region recognized by AP-1 family members, and AP-1’s impact in the malignant phenotype of cervical cancer cell lines.

Using in vivo DNA–protein interaction analysis, we demonstrated that AP-1 transcription factor is recruited to AP1D, AP1M, and AP1P sequences in a differential manner to induce the miR-21 promoter in cervical cancer cells (Fig. 6). Our data show that there are not bound levels of AP-1 in the miR-21 promoter region in HaCaT cells, while in SiHa cells, the promoter region of miR-21 is already bound by AP-1. The transcript levels of miR-21 in SiHa and HaCaT cells is similar in both cell lines without PMA treatment (Fig. 2). However, upon PMA treatment, there is a more than six-fold increase in miR-21 transcript levels in SiHa cells (Fig. 6), whereas in PMA treated HaCaT cells there is a roughly two-fold increase in miR-21. These results suggest that recruitment of AP-1 at the 3AP1DMP site is a requisite for the differences observed in miR-21 expression levels between the SiHa and HaCaT cells, and that AP-1 transcription factor is likely cooperating with other factors to achieve higher levels of expression of this promoter upon PMA treatment. These data suggest that this in vivo interaction contributes to miR-21 gene regulation during the cervical carcinogenesis process. Previous studies have noted that the oncogenic or tumor suppressive activity exhibited by distinct members of AP-1 transcription factor is dependent on cell context, genetic background of the tumor, and on AP-1 composition when it binds to non-canonical recognition sequences [26, 27]. Because of these properties, the AP-1 complex has been described as a double-edged sword in tumorigenesis [28, 29]. Furthermore, it has been reported that miR-21 promoter has recognition sites for other transcription factors such as STAT3 and NF-kB; which can regulate miR-21 gene expression at the transcriptional level during carcinogenesis [30, 31]. Our results confirm that c-Fos binds in vivo to AP1D, AP1M, and AP1P sequences in the miR-21 promoter and we propose that it is involved in the AP-1/miR-21 regulatory genetic network during cervical cancer development.

In this study, we demonstrated that the three AP-1 recognition sequences of the miR-21 promoter have promoter activity in cervical cancer cells, and we found that this activity decreased when the AP-1 sequences are deleted in a progressive manner, demonstrating a synergistic effect since the presence of AP-1 sequences increased miR-21 promoter activity (Fig. 7). The demonstration that AP-1 binding sites belong to the miR-21 signal transduction pathway [32], leads us to raise the hypothesis that these molecules are involved in induction of miR-21 gene expression in cervical cancer cells. The AP-1 transcription factor interacts constitutively with the miR-21 core promoter to generate a level of transcription dependent of external stimuli in the context of regulatory genetic networks [33]. These transcriptional complexes could associate with AP-1 recognition sequences and cooperate with other adjacent transcription factors, enhancing the stability of the transcriptional machinery, or alternatively recruiting co-activators that would increase miR-21 promoter activity throughout the autoregulatory feedback loop, as has been reported previously [32]. In addition, Ferguson et al. [34], demonstrated that the transcriptional activation of recombinant AP-1 in a synthetic promoter that has AP-1 recognition sites, can act on its own or interact with other proteins to increase transcription of the miR-21 promoter. In this context, it has been demonstrated that both c-Fos and c-Jun binding to the miR-21 promoter region increases after treatment with PMA, and that STAT3 and AKT signal transduction pathways can induce miR-21 gene promoter activity in cancer cells [35, 36]. We showed that AP-1 recognition sequences from the miR-21 core promoter functionally cooperate and this scenario was supported by the observation that deletion of three AP-1 binding sites decreased the transcription levels of miR-21 in the context of the AP-1/miR-21 regulatory genetic network. Our findings do not exclude that other molecular pathways mediated by different regulatory binding sites may also be present in the miR-21 core promoter region, and may contribute to its transcriptional regulation. Therefore, there exist excellent opportunities to identify new therapeutic molecular targets against cervical cancer.

Full understanding of the interplay between AP-1-dependent transcriptional control and miR-21-mediated post-transcriptional regulation will require study of the regulatory genetic network in physiological conditions and in development of pathology, as well as the identification of new targets for therapeutic intervention. In this context, several reports have contributed strong evidence in the analysis of gene regulation of the AP-1/miR-21 functional circuit. For instance, Fujita et al. [32], described that AP-1 activates miR-21 transcription in conjunction with SWI/SNF complex. They concluded that a conserved double-negative feedback regulatory mechanism might be occurring to sustain miR-21 gene expression in HL-60 cells. Talotta et al. [37], reported that miR-21 is induced by AP-1 in response to the Ras oncoprotein in rat thyroid cells. The authors provide evidence of a positive autoregulation mechanism of AP-1 in the context of Ras oncogene transformation, as well as of the function of miR-21 as an essential target and regulator of the AP-1 complex in tumorigenesis. Du et al. [38], reported that BMP-6 inhibited miR-21 promoter activity by repressing EF1 and AP-1, and that BMP-6 has a role as an anti-metastasis factor through a mechanism involving transcriptional repression of miR-21 in breast cancer cells. Misawa et al. [39], described the role of miR-21 in maintenance of the chemoresistant phenotype of cancer cells through the involvement of AP-1 and PDCD4. Zhu et al. [40], studied the role of miR-21 in hepatocellular carcinoma cells. They demonstrated that miR-21 promotes the migration and invasion processes in hepatocellular carcinoma cells through the miR-21/PDCD4/AP-1 feedback loop, which may represents a therapeutic target in this malignancy. Zhang et al. [41], examined miR-21 expression in human cirrhotic liver samples. They found that miR-21 expression is maintained itself at constant high levels by using the miR-21/PDCD4/AP-1 circuit reported also by Zhu [40]. These findings indicate that the miR-21/PDCD4/AP-1 regulatory genetic network is one of the main driving forces in progression of hepatic fibrosis. This functional circuit comprises the regulatory genetic network of miR-21 in a positive feedback loop through which down-regulation of PTEN and PCDC4 keeps STAT3 activity under control, as has been previously described [42, 43]. Moreover, it has also been shown that microRNA genes can be regulated by epigenetic mechanisms such as aberrant histone post-transcriptional modification or hypermethylation of CpG islands [44, 45]. Given the existing evidence, it is certainly conceivable that miR-21 gene expression might also be regulated by AP-1 transcription factor via methylated DNA sequence motifs. Our data reported herein in cervical cancer cell lines, and the evidence reported by other groups in tumor tissues, suggest that the regulatory genetic networks such as miR-21/PDCD4/AP-1 and miR-21/Let-7a/STAT3/PTEN, are highly complex and create a balance mechanism [35, 46]. In summary, the evidence described supports the existence of a universal AP-1/miR-21/target genes regulatory genetic network, which very likely also operates in cervical cancer cells. These regulatory genetic networks can be used as biomarkers in cervical cancer; however, it is important to be cautious because gene expression can be highly variable within cervical tissues in different subpopulations. Future studies could build on our results to identify more core factors, relevant motifs, and other parallel regulatory genetic networks in cervical carcinogenesis and/or other tumorigenesis processes.

In conclusion, the oncogenic transformation process is generally associated with increased activity of endogenous transcription factors such as AP-1, which have a role in the regulation of expression of microRNAs such as miR-21 through several signal transduction pathways, and enhancement of miR-21 expression can contribute to oncogenic potential. This scenario supports the idea that microRNAs operate cooperatively with transcription factors in regulating a set of target genes, allowing coordinated regulation of gene expression, both at the transcriptional and post-transcriptional level. Therefore, our data strengthens the role of an interesting regulatory genetic network involving AP-1 and miR-21, which modulates critical genes in cervical cancer cells. Eventually, this knowledge may enable both AP-1 and miR-21 to be exploited as therapeutic targets in the treatment of cervical cancer.