CEBP factors regulate telomerase reverse transcriptase promoter activity in whey acidic protein‐T mice during mammary carcinogenesis
Abstract
Telomerase is activated in the majority of invasive breast cancers, but the time point of telomerase activation during mammary carcinogenesis is not clear. We have recently presented a transgenic mouse model to study human telomerase reverse transcriptase (TERT) gene expression in vivo (hTERTp‐lacZ). In the present study, hTERTp‐lacZxWAP‐T bitransgenic mice were generated to analyze the mechanisms responsible for human and mouse TERT upregulation during tumor progression in vivo. We found that telomerase activity and TERT expression were consistently upregulated in SV40‐induced invasive mammary tumors compared to normal and hyperplastic tissues and ductal carcinoma in situ (DCIS). Human and mouse TERT genes are regulated similarly in the breast tissue, involving the CEBP transcription factors. Loss of CEBP‐α and induction of CEBP‐β expression correlated well with the activation of TERT expression in mouse mammary tumors. Transfection of CEBP‐α into human or murine cells resulted in TERT repression, whereas knockdown of CEBP‐α in primary human mammary epithelial cells resulted in reactivation of endogenous TERT expression and telomerase activity. Conversely, ectopic expression of CEBP‐β activated endogenous TERT gene expression. Moreover, ChIP and EMSA experiments revealed binding of CEBP‐α and CEBP‐β to human TERT‐promoter. This is the first evidence indicating that CEBP‐α and CEBP‐β are involved in TERT gene regulation during carcinogenesis.
Telomerase activity is necessary for the unlimited proliferation capacity of human cells. In the adult human, activity of telomerase is restricted to tissue stem cells, progenitor cells and germ cells and is undetectable in the majority of adult somatic tissues,1-3 primarily due to the repression of its catalytic subunit telomerase reverse transcriptase (TERT). During human tumorigenesis, telomerase becomes reactivated by transcriptional upregulation of TERT. Therefore, downregulation of telomerase activity is thought to function as a tumor suppressor barrier.4, 5
Several studies have described telomerase activity in invasive breast cancers, but the time point and the molecular mechanisms of telomerase reactivation during mammary carcinogenesis are not clear.6 In human breast cancer, some reports hint at the reactivation of telomerase in invasive cancer, but not in DCIS,7 whereas others reported reactivation of telomerase already at the DCIS state.8-10 On the basis of the strong association of telomerase activity with invasive cancer, Umbricht et al.9 argued that the reactivation of telomerase is rather a late event in the progression of mammary cancer. Other studies reported that hTERT mRNA expression is associated with a poor clinical outcome in human breast cancer.11, 12
We have previously published a reporter mouse model (hTERTp‐LacZ) using an 8‐kbp human TERT‐promoter.13 To better understand the mechanism of human telomerase reactivation during carcinogenesis, in the present study, we have generated a bitransgenic mouse model, where telomerase activity can be traced during tumor development. For this purpose, we crossed hTERTp‐lacZ mice with whey acidic protein (WAP)‐T mice. WAP‐T mice carry the SV40‐ER under the control of the murine WAP‐promoter, which is activated by lactotrophic hormones in differentiating mammary epithelial cells (MPCs) during late pregnancy and lactation, resulting in the expression of small‐ and large‐T‐antigens and thus functional inactivation of the p53 and RbB tumor suppressor proteins.14 As a consequence, WAP‐T mice develop multifocal intraepithelial neoplasia, which can further progress to invasive mammary carcinoma. The relevance of this model is emphasized by the close similarity in histology of the mouse tumors with corresponding human tumors.
Importantly, stages of tumor initiation and progression can be precisely followed in a time‐dependent manner after the activation of the SV40 early genome region.15 This allows a time‐dependent analysis of the molecular events occurring during the early steps of tumorigenesis in this animal model. The reactivation of telomerase is one of these events and depends on the loss of tissue‐specific repression mechanisms and/or the induction transcriptional activators.
First, we determined the activity of telomerase at defined time points (i.e., stages of tumor progression) in correlation to histological changes and found that reactivation of telomerase occurs in invasive cancers. At the molecular level, we found that the CEBP factors are involved in the regulation of human and mouse TERT gene regulation. CEBP‐α represses TERT‐promoter activity in normal tissues, whereas CEBP‐β seems to act as a transcriptional activator of the TERT gene in tumor tissue.
Results
Initiation and progression of mammary carcinomas in WAP‐T transgenic mice
The generation and analysis of whey‐acidic‐protein (WAP)‐T mouse lines have been previously described.14 These animals develop multifocal ductal carcinoma in situ (DCIS) and consequently invasive carcinoma within strain‐specific periods of latency. In this study, we used the WAP‐T‐NP8 transgenic mouse line (henceforth called WAP‐T mice). A schematic figure describes the initiation and progression of tumorigenesis in WAP‐T mice (Fig. 1a). In these mice, about 50% of luminal epithelial cells express SV40 Large‐T‐Ag (LT) upon induction by lactotrophic hormones during lactation (Fig. 1Ba). At day 30 postweaning, LT expression is only detectable in a few ductal and acinar cells that survived involution (Fig. 1Bb). At day 50–60 postweaning, additional LT‐positive ducts arise, exhibiting first hyperplastic lesions (mammary intraepithelial neoplasia at early stage, MIN‐I; Fig. 1Bc), some of which later (90–120 days postweaning) develop to prevalent in situ carcinomas (mammary intraepithelial neoplesia at late stage, MIN‐II; Fig. 1Bd). At day 90 postweaning, virtually all terminal end buds in each mamma are converted to intraepithelial neoplasias (here, we refer this stage as the G0 grade or MIN‐II). Occasionally, MIN‐II could be observed at earlier time points. Palpable tumors usually arise between 5 and 10 months after initiation of tumorigenesis (Fig. 1c). Interestingly, however, only few invasive carcinomas (Figs. 1Be–1Bh), which can only be observed in one (occasionally two) mammary glands in each mouse,16 arise from these DCIS in a stochastic manner. The transition from intraepithelial neoplasia to invasive carcinoma thus is the decisive step in the formation of malignant mammary carcinoma, which goes along with additional genetic alterations. Activation of telomerase may be one of these genetic changes.
Initiation and progression of mammary carcinomas in WAP‐T mice. (A) Schematic drawing illustrating the initiation and progression of mammary carcinomas in WAP‐T transgenic mice. Induction of the SV40‐T‐Ags can be detected as early as the second week of pregnancy. To better compare the time points between animals, the day of weaning has been defined as day 0 of sample collection. (B) Histology of WAP‐T‐NP8 tumors and LT expression (red staining) by IHC. Grading of mammary glands was performed according to Annapolis Criteria. Tumors of grades G1 and G2 are well‐differentiated, whereas most tumors display a poorly differentiated (G3 and G4) morphology. (C) Kaplan–Meier curves show severely reduced life spans of WAP‐T mice compared to control BALB/c wt animals.
Human and mouse TERT genes are induced during tumorigenesis
In mouse tissues, endogenous telomerase activity can be detected to some extent, especially in testis and liver. We therefore determined telomerase activity in normal and tumor tissues in hTERTp‐lacZxWAP‐T mice (Figs. 2a and 2b). High‐endogenous telomerase activity can be detected in normal liver samples, but is undetectable (muscle) or low in most normal mouse tissues including normal mammary tissues (Fig. 2a). In contrast, telomerase is active in mouse mammary tumors (Fig. 2b). These results coincide with the previous reports.13, 17
Upregulation of endogenous telomerase activity and Tert‐promoters in tumors of WAP‐T x hTERTp‐LacZ mice. (a) Telomerase activity in normal tissues of the same animal, as determined by TRAP assay. (b) Telomerase activity in normal and tumor tissues of an animal with undifferentiated (G3) tumors in two mammary glands. (c) lacZ, mTert and LT expression by semiquantitative RT‐PCR. (d, e) lacZ and mTert mRNA levels as determined by real‐time qPCR and (f) telomerase activity by real‐time TRAP assay. Telomerase activity was determined either by the gel‐based TRAP assay with the TRAPEZE‐Kit or by a real‐time quantitative TRAP assay established in our laboratory (see Material and Methods section) using 100 ng total cell lysate. (g) Sample preparation procedure for RNA and protein: from virgin mice and from mice at MIN‐I (50 days postweaning) and MIN‐II (90–120 days postweaning) stages, mammary glands (inguinal, upperthoracal, lower thoracal and abdominal) from one half‐site were embedded for histology and mammary glands from the other half‐site were snap‐frozen for later gene expression analyzes. From the MIN‐II stage, only noninvasive samples were analyzed. Invasive mammary carcinomas were grouped into well‐differentiated invasive mammary carcinomas (G1 + G2 stages) and in nondifferentiated invasive mammary carcinomas (G3 + G4 stages). These samples were cut with a scalpel, and a half was embedded for histological analyzes and the other half was snap‐frozen for gene expression analyzes. In the real‐time experiments, two different mammary tissue samples from 12 individual virgin/MIN animals (in total 24 samples) were analyzed, whereas 24 tumor samples were from 24 individual animals in each case.
To analyze the regulation of human and mouse TERT gene promoters during tumor progression, tissues were prepared before and at various time points after initiation of tumorigenesis. First, we used semiquantitative real‐time polymerase chain reaction (RT‐PCR) to determine endogenous mTert mRNA levels and expression of the lacZ reporter gene as a measurement of hTERT‐promoter activity in noninduced mouse mammary samples and at various time points after induction of tumorigenesis (Figs. 2c–2e). Expression of lacZ and mTert genes was undetectable or low in uninduced (virgin) hTERTp‐lacZxWAP‐T mice, as determined by semiquantitative RT‐PCR with a small set of samples (n = 3 or 4). Importantly, no activation of lacZ reporter or of endogenous mTert expression could be detected in samples from 50 days postweaning (MIN‐I) or 120 days postweaning (MIN‐II). These results could be confirmed with a higher number of samples by real‐time quantitative RT‐PCR (Figs. 2d and 2e). In line with telomerase activity, mTert expression is induced in invasive mammary tumors (Fig. 2f). Notably, we also observed a strong increase in lacZ mRNA levels, indicative for hTERT‐promoter reactivation in tumor samples. To exclude that the observed increase in lacZ reporter and the endogenous mTert gene induction is not an epiphenomenon due an increase in the number of epithelial cells in the tumor sample, we performed RT‐qPCR for the epithelial marker EpCAM. Importantly, we observed a strong increase in EpCAM mRNA levels at the MIN‐II stage and a further increase in differentiated invasive tumors (Fig. 2g). Interestingly, EpCAM levels decreased in undifferentiated tumors samples to levels of the MIN‐II stage. The increase of EpCAM at the MIN‐II stage is also compatible with the histological data, where we observe an increase in epithelial cell number (Fig. 1b). Notably, endogenous mTert and lacZ expression is induced only in invasive mammary cancers but not at the MIN‐II stage. The same is true for the induction of CEBP‐β and, on the other hand, for the reduction of CEBP‐α (Figs. 3b and 3c). These data support our conclusion that the induction of TERT gene expression is not merely based on the changes in epithelial‐cell number. This conclusion is further supported by the histological analysis of the reporter β‐Gal protein in normal and tumor samples (Fig. 3). Together, these results indicate that the activation of TERT gene expression is a late step in tumorigenesis.
CEBP‐α downregulation and CEBP‐β up‐regulation during tumorigenesis in hTERTp‐lacZxWAP‐T mice. (a) Human and mouse Tert‐promoters share potential CEBP‐binding sites (containing the consensus binding motif GCAAT). Some of the previously described regulatory elements are also shown. (b, c) Results of real‐time quantitative RT‐PCR indicating the downregulation of CEBP‐α and the upregulation of CEBP‐β mRNA levels in invasive mouse mammary tumor samples. For the gene expression analyzes, the same samples were used as in Figures 2d and 2e. (d) Immunoblot experiments demonstrating the downregulation of CEBP‐α and the upregulation of CEBP‐β protein levels in invasive mouse mammary tumor samples. Representative results from at least three repeat experiments are shown with protein extracts from two animals at each step. The repeat experiments were done with protein extracts from independent mice harvested at the same time points. (e) Immunohistochemistry showing the expression of β‐Gal and CEBP‐α in normal mouse mammary tissue and in tumor samples. Arrows indicate the epithelial cells in the normal mammary epithelium (black) or in the tumor tissue (red).
In accordance with the studies in many human primary cell lines,18 we also found that expression of LT has no direct influence on TERT‐promoter activation. LT mRNA (Fig. 2c) and protein (Fig. 1b) are already detectable shortly after initiation of tumorigenesis, whereas lacZ and mTert activation only occurs in invasive tumors. In total, tissues from 24 mice were analyzed for lacZ and mTert expression as well as telomerase activity. Only one sample of the analyzed tumors did not show telomerase/TERT‐promoter activation in the tumor sample. Interestingly, histological analysis of this tumor sample revealed that it was not a mammary tumor, but a fibrohistiosarcoma (not shown).
In line with the previous results of mouse mammary carcinogenesis,17 we did not observe any significant loss of telomere length during tumorigenesis (not shown).
CEBP‐α downregulation and CEBP‐β upregulation during tumorigenesis in hTERTp‐lacZxWAP‐T mice
TERT gene expression is differentially regulated in most adult mouse and human tissues. In this analysis, we however found that expression of the lacZ reporter gene under the control of the human TERT gene promoter and of the murine Tert gene shows a very similar pattern during mammary carcinogenesis. We thus hypothesized that the human and murine TERT gene promoters may share regulatory elements, which account for their similar expression in the mammary tissue. A comparison of both promoters (Fig. 3a) revealed conserved CEBP‐binding sites at −0.5 and −1.0 kbp positions of mTert‐promoter (0.65 and 1.3 kbp in the hTERT‐promoter, respectively). Of note, CEBP‐α and CEBP‐β have been discussed in the context of mammary tumorigenesis. There is however no information about their involvement in TERT gene regulation.
We therefore measured the expression of these two major CEBP factors during tumorigenesis. We found that CEBP‐α expression was downregulated to background levels in invasive mammary carcinomas and thus inversely correlated with TERT expression (Fig. 3b). Conversely, CEBP‐β expression was elevated in invasive cancers (Fig. 3c). Expression at the mRNA level strongly correlated with CEBP‐α and CEBP‐β protein levels (Fig. 3d). For CEBP‐β, three alternative translation protein products, the 35 and 32 kDA LAP (liver‐enriched activator protein) and the smaller, 20 kDa LIP (liver‐enriched inhibitory protein), were detected. In virgin mammary glands and at the MIN‐I stage, their expression was low. In tumor samples, expression of the larger LAP protein was strongly induced, whereas the expression of the smaller LIP protein was no longer detectable. There was no significant difference in the expression of the 32 kDa LAP protein. We then analyzed CEBP at the cellular level by immunohistochemistry. CEBP‐α was detected in the nuclei of normal epithelial cells of the ducts and of fat cells (Fig. 3e, normal). In contrast, no CEBP‐α was detectable in invasive mammary tumors (Fig. 3e, tumor). This expression pattern inversely correlated with the expression of the β‐Gal reporter protein (Fig. 3e) and lacZ mRNA (Figs. 2c and 2d). Immunohistochemical analysis of CEBP‐β expression revealed no significant difference between virgin/MIN probes and the tumor probes (data not shown). This is likely due to the presence of the smaller, 20 kDa CEBP‐β (LIP) protein, which is expressed only in nontumor tissues (see above, Fig. 3d) and is recognized by the same antibody, the presence of the 32 kDa CEBP‐β LAP isoform, which is expressed both in normal and tumor tissues and the background expression of the 35 kDa CEBP‐β LAP isoform in nontumor tissues.
Ectopic expression of CEBP‐α and CEBP‐β regulates endogenous TERT expression
To analyze whether CEBP‐α represses TERT expression, we transfected MCF7 human mammary carcinoma cells (Figs. 4a and 4b) and 4T1 mouse mammary carcinoma cells (Figs. 4c and 4d) using a CEBP‐α expression vector. We found that CEBP‐α substantially repressed endogenous TERT expression, both in human and mouse mammary tumor cells (Figs. 4a and 4c). As a control, we tested the expression of the established CEBP‐α target gene lactoferrin19 and found strong activation of this gene, both in human and mouse cell lines (Figs. 4b and 4d). On the other hand, CEBP‐β overexpression resulted in significant induction of both, TERT and lactoferrin gene expression in MCF7 cells (Figs. 4e and 4f).
Regulation of endogenous TERT gene expression by ectopic CEBP‐α and CEBP‐β in human and mouse cell lines. MCF7 human mammary carcinoma cells and 4T1 mouse mammary carcinoma cells were transfected with a CEBP‐α expression vector or the control pEGFP‐C3 vector. CEBP‐α expression vector also expresses the GFP reporter gene following an IRES element. Forty‐eight hours after transfection, transfected cells were sorted, RNA isolated and expression of hTERT, mTert or human and mouse Ltf genes were determined by real‐time RT‐PCR (a–d). CEBP‐β expression vector or the control pBABE‐Puro vector was transfected into MCF7 cells, pools of stably expressing cells were selected by puromycin for 2 weeks, and hTERT or Ltf expressions were determined by real‐time RT‐PCR (e,f). Results are from two independent experiments that were performed in duplicate. The PCR assay was always in triplicate.
shRNA‐mediated knockdown of CEBP‐α results in TERT gene expression in primary human mammary epithelial cells with compromised tumor suppressor checkpoints
We next tested whether CEBP‐α‐mediated repression of TERT gene can be reverted in primary mammary epithelial cells (MEC) in the absence of this factor. For this purpose, we used a human primary MEC (huMEC) with intact and compromised p53 and Rb tumor suppressor factors, resembling WAP‐T mouse MPCs before and after induction of tumorigenesis. CEBP‐α knockdown efficiency was confirmed by immunoblot analysis (Fig. 5a, top panel). Interestingly, reactivation of endogenous telomerase activity (Fig. 5b) and hTERT mRNA levels could be observed only in the background of nonfunctional p53 and pRB (Fig. 5c). Of note, reactivation of TERT expression correlated with an upregulation of endogenous CEBP‐β (Fig. 5a, middle panel). On the other hand, induction of CEBP‐β expression is not directly dependent on the presence of LT. These results indicate the need for CEBP‐β and further genetic factors that co‐operate with the CEBPs in TERT gene regulation. In addition, in line with the in vivo results (see Figs. 1b and 2c), the expression of LT has no direct influence on hTERT expression.
Knockdown of CEBP‐α in human MECs results in the reactivaton of hTERT gene expression and telomerase activity. (a) Immunoblot experiments showing the knockdown efficiency of CEBP‐α in primary human mammary epithelial cells (huMEC). A shRNA control vector (shCON) or a shRNA vector directed against CEBP‐α was virally introduced into huMEC cells with intact p53 and Rb (left) or into huMEC cells containing the SV40‐LT (huMEC‐LT). After 1 week of selection with puromycin, cells were expanded, and RNA and protein were isolated for the subsequent experiments. As additional controls, huMEC cells and huMEC‐LT cells were included. (b) TRAP assay and (c) quantitative RT‐PCR showing the reactivation of telomerase activity and hTERT mRNA expression, respectively. The hTERT mRNA levels are relative to GAPDH mRNA levels.
CEBP‐α and CEBP‐β bind and regulate TERT gene promoter
To analyze the presumed association of CEBP factors with the hTERT‐promoter, we performed promoter binding studies by chromatin immunoprecipitation (ChIP) and electrophoretic mobility shift assays (EMSA). We first performed ChIP experiments with MCF7 cells, which had been transiently transfected with CEBP‐α and CEBP‐β expression vectors (Fig. 6a). These ChIP experiments clearly demonstrated the binding of the CEBP factors to two different TERT‐promoter regions (hTERTp‐1 and hTERTp‐2, respectively), encompassing the putative CEBP‐binding sites. The human lactoferrin promoter was used as a positive control as described by Khanna‐Gupta et al.,19 and the c‐myc and H23/27 promoter regions were used as negative controls. These results could be confirmed with endogenous CEBP‐β in MCF7 cells, whereas no or only weak CEBP‐α binding could be observed to the hTERTp‐1 promoter region (Fig. 6b). On the contrary, we could observe binding of CEBP‐α to hTERTp‐1 and hTERTp‐2 regions in TERT/telomerase‐negative huMEC, which express high levels of endogenous CEBP‐α but no/low CEBP‐β (Figs. 5 and 6c). Next, we tested the association of CEBP‐α and CEBP‐β with the human TERT promoter in normal and tumor tissues derived from hTERTp‐lacZxWAP‐T bitransgenic mice (Fig. 6d). CEBP‐α binds to both hTERT‐promoter regions only in normal mouse mammary tissue, whereas strong CEBP‐β association with human TERT‐promoter was observed in tumor samples. As a negative control, we have used a promoter region of the human GAPDH gene (GAPDH‐142).
CEBP‐α and CEBP‐β bind directly to CEBP‐binding sites present in the hTERT‐promoter region and regulate hTERT‐promoter activity. (a–d) Chromatin immunoprecipitation (ChIP), (e–g) EMSA experiments and (h, i) reporter gene analyses. (a) ChIP was performed with MCF7 human mammary cancer cells following transient transfection with CEBP‐α and CEBP‐β expression vectors. Two different hTert‐promoter regions were amplified after ChIP showing the binding of CEBP factors. Human lactoferrin promoter was used as positive control and (b) binding of endogenous CEBP‐β to human TERT‐promoter region 1 (hTERTp‐1) in MCF7 cells and (c) binding of endogenous CEBP‐α to hTERTp‐1 and to hTERTp‐2 in huMEC. (d) Binding of endogenous CEBP‐α and of endogenous CEBP‐β to hTERTp‐1 and to hTERTp‐2 in hTERTp‐lacZxWAP‐T normal mammary gland (left) and in hTERTp‐lacZxWAP‐T tumor tissue (right), respectively. Undiluted input DNA was used as control. (e) CEBP factors specifically bind to wild‐type hTERT CEBP‐binding sites. HEK (293T) cells were transfected by pEGFP (mock), CEBP‐α or CEBP‐β expression vectors, and nuclear extracts were used for the EMSA in combination with oligonucleotides containing either a consensus or hTERT promoter wild type or hTERT promoter mutant CEBP‐binding sites (CON, CEBP‐BS‐1, CEBP‐BS‐2 or mutant CEBP‐BS‐1). (f, g) Supershift experiments indicate CEBP‐binding specificity. Experiments were performed with the CEBP‐BS‐1, and nuclear extracts prepared from HEK (293T) cells transfected with either pEGFP (mock), CEBP‐α or CEBP‐β expression vectors. Arrows indicate specific CEBP–DNA complexes (straight line) or antibody–CEBP–DNA supershift complexes (dashed line). (h) Schematic illustration of a 3‐kb human TERT‐promoter‐luciferase reporter constructs showing the wild‐type and mutated CEBP‐binding sites. (i) Reporter gene analysis in MDA‐MB231 human mammary carcinoma cells shows the changes in reporter protein activity in dependence of the CEBP‐binding site in combination with CEBP‐α or CEBP‐β (see Material and Methods for detail). The value for the firefly luciferase reporter activity was corrected to the Renilla luciferase activity. Experiments were performed in triplicates.
EMSA experiments were performed to examine whether CEBP‐α and ‐β bind to the putative CEBP‐binding sites within the hTERTp‐1 and hTERTp‐2 promoter regions (Fig. 6). Nuclear extracts prepared from CEBP‐α and CEBP‐β transfected 293T cells were used together with oligonucleotides containing either a consensus CEBP‐binding site or the putative CEBP‐binding sites derived from the human TERT promoter or a mutated binding site, which contained two base pair exchanges compared to binding site I (hTERTp‐1; Fig. 6e). Both factors bound the oligonucleotides with consensus or wild‐type CEBP‐binding sites from the TERT promoter, whereas no binding was observed with the mutant oligonucleotide. In addition, we performed supershift experiments (Figs. 6f and 6g). We observed a supershift with the CEBP‐α‐specific antibody, but not with CEBP‐β‐specific antibody with nuclear extracts containing CEBP‐α (Fig. 6f), whereas the opposite was true with nuclear extracts containing CEBP‐β (Fig. 6g). Together, these experiments clearly demonstrate the binding specificity of the CEBP factors to the putative CEBP‐binding sites present within the hTERT promoter.
We next wanted to know whether the binding of the CEBP factors to TERT promoter directly influences TERT‐promoter activity. For this purpose, reporter gene assays were performed with reporter vectors containing base exchanges at each of the CEBP‐binding sites or at both sites (Figs. 6h and 6i). Mutation at each of the single CEBP‐binding site resulted in a slight decrease of the reporter gene activity, probably due to the endogenous CEBP‐β protein. An additive effect was observed when both sites were mutated. Cotransfection of CEBP‐α resulted in approximately three to fourfold reduction of the reporter activity, which was abolished to near wild‐type levels when both sites were mutated. On the other side, cotransfection of CEBP‐β resulted in approximately threefold increase of the reporter gene expression, which was strongly reduced if both CEBP‐binding sites were mutated. Reporter constructs with single‐site mutations showed intermediate repression or activation with CEBP‐α or CEBP‐β, respectively.
Discussion
The present results indicate that telomerase activity is associated with the acquisition of invasive malignancy in WAP‐T mouse mammary cancer involving CEBP‐α and CEBP‐β transcription factors in TERT‐promoter regulation. Although the reactivation of telomerase is one of the critical steps in the tumorigenic conversion of normal cells to tumor cells, the time point of telomerase activation during tumorigenesis is still under debate and may be specific for each tumor type. In mammary tissue, nearly, all invasive cancers are telomerase‐positive. On the other hand, reports on telomerase activity in human DCIS are controversial.7-10 Our studies are in line with the observation that telomerase activity is a late event during mammary carcinogenesis, associated with the invasion of the basement membrane.7 In this regard, one should also consider that DCIS is usually diagnosed at later stages of tumor progression in humans. These samples thus may contain regions with and without invasiveness, whereas we analyzed MIN‐II (DCIS) samples without any sign of invasiveness. This may also explain the heterogeneity in the detection of telomerase activity in human DCIS, whereas nearly all invasive human mammary carcinomas are telomerase‐positive.
It is conceivable that mammary tumors could arise from rare, telomerase positive cells, which are difficult to detect in tissue samples by RT‐PCR. Therefore, we have also performed immunohistochemistry at all stages of tumorigenesis (not shown). However, we did not observe β‐Gal‐positive cells at the earlier stages, that is, MIN‐I or MIN‐II, although this result does not exclude the presence of β‐Gal (i.e., telomerase)‐positive cells, potentially serving as stem/progenitor cells. One possibility is that telomerase/hTERT‐promoter activity is weak and thus below detection limit in these cells. Further studies will be necessary to purify and enrich stem/progenitor cells from normal mammary tissue and from early stages of mammary carcinogenesis to analyze telomerase activity.
The WAP‐T mouse model was shown to reflect histological14 and molecular15 characteristics of human mammary cancer. In the present study, we could show that the WAP‐T mouse model, in combination with in vitro analyses, was useful to unveil regulatory mechanisms, which are involved TERT gene regulation in normal and tumor breast. To date, there is no experimental data on the regulatory mechanisms involved in TERT gene expression in the mammary tissue. Several factors have been described for promoter activation of the human TERT gene, such as c‐myc, E2Fs, estrogen and progesterone receptors, Her2/Neu signal pathway (ER81).20-25 Some of these factors are known to be involved in mammary carcinogenesis and/or are increased or amplified in human breast cancer.26-29
Here, we show for the first time that the CEBP‐α and CEBP‐β transcription factors are involved in TERT gene regulation in the mammary tissue. CEBP‐α functions as a repressor of TERT gene expression in normal mammary tissue, whereas CEBP‐β seems to act as a positive regulator of TERT gene expression in concert with additional factors. CEBP‐α and CEBP‐β were previously described to be involved in breast cancer. Ectopic expression of wild‐type CEBP‐α induces cell‐cycle arrest in a number of human breast cancer cell lines, accompanied by reduced anchorage‐independent cell growth.30 This is compatible with its presumed tumor suppressor role in cancer, including solid tumors.31 The fact that CEBP‐α acts as a transcriptional activator32, 33 and transcriptional repressor, this study and Ref. 34 indicate that it requires co‐operation partners for these antagonistic functions. In fact, Sp1 and E2F factors have been described as co‐operating factors for the action of CEBP‐α.35, 36 Interestingly, Sp1 and E2F factors are involved in TERT‐promoter regulation.37, 38 In contrast to CEBP‐α, CEBP‐β seems to act in promoting tumorigenesis, and its overexpression has been observed in human cancers, including breast tumors.39-41 Furthermore, certain isoforms of CEBP‐β (LAP‐2) can transform MPCs and induce mesenchymal–epithelial transition.42 In our study, ectopic CEBP‐β activated endogenous TERT gene expression in MCF7 tumor‐cell line, but not in huMEC with or without LT (unpublished observation, C.G.) despite strong TERT‐promoter binding, both in vivo and in vitro. This supports the idea that CEBP‐β, similar to CEBP‐α, requires additional genetic alterations and/or co‐operating partners for TERT regulation.
In a preliminary analysis, we have found ∼ 10‐fold increased expression of E2F1, E2F2, E2F7 and E2F8 factors in the WAP‐T mouse mammary tissues (C.G. unpublished results). We have recently shown that E2F‐Rb family proteins are involved in TERT‐promoter regulation during liver regeneration.38 Interestingly, CEBP factors also play an important role in liver regeneration and in liver progenitor cells.43-45 In fact, specific combinations of these regulatory factors might contribute to the species‐specific regulation of telomerase expression, as exemplified by high‐level expression of telomerase in mouse liver, while its expression in human liver is low/absent. Recent studies also support an essential role of CEBP‐α and CEBP‐β in other tissue stem cells.46, 47 One may speculate that these factors also contribute to the regulation of telomerase in progenitor and stem cells. Moreover, it would be important to investigate whether some of the E2Fs also require CEBP factors in liver progenitor cells or in other tissue stem cells for TERT regulation.
Material and Methods
Mice
hTERTp‐lacZ13 and WAP‐T14 mice have been described previously. To analyze the hTERT‐promoter activity in a genetically pure mouse background, transgenic animals showing lacZ gene expression were back‐crossed for more than 10 generations with wild‐type BALB/c mice. These mice were then mated with BALB/c WAP‐T mice to generate hTERTp‐lacZxWAP‐T bitransgenic mice. Animal studies were performed in accordance with the guidelines of the authority of animal use and approved by the local authorities (Behörde für Soziales, Familie, Gesundheit und Verbraucherschutz, Hamburg, Germany).
Cell culture, plasmids and transfection
All cell lines were from ATCC. Hek‐293, MCF7, MDA MB 231 and 4T1 cells were cultured in DMEM‐containing 10% fetal bovine serum and 100 U/ml penicillin and 100 μg/ml streptomycin. Primary human MPCs were purchased from Invitrogen (HuMEC, Cat. No. A10565) and maintained in HuMEC Ready Medium (Cat. No. 12752‐010). HuMEC‐LT cells were generated by retroviral transduction of the SV40‐LT antigen (vector obtained from Addgene: Plasmid 1780: pBABE‐neo largeT cDNA48). CEBP‐α (human) expression vector was a gift of Carol Stocking (Heinrich‐Pette‐Institut, Hamburg, Germany). CEBP‐β (rat) expression vector and shRNA vectors for CEBP‐α and the control shRNA were received from Cornelis Calkhoven (FLI, Jena, Germany). MCF7 and 4T1 cells were transfected with the empty vector, and the CEBP‐α expression vector using the LF2000 reagent and 2 days after transfection GFP‐positive (i.e., CEBP‐α expressing) cells were sorted using the ARIA cell sorter. CEBP‐β vector was transfected into MCF7 cells, and stably expressing cells were selected by puromycin.
Reporter gene analyzes
For the reporter gene analyzes, a 3‐kbp human Tert‐promoter luciferase reporter construct was used. CEBP‐binding site mutations were generated by the QuickChange™ Site‐Directed Mutagenesis Kit (Stratagene, La Jolla, CA). Luciferase activity was determined using the Dual Luciferase Assay Kit (Promega, Madison, WI). Transfections were performed in triplicates in 24‐well plates by Lipofectamin 2000™ according to the recommendations of the supplier (Life Technologies Corporation, Carlsbad, CA). Luciferase reporter gene activity was corrected to Renilla luciferase activity. For each experiment, 300 ng reporter vector, 100 ng renilla luciferase vector and 600 ng expression vector (pEGFP, CEBP‐α or CEBP‐β) were cotransfected.
Retroviral and lentiviral transduction
HEK cells were transfected with gag/pol, env and expression or shRNA vectors. Virus supernatant was filtered and used for the transduction of HuMEC cells according to standard protocols. Two days after transduction, LT‐expressing cells were selected in the presence of 0.5 mg/ml G418, and shRNA‐expressing cells were selected in the presence of 0.5 μg/ml puromycin. In HuMEC cells with LT and sh‐CEBP‐α vectors, the latter was introduced to the HuMEC‐LT cells already containing the LT vector.
SDS–PAGE and Immunobloting
For protein analysis, cells were lysed with 4 × Laemmli buffer, resolved by SDS–PAGE and then transferred to nitrocellulose membranes as described previously.38 The same CEBP‐α and CEBP‐β antibodies were used for immunoblotting, immunohistohemistry and ChIP experiments (for C/EBP‐alpha, sc‐61; for C/EBP‐beta, sc‐150; Santa Cruz Biotechnologies, Santa Cruz, CA). A monoclonal GAPDH antibody was used to control protein loading (sc‐32233, Santa Cruz Biotechnologies).
Immunohistochemistry
Immunohistochemistry was performed as described previously for β‐Gal protein13 and for LT.14 For CEBP‐α, a 1:50 dilution of the sc‐61 antibody was used. Otherwise, the staining procedure was the same as for the β‐Gal protein.
RNA isolation, reverse transcription, semiquatitative and qRT‐PCR
RNA isolation, first‐strand cDNA synthesis and semiquantitative polymerase chain reaction (PCR) were performed as described earlier.13 For quantitative real‐time PCR (qRT‐PCR), cDNA was diluted to 1/10, and 2 μl from this dilution was used in a total reaction volume of 10 μl. qPCRs were performed using iTaq SYBR Green supermix with Rox (BioRad) reagent in AB7300 (Applied Biosystems).
Extract preparation and gel‐based nonradioactive TRAP assay
Telomerase activity was determined by the Telomeric Repeat Amplification Protocol (TRAP) with the TRAPeze kit (Serologicals Corp., Norcross, GA) according to the recommendations of the supplier. TRAP products were separated in a 12.5% nondenaturing PAA‐Gel and visualized by 20‐min incubation in a 1:10,000 dilution of SYBR‐Gold Nucleic Acid Gel Staining solution (Invitrogen). Telomerase activity was allowed to occur for 30 min at 30°C followed by the heat‐inactivation of telomerase for 2 min at 96°C, and PCR amplification of the telomerase elongated TS‐primer products by a three‐step PCR reaction: 30 sec at 94°C, 30 sec at 59°C and 1 min at 72°C for 30 cycles. Relative telomerase activity was then determined in comparison with 3‐[(3‐Cholamidopropyl)dimethylammonio]‐1‐propanesulfonate (CHAPS)‐negative and HeLa‐positive controls.
Real‐time quantitative TRAP assay
Real‐time quantitative Telomeric Repeat Amplification Protocol RT‐qTRAP) assay was established in our laboratory, based on published reports, with some modifications. The qTRAP assay was performed with the ABI Prism 7300 thermal cycler (Applied Biosystems, Foster City, CA) in 15‐μl reaction mixture containing 1:50,000 dilution of SYBR‐Gold, 10 pmol TS‐primer (5′‐ATCCGTCGAGCAGAGTT‐3′), 1 pmol ACX‐primer (5′‐GCGCGG[CTTACC]3CTAACC‐3′), 0.2 mM dNTPs, 1.5 μl 10× TRAP‐buffer (200 mM Tris–HCl, pH 8.3, 15 mM MgCl2, 630 mM KCl, 0.5% Tween, 10 mM EGTA), 0.1 μl Taq‐polymerase [5 U/μl] and 100 ng CHAPS tissue extract. Serial dilution of HeLa cells (equivalent to 1,000, 500, 250, 100, 50 and 10 cells, respectively) was used as telomerase‐positive controls and their heat‐treated samples as negative controls. CHAPS was used as protein‐free negative control. The TRAP reaction was the same as mentioned earlier, except that the amplification was done for 40 cycles.
Chromatin immunoprecipitation
ChIP experiments were performed as described by Weinmann et al.49 Briefly, cells or mammary tissue were crosslinked by incubating with 1% formaldehyde for 10 min at RT. Crosslinking was stopped by adding glycine to a final concentration of 0.125 M. After 5 min incubation, cells were washed twice with phosphate‐buffered saline, and nuclei were released by adding cell‐lysis buffer (5 mM HEPES, pH 8.5; 85 mM KCl; 0.5% NP‐40 and protease inhibitors). After 10 min incubation and centrifugation, chromatin was released with nuclei‐lysis buffer (50 mM Tris–Cl, pH 8.1; 10 mM EDTA, pH 8.0; 1% SDS, protease inhibitors). Chromatin was fragmented by sonicating 6× for 30 sec using a Bandelin Sonopuls sonicator. Chromatin fragments were diluted with IP buffer (0.01% SDS; 1.1% Triton X100; 1.2 mM EDTA, pH 8.0; 16.7 mM Tris–Cl, pH 8.1; 167 mM NaCl, protease inhibitors) and precleared for 30 min with Protein A‐agarose and protease inhibitors to reduce nonspecific background. Chromatin was incubated at 4°C overnight with antibodies against CEBP‐α or CEBP‐β. IgG was used as serum control. Antibody‐coupled chromatin was bound with protein A‐agarose beads. Beads were washed with low‐salt buffer (0.1% SDS; 20 mM Tris–Cl, pH 8.0; 1% Triton; 150 mM NaCl; 2 mM EDTA, pH 8.1) followed by washing in high‐salt buffer (0.1% SDS; 20 mM Tris–Cl, pH 8.0; 1% Triton; 500 mM NaCl; 2 mM EDTA, pH 8.1) and LiCl buffer (250 mM LiCl; 10 mM Tris–Cl, pH 8.0; 1% NP‐40; 1% deoxycholic acid; 1 mM EDTA, pH 8.1) and diluted in TE‐buffer (10 mM Tris–Cl, pH 8.0; 1 mM EDTA, pH 8.1). Digestion of crosslinked proteins was performed by incubation with Proteinase K for 2 h at 65°C. DNA was extracted using QIAquick Nucleotide Removal Kit (Qiagen GmbH) and analyzed by PCR.
The hTERTp‐1 promoter fragment (232 bp) was amplified by primers hTERTp‐1413F (5′‐GATCACTAAGGGGATTTCTAGAAG) and hTERTp‐1181R (5′‐CTTGCAGGGATGCTGTAGCTGAGG). The hTERTp‐2 promoter fragment (210 bp) was amplified by primers hTERTp‐738F (5′‐CCTGCAAAGAGAAATGACGGGC) and hTERTp‐528R (5′‐GCCTGATCCGGAGACCCAGGGC). PCR conditions were as follows: 1 cycle at 94°C for 5 min; 36 cycles at 94°C for 15 sec and at 60°C for 1 min.
Electrophoretic mobility shift assay
Electrophoretic mobility shift assay (EMSA) was performed essentially as described in Calkhoven et al.50 The following 27 or 25‐bp double‐strand oligonucletides from the human TERT‐promoter with putative CEBP‐binding sites were used for EMSA: CEBP‐BS‐1 (−1.3 kbp) 5′‐TGCAGATTGAAGCAATTTCCTCCTGCA (top strand), CEBP‐BS‐2 (−0.65 kbp) 5′‐TGCAGATTGGGAGCAATGCGCTGCA (top strand) and mutant CEBP‐BS‐1 (−1.3 kbp) 5′‐TGCAGATTGAAGtgATTTCCTCCTGCA (top strand). Briefly, protein–DNA complexes were allowed to form for 30 min at RT. Reaction was performed in 15 μl with 20‐μg cellular extract in binding buffer containing 20 mM HEPES pH 7.9, 60 mM KCL, 1 mM EDTA, 2 mM DTT, 2 mM spermidine, 10% glycerol, 2 μg poly dI–dC and 0.5 pmol radioactively labeled oligonucleotides. For the competitive EMSA, 100‐fold excess of unlabeled DNA fragments was added to the binding reaction with the labeled probe. For the supershift EMSA, 5 μg of the specific antibody (for C/EBP‐alpha, sc‐61X; for C/EBP‐beta, sc‐150X; Santa Cruz Biotechnologies) was preincubated on ice with nuclear extracts followed by the addition of the radioactive probes and a further incubation on ice for 30 min. The binding reaction was separated in 6% PAA–Gel for 3 hr at 200 V with 0.5× TBE.
Acknowledgements
We thank Dr. Carol Stocking and Dr. Cornelis Calkhoven for providing vectors. We also thank to Julia Schröder for excellent technical help.
Number of times cited: 8
- Mukesh Kumar, Andre Lechel and Çagatay Güneş, Telomerase: The Devil Inside, Genes, 7, 12, (43), (2016).
- Jitendra K Meena, Aurora Cerutti, Christine Beichler, Yohei Morita, Christopher Bruhn, Mukesh Kumar, Johann M Kraus, Michael R Speicher, Zhao‐Qi Wang, Hans A Kestler, Fabrizio d'Adda di Fagagna, Cagatay Günes and Karl Lenhard Rudolph, Telomerase abrogates aneuploidy‐induced telomere replication stress, senescence and cell depletion, The EMBO Journal, 34, 10, (1371-1384), (2015).
- H Wu, J Zheng, J Deng, L Zhang, N Li, W Li, F Li, J Lu and Y Zhou, LincRNA-uc002yug.2 involves in alternative splicing of RUNX1 and serves as a predictor for esophageal cancer and prognosis, Oncogene, 34, 36, (4723), (2015).
- Roula Tahtouh, Anne-Sophie Azzi, Nada Alaaeddine, Soulaima Chamat, Hasnaa Bouharoun-Tayoun, Layal Wardi, Issam Raad, Riad Sarkis, Najibe Abou Antoun, George Hilal and Diego Calvisi, Telomerase Inhibition Decreases Alpha-Fetoprotein Expression and Secretion by Hepatocellular Carcinoma Cell Lines: In Vitro and In Vivo Study, PLOS ONE, 10, 3, (e0119512), (2015).
- Mingyan Xu, Xihe Chen, Hao Yin, Liqin Yin, Fan Liu, Yucai Fu, Jiangwu Yao, Xiaoling Deng and Andre van Wijnen, Cloning and Characterization of the Human Integrin β6 Gene Promoter, PLOS ONE, 10, 3, (e0121439), (2015).
- Vegard Eldholm, Aage Haugen and Shanbeh Zienolddiny, CTCF mediates the TERT enhancer–promoter interactions in lung cancer cells: Identification of a novel enhancer region involved in the regulation of TERT gene, International Journal of Cancer, 134, 10, (2305-2313), (2013).
- Baiqun Wu, Xiaoming Ji, Ruhui Han, Lei Han, Ting Wang, Jingjin Yang, Baoli Zhu and Chunhui Ni, GITR promoter polymorphism contributes to risk of coal workers’ pneumoconiosis: A case–control study from China, Immunology Letters, 162, 2, (210), (2014).
- Benjamin Otto, Thomas Streichert, Florian Wegwitz, Heidrun Gevensleben, Kristin Klätschke, Christoph Wagener, Wolfgang Deppert and Genrich V. Tolstonog, Transcription factors link mouse WAP‐T mammary tumors with human breast cancer, International Journal of Cancer, 132, 6, (1311-1322), (2012).
