Research Article

Wnt/β-Catenin Signaling Regulates Telomerase in Stem Cells and Cancer Cells

See allHide authors and affiliations

Science  22 Jun 2012:
Vol. 336, Issue 6088, pp. 1549-1554
DOI: 10.1126/science.1218370

From Wnt Signals to Telomerase Activity

Telomerase activity is associated with stem cell renewal and cancers, whereas a decrease in telomerase activity is seen during cell differentiation and senescence. Wnt/β-catenin signaling is also a critical regulator of stem cells, and deregulation of the pathway is associated with cancer. Now, Hoffmeyer et al. (p. 1549; see the Perspective by Greider) have found a link between these two pathways. In embryonic stem cells, β-catenin was able to regulate telomerase expression and activity directly. Similar observations were obtained in adult stem cells, a model of intestinal tumors, and human cancer cells.


Telomerase activity controls telomere length and plays a pivotal role in stem cells, aging, and cancer. Here, we report a molecular link between Wnt/β-catenin signaling and the expression of the telomerase subunit Tert. β-Catenin–deficient mouse embryonic stem (ES) cells have short telomeres; conversely, ES cell expressing an activated form of β-catenin (β-catΔEx3/+) have long telomeres. We show that β-catenin regulates Tert expression through the interaction with Klf4, a core component of the pluripotency transcriptional network. β-Catenin binds to the Tert promoter in a mouse intestinal tumor model and in human carcinoma cells. We uncover a previously unknown link between the stem cell and oncogenic potential whereby β-catenin regulates Tert expression, and thereby telomere length, which could be critical in human regenerative therapy and cancer.

Telomeres are specialized genomic structures that cap linear chromosomes and are essential for genome stability (1). Telomere length is controlled by the telomerase complex comprising an enzymatic subunit, TERT, and a RNA component, Terc (2). Embryonic and other stem cells have long telomeres, which become shorter during differentiation or aging but are stabilized again in tumorigenesis (3). The canonical Wnt signaling pathway plays a major role in regulating pluripotency in embryonic stem (ES) and adult stem cells from various tissues (4). β-Catenin is a central component of the Wnt pathway and forms a complex with members of the TCF family of transcription factors in the nucleus to control the transcription of target genes. Dysregulation of this pathway is frequently observed in human cancer (5). Here we show that TERT is directly regulated by β-catenin. Our results underline the cooperation between Wnt/β-catenin signaling and telomerase in the control of stem cell renewal.

β-Catenin regulates Tert expression in mouse ES cells. Comparing the expression profiles of wild-type and β-catenin–deficient (β-cat−/−) ES cells, we found that Tert, but not Terc, mRNA was significantly reduced in the absence of β-catenin (Fig. 1A and fig. S1A). ES cells harboring a stabilized active form of β-catenin (β-catΔEx3/+) had higher levels of Tert mRNA compared to wild-type cells (Fig. 1A). Those differences in Tert mRNA expression were reflected by TERT protein amounts in Western blot analysis (fig. S1C). Thus, altered levels of β-catenin affect Tert expression and may lead to differences in telomerase activity. Concordantly, telomerase activity was significantly increased in β-catΔEx3/+ compared to the wild type and was reduced in β-cat−/− ES cells (Fig. 1B). Stimulation of wild-type ES cells with Wnt3a led to an increase in Tert expression (Fig. 1C). This supports recent findings that Wnt signaling may regulate Tert protein by sequestration of glycogen synthase kinase 3 (6). Knockdown of β-catenin by small interfering RNA (siRNA) reduced Tert expression and telomerase activity (Fig. 1D and fig. S1, D and E). Changes in Tert expression and telomerase activity resulted in different telomere lengths, with β-cat−/− cells having shortened telomeres (Fig. 1E). Telomeres in wild-type ES cells were on average 50 kb long; in β-catΔEx3/+, 75 kb; and in β-cat−/− ES cells, 24 kb.

Fig. 1

β-Catenin regulates Tert mRNA expression, telomerase activity, and telomere length in genetically modified ES cells. (A) Quantitative polymerase chain reaction analysis of Tert mRNA. (B) Telomerase activity. (C and D) In wild-type (wt) ES cells Wnt3a induces Tert expression, whereas knockdown of β-catenin by siRNA reduces Tert expression and activity (fig. S1). (E) Telomere length determined by quantitative fluorescence in situ hybridization analysis and TFL-Telo software in comparison to reference cell lines. (n = 7, *P < 0.05).

Tert is a direct target of β-catenin. To gain insight into Tert transcriptional regulation by β-catenin, we performed luciferase reporter assays and chromatin immunoprecipitation (ChIP) experiments. Tert promoter fragments of 2.9 kb and 300 base pairs (bp) were equally active in wild-type ES cells (fig. S1F). The 300-bp fragment harbors binding sites for TCF and Klf4 (fig. S3C and S4G). β-Catenin was detected at the transcriptional start site (TSS) of Tert in wild-type ES cells by ChIP (Fig. 2A), which was further enhanced by the addition of Wnt3a (Fig. 2B). Binding of β-catenin to the TSS of Tert was also increased in β-catΔEx3/+ cells, whereas no β-catenin was immunoprecipitated in β-cat−/− ES cells (Fig. 2A). The canonical β-catenin partners TCF3 and TCF4 were not detected at the Tert locus; however, TCF1 was enriched close to the TSS, even in β-cat−/− cells (figs. S2, A and B, and S3A). We postulate that TCF1 may act as a transcriptional repressor of Tert, as knockdown of TCF1 by siRNA and mutational analysis of the 300-bp promoter fragment suggest (fig. S3, B and D). In addition, no Oct3/4 binding at the Tert locus was observed (fig. S2C). c-Myc, a regulator of Tert, bound equally to the Tert promoter region in wild-type, β-cat−/−, and β-catΔEx3/+ cells, although c-Myc protein was reduced in β-cat−/− cells (fig. S2, D and E). Klf4 was of particular interest because of the conserved Klf4-binding site located at the TSS of the Tert promoter and because Klf4 contributes to the maintenance of telomerase activity in human cells (7). β-Catenin and Klf4 coimmunoprecipitate in wild-type ES cells (fig. S1G). Klf4 was identified at the Tert promoter by ChIP (Fig. 2C). Sequential ChIP revealed a Klf4/β-catenin complex at the Tert promoter (Fig. 2D). To study the relationship between β-catenin and Klf4, we inhibited the expression of Klf4 in wild-type ES cells with siRNAs with or without Wnt3a stimulation (Fig. 2E and fig. S4). The accumulation of β-catenin on the Tert promoter upon Wnt3a stimulation was severely reduced after Klf4 knockdown. Therefore, Klf4 is required for β-catenin to localize to the Tert promoter. However, Klf4 alone was insufficient to drive Tert expression, as Klf4 was bound to the Tert promoter in β-cat−/− cells, where Tert expression is low (Fig. 2C). Furthermore, whereas reexpression of β-catenin in β-cat−/− cells led to an increase in Tert mRNA level, overexpression of Klf4 did not (Fig. 2F). Therefore, recruitment of β-catenin is necessary for Tert transcription in mouse ES cells, and Klf4 promotes this binding.

Fig. 2

β-Catenin and Klf4 co-regulate the Tert promoter in ES cells. (A) In ChIP, β-catenin is bound at the transcriptional start site (TSS) of Tert in wt and β-catΔEx3/+ but not in β-cat−/− ES cells. (B) Enrichment of β-catenin at the TSS of Tert after stimulation with Wnt3a. (C) In ChIP, Klf4 is localized at the TSS of Tert in wt, β-catΔEx3/+, and β-cat−/− cells. (D) Re-ChIP, anti-Klf4 followed by anti–β-catenin, demonstrating that both form a complex at the Tert promoter. (E) Wild-type ES cells transfected with siRNA for Klf4 (48 hours) were cultivated with (+) or without (–) Wnt3a (12 hours). β-Catenin binding to the Tert promoter was analyzed by ChIP (fig. S4). (F) β-cat−/− cells were stably transfected with β-catenin or Klf4, and two independent clones were analyzed. Only overexpression of β-catenin resulted in reexpression of Tert mRNA. Axin2 was used as a positive control region and hypoxanthine phosphoribosyltransferase (Hprt) as a negative control region in ChIP experiments. (*P < 0.05, table S3).

β-Catenin regulates Tert promoter activity. Supporting the role of β-catenin in Tert promoter activation, RNA polymerase II (Pol II), Pol II Ser5p (active Pol II), and the active trimethylated lysine-4 on histone-3 (H3K4me3) were detected at the Tert promoter in wild-type and β-catΔEx3/+, but not in β-cat−/−, cells (Fig. 3, A and B, and fig. S5B). Both Pol II and H3K4me3 were reestablished at the Tert promoter in β-cat−/− cells upon transfection with constitutively active β-catenin (Fig. 3C and fig. S6A). These results demonstrate that β-catenin is required for Tert promoter activation. We identified two members of the trithorax group (TrxG) proteins, Ash2l and Setd1a (8), the latter of which exhibited histone methyltransferase (HMT) activity at the TSS of the Tert promoter in ES cells. The localization of Ash2l and Setd1a at the Tert promoter depended on β-catenin, was not detected in β-cat−/− ES cells, and was enhanced in β-catΔEx3/+ ES cells (Fig. 3, E and F, and figs. S5 and 6). Transfection experiments in human embryonic kidney 293 (HEK293) cells followed by coimmunoprecipitation revealed an association of β-catenin with Ash2l and Setd1a (Fig. 3D). These data suggest that β-catenin actively recruits HMTs to regulate the chromatin modifications required for the initiation of Tert transcription.

Fig. 3

β-Catenin at the Tert promoter is associated with H3K4me3, Pol II, Ash2l, and Setd1a. (A and B) ChIP for Pol II and H3K4me3 at the Tert promoter in wt, β-cat−/−, and β-catΔEx3/+ ES cells (fig. S5). (C) ChIP for H3K4me3 and Pol II (fig. S6A) at the Tert promoter in β-cat−/− cells overexpressing β-catenin or Klf4. (D) (Upper panels) β-Catenin and Setd1a (right) and β-catenin and Ash2l (left) association in HEK293 cells, as shown by coimmunoprecipitation (co-IP). In vitro–translated Setd1a associates with glutathione S-transferase (GST)–β-catenin (lower panel). (E and F) Ash2l and Setd1a binding to the Tert promoter correlates with the β-catenin levels in wt, β-cat−/−, and β-catΔEx3/+ ES cells as assessed by ChIP. Surb7, negative control. (*P < 0.05, table S3).

β-Catenin binds to the Tert promoter in adult stem cells. Next, we examined β-catenin at the Tert promoter in adult stem cells. The crypt of the small intestine contains stem cells; the villus contains more differentiated epithelial cells (9). β-Catenin bound to the TSS of Tert was only detected in the crypt cell fraction (Fig. 4A). This was further validated by β-catenin ChIP from isolated Lgr5-positive stem cells using Lgr5::GFP (green fluorescent protein) reporter mice (Fig. 4B) and in organoids from crypt cell cultures (fig. S7D). β-Catenin binding to the Tert promoter (Fig. 4A) correlated with Pol II binding (fig. S9A) and different expression levels of Tert in the crypt versus villus fractions (Fig. 5B).

Fig. 4

Binding of β-Catenin at the Tert promoter in adult stem cells. (A) Mouse small intestines were separated into crypt and villus fractions and subjected to ChIP. β-Catenin at the TSS of Tert was detected only in the crypt cell compartment. (B) Lgr5-positive stem cells were isolated from Lgr5::GFP reporter mice by fluorescence-activated cell sorting (FACS) and subjected to ChIP. (C) The Hes5::GFP reporter mouse was used to isolate neural stem cells by FACS. ChIP for β-catenin in isolated Hes5::GFP neural stem cells (NSC) and in neurospheres (NS) infected with adeno-cre revealed binding of β-catenin at the Tert promoter, which is abolished in neurospheres when β-catenin is deleted.

Fig. 5

β-Catenin regulates Tert in hyperproliferative intestinal epithelial cells and in human cancer cell lines. (A) Hyperplastic lesions in the intestine were induced by the stabilized active form of β-catenin (Villin-cre ERT × β-catExfl3/+), which resulted in hyperproliferative cells all along the crypt-villus axis, as seen by staining for proliferating cell nuclear antigen (PCNA) (upper panels). The stabilized form of β-catenin (β-catΔEx3/+) is smaller in size and equally expressed in the crypt and villus fractions (lower panels). (B) Comparison of Tert mRNA between crypt and villus compartments of wt and β-catΔEx3/+ intestines. (C) In β-catΔEx3/+ intestines, β-catenin is bound to the TSS of Tert in crypt and villus cell compartments (*P < 0.05, table S3). (D) In human NTera2 and SW480 cells, β-catenin is bound at the TSS of hTERT, as detected by ChIP. (E) siRNA-mediated knockdown of β-catenin and Klf4 reduces expression of TERT mRNA. (F) Significant correlation between TERT and β-catenin expression in human colon cancer samples; P < 0.0001, R2 = 0.21068 (for details, see fig. S11A).

To determine whether our findings were relevant in other stem or progenitor cells, we made use of the Hes5::GFP reporter mouse (10) and analyzed primary neurospheres. In isolated Hes5::GFP neural stem cells, as well as in primary neurospheres, β-catenin was enriched at the TSS of the Tert promoter (Fig. 4C). Deletion of β-catenin by adeno-mediated cre resulted in a reduction in Tert mRNA expression and abolished β-catenin binding to the Tert promoter (Fig. 4C and fig. S10).

β-Catenin regulates Tert expression in human cancer cells. Aberrant nuclear activation of β-catenin in cells of the villus leads to increased cell proliferation and the formation of polyps and adenomatous lesions (11). We established a mouse model to analyze Tert in hyperplastic lesions in the small intestine by the conditional activation of one β-catenin gain-of-function allele (villin-creERT × β-catEx3fl/+). Stabilized β-catenin was detected in β-catΔEx3/+ crypt and villus cells (Fig. 5A). In the β-catΔEx3/+ intestine, hyperproliferative cells and induced expression of β-catenin target genes, such as those encoding c-Myc, Axin2, and CD44, were detected in the epithelia all along the crypt-villus axis (Fig. 5A and fig. S8A). Expansion of the Paneth cell–specific marker, lysozyme, and CD44 was also observed in the villi (fig. S8B). Expression of Tert (Fig. 5B), Lgr5, and Klf4 mRNAs (fig.S8A) was induced in the villus fraction of mutants. Concordantly, β-catenin and Pol II binding was increased at the TSS of Tert in villus cells (Fig. 5C and fig. S9B). These results provide strong in vivo evidence for the transcriptional regulation of Tert by β-catenin.

Finally, we examined whether the regulation of Tert expression by β-catenin may also be important in human cancer. We studied the human embryonal carcinoma cell line NTera2 and the human colorectal carcinoma cell line SW480, the latter exhibiting increased amounts of cytoplasmic and nuclear β-catenin due to mutations in adenomatous polyposis coli (APC). β-Catenin was present at the TSS of hTERT (Fig. 5D), and knockdown of β-catenin by siRNA reduced hTERT mRNA levels in both cell lines (Fig. 5E and fig. S11C).

Tert is a β-catenin target. By regulating Tert expression, β-catenin may assure the correct telomere length in stem cells, promoting their genomic stability and maintenance. Telomerase is directly involved in the regulation of Wnt/β-catenin target genes (12). Our findings indicate a regulatory loop between β-catenin and Tert expression. In mouse ES cells, we identified Klf4 as a partner of β-catenin in the regulation of Tert expression, but TCF1 may also be involved. The transcriptional regulation of Tert is very likely complex and combinatorial, and β-catenin may regulate Tert expression in other stem cell compartments in concert with other transcription factors. Clearly, in the absence of β-catenin, the Tert gene is silenced, but Tert transcription is initiated when β-catenin is recruited to the Tert promoter. We show that the initiation of Tert transcription is accompanied by H3K4 trimethylation at the promoter, and we identified the lysine methyltransferase Setd1a as an interaction partner of β-catenin. Our results suggest that β-catenin recruits HMTs to initiate Tert transcription, which supports the increasingly recognized role of β-catenin in chromatin remodeling (13).

Furthermore, we show the localization of β-catenin at the Tert promoter to adult mouse stem cells and to human cancer cell lines, supporting the notion that the regulation of Tert by β-catenin is a general biological feature. Our mouse model (villin-creERT × β-catEx3fl/+) provides in vivo evidence that the aberrant activation of β-catenin in the epithelium of the small intestine leads to Tert expression and binding of β-catenin to the Tert promoter. Exon 3 encodes phosphorylation sites important for β-catenin protein stability (14), and these sites are frequently mutated in human colorectal and other cancers (15). It is notable that in human colorectal cancers, high Tert and β-catenin expression are significantly correlated (P = 0.0001), as taken from publicly available microarray data sets (16) (Fig. 5F and fig. S11A).

Conclusion. By identifying Tert as a target gene of β-catenin, we demonstrate a link between these two key regulators in stem cell biology and cancer. From the results presented here, we propose that mutations in β-catenin can lead to an enhanced Tert expression in human cancer, which results in the stabilization of telomeres, one of the hallmarks of tumorigenesis.

Supplementary Materials

Materials and Methods

Figs. S1 to S11

Tables S1 to S3

References (1725)

References and Notes

  1. Acknowledgments: We thank H. H. Ng (Singapore) for the antibody against Klf4, I. Horikawa (NIH) for the 300-bp promoter construct, and J.-H. Lee (Indiana University) for Ash2l and Setd1a expression vectors. We thank E. M. Varela and M. Blasco (Madrid) for advice and reagents for the Q-fish analysis and for antibodies against Tert. We are grateful to M. M. Taketo (Kyoto) and to S. Robine (Paris) for providing mutant mice. We thank S. Lugert for advice on neurosphere cultures, J. Volkind for technical assistance, V. Taylor and D. Junghans for discussion and critical reading, and R. Schneider for typing of the manuscript. K.H. and A.R. are members of the International Max Planck Research School for Molecular and Cellular Biology (IMPRS-MCB). This work was supported by the Max Planck Society. The authors declare no financial interests.
View Abstract

Navigate This Article