Downregulation of telomerase reverse transcriptase mRNA ... - Nature

Oncogene (2000) 19, 5123 ± 5133 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

Downregulation of telomerase reverse transcriptase mRNA expression by wild type p53 in human tumor cells Dawei Xu*,1, Qian Wang3, Astrid Gruber1, Magnus BjoÈrkholm1, Zhiguo Chen1, Ahmed Zaid4, Galina Selivanova3, Curt Peterson5, Klas G Wiman3 and Pavel Pisa2 1

Department of Medicine, Division of Hematology, Radiumhemmet, Karolinska Hospital, SE-171 76 Stockholm, Sweden; Department of Oncology, Radiumhemmet, Karolinska Hospital, SE-171 76 Stockholm, Sweden; 3Microbiology & Tumor Biology Center, Karolinska Institute, SE-171 77 Stockholm, Sweden; 4Department of Biochemistry, Stockholm University, SE-104 05 Stockholm, Sweden; 5Clinical Pharmacology, Faculty of Health Science, LinkoÈping University, SE-581 85 LinkoÈping, Sweden

2

The p53 tumor suppressor protein inhibits the formation of tumors through induction of cell cycle arrest and/or apoptosis. In the present study we demonstrated that p53 is also a powerful inhibitor of human telomerase reverse transcriptase (hTERT), a key component for telomerase. Activation of either exogenous temperature-sensitive (ts) p53 in BL41 Burkitt lymphoma cells or endogenous wild type (wt) p53 at a physiological level in MCF-7 breast carcinoma cells triggered a rapid downregulation of hTERT mRNA expression, independently of the induction of the p53 target gene p21. Co-transfection of an hTERT promoter construct with wt p53 but not mutant p53 in HeLa cells inhibited the hTERT promoter activity. Furthermore, the activation of the hTERT promoter in Drosophila Schneider SL2 cells was completely dependent on the ectopic expression of Sp1 and was abrogated by wt p53. Finally, wt p53 inhibited Sp1 binding to the hTERT proximal promoter by forming a p53-Sp1 complex. Since activation of telomerase, widely observed in human tumor cell lines and primary tumors, is a critical step in tumorigenesis, wt p53-triggered inhibition of hTERT/telomerase expression may re¯ect yet another mechanism of p53mediated tumor suppression. Our ®ndings provide new insights into both the biological function of p53 and the regulation of hTERT/telomerase expression. Oncogene (2000) 19, 5123 ± 5133. Keywords: hTERT; p53; Sp1; telomerase Introduction Human telomerase, a ribonucleoprotein enzyme that extends chromosome ends with (TTAGGG)n telomeric repeat sequences, is generally inactive in normal human somatic cells (Harley et al., 1994; Kim et al., 1994; Reddel, 1998). Without telomerase activity, telomeres in somatic cells shorten progressively with each round of cell division due to `the end replication problem' and the cells eventually enter senescence when their telomeres reach a critical length. Recent evidence suggests that tumor cells can overcome this intrinsic mitotic clock and escape senescence via activating telomerase and consequently compensating for the

*Correspondence: D Xu, Hematology Laboratory, CMM, L8:03, Karolinska Hospital, SE-171 76 Stockholm, Sweden Received 5 June 2000; revised 21 August 2000; accepted 30 August 2000

attrition of telomeres during cellular replication (Harley et al., 1994; Kim et al., 1994). Activation of telomerase is thus believed to be linked to cellular immortalization and tumorigenesis and has been observed in a variety of human tumor cell lines and in primary malignant cells/tissues (Kim et al., 1994). It is currently unclear how telomerase is activated or regulated in human tumor cells. The human telomerase complex is comprised of multiple components, but telomerase reverse transcriptase (hTERT) is the key component for the control of telomerase activity (Bodnar et al., 1998; Meyerson et al., 1997; Nakamura et al., 1997). Induction of hTERT expression is required for telomerase activation during cellular immortalization and tumor progression (Meyerson et al., 1997). The ectopic expression of hTERT is capable of reconstituting telomerase activity in normal human telomerase-negative ®broblasts and as such immortalizes these transfectants (Bodnar et al., 1998; Vaziri and Benchimol, 1998). Therefore, the insight into the regulatory mechanisms of hTERT expression will contribute to a better understanding of telomerase activation in human tumors. The p53 protein is a transcription factor with multiple biological activities. p53 accumulates in response to cellular stress such as DNA damage, hypoxia and oncogene activation, triggering cell cycle arrest and/or apoptosis (see Asker et al., 1999). This requires transactivation or repression of speci®c target genes and probably transcription-independent events as well (see Ko and Prives, 1996; Gottlieb and Oren, 1998). p53 may also play a role as a regulator of the cellular senescence program (see Levine, 1997; Reddel, 1998). Absence of functional p53 allows cellular immortalization and predisposes cells to neoplastic transformation. p53 is inactivated by mutation and various other mechanisms in more than 50% of all human tumors (see Asker et al., 1999; Hollstein et al., 1994). The fact that hTERT/telomerase and p53 seem to counteract each other in the context of cellular aging and immortalization and that the activation of telomerase and inactivation of p53 are characteristic of human tumors raises the question of whether there exists a link between hTERT/telomerase regulation and p53. p53 has been shown to directly interact with the transcription machinery and to negatively regulate a variety of cellular genes (Ko and Prives, 1996). Introduction of wild type (wt) p53 into tumor cells was recently shown to inhibit telomerase activity/ hTERT expression (Kanaya et al., 2000; Kusumoto

Downregulation of hTERT mRNA by wt p53 D Xu et al

5124

et al., 1999). In the present study, we corroborated this ®nding and further de®ned the regulatory mechanisms by which p53 represses the expression of hTERT.

Results Rapid downregulation of hTERT mRNA expression in BL41-ts p53 cells upon activation of wt p53 The BL41 Burkitt lymphoma cells transfected with a temperature sensitive (ts) p53 construct (BL41-ts p53) are arrested in G1 phase of cell cycle and undergo apoptosis upon activation of wt p53 at 328C (Okan et al., 1995; Ramqvist et al., 1993). The BL41-ts p53 cells were harvested at various time points after temperature shift to 328C and examined for hTERT mRNA expression. A signi®cant decline in hTERT expression occurred in BL41-ts p53 cells 1 ± 2 h after the temperature shift to 328C and only around 10% of the original level of hTERT mRNA remained by 24 h (Figure 1a). Consistent with the long half-life of telomerase activity (524 h) (Holt et al., 1997; Xu et al., 1996), there was no detectable change in telomerase activity within this period (Figure 1a). In contrast, neither hTERT expression nor telomerase activity decreased in parental BL41 cells carrying endogenous mutant p53 cultured at 328C (Figure 1a). To corroborate the wt p53-mediated downregulation of hTERT mRNA shown by the competitive RT ± PCR analysis, we also examined the levels of hTERT mRNA in the BL41-ts p53 cells by Northern blotting. The 377 bp hTERT probe detected two major and one minor RNA species, around 4.4, 9.5 and 6 kb, respectively, in BL41-ts p53 cells grown at 378C. As shown in Figure 1b, hTERT mRNA expression decreased rapidly following a temperature shift to 328C, con®rming the results from the competitive RT ± PCR. Since the BL41-ts p53 cells undergo apoptosis within 24 ± 48 h following the induction of wt p53 at 328C (Ramqvist et al., 1993), the experiment was terminated at 24 h. For the same reason, it was impossible to determine whether the decreased expression of hTERT mRNA is followed by a decline in telomerase activity in these cells. To overcome this problem, we examined a subline of BL41-ts p53 cells in which the EpsteinBarr virus (EBV) encoded latent membrane protein1 (LMP1) had been introduced using a retroviral vector. LMP1 prevents p53-induced apoptosis in BL41-ts p53/ LMP1 cells (Okan et al., 1995), allowing us to study telomerase activity over a longer time period (4 ± 6 days). Induction of wt p53 caused a reduction of hTERT mRNA levels in the BL41-ts p53/LMP1 cells (data not shown), followed by a progressive decline in telomerase activity starting at day 2 after temperature shift (Figure 1c). Activation of endogenous wt p53 induces the inhibition of hTERT expression in MCF-7 breast carcinoma cells To determine whether activation of endogenous wt p53 at physiological levels can also inhibit hTERT expression, we investigated the eects of DNA damage-induced accumulation of wt p53 on hTERT mRNA levels in MCF-7 breast carcinoma cells Oncogene

carrying wt p53. MCF-7 cells were incubated with 10 mg/ml of mitomycin C (MMC) to activate wt p53 (data not shown) and then assayed for hTERT expression and telomerase activity. MMC treatment of MCF-7 cells resulted in a rapid reduction of hTERT expression (Figure 1d), similar to that observed after the activation of exogenous wt p53 in BL41-ts p53 cells. In contrast, only a slight decline in hTERT mRNA level was detected in T-47D breast carcinoma cells carrying mutant p53 and in p53 null MDA-MB157 cells after exposure to MMC (Figure 1d). Thus, hTERT expression is repressed in response to activation of endogenous wt p53. Downregulation of hTERT mRNA mediated by wt p53 is independent of p21 induction Recent observations have suggested that telomerase activity is proliferation-related, with high levels in actively cycling cells and low levels in quiescent cells (Greider, 1998). Moreover, transfection of human tumor cells with p21, the cyclin-cdk inhibitor, was shown to cause cell cycle arrest accompanied by inhibition of telomerase activity (Kallassy et al., 1998). p53 triggers G1 cell cycle arrest through the induction of p21 (Asker et al., 1999). The question arises as to whether wt p53-mediated inhibition of hTERT/ telomerase expression is a consequence of p53-induced p21 expression and/or cell cycle arrest. To address this question, we used antisense oligonucleotides for p21 mRNA (AS/p21) to inhibit the induction of p21 in BL41-ts p53 cells grown at 328C. Following the addition of AS/p21 oligo, the cells were incubated at 328C for 8 h. Compared to the control cells at 328C or the cells treated with sense oligonucleotides for p21 mRNA (SS/p21), up to 80% of the p21 induction observed after temperature shift to 328C was inhibited in the presence of AS/p21, and [3H]-thymidine incorporation demonstrated that the AS/p21-treated cells actively synthesized DNA due to the abrogation of p21 expression (Figure 2a and b). Nevertheless, hTERT and mRNA expression was still downregulated to the same extent in the absence of p21 induction and cell growth arrest (Figure 2c and d). The result suggests that the induction of p21 is not required for the p53-mediated suppression of hTERT expression in BL41-ts p53 cells. However, it is possible that even residual p21 protein in the AS/p21-treated cells was sucient to trigger a decline in the levels of hTERT mRNA. To test this, we treated BL41-ts p53 cells with 0.5 mM mimosine which is known to induce p21 expression and arrest cell growth independently of p53. Mimosine treatment, although inducing p21 expression and inhibiting cell proliferation more eciently than wt p53 activation in BL41-ts p53 cells, did not signi®cantly aect the hTERT mRNA levels (Figure 3). p21 itself, therefore, is unlikely to be a direct regulatory factor that inhibits hTERT expression. wt p53 but not mutant p53 inhibits transcription of hTERT gene We then asked whether hTERT transcription was reduced by overexpression of wt p53. To this end, the eects of p53 on hTERT promoter activities were


5125

Figure 1 hTERT expression and/or telomerase activity in BL41-ts p53, BL41-ts p53/LMP1, BL41 parental cells grown at 328C and breast carcinoma cells with dierent p53 status following the activation of wt p53. Representative experiments are shown. (a) Left panel: hTERT expression in BL41-ts p53 cells and BL41 parental cells incubated at 328C. C: Competitor. Right panel: Densitometric analysis of the PCR products shown in the left panel and the levels of telomerase activity in BL41-ts p53 cells and BL41 parental cells incubated at 328C. The levels of hTERT mRNA and telomerase activity in the cells at 328C were expressed as the percentage of those in each subline at 378C. (b) Left panel: Northern blot analysis of hTERT mRNA in BL41-ts p53 cells at 378C and after incubation for the indicated times at 328C. Right panel: relative levels of hTERT mRNA normalized to GAPDH as determined with densitometry. (c) wt p53-mediated inhibition of telomerase activity in BL41-ts p53/LMP1 cells. Telomerase activity in BL41-ts p53/LMP1 cells grown at 328C for 6 days was analysed daily and expressed as the percentage of that in the same cells grown at 378C. (d) MCF-7 (wt p53), MDA-MB-157 (p53 null) and T-47D (mutant p53) cells were incubated with 10 mg/ml of MMC. hTERT mRNA levels were expressed as the percentage of those in each corresponding cell line without MMC treatment

examined in the cervical carcinoma cell line HeLa that expresses strong telomerase activity and lacks functional p53. We tested three hTERT promoter con-

structs designated p330, p1009 and p3996, containing the ®rst exon, ®rst intron and the part of the second exon, as well as 330, 1009 and 3996 bp fragments Oncogene


5126

Figure 2 Eect of p21 mRNA antisense oligonucleotides on hTERT expression in BL41-ts p53 cells at 328C. N: no treatment; AS or AS/p21: antisense oligonucleotides for p21 mRNA; SS or SS/p21: sense oligonucleotides for p21 mRNA and C: Competitor. The cells were treated with 20 mM of AS/p21 or SS/p21 for 8 h at 328C. (a) Western blot analysis of p21 protein in the control and treated cells. (b) [3H]thymidine incorporation of BL41-ts p53 cells in the presence or absence of AS/p21 and SS/p21 at 328C. (c) Competitive RT ± PCR for hTERT mRNA in BL41-ts p53 cells incubated without and with AS/p21 or SS/p21 at 328C for 8 h. (d) The images in c were analysed with densitometry and the levels of hTERT mRNA in control and AS/p21 or SS/p21-treated BL41-ts p53 cells at 328C were expressed as the percentage of those in the same cells grown at 378C

upstream of the hTERT translation start site ATG, respectively (Cong et al., 1999). The construct p330 represents the hTERT proximal promoter. These three constructs all gave rise to high levels of luciferase activity in HeLa cells (data not shown). When 1 mg p330 reporter plasmids together with 0.5 mg wt p53 vectors were transfected into HeLa cells, hTERT promoter activities were suppressed by up to 90% compared to that in the same cells cotransfected with p330 and the empty vector, demonstrating that the reduction of hTERT mRNA by p53 occurs at transcriptional level (Figure 4a). A similar inhibitory eect was also observed in p1009 or p3996 plus wt p53transfected HeLa cells. This indicates that the hTERT proximal promoter p330, is sucient for p53-mediated inhibition of hTERT transcription. As shown in Figure 4b, the eect of wt p53 was dose-dependent and even 10 ng of wt p53-containing constructs resulted in a signi®cant suppression of hTERT promoter-driven luciferase activity. On the contrary, the His-273 mutant p53 construct, did not signi®cantly inhibit hTERT promoter activity in HeLa cells. Both wt and mutant p53 did not aect b-Gal and GFP expression driven by the CMV promoter (Figure 4c and data not shown), indicating speci®c eects of wt p53 on the hTERT promoter activity. Oncogene

Sp1-mediated transactivation of hTERT promoter is significantly blocked by wt p53 in SL2 cells Like other cellular and viral gene promoters which are repressed by p53, hTERT promoter lacks p53 consensus-binding sites (Cong et al., 1999; Greenberg et al., 1999; Takakura et al., 1999; Wick et al., 1999; Wu et al., 1999a). P53-mediated transcriptional repression of the hTERT gene could result from its interaction with the protein components of transcription machinery such as the transcription activator, Sp1, as demonstrated in several p53-repressing genes (Bargonetti et al., 1997; Ko and Prives, 1996; Perrem et al., 1995; Wang and Beck, 1998b). The hTERT promoter indeed harbors multiple putative Sp1 binding sites (Cong et al., 1999; Takakura et al., 1999; Wick et al., 1999). To investigate whether these Sp1 binding sites could mediate the transcriptional activation of hTERT by Sp1, we cotransfected the P330 construct with the Sp1 expression vector (driven by the Drosophila actin promoter, Pac) into SL2 cells that lack endogenous Sp1 (Courey and Tjian, 1998) and telomerase activity. The P330 construct was inactive in SL2 cells and its activation completely depended on the ectopic expression of Sp1 in a dose-related fashion (Figure 5a). However, the Sp1-mediated transactivation of the p330


5127

Figure 3 Comparison of the eects of mimosine or AS/p21 treatment on p21 expression, cell growth and hTERT mRNA levels in BL41-ts p53 cells. (a) p21 protein expression in mimosine-treated cells at 378C or cells grown at 328C as determined using Western blot. (b) [3H]thymidine incorporation of the cells treated with mimosine or grown at 328C. (c) Competitive RT ± PCR for hTERT mRNA in mimosine-treated BL41-ts p53 cells or cells grown at 328C. C: Competitor. (d) The PCR products shown in (c) were analysed with densitometry and the levels of hTERT mRNA in the treated cells were expressed as the percentage of those in the same cells cultured at 378C without mimosine

reporter was signi®cantly inhibited by co-transfection of wt p53 (Figure 5a). This eect of wt p53 was not due to reduced Sp1 expression as demonstrated by Western blot analysis (Figure 6a). To rule out a possible non-speci®c eect of wt p53, a Pac-driven bGal expression vector was co-transfected with wt p53 and Sp1 vectors into SL2 cells and wt p53 did not aect b-Gal expression (Figure 5b). Moreover, SL2 cells lack p53 homologs (Soussi et al., 1990) and the ectopic expression of p53 did not aect the cell viability and proliferation status (Figure 5c and data not shown). Thus, wt p53 speci®cally inhibited the Sp1mediated activation of the hTERT promoter. Compared to wt p53, His-273 mutant p53 had a much weaker inhibitory eect on the Sp1-dependent hTERT promoter activity (Figure 5a). wt p53 forms a complex with Sp1 and inhibits Sp1 binding to the hTERT proximal promoter Since wt p53 did not aect the Sp1 expression level in SL2 cells (Figure 6a), we examined whether p53 and Sp1 could physically interact with each other. wt p53 was readily immunoprecipitated with an Sp1 antibody in SL2 cells cotransfected with the p53 and Sp1 constructs (Figure 6b), indicating the in vivo formation of p53-Sp1 complex. The nuclear extracts from these cells were further analysed for their binding to Sp1

motifs in the hTERT core promoter. An electrophoretic mobility shift assay (EMSA) was performed with a 400 bp double-stranded probe (Probe 400) containing two consensus and three degenerated Sp1 binding sites upstream of the hTERT transcription start site. Consistent with the p330 reporter data, Sp1-transfected SL2 cells exhibited a speci®c binding activity to the probe while the co-transfection with wt p53 almost completely abolished the binding to the Sp1-binding sites (Figure 6c). In addition, we further determined the potential interaction between Sp1 and wt p53 in MMC-treated MCF-7 cells. Sp1 protein was equally expressed in untreated and treated cells while wt p53 protein was dramatically induced by MMC treatment (Figure 6d). Interestingly, wt p53 could also be co-immunoprecipitated with an Sp1 antibody in MMC-treated MCF-7 cells (Figure 6d), suggesting the possibility that the functional interaction between wt p53 and Sp1 is physiologically relevant in regulating hTERT expression. To further con®rm that wt p53 rather than its downstream eectors directly repressed the formation of Sp1/DNA complex observed in SL2 cells, we performed EMSA in the presence of the puri®ed recombinant human full length of wt p53 protein (wt p53/¯) or the wt p53 core domain (wt p53/core). Neither wt p53/¯ nor wt p53/core bound the hTERT Oncogene


5128

Figure 4 wt p53-mediated suppression of hTERT promoter activity. Bars: s.d. (a) The eects of wt and mutant p53 on hTERT promoter activity. Luciferase activity of each promoter construct in HeLa cells cotransfected with wt p53 or mutant p53 was expressed as the percentage of that in the empty vector-transfected cells. The mean value was obtained from three independent experiments. (b) wt p53 represses the hTERT promoter activity in a dose-related manner. One mg of p330 vector was cotransfected with the wt p53 construct into HeLa cells. Luciferase activity in wt p53-transfected cells was expressed as the percentage of that in the cells transfected with the same amount of the empty construct. (c) The eects of wt and mutant p53 on the b-Gal expression driven by the CMV promoter

proximal promoter (data not shown). The ability of Sp1 to bind the hTERT promoter sequence with a single Sp1 binding site was completely inhibited in the presence of 250 ng wt p53/¯ (Figure 6e). In contrast, wt p53/core did not interfere with Sp1-DNA binding (Figure 6e). These results thus suggest that the full length of p53 protein is required for the inhibition of Sp1-DNA binding. Discussion Our data demonstrate that hTERT expression is transcriptionally downregulated upon the induction of wt p53 in human tumor cells. Similarly, two recent studies showed that the overexpression of wt p53 inhibited telomerase activity and hTERT expression in pancreatic cancer cells (Kanaya et al., 2000; Kusumoto et al., 1999). This ®nding, coupled with the observation that telomerase was activated concomitantly with the loss of the remaining functional wt p53 allele in LiFraumeni syndrome (LFS) skin ®broblasts with heterozygous germline p53 mutations (Gollahon et al., 1998), suggests the existence of a p53-dependent regulatory pathway for hTERT/telomerase control in human cells. In further support of this proposal, a number of clinical studies revealed a close correlation Oncogene

between high levels of telomerase activity and p53 abnormalities (Loveday et al., 1999; Roos et al., 1998; Wu et al., 1999b). One essential question is whether downregulation of hTERT expression is a direct eect of p53 or merely a consequence of other p53-dependent functions. Our results suggest that hTERT downregulation is not due to p53-induced apoptosis, as abrogation of p53-induced apoptosis by the EBV/ LMP1 protein did not block p53-mediated suppression of hTERT expression and telomerase activity. Furthermore, the overexpression of wt p53 did not induce apoptosis in SL2 cells, but nevertheless caused the repression of hTERT promoter activity in these cells. As wt p53 induces p21 expression and cell cycle arrest, while telomerase activity seems growth-regulated, i.e. high in actively proliferating cells and low in quiescent cells (Greider, 1998), we examined the relationship between hTERT mRNA downregulation and p21/cell cycle arrest induced by wt p53 in BL41-ts p53 cells in further detail. The obtained data provide evidence that p21 inducted and cell growth arrest is not required for the onset of hTERT mRNA reduction mediated by wt p53. First, a signi®cant decrease in hTERT expression was already seen within 1 ± 2 h of wt p53 activation, whereas a detectable p21


5129

Figure 5 wt p53 inhibits the transactivation of the hTERT promoter mediated by Sp1 in SL2 cells. Bars: s.d. (a) Complete dependence of the hTERT promoter activity on Sp1 expression in SL2 cells. The cells were transfected with various amounts of Sp1 vector and/or wt p53/mt p53 construct plus p330 as indicated and the luciferase activity was determined. (b) The eects of Sp1, wt p53 and mutant p53 transfection on pPac-b-Gal expression in SL2 cells. 200 ng of Sp1 vector plus 500 ng of pPacU or together with 500 ng of wt p53 and mutant p53 vectors were transfected into Drosophila SL2 cells and b-Gal activity was determined. (c) The proliferation and viability of Sp1, wt p53 and mutant p53 transfected SL2 cells. 16106 cells were transfected with Sp1 (200 ng) plus pPacU (500 ng) or wt p53 (500 ng), or mutant p53 (500 ng) and viable cells were counted 48 h post-transfection

protein and G1 arrest occurred at 4 ± 8 h, and 15 ± 18 h post-wt p53 activation, respectively, in BL41-ts p53 cells (Okan et al., 1995; data not shown). Second, despite the absence of growth arrest in SL2 cells following wt p53 transfection, the transactivation of hTERT promoter by Sp1 was signi®cantly abrogated by wt p53. Finally, hTERT mRNA was still downregulated even when the p53-mediated induction of cell growth arrest was greatly diminished by AS/p21 treatment. One may argue that the residual p21 protein may be sucient to trigger the downregulation of hTERT expression in the AS/p21-treated BL41-ts

p53 cells. However, this seems unlikely because the high levels of p21 protein and cell cycle arrest induced by mimosine in the same cells without wt p53 activation did not signi®cantly aect the expression of hTERT mRNA. In recent studies, hTERT mRNA was found to be expressed in both actively regenerating cells and mitotically inactive breast lobular epithelium (Harle-Bachor and Boukamp, 1996; Kolquist et al., 1998); and the ability of c-myc to induce hTERT expression could be separated from its ability to stimulate cell proliferation (Wu et al., 1999a). Thus, hTERT expression does not strictly correlate with the Oncogene


5130

Figure 6 SP1 and p53 expression, their interaction and the eect of the interaction on the Sp1-binding to the hTERT promoter. (a) Western blot analysis of Sp1 and p53 expression in SL2 cells transfected with Sp1 and/or p53 constructs. Lane 1: pPacU only; lanes 2 ± 5: Sp1 50, 100, 200 and 400 ng, respectively; lane 6: 100 ng of Sp1 plus 500 ng of p53; lane 7: Sp1 200 ng plus 500 ng of p53. (b) Physical interaction between Sp1 and wp p53. The protein complex was precipitated with Sp1 antibody and co-precipitated p53 protein detected with p53 antibody. Lane 1: pPacU only; 2: Sp1 only; 3: Sp1+wt p53 and 4: p53 only. (c) Binding of puri®ed Sp1 protein and Sp1 and/or wt p53-transfected SL2 nuclear extracts to the hTERT core promoter. 0.4 fpu Sp1 or 5 mg of the nuclear extracts from SL2 cells transfected as indicated were incubated with a 400 bp double-stranded hTERT core promoter probe and the Sp1-DNA complex was resolved in 4% of polyacrylamide gels. C: Sp1 competitive oligonucleotides; MC: mutated Sp1 oligonucleotides. (d) wt p53 and Sp1 expression and their physical interaction in MMC-treated MCF-7 cells. wt p53 and Sp1 protein levels were determined using Western blot. For immunoprecipitation (wt p53 IP), the protein extracts were precipitated with Sp1 antibody and co-precipitated p53 protein was detected with p53 antibody. (e) The full length wt p53 but not the p53 core domain protein prevents Sp1 from binding the hTERT proximal promoter. EMSA was performed in the presence of recombinant full length wt p53 or the p53 core domain protein. wt p53/¯: the full length wt p53 and wt p53/core: the p53 core domain

proliferative status of cells. However, transfection of human immortal keratinocytes with p21 was shown to trigger cell cycle arrest accompanied by an inhibition of telomerase activity (Kallassy et al., 1998). A possible explanation for the discrepancy between this result and ours is that p21 is not involved in the early downregulation of hTERT expression observed in Oncogene

BL41-ts p53 cells but may be directly or indirectly involved in hTERT/telomerase regulation over a longer period. Indeed, the suppression of telomerase activity in the human keratinocytes did not occur until 2 weeks after p21 transfection. Alternatively, the observed decline in telomerase activity in the p21transfected cells is simply a consequence of overall low


metabolic activity resulting from p21-induced cell growth arrest. p53 transcriptionally represses a number of cellular and viral genes which lack p53 consensus-binding sites in their promoters (Ko and Prives, 1996; Gottlieb and Oren, 1998). p53-mediated transcriptional repression may result from its direct interaction with transcription factors or protein components of the transcription machinery such as the TATA binding protein (TBP), TATA associated factors (TAFs), CCAAT-binding factor (CBF) and Sp1 (Bargonetti et al., 1997; Ko and Prives, 1996; Perrem et al., 1995). The hTERT promoter does not harbor any classical p53 binding sites, and it lacks a TATA motif. However, the observation that Sp1 binds at least ®ve sites in the hTERT core promoter (Takakura et al., 1999), along with our and other's ®nding that Sp1 is a strong transactivator of the hTERT promoter (Kyo et al., 2000), suggest a plausible pathway: p53 may render Sp1 inactive for hTERT transcription. Kanaya et al. (2000) showed that the mutation of Sp1 motifs on the hTERT promoter led to inability of p53 to repress hTERT transcription. We further demonstrated formation of wt p53-Sp1 complexes, disruption of Sp1hTERT promoter binding and inability of Sp1 to transactivate the hTERT promoter in the presence of wt p53. Moreover, a wt p53-Sp1 complex was formed in the MMC-treated MCF-7 cells, indicating a physiological relevance of the interaction between wt p53 and Sp1 in the regulation of hTERT/telomerase expression in human cells. Similar phenomena have been observed in other Sp1 binding site-containing promoters (Perrem et al., 1995). Using DNase footprinting, Bargonetti et al. (1997) showed that wt p53 not only prevented Sp1 from binding to its consensus motifs but also competed for these binding sites with Sp1, whereas mutant p53 did not change the Sp1-DNA binding. However, Borellini et al. (1993) found that mutant p53 proteins also interact with Sp1 and the resultant mutant p53-Sp1 complexes stimulated rather than inhibited gene transcription. It is currently unclear why wt p53 and mutant p53-Sp1 complexes exhibit dierential eects on the regulation of gene expression. It is well-known that myc oncogene directly activates hTERT gene (Wang et al., 1998a; Wu et al., 199a), and Sp1 seems to be required for this action of myc (Kyo et al., 2000). Although the activation of wt p53 does not lead to changes in c-myc expression in BL41-ts p53 cells (Sangfelt et al., 1996), the interaction between p53 and Sp1 might aect the myc-regulatory pathway of hTERT expression as well. In addition, p53 was shown to interact directly with the telomerase associated protein and consequently inhibit telomerase activity (Li et al., 1999). Therefore, the inhibitory eect of p53 on hTERT and telomerase may occur through multiple pathways. It should be emphasized that inactivation of wt p53 would not necessarily result in telomerase activation. Telomerase activity is stringently controlled at multiple levels by multiple factors and dierent types of cells may display dierent regulatory mechanisms (Cong et al., 1999). Thus p53 is unlikely the only barrier to hTERT/telomerase activation. This is exempli®ed by the human ®broblasts transfected with SV40 large T antigen that binds to and inactivates p53. In these cells,

hTERT/telomerase activation does not occur until further genetic alterations are acquired (Harley et al., 1994; Reddel, 1998). The control of telomerase/hTERT expression may be more stringent in ®broblasts. On the other hand, c-myc and the E6 protein of the human papillomavirus are able to directly induce hTERT expression and activate telomerase in normal human cells (Kiyono et al., 1998; Klingelhutz et al., 1996). Some genetic or molecular events may allow cells to bypass the wt p53-mediated control of hTERT/ telomerase and thus it is not surprising that activation of telomerase also occurs in tumor cells carrying wt p53, such as MCF-7 cells. MCF-7 cells, like other wt p53-carrying tumor cells, have presumably evolved mechanisms to escape many of the eects of wt p53 including growth arrest and apoptosis. Moreover, the maintenance of telomeres by a mechanism (alternative lengthening of telomeres) other than telomerase activation has been observed in a subset of human tumors (Bryan et al., 1995, 1997). Therefore, a simple correlation between hTERT/telomerase activation and p53 status in tumor cells should not be expected. In conclusion, we demonstrate that hTERT expression is downregulated following induction of wt p53 in human tumor cells and that wt p53 directly inhibits hTERT transcription. Given that hTERT/telomerase activation is central for cell immortalization, and tumor development and progression, this novel function of p53 may represent an important mechanism of tumor suppression mediated by p53. Accordingly, our data should contribute to a better understanding of not only telomerase/hTERT regulation but also aging, immortalization and tumorigenesis.

5131

Materials and methods Cell lines and cell culture The BL41 Burkitt lymphoma cell line and the establishment of the BL41-ts p53 (BL41 transfected with a ts p53 construct) and BL41-ts p53/LMP1 (BL41-ts p53 cells infected with an LMP1 retrovirus) sublines have been described previously (Okan et al., 1995; Ramqvist et al., 1993). For the induction of wt p53 in BL41-ts p53 and BL41-ts p53/LMP1 cells, the culture temperature was shifted from 37 to 328C. The BL41ts p53 cells were incubated with 0.5 mM of mimosine (Sigma, St. Louis, MO, USA) to induce p21 and cell cycle arrest. The breast carcinoma cell lines MCF-7 (wt p53), T-47D (mutant p53) and MDA-MB-157 (p53 null) were maintained in IDM medium (Life Technologies, Paisley, UK). MCF-7 cells were incubated with 10 mg/ml mitomycin C (Sigma) to induce wt p53 expression. The cervical carcinoma cell line HeLa cells were grown in RPMI 1640 medium (Life Technologies). The SL2 cells were cultured at 258C in Schneider's Drosophila medium (Life Technologies). All the culture media contained 10% fetal calf serum and 100 U/ml penicillin. Antisense oligonucleotides for p21 mRNA Phosphorothioate-modi®ed oligonucleotides were commercially synthesized (EurogentecBel AS, Seraing, Belgium). The sequence for the p21 antisense oligonucleotide (AS/ P21), 5'-TCC CCA GCC GGT TCT GAC AT-3', is complementary to the region of the inhibition codon of p21 mRNA. The p21 sense oligonucleotide (SS/P21) is 5'-ATG TCA GAA CCG GCT GGG GA-3'. BL41-ts p53 cells were maintained in AIM V medium (Life Technologies) for one week before the experiment was performed. Oligonucleotides Oncogene


5132

were ®rst incubated with lipofectin (Life Technologies) for 30 min and then added to the BL41-ts p53 cell culture (®nal concentration 20 mM) prior to temperature shift from 37 to 328C and the cells harvested at indicated time points at 328C. [3H]thymidine incorporation 26105 cells were cultured in U-shaped 96-well microtiter plates in the presence or absence of AS/p21, SS/p21, and mimosine at 37 or 328C, and 1.0 mCi of [3H]thymidine (Amersham) was added to each well. The cells were harvested with a Tomtec harvesting machine (Wallach, Upplands VaÈsby, Sweden) and the radioactivity was determined in a 1450 Microbeta Trilux Scintillation Counter (Wallach). RNA extraction, reverse transcription and quantitative determination of hTERT mRNA expression and half-life Total cellular RNA was extracted using the ULTRASPECTMII RNA kit (Biotecx Lab., Houston, TX, USA). cDNA was synthesized using random primers (N6) (Pharmacia, Uppsala, Sweden) and MMLV reverse transcriptase, as described (Xu et al., 1998, 1999). The hTERT PCR primer sequences were CGGAAGAGTGTCTGGAGCAA (Upper) and GGATGAAGCGGAGTCTGGA (Lower). The competitive template for hTERT, constructed by inserting a LAC operator sequence in the middle of the wild type template by mimic PCR, was 21 bp longer than its wild type counterpart. cDNA corresponding to 50 ng of RNA was coampli®ed with the competitor (56103 molecules) using 32 cycles (948C 45", 608C 45" and 728C 90"). In addition, b2-Microglobulin (b2M) expression was used as a control for RNA loading and RT eciency and coampli®ed with its competitive template (56105 molecules) using 25 cycles (Xu et al., 1999). PCR products were resolved in 4% Metaphor (Rockland FMC, ME, USA) agarose gels, stained with ethidium bromide, visualized in UV light, and photographed. Volumetric integration of signal intensities was performed by using the NIH Image software (Version 1.58). The relative levels of hTERT mRNA was calculated from the ratio of hTERT and competitor signal density normalized to the ratio of b2-M and its competitor. Northern blot Up to 8 mg of poly A mRNA derived from BL41-ts p53 cells at 37 and 328C were size fractionated in 1% of formaldehyde-agarose gels, transferred to a Hybond nylon membrane (Amersham, UK) and hybridized to a 377 bp hTERT cDNA fragment cut from the pUHD 10-3 vector (Kolquist et al., 1998; Meyerson et al., 1997) (provided by Drs WL Gerald and KA Kolquist, Memorial Sloan-Kettering Cancer Center, New York, USA). The same membrane was then reprobed with a full length of human GAPDH cDNA for loading control. Telomerase activity assay A commercial Telomerase PCR ELISA kit (Roche Diagnostics Scandinavia AB, Stockholm, Sweden) was used to determine telomerase activity in all specimens according to the manufacturer's protocol using 20 PCR cycles (Xu et al., 1998). Protein extraction and measurement was performed as described (Xu et al., 1996). Immunoblotting and immunoprecipitation Protein extracts were prepared by direct lysis into sodium dodecyl sulate (SDS) sample buer. An equal amount of protein extracted from control and treated cells was resolved by SDS-polyacrylamide gel electrophoresis (SDS ± PAGE) and transferred to nitrocellular membrane. p53, p21 and Sp1

Oncogene

proteins were detected with monoclonal (p53 and p21, Oncogene Research Products, Uniondale, NY, USA) or polyclonal (Sp1, Santa Cruz Biotechnology, Santa Cruz, CA, USA) antibodies, respectively and horseradish peroxidase and visualized by enhanced chemiluminescence (ECL Amersham, UK). The blots were then reprobed for b-actin protein for a loading control. For immunoprecipitation, 500 mg of total cellular protein derived from SL2 cells transfected with Sp1 and/or p53 vector or MCF-7 cells treated with MMC or AraC was pre-cleared with protein G-sepharose beads (Pharmacia) 2 h at 48C and immunoprecipitated with 1 mg of Sp1 polyclonal antibody at 48C for 6 h. The antibody-protein complex was then adsorbed onto the beads, washed and denatured. The entire precipitated proteins were resolved in SDS ± PAGE and detected with p53 antibody as described above. Plasmids, transient transfection and reporter gene assay The cytomegalovirus (CMV) enhancer-promoter driven expression vectors containing human wt p53 and mutant p53 have been described (Selivanova et al., 1997). The hTERT promoter deletion mutants p330, p1009 and p3996, cloned upstream of the ®re¯y luciferase reporter in the pGL2Enhancer vector, were described (Cong et al., 1999) and kindly provided by Dr Bacchetti (McMaster University, Canada). Sp1 expression vector pPacSp1 and the empty vector pPacU were generously provided by Dr R Tjian (University of California, Berkeley, USA). To generate the p53 vectors which can be expressed in SL2 cells, the full length of wt and mt p53 cDNA was inserted into pPacU at the BamHI site. All the cells were transfected with Lipofectin (Life Technologies). 105 HeLa cells on 12-well plates were transfected with 1 mg hTERT reporter plasmids in the presence of various amounts of wt p53, mt p53 and empty vectors, respectively. SL2 cells were transfected with p330 together with Sp1 and/or p53 vectors. Cell extracts were prepared 24 h (human cells) or 48 h (SL2 cells) later for luciferase activity determination. For transfection eciency control, all hTERT luciferase vectors were cotransfected with 50 ng of the Renilla reniformis luciferase-containing plasmids, whose gene is under the control of SV40 enhancer/promoter or TK promoter. Fire¯y and Renilla luciferase activity in cell lysates was determined by using the Dual Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. The ®re¯y luciferase activity was normalized to the Renilla reniformis luciferase activity (human cells) or the protein concentration and/or b-Gal (SL2 cells). EMSA The two double-stranded probes, end-labeled with 32P-dATP, were used. One of them was a 400 bp PCR-generated fragment (forward primer: 5'-CTAGGCCGATTCGACCTCTCTC-3' and reverse primer: 5'-CTTCCCACGTGCGCAGCAGGA3') which spans the hTERT core promoter region and contains multiple Sp1 binding sites; the other was a synthetic oligonucleotide with a single Sp1-binding site (corresponding to the region of hTERT transcription start site: 5'TCCTTTCCGTGGCCCCGCCCTCTCCTCGCGGCGCGA -3'). The probe (25 000 c.p.m.) was incubated with 0.4 fpu Sp1 (Promega) in the presence or absence of unlabeled Sp1 competitive oligonucleotides on ice for 30 min and the reaction mixture was electrophoresed in a 4% of polyacrylamide gel in 0.25 6 TBE buer. When the nuclear proteins from SL2 cells were used for EMSA, 5 mg of the extracts were pre-incubated with 2 mg of Poly (dI-dC) for 10 min on ice before the addition of the probe and/or the Sp1 competitors. The Sp1 speci®c competitive oligonucleotide and its mutated sequence were: ATATCTAGAGCGCTGCGGGGCGGAGCGAAG and ATATCTAGAGCGCTGCGGTGTGGAGCGAAG, respectively. The puri®cation of the recombinant human wt p53/¯


and wt p53/core domain proteins has been described (Selivanova et al., 1997). Acknowledgments We thank Dr S Bacchetti (McMaster University, Canada) for providing hTERT promoter vectors and for critical reading of the manuscript, Dr R Tjian (University of California, Berkeley, CA, USA) for providing pPacU and

pPacSp1 expression vectors, Drs WL Gerald and KA Kolquist (Memorial Sloan-Kettering Cancer Center, New York, USA) for hTERT probe. This work was supported by the Swedish Cancer Society, the Cancer Society in Stockholm, Karolinska Institute Funds, Dagmar Ferbs Memorial Fund, Stockholm County Council and the King Gustaf V Jubilee Fund. P Prisa is a research fellow of the Swedish Medical Research Council.

5133

References Asker C, Wiman KG and Selivanova G. (1999). Biochem. Biophy. Res. Commun., 265, 1 ± 6. Bargonetti J, Chicas A, White D and Prives C. (1997). Cell. Mol. Biol., 43, 935 ± 949. Bodnar A, Ouellette M, Frolkis M, Holt S, Chiu C, Morin G, Harley C, Shay J, Lichtsteiner S and Wright W. (1998). Science, 279, 349 ± 352. Borellini F and Glazer RI. (1993). J. Biol. Chem., 268, 7923 ± 7928 (1993). Bryan TM, Englezou A, Gupta J, Bacchetti S and Reddel RR. (1995). EMBO J., 14, 4240 ± 4248. Bryan TM, Englezou A, Dalla-Pozza L, Dunham MA and Reddel RR. (1997). Nat. Med., 3, 1271 ± 1274. Cong Y-S, Wen J and Bacchetti S. (1999). Hum. Mol. Genet., 8, 137 ± 142. Courey AJ and Tjian R. (1998). Cell, 55, 887 ± 893. Gollahon LS, Kraus E, Wu T-A, Yim SO, Strong LC, Shay JW and Tainsky MA. (1998). Oncogene, 17, 709 ± 717. Gottlieb TM and Oren M. (1998). Semin. Cancer Biol., 8, 359 ± 368. Greenberg RA, O'Hagan RC, Deng H, Xiao Q, Hann SR, Adams RR, Lichtsteiner S, Chin L, Morin GB and DePinho RA. (1999). Oncogene, 18, 1219 ± 1226. Greider CW. (1998). Proc. Natl. Acad. Sci. USA, 95, 90 ± 92. Harle-Bachor C and Boukamp P. (1996). Proc. Natl. Acad. Sci. USA, 93, 6476 ± 6481. Harley CB, Kim NW, Prowse KR, Weinreich SL, Hirsch KS, West MD, Bacchetti S, Hirte HW, Counter CM, Greider CW, Piatyszek MA, Wright WE and Shay JW. (1994). Cold Spring Harbor Symposia Quantitative Biology, 59, 307 ± 315. Hollstein M, Rice K, Greenblatt MS, Soussi T, Fuchs R, Sorlie T, Hovig E, Smith-Sorensen B, Montesano R and Harris CC. (1994). Nucleic Acids Res., 22, 3551 ± 3555. Holt S, Aisner D, Shay J and Wright W. (1997). Proc. Natl. Acad. Sci. USA, 94, 10687 ± 10692. Kallassy M, Martel N, Damour O, Yamaskai H and Nakazawa H. (1998). Mol. Carcinog., 21, 26 ± 36. Kanaya T, Kyo S, Hamada K, Takakura M, Kitagawa Y, Harada H and Inoue M. (2000). Clin. Cancer Res., 6, 1239 ± 1247. Kim NW, Piatyszek MA, Prowse KR, Harley CB, West MD, Ho PLC, Coviello GM, Wright WE, Weinrich SL and Shay JW. (1994). Science, 266, 2011 ± 2015. Kiyono T, Foster SA, Koop JI, McDougall JK, Galloway DA and Klingelhutz AJ. (1998). Nature, 396, 84 ± 88. Klingelhutz AJ, Foster SC and McDougall JK. (1996). Nature, 380, 79 ± 82. Ko LJ and Prives C. (1996). Genes Dev., 10, 1054 ± 1072. Kolquist KA, Ellisen LW, Counter CM, Meyerson M, Tan LK, Weinberg RA, Haber DA and Gerald WL. (1998). Nat. Genet., 19, 182 ± 186. Kusumoto M, Ogawa T, Mizumoto K, Ueno H, Niiyama H, Sato N, Nakamura M and Tanaka M. (1999). Clin. Cancer Res., 5, 2140 ± 2147.

Kyo S, Takakura M, Taira T, Kanaya T, Itoh H, Yutsudo M, Ariga H and Inoue M. (2000). Nucleic Acids Res., 59, 669 ± 677. Levine AJ. (1997). Cell, 88, 323 ± 331. Li H, Cao Y, Berndt MC, Funger JW and Liu JP. (1999). Oncogene, 18, 6785 ± 6794. Loveday RL, Greenman J, Drew PJ, Monson JRT and Kerin MJ. (1999). Int. J. Cancer, 84, 516 ± 520. Meyerson M, Counter CM, Eaton EN, Ellisen LW, Steiner P, Caddle SD, Ziaugra L, Beijersbergen RL, Davido MJ, Liu Q, Bacchetti S, Haber DA and Weinberg RA. (1997). Cell, 90, 785 ± 795. Nakamura TM, Morin GB, Chapman KB, Weinrich SL, Andrews WH, Lingner J, Harley CB and Cech TR. (1997). Science, 277, 955 ± 959. Okan I, Wang Y, Chen F, Hu L, Imreh S, Klein G and Wiman K. (1995). Oncogene, 11, 1027 ± 1031. Perrem K, Rayner J, Voss T, Sturzbecher H, Jackson P and Braithwaite A. (1995). Oncogene, 11, 1299 ± 1307. Ramqvist T, Magnusson KP, Wang Y, Szekly L, Klein G and Wiman KG. (1993). Oncogene, 8, 1495 ± 1500. Reddel RR. (1998). BioEssays, 20, 977 ± 984. Roos G, Nilsson P, Cajander S, Nielsen NH, Arnerlov C and Landberg G. (1998). Int. J. Cancer, 79, 343 ± 348. Sangfelt O, Einhorn S, BjoÈrklund AC, Wiman KG, Okan I and GrandeÂr D. (1996). Int. J. Cancer, 67, 106 ± 112. Selivanova G, Iotsova V, Okan I, Fritsche M, Groner B, Grafstrom RC and Wiman KG. (1997). Nat. Med., 3, 632 ± 638. Soussi T, de Fromentel CC and May P. (1990). Oncogene, 5, 945 ± 952. Takakura M, Kyo S, Kanaya T, Hirano H, Takeda J, Yutsudo M and Inoue M. (1999). Cancer Res., 59, 551 ± 557. Vaziri H and Benchimol S. (1998). Curr. Biol., 8, 279 ± 282. Wang J, Xie LY, Allan S, Beach D and Hannon GJ. (1998a). Genes Dev., 12, 1769 ± 1774. Wang Q and Beck WT. (1998b). Cancer Res., 58, 5762 ± 5769. Wick M, Zubov D and Hagen G. (1999). Gene, 232, 97 ± 106. Wu K-J, Grandori C, Amacker M, Simon-Vermot N, Polack A, Lingner J and Dalla-Favera RD. (1999a). Nat. Genet., 21, 220 ± 224. Wu X, Kemp B, Amos CI, Honn SE, Zhang W, Walsh GL and Spitz MR. (1999b). Br. J. Cancer, 80, 453 ± 457. Xu D, Gruber A, Peterson C and Pisa P. (1996). Leukemia, 10, 1354 ± 1357. Xu D, Gruber A, Peterson C and Pisa P. (1998). Br. J. Haematol., 102, 1367 ± 1375. Xu D, Gruber A, BjoÈrkholm M, Peterson C and Pisa P. (1999). Br. J. Cancer, 80, 453 ± 457.

Oncogene