Human Dkk-1, a gene encoding a Wnt antagonist, responds to DNA damage and its overexpression sensitizes brain tumor cells to apoptosis following alkylation damage of DNA

Article metrics


The human Dkk-1 (hDkk-1) gene, a transcriptional target of the p53 tumor suppressor, encodes a powerful inhibitor of the Wnt signaling pathway and regulates the spatial patterning/morphogenesis of the mammalian central nervous system. We investigated the p53-related functions of the hDkk-1 gene by studying its response to DNA damage and its modulation of apoptosis in human glioma cells. Various chemotherapeutic and other agents that induce DNA adducts and compromise its integrity (1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU), cisplatin, H2O2 and UV rays) enhanced the expression of hDkk-1 significantly. The damage-induced increase in hDkk-1 mRNA levels occurred in many human tumor cell lines, irrespective of their p53 gene status. The human glioblastoma cell line, U87MG, which had undetectable hDkk-1 expression, was engineered to express moderate levels of the hDkk protein by stable transfection. The engineered cells did not show any morphological changes, but underwent marked apoptosis after ceramide treatment. Further, the DNA cross-linking drugs BCNU and cisplatin, but not the microtubule poison vincristine, induced significant cell death in U87MG/hDkk cells, and this was accompanied by altered Bcl-2/Bax expression and a reduction in the amount of telomere DNA as visualized by fluorescence in situ hybridization. These results show that hDkk-1 is a pro-apoptotic gene and suggest that it may play important roles in linking the oncogenic Wnt and p53 tumor suppressor pathways.


The Dkk-1 gene, also called Dickkopf-1 (German for ‘big head, stubborn’), is a prototype member of a recently identified gene family that encodes secreted glycoproteins that control cell fate and neural patterning during embryonic development (Glinka et al., 1998; Krupnik et al., 1999; Kazanskaya et al., 2000). In early mouse embryos, Dkk-1 is strongly expressed in the foregut endoderm underlying the head fold (Glinka et al., 1998), and it has been shown that microinjection of Dkk-1 mRNA into Xenopus embryos is sufficient to cause head induction (Glinka et al., 1998). Evidence strongly indicates that Dkk-1 protein functions as a potent inhibitor of the Wnt signaling pathway to enable appropriate positioning and development of the embryonic brain and other organ structures (Glinka et al., 1998; Kazanskaya et al., 2000; Schneider and Mercola, 2001).

The Wnt pathway consists of many highly conserved Wnt protein ligands that bind to the frizzled family of receptors and trigger a cytoplasmic signal transduction cascade in which glycogen synthase kinase3 (GSK3) and β-catenin play essential roles (Dale, 1998; Wodarz and Nusse, 1998). In the absence of Wnt signaling, β-catenin is associated with a cytoplasmic complex containing the GSK-3, axin and the adenomatous polyposis coli protein (APC) (Hayashi et al., 1997; Sakanaka et al., 1998). In this complex, GSK3 constitutively phosphorylates β-catenin protein at serine and threonine residues, and the phosphorylated β-catenin becomes a target for ubiquitination and subsequent degradation by the proteasome (Hart et al., 1998). In the presence of Wnt signaling, GSK3 is inactivated, and this leads to accumulation of free and unphosphorylated β-catenin in the cytoplasm, which then translocates to the nucleus and associates with members of the lymphocyte enhancer factor (LEF)/TCF family of transcription factors to activate a variety of target genes involved in growth, development, and oncogenesis (Polakis, 2000; Barker et al., 2000). Some of the Wnt target genes include cyclin D1, engrailed-2, cyclooxygenase-2, c-myc, and many metalloproteases (McGrew et al., 1999; Howe et al., 1999; Shutman et al., 1999). Therefore, a finely tuned balance between the β-catenin pools in the cytoplasm and in the nucleus determines the final outcome of the Wnt signaling pathway. Such a balance is in part brought about by extracellular Wnt antagonists, including the human Dkk-1 (hDkk-1) protein, which interfere with Wnt signaling and diminish the accumulation of free β-catenin (Fedi et al., 1999; Moon et al., 1997; Molenaar and Destree, 1999). Coexpression of hDkk-1 with Wnt-2 has been shown to reverse the Wnt-2-induced morphological alterations as well (Fedi et al., 1999). The Dkk-1 protein appears to curtail the Wnt-induced signals at an upstream step, not by direct interaction with the fizzled receptor but by binding to an adjacent LDL-receptor-related protein 6 (LRP6) (Mao et al., 2001; Nusse, 2001).

Recently, the hDkk-1 gene was shown to be a direct target for the p53 tumor suppressor. In an elegant study, Wang et al. (2000) isolated the cDNA for hDkk-1 through subtractive hybridization from H1299 lung cancer cells after forcing them to express wild-type p53 in an otherwise p53-null background. Consistent with the specific induction of hDkk-1 by the tumor suppressor, the promoter of this gene was shown to contain consensus p53-binding sites upstream of the transcription start site (Wang et al., 2000). The p53-dependent regulation of hDkk-1 is intriguing, because the developmental functions performed by this gene do not fit the general paradigm of p53-related responses. Since p53 is a critical regulator of the DNA damage response and apoptosis following cell exposure to genotoxic agents, we hypothesized that hDkk-1 may function directly or indirectly in the apoptotic process. In this study, we show that DNA damage significantly increases hDkk-1 expression in a p53-independent manner and that hDkk-1-transfected cells are markedly sensitized for apoptosis by ceramide and alkylation induced DNA damage.


Cell line selection and dose/time-dependent induction of hDkk-1 mRNA by camptothecin in U87MG cells

Because mDkk-1, the mouse homolog of the human Dkk-1 gene plays a pivotal role in embryonic brain development (Glinka et al., 1998), and many brain tumors such as the medulloblastomas have retained embryonic cell properties (Scotting et al., 2000), we initially screened a panel of 10 human glioma cell lines to assess the level of hDkk-1 expression by Northern and Western analyses. A previous study reported marginal expression of hDkk-1 in a number of non-glial human tumor cell lines (Tate and Mitsuya, 1999). In contrast, our data showed that all the glioma cell lines, with the exception of U87MG expressed the hDkk-1 mRNA at moderate to high levels. In U87MG cells, faint bands of hDkk-1 mRNA and protein were detectable only after long film exposures; therefore, this cell line was chosen for studies of hDkk-1 response to DNA damage. Additionally, the fact that U87MG cells harbor a wild-type p53 (Van Meir et al., 1994), and possess an intact Wnt signaling pathway (Roth et al., 2000) provided an appropriate design for our study. Figure 1a shows the kinetics of hDkk–mRNA accumulation following U87MG cell exposure to the DNA strand-breaking drug camptothecin (CPT). A progressive increase of the gene transcripts over a 48-h period and its stabilization at later periods is evident (Figure 1A,B). The induction of the hDkk gene was also dependent on CPT concentration, with 300 nM drug eliciting the highest levels of mRNA during 24 h of incubation (Figure 1C,D). The accumulation of hDkk-1 observed over a long period in this study is analogous to the induction pattern of p53 tumor suppressor and other damage-responsive proteins (Jimenez et al., 1999), suggesting a putative DNA damage response function for the Dkk-1 protein.

Figure 1

Time- and dose-dependent induction of the hDkk-1 gene by camptothecin (CPT) in U87MG cells. (A) Gradual increase of hDkk-1 mRNA after 300 nm CPT treatment. At times specified, total RNA was isolated and subjected to Northern analysis. (B) Densitometric quantitation of the blots shown in panel A. (C) CPT concentration-dependent induction of hDkk-1 mRNA. Cells were exposed for 24 h to CPT at concentrations shown and then subjected to RNA isolation and Northern blot analysis. (D) The blots shown in panel C were quantitated by densitometry

Effect of other DNA damaging agents on hDkk-1 gene expression

We investigated the effects on hDkk-1 expression of hydrogen peroxide, UV radiation (254 nm), and the chemotherapeutic drug Adriamycin, which cause oxidative base damage, thymine dimerization and DNA strand breakage. All three caused a marked upregulation of hDkk mRNA levels at 24 h (Figure 2) in U87MG cells. This was accompanied by the accumulation of p21waf1 gene transcripts, suggesting a role for p53 in the damage-induced hDkk-1 expression at least in this cell line.

Figure 2

Different DNA-damaging treatments induce the hDkk-1 gene transcripts in U87MG cells. Tumor cells were exposed to 100 J of UV rays and incubated for 24 h. Cells were also treated with H2O2 (100 μM) and Adriamycin (5 μM) for 24 h followed by Northern analysis for detecting hDkk-1 mRNA. The membrane was sequentially reprobed with cDNA probes for p21waf1 and GAPDH

hDkk-1 gene responds to alkylation DNA damage by chemotherapeutic agents

The drugs that alkylate and cross-link DNA such as 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) and cisplatin are commonly used in brain tumor therapy. Therefore, we studied hDkk-1 response to the damage induced by these agents in two glioma cell lines, U87MG (wild-type p53) and MGR1, which has a p53 mutation at codon 155 (Ali-Osman et al., unpublished results). Both BCNU and cisplatin significantly increased hDkk-1 mRNA (Figure 3A) and protein (Figure 3B) levels by 24 h. The increase of hDkk mRNA in MGR1 cells suggests the involvement of p53-independent mechanisms in the regulation of this gene.

Figure 3

Alkylation and cisplatin-induced DNA damage enhances hDkk-1 gene expression in U87MG and MGR1 human glioma cells. (A) Tumor cells were treated with 50 μM BCNU or 5 μM cisplatin (cispt.) for 1 h. Cell monolayers were then resuspended in drug-free medium for 24 h followed by Northern analysis for hDkk-1. (B) Cells were treated with 50 μM BCNU for 24 h, and whole-cell extracts (40 μg protein) were processed for Western analysis of the hDkk-1 protein. The Northern and Western blots from three independent experiments were quantitated by densitometry, and the mean±s.e. of the data is shown under the figures

Human tumor cells with mutated or null p53 retain the ability to enhance hDkk-1 expression after DNA damage

To further clarify the relationship of wild-type p53 with the basal hDkk-1 gene transcription, and the induction of this gene after genomic damage, we chose six human cancer cell lines with defined p53 status, and quantitated hDkk-1 mRNA levels in control and camptothecin-treated cells (Figure 4). The panel included two cell lines each with wild-type, mutant and null p53. Northern analysis of hDkk-1 mRNA identified two distinct trends in these cell lines. First, consistent with the transcriptional activation of the hDkk-1 gene by p53 (Wang et al., 2000), the basal levels of hDkk-1 mRNA showed a clear association with a functional p53; thus, the cell lines, A1698 and HCT116, which harbor wild-type p53, expressed higher levels of gene transcripts, while the cells with p53 mutations (T98G and SUSMI), and null p53 cells (EJ, and LNZ-308 to some extent) had low basal levels of hDkk-1 mRNA. Second, in contrast to the inherent steady-state levels, the upregulation of the hDkk gene in response to camptothecin, was however p53-independent. While the extent of increase in hDkk-1 mRNA levels after DNA damage was significant in cell lines with mutant and null p53, such an effect was not evident in wild-type p53 containing cells. Taken together, these data confirm that p53-mediated transactivation is not required for the upregulation of the hDkk-1 gene following DNA damage.

Figure 4

hDkk-1 gene responds to DNA damage in human tumor cell lines irrespective of their p53 gene status. Northern analysis of hDkk-1 mRNA with (+) and without (−) camptothecin treatment for 24 h in cell lines with established p53 status is shown. The cell lines used were as follows: A1698 bladder carcinoma and HCT116 colon cancer with wild-type (Wt) p53, T98G glioblastoma and SUSM1, a carcinogen transformed liver fibroblast with mutant (mt) p53 and LNZ 308 glioblastoma and EJ bladder cancer with no p53 protein expression (null). The p53 status for all cell lines except LNZ 308 has been described by Johnson et al. (1994). LNZ 308 cells have been characterized by Van Meir et al. (1994). The blots were reprobed with cDNA probes for p21waf1 to determine the p53-dependent DNA-damage response in all cell lines; the p53 null-cells had higher basal levels of p21waf1 mRNA. Loading of RNA was assessed by rehybridization with GAPDH cDNA in cell lines with wt and mt p53

Stable transfection of hDkk-1 in U87MG cells promotes ceramide-mediated apoptosis

The dramatic induction of hDkk-1 by p53 in a p53-null cell line (Wang et al., 2000) and the enhancement of its expression by different types of DNA damage (Figures 1,2,3) implied that hDkk may act to promote cell cycle arrest and/or apoptosis. To examine this in detail, we engineered U87MG cells to express moderate levels of hDkk-1 protein by stable transfection, and selected three clones that expressed hDkk mRNA at different levels (Figure 5A) for further analysis. Because of the possibility that endogenous Wnt may interact with the Dkk protein to cause morphological changes (Fedi et al., 1999), the transfected cells were followed for five passages, during which no alterations in cellular morphology occurred. The growth kinetics and doubling time of the hDkk-1 gene-transfected clones remained essentially similar to those of the parental U87MG cells (not shown).

Figure 5

Expression of the hDkk-1 gene in U87MG cells after stable transfection and the anti-proliferative effect of ceramide on hDkk-1 expressing cell clones. (A) Northern analysis of hDkk-1 mRNA. Lane 1, U87MG cells transfected with the pcDNA3.1 vector alone. Lanes 2 to 4 represent clones 1, 2 and 3 respectively obtained after transfecting the hDkk-1 cDNA- pcDNA3.1 constructs. The data shows that clone 1 expressed the highest level of hDkk-1 mRNA, clone 2 the lowest level, with clone 3 showing an intermediate level of gene expression. (B) Time-dependent reduction in the survival of U87MG/hDkk-1, clone 2 after 50 μM ceramide treatment. U87MG cells transfected with the pcDNA3 vector alone served as control. In these assays, petri dishes were plated with similar numbers of cells. After 1 day, the cells were exposed to ceramide. At times specified, adherent and floating cells were pooled, the number of viable cells (those that excluded trypan blue) were counted, and the percentage of cell survival was calculated. (C) Relative survival of the three U87MG clones expressing the hDkk-1 after ceramide treatment. Similar numbers of cells in clones 1, 2 and 3 were plated along with the control cells (U87MG transfected with the vector alone) and exposed to 50 μM ceramide for 24 h. Cell viability and survival were assessed as described above

We used C2-ceramide (N-acetylsphingosine), a cell-permeable lipid second messenger, to activate the antiproliferative and apoptotic signals in hDkk-transfected U87MG cells, because ceramide has been shown to stimulate these processes in a variety of cell types (Hannun, 1996; Jayadev et al., 1995; Senchenkov et al., 2001). Ceramide is the basic structural unit of sphingolipids, and is generated in cells in response to diverse stimuli including p53 activation (Sawada et al., 2001), anticancer agents, cytokines, and stress (Senchenkov et al., 2001; Hannun and Luberto, 2000). Evidence suggests that these lipids modify a number of target proteins, including protein kinases and phosphatases, to induce a cascade of enzymatic and transcriptional activities; however, the exact mechanisms by which they trigger apoptotic signals are unclear. Exogenous ceramide induces efficient antiproliferative signals (Yoshimura et al., 1997; Zinda et al., 2001), and our initial experiments showed that ceramide at 50 μM was required for effective elimination of U87MG cells. Therefore, we examined the survival of U87MG/hDkk cells (clone 2, which expressed the least amount of Dkk-1 mRNA) in the presence of 50 μM ceramide over a 5-day period. This treatment resulted in a time-dependent reduction in cell number, with 20% of cells in the hDkk expressing clone surviving compared with the 75% survival in the vector controls (Figure 5B). Next, we tested the survival of all the three hDkk-1 expressing clones against 50 μM ceramide for 48 h (Figure 5C). Interestingly, the cell survival in different clones showed a close inverse correlation with the levels of hDkk-1 mRNA expressed; thus, 38% of cells survived in clone 1, which had the highest level of ectopic hDkk-1 gene expression, compared to 76% survival in clone 2, which had the lowest mRNA, and 50% survival in clone 3, which had intermediate levels of hDkk mRNA. The quantitative relationship between decreased cell number (shown in the next section as due to apoptosis) and hDkk-1 mRNA suggests that this developmentally important protein also participates directly or indirectly in ceramide-induced cell death.

U87MG cells expressing the hDkk-1 gene are sensitized to apoptosis after treatment with alkylating agents

To determine if the ceramide-induced decrease in cell number resulted from apoptotic cell death and whether the Dkk-1 gene stimulates apoptosis in response to genomic damage induced by alkylating agents used in cancer therapy, we subjected the U87MG clone expressing the highest Dkk mRNA levels (clone 1 in Figure 5A), and the control cells (U87MG transfected with the vector alone) to flow cytometry. These cells were treated with ceramide or DNA-damaging agents (BCNU, cisplatin), and vincristine, a non-DNA targeted drug that disrupts microtubule assembly (Downing, 2000); changes in cell cycle distribution as well as cell death, reflected by the sub G1 DNA content, were then analysed. Previously, we have shown that in U87MG cells, the alkylating drugs BCNU and cisplatin induce detectable levels of DNA damage at concentrations used in this study (Srivenugopal and Ali-Osman, 1996). Figure 6 depicts the representative histograms obtained 24 h after drug treatment; the effect of ceramide was examined at both 24 and 48 h. BCNU and cisplatin did not induce marked changes in cell cycle phase distribution at 24 h in control and Dkk-expressing U87MG cells. However, these drugs evoked reproducible and significant levels of cell death as reflected by sub G1 DNA content of 19.6% for BCNU and 11.5% for cisplatin (Figure 6). In contrast, 10 μM or higher vincristine did not induce significant levels of apoptosis in these cells. Ceramide elicited a gradual but extensive cell death with 16.3 and 88% of cells in apoptotic phase at 24 and 48 h, respectively. The flow cytometric data correlated well with the results obtained with the electrophoretic DNA ladder assay (not shown). Consistent with the drug-induced apoptosis (Figure 6), the clonogenic assay showed U87MG/hDkk cells to be 2.1-fold and 1.6-fold more sensitive to cisplatin and BCNU respectively than U87MG cells transfected with the vector DNA (not shown). These results imply that hDkk-1 may function to amplify or transduce the pro-apoptotic signals emanating from DNA damage or ceramide treatment.

Figure 6

hDkk-1 sensitizes U87MG cells for apoptosis after exposure to DNA damaging chemotherapy drugs and ceramide; analysis by flow cytometry. U87MG parent cells and hDkk-1 expressing U87MG cells were treated for 24 h with 50 μM BCNU or 5 μM cisplatin or 5 μM vincristine, which does not affect DNA but disrupts the microtubule assembly. Ceramide-treated cells were harvested at both 24 and 48 h. Cells were then stained with propidium iodide and subjected to flow cytometry. Appropriate windows in the red channel were selected to differentiate between cells with sub G1 DNA content (hypodiploid) and cells in different phases of cell cycle distribution. Proportion of cells with subG1 DNA content in each treatment is indicated in the histograms. Similar results were obtained in three independent experiments. Previously, we have shown that U87MG cells require >24 h to show significant arrest in S and G2 phases after BCNU or cisplatin treatments (Srivenugopal and Ali-Osman, 1996). Cispt., cisplatin

Alterations of Bax and Bcl2 protein expression in hDkk-1 stimulated cell death

The pro-apoptotic and anti-apoptotic activities of the protoncogene proteins Bax and Bcl2 respectively, along with their related partners, occupy a central position in the apoptotic response leading to mitochondrial membrane permeabilization and caspase activation (Robertson and Orrenius, 2000). To determine whether these proteins are involved in hDkk-1 mediated cell death, we performed Western analysis of Bax and Bcl2 in cells treated with C2-ceramide for 30 h. Bcl2 showed a 70% decrease while the Bax protein levels were consistently upregulated in U87MG/hDkk cells under these conditions (Figure 7A). The Bax/Bcl2 ratio increased progressively and reproducibly after ceramide exposure (Figure 7B). Taken together, these results suggest that an altered balance in Bcl2 and Bax protein content occurs in ceramide treated hDkk-1 expressing cells, and may contribute to the apoptosis observed.

Figure 7

Western analysis of Bax and Bcl2 protein levels in hDkk-1 expressing U87MG cells. (A) Whole-cell extracts (40 μg protein) were electrophoresed, immunoblotted, and probed with monoclonal antibodies to Bax and Bcl2 proteins. Actin was used as the loading control. Similar results were observed in three independent experiments. (B) The Western blots from different experiments were quantitated by densitometry and the relative changes in the levels of Bax and Bcl-2 proteins were assessed

Reduction of telomeric signal in U87MG/hDkk cells treated with ceramide and anticancer agents

Maintenance of a constant and stable telomeric length as reflected by high levels of telomeric associations and intense telomeric signals is a hallmark of proliferating cells (Nagele et al., 2001; De Lange, 1995). On the corollary, dysfunctions involving altered telomerase activity and/or decreased telomeric lengths have been frequently observed in apoptotic and malignant cells that display genomic instability. Reduced telomeric signals have been observed in human tumor cells after exposure to DNA damaging agents (Ishibashi and Lippard, 1998; Multani et al., 1999, 2000) and in cells infected with p53 adenoviral vectors (Mukhopadhyay et al., 1998). Therefore, we examined the telomeric integrity in U87MG/hDkk cells after drug treatment. Cells were treated with ceramide and BCNU for 24 h and the telomeric DNA was assessed by quantitative fluorescence in situ hybridization (FISH) using the telomere-specific DNA probe. Compared to the controls (untreated U87MG/hDkk cells) which showed 1.78% telomeric DNA in relation to the nuclear size, the ceramide and the BCNU treated cells showed 1.08 and 0.8% respectively of telomeric DNA (Figure 8). Cells treated with ceramide for the same length of time showed 17% apoptosis (Figure 5, row 5), thereby suggesting that a reduction of telomeric repeats (TTAGGG)n accompanies the apoptotic process promoted by the hDkk-1 protein.

Figure 8

Significant reduction in telomeric signals accompany the ceramide and BCNU-promoted apoptosis in U87MG cells expressing hDkk-1 gene. U87MG/hDkk cells cultured in T-75 cm2 flasks were treated with 50 μM ceramide or 50 μM BCNU for 24 h, and quantitative FISH assays using a peptide-linked telomeric DNA probe were performed as described in Materials and methods. A minimum of 50 cells were photographed using a Nikon photomicroscope; the telomeric signals observed in interphase nuclei were quantitated and the mean calculated. This data was expressed as per cent telomeric DNA in relation to the total nuclear area, and values are shown on the right side. The microphotographs show markedly reduced signals in the interphase nuclei of ceramide (B) and BCNU (C) treated cells compared to untreated U87MG/hDkk cells (A)

Inhibition of Wnt-signaling by hDkk-1 and ceramide induced apoptosis in C57MG cells

To investigate the possibility that Dkk-1 interaction with the Wnt pathway may have contributed to the promotion of cell death, the Wnt-2 gene expressing C57MG mammary epithelial cells (Shimizu et al., 1997) was used as the experimental model. This cell line is very responsive to Wnt induced morphological and biochemical changes, and has been widely used for engineering stable expression of the Wnt genes (Wong et al., 1994; Naylor et al., 2000). The transformation potential of the Wnt-2 gene, and its ability to increase uncomplexed β-catenin in these cells has been demonstrated (Shimizu et al., 1997). We transfected the Wnt-2 expressing C57MG cells with the hDkk-1 cDNA to obtain a cell clone that expressed both the proteins. The dual transfectant had slightly lower proliferation rate than the parent cells (Fedi et al., 1999). To confirm the activation of the Wnt signaling and its negative regulation by Dkk-1 in our model, we determined the abundance of β-catenin protein in the cytosol and membrane fractions of C57MG cells by Western analysis. An increase in the uncomplexed form of β-catenin without significant alterations of this protein in the membrane pool is a hallmark of Wnt activation (Shimizu et al., 1997; Fedi et al., 1999). As shown in Figure 9A, Dkk-1 expression had no effect on the levels of the membrane-bound β-catenin levels in C57MG–Wnt2 cells. However, the cytosolic β-catenin levels were consistently diminished when the Dkk1 gene was coexpressed with the Wnt-2 gene (Figure 9a). These data agree with the proven interference of the Wnt pathway of Dkk-1 (Fedi et al., 1999), and provided a model for testing the biological effects associated with Wnt inhibition.

Figure 9

Inhibition of Wnt-2 induced signaling by hDkk-1 in C57MG cells and the consequent effect on ceramide-induced apoptosis. (A) Modulation of cytosolic and membrane bound β-catenin levels by stable transfection of hDkk-1 in Wnt-2 expressing C57MG mammary epithelial cells. C57MG cells with ectopic Wnt expression (Shimizu et al., 1997) were stably transfected with hDkk-1 cDNA. Homogenates from C57 cells expressing the Wnt alone and Wnt together with Dkk-1 were fractionated by ultracentrifugation to obtain the cytosolic and membrane fractions according to Shimizu et al. (1997). 30 μg protein samples from each group were analysed by Western blotting with a β-catenin monoclonal antibody. Protein loading was checked by reprobing the blots with an actin antibody. (B) Comparison of ceramide-induced apoptosis in C57 cells expressing the Wnt-2 alone, and Wnt-2+hDkk-1 proteins together. Sub-confluent cultures were treated with 20 μM ceramide for 24 h. Adherent and floating cells were pooled and subjected to flow cytometry. The bar graph represents the per cent of cells with sub G1 DNA content computed from four independent experiments (mean±s.e.). *Significant at P<0.005

Next, we determined the effect of Wnt and Dkk-1 gene coexpression on apoptosis. The C57–Wnt2 cells, with and without Dkk-1 expression were exposed to 20 μM ceramide for 24 h and the sub G1 DNA content was quantitated by flow cytometry. In the absence of ceramide, inhibition of Wnt signaling had only marginal effect on cell death, with approximately threefold increase in the number of apoptotic cells. However, the exposure of cells expressing both Wnt-2 and Dkk-1 to ceramide resulted in a significant and reproducible increase (6–7-fold in different experiments) of apoptotic cell fraction (Figure 9B). These findings suggest that suppression of Wnt-induced signals by hDkk-1 may form the basis for the pro-apoptotic function of the latter gene.


The results presented in this report indicate that the hDkk-1 protein, an established inhibitor of the Wnt signaling pathway, has a pro-apoptotic function as well. hDkk-1 is transcriptionally activated by p53. We demonstrated that this gene, like p53 itself, is inducible to a significant extent by different types of DNA damage including base alkylation, strand breaks, interstrand crosslinks, and lesions with oxidative damage. Initially, our findings of Dkk-1 induction in U87MG cells which contain a wild-type p53, and the fact that the cDNA for the Dkk-1 was cloned from a library of p53-inducible transcripts (Wang et al., 2000), suggested that the damage response of this gene may be exclusively p53-dependent. However, the persistence of this response in a panel of tumor cell lines with no p53 expression, and those expressing mutant p53 clearly suggests that p53-independent mechanisms also regulate Dkk-1 expression after DNA damage. Our studies also imply that wild-type p53 may control the basal transcription of the Dkk-1 gene.

The direct activation of hDkk-1 gene by p53 has been explained as a negative regulatory step in the Wnt signaling circuit for orchestrating appropriate embryonic development (Wang et al., 2000; Ueda et al., 2001). This is consistent with the characteristics of Wnt pathway, which exemplifies a coordinated interplay of oncogenes and tumor suppressor genes (Polakis, 2000). Thus, components of Wnt pathway such as the β-catenin and APC genes, undergo mutations and result in a constitutive Wnt signaling and enhanced expression of the Wnt target genes (cyclin D1, c-myc and others) to establish an oncogenic phenotype (Tetsu and McCormick, 1999). In contrast, the p53 tumor suppressor counteracts the oncogenic and proliferative effects of Wnt signaling through a positive or negative regulation of the genes involved, and through other biochemical events. For example, p53 accumulation resulting from DNA damage or its enforced expression downregulates the β-catenin protein levels (Sadot et al., 2001). Overexpression of mutant β-catenin suppresses the p53-dependent transactivation by competitively binding to CBP/p300 (Miyagishi et al., 2000) as well. Other reports of functional antagonism/cooperation between the Wnt and p53 pathways include (i) the ability of GSK3β to phosphorylate p53 (Turenne and Price, 2001, ii) p53-dependent transcriptional regulation of the APC gene (Jaiswal and Narayan, 2001), and (iii) the activation of c-myc leading to p53 suppression and inhibition of apoptosis (Henriksson et al., 2001). Our findings that hDkk-1 also serves a pro-apoptotic function introduces a new twist in the regulatory network between the Wnt and p53 pathways. For instance, the damage-increased expression of Dkk-1 (mediated by p53-dependent or -independent mechanisms) could contribute significantly to p53-mediated apoptosis. This hypothesis is consistent with the observations that several p53-induced genes, whose functions remain undefined, however, induce apoptosis when overexpressed (Israeli et al., 1997; Deng and Wu, 2000).

We demonstrated the association of hDkk-1 induced cell death with telomere reduction. The telomeric loss appears to be largely independent of the general DNA fragmentation that occurs in apoptotic cells, and merits attention in the context of the alkylating drugs used in the study. The reduction of telomeric signals by ceramide and BCNU has not been reported previously, although, cisplatin has been shown to cause such an effect (Ishibashi and Lippard, 1998; Ishii et al., 2000). It is significant to note that BCNU has a higher affinity for GC-rich DNA and guanine O6-alkylations induced by the drug lead to the formation of cytotoxic interstrand cross-links in DNA (Srivenugopal and Ali-Osman, 1990). Therefore, a direct damage of telomeric DNA (Kruk et al., 1995), the inhibition of the telomerase activity by DNA alkylating drugs (Burger et al., 1997), and/or a transcriptional repression of the telomerase by wild-type p53 (Kanaya et al., 2000) may all contribute to the observed erosion of telomeric repeats in U87MG cells.

Evidence presented in this study (Figure 9) indicates that the ability of Dkk-1 to inhibit Wnt-induced signaling may very well be the basis for its apoptotic potential. The Wnt ligands and other proteins associated with Wnt signaling, by themselves, do not appear to be pro-apoptotic. In fact, the ability of the Wnt pathway to promote cell survival and oncogenesis reflects the operation of anti-apoptotic mechanisms. A recent study showed that Wnt-1 induced signals are indeed anti-apoptotic, and they function to curtail cell death by preventing cytochrome c release and caspase activation in cells exposed to chemotherapeutic drugs (Chen et al., 2001). These findings support the argument that inhibition of Wnt signaling by Dkk-1 may induce sensitivity for apoptosis, as observed in our studies (Figure 9). The apoptotic sensitization induced by Dkk-1 in U87MG cells occurred in the absence of ectopic Wnt expression, and was of higher extent than that observed in C57MG cells; the reason for this discrepancy is unclear, but may relate to the differences in cell types and a differential modulation of Wnt pathway components.

The pro-apoptotic nature of the hDkk-1 resembles similar activities shown by the soluble frizzled-related proteins (SFRPs) and secreted apoptosis-related proteins (SARPs). SARPs and SFRPs contain N-terminal cysteine-rich domains (CRDs) similar to that found in the frizzled receptor and are known to impart pro-apoptotic or anti-apoptotic properties on cells when overexpressed (Leyns et al., 1997; Finch et al., 1997; Wolf et al., 1997; Melkonyan et al., 1997). The hDkk-1 protein also harbors two CRDs near the central and c-terminus of the polypeptide (Fedi et al., 1999; Glinka et al., 1998). The SARPs and SFRPs appear to interfere with Wnt signaling by binding to either Wnt proteins or its receptor, and promote or inhibit cell death (Melkonyan et al., 1997; Zhou et al., 1998). Further understanding of the molecular interactions of Dkk-1 with the Wnt components, and the resulting changes in gene expression may shed light on the nature of pro-apoptotic signals involved. Whether this novel function of Dkk-1 has a role to play in neural development is an important question. Also, any therapeutic significance our findings may have for brain tumors and for targeting cancers that harbor an activated Wnt pathway merits further study.

Materials and methods

Cell lines and stable transfection

The human glioblastoma cell lines used in this study were, U87MG, T98G (ATCC, Rockville, MD, USA), MGR1 (established from a primary specimen in our laboratory), and LNZ 308 (kindly provided by Dr Erwin G Van Meir, Emory University, Atlanta, GA, USA; Van Meir et al., 1994). SUSMI, a transformed liver fibroblast cell line, and the human bladder cancer cell lines, A1698, EJ were provided by Dr J Smith, Baylor College of Medicine, Houston, TX, USA (Burkhart et al., 1999). HCT116 human colon cancer cell line was purchased from ATCC. The C57MG mammary epithelial cell line stably transfected with the Wnt-2 gene (Shimizu et al., 1997) was obtained from Dr J Kitajewski of Columbia University, New York, NY, USA.

All cell lines were cultured in DMEM supplemented with 10% bovine serum and antibiotics at 37°C in a 5% CO2 and air atmosphere. U87MG cells and C57MG–Wnt2 cells were stably transfected to express the human Dkk-1 gene. The cDNA for hDkk (Tate and Mitsuya, 1999; NCBI accession number SEG_AB020314S) was amplified using the primers 5′-GCGAATTCCTTCTGAGATGATG-3′ (sense) and 5′-GCGGTACCGGATCCGTGTCTCTGACA-3′ (antisense), and was cloned into pcDNA3.1/myc-His vector (Invitrogen, Carlsbad, CA) at EcoRI and BamHI sites. Cells cultured in 35-mm petri dishes were transfected with either pcDNA3.1/myc-His vector or the same vector containing the hDkk-1 cDNA using lipofectamine. The plasmids also contained a neomycin marker that allowed selection of neomycin-resistant colonies 2 weeks after transfection. Colonies demonstrating hDkk-1 by Northern analysis were expanded and selected for further studies.

UV irradiation, ceramide, and drug treatment

For UV treatment, cells were suspended in 500 μl of phosphate buffered saline in 35-mm dishes, and irradiated at 254 nm using a germicidal lamp at a fluence rate of 1 J/m2/s. Cells were re-fed with 10% serum in DMEM after irradiation. The drugs camptothecin, BCNU, cisplatin, vincristine, and Adriamycin were obtained from the Drug Development Branch of the National Cancer Institute (Bethesda, Maryland, USA), and stock solutions were prepared in ethanol or dimethyl sulfoxide just before addition to cell cultures. C2-ceramide was obtained from Alexis Biochemicals (San Diego, CA, USA) and dissolved in dimethyl sulfoxide.

Northern blot and Western blot analysis

Total RNA was isolated from cells using the TRIzol reagent as recommended by the manufacturer (GIBCO BRL, Gaithersburg, MD, USA). RNA samples (15 μg) were electrophoresed on 1% agarose-formaldehyde gels, blotted on to nylon membranes, and hybridized with 32P-labeled full-length cDNA probes for hDkk and p21waf1 according to standard procedures (Ausubel et al., 1987). The blots were stripped and reprobed with actin cDNA to control for RNA loading. For Western blot analysis, whole-cell lysates (40 μg protein) were subjected to SDS–PAGE on 12% gels, and the proteins electrotransferred to PVDF membranes. The membranes were blocked with 5% BSA and treated with polyclonal antibodies to hDkk-1 protein (Fedi et al., 1999), or monoclonal antibodies to β-catenin (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) and actin (Chemicon International, Temecula, CA, USA). The immunocomplexes were visualized by chemiluminescence.

Analysis of cell cycle progression and apoptosis by flow cytometry

U87MG cells expressing the hDkk-1 gene were treated with ceramide (50 μM) or the chemotherapeutic agents BCNU (50 μM), cisplatin (10 μM), and vincristine (5 μM) for 24 h. Adherent and floating cells were combined, washed with cold PBS, and fixed in 70% ethanol at 4°C. The cells were pelleted and treated with ribonuclease A (200 μg/ml) and stained with propidium iodide (50 μg/ml) (Srivenugopal and Ali-Osman, 1997). The apoptotic cell population and the proportion of viable cells in different phases of the cell cycle were determined by using an Epics Profile II Flow Cytometer and CELL Quest Software (Becton Dickinson, San Jose, CA, USA). The hypodiploid DNA peak in apoptotic cells was easily distinguishable from the narrow peak of diploid/tetraploid DNA present in cycling cells in the red fluorescence channels.

Quantitative fluorescence in situ hybridization (Q-FISH)

Reduced telomeric signal as a salient feature of apoptotic cells was investigated (Mukhopadhyay et al., 1998; Multani et al., 2000). Cells grown in T-75 cm2 flasks were treated with ceramide (50 μM) or BCNU (50 μM) for 24 h. Cells were exposed to a hypotonic solution (0.075 m KCl) for 20 min and then fixed in methanol/acetic acid mixture (3 : 1 v/v). Air-dried cytological preparations were made (Pathak, 1976), and Q-FISH was performed using a CY-3-conjugated peptide nucleic acid (PNA) probe (Dako Corporation, Carpenteria, CA, USA) as described previously (Multani et al., 2000). The percentage of telomeric area in the interphase nuclei was quantitated by using the Metaview imaging system (Universal Imaging Co., Westchester, PA, USA). For each sample, at least 50 nuclei were analysed for calculating the mean and median percentages of telomeric areas relative to the total nuclear areas (Multani et al., 2000).

Statistical analysis

All experiments were performed in triplicate and repeated at least once. Data from representative experiments are presented and expressed as means and s.e.m. Significance was assessed by using the Students t-test.



adenomatous polyposis coli




fluorescence in situ hybridization


glycogen synthase kinase 3


human Dkk-1


  1. Ausubel FM, Brent R, Kingston RE, Moore DD, Smith JA, Seidman JG, Struhl K (eds) . 1987 Current Protocols in Molecular Biology New York: Wiley Interscience

  2. Barker N, Morin PJ, Clevers H . 2000 Adv. Cancer Res. 77: 1–24

  3. Burger AM, Double JA, Newell DR . 1997 Eur. J. Cancer 33: 638–644

  4. Burkhart BA, Alcorta DA, Chiao C, Isaacs JS, Barrett JC . 1999 Exp. Cell Res. 247: 168–175

  5. Chen S, Guttridge DC, You Z, Zhang Z, Fribley A, Mayo MW, Kitajewski J, Wang C-Y . 2001 J. Cell Biol. 152: 87–96

  6. Dale TC . 1998 Biochem. J. 329: 209–223

  7. De Lange T . 1995 Telomere dynamics and genome instability in human cancer Telomeres Cold Spring Harbor, NY: Cold Spring Harbor Press pp 265–293

  8. Deng Y, Wu X . 2000 Proc. Natl. Acad. Sci. USA 97: 12050–12055

  9. Downing KH . 2000 Annu. Rev. Cell Dev. Biol. 16: 89–111

  10. Fedi P, Bafico A, Nieto Soria A, Burgess WH, Miki T, Bottaro DP, Kraus MH, Aaronson SA . 1999 J. Biol. Chem. 274: 19465–19472

  11. Finch PW, He X, Kelley MJ, Uren A, Schaudies RP, Popescu NC, Rudikoff S, Aaronson SA, Varmus HE, Rubin JS . 1997 Proc. Natl. Acad. Sci. USA 94: 6770–6775

  12. Glinka A, Wu W, Delius H, Monaghan AP, Blumenstock C, Niehrs C . 1998 Nature 391: 357–362

  13. Hannun YA . 1996 Science 274: 1855–1859

  14. Hannun YA, Luberto C . 2000 Trends Cell Biol. 10: 73–80

  15. Hart MJ, de los Santos R, Albert IN, Rubinfeld B, Polakis P . 1998 Curr. Biol. 8: 573–581

  16. Hayashi S, Rubinfeld B, Souza B, Polakis P, Wieschaus E, Levine AJ . 1997 Proc. Natl. Acad. Sci. USA 94: 242–247

  17. Henriksson M, Selivanova G, Lindstrom M, Wiman KG . 2001 Apoptosis 6: 133–137

  18. Howe LR, Subbaramaiah K, Chung WJ, Dannenberg AJ, Brown AM . 1999 Cancer Res. 59: 1572–1577

  19. Ishii K, Yang W-L, Cvijic ME, Kikuchi Y, Nagata I, Chin K-V . 2000 Exp. Cell Res. 255: 95–101

  20. Ishibashi T, Lippard SJ . 1998 Proc. Natl. Acad. Sci. USA 95: 4219–4223

  21. Israeli D, Tessler E, Haupt Y, Elkeles A, Wilder S, Amson R, Telerman A, Oren M . 1997 EMBO J. 16: 4384–4392

  22. Jaiswal AS, Narayan S . 2001 J. Biol. Chem. 276: 18193–18199

  23. Jayadev S, Liu B, Bielawska AE, Lee JY, Nazaire F, Pushkareva MY, Obeid LM, Hannun YA . 1995 J. Biol. Chem. 270: 2047–2052

  24. Jimenez GS, Khan SH, Stommel JM, Wahl GM . 1999 Oncogene 18: 7656–7665

  25. Johnson M, Dimitrov D, Vojta P, Barrett JC, Noda A, Pereira-Smith OM, Smith JR . 1994 Mol. Carcinog. 11: 59–64

  26. Kanaya T, Kyo S, Hamada K, Takakura M, Kitagawa Y, Harada H, Inoue M . 2000 Clin. Cancer Res. 6: 1239–1247

  27. Kazanskaya O, Glinka A, Niehrs C . 2000 Development 127: 4981–4992

  28. Kruk PA, Rampino NJ, Bohr VA . 1995 Proc. Natl. Acad. Sci. USA 92: 258–262

  29. Krupnik VE, Sharp JD, Jiang C, Robison K, Chickering TW, Amaravadi L, Brown DE, Guyot D, Mays G, Leiby K, Chang B, Duong T, Goodearl AD, Gearing DP, Sokol SY, McCarthy SA . 1999 Gene 238: 301–313

  30. Leyns L, Bouwmester T, Kim SH, Piccolo S, De Robertis EM . 1997 Cell 88: 747–756

  31. Mao B, Wu W, Li Y, Hoppe D, Stannek P, Glinka A, Niehrs C . 2001 Nature 411: 255–256

  32. McGrew LL, Takemaru K, Bates R, Moon RT . 1999 Mech. Dev. 87: 21–32

  33. Melkonyan HS, Chang WC, Shapiro JP, Mahadevappa M, Fitzpatrick PA, Kiefer MC, Tomei LD, Umansky SR . 1997 Proc. Natl. Acad. Sci. USA 94: 13636–13641

  34. Miyagishi M, Fuji R, Hatta M, Yoshida E, Araya N, Nagafuchi A, Ishihara S, Nakajima T, Fukamizu A . 2000 J. Biol. Chem. 275: 35170–35175

  35. Molenaar M, Destree O . 1999 Int. J. Dev. Biol. 43: 675–680

  36. Moon RT, Brown JD, Yang-Snyder JA, Miller JR . 1997 Cell 88: 725–728

  37. Multani AS, Chandra J, McConkey DJ, Sen S, Cabral TF, Pathak S . 1999 Oncol Res. 11: 455–460

  38. Multani AS, Ozer M, Narayan S, Kumar V, Chandra J, McConkey D, Newman RA, Pathak S . 2000 Neoplasia 2: 339–345

  39. Mukhopadhyay T, Multani AS, Roth JA, Pathak S . 1998 Oncogene 17: 901–906

  40. Nagele RG, Velasco AQ, Anderson WJ, McMahon DJ, Thomson Z, Fazekas J, Wind K, Lee H-Y . 2001 J. Cell Sci. 114: 377–388

  41. Naylor S, Smalley MJ, Robertson D, Gusterson BA, Edwards PA, Dale TC . 2000 J. Cell Sci. 113: 2129–2138

  42. Nusse R . 2001 Nature 411: 255–256

  43. Pathak S . 1976 J. Reprod. Med. 17: 25–28

  44. Polakis P . 2000 Genes Dev. 14: 1837–1851

  45. Robertson JD, Orrenius B . 2000 Crit. Rev. Toxicol. 30: 609–627

  46. Roth W, Wild-Bode C, Platten M, Grimmel C, Melkonyan HS, Dichgans J, Weller M . 2000 Oncogene 19: 4210–4220

  47. Sadot E, Geiger B, Oren M, Ben-Ze'ev A . 2001 Mol. Cell. Biol. 21: 6768–6781

  48. Sakanaka C, Weiss JB, Williams LT . 1998 Proc. Natl. Acad. Sci. USA 95: 3020–3023

  49. Sawada M, Nakashima S, Kiyono T, Nakagawa M, Yamada J, Yamakawa H, Banno Y, Shinoda J, Nishimura Y, Nozawa Y . 2001 Oncogene 20: 1368–1378

  50. Schneider VA, Mercola M . 2001 Genes Dev. 15: 304–315

  51. Scotting PJ, Thompson SL, Punt JA, Walker DA . 2000 Childs Nerv. Syst. 16: 261–267

  52. Senchenkov A, Litvak DA, Cabot MC . 2001 J. Natl. Cancer Inst. 93: 347–357

  53. Shimizu H, Julius MA, Giarre M, Zheng Z, Brown AM, Kitajewski J . 1997 Cell Growth Differ. 8: 1349–1358

  54. Shutman M, Zhurinsky J, Simcha I, Albanese C, D'Amico M, Petsell RG, Ben-Ze'ev A . 1999 Proc. Natl. Acad. Sci. USA 96: 5522–5527

  55. Srivenugopal KS, Ali-Osman F . 1997 Int. J. Oncol. 11: 1251–1256

  56. Srivenugopal KS, Ali-Osman F . 1996 Proc. Amer. Assoc. Cancer Res. 37: p 5

  57. Srivenugopal KS, Ali-Osman F . 1990 Biochem. Pharmacol. 40: 473–479

  58. Tate G, Mitsuya T . 1999 J. Biochem. Mol. Biol. Biophys. 3: 239–242

  59. Tetsu O, McCormick F . 1999 Nature 398: 422–426

  60. Turenne GA, Price BD . 2001 BMC Cell Biol. 2: 12–16

  61. Ueda Y, Hijikata M, Takagi S, Takada R, Takada S, Chiba T, Shimotohno K . 2001 Biochem. Biophys. Res. Commun. 283: 327–333

  62. Van Meir EG, Kikuchi T, Tada M, Li H, Diserens AC, Wojcik BE, Huang HJ, Friedman T, de Tribolet N, Cavanee WK . 1994 Cancer Res. 54: 649–652

  63. Wang J, Shou J, Chen X . 2000 Oncogene 19: 1843–1848

  64. Wodarz A, Nusse R . 1998 Annu. Rev. Cell. Dev. Biol. 14: 59–88

  65. Wolf V, Ke G, Dharmarajan AM, Bielke W, Artuso L, Saurer S, Friis R . 1997 FEBS Lett. 417: 385–389

  66. Wong GT, Avin BJ, McMahon AP . 1994 Mol. Cell. Biol. 14: 6278–6286

  67. Yoshimura S, Sakai H, Ohguchi K, Nakashima S, Banno Y, Nishimura Y, Sakai N, Nozawa Y . 1997 J. Neurochem. 69: 713–720

  68. Zhou Z, Wang J, Han X, Zhou J, Linder S . 1998 Int. J. Cancer 78: 95–99

  69. Zinda MJ, Vlahos CJ, Lai MT . 2001 Biochem. Biophys. Res. Commun. 280: 1107–1115

Download references


We gratefully acknowledge the supply of cell lines from Drs EG Van Meir, J Kitajevski, and J Smith. This work was supported by grants from the National Institutes of Health (CA-74321), Pediatric Brain Tumor Foundation of the United States, and the Association for Research of Childhood Cancer to KS Srivenugopal and a NIH grant (CA-79644) to F Ali-Osman.

Author information

Correspondence to Kalkunte S Srivenugopal.

Rights and permissions

Reprints and Permissions

About this article


  • Wnt antagonist
  • apoptosis
  • p53
  • alkylating agents
  • brain tumors

Further reading