Advertisement
Cell, Tumor, and Stem Cell Biology

Melanoma Proliferation and Chemoresistance Controlled by the DEK Oncogene

Michael S. Khodadoust, Monique Verhaegen, Ferdinand Kappes, Erica Riveiro-Falkenbach, Juan C. Cigudosa, David S.L. Kim, Arul M. Chinnaiyan, David M. Markovitz and María S. Soengas
DOI: 10.1158/0008-5472.CAN-09-1063 Published August 2009

Abstract

Gain of chromosome 6p is a consistent feature of advanced melanomas. However, the identity of putative oncogene(s) associated with this amplification has remained elusive. The chromatin remodeling factor DEK is an attractive candidate as it maps to 6p (within common melanoma-amplified loci). Moreover, DEK expression is increased in metastatic melanomas, although the functional relevance of this induction remains unclear. Importantly, in other tumor types, DEK can display various tumorigenic effects in part through its ability to promote proliferation and inhibit p53-dependent apoptosis. Here, we report a generalized up-regulation of DEK protein in aggressive melanoma cells and tumors. In addition, we provide genetic and mechanistic evidence to support a key role of DEK in the maintenance of malignant phenotypes of melanoma cells. Specifically, we show that long-term DEK down-regulation by independent short hairpin RNAs resulted in premature senescence of a variety of melanoma cell lines. Short-term abrogation of DEK expression was also functionally relevant, as it attenuated the traditional resistance of melanomas to DNA-damaging agents. Unexpectedly, DEK short hairpin RNA had no effect on p53 levels or p53-dependent apoptosis. Instead, we identified a new role for DEK in the transcriptional activation of the antiapoptotic MCL-1. Other MCL-1–related factors such as BCL-2 or BCL-xL were unaffected by changes in the endogenous levels of DEK, indicating a selective effect of this gene on the apoptotic machinery of melanoma cells. These results provide support for DEK as a long sought-after oncogene mapping at chromosome 6, with novel functions in melanoma proliferation and chemoresistance. [Cancer Res 2009;69(16):6405–13]

  • chemoresistance
  • DEK
  • MCL-1
  • melanoma
  • senescence

Introduction

Metastatic melanomas are invariably chemoresistant and carry a grim prognosis ( 1). Complicating drug development, melanoma cells accumulate a plethora of changes in gene expression with the potential to unleash uncontrolled proliferation, evade senescence programs, and inhibit death pathways at multiple levels ( 2). A hierarchical mapping of these alterations (to distinguish central nodes in tumor cell maintenance from inconsequential byproducts of tumor development) has remained elusive. Equally unclear are the mechanisms underlying common melanoma-associated events that have been ascribed neither to mutations nor to genetic or epigenetic alterations. This is particularly relevant for the antiapoptotic members of the BCL-2 family (BCL-2, BCL-xL, and MCL-1). These proteins are invariably overexpressed in melanoma cells and play key roles in their chemoresistance ( 3). BCL-2 up-regulation may be a consequence of MITF gene amplification ( 4). BCL-xL and MCL-1 can be under the control of the mitogen-activated protein kinase or nuclear factor-κB signaling cascades ( 5). However, neither MITF, mitogen-activated protein kinase, nor nuclear factor-κB inhibition abrogates the expression of these BCL-2 family members ( 6– 8), suggesting alternative, and yet unknown, mechanisms of regulation of the apoptotic machinery in melanoma cells.

Alteration of chromosome 6 is one of the most consistent cytogenetic changes in melanoma ( 9). In particular, gain of the 6p arm is prevalent in different manifestations of the disease, including cutaneous (both sun-exposed and non–sun-exposed), acral, mucosal, and uveal melanomas ( 10, 11). Furthermore, temporal clustering of karyotypic changes in melanoma indicates that 6p gain may represent an early event in the acquisition of chromosomal imbalances ( 12). Adding to the clinical relevance of chromosome 6, 6p copy gains have been associated with poor prognosis in ocular melanomas ( 11). Therefore, it has been long suggested that this genomic region could contain one or multiple pro-oncogenes ( 9, 13, 14). Although the minimal region of 6p gain has not been described in melanoma, several reports in retinoblastoma and bladder cancer have identified a narrow region of gain in 6p22-23 ( 15– 18). This region notably contains the chromatin remodeling factor DEK.

DEK was originally discovered as the target of a chromosomal translocation event [t(6;9)(p23;q34)] in a subset of acute myeloid leukemias ( 19). Subsequent studies have repeatedly identified DEK as a frequently overexpressed gene in several neoplasms ( 16, 20– 22). Furthermore, we and others have shown that DEK has effects on mRNA splicing, transcriptional control, DNA damage repair, differentiation, cell viability, and cell-to-cell signaling ( 23– 31). Pro-oncogenic roles of DEK are supported by its ability to inhibit p53-mediated apoptosis ( 32), cooperate with the viral oncogenes E6 and E7 to overcome senescence ( 29), and promote HRAS-driven keratinocyte transformation ( 28). The mechanism(s) through which DEK mediates its effects, particularly in the context of cellular oncogenes, is not well understood. DEK is a structurally unique protein ( 33) that binds preferentially to euchromatin, although it can also be found within interchromatin granule clusters ( 34). Consequently, it is unclear which DEK targets have a causative role in tumor development and which are passengers of global changes of chromatin architecture.

The contribution of DEK to melanoma progression and chemoresistance has yet to be analyzed. Increased DEK mRNA and protein has been reported in a small number of melanoma cell lines or tissues ( 21, 35). However, DEK function is highly dependent on post-translational modifications ( 34, 36). Therefore, it is unclear whether altered DEK expression is a generalized feature of advanced melanomas and whether this gene has a functional contribution to melanoma cell maintenance and drug response. To assess these questions, we used tissue microarrays to compare the expression of DEK among benign nevi and malignant melanoma specimens. In addition, the role of DEK was defined in a panel of aggressive melanoma cell lines that recapitulate the most frequent melanoma-associated defects. Our studies establish DEK, a gene mapping to the genetically unstable 6p region, as the first chromatin remodeling factor with dual and selective roles in proliferation and apoptotic resistance in melanoma, acting independently of classic tumor suppressor pathways.

Materials and Methods

Cells and reagents. Human melanocytes and melanoma lines were isolated and cultured as described previously ( 6, 37). Doxorubicin hydrochloride, cycloheximide, and 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Sigma. Bortezomib was obtained from Millennium Pharmaceuticals. z-VAD-fmk was purchased from Calbiochem. TW-37 synthesis and validation has been described elsewhere ( 38).

Extract preparation and gene expression analyses. Total cell lysates were obtained by Laemmli extraction, separated by SDS-PAGE, and transferred to Immobilon-P membranes (Millipore) for immunoblotting (antibodies used in this study are listed in Supplementary Materials). Protein expression was quantified using Scion image densitometry software (Scion). Lentiviral-mediated modulation of DEK, MCL-1, BCL-2, and BCL-xL, tissue microarrays, spectral karyotyping, and comparative genomic hybridization procedures are described in Supplementary Information.

Cell proliferation and viability assays. Total cell numbers were estimated at the indicated times by manual counting, and cell death rates were determined by standard trypan blue exclusion assays ( 38). For flow cytometry assays, cells were fixed in 70% ethanol, digested with 0.5 mg/mL RNase A, and stained with 50 μg/mL propidium iodide before analysis in a FACSCalibur (Becton Dickinson Immunocytometry Systems). Data were fitted using ModFit software (Verity Software). Senescence-associated β-galactosidase staining was done as described previously ( 37). Nuclear condensation and fragmentation were assessed by fluorescence microscopy after DAPI staining using an Olympus BX-51 upright microscope.

Quantitative real-time reverse transcription-PCR. RNA purification and real-time reverse transcription-PCR using a Bio-Rad iCycler (Bio-Rad) were done essentially as described previously ( 37, 39) using the following primers: mcl1 forward ATGCTTCGGAAACTGGACAT and reverse TCCTGATGCCACCTTCTAGG and β-actin forward CGCCCAGCACGATGAAA and reverse CCGCCGATCCACACAGA. The assay included a no-template control and a standard curve of six dilutions for both mcl-1 and β-actin. Results are given as the average ± SE of three experiments, and each sample was amplified and analyzed in triplicate.

Luciferase reporter gene assays. The human MCL-1 promoter reporter plasmid −203/+10 was a generous gift of Douglas Cress ( 40). Cells were transfected using Lipofectamine 2000 (Invitrogen) and the lysates were assayed for firefly and Renilla luciferase activity using the Dual-Luciferase Reporter Assay System (Promega). Luminescence was detected with a Tecan GENios plate reader. Data are presented as the mean ± SE of three experiments.

Statistics. Statistical analyses of cell death, quantitative real-time reverse transcription-PCR, and luciferase reporter assays were done using a paired, two-tailed Student's t test.

Results

DEK protein overexpression in metastatic melanoma. We hypothesized that gain of 6p in melanoma may function to promote DEK overexpression. To this end, a panel of 16 metastatic melanoma lines was selected to reflect the most common melanoma-associated alterations in oncogenes or tumor suppressors (see Supplementary Table S1). In parallel, primary human foreskin-derived melanocytes were used as controls for normal cells. Immunoblotting revealed a generalized high expression of DEK, with an average 30-fold induction over the basal levels of normal melanocytes ( Fig. 1A and B ).

Figure 1.
Figure 1.

Overexpression of DEK in melanoma cell lines and tissue specimens. A, immunoblotting of a panel of metastatic melanoma lines and primary foreskin-derived melanocytes (FM293 and FM298), with monoclonal (mAb) and polyclonal (polyAb) anti-DEK antibodies. B, monoclonal anti-DEK band densities were quantified by densitometry, and relative DEK protein levels were normalized to FM293 melanocytes. Horizontal line, mean for each group. C, spectral karyotyping analysis showing increased copy number at chromosome 6 (yellow boxes) in SK-Mel-28 and SK-Mel-29, both expressing high levels of DEK ( Figs. 1A and 2C). D, representative examples of DEK staining in benign nevi (top) and melanoma lesions (bottom) included in tissue microarrays. Right, higher-magnification views of the indicated areas.

Interestingly, spectral karyotyping confirmed that melanoma cell lines with high DEK expression (SK-Mel-28 and -29) contained, among other genetic abnormalities, an increased copy number of chromosome 6 ( Fig. 1C). However, large-scale chromosome 6p amplifications are unlikely to be the sole mechanism underlying DEK up-regulation. Thus, lines such as SK-Mel-103 or SK-Mel-147, with moderate levels of DEK protein ( Fig. 1A), had a normal copy number of chromosome 6 defined by spectral karyotyping (Supplementary Fig. S1A). Comparative genomic hybridization showed focal gains at 6p22.3 but centromeric to DEK (Supplementary Fig. S1B).

To validate the increased DEK protein expression found in cell lines, an extensive immunohistochemical analysis was done on two tissue microarrays, containing surgically removed melanocytic lesions, each spotted in triplicate. Specifically, these tissue microarrays included 87 primary melanomas and 115 metastatic specimens. In addition, 3 dysplastic nevi, 7 benign nevi, and 5 biopsies of normal skin were included as reference. Notably, 90% of the tumors analyzed showed clear positive DEK expression (10% being very marked and 80% with moderate DEK levels; see representative examples in Fig. 1D, bottom). Normal skin melanocytes and benign nevi showed no detectable DEK expression ( Fig. 1D, top; results not shown). Therefore, these results support a broad-spectrum up-regulation of DEK during melanoma progression.

Long-term down-regulation of the endogenous DEK expression in melanoma induces premature senescence. Lentiviral vectors were used to transduce short hairpin RNAs (shRNA) and down-regulate DEK in a stable manner. To avoid complications of off-target effects, three independent shRNAs were tested (see Supplementary Fig. S2). In 6 of 12 lines analyzed (SK-Mel-28, SK-Mel-94, G-361, UACC-62, UACC-257, and MM576), DEK depletion led to a progressive cell cycle arrest, eventually resulting in a near complete inhibition of cell proliferation as determined by total cell counting ( Fig. 2A ) and flow cytometry ( Fig. 2B, top).

Figure 2.
Figure 2.

Induction of senescence-like cell cycle arrest by long-term down-regulation of DEK. A, total cell numbers estimated in SK-Mel-28 (left) and SK-Mel-94 (right) after infection with control shRNA or shDEK-2 (shDEK)–expressing lentiviruses. Bars, SE. B, melanoma cell lines were transduced as in A. Fifteen days after infection, cell cycle profile distribution was analyzed by flow cytometry (top), and senescence features were determined by senescence-associated β-galactosidase activity (bottom, blue signal). C, cells were treated as in A. Whole-cell lysates were processed for detection of the indicated proteins by immunoblotting. NT, no treatment.

The reduction of cells in S phase of the cell cycle ( Fig. 2B, top) was accompanied by cell enlargement, flattening, and vacuolization ( Fig. 2B, bottom) reminiscent of classic senescent phenotypes ( 41). In fact, staining for senescence-associated β-galactosidase activity was positive in DEK-depleted melanoma cells ( Fig. 2B, bottom). Interestingly, cell cycle arrest and the acquisition of senescent-like features occurred in cells expressing either wild-type p53 or mutated p53 (e.g., SK-Mel-94 or SK-Mel-28, respectively; Fig. 2C). Notably, DEK shRNAs did not affect the expression of p16INK4A ( Fig. 2C), a senescence-associated tumor suppressor. These results show that DEK is required for the continued proliferation of metastatic melanoma cells and point to novel p53 and p16 INK4A-independent senescence programs.

Short-term down-regulation of DEK expression bypasses melanoma chemoresistance. Recently, we have shown that DEK can enhance resistance to DNA-damaging agents by promoting efficient resolution of DNA damage foci ( 42). Therefore, we sought to determine whether short term down-regulation of DEK (therapeutically more feasible than sustained gene inactivation) could contribute to the ability of melanoma cells to withstand genotoxic agents. The anthracycline doxorubicin is a potent antineoplastic agent in other tumor types but has only limited efficacy against metastatic melanoma ( 8). This drug was selected to treat cells with the highest levels of DEK, as they may be more dependent on this protein for survival (see Fig. 3 for SK-Mel-19 and Fig. 4 for SK-Mel-29; results with other cell lines are discussed next). SK-Mel-19 was infected with a control shRNA lentiviral vector, a vector expressing a single DEK-specific shRNA (shDEK-1) or a vector expressing two distinct DEK-specific shRNAs (shDEK-2). Two days after lentiviral infection, cells were treated with increasing doses of doxorubicin and cell death was estimated within 24 to 36 h after treatment. Relatively high levels of doxorubicin were required to kill melanoma cells in the presence of DEK (>0.2 μg/mL; Fig. 3A). However, DEK depletion resulted in a marked increase in doxorubicin-driven cell death, particularly on dual-shRNA targeting ( Fig. 3A). Importantly, the increased response of DEK-depleted melanoma cells to doxorubicin was associated with classic apoptotic features such as chromatin condensation ( Fig. 3B) and caspase processing ( Fig. 3C). Therefore, although DEK may not be the sole mediator of melanoma chemoresistance, our results support a key role of this gene contributing to the malignant features of melanoma cells.

Figure 3.
Figure 3.

p53-independent induction of cell death by short-term inactivation of DEK. A, SK-Mel-19 cells transduced with the indicated shRNAs were treated with doxorubicin (Dox) for 30 h, and cell death was determined by 4′,6-diamidino-2-phenylindole staining and counting of condensed and fragmented nuclei. B, fluorescence microscopy illustrating cell nuclear morphology after treatment with 0.3 μg/mL doxorubicin for 30 h in cells transduced with a control vector (left) or shDEK-2 (right). C, SK-Mel-19 cells transduced as in A were treated with 0.2 μg/mL doxorubicin, collected at the indicated times, and analyzed by immunoblotting.

Figure 4.
Figure 4.

Caspase-independent down-regulation of MCL-1 by DEK shRNA. A, indicated melanoma cells were infected with either a control (Ctrl) lentivirus or one of three distinct vectors targeting DEK expression (shDEK-1, shDEK-2, and shDEK-3). Ninety-six hours after infection, cells were collected and their lysates were immunoblotted for DEK and MCL-1 expression. B, SK-Mel-19 cells transduced with the indicated shRNAs were treated with 0.2 μg/mL doxorubicin and collected for immunoblotting analyses. C, relative expression of MCL-1 (left), BCL-2 (middle), and BCL-xL (right) following doxorubicin treatment as estimated by band densitometry. Expression is shown relative to that of untreated control-transduced cells. D, protein immunoblots in control and shDEK-infected SK-Mel-19 cells treated with 0.2 μg/mL doxorubicin alone or with doxorubicin in the presence of 50 μmol/L of the pan-caspase inhibitor z-VAD-fmk.

DEK does not modulate p53 levels in melanoma cells. DEK has been shown previously to mediate antiapoptotic effects through destabilization of p53 protein and inhibition of p53 activity ( 23, 32). However, as with the case of long-term down-regulation of DEK shown above ( Fig. 2C), short-term expression of DEK shRNA was not found to significantly affect total p53 levels ( Fig. 3C). Furthermore, the activation of p53 (detected by phosphorylation of its serine 15) and the induction of p53 targets, such as p21CIP1, were similar in DEK-expressing and DEK-deficient melanoma cells ( Fig. 3C).

DEK-knockdown cells have diminished MCL-1 expression. Overexpression of antiapoptotic BCL-2 family members has been associated with melanoma chemoresistance ( 3). Therefore, we tested whether DEK could control BCL-2 family members. Interestingly, DEK down-regulation with two different shRNAs led to a drastic decrease in MCL-1 expression in SK-Mel-19, SK-Mel-29, SK-Mel-5, and MM576 ( Fig. 4A; data not shown). Notably, DEK shRNA did not alter BCL-2 or BCL-xL levels ( Fig. 4B), suggesting selective effects on the apoptotic machinery.

To further assess the requirement of DEK for high MCL-1 expression, DEK-depleted cells were treated with doxorubicin (which induces MCL-1 levels in melanoma cells; Fig. 4B and C). As shown in Fig. 4B, DEK shRNA cells ultimately expressed minimal levels of MCL-1 after doxorubicin treatment (see quantification in Fig. 4C). Again, the effects of DEK on the apoptotic machinery were selective, as no changes in BCL-2 or BCL-xL levels were observed before or after doxorubicin treatment ( Fig. 4B and C). Importantly, the reduction of MCL-1 expression in DEK-shRNA cells was not due to proteolytic cleavage by apoptotic caspases ( 43). Thus, the pan-caspase inhibitor z-VAD-fmk, although it eliminated caspase-3 cleavage, could not prevent MCL-1 down-regulation in shDEK-treated cells ( Fig. 4D).

Decreased MCL-1 in shDEK-treated cells is not due to increased MCL-1 degradation. To determine if the decrease in MCL-1 protein in DEK-knockdown cells was due to enhanced proteasomal turnover, cells were treated with the proteasome inhibitor bortezomib. As we reported previously ( 39, 44), bortezomib induced a massive MCL-1 stabilization in control melanoma cells within 2 to 3 h of treatment ( Fig. 5A ). A similar rate of MCL-1 accumulation was observed in shDEK-expressing melanoma cells ( Fig. 5B), although these cells have, as indicated above, intrinsically lower basal levels of MCL-1 (note that MCL-1 immunoblots in Fig. 5A were overexposed to favor the visualization of the endogenous low expression of MCL-1 in shRNA-DEK cells). Therefore, the diminished MCL-1 levels in DEK-knockdown cells were not due to increased proteosome-dependent MCL-1 degradation.

Figure 5.
Figure 5.

Transcriptional control of MCL-1 by DEK. A, comparative analysis of MCL-1 protein down-regulation in SK-Mel-19 cells transduced with control or DEK shRNAs and treated for the indicated times with 25 nmol/L bortezomib (Bort). BCL-2, and β-actin are included as examples of DEK-independent proteins and for loading control, respectively. B, quantification by densitometry of MCL-1 levels in A depicted with respect to nontreated control-infected cells. C and D, control or shRNA-transduced SK-Mel-19 cells were treated for 5 h with 25 nmol/L of the proteasome inhibitor bortezomib. Cells were then washed to remove bortezomib and incubated with cycloheximide (CHX) for the indicated times. Following treatment, cells were collected and processed for immunoblotting (C) and subsequent quantification by densitometry (D). E, MCL-1 mRNA expression of control and shDEK-treated SK-Mel-19 cells determined by quantitative real-time reverse transcription-PCR. F, control and shDEK-transduced SK-Mel-19 cells were transfected with either the promoterless pGL2 basic plasmid or a plasmid expressing the firefly luciferase gene under the control of the human MCL-1 promoter (−203/+10 bp). The average luciferase activity of three experiments is shown relative to that of control transduced cells transfected with the pGL2 basic plasmid. Bars, SE. *, P < 0.05.

To assess the stability of MCL-1 protein in DEK-knockdown cells, cycloheximide, a translational inhibitor, was used to inhibit de novo MCL-1 synthesis. To better visualize the decrease in MCL-1 expression, cells were first pretreated with bortezomib to augment MCL-1 levels. Bortezomib was subsequently removed, and cells were incubated in cycloheximide and then processed for immunoblotting analysis (see Materials and Methods for additional details). As shown in Fig. 5C and D, depletion of MCL-1 occurred with similar kinetics in both control and shDEK-treated cells.

Novel role of DEK in MCL-1 transcription. Because DEK was not affecting DEK protein stability, it could be controlling its transcription. To this end, mRNA levels were estimated by quantitative reverse transcription-PCR. As shown in Fig. 5E, DEK down-regulation leads to a 50% reduction in MCL-1 mRNA. To confirm the requirement of DEK for MCL-1 mRNA transcription, a reporter construct driving luciferase expression under the control of the proximal region of the human MCL-1 promoter was transduced into either control or shDEK-transduced SK-Mel-19 cells ( Fig. 5F). We found a 70% to 80% reduction in promoter activity in MCL-1 in the absence of DEK ( Fig. 5F). These results show a new role for DEK in the transcriptional control of MCL-1.

Functional implication of DEK-dependent MCL-1 expression. To define the specific interplay between DEK and MCL-1 in the enhancement of melanoma chemoresistance, we generated isogenic series of melanoma cell lines with differing levels of both proteins and compared their response to doxorubicin. Additional isogenic cell lines were generated to alter (positively and negatively) the endogenous expression of BCL-xL. BCL-xL was chosen as an example of an antiapoptotic BCL-2 factor that is not under the control of DEK ( Fig. 4B) but is also overexpressed in melanoma cells ( 3).

To alter MCL-1 expression, we used two approaches. First, we generated lentiviral vectors to drive MCl-1 from an ectopic promoter that is not sensitive to DEK. Viral titers were selected to restore MCL-1 levels to the basal conditions ( Fig. 6A, left, inset ). In a second set of experiments, the opposite strategy was tested ( Fig. 6B), that is, to abrogate MCL-1 mRNA and protein expression with shRNAs that we have previously validated ( 6, 39, 44). Interestingly, exogenous MCL-1 significantly protected DEK-depleted melanoma cells, reducing the killing activity of doxorubicin in half ( Fig. 6A; P = 0.014 for concentrations of doxorubicin of 0.25 μg/mL). On the other hand, when MCL-1 was depleted, DEK shRNA was not able to further enhance drug response ( Fig. 6B, left; P = 0.441).

Figure 6.
Figure 6.

Functional effect of the selective regulation of BCL-2 family members by DEK. A, restoration of MCL-1 levels using lentiviral vectors coding for MCL-1 under an exogenous promoter that is not sensitive to DEK shRNA down-regulation (left). Control infections were also done with a vector expressing BCL-xL (right). Seventy-two hours after infection with MCL-1 or BCL-xL lentiviruses, cells were treated with the indicated concentration of doxorubicin for 30 h. Following treatment, cells were stained with 4′,6-diamidino-2-phenylindole and apoptotic cells were counted by fluorescence microscopy. B, inactivation of MCL-1 (left) or BCL-xL (right) with the corresponding shRNAs. Insets, shRNA-mediated knockdown of DEK and either knockdown or overexpression of BCL-xL and MCL-1 as determined by immunoblotting. C, SK-Mel-19 cells were transduced as in Fig. 3 and treated with the BH3 mimetic TW-37 for 48 h and then stained with trypan blue. Bars, SE. D, fluorescence microscopy by 4′,6-diamidino-2-phenylindole staining illustrating cell nuclear morphology after treatment with 15 μmol/L TW-37 in cells infected with a control shRNA expressing vector (left) or shDEK-2 (right). P values for the indicated pair-wise comparisons were determined by two-tailed Student's t test.

In contrast to MCL-1, altering BCL-xL levels by either overexpression ( Fig. 6A, right) or shRNAs ( Fig. 6B, right) had no significant effect on melanoma cell viability whether in the presence or absence of DEK (see the corresponding P values in Fig. 6A and B).

From the data presented above, it follows that DEK shRNA may augment the therapeutic efficacy of small molecules, the antitumor activity of which depends on MCL-1 inactivation. To examine this possibility, we tested the small-molecule BH3 mimetic TW-37, which we have shown previously to act primarily by inhibiting MCL-1 ( 6, 38). Interestingly, DEK-depleted SK-Mel-19 cells were significantly more sensitive to TW-37 treatment than their control counterparts, showing an accelerated killing ( Fig. 6C; P = 0.011-0.002) and induction of apoptotic-like features (see chromatin condensation in Fig. 6D). These results provide proof-of-principle for the development of rational therapies targeting DEK/MCL-1–dependent mechanisms of cell survival in melanoma.

Discussion

In this study, we identified new roles for DEK in metastatic melanoma cells. Despite the high degree of heterogeneity characteristic of this tumor type, we found that the DEK protein was invariably overexpressed in all 16 metastatic cell lines investigated independent of their genetic background. Moreover, high-throughput immunohistochemical analyses revealed a marked expression of DEK in 90% of primary and metastatic melanomas studied. In the same conditions, DEK expression in normal skin and benign nevi was nearly undetectable. Importantly, this high expression of DEK was not a simple byproduct of melanoma development. Abrogation of endogenous DEK expression resulted in the disruption of characteristic malignant features of melanomas. Thus, long-term expression of DEK shRNAs led to cell cycle arrest with features of senescence. In turn, acute down-regulation of DEK was sufficient to enhance the response to DNA damaging agents and BH3 mimetics (doxorubicin and TW-37, respectively). Our results additionally unveiled a previously undescribed role for DEK in the regulation of a critical mediator of chemoresistance, MCL-1.

The contribution of DEK to melanoma proliferation and drug resistance provides mechanistic support for the long-suspected presence of oncogene(s) mapping to the short arm of chromosome 6 ( 9, 13, 14). A complication in the identification of the putative onocogene(s) in this region has been its high gene density (624 genes in the 6p21-23 interval). Moreover, this area is heavily enriched in CpG islands; therefore, it can be subject to complex epigenetic regulation. Interestingly, a common area of copy gain at 6p22-p23 in ocular melanomas ( 11) had also been reported in bladder cancer and retinoblastoma ( 16– 18). Of the 7 genes in this region, only DEK was found to have a correlation between copy gain and gene expression ( 18). Our spectral karyotyping and comparative genomic hybridization analyses revealed duplications of chromosome 6 in melanoma cell lines expressing high levels of DEK. However, other mechanisms of regulation may favor DEK up-regulation, as its expression can also be induced in cells with a normal copy number at the DEK locus.

Considering the high number of mutations and alterations in proliferative pathways that melanoma cells express, it is interesting that down-regulation of a single gene (DEK) can ultimately drive them into senescence-like cell cycle arrest. In fact, melanomas are considered to be particularly efficient at overcoming premature senescence, as this is one of the early barriers to melanocyte transformation ( 45). Mechanisms that circumvent premature senescence in melanomas are not well characterized. In fact, a puzzling feature of this disease is that a large fraction of acquired melanomas neither mutate nor delete p53 or p16INK4a, two tumor suppressors with key roles in the induction of senescence programs ( 2). Therefore, DEK overexpression may represent an alternative mechanism to counteract the tumor suppressor activity of p53 and p16 in melanoma cells. The precise effects of DEK in promoting sustained proliferation will require further studies, but it is tempting to speculate that it may cooperate with other oncogenes mapping at chromosome 6p, such as c-MYC, which we have recently shown to control melanoma senescence ( 46).

Although the requirement of sustained expression of DEK to avoid premature senescence has implications for basic aspects of melanoma, the new roles of DEK shown here in the control of the apoptotic machinery are perhaps more significant from a therapeutic perspective. Although DEK has been shown previously to participate in the regulation of gene transcription ( 23– 25, 30, 47), promotion of MCL-1 expression is the first demonstration of a transcriptional target of DEK directly contributing to an oncogenic phenotype. The requirement of DEK for MCL-1 expression is relevant as this protein has an intrinsically short half life ( 5); yet, it is consistently accumulated in melanoma cells ( 3). In addition, MCL-1 transcription can be inhibited by p53 ( 48), E2F1, or E2F2 ( 40), all of which are expressed in melanoma cells. Therefore, the effects of DEK on MCL-1 transcription may serve to compensate for negative regulatory signals intrinsically expressed in melanoma cells. Furthermore, we and others have shown that MCL-1 is essential for long-term survival of melanoma cells, and it is a main determinant of their resistance to dacarbazine, BH3 mimetics, and inhibitors of MEK or the proteasome ( 6, 44, 49). MCL-1 can also control melanoma cell survival under conditions of anoikis ( 50). Thus, enhanced MCL-1 expression caused by DEK overexpression may represent a major mechanism of melanoma cell survival in response to chemotherapeutic agents and other stress-inducing factors.

In conclusion, our results implicate overexpression of the 6p pro-oncogene DEK as an extremely frequent event in metastatic melanomas. Remarkably, this single event can have a dual effect on melanoma cell proliferation and chemoresistance (preventing senescence and inducing the antiapoptotic MCL-1, respectively). Therefore, our results suggest that DEK may be a relevant target for therapeutic intervention in this aggressive disease.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Acknowledgments

Grant support: NIH training grant T32-GM07863 and National Science Foundation and Rackham Predoctoral Fellowships (M.S. Khodadoust); William D. Robinson Fellowship from the Arthritis Foundation/Michigan Chapter and Arthritis Foundation Postdoctoral Fellowship (F. Kappes); NIH Prostate Specialized Program of Research Excellence grant P50CA69568 and NIH grant U01 CA111275 (D.S.L. Kim and A.M. Chinnaiyan); NIH grant R01-AI062248 and Burroughs Wellcome Fund Clinical Scientist Award in Translational Research (D.M. Markovitz); NIH grant R01-CA107237, grant SAF2008-1950 from the Spanish Ministry of Innovation and Science, and Spanish Association Against Cancer development grant (M.S. Soengas); and Fundación Obra Social Caja Navarra post-residency award (E. Riveiro-Falkenbach).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank MaryBeth Riblett (University of Michigan) and Tonantzin G. Calvo (Centro Nacional de Investigaciones Oncológicas) for technical assistance, M. Carmen Martin and Bibiana I. Ferreira (Centro Nacional de Investigaciones Oncológicas) for help with spectral karyotyping and comparative genomic hybridization analyses, and Karolyn Oetjen and Colin Duckett for the BCL-xL expression vector.

Footnotes

  • Received March 24, 2009.
  • Revision received June 11, 2009.
  • Accepted June 18, 2009.

References

  1. ↵
    Gray-Schopfer V, Wellbrock C, Marais R. Melanoma biology and new targeted therapy. Nature 2007; 445: 851–7.
    OpenUrlCrossRefPubMed
  2. ↵
    Chin L, Garraway LA, Fisher DE. Malignant melanoma: genetics and therapeutics in the genomic era. Genes Dev 2006; 20: 2149–82.
    OpenUrlAbstract/FREE Full Text
  3. ↵
    Soengas MS, Lowe SW. Apoptosis and melanoma chemoresistance. Oncogene 2003; 22: 3138–51.
    OpenUrlCrossRefPubMed
  4. ↵
    McGill GG, Horstmann M, Widlund HR, et al. Bcl2 regulation by the melanocyte master regulator mitf modulates lineage survival and melanoma cell viability. Cell 2002; 109: 707–18.
    OpenUrlCrossRefPubMed
  5. ↵
    Kutuk O, Letai A. Regulation of Bcl-2 family proteins by posttranslational modifications. Curr Mol Med 2008; 8: 102–18.
    OpenUrlCrossRefPubMed
  6. ↵
    Verhaegen M, Bauer JA, Martin de la Vega C, et al. A novel BH3 mimetic reveals a mitogen-activated protein kinase-dependent mechanism of melanoma cell death controlled by p53 and reactive oxygen species. Cancer Res 2006; 66: 11348–59.
    OpenUrlAbstract/FREE Full Text
  7. Ueda Y, Richmond A. NF-κB activation in melanoma. Pigment Cell Res 2006; 19: 112–24.
    OpenUrlCrossRefPubMed
  8. ↵
    Gogas HJ, Kirkwood JM, Sondak VK. Chemotherapy for metastatic melanoma: time for a change? Cancer 2007; 109: 455–64.
    OpenUrlCrossRefPubMed
  9. ↵
    Santos GC, Zielenska M, Prasad M, Squire JA. Chromosome 6p amplification and cancer progression. J Clin Pathol 2007; 60: 1–7.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    Curtin JA, Fridlyand J, Kageshita T, et al. Distinct sets of genetic alterations in melanoma. N Engl J Med 2005; 353: 2135–47.
    OpenUrlCrossRefPubMed
  11. ↵
    Namiki T, Yanagawa S, Izumo T, et al. Genomic alterations in primary cutaneous melanomas detected by metaphase comparative genomic hybridization with laser capture or manual microdissection: 6p gains may predict poor outcome. Cancer Genet Cytogenet 2005; 157: 1–11.
    OpenUrlCrossRefPubMed
  12. ↵
    Hoglund M, Gisselsson D, Soller M, Hansen GB, Elfving P, Mitelman F. Dissecting karyotypic patterns in malignant melanomas: temporal clustering of losses and gains in melanoma karyotypic evolution. Int J Cancer 2004; 108: 57–65.
    OpenUrlCrossRefPubMed
  13. ↵
    Balaban G, Herlyn M, Guerry D, 4th, et al. Cytogenetics of human malignant melanoma and premalignant lesions. Cancer Genet Cytogenet 1984; 11: 429–39.
    OpenUrlCrossRefPubMed
  14. ↵
    Pathak S, Drwinga HL, Hsu TC. Involvement of chromosome 6 in rearrangements in human malignant melanoma cell lines. Cytogenet Cell Genet 1983; 36: 573–9.
    OpenUrlCrossRefPubMed
  15. ↵
    Chen D, Gallie BL, Squire JA. Minimal regions of chromosomal imbalance in retinoblastoma detected by comparative genomic hybridization. Cancer Genet Cytogenet 2001; 129: 57–63.
    OpenUrlCrossRefPubMed
  16. ↵
    Evans AJ, Gallie BL, Jewett MA, et al. Defining a 0.5-mb region of genomic gain on chromosome 6p22 in bladder cancer by quantitative-multiplex polymerase chain reaction. Am J Pathol 2004; 164: 285–93.
    OpenUrlPubMed
  17. Grasemann C, Gratias S, Stephan H, et al. Gains and overexpression identify DEK and E2F3 as targets of chromosome 6p gains in retinoblastoma. Oncogene 2005; 24: 6441–9.
    OpenUrlPubMed
  18. ↵
    Orlic M, Spencer CE, Wang L, Gallie BL. Expression analysis of 6p22 genomic gain in retinoblastoma. Genes Chromosomes Cancer 2006; 45: 72–82.
    OpenUrlCrossRefPubMed
  19. ↵
    von Lindern M, van Baal S, Wiegant J, Raap A, Hagemeijer A, Grosveld G. Can, a putative oncogene associated with myeloid leukemogenesis, may be activated by fusion of its 3′ half to different genes: characterization of the set gene. Mol Cell Biol 1992; 12: 3346–55.
    OpenUrlAbstract/FREE Full Text
  20. ↵
    Waldmann T, Scholten I, Kappes F, Hu HG, Knippers R. The DEK protein-an abundant and ubiquitous constituent of mammalian chromatin. Gene 2004; 343: 1–9.
    OpenUrlCrossRefPubMed
  21. ↵
    Carro MS, Spiga FM, Quarto M, et al. DEK Expression is controlled by E2F and deregulated in diverse tumor types. Cell Cycle 2006; 5: 1202–7.
    OpenUrlPubMed
  22. ↵
    Wu Q, Li Z, Lin H, Han L, Liu S, Lin Z. DEK overexpression in uterine cervical cancers. Pathol Int 2008; 58: 378–82.
    OpenUrlCrossRefPubMed
  23. ↵
    Sammons M, Wan SS, Vogel NL, Mientjes EJ, Grosveld G, Ashburner BP. Negative regulation of the RelA/p65 transactivation function by the product of the DEK proto-oncogene. J Biol Chem 2006; 281: 26802–12.
    OpenUrlAbstract/FREE Full Text
  24. Faulkner NE, Hilfinger JM, Markovitz DM. Protein phosphatase 2A activates the HIV-2 promoter through enhancer elements that include the pets site. J Biol Chem 2001; 276: 25804–12.
    OpenUrlAbstract/FREE Full Text
  25. ↵
    Gamble MJ, Fisher RP. SET and PARP1 remove DEK from chromatin to permit access by the transcription machinery. Nat Struct Mol Biol 2007; 14: 548–55.
    OpenUrlCrossRefPubMed
  26. Kappes F, Scholten I, Richter N, Gruss C, Waldmann T. Functional domains of the ubiquitous chromatin protein DEK. Mol Cell Biol 2004; 24: 6000–10.
    OpenUrlAbstract/FREE Full Text
  27. Mor-Vaknin N, Punturieri A, Sitwala K, et al. The DEK nuclear autoantigen is a secreted chemotactic factor. Mol Cell Biol 2006; 26: 9484–96.
    OpenUrlAbstract/FREE Full Text
  28. ↵
    Wise-Draper TM, Morreale RJ, Morris TA, et al. DEK proto-oncogene expression interferes with the normal epithelial differentiation program. Am J Pathol 2009; 174: 71–81.
    OpenUrlCrossRefPubMed
  29. ↵
    Wise-Draper TM, Allen HV, Thobe MN, et al. The human DEK proto-oncogene is a senescence inhibitor and an upregulated target of high-risk human papillomavirus E7. J Virol 2005; 79: 14309–17.
    OpenUrlAbstract/FREE Full Text
  30. ↵
    Kim DW, Chae JI, Kim JY, et al. Proteomic analysis of apoptosis related proteins regulated by proto-oncogene protein DEK. J Cell Biochem 2009.
  31. ↵
    Sitwala KV, Mor-Vaknin N, Markovitz DM. Minireview: DEK and gene regulation, oncogenesis and AIDS. Anticancer Res 2003; 23: 2155–8.
    OpenUrlPubMed
  32. ↵
    Wise-Draper TM, Allen HV, Jones EE, Habash KB, Matsuo H, Wells SI. Apoptosis inhibition by the human DEK oncoprotein involves interference with p53 functions. Mol Cell Biol 2006; 26: 7506–19.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    Devany M, Kappes F, Chen KM, Markovitz DM, Matsuo H. Solution NMR structure of the N-terminal domain of the human DEK protein. Protein Sci 2008; 17: 205–15.
    OpenUrlCrossRefPubMed
  34. ↵
    Cleary J, Sitwala KV, Khodadoust MS, et al. p300/CBP-associated factor drives DEK into interchromatin granule clusters. J Biol Chem 2005; 280: 31760–7.
    OpenUrlAbstract/FREE Full Text
  35. ↵
    Grottke C, Mantwill K, Dietel M, Schadendorf D, Lage H. Identification of differentially expressed genes in human melanoma cells with acquired resistance to various antineoplastic drugs. Int J Cancer 2000; 88: 535–46.
    OpenUrlCrossRefPubMed
  36. ↵
    Kappes F, Damoc C, Knippers R, Przybylski M, Pinna LA, Gruss C. Phosphorylation by protein kinase CK2 changes the DNA binding properties of the human chromatin protein DEK. Mol Cell Biol 2004; 24: 6011–20.
    OpenUrlAbstract/FREE Full Text
  37. ↵
    Denoyelle C, Abou-Rjaily G, Bezrookove V, et al. Anti-oncogenic role of the endoplasmic reticulum differentially activated by mutations in the MAPK pathway. Nat Cell Biol 2006; 8: 1053–63.
    OpenUrlCrossRefPubMed
  38. ↵
    Wang G, Nikolovska-Coleska Z, Yang CY, et al. Structure-based design of potent small-molecule inhibitors of anti-apoptotic Bcl-2 proteins. J Med Chem 2006; 49: 6139–42.
    OpenUrlCrossRefPubMed
  39. ↵
    Fernandez Y, Verhaegen M, Miller TP, et al. Differential regulation of noxa in normal melanocytes and melanoma cells by proteasome inhibition: therapeutic implications. Cancer Res 2005; 65: 6294–304.
    OpenUrlAbstract/FREE Full Text
  40. ↵
    Croxton R, Ma Y, Song L, Haura EB, Cress WD. Direct repression of the Mcl-1 promoter by E2F1. Oncogene 2002; 21: 1359–69.
    OpenUrlCrossRefPubMed
  41. ↵
    Collado M, Blasco MA, Serrano M. Cellular senescence in cancer and aging. Cell 2007; 130: 223–33.
    OpenUrlCrossRefPubMed
  42. ↵
    Kappes F, Fahrer J, Khodadoust MS, et al. DEK is a poly(ADP-ribose)-acceptor in apoptosis and mediates resistance to genotoxic stress. Mol Cell Biol 2008; 28: 3245–57.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    Han J, Goldstein LA, Gastman BR, Froelich CJ, Yin XM, Rabinowich H. Degradation of Mcl-1 by granzyme B: implications for Bim-mediated mitochondrial apoptotic events. J Biol Chem 2004; 279: 22020–9.
    OpenUrlAbstract/FREE Full Text
  44. ↵
    Wolter KG, Verhaegen M, Fernandez Y, et al. Therapeutic window for melanoma treatment provided by selective effects of the proteasome on Bcl-2 proteins. Cell Death Differ 2007; 14: 1605–16.
    OpenUrlCrossRefPubMed
  45. ↵
    Prieur A, Peeper DS. Cellular senescence in vivo: a barrier to tumorigenesis. Curr Opin Cell Biol 2008; 20: 150–5.
    OpenUrlCrossRefPubMed
  46. ↵
    Zhuang D, Mannava S, Grachtchouk V, et al. C-MYC overexpression is required for continuous suppression of oncogene-induced senescence in melanoma cells. Oncogene 2008; 27: 6623–34.
    OpenUrlCrossRefPubMed
  47. ↵
    Ko SI, Lee IS, Kim JY, et al. Regulation of histone acetyltransferase activity of p300 and PCAF by proto-oncogene protein DEK. FEBS Lett 2006; 580: 3217–22.
    OpenUrlCrossRefPubMed
  48. ↵
    Pietrzak M, Puzianowska-Kuznicka M. p53-dependent repression of the human MCL-1 gene encoding an anti-apoptotic member of the BCL-2 family: the role of Sp1 and of basic transcription factor binding sites in the MCL-1 promoter. Biol Chem 2008; 389: 383–93.
    OpenUrlCrossRefPubMed
  49. ↵
    Qin JZ, Xin H, Sitailo LA, Denning MF, Nickoloff BJ. Enhanced killing of melanoma cells by simultaneously targeting Mcl-1 and NOXA. Cancer Res 2006; 66: 9636–45.
    OpenUrlAbstract/FREE Full Text
  50. ↵
    Boisvert-Adamo K, Longmate W, Abel EV, Aplin AE. Mcl-1 is required for melanoma cell resistance to anoikis. Mol Cancer Res 2009; 7: 549–56.
    OpenUrlAbstract/FREE Full Text
View Abstract
Back to top
Cancer Research: 69 (16)
August 2009
Volume 69, Issue 16

Sign up for alerts

View this article with LENS

Open full page PDF
Article Alerts
Email Article
Citation Tools
Melanoma Proliferation and Chemoresistance Controlled by the DEK Oncogene
Michael S. Khodadoust, Monique Verhaegen, Ferdinand Kappes, Erica Riveiro-Falkenbach, Juan C. Cigudosa, David S.L. Kim, Arul M. Chinnaiyan, David M. Markovitz and María S. Soengas
Cancer Res August 15 2009 (69) (16) 6405-6413; DOI: 10.1158/0008-5472.CAN-09-1063
del.icio.us logo Digg logo Reddit logo Twitter logo CiteULike logo Facebook logo Google logo Mendeley logo

Jump to section

Advertisement