Introduction

Bcl-2 functions as an anti-apoptotic protein [1]. Bcl-2, in general, regulates mitochondrial outer membrane permeabilization and thereby determines the cellular commitment to apoptosis [1,2]. The anti-apoptotic function of the Bcl-2 protein depends, at least in part, on its ability to dimerize with another member of the Bcl-2 family, Bax [3]. Overexpression of Bcl-2 has been reported in a wide variety of cancers. 

In some preclinical systems, Bcl-2 overexpression has been shown to attenuate apoptosis, or to restore the clonogenic potential of malignant progenitor cells [4,5]. On the other hand, anti-apoptotic factors, including Bcl-2, impair the ability to achieve remission and cure with chemotherapy, protecting the tumor cells from the apoptotic effects of various antineoplastic agents [6-8]. Therefore, overcoming the resistance conferred by anti-apoptotic factors such as Bcl-2 represents an attractive therapeutic strategy against leukemia cells [3]. Recent study demonstrated that overexpression of Bcl-2 attenuates resveratrol-induced apoptosis in U937 cells by inhibition of caspase-3 activity and sustained expression of the IAP caspase inhibitors [5]. 

Cisplatin is one of the most potent anticancer agents showing significant clinical activity against a variety of solid tumors belonging to a class of platinum containing anti-cancer compounds [9]. It is a representative of anticancer drug used to treat certain types of head and neck cancer, cervical carcinoma, lung cancer, neurologic cancers, and a wide variety of other cancers. However, its resistance remains a significant barrier to the survival of cancer patient. 

Bile acids are polar derivatives of cholesterol essential for the absorption of dietary lipids and regulate the transcription of genes that control cholesterol homeostasis. Different bile acids exhibit distinct biological effects. Importantly, natural bile salts were reported to inhibit cell proliferation and induce apoptosis in various cells [10,11]. Im et al. [12,13] developed several ursodeoxycholic acid (UDCA) and chenodeoxycholic acid (CDCA) derivatives, and it have been reported that they had apoptosis-inducing effect in various cancer cells [14-19]. 

The promyelocytic leukemia protein (PML) is a tumor suppressor that distinctly localizes to a nuclear substructure called the PML nuclear body (PML-NB). Many proteins reside in PML-NBs. Consequently, PML-NBs are implicated in the regulation of various cellular functions, such as the induction of apoptosis and cellular senescence [20]. 

This study was undertaken to determine whether the synthetic chenodeoxycholic acid (CDCA) derivative, HS-1200, treatment can bypass the anti-apoptotic effects of Bcl-2 in human leukemia cells with a specific focus on the involvement of PML-NBs. For this study, we employed Bcl-2overexpressing human leukemic cells (U937/Bcl-2 cells). As will be shown, the synthetic chenodeoxycholic acid (CDCA) derivative, HS-1200, treatment overcomes the resistance conferred by Bcl-2 and is associated with the formation of mature PML nuclear bodies in U937/Bcl-2 cells. 

Materials and Methods

Reagents

The synthetic bile acid derivative, HS-1200 was kindly provided by Professor Young-Hyun Yoo (Department of Anatomy, College of Medicine, Dong-A University, Busan, Korea). 5,5',6,6'-tetrachloro-1,1',3,3'-tetraethylbenzimidazol carbocyanine iodide (JC-1) was from Molecular Probes (Eugene, OR, USA). ApoScreenTM Annexin V-FITC Apoptosis Kit was from Beckman coulter (Fullerton, CA, USA). Dulbecco's modified Eagle's medium (DMEM) and FBS were from Gibco (Gaithersburg, MD, USA). Dimethyl sulfoxide (DMSO), cisplatin, Hoechst 33342, RNase A, proteinase K, aprotinin, leupeptin, PMSF, thiazolyl blue tetrazolium bromide and propidium iodide (PI) were from Sigma (St. Louis, MO, USA); SuperSignal West Femto enhanced chemiluminescence Western blotting detection reagent was from Pierce (Rockford, IL, USA). 

Antibodies

Mouse monoclonal anti-human caspase-3, caspase-7, cas pase-9, poly(ADP-ribose) polymerase (PARP), β-actin antibodies, and FITC-conjugated goat anti-mouse and antirabbit IgGs, and rabbit polyclonal anti-human Bcl-2, p53 (full length), Sp100, PML, CBP, SUMO-1, Daxx and mouse monoclonal anti-human PML and CBP antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Fluorescein isothiocyanate (FITC)-conjugated goat antirabbit and horse anti-mouse IgG antibodies were obtained from Vector (Burlingame, CA, USA). Texas Red-conjugated horse anti-mouse and goat anti-rabbit IgG antibodies were obtained from Molecular Probes (Eugene, OR, USA). HRPconjugated sheep anti-mouse and anti-rabbit IgGs were from Amersham GE Healthcare (Little Chalfront, UK). 

Cell culture and establishment of Bcl-2-overexpressing

Human leukemia U937 cells were obtained from the American Type Culture Collection (ATCC; Rockville, MD). The culture medium used throughout these experiments was RPMI-1640 medium, containing 10% fetal bovine serum (FBS; Gibco), 20 mM HEPES buffer and 100 μg/ml gentamicin. Bcl-2-overexpressing U937 cells were generated using a pMAX vector containing the human Bcl-2 gene (provided by Dr. Rakesh Srivastava, NIH/NIA). U937 cells (400 μl) in RPMI 1640 (2×10⁶ cells/ml) were transfected by pre-incubating with 15 μg Bcl-2 plasmid for 10 min at room temperature and then electroporating at 500 V, 700 μF. The sample was immediately placed on ice for 10 min, then 10 ml of complete medium was added, and the cells were incubated at 37℃ for 24 h. The cells were selected in a medium containing 0.7 μg/ml geneticin (G418; Calbiochem) for 4 weeks. Single cell clones were obtained by limiting dilution and subsequently analyzed for an increase in Bcl-2 protein expression relative to identically cloned empty vector controls. 

HS-1200 and cisplatin treatment and assessment of cell

U937/vector and U937/bcl-2 cells were placed in a 96well plate and were incubated for 24 h. Then cells were treated with 50 μM HS-1200, and 5 μg/ml cisplatin for 24 h. After cells were treated with 500 μg/ml of thiazolyl blue tetrazolium bromide (MTT solution), they were incubated at 37℃ with 5% CO₂  for 4 h. And then the medium was aspirated and formed formazan crystals were dissolved in DMSO. Cell viability was measured by an ELISA reader (Tecan, Männedorf, Switzerland) at 570 nm excitatory emission wavelength. 

Hoechst staining

Cells were harvested and cell suspension was centrifuged onto a clean, fat-free glass slide with a cytocentrifuge. Cells were stained in 4 μg/ml Hoechst 33342 for 10 min at 37℃ in the dark and fixed for 10 min 4% paraformaldehyde. 

DNA electrophoresis

Cells (2×10⁶) were resuspended in 0.5 ml of lysis buffer [10 mM Tris (pH 7.5), 10 mM EDTA (pH 8.0), 10 mM NaCl and 0.5% SDS] into which proteinase K (200 μg/ml) was added. After samples were incubated overnight at 55℃, 200 μl of ice cold 5 M NaCl was added and the supernatant containing fragmented DNA was collected after centrifugation. The DNA was then precipitated overnight at -20℃ in 50% isopropanol and RNase A-treated for 1 h at 37℃. The DNA from 10⁶ cells (15 μl) was equally loaded on each lane of 2% agarose gels in Tris-acetic acid/EDTA buffer containing 0.5 g/ml ethidium bromide at 100 mA for 0.5 h. 

Quantification of DNA hypoploidy by flow cytometry

After treatment for 24 h, cells were harvested by trypsinization and ice cold 95% ethanol with 0.5% Tween 20 was added to the cell suspensions to a final concentration of 70% ethanol. Fixed cells were pelleted, and washed in 1% BSA-PBS solution. Cells were resuspended in 1 ml PBS containing 20 μg/ml RNase A, incubated at 4℃ for 30 min, washed once with BSA-PBS, and resuspended in PI solution (10 μg/ml). After cells were incubated at 4℃for 5 min in the dark, DNA content were measured on a CYTOMICS FC500 flow cytometry system (Beckman Coulter, FL, CA, USA) and data was analyzed using the Multicycle software which allowed a simultaneous estimation of cell-cycle parameters and apoptosis. 

Western blot analysis

Cells (2×10⁶) were washed twice with ice-cold PBS, resuspended in 100 μl RIPA buffer [300 mM NaCl, 50 mM Tris-HCl (pH 7.6), 0.5% TritonX-100, 2 mM PMSF, 2 μl/ ml aprotinin and 2 μl/ml leupeptin] and incubated at 4℃ for 1 h. The lysates were centrifuged at 14,000 revolutions per min for 15 min at 4℃. Protein concentrations of cell lysates were determined with Bradford protein assay (BioRad, Richmond, CA, USA) and 25 μg proteins were loaded onto 7.5-15% SDS/PAGE. The gels were transferred to PVDF membrane (Millipore Corporation, Bedford, USA) and reacted with each antibody. Immunostaining with antibodies was performed using SuperSignal West Femto maximum sensitivity substrate and detected with Alpha Imager HP (Alpha Innotech, Santa Clara, CA, USA). 

Measurement of mitochondrial membrane potential(MMP)

Cells were plated in a standard 6-well plate at a density of 5×10⁵cells/ml. CGM treated cells were incubated for various time points. The cells were harvested and then JC-1 was added directly (1 μM final concentration) and incubated for 15 min at 37℃. Flow cytometry to measure MMP (ΔΨm) was performed on a CYTOMICS FC500 flow cytometry system (Beckman Coulter, Brea, CA USA). The fluorescence emission were collected through a 530/30 band pass filter (FL-1) and through a 585/42 band pass filter (FL-2), both on a log scale. Data were acquired and analyzed using CXP software version 2.2. The analyzer threshold was adjusted on the FSC channel to exclude noise and most of the subcellular debris. 

Immunofluorescent staining and confocal microscopy

A cell suspension was cytospun onto a clean fat-free glass slide. Cells were incubated with each primary antibody for 2 h at 37℃, washed 3 times for 5 min each and then incubated with FITC-conjugated or Texas Red-conjugated secondary antibodies for 1 h at room temperature. Fluorescent images were visualized and analyzed with a Zeiss LSM 700 laser-scanning confocal microscope (Zeiss, Göettingen, Germany). Cells were co-stained with DAPI for 10 min at 37℃ to observe nuclear morphology. 

Co-immunoprecipitation assay

Cells were collected and lysed in 1 ml of immunoprecipitation lysis buffer (300 mM NaCl, 50 mM Tris Cl [pH 7.6], 0.5% Triton X-100, protease inhibitors, 10 mM Na₄P₂O₇, 1 mM Na₃VO₄, 25 mM NaF and 1mM ß-glycerophosphate). Protein concentrations of cell lysates were determined using the Bradford method and 500 μg of protein was precleared and then incubated with PML or p53 antibody in extraction buffer at 4℃ overnight. The immune complexes were precipitated with protein A/G-agarose beads (Sigma) for 2 h and washed 5 times with extraction buffer prior to boiling in SDS sample buffer. Immunoprecipitated proteins were separated by SDS-polyacrylamide gel electrophoresis and Western blot analysis was performed as described above. 

Statistical analysis

Three independent experiments were performed for each experimental group and each experiment was performed in triplicate. The results of the experimental and control groups were cpmpared for statistical significance (p<0.01 and 0.05) using paired T-test statistical method by SPSS for Win 12.0 for summary data. 

Results

HS-1200 reduces cell viability to a similar extent in both U937/vector and U937/Bcl-2cells.

Treatment with 1~10 μg/ml cisplatin efficiently reduced the viability of U937/vector cells in a dose-dependent manner, while overexpression of Bcl-2 significantly attenuated cisplatininduced cell death. In contrast, treatment with 10~ 100 μM HS-1200 reduced cell viability to a similar extent in both U937/vector and U937/Bcl-2 cells (Fig. 1A). These data suggest that HS-1200 overcomes the resistance conferred by Bcl-2 in human leukemic U937 cells. 

HS-1200 overcomes the resistance conferred by Bcl-2 in U937 cells by inducing apoptosis.

Since the dose required for half-maximal inhibition of viability in U937/vector cells was 50 μM for HS-1200 and 5 μg/ml for cisplatin, these concentrations were utilized for further study. To prove that HS-1200 reduced the viability of U937/Bcl-2 cells by inducing apoptosis, cells were examined for the presence of abnormalities in cell cycle status, and atypical nuclear morphology, all of which can indicate that the cells are apoptotic. DNA electrophoresis indicated DNA fragmentation in both HS1200-treated U937/vector and U937/Bcl-2 cells (Fig. 1B). Hoechst staining showed that both U937/vector and U937/Bcl-2 cells treated with HS-1200 had fragmented atypical nuclei (Fig. 1C). A simultaneous estimation of cell cycle parameters and apoptosis by flow cytometry demonstrated that HS-1200 increased the proportion of sub-G1 hypoploid cells in both U937/ vector and U937/Bcl-2 cells (Fig. 1D). 

HS-1200-induced apoptosis is mediated by caspase-3 and Bcl-2 in U937/Bcl-2 cells.

Western blot analysis of HS-1200-treated U937/vector and U937/Bcl-2 cell lysates showed that both pro-caspase-3 and PARP were cleaved and Bcl-2 was down-regulated. This was also observed in cisplatin-treated U937/vector cells but not in cisplatin-treated U937/Bcl-2 cells. It is not able that HS-1200 substantially down-regulated Bcl-2 in U937/ Bcl-2 cells and in U937/vector cells, while down-regulation of Bcl-2 was not observed in cisplatin-treated U937/ Bcl-2 cells (Fig. 2A). Flow cytometry revealed an increase in the portion of cells with depolarized MMP in both U937/ vector and U937/Bcl-2 cells (Fig. 2B). cisplatin also induced these apoptotic manifestations in U937/vector cells, but they were substantially attenuated in cisplatin-treated U937/ Bcl-2 cells. 

Aggregated or mature PML-NBs are formed in U937 cells undergoing apoptosis.

We next asked whether PML-NBs are involved in the resistance to apoptosis conferred by Bcl-2 in human leukemic U937 cells. Visualization of the PML-NBs by confocal microscopy after HS-1200 treatment demonstrated morphological alterations in the nuclear bodies. Whereas PMLNBs were microspeckled in untreated U937/vector or U937/ Bcl-2 cells, HS-1200 induced the formation of mature or aggregated PML-NBs in all three cell types. Although the aggregation of PML-NBs was also observed in cisplatintreated U937/vector cells, this was not observed in cisplatintreated U937/Bcl-2 cells (Fig. 3). 

Fig. 1. HS-1200 overcomes the resistance conferred by Bcl-2 in U937 cells via inducing apoptosis. (A) Cell viability assay 24 h after treatment with cisplatin or HS-1200 shows that the overexpression of Bcl-2 significantly attenuated the cisplatin-induced reduction of viability whereas HS-1200 reduced the viability to a similar extent in both the U937/vector and U937/Bcl-2 cells. Statistical significance: *P<0.05. Three independent assays were performed. Values are means ± SD of triplicates of each experiment. (B) DNA electrophoresis conducted on the cells harvested 24 h after treatment. DNA ladder was demonstrated in 50 μM HS-1200-treated U937/vector and U937/Bcl-2 cells and 5 µg/ml cisplatin-treated U937/vector cells but not in 5µg/ml cisplatin-treated U937/Bcl-2 cells. cis, cisplatin 5 μg/ml. HS, HS-1200 50 µM. (C) Hoechst staining showing nuclear morphology of U937 cells 24 h after cisplatin or HS-1200 treatment. Numbers represent the percentage of apoptotic cells having condensed nuclei. Nuclear condensation or fragmentation was observed in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. Statistical significance: **P<0.01. Bar, 10 µm. (D) Representative histograms showing cell cycle progression and induction of apoptosis. Apoptotic population was increased in 50 µM HS-1200-treated U937/vector and U937/Bcl-2 cells, and 5 µg/ml cisplatin-treated U937/vector cells but not in 5 μg/ml cisplatin-treated U937/Bcl-2 cells. Statistical significance: **P<0.01.

Fig. 2. HS-1200-induced apoptosis was mediated by caspase-3 and Bcl-2 in U937/Bcl-2 cells. (A) Western blots of caspase-3, PARP and Bcl-2. Cleavage of pro-caspase-3 and PARP, and down-regulation of Bcl-2 was demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. GAPDH was used as a loading control. (B) Flow cytometry data showing the reduction of mitochondrial membrane potential in cells undergoing apoptosis. Mitochondrial retention of JC-1 decreased in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. Statistical significance: **P<0.01.

Fig. 3. Aggregated or mature PML-NBs are formed in U937 cells undergoing apoptosis. Confocal microscopy images showing the morphology of PML-NBs after 24 h treatment. Nucleus was stained with DAPI (blue) which is highly specific for DNA. Microspeckled PML-NBs are demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells.

HS-1200 induces the co-localization between PML and SUMO-1, Daxx, or Sp100 in the aggregated PML-NBs of U937/Bcl-2 cells.

Confocal images showed that SUMO-1, Daxx and Sp100 displayed a diffuse nuclear localization pattern in the control cells, although they also co-localized with PML in discrete microspeckled PML-NBs. However, enhanced co- localization of PML with SUMO-1, Daxx and Sp100 was observed in the aggregated PML-NBs of both HS-1200 treated U937/vector and U937/Bcl-2 cells. Although this enhanced co-localization was also observed in cisplatin- treated U937/vector cells, it was not observed in cisplatin- treated U937/Bcl-2 cells (Fig. 4A, 5A and 5C). The profile of PML and PML-NB components showed that SUMO-1, Daxx, and Sp100 proteins are concentrated within PML- NBs in HS-1200- or cisplatin-treated U937/vector and HS- 1200-treated U937/Bcl-2 cells, but not in cisplatin-treated U937/Bcl-2 cells (Fig. 4B, 5B and 5D). Co-immunoprecipitation assays showed that PML physically interacts with SUMO-1, Daxx, and Sp100 in HS-1200- or cisplatin- treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells (Fig. 4C and 5E).

As a control, we examined whether SC35 protein, which is known to be a constituent not of PML-NB but of another nuclear body SC35 speckle, associated with mature PMLNBs. Confocal images and the profile of PML and SC35 showed that SC35 protein was not concentrated on PMLNBs (Fig. 4D and E). 

HS-1200 induces the co-localization of PML with p53 and CBP in the aggregated PML-NBs of U937/Bcl-2 cells and the formation of p53-PML-CBP trimeric complex.

Confocal images of PML, p53, and CBP showed that p53 and CBP display a diffuse nuclear localization pattern in the control cells. Noticeably, both p53 and CBP co- localized with PML in the aggregated PML-NBs of HS- 1200-treated U937/Bcl-2 as well as in U937/vector cells. Although these interactions were also observed in cisplatin- treated U937/vector cells, they were not observed in cisplatin- treated U937/Bcl-2 cells (Fig. 6A, B, C and D) Co- immunoprecipitation experiments were completed to examine the physical interactions between p53, PML and CBP. Results showed that a p53-PML-CBP trimeric complex was formed in both HS-1200-treated U937/vector and U937/ Bcl-2 cells. Although the formation of the trimeric complex was also observed in cisplatin treated U937/vector cells, it was not observed in cisplatin-treated U937/Bcl-2 cells (Fig. 6E and F).

Fig. 4. HS-1200 induces the co-localization between PML and SUMO-1. (A) Confocal microscopy images showing the co- localization of PML and SUMO-1. The enhanced co-localization is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (B) The profile of SUMO-1 and PML fluorescence intensity is depicted. The intensity of SUMO-1 is shown in red, and that of PML is shown in green. In contrast to untreated U937/vector cells, SUMO-1 protein is concentrated within PML-NBs in HS-1200 or cisplatin-treated U937/vector and HS- 1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (C) Co-immunoprecipitation assay data showing the interaction of PML and SUMO-1. Interaction of PML and SUMO-1 is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (D) Confocal microscopy images showing that SC35 does not associate with PML. (E) The profile of SC35 and PML fluorescence intensity is depicted. The intensity of SC35 is shown in red, and that of PML is shown in green. SC35 protein is not concentrated within PML-NBs.

Fig. 5. HS-1200 induces the co-localization between PML and Daxx or sp100 (A, C) Confocal microscopy images showing the co-localization of PML and Daxx or sp100. The enhanced co-localization is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (B, D) The profiles of Daxx or sp100 and PML fluorescence intensity are depicted. The intensity of Daxx is shown in red, and that of PML is shown in green. In contrast to untreated U937/vector cells, Daxx or sp100 proteins are concentrated within PML-NBs in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (E) Co-immunoprecipitation assay data showing the interaction of PML and DAXX. (left panels) or sp100 (right panels) Interaction of PML and DAXX or sp100 is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells.

Fig. 6. HS-1200 induces the co-localization between PML and p53 or CBP. (A, C) Confocal microscopy images showing the co-localization of PML and p53 or CBP. The enhanced co-localization is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (B, D) The profiles of p53 or CBP and PML fluorescence intensity are depicted. The intensity of p53 or CBP are shown in red, and that of PML is shown in green. In contrast to untreated U937/vector cells, p53 or CBP proteins are concentrated within PML-NBs in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells. (E, F) Co-immunoprecipitation assay data showing the formation of p53-PML-CBP trimeric complex. Formation of p53-PML-CBP trimeric complex is demonstrated in HS-1200 or cisplatin-treated U937/vector and HS-1200-treated U937/Bcl-2 cells but not in cisplatin-treated U937/Bcl-2 cells.

Discussion

 The present study has demonstrated that HS-1200 at 50 μM induces apoptosis to a similar extent in both U937 and U937/Bcl-2 cells. These data indicate that HS-1200 at the usual dose overcomes the resistance to apoptosis conferred by Bcl-2-overexpressing human U937 cells. Although the information provided by the present study is still fragmentary and contradictory, these data raise the possibility that HS-1200 could be an effective strategy against Bcl-2-overexpressing human leukemia that have acquired resistance to standard chemotherapeutic agents. Since the mechanism by which the synthetic chenodeoxycholic acid (CDCA) derivative, HS-1200, bypasses the anti-apoptotic effects of Bcl-2 in human leukemic cells remains to be determined, deciphering the molecular mechanism of their actions is an important and challenging task. Arguably, the most intriguing finding of my study is the formation of the mature promyelocytic leukemia (PML) nuclear bodies in HS-1200treated Bcl-2-overexpressing U937 cells. PML-NBs, which are macromolecular nuclear domains are present in virtually every mammalian cell, proteinaceous structures that are found predominantly in the nucleus [21]. Under normal growth conditions, each cell usually harbors several dozen PML-NBs [22]. PML protein is a major constituent of PML-NBs. PML protein was originally identified in leukemic blasts from patients suffering from acute promyelocytic leukemia (APL) [23]. In APL blasts, the expression of the oncogenic PML-RARα fusion protein, the product of a reciprocal chromosomal translocation t(15;17), disrupts the structural integrity of PML-NBs [24]. The formation of PML-NBs is known to be associated with a large number of function of fundamental cellular processes [25].

The assembly of PML into PML-NBs is frequently accompanied by dynamic changes in the structure of PML-NBs and PML gene expression. PML-NB formation is also associated with modifications of PML protein and changes in the association of PML with other proteins [26]. To date, numerous have been reported to be resent in, or associated with, PML-NBs. One group of proteins is constitutively present in PML NBs and includes PML itself, Sp100, and members of the Small Ubiquitin-like Modifier (SUMO) family. Other group of proteins may be readily detected in, or associated with, PML NBs only after specific treatments, which includes p53 [26-28]. 

Various dynamic changes in the molecular structures of PML-NB constituents and their interactions are associated with the regulation of apoptosis. Sumo is an important role in the assembly. The assembly of PML into macromolecular PML-NBs depends on its covalent post-translational modification with SUMO-1. Sumoylation pathway also play an important roles in the assembly of PML-NBs in cells undergoing apoptosis. The covalent post-translational modification of PML with SUMO-1 regulates the interaction between PML and other PML-NB components such as Sp100, Daxx, p53 and CBP [29]. SUMO-conjugated PML recruits Daxx from the chromatin into the PML-NBs [30]. Daxx participates in other apoptosis pathways as a transcriptional co-repressor or co-activator [31]. The interaction between PML and p53 in PML-NBs is also biologically important. PML directly interacts with the DNA-binding domain of p53 and recruits p53 to PML-NBs, which leads to p53 transcriptional activation and apoptosis [32,33]. PML is also involved in the phosphorylation and acetylation of p53 [26,34]. Besides p53 itself, multiple factors that regulate p53, including CBP, are found within, or in association with, PML-NBs [35].

Although previous descriptions of the morphology of PML- NBs are not consistent and morphological characteristics are presumed to be cell specific, PML-NBs are mostly microspeckled in control cells and aggregated in cells undergoing apoptosis. An elaborate study demonstrated that dephosphorylation targets PML to the nuclear matrix which results in the formation of primary PML-NBs. Sumoylation then induces the maturation to secondary, or mature, PML- NBs. In mature PML-NBs, PML forms the outer shell of the structure, and many of the proteins described above are found within the electron clear core of PML [35,36]. The presence of these factors in mature PML-NBs seems to affect cellular sensitivity to apoptosis. Furthermore, SUMO-dependent recruitment of proteasome components into mature PML-NBs was shown to coincide with PML degradation, which plays a key role in the induction of apoptosis [37].

 In the study presented here, we observed the formation of mature PML-NBs in all experimental groups of cells undergoing apoptosis after HS-1200 treatment. Importantly, these data show that HS-1200-induced formation of mature PML-NBs correlates with overcoming the anti-apoptotic effects of Bcl-2-overexpressing U937 cells. These findings suggest that PML and the formation of mature PML-NBs could represent therapeutic targets that could overcome the resistance to apoptosis that is conferred by Bcl-2. However, the mechanism by which PML bypasses the anti-apoptotic effects of Bcl-2 still remains to be elusive. Although some previous studies have shown that proteins localized within the nuclear bodies are implicated in the regulation of transcription, only a few published sources provide information on the regulation of the Bcl-2 gene by PML-NB constituents. One study showed that Daxx protein that is localized to PML-NBs can downregulate Bcl-2 [37]. Based on these studies, we propose that HS-1200 induces the localization of PML-NB protein constituents to the PML-NBs and that the interactions between these proteins affect Bcl-2 expression. We believe that future investigations may provide important information for understating the underlying mechanism.

 In conclusion, HS-1200-induced apoptosis in overcoming the resistance conferred by Bcl-2 is associated with the formation of mature PML nuclear body in human leukemic U937 cells. The exact molecular mechanism by which PML regulates Bcl-2 remains an open question. It may be anticipated that the answer to the question will yield further insights into the mechanism of the resistance conferred by Bcl-2 and the development of therapeutic strategies to overcome this resistance.

Acknowledgement

 This work was supported by for two years Pusan National University research grant.