Knocking-Down Cyclin A2 by siRNA Suppresses Apoptosis and Switches Differentiation Pathways in K562 Cells upon Administration with Doxorubicin

Cyclin A2 is critical for the initiation of DNA replication, transcription and cell cycle regulation. Cumulative evidences indicate that the deregulation of cyclin A2 is tightly linked to the chromosomal instability, neoplastic transformation and tumor proliferation. Here we report that treatment of chronic myelogenous leukaemia K562 cells with doxorubicin results in an accumulation of cyclin A2 and follows by induction of apoptotic cell death. To investigate the potential preclinical relevance, K562 cells were transiently transfected with the siRNA targeting cyclin A2 by functionalized single wall carbon nanotubes. Knocking down the expression of cyclin A2 in K562 cells suppressed doxorubicin-induced growth arrest and cell apoptosis. Upon administration with doxorubicin, K562 cells with reduced cyclin A2 showed a significant decrease in erythroid differentiation, and a small fraction of cells were differentiated along megakaryocytic and monocyte-macrophage pathways. The results demonstrate the pro-apoptotic role of cyclin A2 and suggest that cyclin A2 is a key regulator of cell differentiation. To the best of our knowledge, this is the first report that knocking down expression of one gene switches differentiation pathways of human myeloid leukemia K562 cells.


Introduction
Tumor cells are characterized by deregulation of cell cycle checkpoints, leading to uncontrolled cell division and proliferation under conditions where non-transformed cells cannot enter and pass through the cell cycle. All these may come from overexpression of cyclins and the abnormal activation of cyclindependent kinases (CDKs) [1]. Cyclins are a superfamily of proteins whose levels vary in a cyclical fashion during the cell cycle to activate specific CDK required for the proper progression through the cell cycle. Cyclin A 2 , which is essential for initiation and progression of DNA replication as well as for cell cycle progression through G1/S and G2/M transitions [2][3][4][5], is overexpressed in a variety of human cancers compared with normal cells and tissues [6][7][8][9][10][11][12][13][14][15]. Deregulated expression of cyclin A 2 seems to be closely associated with early events in tumor transformation [6,15]. In addition, its expression level in many types of cancers appears to be of prognostic value such as prediction of aggressiveness, survival or early relapse [9][10][11][12][13][14].
Until recently it has been held that CDK2, presumed master of the known CDK isoforms, is a promising anticancer target for developing small molecule inhibitors. The first-generation CDK inhibitors, flavopiridol and CY-202, are in late-stage clinical trials, and have only modest activity [16]. Recent findings [17], however, suggest that CDK2 may not be a key cell cycle player and question whether selective CDK2 inhibition is a useful cancer therapy strategy. No cell cycle abnormalities are observed in either a CDK2 null mouse or following acute ablation of CDK2 in primary cells, indicating that this gene is not strictly required for cell proliferation [17]. In contrast, deletion of cyclin A 2 in knockout mice is associated with an embryonic lethal phenotype [18]. Moreover, Fine et al. demonstrated that cyclin A 2 and/or cyclin A 2 -CDK2 complex but not CDK2 is a promising anticancer target with a high therapeutic index [19]. Therefore, inhibitors of CDK2 may not be appropriate for cancer therapy and more efforts are focused on inhibition of cyclin A 2 and/or cyclin A 2 -CDK2 complex activity. We have shown that reduction of cyclin A 2 in human chronic myelogenous leukaemia K562 cells using small interfering RNA significantly inhibits cell proliferation [20], further supporting the notion that cyclin A 2 can serve as a novel therapeutic target.
Apoptosis and differentiation are the predominant two mechanisms by which chemotherapeutic agents kill tumor cells. Low dose of doxorubicin (DOX) induces erythroid differentiation in K562 cells, while high concentration of DOX promotes apoptosis [21]. Although many molecular pathways are involved in the apoptosis-regulatory mechanism, evidences suggest that the cell cycle and apoptosis may be interconnected [22][23][24][25][26][27][28]. Several studies have shown that increased expression of cyclin A 2 is found in cells in response to several apoptotic stimuli, but very few studies have dealt with this issue directly [29][30][31][32][33]. So it is important to clarify whether the decreased expression of cyclin A 2 is a cause of cell differentiation or a result of differentiation response.
Carbon nanotubes possess the unique features of being able to enter a living cell without causing its death or without inflicting other damage and can shuttle biological molecules into mammalian cells, indicating their potential application as a vector for the delivery of therapeutic molecules [34][35][36]. Recently we have reported that single wall carbon nanotubes (SWNTs) can induce a sequence-dependent B-A DNA transition [37], selectively induce human telomeric i-motif DNA formation [38], accelerate S1 nuclease cleavage rate [39], cause single-stranded poly(rA) to form a duplex structure and bind to human telomeric i-motif DNA under molecular-crowding conditions [40,41]. Herein, we explore whether altering the levels of cyclin A 2 in K562 cells using RNAi delivered by SWNTs can influence cell apoptosis and differentiation induced by chemotherapeutic agent DOX. K562 cells with reduced cyclin A 2 showed a significant decrease in growth suppression, apoptosis and erythroid differentiation, and were differentiated along megakaryocytic and macrophage-monocytic pathways upon administration with DOX. These findings indicate a positive correlation between cyclin A 2 and apoptosis induced by DOX and suggest that cyclin A 2 is a key regulator of cell differentiation, supporting the notion that cyclin A 2 is an important regulator for cell cycle as well as for cell apoptosis and differentiation. To the best of our knowledge, this is the first report that knocking down expression of one gene can switch K562 cells differentiation pathways.

Upregulation of cyclin A 2 during apoptosis of K562 cells induced by DOX
DOX can inhibit growth of a variety of cancer cells [42]. To measure apoptosis rates induced by varying concentrations of DOX in K562 cells, we use acridine orange (AO)/ethidium bromide (EB) staining assay (Fig. 1). The percentage of apoptotic cells increased with time in a dose dependent fashion. RT-PCR and western blotting were used for studying the effect of DOX on the expression of cyclin A 2 in K562 cells. As shown in Fig. 2, cyclin A 2 expression levels increased with the increase of DOX, and a positive correlation was observed. These results show that expression of cyclin A 2 was significantly up-regulated in DOX-treated K562 cells at time points when DOX caused a significant amount of apoptosis, indicating that the levels of cyclin A 2 are correlated with the ability of DOX to induce apoptosis in K562 cells.

Suppression of DOX-induced growth inhibition by downregulation of cyclin A 2 with siRNA delivered by SWNTs
We have previously demonstrated that functionalized single wall carbon nanotubes (SWNTs) can efficiently deliver siRNA targeting cyclin A 2 into K562 cells, resulting in specific suppression of cyclin A 2 expression [20]. Here, we use SWNTs to transfect cyclin A 2 siRNA into K562 cells and evaluate the effect of cyclin A 2 on growth inhibition induced by DOX. Two hours after transfection, DOX was added. Cell proliferation was examined by trypan blue exclusion method and methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay. Fig. 3 showed cell growth curves of K562 cells after various treatments. It can be seen that depletion of cyclin A 2 inhibited cell proliferation while carbon nanotubes vector had no apparent effect on growth inhibition as well as no additive or synergetic effect on cell toxicology of DOX. DOX (0.4 mM) significantly suppressed cell growth, nevertheless, much less growth inhibition was observed in cells incubated with both cyclin   A 2 siRNA and DOX, supported by microscopic results (Fig. S1). In order to further quantitively investigate the effect of downregulation of cyclin A 2 on the growth inhibition, MTT assays were carried out 24 h after incubation with DOX. Interaction of DOX with MTT was also checked in a cell free system and the results indicated that DOX showed no inference to MTT assay (Fig. S2). As shown in Fig. 4, IC 50 of non-transfected cells was about 2.5 mM. For siRNA transfected K562 cells, IC 50 was about 5.0 mM. The results clearly demonstrate that down-regulation of cellular cyclin A 2 level suppress DOX-induced growth inhibition.
Suppression of DOX-induced apoptotic cell death by down-regulation of cyclin A 2 with siRNA delivered by SWNTs To investigate whether cyclin A 2 participates in cell apoptosis death and determine if reduction of cyclin A 2 has an effect on apoptosis induced by DOX, siRNA specific for cyclin A 2 was transfected into K562 cells by SWNTs, and low dose of DOX (0.4 mM) was administered. As shown in Fig. 5, carbon nanotubes vector showed no apparent cell toxicology, while numerous larger cells and typical lobular nuclei were observed in cells incubated with DOX. For cells coadministered with cyclin A 2 siRNA and DOX, much less apoptotic nuclei were observed, and the size of live cells was much bigger than that of the control cells. Statistics analysis showed that down-regulation of cyclin A 2 significantly reduced the apoptosis rate from more than 60% to about 20%. Similar results were obtained by annexin V-PI double staining flow cytometric technique (as shown in Figure S3). These clearly indicated lowering cyclin A 2 level in K562 cells led to suppression of apoptosis, thus a marked decrease in DOX susceptibility, which provide direct evidence that cyclin A 2 is involved in cellular responses to apoptosis.

The cytoplasmic subcellular distribution of cyclin A 2 correlates with apoptosis of K562 cells induced by DOX
Early reports have demonstrated that there is a link between cyclin A 2 subcellular localization and its cell function, and the level of cyclin A 2 is correlated to cell apoptosis [22,23,26]. For clarifying whether the subcellular distribution of cyclin A 2 in K562 cells correlates with apoptosis induced by DOX, indirect immunofluorescence detection of cyclin A 2 was performed. As shown in Fig. 6A, a significant fraction of cells underwent apoptosis and orange nuclei were observed, where DOX was mostly located. The immunofluorescence labeling of cyclin A 2 showed its presence predominantly in the nucleus of control K562 cells (Fig. S4), whereas in cells administered with DOX, cyclin A 2 was mainly located at the cytoplasm of early and late phases of apoptotic cells (Fig. 6B). Cyclin A 2 labeling was not found in the K562 cells incubated with non-immune serum (Fig. 6C). The results indicated that translocation of cyclin A 2 from the nucleus to cytoplasm was connected with its role in apoptosis.

Suppression of cyclin A 2 by siRNA can switch differentiation pathways of K562 cells induced by DOX
As mentioned above, cells treated with both cyclin A 2 siRNA and DOX were much bigger than the control. Since enlarged phenotype may suggest cell differentiation, we performed the benzidine staining to assess erythroid differentiation, which was the differentiation pathway of K562 cells upon treatment with anthracycline antibiotics including DOX [42,43]. Representative microscopy images of the benzidine staining in K562 cells after various treatments were shown in Fig. S5. For untreated cultures and cells administered with SWNTs, the percentages of benzidine positive cells were very low (less than 2%). Forty hours after incubation with DOX, around 14% benzidine positive cells were observed. However, down-regulation of cyclin A 2 by siRNA in K562 cells substantially suppressed erythroid differentiation upon administration with DOX (less than 1% benzidine positive cells, as shown in Fig. 7).
To determine whether K562 cells with reduced cyclin A 2 upon treatment with DOX underwent megakaryocytic pathway and monocyte-macrophage differentiation, which are the other two differentiation pathways of K562 cells, flow cytometric measurement of megakaryocytic specific surface antigen CD61 (GPIIIa) and nitro blue tetrazolium (NBT) reduction assay were carried out, respectively. As shown in Fig. S6, in comparison with the control, cells co-administered with cyclin A 2 siRNA and DOX showed significant increase in granularity measured by side scatter (side scatter, on Y-axis) and cell size measured by forward scatter (forward scatter, on X-axis), which was in good agreement with our AO/EB staining observation. Cells treated with DOX or SWNTs showed no detectable expression of CD61 (GPIIIa) (shown in Fig.  S7), whereas in K562 cells co-administered with cyclin A 2 siRNA and DOX, the fraction of CD61 (GPIIIa) positive cells (,10%) was clearly observed (Fig. 8). For NBT reduction assay, representative microscopy images of K562 cells after various treatments were shown in Fig. S8. In cells treated with DOX or SWNTs, the percentage of NBT positive cells was very low (1%), whereas a small fraction NBT positive cells (,6%) was observed in K562 cells co-administered with cyclin A 2 siRNA and DOX ( Fig. 9). Although the fractions of cells undergoing megakaryocytic pathway (,10%) and monocyte-macrophage differentiation (,6%) were small, the difference is statistically significant compared to the control group. Considering that DOX is an erythroid differentiation-inducing agent for K562 cells [21,42,43], it is not surprising that only a small fraction of K562 cells with reduced cyclin A 2 underwent megakaryocytic pathway and monocyte-macrophage differentiation.
Taken together, these results indicate that knocking down the expression of cyclin A 2 suppressed DOX-induced erythroid differentiation and a small fraction of K562 cells with reduced cyclin A 2 were differentiated along megakaryocytic and monocyte- macrophage pathways upon treatment with DOX. These findings suggest that cyclin A 2 is an important regulator of cell differentiation.

Discussion
Cyclin A 2 is particularly interesting among the cyclin family because it can activate two different CDKs and functions in both S phase and mitosis. In S phase, phosphorylated cyclin A 2 -CDK2 complexes are suggested to play an important role in the initiation of DNA replication. In mitosis, cyclin A 2 may contribute to the control of cyclin B stability. Consistent with its role as a key cell cycle regulator, overexpression of cyclin A 2 is associated with transformed cells [6][7][8][9][10][11][12][13][14][15]. However, it is difficult to determine whether elevation of cyclin A 2 is a contributing factor or a mere consequence of the increased cell proliferation. In haematological malignancies, cyclin A 2 is associated with proliferation rate of these disorders and can be used for molecular diagnostics as a proliferation marker [13,14].
Single-walled carbon nanotubes (SWNTs) have been considered as the leading candidate for nanodevice applications ranging from gene therapy and novel drug delivery to membrane separations. We have previously showed that siRNA transfection efficiency of lipofectamine 2000 in K562 cells was low (28%) and some cells underwent apoptosis and necrosis during the process [20]. However, SWNTs could efficiently facilitate the coupling of siRNA to form siRNA:SWNTs complexes and carry siRNA into K562 cells, significantly knocking down the expression of target gene. No apparent cell toxicology was observed. Hence, in order to directly probe whether cyclin A 2 participates in cell apoptosis and differentiation, we employed SWNTs as transfection vector to deliver cyclin A 2 siRNA into K562 cells to specifically knock down the expression of cyclin A 2 and found that carbon nanotubes vector showed no additive or synergetic effect on cell toxicology of DOX, which was consistent with our previous report [20].
DOX, a prominent member of anthracycline antibiotics, has been extensively used for treatment of solid tumors and leukemia. It exerts its cytotoxic activity against cancer cells mainly by intercalation into DNA, inhibition of topoisomerase II and helicase activity, leading to cell-cycle arrest at the G2/M phase and apoptosis [42]. In clinical applications, doses of DOX are strictly limited by its cardiotoxicity [43]. It should be noted that the dose of DOX administrated here (0.4 mM) is pharmacological relevant compared to the initial or steady-state plasma concentrations observed in patients after standard bolus infusions (5 mM and 25-250 nM, respectively).
It has been reported that there is a link between cyclin A 2 and apoptosis [22][23][24][25][26][27][28]. Hoang et al. showed that in c-myc overexpressing serum deprived rat 1A fibroblasts undergoing apoptosis, cyclin A 2 mRNA expression was increased, in contrast to the invariant expression for cyclin B, C, D1 and E [24]. Moreover, serumdeprived rat 1A fibroblast stably transfected with cyclin A 2 exhibited  increased apoptosis following stimulation of cyclin A 2 expression. Meikrantz et al. reported that induction of apoptosis was uniformly associated with activation of cyclin A 2 -dependent kinases but not associated with cyclins E or B, and overexpression of the cyclin A 2 could circumvent the anti-apoptosis activity of the oncogene BCL-2 in human Hela cells [25,26]. Furthermore, Hiromura et al. indicated that apoptosis was associated with an increase in cytoplasmic cyclin A 2 -CDK2 activity following UV irradiation, under these conditions, nuclear cyclin A 2 -Cdk2 activity decreased significantly [27]. Our results showed knocking down the expression of cyclin A 2 in K562 cells significantly suppressed the apoptosis induced by DOX and a positive correlation between the levels of cyclin A 2 and apoptosis was observed. The findings also indicate that the cytoplasmic subcellular distribution of cyclin A 2 correlates with its pro-apoptotic role. We speculate that cyclin A 2 associated kinases are involved in DOX-induced apoptotic cell death pathways in K562 cells, although the exact downstream mechanisms are not known.
We have demonstrated that SWNTs could effectively deliver cyclin A 2 siRNA into K562 cells, significantly suppressing the expression of cyclin A 2 with specificity and cell proliferation, and cells with reduced cyclin A 2 showed a decrease in the percentage of cells in S phase [20]. Several studies have indicated that cancer cells with a high S-phase fraction/high proliferative activity are more sensitive to apoptosis induced by chemotherapy [44,45]. As for DOX, it is active throughout the cell cycle, but the effect is most pronounced for cells in S phase-G2 phase, especially in S phase, where it interferes with the DNA replication and transcription [42]. Therefore, it was not surprising that K562 cells with reduced cyclin A 2 showed a marked decrease in DOX susceptibility.
Most of chemotherapeutic agents show significant side effects and not all patients benefit from aggressive chemotherapy. Therefore, searching for tumor biological factors which can predict patient prognosis and chemotherapy response would be of most importance. Several studies have indicated that a high level of cyclin A 2 expression may be a marker of poor prognosis in cancers [10][11][12][13]. Besides, previous studies have shown that cancer patients with high level of cyclin A 2 had better chemotherapy response and survival than those with reduced cyclin A 2 and low expression of cyclin A 2 , indicating that the patients with high expression of cyclin A 2 are more suitable for chemotherapy [46][47][48]. Our results demonstrate the pro-apoptotic role of cyclin A 2 in human myeloid leukemia K562 cells, and indicate that cells with low level of cyclin A 2 were more resistant to chemotherapeutic agent DOX. Poon et al. have suggested that a decrease of cyclin A 2 , rather than increase, promotes tumorigenesis, and once the tumor has developed, high levels of cyclin A 2 simply reflect a high proliferation rate, which can explain this inconsistency [49]. Hence, despite its association with transformed cells, evaluating cyclin A 2 level in patients will be an important prognostic marker for use of chemotherapy. Patients with high level of cyclin A 2 may be more responsive to anticancer drugs through the induction of apoptotic cell death. Moreover, it should be cautious to combine doxorubicin chemotherapy with any small molecule drug targeting cyclin A 2 /cyclin A 2 associated kinases since it can enhance potential drug resistance.
In several systems, it has been reported that down-regulation of cyclin A 2 and its associated CDK 2 activity are important for successful differentiation [29][30][31][32][33]   erythroleukemia cells induced by hexamethylene is accompanied by a decrease in the level of cyclin A 2 and CDK2 proteins and the persistent suppression of cyclin A 2 expression may play a role in HMBA-induced commitment to terminal differentiation [30]. Yoshizumi et al. showed down-regulation of cyclin A 2 gene expression in vivo at both the RNA and protein levels appears to be important in the permanent withdrawal of human and rat cardiomyocytes from the cell cycle during development [32]. Moreover, Rieber et al. demonstrated that the interaction of cyclin A 2 with E2F is the target for tyrosine induction of B16 melanoma terminal differentiation [33]. In this work, we found that knocking down the expression of cyclin A 2 in K562 cells significantly suppressed DOX-induced erythroid differentiation and a small fraction of cells with reduced cyclin A 2 were differentiated along megakaryocytic and monocyte-macrophage pathways upon treatment with DOX. To the best of our knowledge, this is the first report that knocking down expression of one gene can switch K562 cells differentiation pathways. The results suggested that cyclin A 2 is directly involved in the checkpoint of cell differentiation pathways and is a key regulator of this process, although the detail downstream mechanisms are not known. For cancer cells with low level of cyclin A 2 , which are less responsive to chemotherapeutic agents, induction of differentiation might be an alternative strategy. Combination of cyclin A 2 siRNA and DOX may provide a novel option of such therapeutic strategy.
In conclusion, knocking down the expression of cyclin A 2 by siRNA delivered by SWNTs suppresses apoptosis and erythroid differentiation, and promotes megakaryocytic and monocytemacrophage differentiation in human chronic myelogenous leukaemia K562 cells upon administration with DOX. The results demonstrate the pro-apoptotic role of cyclin A 2 and suggest that cyclin A 2 is a key regulator of cell differentiation, supporting the notion that cyclin A 2 is an important regulator for cell cycle as well as for cell apoptosis and differentiation.

Materials and Methods
Ethics statement. The human erythroleukemic cell line K562 [20] was used in this study.

Cell culture
The human erythroleukemic cell line K562 [20] was grown in Iscove's modified Dulbecco's medium (Gibco BRL) supplemented with 10% fetal calf serum in a humidified 37uC incubator with 5% CO 2 . Cells were passed three times per week. DNA fluorochrome staining (DAPI or Hoechst 33258) is used as our routine mycoplasma detection and the cells used for experiments are free of contamination. Exponentially growing cells were used for all experiments described below. Cell viability was determined by trypan blue exclusion in a haemocytometer chamber.

Transfection
The siRNA oligonucleotides were synthesized by Genepharma Corporation (Shanghai, China). The sequence used for targeting silencing of cyclin A 2 was 59-CCAUUGGUC CCUCUU-GAUUTT-39. The nonsilencing control siRNA is an irrelevant siRNA with random nucleotides UUCUCCGAACGUGUCAC-GUTT. Functionalized single wall nanotubes (f-SWNTs) were prepared according to method described in our previous work [20]. Cyclin A 2 siRNA: f-SWNTs complexes (w f-SWNTs /w siRNA = 40) were added at 25 n mol/L (siRNA concentration) to culture of cells.

Reverse transcriptase-polymerase chain reaction (RT-PCR)
Total RNA was extracted using Trizol reagent (Invitrogen) according to the manufacturer's instructions. The primers and conditions for cyclin A 2 were TCCATGTCAGTGCTGA-GAGGA (59), GAAGGTCCATGAGACAAGGC (39); 94uC for 30 seconds , 60uC for 30 seconds, 72uC for 1 minute for 25 cycles. Three introns are present in this pair of primer so that any contaminating genomic DNA would not be amplified. Primers used for the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were ACCTGACCTGCCGTCTAGAA (59), TCCAC-CACCCTGTTGCTGTA (39). Two sets of primers were used for each sample, including primers specific for the gene of cyclin A 2 and primers for GAPDH as an internal control. All PCR products were visualized on 1.5% agarose gel with 0.5 mg mL 21 EB.

MTT assay
K562 cells were plated in 96-well culture plate at a concentration of 0.5610 4 cells/well, and were transfected with cyclin A 2 siRNA or not by SWNTs. Two hours later, the cells were treated with various concentrations of DOX while the blank control wells were added medium without drug. Cells were then cultured for another 24 hours and 20 mL MTT (5 mg/mL) was added in each well, followed by additional four hour incubation. The supernatants were then discarded carefully and 150 mL dimethylsulphoxide was added and shaken vigorously to dissolve the purple precipitation formation. Optical density (OD) of each well was tested using Bio-Rad model-680 microplate reader with a wavelength of 490 nm.

Cell fluorescence staining
Cells were collected by centrifugation at 2006g for 5 minutes, and then washed twice with PBS. Cell concentration was adjusted as 2610 6 25610 6 cells/mL. 1 mL EB/AO dye mix was added in 10 mL cell suspension, followed by 10 minutes of incubation in dark. Stained cells suspension were placed on a clean microscope slide and covered with a cover-slip. Cells were viewed and counted using an Olympus BX-51 optical system microscope (Tokyo, Japan) at 4006 magnification with a blue filter. Pictures were taken with an Olympus digital camera. We note that the definition is sharper by eye through the microscope than in the photo. Tests were done in triplicate, counting a minimum of 300 total cells from at least three random microscope fields each.

Indirect immunofluorescence detection
Cells were harvested, washed three times with PBS, fixed with 4% paraformaldehyde in PBS for 30 min at 4uC and permeabilized with 0.5% Triton X-100 in PBS for 5 min at 4uC. After blocking in 10% goat serum, primary cyclin A 2 antibody (rabbit, polyclonal; Lab Vision Corporation, CA) was diluted to 1:400. Goat anti-rabbit FITC-conjugated secondary antibody (Jackson Immunoresearch, Stratech, UK) was used at a 1:200 dilution. If necessary, DAPI was used to visualize cell nuclei. Cells were observed and photographed by fluorescence microscopy with oil immersion objective and appropriate filters.

Flow cytometry analysis
Expression of GPIIIa is considered the most selective marker of the megakaryocyte lineage since it is not expressed on cells of other hematopoietic lineages from normal human bone marrow. Hence, staining of cells for surface CD61 (GPIIIa) was used to evaluate megakaryocytic differentiation. It employed a mouse monoclonal antibody fluorescein isothiocyanate (FITC)-conjugated anti-CD61 (eBioscience) or isotype-matched immunoglobulin (IgG1-FITC, eBioscience) at a concentration of 0.6 mg/mL. Cells were harvested, washed, resuspended in PBS containing 10% fetus calf serum and 0.1% NaN 3 and incubated for 30 min on ice with antibodies in the dark. After washing three times, cells were resuspended in PBS containing 0.5% formaldehyde and 0.1% NaN 3 , and then analyzed on a FACS Arial cytometer (Becton-Dickinson, San Diego, CA, USA). Viable cells were gated using forward and side scatter characteristics. Fluorescence intensity data were acquired using the BD FACSDIVA TM software.

NBT reduction assay
NBT dye reduction was used to qualitatively monitor monocytemacrophage differentiation. Briefly, cells were collected, washed with PBS and resuspended in IMDM medium without serum. Cells suspension were mixed with an equal volume of 0.1% NBT dissolved in PBS and incubated at 37uC for 40 min. NBT was reduced to insoluble formazan because of the intracellular oxygen radical release in the cells differentiated to monocytes-macro-phages. The percentage of cells containing intracellular reduced blue-black formazan was determined by light microscopy. At least 300 cells per preparation were observed.

Statistical analysis
Data are expressed as mean6s.d. and analysis of variance was carried out using Student's t test with Origin 7.5 (OriginLab Corporation, Northampton, MA, USA), where p,0.05 was considered significant.  Figure S4 Indirect immunofluorescence detection of cyclin A2 in control K562 cells. DAPI was used to visualize cell nuclei. Cells were viewed using an Olympus BX-51 optical system microscope (Tokyo, Japan) with oil lens and appropriate filters. Representative stained fields are shown: (A), DAPI staining (blue); (B), immunofluorescence detection of cyclin A2 (FITC, green); (C) merged image. As indicated, cyclin A2 was located at the nucleus of K562 cells without DOX treatment. Found at: doi:10.1371/journal.pone.0006665.s004 (0.43 MB TIF) Figure S5 Representative microscopy images of the benzidine staining of K562 cells after various treatments. Cells were transfected with cyclin A2 siRNA or not two hours prior to the addition of 0.4 mM DOX. Forty hours later, erythroid differentiation was scored by the benzidine staining method as described in Materials and Methods section. Cells were viewed and counted using an Olympus BX-51 optical system microscope (Tokyo, Japan) at 2006 magnification. Four independent tests were performed. Pictures were taken with an Olympus digital camera. Found at: doi:10.1371/journal.pone.0006665.s005 (0.65 MB TIF) Figure S6 Morphological changes of K562 cells after various treatments. Cells were transfected with cyclin A2 siRNA or not two hours prior to the administration of 0.4 mM DOX. Cells were then cultured for another 32 hours. Flow cytometry was used to show changes in size (forward scatter, on X-axis) and granularity (side scatter, on Y-axis). No significant morphological changes of K562 cells were observed upon RNAi mediated by SWNTs compared to the control. Cells administered with DOX underwent G2/M arrest and changes in granularity and cell size were clearly observed. Although knocking-down the expression of cyclin A2 by RNAi significantly inhibited growth suppression and apoptosis induced by DOX, cells co-administered with siRNA targeting for cyclin A2 and DOX showed similar increases in cell size and granularity compared to those treated with DOX alone. Found at: doi:10.1371/journal.pone.0006665.s007 (1.96 MB TIF) Figure S8 Representative microscopy images of the NBT reduction assay of K562 cells after various treatments. Cells were transfected with cyclin A2 siRNA or not two hours prior to the addition of 0.4 mM DOX. Ninety six hours later, NBT dye reduction was used to qualitatively monitor monocyte-macrophage differentiation. Cells were viewed and counted using an Olympus BX-51 optical system microscope (Tokyo, Japan) at 2006 magnification. Two independent tests were performed. Pictures were taken with an Olympus digital camera. Found at: doi:10.1371/journal.pone.0006665.s008 (0.34 MB TIF)