Hyaluronan-CD44 Interaction with Protein Kinase Cϵ Promotes Oncogenic Signaling by the Stem Cell Marker Nanog and the Production of MicroRNA-21, Leading to Down-regulation of the Tumor Suppressor Protein PDCD4, Anti-apoptosis, and Chemotherapy Resistance in Breast Tumor Cells*

Multidrug resistance and disease relapse is a challenging clinical problem in the treatment of breast cancer. In this study, we investigated the hyaluronan (HA)-induced interaction between CD44 (a primary HA receptor) and protein kinase Cϵ (PKCϵ), which regulates a number of human breast tumor cell functions. Our results indicate that HA binding to CD44 promotes PKCϵ activation, which, in turn, increases the phosphorylation of the stem cell marker, Nanog, in the breast tumor cell line MCF-7. Phosphorylated Nanog is then translocated from the cytosol to the nucleus and becomes associated with RNase III DROSHA and the RNA helicase p68. This process leads to microRNA-21 (miR-21) production and a tumor suppressor protein (e.g. PDCD4 (program cell death 4)) reduction. All of these events contribute to up-regulation of inhibitors of apoptosis proteins (IAPs) and MDR1 (multidrug-resistant protein), resulting in anti-apoptosis and chemotherapy resistance. Transfection of MCF-7 cells with PKCϵ or Nanog-specific small interfering RNAs effectively blocks HA-mediated PKCϵ-Nanog signaling events, abrogates miR-21 production, and increases PDCD4 expression/eIF4A binding. Subsequently, this PKCϵ-Nanog signaling inhibition causes IAP/MDR1 down-regulation, apoptosis, and chemosensitivity. To further evaluate the role of miR-21 in oncogenesis and chemoresistance, MCF-7 cells were also transfected with a specific anti-miR-21 inhibitor in order to silence miR-21 expression and inhibit its target functions. Our results indicate that anti-miR-21 inhibitor not only enhances PDCD4 expression/eIF4A binding but also blocks HA-CD44-mediated tumor cell behaviors. Thus, this newly discovered HA-CD44 signaling pathway should provide important drug targets for sensitizing tumor cell apoptosis and overcoming chemotherapy resistance in breast cancer cells.

of the important surface markers on cancer stem cells (27). HA binding to CD44 is known to be involved in the stimulation of both receptor kinases (e.g. ErbB2, epidermal growth factor receptor, and TGF␤ receptors) and non-receptor kinases (e.g. c-Src and ROK) (28 -34) required for a variety of tumor cellspecific functions leading to tumor progression.
Protein kinase C (PKC), a family of serine-threonine kinases, plays a pivotal role in signal transduction and a number of cellular functions (35). It consists of at least 11 different isoforms, including the novel type of PKC isoforms, such as PKC⑀ (36). A previous study found that PKC⑀ is associated with the anti-apoptotic Bcl-2 family of proteins (37). PKC also functions to prevent apoptosis in a number of cells by up-regulating inhibitors of apoptosis (IAP) proteins (e.g. X-linked IAP (XIAP) and survivin) and by inhibiting caspases (37,38). Down-regulation of PKC⑀ by treating cells with PKC inhibitors sensitizes tumor necrosis factor-␣-mediated cell death in breast tumor cells (39).
Thus, PKC⑀ appears to be functionally linked to anti-apoptotic effects and survival pathways in tumor cells.
In addition, activation of certain PKC isoforms has been implicated in the induction and maintenance of the multidrugresistant (MDR) phenotype (40). Specifically, an increase in PKC⑀ expression is closely associated with the drug-resistant phenotype in epithelial tumor cells (40). P-glycoprotein (P-gp), the product of the MDR1 (ABCB1) gene, is a transmembrane ATP-dependent transporter known to play a role in drug fluxes and chemotherapeutic resistance in a variety of cancers (41). A number of studies have shown that both HA and CD44 are involved in chemotherapeutic drug resistance in many cancer types (5,6,(42)(43)(44)(45)(46)(47)(48). In particular, the stem cell marker, Nanog, appears to interact with Stat-3 (signal transducer and activator of transcription protein 3) in the nucleus, leading to transcriptional activation, MDR1/P-gp expression, and chemotherapy resistance in HA-CD44-activated breast tumor cells (5). The question of whether there is a functional link between PKC⑀ and Nanog signaling in HA-CD44-mediated oncogenesis and drug resistance in breast tumor cells has not yet been addressed.
The miRNAs are evolutionarily conserved and function as negative regulators of gene expression by inhibiting the expression of mRNAs that contain complementary target sites referred to as the "seed region" (49). Previous data have revealed that human miRNAs are processed from capped and polyadenylated transcripts that are precursors to the mature miRNAs (pri-miRNAs) (50). In mammalian miRNA biogenesis, primary transcripts of miRNA genes (pri-mRNAs) are subsequently cleaved to produce an intermediate molecule containing a stem loop of ϳ70 nucleotides (pre-mRNAs) by the nuclear RNase III enzyme DROSHA and exported from the nucleus by exportin 5 (49). A second RNase III enzyme, Dicer, then generates the mature miRNA, which is loaded into the RNA-induced silencing complex in association with the argonaute protein that induces silencing via the RNA interference pathway (51). Although Dicer has an important role in the silencing action of miRNAs, recent studies have shown that silencing can still occur in cells that lack Dicer (52). It has recently been shown that the nuclear p68-RNA helicase is required in the uptake of certain miRNAs into the silencing complex (53). p68 belongs to a family of proteins that are involved in RNA metabolism pro-cesses, such as translation and RNA degradation (54). A previous study showed that miR-21 processing or biogenesis (via the precursor pri-miR-21) required p68 and DROSHA in breast tumor cells (55). Several transcription factors, including Nanog, also appear to be involved in the regulation of pri-miRNA expression during development (56). Whether HA-CD44-mediated signaling is involved in miR-21 maturation/production and chemotherapy resistance in breast tumor cells has not been determined.
Accumulating evidence indicates the involvement of noncoding miRNAs (ϳ22 nucleotides) in both cancer development and multidrug resistance (57). Analysis of the array profile of miRNA expression in normal breast and breast carcinoma tissues reveals that miRNA-21 (miR-21) is abundantly produced in tumors compared with normal tissues (57). The functional significance of miR-21 has been elucidated in several recent studies following the discovery of its specific targets (58). miR-21 is now one of the most studied miRNAs due to its involvement in cancer progression. It has recently been shown that miR-21 plays a role in the inhibition of a tumor suppressor protein, such as PDCD4 (program cell death 4) via a conserved site within the 3Ј-untranslated region of the mRNA (58,82). Down-regulation of PDCD4 expression by miR-21 leads to tissue invasion and metastasis (58,82). Thus, miR-21 is currently considered to be an oncogene.
In this study, we have investigated a novel HA-CD44-mediated PKC⑀ signaling mechanism that regulates the stem cell marker (Nanog)-associated miR-21 production. Our results indicate that HA-CD44-activated PKC⑀ stimulates Nanog phosphorylation, which in turn, activates Nanog signaling-regulated miR-21 production. These events lead to the tumor suppressor protein (PDCD4) reduction, IAP/MDR1 (P-gp) overexpression, anti-apoptosis, and chemoresistance in breast tumor cells. Inhibition of either PKC⑀-Nanog signaling or silencing of miR-21 expression/function by transfecting breast tumor cells with PKC⑀ siRNA (or Nanog siRNA) or anti-miR-21 inhibitor not only results in PDCD4 up-regulation and PDCD4-eIF4A complex formation but also causes a reduction of IAP/MDR1 (P-gp) and an enhancement of apoptosis and chemosensitivity. Our findings provide important new insights into understanding the roles that HA-CD44-mediated PKC⑀ activation and Nanog-regulated miR-21 play in regulating anti-apoptosis and chemotherapy resistance in breast tumor cells.

MATERIALS AND METHODS
Cell Culture-Human breast tumor cell line MCF-7 cells were purchased from ATCC (Manassas, VA) and grown in RPMI 1640 medium supplemented with 10% fetal bovine serum. Cells were routinely serum-starved (and thereby deprived of serum HA) before adding HA.
Antibodies and Reagents-Monoclonal rat anti-CD44 antibody (clone 020; isotype IgG 2b ; obtained from CMB-TECH, Inc., San Francisco, CA) recognizes a determinant of the HAbinding region common to CD44 and its principal variant isoforms (21-24, 28 -34). This rat anti-CD44 was routinely used for HA-related blocking experiments and immunoprecipitation. Immunoreagents, such as rabbit anti-PKC⑀ antibody, goat anti-Nanog antibody, mouse anti-PDCD4 antibody, rabbit anti-MDR1 (P-glycoprotein 170) antibody, and goat anti-actin antibody were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Several immunoreagents, including rabbit anti-DROSHA antibody, mouse anti-p68 antibody, and rabbit anti-phosphoserine antibody, were obtained from Millipore (Billerica, MA). Rabbit anti-eIF4A and rabbit anti-survivin were purchased from Cell Signaling Technology, Inc. (Danvers, MA) and Abcam (Cambridge, MA), respectively. Mouse anti-XIAP antibody was from BD Biosciences. Doxorubicin hydrochloride and paclitaxel (Taxol) were obtained from Sigma. Healon HA polymers (ϳ500,000-dalton polymers), purchased from Amersham Biosciences and The Upjohn Co. were prepared by gel filtration column chromatography, using a Sephacryl S1000 column. The purity of the HA polymers used in our experiments was further verified by anion exchange high performance liquid chromatography, followed by protein and endotoxin analyses, using the BCA protein assay kit (Pierce) and an in vitro limulus amebocyte lysate assay (Cambrex Bio Science, Inc., Walkersville, MD), respectively. No protein or endotoxin contamination was detected in this HA preparation. PKC⑀ substrate peptide-2 was obtained from Millipore (Billerica, MA).
PKC⑀ siRNA-Nanog siRNA Preparations and Transfection-The siRNA sequence targeting human PKC⑀ and Nanog (from mRNA sequence; GenBank TM accession number NM_005400 and NM_024865, respectively) corresponds to the coding region relative to the first nucleotide of the start codon. Target sequences were selected, using the software developed by Ambion Inc. As recommended by Ambion, PKC⑀or Nanogspecific targeted regions were selected beginning 50 -100 nucleotides downstream from the start codon. Sequences close to 50% G/C content were chosen. Specifically, PKC⑀ siRNA, PKC⑀-specific target sequence (AAGATGAAGGAGGCGCT-CAGTT), and scrambled sequences were used. In the case of Nanog, Nanog-specific target sequences (target 1, AATCT-TCACCTATGCCTGTGA; target 2, AATGAAATCTAAGA-GGTGGA; target 3, AAACCATGGATTTATTCCTAA) and scrambled sequences were used. MCF-7 cells were then transfected with siRNA, using siPORT Lipid as transfection reagent (Silencer TM siRNA transfection kit; Ambion) according to the protocol provided by Ambion. Cells were incubated with 50 pmol of PKC⑀ siRNA or Nanog siRNA or 50 pmol of siRNA containing scrambled sequences or no siRNA for at least 24 h before biochemical experiments.
Anti-miR-21 Inhibitor Preparation and Transfection-Anti-miR TM targeting miR-21 (anti-miR-21 inhibitor) (catalogue number 17000; Ambion) and its corresponding negative control (catalogue number 17010; Ambion) were transfected into MCF7 cells, using Lipofectamine 2000 reagent (Invitrogen) for 24 h. Cells were then treated with HA or without HA in various experiments, as described below. The final concentrations of anti-miR-21 and miRNA-negative control used in various experiments were 30 nmol/liter.
In addition, immunoprecipitation was conducted after homogenization of the cell lysate using rat anti-CD44 antibody, followed by goat anti-rat IgG-beads. Subsequently, the immunoprecipitated materials were solubilized in SDS sample buffer, electrophoresed, and blotted onto nitrocellulose. After blocking nonspecific sites with 3% bovine serum albumin, the nitrocellulose filters were incubated with rabbit anti-PKC⑀ antibody (2 g/ml) for 1 h at room temperature. In some cases, the cell lysates were immunoprecipitated with goat anti-Nanog antibody, followed by rabbit anti-goat IgG-beads. Subsequently, the immunoprecipitated materials were processed for immunoblotting using rabbit anti-phosphoserine antibody (2 g/ml).
Northern Blot Analysis-Total RNA was isolated from MCF-7 cells (untreated or pretreated with anti-CD44 antibody or transfected with PKC⑀ siRNA or Nanog siRNA or siRNA with scrambled sequences or anti-miR-21 inhibitor or miRNAnegative control) in the absence or presence of HA for various time intervals (e.g. 0 min, 15 min, 30 min, or 2 h) at 37°C using TriPure Isolation Reagent (Roche Applied Science). The probes were generated using the mirVana miRNA probe construction kit (Ambion), following the manufacturer's instructions. RNA concentrations were verified by measuring absorbance (A 260 ) on the NanoDrop Spectrophotometer ND-1000 (NanoDrop). Total RNA samples (10 g each) and enriched small RNAs (1 g) were electrophoresed on 12% acrylamide, 8 M urea gels, stained with ethidium bromide, and transferred using a capillary blotting method overnight onto Hybond-N ϩ membrane (Amersham Biosciences). 5 S rRNA was used as a loading control. RNA was immobilized by using a UV transilluminator for 10 min. Prehybridization and hybridization were performed at 40°C using the ULTRAHyb buffer from the NorthernMax kit (Ambion) for 30 min. Small RNAs were detected using [␣-32 P]UTP (800 Ci/mmol, 10 mCi/ml), which was used in the transcription reactions to synthesize labeled antisense RNA probe. Radioactively labeled probe was added to the ULTRA-Hyb buffer at 40°C for 16 h. Membranes were washed twice for 5 min at room temperature with Low Stringency Washing Solution (equivalent to 2ϫ SSC, 0.1% SDS) and one more time at a hybridization temperature of 40°C from the NorthernMax kit (Ambion). Sealed blots were exposed to film overnight and visualized using autoradiography.
RNase Protection Assay Analysis of Mature miRNAs-Expression of miRNAs was also qualitatively analyzed by an RNase protection assay. For the RNase protection assay, enriched small RNA isolated from MCF-7 cells (untreated or pretreated with anti-CD44 antibody or transfected with PKC⑀ siRNA, Nanog siRNA, siRNA with scrambled sequences, anti-miR-21 inhibitor, or miRNA-negative control in the presence or absence of HA for various time intervals (e.g. 0 min, 5 min, 10 min, 15 min, 30 min, or 2h) at 37°C) was enriched and purified using the mirVana miRNA isolation kit (Ambion). RNA concentrations were verified by measuring absorbance (A 260 ) on the NanoDrop Spectrophotometer ND-1000 (NanoDrop). The mirVana miRNA probe construction kit (Ambion) was used to synthesize the 32 P-labeled miR-21 antisense probe and miR-191 (60) probe loading control. Probes used were gel-purified using a 12% acrylamide, 8 M urea gel prior to hybridization experiments. Probe hybridization and RNase protection were then carried out using the mirVana miRNA detection kit (Ambion) according to the manufacturer's instructions. After hybridization and RNase treatment, the doublestranded product was resolved in a 12% polyacrylamide, 8 M urea denaturing gel and visualized using autoradiography. All samples were treated under similar conditions, and an additional radioactive-labeled probe, miR-191, was used as a loading control.
PKC⑀-mediated Protein Phosphorylation in Cell-free System-PKC⑀ was prepared by anti-PKC⑀-conjugated immunoaffinity column chromatography. The PKC⑀ kinase reaction was then performed in 50 l of the kinase buffer, containing 25 mM Tris-HCl (pH 7.5), 5 mM ␤-glycerolphosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 , 0.1% CHAPS, 0.1 M calyculin A, 200 M [ 32 P]ATP, 100 ng of PKC⑀ isolated from MCF-7 cells (pretreated with anti-CD44 antibody or transfected with PKC⑀ siRNA or siRNA with scrambled sequences or without any treatment, followed by HA (50 g/ml) addition (or no HA addition) for various time intervals (e.g. 0, 5, 15, or 30 min) at 37°C), and 1 g of PKC⑀ substrate peptide-2 or 1 g of Nanog (obtained from anti-Nanog affinity column). After incubation for 30 min, at 30°C, the reactions were terminated by adding 20% cold trichloroacetic acid, and 2 mg/ml bovine serum albumin was then added as a carrier. Trichloroacetic acid-precipitated proteins were spotted on 3M filter papers followed by an extensive wash with 10% trichloroacetic acid. The radioactivity associated with trichloroacetic acid-precipitated materials was analyzed by liquid scintillation counting.
Immunofluorescence Staining-MCF-7 cells (transfected with PKC⑀ siRNA or siRNA with scrambled sequences) were incubated with HA (50 g/ml) at 37°C for various time intervals (e.g. 0, 10, 30, or 60 min) or with no HA). These cells were then fixed with 2% paraformaldehyde. Subsequently, these cells were rendered permeable by ethanol treatment, followed by incubating with fluorescein (FITC)-conjugated anti-Nanog antibody, followed by a monomeric cyanine nucleic acid staining, Topro-3 (a marker for nucleus). To detect nonspecific antibody binding, Topro-3-labeled cells were incubated with FITC-conjugated normal IgG, respectively. No labeling was observed in control samples. These fluorescence-labeled samples were then examined with a confocal laser-scanning microscope.
Tumor Cell Growth Assays-MCF-7 cells were either untreated or pretreated with anti-CD44 antibody or transfected with PKC⑀ siRNA (or Nanog siRNA, siRNA with scrambled sequences, anti-miR-21, or miRNA-negative control) in the presence or absence of HA, as above. These cells were then plated in 96-well culture plates in 0.2 ml of Dulbecco's modified Eagle's medium/F-12 medium supplement (Invitrogen) containing no serum for 24 h at 37°C in 5% CO 2 , 95% air. In each experiment, a total of five plates (6 wells/treatment (e.g. HA treatment/plate)) were used. Experiments were repeated 5-6 times. The in vitro growth of these cells was determined by measuring increases in cell number using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (CellTiter 96 non-radioactive cell proliferation assay) according to the procedures provided by Promega. Subsequently, viable cellmediated reaction products were recorded by a Molecular Devices (Spectra Max 250) enzyme-linked immunosorbent assay reader at a wavelength of 450 nm.
In some experiments, MCF-7 cells were pretreated with anti-CD44 antibody or transfected with PKC⑀ siRNA (or Nanog siRNA, siRNA with scrambled sequences, anti-miR-21 inhibitor, or miRNA-negative control or without any treatment), as above. These cells (5 ϫ 10 3 cells/well) were then incubated with various concentrations of doxorubicin (4 ϫ 10 Ϫ9 M to 1.75 ϫ 10 Ϫ5 M) or paclitaxel (3.2 ϫ 10 Ϫ9 to M Ϫ1 ϫ 10 Ϫ5 M) with no HA or with HA (50 g/ml). After a 24-h incubation at 37°C, 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide-based growth assays were analyzed as described above. The percentage of absorbance relative to untreated controls (i.e. cells treated with neither HA nor chemotherapeutic drugs) was plotted as a linear function of drug concentration. The 50% inhibitory concentration (IC 50 ) was identified as a concentration of drug required to achieve a 50% growth inhibition relative to untreated controls.

HA-CD44-mediated PKC⑀ Activation and Nanog Phosphorylation/Signaling in MCF-7 Breast Tumor Cells
PKC belongs to a family of isoforms that are involved in a variety of biological activities (35). Previous studies showed that CD44 and PKC are structurally and functionally coupled in a number of cell types (61)(62)(63)(64). As part of our continued effort to investigate CD44-linked PKC activation that correlates with metastatic behaviors, a specific PKC isoform, namely PKC⑀, was identified. In these studies, we performed anti-CD44-mediated immunoprecipitation, followed by anti-PKC⑀ immunoblot (Fig. 1A, a, lane 1) or anti-CD44 immunoblot (Fig. 1A, b, lane 1), using untreated MCF-7 cells. Our results indicate that a very low level of PKC⑀ (Fig. 1A, a, lane 1) is present in the anti-CD44-immunoprecipitated materials (Fig. 1A, b, lane 1). Subsequently, we determined that HA treatment induces the recruitment of a significant amount of PKC⑀ (Fig. 1A, a, lane 2) into the CD44-PKC⑀ complex (Fig. 1A, b, lane 2). Pretreatment of MCF-7 cells with anti-CD44 antibody followed by HA treatment results in a significant reduction of PKC⑀ ( Fig. 1A, a, lane 3) in the anti-CD44-immunoprecipitated materials (Fig. 1A, b, lane 3). These findings establish the fact that PKC⑀ is accumulated in a complex with CD44 (in whole cells) following HA treatment of the MCF-7 breast tumor cells.
In addition, we have measured the PKC⑀ kinase activity in the CD44 complex isolated from MCF-7 cells (Fig. 1B). The kinase activity was determined by the ability of PKC⑀ to phosphorylate a purified PKC-specific substrate peptide-2 (Fig. 1B). Our results indicate that PKC⑀ present in the CD44-associated complex following HA treatment is capable of phosphorylating the PKC substrate peptide-2 (Fig. 1B, bar 2) to a ϳ3-fold higher level compared with the controls (e.g. CD44-associated PKC⑀ activities detected from MCF-7 cells pretreated with anti-CD44 antibody plus HA or treated with no HA (Fig. 1B, bar 1 and bar  3)). These results demonstrate that activation of PKC⑀ is both HA-dependent and CD44-specific in MCF-7 cells.
PKC phosphorylation of cellular proteins plays an important role in regulating tumor cell behaviors (65). Our recent findings indicate that the stem cell marker, Nanog, is closely associated with HA-CD44-mediated tumor cell growth and chemoresistance in MCF-7 cells (5). In searching for a possible linkage between HA-CD44-mediated PKC⑀ signaling and breast tumor cell-specific function, we have demonstrated that CD44-linked PKC⑀ (isolated from MCF-7 cells treated with HA) is capable of phosphorylating Nanog in vitro ( Fig. 2A, bar 2). A ϳ3-fold lower level of Nanog phosphorylation is detected using PKC⑀ isolated from cells that are not treated with HA ( Fig. 2A, bar 1 versus bar 2) or pretreated with anti-CD44 antibody plus HA ( Fig. 2A, bar 3 versus bar 2). Furthermore, we have noted that the ability of PKC⑀ to phosphorylate Nanog is significantly reduced using PKC⑀ isolated from cells treated with PKC⑀ siRNA (but not scrambled sequence siRNA) in the presence or absence of HA ( Fig. 2A, bars 4 -7). These findings clearly indicate that Nanog serves as a specific cellular substrate for HA-activated and CD44-linked PKC⑀ (in cell-free systems).
Further in vivo analysis indicates that significant Nanog phosphorylation occurs in MCF-7 cells treated with HA for 15 min (Fig. 2B, a and b, lane 2). In contrast, Nanog phosphorylation is relatively low in MCF-7 cells without any HA treatment (Fig. 2B, a and b, lane 1) or MCF-7 cells pretreated with anti-CD44 antibody followed by HA treatment (Fig. 2B, a and b, lane  3). Thus, Nanog phosphorylation is both HA-dependent and CD44-specific. Moreover, we have found that a low level of phosphorylated Nanog is present in MCF-7 cells transfected with scrambled sequence siRNA in the absence of HA treatment (Fig. 2B, a and b, lane 4). The unphosphorylated Nanog appears to be localized at the cytosol of MCF-7 cells (Fig. 3, A-C). After HA treatment for 15 min in MCF-7 cells (transfected with scrambled sequence siRNA), Nanog becomes highly phosphorylated (Fig. 2B, a and b, lane 5) and is also translocated from the cytosol to the nucleus (Fig. 3, D-F). Down-regulation of PKC⑀ (by transfecting tumor cells with PKC⑀ siRNA) significantly inhibits HA-CD44-mediated Nanog phosphorylation (Fig. 2B, a and b, lanes 6 and 7) and nuclear translocation (Fig. 3, G-I). These findings indicate that HA-CD44-mediated PKC⑀ activation is required for Nanog phosphorylation and nuclear translocation in MCF-7 cells.
Previous studies showed that activated Nanog functions as a transcription factor that translocates from the cytosol to the nucleus, binds to specific promoter elements of target genes, and regulates gene expression of Rex1 and Sox2 (66 -68). Here, we report that nuclear translocated Nanog (Fig.  4a, lane 2) forms a complex with p68 (an RNA helicase) (Fig.  4b, lane 2) and DROSHA (the nuclear RNase III enzyme) (Fig. 4c, lane 2) in MCF-7 cells treated with HA. It was also determined that a reduced amount of Nanog-p68-DROSHA complex is detected in tumor cells without HA treatment (Fig. 4, a-c, lane 1) or pretreated with anti-CD44 antibody followed by HA addition (Fig. 4, a-c, lane 3). These results indicate that Nanog-p68-DROSHA complex formation requires HA-CD44 interaction.
Further analyses show that the amount of Nanog association with p68 and DROSHA in MCF-7 cells (treated with scrambled   sequence siRNA) in the presence of HA is higher than in these same complexes from cells without HA treatment (Fig. 4, a-c,  lanes 4 and 5). However, HA-induced stimulation of Nanog complex formation with p68 and DROSHA is strongly inhibited in MCF-7 cells transfected with PKC⑀ siRNA (Fig. 4, a-c, lane 6) or Nanog siRNA, (Fig. 4, a-c, lane 7). These observations support the notion that HA-CD44-mediated PKC⑀ activation and Nanog signaling are closely interacting with p68 (a RNA helicase) and DROSHA (the nuclear RNase III enzyme) in MCF-7 breast tumor cells.

HA-CD44-activated PKC⑀ Promotes Nanog-mediated miRNA-21 Production in MCF-7 Cells
The expression of mature miR-21 is detected in various breast cancer-derived cell lines (57) in addition to different tumor cell types. A recent study indicates that both p68 (a RNA helicase) and DROSHA (the nuclear RNase III enzyme) are involved in miR-21 biogenesis (processing the precursor pri-miR-21) in breast tumor cells (55). The question of whether HA-CD44-mediated PKC⑀-Nanog signaling interaction with DROSHA and p68 contributes to an increase in mature miR-21 levels required for oncogenesis and breast tumor progression is the focus of this study.
To test this hypothesis, we first performed a time course experiment to examine the accumulation of pre-miR-21 and mature miR-21 following HA treatment. After extraction of total RNA and isolation of small RNAs, we determined that there was a significant induction of pre-miR-21 (ϳ72 nucleotides) and mature miR-21 (ϳ21 nucleotides) after 2 h of HA treatment by performing Northern Blots with cells that were treated versus untreated with HA (Fig. 5A, lanes 1 and 2). In addition, the mature miR-21 was detected by the miRVana miRNA detection kit, which is based on an RNase protection assay. Again, after 2 h of HA treatment, there was a significant increase in the expression of the mature miR-21 (Fig. 5C, lane 2) compared with that in cells without HA treatment (Fig. 5C, lane  1). The induction of both the pre-miR-21 and the mature-miR-21 levels was specifically a result of the interaction between HA and the CD44. This was concluded because anti-CD44 antibody plus HA treatment significantly reduced the effect of HA on the expression of mature miR-21, as shown by both Northern blot and RNase protection assay (Fig. 5, A, lane  3, and C, lane 3). The increase in miR-21 expression was not due to increased levels of total RNA extracted from each sample, since there were very similar levels of 5 S ribosomal RNA in both HA-activated and control samples (e.g. untreated or pretreated cells followed by HA addition) (Fig. 5A). Our results of Northern blot analyses (Fig. 5B) and RNase protection assays (Fig. 5D) also show that the level of both pre-miR-21 and mature miR-21 production are clearly elevated in MCF-7 cells treated with scrambled sequence siRNA plus HA (Fig. 5, B (lane  2) and D (lane 2)), as compared with those cells without HA addition. In contrast, MCF-7 cells treated with either PKC⑀ siRNA (Fig. 5, B (lane 3) and D (lane 3)) or Nanog siRNA (Fig. 5,  B (lane 4) and D (lane 4)) contain significantly less HA-induced pre-miR-21 and mature miR-21.
Suppression of breast tumor cells by down-regulating miR-21 was previously accomplished by using an anti-miR-21 inhibitor (69,70). In this study, we have observed that the expression of both pre-miR-21 and mature miR-21 production can be induced in cells treated with an miRNA-negative control reagent in the presence of HA (Fig. 5B, lane 6 versus lane 5; Fig.  5D, lane 6 versus lane 5). Finally, we have confirmed that treatment of MCF-7 cells with a miR-21 inhibitor effectively downregulates either pre-miR-21 or mature miR-21 production even in the presence of HA (Fig. 5, B (lane 7) and D (lane 7)). These observations strongly suggest that HA-CD44-mediated PKC⑀ activation and Nanog signaling are tightly linked to the regulation of miR-21 production in MCF-7 cells.

The Effect of HA-CD44-mediated miR-21 (Induced by PKC⑀ and Nanog Signaling) on PDCD4 Expression, Anti-apoptosis, and Chemoresistance in Breast Tumor Cells
The array profile of miRNA expression in normal breast tissues versus breast carcinoma tissue reveals that miR-21 is abundantly produced in tumors compared with normal tissues (57). The functional significance of miR-21 has been emphasized in several recent studies, and the discovery of its specific targets has also been widely investigated. Previous studies indicate that miR-21 may function as an oncogene and play a role in antiapoptosis and chemotherapy resistance, in part through downregulation of several tumor suppressor genes/proteins, including PDCD4 (58,82). However, the identification of miR-21-specific downstream target(s) and oncogenic event(s) that contributed to HA-CD44-dependent breast tumor cell functions has not been established.
Effect of miR-21 on the Expression of PDCD4 and PDCD4-eIF4A Complex Formation-PDCD4 has been identified as one of the tumor suppressor genes regulated by miR-21 (58,82). It inhibits translation by forming a complex with the translation initiation factor eIF4A (an RNA helicase) and interfering with the ability of eIF4A to unwind the secondary structure at the 5Ј-untranslated region of mRNAs (71)(72)(73). In this study, we have found that a basal level of PDCD4 and eIF4A expression is present in cells without HA treatment (Fig. 6A, a and b, lane 1) or in cells pretreated with anti-CD44 antibody followed by HA treatment (Fig. 6A, a and b, lane 3). However, HA treatment promotes down-regulation of the tumor suppressor protein (PDCD4) expression (but not eIF4A expression) in MCF-7 cells (Fig. 6A, a and b, lane 2). Thus, the reduction of PDCD4 expression (but not eIF4A) appears to be HA-dependent and CD44specific in breast tumor cells. Further analyses indicate that a basal level of the PDCD4-eIF4A complex is present in MCF-7 cells without HA treatment (Fig. 6B, a and b, lane 1) or in cells pretreated with anti-CD44 antibody followed by HA treatment (Fig. 6B, a and b, lane 3). However, the amount of the PDCD4-eIF4A complex is significantly decreased in MCF-7 cells treated with HA (Fig. 6B, a and b, lane 2 versus lanes 1 and 3). These findings indicate that HA-CD44 interaction is required for the regulation of PDCD4 expression and PDCD4-eIF4A interaction in breast tumor cells. We have also confirmed that downregulation of miR-21 by treating MCF-7 cells with an anti-miR-21 inhibitor (but not a negative control miRNA) promotes up-regulation of PDCD4 expression (but not eIF4A expression) (Fig. 6A, a and b, lane 10 versus lane 9) and PDCD4-eIF4A association in the presence of HA (Fig. 6B, a and b, lane 10  versus lane 9). These results support the contention that miR-21 (mediated by HA-CD44 binding) is acting as an oncogene by down-regulating the expression of the tumor suppressor, PDCD4, and its interaction with eIF4A. Furthermore, treatment of MCF-7 cells with either PKC⑀ siRNA or Nanog siRNA (in the presence of HA) induces an elevated level of PDCD4 expression (but not eIF4A expression) and PDCD4-eIF4A complex formation (Fig. 6A, a and b, lanes 6 and 7; Fig.  6B, a and b, lanes 6 and 7). Although a basal level of PDCD4 expression and PDCD4-eIF4A complex was detected in cells treated with scrambled sequence siRNA without HA addition (Fig. 6A, a and b, lane 4; Fig. 6B, a and b, lane 4), significant reduction of either PDCD4 expression or PDCD4-eIF4A complex was observed in these cells treated with HA (Fig. 6A, a and  b, lane 5; Fig. 6B, a and b, lane 5). These findings indicate that the signaling network consisting of PKC⑀-Nanog and miR-21 is functionally coupled with the inhibition of the tumor suppressor protein (PDCD4) expression and reduces its association with eIF4A. These effects facilitate the initiation of translation and protein production in HA-CD44-activated breast tumor cells.
Effect of miR-21 on IAPs and MDR1 (P-gp) Expression and Chemotherapeutic Response-To determine how these changes in the tumor suppressor protein (PDCD4) expression and function by miR-21 (via HA-CD44 interaction and PKC⑀-Nanog signaling) may affect breast tumor cell-specific behaviors (e.g. anti-apoptosis and chemoresistance), we decided to analyze the expression of the IAPs and the chemoresistance protein-1 (MDR1/P-gp). The IAPs constitute a family of at least nine proteins, including XIAP and survivin, that block apoptosis by direct binding to caspases (74). Overexpression of IAPs (e.g. XIAP and survivin) is thought to be linked to chemoresistance by suppressing apoptosis (75)(76)(77). MDR1 (P-gp) belongs to the ATP-binding cassette transporters, a superfamily of channel proteins (78 -81). The functions of MDR1 (P-gp) include the efflux and retention of ions, nutrients, lipids, amino acids, peptides, proteins, and drugs (78 -81). HA-CD44 interaction has been shown to induce the expression of survivin and MDR1/ P-gp in tumor cells (5,6,42,45). The question of whether miR-21 (induced by HA-CD44 interaction and PKC⑀-Nanog signaling) regulates the expression of IAPs (e.g. XIAP and survivin) and MDR1 in breast tumor cells has not been investigated previously.
To answer this question, immunoblot analyses using a panel of antibodies (e.g. anti-XIAP antibody, anti-survivin antibody, and anti-MDR1/P-gp antibody) were employed to detect the production of three proteins: survivin, XIAP, and MDR1 (P-gp) in MCF-7 cells. Our data indicate that the expression of both IAPs (e.g. XIAP and survivin) and MDR1/P-gp are significantly increased in MCF-7 treated with HA (Fig. 7, a-c, lane 2 versus  lane 1). In contrast, these three proteins (e.g. XIAP, survivin, and MDR1/P-gp) are present in relatively low amounts in MCF-7 cells treated with no HA (Fig. 7, a-c, lane 1 versus lane  2) or in those cells pretreated with anti-CD44 antibody followed by HA addition (Fig. 7, a-c, lane 3 versus lane 2). These findings support the notion that the expression of IAPs and MDR1 is both HA-and CD44-dependent. Most importantly, down-regulation of miR-21 by treating cells with an anti-miR-21 inhibitor significantly attenuates the HA-CD44-activated expression of IAPs (e.g. XIAP and survivin) and MDR1/P-gp (Fig. 7, a-c,  lane 10). In contrast, MCF-7 cells treated with a miRNA-negative control are capable of inducing the expression of both IAPs (e.g. XIAP or survivin) and MDR1/P-gp in the presence of HA (Fig. 7, a-c, lanes 8 and 9). Furthermore, we have observed that the expression of IAPs (e.g. XIAP and survivin) and MDR1/P-gp are significantly inhibited when MCF-7 cells were pretreated with PKC⑀ siRNA or Nanog siRNA (Fig. 7, a-c, lanes 6 and 7) but not scrambled sequence siRNA followed by HA addition (Fig. 7, a-c, lanes 4 and 5), respectively. The fact that downregulation of both PKC⑀-Nanog signaling and miR-21 production inhibits the expression of IAPs (e.g. XIAP and survivin) and MDR1/P-gp indicates that the HA-CD44-activated PKC⑀-Nanog signaling and miR-21 function actively participate in the up-regulation of IAPs and MDR1/P-gp in breast tumor cells.
To further assess whether the chemotherapeutic drug responses of MCF-7 cells might be regulated by the HA-CD44 interaction with PKC⑀-Nanog signaling and miR-21 production, we performed tumor cell growth and apoptosis assays using two anti-breast cancer chemotherapeutic drugs (e.g. doxorubicin and paclitaxel (Taxol)) in the presence or absence of HA or anti-CD44 antibody plus HA. In the absence of HA, doxorubicin-treated MCF-7 cells displayed an increase of apoptotic tumor cells and a low level of tumor cell survival with IC 50 values of 60 nM (Tables 1 and 2). Paclitaxel-treated MCF-7 cells also exhibit a relatively high level of apoptosis and a low level of tumor cell survival, with IC 50 values of ϳ40 nM (Tables  1 and 2). However, the addition of HA enhances cell survival and reduces apoptosis in untreated controls (i.e. without chemotherapeutic drugs) and decreases the ability of both doxorubicin (IC 50 of 320 nM) and paclitaxel (IC 50 of 160 nM) to induce tumor apoptosis and cell death (Tables 1 and 2). These observations strongly suggest that HA causes a decrease in apoptosis and an increase in tumor cell survival, leading to the enhancement of chemoresistance to both doxorubicin and paclitaxel treatment (Tables 1 and 2). Furthermore, pretreatment of these tumor cells with anti-CD44 antibody followed by HA addition significantly increases tumor cell apoptosis and reduces the HA-mediated drug resistance (Table 1 and 2). This result indicates that HA-CD44 interaction promotes anti-apoptosis and cell survival in the presence of chemotherapeutic drugs, such as doxorubicin and paclitaxel, in breast tumor cells. Moreover, down-regulation of PKC⑀, Nanog, or miR-21 (by transfecting tumor cells with PKC⑀ siRNA or Nanog siRNA or anti-miR-21 inhibitor (but not scrambled sequence siRNA or an miRNA-negative control)) effectively attenuates HA-mediated tumor cell anti-apoptosis/survival and enhances multidrug sensitivity in MCF-7 cells (Tables 1 and 2). Together, these findings indicate that the HA-CD44-mediated PKC⑀-Nanog signaling pathways and miR-21 function provide new drug targets to sensitize tumor cells to undergo apoptosis/ death and to overcome chemotherapy resistance in breast cancer cells.

DISCUSSION
Chemotherapy resistance is one of the primary causes of morbidity in patients diagnosed with solid tumors, such as breast cancer (1)(2)(3). It is now certain that a number of oncogenic signaling pathways are closely involved with multidrugresistant phenotypes (4 -6). In particular, HA-CD44-activated cancer cells have been strongly implicated in the development of chemoresistance (5,6,(42)(43)(44)(45)(46)(47)(48). Specifically, HA is capable of stimulating MDR1 (P-gp) expression and drug resistance in breast tumor cells (5,6). CD44 also interacts with MDR1 (P-gp) to promote cell migration and invasion of breast tumor cells (83). Previously, we have reported that activation of HA-CD44mediated oncogenic signaling events (e.g. intracellular Ca 2ϩ mobilization, epidermal growth factor receptor-mediated ERK signaling, topoisomerase activation, and ankyrin-associated cytoskeleton function) leads to multidrug resistance in a variety of tumor cells (5, 46 -48).
Recently, we have found that the stem cell marker Nanog appears to be closely involved in HA-CD44-mediated chemoresistance in breast tumor cells (5). Nanog is an important transcription factor involved in the self-renewal and maintenance of pluripotency in the inner cell mass of embryos and embryonic stem cells (84). Nanog signaling is regulated by interactions among various pluripotent stem cell regulators (e.g. Rex1, Sox2, and Oct3/4), which together control the expression of a set of target genes required for embryonic stem cell pluripotency (66,85). These findings confirm the essential role of Nanog in regulating a variety of cellular functions. Both breast carcinomas and breast tumor cells have been shown to express several common embryonic stem cell markers, including Nanog (86). The Nanog family of proteins functions as growth-promoting regulators by up-regulating transcriptional activities and gene expression in breast tumor cells (87). It is not well understood how Nanog signaling is regulated by HA-CD44 interaction, which causes chemoresistance in breast tumor cells.
Both HA and CD44 are important activators of oncogenesis and chemoresistance in breast tumor cells (4 -6, 42-48). It has been well documented that the external portion of CD44 binds to HA, whereas the intracellular domain of CD44 interacts with receptor kinases (e.g. ErbB2, epidermal growth factor receptor, and TGF␤ receptors) and non-receptor kinases (e.g. c-Src and ROK) (28 -34). HA-induced CD44 interaction with these kinases plays a pivotal role in promoting breast tumor cell functions (28 -34). Previous studies indicated that the cytoplasmic domain of CD44 can be phosphorylated by PKC (61,62). Most importantly, CD44 phosphorylation by PKC promotes ankyrin binding to CD44 and stimulates a number of biological activities (61,62). These findings indicate that CD44-cytoskeleton interaction and PKC signaling are closely coupled. Based on structural features and activation requirements, the PKC family of proteins has been divided into at least three groups of isoforms: the classical isoforms (␣, ␤I, ␤II, and ␥); the novel isoforms (␦, ⑀, , and ); and the atypical isoforms ( and /) (36). In particular, PKC⑀ appears to play a causative role in establishing breast tumor cell-specific phenotypes (37)(38)(39). PKC⑀ also acts as an anti-apoptotic protein and protects breast cancer MCF-7 cells from tumor necrosis factor-␣-mediated apoptosis, in part through inhibition of Bax activation and translocation to the mitochondria (88).
In addition, PKC⑀ is known to function as a transforming oncogene by interacting with several signaling components, including RhoA/C, Stat-3, and Akt (35). In this study, we have found that HA binding to breast tumor cells (MCF-7 cells) not only recruits PKC⑀ into CD44 complexes but also activates its enzymatic activities (Fig. 1). Most importantly, we have determined that Nanog serves as one of cellular substrates for HA-CD44-activated PKC⑀ (Fig. 2). Nanog has been shown to regulate the expression of pri-miRNA by associating to the 5Ј-regulatory region of the miRNAs that are involved in the most critical molecular processes during development (56). Our data indicate that PKC⑀-activated Nanog is translocated from the cytosol to the nucleus (Fig. 3) and forms a complex with DROSHA/p68 (Fig. 4), resulting in miR-21 production in HA-CD44-activated breast tumor cells (Fig. 5). The fact that down-regulation of either PKC⑀ or Nanog (by transfecting cells with either PKC⑀ siRNA or Nanog siRNA) not only abolishes Nanog association with DROSHA and p68 (Fig. 4) but also inhibits miR-21 production (Fig. 5) in HA-treated MCF-7 cells clearly indicates the importance of PKC⑀ signaling and Nanog function in regulating HA-CD44-regulated miR-21 production.
Our results are consistent with previous reports showing association of DROSHA/p68 microprocessor complex with certain signaling regulators during miRNA production. For example, it has been shown that there is a molecular interaction between p53 and the DROSHA complex in HCT116 cells. This interaction with the processing complex occurs via the RNA helicases (p68/p72) (59). It is also noted that activation of the TGF␤-mediated SMAD-2 signaling stimulates the expression of a subset of miRNAs, including miR-21 (55). Specifically, this TGF␤-mediated signaling event occurs at a post-translational step promoting the processing of primary transcripts of miR-21 (pri-miR-21) into precursor miR-21 (pre-miR-21) by DROSHA complex (55). TGF␤-specific SMAD-2 signaling transducer also becomes recruited to pri-miR-21 in a complex with p68 (the RNA helicase), components of the DROSHA microprocessor complex in human vascular smooth muscle cells (55). Apparently, the DROSHA/p68 microprocessor complex is closely associated with the production of miRNAs, such as miR-21, by a variety of signaling pathways. Since miR-21 has been shown to participate in breast cancer progression (57), the elucidation (in this study) of HA-CD44 signaling pathway-specific mechanisms involved with miR-21 biogenesis is significant for the formulation of future intervention strategies for treating breast cancer.
The ability of certain chemotherapeutic agents (e.g. doxorubicin and paclitaxel) to induce tumor cell death is often counteracted by the presence of anti-apoptotic proteins, leading to chemoresistance (75)(76)(77). Several lines of evidence point toward the IAP family (e.g. survivin and XIAP) playing a role in oncogenesis via their effective suppression of apoptosis (74). The mode of action of IAPs in suppressing apoptosis appears to be through direct inhibition of caspases and procaspases (primarily caspase 3 and 7) (74). IAPs also support chemoresistance by preventing tumor cell death induced by anticancer agents (75)(76)(77). Although certain anti-apoptotic proteins (e.g. Bcl-xL) have been shown to participate in anti-apoptosis and chemoresistance in HA-CD44-activated breast tumor cells (6), the involvement of IAPs in HA-CD44-mediated tumor cell survival and chemoresistance has not been fully elucidated. Multidrug resistance can also be mediated by overexpression of MDR1 (P-gp) (41), which functions as a drug efflux pump actively reducing intracellular drug concentrations in resistant tumor cells (5,41). Previous studies showed that HA-CD44 induces the expression of MDR1 (P-gp) and chemoresistance in breast tumor cells (5,6,(42)(43)(44)(45)(46)(47)(48). In the present study, we have made several important and novel observations. Specifically, our results indicate that HA-CD44-mediated PKC⑀-Nanog signaling mediates miR-21 production, which in turn, exerts its influence on tumor cell-specific functions, including anti-apoptosis and chemoresistance ( Fig. 7 and Tables 1 and 2).
Furthermore, miR-21 down-regulation has been shown to be effective in blocking oncogenesis by up-regulating its known targets, including tumor suppressor proteins, such as PDCD4 (58). In fact, loss of PDCD4 expression occurs during breast tumor progression (89,90). Up-regulation of PDCD4 is closely linked to apoptosis and translation inhibition (via its binding and inhibiting the helicase activity of eIF4A, a component of translation initiation complex) (71)(72)(73). In this study, we have