Hsp27 participates in the maintenance of breast cancer stem cells through regulation of epithelial-mesenchymal transition and nuclear factor-κB

AS-B145 and AS-B244 breast cancer cells, which derived from BC0145 or BC0244 xenograft human breast cancer cells [14], were cultured in MEMα supplemented with 10% fetal bovine serum, bovine insulin (0.1 mg/mL), sodium pyruvate (1 mM), and Glutamax (2 mM). 4T1 mouse breast cancer cells and an MDA-MB-231 human breast cancer cell line were obtained from ATCC (Manassas, VA, USA) and cultivated as ATCC's recommendation. The cells were maintained in a 5% CO₂ air humidified atmosphere at 37°C. Quercetin and JSH-23 were purchased from Calbiochem (San Diego, CA, USA) and dissolved in dimethyl sulfoxide (DMSO). pDsRed-Express2-C1 vector was purchased from Clontech (Mountain View, CA, USA). To construct DsRed tagged Hsp27, the Hsp27 gene was cloned from AS-B145 cDNA by the following primers: 5'-CACTTGAGATCTACCGAGCGCCGCGTCCCCTTC-3' (forward); 5'-CAAGTGGAATTCTTACTTGGCGGCAGTCTCATC-3' (reverse) and inserted into pDsRed-Express2-C1 vector by BglII and EcoRI restriction sites (pDsRed-Hsp27).

MAPK antibody array was purchased from R&D Systems (Minneapolis, MN, USA) and conducted following the manufacturer's protocol. Briefly, the membrane was blocked in blocking buffer and incubated with 150 μg of total cellular protein and detection antibody simultaneously at 4°C overnight. After washing, the membrane was further incubated with streptavidin-HRP at room temperature for 30 minutes and a signal was developed with ECL substrate. For Western blot, cells were lysed with NP-40 lysis buffer and 25 μg of total protein were separated by SDS-PAGE and transferred to polyvinylidene fluoride membrane. Protein detection was conducted by SignalBoost™ Immunodetection Enhancer kit (Calbiochem) according to the manufacturer's recommendation. Hsp27 antibody was purchased from Stressgen (Ann Arbor, MI, USA). IκBα and phosphor-IκBα antibodies were purchased from Cell Signaling Technologies (Boston, MA, USA). NF-κB p65 antibody was purchased from Millipore (Billerca, MA, USA). Snail, twist, vimentin, GAPDH and histone H1 antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). β-actin antibody was purchased from Novus Biologicals (Littleton, CO, USA).

The specific siRNA oligos of Hsp27 (sc-29350) or IκBα (sc-29360), or negative control siRNA oligos was purchased from Santa Cruz Biotechnologies, Inc. The siRNA oligos of Hsp27 or IκBα consisted of pools of three target-specific siRNAs designed to knockdown gene expression and the target sequences were listed below:

sc-29350A: Sense: GAGUGGUCGCAGUGGUUAGtt; Antisense: CUAACCACUGCGACCACUCttsc-29350B: Sense: GACGAGCUGACGGUCAAGAtt; Antisense: UCUUGACCGUCAGCUCGUCttsc-29350C: Sense: CCACGCAGUCCAACGAGAUtt; Antisense: AUCUCGUUGGACUGCGUGGtt

sc-29360A: Sense: CCACACGUGUCUACACUUAtt; Antisense: UAAGUGUAGACACGUGUGGttsc-29360B: Sense: GACGAGAAAGAUCAUUGAAtt; Antisense: UUCAAUGAUCUUUCUCGUCttsc-29360C: Sense: CCUGUCAAGGUUUGUGUUAtt; Antisense: UAACACAAACCUUGACAGGttMetafecteneSI transfection reagent (Biontex, Martinsired, Germany) was used for siRNA transfection following the manufacturer's protocol. To overexpress Hsp27, cells were transfected with pDsRed-Hsp27 by MetafectenePro transfection reagent (Biontex) as a ratio (DNA (μg):reagent (μl)) of 1:3.

An ALDEFLUOR assay kit was purchased from StemCell Technologies, Inc. (Vancouver, BC, Canada) and used following the manufacturer's recommendations. Briefly, 1 × 10⁵ cells were suspended in 50 μl of assay buffer and added to BODIPY-aminoacetaldehyde substrate to a final concentration of 1 μM. For ALDH1 inhibitor control, diethylaminobenzaldehyde (DEAB) was added to the final concentration of 150 μM. Cells were then incubated at 37°C for 45 minutes and stained with 7-AAD on ice for a further 5 minutes. After being washed with PBS, green fluorescence positive cells in live cells (7AAD-) were analyzed by flow cytometry (Epics XL, Beckman Coulter, Brea, CA, USA) by comparing the fluorescence intensity of the DEAB treated sample; these cells will be represented as cells with high ALDH activity (ALDH+ cells).

Cells were harvested from monolayer culture or collected by fluorescence-activated cell sorting (FACS) and prepared at a density of 1 × 10⁴ cells/ml in DMEM/F12 medium contain 0.5% methylcellulose, 0.4% bovine serum albumin, 10 ng/ml EGF, 10 ng/ml bFGF, 5 μg/ml insulin, 1 μM hydrocortisone and 4 μg/ml heparin. A total of 2 ml of cell solution was seeded into wells of ultralow attachment six-well-plate (Corning, Lowell, MA, USA) and incubated for seven days. For secondary spheres, the cells were collected from accutase treated primary spheres, seeded at a density of 2,500 cells/ml and cultivated for a further seven days.

The tumorigenicity of AS-B145 sphere cells was examined by xenograftment assay in NOD/SCID mice. After transfection with ctrl-siRNA or si-Hsp27 for 48 h, an indicated number of AS-B145 cells was mixed with 10⁵ normal breast fibroblasts by 50 μl of MEMα:matrigel (1:1) (BD Biosciences, San Jose, CA, USA) and injected into mammary fat pads of female NOD/SCID mice (BioLASCO Taiwan Co., Ltd. Taipei, Taiwan). The tumor formation was monitored weekly. The CSC frequency was calculated by Extreme Limiting Dilution Assay (Walter-Eliza Hall Bioinformatics, Bundoora Victoria, Australia.) [15].

A cell migration assay was conducted by Oris Universal Cell Migration Assembly kit (Platypus Technologies, LCC, Madison, WI, USA) following the manufacturer's protocol. Briefly, 5 × 10⁴ cells/well/100 μl were loaded into stopper-loaded wells and incubated overnight to permit cell attachment. To start cell migration, the stoppers were removed; wells were gently washed with PBS, then added to complete cell culture medium and incubated for 16 to 18 h. Pictures of wells were captured with inverted microscopy after fixation and stain with 0.5% crystal violet/50% EtOH. Data were analyzed with ImageJ software (National Health Institute, Bethesda, MD, USA).

We have previously established two human breast cancer cells from xenografts of NOD/SCID mice and identified that cells with high intracellular aldehyde dehydrogenase activity (ALDH+) are cancer stem cells [14]. With ALDH activity, proteins extracted from ALDH- or ALDH+ cells of BC0145 or BC0244 xenografts were subjected to mitogen activated protein kinase (MAPK) antibody array and results indicated that the activation of Akt, ERK, p38 MAPK and RSK1 were increased in BSCSs (Additional file 1 Figures S1 and S2). The phosphorylation of Hsp27, which may result from p38 MAPK activity, was also increased in ALDH+ BCSCs from BC0145 or BC0244 xenograft cells (Figure 1A). We also used Western blot to check the level of total Hsp27 protein between ALDH- and ALDH+ AS-B244 cells, which derived from ALDH+ BC0244 xenograft cells [14]. As shown in Figure 1B, the total protein level of Hsp27 was higher in ALDH+ cells than in ALDH- cells. These results indicate that Hsp27and its phosphorylation are up-regulated in BCSCs.

We next investigated the role of Hsp27 in maintenance of BCSCs by siRNA mediated gene silence of Hsp27 expression. After transfection with Hsp27 specific siRNA, the population of ALDH+ cells in AS-B145 or AS-B244 cells was significantly decreased to (50.2 ± 12.2)% or (58.7 ± 3.5)%, respectively, when compared with cells transfected with negative control siRNA (Figure 2A). Knockdown of Hsp27 not obviously caused cell death (less than 22% of all time points) and slowed the cell growth rate of AS-B145 cells (Additional file 1 Figure S3A, C), but caused obvious cell death (35% and 65% at 72 h and 96 h, respectively) and decreased cell number at 72 h and 96 h in AS-B244 cells (Additional file 1 Figure 3B, D). Other than the ALDH+ population of cells, the number of mammospheres as well as the size of formed spheres in AS-B145 or AS-B244 cells were also decreased (Figure 2B). We further examined if Hsp27 was involved in the tumorigenicity of BCSCs. AS-B145 sphere cells were collected for seven days after mammosphere culture, transfected with negative control siRNA or Hsp27 specific siRNA for 48 h and injected into mammary fat pads of female NOD/SCID mice in a serial dilution of injected cell number. As shown in Figure 2C, 10⁵ negative control siRNA transfected AS-B145 sphere cells formed tumors in four out of five mice but 10⁵ Hsp27 knockdown cells only formed tumors in two out of five mice at Day 44 (Figure 2C, left panel). The CSC frequency of Hsp27 knockdown AS-B145 sphere cells was significantly decreased when compared with negative control siRNA groups (1:30,680 versus 1:146,211, P = 0.0206). In addition to RNA interference, we also used quercetin, a plant flavonoid compound which has been reported to suppress the protein level of Hsp27 [16], to treat AS-B145 and AS-B244 cells. Quercetin inhibited the expression of Hsp27 protein (Figure 3A) as well as the population of ALDH+ cells (Figure 3B) in both AS-B145 and AS-B244 cells in a dose-dependent manner (Figure 3C). In order to confirm if the inhibition effect of quercetin is mediated by down-regulation of Hsp27, we next overexpressed Hsp27 in AS-B145 cells and examined the ALDH+ population under quercetin treatment. As shown in Figure 3D and 3E, the inhibitory effect of quercetin could be reversed by overexpression of Hsp27 in AS-B145 cells. We next tested if quercetin also inhibits the self-renewal of BCSCs by mammosphere formation assay. The size and number of primary and secondary mammospheres in AS-B145 (Figure 4A) and AS-B244 (Figure 4B) was suppressed by quercetin in a dose-dependent manner. In addition to human BCSCs, we also tested if quercetin could inhibit self-renewal of Sca-1+ 4T1 mouse BCSCs [17]. As shown in Figure 4C, quercetin decreased primary and secondary mammosphere formation of Sca-1+ 4T1 cells in a dose-dependent manner.

EMT is an important character of cancer stem cells [18]. We next examined if Hsp27 mediates EMT features of BCSCs. With a wound healing based cell migration assay, the cell migration ability of ALDH+ AS-B244, AS-B145, MDA-MB-231 and Sca-1+ 4T1 cells was inhibited by quercetin treatment in a dose-dependent manner (Figure 5A). Furthermore, quercetin treatment dose-dependently inhibited the expression of N-cadherin and twist but increased E-cadherin expression in both AS-B145 and ALDH+ AS-B244 cells (Additional file 1 Figure 4). By siRNA mediated knockdown of Hsp27, the cell migration capacity of AS-B145, MDA-MB-231 or ALDH+ AS-B244 cells was also inhibited in comparison with negative control siRNA (Figure 5B). We also investigated if the Hsp27 pathway also regulates EMT-related molecular signatures. With Western blot analysis, knockdown of Hsp27 in AS-B145 or ALDH+ AS-B244 cells decreased the expression of snail and vimentin and increased the expression of E-cadherin (Figure 5C). These results indicate that Hsp27 may regulate self-renewal of BCSCs through manipulating the EMT process.

It has been reported that Hsp27 enhances the degradation of ubiquitinated proteins by 26S proteasome. Among these ubiquitinated proteins, phosphorylated IκBα could form a complex with Hsp27 and 26S proteasome and Hsp27 could enhance NF-κB activity by facilitating proteasome mediated IκBα degradation [19]. Recently, the NF-κB pathway has been demonstrated to participate in mammary tumorigenesis and cancer stem cell expansion in a transgenic mouse model [20]. We next examined if Hsp27 regulates NF-κB activity in BCSCs. By siRNA mediated knockdown of Hsp27, the expression of IκBα was increased in both AS-B145 and ALDH+ AS-B244 cells and its phosphorylation was decreased (Figure 6A). The nuclear translocation of NF-κB was also inhibited in both AS-B145 and ALDH+ AS-B244 cells when knockdown of Hsp27 occurred (Figure 6B). In the meantime, we also observed that Hsp27 could enter into the nucleus (Figure 6B). With a luciferase based reporter assay, the NF-κB activity was decreased in ALDH+ AS-B244 and AS-B145 cells when knockdown of Hsp27 occurred (Figure 6C). We next used NF-κB inhibitors to examine their effects on BCSCs. In the presence of JSH-23, a NF-κB inhibitor which inhibits the nuclear translocation of NF-κB, the ALDH+ population of AS-B145 and AS-B244 cells was suppressed in a dose-dependent manner (Figure 6D). We further examined if additional activation of NF-κB could diminish the inhibitory effect of ALDH+ cells by Hsp27 knockdown. The increased IκBα, which was caused by knockdown of Hsp27, was suppressed by knockdown of IκBα (Figure 7A) and the NF-κB activity could be restored in Hsp27 knockdown of AS-B145 or AS-B244 cells (Figure 7B). The inhibitory effect of ALDH+ cells by Hsp27 knockdown could be reversed by additional knockdown of IκBα in both AS-B145 and AS-B244 cells (Figure 7C, D). These results suggest that Hsp27 regulates the maintenance of BCSCs through NF-κB activity.