The MCF-7 breast cancer cell line was purchased from the American Type Culture Collection (Manassas, VA, USA) and cultured in Roswell Park Memorial Institute (RPMI)-1640 culture medium (Macgene, Beijing, China) supplemented with 10% (v/v) fetal bovine serum (FBS; Hyclone, Utah, USA) at 37°C in a humidified atmosphere containing 5% CO₂. Hoechst 33342, epidermal growth factor (EGF), human recombinant basic fibroblast growth factor-basic (bFGF), B27 supplement and Lipofectamine^® 2000 were purchased from Invitrogen (Carlsbad, CA, USA). Verapamil was obtained from Sigma (St. Louis, MO, USA) and dissolved in distilled deionized water (ddH₂O). Serum-free Opti-MEM^® I and DMEM/F12 medium were purchased from Gibco (Grand Island, NY, USA).

The SP analysis was performed based on the method reported by Goodell et al(14) with slight modifications. MCF-7 cells were digested with 0.25% trypsin (Sigma), washed twice with PBS, resuspended in pre-warmed RPMI-1640 culture (supplemented with 2% FBS) at a density of 1×10⁶ cells/ml. Then, the cells were pre-incubated with or without 50 μM verapamil for 30 min at 37°C before adding the Hoechst 33342 dye at a final concentration of 5 μg/ml. The mixture was incubated in the dark at 37°C for 90 min with interval mixing. Following incubation, the cells were washed twice with ice-cold PBS, and the cells were then filtered through a 40 μm nylon mesh to obtain single cell suspension and kept at 4°C in the dark. Cell analysis and purification were performed using FACS (FACSAria II; Becton-Dickinson, CA, USA). Hoechst 33342 was excited with UV light at 355 nm and fluorescence emission was measured with 450/BP50 (Hoechst blue) and a 660/BP50 (Hoechst red) optical filters. At the end of sorting, both collected SP and non-SP cells were reanalyzed to evaluate sorting purity and to conduct further experiments. To minimize the non-specific effects of the Hoechst dye on the sorted cells, we cultured both SP and non-SP cells for 24 h to remove dead cells and then performed all experiments described below.

Sorted SP and non-SP cells were cultured in RPMI-1640 (supplemented with 10% FBS) for 28 days. Then the cultured SP and non-SP cells were stained with Hoechst 33342 and analyzed by FACS to determine the differentiation ability of the two subpopulations.

Sorted SP and non-SP cells were cultured in tumor sphere medium consisting of serum-free DMEM/F12 medium, B27 supplement, 20 ng/ml EGF and 10 ng/ml bFGF. Cells were plated at a density of 5×10³ cells/well in ultra-low-attachment 6-well plate triplicates and the medium was changed every other day. After 10 days in culture, colonies that contained >20 cells were counted.

For clone formation assay, sorted SP and non-SP cells were plated in triplicate at 600 cells on each 25-cm² flasks and cultured with RPMI-1640 (supplemented with 10% FBS) for 10 days. Then, cells were fixed and stained with 0.5% crystal violet. Colonies containing >50 cells were manually counted. The clone formation efficiency was the ratio of the clone number to the planted cell number.

Numbers of sorted SP and non-SP cells (1×10⁵) suspended in 200 μl PBS were injected subcutaneously in the flank region of 5-week-old female BALB/c nude mice obtained from the Institute of Laboratory Animal Science of Peking University Health Science Center. The mice were monitored weekly and euthanized 4 weeks after transplantation to assess tumor formation. Tumors were measured using a vernier caliper, weighed and photographed. Tumor volume (TV) was calculated using the following formula: TV (mm³) = (length × width²)/2. A portion of the subcutaneous tumor tissue was collected, fixed in 10% formaldehyde and embedded in paraffin for hematoxylin and eosin (H&E) staining to assess tumor pathology. All animal practices were in accordance with the guidelines of Peking University Health Science Center for the use of laboratory animals.

Subsequently, 600–8,000 cells were plated on 25-cm² flasks in triplicate for each experiment. Twelve hours later, the cells irradiated at room temperature with a ⁶⁰Co laboratory irradiator (Beijing Normal University, Beijing) at a dose rate of 1 Gy/min for the time required to generate a dose curve of 0, 0.5, 1, 2, 4, 6 and 8 Gy. The control was sham irradiated. Following the irradiation, the cells were incubated for an additional 9 days, and the cells were fixed with 100% carbinol and stained with 0.5% crystal violet. Only colonies containing >50 cells were manually counted. The surviving fraction was calculated as follows: plating efficiency (PE) = (colony number/inoculating cell number) × 100%. SF = PE (tested group)/PE (0 Gy group) × 100%. The cell-survival was calculated according to the single-hit multi-target formula: SF = 1 − (1 − e^−D/D0)^N(18). The radiobiological parameters of cellular radiosensitivity (D₀, mean lethal dose), the capacity for sublethal damage repair (D_q, quasi-threshold dose) and the extrapolation number (N) were calculated. Then, those values were used to calculate the SF after irradiation at a dose of 2 Gy (SF2) and the sensitization enhancement ratio (SER).

Total RNA was extracted from newly sorted SP and non-SP cells using Takara RNAiso Plus (Dalian, China) and reverse transcriptions were performed according to the protocol supplied by the manufacturer (Takara, Japan).

Real-time PCR was performed using SYBR-Green I master mix kit (Takara) on a Bio-Rad IQ5 Real-time-PCR Reaction System (Bio-Rad Laboratories, Inc., CA, USA). The relative amounts of mRNA were calculated from the values of comparative threshold cycle by using GAPDH as control (primers are depicted in Table I). The reaction was carried out with the following cycling conditions: 95°C for 2 min followed by 45 cycles of amplification (denaturation at 95°C for 15 sec, annealing at 58°C for 20 sec and extension at 72°C for 30 sec).

To overexpress and silence the GRP78 in SP cells, the plasmid (pcDNA3.1/hGRP78, a gift from Dr RC Austin, McMaster University, Ontario, Canada; pSuper/GRP78 RNAi, designed by our laboratory) which can overexpress and silence the GRP78 in mammalian cells was introduced by transfection. For transient transfection, SP cells were cultured in 6-well plates and transfected at 80% confluence with Lipofectamine^® 2000 according to the manufacturer’s instructions. After transfection, the cells were left for another 36 h before they were harvested by trypsinization and resuspended for clonogenic experiments. The siRNA sequences for human GRP78 are: sense 5′-GATCCCCGATCACAATCACCAATGACTTCAAGAGAGTCATTGGTGATTGTGATCTTTTTA-3′ and antisense 5′-AGCTTAAAAAGATCACAATCACCAATGACTCTCTTGAAGTCATTGGTGATTGTGATCGTG-3′.

Statistical analysis was performed with Sigmaplot 10.0 software. Statistical differences between SP and non-SP cells were analyzed using the Student’s t-test. Data are presented as the means ± SEM. P-values <0.05 were considered to indicate statistically significant differences.

We first attempted to isolate a side-population within MCF-7 breast cancer cell lines. Following trypsinization and staining with fluorescent dye Hoechst 33342, the cells were analyzed by flow cytometry; the P2 gate showed that the SP cells were Hoechst 33342 negative/low and the P3 gate indicated the non-SP cells that were Hoechst 33342 positive (Fig. 1A). MCF-7 cells presented a distinct SP, accounting for 3.3% of the whole population. The percentage of SP cells diminished to ~0.0% of the total cells when pretreated with verapamil (Fig. 1B), confirming that SP cells extrude Hoechst 33342 dye actively via a verapamil-sensitive ABC transporter. Then the SP (P2) and non-SP (P3) cells were sorted separately and applied for further experiments; the purity of SP and non-SP cells was 94.3 and 99.7%, respectively (Fig. 1C and D).

High tumorigenic potential is considered a hallmark of CICs (19). We first performed tumorigenicity assay to compare the tumorigenic potential of SP and non-SP cells, respectively. Both 1×10⁵ numbers of SP and non-SP cells were injected into the flank region of nude mice subcutaneously (n=3). Four weeks after inoculation, both SP and non-SP cells were able to produce tumors as shown in Fig. 2A. Then, the mice were euthanized and the tumors were measured with a vernier caliper. The results showed that SP cells formed a tumor with a mean volume of 1043.12±163.65 mm³ while non-SP cells formed tumor with 204.04±144.86 mm³ mean volume (Fig. 2B). H&E staining results confirmed that the tumors formed by SP and non-SP cells were typical adenomatous carcinoma (Fig. 2C). H&E staining of tumors grown in mice after injection of SP cells showed the presence of malignant cells, with large nuclei and prominent nucleoli; some cells showed a dark basophilic cytoplasm. Moreover, cells were in a chaotic arrangement with necrosis (Fig. 2C).

Additionally, the clone formation assay showed that the mean clone formation efficiency was 77.56±3.67 and 26.39±3.25% in SP and non-SP cells, respectively (Fig. 2D). In vitro clonogenic potential indicated the efficiency of SP and non-SP cells to form a tumor, which is consistent with the in vivo tumor transplant results.

SP cells and non-SP cells were cultured for 4 weeks in normal RPMI-1640 medium, stained again with Hoechst 33342 and analyzed using the cell sorter. As shown in Fig. 3, SP and non-SP fractions were obtained again from the former SP subpopulation after being cultured for 4 weeks (Fig. 3C). By contrast, SP fraction could not be obtained from the former non-SP population (Fig. 3D). This indicated that SP cells may undergo asymmetrical division to generate heterogeneous phenotypes of low-tumorigenic cells, such as non-SP cells that form the bulk of the tumor.

Sphere formation has been well described as a typical characteristic of CICs that reflects the potential for self-renewal (20). We evaluated the ability of SP and non-SP cells to generate spherical colonies in an ultra-low attachment and serum-starved culture system. After 10 days of culture, spherical colonies were counted. As shown in Fig. 4A, SP cells displayed much higher tumor sphere-forming ability than non-SP cells; the sphere formation efficiency was 10.17±2.33% for the SP cells vs. 2.33±0.25% for the non-SP cells (Fig. 4B).

To further confirm the stem phenotype of the SP cells, the expression of embryonic stem cell (ES) marker genes were also investigated, as CICs are considered to share similar characteristics with normal stem cells (21). As shown in Fig. 4C, the q-PCR analysis indicated that the mRNA expression of stemness genes such as Oct4, Nanog, Sox2 and Bmi1 in SP cells was significantly higher than non-SP cells. Furthermore, the results also showed that the mRNA levels of ABCG2 in SP cells were more highly expressed as compared to non-SP cells, according to the sorting phenotype of Hoechst 33342 exclusion.

Current data demonstrate that breast CICs can be enriched after radiation and CSCs/CICs are particularly more resistant to radiotherapy (6,7,22). To compare the SP and non-SP cell response to radiation-induced cytotoxicity, clonogenic assays were performed. The dose-dependent survival curves of SP and non-SP cells are presented in Fig. 4D, the SP cells exhibited more resistance than the non-SP cells. Accordingly, SP cells had larger survival fraction values at 2 Gy irradiation than non-SP cells (Table II). By application of the single-hit multi-target model, the values of D₀, N and D_q of SP and non-SP cells were analyzed. As shown in Table III, SP cells display significantly larger values of D_q than non-SP cells, indicating that enhanced repair of sublethal damage may contribute to higher surviving fraction after irradiation in SP cells.

GRP78 has been hypothesized to be a key regulator of the therapeutic resistance properties of cancer stem-like cells. We then examined the GRP78 expression in SP cells. q-PCR analysis demonstrated that the expression level of GRP78, which is required for survival of embryonic stem cell precursors, was highly expressed in SP cells (Fig. 4C).

To evaluate the correlation between GRP78 expression profile and radiation resistance, the GRP78 gene was silenced using a small interfering RNA expressed in a pSuper vector. The SP cells were transfected with pSuper and pSuper/GRP78 RNAi vector. After 36 h transfection, q-PCR analysis confirmed that the expression of GRP78 was markedly suppressed in pSuper/GRP78 RNAi vector-transfected SP cells (Fig. 5A). Then, the transfected cells were harvested by trypsinization and resuspended for clonogenic experiments. As shown in dose-dependent survival curves (Fig. 5C), the IR effect on SP cells treated with pSuper/GRP78 RNAi vector was significantly stronger as compared to SP cells treated with a scrambled control. These data demonstrated that the silencing of GRP78 significantly enhanced the sensitivity of SP cells to IR.

In order to further verify the role of GRP78 in the resistance of SP cells to irradiation, we performed the gain-of-function approach by transfection with pcDNA3.1(+)/hGRP78 plasmids transient overexpressing GRP78 into SP cells. Total mRNA from SP cells with transfection of GRP78-expressing plasmids displayed elevated expression of GRP78 (Fig. 5B). Then, exposed to IR, the results indicated that GRP78 overexpression resulted in increased radiation resistance in SP cells (Fig. 5D). As shown in Tables II and III, the survival fraction of 2 Gy and radiobiological parameters demonstrated that knockdown of the GRP78 gene increases the effects of IR, while overexpression of GRP78 decreases the radiation-induced cytotoxicity to SP cells. Collectively, our data provide evidence that GRP78 upregulation in SP cells mediates the resistance to IR.