The informed consent form was signed by each participant, and appropriate ethics committee approval was obtained. All the experiments complied with the current laws of China.

The MDA-MB-231 breast carcinoma cell line was purchased from American Type Culture Collection (ATCC; Manassas, VA, USA). The MDA-MB-231-HM breast carcinoma cell line, which readily metastasizes to the lungs, was established previously (23). Briefly, MDA-MB-231 cells (2×10⁷) were orthotopically injected into the exposed axillary mammary fat pad of athymic mice after anesthetization. Mice were sacrificed after 8–10 weeks under anesthetization and pulmonary metastatic lesions tissue samples were removed for primary culture in Leibovitz's L-15 (Gibco) medium supplemented with 10% (v/v) fetal bovine serum (PAN Biotech) at 37°C in a humidified atmosphere containing 5% CO₂. Cells were passaged continuously until there were no apparent mesenchymal cells. Cells grown to ~80% confluence were reinoculated into the mammary fat pad as described above and tissue samples from metastatic lesions were again obtained and treated as above. The procedures were repeated five to six times. A highly lung-metastatic variant of parental MDA-MB-231 cell line, termed MDA-MB-231-HM was established, followed by essential detections, such as Mycoplasma contamination. These two cell lines were cultured in Leibovitz's L-15 (Gibco) medium supplemented with 10% (v/v) fetal bovine serum (PAN Biotech) at 37°C in a humidified atmosphere containing 5% CO₂. The pLKO.1-TRC cloning vector (Addgene plasmid 10878) was utilized to construct two shRNA plasmids, shCD59 and shCD59', via insertion of shCD59F 5′-CCGGCCGTCAATTGTTCATCTGATTCTCGAGAATCAGATGAACAATTGACGGTTTTTG-3′ and shCD59R 5′-AATTCAAAAACCGTCAATTGTTCATCTGATTCTCGAGAATCAGATGAACAATTGACGG-3′ or shCD59′F 5′-CCGGGCTAACGTACTACTGCTGCAACTCGAGTTGCAGCAGTAGTACGTTAGCTTTTTG-3′ and shCD59'R 5′-AATTCAAAAAGCTAACGTACTACTGCTGCAACTCGAGTTGCAGCAGTAGTACGTTAGC-3′, respectively. Lentiviruses containing one of 2 control vectors or one of 2 CD59-targeting shRNAs were added to MDA-MB-231-HM culture medium with 8 μg/ml polybrene (Sigma-Aldrich), and the cells were incubated with the indicated lentiviruses for 24 h. The infected cells were selected using 1 μg/ml puromycin (Sigma-Aldrich).

Anti-β-actin (C4), anti-TFIIB (D-3), anti-CD59 (H-7), anti-CD46 (M177), anti-CD55 (H-7), anti-phospho-ERK1/2 (E-4), anti-p65 (F-6), anti-p50 (E-10), anti-cRel (B-6), anti-Sp1 (E-3), anti-CREB-1 (24H4B), and anti-phosphorylated CREB-1 (Ser-133) antibodies were purchased from Santa Cruz Biotechnology. Anti-Akt (10176-2-AP) was purchased from ProteinTech Group. Phospho-Akt^Ser473 (2118-1) antibody was purchased from Epitomics. FITC-conjugated mouse anti-human CD59 mAb (p282/H19) was purchased from BD Pharmingen. A rabbit anti-CD59 polyclonal antibody (10742-1-AP) was obtained from Proteintech Group. Polybrene, puromycin, and CelLytic M (for mammalian cell lysis and extraction) were purchased from Sigma-Aldrich. Cell Counting Kit-8 reagents were purchased from Dojindo. U0126 and MK2206 were purchased from Selleck Chemicals and used at terminal concentrations of 1 and 2 μM.

Total cell lysates were extracted using a mammalian cell lysis/extraction reagent (Sigma-Aldrich). Nuclear and cytoplasmic proteins were isolated using a nuclear protein extraction kit (Active Motif). Immunoblotting was done, essentially as described (24). The primary antibodies used were as follows: anti-β-actin (C4; Santa Cruz Biotechnology), anti-TFIIB (D-3; Santa Cruz Biotechnology), anti-CD59 (H-7; Santa Cruz Biotechnology), anti-CD46 (M177; Santa Cruz Biotechnology), anti-CD55 (H-7; Santa Cruz Biotechnology), anti-ERK1/2 (C-9; Santa Cruz Biotechnology), anti-anti-phospho-ERK1/2 (E-4; Santa Cruz Biotechnology), anti-p65 (F-6; Santa Cruz Biotechnology), anti-p50 (E-10; Santa Cruz Biotechnology), anti-cRel (B-6; Santa Cruz Biotechnology), anti-Sp1 (E-3; Santa Cruz Biotechnology), anti-phosphorylated Sp1 (phospho-T453; Bioworld Technology), anti-CREB-1 (24H4B; Santa Cruz Biotechnology), anti-phosphorylated CREB-1 (Ser-133; Santa Cruz Biotechnology), rabbit anti-CD59 polyclonal antibody (10742-1-AP; ProteinTech Group). The quality of loading and transferring was assessed by immunostaining with the β-actin antibody. The dilutions for the antibodies were as follows: 1:500 for CD59, 1:500 for CD55, 1:500 for CD46, 1:1,000 for TFIIB, 1:1,000 for p65, 1:500 for p50 1:1,000 for cRel, 1:1,000 for Sp1, 1:500 for phosphorylated Sp1, 1:1,000 for CREB-1, 1:1,000 for phosphorylated CREB-1, 1:500 for ERK1/2, 1:500 for phospho-ERK1/2, 1:1,000 for rabbit anti CD59 polyclonal antibody and 1:2,000 for β-actin.

CCK8 assays were used to measure cell proliferation. The indicated cells were seeded at a density of 2×10³ cells in a volume of 200 μl per well in 96-well plates (Nunc). Before detection, 100 μl of fresh culture medium was added into each well after the cells had been washed twice with PBS. Then, 10 μl of CCK8 solution (Dojindo) was added to each well, and the plates were incubated for 7 days at 37°C. Cell viability was measured as the absorbance at 450 nm with a microplate reader (BioTek), and the mean optical density (OD450) values from triplicate wells for each cell line were used as an index of cell viability. All the experiments were performed in triplicate.

CD59 expression at the cell membrane was detected by flow cytometry as described previously (24). Briefly, the harvested cells were incubated with a FITC-conjugated anti-human CD59 antibody (BD Pharmingen) and then analyzed using a Cytomics FC 500 MPL (Beckman Coulter).

For cell cycle analysis, cells were digested into single-cell suspension and 1×10⁶ cells were suspended by 1 ml PBS. Three milliliters Pre-cold ethanol (3 ml) at −20°C was added into cell suspension with vortex and then the sample was fixed at −20°C overnight. Fixed cells were collected by centrifugation at 2,000 rpm for 5 min and the supernatant was discarded. Cells were suspended by PBS for 15 min at room temperature. Cells were then recollected by centrifugation at 800 rpm for 5 min and stained by Cell Cycle Staining kit [MultiSciences (Lianke) Biotechnology] as instructed. Then cell cycle distribution was analyzed by flow cytometry.

Total RNA was extracted from cells with or without relative treatments using TRIzol reagent (Invitrogen) and reverse transcribed into cDNA using PrimeScript™ RT Master Mix (Takara) following the manufacturer's instructions. Standardized cDNA was used as template for amplification of 40 cycles with SYBR Green Master Mix and specific primers of different genes on ABI PRISM 7900HT (Applied Biosystems). β-actin was used as endogenous control, and samples were analyzed in triplicate. The primers for amplification of β-actin were 5′-ACC GAGCGCGGCTACAG-3′ (forward) and 5′-CTTAATGTCACGCACGATTTCC-3′ (reverse). The primers for amplification of CD59 were 5′-ATGGGAATCCAAGGAGGGTC-3′ (forward) and 5′-ATTGACGGCTGTTTTGCAGT-3′ (reverse).

Single-cell suspension was obtained by 0.25% trypsin digestion and 500 cells were seeded in each well of 6-well plates. After incubation at an atmosphere of 5% CO₂ at 37°C for 15 days, cells were washed with PBS twice and fixed with 4% paraformaldehyde/PBS (Solarbio) for 30 min. Then cells were stained by 0.1% crystal violet staining solution (Solarbio) for 30 min and washed with PBS 3 times. The number of colonies was counted and the colony formation efficiency was calculated as number of colonies/number of cells seeded x 100%.

Immunohistochemistry for CD59, Akt, and phospho-Akt was performed on the slides, using a two-step protocol (EnVision™ Detection Systems, Dako, Glostrup, Denmark). The slides were dewaxed with xylene, gradually hydrated with gradient ethanol, and washed with phosphate-buffered saline (PBS). Antigens were retrieved by boiling the slides in citrate buffer (pH 6.0) at 121°C for 15 min. Endogenous peroxidase was blocked by incubating the slides in 3% H₂O₂ solution. Then, the slides were washed in PBS for 10 min, blocked with 10% normal goat serum (CW0130; CWBIO, Beijing, China) for 30 min at 37°C, and incubated with anti-CD59 (H-7; Santa Cruz Biotechnology) at 4°C overnight. The antibody dilution was 1:100 for the CD59 antibodies and 1:100 for Akt and phospho-Akt antibody. The immunoreactive products were visualized by the catalysis of 3, 3-diaminobenzidine by horseradish peroxidase (EnVision™ Detection Systems), following extensive washing. The slides were then counterstained with Gill hematoxylin (BA-5021, Zhuhai, China) and dehydrated in ascending grades of ethanol, and finally, cleared in xylene and mounted under a coverslip.

As in our study, the expression of CD59 and phospho-Akt was semi-quantitatively classified according to the staining index (SI; range 0–9), which was calculated by multiplying the staining intensity by the proportion score. The staining intensities were classified according to four grades (0 denoting negative; 1, weak; 2, moderate; and 3, strong), and the proportion score was graded as the percentage of cells that were stained (0 denoting 0 to <10% of cells; 1, between 10 and 30% of cells; 2, between 31 and 60% of cells; and 3, staining of >60% of cells). In this study, SI 3 was defined as CD59-positive staining, whereas SI <3 was defined as negative staining. The IHC scores were independently determined by two experienced pathologists (Yiqun Du and Yun Xia) as described previously with slight modifications (25,26).

All the mouse experiments were performed according to a protocol approved by the Fudan University Shanghai Cancer Center (Shanghai Cancer Center Ethics Committee) Animal Care and Use Committee. Female Balb/C nude mice, aged 8 weeks, were purchased from SLAC Laboratory Animals, and were maintained under specific pathogen-free conditions. The mice were acclimated in-house for ≥7 days prior to use. A total of 1.0×10⁷ cancer cells were suspended in complete culture medium containing 10% FBS and then injected into the left fourth mammary fat pad of each female BALB/c nude mouse; a total of seven mice were included in each group. Tumor size was measured twice weekly with calipers, and tumor volume was calculated using the following formula: (width)² x length/2. At the endpoint of the experiment, the tumor tissues were carefully excised and weighed after the mice were anesthetized and sacrificed with CO₂.

Between May 2008 and July 2008, 120 women with breast cancer, as confirmed by core needle biopsy, with no prior treatment were treated by surgical resection at the Shanghai Cancer Center, Fudan University. This study was approved by the independent ethics committee/institutional review board of the Fudan University Shanghai Cancer Center (Shanghai Cancer Center Ethics Committee). All patients provided written informed consent before they were included in this study.

All the data are presented as the mean and standard deviation (SD). The χ² test was used to evaluate the relationship between patient characteristics and body mass index (BMI) and to compare groups based on CD59 expression. Fisher's exact test was performed when necessary. A multivariate logistic regression model to predict metastasis was utilized based on both categorical variables (CD59 expression, tumor size, lymph node status, cancer embolus, ER status, PR status, and HER-2 status) and continuous variables (age) that were evaluated at the time of diagnosis. All the statistical tests were two-sided, and P-values <0.05 were considered significant. All the analyses were performed using SPSS (version 13.0; SPSS Inc., Chicago, IL, USA).

We performed immunoblotting assays with total cell lysates to elucidate the expression levels of CD59 in breast cancer cell lines, such as ZR-75-30, T47D, MCF-7, MDA-MB-453, SK-BR-3, ZR-75-1, MDA-MB-468, MDA-MB-231 cells and the derived MDA-MB-231-HM cells, which are prone to lung metastasis. CD59 expression was significantly upregulated in MDA-MB-231-HM cells and ZR-75-30 cells (Fig. 1A). CD59, but not that of two other mCRPs (CD46 and CD55) was upregulated in MDA-MB-231-HM compared with MDA-MB-231 (Fig. 1B). Flow cytometry assays and real-time PCR further confirmed that CD59 expression was significantly higher in MDA-MB-231-HM cells compared to the parental cells (Fig. 1C and D).

We recently reported that the constitutive expression of CD59 is controlled by Sp1, whereas the inducible expression of CD59 is regulated by NF-κB and enhancer-binding CREB (24). To investigate the underlying mechanisms by which MDA-MB-231-HM cells upregulate CD59, we first examined the nuclear levels of total and phosphorylated Sp1 and the nuclear translocation of NF-κB. There were no significant differences between these two cell lines with respect to Sp1 and NF-κB signaling (Fig. 1E). However, there was increased phosphorylation of CREB, another transcription regulator, in MDA-MB-231-HM cells compared to MDA-MB-231 cells (Fig. 1F). CREB can be directly or indirectly phosphorylated by phospho-Akt and phospho-ERK (27–29), so we determined the levels of Akt and ERK phosphorylation. Both phospho-Akt and phospho-ERK were significantly increased in MDA-MB-231-HM cells compared to MDA-MB-231 cells (Fig. 1G). The same result failed to apply to the other group of breast cancer cells: ZR-75-1 and ZR-75-30 (data not shown), as the levels of phosphorylated CREB between ZR-75-1 and ZR-75-30 showed no significant difference. This is understandable considering the complicated regulation of CD59 and that MDA-MB-231-HM was derived from MDA-MB-231. To investigate the mechanism underlying the upregulation of CD59 in cell MDA-MB-231-HM, we inhibited ERK phosphorylation and Akt phosphorylation using U0126 and MK2206, respectively. The inhibition of ERK phosphorylation downregulated CD59 expression both at mRNA level and protein level (Fig. 1H and I), while the inhibition of Akt phosphorylation failed (Fig. 1J). These findings suggested that CD59 was upregulated in breast cancer MDA-MB-231-HM cells and that this upregulation may occur via the ERK pathway.

Breast cancer lung metastasis leads to a poor prognosis (6,7). We utilized MDA-MB-231-HM cells, which are prone to lung metastasis, to explore the role of CD59 in the highly metastatic MDA-MB-231-HM cells. We knocked down CD59 expression in MDA-MB-231-HM cells using two vectors expressing shRNA against CD59, referred to as 231-HM-shCD59 and 231-HM-shCD59' (the corresponding controls were designated 231-HM-Vector and 231-HM-Vector', respectively). As shown in Fig. 2A, CD59 was robustly knocked down in MDA-MB-231-HM cells. 231-HM-shCD59 was more efficient at knocking down CD59 expression, and we therefore selected this construct for use in subsequent studies.

We then used CCK8 assays to determine the effect of CD59 knockdown on cell proliferation. As shown in Fig. 2B, 231-HM-shCD59 cells lacking CD59 expression grew more slowly than 231-HM-Vector cells beginning on day 4 after cell plating (P<0.01). Colony formation assay showed similar result (P<0.01) (Fig. 2C), and cell cycle analysis indicated that this growth arrest may due to the cell cycle arrest in 231-HM-shCD59 cells (Fig. 2D). We next implanted 231-HM-Vector and 231-HM-shCD59 cells into the left fourth mammary fat pads of female athymic nude mice. Primary tumors developed successfully in all mice, and CD59 knockdown markedly decreased tumor growth, as evidenced by tumor size (Fig. 2E, P<0.02, P<0.01). At 5 weeks after injection, all the mice were sacrificed, and their tumors and lungs were resected, as MDA-MB-231-HM is prone to lung metastasis. The tumors were weighed, and the lungs were examined by hematoxylin and eosin (H&E) staining. The tumors in the 231-HM-shCD59-injected mice were clearly much smaller than those in the control mice, and the corresponding tumor weights were consistent with this observation (representative images are shown in Fig. 2F, P<0.01). In the H&E-stained slices, we found metastases in 3/10 lungs from the 231-HM-Vector group, whereas none of the 10 lungs from the 231-HM-shCD59 group had meta-static lesions; representative images are shown in Fig. 2G. These data indicated that CD59 knockdown suppressed cell proliferation both in vitro and in vivo and may decrease lung metastasis in vivo.

To explore the clinical significance of CD59 expression in prognosis, we analyzed a cohort of 120 breast cancer patients, 58 with lung metastases and 62 without metastatic disease. We performed IHC staining for CD59 expression, and the expression scores were ranked by two independent pathologists who were blinded to the identity of each sample. Representative images of CD59 in 3 patients with metastasis and in 3 patients without metastasis are shown in Fig. 3. We retrospectively collected complete clinical data, including age, tumor size, pathological assessment and follow-up visit details. Table I shows the characteristics of the patients grouped based on CD59 expression. Only the presence of metastasis was significantly associated with CD59 expression (P=0.001). There was a higher percentage of HER-2-negative patients among those with high CD59 expression. However, age, tumor size, node status, tumor histology, ER status, PR status and HER-2 status were not significantly correlated with CD59 expression.

The metastasis rate was 48.3% (58/120) in the study cohort. Table II shows the univariate and multivariate analyses of the metastasis rate predictors using a logistic regression model. In the univariate model, there were no significant relationships between metastasis rate and tumor size, tumor histology, ER status, PR status or HER-2 status (P>0.05). However, metastasis was significantly associated with age, lymph node status, cancer embolus and CD59 expression (P<0.05). Patients 35 years or older, as well as those with positive lymph node status, cancer embolus or high CD59 expression were more likely to develop metastases. In the multivariate model, CD59 expression, lymph node status and cancer embolus were independent predictors of metastasis, with P-values of 0.041, 0.032 and 0.008, respectively. Notably, a Kaplan-Meier survival analysis revealed that CD59 expression levels were negatively correlated with relapse-free survival among breast cancer patients (P<0.001). The analysis indicated that patients with higher CD59 expression would have a worse prognosis compared to patients with lower CD59 expression (Fig. 4). These clinical data suggested that CD59 expression may be a prognostic biomarker for metastasis and outcome in breast cancer patients.