• ovarian cancer;
  • membrane proteome profiling;
  • 2-D PAGE;
  • selenium binding protein 1;
  • prognostic marker;
  • androgen;
  • methylselenocysteine


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Selenium binding protein 1 (SELENBP1) was identified to be the most significantly down-regulated protein in ovarian cancer cells by a membrane proteome profiling analysis. SELENBP1 expression levels in 4 normal ovaries, 8 benign ovarian tumors, 12 borderline ovarian tumors and 141 invasive ovarian cancers were analyzed with immunohistochemical assay. SELENBP1 expression was reduced in 87% cases of invasive ovarian cancer (122/141) and was significantly reduced in borderline tumors and invasive cancers (p < 0.001). Cox multivariate analysis within the 141 invasive cancer tissues showed that SELENBP1 expression score was a potential prognostic indicator for unfavorable prognosis of ovarian cancer (hazard ratio [HR], 2.18; 95% CI = 1.22–3.90; p = 0.009). Selenium can disrupt the androgen pathway, which has been implicated in modulating SELENBP1 expression. We investigated the effects of selenium and androgen on normal human ovarian surface epithelial (HOSE) cells and cancer cells. Interestingly, SELENBP1 mRNA and protein levels were reduced by androgen and elevated by selenium treatment in the normal HOSE cells, whereas reversed responses were observed in the ovarian cancer cell lines. These results suggest that changes of SELENBP1 expression in malignant ovarian cancer are an indicator of aberration of selenium/androgen pathways and may reveal prognostic information of ovarian cancer. © 2005 Wiley-Liss, Inc.

Despite advances in cancer therapeutic agents in recent years, approximately 14,000 women still die of ovarian cancer each year in the United States, making it the most lethal of the gynecological malignancies.1 Currently, surgical debulking followed by chemotherapy is the major treatment approach for ovarian cancer. Most cases of ovarian cancer are diagnosed at advanced stages when the prognosis for 5-year survival is poor.2 Developing new diagnostic technique and improving current therapeutic strategy are the main directions to fight this morbid disease.

Multiple genomic and proteomic approaches have been applied to identify disease-associated biomarkers for diagnosis and disease management. Applying cDNA and oligo microarray technology have enabled scientists to look into the global differences in gene expression between normal and cancer cells.3 For proteomic approaches, two-dimensional polyacrylamide gel electrophoresis (2-D PAGE) followed by protein identification using mass spectrometry has been the primary technique for biomarker discovery.4, 5 Membrane-associated proteins have been suggested to be a potential resource for biomarker screening.6 In our pilot study of the use of 2-D PAGE to profile membrane-associated proteins of normal ovarian surface epithelial cells and ovarian cancer cell lines, selenium binding protein 1 (SELENBP1) was identified as the most significantly down-regulated protein in the cancer cells.

The human selenium binding protein gene (SELENBP1) was mapped at chromosome 1q21–22.7 It has been noted that high levels of SELENBP1 transcripts were detected in normal tissues that appeared to benefit from cancer-preventive action of dietary selenium, such as prostate, colon, lung, liver, kidney and pancreas.8 Reduced expression of SELENBP1 in lung9, 10 and gastric11 adenocarcinomas has also been reported. The molecular mechanism of down-regulation of SELENBP1 in cancer cells is not clear at present. Promoter hypermethylation and gene deletion are not the mechanisms in lung adenocarcinomas.10 However, androgen treatment can reduce the level of SELENBP1 transcript in a concentration-dependent manner in a human prostate cancer cell line LNCaP.8 In a recent report, selenium has been shown to disrupt the androgenic signaling pathway in LNCaP cells.12 Such evidence suggests that selenium may also modulate the expression status of SELENBP1.

This study was undertaken to determine the expression of SELENBP1 in normal and tumor tissues of ovary. The prognostic value of SELENBP1 for invasive ovarian cancer and the effects of androgen and selenium on the expression level of SELENBP1 in normal human ovarian surface epithelial (HOSE) and ovarian cancer cells were also investigated.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Tissue specimens

Archived formalin-fixed, paraffin-embedded tumor tissues that had been collected from patients undergoing primary laparotomy at the Brigham and Women's Hospital during 1986–2003 were used in the study. All patient-derived biologic specimens were collected and archived under protocols approved by the Human Subjects Committee of the Brigham and Women's Hospital, Boston, Massachusetts. All cases were staged according to International Federation of Gynecology and Obstetrics (FIGO) system.

Cell lines

Three ovarian cancer cell lines, DOV13, OVCA429 and OVCA882, were established previously in our laboratory from serous ovarian cancers. TOV112D and SKOV3 were 2 ovarian cancer cell lines from American Tissue Culture Collection (ATCC, Manassas, VA). Normal human ovarian surface epithelial (HOSE) cells were collected by scraping the ovary surface of the control subjects who were undergoing the procedure of hysterectomy or oophorectomy for benign diseases. HOSE2089 is a spontaneously immortalized cell line from a HOSE primary culture. All normal cells and cancer cells were established and grown in a mixture of medium 199 and MCDB105 medium (1:1) (Sigma, St. Louis, MO) supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, CA) as described previously.13

Preparation of total membrane proteins

Total membrane proteins were prepared according to the manufacturer's recommendations (2-D Sample Prep for Membrane Proteins, Pierce Biotechnology, Rockford, IL). Briefly, 106 normal or cancer cells were resuspended in 150 μl of Mem-PER Cell Lysis Reagent and incubated for 10 min at room temperature with occasional vortex mixing. Precooled solubilization reagent (450 μl) was then added into the sample and vortexed. The mixture was kept on ice for 30 min and vortexed every 5 min. After incubation, the undissolved cell debris was removed by centrifugation (10,000g for 3 min). The supernatant was then transferred to another clean tube and incubated at 37°C for 10 min. The hydrophobic and hydrophilic phases were partitioned by centrifugation (10,000g for 2 min), and the hydrophilic layer (upper layer) was then removed. The hydrophobic layer (enriched with membrane-associated proteins) was diluted with 100 μl of Mem-PER buffer and subsequently purified with the 2-D PAGEprep Resin. The isolated membrane proteins were further desalted with a protein desalting spin column, and the protein concentrations of the purified membrane proteins were determined with the 2-D Quant kit (Amersham Biosciences, San Francisco, CA).

2-D PAGE and comparative analysis of membrane proteins

For each profiling, 100 μg of total membrane proteins was loaded. The sample volume was adjusted to 185 μl by adding a sufficient amount of 1× rehydration buffer (Bio-Rad Laboratories, Hercules, CA). Isoelectrical focusing (IEF) was carried out with the ReadyStrip IPG Strip 5–8 (11 cm) (Bio-Rad Laboratories, Hercules, CA) and run for 35,000 V/hr at room temperature on the PROTEIN IEF Cell (Bio-Rad Laboratories, Hercules, CA). The strip was then equilibrated with the equilibration buffer I for 10 min and with buffer II for another 10 min (Bio-Rad Laboratories, Hercules, CA). The strip was then loaded on a 10% Tris-SDS gel for the second-dimensional separation. The sample was run for 1 hr at 200 V at 4°C. After electrophoresis, the gels were fixed with 10% methanol/7% acetic acid solutions for 30 min and stained with 1× Sypro Ruby protein stain solution (Bio-Rad Laboratories, Hercules, CA) overnight with gentle agitation. After the gels were destained with 10% methanol/7% acetic acid solutions for 30 min, they were scanned with the Molecular Imager FX (Bio-Rad Laboratories, Hercules, CA) and assayed with PD Quest software (Bio-Rad Laboratories, Hercules, CA). Approximately 800 spots were identified from each gel by the software. The cutoff value for differential expression was set to 1.5. Forty spots showed at least 1.5-fold differences between normal and cancer gels. Spot number 5,603 showed the most significant fold-change among these spots and was chosen for protein identification.

Mass spectrometry

SSP5603 was picked from the 2-D gel and submitted to Taplin Biological Mass Spectrometry Facility for protein identification. The protein of interest was trypsin-digested (in-gel), eluted and separated with high-pressure liquid chromatography (HPLC). As each peptide eluted from HPLC, it was subjected to electrospray ionization and then entered into a LCQ DECAXP Plus ion-trap mass spectrometer (ThermoFinnigan, San Jose, CA). Each ionized peptide was detected, isolated and fragmented to produce a tandem mass spectrum. Protein identity was determined by matching the acquired fragmentation pattern with the protein databases by Sequest software program (ThermoFinnigan, San Jose, CA).14

Western blot

Total membrane (or total cell) lysates were resolved by 1-D SDS PAGE (20 μg per sample), and the resolved proteins were transferred to a polyvinylidene fluoride (PVDF) membrane (Perkin Elmer, Boston, MA) with the SEMI-DRY Transfer cell (Bio-Rad Laboratories, Hercules, CA). After the transfer, the membrane was stained with Ponceau S solution (Sigma, St. Louis, MO) for 10 min to validate the transfer efficiency and serve as a loading control. After a brief washing step with 1× TBS (Tris-buffered saline) buffer, the membrane was blocked with 5% nonfat dry milk in 1× TBS at room temperature for 1 hr. The membrane was then incubated with anti-SELENBP1 antibody (1:1,000 dilution in blocking solution; Clone 4D4, MBL, Nagoya, Japan) at 4°C overnight. The bound antibody was detected by the anti-mouse secondary antibodies conjugated with peroxidase (Pierce Biotechnology, Rockford, IL). For signal detection, the membrane was incubated with SuperSignal ECL substrate (Pierce Biotechnology, Rockford, IL) for 5 min at room temperature. The membrane was then exposed to a XAR-5 X-ray film (Marsh Bio Products, Rochester, NY), and protein bands were visualized after developing the film. For protein bands quantitation, GS-700 imaging densitometer (Bio-Rad Laboratories, Hercules, CA) and Quantity one software (Bio-Rad) were used.

Quantitative PCR analysis

Total RNA was isolated from normal and cancer cells with RNeasy Mini Kit (Qiagen, Valencia, CA), according to the manufacturer's recommendations. The TaqMan Reverse Transcription reagent kit (Applied Biosystems, Foster City, CA) was applied for the cDNA synthesis step, and iTaq SYBR Green Supermix with ROX (Bio-Rad Laboratories, Hercules, CA) was used for real-time quantitative PCR analysis. All protocols were conducted according to the manufacturer's recommendations. During PCR, reactions were monitored continuously with the ABI Prism 5700 Sequence Detector (Applied Biosystems, Foster City, CA). The PCR reaction consisted of 1 DNA denaturation step (95°C, 3 min) and 40 cycles of amplification step (95°C, 15 sec; 60°C, 1 min). The dissociation curve was also generated for every run to validate the specificity of amplification. Cyclophilin A was chosen as the internal control because of its stable expression in the cells. The sequences of the primers used were described as follows: SBP1-1208F (5′-TCA GAT GAT CCA GCT CAG CCT-3′), SBP1-1317R (TCA CAG AGC CTT CCC TGA TGA-3′), CycA-432F (5′-AGA CTG AGT GGT TGG ATG GCA-3′), CycA-567R (5′-TGT CCA CAG TCA GCA ATG GTG-3′). Primers were designed with PrimerExpress 1.5 software (Applied Biosystems, Foster City, CA). The forward and reverse primers were located on different exons separated by long introns or on the exon–intron junction to prevent the generation of noise signals from contaminated genomic DNA during the PCR reactions. For calculation of relative mRNA level, the ΔΔCT method was used as described in the manufacturer's technical note (Applied Biosystems, Foster City, CA).


For immunohistochemistry, 7-μm sections were cut from the paraffin-archived tissues, mounted on Superfrost Plus microscopic slides (Fisher Scientific, Pittsburgh, PA) and incubated at 50°C for 4 hr. The sections were deparaffinized in xylene and rehydrated with a descending series of ethanol. For antigen unmasking, sections were immersed in antigen-unmasking solution (Vector Laboratories, Burlingame, CA) and boiled in a microwave oven for 10 min. Endogenous peroxidase activity was quenched using 0.3% H2O2 in methanol for 20 min. The sections were then blocked with normal horse serum for 20 min and were subsequently incubated overnight with anti-SELENBP1 antibody (1:500; Clone 4D4, MBL, Nagoya, Japan) at 4°C. After incubation, VECTASTAIN Elite Avidin-Biotin Complex kit (Vector Laboratories, Burlingame, CA) was used for the detection of the bound anti-SELENBP1 antibody. Diaminobenzidine (DAB) (Vector Laboratories, Burlingame, CA) was used for the color development. The sections were counterstained lightly with hematoxylin (Vector Laboratories, Burlingame, CA), dehydrated with an ascending series of ethanol, cleared in xylene and mounted in Permount (Fisher Scientific, Pittsburgh, PA). The staining was quantified with a semiquantitative scoring system.15 The weighted score was obtained by multiplying the staining intensity score (3+, strong positive stain in cells; 2+, moderate stain in cells; 1+, weak stain in cells; 0, no evidence of stain) and score for the percentage of positive cells (3+, most of cells stained; 2+, half of cells stained; 1+, few cells stained; 0, no cells stained). Two observers scored the slides independently, and the scores for all cases were compared for discrepancies. Final scores were assigned after discussion by the 2 observers. Representative photomicrographs of tumor tissues showing positive and negative staining for SELENBP1 are presented in Figure 3.

thumbnail image

Figure 3. Immunohistochemical staining of SELENBP1. (a), Positive staining of SELENBP1 in the normal ovary epithelium. The arrow sign indicates the layer cells of ovary surface epithelium. (b), Positive staining of SELENBP1 in a benign ovarian tumor. (c), Negative staining of SELENBP1 in a mucinous borderline tumor. (d), Positive staining of SELENBP1 in a mucinous borderline tumor. (e), Negative staining of SELENBP1 in a serous adenocarcinoma. (f), Positive staining of SELENBP1 in a serous adenocarcinoma. (g), Negative staining of SELENBP1 in a mucinous adenocarcinoma. (h), Positive staining of SELENBP1 in a mucinous adenocarcinoma. (i), Negative staining of SELENBP1 in an endometrioid adenocarcinoma. (j), Positive staining of SELENBP1 in an endometrioid adenocarcinoma. (k), Negative staining of SELENBP1 in a clear cell carcinoma. (l), Positive staining of SELENBP1 in a clear cell carcinoma. Scale bar represents 50 μm.

Download figure to PowerPoint

Statistical analysis

The nonparametric Kruskal–Wallis test was applied to test the equality of median scores of SELENBP1 among different age, diagnosis, grades, stage of disease and histopathological subtypes. Under the null hypothesis, the Kruskal–Wallis statistic has an asymptotic χ2 distribution with k-1 degrees of freedom, where k is the number of groups in the comparisons. In cases where the sample sizes of some groups are smaller than 5, the incomplete β approximation is adopted to calculate the p-value of the test statistic.16, 17 Significance of the test was considered at the 5% level (i.e. p ≤ 0.05). All calculations were performed with MINITAB statistical software (Minitab, State College, PA). For the survival analysis, 10 patients were excluded from the analysis because of the lack of follow-up information, as these patients were transferred to other hospitals after surgery. The length of overall survival was defined from the date of primary laparotomy to the date of the patient's death (uncensored) or the date of last visit (censored). Our objective was to study whether the expression level of SELENBP1 was linked to the patient's outcome (survival). We defined the low and high SELENBP1 expression groups using the cutoff score given by the median score of all samples plus one standard deviation (Score 4.6). The impact of SELENBP1 expression on patient survival was studied using the Cox proportional hazards regression model. The variables applied for adjustment in the multivariate analysis included tumor grade, stage of disease and patient's age. In the regression analysis, patient's age was treated as a continuous variable, while tumor grade and stage of disease were analyzed as ordinal variables. Grade 1 and 2 were grouped together to compare to grade 3 cases. Stage 1 and 2 cases were grouped together to compare to the group of stage 3 and 4 cases. The impact of SELENBP1 expression on patient survival was determined by examining the relative hazard ratios with respect to the low/high SELENBP1 expression groups. The significance of estimated hazard ratios was tested using the Wald test. A significant result implies a difference in overall survival between the low and high SELENBP1 expression groups. The Kaplan–Meier method was used to estimate the overall survival of patients belonging to the low and high SELENBP1 expression groups, respectively. Kaplan–Meier survival curves of these two groups were compared with the log-rank test. SELENBP1 immunoreactivity score was also applied as a continuous variable in a separate regression analysis to evaluate its prognostic value. All calculations within the survival analysis were performed with the S-PLUS software (Insightful, Seattle, WA).

MTT assay (cytotoxicity assay)

Cells were seeded in 100 μl of medium (5,000 cells) per well in 96-well plates. After overnight incubation, 100 μl of medium with different dilutions of selenium-(methyl)-selenocysteine (methylselenocysteine) (Sigma, St. Louis, MO) were prepared and added to the cells for 48 hr. Cell viability was quantified with the MTT cell proliferation assay kit (Roche Diagnostics, Indianapolis, IN) as described in the manufacturer's manual. The IC10 and IC25 value are the concentrations of methylselenocysteine resulting in a 10 and 25% reduction in absorbance, respectively, at 550 nm compared with that of mock-treated wells. Triplicate wells were used for each concentration in the assay.

Methylselenocysteine and dihydrotestosterone treatment

Cells were seeded and grown in a 37°C incubator for 6 hr before methylselenocysteine was added. For the mock-treated control plate, the same volume of culture medium was added. After 18 hr of treatment, cells were harvested for quantitative PCR analysis. For the dihydrotestosterone (DHT) (Sigma, St. Louis, MO) experiment, cells were grown in medium supplemented with hormone reduced serum (HyClone, Logan, UT) for 24 hr before adding DHT (final concentration 10 nM). The same volume of ethanol was added to the mock-treated control plate. Cells were harvested for quantitative PCR analysis after 24 hr of treatment.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

SELENBP1 protein identification and verification

Membrane proteins of normal HOSE cells and 3 ovarian cancer cell lines were isolated, profiled and analyzed as described in the “Methods” section. The difference between HOSE and cancer profiles was greatest for SSP5603 (Fig. 1). Protein identification analysis of SSP5603 showed coverage of 26.5% by amino acid count to SELENBP1 (Fig. 2a). SELENBP1 is a cytoplasmic protein, but a significant amount of SELENBP1 was also membrane associated, predominantly on the Golgi membrane.18 To validate our findings, SELENBP1 protein levels in the membrane fraction of HOSE and cancer cells were assayed with the SELENBP1-specific antibody. As expected, SELENBP1 was detected in the HOSE cells (Fig. 2b, lane 1) and was not detectable in the cancer cells (Fig. 2b, lane 3–5). However, HOSE2089, a spontaneous immortalized HOSE cell line, also had no detectable SELENBP1 proteins (Fig. 2b, lane 2). Similar results were obtained with total cell lysates (Fig. 2c). SELENBP1 mRNA levels in HOSE and cancer cell lines were also determined by quantitative PCR analysis. Down-regulation of SELENBP1 mRNA in immortalized HOSE2089 and cancer cell lines were identified (Table I).

thumbnail image

Figure 1. Sypro Ruby-stained 2-D PAGE of membrane proteomes. (a), Total membrane proteome of human ovary surface epithelial (HOSE) cells. (b), Enlarged image of the rectangle area in (a). SSP5603 was circled in the figure, and the two protein spots marked with a small rectangle are the landmark spots to position the gel. (c)–(e), The corresponding area containing SELENBP1 from DOV13 gel, OVCA429 gel and OVCA882 gel, respectively.

Download figure to PowerPoint

thumbnail image

Figure 2. Identification of SELENBP1. (a), Identification of SELENBP1 by the tandem mass spectrometry. Eight peptides that matched to SELENBP1 protein sequence are underlined and in bold font. (b), Western blot of total membrane proteins with the SELENBP1 antibody. Lane 1, HOSE cells; Lane 2, HOSE2089; Lane 3, DOV13; Lane 4, OVCA429 and Lane 5, OVCA882. Ponceau S-staining image shown at the bottom of the membrane protein western blot is the loading control. It has been validated as a sensitive and an quantitative assay for protein loading control.35, 36 (c), Western blot of total cell lysates with the SELENBP1 antibody. The order of loaded lysates is the same as in (b). β-actin was used as the loading control for the total lysate Western blot assay.

Download figure to PowerPoint

Table I. Relative Levels of SELENBP1 mRNA in Normal and Cancer Cells
  • 1

    Pooled human ovary surface epithelial cells.

  • 2

    SELENBP1 mRNA level is presented as the relative level to that of TOV112D.

  • 3

    SELENBP1 mRNA is not detectable with quantitative real time PCR analysis.


SELENBP1 expression levels in ovarian cancer

Levels of SELENBP1 expression in 4 normal ovaries, 8 benign ovarian tumors, 12 borderline ovarian tumors and 141 invasive ovarian cancers were analyzed by immunohistochemistry. SELENBP1 expression was reduced in 87% of invasive cancer cases (122/141). Representative photomicrographs of tumor tissues showing positive or negative staining of SELENBP1 are presented in Figure 3. SELENBP1 was localized in both the cytoplasm and nucleus as described previously.10 High levels of SELENBP1 proteins were identified in normal ovarian surface epithelium and benign tumors, while levels were reduced or diminished in most borderline tumors and invasive cancers (Table II). By using the incomplete β approximation16 to adjust for the small sample size, the differences among different diagnostic groups were statistically significant (p < 0.001). By performing pairwise comparisons between different diagnostic groups with the Bonferroni adjustment for multiple comparisons, it is found that there is significant evidence at 5% level of a difference between the median SELENBP1 expression level of invasive cancer and normal group. Significant difference was also found between borderline tumor and normal group. However, there was no significant difference between invasive cancer and borderline tumor group, or between normal and benign tumor group. No significant differences in SELENBP1 immunoreactivity among histopathologic subtypes, grades and stages were found (Table II).

Table II. Association Between SELENBP1 Immunoreactivity and Clinicopathological Parameters in Human Epithelial Ovarian Cancer
 No. of casesScore (mean ± SD)1p-value2
  • Note: All scores were obtained from formalin-fixed and paraffin-embedded tissue sections.

  • 1

    Mean with standard deviation.

  • 2

    p-value of Kruskal-Wallis test.

  • 3

    Statistically significant differences were found among groups.

Age (years)
 <65115127 ± 2.410.053
 ≥65503.2 ± 3.56
 Normal48.25 ± 1.50<0.0013
 Benign85.50 ± 3.25
 Borderline121.58 ± 2.15
 Invasive141233 ± 2.34
 Serous722.17 ± 2.360.37
 Mucinous242.42 ± 2.55
 Endometricid192.26 ± 2.6
 Clear cell83.13 ± 1.25
 Mixed182.56 ± 2.23
 I432.09 ± 2.060.36
 II233 ± 2.7
 III752.25 ± 2.38
Stage of disease
 I & II492.45 ± 2.40.56
 III & IV922.26 ± 2.33

Survival analysis

We also tested whether SELENBP1 staining provides any prognostic value within the invasive ovarian cancer group by performing survival analysis. Ten cases were excluded from our analysis because of the lack of follow-up information. Table III summarizes the association between SELENBP1 expression and the overall survival of patients with invasive ovarian cancer. In univariate Cox regression analysis, patients with higher levels of SELENBP1 had a higher risk of death than did patients with lower levels of SELENBP1 (hazard ratio [HR], 2.18; 95% CI = 1.24–3.83; p = 0.007). When SELENBP1 immunoreactivity was applied as a continuous variable in the survival analysis, a statistically significantly prognostic value for overall survival was still observed (HR, 1.17; 95% CI = 1.06–1.28; p = 0.001). In multivariate Cox regression analysis, higher risk of death for the patients with high expression level of SELENBP1 was found after adjustment for stage of disease, tumor grade and patient's age (HR, 2.18; 95% CI = 1.22–3.90; p = 0.009). Kaplan–Meier survival curves also showed that the small group of invasive cancer patients who had high expression of SELENBP1 had a significantly shorter overall survival (Fig. 4, log-rank, p = 0.006). When score of SELENBP1 immunoreactivity was treated as a continuous variable in a separate multivariate Cox regression analysis, it was also a significant predictor for disease prognosis after the adjustment (HR, 1.17; 95% CI = 1.06–1.29; p = 0.002).

thumbnail image

Figure 4. Kaplan–Meier curves for overall survival of patients with low and high expression levels of SELENBP1.

Download figure to PowerPoint

Table III. Univariate and Multivariate Survival Analysis
VariableHR195% CI2p
  • Abbreviations: HR, hazard ratio; CI, confidence interval; SELENBPI, selenium binding protein 1.

  • 1

    Hazard ratio estimated from Cox proportional hazard regression model.

  • 2

    Confidence interval of the estimated hazard ratio.

  • 3

    Median plus 1 standard deviation was applied for the cutoff value.

  • 4

    Score of SELENBP1 immunoreactivity was used as a continuous variable for the analysis.

  • 5

    Multivariate analysis was adjusted for stage of disease, tumor grade and age.

Univariate analysis
 SELENBP1, n = 131
  High expression level32.181.24–3.830.007
  SELENBP1, continuous variable41.171.06–1.280.001
  Stage of disease, ordinal6.113.05–12.20<0.0001
  Grading, ordinal2.401.50–3.850.0003
Multivariate analysis5
 SELENBP1, n = 131
  High expression level2.181.22–3.900.009
  Stage of disease, ordinal4.862.25–10.50<0.0001
  Grading, ordinal1.20.72–2.010.48
Multivariate analysis
 SELENBP1, n = 131
  SELENBP1, continuous variable1.171.06–1.290.002
  Stage of disease, ordinal5.102.37–11.0<0.0001
  Grading, ordinal1.190.71–2.000.50

Effect of androgen on SELENBP1 expression

The results of down-regulation of SELENBP1 in ovarian cancer cells, and the association of high levels of SELENBP1 expression with poor prognosis of invasive ovarian cancer patients suggest alteration of SELENBP1 expression in invasive ovarian cancer. To gain some insights into these observations, we investigated whether there are differences of SELENBP1 expression response to androgen treatment between HOSE cells and ovarian cancer cells, as other studies have shown that androgen can modulate SELENBP1 levels8 and androgen may be associated with ovarian cancer risk.19 Because the base levels of SELENBP1 protein in the immortalized HOSE and cancer cell lines are beyond detection, we determined the change of SELENBP1 expression in these cells by measuring SELENBP1 transcripts using quantitative real-time polymerase chain reaction. The results were summarized in Table IV. Treatment with dihydrotestosterone (DHT, active form of androgen) reduced the level of SELENBP1 mRNA in HOSE2089 cells, whereas it increased the levels of SELENBP1 mRNA in 3 ovarian cancer cell lines (Table IV). No significant change of the level of SELENBP1 mRNA was observed in DOV13 cells. To demonstrate that the SELENBP1 protein expression has the similar responses, we repeated the experiment using a normal HOSE primary culture (non-immortalized cells) and measure the changes of SELENBP1 protein level by Western blot analysis (Fig. 5a). As shown in Figure 5a, the SELENBP1 protein level in the normal HOSE primary culture reduced in response to the DHT treatment, similar to the mRNA response shown by the immortalized HOSE cells. Hence, our androgen treatment studies illustrated that there is a difference of SELENBP1 expression in response to androgen between HOSE cells and ovarian cancer cells. To investigate whether this is a general androgenic response, we determined the transcript levels of another androgen-regulated gene, human kallikrein 13 (hK13),20 by the same treatment. Hk13 mRNA levels have similar fold changes as SELENBP1 after DHT treatment in all cell lines, except DOV13, in which hk13 mRNA level increased after DHT treatment (data not shown), suggesting that there is a genuine difference in the androgen pathway between normal HOSE and ovarian cancer cells.

thumbnail image

Figure 5. Effect of dihydrotestosterone (DHT) (a), and methylselenocysteine (b) on SELENBP1 expression in non-immortalized human ovarian surface epithelial cells. Cells were treated with DHT and methylselenocysteine as described in the Methods. Cells were harvested by adding SDS lysis buffer in the culture dishes, and 20 μg of total lysate was loaded onto each lane for Western blot analysis. The SELENBP1 protein levels were normalized with the housekeeping gene β-actin levels for the calculation of percentage of control. The percentage of control is presented as the mean ± SD of 3 independent experiments; * indicates significantly different from the control (100%) at 5% level (p < 0.05) with the one sample t test and the nonparametric one sample sign test.

Download figure to PowerPoint

Table IV. Effects of DHT and Methylselenocysteine on SELENBP1 mRNA Levels in Immortalized Hose and Ovariancancer Cell Lines
Cell lineDHTMethylselenocysteine
10 nM1 μM2.5 μMIC10IC25
  • Abbreviations: DHT, dihydrotestosterone.

  • Note: Representative results from 2 independent experiments are presented. All data points were performed in triplicate.

  • 1

    Mean ± standard deviation (n = 3), SELENBP1 mRNA level is presented as the relative level of mock-treated control.

  • 2

    The IC10 or IC25 concentrations of methylselenocysteine for each cell line were given in the parentheses.

 HOSE20890.55 ± 0.l011.41 ± 0.091.57 ± 0.171.43 ± 0.03 (0.2 μM)21.50 ± 0.16 (5 μM)
 OVCA8821.90 ± 0.040.67 ± 0.170.57 ± 0.060.37 ± 0.01 (6 μM)0,52 ± 0.06 (100 μM)
 DOV131.09 ± 0.090.80 ± 0.080.64 ± 0.060.54 ± 0.02(50 μM)0.68 ± 0.03 (150 μM)
 SKOV31.90 ± 0.231.07 ± 0.080.56 ± 0.020.98 ± 0.01 (0.1 μM)0.55 ± 0.08 (5 μM)
 TOV112D2.20 ± 0.211.01 ± 0.031.01 ± 0.090.99 ± 0.1 (0.l μM)1.01 ± 0.09 (2.5 μM)

Effect of selenium on SELENBP1 expression

Selenium has been shown to modulate the expression of selenoprotein glutathione peroxidase and thioredoxin reductase.21, 22 Selenium has also been shown to disrupt the androgenic signaling pathway,12 which has been shown to regulate the expression level of SELENBP1. Hence, selenium effects on the SELENBP1 expression in HOSE cells and ovarian cancer cells were investigated. Selenium at the level of IC25 concentration was sufficient to significantly affect androgenic signaling in the previous report.12 Methylselenocysteine at the levels of 1 and 2.5 μM (equivalent to 79 and 197.5 ng/ml selenium, respectively) are the physiological range of selenium in serum. Therefore, each cell line was treated with methylselenocysteine in the concentration of 1 μM, 2.5 μM, IC10 and IC25. Our results showed that the physiological level of methylselenocysteine was sufficient to elevate SELENBP1 expression in immortalized HOSE cells (Table IV). However, reduced SELENBP1 expression was identified after methylselenocysteine treatment in 3 out of 4 ovarian cancer cell lines we assayed. Dose-dependent response was identified at the range of 0–6 μM for OVCA882 and DOV13, while SKOV3 has its maximal response at 2.5 μM. For HOSE2089, 0.1 μM methylselenocysteine is sufficient to induce the maximal response. In the repeating experiment with a normal HOSE primary culture, methylselenocysteine treatment also increased the protein level of SELENBP1 in a dose-dependent response (Fig. 5b).

In conjunction with the DHT experiment, the results of methylselenocysteine experiment illustrate that the SELENBP1 levels change in response to androgen and selenium treatment, which may be the general response of other androgen-regulated genes such as hK13. More importantly, our DHT and methylselenocysteine experiments suggest that there may be a different androgen/selenium-regulated pathway between normal HOSE and ovarian cancer cells.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

In this present study, we used a comparative strategy to identify potential biomarkers in membrane proteomes. In our preliminary analysis of normal ovarian surface epithelial cells and 3 ovarian cancer cell lines, we found that SELENBP1 was down-regulated in ovarian cancer cell lines. Immunohistochemical analysis of normal and tumor tissues validated the down-regulation of SELENBP1 in ovarian cancers. Survival analysis with Cox proportional hazard regression analysis and Kaplan–Meier method demonstrated that SELENBP1 was a potential prognosis predictor within the invasive ovarian cancer group. Further analysis of SELENBP1 expression in cultured ovarian epithelial cells suggested that SELENBP1 expression was regulated by selenium and androgen, and different responses were found between HOSE and ovarian cancer cells.

Selenium is an important micronutrient with novel anticancer activities.23 There are convincing epidemiologic data showing a statistically significant inverse relationship between selenium level and cancer risk. A large completed prevention trial, the Nutritional Prevention of Cancer (NPC) further strengthened the cancer preventive activity of selenium.24, 25 The outcomes and prospects of the completed or ongoing clinical trials for selenium were reviewed in a recent study.26 Therefore, the identification of SELENBP1 as one of the target genes of selenium and the close association of SELENBP1 with ovarian cancer suggest a possible involvement of SELENBP1 in the pathway of selenium-mediated anticancer activity.

Greatly reduced SELENBP1 expression levels were observed not only in ovarian cancer cells but also in immortalized HOSE cells, suggesting that loss of SELENBP1 begins during the immortalization process, an early event in tumorigenesis. Hence, the molecular mechanism responsible for the loss of SELENBP1 expression may be a key step in the transition from normal to cancerous cells.

It is intriguing that although SELENBP1 is down-regulated in the majority of ovarian cancer cases, a significant relationship between poor prognosis and high SELENBP1 expression was identified within the invasive ovarian cancer cases. Our further investigation suggests that SELENBP1 level may be an indicator for the cellular selenium level and the diagnostic and prognostic value of SELENBP1 may come from its ability to represent the selenium activity in the cells. Further investigation of the biological activity of SELENBP1 will be required to understand the underlying molecular mechanism.

Androgen receptors are expressed in the surface epithelia of ovary and most ovarian cancers,27, 28, 29 and androgen has been suggested to be a risk factor for ovarian cancer.19 Antiandrogen can inhibit the growth of ovarian cancer cells.30 Oral contraceptives, the most effective chemopreventive agent against ovarian cancer, suppress ovarian testosterone production by 35–70%.31, 32, 33, 34 Hence, the role of procancerous action of androgen in ovarian cancer is suggested. On the basis of the association of high SELENBP1 level and poor clinical outcome of patients with ovarian cancer, the level of SELENBP1 may reflect the cellular level of androgen. Therefore, androgen activity may also play some role in the diagnostic and prognostic value of SELENBP1.

In a prior study with lung cancer patients, low levels of SELENBP1 were associated with poor clinical outcome.10 These different findings may be due to tissue specificity, particularly in the context of hormonal response. Detailed characterization of the selenium and androgen responses in SELENBP1 expression with cell lines derived from different types of tissues will be required to understand whether SELENBP1 expression is regulated differently in different tissues.

In conclusion, this is the first study to report SELENBP1 as a potential predictor of unfavorable prognosis in patients with invasive ovarian cancer. Because of the poor understanding of the biological activity of SELENBP1, further functional studies will be required to understand the roles of SELENBP1 in ovarian cancer. Finally, SELENBP1 represents a unique selenium containing protein with the prognostic value in ovarian cancer. Its potential association with selenium and androgen makes it different from other known biomarkers and warrants future studies. Characterization of SELENBP1 and identification of the regulatory mechanism for its expression may advance our understanding of anticancer activity of selenium.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

We thank all the patients for their participation in the study.


  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  • 1
    Jemal A, Thomas A, Murray T, Thun M. Cancer statistics. CA Cancer J Clin 2002; 52(1): 2347.
  • 2
    Holschneider CH, Berek JS. Ovarian cancer: Epidemiology, biology, and prognostic factors. Semin Surg Oncol. 2000; 19(1): 310.
  • 3
    Kiviat NB, Critchlow CW. 2002. Novel approaches to identification of biomarkers for detection of early stage cancer. Disease Markers 18(2): 7381.
  • 4
    Jones MB, Krutzsch H, Shu H, Zhao Y, Liotta LA, Kohn EC, Petricoin EF,III. Proteomic analysis and identification of new biomarkers and therapeutic targets for invasive ovarian cancer. Proteomics 2002; 2(1): 7684.
  • 5
    Chen G, Gharib TG, Huang CC, Thomas DG, Shedden KA, Taylor JM, Kardia SL, Misek DE, Giordano TJ, Iannettoni MD, Orringer MB, Hanash SM, et al. Proteomic analysis of lung adenocarcinoma: Identification of a highly expressed set of proteins in tumors. Clin Cancer Res 2002; 8(7): 22982305.
  • 6
    Hopkins AL, Groom CR. The druggable genome. Nat Rev Drug Discov 2002; 1(9): 727730.
  • 7
    Chang PW, Tsui SK, Liew C, Lee CC, Waye MM, Fung KP. Isolation, characterization, and chromosomal mapping of a novel cDNA clone encoding human selenium binding protein. J Cell Biochem 1997; 64(2): 217224.
  • 8
    Yang M, Sytkowski AJ. Differential expression and androgen regulation of the human selenium-binding protein gene hSP56 in prostate cancer cells. Cancer Res 1998; 58(14): 31503.
  • 9
    Li LS, Kim H, Rhee H, Kim SH, Shin DH, Chung KY, Park KS, Paik YK, Chang J. Proteomic analysis distinguishes basaloid carcinoma as a distinct subtype of nonsmall cell lung carcinoma. Proteomics 2004; 4(11): 33943400.
  • 10
    Chen G, Wang H, Miller CT, Thomas DG, Gharib TG, Misek DE, Giordano TJ, Orringer MB, Hanash SM, Beer DG. Reduced selenium-binding protein 1 expression is associated with poor outcome in lung adenocarcinomas. J Pathol 2004; 202(3): 321329.
  • 11
    He QY, Cheung YH, Leung SY, Yuen ST, Chu KM, Chiu JF. Diverse proteomic alterations in gastric adenocarcinoma. Proteomics 2004; 4(10): 32763287.
  • 12
    Dong Y, Lee SO, Zhang H, Marshall J, Gao AC, Ip C. Prostate specific antigen expression is down-regulated by selenium through disruption of androgen receptor signaling. Cancer Res 2004; 64(1): 1922.
  • 13
    Tsao SW, Mok SC, Fey EG, Fletcher JA, Wan TS, Chew EC, Muto MG, Knapp RC, Berkowitz RS. Characterization of human ovarian surface epithelial cells immortalized by human papilloma viral oncogenes (HPV-E6E7 ORFs). Exp Cell Res 1995; 218(2): 499507.
  • 14
    Yates JR,III, Eng JK, McCormack AL, Schieltz D. Method to correlate tandem mass spectra of modified peptides to amino acid sequences in the protein database. Anal Chem 1995; 67(8): 14261436.
  • 15
    Yiu GK, Chan WY, Ng SW, Chan PS, Cheung KK, Berkowitz RS, Mok SC. SPARC (secreted protein acidic and rich in cysteine) induces apoptosis in ovarian cancer cells. Am J Pathol 2001; 159(2): 609622.
  • 16
    Kruskal WH, Wallis WA. Use of ranks in one-criterion variance analysis. J Am Stat Assoc 1952; 47(260): 583621.
  • 17
    Spurrier JD. On the null distribution of the Kruskal–Wallis statistic. Nonparametric Statistics 2003; 15(6): 685691.
  • 18
    Porat A, Sagiv Y, Elazar Z. A 56-kDa selenium-binding protein participates in intra-Golgi protein transport. J Biol Chem 2000; 275(19): 1445714465.
  • 19
    Risch HA. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst 1998; 90(23): 17741786.
  • 20
    Yousef GM, Chang A, Diamandis EP. Identification and characterization of KLK-L4, a new kallikrein-like gene that appears to be down-regulated in breast cancer tissues. J Biol Chem 2000; 275(16): 1189111898.
  • 21
    Hadley KB, Sunde RA. Selenium regulation of thioredoxin reductase activity and mRNA levels in rat liver. J Nutr Biochem 2001; 12(12): 693702.
  • 22
    Sunde RA, Thompson BM, Palm MD, Weiss SL, Thompson KM, Evenson JK. Selenium regulation of selenium-dependent glutathione peroxidases in animals and transfected CHO cells. Biomed Environ Sci 1997; 10(2–3): 346355.
  • 23
    Combs GF,Jr. Current evidence and research needs to support a health claim for selenium and cancer prevention. J Nutr 2005; 135(2): 343347.
  • 24
    Duffield-Lillico AJ, Reid ME, Turnbull BW, Combs GF,Jr, Slate EH, Fischbach LA, Marshall JR, Clark LC. Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomized clinical trial: a summary report of the Nutritional Prevention of Cancer Trial. Cancer Epidemiol Biomarkers Prev 2002; 11(7): 630639.
  • 25
    Clark LC, Combs GF,Jr, Turnbull BW, Slate EH, Chalker DK, Chow J, Davis LS, Glover RA, Graham GF, Gross EG, Krongrad A, Lesher JL,Jr, et al. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 1996; 276(24): 19571963.
  • 26
    Taylor PR, Greenwald P. Nutritional interventions in cancer prevention. J Clin Oncol 2005; 23(2): 333345.
  • 27
    Lau KM, Mok SC, Ho SM. Expression of human estrogen receptor-alpha and -beta, progesterone receptor, and androgen receptor mRNA in normal and malignant ovarian epithelial cells. Proc Natl Acad Sci USA 1999; 96(10): 57225727.
  • 28
    Chadha S, Rao BR, Slotman BJ, van Vroonhoven CC, van der Kwast TH. An immunohistochemical evaluation of androgen and progesterone receptors in ovarian tumors. Hum Pathol 1993; 24(1): 9095.
  • 29
    Cardillo MR, Petrangeli E, Aliotta N, Salvatori L, Ravenna L, Chang C, Castagna G. Androgen receptors in ovarian tumors: Correlation with oestrogen and progesterone receptors in an immunohistochemical and semiquantitative image analysis study. J Exp Clin Cancer Res 1998; 17(2): 231237.
  • 30
    Slotman BJ, Rao BR. Response to inhibition of androgen action of human ovarian cancer cells in vitro. Cancer Lett 1989; 45(3): 213220.
  • 31
    Browning MC, Anderson J. Effect of oral contraceptives on plasma testosterone concentration. Br Med J 1977; 1(6053): 107.
  • 32
    Murphy A, Cropp CS, Smith BS, Burkman RT, Zacur HA. Effect of low-dose oral contraceptive on gonadotropins, androgens, and sex hormone binding globulin in nonhirsute women. Fertil Steril 1990; 53(1): 3539.
  • 33
    van der Vange N, Blankenstein MA, Kloosterboer HJ, Haspels AA, Thijssen JH. Effects of seven low-dose combined oral contraceptives on sex hormone binding globulin, corticosteroid binding globulin, total and free testosterone. Contraception 1990; 41(4): 345352.
  • 34
    Kuhnz W, Sostarek D, Gansau C, Louton T, Mahler M 1990. Single and multiple administration of a new triphasic oral contraceptive to women: Pharmacokinetics of ethinyl estradiol and free and total testosterone levels in serum. Am J Obstet Gynecol 1990; 165(3): 596602.
  • 35
    Moore MK, Viselli SM. Staining and quantification of proteins transferred to polyvinylidene fluoride membranes. Anal Biochem 2000; 279(2): 241242.
  • 36
    Klein D, Kern RM, Sokol RZ. A method for quantification and correction of proteins after transfer to immobilization membranes. Biochem Mol Biol Int 1995; 36(1): 5966.