Human tissue specimens

A total of 237 human ovarian tissue samples were taken on surgical resection in the Cancer Center and the First Affiliated Hospital, Sun Yat-sen University, Guangzhou, China. These samples included 30 ovarian tissues, 30 cystadenoma, 30 borderline, and 147 ovarian cancers tissues. All the samples were made into paraffin-embedded tissue specimens after normal dehydration and dehydration processing. All patients had not received any preoperative chemo-radiation, and patients had complete clinical follow-up data available for review. Additional clinico-pathological characteristics are summarized in Table 1. We obtained written, informed consent from all patients as well as approval from the Institute Research Medical Ethics Committee of Sun Yat-sen University prior to beginning the study.

Construction of tissue microarray and immunohistochemistry

The tissue microarray was constructed with 0.6 mm cores selected from the most representative tumor areas using a manual tissue array instrument (Beecher Instruments, Silver Spring, MD, USA), according to previously described methods.9 10 Immunohistochemical analysis of SLC12A5 was performed using a standard streptavidin-biotin-peroxidase complex method that has been previously described.5 Briefly, tissue sections were deparaffinized and rehydrated. The endogenous peroxidase activity was blocked with 3% H₂O₂ for 10 min. Tissue was then subjected to antigen retrieval, during which slides were immersed in 10 mM citrate buffer (pH 8.0) and boiled for 3 min in a pressure cooker. Non-specific binding was blocked by 5% normal goat serum for 10 min. The slides were incubated with a 1:100 dilution of polyclonal antibody against SLC12A5 (Abcam) at 4°C overnight in a moist chamber. Slides were then sequentially incubated with biotinylated goat anti-mouse IgG (1:100 dilution; Santa Cruz Biotechnology) and then streptavidin-peroxidase conjugate, each for 30 min at room temperature. Isotope-matched human IgG was used in each case as a negative control. Finally, the 3, 5-diaminobenzidine (DAB) Substrate Kit (Dako) was used for color development followed by Mayer hematoxylin to provide a counterstain.

To evaluate SLC12A5 staining, its extent was scored by assigning a staining percentage to positive tumor cells (0, none; 1,<20% of positive staining cells; 2, 20–50% of positive staining cells; 3,>50% of positive staining cells). Samples with scores of 0, 1, or 2 were defined as low expression, while cases with a score of 3 were defined as having high expression,5 All histological evaluations were carried out in a double-blind manner by two expert pathologists (C-JW and DX).

Fluorescence in situ hybridization

Two-color fluorescence in situ hybridization (FISH) was used to study the genetic amplification of SLC12A5. A Spectrum BAC clone (RP11-465L10) containing the SLC12A5 gene was labeled with Spectrum Red (Vysis, Downers Grove, IL). Simultaneously, a chromosome 20 centromere probe labeled with Spectrum Green (Vysis, Downers Grove, IL) was used as an internal control. The FISH reaction was carried out as previously described with slight modifications. Briefly, deparaffinized ovarian carcinoma tissue sections were treated with proteinase K (400 µg/mL) at 37°C for 45 min, followed by denaturation in 70% formamide, 2×SSC at 75°C for 8 min. Fifty nanograms of each probe were mixed in a 20 µl hybridization mixture (containing 55% formamide, 2×SSC, and 2 µg human Cot1 DNA), denatured at 75°C for 6 min, and then hybridized to the denatured tissue sections at 37°C for 24 hours. Slides were counterstained with 1 µg/mL DAPI in an anti-fade solution and were examined with a Zeiss Axiophot microscope equipped with a triple-band pass filter. FISH signals from a minimum of 300 tumor cells were evaluated per specimen. SLC12A5 gene amplification in tumor cells was defined as either: the presence of six (or more) SLC12A5 gene signals; or more than three times as many gene signals than seen in the reference signals (Figure 1D).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Immunohistochemical staining of SLC12A5 protein and FISH assay of SLC12A5 gene in human ovarian carcinoma tissue. (A) An ovarian carcinoma (case 24) showed normal expression of SLC12A5 protein with a negative staining of SLC12A5 protein. (B) Representative image of primary ovarian carcinoma (case 11) showing low SLC12A5 protein expression. Less than 10% of tumor cells showed high SLC12A5 nuclear immunoreactivity (original magnification,×100). (C) Representative image of SLC12A5 protein overexpression in primary ovarian carcinoma (case 145). All of the tumor cells revealed positive SLC12A5 immunostaining (original magnification,×100). (a), (b), and (c) are higher magnifications of the indicated areas from the representative images presented in (A), (B), and (C), respectively (original magnification,×400). (D) Amplification of the SLC12A5 gene was observed using FISH in the same ovarian carcinoma case (case 145). SLC12A5 gene amplification (red) was at least three times stronger than the signals for chromosome 20 centromere (green) (original magnification,×1000). (E) Probability of patient survival: low expression of SLC12A5, n=92; overexpression of SLC12A5, n=55 (P<0.001). (F) FIGO stage (AUC)=0.737,P<0.001), SLC12A5 protein expression ((AUC)=0.655, P=0.002), pT ((AUC)=0.641, P=0.004), pN ((AUC)=0.601, P=0.039), pM ((AUC)=0.607, P=0.028), Pathological grading ((AUC)=0.612, P=0.021) was significantly associated with survival.

Statistical analyses

All statistical analyses were executed using SPSS software (SPSS Standard version 20.0, SPSS Inc.). A chi-square test was used to assess the association of SLC12A5 expression with clinico-pathological features of ovarian carcinoma patients. The Kaplan–Meier method was used for univariate survival analysis and a log-rank test was used to compare different survival curves. Multivariate Cox regression analysis was used to assess the overall survival power of these parameters. Receiver operating characteristic (ROC) curve analysis were used to assess differences between the clinico-pathological features and estimation of survival predictions. The P value of<0.05 was considered statistically significant.