Purpose: Patients with ovarian clear cell adenocarcinoma generally have a poor response to combination chemotherapy and have overall poorer prognosis than patients with other histologic types of ovarian cancer. Genetic changes in this group of cancer have not been thoroughly explored. Identification of these changes may provide us new therapeutic targets to treat this disease.

Experimental Design: Genomic and expression array analyses were applied on 30 clear cell ovarian cancer cases and 19 serous cases using a 10,816-element cDNA microarray platform. Further validation and clinical correlation studies were done on differentially expressed genes that are related to chemoresistance.

Results: Based on array analyses, 12 genes showed a significant increase in DNA and mRNA copy number and 5 genes showed a significant decrease in DNA and RNA copy number in clear cell tumors compared with those in the serous type. One of the genes was ABCF2, which belongs to the ATP-binding cassette gene superfamily and has been shown to amplify in other tumor types. Validation studies were done using real-time quantitative PCR and immunohistochemistry. The results showed significantly higher ABCF2 DNA and mRNA copy number and protein levels in clear cell cases compared with those in serous cases. Furthermore, in 20 clear cell cases with chemoresponse data available, ABCF2 cytoplasmic staining was significantly higher in nonresponders than that in the responders (60.0% versus 28.5%; P = 0.0002).

Conclusions: These data suggest that ABCF2 protein may be a prognostic marker for ovarian clear cell ovarian adenocarcinoma.

Ovarian cancer is the fifth most common form of cancer in women in the United States, accounting for 4% of the total number of cancer cases and 25% of those cases occur in the female genital tract. It was estimated that 14,300 deaths would be caused by ovarian cancer in the year 2005 (1). Ovarian cancer can be subdivided into four major histologic types. Among them, clear cell ovarian cancer, which constitutes 5% to 10% of ovarian cancer cases, differs from the other histologic types with respect to its clinical characteristics (2, 3). This type of tumor is thought to sometimes arise from endometriosis and many of the patients present the disease at early stages (2, 3). Clear cell type is usually more resistant to systemic chemotherapy than other types and has a worse prognosis (4, 5). In fact, in current clinical practice, patients with clear cell type ovarian cancer are treated as those with high-grade neoplasms (6). The molecular pathobiology of clear cell type ovarian cancer remains largely unknown.

Recent studies showed that 25% to 75% of clear cell type ovarian cancer showed increased DNA copy numbers on 8q11-q13, 8q21-q22, 8q23, 8q24-qter, 17q25-qter, and 20q13-qter and decreased copy number on 19p by chromosome comparative genomic hybridization (CGH; ref. 7). However, changes in the DNA copy number on the gene level have not been identified. Using oligonucleotide array expression profiling, Schwartz et al. reported that clear cell type ovarian cancer has a molecular signature that distinguishes it from other histologic types. A total of 73 genes, with expression levels 2- to 29-fold higher in clear cell type than in other histologic types, were identified (8). However, in this report, only eight clear cell specimens were included and DNA copy number changes were not analyzed. The cDNA array platform for CGH analysis is a powerful tool to identify DNA copy number changes on the gene-by-gene basis (9, 10). In addition, we can use the same platform to analyze RNA levels simultaneously. Previously, we have successfully adopted this method to identify changes in both DNA and RNA levels from microdissected tumor specimens (11). In this study, we reported the use of the same approach to identify differential DNA copy number abnormalities (CNA) and expression patterns in clear cell ovarian cancer compared with those identified in the serous type. We further validated the expression of one of the genes called ABCF2 and correlated its expression with clinical outcomes in patients with clear cell ovarian cancers.

Clinical samples. A total of 27 clear cell ovarian adenocarcinomas and 19 serous ovarian adenocarcinomas were included in this study. All patient-derived specimens were collected and archived under protocols approved by the institutional review boards of the parent institutions. Clinical data of these samples are shown in Table 1. Normal female DNA (Promega, Madison, WI) was used as the reference for CGH analysis, and RNAs isolated from a pool of 10 normal ovarian epithelial cell (human ovarian surface epithelium) cultures were used as the reference for expression profiling.

Table 1.

Clinical features of ovarian cancer cases

Case no.HistologyAge (y)Clinical stageHistologic grade
BWH-451 Clear cell adenocarcinoma 54 IV 
BWH-539 Clear cell adenocarcinoma 57 Ic 
BWH-671 Clear cell adenocarcinoma 37 IIc 
BWH-725 Clear cell adenocarcinoma 62 Ia 
BWH-805 Clear cell adenocarcinoma 67 III 
BWH-810 Clear cell adenocarcinoma 53 IIIc 
BWH-855 Clear cell adenocarcinoma 51 IIIc 
BWH-857 Clear cell adenocarcinoma 39 IIIc 
BWH-1048 Clear cell adenocarcinoma 40 IIIc 
BWH-1288 Clear cell adenocarcinoma 67 Ic 
BWH-1411 Clear cell adenocarcinoma 78 IIIc 
BWH-1484 Clear cell adenocarcinoma 51 Ib 
BWH-1640 Clear cell adenocarcinoma 49 IIIc 
BWH-1638 Clear cell adenocarcinoma 50 III 
OCGH-63 Clear cell adenocarcinoma 55 Ic 
OCGH-65 Clear cell adenocarcinoma 57 Ic 
OCGH-74 Clear cell adenocarcinoma 55 Ia 
OCGH-92 Clear cell adenocarcinoma 39 IIb 
OCGH-98 Clear cell adenocarcinoma 55 IIIc 
OCGH-108 Clear cell adenocarcinoma 60 Ic 
OCGH-110 Clear cell adenocarcinoma 52 Ic 
OCGH-113 Clear cell adenocarcinoma 62 IIc 
OCGH-120 Clear cell adenocarcinoma 54 IIIc 
OCGH-122 Clear cell adenocarcinoma 56 Ic 
OCGH-123 Clear cell adenocarcinoma 59 Ia 
OCGH-126 Clear cell adenocarcinoma 54 Ic 
BWH403 Serous cystadenocarcinoma 67 Ia 
BWH437 Serous cystadenocarcinoma 42 Ia 
BWH662 Serous cystadenocarcinoma 59 III 
BWH800 Serous cystadenocarcinoma 59 IIIc 
BWH 1231 Serous cystadenocarcinoma 56 III 
BWH715 Serous cystadenocarcinoma 61 IIIc 
BWH879 Serous cystadenocarcinoma 46 III 
BWH489 Serous cystadenocarcinoma 73 IIIb 
BWH765 Serous cystadenocarcinoma 73 III 
BWH333 Serous cystadenocarcinoma 59 Ic 
BWH398 Serous cystadenocarcinoma 52 Ic 
BWH416 Serous cystadenocarcinoma 42 Ia 
BWH484 Serous cystadenocarcinoma 54 Ib 
BWH565 Serous cystadenocarcinoma 68 Ic 
BWH627 Serous cystadenocarcinoma 58 Ia 
BWH689 Serous cystadenocarcinoma 43 Ic 
BWH808 Serous cystadenocarcinoma 57 Ia 
OCGH-130 Serous cystadenocarcinoma 55 II 
Case no.HistologyAge (y)Clinical stageHistologic grade
BWH-451 Clear cell adenocarcinoma 54 IV 
BWH-539 Clear cell adenocarcinoma 57 Ic 
BWH-671 Clear cell adenocarcinoma 37 IIc 
BWH-725 Clear cell adenocarcinoma 62 Ia 
BWH-805 Clear cell adenocarcinoma 67 III 
BWH-810 Clear cell adenocarcinoma 53 IIIc 
BWH-855 Clear cell adenocarcinoma 51 IIIc 
BWH-857 Clear cell adenocarcinoma 39 IIIc 
BWH-1048 Clear cell adenocarcinoma 40 IIIc 
BWH-1288 Clear cell adenocarcinoma 67 Ic 
BWH-1411 Clear cell adenocarcinoma 78 IIIc 
BWH-1484 Clear cell adenocarcinoma 51 Ib 
BWH-1640 Clear cell adenocarcinoma 49 IIIc 
BWH-1638 Clear cell adenocarcinoma 50 III 
OCGH-63 Clear cell adenocarcinoma 55 Ic 
OCGH-65 Clear cell adenocarcinoma 57 Ic 
OCGH-74 Clear cell adenocarcinoma 55 Ia 
OCGH-92 Clear cell adenocarcinoma 39 IIb 
OCGH-98 Clear cell adenocarcinoma 55 IIIc 
OCGH-108 Clear cell adenocarcinoma 60 Ic 
OCGH-110 Clear cell adenocarcinoma 52 Ic 
OCGH-113 Clear cell adenocarcinoma 62 IIc 
OCGH-120 Clear cell adenocarcinoma 54 IIIc 
OCGH-122 Clear cell adenocarcinoma 56 Ic 
OCGH-123 Clear cell adenocarcinoma 59 Ia 
OCGH-126 Clear cell adenocarcinoma 54 Ic 
BWH403 Serous cystadenocarcinoma 67 Ia 
BWH437 Serous cystadenocarcinoma 42 Ia 
BWH662 Serous cystadenocarcinoma 59 III 
BWH800 Serous cystadenocarcinoma 59 IIIc 
BWH 1231 Serous cystadenocarcinoma 56 III 
BWH715 Serous cystadenocarcinoma 61 IIIc 
BWH879 Serous cystadenocarcinoma 46 III 
BWH489 Serous cystadenocarcinoma 73 IIIb 
BWH765 Serous cystadenocarcinoma 73 III 
BWH333 Serous cystadenocarcinoma 59 Ic 
BWH398 Serous cystadenocarcinoma 52 Ic 
BWH416 Serous cystadenocarcinoma 42 Ia 
BWH484 Serous cystadenocarcinoma 54 Ib 
BWH565 Serous cystadenocarcinoma 68 Ic 
BWH627 Serous cystadenocarcinoma 58 Ia 
BWH689 Serous cystadenocarcinoma 43 Ic 
BWH808 Serous cystadenocarcinoma 57 Ia 
OCGH-130 Serous cystadenocarcinoma 55 II 

Microdissection and RNA and DNA extraction. Microdissection was done as described previously (11). Approximately 15,000 tumor cells were used in each case. For DNA extraction, dissected cells were collected into 50 μL cell lysis buffer [1× expand high-fidelity buffer (Boehringer Mannheim, Mannheim, Germany) containing 4 mg/mL proteinase K] and incubated for 72 hours at 55°C. The proteinase K was inactivated by heating at 95°C for 10 minutes before PCR. Total RNA extraction was done using the RNeasy kit according to the manufacturer's instructions (Qiagen, Valencia, CA).

DNA and RNA amplification. DNA amplification was done using the DOP-PCR master kit (Roche, Indianapolis, IN) according to the manufacturer's instructions. PCR was done in a ABI 9600 thermocycler (Applied Biosystems, Foster City, CA). RNA was amplified using a modified single-round T7 RNA amplification protocol. In brief, total RNA (600 ng) was first incubated with 1 μL T7 primer (5′-GCATTAGCGGCCGCGAAATTAATACGACTCACTATAGGGAGATTTTTTTTTTTTTTTTTTVN-3′, 200 ng/μL) in a total volume of 50 μL for 3 minutes at 70°C. First-strand cDNA synthesis was then done by incubating 5 μL primer annealed sample and 5 μL first-strand master mix containing 2 μL of 5× first-strand buffer, 1 μL of 0.1 mol/L DTT, 0.5 μL DEPC water, 0.5 μL of 10 mmol/L deoxynucleotide triphosphate mix, 0.5 μL RNase inhibitor, and 0.5 μL Moloney murine leukemia virus (200 units/μL) for 1 hour and 15 minutes at 37°C. Subsequently, second-strand cDNA synthesis was done by incubating the10 μL first-strand reaction with 65 μL second master mix, which contained 46 μL DEPC water, 15 μL of 5× second-strand buffer, 1.5 μL of 10 mmol/L deoxynucleotide triphosphate mix, 0.5 μL Escherichia coli DNA ligase (10 units/μL), 1.5 μL E. coli DNA polymerase I (10 units/μL), and 0.5 μL E. coli RNase H (2 units/μL) for 2 hours at 16°C and then for 15 minutes at 70°C. The entire 75 μL cDNA sample was loaded onto a ChromaSpin TE-200 spin column (BD Biosciences, San Diego, CA), which was centrifuged for 5 minutes at 2,900 rpm (700 × g) in an Eppendorf centrifuge. Purified cDNA collected was lyophilized, dissolved in 8 μL RNase-free water, and incubated at 70°C for 10 minutes. In vitro transcription was subsequently done by incubating the 8 μL postlyophilization cDNA product with 12.2 μL master mix containing 2 μL of 10× T7 reaction buffer, 6 μL of 25 mmol/L rNTP mix, 2 mL of 100 mmol/L DTT, 0.2 μL RNase inhibitor (40 units/mL), and 2 μL T7 RNA polymerase for 3 hours at 37°C. The amplified RNA was purified on a RNeasy Mini column (Qiagen) according to the manufacturer's protocol. The purified amplified RNA was quantified by the RiboGreen RNA Quantitation Reagent (Molecular Probes, Eugene, OR).

Microarray analyses. Tumor DNAs (1.5 μg) and normal female control DNAs were labeled with fluorescein-12-dCTP and biotin-11-dCTP, respectively, using the Random Primers DNA Labeling System (Invitrogen, Carlsbad, CA) as described (11). Amplified RNA (1 μg) prepared from normal human ovarian surface epithelium cells and microdissected tumor tissue specimens were labeled with the MicroMax TSA labeling and detection system (Perkin-Elmer, Boston, MA) as described (12). The relative fluorescent level or fluorescent ratio, which represents the relative amount of target sequences in the probe mix, is analyzed by comparing the fluorescent intensity of corresponding individual spots after local background subtraction and normalization. The average local background and SD over all the array spots were also calculated. A 1.4-fold cutoff for amplification and a 0.7-fold cutoff for deletion, which could achieve a 95% specificity in identifying differential DNA CNAs, were used as described previously (11).

Real-time quantitative PCR. Both DNA and mRNA copy numbers were validated using TaqMan real-time PCR amplification with TaqMan Universal PCR Master Mix or TaqMan One-Step RT-PCR Master Mix and an ABI Prism 7000 Sequence Detection System (Applied Biosystems). All results were normalized to the amount of β-actin DNA or rRNA (Applied Biosystems). Primers and probes used were as follows: β-actin, probe: 6FAM-CTACGAGCTGCCTGACGGCCAGG-TAMRA, primers: forward 5′-GATGGCCACGGCTGCTT and reverse 5′-ACCGCTCATTGCCAATGG. TaqMan rRNA control reagents (Applied Biosystems) were used as an internal control for DNA validation. Primers and probes used for each analysis were as follows: ABCF2, probe: 6FAM-CCTCGCGGATCTTGCATGGACTG-TAMRA, primers: forward 5′-GGAGCTGGATGCCGACAA-3′ and reverse 5′-CTGCATGGCAGGTGTGAAAC-3′. For DNA, amplification was done using 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. For RNA, amplification was done using 48°C for 30 minutes, 95°C for 10 minutes followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. DNA and mRNA quantification was assessed by the fluorescence intensity emitted after PCR reaction. The difference in the fluorescence between tubes with both the internal control amplification and the test amplification was compared for tumor and normal control samples.

Establishment of a polyclonal anti-ABCF2 antibody. The polyclonal anti-ABCF2 antibody was generated by injecting the purified full-length ABCF2 fusion protein into two rabbits. The specificity of the antibody was determined by Western blot and immunohistochemical analyses on human embryonic kidney cells (293T) transfected with an expression vector pcDNA3.1 or with the vector containing either a full-length ABCF2 coding sequence or a full-length ABCF2 coding sequence with a Myc and a His tag. In brief, the open reading frame encoding the human ABCF2 gene was amplified from the pET28a(+) containing ABCF2 using the forward primer 5′-AATAGGATCCACCATGCCCTCCGACCTGGC-3′ and the reverse primer 5′-AATAACTAGTCACGTTGTGGGTCCTCTTGG-3′. The PCR product was ligated in frame into the BamHI and SpeI sites of the mammalian expression vector pcDNA3.1/Myc-His A (Invitrogen), which encodes a COOH-terminal Myc epitope and six His polypeptides. To exclude the His tag, the reverse primer 5′-AATAACTAGTTCACACGTTGTGGGTCCTCTTG-3′, including the stop codon in front of SpeI sites, was designed. The pcDNA3.1 vector alone or vectors containing the ABCF2 sequence and the Myc-His-tagged ABCF2 sequence, pcDNA3.1/ABCF2 and pcDNA3.1/ABCF2mH, respectively, were then transiently transfected in human embryonic kidney cells (293T) grown in DMEM supplemented with 10% FCS, 1 unit/mL penicillin, and 1 μg/mL streptomycin using LipofectAMINE (Invitrogen). After 3 days, cells were lysed according to the method of Laemmli (13). SDS-PAGE was done with a 12.5% polyacrylamide gel. Electrophoresed proteins were transferred to a polyvinylidene difluoride membrane (Millipore, Bedford, MA). After blocking with PBS containing 5% bovine serum albumin and 0.1% NaN3, the membrane was reacted with an anti-human ABCF2 polyclonal antibody or a mouse anti-Myc antibody (MBL, Nagoya, Japan) for 1 hour at room temperature. Peroxidase-conjugated anti-rabbit IgG or peroxidase-conjugated anti-mouse IgG (MBL) was then added and incubated for 1 hour at room temperature. Finally, the enzyme activity was detected using the enhanced chemiluminescence Western blotting detection reagents (Amersham Pharmacia Biotech, Piscataway, NJ).

For immunolocalization of ABCF2 protein in vitro, cells transfected with different constructs as described above in air-dried cytocentrifuge preparations were fixed in 4% paraformaldehyde at room temperature for 15 minutes. After pretreatment with 0.2% Triton X-100, cells were incubated with a rabbit anti-ABCF2 polyclonal antibody or normal rabbit serum at room temperature for 30 minutes. Cells were washed with PBS and incubated with biotin-labeled anti-rabbit IgG (American Gualex, San Clemente, CA) at room temperature for 30 minutes. After washing with PBS, cells were incubated with phycoerythrin-conjugated streptavidin (Beckman Coulter, Fullerton, CA) at room temperature for 30 minutes. Cells were then washed with PBS and water and were mounted in ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole (Invitrogen). In addition, cells were also incubated with a mouse anti-Myc tag monoclonal antibody (MBL) or normal mouse IgG1 at room temperature for 30 minutes. Cells were washed with PBS and incubated with phycoerythrin-conjugated anti-mouse IgG (Beckman Coulter) at room temperature for 30 minutes. Cells were then washed with PBS and water and were mounted in ProLong Gold antifade reagent with 4′,6-diamidino-2-phenylindole. Cells were examined under a Leica DMIRE2 fluorescent microscope.

Immunohistochemistry. Immunolocalization of the ABCF2 protein was done using a polyclonal anti-ABCF2 antibody generated by injecting the purified full-length ABCF2 fusion protein into the rabbits as described above. In brief, histologic sections (4 μm) were affixed to glass slides, dewaxed, and rehydrated. The sections were then incubated in 3% hydrogen peroxide for 10 minutes at room temperature to quench endogenous peroxidase activity. The sections were reacted with the ABCF2 antibody at 4°C overnight. The peroxidase activity for all proteins was visualized by applying diaminobenzidine chromogen containing 0.05% hydrogen peroxide for 2 to 10 minutes at room temperature. The sections were then counterstained with hematoxylin. The slides were read by two independent pathologists, who were blinded to the clinical background of the patients. Positive cells were counted for ABCF2 protein in nuclear or cytoplasm by examining at least 1,000 tumor cells. Levels of ABCF2 were scored based on the percentage of cells with positive nuclear or cytoplasmic staining.

Evaluation of ABCF2 expression and chemotherapy response in patients with clear cell adenocarcinoma. Chemorespond data from a total of 20 patients with clear cell ovarian cancer were available. Among them, 17 cases had at least one measurable tumor lesion documented radiographically after primary surgery. The remaining three cases were recurrent cases who had at least one measurable tumor lesion documented radiographically. Tumor response was evaluated according to WHO criteria. The response was assessed from the product of the two largest perpendicular diameters using the following criteria: complete response was defined as the disappearance of all detectable lesions with no new lesions for at least 4 weeks; partial response was defined as ≧50% reduction of the sum of the products of measurable lesions for at least 4 weeks. Progressive disease was defined as a ≧25% increase in the sum of the products of all measurable lesions, reappearance of any lesion that had disappeared, or appearance of a new lesion. No change was defined as any outcome that did not qualify as response or progression. Both complete response and partial response patients were defined as responder.

Statistical analysis. Comparison of the mean of log expression between clear cell type and serous type was done by unpaired t test at the 3% level of significance. When the test statistic of a gene was positive and significant, that gene was defined as the overexpressed gene. When the test statistic of a gene was negative and significant, that gene was defined as the underexpressed gene. Quantitative PCR data and immunochemistry scores were compared using nonparametric Mann-Whitney U test. The level of critical significance was considered to be P < 0.05. Significances in the correlation between array and quantitative PCR data and between DNA and mRNA copy numbers were determined by Spearman's correlation analysis.

Identification of amplification and overexpression of ABCF2 in clear cell tumors by cDNA microarray. Using the same cDNA array platform, we first identified genes with CNA, which significantly overrepresented in clear cell tumors compared with those in the serous tumors. Further selection of genes that were simultaneously overexpressed relative to that in human ovarian surface epithelium was done. A total of 12 genes were identified (Table 2A). Subsequently, we identified genes with CNA, which significantly underrepresented in clear cell tumors compared with those in the serous tumors. Further selection of genes that were simultaneously underexpressed relative to that in human ovarian surface epithelium was done. A total of five genes were identified (Table 2B). Using the same approach, we next identified genes with CNA, which significantly overrepresented in serous tumors compared with those in clear cell tumors. Further selection of genes that were simultaneously overexpressed relative to that in human ovarian surface epithelium was done. A total of eight genes were identified (Table 2A). Finally, we identified genes with CAN, which significantly underrepresented in serous tumors compared with those in clear cell tumors and a total of 15 genes were identified (Table 2B).

Table 2.

Gene nameGene symbolMapAccession no.CGH analysis ratioExpression analysis t (P)
A. Amplified and overexpressed genes      
    Clear cell adenocarcinoma      
        Neurogranin (protein kinase C substrate, RC3) NRGN 11q24 NM_006176 3.03 2.78 (0.004) 
        ADP-ribosyltransferase 3 ART3 4p15.1-p14 NM_001179 2.68 3.37 (0.01) 
        ATP-binding cassette, subfamily F (GCN20) ABCF2 7q35-q36 NM_007189 2.58 2.21 (0.02) 
        Zinc finger protein 207 ZNF207 17q11.2 NM_003457 2.27 3.55 (<0.001) 
        Cytochrome P450, subfamily VII, subfamily CYP7B1 8q21.3 NM_004820 1.65 2.46 (0.001) 
        KIAA1140 protein KIAA1140 2p21 AK_046740 1.59 2.15 (0.002) 
        Leucine zipper–like transcriptional regulator 1 LZTR1 22q11.21 NM_006767 1.48 4.94 (<0.001) 
        Hypoxia-inducible protein 2 HIG2 7q32.2 NM_013332 1.48 2.11 (0.021) 
        Kinesin family member 5A KIF5A 12q13 NM_004984 1.47 2.88 (0.004) 
        ATPase, Na+/K+ transporting, β3 polypeptide ATP1B3 3q22-q23 NM_001679 1.43 2.18 (0.018) 
        Transcription factor 21 TCF21 6q23 NM_198392 1.42 2.62 (0.007) 
        KIAA0955 protein KIAA0955 19q13.33 AB_023172 1.41 2.12 (0.002) 
    Serous cystadenocarcinomas      
        Interleukin-15 receptor, α IL15RA 10p15-p14 NM_002189 2.20 −2.23 (0.017) 
        Hepatocyte nuclear factor 3, α HNF3A 14q12-q13 NM_004496 1.93 −3.06 (0.002) 
        Small proline-rich protein SPRK SPRK 1q21 AW_135226 1.49 −2.50 (0.009) 
        Cysteine-rich protein 1 (intestinal) CRIP1 7q11.23 NM_001311 1.44 −2.93 (0.003) 
        Podocalyxin-like PODXL 7q32-q33 NM_005397 1.42 −2.95 (0.003) 
        GCIP-interacting protein p29 p29 1p36.13-p35.1 NM_015484 1.42 −2.61 (0.007) 
        Neuron-specific protein D4S234E 4p16.3 NM_014392 1.41 −2.21 (0.017) 
        DKFZP564D177 protein NIPSNAP3A 9q31 NM_015469 1.41 −2.71 (0.005) 
B. Deleted and underexpressed genes      
    Clear cell adenocarcinoma      
        Kallikrein 6 (neurosin, zyme) KLK6 19q13.3 NM_0002774 0.63 −3.22 (0.002) 
        KIAA0575 gene product GREB1 2p25.1 NM_148903 0.69 −2.90 (0.003) 
        ESTs — 12q24.1 — 0.69 −3.72 (<0.001) 
        Prepronociceptin PNOC 8q21 NM_006228 0.71 −3.87 (<0.001) 
        ESTs — 19q13.4 — 0.71 −3.30 (0.001) 
    Serous cystadenocarcinomas      
        Fibrinogen, A α polypeptide FGA 4q28 NM_000508 0.09 2.28 (0.015) 
        Meningioma expressed antigen 6 MGEA6 14q13.3 NM_203356 0.12 4.07 (<0.001) 
        Aldo-keto reductase family 1, member C4 AKR1C4 10p15-p14 NM_001818 0.12 3.97 (<0.001) 
        Ligase III, DNA, ATP dependent LIG3 17q11.2-q12 NM_013975 0.13 3.66 (0.001) 
        Methyl-CpG-binding domain protein 1 MBD1 18q21 NM_015846 0.18 2.71 (0.005) 
        Retinol-binding protein 4, plasma RBP4 10q23-q24 NM_006744 0.20 3.28 (0.001) 
        Insulin-like growth factor 1 receptor IGF1R 15q26.3 NM_000875 0.22 2.62 (0.007) 
        Insulin-degrading enzyme IDE 10q23-q25 NM_004969 0.24 2.21 (0.017) 
        Glypican 3 GPC3 Xq26.1 NM_004484 0.26 3.01 (0.003) 
        Spectrin, β, nonerythrocytic 1 SPTBN1 2p21 NM_003128 0.28 322 (0.001) 
        LIM and senescent cell antigen-like domains 1 LIMS1 2q12.3 NM_004987 0.29 2.22 (0.017) 
        Mutated in colorectal cancers MCC 5q21-q22 NM_002387 0.29 2.11 (0.022) 
        Carboxypeptidase M CPM 12q15 NM_001874 0.29 3.16 (0.002) 
        Hexosaminidase A (α polypeptide) HEXA 15q23-q24 NM_000520 0.29 2.40 (0.01) 
        Homo sapiens mRNA; cDNA DKFZp434K152 DKFZp434K152 1p34.2 AL_157461 0.30 2.42 (0.01) 
Gene nameGene symbolMapAccession no.CGH analysis ratioExpression analysis t (P)
A. Amplified and overexpressed genes      
    Clear cell adenocarcinoma      
        Neurogranin (protein kinase C substrate, RC3) NRGN 11q24 NM_006176 3.03 2.78 (0.004) 
        ADP-ribosyltransferase 3 ART3 4p15.1-p14 NM_001179 2.68 3.37 (0.01) 
        ATP-binding cassette, subfamily F (GCN20) ABCF2 7q35-q36 NM_007189 2.58 2.21 (0.02) 
        Zinc finger protein 207 ZNF207 17q11.2 NM_003457 2.27 3.55 (<0.001) 
        Cytochrome P450, subfamily VII, subfamily CYP7B1 8q21.3 NM_004820 1.65 2.46 (0.001) 
        KIAA1140 protein KIAA1140 2p21 AK_046740 1.59 2.15 (0.002) 
        Leucine zipper–like transcriptional regulator 1 LZTR1 22q11.21 NM_006767 1.48 4.94 (<0.001) 
        Hypoxia-inducible protein 2 HIG2 7q32.2 NM_013332 1.48 2.11 (0.021) 
        Kinesin family member 5A KIF5A 12q13 NM_004984 1.47 2.88 (0.004) 
        ATPase, Na+/K+ transporting, β3 polypeptide ATP1B3 3q22-q23 NM_001679 1.43 2.18 (0.018) 
        Transcription factor 21 TCF21 6q23 NM_198392 1.42 2.62 (0.007) 
        KIAA0955 protein KIAA0955 19q13.33 AB_023172 1.41 2.12 (0.002) 
    Serous cystadenocarcinomas      
        Interleukin-15 receptor, α IL15RA 10p15-p14 NM_002189 2.20 −2.23 (0.017) 
        Hepatocyte nuclear factor 3, α HNF3A 14q12-q13 NM_004496 1.93 −3.06 (0.002) 
        Small proline-rich protein SPRK SPRK 1q21 AW_135226 1.49 −2.50 (0.009) 
        Cysteine-rich protein 1 (intestinal) CRIP1 7q11.23 NM_001311 1.44 −2.93 (0.003) 
        Podocalyxin-like PODXL 7q32-q33 NM_005397 1.42 −2.95 (0.003) 
        GCIP-interacting protein p29 p29 1p36.13-p35.1 NM_015484 1.42 −2.61 (0.007) 
        Neuron-specific protein D4S234E 4p16.3 NM_014392 1.41 −2.21 (0.017) 
        DKFZP564D177 protein NIPSNAP3A 9q31 NM_015469 1.41 −2.71 (0.005) 
B. Deleted and underexpressed genes      
    Clear cell adenocarcinoma      
        Kallikrein 6 (neurosin, zyme) KLK6 19q13.3 NM_0002774 0.63 −3.22 (0.002) 
        KIAA0575 gene product GREB1 2p25.1 NM_148903 0.69 −2.90 (0.003) 
        ESTs — 12q24.1 — 0.69 −3.72 (<0.001) 
        Prepronociceptin PNOC 8q21 NM_006228 0.71 −3.87 (<0.001) 
        ESTs — 19q13.4 — 0.71 −3.30 (0.001) 
    Serous cystadenocarcinomas      
        Fibrinogen, A α polypeptide FGA 4q28 NM_000508 0.09 2.28 (0.015) 
        Meningioma expressed antigen 6 MGEA6 14q13.3 NM_203356 0.12 4.07 (<0.001) 
        Aldo-keto reductase family 1, member C4 AKR1C4 10p15-p14 NM_001818 0.12 3.97 (<0.001) 
        Ligase III, DNA, ATP dependent LIG3 17q11.2-q12 NM_013975 0.13 3.66 (0.001) 
        Methyl-CpG-binding domain protein 1 MBD1 18q21 NM_015846 0.18 2.71 (0.005) 
        Retinol-binding protein 4, plasma RBP4 10q23-q24 NM_006744 0.20 3.28 (0.001) 
        Insulin-like growth factor 1 receptor IGF1R 15q26.3 NM_000875 0.22 2.62 (0.007) 
        Insulin-degrading enzyme IDE 10q23-q25 NM_004969 0.24 2.21 (0.017) 
        Glypican 3 GPC3 Xq26.1 NM_004484 0.26 3.01 (0.003) 
        Spectrin, β, nonerythrocytic 1 SPTBN1 2p21 NM_003128 0.28 322 (0.001) 
        LIM and senescent cell antigen-like domains 1 LIMS1 2q12.3 NM_004987 0.29 2.22 (0.017) 
        Mutated in colorectal cancers MCC 5q21-q22 NM_002387 0.29 2.11 (0.022) 
        Carboxypeptidase M CPM 12q15 NM_001874 0.29 3.16 (0.002) 
        Hexosaminidase A (α polypeptide) HEXA 15q23-q24 NM_000520 0.29 2.40 (0.01) 
        Homo sapiens mRNA; cDNA DKFZp434K152 DKFZp434K152 1p34.2 AL_157461 0.30 2.42 (0.01) 

NOTE: t = [log2(clear cell) − log2(serous)] / SE[log2(clear cell) − log2(serous)].

Because clear cell tumors in general have a more chemoresistant phenotype compared with the serous type, we therefore selected genes that are have been shown to be related to chemoresistance for further validation. ABCF2, a member of the ATP-binding cassette family, was selected because overexpression and amplification of several members of this family have been shown to correlate with chemoresistance in different cancer types (14). Array CGH analysis showed a significantly higher fold increase in CNA in clear cell tumors than in serous tumors (P = 0.016; Table 2; Fig. 1A). Expression profiling also showed a significantly higher level of ABCF2 expression in clear cell tumors than in the serous type (P = 0.02; Fig. 1B).

Fig. 1.

A, array CGH profile of chromosome 7 in serous cystadenocarcinoma and in clear cell adenocarcinoma. B, expression profile of genes located chromosome 7 in serous cystadenocarcinoma and in clear cell adenocarcinoma.

Fig. 1.

A, array CGH profile of chromosome 7 in serous cystadenocarcinoma and in clear cell adenocarcinoma. B, expression profile of genes located chromosome 7 in serous cystadenocarcinoma and in clear cell adenocarcinoma.

Close modal

Quantitative PCR and quantitative reverse transcription-PCR were used to validate the microarray data. Significantly higher fold increase in CNA was identified in clear cell tumors than in serous tumors (P < 0.001; Fig. 2A). When 1.5 for amplification and 0.5 for deletion were used as cutoffs, ABCF2 gene was amplified in 57.1% (16 of 28) of clear cell types and 11.1% (2 of 18) of serous type (P = 0.002). Relative gene expression of ABCF2 was significantly higher in clear cell type than it in serous type (P < 0.001; Fig. 2B).

Fig. 2.

A, comparison of DNA copy number of ABCF2 gene between clear cell adenocarcinoma and serous cystadenocarcinoma. B, comparison of mRNA expression of ABCF2 between clear cell adenocarcinoma and serous cystadenocarcinoma. The box is bounded above and below by the 75th and 25th percentiles, and the median is the line in the box. Whiskers are drawn to the nearest value not beyond a standard span from the quartiles; points beyond (outliers) are drawn individually, where the standard span is 1.5 × (interquartile range).

Fig. 2.

A, comparison of DNA copy number of ABCF2 gene between clear cell adenocarcinoma and serous cystadenocarcinoma. B, comparison of mRNA expression of ABCF2 between clear cell adenocarcinoma and serous cystadenocarcinoma. The box is bounded above and below by the 75th and 25th percentiles, and the median is the line in the box. Whiskers are drawn to the nearest value not beyond a standard span from the quartiles; points beyond (outliers) are drawn individually, where the standard span is 1.5 × (interquartile range).

Close modal

The relationship between DNA and RNA levels of ABCF2 identified by the array and the quantitative PCR platform was evaluated. There was a significant correlation between CNA and expression levels of ABCF2 in the microarray analysis (r = 0.418; P = 0.027). Quantitative PCR also showed a significant correlation between DNA and RNA copy number (r = 0.426; P = 0.006).

Immunolocalization of ABCF2. To confirm the specificity of the anti-ABCF2 antibody, Western blot analysis was done on cell lysates prepared from 293T cells transfected with pcDNA3.1 vector alone or with vector containing the ABCF2 sequence and the Myc-His-tagged ABCF2 sequence, pcDNA3.1/ABCF2 and pcDNA3.1/ABCF2mH, respectively. The results showed that the anti-ABCF2 antibody recognized a single 70- to 75-kDa protein band in 293 cells transfected with pcDNA3.1/ABCF2 and a band with slightly higher molecular weight in cells transfected with pcDNA3.1/ABCF2mH. Furthermore, when the anti-Myc antibody was used, a protein band was only observed in cells transfected with pcDNA3.1/ABCF2mH (Fig. 3A). Immunohistochemical analysis showed strong cytoplasm ABCF2 staining as well as weak nuclear staining in cells transfected with the pcDNA3.1/ABCF2 and the pcDNA3.1/ABCF2mH constructs compared with the mock transfectants using the anti-ABCF2 antibody. Furthermore, strong cytoplasm Myc staining and weak nuclear staining were observed only in cells transfected with pcDNA3.1/ABCF2mH using the anti-Myc tag antibody (Fig. 3B). These data strongly suggest that the anti-ABCF2 antibody is specific to the ABCF2 protein, which is predominantly located in the cytoplasm of the cells.

Fig.3.

A, Western blot analyses on cell lysates prepared from 293T wild-type (lane I) or 293T transfected by pcDNA3.1 carrying ABCF2 (lane II) or Myc-His-tagged ABCF2 (lane III) using an anti-ABCF2 polyclonal antibody or an anti-Myc tag monoclonal antibody. B, immunolocalization of ABCF2 and Myc tag proteins in 293T cells transfected with pcDNA3.1 vector alone or with vectors containing ABCF2 or Myc-His-tagged ABCF2. C, immunolocalization of ABCF2 protein in normal ovaries, endometriotic cysts, and ovarian cancer tissue. a, negative immunostaining of ABCF2 in the surface epithelium of a normal ovary (arrowhead). b, negative immunostaining of ABCF2 in the epithelial lying of an endometriotic cyst (arrowhead). c, positive immunostaining of ABCF2 in serous cystadenocarcinoma. d, negative immunostaining of ABCF2 in serous cystadenocarcinoma. e, positive immunostaining of ABCF2 in clear cell adenocarcinoma. f, negative immunostaining of ABCF2 in clear cell adenocarcinoma. Bar, 50 μm. D, Comparison of protein expression of ABCF2 between clear cell adenocarcinoma and serous cystadenocarcinoma. The box is bounded above and below by the 75th and 25th percentiles, and the median is the line in the box. Whiskers are drawn to the nearest value not beyond a standard span from the quartiles; points beyond outliers are drawn individually, where the standard span is 1.5 × (interquartile range).

Fig.3.

A, Western blot analyses on cell lysates prepared from 293T wild-type (lane I) or 293T transfected by pcDNA3.1 carrying ABCF2 (lane II) or Myc-His-tagged ABCF2 (lane III) using an anti-ABCF2 polyclonal antibody or an anti-Myc tag monoclonal antibody. B, immunolocalization of ABCF2 and Myc tag proteins in 293T cells transfected with pcDNA3.1 vector alone or with vectors containing ABCF2 or Myc-His-tagged ABCF2. C, immunolocalization of ABCF2 protein in normal ovaries, endometriotic cysts, and ovarian cancer tissue. a, negative immunostaining of ABCF2 in the surface epithelium of a normal ovary (arrowhead). b, negative immunostaining of ABCF2 in the epithelial lying of an endometriotic cyst (arrowhead). c, positive immunostaining of ABCF2 in serous cystadenocarcinoma. d, negative immunostaining of ABCF2 in serous cystadenocarcinoma. e, positive immunostaining of ABCF2 in clear cell adenocarcinoma. f, negative immunostaining of ABCF2 in clear cell adenocarcinoma. Bar, 50 μm. D, Comparison of protein expression of ABCF2 between clear cell adenocarcinoma and serous cystadenocarcinoma. The box is bounded above and below by the 75th and 25th percentiles, and the median is the line in the box. Whiskers are drawn to the nearest value not beyond a standard span from the quartiles; points beyond outliers are drawn individually, where the standard span is 1.5 × (interquartile range).

Close modal

Using the anti-ABCF2 antibody, immunolocalization of ABCF2 protein was then done on 5 normal ovaries, 5 endometriotic cysts, 22 clear cell cases, and 15 serous cases. Epithelial cells lying the surface of all the normal ovaries and endometriotic cysts had undetectable ABCF2 staining in both the nuclei and the cytoplasm. In 86.4% (19 of 22) of clear cell cases, ABCF2 immunoreactivity was detected in both nuclei and cytoplasm of cancer cells (Fig. 3C). In 13.6% (3 of 22) of clear cell cases, ABCF2 protein was negative in both nuclei and cytoplasm of cancer cells. In 20% (3 of 15) of serous cell cases, ABCF2 immunoreactivity was detected in both nuclei and cytoplasm of cancer cells. In 46.7% (7 of 15) of serous cell cases, ABCF2 protein staining was negative. In 26.7% (4 of 15) of serous cell cases, ABCF2 protein was detected only in nuclei of cancer cells, and in 6.7% (1 of 15) of serous cell cases, ABCF2 protein was detected only in the cytoplasm of cancer cells.

At least 1,000 tumor cells were examined for positive nuclear or cytoplasm ABCF2 protein staining. There was no significant difference in the number cells with positive nuclear ABCF2 staining between clear cell and serous cases (P = 0.181). However, levels of cytoplasmic ABCF2 staining were significantly higher in the clear cell cases compared with that in serous cases (P < 0.001; Fig. 3D).

In addition, the relationship between ABCF2 protein expression and both DNA and RNA copy number was determined. Both DNA copy number and relative mRNA expression significantly correlated with levels of ABCF2 protein cytoplasmic expression (r = 0.483; P = 0.005 and r = 504; P = 0.006, respectively).

Relationship between ABCF2 expression and chemotherapy response. ABCF2 protein levels in 20 patients were used to correlate with chemotherapy response. Median age was 55 years (range, 37-82 years). Sixteen cases were stage III, 1 case was stage IV, and 3 cases were recurrent cases. Seventeen primary cases had at least one measurable tumor lesion documented radiographically after primary surgery. Three recurrent cases had at least one measurable tumor lesion documented radiographically after second cytoreductive surgery. All patients received platinum-based chemotherapy. The relationship between ABCF2 expression and chemotherapy response was shown in Fig. 4. ABCF2 cytoplasmic staining was significantly higher in nonresponder than that in responder [60.0% (95% CI, 51.7-68.3) versus 28.5% (95% CI, 18.7-38.3); P < 0.0001]. However, there was no significant relationship between ABCF2 nuclear staining and chemotherapy response [48.0% (95% CI, 22.0-74.0) versus 35.5% (95% CI, 14.5-56.5)].

Fig. 4.

Comparison of ABCF2 protein expression between responders and nonresponders.

Fig. 4.

Comparison of ABCF2 protein expression between responders and nonresponders.

Close modal

Clear cell adenocarcinoma is usually more resistant to systemic chemotherapy than other histologic types of ovarian cancer and patients with clear cell type ovarian cancer have worst prognosis (4, 5). In fact, in current clinical practice, clear cell ovarian cancers are treated as high-grade neoplasms (6). The overall clinical response rate to platinum-based chemotherapy was reported to be 72.5% in patients with serous type; however, only 11.1% of the patients with clear cell type responded to platinum-based chemotherapy (15). Schwartz et al. reported that GPX3, GLRX, and SOD2 were highly expressed in clear cell type ovarian cancer. They suggested that high levels of these proteins and perhaps other antioxidant proteins in clear cell tumors may render these tumors to be more resistant to chemotherapy (8). One of the important mechanisms of drug resistance is a decrease in the accumulation of the drug caused by enhanced drug efflux mediated by ABC transporter. The correlation between expression of these ABC transporter genes or proteins and drug response has been examined in epithelial ovarian cancer. However, the conclusions are conflicting (1620). In addition, to our knowledge, there are few reports that examine the mechanism of chemoresistance in clear cell type of ovarian cancer. In this type of ovarian cancer, low proliferation rates of tumor cells were reported to contribute to its survival. In addition, both MDR1 and MRP3 expression have not been shown to be related to the prognosis of patients with clear cell ovarian cancer (21, 22). Ohishi et al. showed that the mRNA levels of MRP1 and MRP3 were related to the prognosis of serous type ovarian cancer and mRNA levels of MRP3 were significantly higher in clear cell type than that in serous type (23). In spite of all these studies, the mechanism of chemoresistance in clear cell type ovarian cancer still remains unclear.

In this study, we showed that ABCF2 protein is predominantly located in the cytoplasm of cells. We showed that both ABCF2 gene and protein expression were significantly correlated with gene amplification especially in clear cell ovarian cancer. Furthermore, we also showed that cytoplasmic ABCF2 expression was significantly correlated with chemotherapy response despite the small number of cases. These data suggest that ABCF2 expression may contribute to the chemoresistant phenotype of clear cell ovarian cancer. However, the role of ABCF2 in conferring chemoresistance in cancer cells is unclear. Yasui et al. reported that ABCF2 gene is amplified in a chemoresistant ovarian cell line (t24/cDDp10), which had chromosome gain at 7q34-36 (14). Besides, there are no other reports to our knowledge showing the mechanism of ABCF2 in chemoresistance. ABCF2 protein is a member of the ABCF transporter superfamily and the GCN20 subfamily (24). Like other members of the ABCF family, ABCF2 contains a pair of nucleotide-binding domain but without any transmembrane domains (25, 26), suggesting that it unlikely functions as a transporter located on the cell membrane as other members of the ABC family. This is further confirmed by our immunohistochemistry data showing predominantly cytoplasmic localization of the protein. The functions of many of these twin nucleotide-binding domain proteins remain unknown. Kerr suggested that a mechanistic similarity exists between eukaryotic members of the ABCF family, which are involved in the control of translation initiation and elongation. These proteins may also have functional similarities to prokaryotic ABCF proteins, which have been shown to be involved in translational control, antibiotic resistance, and RNase L inhibition (26). ABCF may induce factors related with chemoresistance. Further study will be required to delineate the role of ABCF2 in chemoresistance.

In conclusion, ABCF2 protein is a potential prognostic marker for clear cell ovarian cancer and its expression correlates with chemoresponse in patients with clear cell ovarian cancer. Further functional studies of ABCF2 in clear cell ovarian cancer pathogenesis are under way.

Grant support: Dana-Farber/Harvard Cancer Center Ovarian Cancer Specialized Programs of Research Excellence grants P50CA165009 and R33CA103595 from NIH, Department of Health and Human Services, Gillette Center for Women's Cancer, Adler Foundation, Inc., Edgar Astrove Fund, Ovarian Cancer Research Fund, Inc., Morse Family Fund, Natalie Pihl Fund, Ruth N. White Research Fellowship, and Friends of Dana-Farber Cancer Institute and Osaka City General Hospital grant.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
American Cancer Society I. American Cancer Society. Cancer facts and figures 2005; 2005. Available from: http://www.cancer.org/docroot/STT/stt_0.asp).
2
Russel P. Surface epithelial-stroma tumors of the ovary. In: Kurman RJ, editor. Blaustein's pathology of the female genital tract. 4th ed. New York: Springer-Verlag; 1994. p. 705–82.
3
Scully RE. Ovary. In: Henson DE, Albores-Saavedra F, editors. The pathology of incipient neoplasia. Philadelphia: W.B. Saunders; 1993. p. 179–293.
4
Behbakht K, Randall TC, Benjamin I, Morgan MA, King S, Rubin SC. Clinical characteristics of clear cell carcinoma of the ovary.
Gynecol Oncol
1998
;
70
:
255
–8.
5
Goff BA, Sainz de la Cuesta R, Muntz HG, et al. Clear cell carcinoma of the ovary: a distinct histologic type with poor prognosis and resistance to platinum-based chemotherapy in stage III disease.
Gynecol Oncol
1996
;
60
:
412
–7.
6
Morgan RJ, Alvarez RD, Armstrong DK, et al. NCCN practice guide for the ovary. Complete library of NCCN oncology practice guidelines. National Comprehensive Cancer Network 2000.
7
Suehiro Y, Sakamoto M, Umayahara K, et al. Genetic aberrations detected by comparative genomic hybridization in ovarian clear cell adenocarcinomas.
Oncology
2000
;
59
:
50
–6.
8
Schwartz DR, Kardia SL, Shedden KA, et al. Gene expression in ovarian cancer reflects both morphology and biological behavior, distinguishing clear cell from other poor-prognosis ovarian carcinomas.
Cancer Res
2002
;
62
:
4722
–9.
9
Pollack JR, Perou CM, Alizadeh AA, et al. Genome-wide analysis of DNA copy-number changes using cDNA microarrays.
Nat Genet
1999
;
23
:
41
–6.
10
Pollack JR, Sorlie T, Perou CM, et al. Microarray analysis reveals a major direct role of DNA copy number alteration in the transcriptional program of human breast tumors.
Proc Natl Acad Sci U S A
2002
;
99
:
12963
–8.
11
Tsuda H, Birrer MJ, Ito YM, et al. Identification of DNA copy number changes in microdissected serous ovarian cancer tissue using a cDNA microarray platform.
Cancer Genet Cytogenet
2004
;
155
:
97
–107.
12
Wong KK, Cheng RS, Mok SC. Identification of differentially expressed genes from ovarian cancer cells by MICROMAX cDNA microarray system.
Biotechniques
2001
;
30
:
670
–5.
13
Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4.
Nature
1970
;
227
:
680
–5.
14
Yasui K, Mihara S, Zhao C, et al. Alteration in copy numbers of genes as a mechanism for acquired drug resistance.
Cancer Res
2004
;
64
:
1403
–10.
15
Sugiyama T, Kamura T, Kigawa J, et al. Clinical characteristics of clear cell carcinoma of the ovary: a distinct histologic type with poor prognosis and resistance to platinum-based chemotherapy.
Cancer
2000
;
88
:
2584
–9.
16
Arts HJ, Katsaros D, de Vries EG, et al. Drug resistance-associated markers P-glycoprotein, multidrug resistance-associated protein 1, multidrug resistance-associated protein 2, and lung resistance protein as prognostic factors in ovarian carcinoma.
Clin Cancer Res
1999
;
5
:
2798
–805.
17
Bourhis J, Goldstein LJ, Riou G, Pastan I, Gottesman MM, Benard J. Expression of a human multidrug resistance gene in ovarian carcinomas.
Cancer Res
1989
;
49
:
5062
–5.
18
Izquierdo MA, van der Zee AG, Vermorken JB, et al. Drug resistance-associated marker Lrp for prediction of response to chemotherapy and prognoses in advanced ovarian carcinoma.
J Natl Cancer Inst
1995
;
87
:
1230
–7.
19
Kigawa J, Minagawa Y, Cheng X, Terakawa N. γ-Glutamyl cysteine synthetase up-regulates glutathione and multidrug resistance-associated protein in patients with chemoresistant epithelial ovarian cancer.
Clin Cancer Res
1998
;
4
:
1737
–41.
20
Nakayama K, Kanzaki A, Ogawa K, Miyazaki K, Neamati N, Takebayashi Y. Copper-transporting P-type adenosine triphosphatase (ATP7B) as a cisplatin based chemoresistance marker in ovarian carcinoma: comparative analysis with expression of MDR1, MRP1, MRP2, LRP and BCRP.
Int J Cancer
2002
;
101
:
488
–95.
21
Itamochi H, Kigawa J, Akeshima R, et al. Mechanisms of cisplatin resistance in clear cell carcinoma of the ovary.
Oncology
2002
;
62
:
349
–53.
22
Itamochi H, Kigawa J, Sugiyama T, Kikuchi Y, Suzuki M, Terakawa N. Low proliferation activity may be associated with chemoresistance in clear cell carcinoma of the ovary.
Obstet Gynecol
2002
;
100
:
281
–7.
23
Ohishi Y, Oda Y, Uchiumi T, et al. ATP-binding cassette superfamily transporter gene expression in human primary ovarian carcinoma.
Clin Cancer Res
2002
;
8
:
3767
–75.
24
Vazquez de Aldana CR, Marton MJ, Hinnebusch AG. GCN20, a novel ATP binding cassette protein, and GCN1 reside in a complex that mediates activation of the eIF-2α kinase GCN2 in amino acid-starved cells.
EMBO J
1995
;
14
:
3184
–99.
25
Allikmets R, Gerrard B, Hutchinson A, Dean M. Characterization of the human ABC superfamily: isolation and mapping of 21 new genes using the expressed sequence tags database.
Hum Mol Genet
1996
;
5
:
1649
–55.
26
Kerr ID. Sequence analysis of twin ATP binding cassette proteins involved in translational control, antibiotic resistance, and ribonuclease L inhibition.
Biochem Biophys Res Commun
2004
;
315
:
166
–73.