Overexpression of EIF-5A2 Is an Independent Predictor of Outcome in Patients of Urothelial Carcinoma of the Bladder Treated with Radical Cystectomy

Bladder cancer is a major cause of morbidity and mortality in Western countries (1). Bladder tumorigenesis is believed to be a multistep process driven by an accumulation of genetic alterations. The acquisition of loss of tumor suppressor genes and activation of oncogenes by tumor cells is a central event in the development and progression of bladder cancer and one that may frequently decide this tumors future malignant potential (2). Although bladder cancer has been widely studied, the search for specific gene alterations associated with tumorigenesis of this cancer has been of insufficient magnitude and the identification of the prevalence and clinical/prognostic significance of genetic markers in bladder cancer is still substantially limited.

Urothelial carcinoma of the bladder (UC) is the most common histopathologic type of bladder cancer. Chromosomal aberrations of UC have been extensively analyzed by comparative genomic hybridization with several amplified regions including 3q or parts of 3q, being reported (3-5). Amplification of 3q also has been detected frequently in ovarian cancer (6), lung cancer (7), colorectal carcinoma (8), and other neoplasms (9-12), suggesting that human chromosome 3q contains oncogenes related to tumorigenesis and progression of a number of different solid tumors. Using a chromosome microdissection-hybrid selection method, we have previously isolated a novel candidate oncogene, eIF-5A2 (eukaryotic initiation factor 5A2), from a primary ovarian cancer cell line containing a high copy number amplification of 3q26 (13). EIF-5A2 protein shares 82% identical amino acid sequence with EIF-5A1. A functional characterization study of human EIF-5A1 isoforms has revealed that either the eIF-5A1 gene or the eIF-5A2 gene could complement growth of a yeast strain in which the yeast eIF-5A genes were disrupted (14). Recently, oncogenic characteristics of eIF-5A2 have been shown by in vitro assays and overexpression of EIF-5A2 has been observed to correlate with late clinical stage in ovarian cancer (15). More recently, one report showed that overexpression of eIF-5A2 mRNA was associated with a higher risk of lymph node metastasis in human gastric adenocarcinomas (16). To date, the status of abnormalities of the eIF-5A2 gene in UC, however, has not been examined.

The purpose of the present study was to investigate the status of abnormalities of the eIF-5A2 gene in UC and the clinicopathologic and prognostic significance of these abnormalities. In this study, reverse transcription-PCR, immunohistochemistry (IHC), and fluorescence in situ hybridization (FISH) were used to examine mRNA/protein expression and amplification of eIF-5A2 in a cohort of UCs treated with radical cystectomy.

In this study, for eIF-5A2 mRNA expression analysis, sixteen patients with UC that underwent radial cystectomy were collected from Department of Urology, Cancer Center, and the First Affiliated Hospital, Sun Yat-Sen University, and the Hospital of Guangdong province (Guangzhou, China) in 2007. Age of the 16 UC patients ranged from 47 to 71 y (mean, 63 y), with a male to female ratio of 4.3:1. The tumor specimens encompassed 13 cases of invasive UC (pT2-pT4) and 3 cases of high-risk superficial UC (grade 3/pT1). The primary UC tissue and its adjacent normal bladder tissue specimens were snap-frozen in chilled liquid nitrogen and stored at −80°C until further processing.

For preparation of the bladder tissue microarray (TMA), 90 patients with UC that underwent radial cystectomy were selected from the surgical pathology archives of the Department of Pathology, the First Affiliated Hospital and Cancer Center, Sun Yat-Sen University and the Hospital of Guangdong province (Guangzhou, China) between 1996 and 2005. The cases selected were based on availability of resection tissue and follow-up data. The follow-up period ranged from 11 to 64 mo (median, 34 mo). The medical records of all patients were retrospectively reviewed with emphasis on cancer-specific survival. Cancer-specific survival was determined from the date of surgery to the date of death from bladder urothelial carcinoma or last follow-up. Patients whose cause of death remained unknown were excluded in this study. Of the 90 UC patients, 69 had invasive UCs (pT2-pT4). The remaining 21 UC patients had high-risk superficial UCs, in which 6 UCs were multifocal grade 2/pT1 tumor, 15 were grade 3/pT1 tumor. The age of these UC patients ranged from 35 to 88 y at the time of surgery (median age, 62 y) and the male/female ratio was 4.6:1. The clinicopathologic characteristics are summarized in Table 1. The tumor specimens were recruited from paraffin blocks of 90 primary UCs. Fifty cases of normal bladder mucosa from adjacent nonneoplastic bladder tissue of the same UC patients, in paraffin blocks, were also obtained.

None of the UC patients included in this study had received preoperative radiation or chemotherapy. Tumor grade and stage were defined according to the criteria of the WHO and the 6th edition of the pTNM classification of the International Union Against Cancerf (UICC, 2002). The study was approved by the medical ethics committee of our institute.

Total RNA was isolated from 16 pairs of UC tissues and their adjacent normal bladder tissues using TRIZOL reagent (Invitrogen). RNA was reverse-transcribed using SuperScript First Strand cDNA System (Invitrogen). The first strand cDNA products were then amplified with GAPDH-specific (F: 5′-TTTGGTATCGTGGAAGGAC-3′ and R: 5′- AAGGTGGAGGAGTGGGT -3′) and eIF-5A2–specific (F: 5′-TTCCAGCACTTACCCTAT-3′ and R: 5′-TTTCCCTCTATTTCTTTG-3′) primers by PCR. The PCR condition for GAPDH was as follows: 94°C for 30 s, 55°C for 30 s, and 72°C for 30 s; 27 cycles. The PCR condition for eIF-5A2 was as follows: 94°C for 30 s, 48°C for 30 s, and 72°C for 30 s; 30 cycles. The PCR products were analyzed by 2% agarose gel electrophoresis.

The TMA was constructed according to a method described previously (17). Briefly, the individual donor tissue block and the corresponding histologic H&E-stained slides were overlaid for tissue TMA sampling. The tissues (90 UC and 50 normal bladder tissues) were sampled using a tissue arraying instrument (Beecher Instruments); a 0.6-mm diameter cylinder of tissue was removed. Subsequently, the tissue cylinder was re-embedded into a predetermined position in a recipient paraffin block. In our constructed bladder tissue-TMA, three cores of sample were selected from each primary UC and normal bladder tissue. Multiple sections (5-μm thick) were cut from the TMA block and mounted on microscope slides.

IHC studies were done using a standard streptavidin-biotin-peroxidase complex method (18). In brief, TMA sections were deparaffinized and rehydrated. Endogenous peroxidase activity was blocked with 0.3% hydrogen peroxide for 20 min. For antigen retrieval, TMA slides were microwave-treated and boiled in a 10 mmol/L citrate buffer (pH 6.0) for 10 min. Nonspecific binding was blocked with 10% normal rabbit serum for 20 min. The TMA slides were incubated with either monoclonal mouse anti-human EIF-5A2 (kindly provide by Dr. Geng-Xi Hu, the Institute of Shanghai Science and Technology, Shanghai, China; 1:100 dilution) overnight at 4°C or anti–Ki-67 (Dako; 1:100 dilution) for 30 min at 37°C in a moist chamber. The slides were sequentially incubated with biotinylated rabbit antimouse immunoglobulin at a concentration of 1:100 for 30 min at 37°C and then reacted with a streptavidin-peroxdase conjugate for 30 min at 37°C and 3′-3′ diaminobenzidine as a chromogen substrate. The nucleus was counterstained using Meyer's hematoxylin. A negative control was obtained by replacing the primary antibody with normal murine IgG. Known immunostaining positive ovarian cancer slides were used as positive controls.

For evaluation of the EIF-5A2 IHC staining in different bladder tissues, the positive expression of EIF-5A2 in normal and malignant bladder tissues was a cytoplasmic pattern. Because all tissues are present on one TMA slide, this ensures equivalent staining conditions. Hence, a semiquantitative scoring criterion for IHC of EIF-5A2 was used, in which both staining intensity and positive areas were recorded. A staining index (values 0-12), obtained as the intensity of EIF-5A2–positive staining (negative, 0; weak, 1; moderate, 2; or strong, 3 scores) and the proportion of immunopositive cells of interest (<25%, 1; 25-50%, 2, >50-75%, 3; ≥75%, 4 scores) were calculated. For Ki-67, the status of Ki-67 nuclear expression was assessed as the percentage of Ki-67–positive cells stained in each tumor. For cell counting of both EIF5A2 and Ki-67 IHC staining in our TMAs, the whole area of each TMA sample was observed. For each case, a total of at least 300 epithelial cells from 3 TMA samples were evaluated and mean values of the 3 replicates were calculated and recorded as a staining index.

Two-color FISH was done using a Spectrum Orange–labeled BAC clone (RP11-115J24) at 3q26.2 containing the eIF-5A2 gene and a Spectrum Green–labeled reference centromeric probe on chromosome 3 (Vysis). The FISH reaction was done as described previously (19) with slight modification. Briefly, deparaffinized TMA sections were treated with proteinase K (400 μg/mL) at 37°C for 30 min, followed by denaturing in 70% formamide, 2× SSC at 75°C for 6 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 TMA sections at 37°C for 24 h. After hybridization, TMA sections were washed thrice in 50% formamide and 2× SSC at 45°C for 3 min each. The slides were counterstained with 1 μg/mL 4′,6-diamidino-2-phenylindole in an antifade solution and were examined with a Zeiss Axiophot microscope equipped with a triple-band pass filter. A minimum of 300 tumor cells was evaluated per specimen. Amplification of eIF-5A2 was defined as presence of either 6 (or more) eIF-5A2 gene signals or at least thrice as many gene signals than centromere signals of chromosome 3 in >20% of cells.

Statistical analysis was done with the SPSS software (SPSS Standard version 10.0; SPSS, Inc.). The association of EIF-5A2 protein expression with UC patient's clinicopathologic features was assessed by the χ² test. For survival analysis, we analyzed all UC patients by Kaplan-Meier analysis. Log-rank test was used to compare different survival curves. Multivariate survival analysis was done on all variables that were found to be significant on univariate analysis using the Cox regression model. An unpaired sample t test was used to assess the expression of Ki-67 between groups with overexpression and normal expression of EIF-5A2. P values of <0.05 were considered significant.

To investigate the status of gene expression of eIF-5A2 in UC, we studied the eIF-5A2 mRNA expression in 16 pairs of primary UC and adjacent normal bladder specimens using reverse transcription-PCR. A total of 8 of 16 (50%) UCs showed up-regulated eIF-5A2 expression, when compared with their adjacent normal bladder tissues (Table 1; Fig. 1).

In the present study, the protein expression of eIF-5A2 was investigated by IHC in a bladder tissue-TMA, which contained 90 cases of primary UC and 50 cases of normal bladder mucosa. Positive expression of EIF-5A2 in epithelial tissue cells was primarily a cytoplasmic pattern. EIF-5A2 expression could be evaluated informatively in 86 of 90 (95.6%) of the UCs and 41 of 50 (82.0%) of normal bladder mucosa (Table 1). The noninformative samples included unrepresentative samples, samples with too few tumor cells (<300 cells per case) and lost samples; such were not used in data compilation.

Because the expression of EIF-5A2 in each of 41 informative normal bladder mucosa was negative (0) or weak (1; Fig. 2A), the staining index in the normal bladders was determined to be less or equal to 3. Therefore, we designated the staining index of 0 to 3 as the normal expression of EIF-5A2 (Fig. 2A), whereas staining index of 4 to 12 was depicted as overexpression of this protein (Fig. 2B and C). Using this designation, the overexpression of EIF-5A2 was observed in 39 of 86 (45.3%) informative UCs. In addition, a potential association between EIF-5A2 expression and tumor clinicopathologic features in UCs was further evaluated. Results show that no significant association was found between EIF-5A2 expression and the clinicopathologic features of the UC cohorts, such as patient's age and gender and tumor grade, pT status, pN status, and multiplicity (P >0.05; Table 1).

In univariate survival analyses, cumulative survival curves were calculated according to the Kaplan-Meier method. Differences in survival times were assessed with the log-rank test. Kaplan-Meier analysis showed a significant effect of certain clinicalpathologic prognostic variables, such as tumor pT status (P = 0.001) and pN status (P = 0.002) on patient survival (Table 2). With regard to EIF-5A2 expression, the mean survival time for patients with UCs having EIF-5A2 overexpression was 38.2 months compared with 52.9 months for patients with UCs having normal EIF-5A2 expression (P = 0.001; Fig. 3; Table 3). Additionally, survival analysis was done with regard to EIF-5A2 expression in subsets of patients with different tumor histopathologic grades, pT, and pN stages. The results show that overexpression of EIF-5A2 was a prognostic factor in UC patients having tumor grade 1 of 2 (P = 0.0009) and grade 3 (P = 0.016), pT1 (P = 0.0089), pT2 (P = 0.0354), pT3/4 (P = 0.0058), pN0 (P = 0.0039), and pN1-2 (P = 0.0093; Fig. 3; Table 3).

Because variables observed to have prognostic influence by univariate analysis may covariate, the expression of EIF-5A2 as well as other clinical pathologic variables that were significant in univariate analysis (tumor pT status and pN status) were examined in multivariate analysis (Table 4). The expression of EIF-5A2 was found to be an independent prognostic factor for poor overall survival (relative risk, 2.777; confidence interval, 1.398-5.517; P = 0.004). Of the other variables, only tumor pT status was shown as an independent prognostic factor (P = 0.005) for overall survival.

In our FISH study, the FISH analysis was informative in 23 of 50 (46.0%) of the normal bladder tissues and 47 of 90 (52.2%) of the UCs. Samples without FISH signal and samples with weak target signals or those with a strong signal background were the main reasons for most of the noninformative cases. Other reasons for noninformative cases were due to problems related to TMA technology as described in the IHC study. FISH results showed that the amplification of eIF-5A2 was not detected in any of the normal bladder tissues but was detected in 5 of 47 (10.6%) of the informative UCs; in each of the 5 cases with eIF-5A amplification, overexpression of EIF-5A2 was observed (Fig. 2C and D). In the remaining 42 informative cancers without amplification of eIF-5A2, 27 (64.3%) cases showed normal expression of EIF-5A2, whereas 15 (35.7%) cases were observed overexpression of EIF-5A2.

To address whether or not EIF-5A2 expression in UC is associated with cell proliferation, the expression of Ki-67, a widely used cellular proliferation marker, was examined by IHC in our UC-TMA. Among the 90 UCs, in 83 samples, EIF-5A2 and Ki-67 IHC was detected successfully and simultaneously. For the 37 cases with overexpression of EIF-5A2, an average of 38.2% of the UC cells stained positive with Ki-67 antibody, a percentage of cancer cells that was significantly larger than that (25.1%) in the remaining 46 cancers with a normal expression of EIF-5A2 (P < 0.05, unpaired sample t test; Fig. 2E).

At present, current pTNM staging system is established and is useful prognostic indicators for UCs (20). This clinicopathologic system, however, based on clinicopathology and extent of disease at presentation, has reached its limit in providing critical information that may influence the strategy of treatment. Patients with the same clinicopathologic stage of UC treated with radical cystectomy display considerable variability in survival (21, 22). Therefore, there is an urgent need for new variables that can distinguish between patients with unfavorable prognosis and others with better prognosis. It is known, if individuals with different prognosis could be identified by molecular biomarkers at the time of surgery, their survival might be prolonged by more effective adjuvant therapy (21). Although UC has been widely studied, however, the identification of specific genetic alterations associated with tumorigenesis of UC and their clinical/prognostic significance is still substantially limited. Thus, further investigations are needed to develop appropriate biomakers, as well as good control in examine of the disease.

Recently, we have isolated a novel candidate oncogene, eIF-5A2, from a primary ovarian cancer cell line containing a high-copy-number amplification of 3q26 (13). The eIF-5A2 gene is commonly amplified and/or overexpressed in several types of human cancers, including ovarian (15), gastric (16), and colorectal cancer (23). To investigate the potential oncogenic role of eIF-5A2 in UC, we examined first the expression of eIF-5A2 by reverse transcription-PCR, in a panel (16 cases) of primary UC and adjacent normal bladder tissues. The results showed that up-regulated expression of eIF-5A2 mRNA was observed in a half (8 of 16) of UCs, when compared with their adjacent normal bladder tissues. Next, we examined protein expression of eIF-5A2 by IHC, in a TMA of a large cohort of normal and cancerous bladder tissues. Our results show that the expression of EIF-5A2 in all of the normal bladder tissue specimens was absent or at low levels. In many of our UC specimens, in contrast, an overexpression of EIF-5A2 was frequently detected. These findings suggest the possibility that up-regulated expression of eIF-5A2 may provide a selective advantage in UC tumorigenic processes.

Our previous studies have observed that in certain human solid tumors, such as ovarian and colorectal cancers, overexpression of EIF-5A2 protein was positively correlated with an ascending clinical stage of the tumor (15, 23). In addition, more recent report has provided evidences that up-regulated expression of eIF-5A2 mRNA was associated with a higher risk of lymph node metastasis of gastric carcinomas (16). These data provide evidence that increased expression of eIF-5A2 may involve in the invasive and/or metastatic processes of several types of human cancer. In the present study of a large series of UC tissues, however, no significant association was observed between EIF-5A2 expression and any of the UC patient's clinicopathologic features, including tumor grade and clinical stage. Interestingly, it is noteworthy that overexpression of EIF-5A2 was a strong and independent predictor of short overall survival as evidenced by Kaplan-Meier curves and multivariate Cox proportional hazards regression analysis in our present study. Most importantly, stratified survival analysis of UC histopathologic grade and/or pTN stage showed EIF-5A2 expression to be closely correlated to survival of different subsets of UC patients. Thus, EIF-5A2 expression seems to have the potential to predict UC patient outcome. The examination of EIF-5A2 expression by IHC may, therefore, be used as an additional tool in identifying those patients at risk of tumor recurrence and/or progression, and it may be a helpful criterion to optimize individual therapy management. These findings raise the question of a potentially important role of eIF-5A2 as an underlying biological mechanism in the development and/or growth of UCs.

The gene eIF-5A2 is located at chromosome 3q26.2 and was recently recognized as a novel member of the eIF-5A gene family (13, 14). EIF-5A2 shares 82% identical amino acid sequence with its family members, EIF-5A1, including the minimum domain needed for EIF-5A maturation by hypusine modification at lysine-50 residue. Multiple studies have shown that eIF-5A1 is involved in many biological processes such as cell proliferation and apoptosis (24-26). Functional characterization studies of the human EIF-5A isoforms have revealed that either the eIF-5A1 or the eIF-5A2 gene could complement growth of a yeast strain in which the yeast eIF-5A genes were disrupted (14). Our previous in vitro study has found that antisense DNA against the eIF-5A2 gene could inhibit cell growth of ovarian cancer cell line UACC-1598 that has amplification of eIF-5A2. In addition, anchorage-independent growth in soft agar was observed in eIF-5A2–transfected NIH3T3 and LO2 cells, and tumor formation in athymic nude mice was induced in eIF-5A2–transfected LO2 cells (15). These findings provide further evidence that the eIF-5A2 may function as a putative oncogene in tumorigenetic processes. In this study, we observed an overall significant positive association of up-regulated expression of EIF-5A2 and increased UC cellular proliferation. These results suggest a potential important role of eIF-5A2 in the control of cell proliferation, an activity that might be responsible, at least in part, for tumorigenesis and/or progression of UC. Recently, a gene-expression signature that may distinguish primary from metastatic tumors in adenocarcinomas was proposed. Identifying the metastatic signature gene was accomplished by comparing the gene-expression profiles between primary and metastatic tumors of multiple adenocarcinomas and deoxyhypusine synthase, the enzyme involved in the hypusination reaction of EIF-5A1 and EIF-5A2 (27). Hypusine modification at the lysine-50 residue is necessary for the maturation of EIF-5A1 and EIF-5A2 proteins. The role of deoxyhypusine synthase in tumor metastasis might be associated with EIF-5A2 because EIF-5A2 is one of the two only known cellular substrates for deoxyhypusine synthase. Clearly, further work needs to be done to more precisely understand the potential oncogenic function of eIF-5A2 in human cancer pathogenesis and which signaling pathway is involved during the tumorigenic process of cancer cells caused by increased expression of EIF-5A2.

With regard to the mechanism of up-regulated protein expression of eIF-5A2 in UCs, it is known that gene amplification is a common pathologic mechanism of gene overexpression in human cancers (28). To determine whether the overexpression of EIF-5A2 in UCs was caused by gene amplification, the amplification status of eIF-5A2 was examined by FISH. In our 47 informative cases of UC by both IHC and FISH simultaneously, overexpression of EIF-5A2 was detected in all (5 of 5) UCs that had eIF-5A2 amplification. However, amplification of eIF-5A2 was not observed in 15 other UCs with overexpression of EIF-5A2. These results indicate that the protein expression level of eIF-5A2 in UC does not always coincide with gene amplification. On other hand, there was a recent report showing that in a large series of human cancer cell lines, in which eIF-5A mRNA was demonstrable in most cells of these cell lines, EIF-5A2 protein was detectable only in two of the cell lines, colorectal-SW-480 and ovarian-UACC-1598 (29). These data suggest that the regulation of protein expression of EIF-5A2 is complicated and it might be regulated not only by gene amplification but also by other molecular mechanisms including transcriptional regulation and posttranslational regulation (e.g., microRNA).

In summary, in this study, we describe, for the first time, mRNA/protein expression and amplification patterns of eIF-5A2 in normal human bladder tissues and in UC tissues. Our results provide a basis for the concept that up-regulated expression of EIF-5A2 in human UC may be important in the acquisition of a poor prognostic phenotype, suggesting that the expression of EIF-5A2, as detected by IHC, is an independent molecular marker for shortened survival time of UC patients treated with radical cystectomy, and it might be a helpful criterion to optimize individual therapy management. Because the recurrence-free survival data of most UCs studied in this report was not available, we failed to perform this kind of prognosis analysis. This was indeed a limitation of our present study. Further studies designed to determine whether or not there is an association between expression of EIF-5A2 in UCs and patient recurrences are in order.

No potential conflicts of interest were disclosed.

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.