Human Pathology (2009) 40, 631–638
www.elsevier.com/locate/humpath
Original contribution
Aurora kinases as prognostic biomarkers in ovarian carcinoma☆,☆☆ Marta Mendiola PhD a,b,c,d , Jorge Barriuso MD, PhD b,c,e,f , Adrián Mariño-Enríquez MD a , Andrés Redondo MD, PhD c,f , Aurora Domínguez-Cáceres PhD a,b,c,g , Ginés Hernández-Cortés MD h , Elia Pérez-Fernández BsC b,i , Iker Sánchez-Navarro BsC b,c,e , Juan Ángel Fresno Vara PhD b,c,e , Asunción Suárez MD a , Enrique Espinosa MD, PhD c,f , Manuel González-Barón MD, PhD c,f , José Palacios MD, PhD j , David Hardisson MD, PhD a,c,⁎ a
Department of Pathology, Hospital Universitario La Paz, Universidad Autónoma de Madrid, Spain Fundación para la Investigación Biomédica del Hospital Universitario La Paz (FIBHULP) c Laboratory of Pathology and Oncology, Research Unit, Hospital Universitario La Paz, Universidad Autónoma de Madrid, Spain d Fundación Científica de la Asociación Española de Lucha Contra el Cáncer (AECC) e Fondo de Investigación Sanitaria (FIS), Instituto de Salud Carlos III, Ministerio de Sanidad y Consumo, Spain f Department of Medical Oncology, Hospital Universitario La Paz, Universidad Autónoma de Madrid, Spain g Programa Juan de la Cierva, Ministerio de Educación y Ciencia, Spain h Department of Obstetrics and Gynecology, Hospital Universitario La Paz, Universidad Autónoma de Madrid, Spain i Unit of Biostatistics, Hospital Universitario La Paz, Universidad Autónoma de Madrid, Spain j Department of Pathology, Hospital Universitario Vírgen del Rocío, Sevilla, Spain b
Received 28 July 2008; revised 16 October 2008; accepted 20 October 2008
Keywords: Ovarian carcinoma; Aurora kinases; AURKA; Prognosis
Summary We investigated the expression of Aurora kinases A and B by immunohistochemistry in 68 ovarian carcinomas to analyze their prognostic value. The amplification of AURKA gene by fluorescence in situ hybridization was also assessed. Overall, 58.8% and 85.3% of ovarian carcinomas showed expression of Aurora A and B, respectively. Amplification of AURKA was found in 27.6% of cases examined. Tumors with Aurora A expression showed a lower rate of recurrence than those tumors without Aurora A expression (65% versus 91.7%, P = .019). In the univariate analysis, patients with Aurora A and B expression showed an increased progression-free survival (P = .023 and .06, respectively, log-rank test) and overall survival (P = .03 and .02, respectively, log-rank test). The
☆
This study was supported by grants SAF2004-0825-C02-02 (Ministerio de Educación y Ciencia, Spain; D.H., A.D.-C., and G.H.-C.), SAF2004-0825C02-01 (Ministerio de Educación y Ciencia, Spain; J.P.), and RETICS (RD06/0020/0013; J.P.). M.M. is supported by the Scientific Foundation of Asociación Española Contra el Cáncer. J.B. is supported by a junior research contract of FIS 2006 (Instituto de Salud Carlos III, Ministerio de Sanidad y Consumo, Spain). A.D.-C. is supported by Programa Juan de la Cierva, (Ministerio de Educación y Ciencia, Spain). J.A.F.V. is supported by grant award CP05/00248 (FIS). ☆☆ This article was presented in part at the 21st European Congress of Pathology, Istanbul (Turkey), September 8-13, 2007 (abstract PP4-2). ⁎ Corresponding author. Departamento de Anatomia Patologica; Hospital Universitario La Paz; E-28046 Madrid, Spain. E-mail address:
[email protected] (D. Hardisson). 0046-8177/$ – see front matter © 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.humpath.2008.10.011
632
M. Mendiola et al. multivariate analysis adjusted to optimal surgery by Cox proportional hazards regression showed Aurora A expression as an independent prognostic factor for progression-free survival (P = .03) and overall survival (P = .02). In conclusion, Aurora A expression seems to have a prognostic value in ovarian carcinoma. © 2009 Elsevier Inc. All rights reserved.
1. Introduction Ovarian carcinoma is the most common cause of death from a gynecological malignancy [1]. Most patients present with advanced disease, which is managed with surgical resection and a combination of paclitaxel and platinum-based chemotherapy. Most patients present with advanced disease (stage III/IV), and although standard treatment will result in an initial response rate of more than 70%, only 50% of these patients will still be alive after 5 years [2]. Unrestrained cell division in cancer cells is dependent upon mitosis and its related processes. A proven effective strategy in cancer treatment has been to interfere with the function of the mitotic spindle. Tubulin remains the only spindle-associated protein targeted by clinically approved agents with relative success. In recent years, major advances have been made in targeting proteins that associate with tubulin and the mitotic spindle; mitotic kinases such as the Aurora family are receiving significant attention due to their vital roles in assuring proper centrosome separation and chromosome segregation. Aurora A (also known as AURKA, STK15, BTAK, AIK, AURA, and AURORA2) is a serinethreonine kinase essential for mitotic spindle formation and accurate chromosome segregation. Overexpression of Aurora A at mRNA and/or protein level has been described in a variety of human neoplasms, including breast [3], gastric [4], ovarian [5], pancreatic [6], and hepatocellular carcinomas [7], with positive rates ranging from 26% to 94%. Recent in vitro studies suggest that overexpression of Aurora A plays a role in resistance to taxanes. In this sense, it has been demonstrated that inhibition of Aurora A expression by small interfering RNA enhanced the chemosensitivity of pancreatic cancer cells to taxanes [8]. Moreover, Aurora A overexpression induced resistance to taxol in another in vitro study [9]. In the case of Aurora B, the situation is unclear. Aurora B (also known as AURKB and AIK2) plays an essential role in chromosome segregation and cytokinesis, and its kinase activity is required for bipolar chromosome orientation and condensation [10]. Aurora B kinase is overexpressed in cancer cells, and an increased level of Aurora B correlates with advanced stages of colorectal cancer [11]. A few reports strongly suggest a direct link between Aurora B and carcinogenesis, proposing that Aurora B overexpression might be a secondary event in p53-defective cells that leads to malignancy [12]. In both Aurora A and Aurora B overexpressing cells, defects in p53 seem to play an essential role in stabilizing polyploidy.
In this study, we examined the immunohistochemical expression of Aurora A, Aurora B, and p53 in a series of ovarian carcinomas. AURKA gene amplification was analyzed using fluorescence in situ hybridization (FISH). The mutational status of TP53 and the proliferation marker Ki67 were also assessed. Finally, the prognostic significance of the expression of Aurora A and Aurora B kinases, AURKA gene amplification, and TP53 mutations was examined using the follow-up data.
2. Materials and methods 2.1. Patients selection We included in our study 68 nonconsecutive ovarian carcinomas from patients treated at the Hospital Universitario La Paz (Madrid, Spain) between February 1996 and December 2003. All patients underwent exploratory laparotomy for diagnosis, staging, and debulking, followed by platinum/taxane-based chemotherapy. Patients were staged according to the International Federation for Gynecology and Obstetrics (FIGO) classification. Optimal debulking was defined as 1 cm or less (diameter) residual disease. Progression-free survival (PFS) was defined as the time interval between the start of the treatment and the first confirmed sign of disease recurrence or progression. Overall survival (OS) was defined as the time interval between the start of the treatment and the date of death or end of followup. Follow-up data were obtained by retrospective chart review at Hospital Universitario La Paz. Approval for the study was obtained from the Local Ethics Committee.
2.2. Tissue microarray construction Representative areas of the tumors were selected on hematoxylin and eosin–stained sections and marked on individual paraffin blocks. Two tissue cores (1 mm in diameter) were obtained from each specimen. The tissue cores were arrayed into a receptor paraffin block using a tissue microarray workstation (Beecher Instruments, Silver Spring, MD) as described previously [13]. A hematoxylin and eosin–stained section of the array was reviewed to confirm the presence of morphologically representative areas of the original lesions. A tissue core was considered informative if at least 50% of the sample contained tumor tissue.
Aurora kinases in ovarian carcinoma
2.3. Immunohistochemistry Immunohistochemistry was performed on 4-μm sections of formalin-fixed, paraffin-embedded tissues. Briefly, the tissue sections were deparaffinized and rehydrated in water, after which antigen retrieval was carried out by incubation in EDTA solution, pH 8.2 at 50°C for 45 minutes in an autoclave. Endogenous peroxidase and nonspecific antibody reactivity was blocked with peroxidase blocking reagent (Dako, Glostrup, Denmark) at room temperature for 15 minutes. The sections were then incubated for 60 to 90 minutes at 4°C with the following antibodies: Aurora A monoclonal antibody (clone JLM28, Novocastra Laboratories, Newcastle Upon Tyne, UK; dil. 1:50), Aurora B polyclonal antibody (Abcam pcl, Cambridge, UK; dil. 1:50), p53 monoclonal antibody (clone DO-7, Novocastra Laboratories; dil. 1:100), and Ki67 monoclonal antibody (clone MIB1, Dako; dil. 1:100). Detection was performed with Envision Plus Detection System (Dako). Negative controls were used with goat serum replacing the primary antibody.
2.4. Analysis of immunohistochemical stains Immunohistochemical staining was evaluated by visual counting of the cells. Aurora A staining was predominantly seen in the cytoplasm. For Aurora B, p53, and Ki67, the staining was nuclear. For all the markers, immunoreactivity was expressed as the percentage of tumor cells that exhibited any staining, regardless of intensity. Given that Aurora A and Aurora B are normally undetectable by immunohistochemistry in normal nonmitotic cells [14], and as described by others, any expression of the protein could be considered positive [15,16]. We established a cutoff rate of 5% to consider the expression of both, Aurora A and Aurora B, as positive. p53 was considered positive if more than 10% of the tumor cells showed nuclear-positive immunostaining, according to previous studies on ovarian carcinoma [17-20]. The percentage of tumor cells with positive Ki67 nuclear staining was interpreted as the proliferation index. Proliferation index was classified as high (N30%) or low (≤30%) according to the median value of the registered scores.
2.5. FISH analysis FISH analysis and detection of AURKA amplification was performed with the Bacterial Artificial Chromosome (BAC) BAC RP5-1167H4, from the Human BAC Clone Library RPC5 (Children's Hospital Oakland Research Institute, CA), which spans the entire AURKA genomic region, and a commercial probe for chromosome 20 (CEP20; Vysis, Downer's Grove, IL) as a control for the ploidy level, as previously described [21]. Fluorescence signals were scored in each sample by counting the number of single-copy gene and centromeric signals in 100 welldefined nuclei. Amplification was defined as the presence (in N5% of tumor cells) of either more than 10 gene signals or more than 3 times as many gene signals as centromere
633 signals of chromosome 20 [21]. All images were collected on a Nikon Eclipse 90i fluorescence microscope equipped with a high-resolution, high-sensitivity Nikon DS-Fi1 camera and were digitally processed by using NIS elements F220 Imaging software (Nikon).
2.6. DNA isolation and TP53 mutational status analysis TP53 gene status was analyzed on formalin-fixed, paraffin-embedded tumor specimens. Representative tumor tissue sections were cut and placed directly into a sterile microfuge tube. DNA was extracted using MasterPure™ DNA Purification Kit (EPICENTRE Biotechnologies). All procedures were performed according to manufacturer's protocols. Polymerase chain reaction (PCR) was performed in 25 μL final volume, containing 5 μL of DNA (with an approximately concentration of 250 ng), 1 mmol/L dNTP (Applied Biosystems), 1.5 mmol/L MgCl2, 1× PCR buffer and 1 U AmpliTaq Pol (Applied Biosystems), and 0.5 to 0.8 μmol/L of each primer (MWG-Biotech AG, Ebersberg, Germany). The sequences of PCR primers used are as follows (exons 5 through 8): E5F (5′- CCG TGT TCC AGT TGC TTT ATC), E5R (5′- AGC CCT GTC GTC TCT CCA), E6F (5′- GGG CTG GTT GCC CAG GGT), E6R (5′- AGT TGC AAA CCA GAC CTCA), E7F (5′-CCA CAG GTC TCC CCA AGG), Table 1
Clinicopathological data n (%)
Age, median (range), y Histological subtype (n = 68) Serous Other a Tumor grade (n = 67) Well differentiated (grade 1) Moderately differentiated (grade 2) Poorly differentiated (grade 3) FIGO stage (n = 68) I-II III IV Debulking status (n = 55) Optimal (≤1 cm) Suboptimal (N1 cm) Clinical follow-up (n = 68) Recurrence Yes No PFS, median (range), mo Final status Dead Alive OS, median (range), mo a
54 (21-82) 47 (69.1) 21 (30.9) 5 (7.5) 25 (37.3) 37 (55.2) 7 (10.3) 51 (75) 10 (14.7) 23 (41.8) 32 (58.2)
50 (73.5) 18 (26.5) 28.7 (1.37-124.47) 42 (61.8) 26 (38.2) 47.8 (1.57-127.47)
Endometrioid, clear cell, mucinous, undifferentiated.
634 E7R (5′- TGG CAA GTG GCT CCT GAC), E8F (5′-CCT ATC CTG AGT AGT GGT AA), E8R (5′- TCC TCC ACC GCT TCT TGT). The target DNAwas denatured at 93°C for 5 minutes, whereafter, 40 cycles of amplification were performed in the PX2 thermal cycler (Thermo Electron Corp, Waltham, MA) under the following conditions: DNA denaturation at 95°C for 30 seconds, primer annealing at 58°C (exon 5), 56°C (exon 6), 55°C (exons 7-8) for 45 seconds, and primer extension at 72°C for 1 minute. For all reactions, the last extension step was prolonged with 7 minutes at 72°C. Before further use, 5 μL of the PCR product was run on an agarose gel to confirm the existence of a single product of the expected size. Denaturing high-performance liquid chromatography (DHPLC) was performed on a WAVE DNA fragment analysis system (Transgenomic, San Jose, CA). To enhance heteroduplex formation, we denature untreated PCR product at 95°C for 5 minutes, followed by and incubation at 65°C for 60 minutes. Five microliters were automatically loaded on the column and eluted with a linear acetonitrile gradient in 0.1 mmol/L triethylamine acetate buffer (pH 7.0) at a constant flow rate. Column temperatures were determined by a melting curve (Exon 5, 60.8°C; exon 6, 61.5°C; exon 7, 62.1°C; exon 8, 62.0°C). Eluted DNA fragments were detected by an UV-C detector. PCR products, which had shown a potential variant with denaturing high-performance liquid chromatography, were sequenced in both directions starting from a fresh PCR product. Before sequencing, the PCR products were purified using the Invisorb Spin PCRapid kit (Invitek, Berlin, Germany). Sequencing was
M. Mendiola et al. then performed using the BigDye Terminator Cycle Sequencing Kit and analyzed on an ABIPRISM 3100 (Applied Biosystems).
2.7. Statistical analysis Statistical analysis was carried out using 9.0 SPSS software for Windows (SPSS Inc, Chicago, IL). All tests were 2-sided and used a significance level of .05. Qualitative data were registered as absolute frequencies and percentages; quantitative data were expressed as median, range, and/or mean and standard deviation. Continuous variables were analyzed by analysis of variance and t test. Frequency tables were tested by Fisher test for comparison of discrete variables. Analysis of progression-free and OS data were carried out using Kaplan-Meier plots and log-rank test. The Cox proportional hazards model was used to evaluate the prognostic significance of pathological variables analyzed.
3. Results 3.1. Patients characteristics Characteristics of the 68 patients are shown in Table 1. Optimal surgery was a potent predictor of PFS and OS. Mean time to treatment failure was 50.5 months for patients with optimal surgery versus 31 months for patients with residual disease after surgery (P = .014). Patients with optimal
Fig. 1 Expression of Aurora kinases A and B in ovarian carcinoma. A, Moderately differentiated serous carcinoma (hematoxylin-eosin ×400). B, Expression of Aurora A in the tumor cells (×400). C, Expression of Aurora A is observed in the cytoplasm of the tumor cells (×400). D, Poorly differentiated ovarian carcinoma (hematoxylin-eosin ×400). E, Expression of Aurora B protein in the nuclei of the tumor cells (×400). F, Expression of Aurora B is frequently observed in mitotic cells (×400).
Aurora kinases in ovarian carcinoma
635
surgery had a mean OS of 77.26 months versus 46.68 months for patients with residual disease after surgery (P = .0006).
index (P = .64), overexpression of p53 (P = 1.000), and TP53 gene status (P = .32).
3.2. Immunohistochemistry
3.3. AURKA gene amplification analysis
Aurora A was expressed in 40 (58.8%) of 64 tumors, whereas Aurora B was expressed in 58 (85.3%) of 63 ovarian carcinomas in our study (Fig. 1). Aurora A and B expressions were not related to the histological type or the tumor grading. Of 64 tumors, 38 (59.4%) showed overexpression of p53 protein. Among the 38 patients with p53 overexpression, 29 (60.4%) had TP53 mutations. However, no significant correlations were found between p53 expression and TP53 gene status (P = .77). p53 protein expression was not related to the histological tumor type (P = .411), tumor grading (P = .452), PFS (P = .796), or OS (P = .406). The expression of Aurora A was associated with the proliferation index (Ki67). Thus, 80% of tumors with expression of Aurora A showed a high proliferation index (P = .033). The expression of Aurora A was not associated with overexpression of p53 (P = 1.000) or TP53 gene status (P = 1.000). Expression of Aurora B was frequently observed in mitotic cells but was not associated with the proliferation
We screened 58 ovarian carcinomas for AURKA amplification, 37 and 21 tumors with and without Aurora A protein expression, respectively. Overall, AURKA amplification was found in 16 (27.6%) carcinomas. Twenty-six (61.9%) cases without gene amplification showed expression of the protein. Amplification of AURKA was not related to the histological tumor type (P = .760) or the tumor grading (P = .571). No relation was found between AURKA amplification and expression of Aurora A (P = .63), Aurora B (P = .77), p53 (P = 1.000), TP53 gene status (P = .32), and proliferation index (P = .76).
3.4. TP53 mutations Of 68 patients, 19 (27.9%) showed mutant TP53. Most mutations (84.2%) were single nucleotide substitutions (point mutations). In this group, missense mutations were the most common (11 of 16 cases, 57.8%) followed by nonsense
Fig. 2 Kaplan-Meier curves showing associations statistically significant (P b .05) between expression of (A) Aurora A and PFS, (B) Aurora B and PFS, (C) Aurora A and OS, and (D) Aurora B and OS.
636 Table 2
M. Mendiola et al. Prognostic value of Aurora A and B expression adjusted for optimal surgery by Cox proportional hazards regression
Prognostic factor
Univariate Optimal surgery Aurora A expression Aurora B expression Multivariate Optimal surgery Aurora A expression Optimal surgery Aurora B expression
PFS
OS
P
HR
CI (95%)
P
HR
CI (95%)
.017 .025 .074
0.445 0.522 0.426
0.229-0.865 0.295-0.922 0.167-1.087
.001 .034 .027
0.270 0.506 0.300
0.122-0.600 0.270-0.949 0.103-0.869
.015 .029 .028 .348
0.433 0.490 0.456 0.597
0.221-0.850 0.259-0.929 0.227-0.917 0.204-1.752
.002 .023 .003 .075
0.281 0.447 0.291 0.369
0.126-0.629 0.222-0.896 0.128-0.660 0.123-1.108
mutations (4 cases, 21%). Transitions (12 of 16 cases, 75%) were more frequent than transversions (4 of 16 cases, 25%). G:C to A:T was the most frequent pattern of transition found in our series (8 cases, 42.1%). Of 8 G:C to A:T transitions, 4 were located in CpG sites that are known to be potential sites of DNA methylation. We also found 3 (15.7%) deletions that produce a frameshift mutation and 1 silent mutation (5.3%). In detail, 9 mutations were found in exon 5, 3 in exon 6, 4 in exon 7, and 3 in exon 8. In addition, we found a previously undescribed polymorphism (A:T N G:C) at codon 213 in exon 6 in 1 of the carcinomas (Appendix A). Mutations of the TP53 gene were not related to the histological tumor type (P = .55), tumor grading (P = .57), tumor recurrence (P = .53), Aurora A expression (P = .14), Aurora B expression (P = 1.0), PFS (P = .24), or OS (P = .62).
3.5. Association between Aurora A and B kinases and clinicopathological variables Tumors with Aurora A protein expression showed a lower rate of recurrence than those tumors without Aurora A expression (65% versus 91.7%, P = .019). In the univariate analysis, Kaplan-Meier method showed that patients with expression of Aurora A had an increased PFS compared with patients whose tumors did not express Aurora A protein (mean survival of 52.13 months: 95% confidence interval [CI], 35.56-68.70, versus 22.44 months: 95% CI, 14.7030.19; P = .023) (Fig. 2). Regarding OS, patients with expression of Aurora A showed a significant increased survival time compared to those patients with absence of Aurora A expression (mean survival of 39.81 months: 95% CI, 27.72-51.90, versus 73.18 months: 95% CI, 57.84-88.53; P = .03). The multivariate analysis using the Cox regression model adjusted to optimal surgery showed Aurora A protein expression as an independent prognostic factor for both PFS (P = .03, hazard ratio [HR] = 0.49 [0.26-0.93]) and OS (P = .02, HR = 0.45 [0.22-0-90]) (Table 2). Although the expression of Aurora B was not significantly associated to tumor recurrence (P = .32), patients with expression of Aurora B showed an increased PFS compared to those patients without expression of Aurora B (43.1 months [95%
CI, 30.6-55.6] versus 15.4 months [95% CI, 4.0-26.9]; P = .06). Moreover, those patients with expression of Aurora B showed and increased OS compared to patients without expression of Aurora B (66.1 months [95% CI, 53.4-78.8] versus 23.5 months [95% CI, 8.5-38.6]; P = .02) (Fig. 2). However, multivariate analysis using the Cox regression model adjusted to optimal surgery did not show Aurora B as an independent prognostic factor for PFS and OS (Table 2). Tumors with AURKA gene amplification showed an increased PFS compared to those tumors without AURKA gene amplification, although this difference was not statistically significant (45.3 months [95% CI, 31.2-59.4] versus 29.9 months [95% CI, 14.2-45.7]; P = .28). Patients with AURKA gene amplification showed a decreased OS compared to those patients without AURKA gene amplification. However, these differences were not statistically significant (82.51 months [95% CI, 71.9-93] versus 88.73 months [95% CI, 71.7-105.7]; P = .83).
4. Discussion In the present study, we have analyzed the prognostic value of the expression of Aurora kinases A and B at the DNA and protein levels in a series of ovarian carcinomas homogeneously treated with a combination of surgery and carboplatin/taxane-based chemotherapy. The expression and mutational status of TP53 and the proliferation index (Ki67) were also assessed in these cases. In our study, 58.8% of ovarian cancer specimens showed expression of Aurora A protein. There were no statistically significant differences in Aurora A protein expression among the different histopathological types of ovarian carcinomas. These results are in agreement with those previously reported in ovarian carcinoma that showed that expression of Aurora A protein is observed in 45% to 67% of these tumors [5]. AURKA gene amplification was detected in 27.6% of ovarian carcinomas examined. Previous studies reported that AURKA is amplified in 15% to 25% of ovarian cancer cell lines and primary tumors [21,22]. In our series, 61.9% of cases without gene amplification showed expression of the
Aurora kinases in ovarian carcinoma protein, suggesting that the expression of Aurora A is likely to be regulated not only by gene amplification but also by other mechanisms such as transcriptional activation and/or suppression of protein degradation, as it has been demonstrated in previous studies [23]. Our study demonstrated that Aurora B is frequently expressed in ovarian carcinomas (85.3%). Aurora B is reported to form complexes with inner centromere protein and survivin, and these complexes are thought to be involved in the regulation of chromosome alignment, segregation, and cytokinesis [24,25]. In the present study, the immunohistochemical expression of Aurora B was observed predominantly in the nucleus. Because recent studies have identified the histone H3 protein as an important substrate of Aurora B kinase [12,26], nuclear localization of Aurora B seems functionally important. Moreover, in our cases, mitotic cells were positive for Aurora B, indicating the functional involvement of Aurora B in the replication of the tumor cells. The relationship between Aurora A and p53 is an important factor in carcinogenesis. It has been reported that p53 interacts with Aurora A and suppresses its oncogenic activity in a transactivation-independent manner [27]. Aurora A is a key regulatory component of the p53 pathway, and previous studies have shown that high expression of Aurora A phosphorylates p53 and leads to an increased p53 degradation, facilitating oncogenic transformation [28]. Phosphorylation of p53 is associated with Aurora A regulated cycle progression, cell survival, and transformation. Thus, the deregulation of this mutual suppression mechanism between Aurora A and p53 may trigger checkpoint abnormalities and centrosome instability. Recent reports showed that the effects of Aurora A in cell growth could be highly variable depending on p53 status and other molecular partners [29] and that high p53 expression levels were correlated with a high level of Aurora A expression [30]. However, in our study, we found that Aurora A expression was not correlated with TP53 mutation or p53 protein overexpression. Finally, we examined the prognostic value of Aurora A and Aurora B expression in our series. In contrast with the assumption that Aurora A expression is a prognostic factor for poor survival in many tumor types [3,7], including ovarian carcinoma [5,31,32], we found that, paradoxically, patients with expression of Aurora A had longer PFS and OS. A possible explanation for the better outcome for the patients in the group of tumors with expression of Aurora A protein could be based on the fact that high Aurora A expression correlates with a higher proliferation index, and therefore, these high proliferative tumors could better respond to chemotherapy. However, our data should be interpreted warily because Kulkarni et al [32] have recently demonstrated that expression of Aurora A was strongly predictive of shorter disease-free survival, specially in early stage ovarian carcinomas. In the univariate analysis, patients with Aurora A and B expression showed an increased PFS (P = .023 and .06, respectively, log-rank test) and OS (P =
637 .03 and .02, respectively, log-rank test). Moreover, the multivariate analysis adjusted to optimal surgery by Cox proportional hazards regression showed Aurora A expression as an independent prognostic factor for PFS (P = .03) and OS (P = .02). Similar results have been recently reported by Lassmann et al [33] showing that high Aurora A protein expression was associated with improved OS in patients with stage III ovarian cancer with optimal debulking and receiving taxol/carboplatin therapy. Interestingly, these authors found that the expression of Aurora A protein was associated with poor prognosis only in patients receiving non–taxane-based chemotherapy. These findings are of great interest because in vitro studies have previously found that overexpression of Aurora A induced chemoresistance to taxanes and platinum agents [8,9]. Our report matches Lassmann et al [33], probably because all the patients in our study have received a taxane agent, and the distribution by FIGO stage includes predominantly stage III and IV tumors, but we agree with this group in the necessity for investigating predictive molecular marker candidates in situ to complement functional in vitro testing with the clinicopathological variables associated with individual cancer patients. Therefore, our results should be confirmed in a larger series of ovarian carcinomas.
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.humpath. 2008.10.011.
References [1] Cannistra SA. Cancer of the ovary. N Engl J Med 2004;351:2519-29. [2] McGuire WP, Hoskins WJ, Brady MF, et al. Cyclophosphamide and cisplatin compared with paclitaxel and cisplatin in patients with stage III and stage IV ovarian cancer. N Engl J Med 1996;334:1-6. [3] Miyoshi Y, Iwao K, Egawa C, Noguchi S. Association of centrosomal kinase STK15/BTAK mRNA expression with chromosomal instability in human breast cancers. Int J Cancer 2001;92:370-3. [4] Sakakura C, Hagiwara A, Yasuoka R, et al. Tumour-amplified kinase BTAK is amplified and overexpressed in gastric cancers with possible involvement in aneuploid formation. Br J Cancer 2001;84: 824-31. [5] Gritsko TM, Coppola D, Paciga JE, et al. Activation and overexpression of centrosome kinase BTAK/Aurora-A in human ovarian cancer. Clin Cancer Res 2003;9:1420-6. [6] Li D, Zhu J, Firozi PF, et al. Overexpression of oncogenic STK15/ BTAK/Aurora A kinase in human pancreatic cancer. Clin Cancer Res 2003;9:991-7. [7] Jeng YM, Peng SY, Lin CY, Hsu HC. Overexpression and amplification of Aurora-A in hepatocellular carcinoma. Clin Cancer Res 2004;10:2065-71. [8] Hata T, Furukawa T, Sunamura M, et al. RNA interference targeting aurora kinase a suppresses tumor growth and enhances the taxane chemosensitivity in human pancreatic cancer cells. Cancer Res 2005;65:2899-905.
638 [9] Anand S, Penrhyn-Lowe S, Venkitaraman AR. AURORA-A amplification overrides the mitotic spindle assembly checkpoint, inducing resistance to Taxol. Cancer Cell 2003;3:51-62. [10] Dewar H, Tanaka K, Nasmyth K, Tanaka TU. Tension between two kinetochores suffices for their bi-orientation on the mitotic spindle. Nature 2004;428:93-7. [11] Katayama H, Ota T, Jisaki F, et al. Mitotic kinase expression and colorectal cancer progression. J Natl Cancer Inst 1999;91:1160-2. [12] Ota T, Suto S, Katayama H, et al. Increased mitotic phosphorylation of histone H3 attributable to AIM-1/Aurora-B overexpression contributes to chromosome number instability. Cancer Res 2002;62:5168-77. [13] Hardisson D, Moreno-Bueno G, Sanchez L, et al. Tissue microarray immunohistochemical expression analysis of mismatch repair (hMLH1 and hMSH2 genes) in endometrial carcinoma and atypical endometrial hyperplasia: relationship with microsatellite instability. Mod Pathol 2003;16:1148-58. [14] Tanaka T, Kimura M, Matsunaga K, Fukada D, Mori H, Okano Y. Centrosomal kinase AIK1 is overexpressed in invasive ductal carcinoma of the breast. Cancer Res 1999;59:2041-4. [15] Keen N, Taylor S. Aurora-kinase inhibitors as anticancer agents. Nat Rev Cancer 2004;4:927-36. [16] López-Ríos F, Chuai S, Flores R, et al. Global gene expression profiling of pleural mesotheliomas: overexpression of aurora kinases and P16/CDKN2A deletion as prognostic factors and critical evaluation of microarray-based prognostic prediction. Cancer Res 2006;66:2970-9. [17] Anttila MA, Ji H, Juhola MT, Saaeikoski SV, Syrjänen KJ. The prognostic significance of p53 expression quantified by computerized image analysis in epithelial ovarian cancer. Int J Gynecol Pathol 1999;18:42-51. [18] Sengupta PS, McGown AT, Bajaj V, et al. p53 and related proteins in epithelial ovarian cancer. Eur J Cancer 2000;36:2317-28. [19] Dogan E, Saygili U, Tuna B, et al. p53 and mdm2 as prognostic indicators in patients with epithelial ovarian cancer: a multivariate analysis. Gynecol Oncol 2005;97:46-52. [20] Green JA, Berns EM, Coens C, et al. Alterations in the p53 pathway and prognosis in advanced ovarian cancer: a multi-factorial analysis of the EORTC Gynaecological Cancer group (study 55865). Eur J Cancer 2006;42:2539-48. [21] Moreno-Bueno G, Sanchez-Estevez C, Cassia R, et al. Differential gene expression profile in endometrioid and nonendometrioid endometrial carcinoma: STK15 is frequently overexpressed and
M. Mendiola et al.
[22]
[23] [24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
amplified in nonendometrioid carcinomas. Cancer Res 2003;63: 5697-702. Tanner MM, Grenman S, Koul A, et al. Frequent amplification of chromosomal region 20q12-q13 in ovarian cancer. Clin Cancer Res 2000;6:1833-9. Fu S, Hu W, Kavanagh JJ, Bast Jr RC. Targeting Aurora kinases in ovarian cancer. Expert Opin Ther Targets 2006;10:77-85. Adams RR, Maiato H, Earnshaw WC, Carmena M. Essential roles of Drosophila inner centromere protein (INCENP) and aurora B in histone H3 phosphorylation, metaphase chromosome alignment, kinetochore disjunction, and chromosome segregation. J Cell Biol 2001;153:865-80. Wheatley SP, Henzing AJ, Dodson H, Khaled W, Earnshaw WC. Aurora-B phosphorylation in vitro identifies a residue of survivin that is essential for its localization and binding to inner centromere protein (INCENP) in vivo. J Biol Chem 2004;279:5655-60. Crosio C, Fimia GM, Loury R, et al. Mitotic phosphorylation of histone H3: spatio-temporal regulation by mammalian Aurora kinases. Mol Cell Biol 2002;22:874-85. Chen SS, Chang PC, Cheng YW, Tang FM, Lin YS. Suppression of the STK15 oncogenic activity requires a transactivation-independent p53 function. EMBO J 2002;21:4491-9. Katayama H, Sasai K, Kawai H, et al. Phosphorylation by aurora kinase A induces Mdm2-mediated destabilization and inhibition of p53. Nat Genet 2004;36:55-62. Mao JH, Wu D, Perez-Losada J, et al. Crosstalk between Aurora-A and p53: frequent deletion or downregulation of Aurora-A in tumors from p53 null mice. Cancer Cell 2007;11:161-73. Zhang H, Jin Y, Chen X, et al. Papillomavirus type 16 E6/E7 and human telomerase reverse transcriptase in esophageal cell immortalization and early transformation. Cancer Lett 2007;245:184-94. Landen Jr CN, Lin YG, Immaneni A, et al. Overexpression of the centrosomal protein Aurora-A kinase is associated with poor prognosis in epithelial ovarian cancer patients. Clin Cancer Res 2007;13: 4098-104. Kulkarni AA, Loddo M, Leo E, et al. DNA replication licensing factors and aurora kinases are linked to aneuploidy and clinical outcome in epithelial ovarian carcinoma. Clin Cancer Res 2007;13:6153-61. Lassmann S, Shen Y, Jutting U, et al. Predictive value of Aurora-A/ STK15 expression for late stage epithelial ovarian cancer patients treated by adjuvant chemotherapy. Clin Cancer Res 2007;13: 4083-91.
