Clinical Cancer Research The Future of Cancer Research: Science and Patient Impact
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online
Clinical Cancer Research 13, 4083-4091, July 15, 2007. doi: 10.1158/1078-0432.CCR-06-2775
© 2007 American Association for Cancer Research
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lassmann, S.
Right arrow Articles by Werner, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lassmann, S.
Right arrow Articles by Werner, M.

Imaging, Diagnosis, Prognosis

Predictive Value of Aurora-A/STK15 Expression for Late Stage Epithelial Ovarian Cancer Patients Treated by Adjuvant Chemotherapy

Silke Lassmann1, Yi Shen1,5, Uta Jütting3, Philipp Wiehle2, Axel Walch4, Gerald Gitsch2, Annette Hasenburg2 and Martin Werner1

Authors' Affiliations: 1 Institute of Pathology, and 2 Department of Obstetrics and Gynecology, University Hospital, Albert-Ludwigs-University Freiburg, Freiburg, Germany; 3 Institute for Biomathematics and Biometry, 4 Institute of Pathology, GSF-National Research Center for Environment and Health, Neuherberg, Germany; and 5 Department of Obstetrics and Gynecology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China

Requests for reprints: Silke Lassmann, Institut für Pathologie, Universitätsklinikum Freiburg, Breisacherstr. 115a, 79106 Freiburg. Phone: 49-761-270-8031; Fax: 49-761-270-8004; E-mail: silke.lassmann{at}uniklinik-freiburg.de.


   Abstract
Top
 Abstract
Materials and Methods
 Results
 Discussion
 References
 
Purpose: To investigate the expression and regulation of the centrosomal kinase Aurora-A/STK15 (AURKA) in epithelial ovarian cancers and to determine the prognostic and predictive value of this marker for patients with late stage epithelial ovarian cancer treated by distinct adjuvant chemotherapies.

Experimental Design: Archival resection specimens of epithelial ovarian cancers (n = 115) and nonneoplastic ovaries (n = 28) were analyzed for AURKA mRNA and protein expression by microdissection and quantitative reverse transcriptase-PCR and immunohistochemistry. AURKA DNA copy numbers were measured by fluorescence in situ hybridization in 37 cases. Statistical evaluation was done with respect to clinicopathologic variables, disease-free survival, and overall survival.

Results: AURKA mRNA expression was significantly elevated in cancers (P < 0.001) and correlated with AURKA protein expression (P = 0.0134). Overexpression of AURKA protein was detected in 68 of 107 (63.5%) cases and was linked with increased AURKA DNA copy numbers (P = 0.0141) and centromere 20 aneusomy (P = 0.0137). Moreover, AURKA overexpression was associated with improved overall survival in optimal debulked patients receiving taxol/carboplatin therapy (n = 43, P = 0.018). Finally, in an exploratory approach, patients receiving non–taxane-based therapy, AURKA overexpression was predictive for worse overall survival (n = 30, P = 0.049).

Conclusions: AURKA overexpression is seen in the majority of late stage epithelial ovarian cancers, most likely due to increased AURKA DNA copy numbers and/or chromosome 20 aneusomy. Importantly, AURKA overexpression may differentially affect taxane and non–taxane-based adjuvant therapy responses. The study sheds new light on AURKA expression and regulation in epithelial cancers in vivo and specifically shows its value as a clinically relevant marker and as a potential therapeutic target per se.

Ovarian cancer represents the second major gynecological malignancy in women, is diagnosed in most patients at a late stage, and has an associated 5-year survival rate of only 20% to 30% (1–3). The majority of ovarian cancers are derived from the epithelium and are histopathologically classified into eight major subgroups, primarily including the serous tumors (4–6). Standard treatment involves surgical resection of the tumor mass, with complete reduction of the tumor (i.e., optimal debulking) not being possible in a minor fraction of patients due to the advanced stage (7, 8). Most patients receive subsequent adjuvant chemotherapy, with a combination of taxane (e.g., paclitaxel, docetaxel) and platinum (e.g., carboplatin, cisplatin) compounds representing the most frequently applied regimens (3, 5, 7–10). Taxanes stabilize microtubules, resulting in a G2-M cell cycle arrest and induction of apoptosis. Platinum drugs induce DNA cross-links, which are recognized by the DNA repair factors, and also result in apoptosis. Taxol/carboplatin or carboplatin therapy alone or in combination with other drugs, such as topotecan, an inhibitor of topoisomerase I and hence DNA synthesis, are widely used for patients with ovarian cancer due to their highly cytotoxic effects on proliferating and cycling tumor cells. Although the debulking status of the primary surgical resection is a potent prognostic factor, thus far there are no other validated prognostic variables to allow improved patient stratification for adjuvant treatment, especially none associated with a predictive effect for the specific chemotherapeutic regimens (11).

Recent studies have shown that DNA gains of the chromosomal region 20q13 and overexpression of the centrosomal kinase Aurora-A/STK15 (AURKA), the gene of which is found within the 20q13 region, are hallmarks of ovarian cancer (12). AURKA is a putative oncogene (13) and is involved in centrosome dysregulation and genomic instability in various human cancers (14, 15). Overexpression of AURKA has been associated with tumor development and progression in gastrointestinal cancers (16–19), bladder cancer (20), and gynecological cancers (21) including ovarian cancer (22) as well as ovarian cancer cell lines (23, 24). However, the clinical relevance of AURKA overexpression for patient outcome has only been addressed by few studies: thus, in epithelial cancers of the esophagus (25) and of the head and neck (26), AURKA (over-)expression represented a survival factor for tumor cells and a negative prognostic molecular marker. However, these studies did not allow for the evaluation of a predictive effect of AURKA expression in view of the specific, but diverse, adjuvant chemotherapeutic drugs given to these patients.

Interestingly, recent in vitro data showed the interaction of AURKA with taxanes: (a) inhibition of AURKA expression by small interfering RNA enhanced the chemosensitivity of pancreatic cancer cells to taxanes (27) and (b) AURKA overexpression induced resistance to taxol in another in vitro study (28). Moreover, high levels of AURKA in ovarian cancer cells protected the cells from apoptosis after treatment with chemotherapeutic agents, such as paclitaxel or cisplatin (29). These studies imply that overexpression of AURKA may overcome mitotic spindle checkpoint controls induced by the chemotherapeutic agents, overriding apoptotic mechanisms and inducing chemotherapeutic resistance. In view of the lack of prognostic/predictive factors for ovarian cancer patients treated with taxane and/or platinum therapy, these in vitro aspects are a particular hint for the potential importance of AURKA expression as a clinically relevant marker for patients with ovarian cancer. Moreover, with the current efforts to develop Aurora kinase inhibitors (30–37), there is also the question of whether or not the application of Aurora inhibitors may provide valuable chemotherapeutic agents for ovarian cancer, especially as a combinatorial therapy with existing agents (38).

The present study addressed the issue of whether AURKA represents a clinically relevant molecular marker for patients with late stage epithelial ovarian cancer, in particular, with respect to the predictive value of AURKA for the response of patients to different adjuvant chemotherapy protocols. For this purpose, we preferentially selected cases with late stage (International Federation of Gynecology and Obstetrics III) epithelial ovarian cancers of the serous type (6). AURKA mRNA and protein expression were thus examined in archival resection specimens of 115 patients with ovarian cancer, using AURKA-specific quantitative reverse transcriptase PCR (QRT-PCR) and immunohistochemistry on serial sections of formalin-fixed and paraffin-embedded tissue specimens, respectively. In addition, AURKA DNA copy numbers were determined in 37 cases on serial sections by fluorescence in situ hybridization (FISH) to examine the regulation of AURKA expression at the DNA level. Finally, AURKA expression was correlated with clinicopathologic variables and the clinical outcome in patients with ovarian cancer, by stratifying patients according to known prognostic factors such as debulking status and different adjuvant treatment protocols.


   Materials and Methods
Top
 Abstract
 Materials and Methods
Results
 Discussion
 References
 
Patients and tissues. The study included a total of 143 patients treated at the Department of Obstetrics and Gynecology, University Hospital Freiburg (Freiburg, Germany), with a median follow-up time of 22.5 months (range, 1-82 months). Cases with International Federation of Gynecology and Obstetrics stage III and epithelial ovarian cancers of the serous type (6) were preferentially selected to obtain a "homogeneous" patient group for statistical evaluation of prognostic/predictive associations. Formalin-fixed and paraffin-embedded tissue specimens from all cases were derived from routinely processed primary resection specimens (archive, Institute of Pathology, University Hospital Freiburg). Of the 143 cases, resection specimens from 115 ovarian cancers and a control group of 28 nonneoplastic ovaries were analyzed. The study was approved by the local ethics committee (no. 97/05; Ethik-Kommission der Albert-Ludwigs-Universität, Freiburg, Germany). Clinicopathologic data from all cases are summarized in Table 1 (note that the distribution of variables was influenced by the specific preselection of cases).


View this table:
[in this window]
[in a new window]

  		Table 1. Clinicopathologic data

 
Routinely processed H&E-stained tissue sections of the invasive carcinomas were reevaluated (6, 39), and representative tissue areas were selected for further analyses, as detailed below. First, two 10-µm sections were cut from the selected tissue blocks for tumor cell microdissection and QRT-PCR. Second, the same tissue blocks were then used for manual preparation of tissue microarrays (TMA) of core biopsies (0.2 mm diameter) from the invasive carcinomas.

Immunohistochemistry. For detection of AURKA protein expression as previously reported by Reiter et al. (26), 4-µm-thick sections of the TMAs were cut, dried overnight at 60°C, and subjected to deparaffinization in xylene, graded alcohols (2x 100%, 95%, 75%, and 50%), and water. Antigen retrieval was done by boiling TMA sections in Tris-EDTA buffer (pH 9) using a pressure cooker. Incubation time with primary antibody (1:50, Aurora Kinase 2, clone JLM28; Loxo/Novocastra) was 60 min, followed by secondary antibody and detection using the LSAB-Fast red system (DakoCytomation). Stainings were done on a Dako Autostainer (DakoCytomation). The scoring of cytoplasmic AURKA staining patterns was done according to the study by Reiter et al. (26), with minor modifications: three high-power fields (40x magnification) of representative areas of each TMA core biopsy were selected for scoring of AURKA-positive tumor cells. The mean count of AURKA-positive tumor cells was grouped as follows: score 0, all tumor cells were negative; score 1, faint cytoplasmic expression in tumor cells with elevated expression in <10% of individual tumor cells; score 2, cytoplasmic expression in tumor cells with strong overexpression in 10% to 30% of individual tumor cells; score 3, cytoplasmic expression in tumor cells with frequent (>30%) and strong overexpression of individual tumor cells. All TMA sections were scored by two investigators (S. Lassmann and Y. Shen) in a blinded fashion. Correlations between the immunohistochemical scoring of the two investigators for ovarian cancers was significant at P < 0.0001 (r = 0.77391). The same scoring system was applied for the analysis of 24 of 28 nonneoplastic ovaries and showed negative (score 0) AURKA protein expression in 22 of 24 (91.7%) and focal faint cytoplasmic AUKRA protein expression (score 1) in 2 of 24 (8.3%) nonneoplastic ovaries.

Microdissection and QRT-PCR. Microdissection and QRT-PCR were essentially done as described before (40). In brief, for the evaluation of AURKA mRNA expression, two 10-µm sections of the tissue specimens were deparaffinized using xylene and graded alcohols (2x 100%, 95%, 75%, and 50%) and water. Subsequently, tissue sections were briefly stained with instant hematoxylene (Shandon) and invasive tumor cells were microdissected under microscopic surveillance using fine needles. Microdissected tumor cells were then placed into an Eppendorf tube and RNA isolation was done using the "Masterpure RNA Isolation Kit" (Epigenomics/Biozym) according to the instructions of the manufacturer. A 30-min DNase digestion step was included to avoid the contamination of RNA samples with genomic DNA. Purified total RNA was measured on a conventional spectrophotometer (ND1000, Peqlab) and 1 µg of RNA was used for reverse transcription by Moloney murine leukemia virus reverse transcriptase (Invitrogen) into 30 µL of cDNA. A negative control reaction (water instead of RNA) was included for each set of cDNA syntheses. For QRT-PCR using the ABI7000 system (TaqMan, AppliedBiosystems), 3 µL of each cDNA was incubated in a 30 µL reaction volume, containing primer and probes for the target gene AURKA (40) and the reference gene TBP (40–42), as well as universal Master Mix (AppliedBiosystems), with cycling conditions as follows: 50°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 15 s and 60°C for 1 min. The QRT-PCR assays were previously validated in our laboratory for specificity, sensitivity, and efficiency (40). The comparative Ct-method was used to determine target gene expression levels (43).

FISH. For examination of AURKA DNA copy numbers, serial sections (5 µm) of the TMAs were hybridized with a combination of custom-made fluorescent hybridization probes specific for AURKA and for centromere 20 (chromosome enumeration probe, CEP20), according to previously published protocols (44). The AURKA-specific probe (red, Cy3-labeled) was prepared from two overlapping Pac/BAC clones (RP11-61B1 and RP5-1167H4) and the CEP20-specific probe (green, FITC-labeled) was prepared from two adjacent BAC clones (RP11-815L24 and RP11-75I24; Chrombios).

Prior to probe hybridization, TMA sections were subjected to deparaffinization (2x xylene, 100%, 95%, 70%, and 50% ethanol) and pretreatment in citrate buffer (pH 6.0) in a microwave oven (180 W) for 20 min followed by Pronase E (0.05%) digestion for 3 min at 37°C. Subsequently, TMA sections were denatured in 50% formamide at room temperature for 15 min and in 70% formamide at 75°C for 5 min. Slides were immediately immersed in ice-cold ethanol (70%, 95%, and 100%, 5 min each) and dried at 37°C. In the meantime, fluorescent hybridization probes had been denatured at 75°C for 5 min and were pipetted onto dried TMA sections, followed by incubation for 16 h at 37°C. Slides were then washed in 2x SSC at room temperature and in 2x SSC at 37°C each for 2 min and finally counterstained with 4',6-diamidino-2-phenylindole for 3 min. Coverglassed (VectaShield, Molecular Probes) slides were stored at –20°C until analysis.

Microscopic analysis of FISH sections was done using a fluorescence microscope with ApoTome imaging system for three-dimensional visualization ("Zeiss Axioplan2 Imaging Microscope" equipped with a PlanApochromat x63/NA1.4 oil objective; Carl Zeiss MicroImaging GmbH). Image stacks were taken at intervals from representative tumor areas and the AxioVision software converted the image stacks into a three-dimensional view, which was then assessed for gene- and centromere-specific signals. For each case, at least 60 cells were counted for AURKA- and CEP20-specific signals (presented as mean signals per cell for each case). The mean ratio of AURKA- to CEP20-specific signals was calculated from the mean AURKA- and CEP20-specific signals per cell in each case.

Statistical evaluation. Clinicopathologic variables of all cases (patient age, tumor stage, histologic tumor type, and tumor grading, debulking status, and first-line adjuvant chemotherapy), experimental data of AURKA protein (immunohistochemistry, scores 0-3) and mRNA (QRT-PCR, defined by raw dCT-values; ref. 43) expression as well as AURKA DNA copy numbers and CEP20 signals (FISH, mean AURKA and CEP20 signals, and mean AURKA/CEP20 ratios) were evaluated.

Correlations between variables were done using the Pearson correlation coefficient. ANOVA or t test was applied to compare all continuous variables either for different classes or for two-class cases, respectively. Frequency tables were tested by {chi}2 test for comparison of discrete variables. Step-wise Cox regression analysis was applied to test the significance of variables with survival time or disease-free survival time. For discrete variables, Kaplan-Meier curves for different strata were plotted for disease-free and overall survival. Log-rank test was used to test the significance of different survival curves. All statistical evaluations were done at the 95% level.


   Results
Top
 Abstract
 Materials and Methods
 Results
Discussion
 References
 
AURKA mRNA and protein expression. AURKA mRNA expression was determined by QRT-PCR of microdissected tumor cells in 115 patients with ovarian cancer and in nonneoplastic ovaries from 28 patients undergoing non–cancer-related surgery. QRT-PCR was successful in 101 of 115 (87.8%) of the patients with ovarian cancers and in 28 of 28 (100%) of the patients with nonneoplastic ovaries. Relative AURKA mRNA expression levels differed significantly between nonneoplastic ovaries and ovarian cancers irrespective of stage (P < 0.001; Fig. 1A ). AURKA mRNA expression also varied within the group of nonneoplastic ovaries or, more prominently, in the group of ovarian cancers, but this was independent of the patient's age (P = 0.2626). The range of AURKA mRNA expression in ovarian cancers was not significantly related to tumor stage (P = 0.122; Fig. 1A) or the tumor grading (P = 0.3726; Fig. 1B), despite a tendency for higher AURKA mRNA levels with increasing grade. No correlation was done between histologic tumor types, as cases with serous histology had been specifically selected for the study (Materials and Methods; Table 1).


Figure 1
View larger version (7K):
[in this window]
[in a new window]

  		Fig. 1. AURKA mRNA expression levels in nonneoplastic ovaries and ovarian cancers. AURKA-specific QRT-PCR was done using microdissected tumor cells of ovarian resection specimens as in Materials and Methods. Relative AURKA mRNA levels for each case ({circ}, bullet), for which QRT-PCR had been successful (129 of 143 cases). A, comparison of nonneoplastic ovaries (No) with ovarian cancers of stage I, stage II, and stage III. B, comparison of ovarian cancers between tumor grading, with grade 1 (G1), grade 2 (G2), grade 3 (G3), and grade 4 (G4).

 
AURKA protein expression was evaluated in ovaries from 24 of 28 patients undergoing non–cancer-related surgery and 107 of 115 (93.0%) patients with ovarian cancer using immunohistochemical scoring as specified in Materials and Methods. In contrast with nonneoplastic ovaries (Materials and Methods), immunohistochemical staining of AURKA protein in ovarian cancers revealed distinct AURKA protein expression patterns (Fig. 2, left ): negative staining (score 0) was observed in 16 of 107 (14.9%) patients and positive staining was seen in 91 of 107 (85.1%) ovarian cancer cases. For positively stained cases, 23 of 107 (21.5%) had a score of 1, 44 of 107 (41.1%) had a score of score 2, and 24 of 107 (22.4%) had a score of 3. AURKA protein expression was not related to the histologic tumor type (P = 0.1100; due to specific selection of serous type ovarian cancers) or the tumor grading (P = 0.6830). Except for the occasional weakly stained lymphocytes or granulocytes, AURKA staining was negative for the tumor surrounding stromal cells (Fig. 2, left). In ovarian cancers, the levels of AURKA mRNA and protein expression were significantly correlated (P = 0.0134) in 101 cases in which both QRT-PCR and immunohistochemistry had been successful.


Figure 2
View larger version (123K):
[in this window]
[in a new window]

  		Fig. 2. AURKA protein expression in ovarian cancer and associated DNA copy numbers. Three representative cases for which matching AURKA protein analysis (immunohistochemistry, left; magnification, x40) and analysis of AURKA DNA copy numbers (FISH, right; magnification, x63) were done (Materials and Methods; cases are also listed in Table 2, including detailed FISH, immunohistochemistry, and QRT-PCR data). Patient 93 has no AURKA protein expression within tumor cells (score 0) and shows normal AURKA DNA copy numbers (FISH with two AURKA/red and CEP20/green-specific signals). Patient 136 had weak cytoplasmic AURKA protein expression in most tumor cells and a low number of tumor cells overexpressing AURKA protein (score 2). Corresponding FISH analysis showed increased AURKA DNA copy numbers (red) and concomitant centromere 20 aneusomie (green) in some tumor cells. Patient 75 had weak cytoplasmic AURKA protein expression in most tumor cells and a high number of tumor cells overexpressing AURKA protein (score 3). Corresponding FISH analysis shows frequent AURKA DNA copy number increases (red) paralleled by centromere 20 aneusomie (green). Right, for all FISH images, cell nuclei were counterstained by 4',6-diamidino-2-phenylindole (blue).

 

View this table:
[in this window]
[in a new window]

  		Table 2. Results of FISH analysis of AURKA copy numbers in 37 ovarian cancer cases with corresponding AURKA mRNA and protein expression levels

 
Examination of AURKA DNA copy numbers and correlation with AURKA expression. In order to examine whether high AURKA expression was caused by AURKA gene amplification, we investigated AURKA-specific DNA copy numbers in 37 ovarian cancers using FISH (Materials and Methods). All FISH results are listed in detail in Table 2 , together with corresponding AURKA mRNA and protein levels. Representative FISH images are shown in Fig. 2 (right).

Detailed evaluation of AURKA- and CEP20-specific signals in at least 60 tumor cells per case revealed >2 AURKA copies for all cases, ranging from 2.33 to 6.18 mean AURKA-specific signals per tumor cell. Similarly, mean CEP20 signals were also >2, but the range was smaller (1.92-4.25 mean CEP20-specific signals per tumor cell). This resulted in mean FISH ratios (AURKA- to CEP20-specific signals per case) ranging from 1.08 to 2.0. According to a previously published cutoff for FISH evaluation in ovarian cancer (12), only low-level (FISH ratio, 1.5-3), but not high-level (FISH ratio, >3) amplification of AURKA was seen in our group of patients with ovarian cancer. Mean AURKA- and CEP20-specific FISH signals were closely correlated (P < 0.0001).

Whereas the mean AURKA-specific FISH ratio did not correlate with AURKA mRNA (P = 0.1585) or protein (P = 0.8611) expression, separate analyses of the mean AURKA- and CEP20-specific signals revealed a significant association with AURKA protein expression (P = 0.0141 and P = 0.0137, respectively). No association of mean AURKA (P = 0.1592) or CEP20-specific (P = 0.5525) signals with AURKA mRNA expression was seen, even though AURKA mRNA and protein expression remained significantly correlated within this subset of 37 cases (P = 0.0296).

Evaluation of the prognostic and predictive effects of AURKA expression. In order to test whether AURKA mRNA and/or protein levels may represent a prognostic and/or predictive marker for the clinical outcome of patients with ovarian cancer, AURKA expression levels were correlated with disease-free and overall survival in patients with International Federation of Gynecology and Obstetrics stage III ovarian cancer (n = 101) by stratifying protein expression into none or marginal (score 0 + 1), low (score 2), and high (score 3) protein expression.

No association was seen for AURKA mRNA expression with disease-free survival, neither in the entire group of patients (n = 100; P = 0.6977) nor in patients having been optimally debulked and receiving adjuvant taxol/carboplatin therapy (n = 36; P = 0.7361). Similarly, no association was seen for AURKA mRNA expression with overall survival in the same patient groups (stage III, n = 80, P = 0.5343; stage III, optimal debulking, taxol/carboplatin, n = 33, P = 0.9851).

Similarly, AURKA protein expression within the entire group of stage III patients was not associated with disease-free (n = 76, P = 0.690) or overall (n = 95, P = 0.426) survival (Fig. 3A and C ). Stratifying the group of stage III patients with ovarian cancer into those with optimal debulking and taxol/carboplatin therapy did not reveal the effect of AURKA protein expression on disease-free survival (n = 40, P = 0.157; Fig. 3B). However, a significant effect was seen in this patient group for high AURKA protein expression and improved overall survival (n = 43, P = 0.018; Fig. 3D).


Figure 3
View larger version (13K):
[in this window]
[in a new window]

  		Fig. 3. Association of AURKA protein expression with the clinical outcome of patients with ovarian cancer. The graphs show Kaplan-Meier curves for disease-free survival of patients with stage III ovarian cancer (A) and the subgroup of stage III patients with optimal debulking and adjuvant taxol/carboplatin chemotherapy (B). Kaplan-Meier curves for overall survival for patients with stage III ovarian cancer (C) and the subgroup of stage III patients with optimal debulking and adjuvant taxol/carboplatin chemotherapy (D). Coding of curves was according to immunohistochemistry scores, with full lines (score 0 + 1), interrupted lines (score 2), and dotted line (score 3). P values of the log-rank test are indicated. Circles, censored patients.

 
Finally, in an exploratory approach in view of low patient numbers, stage III patients treated with taxol/carboplatin chemotherapy were compared with those having received carboplatin-based chemotherapy not including taxanes with respect to overall survival. In the taxol/carboplatin-treated group (Fig. 4A ), patients with AURKA protein overexpression (immunohistochemistry score, 2 + 3) seemed to perform better than those with no or marginal AURKA protein expression (immunohistochemistry score, 0 + 1), but this was not significant (n = 56, P = 0.0997). In contrast, in the carboplatin-based treatment group (Fig. 4B), patients with AURKA protein overexpression (immunohistochemistry score, 2 + 3) had a significantly worse survival as those with no or marginal (immunohistochemistry score, 0 + 1) AURKA protein expression (n = 30, P = 0.0496).


Figure 4
View larger version (9K):
[in this window]
[in a new window]

  		Fig. 4. Association of AURKA protein expression with the clinical outcome in ovarian cancer patients receiving different adjuvant chemotherapy treatment. The graphs show Kaplan-Meier curves of an exploratory approach to test overall survival in patients with stage III ovarian cancer and either taxol/carboplatin chemotherapy (A) or carboplatin-based chemotherapy not containing taxanes (B). Patients were stratified according to immunohistochemistry scores into those with none or marginal AURKA protein expression (scores 0 + 1; full line) and those with AURKA protein overexpressing tumor cells (immunohistochemistry scores 2 + 3; interrupted line).

 

   Discussion
Top
 Abstract
 Materials and Methods
 Results
 Discussion
References
 
To our knowledge, the present study is the first to have analyzed the role of AURKA in epithelial ovarian cancer at the DNA, mRNA, and protein levels as well as to have examined the prognostic/predictive value of AURKA expression in a well-defined group of patients with ovarian cancer.

AURKA is a centrosomal kinase associated with various human cancers (13–21), including ovarian cancer (12, 22–24), and its overexpression is associated with poor disease outcome in esophageal (25) and head and neck tumors (26). Moreover, polymorphisms of AURKA seem to be linked to the risk of several cancers (45), including ovarian cancer (46). The question of whether or not AURKA expression represents a prognostic/predictive marker for patients with ovarian cancer has, however, not been addressed. In addition, it is thus far not clearly established whether AURKA in ovarian cancers may represent a therapeutic target per se (38).

In contrast to the only previous detailed study on AURKA protein expression and activity in ovarian cancer tissue specimens (22), we therefore specifically selected a patient group with stage III (n = 101) epithelial ovarian cancers and examined AURKA mRNA and protein expression as well as the regulation of AURKA expression at the DNA level. Moreover, of the 101 patients, 68 had been treated by the standard adjuvant taxol/carboplatin chemotherapy whereas the others received carboplatin/topotecan (n = 15, participating in a clinical trial), carboplatin alone (n = 10), or carboplatin with cyclophosphamide (n = 3), and a median of 22.5 months of follow-up was available for investigating the role of AURKA for therapy-related survival.

We found significantly increased AURKA mRNA expression in ovarian cancers as compared with nonneoplastic ovaries. This was independent of the patient's age, and hence, a potential hormonal regulation of AURKA expression (46) was not further assessed. AURKA mRNA levels in ovarian cancers seemed to increase with stage and tumor grading and were significantly correlated with protein levels (P = 0.0134). Whereas no AURKA protein expression was observed in 91.7% of the investigated nonneoplastic ovaries of patients undergoing non–cancer-related surgery, AURKA protein expression was detected in the tumor cells of the majority (85.1%) of ovarian cancer cases. For the latter, AURKA protein expression varied between marginal cytoplasmic expression in 21.5% of the cases and cytoplasmic expression together with strongly overexpressing tumor cells in 63.5% of the cases. AURKA protein expression patterns were not linked to the histologic tumor type (due to selection of serous type ovarian cancers) or the tumor grading.

At the DNA level, we found low-level amplification of the AURKA gene locus in 11 of 37 cases, using previously defined cutoffs for low-level (FISH ratio, 1.5-3) and high-level (FISH ratio, < 3) amplification for ovarian cancers (12). Interestingly, the mean FISH ratio of AURKA to centromere 20-specific (CEP20) signals were not correlated with AURKA mRNA or protein expression. In contrast, the mean AURKA- and CEP20-specific FISH signals were linked to AURKA protein expression. This may be explained by the fact that, first, both FISH and immunohistochemistry in situ analyses were done on serial sections of TMAs and QRT-PCR was done using extracts of microdissected tumor cells from entire sections of the tumor blocks used for TMA construction. Thus, cell populations analyzed by FISH and immunohistochemistry match directly and corresponded with the tumor areas investigated by QRT-PCR, but the number of investigated tumor cells were different for FISH, immunohistochemistry, and QRT-PCR. In addition, FISH, immunohistochemistry, and QRT-PCR values differ per se and correlations may become more precise on single cell correlation of continues values, e.g., using image software for FISH and immunohistochemistry analyses. Second, differences in DNA copy numbers and AURKA mRNA expression may reflect the fact that polyploidy or DNA amplification does not necessarily translate 1:1 into stable increased AURKA mRNA expression. It is therefore possible that mRNA expression (or protein phosphorylation and/or activity) are more fragile in tissue analysis than DNA copy numbers or protein expression. Interestingly, for Her2/neu testing in breast cancer, it was FISH and immunohistochemistry testing that is of diagnostic significance and not QRT-PCR analysis.

Irrespective of the detailed mechanisms of AURKA regulation, the data on AURKA DNA copy numbers and protein expression suggest that elevated AURKA protein expression may be closely linked to chromosome 20 aneusomy and polyploidization. In fact, we frequently observed the overexpression of AURKA in ovarian tumor cells with mitotic features indicating tetraploidy (data not shown). Although this may support the role of AURKA (over)expression and associated genomic instability in tumor development and progression, AURKA overexpression in late stage ovarian cancer may simply be a reflection of the progressed cancer, in which tumor cells display high genomic instability not necessarily linked to AURKA-specific gene amplification (22–24, present study). A precise analysis of AURKA expression and regulation in early ovarian cancer is needed to support the data from the 13 early stage cases presented here and to combine the findings with those of Gritsko et al. (22). Furthermore, the investigation of the effect of AURKA expression on centrosome changes and polyploidization in especially early ovarian cancer is necessary, in particular, as this may also affect the response to drugs which target the mitotic spindle, such as taxanes.

Recently, in vitro studies showed that small interfering RNA–mediated knock-down of AURKA expression increased the efficacy of taxane treatment in pancreatic cancer cell lines (27). Conversely, targeted AURKA overexpression in mouse embryonic fibroblasts was associated with resistance to taxol treatment (28). Similarly, ovarian cancer cell lines exhibiting high AURKA levels were resistant to apoptosis induced by common taxane (paclitaxel) and platinum (cisplatin) drugs (29).

In order to find out whether these correlations defined in vitro could also be translated into the clinical setting, we therefore examined the prognostic/predictive value of AURKA expression assessed in routinely processed ovarian cancer resection specimens derived from 115 patients. Specifically, we focused on those patients with adjuvant treatment with taxane- and non–taxane-based chemotherapy.

In contrast with the assumption that high AURKA expression is a prognostic factor for poor survival in other epithelial cancers (25, 26), our data for all stage III ovarian cancer patients (i.e., irrespective of debulking status and adjuvant chemotherapy) indicates that AURKA expression was not prognostic for disease-free or overall survival.

However, high AURKA expression is associated with improved overall survival in patients with stage III ovarian cancer with optimal debulking and receiving taxol/carboplatin therapy. This is particularly interesting as previous in vitro studies showed that high AURKA expression induced resistance to taxanes or platinum drugs (27–29). Moreover, the results of our exploratory approach comparing patients with stage III ovarian cancer treated by taxane- and non–taxane-based adjuvant therapy (irrespective of debulking status), showed that high AURKA protein expression was associated with a worse disease outcome only in patients receiving non–taxane-based adjuvant therapy. Thus, our data suggests that AURKA protein expression has a distinct role and predictive value for stage III ovarian cancer patients receiving taxane and non–taxane-based adjuvant therapy regimens. In addition, these results underline the necessity for investigating candidate predictive molecular markers in situ to complement functional in vitro testing with the unique clinicopathologic variables associated with individual cancer patients.

Irrespective of the differential effects of AURKA expression on the efficacy of different adjuvant chemotherapies, our data suggest that with its overexpression in 63.5% of ovarian cancers, AURKA represents a target for therapeutic intervention (30–38), either alone or in combination with other (non–taxane-based) chemotherapeutic regimens. However, it has to be emphasized that the current Aurora inhibitors are not specific for Aurora-A, so that future work with these inhibitors should be only done after further detailed investigation of Aurora-B and Aurora-C in ovarian cancer.

In summary, our data support the hypothesis that AURKA overexpression is a characteristic feature of epithelial ovarian cancer and is linked to increased AUKRA copy numbers and/or chromosome 20 aneusomy. AURKA expression represents a predictive marker for the clinical outcome of patients with late stage ovarian cancer, whereby AURKA overexpression may differentially affect taxane and non–taxane-based adjuvant therapy responses. The study sheds new light on AURKA expression and regulation in epithelial cancers in vivo and specifically shows its value as a clinically relevant marker and as a potential therapeutic target per se.


   Acknowledgments
 
We thank Evelyn Wätzig, Anja Schöpflin, and Jasmine Roth for expert technical assistance.


   Footnotes
 
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.

Received 11/27/06; revised 3/ 3/07; accepted 5/ 3/07.


   References
Top
 Abstract
 Materials and Methods
 Results
 Discussion
 References
 

  1. Runnebaum IB, Stickeler E. Epidemiological and molecular aspects of ovarian cancer risk. J Cancer Res Clin Oncol 2001;127:73–9.[CrossRef][Medline]
  2. Jemal A, Murray T, Ward E, et al. Cancer statistics, 2005. CA Cancer J Clin 2005;55:10–30.[Abstract/Free Full Text]
  3. Bhoola S, Hoskins WJ. Diagnosis and management of epithelial ovarian cancer. Obstet Gynecol 2006;107:1399–410.[Abstract/Free Full Text]
  4. Scott M, McCluggage WG. Current concepts in ovarian epithelial tumourigenesis: correlation between morphological and molecular data. Histol Histopathol 2006;21:81–92.[Medline]
  5. Cannistra SA. Cancer of the ovary. N Engl J Med 2004;351:2519–29.[Free Full Text]
  6. Tavassoli FA, Deville P, editors. WHO classification of tumours: tumours of the breast and female genital organs. Lyon: IARC Press; 2003.
  7. du Bois A, Quinn M, Thigpen T, et al. 2004 consensus statements on the management of ovarian cancer: final document of the 3rd International Gynecologic Cancer Intergroup Ovarian Cancer Consensus Conference (GCIG OCCC 2004). Ann Oncol 2005;16 Suppl 8:viii7–12.[Free Full Text]
  8. Pfisterer J, Harter P, Canzler U, et al. AGO Ovarian Cancer Study Group (AGO-OVAR). The role of surgery in recurrent ovarian cancer. Int J Gynecol Cancer 2005;15 Suppl 3:195–8.[CrossRef][Medline]
  9. Pignata S, Pisano C, Di Maio M, et al. Medical treatment of resistant or recurrent epithelial ovarian cancer. Ann Oncol 2006;17 Suppl 7:vii49–50.[Abstract/Free Full Text]
  10. Agarwal R, Kaye SB. Ovarian cancer: strategies for overcoming resistance to chemotherapy. Nat Rev Cancer 2003;3:502–16.[CrossRef][Medline]
  11. Friedlander ML. Prognostic factors in ovarian cancer. Semin Oncol 1998;25:305–14.[Medline]
  12. Tanner MM, Grenman S, Koul A, et al. Frequent amplification of chromosomal region 20q12-13 in ovarian cancer. Clin Cancer Res 2000;6:1833–9.[Abstract/Free Full Text]
  13. Zhou H, Kuang J, Zhong L, et al. Tumour amplified kinase STK15/BTAK induces centrosome amplification, aneuploidy and transformation. Nat Genet 1998;20:189–93.[CrossRef][Medline]
  14. Marumoto T, Zhang D, Saya H. Aurora-A—a guardian of poles. Nat Rev Cancer 2005;5:42–50.[CrossRef][Medline]
  15. Giet R, Petretti C, Prigent C. Aurora kinases, aneuploidy and cancer, a coincidence or a real link? Trends Cell Biol 2005;15:241–50.[CrossRef][Medline]
  16. Tong T, Zhong Y, Kong J, et al. Overexpression of Aurora-A contributes to malignant development of human esophageal squamous cell carcinoma. Clin Cancer Res 2004;10:7304–10. Erratum in: Clin Cancer Res 2005;11:4635.[Abstract/Free Full Text]
  17. Li D, Zhu J, Firozi PF, Abbruzzese JL, et al. Overexpression of oncogenic STK15/BTAK/Aurora A kinase in human pancreatic cancer. Clin Cancer Res 2003;9:991–7.[Abstract/Free Full Text]
  18. 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.[CrossRef][Medline]
  19. Bischoff JR, Anderson L, Zhu Y, et al. A homologue of Drosophila aurora kinase is oncogenic and amplified in human colorectal cancers. EMBO J 1998;17:3052–65.[CrossRef][Medline]
  20. Sen S, Zhou H, Zhang RD, et al. Amplification/overexpression of a mitotic kinase gene in human bladder cancer. J Natl Cancer Inst 2002;94:1320–9.[Abstract/Free Full Text]
  21. 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.[CrossRef][Medline]
  22. 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.[Abstract/Free Full Text]
  23. Chung CM, Man C, Jin Y, et al. Amplification and overexpression of aurora kinase A (AURKA) in immortalized human ovarian epithelial (HOSE) cells. Mol Carcinog 2005;43:165–74.[CrossRef][Medline]
  24. Hu W, Kavanagh JJ, Deaver M, et al. Frequent overexpression of STK15/Aurora-A/BTAK and chromosomal instability in tumourigenic cell cultures derived from human ovarian cancer. Oncol Res 2005;15:49–57.[Medline]
  25. Tanaka E, Hashimoto Y, Ito T, et al. The clinical significance of Aurora-A/STK15/BTAK expression in human esophageal squamous cell carcinoma. Clin Cancer Res 2005;11:1827–34.[Abstract/Free Full Text]
  26. Reiter R, Gais P, Jutting U, et al. Aurora kinase a messenger RNA overexpression is correlated with tumour progression and shortened survival in head and neck squamous cell carcinoma. Clin Cancer Res 2006;12:5136–41.[Abstract/Free Full Text]
  27. Hata T, Furukawa T, Sunamura M, et al. RNA interference targeting aurora kinase a suppresses tumour growth and enhances the taxane chemosensitivity in human pancreatic cancer cells. Cancer Res 2005;65:2899–905.[Abstract/Free Full Text]
  28. 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.[CrossRef][Medline]
  29. Yang H, He L, Kruk P, Nicosia SV, Cheng JQ. Aurora-A induces cell survival and chemoresistance by activation of Akt through a p53-dependent manner in ovarian cancer cells. Int J Cancer 2006;119:2304–12.[CrossRef][Medline]
  30. Andrews PD. Aurora kinases: shining lights on the therapeutic horizon? Oncogene 2005;24:5005–15.[CrossRef][Medline]
  31. Keen N, Taylor S. Aurora-kinase inhibitors as anticancer agents. Nat Rev Cancer 2004;4:927–36.[CrossRef][Medline]
  32. Warner SL, Bearss DJ, Han H, Von Hoff DD. Targeting Aurora-2 kinase in cancer. Mol Cancer Ther 2003;2:589–95.[Abstract/Free Full Text]
  33. Harrington EA, Bebbington D, Moore J, et al. VX-680, a potent and selective small-molecule inhibitor of the Aurora kinases, suppresses tumour growth in vivo. Nat Med 2004;10:262–7.[CrossRef][Medline]
  34. Lee EC, Frolov A, Li R, Ayala G, Greenberg NM. Targeting Aurora kinases for the treatment of prostate cancer. Cancer Res 2006;66:4996–5002.[Abstract/Free Full Text]
  35. Hauf S, Cole RW, LaTerra S, et al. The small molecule Hesperadin reveals a role for Aurora B in correcting kinetochore-microtubule attachment and in maintaining the spindle assembly checkpoint. J Cell Biol 2003;161:281–94.[Abstract/Free Full Text]
  36. Jung FH, Pasquet G, Lambert-van der Brempt C, et al. Discovery of novel and potent thiazoloquinazolines as selective Aurora A and B kinase inhibitors. J Med Chem 2006;49:955–70.[CrossRef][Medline]
  37. Carvajal RD, Tse A, Schwartz GK. Aurora kinases: new targets for cancer therapy. Clin Cancer Res 2006;12:6869–75.[Abstract/Free Full Text]
  38. Fu S, Hu W, Kavanagh JJ, Bast RC, Jr. Targeting Aurora kinases in ovarian cancer. Expert Opin Ther Targets 2006;10:77–85.[CrossRef][Medline]
  39. Sobin LH, Wittekind C, editors. TNM—classification of malignant tumours. UICC, 6th Edition, Wiley Liss, Inc., 2002.
  40. Gerlach U, Kayser G, Walch A, et al. Centrosome-, chromosomal-passenger- and cell-cycle-associated mRNAs are differentially regulated in the development of sporadic colorectal cancer. J Pathol 2006;208:462–72.[CrossRef][Medline]
  41. Specht K, Richter T, Muller U, Walch A, Werner M, Hofler H. Quantitative gene expression analysis in microdissected archival formalin-fixed and paraffin-embedded tumour tissue. Am J Pathol 2001;158:419–29.[Abstract/Free Full Text]
  42. de Kok JB, Roelofs RW, Giesendorf BA, et al. Normalization of gene expression measurements in tumour tissues: comparison of 13 endogenous control genes. Lab Invest 2005;85:154–9.[Medline]
  43. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-{Delta}{Delta}C(T)) method. Methods 2001;25:402–8.[CrossRef][Medline]
  44. Lassmann S, Weis S, Makowiec F, et al. Array CGH identifies distinct DNA copy number profiles of oncogenes and tumoursuppressorgenes in chromosomal- and microsatellite-unstable sporadic colorectal carcinomas. J Mol Med 2007;85:289–300.[Medline]
  45. Ewart-Toland A, Dai Q, Gao YT, et al. Aurora-A/STK15 T+91A is a general low penetrance cancer susceptibility gene: a meta-analysis of multiple cancer types. Carcinogenesis 2005;26:1368–73.[Abstract/Free Full Text]
  46. Li JJ, Weroha SJ, Lingle WL, Papa D, Salisbury JL, Li SA. Estrogen mediates Aurora-A overexpression, centrosome amplification, chromosomal instability, and breast cancer in female ACI rats. Proc Natl Acad Sci U S A 2004;101:18123–8.[Abstract/Free Full Text]



This article has been cited by other articles:


Home page
 Clin. Cancer Res.Home page
Y. Nadler, R. L. Camp, C. Schwartz, D. L. Rimm, H. M. Kluger, and Y. Kluger
Expression of Aurora A (but Not Aurora B) Is Predictive of Survival in Breast Cancer
Clin. Cancer Res., July 15, 2008; 14(14): 4455 - 4462.
[Abstract] [Full Text] [PDF]

Home page
 Clin. Cancer Res.Home page
O. Gautschi, J. Heighway, P. C. Mack, P. R. Purnell, P. N. Lara Jr., and D. R. Gandara
Aurora Kinases as Anticancer Drug Targets
Clin. Cancer Res., March 15, 2008; 14(6): 1639 - 1648.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Lassmann, S.
Right arrow Articles by Werner, M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Lassmann, S.
Right arrow Articles by Werner, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Cancer Research Clinical Cancer Research
Cancer Epidemiology Biomarkers & Prevention Molecular Cancer Therapeutics
Molecular Cancer Research Cancer Prevention Research
Cancer Prevention Journals Portal Cancer Reviews Online
Annual Meeting Education Book Meeting Abstracts Online