- Email: [email protected]
TPX2 as a prognostic indicator and potential therapeutic target in clear cell renal cell carcinoma
Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
Original article
TPX2 as a prognostic indicator and potential therapeutic target in clear cell renal cell carcinoma Zachary A. Glaser, B.A.a,*, Harold D. Love, Ph.D.a, Shunhua Guo, B.M.b, Lan Gellert, M.D., Ph.D.b, Sam S. Chang, M.D., M.B.A.a, Stanley Duke Herrell, M.D.a, Daniel A. Barocas, M.D.a, David F. Penson, M.D., M.P.H.a, Michael S. Cookson, M.D.a, Peter E. Clark, M.D.a a
Department of Urologic Surgery, Vanderbilt University Medical Center, Nashville, TN b Department of Pathology, Vanderbilt University Medical Center, Nashville, TN
Received 4 September 2016; received in revised form 5 December 2016; accepted 19 December 2016
Abstract Objectives: Our aims were to determine if targeting protein for Xklp2 (TPX2) is correlated with clear cell renal cell carcinoma (ccRCC) histology and oncologic outcomes using The Cancer Genome Atlas (TCGA) and an institutional tissue microarray (TMA). Methods: Clinicopathological data obtained from the TCGA consisted of 415 samples diagnosed with ccRCC. A TMA was constructed from tumors of 207 patients who underwent radical nephrectomy for ccRCC. TPX2 expression by immunohistochemistry on TMA was assessed by a genitourinary pathologist. Clinical data were extracted and linked to TMA cores. TPX2 and Aurora-A mRNA coexpression were evaluated in the TCGA cohort. Overall survival (OS), cancer-specific survival, and recurrence-free survival (RFS) were analyzed using the Kaplan-Meier method and log-rank statistics. Univariate and multivariate analyses were conducted using Cox proportional hazard models. Results: Median follow-up time for the TCGA cohort was 3.07 years. Aurora-A and TPX2 mRNA coexpression were significantly correlated (Pearson correlation ¼ 0.918). High TPX2 mRNA expression was associated with advanced stage, metastasis, poor OS, and RFS. Median follow-up time for the TMA cohort was 5.3 years. Elevated TPX2 protein expression, defined as greater than 75th percentile staining intensity, was identified in 47/207 (22.7%) patients. Increased TPX2 immunostaining was associated with poor OS (P ¼ 0.0327, 53% 5-year mortality), cancer-specific survival (P o 0.01, 47.8% 5-year cancer-specific mortality), RFS (P ¼ 0.0313, 73.6%, 5-year recurrence rate), grade, T stage, and metastasis. Multivariate analysis demonstrated elevated expression served as an independent predictor of RFS (hazard ratio ¼ 3.62 (1.13–11.55), P ¼ 0.029). Conclusions: We show TPX2, a regulator of Aurora-A, is associated with high grade and stage of ccRCC, and is an independent predictor of recurrence. Future studies are warranted testing its role in ccRCC biology, and its potential as a therapeutic target. r 2017 Elsevier Inc. All rights reserved. Keywords: TPX2; Clear cell renal cell carcinoma; The Cancer Genome Atlas
1. Introduction Renal cell carcinoma (RCC) is the most commonly found malignant renal tumor in adults, accounting for 62,700 newly diagnosed cases and roughly 14,000 deaths annually in the United States [1,2]. Clear cell RCC (ccRCC) comprises the most common RCC histological subtype [1]. Surgical excision is the standard of care for initial treatment. Unfortunately, up to one-third of patients have metastatic Corresponding author. Tel.: þ1-314-580-4008; fax: þ1-314-991-8610. E-mail address: [email protected] (Z.A. Glaser). *
http://dx.doi.org/10.1016/j.urolonc.2016.12.012 1078-1439/r 2017 Elsevier Inc. All rights reserved.
disease at initial presentation, and many with localized disease carry a considerable risk of recurrence [1]. Although considerable progress has been made with novel therapeutic approaches, response remains variable and few, if any, patients are cured [1–3]. Identification of novel biomarkers has important implications for the development targeted therapy to curtail patient risk and improve survival. This relies on investigation of the underlying biology driving ccRCC tumorigenesis [4]. Development of ccRCC is typically dependent on aberrant function of the von Hippel-Lindau gene [1,5]. Loss of von Hippel-Lindau gene expression results in
2
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
elevated hypoxia inducible factor alpha, which in turn leads to increased expression of multiple proteins including betacatenin [5]. Increased beta-catenin expression up-regulates the oncoprotein, Aurora-A, and a serine/threonine kinase involved in cell cycle progression whose overexpression is correlated with increased Fuhrman grade in ccRCC [3,6]. It has 2 functions namely (1) disassembly of primary cilium and (2) promotion and maintenance of mitotic spindle assembly. This latter function is dependent on its recruitment to the mitotic spindle and subsequent activation by the targeting protein for Xklp2 (TPX2) [7]. Targeted Aurora-A inhibitor therapy has recently been of clinical interest in various cancers including hepatocellular carcinoma, small cell lung cancer, breast cancer, non-Hodgkin's lymphoma (NHL), and RCC [3,8,9]. TPX2 is a microtubule-associated protein first found by Heidebrecht et al. in 1997 that is located in chromosome 20q 11.2. TPX2 modulates mitotic spindle assembly and plays an important role in regulating Aurora-A activity [10,11]. In its inactive state, TPX2 resides in the cytoplasm bound to importin alpha, which is in turn bound to importin beta. This inactive complex is shuttled into the nucleus during interphase, where the importin beta complex interacts with RanGTP. This interaction permits TPX2 dissociation from importin alpha, resulting in subsequent activation [12,13]. Within the nucleus, RanGTP is concentrated highest in the vicinity of chromosomes, providing optimal location for TPX2-mediated recruitment, and activation of Aurora-A [12]. In absence of this sequence of events, Aurora-A fails to localize to mitotic spindles [7,13]. Interestingly, degree of TPX2 expression may serve as a predictor of response to Aurora-A inhibition in patients with NHL [14]. The Cancer Genome Atlas (TCGA) is an ongoing collaboration initiated in 2006 between the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI). The program has cataloged comprehensive, multidimensional genomic maps of 33 major types, and subtypes of cancer including ccRCC [15–17]. In this article, we explore TPX2's clinical association with ccRCC histology and oncologic outcomes using the TCGA ccRCC cohort, and validate these findings in an institutional tissue microarray (TMA).
2. Materials and methods 2.1. TCGA patient cohort Clinicopathological data were obtained employing the cBIO cancer genome portal (Memorial Sloan-Kettering Cancer Center, New York, NY). The TCGA cohort consisted of 415 samples diagnosed with ccRCC. These biospecimens were obtained and processed by the Biospecimen Core Resource (Research Institute at Nationwide Children's Hospital, Columbus, OH). Data generation and
analyses were performed by the Genome Characterization Centers, Genome Sequencing Centers, and Genome Data Analysis Centers. Methylation profiling was performed by the TCGA using the Illumina HM450K array. The TCGA employed RNASeq V2 software to determine mRNA expression levels. Patient demographics, pathological, and clinical outcomes were linked to genomic data in a deidentified manner through the Data Coordinating Center (http://cancergenome.nih.gov/abouttcga/overview). 2.2. Tissue microarray A TMA was constructed from the tumors of 207 patients who underwent radical nephrectomy at Vanderbilt University Medical Center between 1994 and 2006 with confirmed histologic diagnosis of ccRCC. Two punches from areas of histologically viable tumor without evidence for necrosis or hemorrhage and 2 punches from histologically normal renal cortex from each patient were embedded in the TMA block (828 cores total). After TMA construction, a hematoxylin and eosin-stained slide of each TMA was reviewed by a genitourinary pathologist (L.G.) to confirm the appropriate tissue was represented. Patient demographic, pathologic, and clinical outcomes data were extracted from the electronic medical records and anonymously linked to the TMA cores under an Institutional Review Board approved protocol. All cases were assigned a Fuhrman nuclear grade and staged in accordance with the 2010 American Joint Committee on Cancer (AJCC) guidelines [18,19]. 2.3. Immunohistochemistry To detect the expression of TPX2, TMA slides were immunostained with polyclonal anti-TPX2 antibody (1:100 dilution; Sigma Aldrich, St. Louis, MO) using standardized technique [20,21]. Sections of TMAs were deparaffinized and rehydrated in a graded alcohol series. Slides were then incubated with primary TPX2 antibodies and exposed to a biotinylated secondary antibody, DAB. The immunostained sections were evaluated in a blinded fashion by a pathologist (S.G.) using an Olympus BX43 microscope (Olympus, Center Valley, PA). Sections were examined for immunoreactivity and graded semiquantitatively using the H scoring method, which reflects antigen concentration [22]. The H-score system was determined according to the formula: ([% weak staining] 1) þ ([% moderate staining] 2) þ ([% strong staining] 3), obtaining a value from 0 to 300. 2.4. Statistical analysis Statistical analyses were performed using StataSE software version 12.1 (College Station, TX). Group differences for qualitative variables were analyzed using chi-square, t-test, Mann-Whitney U, and Pearson correlation tests as appropriate. Overall survival (OS), cancer-specific survival
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
(CSS) and recurrence-free survival (RFS) rates were stratified using the Kaplan-Meier method and were analyzed using log-rank statistics. Univariate and multivariate analyses using Cox proportional hazard models were performed to measure correlations among clinicopathological data, OS, CSS, and RFS. We considered a P o 0.05 to be statistically significant.
3. Results 3.1. TCGA ccRCC cohort analysis To determine whether TPX2 expression was associated with pathological stage and clinical course, we accessed the TCGA portal to evaluate genomic alterations, DNA methylation profiles, and mRNA expression. Available demographic and clinical characteristics are summarized in Table 1. Mean follow-up time was 3.07 years for the 415 patients. Among these, 109/415 (26.3%) had disease recurrence. Tumor characteristics included 200/415 (48.2%) with T1 disease, and 67/415 (16.1%) with metastasis. A missense mutation (E280A) upstream of TPX2 was observed in our cohort and was not obviously contributory to disease histology or course. To evaluate the influence of DNA methylation on clinicopathological course, we stratified our cohort by median TPX2 methylation levels (β ¼ 0.0404), and demonstrated that TPX2 hypomethylation was associated with advanced T stage and metastasis (P ¼ 0.049, Table 2A). Furthermore, on univariate survival analysis, hypomethylation was associated with poor OS and RFS (P ¼ 0.0068 and P o 0.001, Fig. 1A and B, respectively). We next investigated the effect of mRNA expression. Consistent with our findings with DNA hypomethylation, we found elevated TPX2 mRNA expression was associated with advanced T stage and metastasis (both P o 0.001, Table 2B). To facilitate the evaluation of OS and RFS, TPX2 mRNA expression was divided into categorical variables defined as greater or less than 75th percentile. On univariate analysis, OS and RFS were significantly worse in patients in the highest quartile for TPX2 mRNA expression (both P o 0.001, Fig. 1C and D). Actuarial estimated overall survival at 5 years for the subset with the highest quartile of TPX2 expression was 36.8% (95% CI: 25.9%–47.6%) when compared with 67.1% (95% CI: 60.0%–73.3%) for those with lower expression. Prior studies revealed that TPX2 is necessary for recruitment and activation of oncoprotein Aurora-A, whose upregulation has been associated with aggressive ccRCC [3,6,10–13]. We, therefore, compared mRNA expression of TPX2 with Aurora-A. A strongly positive correlation for expression of TPX2 and Aurora-A mRNA was observed with a Pearson correlation of 0.918 and Spearman correlation of 0.812 (both P o 0.001, Fig. 2).
3
Table 1 Baseline characteristics of TCGA (n ¼ 415) and TMA (n ¼ 207) patient cohorts Clinical information Number of patients (%) TCGA cohort (n ¼ 415) TMA cohort (n ¼ 207) Age, yr Mean Range Sex Male Female Race White Black Hispanic Asian Unknown BMI Mean Margins Positive Negative Sarcomatoid features Yes No Fuhrman grade 1 2 3 4 T stage T1 T2 T3 T4 Lymph nodes N1 N0/Nx Metastasis Yes No Disease recurrence Yes No Disease course Alive Dead Died of disease
60.7 26–90
61 34–87
270 (65.1) 145 (34.9)
151 (72.9) 56 (27.1)
373 14 21 7 0
187 17 2 0 1
(89.8) (3.37) (5.10) (1.69)
–
(90.3) (8.21) (0.97) (0.01)
31
– –
14 (6.76) 193 (93.2)
– –
16 (7.73) 191 (92.3)
– – – – 200 49 157 6
(48.2) (11.8) (37.8) (1.45)
19 87 71 30
(9.18) (42.0) (34.3) (14.5)
73 32 90 12
(35.3) (15.5) (43.5) (5.80)
11 (2.65) 404 (97.3)
13 (6.28) 194 (93.7)
67 (16.1) 346 (83.3)
44 (21.3) 163 (78.7)
109 (26.2) 227 (54.7)
42 (20.3) 165 (79.7)
268 (64.6) 147 (35.4) –
103 (49.7) 104 (50.3) 67 (32.4)
BMI ¼ body mass index.
3.2. Clinical course and TPX2 expression in ccRCC TMA cohort To validate our findings from the TCGA cohort, we constructed a TMA of 207 patients diagnosed with ccRCC. The demographics and clinical characteristics of the TMA cohort are summarized in Table 1. The median follow-up time was 5.3 (0–20.8) years, and median times to recurrence and death were 12.8 (1–199) and 24.5 (1–146) months, respectively. Our cohort included 19/207 (9.18%) patients with Fuhrman grade 1 disease, 73 (35.3%) with T1 tumors, and 44 (21.3%) with metastasis. Recurrence occurred in 42
4
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
Table 2 Association of (A) TPX2 methylation with T stage (β ¼ 0.0403765, n ¼ 240) and (B) TPX2 mRNA expression with T stage (n ¼ 412) and metastasis (n ¼ 414) T stage
Low methylation (%)
High methylation (%)
P
(A) Association of TPX2 methylation with T stage (β ¼ 0.0403765, n ¼ 240) Patients 118 (49.1) 122 (50.9) 0.049 T1 44 (37.2) 67 (54.9) T2 18 (15.2) 11 (9.02) T3 53 (44.9) 42 (34.4) T4 3 (2.54) 2 (1.63) T and M stage Low mRNA High mRNA P expression (%) expression (%) (B) Association of TPX2 mRNA expression with T stage (n ¼ 412) and metastasis (n ¼ 414) Patients 308 (74.8) 104 (25.2) o0.001 T1 170 (55.2) 30 (28.8) T2 34 (11.0) 15 (14.4) T3 103 (33.4) 54 (51.9) T4 1 (0.32) 5 (4.81) Patients 309 (74.6) 105 (25.4) o0.001 M0 275 (80.9) 72 (68.6) M1 34 (11.0) 33 (31.4)
(20.3%) patients. At the time of review, 65/207 (31.4%) patients were alive with no evidence of disease, 15 (7.25%) were alive with disease, and 23 (11.1%) were alive with unknown status. A total of 70 (33.9%) patients died of disease, 16 (7.73%) died with no evidence of disease, 17 (8.2%) died with unknown disease status, and 1 patient died of other causes. Immunohistochemical analysis demonstrated TPX2 protein is expressed in the cytoplasm, consistent with The Human Protein Atlas tissue repository (Fig. 3) [13,23]. To stratify patients by TPX2 protein expression, we defined high expression as greater than the 75th percentile staining intensity according to mean H-score of the paired tissue cores. High expression of TPX2 was identified in 47 (22.7%) patients (Table 3). We found a significant correlation of multiple adverse clinical and pathological features with high expression of TPX2 (Table 3). Elevated TPX2 expression was associated with advanced Fuhrman grade and AJCC stage (P ¼ 0.001 and P ¼ 0.005, respectively; Table 3). Additionally, a higher incidence of metastatic disease was observed in the elevated TPX2 cohort (P ¼ 0.016, Table 3).
Fig. 1. Kaplan-Meier plot of overall and recurrence-free survival according to high (red) vs. low (blue) TPX2 methylation and mRNA expression in the TCGA cohort. (A) OS, TPX2 methylation (n ¼ 240, P ¼ 0.0223). (B) RFS, TPX2 methylation (n ¼ 195, P ¼ 0.0001). (C) OS, TPX2 mRNA expression (n ¼ 415, P o 0.001). (D) RFS, TPX2 mRNA expression (n ¼ 344, P o 0.001).
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
Fig. 2. Comparison of TPX2 and Aurora-A mRNA expression in TCGA cohort (n ¼ 415, Pearson ¼ 0.918, Spearman ¼ 0.812, P o 0.001).
To determine whether TPX2 protein was elevated in the histologically normal kidney in patients with high TPX2 expressing tumors, we evaluated their corresponding benign tissue cores. A Student's t-test demonstrated mean TPX2 protein expression in those with TPX2-rich tumors was
5
significantly increased in the corresponding benign tissue compared to the rest of the cohort (173.89 vs. 142.18, P ¼ 0.0028, Table 3). Thus, those patients with the highest expression of TPX2 in their tumor also had higher expression levels in their histologically benign renal cortex from the same kidney. Univariate analysis demonstrated high TPX2 protein expression in the tumor was associated with poor OS, CSS, and RFS (P ¼ 0.0327, P o 0.01, and P ¼ 0.0313, respectively; Fig. 4, Supplementary Fig. 1, and Supplementary Table 1). Actuarial estimated OS at 5 years for the highly expressed TPX2 subset was 47% (95% CI: 31.9%–60.6%) compared with 62.3% (95% CI: 53.4%– 69.6%) in the remaining group. The 5-year CSS rate for this same subset was 52% (95% CI: 36.5%–65.8%). The 5-year recurrence-free rate for the highly expressed TPX2 cohort was 73.6% (95% CI: 55.3%–85.3%) compared with 86.9% (95% CI: 78.4%–92.3%) in the lower expressing cohort. We constructed a multivariable model to determine whether TPX2 protein expression served as an independent predictor for overall and RFS. After correcting for age, Fuhrman grade, stage, lymph node status, and metastasis (the latter only for OS
Fig. 3. Weak (A, H-score ¼ 0), intermediate (B, H-score ¼ 40), and strong (C, H-score ¼ 100, 475th percentile) TPX2 immunostaining in ccRCC specimen (20). H-score, determined by DAB staining, reflects antibody binding and intensity of staining.
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
6
Table 3 Comparison of normal vs. high TPX2 immunostaining. High TPX2 staining intensity was defined as greater than the 75th percentile mean H-score Clinical information
Normal TPX2 protein expression (%)
Elevated TPX2 protein expression (%)
Patients Age (mean, yr) Sex Male Female Race White Black Hispanic Unknown BMI (mean) Margins Positive Negative Sarcomatoid Lymph node status Fuhrman grade 3 or 4 AJCC stage III or greater Metastasis Recurrence Corresponding benign tissue TPX2 expression
159 (76.8) 61.72
47 (22.7) 59.225
122 (76.7) 37 (33.3)
29 (61.7) 18 (38.3)
146 (91.8) 11 (18.6) 2 (3.3) 0 31.5
40 (85.1) 6 (12.7) 0 1 (2.1) 28.5
8 151 11 7 68 75
6 41 5 6 33 33
P
0.201 0.041
0.129
(5.0) (95.0) (6.9) (4.4) (42.7) (47.2)
28 (17.6) 26 (16.4) 142.18
(12.7) (87.3) (10.6) (2.1) (70.2) (70.2)
16 (34.0) 16 (34.0) 173.89
0.472 0.064
0.403 0.165 0.001 0.005 0.016 0.002 0.0028
evaluation), our model demonstrated high TPX2 expression is an independent predictor of RFS (hazard ratio ¼ 3.62, P ¼ 0.029, 95% CI: 1.13–11.55; Table 4A), but not for OS or CSS (P 4 0.05, Table 4B and C). 4. Discussion Before our study, TPX2 had not been investigated in the context of ccRCC. Here, we show in 2 independent cohorts
that elevated mRNA expression and TPX2 protein levels are associated with more aggressive ccRCC and worse outcomes. TPX2 mRNA expression is also positively correlated in the TCGA with Aurora-A levels, which suggests this could be one such mechanism by which TPX2 could influence oncologic outcomes. Immunohistochemical analysis of our TMA cohort validated our TCGA findings demonstrating increased protein expression of TPX2 is associated with advanced Fuhrman grade, AJCC T stage, metastatic potential, worse overall survival, CSS, and is a significant predictor of recurrence. TPX2 has gained attention as a possible prognostic indicator and therapeutic target in human malignancies such as pancreatic adenocarcinoma, hepatocellular carcinoma, and NHL [11,14,24]. Although our study demonstrates an association of TPX2 with ccRCC disease severity, this does not imply a direct causal relationship exists. Previous studies investigating TPX2 in malignant processes demonstrated causal tumorigenic relationship via knockdown with specific TPX2-siRNA in mice [24–26]. It was demonstrated that TPX2 activates Aurora-A by binding an allosteric site, thereby increasing Aurora-A's binding affinity to ATP and substrate [27]. Furthermore, the binding of TPX2 to Aurora-A reduces the affinity of VX-680/MK0457, an Aurora-A inhibitor [25,27,28]. Alternatively, it was shown that TPX2-siRNA transfected colorectal carcinoma cells up-regulated AKT expression, suggesting TPX2 may induce tumorigenesis via the PI3K/Akt/mTOR pathway [21]. Similar studies with ccRCC would be useful for creating next steps to determine if this disease exhibits similar tumorigenic routes, and also if targeting TPX2 may be a superior approach to Aurora-A targeted therapy. If TPX2 does directly propagate ccRCC tumorigenesis, this may open the door to multiple novel therapeutic approaches currently under investigation. In pancreatic adenocarcinoma, TPX2-siRNA transfected cells demonstrated unusually high sensitivity to paclitaxel therapy [24]. Targeted TPX2 inhibition in ccRCC cells may provide
Fig. 4. Kaplan-Meier plot showing overall and recurrence-free survival according to high (red) vs. low (blue) TPX2 protein expression in the TMA cohort. (A) OS, TPX2 protein expression (n ¼ 203, P ¼ 0.0327). (B) RFS, TPX2 protein expression (n ¼ 163, P ¼ 0.0313).
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎ Table 4 Multivariate analysis using Cox proportional hazards evaluating TPX2 protein expression's potential role as a biomarker in RFS (A, n ¼ 163), OS (B, n ¼ 207), and CSS (C, n ¼ 207) Variable
HR
95% CI
P 4 |z|
(A) RFS (n ¼ 163) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status
3.63 2.12 3.41 1.00 0.73
1.14–11.55 0.66–6.78 0.87–13.39 0.95–1.06 0.14–3.93
0.029 0.205 0.079 0.889 0.714
(B) OS (n ¼ 207) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status Metastasis
1.29 2.09 3.32 1.00 0.76 2.30
0.58–2.90 0.84–5.19 0.92–11.93 0.97–1.04 0.28–2.12 0.86–6.15
0.534 0.114 0.066 0.98 0.605 0.099
(C) CSS (n ¼ 207) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status Metastasis
1.58 2.60 4.21 1.00 0.80 2.25
0.63–3.96 0.91–7.46 0.90–19.73 0.96–1.04 0.27–2.30 0.82–6.17
0.327 0.075 0.068 0.845 0.674 0.117
7
potential to have unmeasured confounders bias our results. Although both cohorts are relatively large by the standards of biomarker studies, nevertheless a larger cohort may demonstrate additional or alternate associations. Furthermore, the number of lymph nodes removed and the locations of metastatic disease were not recorded in both cohorts, which may have provided greater insight to the association of TPX2 with advanced disease. Additionally, this study demonstrates an association between TPX2 and Aurora-A and outcomes in ccRCC, but we do not have specific demonstration of a mechanistic link between TPX2 and ccRCC. 5. Conclusion In conclusion, we show in 2 disparate cohorts that TPX2, a regulator of Aurora-A, is associated with high grade and stage of ccRCC, and is an independent predictor of disease recurrence. Further studies are warranted to elucidate its role in ccRCC tumor biology and potential as a therapeutic target. Acknowledgments
HR ¼ hazard ratio.
a similar optimized therapeutic environment, though admittedly cytotoxic chemotherapy is not standard of care in ccRCC. In hepatocellular carcinoma, T-cell mediated immunotherapy using cytotolytic T lymphocytes (CTLs) targeting TPX2-specific peptides has demonstrated early promise [11]. In addition, the TPX2-Aurora-A complex was identified as a likely cellular target of withanone, and may be of therapeutic benefit [29]. These studies highlight the potential of TPX2-targeting therapies and demonstrate the potential application to ccRCC management. Moreover, staining histologically normal tissue for TPX2 may serve as a companion diagnostic test if targeted therapy for this pathway is considered for a patient. Interestingly, evaluation of the corresponding histologically benign tissue in patients with ccRCC showed that the cohort with elevated TPX2 in their tumor also demonstrated significantly higher protein expression in the histologically normal renal parenchyma. As these individuals had a greater propensity to develop high-grade tumors and recurrent disease, this observation raised the intriguing hypothesis that TPX2 expression may help drive the development of aggressive ccRCC. As TPX2 was not a significant predictor of survival in these cohorts, this may imply that although factors such as age, Fuhrman grade, and AJCC stage better predict overall and cancer-specific mortality, TPX2 may be a primary driver of recurrence. In this scenario, staining tumor specimens for TPX2 may inform providers if closer follow-up is warranted. Our study does have several limitations. Both the TCGA and our own cohort are retrospective studies, so there is the
The results published here are in whole or part based on data generated by the TCGA Research Network in adherence to the TCGA publication guidelines: http://cancer genome.nih.gov/. Whole slide imaging was performed in the Digital Histology Shared Resource at Vanderbilt University Medical Center (www.mc.vanderbilt.edu/dhsr). This project was supported by the Inquiry Program of the Office of Medical Student Research at Vanderbilt University School of Medicine and the Vanderbilt University Medical Center. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2016.12.012.
References [1] Clark PE. The role of VHL in clear-cell renal cell carcinoma and its relation to targeted therapy. Kidney Int 2009;76:939–45. [2] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016;66:7–30. [3] Ferchichi I, Kourda N, Sassi S, Romdhane KB, Balatgi S, Cremet JY, et al. Aurora A overexpression and pVHL reduced expression are correlated with a bad kidney cancer prognosis. Dis Markers 2012;33:333–40. [4] Calvo E, Schmidinger M, Heng DY, Grunwald V, Escudier B. Improvement in survival end points of patients with metastatic renal
8
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎ cell carcinoma through sequential targeted therapy. Cancer Treat Rev 2016;50:109–17. Schonenberger D, Harlander S, Rajski M, Jacobs RA, Lundby AK, Adlesic M, et al. Formation of renal cysts and tumors in Vhl/Trp53deficient mice requires HIF-1alpha and HIF-2alpha. Cancer Res 2016;76:2025–36. Dere R, Perkins AL, Bawa-Khalfe T, Jonasch D, Walker CL. betacatenin links von Hippel-Lindau to aurora kinase A and loss of primary cilia in renal cell carcinoma. J Am Soc Nephrol 2015;26:553–64. Kufer TA, Sillje HH, Korner R, Gruss OJ, Meraldi P, Nigg EA. Human TPX2 is required for targeting Aurora-A kinase to the spindle. J Cell Biol 2002;158:617–23. Jin S, Wang X, Tong T, Zhang D, Shi J, Chen J, et al. Aurora-A enhances malignant development of esophageal squamous cell carcinoma (ESCC) by phosphorylating beta-catenin. Mol Oncol 2015;9:249–59. Li Y, Zhou W, Wei L, Jin J, Tang K, Li C, et al. The effect of Aurora kinases on cell proliferation, cell cycle regulation and metastasis in renal cell carcinoma. Int J Oncol 2012;41:2139–49. Gruss OJ, Wittmann M, Yokoyama H, Pepperkok R, Kufer T, Sillje H, et al. Chromosome-induced microtubule assembly mediated by TPX2 is required for spindle formation in HeLa cells. Nat Cell Biol 2002;4:871–9. Aref AM, Hoa NT, Ge L, Agrawal A, Dacosta-Iyer M, Lambrecht N, et al. HCA519/TPX2: a potential T-cell tumor-associated antigen for human hepatocellular carcinoma. Onco Targets Ther 2014;7:1061–70. Neumayer G, Belzil C, Gruss OJ, Nguyen MD. TPX2: of spindle assembly, DNA damage response, and cancer. Cell Mol Life Sci 2014;71:3027–47. Schatz CA, Santarella R, Hoenger A, Karsenti E, Mattaj IW, Gruss OJ, et al. Importin—alpha-regulated nucleation of microtubules by TPX2. EMBO J 2003;22:2060–70. Chowdhury A, Chowdhury S, Tsai MY. A novel Aurora kinase A inhibitor MK-8745 predicts TPX2 as a therapeutic biomarker in nonHodgkin lymphoma cell lines. Leuk Lymphoma 2012;53:462–71. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov 2012;2:401–4. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 2013:6:pl1.
[17] Cancer Genome Atlas Research Network: Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 2013;499:43–9. [18] Fuhrman SA, Lasky LC, Limas C. Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol 1982;6:655–63. [19] Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. [20] Yi Y, Polosukhina D, Love HD, Hembd A, Pickup M, Moses HL, et al. A Murine model of K-RAS and beta-catenin induced renal tumors expresses high levels of E2F1 and resembles human Wilms tumor. J Urol 2015;194:1762–70. [21] Wei P, Zhang N, Xu Y, Li X, Shi D, Wang Y, et al. TPX2 is a novel prognostic marker for the growth and metastasis of colon cancer. J Transl Med 2013;11:313. [22] Detre S, Jotti GS, Dowsett M. A “quick score” method for immunohistochemical semiquantitation: validation for oestrogen receptor in breast carcinomas. J Clin Pathol 1995;48:876–8. [23] Uhlen M., et al. Proteomics. Tissue-based map of the human proteome. Science 2015;347:1260419. [24] Warner SL, Stephens BJ, Nwokenkwo S, Hostetter G, Sugeng A, Hidalgo M, et al. Validation of TPX2 as a potential therapeutic target in pancreatic cancer cells. Clin Cancer Res. 2009;15:6519–28. [25] Satow R, Shitashige M, Kanai Y, Takeshita F, Ojima H, Jigami T, et al. Combined functional genome survey of therapeutic targets for hepatocellular carcinoma. Clin Cancer Res 2010;16:2518–28. [26] Perez de Castro I, Malumbres M. Mitotic stress and chromosomal instability in cancer: the case for TPX2. Genes Cancer 2012;3: 721–30. [27] Eyers PA, Erikson E, Chen LG, Maller JL. A novel mechanism for activation of the protein kinase Aurora A. Curr Biol 2003;13:691–7. [28] Anderson K, Yang J, Koretke K, Nurse K, Calamari A, Kirkpatrick RB, et al. Binding of TPX2 to Aurora A alters substrate and inhibitor interactions. Biochemistry 2007;46:10287–95. [29] Grover A, Singh R, Shandilya A, Priyandoko D, Agrawal V, Bisaria VS, et al. Ashwagandha derived withanone targets TPX2-Aurora A complex: computational and experimental evidence to its anticancer activity. PLoS One 2012;7:e30890.
Original article
TPX2 as a prognostic indicator and potential therapeutic target in clear cell renal cell carcinoma Zachary A. Glaser, B.A.a,*, Harold D. Love, Ph.D.a, Shunhua Guo, B.M.b, Lan Gellert, M.D., Ph.D.b, Sam S. Chang, M.D., M.B.A.a, Stanley Duke Herrell, M.D.a, Daniel A. Barocas, M.D.a, David F. Penson, M.D., M.P.H.a, Michael S. Cookson, M.D.a, Peter E. Clark, M.D.a a
Department of Urologic Surgery, Vanderbilt University Medical Center, Nashville, TN b Department of Pathology, Vanderbilt University Medical Center, Nashville, TN
Received 4 September 2016; received in revised form 5 December 2016; accepted 19 December 2016
Abstract Objectives: Our aims were to determine if targeting protein for Xklp2 (TPX2) is correlated with clear cell renal cell carcinoma (ccRCC) histology and oncologic outcomes using The Cancer Genome Atlas (TCGA) and an institutional tissue microarray (TMA). Methods: Clinicopathological data obtained from the TCGA consisted of 415 samples diagnosed with ccRCC. A TMA was constructed from tumors of 207 patients who underwent radical nephrectomy for ccRCC. TPX2 expression by immunohistochemistry on TMA was assessed by a genitourinary pathologist. Clinical data were extracted and linked to TMA cores. TPX2 and Aurora-A mRNA coexpression were evaluated in the TCGA cohort. Overall survival (OS), cancer-specific survival, and recurrence-free survival (RFS) were analyzed using the Kaplan-Meier method and log-rank statistics. Univariate and multivariate analyses were conducted using Cox proportional hazard models. Results: Median follow-up time for the TCGA cohort was 3.07 years. Aurora-A and TPX2 mRNA coexpression were significantly correlated (Pearson correlation ¼ 0.918). High TPX2 mRNA expression was associated with advanced stage, metastasis, poor OS, and RFS. Median follow-up time for the TMA cohort was 5.3 years. Elevated TPX2 protein expression, defined as greater than 75th percentile staining intensity, was identified in 47/207 (22.7%) patients. Increased TPX2 immunostaining was associated with poor OS (P ¼ 0.0327, 53% 5-year mortality), cancer-specific survival (P o 0.01, 47.8% 5-year cancer-specific mortality), RFS (P ¼ 0.0313, 73.6%, 5-year recurrence rate), grade, T stage, and metastasis. Multivariate analysis demonstrated elevated expression served as an independent predictor of RFS (hazard ratio ¼ 3.62 (1.13–11.55), P ¼ 0.029). Conclusions: We show TPX2, a regulator of Aurora-A, is associated with high grade and stage of ccRCC, and is an independent predictor of recurrence. Future studies are warranted testing its role in ccRCC biology, and its potential as a therapeutic target. r 2017 Elsevier Inc. All rights reserved. Keywords: TPX2; Clear cell renal cell carcinoma; The Cancer Genome Atlas
1. Introduction Renal cell carcinoma (RCC) is the most commonly found malignant renal tumor in adults, accounting for 62,700 newly diagnosed cases and roughly 14,000 deaths annually in the United States [1,2]. Clear cell RCC (ccRCC) comprises the most common RCC histological subtype [1]. Surgical excision is the standard of care for initial treatment. Unfortunately, up to one-third of patients have metastatic Corresponding author. Tel.: þ1-314-580-4008; fax: þ1-314-991-8610. E-mail address: [email protected] (Z.A. Glaser). *
http://dx.doi.org/10.1016/j.urolonc.2016.12.012 1078-1439/r 2017 Elsevier Inc. All rights reserved.
disease at initial presentation, and many with localized disease carry a considerable risk of recurrence [1]. Although considerable progress has been made with novel therapeutic approaches, response remains variable and few, if any, patients are cured [1–3]. Identification of novel biomarkers has important implications for the development targeted therapy to curtail patient risk and improve survival. This relies on investigation of the underlying biology driving ccRCC tumorigenesis [4]. Development of ccRCC is typically dependent on aberrant function of the von Hippel-Lindau gene [1,5]. Loss of von Hippel-Lindau gene expression results in
2
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
elevated hypoxia inducible factor alpha, which in turn leads to increased expression of multiple proteins including betacatenin [5]. Increased beta-catenin expression up-regulates the oncoprotein, Aurora-A, and a serine/threonine kinase involved in cell cycle progression whose overexpression is correlated with increased Fuhrman grade in ccRCC [3,6]. It has 2 functions namely (1) disassembly of primary cilium and (2) promotion and maintenance of mitotic spindle assembly. This latter function is dependent on its recruitment to the mitotic spindle and subsequent activation by the targeting protein for Xklp2 (TPX2) [7]. Targeted Aurora-A inhibitor therapy has recently been of clinical interest in various cancers including hepatocellular carcinoma, small cell lung cancer, breast cancer, non-Hodgkin's lymphoma (NHL), and RCC [3,8,9]. TPX2 is a microtubule-associated protein first found by Heidebrecht et al. in 1997 that is located in chromosome 20q 11.2. TPX2 modulates mitotic spindle assembly and plays an important role in regulating Aurora-A activity [10,11]. In its inactive state, TPX2 resides in the cytoplasm bound to importin alpha, which is in turn bound to importin beta. This inactive complex is shuttled into the nucleus during interphase, where the importin beta complex interacts with RanGTP. This interaction permits TPX2 dissociation from importin alpha, resulting in subsequent activation [12,13]. Within the nucleus, RanGTP is concentrated highest in the vicinity of chromosomes, providing optimal location for TPX2-mediated recruitment, and activation of Aurora-A [12]. In absence of this sequence of events, Aurora-A fails to localize to mitotic spindles [7,13]. Interestingly, degree of TPX2 expression may serve as a predictor of response to Aurora-A inhibition in patients with NHL [14]. The Cancer Genome Atlas (TCGA) is an ongoing collaboration initiated in 2006 between the National Cancer Institute (NCI) and the National Human Genome Research Institute (NHGRI). The program has cataloged comprehensive, multidimensional genomic maps of 33 major types, and subtypes of cancer including ccRCC [15–17]. In this article, we explore TPX2's clinical association with ccRCC histology and oncologic outcomes using the TCGA ccRCC cohort, and validate these findings in an institutional tissue microarray (TMA).
2. Materials and methods 2.1. TCGA patient cohort Clinicopathological data were obtained employing the cBIO cancer genome portal (Memorial Sloan-Kettering Cancer Center, New York, NY). The TCGA cohort consisted of 415 samples diagnosed with ccRCC. These biospecimens were obtained and processed by the Biospecimen Core Resource (Research Institute at Nationwide Children's Hospital, Columbus, OH). Data generation and
analyses were performed by the Genome Characterization Centers, Genome Sequencing Centers, and Genome Data Analysis Centers. Methylation profiling was performed by the TCGA using the Illumina HM450K array. The TCGA employed RNASeq V2 software to determine mRNA expression levels. Patient demographics, pathological, and clinical outcomes were linked to genomic data in a deidentified manner through the Data Coordinating Center (http://cancergenome.nih.gov/abouttcga/overview). 2.2. Tissue microarray A TMA was constructed from the tumors of 207 patients who underwent radical nephrectomy at Vanderbilt University Medical Center between 1994 and 2006 with confirmed histologic diagnosis of ccRCC. Two punches from areas of histologically viable tumor without evidence for necrosis or hemorrhage and 2 punches from histologically normal renal cortex from each patient were embedded in the TMA block (828 cores total). After TMA construction, a hematoxylin and eosin-stained slide of each TMA was reviewed by a genitourinary pathologist (L.G.) to confirm the appropriate tissue was represented. Patient demographic, pathologic, and clinical outcomes data were extracted from the electronic medical records and anonymously linked to the TMA cores under an Institutional Review Board approved protocol. All cases were assigned a Fuhrman nuclear grade and staged in accordance with the 2010 American Joint Committee on Cancer (AJCC) guidelines [18,19]. 2.3. Immunohistochemistry To detect the expression of TPX2, TMA slides were immunostained with polyclonal anti-TPX2 antibody (1:100 dilution; Sigma Aldrich, St. Louis, MO) using standardized technique [20,21]. Sections of TMAs were deparaffinized and rehydrated in a graded alcohol series. Slides were then incubated with primary TPX2 antibodies and exposed to a biotinylated secondary antibody, DAB. The immunostained sections were evaluated in a blinded fashion by a pathologist (S.G.) using an Olympus BX43 microscope (Olympus, Center Valley, PA). Sections were examined for immunoreactivity and graded semiquantitatively using the H scoring method, which reflects antigen concentration [22]. The H-score system was determined according to the formula: ([% weak staining] 1) þ ([% moderate staining] 2) þ ([% strong staining] 3), obtaining a value from 0 to 300. 2.4. Statistical analysis Statistical analyses were performed using StataSE software version 12.1 (College Station, TX). Group differences for qualitative variables were analyzed using chi-square, t-test, Mann-Whitney U, and Pearson correlation tests as appropriate. Overall survival (OS), cancer-specific survival
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
(CSS) and recurrence-free survival (RFS) rates were stratified using the Kaplan-Meier method and were analyzed using log-rank statistics. Univariate and multivariate analyses using Cox proportional hazard models were performed to measure correlations among clinicopathological data, OS, CSS, and RFS. We considered a P o 0.05 to be statistically significant.
3. Results 3.1. TCGA ccRCC cohort analysis To determine whether TPX2 expression was associated with pathological stage and clinical course, we accessed the TCGA portal to evaluate genomic alterations, DNA methylation profiles, and mRNA expression. Available demographic and clinical characteristics are summarized in Table 1. Mean follow-up time was 3.07 years for the 415 patients. Among these, 109/415 (26.3%) had disease recurrence. Tumor characteristics included 200/415 (48.2%) with T1 disease, and 67/415 (16.1%) with metastasis. A missense mutation (E280A) upstream of TPX2 was observed in our cohort and was not obviously contributory to disease histology or course. To evaluate the influence of DNA methylation on clinicopathological course, we stratified our cohort by median TPX2 methylation levels (β ¼ 0.0404), and demonstrated that TPX2 hypomethylation was associated with advanced T stage and metastasis (P ¼ 0.049, Table 2A). Furthermore, on univariate survival analysis, hypomethylation was associated with poor OS and RFS (P ¼ 0.0068 and P o 0.001, Fig. 1A and B, respectively). We next investigated the effect of mRNA expression. Consistent with our findings with DNA hypomethylation, we found elevated TPX2 mRNA expression was associated with advanced T stage and metastasis (both P o 0.001, Table 2B). To facilitate the evaluation of OS and RFS, TPX2 mRNA expression was divided into categorical variables defined as greater or less than 75th percentile. On univariate analysis, OS and RFS were significantly worse in patients in the highest quartile for TPX2 mRNA expression (both P o 0.001, Fig. 1C and D). Actuarial estimated overall survival at 5 years for the subset with the highest quartile of TPX2 expression was 36.8% (95% CI: 25.9%–47.6%) when compared with 67.1% (95% CI: 60.0%–73.3%) for those with lower expression. Prior studies revealed that TPX2 is necessary for recruitment and activation of oncoprotein Aurora-A, whose upregulation has been associated with aggressive ccRCC [3,6,10–13]. We, therefore, compared mRNA expression of TPX2 with Aurora-A. A strongly positive correlation for expression of TPX2 and Aurora-A mRNA was observed with a Pearson correlation of 0.918 and Spearman correlation of 0.812 (both P o 0.001, Fig. 2).
3
Table 1 Baseline characteristics of TCGA (n ¼ 415) and TMA (n ¼ 207) patient cohorts Clinical information Number of patients (%) TCGA cohort (n ¼ 415) TMA cohort (n ¼ 207) Age, yr Mean Range Sex Male Female Race White Black Hispanic Asian Unknown BMI Mean Margins Positive Negative Sarcomatoid features Yes No Fuhrman grade 1 2 3 4 T stage T1 T2 T3 T4 Lymph nodes N1 N0/Nx Metastasis Yes No Disease recurrence Yes No Disease course Alive Dead Died of disease
60.7 26–90
61 34–87
270 (65.1) 145 (34.9)
151 (72.9) 56 (27.1)
373 14 21 7 0
187 17 2 0 1
(89.8) (3.37) (5.10) (1.69)
–
(90.3) (8.21) (0.97) (0.01)
31
– –
14 (6.76) 193 (93.2)
– –
16 (7.73) 191 (92.3)
– – – – 200 49 157 6
(48.2) (11.8) (37.8) (1.45)
19 87 71 30
(9.18) (42.0) (34.3) (14.5)
73 32 90 12
(35.3) (15.5) (43.5) (5.80)
11 (2.65) 404 (97.3)
13 (6.28) 194 (93.7)
67 (16.1) 346 (83.3)
44 (21.3) 163 (78.7)
109 (26.2) 227 (54.7)
42 (20.3) 165 (79.7)
268 (64.6) 147 (35.4) –
103 (49.7) 104 (50.3) 67 (32.4)
BMI ¼ body mass index.
3.2. Clinical course and TPX2 expression in ccRCC TMA cohort To validate our findings from the TCGA cohort, we constructed a TMA of 207 patients diagnosed with ccRCC. The demographics and clinical characteristics of the TMA cohort are summarized in Table 1. The median follow-up time was 5.3 (0–20.8) years, and median times to recurrence and death were 12.8 (1–199) and 24.5 (1–146) months, respectively. Our cohort included 19/207 (9.18%) patients with Fuhrman grade 1 disease, 73 (35.3%) with T1 tumors, and 44 (21.3%) with metastasis. Recurrence occurred in 42
4
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
Table 2 Association of (A) TPX2 methylation with T stage (β ¼ 0.0403765, n ¼ 240) and (B) TPX2 mRNA expression with T stage (n ¼ 412) and metastasis (n ¼ 414) T stage
Low methylation (%)
High methylation (%)
P
(A) Association of TPX2 methylation with T stage (β ¼ 0.0403765, n ¼ 240) Patients 118 (49.1) 122 (50.9) 0.049 T1 44 (37.2) 67 (54.9) T2 18 (15.2) 11 (9.02) T3 53 (44.9) 42 (34.4) T4 3 (2.54) 2 (1.63) T and M stage Low mRNA High mRNA P expression (%) expression (%) (B) Association of TPX2 mRNA expression with T stage (n ¼ 412) and metastasis (n ¼ 414) Patients 308 (74.8) 104 (25.2) o0.001 T1 170 (55.2) 30 (28.8) T2 34 (11.0) 15 (14.4) T3 103 (33.4) 54 (51.9) T4 1 (0.32) 5 (4.81) Patients 309 (74.6) 105 (25.4) o0.001 M0 275 (80.9) 72 (68.6) M1 34 (11.0) 33 (31.4)
(20.3%) patients. At the time of review, 65/207 (31.4%) patients were alive with no evidence of disease, 15 (7.25%) were alive with disease, and 23 (11.1%) were alive with unknown status. A total of 70 (33.9%) patients died of disease, 16 (7.73%) died with no evidence of disease, 17 (8.2%) died with unknown disease status, and 1 patient died of other causes. Immunohistochemical analysis demonstrated TPX2 protein is expressed in the cytoplasm, consistent with The Human Protein Atlas tissue repository (Fig. 3) [13,23]. To stratify patients by TPX2 protein expression, we defined high expression as greater than the 75th percentile staining intensity according to mean H-score of the paired tissue cores. High expression of TPX2 was identified in 47 (22.7%) patients (Table 3). We found a significant correlation of multiple adverse clinical and pathological features with high expression of TPX2 (Table 3). Elevated TPX2 expression was associated with advanced Fuhrman grade and AJCC stage (P ¼ 0.001 and P ¼ 0.005, respectively; Table 3). Additionally, a higher incidence of metastatic disease was observed in the elevated TPX2 cohort (P ¼ 0.016, Table 3).
Fig. 1. Kaplan-Meier plot of overall and recurrence-free survival according to high (red) vs. low (blue) TPX2 methylation and mRNA expression in the TCGA cohort. (A) OS, TPX2 methylation (n ¼ 240, P ¼ 0.0223). (B) RFS, TPX2 methylation (n ¼ 195, P ¼ 0.0001). (C) OS, TPX2 mRNA expression (n ¼ 415, P o 0.001). (D) RFS, TPX2 mRNA expression (n ¼ 344, P o 0.001).
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
Fig. 2. Comparison of TPX2 and Aurora-A mRNA expression in TCGA cohort (n ¼ 415, Pearson ¼ 0.918, Spearman ¼ 0.812, P o 0.001).
To determine whether TPX2 protein was elevated in the histologically normal kidney in patients with high TPX2 expressing tumors, we evaluated their corresponding benign tissue cores. A Student's t-test demonstrated mean TPX2 protein expression in those with TPX2-rich tumors was
5
significantly increased in the corresponding benign tissue compared to the rest of the cohort (173.89 vs. 142.18, P ¼ 0.0028, Table 3). Thus, those patients with the highest expression of TPX2 in their tumor also had higher expression levels in their histologically benign renal cortex from the same kidney. Univariate analysis demonstrated high TPX2 protein expression in the tumor was associated with poor OS, CSS, and RFS (P ¼ 0.0327, P o 0.01, and P ¼ 0.0313, respectively; Fig. 4, Supplementary Fig. 1, and Supplementary Table 1). Actuarial estimated OS at 5 years for the highly expressed TPX2 subset was 47% (95% CI: 31.9%–60.6%) compared with 62.3% (95% CI: 53.4%– 69.6%) in the remaining group. The 5-year CSS rate for this same subset was 52% (95% CI: 36.5%–65.8%). The 5-year recurrence-free rate for the highly expressed TPX2 cohort was 73.6% (95% CI: 55.3%–85.3%) compared with 86.9% (95% CI: 78.4%–92.3%) in the lower expressing cohort. We constructed a multivariable model to determine whether TPX2 protein expression served as an independent predictor for overall and RFS. After correcting for age, Fuhrman grade, stage, lymph node status, and metastasis (the latter only for OS
Fig. 3. Weak (A, H-score ¼ 0), intermediate (B, H-score ¼ 40), and strong (C, H-score ¼ 100, 475th percentile) TPX2 immunostaining in ccRCC specimen (20). H-score, determined by DAB staining, reflects antibody binding and intensity of staining.
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎
6
Table 3 Comparison of normal vs. high TPX2 immunostaining. High TPX2 staining intensity was defined as greater than the 75th percentile mean H-score Clinical information
Normal TPX2 protein expression (%)
Elevated TPX2 protein expression (%)
Patients Age (mean, yr) Sex Male Female Race White Black Hispanic Unknown BMI (mean) Margins Positive Negative Sarcomatoid Lymph node status Fuhrman grade 3 or 4 AJCC stage III or greater Metastasis Recurrence Corresponding benign tissue TPX2 expression
159 (76.8) 61.72
47 (22.7) 59.225
122 (76.7) 37 (33.3)
29 (61.7) 18 (38.3)
146 (91.8) 11 (18.6) 2 (3.3) 0 31.5
40 (85.1) 6 (12.7) 0 1 (2.1) 28.5
8 151 11 7 68 75
6 41 5 6 33 33
P
0.201 0.041
0.129
(5.0) (95.0) (6.9) (4.4) (42.7) (47.2)
28 (17.6) 26 (16.4) 142.18
(12.7) (87.3) (10.6) (2.1) (70.2) (70.2)
16 (34.0) 16 (34.0) 173.89
0.472 0.064
0.403 0.165 0.001 0.005 0.016 0.002 0.0028
evaluation), our model demonstrated high TPX2 expression is an independent predictor of RFS (hazard ratio ¼ 3.62, P ¼ 0.029, 95% CI: 1.13–11.55; Table 4A), but not for OS or CSS (P 4 0.05, Table 4B and C). 4. Discussion Before our study, TPX2 had not been investigated in the context of ccRCC. Here, we show in 2 independent cohorts
that elevated mRNA expression and TPX2 protein levels are associated with more aggressive ccRCC and worse outcomes. TPX2 mRNA expression is also positively correlated in the TCGA with Aurora-A levels, which suggests this could be one such mechanism by which TPX2 could influence oncologic outcomes. Immunohistochemical analysis of our TMA cohort validated our TCGA findings demonstrating increased protein expression of TPX2 is associated with advanced Fuhrman grade, AJCC T stage, metastatic potential, worse overall survival, CSS, and is a significant predictor of recurrence. TPX2 has gained attention as a possible prognostic indicator and therapeutic target in human malignancies such as pancreatic adenocarcinoma, hepatocellular carcinoma, and NHL [11,14,24]. Although our study demonstrates an association of TPX2 with ccRCC disease severity, this does not imply a direct causal relationship exists. Previous studies investigating TPX2 in malignant processes demonstrated causal tumorigenic relationship via knockdown with specific TPX2-siRNA in mice [24–26]. It was demonstrated that TPX2 activates Aurora-A by binding an allosteric site, thereby increasing Aurora-A's binding affinity to ATP and substrate [27]. Furthermore, the binding of TPX2 to Aurora-A reduces the affinity of VX-680/MK0457, an Aurora-A inhibitor [25,27,28]. Alternatively, it was shown that TPX2-siRNA transfected colorectal carcinoma cells up-regulated AKT expression, suggesting TPX2 may induce tumorigenesis via the PI3K/Akt/mTOR pathway [21]. Similar studies with ccRCC would be useful for creating next steps to determine if this disease exhibits similar tumorigenic routes, and also if targeting TPX2 may be a superior approach to Aurora-A targeted therapy. If TPX2 does directly propagate ccRCC tumorigenesis, this may open the door to multiple novel therapeutic approaches currently under investigation. In pancreatic adenocarcinoma, TPX2-siRNA transfected cells demonstrated unusually high sensitivity to paclitaxel therapy [24]. Targeted TPX2 inhibition in ccRCC cells may provide
Fig. 4. Kaplan-Meier plot showing overall and recurrence-free survival according to high (red) vs. low (blue) TPX2 protein expression in the TMA cohort. (A) OS, TPX2 protein expression (n ¼ 203, P ¼ 0.0327). (B) RFS, TPX2 protein expression (n ¼ 163, P ¼ 0.0313).
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎ Table 4 Multivariate analysis using Cox proportional hazards evaluating TPX2 protein expression's potential role as a biomarker in RFS (A, n ¼ 163), OS (B, n ¼ 207), and CSS (C, n ¼ 207) Variable
HR
95% CI
P 4 |z|
(A) RFS (n ¼ 163) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status
3.63 2.12 3.41 1.00 0.73
1.14–11.55 0.66–6.78 0.87–13.39 0.95–1.06 0.14–3.93
0.029 0.205 0.079 0.889 0.714
(B) OS (n ¼ 207) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status Metastasis
1.29 2.09 3.32 1.00 0.76 2.30
0.58–2.90 0.84–5.19 0.92–11.93 0.97–1.04 0.28–2.12 0.86–6.15
0.534 0.114 0.066 0.98 0.605 0.099
(C) CSS (n ¼ 207) Elevated TPX2 protein expression Fuhrman grade III or IV AJCC stage T2 or greater Age Lymph node status Metastasis
1.58 2.60 4.21 1.00 0.80 2.25
0.63–3.96 0.91–7.46 0.90–19.73 0.96–1.04 0.27–2.30 0.82–6.17
0.327 0.075 0.068 0.845 0.674 0.117
7
potential to have unmeasured confounders bias our results. Although both cohorts are relatively large by the standards of biomarker studies, nevertheless a larger cohort may demonstrate additional or alternate associations. Furthermore, the number of lymph nodes removed and the locations of metastatic disease were not recorded in both cohorts, which may have provided greater insight to the association of TPX2 with advanced disease. Additionally, this study demonstrates an association between TPX2 and Aurora-A and outcomes in ccRCC, but we do not have specific demonstration of a mechanistic link between TPX2 and ccRCC. 5. Conclusion In conclusion, we show in 2 disparate cohorts that TPX2, a regulator of Aurora-A, is associated with high grade and stage of ccRCC, and is an independent predictor of disease recurrence. Further studies are warranted to elucidate its role in ccRCC tumor biology and potential as a therapeutic target. Acknowledgments
HR ¼ hazard ratio.
a similar optimized therapeutic environment, though admittedly cytotoxic chemotherapy is not standard of care in ccRCC. In hepatocellular carcinoma, T-cell mediated immunotherapy using cytotolytic T lymphocytes (CTLs) targeting TPX2-specific peptides has demonstrated early promise [11]. In addition, the TPX2-Aurora-A complex was identified as a likely cellular target of withanone, and may be of therapeutic benefit [29]. These studies highlight the potential of TPX2-targeting therapies and demonstrate the potential application to ccRCC management. Moreover, staining histologically normal tissue for TPX2 may serve as a companion diagnostic test if targeted therapy for this pathway is considered for a patient. Interestingly, evaluation of the corresponding histologically benign tissue in patients with ccRCC showed that the cohort with elevated TPX2 in their tumor also demonstrated significantly higher protein expression in the histologically normal renal parenchyma. As these individuals had a greater propensity to develop high-grade tumors and recurrent disease, this observation raised the intriguing hypothesis that TPX2 expression may help drive the development of aggressive ccRCC. As TPX2 was not a significant predictor of survival in these cohorts, this may imply that although factors such as age, Fuhrman grade, and AJCC stage better predict overall and cancer-specific mortality, TPX2 may be a primary driver of recurrence. In this scenario, staining tumor specimens for TPX2 may inform providers if closer follow-up is warranted. Our study does have several limitations. Both the TCGA and our own cohort are retrospective studies, so there is the
The results published here are in whole or part based on data generated by the TCGA Research Network in adherence to the TCGA publication guidelines: http://cancer genome.nih.gov/. Whole slide imaging was performed in the Digital Histology Shared Resource at Vanderbilt University Medical Center (www.mc.vanderbilt.edu/dhsr). This project was supported by the Inquiry Program of the Office of Medical Student Research at Vanderbilt University School of Medicine and the Vanderbilt University Medical Center. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j. urolonc.2016.12.012.
References [1] Clark PE. The role of VHL in clear-cell renal cell carcinoma and its relation to targeted therapy. Kidney Int 2009;76:939–45. [2] Siegel RL, Miller KD, Jemal A. Cancer statistics, 2016. CA Cancer J Clin 2016;66:7–30. [3] Ferchichi I, Kourda N, Sassi S, Romdhane KB, Balatgi S, Cremet JY, et al. Aurora A overexpression and pVHL reduced expression are correlated with a bad kidney cancer prognosis. Dis Markers 2012;33:333–40. [4] Calvo E, Schmidinger M, Heng DY, Grunwald V, Escudier B. Improvement in survival end points of patients with metastatic renal
8
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
Z.A. Glaser et al. / Urologic Oncology: Seminars and Original Investigations ] (2017) ∎∎∎–∎∎∎ cell carcinoma through sequential targeted therapy. Cancer Treat Rev 2016;50:109–17. Schonenberger D, Harlander S, Rajski M, Jacobs RA, Lundby AK, Adlesic M, et al. Formation of renal cysts and tumors in Vhl/Trp53deficient mice requires HIF-1alpha and HIF-2alpha. Cancer Res 2016;76:2025–36. Dere R, Perkins AL, Bawa-Khalfe T, Jonasch D, Walker CL. betacatenin links von Hippel-Lindau to aurora kinase A and loss of primary cilia in renal cell carcinoma. J Am Soc Nephrol 2015;26:553–64. Kufer TA, Sillje HH, Korner R, Gruss OJ, Meraldi P, Nigg EA. Human TPX2 is required for targeting Aurora-A kinase to the spindle. J Cell Biol 2002;158:617–23. Jin S, Wang X, Tong T, Zhang D, Shi J, Chen J, et al. Aurora-A enhances malignant development of esophageal squamous cell carcinoma (ESCC) by phosphorylating beta-catenin. Mol Oncol 2015;9:249–59. Li Y, Zhou W, Wei L, Jin J, Tang K, Li C, et al. The effect of Aurora kinases on cell proliferation, cell cycle regulation and metastasis in renal cell carcinoma. Int J Oncol 2012;41:2139–49. Gruss OJ, Wittmann M, Yokoyama H, Pepperkok R, Kufer T, Sillje H, et al. Chromosome-induced microtubule assembly mediated by TPX2 is required for spindle formation in HeLa cells. Nat Cell Biol 2002;4:871–9. Aref AM, Hoa NT, Ge L, Agrawal A, Dacosta-Iyer M, Lambrecht N, et al. HCA519/TPX2: a potential T-cell tumor-associated antigen for human hepatocellular carcinoma. Onco Targets Ther 2014;7:1061–70. Neumayer G, Belzil C, Gruss OJ, Nguyen MD. TPX2: of spindle assembly, DNA damage response, and cancer. Cell Mol Life Sci 2014;71:3027–47. Schatz CA, Santarella R, Hoenger A, Karsenti E, Mattaj IW, Gruss OJ, et al. Importin—alpha-regulated nucleation of microtubules by TPX2. EMBO J 2003;22:2060–70. Chowdhury A, Chowdhury S, Tsai MY. A novel Aurora kinase A inhibitor MK-8745 predicts TPX2 as a therapeutic biomarker in nonHodgkin lymphoma cell lines. Leuk Lymphoma 2012;53:462–71. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov 2012;2:401–4. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Sci Signal 2013:6:pl1.
[17] Cancer Genome Atlas Research Network: Comprehensive molecular characterization of clear cell renal cell carcinoma. Nature 2013;499:43–9. [18] Fuhrman SA, Lasky LC, Limas C. Prognostic significance of morphologic parameters in renal cell carcinoma. Am J Surg Pathol 1982;6:655–63. [19] Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th edition of the AJCC cancer staging manual and the future of TNM. Ann Surg Oncol 2010;17:1471–4. [20] Yi Y, Polosukhina D, Love HD, Hembd A, Pickup M, Moses HL, et al. A Murine model of K-RAS and beta-catenin induced renal tumors expresses high levels of E2F1 and resembles human Wilms tumor. J Urol 2015;194:1762–70. [21] Wei P, Zhang N, Xu Y, Li X, Shi D, Wang Y, et al. TPX2 is a novel prognostic marker for the growth and metastasis of colon cancer. J Transl Med 2013;11:313. [22] Detre S, Jotti GS, Dowsett M. A “quick score” method for immunohistochemical semiquantitation: validation for oestrogen receptor in breast carcinomas. J Clin Pathol 1995;48:876–8. [23] Uhlen M., et al. Proteomics. Tissue-based map of the human proteome. Science 2015;347:1260419. [24] Warner SL, Stephens BJ, Nwokenkwo S, Hostetter G, Sugeng A, Hidalgo M, et al. Validation of TPX2 as a potential therapeutic target in pancreatic cancer cells. Clin Cancer Res. 2009;15:6519–28. [25] Satow R, Shitashige M, Kanai Y, Takeshita F, Ojima H, Jigami T, et al. Combined functional genome survey of therapeutic targets for hepatocellular carcinoma. Clin Cancer Res 2010;16:2518–28. [26] Perez de Castro I, Malumbres M. Mitotic stress and chromosomal instability in cancer: the case for TPX2. Genes Cancer 2012;3: 721–30. [27] Eyers PA, Erikson E, Chen LG, Maller JL. A novel mechanism for activation of the protein kinase Aurora A. Curr Biol 2003;13:691–7. [28] Anderson K, Yang J, Koretke K, Nurse K, Calamari A, Kirkpatrick RB, et al. Binding of TPX2 to Aurora A alters substrate and inhibitor interactions. Biochemistry 2007;46:10287–95. [29] Grover A, Singh R, Shandilya A, Priyandoko D, Agrawal V, Bisaria VS, et al. Ashwagandha derived withanone targets TPX2-Aurora A complex: computational and experimental evidence to its anticancer activity. PLoS One 2012;7:e30890.