T-box Transcription Factor Brachyury Is Associated with Prostate Cancer Progression and Aggressiveness

Prostate cancer is the most common malignancy in men and the second leading cause of cancer-related deaths worldwide. In the United States, prostate cancer is the leading cause of cancer-related mortality (1). Despite advances in prevention and early detection, refinements in surgical techniques, and improvements in adjuvant radiotherapy and chemotherapy, metastasis is a frequent event that hinders patients' cure. One important mechanism that governs cancer cell invasion and further metastasis is cellular epithelial–mesenchymal transition (EMT; ref. 2). The EMT is a complex process that involves downregulation of epithelial markers, such as E-cadherin, and upregulation of mesenchymal markers, such as Snail, Slug, and N-cadherin, among other alterations. These lead to loss of epithelial cell polarity and acquisition of more motile and invasive phenotypes, promoting cancer cell dissemination into distant sites (3).

The T-box protein Brachyury is a transcription factor required for mesoderm specification during embryo development (4), which is widely expressed in notochord cells and plays a pivotal role in notochord development (5). Recently, Brachyury was associated with tumor aggressiveness in several tumor types (6–11) and was found to be a significant predictor of poor prognosis in early colon cancer (8) and lung cancer (6). In vitro studies suggested that these associations are driven by EMT, accomplished by increased migratory and invasion capacity (12–14) and increased cancer stem cell features (10, 11). Different studies have reported divergent effects of Brachyury expression on cell proliferation. In lung cancer cell lines, it was demonstrated that Brachyury blocks cell-cycle progression and, therefore, mediates tumor resistance (15). However, in adenoid cystic carcinoma cells, Brachyury promoted tumor growth and metastasis formation in vivo (11). Therefore, despite the described oncogenic role of Brachyury, some authors suggest that it can also act as a tumor suppressor gene (16).

A recent in vitro study showed that Brachyury overexpression promoted cell invasion in prostate cancer, probably mediated by TGFβ1 production (13). However, knowledge on the role of Brachyury in prostate cancer progression remains very limited. In the present work, we investigated the clinical impact of Brachyury expression in a well-characterized cohort of human prostate cancer samples and evaluated its biologic role in prostate cancer cell proliferation and invasiveness. We report that Brachyury is overexpressed in primary prostate cancer and metastatic tissues and that Brachyury expression is correlated with classic parameters of prostate cancer progression and aggressiveness. We also provide data that suggest Brachyury as a therapeutic target in prostate cancer treatment.

Prostate tissues were obtained from 480 patients with a 64-year-old median age (range, 46–74), who performed radical prostatectomy as primary therapy (no preceding hormonal or radical therapy) from 1993 to 2010 at Centro Hospitalar do Porto and Centro Hospitalar do Alto Ave-Guimarães, Portugal. The series included a total of 211 nonneoplastic tissue, 143 high-grade prostate intraepithelial neoplasia (PIN) lesions, and 409 primary prostate carcinomas. High-grade PIN lesions and nonneoplastic tissues were obtained from tumor adjacency. Thirteen normal samples were obtained from patients undergoing radical cystoprostatectomy for transitional cell carcinoma of the bladder. Nine metastatic tissues were obtained from patients who performed biopsies for metastatic prostate cancer. Formalin-fixed and paraffin-embedded tumors and clinicopathologic data were retrieved from the files of the Department of Pathology of both the hospitals. Tumors were staged using the 2010pTNM American Joint Committee on Cancer (AJCC) classification (17) and graded using the Gleason grading system 2005 (18). Samples were organized into tissue microarray (TMA) as previously described (19). The histologic features of the sampled areas were representative of the final Gleason score for the case. The study was previously approved by Local Ethical Review Committee of Centro Hospitalar do Porto (ref. no. 017/08-010-DEFI/015-CES).

Five human prostate cell lines representing in vitro models of prostate cancer progression and aggressiveness, PNT2, 22RV1, LNCaP, PC3, and DU145 (ATCC), were grown in RPMI-1640 medium supplemented with 10% FBS (GIBCO, Invitrogen) and 1% penicillin/streptomycin (P/S; GIBCO, Invitrogen). PNT2 is a normal prostate cell line, 22RV1 is a prostate epithelial carcinoma cell line, LNCaP is derived from lymph node metastasis and is hormone-sensitive, and DU145 and PC3 cell lines are derived from brain and bone metastasis, respectively, and represent poorly differentiated tumors.

22RV1 and DU145 cell lines were transfected with full-length human Brachyury in pcDNA4/TO vector, thus designated pcBrachyury. The empty vector (designated 4/TO) was used as control (12). Stable 22RV1 and DU145 cell pools with pcBrachyury expression were obtained following treatment with 50 μg/mL zeocin (Invitrogen). PC3 cells were transfected with Brachyury-specific shRNA construct (shBrachy.1) or empty vector alone (pLKO.1; Sigma-Aldrich) using X-tremeGENE HP transfection reagent (Roche) as recommended by the manufacturer. Stable PC3 cells with depleted endogenous Brachyury expression were obtained following treatment with 5 μg/mL of puromycin (Sigma-Aldrich).

Total RNA was extracted from cell lines using TRIzol Reagent (Invitrogen S.A.). One microgram of RNA was reverse-transcribed using Phusion RT-PCR Kit (Finnzymes), as recommended by the manufacturer. The primers used are presented in Supplementary Table S1. No amplification was obtained when RNA was mock-transcribed without adding reverse transcriptase.

Cells were lysed in buffer containing 50 mmol/L Tris, pH 7.6–8, 150 mmol/L NaCl, 5 mmol/L EDTA, 1 mmol/L Na₃VO₄, 10 mmol/L NaF, 10 mmol/L Na pyrophosphate, 1% NP-40, and 1/7 of protease cocktail inhibitors (Roche). Proteins were resolved on standard 12% SDS-PAGE gels, transferred onto nitrocellulose membranes, and probed with antibody against Brachyury (AF2085, R&D Systems) and GAPDH (sc-69778, Santa Cruz Biotechnology, Inc.) at 4°C overnight. Blot detection was done by chemiluminescence (ECL Western Blotting Detection Reagents, GE Healthcare) using Chemidoc (Bio-Rad).

Cells were plated on glass coverslips placed into 12-well plates and allowed to adhere overnight. Cells were fixed with 4% paraformaldehyde (PFA) in 1X PBS, washed, and permeabilized with 0.1% Triton X-100. Then, cells were blocked in 10% FBS, labeled for 1 hour at room temperature with primary anti-Brachyury antibody (sc-20109, Santa Cruz Biotechnology, Inc.), washed, and incubated at room temperature for 1 hour with a secondary anti-rabbit Alexa-488 antibody (Invitrogen-Molecular Probes). Coverslips were mounted on microscope slides with Vectashield Mounting Medium with DAPI (Vector Laboratories). Digital images were recorded with Olympus BX61 (Olympus Corporation).

Colony formation assays were used to assess the survival capacity of 22RV1, DU145, and PC3 cells with and without Brachyury. A total of 1 × 10³ cells per well were seeded into 6-well plates. After 15 days of culture, colonies formed were fixed in 4% PFA, washed, stained with 0.05% crystal violet, and manually counted.

MTS and bromodeoxyuridine (BrdUrd) assays were used to evaluate the viability and proliferation capacity over time. A total of 2 × 10³ cells per well for 22RV1 and 1 × 10³ cells per well for DU145 and PC3 were plated into 96-well plates in triplicate and allowed to adhere overnight. After 6 hours of starvation (RPMI only), viable or proliferative cells were quantified using the Cell Titer96 Aqueous cell proliferation (MTS, Promega) or Cell Proliferation ELISA, BrdUrd (colorimetric, Roche Applied Science) assay and this was the value for time 0. After 24, 48, and 72 hours, cell viability and proliferation were again assessed. The results were calibrated to the starting value (time 0 hours, considered as 100% of viability/proliferation) as previously described (20).

Cells were seeded in 12-well plates and cultured to at least 95% of confluence. Monolayer cells were washed with 1X PBS and scraped with a plastic pipette tip and then incubated with fresh RPMI medium. The “wounded” areas were photographed by phase-contrast microscopy at different time points. The relative migration distance was calculated as described (20, 21).

Matrigel invasion assays were performed using 8-μm pore size BD BioCoat Matrigel Invasion Chambers (BD Biosciences). Briefly, after rehydration with RPMI, 10% FBS, the upper compartment of the chamber received 2.5 × 10⁴ cells per well grown in RPMI only, whereas the lower compartment contained fresh medium supplemented with 10% FBS and 10 ng/mL of EGF (Prepotech). After 22 hours of incubation, the upper surface of the filter was washed with 1X PBS and fixed with 4% PFA. Then, residual cells were cleared with a cotton swab, the filter washed with 1X PBS, and invasive cells attached to the lower filter surface were mounted in Vectashield Mounting Medium with DAPI (Vector Laboratories). Images were recorded on an Olympus BX61 microscope (Olympus Corporation), and invasive cells counted using ImageJ software.

Histologic slides with 4-μm-thick tissue sections were subjected to IHC analysis according to the streptavidin–biotin peroxidase complex system (UltraVision Large Volume Detection System Anti-Polyvalent, HRP; LabVision Corporation), using the primary antibody raised against Brachyury (diluted 1:200; sc-20109, Santa Cruz Biotechnology, Inc.) or against AMACR (diluted 1:50; 504R-16, Cell Marque). CD44 staining was performed using an anti-human CD44 antibody (diluted 1:100; 156-3C11, AbD Serotec) and detected using Vectastain Universal Elite ABC kit PK-6200 (Vector Laboratories). The negative control was treated identically but with omitted primary antibody. Sections were scored in a double-blind fashion for cytoplasm expression following a semiquantitative criterion based on the intensity (0, negative; 1, weak; 2, moderate; 3, strong) and percentage of cells stained (0, 0% of immunoreactive cells; 1, <25% of immunoreactive cells; 2, 25%–50% of immunoreactive cells; and 3, >50% of immunoreactive cells). Both components were considered for an overall semiquantitative staining score (range, 0–6). Samples with scores 0, 1, and 2 were considered negative and those with scores 3–6 were considered positive. Tissues sections were separately evaluated for expression in the nucleus (≥25% nuclear staining was considered positive and cases with <25% of nuclear staining were considered negative).

Brachyury mRNA expression was assessed in 7 prostate cancer datasets (LaTulippe, ref. 22; Varambally, ref. 23; Grasso, ref. 24; Taylor, ref. 25; Glinsky, ref. 26; Yu, ref. 27; TCGA, ref. 28; and Arredouani, ref. 29) from the Oncomine database (30, 31). Categorization of patients with Brachyury-positive and Brachyury-negative prostate cancer was based on the log₂ median-centered intensity values of Brachyury probes per study, and a linear model was fitted to estimate the association significance. Patient samples in each study with Brachyury expression values greater than its median intensity were grouped as Brachyury-positive and others were grouped as Brachyury-negative. Brachyury expression was further correlated with corresponding patient clinical data available.

Microarray co-expression studies were extracted from the Oncomine database. Microarray expression profiles were clustered by functional importance and signaling pathways using DAVID v6.7 bioinformatic tool (The Database for Annotation, Visualization and Integrated Discovery; refs. 32, 33).

Univariate (χ² test) and multivariate analyses (linear regression model) were used to assess the correlations between Brachyury expression and clinicopathologic features from primary specimens. Pearson test was used to evaluate the correlation between gene expression profiles. Simple comparisons between 2 different conditions were analyzed using the Student t test and, for comparison of 2 conditions over time, we used the 2-way ANOVA (Bonferroni post-test). The statistical analysis was performed using SPSS software (version 19.0) or using Prism GraphPad software (version 5.0a). The level of significance in the statistical analyses is indicated as *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Brachyury protein expression was assessed by IHC in a series of 784 prostate tissues, including normal tissues, PIN lesions, primary prostate cancer samples with different Gleason scores and prostate cancer metastasis. Figure 1 shows representative results of intensity scores observed for Brachyury expression. Normal prostate gland and adjacent nonneoplastic tissues presented absence or low levels of Brachyury staining when compared with neoplastic tissues (Fig. 1A; Supplementary Fig. S1) and were therefore clustered in a single group, designated nonneoplastic tissues. Brachyury was expressed in the nuclei and/or cytoplasm of epithelial cells in nonneoplastic tissues, PIN lesions, prostate cancer, and metastatic tissues (Fig. 1B–D; Supplementary Fig. S1). Overall, the number of cases presenting cytoplasm protein expression increased from nonneoplastic to prostate cancer and PIN lesions and to metastasis (33.9%, 55.2%, 61.5%, and 100% of positive cases, respectively; P < 0.001; Fig. 2A). Brachyury nuclear staining was present in a comparable number of cases in nonneoplastic (25.0%), PIN lesions (38.6%), and prostate cancer cases (25.4%), in contrast to 100% of metastatic tissue samples (P < 0.001; Figs. 1D and 2A). Interestingly, Brachyury was also detected in the stroma (Supplementary Fig. S1) with a significant reduction of stroma-positive cases from nonneoplastic tissues (52.6%), to PIN lesions (44.2%), and to prostate cancer (14.2%; P < 0.001; Fig. 2A), indicating a possible role for Brachyury in prostate cancer tumor tissue microenvironment.

Heatmap analysis of overall Brachyury protein expression showed that Brachyury is remarkably overexpressed in PIN, prostate cancer, and metastatic prostate tissues when compared with nonneoplastic tissues (Fig. 2B). The higher expression profile was found in metastasis with scores consistently ≥4.

The clinical impact of Brachyury protein expression levels was further explored in our cohort of 409 primary prostate cancer. Univariate outcome analysis showed that Brachyury-positive cases (scores ≥ 3) are significantly (P < 0.001) associated with the prostate cancer biomarker α-methylacyl-CoA racemase (AMACR; Supplementary Table S2). Brachyury-positive cases correlated with highly undifferentiated prostate cancer tumors (P = 0.007, Table 1) and, concordantly, a strong tendency to be associated with stem cell marker CD44 was observed (P = 0.054; Supplementary Table S2). Importantly, Brachyury protein levels increased with the Gleason score (P < 0.027, Table 1; P < 0.01, Supplementary Fig. S2). To evaluate the clinical impact of the presence of Brachyury in the nucleus, a comparison between Brachyury nuclei–positive and nuclei-negative in prostate cancer–positive cases was performed (nonneoplastic tissues, n = 76; PIN lesion, n = 88; prostate cancer, n = 228; metastasis, n = 9). Primary prostate cancer tumors with nuclear Brachyury staining were significantly associated with perineural invasion (P = 0.046) and with capsular invasion (P = 0.025; Table 1), which is in agreement with predominant nuclear expression in metastasis (Figs. 1 and 2). Because Gleason scores and pT stage are known prognostic biomarkers, we performed multivariate analysis to determine whether high Brachyury expression has an independent statistical value. We observed that Brachyury is significantly associated with capsular invasion (P = 0.027, Table 1) on primary prostate cancer samples.

No significant correlations were found for Brachyury staining in stroma with clinicopathologic parameters by univariate analysis (Table 1). Yet, the multivariate analysis showed a significant association with capsular invasion (P = 0.030, Table 1), indicating the possible role of Brachyury on tumor microenvironment.

Altogether, these data suggest that high Brachyury levels are associated with patient's poor outcome and indicate that nuclear Brachyury staining in prostate cancer is an independent prognostic factor.

To corroborate our findings, we extended the analysis to microarray profiling datasets of prostate cancer tissues available on the Oncomine database (30, 31). Brachyury mRNA expression was analyzed in 6 independent prostate cancer datasets (LaTulippe, ref. 22; Varambally, ref. 23; Grasso, ref. 24; Taylor, ref. 25; Yu, ref. 27; and Arredouani, ref. 29) comprising a total of 97 normal prostate gland, 304 prostate cancer, and 83 prostate cancer metastasis samples. We found that Brachyury was significantly overexpressed in prostate tissues in multiple microarray cancer profiling datasets, in particular in metastatic prostate cancer (Fig. 2C). This was concordant with our protein analysis reported above. Importantly, although multiple probes were used to determine Brachyury mRNA levels in these datasets (23996_at, 206524_at, A_24_P63642, 7679), they all consistently showed that Brachyury overexpression is a common event in primary and metastatic prostate cancer (Fig. 2C).

We next assessed the impact of Brachyury in prostate cancer prognosis at the mRNA level, exploring the microarray profiling datasets of localized prostate tumors with clinical data from Oncomine (Table 2). Univariate statistical analysis revealed that high levels of Brachyury expression correlated with higher (>7) Gleason scores for the Taylor (25), Glinsky (26), and LaTulippe (22) datasets (P = 0.043, P = 0.042, and P = 0.049, respectively; Table 2). In agreement with protein IHC levels, the percentage of Brachyury-positive cases directly increased with the Gleason score (P < 0.05; Supplementary Fig. S2). In addition, high Brachyury mRNA levels correlated with pT (Yu, ref. 27; P = 0.033) and N stage (Taylor, ref. 25; P = 0.043), biochemical recurrence (Taylor, ref. 25; P = 0.048), capsular invasion and extraprostatic extension (Glinsky, ref. 26; P = 0.002 and 0.007, respectively; Table 2). A similar tendency could also be observed in the other datasets. In the multivariate analysis, we found that high Brachyury mRNA levels still correlate with capsular invasion and extraprostatic extension (Glinsky, ref. 26; P = 0.001 and 0.032, respectively) and with biochemical recurrence (TCGA, ref. 28; P = 0.004; Table 2).

To explore the biologic role of Brachyury in prostate cancer aggressiveness, 5 prostate cancer cell lines (PNT2, 22RV1, LnCaP, PC3, and DU145), representative of different degrees of prostate cancer progression, were screened for Brachyury expression by semiquantitative RT-PCR and Western blot analyses. Brachyury protein subcellular localization was additionally evaluated by immunofluorescence. We observed an absence of Brachyury at both mRNA and protein levels in the nonmalignant prostate cell line (PNT2) and in the primary prostate cancer cell line (22RV1; Supplementary Fig. S3A). In contrast, the metastatic cell lines LNCaP and PC3 showed strong nuclear and cytoplasm Brachyury expression both at mRNA and at protein levels (Supplementary Fig. S3A). The metastatic DU145 cell line was negative for mRNA by conventional RT-PCR but still exhibited low levels of nuclear protein expression (Supplementary Fig. S3A). These findings indicate that prostate cell lines are good models to study the functional role of Brachyury in prostate cancer cells, as they recapitulate the expression profiles found in human clinical samples.

To address whether the modulation of Brachyury expression influences the tumorigenic properties of prostate cancer cells, Brachyury was overexpressed in primary (22RV1) and metastatic (DU145) prostate cancer cell lines. Successful ectopic overexpression was obtained upon transfection of both cell lines with the pcBrachyury expression vector and Brachyury protein exhibited nuclear localization (Supplementary Fig. S3B). To investigate the effect of Brachyury inhibition, a specific short-hairpin clone (shBrachy.1) was used to deplete Brachyury in a positive metastatic prostate cell line (PC3; Supplementary Fig. S3B).

We initially studied the biologic role of Brachyury on prostate cancer cell viability and proliferation (Fig. 3A; Supplementary Fig. S4A). pcBrachyury prostate cells had a significant (P < 0.05) viability advantage over time (MTS assay) compared with the cells transfected with the empty vector (4/T0). Colony formation assays revealed a significant (P < 0.05) increase in the number of the colonies formed in the pcBrachyury-transfected cells when compared with the control cells (Fig. 3A, Supplementary Fig. S4A). The opposite findings were obtained with Brachyury depletion in shBrachy.1-PC3 cells (P < 0.05; Fig. 3A). To determine whether this viability advantage was due to higher proliferation rates, we analyzed BrdUrd incorporation during S-phase of the cell cycle. The presence of Brachyury, whether endogenous or exogenously overexpressed, promoted higher rates of proliferation over time (P < 0.05; Fig. 3A; Supplementary Fig. S4A). We further performed wound migration and Matrigel invasion assays in the transfected cell lines and observed that both 22RV1 and DU145 pcBrachyury cells had a higher migratory capacity over time and increased cell invasion capability than the empty vector cells (P < 0.05; Fig. 3B; Supplementary Fig. S4B). When Brachyury was depleted in an endogenously positive cell line, we were able to attenuate the aggressive behavior (P < 0.05; Fig. 3B).

To characterize the molecular players underlying prostate aggressiveness in vitro, we studied the expression profile of some key genes involved in EMT, migration, and stemness processes. We observed that Brachyury expression was associated with a decrease of the epithelial marker E-cadherin and concomitant increased expression of mesenchymal genes (N-cadherin, fibronectin, and Snail), as well as upregulation of metalloprotease MMP14 (Supplementary Fig. S5). Concordantly with IHC analysis in human prostate cancer (Supplementary Table S2), Brachyury overexpression was associated with an increased expression of the stem cell marker CD44 (Supplementary Fig. S5).

We clustered the genes co-expressed with Brachyury in prostate tissues available at the Oncomine database (30, 31) by their functional role and importance in signal transduction pathways using the DAVID bioinformatic tool (Fig. 4A). We found that the majority of genes co-expressed with Brachyury were functionally clustered in the categories of immune response, cell membrane/receptor activity, development, cell motility, and chemotaxis pathways in cancer and response to hormone stimulus. A subanalysis by KEGG signaling pathways revealed that Brachyury co-expressed genes are grouped in pathways associated with higher aggressiveness, namely, pathways in cancer, positive cell-cycle regulation, and immune response (Fig. 4B). These analyses strongly point to a role of Brachyury, not only in cell migration and invasion but also as a regulator of the cell cycle and in cancer microenvironment metabolism. Importantly, we found that the levels of Brachyury expression in prostate cancer tissues are directly correlated with those of IL8 and TGFβ1 (Fig. 4C), which are involved in EMT and cancer microenvironment modulation, as previously described (13, 14). Accordingly, there is a correlation of Brachyury expression with the expression of genes associated with EMT process (like fibronectin) and migration (MMP14; MMP24; Fig. 4C) that support our expression analysis (Supplementary Fig. S5).

The T-box transcription factor Brachyury was initially discovered for its role in mouse mesoderm development and differentiation (34). This involves massive conversion of epithelial cells into migratory and invasive mesenchymal cells during gastrulation via a process known as EMT (2, 4). Several reports have demonstrated that EMT is critical for prostate cancer progression, as acquisition of mesenchymal features may favor dissemination and resistance to therapy (35). High levels of Brachyury have previously been reported in various types of cancer (5–9, 11) and a phase I clinical trial of a vaccine targeting Brachyury-positive tumors (GI-6301) is currently under way (36, 37). Although prostate cancer is a leading cause of cancer-related deaths in men worldwide, a characterization of Brachyury biologic role in prostate tumorigenesis is missing.

Our study reports for the first time that the transcription factor Brachyury is aberrantly overexpressed across prostate malignancy and, in particular, nuclear Brachyury staining is associated with prostate invasion and prostate metastatic tissues. Our findings indicate Brachyury as an independent prognostic factor in prostate cancer. The role of Brachyury nuclear staining in metastasis was demonstrated in other tumor types, such as colorectal, lung cancer, and oral squamous cell carcinoma (6–9). Herein, we also found that cytoplasm immunostaining in prostate cancer is associated with prostate cancer biomarker AMACR and with highly aggressive tumors. The role of Brachyury in the cell cytoplasm remains to be elucidated, yet, we can hypothesize that it interacts with other proteins and in this way regulates cell behavior in a nontranscription manner.

It has been shown that Brachyury expression can influence tumor microenvironment through the release of soluble factors that could induce adjacent epithelial tumor cells to undergo an EMT and acquire metastatic potential (14). In prostate cancer, the activation and secretion to the extracellular environment of soluble factors that mediate the cross-talk between tumor cells and tumor stroma, such as interleukins and growth factors, has been reported to play a role in tumor progression (13, 14, 38–40). Our analysis confirm a previously result that demonstrated that Brachyury increases the expression and secretion of TGFβ1 in a prostate cell line (13). However, the influence of Brachyury expression in stromal cells has not yet been characterized. To our knowledge, we provide the first evidence for decreased Brachyury expression in stromal cells with prostate malignancy, at variance with the reported upregulation in tumor cells. Therefore, we can hypothesize that Brachyury has different roles in stromal and tumor cells and that it could be involved in the regulation of tumor microenvironment. In addition, we found that the majority of Brachyury co-expressed genes are involved in immune or metabolic processes.

By Brachyury overexpression and downregulation in prostate cancer cell lines, we demonstrated its role in tumor cell migration and invasion, as well as in cell viability and proliferation. Our findings were further corroborated by an in silico analysis with multiple genes functionally clustered in pathways related with cell motility and cell proliferation. A study performed by Shomoda and colleagues demonstrated that ablation of Brachyury in adenoid cystic carcinoma (ACC) cells decreased the number of metastasis and tumor size in vivo (11). Moreover, depletion of Brachyury in chordoma cells promotes a complete block of cell proliferation (41). An opposite role for Brachyury in cell proliferation was demonstrated in lung and colorectal cell lines by Huang and collaborators, where Brachyury blocks cell cycle progression and mediates tumor resistance to conventional antitumor therapies (15). Therefore, it can be deduced that the role of Brachyury may be tissue-specific or cell-type–dependent.

Brachyury seems to be a key driver of EMT in various human tumors by increasing expression of genes such as Slug, Snail, MMPs, IL8, and TGFβ1 (6, 9, 13–15, and current study). A possible link between cells undergoing EMT and cells with “stem cell–like” properties was recently described (42). The role of Brachyury in conferring stemness properties was already demonstrated in colorectal cancer cells (10) and in ACC cells (11). The present study shows that Brachyury is more represented in CD44⁺ prostate tissues, and ectopic Brachyury overexpression in vitro promotes CD44 expression. Because CD44⁺ prostate tumors are more resistant to the currently used therapies (43, 44), we speculate that Brachyury could have a role in prostate cancer therapy resistance. Future studies are warranted to elucidate this hypothesis.

In conclusion, the present work reports increased levels of Brachyury expression in localized and metastatic prostate cancer, with clinicopathologic significance and evidences a role for Brachyury in promoting prostate cancer cell growth and invasion. Our work further suggests new roles for Brachyury in prostate cancer, namely, in tumor microenvironment regulation and possibly in immune response. Clinical applicable prognostic biomarkers are needed for clinical management of patients with prostate cancer and our study positions Brachyury as a putative independent prognostic biomarker in prostate cancer and a possible therapeutic target for advanced prostate tumor patients.

No potential conflicts of interest were disclosed.

Conception and design: F. Pinto, R.P. Andrade, R.M. Reis

Development of methodology: F. Pinto, R.M. Reis

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Pinto, N. Pértega-Gomes, M.S. Pereira, P. Monteiro, R.M. Henrique, F. Baltazar

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Pinto, N. Pértega-Gomes, M.S. Pereira, J.R. Vizcaíno, R.M. Reis

Writing, review, and/or revision of the manuscript: F. Pinto, N. Pértega-Gomes, F. Baltazar, R.P Andrade, R.M. Reis

Study supervision: R.M. Reis

The authors thank Dr. Jeffrey Scholm for providing 4/T0 and pcBrachyury expressing vectors and to the journal's anonymous reviewers for helpful comments and suggestions.

This study was supported by the ICVS internal research funds of participating authors and by the FCT project, ref. PTDC/SAU-MET113415/2009. F. Pinto and N. Pértega-Gomes received fellowships from the FCT, ref. SFRH/BD/81369/2011 and SFRH/BD/61027/2009, respectively. R.P. Andrade was funded by Ciencia2007 Program Contract and Programa Operacional Regional do Norte (ON.2) — NORTE-07-0124-FEDER-000017.

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.