Purpose: Identification of tumor antigens is essential in advancing immune-based therapeutic interventions in cancer. Particularly attractive targets are those molecules that are selectively expressed by malignant cells and that are also essential for tumor progression.

Experimental Design and Results: We have used a computer-based differential display analysis tool for mining of expressed sequence tag clusters in the human Unigene database and identified Brachyury as a novel tumor antigen. Brachyury, a member of the T-box transcription factor family, is a key player in mesoderm specification during embryonic development. Moreover, transcription factors that control mesoderm have been implicated in the epithelial-mesenchymal transition (EMT), which has been postulated to be a key step during tumor progression to metastasis. Reverse transcription-PCR analysis validated the in silico predictions and showed Brachyury expression in tumors of the small intestine, stomach, kidney, bladder, uterus, ovary, and testis, as well as in cell lines derived from lung, colon, and prostate carcinomas, but not in the vast majority of the normal tissues tested. An HLA-A0201 epitope of human Brachyury was identified that was able to expand T lymphocytes from blood of cancer patients and normal donors with the ability to lyse Brachyury-expressing tumor cells.

Conclusions: To our knowledge, this is the first demonstration that (a) a T-box transcription factor and (b) a molecule implicated in mesodermal development, i.e., EMT, can be a potential target for human T-cell–mediated cancer immunotherapy.

Immunotherapeutic interventions against cancer depend on the identification of tumor antigens able to elicit a host immune response against the tumor cells. Particularly attractive targets are those molecules that are selectively expressed by malignant cells and that are also essential for malignant transformation and/or tumor progression. In recent years, the epithelial-mesenchymal transition (EMT) has been recognized as a key step during the progression of primary tumors into metastases (1). EMT describes a series of events during which cells lose epithelial characteristics, such as cell-layer organization and apical-basolateral polarization, and acquire properties of mesenchymal or fibroblastoid cells, including motility (2). Several molecules have been recently identified that play a key role in EMT during tumor progression (3), among them the transcription factors Twist, Snail, and Slug (4, 5). In the context of cancer immunotherapy, molecules that trigger EMT might function effectively because targeting them might prevent tumor invasion and metastasis.

In the present work, we have used a computer-based differential display (CDD) analysis tool, designated HSAnalyst, to conduct global comparison of expressed sequence tag (EST) clusters belonging to the human Unigene database (6, 7). Among a list of various genes highly represented in tumor-derived libraries and rarely observed in normal tissue-derived libraries was the Unigene cluster Hs.389457, which contains the complete mRNA sequence of human Brachyury, a homologue of the mouse Brachyury gene (8). The product of the Brachyury gene is a member of the T-box family of transcription factors, characterized by a highly conserved DNA-binding domain designated as T-domain (9). Orthologs of Brachyury have been identified in a vast variety of multicellular organisms, such as Caenorhabditis elegans, Xenopus, mouse, and human, among others (911). All T-box proteins studied thus far play key roles during early development, mostly in the formation and differentiation of normal mesoderm (12, 13). Mesoderm formation during embryo gastrulation is a typical example of an EMT process where the transcription factor Brachyury plays a crucial role (14). Brachyury is also transiently induced in vitro in rhesus monkey embryonic stem cells and mouse embryonal carcinoma cells undergoing mesodermal differentiation (15, 16). In light of these implications of Brachyury in the process of EMT and, thus, potentially in the progression of tumors, we have examined Brachyury as a target for T-cell–mediated cancer immunotherapy. Our analysis of the expression of the human Brachyury gene showed that Brachyury mRNA is present in various human tumor tissues and cancer cell lines. By using a predictive algorithm, we have identified an HLA-A0201–restricted CTL epitope of Brachyury that was successfully used for the generation of Brachyury-specific T-cell lines from both normal donors as well as cancer patients capable of lysing Brachyury-expressing tumor cells. We show that up-regulation of Brachyury expression occurs in several human tumor types and propose Brachyury as a candidate tumor antigen for immunotherapeutic approaches against cancer.

Computer-based differential display analysis. Comparison of all EST clusters on the human Unigene Built 1714

was conducted by using the HSAnalyst program as previously described (6). Unigene EST cluster Hs.389457 corresponds to accession number NM_003181.

Source of cDNA. Expression in normal tissues was studied by using multiple tissue cDNA (MTC) panels containing sets of normalized cDNAs from pooled normal tissues from several individuals (Clontech, Mountain View, CA). The following panels were used: human MTC panel I, panel II, and blood fractions panel. Commercially available tumor tissue–derived cDNAs, prepared from different individuals with different tumor types, were obtained from BioChain Institute Inc. (Hayward, CA). Total RNA from human cancer cell lines and normal CD19+ isolated B cells were prepared by using the RNeasy extraction kit (Qiagen Inc., Valencia, CA).

PCR analysis. PCR amplification of cDNA panels was carried out with the following primers specific for NM_003181: E7F 5′-GGGTGGCTTCTTCCTGGAAC-3′ and E7R 5′-TTGGAGAATTGTTCCGATGAG-3′. GAPDH-specific primers were forward 5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′, reverse 5′-CATGTGGGCCATGAGGTCCACCAC-3′. The following conditions were used: 1 min at 95°C, 35 cycles consisting of 30 s at 95°C, 30 s at 58°C, and 1 min at 72°C, and 5 min elongation at 72°C. The expected size for the Brachyury and GAPDH products was 172 and 983 bp, respectively. Total RNA derived from human cancer cell lines and normal CD19+ isolated B cells were amplified by using the TITANIUM One-Step RT-PCR kit (Clontech), following the manufacturer's instructions. Primer sequences were as follows: Brachyury, E3F 5′-ACTGGATGAAGGCTCCCGTCTCCTT-3′, and E8R 5′-CCAAGGCTGGACCAATTGTCATGGG-3′ (8); and β-actin, forward 5′-ATCTGGCACCACACCTTCTACAATGAG-3′, and reverse 5′-CGTGGTGGTGAAGCTGTAGCCGCGCTC-3′. The expected size of the PCR products was 568 and 356 bp, respectively.

Cell cultures. The human carcinoma cell lines used in this study were maintained free of Mycoplasma in RPMI 1640 (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum, 2 mmol/L glutamine, and 1× solution of antibiotic/antimycotic (Invitrogen). Additional cell lines used in this study were the C1R-A2 cell line, which is a human B-cell lymphoblastoid line transfected to express surface HLA-A2 antigen (17), and the T2 (HLA-A2+) transport-deletion mutant cell line (18).

Peptides. The computer algorithm from the Bioinformatics and Molecular Analysis Section of NIH (BIMAS) developed by Parker et al. (19) was used. A panel of 9-mer and 10-mer peptides was synthesized at >90% purity (Biosynthesis, Lewisville, TX). The carcinoembryonic antigen peptide CAP1-6D (YLSGADLNL), the HIV peptide (ILKEPVHGV), and a CEA peptide specific for HLA-A3 were used as controls.

HLA-A2 binding assay. Binding of Brachyury-specific peptides (T-p1, T-p2, T-p3, and T-p4) to HLA-A02 molecules was evaluated by flow cytometry analysis of HLA-A02 surface expression on T2 cells. T2 cells (1 × 106) in serum-free Iscove's modified Dulbecco's medium were incubated in the presence of various concentrations of each peptide in 24-well culture plates at 37°C with 5% CO2. After 18 h in culture, T2 cells were harvested, washed with 1× PBS (Invitrogen) and stained with 20 μL of a FITC-conjugated anti–HLA-A02–specific monoclonal antibody (MAb) (One Lambda, Inc., Canoga Park, CA). A FITC-conjugated immunoglobulin G2a (IgG2a) MAb (BD Biosciences, San Jose, CA) was used as an isotype control. Data acquisition and analysis were conducted on a FACSCalibur system using the CELLQuest software (BD Biosciences). Results were expressed as mean fluorescence intensity (MFI) collected on a log scale. To measure the half-life of MHC-peptide complexes, T2 cells were incubated for 18 h in the presence of 25 μmol/L of each peptide, subsequently washed free of unbound peptides, and incubated for various time points in the presence of 10 μg/mL of brefeldin A. Flow cytometry was conducted as described above. Assuming first-order kinetics, the log2 of MFI/MFI0 (MFI is the fluorescence at each time point, and MFI0 is the initial fluorescence at time 0) was plotted against time (minutes). The decay rate constant was calculated as the slope of the linear regression for each curve, and the half-life of each peptide-MHC complex was calculated as the inverse of the ratio 1/decay rate constant.

Culture of DCs from PBMCs. Peripheral blood used in this study was collected from healthy donors and cancer patients after Institutional Review Board approval, and informed consent was obtained. Peripheral blood mononuclear cells (PBMC) were isolated from leukapheresis samples by centrifugation on a Ficoll density gradient (Lymphocyte Separation Medium, ICN Biochemicals Inc., Aurora, OH). For the preparation of dendritic cells (DC), PBMCs were resuspended in AIM-V medium (Invitrogen) and allowed to adhere to the surface of T-150 flasks (Corning Costar Corp., Cambridge, MA). After 2 h at 37°C, the nonadherent cell fractions were removed, and the adherent cells were cultured in AIM-V medium containing 100 ng/mL of recombinant human granulocyte macrophage colony-stimulating factor and 20 ng/mL of recombinant human interleukin 4 (rhIL-4) for 7 days.

Generation of T-cell lines. To generate Brachyury-specific CTLs, peptide-pulsed irradiated (30 Gy) DCs were used as antigen-presenting cells with autologous, nonadherent cells used as effector cells at an effector–to–antigen-presenting cell ratio of 10:1. Cultures were maintained for 3 initial days in medium containing 10% human AB serum, and 4 additional days in the same medium supplemented with 20 units/mL of recombinant human IL-2. After a 7-day culture period, designated as an in vitro stimulation (IVS) cycle, cells were restimulated as described above.

Detection of cytokines. After three IVS cycles, CD8+ T cells that were negatively isolated by using a CD8+ isolation kit (Miltenyi Biotec, Auburn, CA) were stimulated for 24 h in the presence of peptide-pulsed autologous DCs. Culture supernatants were analyzed for the presence of IFN-γ by using an ELISA kit (Biosource International Inc., Camarillo, CA). Results were expressed in picograms per milliliter.

Cytotoxic assay. Target cells were labeled with 50 μCi of 111indium-labeled oxyquinoline (Amersham Health, Silver Spring, MD) for 15 min at room temperature. Target cells in medium containing 10% human AB serum were plated at 3 × 103 cells per well in 96-well rounded-bottom culture plates. Labeled C1R-A2 or T2 cells were incubated with peptides at the indicated concentrations for 60 min at 37°C in 5% CO2 before the addition of effector cells. No peptide was added when carcinoma cells or CD19+ B cells were used as targets. CD8+ T cells negatively isolated from T-cell cultures were used as effector cells at various effector-to-target (E:T) cell ratios. When target cells were C1R-A2 or T2, cocultures were incubated at 37°C in a 5% CO2 atmosphere for 6 h as previously described (20); when carcinoma cell lines were used as targets, cocultures were incubated as previously described (20) in the same conditions for a period of 16 h. Cytotoxic assays employing normal donor CD19+ B cells as targets were conducted for 5 h as previously described (21) due to the high levels of spontaneous release observed after a 16-h incubation period. Supernatants were harvested, and the 111In released was measured by gamma counting. Spontaneous release was determined by incubating the target cells with medium alone, and complete lysis was determined by incubating the target cells with 2.5% Triton X-100. All determinations were done in triplicate, and SDs were calculated. Specific lysis was calculated as follows: specific lysis (%) = [(observed release − spontaneous release)/(complete release − spontaneous release)] × 100.

In silico profiling of gene expression in the human Unigene Built 171 was conducted as previously described (6) by using the HSAnalyst software tool. An algorithm executed by the program returned a list of candidate EST clusters that contained >10 ESTs with >90% of the ESTs derived from tumor libraries. Among them, the cluster Hs.389457 contained the whole mRNA sequence encoding for the human Brachyury gene (mouse Brachyury homologue). From a total of 55 ESTs included in this cluster, 50 ESTs corresponded to tumor-derived libraries constructed from lung carcinoma cell lines, germ-cell tumors, chronic lymphocytic leukemia B cells, and breast cancer. Two normal tissue-derived ESTs found in the cluster Hs.389457 belonged to a library constructed from pooled RNA from fetal lung, testis, and normal B cells. The other three ESTs in the cluster were designated as “undefined” because they lacked tissue origin descriptions.

The computer-based predictions of the expression of Brachyury mRNA were then verified by reverse transcription-PCR (RT-PCR) analysis of Brachyury expression in a range of normal and malignant human tissues. Most normal tissue-derived cDNA samples, as predicted by the algorithm, showed no Brachyury mRNA expression (Fig. 1A and B). Very weak signals, however, were observed with cDNA derived from normal testis, spleen (Fig. 1A), and resting CD19+ purified cells (Fig. 1B). These results were also in accordance with the software's prediction that two out of 55 ESTs in the cluster belonged to a library prepared from pooled testis, fetal lung, and normal B lymphocytes. The expression of Brachyury in normal B cells was further evaluated in CD19+ samples isolated from various healthy donors; weak amplification was observed in four out of nine samples analyzed when using 1 μg of total RNA and 35 cycles of PCR amplification.

Fig. 1.

RT-PCR analysis of Brachyury expression in human normal and tumor tissues. Human multiple tissues cDNA panels I and II (A), various human blood fraction cDNAs (B), and cDNA from tumor tissues (each tissue from an individual cancer patient) (C) were amplified for expression of Brachyury (top) and GAPDH (bottom). Human DNA was used as a positive control for the PCR reaction; water was added to the tubes labeled as negative control.

Fig. 1.

RT-PCR analysis of Brachyury expression in human normal and tumor tissues. Human multiple tissues cDNA panels I and II (A), various human blood fraction cDNAs (B), and cDNA from tumor tissues (each tissue from an individual cancer patient) (C) were amplified for expression of Brachyury (top) and GAPDH (bottom). Human DNA was used as a positive control for the PCR reaction; water was added to the tubes labeled as negative control.

Close modal

In contrast, RT-PCR amplification of cDNA samples derived from tumor tissues showed relatively high levels of Brachyury mRNA expression in carcinomas of the esophagus, stomach, small intestine, kidney, bladder, uterus, ovary, and testis (Fig. 1C) and a weak signal in a lung carcinoma–derived sample. PCR products derived from two of the reactions were subsequently sequenced to confirm the gene and rule out the possibility of nonspecific amplification (data not shown). The expression of Brachyury was further analyzed in total RNA derived from 30 human carcinoma cell lines (Table 1). Brachyury mRNA expression was observed in most of the lung cancer–derived, colon cancer–derived, and prostate cancer–derived tumor cell lines (Table 1). These results thus validated the CDD predictions through RT-PCR and confirmed expression of Brachyury in several tumors but not in normal tissues.

Table 1.

RT-PCR expression of human Brachyury in human tumor cell lines

Tumor typeTumor cell lineBrachyury mRNA
Lung H441 ++ 
 NCI-H460 ++ 
 H226 
 NCI-H520 
 SW900 − 
Colon SW480 ++ 
 SW620 ++ 
 Colo 201 
 Colo 205 
 CaCo2 
 SW403 
 T-84 
 SW948 − 
 SW1463 − 
 HT-29 − 
 SW1116 − 
Prostate LNCAP 
 PC-3 
 DU145 
Pancreatic Capan-2 
 Paca-2 − 
 BxPC3 − 
 PANC-1 − 
 ASPC-1 − 
Breast MCF-7 − 
 MA-MB-231 − 
Ovarian SW626 ++* 
 NIH-OVCAR3 − 
 SK-OV3 − 
Osteosarcoma U2OS − 
Tumor typeTumor cell lineBrachyury mRNA
Lung H441 ++ 
 NCI-H460 ++ 
 H226 
 NCI-H520 
 SW900 − 
Colon SW480 ++ 
 SW620 ++ 
 Colo 201 
 Colo 205 
 CaCo2 
 SW403 
 T-84 
 SW948 − 
 SW1463 − 
 HT-29 − 
 SW1116 − 
Prostate LNCAP 
 PC-3 
 DU145 
Pancreatic Capan-2 
 Paca-2 − 
 BxPC3 − 
 PANC-1 − 
 ASPC-1 − 
Breast MCF-7 − 
 MA-MB-231 − 
Ovarian SW626 ++* 
 NIH-OVCAR3 − 
 SK-OV3 − 
Osteosarcoma U2OS − 

NOTE: Expression of Brachyury mRNA is shown relative to the expression of β-actin as being negative (−), positive (+), or strongly positive (++).

*

There is conflicting evidence that this cell line may be of colonic origin. See ref. 35.

The amino acid sequence of the Brachyury protein was then analyzed for HLA-A0201 peptide-binding prediction by using a computer algorithm from BIMAS. The top-ranking candidate peptides generated by the program, including three 9-mers and a 10-mer, were selected for further studies (see Table 2). In silico predicted epitopes were then assessed for binding to the MHC molecules in a cell-based assay. TAP-deficient T2 (HLA-A2+) cells were incubated in the presence of 25 μmol/L of each peptide and subsequently tested for cell surface MHC stabilization by bound peptides. Flow cytometry staining of HLA-A02 (Fig. 2A) showed that all four candidate peptides predicted by the algorithm efficiently bound to HLA-A02 molecules when compared with positive and negative control peptides. Peptides with the highest binding to T2 cells (T-p2, T-p3, and T-p4) were selected for further studies. The half-life of each peptide-MHC complex was determined; T2 cells were incubated overnight in the presence of 25 μmol/L of each peptide followed by the addition of brefeldin A and subsequent evaluation of cell surface staining of HLA-A02 at various time points. MHC-peptide complexes involving Tp-2 have a half-life of 514 min, similar to that of the positive control peptide (CAP1-6D). In contrast, MHC-peptide complexes involving T-p3 and T-p4 showed shorter half-lives of 225 and 312 min, respectively (Fig. 2B).

Table 2.

HLA-A0201 peptide motif search using BIMAS software

PeptideResiduesStart position*SequenceScore
T-p1 9-mer 345 SQYPSLWSV 389.26 
T-p2 9-mer 246 WLLPGTSTL 363.59 
T-p3 9-mer 422 RLIASWTPV 118.24 
T-p4 10-mer 86 AMYSFLLDFV 996.36 
PeptideResiduesStart position*SequenceScore
T-p1 9-mer 345 SQYPSLWSV 389.26 
T-p2 9-mer 246 WLLPGTSTL 363.59 
T-p3 9-mer 422 RLIASWTPV 118.24 
T-p4 10-mer 86 AMYSFLLDFV 996.36 
*

Start position corresponds to the amino acid position in the protein sequence.

Estimate of half-time disassociation of a molecule containing this subsequence.

Fig. 2.

Binding of predicted peptides to HLA-A0201 molecules. A, peptides at 25 μmol/L were analyzed for binding to T2 cells; a positive control (CAP1-6D) and an HLA-A3 binding peptide (negative control) were used at the same concentration. B, analysis of half-life of peptide-MHC complexes was conducted as described in Materials and Methods. For each peptide and the positive control CAP1-6D, half-life time is given in minutes.

Fig. 2.

Binding of predicted peptides to HLA-A0201 molecules. A, peptides at 25 μmol/L were analyzed for binding to T2 cells; a positive control (CAP1-6D) and an HLA-A3 binding peptide (negative control) were used at the same concentration. B, analysis of half-life of peptide-MHC complexes was conducted as described in Materials and Methods. For each peptide and the positive control CAP1-6D, half-life time is given in minutes.

Close modal

Once the ability of the predicted peptides to bind HLA-A02 molecules was shown, we next investigated the immunogenicity of peptides T-p2, T-p3, and T-p4 by evaluating their ability to induce specific CTLs in vitro. Irradiated DCs pulsed with 25 μmol/L of each peptide were used to stimulate autologous T cells from a healthy donor's PBMCs. After three IVS, isolated CD8+ T cells were subsequently stimulated for 24 h in the presence of autologous DCs alone or DCs pulsed with each of the “inducer” peptides (T-p2, T-p3, or T-p4) or an irrelevant HIV peptide. Of the three peptides tested, T-p2 and T-p3 induced antigen-specific CTLs able to release IFN-γ upon stimulation with the specific peptide (Fig. 3A). Both CTL lines were then tested for their cytotoxic activity against peptide-pulsed HLA-A02+ targets. As shown in Fig. 3B, only T cells generated with the T-p2 peptide were able to specifically lyse peptide-pulsed target cells, consistent with the peptide's ability to form stable MHC complexes compared with T-p3 and T-p4. Titration of the cytotoxic activity of the T-p2 CTLs showed cytotoxic responses at peptide concentrations as low as 1 nmol/L (Fig. 3C). Cytotoxic lysis of normal B lymphocytes was also analyzed because low expression of Brachyury was detectable in CD19+ cells isolated from various healthy donors. No lysis was observed with any of the normal B cells analyzed from five different healthy donors (data not shown).

Fig. 3.

Cytokine production and cytotoxic activity of CTLs specific for three Brachyury-derived peptides. A, CD8+ T cells generated from PBMC of a healthy donor against peptides T-p2, T-p3, and T-p4 were stimulated for 24 h in the presence of Brachyury (T)-specific peptides or irrelevant peptide-pulsed autologous DCs. IFN-γwas evaluated in the supernatants by ELISA. B, cytotoxic activity (6-h assay) of CTLs generated with peptides T-p2 and T-p3 against peptide-pulsed C1R-A2 targets. Two effector-to-target ratios (E:T) were used as indicated. C1R-A2 cells were pulsed with 25 μmol/L of T-p2 peptide (•), T-p3 peptide (○), irrelevant CAP1-6D peptide (▴), and without peptide (△). C, T2 cells were pulsed with various concentrations of T-p2 peptide as indicated and used as targets with T-p2 CTLs (at an effector-to-targets ratio equal to 12.5:1).

Fig. 3.

Cytokine production and cytotoxic activity of CTLs specific for three Brachyury-derived peptides. A, CD8+ T cells generated from PBMC of a healthy donor against peptides T-p2, T-p3, and T-p4 were stimulated for 24 h in the presence of Brachyury (T)-specific peptides or irrelevant peptide-pulsed autologous DCs. IFN-γwas evaluated in the supernatants by ELISA. B, cytotoxic activity (6-h assay) of CTLs generated with peptides T-p2 and T-p3 against peptide-pulsed C1R-A2 targets. Two effector-to-target ratios (E:T) were used as indicated. C1R-A2 cells were pulsed with 25 μmol/L of T-p2 peptide (•), T-p3 peptide (○), irrelevant CAP1-6D peptide (▴), and without peptide (△). C, T2 cells were pulsed with various concentrations of T-p2 peptide as indicated and used as targets with T-p2 CTLs (at an effector-to-targets ratio equal to 12.5:1).

Close modal

The cytolytic activity of the T-p2–specific CTLs was then tested against several tumor targets. Tumor cell lines used as targets included the lung carcinoma cells H441 (HLA-A02+/T antigen+) and NCI-H460 (HLA-A24, 68+/T antigen+), the colorectal carcinoma line SW1463 (HLA-A02+/T antigen−), and the pancreatic carcinoma cells AsPC-1 (HLA-A02−). CTLs derived with the T-p2 epitope were highly efficient at killing H441 tumor cells, whereas no lysis was observed against the other cell lines. MHC restriction was shown by the observation that the H460 tumor cell line that is highly positive for Brachyury but HLA-A02 negative was not killed by the Tp-2 CTLs (Fig. 4A). Conversely, the tumor cell line SW1463 served as an antigen-specific control because it is negative for the expression of Brachyury but positive for the expression of HLA-A02. Similarly, the control AsPC-1 (HLA-A02−) cells were also not killed by the Brachyury-specific T cells. These results indicate that T cells that have been expanded in vitro in the presence of the T-p2 peptide are able to specifically lyse those tumor cells that express Brachyury within the correct MHC class I context. As shown in Fig. 4B, T-p2 CTL-mediated killing of H441 tumor cells was blocked by antibodies directed against the MHC class I molecules but not the MHC class II molecules, further confirming the MHC class I restriction of the observed lysis. The Tp-2 peptide was then tested for in vitro expansion of Brachyury-specific T cells from PBMC of four additional healthy donors. We have been able to induce Tp-2–specific CTLs from two out of five healthy donors tested (data not shown).

Fig. 4.

Cytotoxic activity of Brachyury-specific CTLs against tumor targets. A, T-p2 CTLs from a normal donor were used as effectors against various tumor targets in an 111In 16-h release assay, as indicated. B,111In-labeled H441 tumor cells were incubated with 25 μg/mL of a control IgG, anti–HLA class I, or anti–HLA class II MAb for 1 h before the addition of T-p2 T cells. The E:T ratio was 20:1. CTLs established from the blood of (C) a colorectal cancer patient (patient 1) and (D) an ovarian cancer patient (patient 2) were used after three IVS for cytotoxic killing of H441 and AsPC-1 tumor cells. E, cytotoxic killing of LNCAP tumor cells by T-p2 T cells derived from patient 1. 111In-labeled LNCAP tumor cells were incubated with 25 μg/mL of a control IgG or an anti–HLA-A2,28 MAb for 1 h before the addition of T-p2 T cells. F, T cells derived from patient 2 were used as effectors against various tumor targets, as indicated. Also shown is the expression of Brachyury and β-actin mRNA by RT-PCR in each tumor cell line.

Fig. 4.

Cytotoxic activity of Brachyury-specific CTLs against tumor targets. A, T-p2 CTLs from a normal donor were used as effectors against various tumor targets in an 111In 16-h release assay, as indicated. B,111In-labeled H441 tumor cells were incubated with 25 μg/mL of a control IgG, anti–HLA class I, or anti–HLA class II MAb for 1 h before the addition of T-p2 T cells. The E:T ratio was 20:1. CTLs established from the blood of (C) a colorectal cancer patient (patient 1) and (D) an ovarian cancer patient (patient 2) were used after three IVS for cytotoxic killing of H441 and AsPC-1 tumor cells. E, cytotoxic killing of LNCAP tumor cells by T-p2 T cells derived from patient 1. 111In-labeled LNCAP tumor cells were incubated with 25 μg/mL of a control IgG or an anti–HLA-A2,28 MAb for 1 h before the addition of T-p2 T cells. F, T cells derived from patient 2 were used as effectors against various tumor targets, as indicated. Also shown is the expression of Brachyury and β-actin mRNA by RT-PCR in each tumor cell line.

Close modal

Generation of T-p2–specific CTLs was also successfully carried out from PBMCs of two cancer patients. T cells isolated from PBMCs of a colorectal cancer patient (designated as patient 1) and an ovarian cancer patient (designated as patient 2) were stimulated in vitro for three cycles in the presence of autologous, irradiated T-p2–pulsed DCs as described in Materials and Methods. CD8+ T cells negatively isolated from these cultures were assayed for cytotoxic activity against tumor cells. As shown in Fig. 4C and D, after three IVS, both CTL lines were able to lyse H441 tumor cells. After five IVS, CTLs derived from both patients were tested for their ability to lyse additional tumor cell lines positive for the expression of Brachyury. As shown in Fig. 4E, T-p2–specific CTLs derived from patient 1 were able to lyse LNCAP cells (HLA-A2+/Brachyury+) in an HLA-A02–restricted way, as denoted by the blocking of cytotoxic killing in the presence of anti–HLA-A02, but not in the presence of a control IgG. Figure 4F shows that T-p2 cells that expanded from the blood of patient 2 were able to lyse H441, SW620, and SW480 tumor cells, all of them being Brachyury+ and HLA-A2+. On the other hand, lysis of SW403 cells, which are HLA-A2+ and express lower levels of Brachyury mRNA (Fig. 4E), was only minimal. Altogether, Tp-2 cells derived from healthy individuals and cancer patients were able to lyse 4/5 Brachyury-positive tumors, whereas no lysis was observed for control tumor cells that were (a) HLA-A2−/Brachyury+ (NCI-H460) or (b) HLA-A2+/Brachyury− (SW1463).

In conclusion, our results showed that T-p2–specific T cells generated from both healthy donors and cancer patients were able to recognize and mediate cytotoxic lysis of tumor cells that endogenously express the Brachyury protein.

High-throughput gene expression analysis in tumors versus normal tissues constitutes a relatively new approach for the identification of therapeutic cancer targets. Computer programs have been emerging for mining of EST databases that use publicly available information from the vast collection of ESTs (22). Because the frequency of ESTs in a cDNA library seems to be proportional to the abundance of associated transcripts in the tissue from which the library was prepared (23), data on EST expression can be correlated with tissue-related or disease-related gene expression signatures. Here, we have successfully used a data mining software tool (HSAnalyst) for the identification of Unigene EST cluster Hs.389457, corresponding to the human gene Brachyury, as a tumor antigen, and validated the in silico prediction by RT-PCR in a set of normal and tumor tissues and cancer cell lines. Expression of Brachyury was elevated in tumors of the small intestine, stomach, kidney, bladder, uterus, ovary, and testis and in the majority of cell lines derived from lung, colon, and prostate carcinomas. Because of the high grade of conservation among members of the T-box family, BLAST analysis of the primer sequences was conducted to discard any possible amplification of sequences derived from other members of the T-box family, and the fidelity of the amplified band was confirmed by DNA sequencing. The high levels of expression of Brachyury in tumors contrasted with its lack of expression in most normal adult tissues, with the exception of low levels observed in testis, spleen, and CD19+ (resting) lymphocytes. The weak signal in spleen could be attributable to the presence of CD19+ cells.

The Brachyury gene was initially cloned from mouse developmental mutants characterized by an arrest in mesoderm formation (11) and was recognized as a relevant gene for the development of mesoderm during gastrulation. In fact, Brachyury is the founding member of a family of transcription factors, designated T-box transcription factors, characterized by a conserved DNA-binding domain (9), that has an essential role in the formation and organization of mesoderm in vertebrates (8, 11, 24, 25). For example, in Xenopus, Brachyury is an early-immediate response gene of mesoderm inducers, such as activin or transforming growth factor-β, and injection of Brachyury mRNA in embryos is sufficient to induce ectopic mesoderm development (26). In addition to the fundamental role of the T-box proteins in the control of developmental processes, several members of this family seem to be deregulated in cancer. The human Tbx2 gene has been reported to be amplified in pancreatic cancer cell lines (27) and overexpressed in BRCA-1– and BRCA-2–mutated breast tumors (28). Likewise, Tbx3 expression has been shown to be augmented in certain human breast cancer cell lines (29). Brachyury expression has been previously reported in human teratocarcinoma lines: a subset of germ cell tumors, teratocarcinomas are embryonal carcinoma cells with competence for mesoderm differentiation (30) as well as in chordomas (31, 32).

As a hallmark of mesoderm development, Brachyury is a particularly attractive target for cancer intervention. Formation of mesoderm during embryogenesis is a manifestation of epithelial plasticity where epithelial cells change into mesenchymal cells by triggering an EMT process (2). Recently, EMT has been recognized as a central process in the progression of primary tumors into metastasis, which involves the acquisition of motility and invasiveness by the tumor cells (3, 4). Several transcription factors involved in the control of mesoderm have been identified that are able to control or trigger EMT, such as the basic helix-loop-helix protein Twist and the zinc finger proteins Snail and Slug (4, 5, 33). The role of Brachyury in the EMT process during tumor progression has not yet been shown, and future studies are necessary to elucidate its potential role in the metastatic progression of tumors. The high expression of Brachyury in certain tumor tissues and cancer cell lines, together with its known involvement in the EMT process during mesodermal development in the embryo and in in vitro EMT processes, have all contributed to our interest in studying Brachyury as a potential target for immune interventions in cancer.

As previously shown, algorithms available on the web can be successfully applied to the selection of short peptide sequences with high-affinity binding to MHC class I molecules (34). The affinity prediction method from BIMAS was applied here in searching Brachyury peptides with high-affinity binding for HLA-A0201. All four top-ranked peptides effectively bound to HLA-A02 molecules, although peptide-MHC complexes showed differences in their decay rate. Tp-2 was the only peptide, however, able to expand CTLs in vitro that are capable of releasing IFN-γ in response to peptide-specific stimulation and lysing peptide-pulsed targets with high efficiency. This peptide also showed the maximum stability of binding to HLA-A02, a possible explanation of its increased immunogenicity. Because the ability of CTLs to lyse peptide-pulsed target cells does not necessarily correlate with their ability to mediate cytotoxic killing of tumor cells, it was important to confirm the capacity of Tp-2–specific T cells for specific tumor recognition. The lung carcinoma cell line H441 was effectively lysed in the presence of Brachyury-specific CTLs even at a low ratio of effector T cells to targets in an antigen-specific and MHC-restricted manner. Furthermore, we were also able to show that Brachyury–T-p2–specific CTLs can be expanded in vitro from PBMCs of a colorectal cancer patient and an ovarian carcinoma patient, satisfying a critical prerequisite for the use of Brachyury as a therapeutic target for cancer vaccine regimens. We believe this is the first study to show that (a) a T-box transcription factor and (b) a molecule implicated in mesodermal development, i.e., EMT, can be a potential target for human T-cell–mediated cancer immunotherapy.

In conclusion, our results show that up-regulation of Brachyury occurs in certain tumor tissues and cancer cell lines, and that Brachyury-specific CTLs can be generated from the blood of cancer patients and normal donors, which, in turn, can effectively lyse Brachyury-expressing tumor cells.

Grant support: NIH Intramural Research Program, National Cancer Institute, Center for Cancer Research, and The Biomedical Center, St. Petersburg, Russia.

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.

We thank Diane Poole, Margie Duberstein, and Neema Soares of the Laboratory of Tumor Immunology and Biology, and Julia Nosova, of The Biomedical Center, for their technical assistance. We also thank Debra Weingarten for her editorial assistance in the preparation of the manuscript.

1
Thiery JP. Epithelial-mesenchymal transitions in tumour progression.
Nat Rev Cancer
2002
;
2
:
442
–54.
2
Thiery JP. Epithelial-mesenchymal transitions in development and pathologies.
Curr Opin Cell Biol
2003
;
15
:
740
–6.
3
Huber MA, Kraut N, Beug H. Molecular requirements for epithelial-mesenchymal transition during tumor progression.
Curr Opin Cell Biol
2005
;
17
:
548
–58.
4
Yang J, Mani SA, Donaher JL, et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis.
Cell
2004
;
117
:
927
–39.
5
Cano A, Perez-Moreno MA, Rodrigo I, et al. The transcription factor snail controls epithelial-mesenchymal transitions by repressing E-cadherin expression.
Nat Cell Biol
2000
;
2
:
76
–83.
6
Baranova AV, Lobashev AV, Ivanov DV, Krukovskaya LL, Yankovsky NK, Kozlov AP. In silico screening for tumour-specific expressed sequences in human genome.
FEBS Lett
2001
;
508
:
143
–8.
7
Krukovskaja LL, Baranova A, Tyezelova T, Polev D, Kozlov AP. Experimental study of human expressed sequences newly identified in silico as tumor specific.
Tumor Biol
2005
;
26
:
17
–24.
8
Edwards YH, Putt W, Lekoape KM, et al. The human homolog T of the mouse T(Brachyury) gene; gene structure, cDNA sequence, and assignment to chromosome 6q27.
Genome Res
1996
;
6
:
226
–33.
9
Papaioannou VE, Silver LM. The T-box gene family.
Bioessays
1998
;
20
:
9
–19.
10
Kispert A, Herrmann BG, Leptin M, Reuter R. Homologs of the mouse Brachyury gene are involved in the specification of posterior terminal structures in Drosophila, Tribolium, and Locusta.
Genes Dev
1994
;
8
:
2137
–50.
11
Herrmann BG, Labeit S, Poustka A, King TR, Lehrach H. Cloning of the T gene required in mesoderm formation in the mouse.
Nature
1990
;
343
:
617
–22.
12
Kispert A, Herrmann BG. Immunohistochemical analysis of the Brachyury protein in wild-type and mutant mouse embryos.
Dev Biol
1994
;
161
:
179
–93.
13
Showell C, Binder O, Conlon FL. T-box genes in early embryogenesis.
Dev Dyn
2004
;
229
:
201
–18.
14
Technau U, Scholz CB. Origin and evolution of endoderm and mesoderm.
Int J Dev Biol
2003
;
47
:
531
–9.
15
Behr R, Heneweer C, Viebahn C, Denker HW, Thie M. Epithelial-mesenchymal transition in colonies of rhesus monkey embryonic stem cells: a model for processes involved in gastrulation.
Stem Cells
2005
;
23
:
805
–16.
16
Vidricaire G, Jardine K, McBurney MW. Expression of the Brachyury gene during mesoderm development in differentiating embryonal carcinoma cell cultures.
Development
1994
;
120
:
115
–22.
17
Shimojo N, Maloy WL, Anderson RW, Biddison WE, Coligan JE. Specificity of peptide binding by the HLA-A2.1 molecule.
J Immunol
1989
;
143
:
2939
–47.
18
Salter RD, Cresswell P. Impaired assembly and transport of HLA-A and -B antigens in a mutant TxB cell hybrid.
EMBO J
1986
;
5
:
943
–9.
19
Parker KC, Bednarek MA, Coligan JE. Scheme for ranking potential HLA-A2 binding peptides based on independent binding of individual peptide side-chains.
J Immunol
1994
;
152
:
163
–75.
20
Tsang KY, Palena C, Yokokawa J, et al. Analyses of recombinant vaccinia and fowlpox vaccine vectors expressing transgenes for two human tumor antigens and three human costimulatory molecules.
Clin Cancer Res
2005
;
11
:
1597
–607.
21
Palena C, Foon KA, Panicali D, et al. Potential approach to immunotherapy of chronic lymphocytic leukemia (CLL): enhanced immunogenicity of CLL cells via infection with vectors encoding for multiple costimulatory molecules.
Blood
2005
;
106
:
3515
–23.
22
Scheurle D, De Young MP, Binninger DM, Page H, Jahanzeb M, Narayanan R. Cancer gene discovery using digital differential display.
Cancer Res
2000
;
60
:
4037
–43.
23
Audic S, Claverie JM. The significance of digital gene expression profiles.
Genome Res
1997
;
7
:
986
–95.
24
Kispert A, Herrmann BG. The Brachyury gene encodes a novel DNA binding protein.
EMBO J
1993
;
12
:
3211
–20.
25
Wilkinson DG, Bhatt S, Herrmann BG. Expression pattern of the mouse T gene and its role in mesoderm formation.
Nature
1990
;
343
:
657
–9.
26
Smith JC, Price BM, Green JB, Weigel D, Herrmann BG. Expression of a Xenopus homolog of Brachyury (T) is an immediate-early response to mesoderm induction.
Cell
1991
;
67
:
79
–87.
27
Mahlamaki EH, Barlund M, Tanner M, et al. Frequent amplification of 8q24, 11q, 17q, and 20q-specific genes in pancreatic cancer.
Genes Chromosomes Cancer
2002
;
35
:
353
–8.
28
Sinclair CS, Adem C, Naderi A, et al. TBX2 is preferentially amplified in BRCA1- and BRCA2-related breast tumors.
Cancer Res
2002
;
62
:
3587
–91.
29
Fan W, Huang X, Chen C, Gray J, Huang T. TBX3 and its isoform TBX3+2a are functionally distinctive in inhibition of senescence and are overexpressed in a subset of breast cancer cell lines.
Cancer Res
2004
;
64
:
5132
–9.
30
Gokhale PJ, Giesberts AM, Andrews PW. Brachyury is expressed by human teratocarcinoma cells in the absence of mesodermal differentiation.
Cell Growth Differ
2000
;
11
:
157
–62.
31
Vujovic S, Henderson S, Presneau N, et al. Brachyury, a crucial regulator of notochordal development, is a novel biomarker for chordomas.
J Pathol
2006
;
209
:
157
–65.
32
Romeo S, Hogendoorn PC. Brachyury and chordoma: the chondroid-chordoid dilemma resolved?
J Pathol
2006
;
209
:
143
–6.
33
Kang Y, Massague J. Epithelial-mesenchymal transitions: twist in development and metastasis.
Cell
2004
;
118
:
277
–9.
34
Lu J, Celis E. Use of two predictive algorithms of the World Wide Web for the identification of tumor-reactive T-cell epitopes.
Cancer Res
2000
;
60
:
5223
–7.
35
Furlong MT, Hough CD, Sherman-Baust CA, Pizer ES, Morin PJ. Evidence for the colonic origin of ovarian cancer cell line SW626.
J Natl Cancer Inst
1999
;
91
:
1327
–8.