Six-transmembrane epithelial antigen of the prostate and enhancer of zeste homolog 2 as immunotherapeutic targets for lung cancer

Mouse fibroblast cell lines (L-cells) transfected and expressing individual human HLA-DR molecules were kindly provided by Dr. Robert W. Karr (Karr Pharma, Saint. Louis, MO, USA) and by Dr. Takehiko Sasazuki (Kyushu University, Fukuoka, Japan). The LC cell lines PC14, A549, LC-2/ad, LCAM1 (adenocarcinomas), LK2, RERF-LC-AI, EBC1 (squamous cell carcinomas), LU65, and LU99 (large cell carcinomas) were supplied by the RIKEN Bio-Resource Center (Tsukuba, Japan). The LC cell lines LHK2 and 1-87 (adenocarcinomas) were kindly provided by Dr. Yasuaki Tamura (Sapporo Medical University, Sapporo, Japan). Tumor cell lines H23, H441 (lung adenocarcinomas), H520, SK-MES-1, Calu-1 (lung squamous cell carcinomas), PC3 (prostate cancer), MCF7 (breast cancer), WiDr (colon carcinoma) and Jurkat (T cell lymphoma) were purchased from ATCC (Manassas, VA, USA). All cell lines were maintained in tissue culture as recommended by supplier.

An indirect immunoperoxidase technique (the streptavidin-biotin method) was performed. To detect STEAP, polyclonal rabbit anti-human STEAP (ZMD.265, Zymed Laboratories, Inc., South San Francisco, CA, USA), diluted 1:200, served as the primary antibody. To detect EZH2, monoclonal mouse anti-human EZH2 (BD.612666, BD Bioscience, San Jose, CA, USA), diluted 1:200, served as the primary antibody. To detect HLA-DR, monoclonal mouse anti-human HLA-DR α chain (TAL.1B5, DakoCytomation, Denmark), diluted 1:10000, served as the primary antibody. We assessed STEAP, EZH2 and HLA-DR expression in surgically resected LC specimens.

One million cells were washed in phosphate buffered saline (PBS) and lysed in NuPAGE LDS sample buffer (Invitrogen). The cell lysate was subjected to electrophoresis in a 4-12% NuPAGE bis-Tris SDS-PAGE gel (Invitrogen) under reducing condition and transferred to Immobilon-P (Millipore, Bedford, MA, USA) membrane. The membrane was blocked in PBS containing 0.01% Tween 20 and 5% non-fat dry milk for 1 h at room temperature and incubated with rabbit anti-human STEAP polyclonal antibody (H-105, Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted 1:200 in blocking buffer, anti-human EZH2 mouse monoclonal antibody (mAb) (BD.612666, BD Bioscience, San Jose, CA, USA) diluted 1:200, or anti-β-actin mAb (C4, Santa Cruz Biotechnology, Santa Cruz, CA, USA) diluted 1:1,000 in blocking buffer as the control, for 18 h at 4°C. After washing, the membrane was incubated with horseradish peroxidase-labeled sheep anti-rabbit or -mouse IgG and subjected to the enhanced chemiluminescence assay using the ECL detection system (Amersham, Buckinghamshire, UK).

Potential HLA-DR-restricted CD4 T-cell epitopes were selected from the amino acid sequence of STEAP and EZH2 using the computer-based algorithms peptide/MHC binding for three common HLA-DR alleles (DRB1*0101, DRB1*0401, DRB1*0701) developed by Southwood et al. [14]. The predicted peptide epitopes were synthesized by solid phase organic chemistry and purified by high-performance liquid chromatography (HPLC). The purity (> 80%) and identity of peptides were assessed by HPLC and mass spectrometry, respectively. The synthetic peptides used throughout this study were STEAP_261-275 (SLLLGTIHALIFAWN), STEAP_281-296 (KQFVWYTPPTFMIAVF), EZH2_95-109 (VIPLKTLNAVASVPI), EZH2_220-234 (SDKIFEAISSMFPDK), EZH2_693-704 (PNCYAKVMMVNGDHR). These peptides were selected on the basis of having top 10 scores for at least 2 of the three HLA-DR alleles. The tetanus toxoid (TT) _830-843 (QYIKANSKFIGITE) peptide was used as a control universal epitope peptide, since it is presented by multiple HLA-DR alleles [15].

The procedure utilized for the generation of STEAP and EZH2 reactive CD4 T-cell clones using peptide-stimulated lymphocytes from peripheral blood mononuclear cells (PBMCs) of human healthy individuals has been described in detail [16]. Briefly, dendritic cells (DCs) were produced from purified CD14 monocytes (using antibody-coated magnetic microbeads from Miltenyi Biotech, Auburn CA, USA) cultured for 7 days at 37°C in a humidified CO₂ (5%) incubator in the presence of 50 ng/ml GM-CSF and 1,000 IU/ml IL-4. Peptide-pulsed DCs (3 μg/ml for 2 h at room temperature) were irradiated (4,200 rad) and co-cultured with autologous purified CD4 T-cells in 96-flat-bottomed-well culture plates. One week after peptide stimulation, the CD4 T-cells were restimulated in individual microcultures with peptide-pulsed irradiated autologous PBMCs and 2 days later, recombinant human IL-2 was added at a final concentration of 10 IU/ml. One week later, the T-cells were tested for antigen reactivity using a cytokine-release assay as described below. Those microcultures exhibiting a significant response of cytokine-release to peptide (at least 2.5-fold over background) were cloned by limiting dilution and expanded in 24-well plates by weekly restimulation with peptides and irradiated autologous PBMCs. Complete culture medium for all procedures consisted of AIM-V medium (Invitrogen/GIBCO, Carlsbad, CA, USA) supplemented with 3% human male AB serum. All blood samples were obtained after the appropriate informed consent.

CD4 T-cells (3 × 10⁴/well) were mixed with irradiated antigen-presenting cells (APCs) in the presence of various concentrations of antigen (peptides, tumor lysates) in 96-well culture plates. APCs consisted of autologous PBMCs (1 × 10⁵/well), HLA-DR expressing L-cells (3 × 10⁴ well), autologous DCs (5 × 10³/well) or LC cell lines (3 × 10⁴/well). The LC cell lines were previously treated with IFN-γ at 500 U/ml for 48 h to enhance HLA-DR expression. The expression of HLA-DR molecules on tumor cells was evaluated by flow cytometry using anti-HLA-DR mAb conjugated with fluorescein isothiocyanate (BD Pharmingen, San Diego, CA). Tumor cell lysates were prepared by 3 freeze-thaw cycles of 1 × 10⁸ tumor cells, resuspened in 1 ml of serum-free AIM-V medium. Tumor cell lysates served as an antigen at 5 × 10⁵ cell equivalents/ml. Culture supernatants were collected after 48 h for measuring antigen-induced lymphokine (IFN-γ or GM-CSF) or granzyme B production by the CD4 T-cells using ELISA kits (IFN-γ and GM-CSF; BD Pharmingen, San Diego, CA, USA, granzyme B; Mabtech AB, Nacka Strand, Sweden). To demonstrate antigen-specificity and HLA-DR restriction, blocking of antigen-induced responses were assessed by adding anti-HLA-DR mAb L243 (IgG2a prepared from supernatants of the hybridoma HB-55 obtained from the ATCC) or the anti-HLA-A/B/C mAb W6/32 (IgG2a, ATCC) at 10 μg/ml throughout the 48 h incubation period.

Cytotoxic activity of CD4 T-cell clones was measured using a colorimetric CytoTox 96 assay (Promega, Madison, WI, USA). This system quantifies the release of lactate dehydrogenase (LDH) from target cells. T-cells were mixed with 2 × 10⁴ target cells at different effector to target (E:T) ratios in 96-round-bottomed-well plates. After 6 h incubation at 37°C, 50 μl supernatant samples were collected from each well to measure LDH content.

PBMCs were isolated from fresh heparinized blood by Ficoll-Hypaque centrifugation and washed twice with RPMI 1640. PBMCs were resuspended in complete medium and placed at 1.5 × 10⁵ cells/well and cultured in triplicate in 96-round-bottomed well plates in the presence of 10 μg/ml peptide. Negative controls (in the absence of peptide), were done in eight replicate samples. One week later, the cultures were restimulated with peptide-pulsed (10 μg/ml) irradiated autologous PBMCs (5 × 10⁴/well). After 7 days of restimulation, supernatants were harvested for examining the ability of peptide-induced lymphokine production by the LC patient's PBMCs. The institutional ethics committee had approved the study protocol and the appropriate informed consent for blood donation was obtained from all patients before blood sampling.

In immunohistochemistry studies with several surgical specimens of LC, we observed moderate to strong staining with anti-STEAP antibody and anti-EZH2 mAbs. Conversely, staining of adjacent normal tissues with these antibodies was much weaker. STEAP was detected in 8/8 cases of the adenocarcinoma and 6/6 squamous cell carcinoma cases. EZH2 was detected in 5/5 cases of the adenocarcinoma and in 5/5 squamous cell carcinoma cases (examples are shown in Figure 1). In addition, we performed immunohistochemical staining for expression of HLA-DR in 5 lung adenocarcinomas and 4 lung squamous cell carcinomas. We have found the expression of HLA-DR in the adenocarcinoma cells (4/5 cases) and in the infiltrating APCs, lymphocytes and alveolar macrophages (Figure 2). These observations suggest that tumor cells could be logical targets of STEAP- or EZH2-specific CD4 T-cells in situ.

Western blot analysis revealed that STEAP was present in several tumor cell lines including LCs (except for Jurkat T cell lymphoma, which served as a negative control), but was absent in PBMC samples from healthy individuals (Figure 3A). EZH2 was detected in all the tumor cell lines (including Jurkat), but not in PBMCs (Figure 3B). These observations provide a rationale for exploring the use of STEAP and EZH2 as TAAs for developing anti-tumor T-cell based immunotherapy for LC (and the other tumor types).

To identify the promiscuous (broadly degenerate) HLA-class II binding peptide epitopes is very important since using the degenerate epitope-based vaccines in applicable to general population. For predicting promiscuous HLA-class II binding peptides, we used computer-based MHC peptide binding motifs for three common HLA-class II molecules, HLA-DR1, DR4 and DR7 [14]. In addition, this algorithm indicate that some peptides that score high for the DR1, DR4 and DR7 alleles also have the capacity to bind to additional HLA-class II alleles such as DR9, DR13, DR15 and DR53. Therefore, we have reported that some peptides predicted by additional HLA-class II alleles, such as DR9, DR13, DR15 or DR53 [17]. In this study, the prediction system could select nine peptide sequences from EZH2 (data not shown) and we selected the three highest potentially promiscuous peptides, EZH2_95-109, EZH2_220-234 and EZH2_693-704. In the case of STEAP, the algorithm system suggested 28 peptide sequences (data not shown) as potentially promiscuous HLA-class II binders and we have reported that two STEAP peptides, STEAP_102-116 and STEAP_192-206 were effective in stimulating in vitro anti-tumor T helper responses [18, 19]. In the present study, considering that an ideal TAA should contain peptide epitopes to stimulate both CD8 and CD4 T-cells, we selected 2 STEAP peptides from these promiscuous binders, STEAP_261-275, which overlaps with HLA-A2-restricted CD8 T-cell epitope STEAP_262-270 [20] and STEAP_281-296, which lies proximal to HLA-A2-restricted CD8 epitope, STEAP_292-300 [21]. We synthesized the three EZH2 and two STEAP epitope peptides and proceeded to determine whether these peptides could induce CD4 helper T-cell responses in vitro.

Next, we evaluated whether these peptides were capable of stimulating CD4 T-cells obtained from 6 healthy individuals (HLA-DR4/15, HLA-DR4/9, HLA-DR4/15, HLA-DR1/15, HLA-DR9/13 and HLA-DR9/14) using autologous DCs as APCs. After 2-3 peptide restimulation cycles, 2 of 5 peptides (STEAP_281-296 and EZH2_95-109) were able to induce peptide-specific CD4 T-cell responses. To carry out more detailed antigen specificity and HLA-DR restriction analyses, and to assess tumor cell recognition, T-cell clones (4 reactive with STEAP_281-296 and 6 reactive with EZH2_95-109) were isolated by limiting dilution. As shown in Figure 4, all of these clones recognized their respective peptide epitopes in a dose dependent manner and this recognition was inhibited by anti-HLA-DR (L243) mAb but not by anti-HLA-class I (W6/32) mAb. Since mAb L243 is specific for HLA-DR, these results indicate that the presentation of STEAP_281-296 and EZH2_95-109 to the T-cells is via HLA-DR molecules and not via other HLA-class II molecules (HLA-DQ or DP). To more specifically delineate the HLA-DR restriction alleles, we studied the T-cells responses using a panel of HLA-DR transfected mouse L-cells and HLA-typed human EBV-LCLs. As shown in Figure 5A, the STEAP_281-296-reactive clone SH31 was restricted by HLA-DR15 molecules, and clones SH6, HK7 and Sa12 recognized the same peptide but in the context of HLA-DR53. The EZH2_95-109 epitope was presented to clone SK31 via HLA-DR1, to clones SK32 and Sa361 by HLA-DR15 and to clones MT10, Sa362 and YM18 by HLA-DR53 (Figure 5B, C and 5D). These results indicate that STEAP_281-296 and EZH2_95-109 can be presented to CD4 T-cells by more than one HLA-DR allele, behaving as classical promiscuous epitopes.

We then evaluated whether STEAP or EZH2 reactive CD4 T-cell clones could directly recognize STEAP or EZH2 expressing LC cells. Before this study, we first measured cell surface HLA-DR expression in several tumor cell lines that would be used in these studies. The LC cell lines (LU65, LC-2/ad, RERF-LC-AI, EBC1, Calu-1) displayed measurable levels of surface HLA-DR after IFN-γ treatment. However, there was no HLA-DR expression in A549 lung adenocarcinoma and Jurkat T-cell lymphoma (data not shown). The peptide-reactive CD4 T-cell clones were then tested for reactivity (IFN-γ production) against the HLA-DR expressing, STEAP/EZH2 positive LC cell lines. As shown in Figure 6A, CD4 T-cell clone SH31 (STEAP_281-296-reactive, DR15-restricted) recognized two STEAP+/DR15+ LC tumors (LU65, RERF-LC-AI). Similarly, the HLA-DR53-restricted, STEAP_281-296-reactive CD4 T-cell clones, SH6, HK7 and Sa12 recognized STEAP+/DR53+ LC tumors (LU65, Calu-1, Figure 6B, C and 6D). In addition, the STEAP_281-296-reactive CD4 T-cell clones SH31 and Sa12 produced granzyme B (serine protease cytotoxic molecule) when stimulated with LC tumors in an HLA-DR dependent manner (Figure 7A). In the case of EZH2-reactive CD4 T-cell responses, clone SK31 (EZH2_95-109-reactive, DR1-restricted) recognized several EZH2+/DR1+ LC tumors (LU65, EBC1 and LC-2/ad, Figure 8A). CD4 T-cell clones SK32 and Sa361 (EZH2_95-109-reactive, DR15-restricted) recognized EZH2+/DR15+ LC tumors (LU65, RERF-LC-AI, Figure 8B and 8C). Lastly, as shown in Figure 8D and 8E, CD4 T-cell clones Sa362 and YM18 (EZH2_95-109-reactive, DR53-restricted) were able to react with EZH2+/DR53+ LC tumors (LU65, Calu-1). Furthermore, the EZH2_95-109-reactive CD4 T-cell clones produced granzyme B when stimulate with LC cells (Figure 7B). Direct tumor recognition in all cases was antigen-specific and HLA-DR-restricted since tumor cell lines not expressing the appropriate antigen (STEAP, EZH2), and the corresponding HLA-DR molecule failed to stimulate the CD4 T-cell responses. Furthermore, the capacity of the STEAP- and EZH2-expressing LC cells to activate the CD4 T-cell clones was significantly inhibited by the addition of anti-HLA-DR L243 mAb confirming that the peptide epitopes were presented via HLA-DR molecules.

In addition, we assessed whether APCs such as DCs could capture and process antigens derived from dead STEAP expressing tumor cells (freeze-thaw cell lysates) and present the STEAP epitope to the CD4 T-cell clones. As shown in Figure 9, the STEAP_281-296-reactive CD4 T-cell clones SH31, SH6 and Sa12 recognized naturally processed epitope presented by DCs derived from STEAP positive LC cell lysates (LU65, A549, RERF-LC-AI) but not STEAP negative, Jurkat cell lysates. These responses were inhibited by L243 mAb (anti-HLA-DR) but not by W6/32 mAb (anti-HLA-class I), indicating that CD4 T-cell responses were antigen-specific and via the interaction of T-cell receptor with HLA-DR molecules. These results indicate that the STEAP_281-296 epitope can be presented to T-cells via either the endogenous (direct tumor recognition) or exogenous (DCs) antigen processing pathways.

Conversely, the EZH2-reactive CD4 T-cell clones were unable to recognize naturally processed epitopes derived from tumor cell lysates via DCs (data not shown).

Having observed that peptide STEAP_281-296 and EZH2_95-109 elicited CD4 T-cell clones produced granzyme B when stimulated with STEAP or EZH2 expressing LC cells (Figure 7), we evaluated the cytotoxic activity of the CD4 T-cell clones. The STEAP_281-296-reactive, DR15-restricted clones SH31 and Sa12 effectively lysed STEAP expressing LC cell lines in a dose dependent manner (Figure 10A). Similarly, the EZH2_95-109-reactive clones SK32 and Sa361 and Sa362 were efficient in killing the EZH2 expressing LC cell lines (Figure 10B).

Lastly, to further evaluate the capacity of LC patients to exhibit CD4 T-cell responses to STEAP_281-296 and EZH2_95-109, we stimulated PBMCs from 10 LC patients and 3 healthy donors with the corresponding peptides in short-term cultures. Seven days after the second stimulation, culture supernatants were collected and IFN-γ or GM-CSF production was assessed. Peptide TT_830-843 was used as a positive control since this antigen has ability to induce strong CD4 T-cell responses in the majority individuals regardless of their HLA-DR alleles. The results presented in Table 1 show significant T-cell responses to STEAP_281-296 in 9/10 patients, to EZH2_95-109 in all patients, and to the positive control TT_830-843 in all patients. These results suggest that T-cell precursors that are reactive with the STEAP and EZH2 epitopes exist in the peripheral blood of patients with LC.