Expression of tumor rejection antigens in colorectal carcinomas

Authors


Abstract

BACKGROUND

The authors recently reported that the SART2 and SART3 antigens encode tumor epitopes recognized by HLA-A24-restricted and tumor-specific cytotoxic T lymphocytes (CTLs) established from esophageal carcinoma patients. The current study investigated these antigens to explore a potential molecule for specific immunotherapy for colorectal carcinoma patients.

METHODS

The SART2 and SART3 antigens were investigated by Western blotting in colorectal carcinoma cell lines and in cancer tissues. For induction of CTLs, peripheral blood mononuclear cells (PBMCs) of HLA A-24-positive cancer patients were stimulated in vitro with peptides.

RESULTS

The 140 kD SART3 antigen was expressed in both the cytosol and nuclear fractions of all six colon carcinoma cell lines, 27 of 41 (65.9%) cytosol fractions, 30 of 41 (73.2%) nuclear fractions of colorectal carcinoma tissue samples, and in 0 of 7 non-tumorous tissues. The 100 kD SART2 antigen was expressed in the cytosol fractions of 2 of 6 colon carcinoma cell lines, 5 of 20 (25%) cytosol fractions of colorectal carcinoma tissue samples, and in 0 of 7 non tumorous tissues. HLA-A24-restricted CTLs cytotoxic to colon carcinoma cells were induced from PBMCs of colon carcinoma patients by stimulation with the two immunogenic peptides of SART3.

CONCLUSIONS

The SART3 antigen could be an appropriate target molecule for specific immunotherapy for colorectal carcinoma patients. Cancer 2002;94:1636–41. © 2002 American Cancer Society.

DOI 10.1002/cncr.10421

Colorectal carcinoma is a leading cause of morbidity and mortality, with about 300,000 new cases and 200,000 deaths in Europe, the U.S. and Japan each year.1 The prognosis of colorectal carcinoma at advanced stages is extremely low regardless of recent clinical trials with many chemotherapeutic agents. Attempts at nonspecific immunotherapy using agents such as BCG (bacillus Calmette–Guerin) or levamisole have proven ineffective in preventing the recurrence of colorectal carcinoma.2, 3 Therefore, new treatment modalities are clearly needed, and peptide-based immunotherapy is one of the modalities currently under consideration.4, 5 Many genes encoding tumor-rejection antigens recognized by cytotoxic T lymphocytes (CTLs) have been identified from the cDNA of melanomas.6–9 Some peptides encoded by these genes are under clinical trials as cancer vaccines, and major tumor regressions have been seen in some of these melanoma patients.4, 5, 10 Although infiltration of CD8+ T cells in colorectal carcinoma tissue has been reported,11, 12 little is known about the molecular basis of the host defense against cancer cells in colorectal carcinoma patients. Subsequently, little information is available on tumor antigens for specific colorectal carcinoma immunotherapy.13–15 We recently identified the SART2 and SART3 genes, which encode tumor antigens recognized by the HLA-A26 and HLA-A24 restricted CTLs.16, 17 The SART3 antigen was expressed in the nucleus of all the malignant tumor cell lines tested in the current study and the majority of cancer tissues with different histologic types. It was also expressed in the cytosols of all the proliferating cells, including normal T cells and malignant cells. However, it was not expressed in normal tissues, except for testis and fetal liver. The SART3 antigen possesses two epitopes that are able to induce HLA-A 24 restricted CTLs from the majority of cancer patients with various histologic types.17 We investigated in the current study the expression of these two antigens in colorectal carcinomas and report that the SART3 antigen could be an appropriate vaccine candidate for specific immunotherapy for colorectal carcinoma patients.

MATERIALS AND METHODS

Cell Lines and Tissue Samples

Primary colorectal carcinoma tissue samples (n = 41) and nontumorous colorectal tissue samples (n = 7) were obtained at the Kurume University Hospital at the time of surgery. A portion of these tissue samples was minced and frozen at −80 °C until use. The mean age of the 41 patients was 74 years, with a range of 45–86 years. The male to female ratio was 26:15. Among the 41 tumors, 22 were rectal carcinomas and 19 were colon carcinomas (2 caecal, 8 ascending, 3 transversal, 3 descending, and 3 sigmoid colon carcinomas). Histologically, all 41 tumors were adenocarcinomas. Six colorectal carcinoma cell lines (Colo201, Colo205, Colo320, HCT116, KM12LM, and SW620) were provided by Japan collection of research bioresources (Tokyo, Japan) or purchased from American Type Culture Collection (Rockville, MD). The HLA-A24 restricted and tumor-specific CTL lines (CD3+CD4, CD8+) were established from peripheral blood mononuclear cells (PBMCs) of an esophageal carcinoma patient (KE-4, HLA-A2402/2601) after stimulation with autologous KE-4 tumor cells by standard mixed lymphocyte-tumor cell culture; its characteristics have been described elsewhere.18, 19 The KE4 esophageal carcinoma cell line (HLA-A2402/2601) from which the SART2 and SART3 genes were cloned was used as a positive control as target cells, while VA13 fibroblast cells were used as a negative control. Tumor cells were cultured with the RPMI1640 medium, MEM medium, or Dulbeccos modified eagle medium supplemented with 10% fetal calf serum (FCS). The CIR cells expressing HLA-A*2402 (CIR-A*2402 [asterisk denotes CIR cell line that transfected by the HLA-A2402 gene]) for peptide loading were provided by Dr. Masahumi Takiguchi, M.D., Ph.D. (Kumamoto University, Kumamoto, Japan) and maintained in RPMI 1640 medium containing 10% FCS and 0.5 mg/mL of hygromycin B (Wako, Tokyo, Japan).

Detection of the SART2 and SART3 Antigens

Tissues were sonicated for 60 to 90 seconds by an Astron ultrasonic processor (Heat Systems, Farmingdale, NY). The samples were lysed with a buffer containing 10 mM Tris-HCI (pH 7.4), 150 mM NaCl, 0.5% Triton X-100, 0.2 mM phenylmethylsulfonyl fluoride (Sigma Chemical Co., St. Louis, MO), and 0.03 1 trypsin inhibitor unit/mL aprotinin, sonicated and centrifuged and collected the supernatant to use as the cytosol fraction. The precipitation was then lysed with buffer consisting of 7.2 nM urea, 1.6% Triton X-100, 0.8% dithiothreitol, and 2% lithium dodecyl sulfate. Following that lysing, it was centrifuged, and the supernatant was used as the nuclear fraction. Expression of the SART2 and SART3 antigen in the samples was investigated by Western blot analysis with the polyclonal anti-SART2 antibody and anti-SART3 antibodies respectively.16–19

Expression of HLA-A24 Molecule

Expression of HLA class I or HLA-A24 antigens on colorectal carcinoma cell lines was studied by staining the cells with anti-class I (w6/32) monoclonal antibody (mAb) or anti-HLA-A24 mAb (one Lambda, Inc., Canoga Park, CA). It was measured by fluorescence-activated cell sorting scan (FACScan, Becton Dickinson, San Jose, CA).19

Induction of CTLs by the Peptides

The SART3109–118 (VYDYNCHVDL) and SART3315–323 (AYIDFEMKI) peptides capable of inducing HLA-A24 restricted CTLs were synthesized. HLA-A2402-binding human immunodeficiency virus peptide (RYLRDQQLLGI) was also prepared as a negative control. These peptides were kindly provided by Dr. Masahara Kanaoka, M.D., Ph.D. (Sumitomo, Osaka, Japan), and they were over 95% pure. HLA-A24+ PBMCs of the colorectal carcinoma patients were provided for the current study, and their histology was well-differentiated adenocarcinoma. A detailed method for CTL induction by a peptide has been described elsewhere.16–19 Briefly, PBMCs from colorectal carcinoma patients were incubated with 10 μM of peptide in one well of a 24-well plate containing 2 mL of culture medium (45% RPMI1640 medium, 45% AIM-V medium [GIBCO BRL, Walkersville, MA], and 10% FCS [EQUITECH BIO, Ingram, TX] with 100 IU/mL of interleukin (IL)-2 [Shiongi Pharm. Co., Osaka, Japan]). At Days 7 and 14, the cells were incubated with the irradiated (50 Gry) autologous PBMCs as antigen presenting cells that had been pre-incubated with the same peptide at the same dose for two hours. The cells were harvested at culture Day 21 and were immediately tested for their ability to produce interferon-(IFN)-γ in response to various target cells by enzyme-linked immunosorbent assay at various effector to target cell (E/T) ratios in triplicate assays. For a 6 hour 51Cr-release assay, PBMCs (5,000 cells/well) that had been stimulated 3 times with the peptide were further cultured in a 96-well U-bottom microculture plate in the presence of feeder cells consisting of irradiated HLA-A24+ allogenic PBMCs (2 × 105 cells/well) that had been pre-pulsed with a corresponding peptide. Seven to 10 days later, the expanded cells were transferred to a 24-well plate and were incubated in the absence of either a peptide or feeder cells for 14–20 days. The CTL activity of these cells was re-checked by an IFN-γ production assay, and the cells were then tested for their cytotoxicity against 51Cr-labeled target cells at different E/T ratios. The surface phenotype of effector cells was investigated using an immunofluorescence assay with fluorescein isothiocyanate (FITC)-conjugated anti-CD3, -CD4, or -CD8 mAb. For inhibition of CTL activity, 2 mg/mL of anti-class I (W6/32, IgG2a) or anti-CD8 (IgG2a), anti-class II (H-DR-1, IgG2a), anti-CD4 (IgG1) mAb were used as described previously.17 Anti-CD13 (MCS, IgG2a) and anti-CD14 mAb (H14, IgG2a) served as a control mAb. The two-tailed Student t test was employed for the statistical analysis.

RESULTS

SART2 and SART3 Expression

The expression of the SART2 and SART3 antigen in the malignant and normal colorectal cells and tissues was investigated by Western blot analysis. Representative results are shown in Figure 1, and the summary is shown in the Table 1. The SART2 antigen was expressed in the cytosol fraction of 2 of 6 colorectal carcinoma cell lines, 5 of 20 (25%) colorectal carcinoma tissues, and 0 of 7 nontumorous colorectal tissues. It was not expressed at all in the nuclear fraction of colorectal carcinoma cell lines or colorectal carcinoma tissues tested. The SART3 antigen was expressed in the cytosol fraction of all 6 colorectal carcinoma cell lines, 27 of 41 (65.9%) colorectal carcinoma tissues, and 0 of 7 nontumorous colorectal tissues. This antigen was also expressed in the nuclear fraction of all 6 colorectal carcinoma cell lines, 30 of 41 (73.2%) colorectal carcinoma tissues, and 0 of 7 nontumorous colorectal tissues.

Figure 1.

Expression of the SART2 and SART3 antigens. Expression of protein levels of the SART2 and SART3 antigens in the samples was investigated by Western blot analysis using the anti-SART2 or the anti-SART3 polyclonal antibodies, respectively, as previously described.15, 16 Both the cytosol and nuclear fractions were investigated. Representative results are shown in the figure (PBMC [peripheral blood mononuclear cell] MKN45: a negative control; KE4: an esophageal squamous cell carcinoma as a positive control; Colo201, Colo205, Colo320, Colo201, HCT116, SW620: colorectal carcinoma cell lines).

Table 1. Expression of the SART2 and SART3 Antigens in Colorectal Carcinomas
SamplesSART3SART2
Cytosol (%)Nucleus (%)Cytosol (%)
Colorectal carcinoma cell lines6/6 (100)6/6 (100)2/6 (33)
Tumor tissues27/41 (65)30/41 (73)5/20 (25)
Colon carcinoma16/19 (64)14/19 (74)2/10 (20)
Rectal carcinoma11/22 (50)16/22 (73)3/10 (30)
Nontumorous Colorectal tissues0/7 (0)0/7 (0)0/7 (0)

Induction of CTLs by the SART3 Peptides

HLA-A24+ PBMCs of the three colorectal cancer patients were provided for stimulation in the presence of interleukin-2 (100 IU/mL) by the SART3109–118 and SART3315–323 peptides (10 mM), which were previously identified as peptides capable of inducing CTLs from the PBMCs of epithelial carcinoma patients.17 The percentages of CD8+ cells in these PBMCs stimulated by the SART3109–118 peptide were 32.2, 27.2, and 17.9, respectively (Table 2). The percentages of CD8+ cells in the PBMCs stimulated by the SART3315–323 peptide were 20.5, 8.8, and 58. The peptide-stimulated PBMCs from colorectal carcinoma patients were tested for CTL activity. PBMCs from all three colorectal carcinoma patients tested showed low but significant amounts of INF-γ production by recognition of KE4 and SW620 tumor cells when stimulated three times in vitro by either SART3109–118 or SART3315–323 peptide (Table 2). These peptide-stimulated PBMCs showed significant levels of cytotoxicity against KE4 and SW620 tumor cells (HLA-2402+), but not against Colo201 tumor cells, VA13 fibroblast cells (HLA-A2402), or HLA-2402+ PHA-blastoid cells. Representative cytotoxicity results are shown in Figure 2. These CTL activities were inhibited by anti-CD8 but not anti-CD4 mAb (data not shown). In addition, CTL sublines were established and tested for their reactivity to a peptide in order to confirm the peptide specificity in the CTLs (Fig. 3). Sublines from PBMCs of Patient 1 or Patient 2 stimulated with the SART3109–118 peptide or SART3315–323 peptide produced significant amounts of INF-γ by recognition of CIR-A*2402 cells pulsed with a corresponding peptide, respectively. The other sublines failed to show peptide specificity (data not shown).

Table 2. Cytotoxic T-lymphocyte Induction by the SART3 Peptides in PBMCs from HLA-A24+ Patients with Colorectal Carcinoma
  1. NS: not significant; PBNCs: peripheral blood mononuclear cells.

original image
Figure 2.

Cytotoxic T lymphocyte (CTL) induction by SART3-derived peptides. PBMCs from HLA–A 24+ colorectal carcinoma patients were stimulated in vitro with 10 μM of the SART3109–118 (VYDYNCHVDL) or SART3315–323 (AYIDFEMKI) peptides as described in materials and methods. These stimulated PBMCs were tested for their ability to produce interferon (INF)-γ by various target cells. Interferon-γ production at Day 21 is summarized in Table 2. (Values represent the means of triplicate deteterminants at an effector to target cell ratio of four. The background levels of INF-γ production of effector cells without target cells were subtracted from the values.) These peripheral blood mononuclear cells were further cultured for an additional 14–21 days followed by testing of their surface phenotypes and CTL activities against various target cells shown in the figure by a six hour 51Cr-release assay at different effector to target ratios.

Figure 3.

Peptide specificity of the peptide-induced cytotoxic T lymphocytes (CTLs). The CTL sublines were established by incubation of the peripheral blood mononuclear cells of patients that had been stimulated with the SART3109–118 or SART3315–323 peptide. HLA-A2402-binding human immunodeficiency virus peptide (RYLRDQQLLGI) was also prepared as a negative control. These sublines stimulated with the SART3109–118 peptide or SART3315–323 peptide produced significant amounts of interferon-γ by recognition of CIR-A*2402 cells pulsed with a corresponding peptide, respectively.

DISCUSSION

The SART3 antigen was detectable in the majority of colorectal carcinoma tissues. In contrast, the SART2 antigen was barely detected in the colorectal carcinoma tissues. We previously reported that it was expressed in the majority of head and neck squamous cell carcinomas (SCCs), 80% of esophageal SCCs, and 50% of lung SCCs and adenocarcinomas,17 while the SART2 antigen was detected in the majority of SCCs from various tissues, in the majority of renal cell carcinomas and brain tumors, and in about one-third of melanomas and adenocarcinomas from various organs other than the breast.16, 18 These results indicate that the SART3 antigen, but not the SART2 protein, could be an appropriate molecule for use in specific immunotherapy for colorectal carcinomas. The tumor cells (KE4, Colo320, and SW620) expressing both the SART3 and HLA-A24 molecules stimulated the HLA-A24-restricted and SART3 specific CTLs to produce significant levels of IFN-γ. In contrast, the others, including Colo201, Colo205, HCT116, and KM12LM cells (SART3+ and HLA-A24), failed to do so. Both SART3109–118 and SART3315–323 peptides have the ability to induce HLA-A24 restricted and tumor-specific CTLs in PBMCs.17 These peptide-stimulated PBMCs from colorectal carcinoma patients showed capability against the SART3+ and HLA-A24+ colorectal carcinoma cells, but failed to lyse any HLA-A24 SART3+ colorectal carcinoma cells or HLA-A24+ SART3+ PHA-blastoid cells. These results suggest the presence of CTL precursors reacting to these SART3 epitopes on cancer cells in the circulations of colorectal carcinoma patients.17, 20 With the exception of peptides for melanoma patients, the peptides capable of inducing CTLs against colorectal carcinoma have not yet been fully identified. The results in the current study indicate that administration of the SART3109–118 or SART3315–323 peptide could induce HLA-A24 restricted CTLs against colorectal carcinoma cells in HLA-A24+ cancer patients.

We have recently identified tumor-epitopes of SART3 antigen on HLA-A0207 molecules which are able to induce HLA-A2 restricted and tumor-specific CTLs in the PBMCs of epithelial carcinoma patients.21

The HLA-A24 allele is found in 60% of the Japanese, 20% of the Caucasian, and 12% of the African populations.22 The HLA-A2 allele is found in 40% of the Japanese and 50% of the Caucasian populations.22 These results suggest that SART3 peptides could be appropriate molecules for use in specific immunotherapy for HLA-A24+ or -A2+ colorectal carcinoma patients. Therefore, the SART3 antigen and its peptides could be an appropriate vaccine candidate for a relatively large number of colorectal carcinoma patients throughout the world.

Acknowledgements

The authors thank Kunzo Orita, M.D., Ph.D., for providing the natural human interferon-γ used for obtaining anti-interferon-γ polyclonal antibody.

Ancillary