Identification of HLA-A24 restricted shared antigen recognized by autologous cytotoxic T lymphocytes from a patient with large cell carcinoma of the lung

Authors


Abstract

The aim of the present study was to elucidate the tumor-specific cellular immunological responses occurring in a patient with large cell carcinoma of the lung who had no evidence of recurrence following surgical resections of both a primary lung lesion and a metastatic adrenal lesion. We analyzed an autologous tumor-specific cytotoxic T lymphocytes (CTL clone F2b), which were HLA-A*2402 restricted from regional lymph node lymphocytes. The F2b possessed T cell receptor (TCR) using the Vα5 and Vβ7 gene segment. The existence of precursor CTL (pCTL) against autologous tumor cells (A904L) was analyzed using CTL clone-specific PCR. Lymphocytes with the same TCR as F2b were detected in the primary tumor tissue, regional lymph node and the peripheral blood collected from the patient 3 years after the operation. Using the F2b, we identified a cDNA clone encoding the tumor antigen using cDNA expression cloning method. The gene was found to encode splicing variant of the Tara gene. Finally, we identified the 9-mer Ag peptide, using constructions of mini-genes. The F2b recognized 3 out of 7 HLA-A24 positive allogeneic tumor cell lines and in 1 out of 7 HLA-A24 negative allogeneic tumor cell lines when transfected with HLA-A24. This peptide is therefore considered to be potentially useful for performing specific immunotherapy in a significant proportion of lung cancer patients bearing HLA-A24. © 2006 Wiley-Liss, Inc.

Many genes encoding tumor antigens recognized by cytotoxic T lymphocytes (CTLs) have been identified from malignant melanoma since the identification of the MAGE gene.1, 2, 3 Synthetic peptides derived from these identified tumor antigens have been used as vaccine therapy for patients with malignant melanoma, and clinical responses in a significant proportion of patients have been reported.4, 5 Although most of these pioneering studies have focused on malignant melanoma, tumor antigens have also been identified in other malignant tumors including lung cancer.6, 7, 8, 9, 10 However, the expression of tumor antigens is heterogeneous among patients and also among tumor cells even in a single tumor.

Little is still known about the specific immune response against tumor antigens in lung cancer because of difficulties in both establishment of lung cancer cell lines and induction of CTLs. The identification of tumor antigens recognized by CTL is thus of critical importance in the development of vaccine-based immunotherapy. For this purpose, we previously reported the induction of tumor-specific CTL from patients with lung cancer (large cell carcinoma, adenocarcinoma, squamous cell carcinoma and adenosquamous cell carcinoma).8, 9, 10, 11, 12 We successfully induced tumor-specific CTLs in all 8 patients when their regional lymph node lymphocytes (RLNL) were stimulated with CD80-transfected autologous tumor cells.13 These findings therefore indicate the presence of tumor antigens on the lung cancer cells recognized by CTL in a significant proportion of patients with lung cancer.

We herein report analysis of 1 of 3 autologous tumor-specific CTL clones (clone F2b) against a lung cancer cell line from a patient (A904) with large cell carcinoma. Studies about the other 2 CTL clones have been reported previously.9, 14 The patient A904 underwent surgical resection of primary lung cancer and thereafter had a right adrenalectomy for adrenal metastasis. The patient is presently alive without any clinical evidence of recurrence at 8 years after the second surgery. To elucidate the immune-response to lung cancer, it is very important to monitor the precursor CTL (pCTL) frequencies. Many investigators have so far performed immuno-monitoring by limiting dilution methods, elispot assay and tetramer staining of HLA-class I and Ag peptide complex.15, 16, 17 In this experiment, the precursors of the CTL were evaluated by a CTL clone-specific PCR for T cell receptor (TCR) with special reference to the clinical course of the patient. Moreover, we also have identified HLA-A24 restricted shared tumor antigen, which encodes the Tara gene, recognized by the CTL clone. Regarding the HLA-A locus haplotypes among Japanese, the HLA-A24 positive population comprises about 60%. As a result, the development of a peptide vaccination using this tumor antigen should be thus potentially useful for the treatment of a large number of HLA-A24+ lung cancer patients.

Abbreviations:

AT, autologous tumor cells; CD80-AT, CD80-transfected autologous tumor cells; CM, culture medium; CTL, cytotoxic T lymphocytes; EBV-B, Epstein-Barr virus transformed B cells; mAb, monoclonal antibody; RLNL, regional lymph node lymphocytes.

Material and methods

Culture medium

The culture medium (CM) consisted of RPMI 1640 (GIBCO-BRL, Grand Island, NY) supplemented with 5 or 10% heat-inactivated fetal calf serum (FCS) (Equitech-Bio, Ingram, TX), 10 mM HEPES, 100 U/ml of penicillin G and 100 mg/ml of streptomycin sulfate.

Patient A904

The patient (A904) was a 51-year-old man who had undergone a the right upper lobectomy and a combined partial resection of the middle lobe in September 1996. The pathological stage of the lung cancer was T2N0M0, stage IB according to the Union Internationale Contore le Cancer (UICC) TNM classification.18 The clinical course of the patient is shown in Figure 1. In December 1996, computed tomography demonstrated a slight swelling of the right adrenal gland. Because the adrenal gland had gradually increased in size, in October 1997 (13 months after the first operation), a right adrenalectomy was performed and adrenal metastasis was pathologically confirmed. The patient received antitumor systemic chemotherapy using cisplatin, mitomycin C and vindecine sulfate until December 1997. He has no evidence of recurrence at 8 years after the 2nd operation.

Figure 1.

Clinical course of lung cancer patient A904. In 1996, the primary tumor was resected. In 1997, a right adrenalectomy was performed, and the patient was treated with systemic chemotherapy. Peripheral blood specimens of the patient were collected in 1999 and 2001.

Cell lines

A lung large cell carcinoma cell line (A904L) was established from an operative specimen of A904. The HLA class I genotype of A904L is HLA-A*2402, B*0702, Cw*0702 with haplotype loss. Lung adenocarcinoma cell lines A110L (HLA-A*2402, B*5201, Cw*1202), A129L (HLA-A*0206, B*0702, Cw*0702), B203L (HLA-A*2402, 31012, B*1501, 5401, Cw*0102, 0304), F1121L (HLA-A*2402, 0201, B*1507, 4006, Cw*0303, 0801), G821 (HLA-A*2602, B*5101, Cw*1402), lung squamous cell carcinoma cell line B1203L (HLA-A*2402, 2402, B*5201, 5401, Cw*1202, 0102), QG56 (HLA-A*2601, B*4601, Cw*0102) pleomorphic carcinoma cell line G603L (HLA-A*2602, 31012, B*1501, 4002, Cw*0303, 0304) and melanoma cell line G613M (HLA-A*2402, 1101, B*5201, 5502, Cw*1202, 1203) were established in our laboratory. Breast cancer cell line MCF-7 (HLA-A*0201, B*1802, Cw*0501) was purchased from American Type Culture Collection (American Type Cell Culture, Rockville, MD). Non-small cell lung cancer cell lines SQ-1 (HLA-A*1101, 2402, B*1501, 5201, Cw*0401, 1202,), PC-9 (HLA-A*0602, 2402, B*0702, 5502, Cw*0302, 0702) and PC13 (HLA-A*2601, B*3901, Cw*0702) were kindly donated by Dr. K. Itoh, Kurume University, Kurume. The identification of their HLA genotypes was performed by polymerase chain reaction (Shionogi Co., Osaka, Japan). K562 is an erythroleukemia cell line lacking MHC class I expression on the cell surface, which is sensitive to natural killer cell cytotoxicity. A904 B cells immortalized by Epstein-Barr virus (EBV-B) (HLA-A*2402, 2603, B*0702, 3901, Cw*0702, 0702) were produced from this patient's PBMC by an infection with the supernatant from the EBV producer line B95.8. A904-PHA-blast (HLA-A*2402/2603, B*0702/3901, Cw*0702) was induced by stimulation with 10 ng/ml PHA and 200 U/ml rIL-2 (kindly donated by Takeda Chemical Industries, Osaka, Japan). Rosi-EBV-B (HLA-A*2402, 3201, B*3503, 44031, Cw*0401, 0401), TNF-sensitiveWEHI-164cl3 cells and 293-EBNA cells were kindly donated by Dr. Coulie PG (Cellular Genetics Unit, Université Catholique de Louvain, Brussels, Belgium). WEHI-164cl3 cells were maintained in CM with 5% FCS, 293-EBNA cells were maintained in DMEM (Life Technologies, Gaithersburg, MD) with 5% FCS, while the other cell lines were maintained in CM with 10% FCS.

Transfection of CD80 into tumor cells

A904L was transfected with 2 μg of the pBj/hCD80 plasmid containing human CD80 gene (kindly provided by Dr. M. Azuma, National Children's Medical Research Center, Department of Immunology) using a lipofectamin reagent (GIBCO-BRL) according to an instruction manual as described previously.19

Isolation of lymphocytes

The regional lymph nodes from the lung cancer patient were obtained at the time of surgery. RLNL were prepared as described previously.20 RLNL were frozen in a deep freezer at −130°C until use.

Induction of CTL

RLNL (2.6 × 107) obtained at the time of surgery were rapidly thawed and stimulated with irradiated (100 Gy) CD80-transfected A904L weekly at a tumor cell-to-lymphocyte ratio of 1:10 in CM with 10% FCS with 25 U/ml rIL-2 (kindly donated by Takeda Chemical Industries), 5 ng/ml IL-4 (Serotec, Oxford, U.K.) and 5 ng/ml IL-7 (Genzyme Techne, Cambridge, MA) for 4 weeks in 24-well plates (Iwaki glass, Tokyo, Japan) at 37°C in a 5% CO2 atmosphere as previously reported.21 The CTL activity was assessed at 1 week after the 4th stimulation. To generate T cell clones, limiting dilution was performed from the bulk CTL line 7 days after the last stimulation as described previously.8, 9, 10, 12

mAb (monoclonal antibody)

Hybridomas (HB-145, HB-95) were purchased from the American Type Culture Collection (American Type Cell Culture). C7709.A2.6 (anti-HLA-A24) and B 1.23.2 (anti-HLA-B, C) were kindly donated by Dr. Coulie PG (Cellular Genetics Unit). The culture supernatants of ATCC HB-145 (IVA12; anti-HLA-DR, DP, DQ), HB-95 (W6/32; anti-HLA-A, B, C), C7709.A2.6 and B 1.23.2 were used for analyzing the HLA restriction of T cell clones.

Cytotoxicity assay and cytokine production of CTL

The cytotoxicity of CTL was assessed by a standard 4 hr 51Cr release assay as described previously.20 TNF production of CTL was measured by WEHI assay using TNF-sensitive WEHI cells.22 Briefly, CTL clone F2b (6 × 104/ml) was incubated with tumor cell lines (6 × 105/ml) in CM with 10% FCS overnight, and the amount of TNF in the culture supernatant was measured by measurement of the cytotoxic effect on WEHI-164cl3 cells in an MTT colorimetric assay.23 To asses involvement of standard proteasome, tumor cell line was treated with/without IFN-γ (50 U/ml) for 1 week before CTL assay. In the blocking assay using mAbs, a 1/4-diluted culture supernatant of hybridomas was added into the co-culture of CTL and A904L as described previously.8, 9, 10, 12

Cloning of HLA-A24, B07, Cw07 cDNAs from A904L

The total RNA extracted from A904L using the Rneasy Mini Kit (Qiagen, Hilden, Germany) was converted to cDNA using an oligo (dT) primer. The cDNA served as a template for PCR amplification using each HLA-class I specific forward primer and reverse primer and was then cloned into pcDNA3.24

TCR usage analysis by direct sequence

The TCR Vα and Vβ usage of CTL clones (F2b) was assessed by RT-PCR and sequencing. RNA extracted from each clone using the Rneasy Mini Kit (Qiagen) was converted to cDNA using an oligo(dT) primer. cDNA served as a template for PCR amplification using a panel of Vα and Vβ specific forward primers and reverse Cα and Cβ primers.25, 26 The PCR products were purified using the QIAquick Gel extraction kit, the QIA quick PCR purification Kit (Qiagen) and then were sequenced with the BigDye Terminator Cycle Sequence kit (Perkin-Elmer PE Applied Biosystems, Foster City, CA). The products of the sequencing reactions were analyzed on an ABI 310 Sequencer (Perkin-Elmer PE Applied Biosystems).

PCR analysis for the detection of CTL clone F2b

To prove whether the CTL clone F2b with the same TCRs was present in the primary tumor tissue, a regional lymph node was collected at the first surgery and peripheral blood specimen were collected at either 3 or 5 years after the first operation. RNA extracted from each sample was converted to cDNA. The cDNA served as templates for PCR amplification using Vα and equation image specific forward primers (OKY 69: 5′-TTGCAGACACCGAGACTGGGGACT-3′, OKY 49: 5′-GGTGTGATCCAATTTCGGGTCATG-3′) and TCR-D-J jointing region-specific reverse primer (OKY 487: 5′-CTGCGGCTGTGGTCCAGCTGA-3′, OKY 119: 5′-CTCCAAAATATATGGTGTTTCTTAGGTCG-3′). Second PCR was performed using diluted 1st PCR products (1:1 × 104) as a template with using the same forward primer and a newly designed TCR-D-J jointing region-specific reverse primer (OKY 811: 5′-TGCTCTCCTGCCCGTACT-3′, OKY 118: 5′-CCTCTCCAAAATATATGGTGTTTCTT-AGG-3′). Complementarity-determining region 3 (CDR3) region of TCR was confirmed by direct sequencing.

Construction and screening of the cDNA library

About 5 μg of poly (A)+ RNA extracted from A904L using the Fastrack kit (Invitrogen, San Diego, CA) was converted to cDNA with the Superscript Choice System (Life Technologies) using an oligo (dT) primer [5′-ATAAGAATGCGGCCGCTAAACTA (T) 18VZ-3′; V = G, A or C; Z = G, A, T or C] containing a Not I site at its 5′ end. The cDNA was ligated to Hind III - EcoRI adoptors (Stratagene, Heidelberg, Germany), phosphorylated, digested with Not I and inserted at the Hind III and Not I site of expression vector pCEP4 (Invitrogen). Escherichia coli TOP10 (Invitrogen) were transformed by electroporation with the recombinant plasmids and selected with ampicillin (50 μg/ml) as described previously.8 The library was divided into 734 pools each containing 100 cDNA clones. Each pool was amplified for 4 hr and plasmid DNA was extracted using QIA prep 8 plasmid Kit (Qiagen). Duplicate microcultures of 293-EBNA cells, plated in flat-bottomed 96 microwells (3.5 × 104/well) 24 hr before transfection, were cotransfected with 1.3 μl of lipofectAMINE reagent (Life Technologies), 100 ng of plasmid DNA of each pool of cDNA library and 50 ng of plasmid pcDNA3 (Invitrogen) containing the HLA-A*2402 cDNA. After 24 hr, CTLs (3,000 cells/ well) were added, and the supernatant was collected after another 24 hr. The TNF content of each well was measured by testing the cytotoxic effect on WEHI-164cl3 cells in an MTT colorimetric assay.25, 26, 27, 28

Quantitative RT-PCR

Quantitative RT-PCR was carried out in ABI PRISM 7000 (Applied Biosystems, Foster, CA). The relative amount of Tara mRNA was measured by means of detection of intercalated SYBR green. PCR was performed with the 10 μl SYBR GREEN PCR Master Mix (Applied Biosystems), either 1 μl of cDNA or 1 μl of water and each primer set was described below in a total volume of 20 μl. The PCR conditions were 10 min at 95°C for denature, followed by 50 cycles of 15 sec at 95°C for and 1 min at 60°C. The primer sequences of Tara for quantitative RT-PCR were OKY501: 5′-TCGCTCACCACCACCTCTACT-3′ and OKY 511: 5′-GACAGCATCCTTGGTGTGGAT-3′.

The quantitative PCR primer sequences of β-actin were CCTGGCACCCAGCACAAT and GCCGATCCACACGGAGTACT. The concentration of each primer set was 200 nM for the identified gene and β-actin. The threshold cycle number (CT) was defined as a fractional cycle number at which the amount of amplified target product reaches a fixed threshold. ΔCT was obtained by comparing CT of Tara with CT of β-actin in same amount of templates. Relative quantification was achieved by comparisons with ΔCT of A904L. The relative expression was calculated by the following formula:

equation image

Peptide binding assay

C1R-A24 cells were used for binding studies. A 1 mM peptide stock solution in PBS 10% DMSO was made, and cells were incubated with peptide at a final concentration of 50 μM in RPMI 1640 overnight at 37°C. HLA surface expression was monitored after staining with primary antibody C7709.A2.6 (antiHLA-A24mAb) and secondary antibody of a FITC-coated goat antimouse IgG (Dianova, Hamburg, Germany) on a FACSCalibur cytometer (Becton-Dickinson, San Jose, CA).

Results

Generation of a CTL clone recognizing autologous large cell carcinoma cells

RLNL (2.6 × 107) isolated and cryopreserved at the time of surgery were thawed and stimulated with 2.6 × 106 of irradiated CD80-A904L in the presence of IL-2, IL-4 and IL-7. The A904L-specific bulk CTL line was induced after 4 times of tumor stimulations. The CTL line was cloned by limiting dilution and finally the CTL clone F2b was established. The CTL clone F2b lysed AT, but did not lyse autologous EBV-B, K562 or A904-PHA-blast (Fig. 2).

Figure 2.

Characterization of CTL clone F2b. The cytotoxic activities of CTL clone F2b derived from patient A904. CTL clones were obtained from CTL line by limiting dilution. Target cells included AT (A904L), autologous EBV-B, K562 and autologous PHA-blast. Open squares indicate AT; closed squares indicate autologous EBV-B; open triangles indicate K562; closed triangle indicate autologous PHA-blast. The blocking effects of mAbs on the CTL (F2b) activity. The CTL was cocultured with AT overnight, and the amount of TNF in the culture supernatant was measured by WEHI assay.

To analyze the HLA restriction, we measured the cytokine production by CTL clone F2b in response to A904L in the presence of mAbs against various HLA molecules. CTL clone F2b produced TNF in response to A904L. The production of TNF by CTL clone F2b was inhibited by the addition of anti-MHC class I mAb (W6/32) or anti-HLA-A24 mAb (C7709.A2.6) but not anti-MHC class II mAb (IVA12) nor anti-HLA-B/C mAb (B 1.23.2), thus suggesting that the response of CTL clone F2b against A904L was restricted by HLA-A*2402. The cross-reactivity of CTL clone F2b against 7 HLA-A24 positive allogeneic tumor cell lines and against 7 HLA-A24 negative tumor cell lines transfected with HLA-A24 were tested. As shown in Figure 3, CTL clone F2b showed TNF production against 3 of 7 HLA-A24 positive allogeneic tumor cell lines other than autologous A904L (Fig. 3a) and 1 out of 7 HLA-A24 negative tumor cell lines (Fig. 3b). These CTL responses were not affected by IFN-γ treatment of target cells at all as shown in Figure 3a. These findings suggested that the Ag recognized by CTL clone F2b in the context of HLA-A24 is a shared Ag among autologous A904L and allogeneic tumor cell lines.

Figure 3.

Cross-reactive TNF production of CTL clone F2b against HLA-A locus matched and un-matched allogeneic tumor cell lines. The CTL was cocultured with the indicated HLA-A24 positive tumor cell lines (B1203L, B203L, A110L, F1121L, G613M, SQ-1 and PC-9), treated with/without IFN-γ (50 U/ml) for 1 week before CTL assay; (a) and HLA-A24 negative tumor cell lines (A129L, G831L, A549, G603L, QG56, PC13 and MCF7); (b) HLA-A24 negative tumor cell lines transfected with HLA-A24, overnight, and the amount of TNF in the culture supernatant was measured by WEHI assay. Relative expression of Tara mRNA was calculated on the base of Tara mRNA expression in normal lung tissue as 1.0.

Clonotypic PCR and detection of the CTL clone F2b

CTL clone F2b was shown to express the Vα5 and Vβ7 gene segments (Fig. 4a). To determine whether or not T lymphocytes expressing the same TCR as CTL clone F2b were present in the fresh tumor tissue, mediastinal regional lymph node collected at the first surgery and then peripheral blood collected 3 and 5 years after the first operation, clone specific PCR was performed using Vα and equation image specific forward primers and TCR-D-J jointing region-specific reverse primer as described in the Material and methods. The detection limit of the clonotypic PCR was 1 cell in 1 × 105 cells. The cDNA from each sample served as templates for PCR amplification followed by DNA sequence analysis. As shown in Figure 4b, the same band for Vα 5 and Vβ7 was detected in 20 mg of the primary tumor tissue, in 1 × 105 of lymphocytes from the regional lymph node and furthermore, in 1 × 106 of lymphocytes from the peripheral blood at 3 years after the first operation, but not in the peripheral blood (1 × 106) at 5 years after the first operation. The same PCR products for Vα5 and Vβ7 were not detected in samples of PBMC from 3 healthy donors. These results indicated that the pCTL still circulated in the peripheral blood even 3 years after the operation.

Figure 4.

The existence of the CTL clone F2b by a TCR analysis. (a) An analysis of TCR usage of CTL clone F2b. The CTL clone F2b possessed Vα5 and Vβ7. (b) The existence of same PCR product for Vα5 and Vβ7 of F2b in the primary tumor tissue, regional lymph node and peripheral blood. lane 1: F2b (positive control); lane 2: primary tumor tissue; lane 3: 1 × 105 of lymphocytes from regional lymph node; lane 4: 1 × 106 of lymphocytes from peripheral blood collected at 3 years after the first operation; lane 5: 1 × 106 of the lymphocytes from the peripheral blood at 5 years after the first operation; lane 6: water (negative control). The β-actin cDNA of each sample was detected at the same level by RT-PCR.

Identification of the cDNA clone encoding specific antigen

A cDNA library prepared with total mRNA from A904L was cloned into the expression vector pCEP4. This plasmid contains the EBV origin of replication, thus resulting in the episomal multiplication of the transfected with EBV EBNA-1 gene. The cDNA library was divided into 734 pools each containing 100 cDNA clones. The plasmid DNA was extracted from each pool and then was cotransfected into duplicate microcultures of 293 EBNA cells with an HLA-A*2402 construct. After 48 hr, the cells cotransfected with HLA-A24 were screened for the expression of the antigen by adding CTL clone F2b and measuring the production of TNF. Only 1 pool of cDNA proved to be positive for TNF production. The positive pool was subcloned from 100 to 12 clones, and finally a single cDNA clone 214.2 was isolated (Fig. 5a). CTL clone F2b showed specificity for the 293EBNA cells co-transfected with cDNA clone 214.2 and the HLA-A24 construct. The transfectants did not secrete TNF by themselves. The CTL clone F2b did not respond to the transfectants with HLA-A24 alone, or cDNA clone 214.2 alone, or combination of cDNA clone 214.2 and unrelated HLA-A2. The cDNA clone 214.2 was 3339-bp long and contained a poly (A) tail. Its sequence encoded for splicing variant of known gene, named Tara, in the databanks (GenBank Accession No.: AF281030). The splicing variant contained an open reading frame and was found to consist of 617 amino acids.

Figure 5.

Identification of a cDNA clone encoding antigen and the antigenic peptide recognized by CTL clone F2b. (a) 293-EBNA cells (5 × 104/well) were co-transfected using lipofectamine with 50 ng plasmid vector of HLA-A24 and cDNA clone 214.2. CTL clone (3 × 103/well) was added after 24 hr, and the medium was collected to measure the TNF after another 24 hr. (b) cDNA clone 214.2 is represented as a box with a open reading frame and a 5′ and 3′ untranslated region. Minigenes were amplified by PCR with a common sense primer (OKY167) and different antisense primers (OKY164, OKY192 and OKY193). The products were cloned into pcDNA3, and the constructs were cotransfected with HLA A24 cDNA into 293-EBNA cells. The recognition of the transfectants by CTL clone F2b was evaluated as TNF production by using WEHI assay.

Identification of the antigenic peptide

The cDNA clone 214.2 was sequenced as described earlier. To identify the antigenic peptide, several cDNA fragments (mini-genes) were amplified from the cDNA clone 214.2 with PCR. The PCR conditions were 5 min at 94°C followed by 30 cycles consisting of 1 min at 94°C, 2 min at 59°C and 3 min at 72°C, followed by final elongation step of 10 min at 72°C. The PCR products were purified using QIAquick Gel extraction kit (Qiagen) and cloned into expression plasmid pcDNA3.1 using the Eukaryotic TOPO TA Cloning Kit (Invitrogen). The constructs were cotransfected, using LipofectAMINE as earlier, with the HLA-A24 cDNA clone into 293-EBNA cells.11 Transfections of truncated forms of cDNA clone 214.2 indicated that the sequence coding for the antigenic peptide was located between nt 64 (OKY167) and 871 (OKY164) of the cDNA (Fig. 5b). With a set of minigenes covering this region, the portion of cDNA encoding the antigen epitope was narrowed down to a 6-bp sequence between OKY193 and OKY191. Ten nonapeptides covering this region were synthesized. The autologous EBV-B cells were loaded with each nonapeptide. The cytolytic activity of CTL clone F2b was determined against the EBV-B loading each peptide as shown in Figure 6. CTL clone F2b responded to the EBV-B loading the peptide NYGFQIHTK alone. Although the binding affinity of identified antigenic peptide to stabilize HLA-A*2402 was very low as shown in Figure 7, one-half of maximal lysis was obtained at 20 ng/ml.

Figure 6.

Identification of the antigenic peptide. 51Cr-labeled A904-EBV-B cells incubated for 20 min at room temperature with the indicated concentrations of the each peptide, and CTL clone was added at an E:T ratio of 10. Chromium release was measured after 4 hr.

Figure 7.

Peptide binding assay. Each peptide was compared for their ability to stabilize HLA-A*2402 on the surface of C1RA24 cells. The mean fluorescence of C1RA24 cells incubated overnight with 50 μM peptide was examined by flow cytometry.

Expression of the Tara gene

Expression of this gene at mRNA level was investigated by RT-PCR methods in normal tissues. The specific band for Tara gene was observed in all of 20 normal tissues and this gene was expressed all of 15 tumor cell lines tested (data not shown). Among normal lung tissue, A904EBV-B and tumor cell lines, relative expression levels of Tara gene were determined by quantitative RT-PCR. Relative expression levels of Tara gene in A904L and A904EBV-B were 37 and 0.02, respectively on the basis of Tara expression in normal lung tissue as 1.0. The relative expression levels of this gene in tumor cell lines were not correlated with magnitude of response of the CTL clone F2b. However, it is likely that Tara was overexpressed in some tumor cell lines as B203L, G613M, SQ-1 and PC9 other than autologous A904L (Fig. 3).

Discussion

Lung cancer is one of the tumors with the worst prognosis among human cancers. Since permanent lung cancer cell lines are extremely difficult to establish, little is still known about the tumor-specific immune response and tumor antigens specific for lung cancer. There have only been a several reports concerning the CTL against autologous lung cancer cells.8, 9, 10, 11, 12, 27 From these studies, such mutated Ags as elongation factor 2 gene,28 α-actinin-4, malic enzyme17 and NF-YC10 have been identified.

As shown in Figure 1, the patient A904 underwent a surgical resection of primary lung cancer and 13 months later further underwent a resection of right adrenal metastasis. Thereafter, the patient has remained alive without any clinical evidence of recurrence 8 years after the second surgery. Three CTL clones were induced against A904L. Two of them have been reported previously.9, 14 Concerning 2 other CTL clones, the frequency against AL137255 was estimated 1/2.6 × 107 and the frequency against mutated p53 was estimated 1/2.1 × 107. However, the frequency of this CTL clone F2b was estimated to be 1/1.0 × 105. It is possible that these CTL clones played an important role in the clinical outcome of this patient.17 Hishii M. et al.29 reported that the identical antimelanoma T cell clone accumulated in 2 different metastatic tumor tissue specimens in a melanoma patient.30 We thus tried to determine whether CTL clone F2b existed in the primary tumor tissue, regional lymph node and in the peripheral blood of patient A904 collected at 3 and 5 years after the first operation. CTL clone F2b expressed TCR using the Vα5 and Vβ7 gene segment and PCR amplification was performed using Vα5 and Vβ7-specific forward primer and each TCR-D-J jointing region-specific unique reverse primer. The PCR product specific for Vβ7 was detected in the primary tumor tissue, regional lymph node (1 × 105 cells) and in the peripheral blood of the patient (1 × 106 cells) collected at 3 years after the first operation (Fig. 4b). It is probable that the circulating CTL may be continuously re-stimulated by tumor antigens, because micrometastatic tumor cells were frequently detected in the bone marrow and lymph nodes even in the patients who underwent a complete resection of tumors and who are at a high risk of experiencing recurrent disease as reported by us.9, 31 Melanoma and lung cancer patients who have high frequencies of tumor-specific CTLs in the peripheral blood have been reported to enjoy an unusually favorable clinical evolution.16 The detection limit of the clone-specific PCR was less than 1 × 10−5. The sensitivity of a limiting dilution analysis, ELISPOT assay or tetramer technique is reported to be around 1 × 10−4 of CD8+ cells.32 In this study, HLA and epitope peptide complex tetramer detected the CTL clone F2b, but could not detect any cluster of cells from the peripheral blood of the patient (1 × 106 cells) collected at 3 and 5 years after the first operation (data not shown). The sensitivity of the clone-specific PCR might be higher than that of tetramer technique.

The identified protein, Tara, has a molecular weight of about 80,000. Tara consists of an N-terminal pleckstrin homology domain and C-terminal coiled-coil region and is located on chromosome 22q. The PH domain could mediate protein–protein interactions, including the direct binding to actin. The coiled-coil region dimerizes and enables Tara dimerization or the formation of higher order structures. Coiled-coils can form rod-like structures that are important for specific functions and they also facilitate binding to other proteins.33Tara regulates actin cytoskeletal organization by the direct bind and stabilization of F-actin. The localized formation of Tara and Trio complexes functions to coordinate actin remodeling.

Although Tara mRNA is ubiquitously expressed in human tissues including normal tissues and cancer cell lines by RT-PCR,34 CTL clone F2b recognized autologous A904L and 4 of 14 allogeneic cancer cells in the context of HLA-A24. Relative expression levels of Tara in A904L and A904EBV-B were 37 and 0.02, respectively, on the basis of Tara expression in normal lung tissue as 1.0. As shown in Figure 3, the relative expression levels of this gene in tumor cell lines were not correlated with magnitude of response of the CTL clone F2b. The reason why CTL clone F2b did not recognize either autologous EBV-B or PHA blast might be that Tara was minimally expressed in autologous EBV-B or PHA blast when compared with autologous tumor cells. An another possible explanation for the discrepancy that CTL clone F2b lysed only tumor cells but not normal cells is (i) the posttranscriptional regulation of the gene expression as well as alterations in processing of the protein caused higher expression levels of epitope and (ii) over expression in transient expression system might result in presentation of peptide on the cell surface at sufficient levels to be recognized by CTL.35 Engelhard reported about the unusual peptides derived from ubiquitous protein through posttranslational modification,36 for more detail, Hanada et al. reported a peptide derived from the fibroblast growth factor-5 protein by the posttranslational splicing of 2 discontiguous segments.37 The other possibility is (iii) the existence and involvement of different antigen cleavelage and processing machinery such as immunoproteasome that works in some tumor cells but not in normal cells, EBV-B cells, and some of cell lines.38 As shown in Figure 3a, CTL activity was not affected by the treatment of target cells with IFN-γ, therefore, the antigen processing was not ascribed to standard proteasome.

To ascertain whether the peptide-specific CTL precursors exist in other patients with lung cancer, RLNL from such patients were stimulated with the peptide (NYGFQIHTK) in vitro. The peptide-specific CTL have been established from 1 of 3 lung cancer patients with HLA-A24 (data not shown).

Regarding the HLA-A locus haplotypes among Japanese, the HLA-A24 positive population comprises about 60%.39 Since the CTL clone F2b recognized in 4 out of 14 allogeneic tumor cell lines in HLA-A24-restricted manner, then the peptide derived from Tara should be a shared antigen among lung cancer. As a result, a peptide vaccination could therefore be potentially useful for performing specific immunotherapy in a significant proportion of lung cancer patients with HLA-A24.

Acknowledgements

We thank Ms. Yumiko Hase, Ms. Kahoru Noda and Ms. Miki Shimada for their expert help.

Ancillary