INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

When blood lymphocytes or tumor-infiltrating lymphocytes (TIL) of cancer patients are incubated with autologous tumor cells in the presence of interleukin-2 (IL-2), they often proliferate, with the responder cells being capable of lysing the tumor cells (1). From populations of such cytolytic T lymphocytes (CTL) obtained in mixed lymphocyte-tumor cell culture (MLTC), it is possible to derive stable CTL clones that specifically recognize antigens presented by human leukocyte antigen (HLA) class I molecules on the autologous tumor cells. Owing to the relative ease with which melanoma cell lines may be obtained, most antitumor CTL clones have been derived from melanoma patients. Gene transfection approaches have been used to identify the genes that encode the antigens recognized by such antitumor CTL clones. The genes so far identified as encoding such antigens fall into four groups (5, 6). A first group comprises families of genes, such as MAGE, BAGE, or GAGE, that are not expressed in normal tissues except for male germinal cells, but which are expressed in a significant proportion of tumors of different histologic types. A second group contains genes such as tyrosinase, which are expressed only in normal melanocytes and in melanoma cells. Genes that are mutated in tumor cells as compared with the normal cells of a patient constitute a third group, and the final group consists of genes that are overexpressed in tumor as compared with normal cells.

Little is known about the presence on lung tumors of antigens that can be recognized by autologous T lymphocytes and that could be the targets of tumor rejection responses. Most lung tumors belong to the categories of non-small-cell lung cancer (NSCLC) or small-cell lung cancer (SCLC). NSCLC includes squamous, adeno-, and large-cell carcinomas. These malignancies can often be cured by surgical resection before the appearance of locoregional or distant metastases, but are weakly sensitive to chemotherapy and radiotherapy (7). In contrast, SCLC cells are highly sensitive to chemotherapy and radiotherapy but metastasize earlier than NSCLC cells (8). No information is currently available about antigens recognized by T lymphocytes on SCLC cells, whereas two reports have described CTL recognizing NSCLC cells. HLA-A28-restricted CTL were obtained in an autologous MLTC experiment with a squamous-cell lung carcinoma (4, 9). The corresponding antigens have not yet been identified. HLA-A2-restricted CTL, derived by stimulating TIL with anti-CD3 antibodies and autologous NSCLC cells, were shown to recognize the HER-2/ neu gene product (10). We report here the generation by autologous MLTC of CTL clones showing specificity for SCLC or NSCLC cells.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

Patient LB129 (HLA-A2, B44, Cw5, Cw7), a 58-yr-old man, presented in December 1988 with an SCLC invading the mediastinum. No extrathoracic dissemination was detected. A complete tumor response was obtained after six courses of chemotherapy, and the patient has been disease-free since then (Figure 1).

View larger version (17K): [in this window] [in a new window]   Figure 1.   Clinical evolution of patients LB129 and LB37.

Patient LB37 (HLA-A2, A26, B8, B51, Cw2, Cw7), a 55-yr-old man, presented in January 1990 with a squamous-cell carcinoma of the lung. The tumor was resected, but only partially because of a mediastinal invasion. The patient received two courses of cisplatin, vindesin, and mitomycin as adjuvant chemotherapy, and a complete response was obtained. Subsequent to October 1990, the patient was vaccinated with intradermal inoculations of lethally irradiated autologous tumor cells, consisting of LC-5 cells and five clones derived from a population of LC-5 cells that survived mutagenic treatment with N-methyl-N-nitroguanidine (MNNG). In May 1993, computed tomography (CT) indicated the presence of enlarged mediastinal lymph nodes, of which one was biopsied through mediastinoscopy and proved to have been invaded by squamous carcinoma cells. The patient was treated with radiotherapy and vaccinations with autologous tumor cells. He has had no evidence of disease since then (Figure 1).

Cells

Tumor cell line LB-129-SCLC was derived from a mediastinal lymph node collected from patient LB129 through mediastinscopy on January 3, 1989. These cells grow in RPMI-1640 medium (Gibco Laboratories, Grand Island, NY) supplemented with 10% fetal calf serum (FCS), L-arginine (116 mg/L), L-asparagine (36 mg/L), L-glutamine (216 mg/L) (AAG), and hydrocortisone (10 mM), insulin (5 µg/ml), transferrin (100 µg/ml), 17-estradiol (10 mM), and selenium (30 mM), a mixture reported to improve the growth of SCLC cells in vitro (11). It is difficult to prove formally that these permanently growing cells are SCLC cells. However, they grow in vitro as loose clumps in suspension, like other cell lines from SCLC patients; they are labeled in flow cytometry by monoclonal antibodies LCA-1, -2, and -3, which have been shown to recognize SCLC cell lines (12); and they are labeled in histochemistry by monoclonal antibodies recognizing the CD57 antigen (Leu-7; Becton Dickinson, Mountain View, CA) and the neuroendocrine antigens neuron-specific enolase (BBS/NC/VI-H14; DAKO, Glostrup, Denmark) and chromogranin A (DAK-A3, DAKO), which have been shown to be present in SCLC cells (13, 14).

Tumor cell line LB37-NSCLC was derived from a metastatic lymph node resected from patient LB37 during the thoracotomy performed in 1990. The cell line was established in serum-free RPMI-1640 supplemented with insulin, transferrin, sodium selenite, hydrocortisone, epidermal growth factor, ethanolamine, triiodothyronine, phosphorylethanolamine, bovine serum albumin, 4-(2-hydroxyethyl)- 1-piperazine-N-2-ethanesulfonic acid (Hepes), sodium pyruvate, and glutamine, as described for the ACL-4 medium (15), and was further maintained in Iscove's medium plus 10% FCS and AAG. The cells were stained by antibodies against keratin (polyclonal rabbit antihuman keratin [DAKO]), by a monoclonal antibody recognizing human cytokeratin (CAM 5.2 [DAKO]), and by a monoclonal antibody recognizing the human epithelial membrane antigen (E29 [DAKO]), confirming that LB37-NSCLC cells are of epithelial origin. LB37-NSCLC-5 (LC-5) is a clonal subline derived from LB37-NSCLC by limiting dilution. LB37-NSCLC-209 (LC-9) is one of the clonal sublines derived from LB37-NSCLC cells that survived two treatments with MNNG. LC-9 cells were cotransfected with the expression plasmid pcDNA1/Amp (InVitrogen, Leek, The Netherlands) containing a complementary DNA (cDNA) encoding the B7-1 costimulatory molecule, and with plasmid pSVtkneo containing a gene conferring resistance to neomycin. Transfectants were selected with the neomycin analogue G418 (Gibco) added at 0.75 mg/ml. G418-resistant cells were cloned, and clone LC-9.B7 was derived. These cells were labeled with an anti-B7-1 antibody (Becton Dickinson).

Restriction fragment-length polymorphism (RFLP) analyses confirmed that the LB129-SCLC and LB37-NSCLC cells were derived from patients LB129 and LB37, respectively.

Lines of activated T lymphocytes were established by stimulating blood mononuclear cells with phytohemagglutinin-A (PHA-P; Difco Laboratories [Detroit, MI]; 1/1,000 [vol/vol] and IL-2 (100 U/ml, with 1 unit/ml giving 50% maximal proliferation of CTLL-2 cells).

Fibroblasts were derived from a skin biopsy of patient LB37.

Derivation and Culture of CTL Clones

To obtain the CTL clone recognizing LB129-SCLC cells, blood mononuclear cells collected from patient LB129 in January 1989 were labeled with an anti-CD8 antibody coupled to fluorescein, and were sorted with a flow cytometer. The purified CD8⁺ lymphocytes (10⁵/ well) were stimulated with irradiated (100 Gy) LB129-SCLC cells (10⁵/ well) that had been incubated for 2 d with interferon- (IFN-) (Boehringer-Mannheim, Mannheim, Germany; 50 U/ml), in 2 ml Iscove's medium with AAG plus 5 × 10⁵ M 2-mercaptoethanol (AAGM) supplemented with 10% human serum (HS: a pool of serum from blood donors), recombinant IL-1 (rIL-1) (3 U/ml; a gift of P. Lomedico of the Roche Research Center, Nutley, NJ), and rIL-4 (approximately 5 U/ml, with 1 U/ml being the concentration that gives 50% maximal proliferation of human T cells previously activated with PHA and IL-2). IL-2 (30 U/ml) was added 3 d later. The lymphocytes were restimulated each week by the addition of irradiated IFN--treated tumor cells in medium with IL-1, IL-2, and IL-4. On Day 24 the population of CTL was cloned by limiting dilution in microwells, and was stimulated each week with irradiated IFN--treated tumor cells (3,000/well), and irradiated allogeneic Epstein-Barr virus (EBV)- transformed B cells LG2-EBV (10⁵/well) in medium (Iscove's medium plus AAGM and HS) with IL-1, IL-2, and IL-4 as above. After 4 wk of stimulation, 24 clonal populations could be tested for their lytic activity against the autologous tumor cells. Three of clonal populations lysed LB129-SCLC cells specifically. One of the populations, CTL clone LB129-CTL8/10, proliferated sufficiently to be amplified and analyzed further.

To obtain CTL clone 1/7, blood mononuclear cells collected from patient LB37 in September 1995 were incubated with magnetic microbeads coated with an anti-CD8 antibody (Miltenyi Biotech, Auburn, CA), and CD8⁺ lymphocytes were isolated with a magnet. The lymphocytes were cultured in Iscove's medium supplemented with AAGM and 3% autologous serum. The lymphocytes were stimulated with irradiated (150 Gy) LC-5 cells that had been incubated for 7 d in AAGM supplemented with autologous serum as described earlier, and to which IFN- (50 U/ml) had been added 2 d before the stimulation. IL-2 was added at a concentration of 5 U/ml. The lymphocytes were restimulated each week by the addition of irradiated LC-5 cells in fresh medium with IL-2. The treatment of the autologous tumor cells (replacement of FCS by autologous serum, and addition of IFN-) was used for each restimulation of the lymphocytes throughout the MLTC, and also for the first three restimulations of the CTL clones. On Day 21 the responder cells were cloned by limiting dilution in microwells containing LC-5 cells (10⁴/well) treated as previously described, and with irradiated (150 Gy) LG2-EBV cells (10⁵/well) incubated for 5 d in medium with 3% autologous serum. The incubation medium was Iscove's medium plus AAGM, 3% autologous serum, and IL-2 (50 U/ml). CTL clone 1/7 was later restimulated with LC-5 and LG2-EBV cells cultured as described in medium with FCS and in Iscove's medium plus AAGM and 10% human serum from a pool of blood donors.

CTL clone 110 was derived as follows: CD8⁺ lymphocytes (9 × 10⁵ in a 2-ml well), magnetically sorted from blood mononuclear cells collected from patient LB37 in September 1994, were stimulated with irradiated LC-9.B7 cells (2 × 10⁵) that had been incubated for 1 wk in serum-free medium. The LC-9.B7 cells had been maintained in Iscove's medium plus 10% FCS and AAG, and were seeded in TC-flasks 10 d before being used as stimulator cells. Three days after seeding, the medium was replaced with serum-free X-VIVO 10 medium (BioWhitaker, Walkersville, MD). After a further 4 d, the medium was replaced and IFN- was added at 50 U/ml. Lymphocytes and tumor cells were mixed in Iscove's medium with AAGM, 10% HS, rIL-6 (1,000 U/ml, with 1 U/ml producing 50% maximal proliferation of 7TD1 cells [16]), and rIL-12 (10 ng/ml; specific activity 5 × 10⁶ U/mg; donated by S. Wolf of the Genetics Institute, Cambridge, MA). On Days 7 and 14, responder lymphocytes were restimulated with tumor cells treated as previously described, in fresh medium containing rIL-2 (10 U/ml) and rIL-7 (5 ng/ml; R&D Systems Europe, Abingdon, Oxon, UK). A fraction of the CTL population was frozen at Day 14 of the MLTC. A lysis assay was performed on Day 21. Later, the CTL from Day 14 were thawed, cloned by limiting dilution, and stimulated with irradiated LC-5 cells (10⁴/well, incubated in serum-free medium with INF- as previously described) and irradiated autologous EBV-transformed B cells (10⁵/well), in Iscove's medium plus AAGM and 10% HS with 50 U/ml rIL-2. CTL clone 110 could be maintained in culture with restimulations performed each week in 2 ml wells with 3 × 10⁵ CTL, 10⁵ tumor cells, and 10⁶ feeder cells. After assessment of their specificity, the CTL were restimulated with LC-5 cells that had not been incubated in serum-free medium.

Functional Assays

Cytolytic activities were tested in standard ⁵¹Cr release assays. Production of tumor necrosis factor (TNF) was measured as described (17) with the TNF-sensitive clonal cell line WEHI-164c13 (18).

The antibodies used for analyzing the HLA restriction of the CTL included W6/32 (IgG2a anti-HLA class I [19]), L243 (IgG2a anti-HLA-DR [20]), BB7.2 (IgG2a anti-HLA-A2 [21]), and B1.23.2 anti-HLA-B/C IgG2a (22, 23).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Recognition of Cell Line LB129-SCLC by an Autologous CTL Clone

SCLC cell lines have been shown to be poorly labeled by anti-HLA class I antibodies (24). We previously described three SCLC cell lines that could neither stimulate nor be lysed by allogenic CTL but which became efficient stimulator and target cells for these CTL after an incubation with IFN-, which increased their level of expression of HLA class I molecules (25). The LB129-SCLC cell line was derived from a mediastinal lymph node of patient LB129. These cells were poorly labeled with the anti-HLA-A, B, C monoclonal antibody W6/32 (Figure 2). Incubation of the tumor cells with IFN- increased their expression of HLA class I molecules without significantly affecting their expression of HLA-DR antigens or of the adhesion molecules intercellular adhesion molecule-1 (ICAM-1) and lymphocyte function-associated antigen-3 (LFA-3).

View larger version (26K): [in this window] [in a new window]   Figure 2.   Expression of surface molecules by LB129-SCLC cells. Tumor cells were incubated for 72 h with or without IFN- (50 U/ml), and then incubated with mouse monoclonal antibodies against the indicated surface antigens, washed, and labeled with goat antimouse Ig antibodies coupled to fluorescein (GAM-FITC). The positive control consisted of autologous blood mononuclear cells stimulated with PHA in the presence of IL-2. Shaded curves correspond to cells labeled with GAM-FITC alone.

Peripheral blood mononuclear cells (PBMC) or purified CD8⁺ lymphocytes of patient LB129 were stimulated in medium containing IL-2 and IL-4 with irradiated LB129-SCLC cells that had been incubated with IFN-. The responder lymphocytes were restimulated each week by the addition of tumor cells, IL-2, and IL-4. On Day 28, the effector cells derived from the PBMC lysed LB129-SCLC cells, whether or not the latter had been incubated with IFN-, and they also lysed K562 cells, the prototype target of natural killer (NK)-like lytic effector cells (Figure 3). In contrast, the effector cells derived from the CD8⁺ lymphocytes lysed only the tumor cells that had been incubated with IFN-, and did not lyse K562 cells.

View larger version (22K): [in this window] [in a new window]   Figure 3.   Lytic activity of responder cells obtained in autologous MLTC. PBMC or purified CD8+ blood lymphocytes of patient LB129 were stimulated every week with irradiated LB129-SCLC cells that had been incubated with IFN-. The lytic activity of the population of responder lymphocytes was tested on Day 28 against the autologous LB129-SCLC cells, which had been incubated for 72 h with or without IFN-, and against K562 cells.

The CTL population derived from the CD8⁺ lymphocytes was cloned by limiting dilution, and CTL clone 8/10 was derived. CTL clone 8/10 lysed IFN--treated autologous tumor cells but did not lyse untreated tumor cells, autologous T-lymphocytes, an allogenic IFN--treated SCLC cell line (LB92-SCLC), or K562 (Figure 4). CTL clone 8/10 produced TNF after stimulation with autologous SCLC cells incubated with IFN- (Figure 5A). It is our experience that this production of TNF by CTL clones can be obtained with a lower concentration of antigen than that needed to trigger the clones' lytic activity. We observed that despite their low level of HLA class I molecules, LB129-SCLC cells that had not been incubated with IFN- could stimulate the production of a detectable amount of TNF by CTL clone 8/10 (Figure 5A).

View larger version (14K): [in this window] [in a new window]   Figure 4.   Lytic activity of CTL clone 8/10. 51Cr-labeled target cells included autologous LB129-SCLC cells incubated with or without IFN-, autologous T lymphocytes stimulated with PHA and IL-2, allogenic LB92-SCLC cells treated with IFN-, and NK-target K562 cells.

View larger version (21K): [in this window] [in a new window]   Figure 5.   Specificity of CTL clone 8/10 analyzed in a TNF release assay. (A) Effect of IFN- on the capacity of LB129-SCLC cells to stimulate CTL clone 8/10. The indicated numbers of LB129-SCLC cells, preincubated for 72 h with or without IFN-, were added to CTL clone 8/10 (2,000 cells/well). After 24 h, the amount of TNF produced by the CTL was measured by testing the killing activity of the culture supernatant on TNF-sensitive WEHI-164c13 cells. (B) Inhibition of the stimulation of CTL clone 8/10 by anti-HLA antibodies. IFN--treated LB129-SCLC cells (2,000 cells/well) were incubated for 30 min with monoclonal antibodies recognizing HLA class I (W6/32, IgG2a), HLA-DR (L243, IgG2a), HLA-A2 (BB7.2, IgG2a), and HLA-B/C (B1.23.2, IgG2b) molecules. CTL clone 8/10 (2,000 cells/well) was added and the production of TNF was measured after 24 h.

Recognition of the tumor cells by CTL clone 8/10 was inhibited by the addition of monoclonal anti-HLA-A/B/C or anti-HLA-B/C antibodies, indicating that the antigen recognized by the CTL was presented by HLA-B or HLA-C molecules (Figure 5B).

Recognition of Cell Line LB37-NSCLC by Autologous CTL Clones

The NSCLC cell line LB37-NSCLC was derived from a mediastinal lymph node of patient LB37. Clonal line LB37-NSCLC-5 (LC-5) was derived from LB37-NSCLC cells by limiting dilution. LC-5 cells expressed HLA class I molecules at the same level as did blood mononuclear cells, and incubating the tumor cells with IFN- did not increase this expression. Autologous blood CD8⁺ lymphocytes were stimulated with irradiated LC-5 cells in the presence of IL-2. After 3 wk of stimulation the responder lymphocytes displayed a low lytic activity (< 10% specific lysis) against the LC-5 cells. These lymphocytes were cloned by limiting dilution, and CTL clone 1/7 was derived. This clone lysed LC-5 cells very efficiently, and did not lyse K562 cells, autologous PHA-activated T lymphocytes, or autologous fibroblasts (Figure 6). LC-5 target cells that had been incubated with IFN- over a period of 3 d were lysed slightly better than untreated cells. Lysis of LC-5 cells was inhibited by an anti-HLA-A2 antibody, indicating that CTL clone 1/7 recognizes an antigen presented by HLA-A2 molecules (Figure 7).

View larger version (20K): [in this window] [in a new window]   Figure 6.   Cytolytic activity of antitumor CTL clones derived from NSCLC patient LB37. Target cells included control LC-5 cells (closed symbols), LC-5 cells incubated for 3 d with IFN- (50 U/ml) (open symbols), autologous T lymphocytes activated with PHA and IL-2, autologous fibroblasts, and K562 cells.

View larger version (16K): [in this window] [in a new window]   Figure 7.   Inhibition by anti-HLA antibodies of the lytic activity of CTL clone 1/7. 51Cr-labeled LC-5 cells (1,000 cells/well), which had been treated with IFN-, were incubated for 15 min at 20° C with monoclonal antibodies recognizing HLA class I molecules or HLA-A2 molecules (W6/32 and BB7.2, respectively; 1:30 dilution of ascites fluid from mice inoculated with the hybridoma cells), or recognizing HLA-B and HLA-C molecules (B1.23.2; 1:3 dilution of culture medium of the hybridoma cells). CTL were added and lysis was measured after 4 h.

Another autologous MLTC experiment was set up using as stimulator cells another subclone, which had been derived from a mutagenized population of LB37-NSCLC cells and which had been transfected with a cDNA clone encoding the B7-1 costimulatory molecule. Autologous CD8⁺ lymphocytes were stimulated each week with these LC9.B7 cells and IL-2, and the lytic activity of the responder population was tested on Day 21 on the autologous LC-5 cell line. The lymphocytes lysed these cells, but also lysed the NK target cells K562 (Figure 8). The addition of a 50-fold excess of unlabeled K562 cells completely abolished the lysis of K562 and partly inhibited the lysis of LC-5 cells, suggesting that the responder cell population contained both antitumor CTL and NK-like lytic effectors. These antitumor lytic activities were not significantly greater that those obtained previously by using LC-5 cells as stimulator cells. The CTL population was cloned by limiting dilution, and CTL clone 110 was obtained. It lysed LC-5 cells, and did not recognize autologous fibroblasts, autologous T lymphoblasts, or K562 cells (Figure 6). Its recognition of the tumor cells was also restricted by HLA-A2 (data not shown).

View larger version (16K): [in this window] [in a new window]   Figure 8.   Cytolytic activity of responder cells obtained in autologous MLTC after 21 d of stimulation. CD8+ blood lymphocytes of patient LB37 were stimulated each week with IFN--treated LC-9.B7 cells. The cytolytic activity of the responder cells was tested against LC-5 cells that had been incubated with IFN-. Effector and target cells were incubated for 4 h at 37° C in the absence or in the presence of a 50-fold excess of unlabeled K562 cells in order to inhibit lysis by NK-like lytic effector cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The identification of tumor-specific antigens recognized by CTL on melanomas opened the possibility of immunizing melanoma patients whose tumor cells carry the relevant antigens. Tumor regressions have been observed in some patients after injections of a peptide encoded by gene MAGE-3 and presented by HLA-A1 molecules (26). The extension of this specific immunotherapeutic approach to lung cancer patients requires the identification of tumor-specific antigens on lung carcinoma cells. Our initial approach was to test whether or not genes that have been shown to encode antigens on melanomas are also expressed in SCLC or NSCLC cells. We found that the MAGE genes, which encode multiple antigens on melanomas and other types of tumors, are expressed in lung cancers: approximately 45% of surgical samples of NSCLC express genes MAGE-1 or MAGE-3, and 38% of NSCLC cells express genes GAGE-1 or 2 (27, and Weynants, unpublished results).

An alternative strategy is to derive new CTL that specifically recognize lung cancer cells and to then identify the target antigens of these CTL. This may lead to the identification of new antigens that are selectively carried by lung carcinoma cells. There are, however, only a few descriptions of CTL clones that specifically lyse autologous lung cancer cells. This scarcity of CTL that act against NSCLC or SCLC cells is probably due to several factors. First, NSCLC cell lines are difficult to derive. We failed to establish a single cell line from 16 samples of primary NSCLC. In contrast, three of 12 attempts with metastatic tissue proved successful. This greater proportion of success with metastatic tumor samples is in accordance with results reported in the literature (28, 29). SCLC cells, on the contrary, tend to grow well in culture: we obtained 16 cell lines from 18 tumor samples. Culture media supplemented with hormones and growth factors were reported to improve the establishment of lung cancer cell lines (11, 15, 30, 31). In accord with these reports, we observed with the tumor material from patients LB129 and LB37 that these additions to the medium were critical for the establishment of a cell line.

A second factor that may explain the scarcity of CTL clones reported to lyse lung cancer cells is the low level of expression of HLA class I molecules on lung cancer cell lines. This is clearly the case with SCLC cells (24), which carry no detectable HLA class I molecules on their surface and which often express the transporters associated with antigen processing (TAP) genes at a low level (32, 33). We observed that for SCLC cells the expression of HLA class I molecules could be increased by pretreating the cells with IFN- before using them as stimulator or target cells for the CTL (25). An alternative strategy that has been applied is the transfer of an IFN- gene with retroviral vectors (34). Our results obtained by stimulating anti-SCLC CTL clone 8/10 to produce TNF indicated that LB129-SCLC cells were capable of stimulating the CTL without pretreatment with INF-. Although this stimulation did not trigger the lytic activity of the CTL, the results suggest that SCLC cells may be capable of stimulating autologous T lymphocytes in vivo.

We succeeded in obtaining CTL clones against SCLC and NSCLC cells, and it may not be a coincidence that these CTL clones were derived from two patients who at the time of this writing have no evidence of disease at 8 and 10 yr after diagnosis. This clinical evolution of patients LB37 and LB129 is quite unusual: less than 5% of lung cancer patients with a similar tumor extension are found to be alive after 5 yr (7, 8). It is worth noting that with the same experimental approach of autologous MLTC, we failed to obtain CTL that were active against the tumor cell lines of six SCLC patients and one NSCLC patient, none of whom survived more than 2 yr after diagnosis. These results suggest that the CTL described here were relevant to the clinical evolution of the patients. It may be worth noting that patient LB37 received intradermal injections of lethally irradiated autologous tumor cells (Figure 1). Although we have no evidence for a role of this treatment in the favorable clinical course of the patient, it is possible that the vaccinations contributed to priming or to maintaining an antitumor immune response. The generation of CTL clones recognizing lung cancer cells is a first step toward the identification of new tumor antigens, which may be of future use for the specific immunotherapy of lung cancer patients.