We have identified an antigen recognized on a human large cell carcinoma by an autologous tumor-specific CTL clone that was derived from mononuclear cells infiltrating the primary tumor. The antigenic peptide is presented by HLA-A2 molecules and is encoded by the α-actinin-4 gene, which is expressed ubiquitously. In the tumor cells, a point mutation generates an amino-acid change that is essential for recognition by the CTLs. The mutation was not found in α-actinin-4 cDNA sequences from about 50 lung carcinoma cell lines, suggesting that it is unique to this patient. Although he did not receive chemotherapy or radiotherapy, the patient has been without evidence of tumor since the resection of the primary lesion in 1996. Using tetramers of soluble HLA-A2 molecules loaded with the mutated antigenic peptide, anti-α-actinin-4 CTLs could be derived from blood samples collected from the patient in 1998 and 2000. It is possible that these CTLs, recognizing a truly tumor-specific antigen, play a role in the clinical evolution of this lung cancer patient.
A large number of antigens recognized by CTL derived from blood lymphocytes or TILs4 were identified over the last 10 years (1). Some of these antigens, such as those encoded by the MAGE-A genes, are strictly tumor-specific and shared by many tumors and, therefore, constitute promising targets for anticancer immunization. Antigenic peptide MAGE-A3.A1, encoded by gene MAGE-A3 and presented to CTL on HLA-A1 molecules, has been used to vaccinate tumor-bearing metastatic melanoma patients. Tumor regressions were observed in a significant proportion (7 of 25) of the patients (2). Regressions of metastases were also observed in melanoma patients who received injections of autologous dendritic cells incubated with the MAGE-A3.A1 peptide and a recall antigen (3). Although the mechanism underlying these tumor regressions is still unclear, these results suggest that defined tumor-specific antigens recognized by CTL may be used in therapeutic vaccines, without significant toxicity.
The majority of tumor antigens have been identified with CTL clones that recognized autologous melanoma cells. Much less is known about tumor-specific antigens recognized by CTL on lung tumors. Most of the latter belong to the categories of NSCLC or SCLC. NSCLC includes squamous carcinomas, adenocarcinomas, and large cell carcinomas and represents 80% of lung cancers. These tumors are often infiltrated by T cells, most of which are T-cell receptor α/β+, CD8+, CD28− (4).5 The role and target antigens of these T cells are not known. NSCLC cell lines are difficult to establish, but a few of them could be used to stimulate autologous lymphocytes and derive tumor-specific CTLs (5, 6, 7). Two tumor-specific antigens were identified. One of them is encoded by HER2/neu, which is overexpressed in many tumors (8). The antigenic peptide, presented by HLA-A2 molecules, was initially identified with CTLs derived from TILs. The other antigenic peptide that was found to be recognized by CTLs on NSCLC is encoded by a mutated elongation factor 2 gene and presented by HLA-A68.2 molecules (9). Recently, two antigenic peptides presented by HLA-A24 molecules and reported to be recognized by tumor-specific CTLs were found to be encoded by the cyclophilin B gene, which is expressed ubiquitously (10), and peptides derived from an endoplasmic reticulum-resident protein were shown to be presented on HLA-A24 molecules to CTLs derived from TILs (11).
We have derived antitumor CTL clones by stimulating lymphocytes infiltrating the large cell carcinoma of a patient named Heu (4) with the autologous tumor cell line. In this study, we report on the identification of an antigen recognized by one of these CTL clones.
MATERIALS AND METHODS
Patient Characteristics and Clinical Course.
Patient Heu, a 65-year old Caucasian man, presented in 1996 with an undifferentiated large cell carcinoma of the left lung. No other tumoral localizations were found, and the primary tumor (pT2; N0) was resected. The tumor lesion was massively necrotic and infiltrated with T lymphocytes (4). The patient received no other treatment and has been without evidence of cancer since then. At the time of surgery and 1, 2.5, 3, and 4 years later, blood samples were collected, and mononuclear cells were isolated and frozen.
Derivation and Culture of CTL Clone Heu127.
NSCLC cell line IGR-Heu was derived from a biopsy of the large cell lung carcinoma of patient Heu (HLA-A2, A68, B7, B35, C4, C7; Ref. 12). IGR-Heu cells were maintained in DMEM:Ham’s F-12 medium supplemented with 10% FCS (Seromed, Berlin, Germany). CTL clone Heu127 was derived as described (4). Briefly, a fresh tumor sample from patient Heu was dissociated in DMEM medium containing 1 mm HEPES, 0.3 units/ml DNase, 0.5 units/ml collagenase, and 0.28 units/ml hyaluronidase (Life Technologies, Inc., Cergy Pontoise, France), and the resulting cell suspension was frozen. After thawing, viable TILs and tumor cells were isolated using Ficoll-Hypaque (Pharmacia Fine Chemicals, Uppsala, Sweden) density gradient centrifugation. The lymphocytes were seeded at 104 cells/microwell and stimulated by the addition of irradiated (10,000 rads) autologous tumor cells (3.103/well) and irradiated autologous EBV-transformed B cells (Heu-EBV, 4.104/well) in RPMI 1640 medium supplemented with 10% human AB serum (Institut Jacques Boy, Reims, France) and rIL-2 (20 units/ml; Roussel-Uclaf, Romainville, France). Cells were fed every 3 days with medium and IL-2 and restimulated every other week with irradiated IGR-Heu cells and irradiated Heu-EBV cells. After 3 weeks, the resulting cell line was cloned by limiting dilution (0.5 cell/well) in 96-well V-shaped microtiter plates (Nunc, Roskilde, Denmark) in the presence of irradiated IGR-Heu (3.103/well) and Heu-EBV (4.104/well) cells, rIL-2 (100 units/ml), and 3% of conditioned medium from phytohemagglutinin-activated lymphocytes. CTL clone Heu127 was restimulated every other week with this same protocol.
Lytic activity was measured with a conventional 51Cr-release assay of 4 h using 1000 target cells/well. Inhibition of lysis by anti-HLA-A2 antibody MA2.1 (American Type Culture Collection) was tested by preincubating target cells for 2 h at 37°C with a saturating amount of ascitic fluid from mice inoculated with the hybridoma cells. TNF-α was detected by measuring the cytotoxicity of the culture medium on the TNF-sensitive WEHI-164c13 cells (13) with a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide colorimetric assay (14).
Construction and Screening of the cDNA Library.
About 3 μg of poly(A)+ RNA extracted from IGR-Heu cells with the Fastrack kit (Invitrogen, Groningen, the Netherlands) was converted to cDNA with the Superscript Choice System (Life Technologies, Inc., Gaithersburg, MD) using the oligodeoxythymidylate primer 5′-ATAAGAATGCGGCCGCTAAACTA(T)18VN (V = G, A, or C; N = G, A, T, or C), which contains a NotI site at its 5′ end. The degenerate 3′ end of the oligonucleotide favors annealing at the 5′ end of the poly(A) tail of the mRNA molecules, reducing the proportion of cDNA clones containing long poly(A) sequences, which are difficult to sequence. The cDNA was ligated to HindIII-EcoRI adapters (Stratagene, La Jolla, CA), phosphorylated, digested with NotI, and ligated to plasmid pCEP4 (Invitrogen), which had been digested previously with HindIII and NotI and dephosphorylated. Recombinant plasmids were electroporated into Escherichia coli DH5α and selected with ampicillin (50 μg/ml). The library was divided into pools of approximately 100 cDNA clones. Each pool was amplified, and plasmid DNA was extracted using the QIAprep8 plasmid kit (Qiagen, Hilden, Germany). Duplicate microcultures of 293-EBNA cells (Invitrogen), plated in flat-bottomed 96 microwells (5–7 × 104/well) 24 h before transfection, were cotransfected with about 100 ng of plasmid DNA of each pool of the cDNA library, 50 ng of expression vector pcDNA1/Amp (Invitrogen) containing an HLA-A*0201 cDNA clone, and 1.3 μl of LipofectAMINE reagent (Life Technologies, Inc.). After 24 h, CTL clone Heu127 (3000 cells/well) was added. After another 24 h, half of the medium was collected, and its TNF content measured with the WEHI-164c13 cells.
Sequence Analysis and Localization of the Antigenic Peptide.
cDNA clone 57 was sequenced using dideoxy chain method in an ABI 310 automated DNA sequencer (Perkin-Elmer Applied Biosystems, Warrington, United Kingdom).6 Sequence alignments were performed with Geneworks software (Intelligenetics, Mountain View, CA). To identify the antigenic peptide, two cDNA fragments were amplified from cDNA clone 57 with PCR using forward primer OPC1059 (5′-AAGATGAGAGTGCACAAAATCAAC), which contains a Kozak ATG in frame with the longest open reading frame of the cDNA, and reverse primers OPC1061 (5′-GAGCCCTTCCTTGGCCGA) and OPC1062 (5′-GCGGTCCTGGTGCGCA). PCR conditions were 3 min at 94°C followed by 30 cycles consisting of 1 min at 94°C, 2 min at 61°C (OPC1061) or 67°C (OPC1062), and 3 min at 72°C, followed by a final elongation step of 15 min at 72°C. PCR products were purified using QIAEX II agarose 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 above, with the HLA-A2 cDNA clone into 293-EBNA cells.
Screening for α-Actinin-4 Mutations.
Total RNA extracted from 56 tumor cell lines (NSCLC, SCLC, melanoma, and thyroid) was converted to cDNA with the M-MLV reverse transcriptase (Boehringer Mannheim, Mannheim, Germany) using an oligodeoxythymidylate primer. The cDNA were used as templates for PCR amplification using primers OPC1216 (5′-ATGGGCGACTACATGGCC) and OPC1217 (5′-CGGTTGGCGGCAGTTTCA) for 30 cycles of 1 min at 94°C, 2 min at 63°C, and 2 min at 72°C. PCR products were purified using QIAquick PCR purification Kit (Qiagen) and sequenced.
An HLA-A*0201 cDNA clone was used as a template to amplify the sequence coding for the extracellular domains (amino acids 1–276 of the mature protein) of the HLA-A*0201 heavy chain with primers A1M8 (5′-AAGAAGGAGATATACCATGGGtTCaCACagtATGcgcTATTTtTTtACATCCGTGTCCCGG) and A2b (5′-ATGATGCAGGGATCCTTCGAAGATGTCGTTCAGACCACCACCCGGCTCCCATCTCAGGGTG). A1M8 contains several base changes (small letters) designed to optimize protein expression in øutl E. coli BL21(DE3)pLysS. The PCR product was digested with NcoI and SfuI and cloned into a vector derived from pET3D (Stratagene) and containing a BirA biotinylation site in frame with the 3′ end of the HLA sequence. Recombinant HLA-A*0201 molecules were folded in vitro with β2-microglobulin (produced from pHN1-β2m, kindly provided by P. Moss, Oxford University) and the mutated α-actinin-4 peptide, as described (15). Soluble complexes purified by gel filtration were biotinylated using the BirA enzyme (Avidity LCC, Denver, CO). Phycoerythrin-labeled tetramers were produced by mixing the biotinylated complexes with extravidin-phycoerythrin (Sigma Chemical Co., St. Louis, MO). For tetramer-staining experiments, PBMC (1–2 × 106) were thawed, an aliquot was taken for staining, and the remaining cells were stimulated as follows. Cells were incubated over 1 h at ambient temperature with the mutated α-actinin-4 peptide (20 μm) in Iscove’s medium containing 1% HS, washed, and cultured at about 106 cells/1 ml wells in Iscove’s medium containing 10% HS, IL-2 (20 units/ml), IL-4 (10 ng/ml), and IL-7 (10 ng/ml). After 7 days, an aliquot of the responding cells was used for tetramer staining, and the remaining cells were restimulated and cultured as above. After an additional 7 days, the cells were collected, washed, and stained. Positive cells were sorted and seeded at 1 cell/well in round-bottomed microplates using a FACS-VANTAGE (Becton Dickinson, San Jose, CA). The lymphocytes were restimulated by the addition of irradiated allogeneic HLA-A2-matched EBV-B cells (2 × 104/well), incubated over 1 h at ambient temperature with antigenic peptide (20 μm) and washed, 100 units/ml IL-2, 10 ng/ml IL-4, 10 ng/ml IL-7, and irradiated feeder cells consisting of allogeneic PBMC (8 × 104/well). The same protocol of stimulation was applied after 1 week of culture, and 7–15 days later, clonal populations were labeled with tetramer. Some of them, chosen randomly, were used in a stimulation assay (production of TNF) with an HLA-A2 melanoma cell line (MZ-2-MEL cells transfected with an HLA-A2 construct) incubated with or without 1 nm of the mutated α-actinin-4 peptide. For the tetramer labeling, cells were resuspended in PBS + 1% HS and incubated for 15 min at ambient temperature with tetramer coupled to phycoerythrin. Antibodies (anti-CD3 coupled to FITC and anti-CD8 coupled to PerCP, Becton Dickinson) were then added for an additional 15 min at 4°C.
Assessment of Peptide/HLA-A*0201 Complex Stability.
T2 cells (106/ml) were incubated overnight at 37°C with 100 μm of each peptide in serum-free RPMI 1640 medium supplemented with 100 ng/ml of β2m. Cells were then washed four times to remove free peptides, incubated with brefeldin A (10 μg/ml) for 1 h to block cell surface expression of newly synthesized HLA-A*0201 molecules, washed, and incubated at 37°C for 0, 2, 4, and 6 h. Subsequently, the cells were stained with mAb BB7.2 followed by FITC-conjugated goat-antimouse immunoglobulin mAb. For each time point, the expression of HLA-A*0201 induced by the peptide was calculated as mean fluorescence of cells preincubated with peptide − mean fluorescence of cells treated in similar conditions but in the absence of peptide.
A CTL Clone Recognizing Autologous Lung Tumor Cells.
The NSCLC cell line IGR-Heu was derived from a tumor resected from patient Heu in April 1996. Mononuclear cells that infiltrated the tumor were isolated and stimulated on day 0 and 14 with irradiated IGR-Heu cells, irradiated autologous EBV-transformed B cells, and IL-2. On day 21, the responder lymphocytes were cloned by limiting dilution. A panel of tumor-specific CTL clones was obtained, which could be classified into three groups on the basis of the sequence of their T-cell receptor β chain (4). CTL clone Heu127 expressed a Vβ22-Jβ1.4 rearrangement. It lysed the autologous tumor cells IGR-Heu but not autologous EBV-B cells or the NK-target cells K562 (Fig. 1 A). Patient Heu was typed HLA-A2 by serology, and the recognition of IGR-Heu cells by CTL Heu127 was inhibited by the anti-HLA-A2 mAb MA2.1 (data not shown).
Identification of the Gene Encoding the Antigen.
A cDNA library prepared with poly(A)+ RNA extracted from the IGR-Heu tumor cells was cloned into expression vector pCEP4. The library was divided into 350 pools of approximately 100 recombinant clones, and DNA was prepared from each pool. 293-EBNA cells were cotransfected with DNA from each pool and with an HLA-A*0201 construct. CTL clone Heu127 was added to the transfectants after 24 h. After another 24 h, the supernatant was collected, and its TNF content was measured with the TNF-sensitive WEHI-164c13 cells. Only one pool of cDNA proved positive. It was subcloned, and a cDNA clone, named 57, was isolated (Fig. 1 B). cDNA 57 was 3161-bp long and contained a poly(A) tail. Its sequence corresponded to that of gene ACTN4, which codes for α-actinin-4, one of the four human α-actinin products described thus far (16). cDNA 57 was incomplete and contained an open reading frame coding for residues 58–884 of the protein, which contains 884 amino acids.
Presence of a Point Mutation in the cDNA.
The sequence of the protein encoded by cDNA 57 matched exactly that of the α-actinin-4 proteins present in databanks (D89980, NM004924), except for one amino acid change. An adenine → thymine substitution at position 126 of the cDNA modifies codon AAA (lysine), present in recorded sequences, into AAT (asparagine) in cDNA 57 (Fig. 2). To demonstrate that this difference in the sequence resulted from a mutation in the tumor cells, RNA extracted from IGR-Heu cells and from autologous Heu-EBV cells was converted into cDNA. A fragment was amplified by PCR, using primers OPC1216 and OPC1217 shown on Fig. 2, and sequenced. The sequence derived from the tumor cells corresponded to that of cDNA 57, with T at position 126 of the cDNA, whereas that derived from the EBV-B cells contained A at that position. These results indicated that the ACTN4 gene was mutated in the IGR-Heu tumor cell line.
The same 758-bp fragment of the ACTN4 message was amplified with PCR on cDNA extracted from tumor cell lines (26 NSCLC, 23 SCLC, 3 thyroid carcinoma, and 4 melanoma), and the amplified products were sequenced. All of the sequences were identical to the wild-type sequence, indicating that this region of the ACTN4 gene is not mutated frequently in tumors.
Identification of the Antigenic Peptide.
The mutated asparagine residue, at position 95 in the complete α-actinin-4 protein, is part of a ten amino acids peptide, FIASNGVKLV, which contains the HLA-A2-binding motif: leucine, isoleucine, or methionine in position 2 and leucine or valine in position 10. This peptide was used to sensitize autologous EBV-B cells to lysis by CTL clone Heu127. It was recognized with a half-maximal effect at 3 nm (Fig. 2). Peptides with shorter or longer NH2 or COOH termini were not recognized, suggesting that FIASNGVKLV was the optimal antigenic peptide. The normal peptide, with lysine instead of asparagine at position 5, was not recognized, even at 1 μm. Competition assays indicated that the normal and mutated peptides bound to HLA-A2 molecules with similar affinities (Fig. 3), indicating that the asparagine residue is part of the epitope recognized by CTL Heu127.
Analysis of Anti-α-actinin-4 T Cells with Soluble HLA-peptide Complexes.
Soluble recombinant HLA-A2 molecules were folded with the mutated α-actinin-4 peptide, biotinylated, and multimerized with avidin conjugated to phycoerythrin. The mutated α-actinin-4 tetramers labeled CTL clone Heu127 but not CTL clone F10, which is also restricted by HLA-A2 molecules but recognizes a peptide from the BMLF1 protein of EBV (Fig. 4).
α-Actinin-4-specific T cells were undetectable (<0.01% of CD3+ CD8+ cells) in PBMCs collected from patient Heu in October 1998 and April 2000. These PBMCs were stimulated with the antigenic peptide and IL-2, IL-4, and IL-7. After 1 week, CD8+ α-actinin-4-specific cells could be detected with frequencies of 0.7 and 3.6% of the CD8 cells in blood collected in 1998 and 2000, respectively (Fig. 5). After another restimulation with the antigenic peptide, the proportion of tetramer positive cells increased to 56% of the CD8 cells derived from the 1998 blood collection. To verify the specificity of the staining, the tetramer-positive cells were seeded at one cell/well and stimulated with irradiated HLA-A2 EBV-B cells incubated with the antigenic peptide, irradiated feeder cells, IL-2, IL-4, and IL-7. The CTL clones that were obtained proved to be specific for the mutated α-actinin-4 peptide, as shown in a TNF secretion assay (Fig. 5).
α-Actinins are proteins of about Mr 100,000 encoded by members of the spectrin gene superfamily. They contain an NH2-terminal actin-binding domain, a central rod composed of four spectrin-like repeats (17), and a calmodulin-like domain at the COOH terminus. Organized as antiparallel homodimers, they form a rod-shaped complex with an actin-binding domain at each end (18). Four human α-actinin genes have been described. The ACTN2 and ACTN3 genes are expressed in skeletal muscle (19), and ACTN2 is also expressed in cardiac muscle. The ACTN2 and ACTN3 proteins cross-link actin filaments into tight bundles (18). The ACTN1 gene is expressed ubiquitously (20), and its product interacts with adhesion molecules such as β1 integrins (21, 22), ICAM-1 (23), or the cadherin/catenin complex (24). It seems to play a role in stabilizing cell adhesion and regulating cell shape and cell motility (22). The ACTN4 gene, which is also expressed ubiquitously, was cloned recently by screening a λgt11 cDNA expression library with a new anti-α-actinin mAb, the pattern of staining of which differed from that of an anti-α-actinin-1 antibody (16).
The precise role of α-actinin-4 is still unclear. Mutations in the gene were recently reported to cause familial focal segmental glomerulosclerosis, a common, nonspecific renal lesion characterized by increased urinary protein excretion and decreasing kidney function leading to renal failure. A lysine-to-glutamate substitution at residue 228, a threonine-to-isoleucine substitution at residue 232, and a serine-to-proline substitution at residue 235 were identified in three distinct families affected with an autosomal dominant form of the disease. In vitro, filamentous actin was shown to bind mutated α-actinin-4 more strongly than wild-type α-actinin-4 (25).
Cell adhesion plays a critical role in malignant transformation through dynamic interactions between the extracellular matrix and the actin cytoskeleton (26). Early studies (27) of transformed cells showed a direct association between disorganization of the cytoskeleton and tumorigenicity or invasive capacity. It has been shown that the expression of actin-associated proteins such as α-actinin-1 was decreased in cancer cells and that increased expression after transfection could reduce or abrogate metastatic potential (28, 29). Along the same line, the human ACTN4 gene was recently reported (30) to suppress tumorigenicity of human neuroblastoma cells, suggesting that it could be a tumor suppressor gene. A high level of ACTN4 expression was found in nontumorigenic neuroblastoma cell variants, but little or no expression was observed in malignant neuroblasts. In addition, transfected clones of highly malignant neuroblastoma stem cells that expressed higher levels of ACTN4 showed decreased or abrogated tumorigenicity, together with decreased anchorage-independent growth ability and decreased expression of the N-myc proto-oncogene (30).
The point mutation that we describe here in the ACTN4 gene is the second mutation that is found in a gene encoding an actin-binding protein and that leads to the expression of a tumor-specific antigen recognized by autologous CTL. A mutation in the gene coding for β-catenin, which binds actin and members of the cadherin family of cell surface adhesion, was shown to encode an antigen that was specifically recognized on a melanoma by CTL derived from TIL (31). Subsequently, abnormally high amounts of β-catenin and missense mutations in the β-catenin gene resulting in stabilization of the protein were identified in several melanoma cell lines (32). Furthermore, mutations in the α-catenin (33) and β-catenin (34) genes have been found in tumor cells and appeared to be associated with a loss of cell adhesiveness. Mutations in actin-binding proteins, including α-actinins, may be involved, therefore, in the development of a malignant phenotype. We do not know whether the α-actinin-4 mutation described in this report affects the function of the protein, but a participation of this mutation in the phenotype of IGR-Heu cells cannot be excluded. However, it should be noted that similar amounts of α-actinin-4 protein were detected in IGR-Heu and Heu-EBV B-cell lines as shown by intracytoplasmic immunofluorescence analysis with the anti-α-actinin-4 mAb NCC-Lu-632, (generously provided by Dr. S. Hirohashi; Ref. 16; data not shown).
In patient Heu, the tumor-specific anti-α-actinin-4 CTLs were found in TILs, where they appeared to be clonally expanded (4), and in blood several years later. It is worth noting that, although the frequency of anti-α-actinin-4 CTLs found in blood samples was low (less than 10−4 among the CD8 cells), these CTLs were repeatedly detectable in the absence of clinical signs of cancer. They may correspond to long-lived memory cells, derived from a much larger pool of CTLs that was present in the tumor and possibly also in blood. Alternatively, they may persist because they are regularly restimulated by micrometastases that remain undetected. An analysis of the phenotype of these circulating CTLs may help to clarify this issue. The presence of CTLs directed against a truly tumor-specific antigen in the blood of a disease-free lung cancer patient 4 years after resection of the tumor is compatible with the hypothesis that these CTLs participate in controlling the tumor.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Supported by Grants from the Institut National de la Santé et de la Recherche Médicale, the Institut Gustave Roussy, the Association de la Recherche contre le Cancer (Grants 9307), the Ligue Nationale Française de Recherche contre le Cancer, GEFLUC (France), Fondation de France, the Fédération Belge contre le Cancer (Belgium), the Fonds Jean Maisin (Belgium), and the FB Assurances and VIVA (Belgium). H. E. was supported by a fellowship from the Ligue Nationale Française de Recherche contre le Cancer (comité Val de Marne).
The abbreviations used are: TIL, tumor-infiltrating lymphocyte; NSCLC, non-small cell lung cancer; IL, interleukin; TNF, tumor necrosis factor; PBMC, peripheral blood mononuclear cell; HS, human serum; mAb, monoclonal antibody; poly(A), polyadenylate.
H. Echchakir, I. Vergnon, F., and Mami-Chouaib, unpublished data.
A computer search for sequence homology was performed using programs available at http://www.ncbi.nlm.nih.gov/blast/blast.cgi.
We thank Dr. D. Grunenwald from the Thoracic Department (Institut Montsouris, Paris, France) and A. Scardino for his help with the peptide-binding assay.