Portal de Periódicos da CAPES

Direct Evidence to Support the Role of Antigen-Specific CD8+ T Cells in Melanoma-Associated Vitiligo

2001; Elsevier BV; Volume: 117; Issue: 6 Linguagem: Inglês

10.1046/j.0022-202x.2001.01605.x

1523-1747

Frédérique‐Anne Le Gal, Philippe Lefèbvre, Jean‐Christophe Deschemin, Muriel Andrieu, Jean-Gérard Guillet, Marie‐Françoise Avril, Jacques Bosq, Martin-Xavier Dore,

Immunotherapy and Immune Responses

Vitiligo is a cutaneous pigmentary disorder characterized by the loss of melanocytes. An autoimmune mechanism is strongly suspected to be involved in this affection given that it is frequently associated with autoimmune hormonal disorders, and because antibodies directed against melanocytic antigens are found in the serum of patients with vitiligo. We examined the role of cellular immunity in melanoma-associated vitiligo by expanding infiltrating lymphocytes from fresh biopsy specimens of vitiligo patches in melanoma patients. The vitiligo-infiltrating lymphocytes were almost exclusively T lymphocytes, and most were CD8+. Following in vitro expansion, vitiligo-infiltrating lymphocytes remained predominantly CD8+ and expressed the cutaneous homing receptor CLA. Furthermore, vitiligo-infiltrating lymphocytes had a clonal or oligoclonal T cell receptor profile, possibly reflecting specific antigenic stimulation. Finally, vitiligo- infiltrating lymphocytes specifically recognized differentiation antigens shared by normal melanocytes and melanoma cells. This direct demonstration of CD8+ T cell involvement in vitiligo suggests that, in melanoma patients, vitiligo may be a visible effect of a spontaneous antitumoral immune response. Vitiligo is a cutaneous pigmentary disorder characterized by the loss of melanocytes. An autoimmune mechanism is strongly suspected to be involved in this affection given that it is frequently associated with autoimmune hormonal disorders, and because antibodies directed against melanocytic antigens are found in the serum of patients with vitiligo. We examined the role of cellular immunity in melanoma-associated vitiligo by expanding infiltrating lymphocytes from fresh biopsy specimens of vitiligo patches in melanoma patients. The vitiligo-infiltrating lymphocytes were almost exclusively T lymphocytes, and most were CD8+. Following in vitro expansion, vitiligo-infiltrating lymphocytes remained predominantly CD8+ and expressed the cutaneous homing receptor CLA. Furthermore, vitiligo-infiltrating lymphocytes had a clonal or oligoclonal T cell receptor profile, possibly reflecting specific antigenic stimulation. Finally, vitiligo- infiltrating lymphocytes specifically recognized differentiation antigens shared by normal melanocytes and melanoma cells. This direct demonstration of CD8+ T cell involvement in vitiligo suggests that, in melanoma patients, vitiligo may be a visible effect of a spontaneous antitumoral immune response. vitiligo-infiltrating lymphocyte Vitiligo is a skin disorder characterized by patchy depigmented macules resulting from a loss of melanocytes. Pathophysiologic hypotheses to explain this pathology include autocytotoxic, neural, and immunologic mechanisms. The association of vitiligo with autoimmune diseases, as well as the presence of antibodies directed against melanocyte antigens involved in melanin synthesis (Naughton et al., 1983Naughton G. Eisinger M. Bystryn J.C. Antibodies to normal human melanocytes in vitiligo.J Exp Med. 1983; 158: 246-251Crossref PubMed Scopus (191) Google Scholar;Cui et al., 1992Cui J. Harning R. Henn M. et al.Identification of pigment cell antigens defined by vitiligo antibodies.J Invest Dermatol. 1992; 98: 162-165Crossref PubMed Scopus (100) Google Scholar;Song et al., 1994Song Y.H. Connor E. Li Y. et al.The role of tyrosinase in autoimmune vitiligo.Lancet. 1994; 344: 1049-1052Abstract PubMed Scopus (161) Google Scholar), favor an immunologic mechanism. Moreover, the recent observation of the adoptive transfer of vitiligo following allogeneic bone marrow transplantation for non-Hodgkin's lymphoma represents a new argument for an immunologic origin of vitiligo (Neumeister et al., 2000Neumeister P. Strunk D. Apfelbeck U. et al.Adoptive transfer of vitiligo after allogeneic bone marrow transplantation for non-Hodgkin's lymphoma.Lancet. 2000; 355: 1334-1335Abstract Full Text Full Text PDF PubMed Scopus (44) Google Scholar). The statistically significant association between vitiligo and melanoma is generally considered as a sign of good prognosis. This observation suggests that vitiligo might result from an antitumoral response directed against differentiation antigens shared by normal melanocytes and melanoma cells. The particularly high frequency of vitiligo in melanoma patients treated with recombinant cytokines (Le Gal et al., 1996Le Gal F.A. Paul C. Chemaly P. et al.More on cutaneous reactions to recombinant cytokine therapy.J Am Acad Dermatol. 1996; 35: 650Abstract Full Text PDF PubMed Scopus (11) Google Scholar;Rosenberg and White, 1996Rosenberg S.A. White D.E. Vitiligo in patients with melanoma: normal tissue antigens can be targets for cancer immunotherapy.J Immunother Emphasis Tumor Immunol. 1996; 19: 81-84Crossref PubMed Scopus (292) Google Scholar) further supports this hypothesis, and is indicative of the involvement of cellular immune effectors. Although recent studies have highlighted intralesional lymphocytic infiltrates (Badri et al., 1993Badri A.M. Todd P.M. Garioch J.J. et al.An imunohistological study of cutaneous lymphocytes in vitiligo.J Pathol. 1993; 170: 149-155Crossref PubMed Scopus (131) Google Scholar;Le Poole, 1996Le Poole I.C. et al.Presence of T cells and macrophages in inflammatory vitiligo skin parallels melanocyte disappearance.Am J Pathol. 1996; 148: 1219-1228PubMed Google Scholar) and melanocyte-antigen-specific circulating CD8+ T cells in nonmelanoma vitiligo patients (Ogg et al., 1998Ogg G.S. Dunbar P.R. Romero P. et al.High frequency of skin-homing melanocyte-specific cytotoxic T lymphocytes in autoimmune vitiligo.J Exp Med. 1998; 188: 1203-1208Crossref PubMed Scopus (365) Google Scholar), the functional role of vitiligo-infiltrating T lymphocytes (VIL) in melanocyte destruction has not yet been directly demonstrated. We therefore determined the phenotype and antigenic specificity of VIL expanded from depigmented skin samples of patients with melanoma. We prospectively included in this study five patients with histologically documented melanoma, who developed vitiligo at various time points in the course of the malignancy. Following the approval of this research project by the local ethics committee, written informed consent was obtained from each patient. Patient 1 (HLA-A2/-A68, -B18/-B35), a 32-y-old male, presented with a metastatic melanoma. He developed vitiligo at the end of fotemustine therapy, which led to the complete regression of all identified metastases. Patient 2 (HLA-A3/-A24, -B7/-B18), a 63-y-old male, presented with a vitiligo that developed 2 y after surgical resection of a primary melanoma in 1985. The malignancy has not relapsed. In patient 3 (HLA-A3/-A29, -B7/-B44), a 62-y-old male, the diagnosis of melanoma was based on lymph node metastases from an unidentified primary tumor. Vitiligo occurred about a year prior to the diagnosis, possibly coinciding with regression of the primary tumor. The depigmentation increased during interferon-α (IFN-α) therapy. Patient 4 (HLA-A2/-A29, -B44/-), a 55-y-old male, presented a metastatic melanoma. Vitiligo developed after 1 y of cisplatin/interleukin-2 (IL-2) therapy, which resulted in stabilization of the metastatic disease. In patient 5 (HLA-A1/-A26, -B44/-), a 73-y-old male, the diagnosis of melanoma was based on a colon metastase from an unidentified primary tumor. Vitiligo developed about 5 y prior to the diagnosis and spread until 6 mo before this study. Skin biopsy specimens were taken from the inner border of vitiligo patches for histologic analysis and expansion of VIL. Likewise, 20 ml of blood were withdrawn from each patient and anticoagulated with lithium heparinate. Control biopsies from other skin conditions were taken from a 21-y-old male patient (HLA-A1/-A33, -B14/-) with diffuse lichen planus and a 25-y-old male patient (HLA-A3/-B27/-) with a recent eruption of secondary syphilis. HLA class I phenotyping was performed on fresh peripheral blood mononuclear cells (PBMC) using serologic methods (One Lambda, Canoga Park, CA). Thirteen melanoma-associated antigenic epitopes known to be presented by the patients' HLA class I molecules were synthesized and characterized by Neosystem (Strasbourg, France): tyrosinase 1–9 (HLA-A2) (Wölfel et al., 1994Wölfel T. Van Pel A. Brichard V. et al.Two tyrosinase nonapeptides recognized on HLA-A2 melanomas by autologous cytolytic T lymphocytes.Eur J Immunol. 1994; 24: 759-764Crossref PubMed Scopus (379) Google Scholar), tyrosinase 368–376 (HLA-A2, with a post-translational conversion of asparagine to aspartic acid at position 370) (Skipper et al., 1996aSkipper J.C.A. Hendrickson R.C. Gulden P.H. et al.An HLA-A2-restricted tyrosinase antigen on melanoma cells results from posttranslational modification and suggests a novel pathway for processing of membrane proteins.J Exp Med. 1996; 183: 527-534Crossref PubMed Scopus (368) Google Scholar), tyrosinase 206–214 (HLA-A24) (Kang et al., 1995Kang X. Kawakami Y. El-Gamil M. et al.Identification of a tyrosinase epitope recognized by HLA-A24-restricted, tumor infiltrating lymphocytes.J Immunol. 1995; 155: 1343-1348PubMed Google Scholar), tyrosinase 192–200 (HLA-B44) (Brichard et al., 1996Brichard V. Herman J. Van Pel A. et al.A tyrosinase nonapeptide presented by HLA-B44 is recognized on a human melanoma by autologous cytolytic T lymphocytes.Eur J Immunol. 1996; 26: 224-230Crossref PubMed Scopus (115) Google Scholar), tyrosinase 243–251 and its two modified counterparts C244A and C244S (HLA-A1) (Kittlesen et al., 1998Kittlesen D.J. Thompson L.W. Gulden P.H. et al.Human melanoma patients recognize an HLA-A1-restricted CTL epitope from tyrosinase containing two cysteine residues: implication for tumor vacine development.J Immunol. 1998; 160: 2099-2106PubMed Google Scholar), gp100 17–25 (HLA-A3) (Skipper et al., 1996bSkipper J.C. Kittlesen D.J. Hendrickson R.C. et al.Shared epitopes for HLA-A3-restricted melanoma-reactive human CTL include a naturally processed epitope from pMel-17/gp100.J Immunol. 1996; 157: 5027-5033PubMed Google Scholar), gp100 87–95 (HLA-A3) (Kawashima et al., 1998Kawashima I. Tsai V. Southwood S. et al.identification of gp100-derived, melanoma-specific cytotoxic T-lymphocyte epitopes restricted by HLA-A3 supertype molecules by primary in vitro immunization with peptide-pulsed dendritic cells.International J Cancer. 1998; 78: 518-524Crossref PubMed Scopus (40) Google Scholar), gp100 209–217 T210M (HLA-A2) (Parkhurst et al., 1996Parkhurst M.R. Salgaller M.L. Southwood S. et al.Improved induction of melanoma-reactive CTL with peptides from the melanoma-antigen gp100 modified at HLA*0201-binding sites.J Immunol. 1996; 157: 2539-2548PubMed Google Scholar), gp100 280–288 A288V (HLA-A2) (Parkhurst et al., 1996Parkhurst M.R. Salgaller M.L. Southwood S. et al.Improved induction of melanoma-reactive CTL with peptides from the melanoma-antigen gp100 modified at HLA*0201-binding sites.J Immunol. 1996; 157: 2539-2548PubMed Google Scholar), Melan-A/MART-1 26–35 A27L (HLA-A2) (Valmori et al., 1998Valmori D. Fonteneau J.F. Lizana C.M. et al.Enhanced generation of specific tumor-reactive CTL in vitro by selected melan-A/MART-1 immunodominant peptide analogues.J Immunol. 1998; 160: 1750-1758PubMed Google Scholar), Melan-A/MART-1 32–40 (HLA-A2) (Castelli et al., 1995Castelli C. Storkus W.J. Mauerer M.J. et al.Mass spectrometric identification of a naturally processed melanoma peptide recognized by CD8+ cytotoxic T lymphocytes.J Exp Med. 1995; 181: 363-368Crossref PubMed Scopus (180) Google Scholar). The amino acid sequences of these peptides have been reported in the above-cited references. The control peptide was derived from HIV NEF protein: NEF 186–194, DSRLAFHHV (HLA-B51). The degree of purity of each peptide was above 90%. The lyophilized peptides were diluted to 1 mg per ml in distilled water containing 10% dimethylsulfoxide, and then aliquoted and stored at -20°C. Melanoma cell line BOU (HLA-A2/-A66, -B50/-B41) was obtained by enzymatic digestion of a subcutaneous metastasis. DEL (HLA-A1/-A3, -B35/-B39) and CHO (HLA-A1/-A11, -B42/-B44) were provided by Dr. P. Paul. IGR1/54 was provided by Dr. F. Jotereau (HLA-A2/-A3, -B58/-). These cell lines were grown in RPMI 1640 medium with Glutamax supplemented with 10% fetal bovine serum, nonessential amino acids, HEPES, and antibiotics. BOU, IGR1/54, CHO, and DEL expressed Melan-A/MART-1, gp100, and tyrosinase, confirmed by reverse transcription polymerase chain reaction (RT-PCR) and immunostaining. Biopsy samples were taken from the inner periphery (active margin) of vitiligo patches from five melanoma patients. They were fixed in alcohol-formalin-acetic acid and then embedded in paraffin. Sections 5 μm thick were stained with hematoxylin-eosin-saffron for standard histologic analysis. Immunohistochemical staining was performed on nonadjacent sections after microwave pretreatment, using the following primary antibodies: polyclonal anti-CD3 (Dako-A0452, Glostrup, Denmark), monoclonal anti-CD4 (Novocastra-1F6, Newcastle, U.K.), monoclonal anti-CD8 (Novocastra-AB11), and monoclonal anti-CD56 (Zymed Laboratory-123C3, San Francisco, CA). Anti-CD3 was revealed with the Envision polyclonal system (Dako). The Envision monoclonal system (Dako) was used to reveal anti-CD8 and anti-CD56. Finally, anti-CD4 was revealed using the CSA system (Dako). Biopsy specimens were minced and placed in culture medium (RPMI supplemented with 10% human serum) in six-well plates. IL-2 (100 IU per ml, Boehringer, Mannheim, Germany) was added to the cell cultures once a week. VIL were first harvested within 1–3 wk of culture. The same method was used for the expansion of lymphocytes infiltrating cutaneous lesions of lichen planus and secondary syphilis. PBMC were isolated by density gradient centrifugation through a lymphocyte separation medium (Ficoll-Hypaque, Pharmacia, Uppsala, Sweden). In order to expand peptide-specific CD8+ T cells, fresh unfractionated PBMC were seeded in 24-well plates (4 × 106 per well) in RPMI medium with Glutamax (Gibco Laboratories, Grand Island, NY), supplemented with 10% human serum, nonessential amino acids, HEPES, and antibiotics. To each well were added the various peptides (three to six peptides per well) at a final concentration of 1 µg per ml. IL-2 (10 IU per ml) was added on days 3 and 7. Stimulated PBMC were tested in an IFN-γ Elispot assay on day 12. The IFN-γ Elispot assay was adapted from that ofel Ghazali et al., 1993el Ghazali G.E. Paulie S. Andersson G. et al.Number of interleukin-4- and interferon-gamma-secreting human T cells reactive with tetanus toxoid and the mycobacterial antigen PPD or phytohemagglutinin: distinct response profiles depending on the type of antigen used for activation.Eur J Immunol. 1993; 23: 2740-2745Crossref PubMed Scopus (97) Google Scholar. Briefly, 96-well nitrocellulose plates (Millipore, Bedford, MA) were coated with 1 µg per ml of mouse antihuman IFN-γ capture monoclonal antibody (MoAb) (1-D1K, Mabtech, Nacka, Sweden) diluted with phosphate-buffered saline (PBS) and incubated overnight at 4°C. The wells were washed with PBS-Tween 0.05% and saturated with complete RPMI medium. VIL or stimulated PBMC (4 × 104-105 per well) were plated in duplicate or triplicate with various CD8+ epitopic peptides (1 µg per ml) or with target cells (5 × 103 per well) and incubated for 20 h at 37°C under 5% CO2. The wells were then washed and incubated for 1 h with 100 µl of biotinylated mouse antihuman IFN-γ (1 µg per ml; 7-B-6 Mabtech, Nacka, Sweden). Alkaline-phosphatase-labeled Extravidin (Sigma, St. Louis, MO) was then added for 1 h. Finally, 100 µl of a chromogenic alkaline phosphatase conjugate substrate (Bio-Rad, Hercules, CA) was added to each well to reveal the spots. Finally, spots were counted using a computer-assisted video image analysis (Zeiss, Jena, Germany). For the first flow cytometry analysis (Figure 2a later), anti-CD3-FITC, anti-CD4-PE, anti-CD8-FITC, anti-CD16 + CD56-PE, and anti-CLA HECA-452 (Becton-Dickinson, Le Pont de Claix, France) were used according to the manufacturer's instructions. Fluorescein isothiocyanate rabbit antirat Ig antibody (Dako) was used as second-step reagent for indirect immunofluorescence staining with HECA-452. Fluorescence-activated cell sorter (FACS) analysis was performed using a FACScan apparatus (Becton-Dickinson). For the second flow cytometry analysis (Figure 2b later) the cells were initially surface stained for 30 min at room temperature with anti-CD8-APC (clone SK1, IgG1), anti-CD56-PE (clone NCAM16.2, IgG2b), and anti-CD45-FITC (clone HI100, IgG2b) (Becton-Dickinson). The cells were then washed and resuspended in 1% paraformaldehyde. Samples were evaluated using a FACScalibur flow cytometer (Becton-Dickinson) and data were analyzed with Lysis software (Becton-Dickinson). For each sample 104 events were acquired, gated on CD8 expression with a scatter gate designed to include only viable lymphocytes. Total RNA was obtained from thawed ex vivo PBMC, in vitro cultured PBMC, and VIL, using the method ofChomczynsky and Sacchi, 1987Chomczynsky P. Sacchi N. Single-step method of RNA isolation by acid guanidinium thicyanate-phenol-chloroform extraction.Anal Biochem. 1987; 162: 156-159https://doi.org/10.1006/abio.1987.9999Crossref PubMed Scopus (0) Google Scholar (RNAble kit; Eurobio, Paris, France). Pellet Paint was used as a coprecipitant during isopropanol precipitation. RNA was resuspended in 30 µl of RNAse-free water and stored at -80°C until use. TCR-BV repertoire analysis was performed as previously described (Prevost-Blondel et al., 1995Prevost-Blondel A. Chassin D. Zeliszewski D. et al.Preferentiel usage of the T-cell repertoire by influenza virus hemaglutinin-specific human specific CD4+ T lymphocytes: in vitro life span of clonotypic cells.J Virol. 1995; 69: 8046-8050PubMed Google Scholar). Briefly, total RNA was reverse transcribed with random hexamers to which were added oligo(dT) 16 primers and murine leukemia virus reverse transcriptase (Perkin Elmer, Foster City, CA) according to the manufacturer's instructions. cDNAs were PCR-amplified using the 24 Vβ-specific probes and a common Cβ-specific probe labeled with Fam fluorescent dye (Genset, Paris, France). All primers used in this study have been described elsewhere (Martinon et al., 1999Martinon F. Michelet C. Peguillet I. et al.Persistent alterations in T-cell repertoire, cytokine and chemokine receptor gene expression after 1 year of highly reactive antiretroviral therapy.AIDS. 1999; 13: 185-194https://doi.org/10.1097/00002030-199902040-00006Crossref PubMed Scopus (50) Google Scholar). Each Vβ-Cβ PCR product was analyzed by electrophoresis in agarose gel. For the analysis of CDR3 length diversity, PCR products were electrophoresed in an ABI 373 A DNA sequencer (Perkin Elmer). Data were analyzed for fragment size and fluorescence intensity with Immunoscope 3.01d software. Skin biopsy samples were taken from the active margin (inner border) of vitiligo patches of five melanoma patients. All the specimens had the histopathologic features of vitiligo, such as loss of melanocytes in the basal cell layer of the epidermis (revealed by Fontana staining) and presence of an inflammatory infiltrate in the superficial dermis and in the basal layer of the epidermis, composed of small, well-differentiated lymphocytes often located in perivascular and periannexial areas Figure 1. Immunohistochemical analysis of the biopsy specimen from all five patients showed that the infiltrating lymphocytes, particularly in the basal layer of the epidermis, were almost exclusively CD3+, and predominantly CD8+ Figure 1. Very few cells were stained with anti-CD56 (not shown). Cutaneous lymphocytes gradually proliferated and remained in close contact with the skin fragments when minced vitiligo biopsy specimens were placed in culture. No lymphocyte expansion was observed with a normal skin specimen from patient 1 (not shown). Sufficient cells were recovered from vitiligo samples from patient 3 only, in order to determine their phenotype by FACS analysis Figure 2. Following 3 wk of culture, 87% of expanded cells expressed CD3 antigen, showing that they were predominantly T lymphocytes. These VIL were predominantly CD8+. Only 11% of expanded VIL from this patient expressed CD56 and/or CD16 antigens (markers of natural killer cells). Immunohistochemical analysis of the biopsy specimens from the five patients also showed very slight staining by anti-CD56 and anti-CD16 monoclonal antibodies (not shown). To further determine the phenotype of VIL, we performed a triple staining of CD8/CD56 and CD45RA expression and a cytometric analysis Figure 2b. Sixty-four percent of CD8+ VIL were CD45RA–/CD56–, indicating a memory phenotype, and 32% were CD8+/CD45RA+/CD56–, suggesting a naive phenotype (although some memory cells may revert to CD45RAhigh:Bell et al., 1998Bell E.B. Sparshott S.M. Bunce C. CD4+ T-cell memory, CD45R subsets and the persistence of antigen – a unifying concept.Immunol Today. 1998; 19: 60-64Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar). Only 4% of the CD8+ VIL were CD56+ and 1% expressed both CD56 and CD45RA. Half of the CD56+ VIL were CD8–, i.e., natural killer cells (not shown). Altogether, these results suggest that very few in vitro expanded VIL have a CD8+/CD56+ cytolytic phenotype. Almost all (97%) expanded VIL expressed CLA, ruling out expansion of contaminating blood T cells. Indeed, merely 26% of freshly isolated autologous PBMC were found to express CLA (not shown). When expanded VIL were separated from the skin fragments and seeded in microwells, CLA expression fell to 34% after 3 d (not shown). This observation is in agreement with the results fromFuhlbrigge et al., 1997Fuhlbrigge R.C. Kieffer J.D. Armerding D. et al.Cutaneous lymphocyte antigen is a specialized form of PSGL-1 expressed on skin-homing T cells.Nature. 1997; 389: 978-981Crossref PubMed Scopus (452) Google Scholar who used freshly isolated and cultured PBMC. In order to investigate whether VIL were antigenically stimulated, we studied their TCR repertoire. In patient 3, Vβ usage was studied in VIL using RT-PCR (see Materials and Methods). As control, the same experiment was performed using ex vivo (thawed) PBMC and in vitro expanded PBMC cultured in the same conditions as VIL (3 wk with 100 UI per ml IL-2). In vitro expanded PBMC, as well as nonexpanded PBMC, showed an amplification of all Vβ-Cβ families [except for Vβ11, which is associated with natural killer cells (Dellabona et al., 1994Dellabona P. Padovan E. Casorati G. et al.An invariant Vα24-JαQ/Vβ11 T cell receptor is expressed in all individuals by clonally expanded CD4-8-T cells.J Exp Med. 1994; 180: 1171-1176Crossref PubMed Scopus (377) Google Scholar)]. In contrast, not all Vβ families were expressed by VIL, further confirming that expanded cells were not derived from contaminating circulating T cells Figure 3. Moreover, immunoscope analysis of the VIL TCR repertoire showed a clonal or oligoclonal profile of CDR3 length distribution (refer to Figure 3; one representative experiment out of three experiments with the same pattern of responses). In contrast, CDR3 size distribution in autologous ex vivo PBMC was Gaussian. In cultured PBMC, oligoclonal or clonal distributions of CDR3 size were also observed in some Vβ families, probably due to a certain in vitro selection. Nevertheless, although not proving it, the biased TCR repertoire of VIL could suggest their antigenic stimulation. The antigen specificity of VIL was assessed by an IFN-γ Elispot assay following 1–3 wk of culture. Unfractionated VIL were tested against peptides derived from the melanocyte-lineage-specific antigens Melan-A/MART-1, gp100, and tyrosinase, depending on the patients' HLA phenotype Figure 4a. An HIV-derived peptide was used as negative control. For each assay, a cut-off value of specific secretion of IFN-γ was determined, based on the number of nonspecific spots obtained after stimulation with the control peptide (cut-off = mean background + 3 SEM). Because PBMC from patients with melanoma can contain melanoma differentiation antigen specific CTL precursors, VIL and peptide-activated PBMC (see Materials and Methods) were compared for their reactivity against the same set of peptides Figure 4a, b. CD8+ VIL recognized peptides derived from pigment cell lineage specific proteins involved in melanin synthesis, such as tyrosinase, Melan-A/MART-1, or gp100. As controls, several samples of skin lesions from other skin conditions were placed in a culture medium. Lymphocytes extracted from lesions of lichen planus and of cutaneous secondary syphilis did not recognize any of the peptides derived from melanocyte antigens Figure 4c. In addition, PBMC and VIL from patient 4 had a similar recognition pattern. In patients 1 and 3, certain peptides were recognized by both VIL and PBMC (MART 26–35 27L in patient 1, Tyr192–200 in patient 3); others were recognized by either PBMC or VIL. In PBMC from patients 2 and 5, a high background noise was observed and no specific response to the tested melanocyte-derived peptides could be detected. To further characterize the CD8+ VIL of patients 1, 3, and 4, we tested their ability to recognize naturally processed antigens on melanoma cells. An IFN-γ Elispot assay was performed with several melanoma cell lines as targets, sharing only one HLA-A class I locus with the patients' VIL, or having no matches (control). VIL from patients 1, 3, and 4 recognized one or several peptides naturally presented by melanoma cells in association with HLA-A2 (patients 1 and 4) or HLA-A3 (patient 3) Figure 5. Weak IFN-γ secretion in response to K562 stimulation, attributable to natural killer activity, was consistent with the very small proportion of CD16+ CD56+ cells identified by immunohistology and FACS analysis. The increased frequency of vitiligo in patients with autoimmune hormonal disorders has long suggested that autoimmunity could be responsible for melanocyte destruction in this setting. Early studies focused on the identification of a humoral response against melanocytic antigens. Since the early 1980s, several authors have reported the detection of antibodies against melanocyte-specific antigens (Naughton et al., 1983Naughton G. Eisinger M. Bystryn J.C. Antibodies to normal human melanocytes in vitiligo.J Exp Med. 1983; 158: 246-251Crossref PubMed Scopus (191) Google Scholar;Cui et al., 1992Cui J. Harning R. Henn M. et al.Identification of pigment cell antigens defined by vitiligo antibodies.J Invest Dermatol. 1992; 98: 162-165Crossref PubMed Scopus (100) Google Scholar), tyrosinase in particular (Song et al., 1994Song Y.H. Connor E. Li Y. et al.The role of tyrosinase in autoimmune vitiligo.Lancet. 1994; 344: 1049-1052Abstract PubMed Scopus (161) Google Scholar), in the serum of patients with vitiligo. The better prognosis of melanoma patients who develop vitiligo (Nordlund et al., 1983Nordlund J.J. Kirkwood J.M. Forget B.M. et al.Vitiligo in patients with metastatic melanoma: a good prognostic sign.J Am Acad Dermatol. 1983; 9: 689-696Abstract Full Text PDF PubMed Scopus (259) Google Scholar;Bystryn et al., 1987Bystryn J.C. Rigel D. Friedman R.J. et al.Prognostic significance of hypopigmentation in malignant melanoma.Arch Dermatol. 1987; 123: 1053-1055Crossref PubMed Scopus (184) Google Scholar;Duhra and Ilchyshyn, 1991Duhra P. Ilchyshyn A. Prolonged survival in metastatic malignant melanoma associated with vitiligo.Clin Exp dermatol. 1991; 16: 303-305Crossref PubMed Scopus (58) Google Scholar) is also indicative of an autoimmune mechanism, in which the immune response against malignant melanocytes would also destroy normal “bystander” melanocytes by targeting antigens shared by the two cell types. Antibodies directed against shared melanocyte antigens have indeed been found in patients with isolated or melanoma-associated vitiligo (Cui and Bystryn, 1995Cui J. Bystryn J.C. Melanoma and vitiligo are associated with antibody responses to similar antigens on pigment cells.Arch Dermatol. 1995; 131: 314-318Crossref PubMed Scopus (86) Google Scholar;Baharav et al., 1996Baharav E. Merimski O. Shoenfeld Y. et al.Tyrosinase as an autoantigen in patients with vitiligo.Clin Exp Immunol. 1996; 105: 84-88Crossref PubMed Scopus (98) Google Scholar;Merimski et al., 1996Merimski O. Baharav E. Shoenfeld Y. et al.Anti-tyrosinase antibodies in malignant melanoma.Cancer Immunol Immunother. 1996; 42: 297-302https://doi.org/10.1007/s002620050286Crossref PubMed Scopus (29) Google Scholar). Little work has been done on the possible role of T cells in vitiligo. Recently, Ogg et al showed that cytolytic CD8+ T cells directed against the melanocytic differentiation antigen derived epitope MART 26–35 were more frequent in peripheral blood of nonmelanoma vitiligo patients than in healthy subjects (Ogg et al., 1998Ogg G.S. Dunbar P.R. Romero P. et al.High frequency of skin-homing melanocyte-specific cytotoxic T lymphocytes in autoimmune vitiligo.J Exp Med. 1998; 188: 1203-1208Crossref PubMed Scopus (365) Google Scholar). Becker et al identified clonally expanded T cells with identical BV regions in melanoma and surrounding vitiligo, i.e., in areas of destruction of both normal and neoplastic cells. This observation strongly suggested a role of T-cell-mediated antitumoral response in the development of melanoma-associated vitiligo (Becker et al., 1999Becker J.C. Guldberg P. Zeuthen J. et al.Accumulation of identical T cells in melanoma and vitiligo-like leukoderma.J Invert Dermatol. 1999; 113: 1033-1038Crossref PubMed Scopus (66) Google Scholar). More recently, Yee et al identified Melan-A/MART-1-specific CD8+ T cells in inflammatory lesions following infusion of Melan-A/MART-1-specific CD8+ T cell clones in melanoma patients (Yee et al., 2000Yee C. Thompson J.A. Roche P. et al.Melanocyte destruction after immunotherapy of melanoma: direct evidence of T-cell-mediated vitiligo.J Exp Med. 2000; 192: 1637-1643Crossref PubMed Scopus (368) Google Scholar). In this study, we addressed the role of CD8+ T cells in melanoma-associated vitiligo using functional tests. For this purpose, we preferred to focus on the local immune response rather than to limit our analysis to the systemic response. Histologic analysis of vitiligo biopsy specimens always showed a lymphocytic infiltrate. This was also previously observed at the active margins of vitiligo lesions (Nordlund and Lerner, 1982Nordlund J.J. Lerner A.B. Vitiligo: it is important.Arch Dermatol. 1982; 118: 5-8Crossref PubMed Scopus (127) Google Scholar;Badri et al., 1993Badri A.M. Todd P.M. Garioch J.J. et al.An imunohistological study of cutaneous lymphocytes in vitiligo.J Pathol. 1993; 170: 149-155Crossref PubMed Scopus (131) Google Scholar). In keeping with the results of previous studies, we observed a strong predominance of CD3+ lymphocytes and an increased CD8/CD4 ratio, especially in the basal layer of the epidermis where melanocyte destruction takes place. These results outline a role of CD8+ T cells in vitiligo (Badri et al., 1993Badri A.M. Todd P.M. Garioch J.J. et al.An imunohistological study of cutaneous lymphocytes in vitiligo.J Pathol. 1993; 170: 149-155Crossref PubMed Scopus (131) Google Scholar;Le Poole, 1996Le Poole I.C. et al.Presence of T cells and macrophages in inflammatory vitiligo skin parallels melanocyte disappearance.Am J Pathol. 1996; 148: 1219-1228PubMed Google Scholar). Following a 3 wk culture, the in vitro expanded VIL were almost exclusively CD8+ T cells. At this time, their CD56/CD45RA expression did not reveal a cytolytic profile, according to the recent work of Pittet et al (Pittet et al., 2000Pittet M.J. Speiser D.E. Valmori D. et al.Cytolytic effector function in human circulating CD8+ T cells closely correlates with CD56 surface expression.J Immunol. 2000; 164: 1148-1152Crossref PubMed Scopus (220) Google Scholar;Pittet et al., 2001Pittet M.J. Zippelius A. Speiser D.E. et al.Ex vivo IFN-γ secretion by circulating CD8 T lymphocytes: implications of a novel approach for T cell monitoring in infectious and malignant diseases.J Exp Med. 2001; 166: 7634-7640Google Scholar), but rather corresponded to memory (64%) and probably naive (32%) populations. We cannot rule out, however, that the CD45RAhigh VIL are memory cells (i.e., CD45RAlow cells) that have switched back to a CD45RAhigh phenotype (Bell et al., 1998Bell E.B. Sparshott S.M. Bunce C. CD4+ T-cell memory, CD45R subsets and the persistence of antigen – a unifying concept.Immunol Today. 1998; 19: 60-64Abstract Full Text Full Text PDF PubMed Scopus (179) Google Scholar). Moreover, these findings concern in vitro cultured cells and may not reflect the in vivo situation. Nevertheless, the scarce presence of CD56+ cells revealed by the immunohistochemistry analysis of the initial biopsies suggests that destruction of melanocytes could occur through a noncytolytic pathway. As long as they were allowed to remain in close contact with the skin fragments in the tissue-culture plates, in vitro expanded VIL expressed the CLA homing molecule. This is in keeping with immunohistologic results ofBadri et al., 1993Badri A.M. Todd P.M. Garioch J.J. et al.An imunohistological study of cutaneous lymphocytes in vitiligo.J Pathol. 1993; 170: 149-155Crossref PubMed Scopus (131) Google Scholar. Expanded T cells were not derived from contaminating blood, as confirmed by the high proportion of VIL expressing CLA and the altered TCR-BV repertoire of expanded VIL. Furthermore, the restricted Vβ usage and the clonal or oligoclonal profile of the CDR3 length distribution in expanded VIL from patient 3 may suggest their activation or recruitment at the lesion site. Nevertheless, as demonstrated byDietrich et al., 1997Dietrich P.Y. Walker P. Schnuriger V. et al.TCR analysis reveals significant repertoire selection during in vitro lymphocyte culture.Int Immunol. 1997; 9: 1073-1083Crossref PubMed Scopus (39) Google Scholar, a repertoire selection may occur during in vitro lymphocyte culture. This was confirmed by the CDR3 size analysis of in vitro expanded PBMC compared to that of noncultured PBMC of patient 3, showing that our culture conditions were, to some extent, responsible for an in vitro selection. It was therefore essential to analyze the antigenic reactivity of VIL by means of indisputable functional tests. IFN-γ Elispot assays showed that VIL specifically secreted IFN-γ in response to several melanocyte-derived antigenic epitopes that were expressed on both normal melanocytes and melanoma cells. Moreover, these VIL recognized antigenic peptides naturally processed by melanoma cells Figure 5. In contrast, T cells expanded from lichen planus and secondary syphilis lesions in the same culture conditions as VIL did not recognize any melanocyte-derived antigens. In two patients (2 and 5), no melanocyte-specific response could be demonstrated in the peripheral blood. Therefore, the response of these patients would have escaped detection by merely analyzing the systemic immune response. Altogether, PBMC and VIL did not present the same pattern of reactivity in four out of five patients Figure 4a, b. This underlines the necessity of analyzing immune responses at the site in which they are likely to occur. To our knowledge, this work is the first to show the direct and functional evidence that CD8+ T cells, directed against several melanocytic antigens, can play a role in vitiligo naturally occurring in melanoma patients. Ogg et al showed the increased frequency of MART 26–35 specific cytolytic CD8+ T cells in peripheral blood of patients having vitiligo in the absence of melanoma (Ogg et al., 1998Ogg G.S. Dunbar P.R. Romero P. et al.High frequency of skin-homing melanocyte-specific cytotoxic T lymphocytes in autoimmune vitiligo.J Exp Med. 1998; 188: 1203-1208Crossref PubMed Scopus (365) Google Scholar). We further observed a specific reactivity toward several other melanocyte-derived antigenic epitopes (five peptides derived from gp100, four peptides derived from Melan-A/MART-1) in the peripheral blood of a 65-y-old patient with “spontaneous” widespread vitiligo (data not shown). These observations suggest that the tumor is not necessary to prime the melanoma-specific T cells and that, in nonmelanoma patients, this autoimmune cellular reactivity has its origins in another phenomenon. Nevertheless, several lines of evidence suggest that vitiligo is probably linked with an antitumoral response in melanoma patients. For example, we have observed a higher frequency of melanoma-associated antigen-specific CD8+ reactivity in the peripheral blood of melanoma patients with vitiligo than in melanoma patients without vitiligo (manuscript in preparation). This observation is in accordance with the frequent occurrence of vitiligo in melanoma patients receiving immunotherapy (Le Gal et al., 1996Le Gal F.A. Paul C. Chemaly P. et al.More on cutaneous reactions to recombinant cytokine therapy.J Am Acad Dermatol. 1996; 35: 650Abstract Full Text PDF PubMed Scopus (11) Google Scholar;Rosenberg and White, 1996Rosenberg S.A. White D.E. Vitiligo in patients with melanoma: normal tissue antigens can be targets for cancer immunotherapy.J Immunother Emphasis Tumor Immunol. 1996; 19: 81-84Crossref PubMed Scopus (292) Google Scholar;Buendia-Eisman et al., 1997Buendia-Eisman A. Serrano-Ortega S. Ortega del Omo R. Vitiligo-like lesions in patients with melanoma treated with immunotherapy.Eur J Dermatol. 1997; 7: 522-524Google Scholar) and also with its possible link with better prognosis (Nordlund et al., 1983Nordlund J.J. Kirkwood J.M. Forget B.M. et al.Vitiligo in patients with metastatic melanoma: a good prognostic sign.J Am Acad Dermatol. 1983; 9: 689-696Abstract Full Text PDF PubMed Scopus (259) Google Scholar;Bystryn et al., 1987Bystryn J.C. Rigel D. Friedman R.J. et al.Prognostic significance of hypopigmentation in malignant melanoma.Arch Dermatol. 1987; 123: 1053-1055Crossref PubMed Scopus (184) Google Scholar). These observations, together with the results of this study, suggest that vitiligo in melanoma patients could be the visible consequence of a cellular antitumoral response, although not predictive of the clinical outcome of this immune reaction. We wish to thank the patients for their generous and enthusiastic collaboration, and the staff of the Gustave Roussy Institute for performing the skin biopsies and blood sampling. We are also grateful to Dr. Valérie Molinier-Frenkel for her help with FACS analysis, and Marylène Garcette for her excellent technical assistance. We also wish to thank David Young, Renaud Fortuner, and Zoé Coutsinos for editing the English text. This work was supported by the ARC (Association pour la Recherche contre le Cancer).