COMMENTARY

In This Issue

James T. ElderUniversity of Michigan Medical Center, Ann Arbor, Michigan, USA

T-CELLRECEPTOR^ANTIGEN INTERACTION DEFINEDIN AUTOIMMUNE VITILIGO

Vitiligo is a common disorder of cutaneous depigmentation,which, though acquired, has a de¢nite genetic predisposition.While the endpoint of vitiligo is clearly loss of melanocytes, thepathomechanism of vitiligo has been a subject of longstandingdebate. Textbooks present the neural hypothesis, the self-destructhypothesis, and the autoimmune hypothesis as possible explana-tions for vitiligo; these mechanisms may not be mutually exclu-sive. It is clear that skin-homing T-cells can be found at themargins of depigmented vitiligo lesions. Autoimmune mechan-isms for attacking the melanocyte include in£ux of skin-homingautoreactive T-cells, aided by cytotoxic antibodies, complementfactors, and nitric oxide (NO).These mechanisms may be furtherabetted by toxic by-products of melanin biochemistry. In this is-sue, Mantovani et al (p. 308) present some exciting new insightsinto the autoimmune hypothesis of vitiligo. Studying a singlesubject over several years, they characterized the T-cell cytotoxicresponse to an antigen called Melan-A/MART-1 (MelanomaAntigen Recognized byT cells-1). Cytotoxic CD8þ T-cells rec-ognize peptide antigens bound to MHC Class I molecules ontarget cells via the T-cell receptor (TCR), which is comprised ofa- and b- chains. Each b-chain is composed of somatically rear-ranged V (variable), D (diversity), and J (joining) germline DNAsegments that confer antigen speci¢city. Many recent studies havefocused on the hypervariable D region of the b chain as the focalpoint of antigen recognition by theTCR. However, recent X-raycrystallography studies have revealed that TCR-a actually con-tacts the peptide antigen more extensively than doesTCR-b, sug-gesting the possibility that peptide recognition may rely stronglyon TCR-a, even though TCR-a does not contain a D region. Inprevious work (Mantovani et al, 2002) these authors found thatHLA-A2-positive subjects markedly over-utilize a particularV re-gion segment of the TCR a chain called AV2 in immatureCD8þ thymocytes capable of recognizing Melan-A peptides.The same phenomenon was true of mature CD8þ cells in theirvitiligo patient: 0.16% of all blood-derived CD8þ T-cells recog-nized a Melan-A peptide, as determined by binding of a tetra-meric MHC-peptide complex. Twenty eight of 29 T-cell clonesderived from this population expressed the AV2 TCR-a V regionsegment. In contrast, these T-cell clones expressed many di¡erentrearrangements of the TCR b chain. Melan-A-recognizing cellsexpressed the skin homing receptor CLA and displayed highavidity for the Melan-A antigen, suggesting that they had beenselected for via repeated antigenic stimulation and could tra⁄cto the skin. They were capable of killing normal melanocytes,and they persisted for more than three years in the patient’s blood.Taken together, these data strongly suggest that these antigen-speci¢c, AV2-restricted but polyclonal T-cells play an importantrole in at least some forms of vitiligo. Presumably, repeated expo-sure of AV2-bearing ab-TCR to phagocytosed melanocytes leadsto clonal expansion of many di¡erent T-cells recognizing various

Melan-A-derived peptides. Even if a toxic process begins thechain of melanocyte destruction, it may be completed by a cyto-toxic T-cell response. Thus, the neural hypothesis, the self-de-struct hypothesis, and the autoimmune hypothesis need not bemutually exclusive.Why should an entire family of TCRs characterized by the

AV2 variable region segment recognize Melan-A? Interestingly,Melan-A displays strong homologies to certain microbial peptides,and it is likely that over evolutionary time, responses to microbeswere critical in shaping theT-cell repertoire. Thus, similarities be-tween bacterial and melanocyte peptides melanocytes may haveplayed an evolutionary role in the emergence of vitiligo.

INNATE IMMUNE RESPONSES TO BACTERIAL DNA INTHE SKIN

Eukaryotic DNAs undergo enzymatic methylation of cytosine atCpG residues, and CpG dinucleotides are relatively uncommonin eukaryotes due to deamination of methyl-cytosine to thymi-dine. Methylation of DNA is correlated with inaccessibility ofDNA in chromatin, and this methylation mechanism may havearisen as a silencing mechanism defense against genomic patho-gens such as retroviruses. Bacteria and viruses lack this methyla-tion mechanism and, as a consequence, are not de¢cient in CpGdinucleotides. Thus, this mechanism may have also have arisen asa way for eukaryotes to mount an immune response against non-self DNA. Bacterial and viral DNAs, and oligonucleotides con-taining CpG residues, are stimulatory to the normal humanimmune system and have been used as adjuvants for immuniza-tion, whereas immune responses to human DNA arise only inpathogenic states such as lupus erythematosus (LE). In this issue,M˛lne, Collins, and Tarkowski (p. 294) injected DNA derivedfrom the common skin pathogen Staphylococcus aureus intomice, and characterized the duration and composition of the re-sulting cutaneous in¢ltrate. They also compared these parametersafter injection of synthetic CpG-containing oligonucleotides (20^24 nucleotides long) linked either by nuclease-sensitive phospho-diester bonds or nuclease-resistant phosphorothioate bonds. Sta-phylococcal DNA and phosphodiester oligos produced a briskbut short-lived in£ammatory response, which peaked at two daysand was characterized by a macrophage-rich in¢ltrate. Phosphor-othioate oligos produced a slower and more pronounced re-sponse, which peaked at 7 days. The response was clearlydependent on the presence of CpG, as substitution of GpC forCpG in either type of oligonucleotide reduced the peak responseby over 80%. By selective depletion of macrophages, polymor-phonuclear leukocyte (PMNs), or T-cells in various ways, theyshowed that the clinical severity of the reaction was moststrongly dependent on macrophages, followed by PMNs and T-cells. Given the rapidity of response and the relatively larger de-pendence on macrophages, this reaction has the hallmarks of aninnate, rather than an acquired immune response. This would be

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

vi

consistent with recent observations that certain members of theToll family of receptors recognize unmethylated CpG residues.These receptors have been strongly implicated in innate immu-nity via activation of NF-kB signaling. Other workers haveshown that CpG oligonucleotides can serve as adjuvants in theskin, at least in part by accelerating the tra⁄cking of Langerhanscells to draining lymph nodes and by encouraging aTh1 response.Thus, one can envision a day in the not-too-distant future inwhich CpG oligonucleotides might be tailored not only to boostimmuneTh1 immune responses in immunocompromised indivi-duals, but also to deviate the immune response fromTh2 toTh1in conditions such as acute £ares of Th2-mediated diseases such asatopic dermatitis, pemphigus, pemphigoid, and cutaneous lupuserythematosus. As ‘‘timing is everything’’ in the immune response,the ability to vary the duration of the innate immune response byaltering the chemical structure of the oligos used may be an im-portant means of ¢ne-tuning cutaneous immunotherapy.

DUST MITES: HOW THEY GET CAN ON YOUR NERVES

Speci¢c allergenic peptides derived from the house dust mitehave been suggested to play a direct pathophysiological role inatopic dermatitis (AD), because many individuals display highlevels of speci¢c IgE directed against this ubiquitous ectoparasite.In this issue of the Journal, Huang et al (p. 289) develop an epicu-

taneous sensitization model of atopic dermatitis in Balb/c miceusing Der p 8, a glutathione-S-transferase. Patches containing al-lergen or saline were applied three times for 4 days each, spaced 17days apart. The animals developed a localized subacute dermatitiswith both spongiosis and epidermal hyperplasia, along with in-¢ltration of eosinophils, neutrophils, degranulated mast cells,CD4 and CD8 T-cells, and dendritic cells. Interestingly, therewas also evidence of increased neurocutaneous in£ammation,with ingrowth of nerve ¢bers making contact with mastcells and expressing CGRP and substance P. There was alsoevidence of a systemic Th2 response. As is the case in humanAD, genetics may have played an important role in this model,as Balb/c mice are known to be biased towards Th2 responses incertain settings such as Leishmania infection. It would be inter-esting to know if the same allergen produces the same type ofin£ammation in other strains of mice.This is the ¢rst report of anAD-like rash after epicutaneous sensitization with a relevanthuman skin allergen, and it should be a valuable tool for cellularand genetic dissection of the complex neuro-immunopathogen-esis of AD.

REFERENCES

Mantovani S, Palermo B, Garbellis, et al: Dominant TCR-alpha requirements for selfantigen recognition in humans. J Immunol 169:6253^6260, 2002

IN THIS ISSUEVOL. 121, NO. 2 AUGUST 2003