Immunobiology of Mouse Dendritic Epidermal T Cells: A Decade Later, Some Answers, But Still More Questions

Immunobiology of Mouse Dendritic Epidermal T Cells: A Decade Later, Some Answers, But Still More Questions

DENDRITIC EPIDERMAL T CELLS Immunobiology of Mo u se Dendritic Epidermal T Cells: A Decade Later, Some Answers, But Still More Questions Robert E. Ti...

2MB Sizes 2 Downloads 14 Views


Immunobiology of Mo u se Dendritic Epidermal T Cells: A Decade Later, Some Answers, But Still More Questions Robert E. Tigelaar*t and Julia



"Department of Dennat ology

and tSection of Immunobiology, Yale Skin Diseases Research Center, Yale University School of Medicine, New Haven, Cormecticut, U.S.A.

,,8 cells

Over the past decade, overwhelming evidence has

substantive advances in our knowledge about

accumulated in many species, most notably in mice,

in general (e.g., recent evidence that their manner of

that epithelial sites such as skin, intestine, and repro­

antigen recognition may be fundamentally different

ductive tract are populated with relatively discrete

from that used by conventional afJ T cells) and about

subsets of ,,8 cells. Such studies have identified sev­

epithelial-specific subsets such as murine

eral distinguishing and, in some cases, unique fea­

particular, it is clear that, compared with our under­


of the


standing of afJ cells, major gaps still exist

dendritic epidermal T cells (DETC) pop­

in our

ulating the skin of

all normal mice: homogeneous VS-Jl-C"lNl-D2-J2-C8 T-cell receptors devoid of

understanding of these cells. Persisting questions about DETC include: precise identification of the


ligands for their homogenous T-cell receptors, the

diversity, apparent tissue restriction in

adult mice to the skin, an important role for active

cellular and molecular requirements for their activa­

hair growth in their localization and/or proliferation

tion, their full range of functional activities, the



reason(s) for the absence in normal human skin of a precise morphologic and phenotypic homologue,

skin, and a capacity to recognize an antigen

expressed on stressed epidermal cells. These proper­

ties have

led to the hypothesis that

tinctive roles



DETC play dis­


and, perhaps most important, their biologically rele­ vant






and/or immunoregulation via recognition of a com­

and/or pathology. K£y words: ,,8

mon self-antigen expressed by adjacent cells under

105:43S-49S, 1995


cells. ] Invest Dermatol

various potentially harmful circumstances. Despite




non-germline- encoded (N) lating

It is now common knowledge (reviewed in [1]) that in most


chain . A minor proportion (less than


with the external environment (e.g., skin, reproductive tract,

gastrointestinal tract), particularly in rodents such as mice and rats,

div ersity seen among both recirculating a(3 and y8 TCRs is generated through a variety of mechanisms during the process of recombination. Combinatorial diversity, via somatic rearrangement (V, J, and C for (3 and

of different germline-encoded gene segments

furthermore, that the y8 TCRs expressed in a particular epithelium

are characteristic of and, in some cases, unique to that site.



DETC were initially named Thy-l + dendritic epidermal cells because of their characteristic expression of the Thy-1 alloantigen and formation of a conspicuous dendritic network among basal layer keratinocytes [2,3]. It soon be came clear that the adult epidermis of all normal strains of mice (but not immunodeficient strains suc h as scid and ntl/ntl) was populated b y Ia-, CD4-, CDS- cells that could display functional properties ofT cells, e.g., mitogen and interleukin-2 (IL-2) responsiveness, secretion ofIL-2, and various forms of cytotoxicity [4-6]. With the demonstration that the overwhelming majority of such cells from normal adult

Reprint requests to: Dr . Robert E. Tigelaar, Department of Dennatol­ ogy, Box 208059, Yale University, New Haven, CT 06520-8059. Abbreviations: HSP, hea t shock protein; KGF, kerarino cyte growth factor.

are populated preferentially by relatively sessile y8 T cells , and


1 chains; V, D, J, and C for a and 8 chains), is substantial. But the vast majority of TCR diver sity among such cells is concentrated within the V-J or V-D-J junctional region (also known as the CDR3 region) through the activities of two enzymes: 1) variable nibbling of the recombining gene segment ends by a DNA exonu­ clease, and 2) in sertion into these junctions of variable numbers of


gives recircu­

to recognize a vast array of

become clear over the past decade that various epi thelia interfacing


typically expre ss neither CD4 nor CDS. The enormous collective

capa city

particular antigen. In striking contrast to r ecirculating T cells, it has

CD3 -associated TCRs that are 18 heterodimers; such T cells


T cells the collec tive

enhances the likelihood that a given T cell will confront its

peripheral lymphoid tissues express either CD4 or CDS, as well as a CD3-associated, heterodimeric TCR for antigen composed of an and a

terminal deoxynucle­

foreign antigens; at the same time, their c apacity to recircula te

mammals the large majority of recirculating T cells in the blood and


nucleotides via

otidyl transferase . Such marked TCR heterogeneity

Copyright © 1995 by The S ociety for Investigative Dennatology, Inc. 43S




Table I.

Distinguishing or Unique



Features of Mouse DETC

1. Striking TCR homogeneity a. >95% ofDETC in nonnal adult mice (regardless of strain) express identical,,8 TCRs, all using the same" gene (V"5J,,lC,,1) and 8 gene (V81D82J82C8) coding segments and devoid of significant junctional diversity b. Implication: unlike conventional, recirculating 1l'f3 and ,,8 cells, whose collectively enonn ous TCR heterogeneity permits recognition of a broad array of antigens, DETC recognition of antigen is extremely limited (perhaps to a single ligand) 2. Restricted tissue distribution a. In adults, cells expressh.g homogeneous V,,51V81 TCRs have only been found in the skin (not isolated from blood, lymphoid tissues, or even other epithelial sites such as gastrointestinal tract, lungs, or reproductive tract) b. Cells with identical TCRs are present in early fetal thymus (13-17 d of gestation) but not in the thymus at later times 3. Critical role for activ e hair growth in DETC localization and/or proliferation in the skin 4. Capacity to recognize via their TCRs an antigen expressed on the surface of stressed or altered keratinocytes

mice expressed CD3-associated ya TCRs [7], they were renamed

identical to that on DETC (V1-D2-J2-Ca), the y chain is distinct

DETC. The results of more precise characterization of the ya TCRs

(Vy6 instead of V y5) [16].

on DETC were totally unanticipated. Not only did all of several


independent long-term DETC lines from AKRIJ mice express



TCRs with identical y coding segments (Vy5J yl Cyl) and identical

a coding segments (ValDa2]a2Ca), but the y and a junctional and devoid of either significant exonucleolytic nibbling or N nucleotide insertions [8]. Compatible results have been obtained on DETC freshly isolated from a variety of mouse strains by both flow cytometric staining and polymerase chain reaction analysis [9,10], as well as by our own studies of epidermal sheets demonstrating that more than 95% ofCD3+ cells also stain with monoclonal antibody 17D1, which reacts only with TCRs containing both Vy5 and Val [11]. Collectively, these results indicate that the TCR repertoire of the D ETC from normal mice is virtually monomorphic, certainly one of their most distin­ guishing features (Table I). It should be pointed out that T cells not expressing the canonical Vy5/Va1 sequence of prototypic DETC TCRs have been identi­ fied in epi dermal cell suspensions prepared from normal adult mice.

Early Fetal Thymocytes as

Such TCRs have been of both the ya and a{3 type and typically

izing on the absence ofCD3+ Vy5/Va1 + DETC in the epidermis

regions each were homogeneous

have shown considerable jun ctional diversity, raising the possibility that they actually represent recirculating cells isolated as


result of

either dermal "contamination" of the epidermal cell suspensions or their transient migration ("passenger lymphocytes") into the epi­

dermis. Cells with non-prototypic DETC TCRs have also been

observed in the epidermis of mice lacking a normal resident network of homogeneous Vy5/V()1 + DETC ; e.g., various non­

Vy5 ya cells in athymic (nulnu) mice [12], and ct{3+ cells in lethally irradiated, bone-marrow-reconstituted mice [13] as well as in mice genetically deficient in all ya T cells by virtue of homologous recombination into their genome of a nonfunctional TCR a gene

[14]. Whethe r

such non-prototypic DETC exhibit identical, over­

lapping, or distinct functional activities both

in vitro

and, more

DETC Precursors

Although few,

if any, T cells outside the epidermis of adult mice express DETC­ type, junctionally homogeneous Vy5/Val + TCRs, identical TCRs are seen on the earliest fetal thymocytes to express


CD3/TCR complex (12-13 d of gestation). Such Vy5/Val + cells remain the predominant CD3 + fe tal thymocyte until 17 d of gestation, after which time their proportion declines precipitously, coincident with the appearance of


cell s and a second , also

transient, "wave" of ya cells with TCRs identical to those on the T cells in reproductive epithelia [9,17]. Expression of either the

D ETC-type or reproductive-type y or a coding segments is rarely seen in thymocytes examined after birth [18]. Identical TCRs on adult DE TC and early fetal thymocytes strongly suggested that early fetal thymocytes were DETC precur· sors; data consistent with this possibility were obtained by capital­

nul"u mice. Transfer of either unfractionated (as thymus grafts or cell suspensions) or purified CD3+ or 17D1 + (Vy5/Va1+) day 16-17 fetal thymocytes into young nulnu (or Thy-1 congenic) recipients was followed by the gradual accumu­ of young athymic

lation in the epidermis ofCD3+ Vy5/Val + cells, initially as round

and relatively disper s ed cells which expanded in clusters (consis tent with in situ proliferation) and became more dendritic [19-211-* Transfer ofCD3- or VyS /V 81- fetal thymocytes failed to repop· ulate recipient epidermis with CD3+ Vy5/Val + DETC. These data indicate that fetal thymocyte precursors of DETC rearrange and express their that


TCR genes before they enter the e pidermis and

epidermis, despite its obvious abnormality, can receive

DETC precursors and permit their post-thymic expansion. Al­

important, in vivo, remains to be clarified (see also below).

though such data are consistent with the hypothesis that the DETC

In normal adult mice, T cells with DETC-type, junctionally homogeneous Vy5/Val + TCRs are apparently restricted to the

prov e that all DETC must be derived from fetal thymic precursors.

skin, as this is the only site reliably shown to contain T cells with

found in normal mice are (fetal) thymic

dependent ,

they do


In fact, grafts of 16-d fetal skin containing scattered Thy-l + CD3-

this specific receptor. Although it has been suggested that DETC


may migrate from the skin after perturbations such as the a pplica­

skin graft donor-type CD3+ VyS+ DETC, consistent with intra­

cells have been reported to become gradually populated by

tion of contact irritants or allergens, to date, such VyS/Va1 + cells

cutaneous ,

have not been isolated from either peripheral lymphoid tissues

Furthermore, adult bone marrow reconstitution of thymectomized,

(e.g., spleen, lymph nodes) or the thymus of adult mice. Cells

lethally irradiated recipients results in the gradual appearance in the

expressing the DETC-type TCR also have not been isolated from

epidermis (particularly in footpads, which normally have a very low

potentially thymic-independent di1ferentiation


other epithelial sites known to be populated with other distinct

density of prototypic VyS/V81 + DETC) of donor-type CD3t

subsets of y8 cells. For example , the TCR repertoire of the y8 cells


in the gut epithelium is strongly biased toward selective expression

some DETC can be derived from thymic-independent precursors.

ofVy7+ cells; however, unlike DETC TCRs, there is no prefer­

Evidence has been obtained recently that the first two waves of Vy5+ DETC-type and Vy6+ reproductive-type fetal thymocytes are derived from a common lineage distinct from other ya cells and

ential pairing with a particular a chain, and both the V-y7 and the paired a chain junctional sequences exhibit considerable diversity



DETC [13], again consistent with the possibility that

The TCR homogeneity and tissue restriction of DETC are

mimicked by the y8 cells

in the epithelium of the vagina, uterus,

and tongue. Like DETC, these cells express a monomorphic TCR

* Lewis

devoid of junctional diversity; however, whereas the 8 chain is


JM. Tigelaar

RE: Thymic dependence of Thy-l + dendritic

cells (abstr).] Invest Dernlatol 92:471A, 1989.


105, NO.






that the rearrangement ofspecific TCR V-region coding segments is programmed developmentally. In nine V'Y5+ hybridomas (in­ cluding ones derived from the skin), mRNA for V')'6-C'Y1 was found in only one, and it was fOWld in none of 10 other 'Y8 hybridomas. C onversely, all 25 V'Y6+ hybridomas (derived from reproductive tract) examined expressed V')'5-C'Y1 mRNA, whereas 33 other V')'6- hybridomas expressed no V')'5 message [23]. These findings extend previous studies that showed that fetal liver stem. cells , but not adult bone marrow stem cells, could give rise to V'YS+ cells in fetal thymic organ cultures [24]; and that the V')'6 gene was not accessible to transcription factors in precursors of Vy5+ cells [2S]. Recendy, studies using fetal thymic organ culture have clarified alterations in cell surface phenotype that accompany the intrathy­ mic differentiation of the precursors of V'Y51V81+ DETC [26,2 7]. Immature DETC precursors had the following phenotype: V'Y S1ow, heat-stable antigenhigh, MEL14high (a homing receptor comple­ mentary to L-se1ectin-type addressins expressed on peripheral lyn;ph node endothelial cells), and CDShigh• It is interesting that nearly all these thymocytes were CD8+, and about 50% expressed low levels ofCD4. This phenotype gradually changed with time to V-yShigb, heat-stable antigenlow, MEL141ow, CDSlow, C D8 , and CD4-. Recent experiments have also definitively addressed the role of intrathymic selection (during which cells expressing low levels of TCR on their surface are selected on the basis of their affinities toward ligands on other cells within the thymus) in generating DETC precursors with homogeneous, junctionally nondiverse VyS/V81 TCRs [14]. These elegant studies used mice with a targeted mutation (via the process of homologous recombination) in the TCRB locus such that no full-length 8 chain protein could be made, and hence no')'8 TCRs could be expressed on developing T cells. When the'Y and8 gene jWlctional sequences were analyzed in these 8 gene "knockout" mice, they were similar to those of'Y8+ "wild-type" littermates, i.e., programme d rearrangements of spe­ cific V , D, and J coding segments; limited exonuclease-mediated trimming of coding segment ends; absence of N-region additions (probably the result of lack of terminal transferase activity in these particular precursors); and site-specific recombination guided by short homologous regions at the ends of adjacent coding segments. These results convincingly argue that TCR homogeneity inDETC is best explained not by intrathymic selection but rather by a "molecular constraint model" (restricted recombination). -

lloles of Various Cell Surface Molecules and Cytokines in DETC Localization and/or ProHferation Daily exposure of embryos in utero (from day 10 of gestation Wltil birth) to an anti-V')'5 monoclonal antibody (MoAb) given to the pregnant female resulted in the absence of V')'5 + DETC in the skin of the offspring 4 months later [19]. These results are consistent with a crucial role for the V'Y51VB1 TCR in the localization of DETC precursors to the skin. However, the filct that such in utero treatment also resulted in the total depletion of V'Y5 + cells from day-16 fetal thymus strongly points to a mechanism of antibody­ mediated ablation of precursors within the developing thymus. That the Vy5/VB1 TCR is not absolutely required for the migra­ tion of'YB cells to the skin has been shown by studies of mice in which all the 'YB cells expressed a transgenic V'Y41V85 TCR normally expressed on some adult thymocytes and lymph node ')'8 cells; transgene+, V')'S/V81- DETC were readily seen in the epidermis [28]. Of note, however, the density of V')' 4VB5+ DETC in these transgenic animals was only 5% to 20% of the density of V151V8 1 + DETC seen in normal mice; these results are consistent with a role for the V'Y5V81 TCR in the proliferation ofDETC after their localization in the skin (see also below). In utero treatment with antibody specific for the intermediate­ affinity (P75) IL-2 receptor {3 chain (which appears to be preferen­ tially expressed on V')'S+ fetal thymocytes) also resulted in the absence ofV'Y5 + DETC in the skin of oftSpring1 month after birth.. This absence was highly selective; Langerhans cell densities in





treated mice were normal, as were the numbers of 18 and af3 T cells in the spleen and gastrointestinal tract [29]. These results suggest an important role for IL-2 in the development. localization, and/or proliferation of DETC in the fetal thymus or skin, as do studies demonstrating a selective increase in DETC in transgenic mice expressing the human IL-2 gene [30]. Furthermore, virtually all V')' SV81 + DETC lines can be expanded and maintained in the presence of only IL-2 and respond to TCR cross-linking by producing large amoWlts of this lymphokine. However, the fact (A. Elbe and G. Stingl, personal commWIication) that DETC densities are normal in mice rendered totally deficient in IL-2 via the strategy ofhomologous recombination [31] argues strongly that this lymphokine is not absolutely required. Which cytokines the DETC from such mice produce and utilize in vivo and in vitro has yet to be reported. Given the documented capacity of DETC to respond to IL-7, a cytokine known to be produced by keratinocytes [32], characterization ofDETC in mice deficient in both IL-2 and IL-7 could be very informative. IL-4 has been shown to act synergistically with IL-2 in inducing proliferation of V'Y 5+ cells in fetal thymic organ cultures [33]. On the other hand, examinations of IL-4 transgenic mice that overex­ press this cytokine in a variety of tissues have shown that although freshly isolated 14-d transgenic fetal thymus had normal numbers of V')'5 + cells, the skin of adult transgenies was totally devoid of V')'5+ DETC (A. Elbe, personal communication). Although the mechanism by which IL-4 acts in these mice has yet to be clarified, these results strongly suggest that IL-4 is involved in the maturation and/or migration ofDETC. LeFran�ois et al [34] demonstrated that DETC in adult skin and their fetal thymocyte precursors express high levels of the integrin aEf37, also known as CD103. V ery recendy, studies by C epek et al [35] of d'f37-mediated adhesion of human intraepithelial (gut­ derived) lymphocytes to epithelial cells in vitro have demonstrated elegandy that this adhesion is dependent upon the epithelial adhesion molecule E-cadherin, the first data to suggest strongly that the d'{37 and E-cadherin molecules can act as cOWlter-receptors. Interactions between aB{37 on DETC and E-cadherin on keratino­ cytes may be responsible, at least in part, not only for the intraepidermal localization of these 'Y8 cells, but also for their characteristic dendritic morphology.

Role of Anagen Hair Follicles Site-to-site variations in DETC density in normal inbred strains of mice correlate at least in some cases with the density of hair follicles and/or hair growth; e.g., normal footpad epidermis (devoid of hair) has an extremely low density ofDETC [36]. Despite the obvious abnormality in follicular keratinization in nu lnu mice, hair follicles are present in normal numbers and undergo normal cycling. Hair follicle cycling in mice is characterized by synchronized entry into anagen, which proceeds from one contiguous region to another, producing "waves" ofhair growth in cyclic, head-ta-tail patterns [37,38]. In nude mice, these waves are visible as areas of thickened skin with barely visible, broken hairs. Although newborn mice (including nulnu) are clini­ cally hairless, the initial wave of hair growth begins about the time of birth, rapidly progresses over the entire skin surface, and then terminates at about 3 weeks of age. This is followed at several-week intervals by slower, periodic cephalocaudal waves, making it likely that at any given time, the follicles at some (but not all) sites on adult mice will be in anagen. Our initial analysis of adult nulnu mice 8-12 weeks after reconstitution with day-16 fetal thymocytes revealed a very irreg­ ular distribution of CD3+ V'Y5/V 81 + DETC. Although occa­ sional, randomly analyzed (i.e., without regard to specific site) pieces of body wall skin had relatively normal DETC densities, many pieces, representing a substantial proportion of the trunk skin, were totally devoid ofV')'5/V81 + DETC. To test the hypothesis that DETC localization and/or proliferation were associated pref­ erentially with anagen follicles, those regions of trurJk skin of prospective adult nulnu recipients that were obviously in anagen hair growth were outlined with a permanent marker immediately




before injection of day-16 fetal thymocytes. Eight weeks later, Vy5V/)1 + DETC were observed only in skin obtained from sites that had been in active hair growth at the time of injection. To test further the relation between DETC density and hair follicles, we have begun examining mice with various abnormalities in hair patterns or growth. A recently characterized mouse line transgenic for a parathyroid hormone-related peptide gene driven by the keratin 14 promote, was noted to be virtually devoid of hair on the abdomen [39]. Our subsequent examination of the densities of bone-marrow-derived intraepithelial cells in these mice revealed that although Langerhans cell densities were equivalent in the "bald" area on the abdomen and in adjacent normal skin, the density of Vy5/V/)1 + DETC in the hairless area was only 8% of their density in normal skin. Furthermore, DETC density in the somewhat sparsely coated skin of "crinkled" (crlcr) mice was only 60% of that observed in clinically normal skin of littermate controls; crier mice have sparse coats because hair follicle formation is restricted to the period from fetal day 17 to birth (compared with fetal day 14 to 3 d after birth in normals) [40]. In contrast, DETC densities are significantly increased in two hair-mutant strains (angora and hairy ears) in which the anagen cycle is prolonged by several days [41,42]. The mechanism(s) by which active hair growth affects adult DETC density remains to be determined, but possibilities include affecting either the initial localization of fetal thymocyte precursors and/or their subsequent in situ proliferation. An intriguing possi­ bility is that active hair growth stimulates the in situ proliferation of DETC via the localized, transient expression of the physiologic ligand (antigen) for the Vy5/V/)1 + TCR selectively expressed by DETC. In fact, comparison of the densities of intraepidermal Vy5/V81 + cells and intraepidermal CD3+ cells during the neo­ natal period (i.e., during the initial anagen cycle) provides data consistent with this possibility. In newborn C57B1I6 epidermis, the density of all T cells regardless of their specific TCRs (assessed by staining with anti-CD3) was only 10% of that seen in 3-week-old mice, the age at which adult densities are first reached in this strain; furthermore, only about 50% of the CD3 + cells in newborn epidermis were Vy5/V/)1+. By 3 weeks, however, more than 95% of the intraepidermal CD3 + cells were V y5/V /)1 +. Calculation of the changing densities of CD3 + Vy5/V/)1 + versus CD3 + Vy51 V/)l- cells during this newborn period revealed that while there had been a 10-fold increase in the density of cells with prototypic DETC TCRs, the density of CD3+ Vy5/V/)1- cells remained essentially unchanged. Such results are consistent with the possi­ bility that the density and TCR homogeneity of the DETC present in normal adult skin are due, at least in part, to preferential proliferation of Vy5/V/)1 + DETC, conceivably in response to expression of the physiologic ligand for this TCR in the microen­ viromnent· of anagen hair follicles. The previously noted [28] low density of intraepidermal T cells seen in adult Vy4/V/)5 transgenic mice (in which virtually all T cells, including those in the skin, expressed the transgenic Vy4/V/)5 TCR) is also consistent with the above hypothesis. Given the substantial differences in hair patterns and follicle cycling between mice and man, further analysis of the role(s) of anagen follicles in DETC differentiation, localization, and proliferation is clearly indicated and may provide insights into why normal human skin lacks an obvious morphologic and phenotypic counterpart to mouse DETC [43,44].

ANTIGEN RECOGNITION BY DETC The monomorphic nature of the TCRs on the vast majority of DETC in normal adult skin, coupled with the apparent restriction of cells expressing such invariant Vy5/V81 + TCRs to adult epidermis, strongly suggested that DETC primarily recognize only one ligand and do so primarily, if not exclusively, in the skin. This reasoning led to the hypothesis that DETC function in immune surveillance and/or immunoregulation by recognizing a common self-antigen expressed by transformed, damaged, or altered cells in the epidermal microenvironment [8 45] . One obvious candidate for such a ligand would be a stress protein such as a heat shock protein ,



(HSP). HSPs are phylogenetically highly conserved families nor­ mally expressed at modest levels, but markedly up-regulated after stimuli such as heat shock, nutrient deprivation, intracellular infec­ tion, or malignant transformation, and whose vital functions in­ clude acting as chaperones in protein folding and intracellular transport (reviewed in [46]). Increased proportions of y/) cells have been found in blood and/or tissues of humans and animals infected with a broad range of microorganisms (reviewed in [47]), suggest­ ing that such y/) cells react against either microbial HSPs and/or self-HSPs expressed by "stressed" autologous cells. Reactivity against both mycobacterial-derived and autologous HSP-65 (or -60) has been reported both in human y/) cells and in murine hybridomas produced from newborn thymocytes or adult spleen cells (reviewed in [48]). However, the responding cells express junctionally diverse TCRs [49], consistent with the possibility that HSPs in these cases may be acting as superantigens (i.e., directly binding to V-, but not to junctional regions) and not as conven­ tional peptide antigens. Unlike the above cells, Vy5/V81 + DETC have not been found to be stimulated by crude mycobacterial extracts or purified protein derivative. However, recent data are consistent with expression of the ligand for the DETC TCR on keratinocytes stressed by heat shock, nutrient deprivation, or malignant transformation [50,51). Purified stressed keratinocytes (in the absence of any added source of antigen-presenting cells) activated DETC lines via their Vy51 V/)l TCRs to proliferate and secrete IL-2; similar activation was not seen with normal keratinocytes or with normal or heat-shocked fibroblasts or spleen cells, consistent with the possibility that DETC ligand expression may be tissue restricted. Furthermore, in vivo application of contact allergens or irritants increases DETC density with morphologic changes suggestive of cell division; activates DETC, as assessed by their enhanced responsiveness to exogenous IL-2 [52]; and induces expression of the DETC ligand on kerati­ nocytes, as such cells required no additional stress (e.g., heat shock) to stimulate DETC proliferation in vitro. More precise identification of the ligand(s) on stressed keratinocytes for the DETC Vy5/V/)1 TCR has not been published despite substantial efforts in several laboratories. Although the stimulatory activity appears to be con­ tained in a low-molecular-weight fraction of acid-eluted material (W. Havran, unpublished data), other studies have shown that such material may be mitogenically active on a variety of target cells, including TCR loss mutants (C. Reardon, personal communica­ tion). Implicit in the issue of the physiologic ligand for the DETC TCR is the still unsettled question of whether such recognition involves a restricting element, as is required for presentation of antigenic peptides to conventional a{3 T cells [53]. The observations that DETC were equivalently activated by stressed epidermal cells from syngeneic or allogeneic donors [50,51] indicate that if a restriction element is used, it is distinct from the usual polymorphic class I and II major histocompatibility complex (MHC) molecules. Other less polymorphic, f32-microglobulin-associated MHC-like molecules, such as class Ib molecules CD1, Qa, or Tla, whose expression has been shown to be tissue restricted [54,55], had been considered theoretically attractive alternatives. However, homologous recom­ binant mice lacking a functional gene for {32-microglobulin (and hence markedly deficient in cell surface expression of class I and class Ib) have normal numbers of functional DETC (reviewed in [56]). Until very recently, the experimental design of studies of y/) TCR specificity has generally used the paradigm of antigen recog­ nition by conventional a{3 T cells, i.e., recognition of a peptide in the groove of an MHC-like restricting element. However, evi­ dence is mounting that y/) cell antigen recognition is fundamentally different from recognition by a{3 cells and may in fact resemble much more closely recognition of antigen by immunoglobulin, i.e., conformational determinants on intact molecules that are not necessarily proteins [57]. Recently, a y/) clone isolated from the lymph node of a Herpes simplex virus-1-immunized mouse was shown to recognize a Herpes simplex virus-1 transmembrane






glycoprotein (gl) independently of MHC class I or class II; this recognition was blocked by anti-gI antibodies, which themselves recognize a conformation epitope on the gl molecule. The y8 clone Wled to recognize cells transfected with a transmembrane region loss mutant of gl, whereas wild-type gl transfectants were recog­ nized. Finally, a soluble recombinant gl molecule in its whole, unprocessed form was recognizable by the clone in the absence of added antigen-presenting cells [58]. Additional recent data suggest­ ing that y8 TCR recognition may be unique includes demonstra­ tions that some human y8 cells are activated by nonpeptide ligands, including monoethyl phosphates [59] and a triphosphorylated thy­ midine-containing compound [60). Another persisting question is whether, as for conventional a{3 T cells [61], y8 TCR-mediated triggering in DETC involves the crucial interactions of accessory and!or co stimulatory molecules on the responding DETC and antigen-presenting cell. Roberts et al [62] have demonstrated a critical role for vitronectin receptors during activation of Vy1.1-Cy4 cells. However, there is little knowledge about the range of accessory molecules expressed by V'Y5/V81 + DETC, whether expression levels vary with the acti­ vation state, and the role(s) of such molecules in TCR-mediated triggering. Takashima et al (personal communication) found re­ cently that DETC express a molecule named 2B4. Not only does expression of 2B4 seem to correlate with the capacity to exhibit Iymphokine-activated killer and/or natural killer-like cytotoxicity, but DETC also are stimnlated to proliferate and secrete lympho­ kines after activation with anti-2B4 MoAb. The relative ease with which one can establish DETC lines in vitro by stimulation of epidermal cell suspensions with mitogens or immobilized anti-TCR MoAbs, compared with the difficulties in establishing long-term lines of mouse y8 cells from other sites, is consistent with the possibility that DETC may have unusual, if not unique, activation requirements and!or signaling pathways after activation. In this regard, recent analysis of the TCR complexes on T cells in homologous recombinant mice lacking the CD3 zeta chain (and markedly deficient in conventional a{3 cells) revealed normal numbers of DETC with high levels of expression of TCRs associ­ ated with the FcRy chain [63]. FUNCTIONAL ACTIVlTmS OF DETC IN VITRO As noted previously, activated (but not freshly isolated) Vy5/V81 + DETC can kill a variety of target cells (including transformed keratinocytes) and can also secrete a number of lymphokines. All long-term DETC lines examined have exhibited the lymphokine mRNA and/or protein profile characteristic of CD4+ TCR a{3+ THI cells, i.e., IL-2 and interferon-y (IFN-y), but not IL-4 or IL-5 [6,64]. On the other hand, activation of short-term lines of DETC containing about 95% Vy5+ cells was shown recently to result in the appearance of IL-4 mRNA [64]. A variety of data indicate that the lymptlOkine profile of CD4+ TCR a{3+ cells is dependent on a number of factors, including antigen dose (strength of the TCR activating signal) and the cytokine microenvironment in which the activated cells proliferate [65]. We recently analyzed (by bioassay or enzyme-linked immunosorbent assay) the capacities of a number of DETC lines to secrete IL-2, IFN-y, and IL-4. After activation with anti-TCR MoAbs, all short-term DETC lines unequivocally produced small amounts of IL-4 and IFN-y protein and large amounts of IL-2, whereas two long-term DETC lines produced comparable amounts of IL-2 and IFN-y but no detectable IL-4. These results confirm and extend previous studies in which mRNA levels were assessed qualitatively [64]. However, the amounts of IFN-"I and IL-4 secreted after activation were apparently unaffected by the culture conditions used initially to stimulate and then to maintain the short-term DETC lines. That is, lOOO-fold variations in the concentrations of immobilized anti-TCR MoAb used for initial stimulation, and/or addition of exogenous IL-4' and anti­ IFN-"I MoAbs during the first several weeks of culture in addition to the routinely added IL-2, did not result in lines with a TH2-like profile (increased IL-4 and decreased IFN-y). Very recently, Boismenu and Havran [66] reported that a



long-term DETC line, particularly after it was activated with mitogen or anti-TCR MoAb, expressed mRNA for and secreted a factor which, by a variety of rather stringent criteria, was indistin­ guishable from keratinocyte growth factor (KGF). Freshly isolated, resting DETC did not produce KGF; however, DETC cultured for several days or activated with anti-TCR MoAb did express KGF. Although similar KGF expression was detected in activated intra­ epithelial '18 cells isolated from the gastrointestinal tract, several other populations examined, including freshly isolated as well as long-term lines of spleen- and thymus-derived a{3 and '18 cells, and even intraepithelial a{3 cells isolated from the gastrointestinal tract, failed to express KGF. IN VIVO FUNCTIONS OF DETC AND OTHER '18 CELLS

Perhaps the most persistent and obvious questions about '18 cells in general and DETC in particular relate to their relevant biologic (physiologic and/or pathologic) functions in vivo. Speculation re­ garding the functional importance of '18 cells in general has centered around several nonexclusive possibilities, including early responses to various infectious agents (i.e., a "first line of de­ fense"), elimination of altered (e.g., stressed or transformed) cells, initiation of autoimmunity, and down-regulation of responses mediated by conventional a{3 T cells. Commonly, such speculation has been based on correlative data showing increased numbers of "18 cells during particular infections or in particular sites of disease, e.g., the reports of increased numbers of '18 cells in granulomatous skin lesions [47] and in discoid lupus erythematosus [67]. With the increasing use of both anti-TCR antibodies to selectively deplete animals of discrete subsets of T cells, and of homologous recom­ binant mice genetically deficient in either a {3 or "18 cells, studies have begun to evaluate more pointedly the roles of these cells in systemic immunity/immunopathology, particularly in resistance to various infectious agents. For example, depletion studies using MoAbs or homologous recombinant mice have both been per­ formed in a mouse model of Listeria monocytogenes infection. The results indicated that either a{3 or ,,/8 cells were sufficient for early protection in primary infection. Resistance to secondary infection was mediated mainly by a{3 T cells but also involved "18 cells, which apparently regulate a{3 cells; i.e., in the absence of y8 cells, a {3 cells induce multiple abscess-like hepatic lesions rather than granuloma­ tous lesions ([68]; also reviewed in [56]). Most studies of mice deficient in a{3 cells have revealed a profound defect in their capacities to make humoral antibodies to a variety of antigens [56,69]; however, recent observations in TCR a chain "knockout mice" (-/-) have suggested that some '18 cells can make IL-4 and help B cells make IgG antibodies, including autoantibodies against DNA (L. Wen and A. Hayday, personal communication). And although rejection of histoincompatible skin is usually the almost exclusive functional domain of a{3 cells, rejection of grafts express­ ing nonclassic MHC antigens (such as Qa and Tla) in addition to classic MHC alloantigens in TCR a -/- mice is consistent with the view that in some mice, ,,/8 cells are sufficient to cause skin graft rejection, perhaps via recognition of such nonclassic MHC antigens [69]. Consideration of the biologic role(s) for specific subsets of '18 cells such as DETC centered initially around their potential rele­ vance in what can broadly be termed "immune surveillance" (Table 0). This postulated role for DETC focuses on their potential to e1fect the removal of epidermal cells altered in such a way (e.g., infected, damaged, or transformed) as to threaten the skin's functional integrity; such DETC-mediated removal of altered cells could be direct (e.g., via the known cytotoxic potential of activated DETC) and!or indirect (e.g., via secretion of lymphokines that could recruit and activate other more conventional pro­ inflammatory cells into the site). The recent discovery that acti­ vated DETC can secrete KGF in vitro [66] has raised the possibility that an important function of DETC in vivo may be to stimulate keratinocyte proliferation after insults that disrupt the epidermal barrier. Although the amount of KGF produced by dermal cells such as fibroblasts after full-thickness wounding of the skin is likely



Table II.


Potential Biologic Roles of Intracutaneous yo Cells as Protectors of the Skin's Functional Integrity

Immune surveillance a.

2. 3.


Against intracutaneous pathogens (e.g. ,



b . Against altered (e.g. , damaged or transformed) epidermal cells Stimulation of keratinocyte proliferation via elaboration of growth factors (e.g., KGF) in response to epidermal damage Regulation of intracutaneous immune responses mediated by other cells a.

Down-regulation of contact hypersensitivity reactions

b. Down-regulation of im:'iepidermal autoimmune reactions

to be quantitatively greater than that produced by activated DETC, the intimate association of DETC and keratinocytes may render the DETC-derived KGF contribution biologically relevant. Further­ more, after stresses or damage confined to the epidermis, KGF

We acknowledge Adrian Hayday for stimulating and helpful discussions in preparing this review. This work was suppolted in palt by National Institutes of Health grants AI2 7404 and AR4 1 942 (Yale Skin Diseases Research Core Center) .

might be produced selectively by activated DETC . In considering other potential roles for DETC as protectors of the functional integrity of the skin, it is important to realize that epidermal integrity and normal function can also be impaired by inadequately suppressed intracutaneous immune responses medi­ ated by conventional T cells, including autoreactive T cells. Several studies , commonly using the model of contact hypersensitivity to allergens such as dinitrofluorobenzene, are consistent with the hypothesis that DETC may function as down-regulators of intra­ cutaneous immune responses mediated by conventional T cells (reviewed in [70] ) .


recent report has shown that allergen­


Hedrick SM, Eidelman FJ: In: T Lymphocyte Antigen Receptors. Paul WE (ed.).


Bergstresser PR, Tigelaar RE, Dees JH, Streilein JW: Thy-1 antigen-bearing


Tschachler E, Schuler G, Hutterer J, Leibl H , Wolff K, Stingl G: Expression of


Nixon-FultonJL, Hackett J, Bergstresser PR, Kumar V, Tigelaar RE: Phenotypic

Fundamental Immunology. Raven, New York, 1 99 3 , pp 383-420 dendritic cells populate murine epidennis . ] Invest Dermatol 81 : 2 8 6 -288, 1 983 Thy-l antigen by murine epidermal cells. ] Invest Dermatol 8 1 : 282-2 8 5 , 1 983 heterogeneity and cytotoxic activity of Con A and IL-2-stimulated cultures of mouse Thy-l + epidermal cells . ] Invest DermatoI 9 1 : 6 2- 68 , 1 988 5.

epidennal cells in mice: precursor frequency analysis and cloning of con­

conjugated activated DETC can inhibit induction of contact hyper­ sensitivity to the same allergen in vivo and that the mechanism of this down-regulation may involve killing of allergen-specific Shiohara CD4 +


a l [36] established a model o f autoimmunity in which




canavalin A-reactive cells. ] Invest DermatoI 90:671-678, 1 9 88 6.

clones. ] ImmunoI 1 4 2 : 1 422-1 428, 1 9 89 7.

only an intense delayed type hypersensitivity reaction in the dermis,

epidermal cells express T3 antigen and the T-cell receptor l' chain. Proc Natl Acad Sci USA 8 4 : 4 5 8 6 - 4 5 9 0 , 1 98 7 8.

lular adhesion molecule- 1 -mediated epidermotropism of injected cells occurred in a graft-versus-host-like (lichenoid) pattern. In

cells. Cell 5 5 : 83 7- 847, 1 988 9.

Havran WL, Allison JP: Developmentally ordered appearance of thymocytes


Asamow DM, Goodman T, Lefrancois L, Allison JP: Distinct antigen receptor

expressing different T-cell antigen receptors. Nature 3 3 5 :443- 4 4 5 , 1 988 repertoires of two classes of murine epithelium-associated T cells. Nature

normal mice, for a period of many months after the initial delayed type hypersensitivity/graft-versus-host reaction has cleared in both

3 4 1 : 60 - 62 , 1 989 11.

hypersensitivity response, but virtually no intraepidermal compo­

5 8 S- 63 S , 1 990 12.

gous recombinant ("knockout") mice, which are totally deficient in all 'Y8 cells, including V'YSVa1 + numbers of


however, TCR

8 -1- mice previously reconstituted with day- 1 6

develop intraepidermal resistance after initial inj ection of autoreac­ tive cells [72 ] . These results represent the strongest evidence to date functioning to maintain epidermal integrity in the face of autoim­ (or ideally, of homologous recombinant mice selectively deficient

and chemical carcinogenesis, should further enhance our under­ standing of this unique intraepidermal T-cell subset in mice and refocus efforts to identifY the cell(s) that can perform analogous functions in human skin.

Kyes S , Carew E, Carding SR, Janeway CA Jr, Hayday A: Diversity in T-cell receptor l' gene usage in intestinal epithelium. Proc Natl Acad Sci USA 8 6 : 5 5 27-5 53 1 , 1 989


Itohara S , Farr AG, LafailIe JJ, Bonneville M, Takagaki Y, Haas W, Tonegawa S : Homing of a 1'/) thymocyte subset with homogeneous T-cell receptors to mucosal epithelia. Nature 3 4 3 : 7 5 4 -757, 1 9 90

1 7.

!to ' K, Bonneville M, Takagaki Y, Nakanishi


Kanagawa 0 , Krecko EG,

Tonegawa S : Different 'Yl> T-cell receptors are expressed on thymocytes at different stages of development. Proc Natl Acad Sci USA 8 6 : 6 3 1 - 6 3 5 , 1 989 18.



DeClonx A , Bonneville M , Takagaki Y, Tonegawa S : Junctional

sequences of T cell receptor 1'/) genes: implications for 1'/)


cell lineages and

for a novel intermediate of V-(D)-J joining. Cell 5 9 : 8 5 9 - 870, 1 989 19.

Havran WL, Allison JP: Origin ofThy-l + dendritic epidermal cells of adult mice


Payer E, Elbe A, Stingl G : Circulating CD3 + IT cell receptor V1'3 + fetal mutine

from fetal thymic precursors. Nature 344:68-70, 1 990 thymocytes home to the skin and give rise to proliferating dendritic epidermal

just in V'YS /V81 + DETC) to other models of cutaneous immunity and/or pathology, e . g . , contact hypersensitivity, viral infections,

Clarke AR, Hooper

Cell 7 2 : 3 3 7-34 8 , 1 993 15.

that the DETC in normal mice are biologically relevant, in this case mune attack. Future studies extending the use of such 8 - 1 - mice


generation of a i3 T cells and programmed rearrangements of 1'/) TCR genes.

resistance after their initial challenge with autoreactive T cells; fetal thymocytes (containing V'YS/V81 + DETC precursors) did

!tohara S , Mombaerts P, Lafaille J, Iacomini J, Nelson

ML, Farr A, Tonegawa S : T cell receptor /) gene mutant mice: independent

DETC, but have normal

cells . These mice failed to develop intraepidermal

Shiohara T, Moriya N, Hayakawa J, Arahari K, Yagita H , Nagashima M, receptor a/3/CD3 and CD8 . ] ImmunoI 1 5 0:4323-4330, 1 993


autoimmune attack. Very recently, this postulate has been strongly

8 chain homolo­

Demonstration of a CD3 +

Ishikawa H : Bone marrow-derived dendritic epidermal T cells express T cell

for DETC in the acquisition of such resistance to intraepidermal supported by similar studies carried out in TCR


receptor diversity. ] ImmunoI 1 4 9 : 4 1 3- 420, 1 9 92 13.

with the markedly increased numbers of DETC in the epidermis of "resistant" mice, led these authors to postulate an important role

Payer E, Strohal R, Kutil R, Elbe A, Sting!

lymphocyte subset in the epidennis of athymic nude mice: evidence for T cell

nent. The absence of such intraepidermal "resistance " in nulnu or adult thymectomized, bone-marrow-reconstituted mice , coupled

Tigelaar RE, Lewis JM, Bergstresser PRo TCR 1'/ /)+ dendritic epidermal T cells as constituents of skin-associated lymphoid tissue . ] Invest Dermatol 94(Suppl):

the dermis and epidermis, rechallenge of the same site with the autoreactive clone results in an intense intradermal delayed type

Asarnow DM, Kuziel WA, Bonyhadi M, Tigelaar RE, Tucker P W , Allison JP: Limited diversity of 1'/) antigen receptor genes of Thy-l + dendritic epidermal

but also a prominent concurrent intraepidermal inflammatory com­ ponent, in which leukocyte function-associated antigen-1 /intracel­

Stingl G, Koning F, Yamada H , Yokoyama WM, Tschachler E, Bluestone JA, Steiner G, Samelson LE, Lew AM, Coligan JE, Shevach EM: Thy-l + dendritic

MHC-autoreactive clones, when inj ected

into the footpad (a site normally deficient in DETC) , elicited not

Havran WL, Poenie M, Tigelaar RE, Tsien RY, Allison JP: Phenotypic and functional analysis of ')'8 T cell receptor-positive murine dendritic epidermal


cells by allergen-conjugated DETC [ 7 1 ] .

Takashima A, Nixon-Fuiton JL, Bergstresser PR, Tigelaar RE: Thy-l + dendritic

T cells. ] ImmunoI 1 4 6 : 2 5 3 6 -2543, 1 99 1 21.

Modlin RL, Lewis J, Uyemura K, Tigelaar RE: T lymphocytes bearing 1' /) antigen


Elbe A, Kilgus 0 , Strohal R, Payer E, Schreiber S , Stingl G : Fetal skin: a site of


Heybome K , F u Y-X, Kalataradi H, Reardon C, Roark C, Eyster C, Vollmer M ,

receptors in skin. Chem Immunol 5 3 : 6 1 -74, 1 992 dendritic epidermal T cell development. ] ImmunoI 1 4 9 : 1 6 94-1 70 1 , 1 99 2 B o rn W , O'Brien R : Evidence that murine V 1'5 and V1' 6 1'/)..T C R+ lympho-





1 995

1 5 1 :4523- 4 527,



Ikuta K, Kina T, MacNeil I, Uchida N, Peault B, Chien Y-H, Weissman IL: Development of -y8 T-cell subsets from fetal hematopoietic stem cells. Ann NY



Goldman JP, Spencer DM, Raulet DH: Ordered rearrangement of variable region genes of the T cell receptor 'Y locus correlates with transcription of the unrearranged genes. ] Exp Med 1 77 : 729-739, 1 993 Leclercq G, Plum J, Nandi D, De Smedt M, Allison JP: Intrathymic differentia­ tion of Vy3 T cells. ] Exp Med 1 78 : 3 0 9 -3 1 5 , 1 993 Payer E, Kutil R, Stingl G: CD5 - dendritic epidermal T cells are derived from CD5 + precursor cells. Eur] Immunol 2 4 : 1 3 1 7-1 3 2 2 , 1 9 9 4 Bonneville M, Itohara S, Krecko EG, Mombaerts P, Ishida I, Katsuki M, Berns A, Farr AG, Janeway CA Jr, Tonegawa S: Transgenic mice demonstrate that epithelial homing of 1'/0 T cells is determined by cell lineages ind�pendent of T cell receptor specificity. ] Exp Med 1 7 1 : 1 0 1 5-1026, 1 990 Tanaka T, Takeuchi Y, Shiohara T, Kitamura F, Nagasaka Y, Hamamura K, Yagita H, Miyasaka M: In utero treatment with monoclonal antibody to IL-2 receptor {3-chain completely abrogates development of Thy-I + dendritic epidermal cells. Int Immunol 4 : 4 8 7- 4 9 1 , 1 99 2 Ishida Y, Nishi M, Taguchi 0, lnaba K, Minato N, Kawaichi M, Honjo T: Effects of the deregulated expression of human interleukin-2 in transgenic mice. Int

cytes are derived from a common disrinct lineage . ] ImmunoI 1 993

Acad Sci 6 5 1 : 2 1-3 2 , 1 992

26. 27. 28.



Immunol 1 : 1 1 3-1 2 1 , 1 98 9 31.

Schorle H, Holtschke T, Hiinig T, Schimpl A, Horak I: Development and fimction of T cells in mice rendered interleukin-2 deficient by gene targeting. Nature 3 5 2 : 6 2 1 - 62 4 , 1 9 9 1


33. 34.

Matsue H, Bergstresser PR, Takashima A: Reciprocal cytokine-mediated cellular interactions in mouse epidermis: promotion of yo T-cell growth by IL-7 and TNFll' and inhibition of keratinocyte growth by y-IFN. ] Invest Dermatol







1 99 1

Eghtesady P, Brorson KA, Cheroutre H, Tigelaar RE, Hood L, Kronenberg M : Expression of mouse Tl a region class I genes in tissues enriched for y o cells.

56 .

Haas W, Pereira P, Tonegawa S: Gamma/delta cells.

Immunogenetics 3 6 : 3 77-388, 1 992

KL ,


Eaton GJ: The growth of hair follicles in waves.

Ann NY Acad Sci

83: 3 65-3 6 8 , 1 9 5 9 38.


Eaton GJ: Hair growth waves and cycles i n nude mice. Proceedings o f the Second International Workshop on Nude Mice. University of Tokyo Press/Gustav Fischer Verlag, Stuttgart, 1 9 77, pp 8 9 -93 Wysolmerski JJ, Broadus AE, Zhou J , Fuchs E, Milstone LM, Philbrick WM: Overexpression of parathyroid hormone-related protein in the skin of trans­ genic mice interferes with hair follicle development. Proc Natl Acad &i USA 9 1 : 1 1 33-1 1 3 7, 1 994


Sundberg JP: The crinkled


mutation, chromosome


In: Sundberg JP (ed.). Boca Raton, FL,

Handbook of Mouse Mutations With Skin a n d Hair Abnormalities. 41.


43 . 44.

CRC, 1 994, pp 221-229 Sundberg JP: The angora (go) mutation, chromosome 5 . In: Sundberg JP (ed.). Handbook of Mouse Mutations With Skin and Hair Abnormalities. Boca Raton, FL, CRC, 1 9 94, pp 1 6 5-1 70 Sundberg JP, Orlow SJ: The hairy ears (Eh) mutation, chromosome 1 5. In: Sundberg JP (ed.) . Handbook of Mouse Mutations With Skin a nd Hair Abnormal­ ities. Boca Raton, FL, CRC, 1 994, PP 3 1 9 -3 2 2 Foster CA, Yokozeki H, Rappersberger K, Koning F, Vole-Platzer B, Rieger A, Coligan JE, WolffK, Stingl G: Human epidermal T cells predominately belong to the lineage expressing ll'/f3 T cell receptor. ] Exp Med 1 7 1 :997-1 0 1 3 , 1 99 0 Dupuy P, Heslan M, Fraitag S, Hercend T, Dubertret L, Bagot M: T-cell receptor-y/o bearing lymphocytes in normal and inflammatory human skin. J




C U lT Opin

Sciammas R, Johnson RM, Sperling AI, Brady W, Linsley PS, Spear PG, Fitch FW, Bluestone JA: Unique antigen recognition by a Herpesvirus-specific TCR-yo cell. ] ImmunoI 1 52 : 5392-5 3 9 7 , 1 99 4 Tanaka Y, Sano S, Nieves E, De Libero G, Rosa D, Modlin RL, Brenner MB , Bloom B R , Morita CT: Nonpeptide ligands for human y o T cells. Proc Natl


Constant P, Davodeau F, Peyrat M-A, Poquet Y, Puzo G, Bonneville M , Fournie J-J: Stimulation of human yo T cells by nonpeptidic mycobacterial ligands.


Janeway CA Jr, Bottomly K: Siguals and sigus for lymphocyte responses.

Science 264:267-270, 1 994

Born WK, Harshan K, Modlin RL, O'Brien RL: The role of yo T lymphocytes in infection. CUlT Opin Immunol 3 : 4 5 5- 4 5 9 , 1 9 9 1


76:275-2 8 5 , 1 9 9 4 62.

Roberts K, Yokoyama W, Kehn PJ, Shevach E: The vitronecrin receptor serves as an accessory molecule for the activation of a subset of yo T cells. ] Exp Med

63 .

Ohno H, Ono S, Hirayama N, Shimada S, Saito T: Preferential usage of the Fe receptor l' chain in the T cell antigen receptor complex by 1'/0 T cells localized in epithelia. ] Exp Med 1 7 9 : 3 6 5 , 1 99 4 Matsue H, Cruz PD Jr , Bergstresser PR, Takashima A: Profiles of cytokine mRNA expressed by dendritic epidermal T cells in mice. ] Invest Dermatol

1 73 : 2 3 1-240, 1 9 9 1


1 0 1 : 5 3 7-542, 1 993 65.

Paul WE, Seder


Lymphocyte responses and cytokines.

Cell 7 6 : 2 4 1 -2 5 1 ,

1 994

66. 67.

Boismenu R, Havran WL: Modulation of epithelial cell growth by intraepithelial yo T cells. &ience 2 66 : 1 253-1 2 5 5 , 1 9 9 4 Vole-Platzer B, Anegg B, Milota S, Pickl W, Fischer G: Accumulation of yo T cells in chronic cutaneous lupus erythematosus.] Invest Dermatol l 0 0 : 84S-9 1 S , 1993


Mombaerts P, ArnoldiJ, Russ F, Tonegawa S, Kaufmann SHE: Different roles of 0if3 and yo T cells in immunity against an intracellular bacterial pathogen.


Chandler P, Frater AJ, Viney JL, Philpott KL , Kay G, Millrain M, Hayday AC, Owen MJ, Simpson E. Altmann OM: Immune responsiveness in mutant mice lacking T cell receptor ll'f3+ cells. Int Immunol (in press) Tigelaar RE, Lewis JM, Nixon-Fulton JL, Bergstresser PR: Thy-1 antigen­ bearing dendritic epidermal cells in mice. In: Goldsmith L (ed. ) . Physiology, Biochemistry, and Molecular Biology of the Skin. Oxford University Press, Oxford, 1 99 1 , pp 1 1 64 -1 1 8 7 Love-Schimenti CD, Kripke ML: Dendritic epidermal T cells inhibit T-cell proliferation and may induce tolerance by cytotoxicity. ] Immunol 1 5 3 : 3 4 5 0 -

Nature 3 6 5 : 5 3-56, 1 993



3 4 5 6 , 1 994

A nnu Rev Immunol 8 : 4 0 1 - 420,

1 990 47.

Havran WL, Boismenu R: Activation and function of yo T cells. Immunol 6:442- 446, 1 9 94

Invest Dermatol 94:764 -768, 1 990

45. Janeway CAJr, Jones B, Hayday A: Specificity and function ofT cells bearing yo receptors. Immunol Today 9 : 73-76, 1 9 88 Young RA : Stress proteins and inununology.

Annu Rev Immunol

1 1 :637-68 5 , 1 9 9 3

Acad Sci USA 9 1 : 8 1 7 5- 8 1 79 , 1 994

Shaw SK, Parker CM, Russell GJ, Morrow JS, Rimm DL, Brenner MB: Adhesion between epithelial cells and T lymphocytes mediated by E-cadherin and the ll'Ef37 integrin. Nature 3 7 2 : 1 9 0 -1 9 3 , 1 994 Shiohara T, Moriya N, Gotoh C, Hayakawa J, Nagashima M, Saizawa K, Ishikawa H: Loss of epidermal integrity by T cell-mediated attack induces long-tenn local resistance to subsequent attack. I. Induction of resistance correlates with increases in Thy-l + epidermal cell numbers. ] Exp Med Chase

Germain RN: MHC-dependent antigen processing and peptide presentation: providing ligands for T lymphocyte activation. Cell 7 6 : 2 87-299, 1 994 Wu M, Van Kaer L, Itohara S, Tonegawa S: Highly restricted expression of the thymus leukemia antigens on intesrinal epithelial cells.] Exp Med 1 74 : 2 1 3-2 1 8,


1 7 1 : 1 027-1 04 1 , 1 990


Havran WL, Chien Y-H, AllisonJP: Recognition of self antigens by skin-derived T cells with invariant yo antigen receptors. ScielUe 2 5 2 : 1 43 0 -1 4 3 2 , 1 9 9 1 Tigelaar RE, Lewis JM: Factors involved in the localization and activation of murine 1'/0 dendritic epidermal T cells. In: Lambert WC, Giannotti B, van Vloten W (eds.). Basic Mechanisms of Physiologic and Aberrant LymphoproliJeration in the Skin. Plenum Press, New York, 1 994, pp 3 9 -5 5 Kaminski MJ, Bergstresser PR, Takashima A: In vivo activation o f mouse dendritic epidermal T cells in sites of contact dermatitis. Eur ] Immunol 2 3 : 1 7 1 5-1 7 1 8, 1 9 9 3


Immunol 2 4 : 63 5- 640, 1 9 94


O'Brien RL, Born W: Heat shock proteins as antigens for yo T cells. In: Allison J (ed.). Seminars in Immunology: yo T Cells. Philadelphia, WB Saunders, 1 9 9 1 , pp 81-88 O'Brien RL, Fu Y-X, Cranfill R, Dallas A, Ellis C, Reardon C, Lang J, Carding SR, Kubo R, Born W: Heat shock protein Hsp60-reactive yo cells: a large, diversified T-lymphocyte subset with highly focused specificity. Proc Natl Acad Sci USA 8 9 : 4 3 4 8 - 4 3 5 2 , 1 992



1 0 1 : 543-5 4 8, 1 993

Leclercq G, De Smedt M, Plum J: Cytokine production and responsiveness of fetal T-cell receptor Vy3 thymocytes. &and] ImmunoI 36:833- 8 4 1 , 1 9 92 Lefranl'ois L, Barrett TA, Havran WL, Puddington L: Developmental expression of the ll'IELf37 integrin on T cell receptor yo and T cell receptor ll'f3 T cells. Eur



Shiohara T, Moriya N, Hayakawa J , Ishikawa H, Itohara S : Resistance to cutaneous graft-versus-host disease is not induced in TCR S gene mutant mice. ] Exp Med (in press)