Skin-Associated Lymphoid Tissues (SALT): Origins and Functions

Transcription

1 REPORTS Skin-Associated Lymphoid Tissues (SALT): Origins and Functions J. WAYNE STREILEIN, M.D. The skin has an unusual set of immunologic requirements. It is confronted by a specialized set of pathogenic organisms and environmental chemicals that represent a distinctive spectrum of antigenic specificities. Skin is subjected to physicochemical stresses such as irradiation with ultraviolet light that alter dramatically its immunologic properties. It is proposed that nature has provided skin with a unique collection of lymphoid cells, reticular cells, and organized lymphoid organs to deal with these special demands. Evidence in favor of the existence of skin-associated lymphoid tissues (SALT) includes (1) the cutaneous microenvironment is capable on its own of accepting, processing, and presenting nominal antigen; (2) strategically located peripheral lymph nodes are able to accept immunogenic signals derived from skin; (3) subsets of T lymphocytes display differential affinity for skin and its associated peripheral nodes; and (4) acquisition of this affinity by T cells is determined at least in part by differentiation signals received in situ from resident cutaneous cells. Responsibility for the establishment and integration of SALT rests with keratinocytes, Langerhans cells, and immunocompetent lymphocytes, each of which contributes uniquely to the synthesis. Together they provide skin with immune surveillance that effectively prejudices against the development of cutaneous neoplasms and persistent infection with intracellular pathogens. In patients who have been under long-term immunosuppressive therapy, the large majority of nonlymphoid malignancies arise within the skin, rather than other types of tissues. These data suggest that immune surveillance, once thought to be an immune defense operative in all somatic tissues, is a specialized immune function dedicated to the skin and mediated by SALT. The concept of immune surveillance, formulated initially by Thomas [1] and amplified subsequently by Burnet [2], ascribes to the immune system a function that transcends mere recognition and response to exogenous pathogenic agents. It proposes that recirculating immunocompetent lymphocytes continually survey somatic tissues for antigenic evidence of malignant transformation of cells. When neoplastically transformed cells are recognized by immunocompetent cells, the former are swiftly destroyed, thereby quenching potential neoplasms before they even reach clinical expression. At a time when immunologic were reeling from the discovery of neonatal transplantation tolerance, with all that it implied for autoimmunity and active selfrecognition, immune surveillance offered new hope for a cent ralizing synthesis. The attractiveness of this hypothesis has permitted it to survive even today despite apparently lethal blows to its validity. Not only is there virtually no definitive experimental evidence to verify the surveillance hypothesis, there is strong circumstantial evidence that no such general mechanism operates. In immunodefieient individuals made thus by prolonged immunosuppressive therapy or on a genetically determined basis there is no widespread expression of malignancies of various and differing somatic tissues as one might have predicted if the surveillance hypothesis were correct. Instead, the immunodefieient state seems to display a high incidence of malignancies only for cells of the lymphoreticular system. Departments of Cell Biology and Internal Medicine, Southwestern Medical School and University of Texas Health Science Center, Dallas, Texas, U.S.A. This work was supported in part by U.S. Public Health Service Grants AI and AM Reprint requests to: Dr. J. Wayne Streilein, Department of Cell Biology, University of Texas Health Science Center, 5323 Harry Hines Blvd., Dallas, Texas Abbreviations: CTCL, cutaneous T-cell lymphomas/leukernias; DNFB, dinitrofluorobenzene; DTH, delaved-type hypersensitivity; ETAF, epidermalcell-derived thymocyte-activating factor; SALT, skin-associated lymphoid tissue; UV-B, ultraviolet light B Recently, careful examination of the records of the Kidney Transplantation Registry has revealed that patients who have withstood prolonged immunosuppressive therapy actually have increased risk of the development of two types of malignancies: lymphoreticular and cutaneous [3]. Capitalizing on this realization, Mitchison [4] has suggested that immune surveillance may exist, but that its scope is limited to the integument. Demographic evidence combined with knowledge about the distributions of cutaneous neoplasms in these patients further suggest that exposure to sunlight is an important contributing factor. Thus the idea was spawned that a special relationship may exist between elements of the immune system and the skin with relevance to its unique exposure to actinic radiation and environmental pathogens. In this review I would like to assemble some of the important observations that bear on this special relationship. In so doing, it is proposed that skin-associated lymphoid tissues (SALT) exist as a functional unit that provides skin with a unique surveillance mechanism [5]. The evidence to be presented will make the following general points: 1. Some, but not all, T lymphocytes display special affinity for skin and draining lymph nodes. 2. Skin contains immunocompetent lymphocytes and antigen-presenting cells and produces immunoregulatory molecules. 3. Immune recognition of antigen takes place within skin. 4. Antigen that escapes intracutaneous recognition induces specific unresponsiveness. EPIDERMOTROPISM OF T LYMPHOCYTES Within the past 5 years, especially now that reagents exist to accurately identify types of blood lymphocytes, it has been determined that certain T-lymphocyte malignancies show a distinct propensity for infiltration of the skin. The term cutaneous T-cell lymphomas/ leukemias (CTCL) has been used to identify this phenomenon [6]. By implication, non-t-cell neo-plastic disorders of lymphocytes show no 12s

2 June 1983 SKIN-ASSOCIATED LYMPHOID TISSUES special affinity for skin. Even among T cells, the affinity for skin is not generic, in that in the vast majority of cases in which malignant T cells involve the epidermis, the surface phenotype of the infiltrating cells coincides with the helper/inducer subpopulation [7]. In lesions examined microscopically, these cells demonstrate a striking tendency to localize around marrow-derived dendritic cells that express Ia-like surface antigens [8]. These observations are consistent with the hypothesis that CTCLs represent the neoplastic amplification of T lymphocytes that recirculate; normally through the blood and into the skin. Presumably, other types of normal T cells do not possess this trafficking capability; consequently, one might expect their neoplastic representatives to infiltrate other types of tissues. Based in part on these considerations, several groups of investigators have attempted to collect data using nonneoplastic (i.e., normal) T cells seeking evidence for physiologic epider-motropism of normal T lymphocytes. Radiolabeled unfraction-aterl normal lymphoid cells from mesenteric and peripheral lymph nodes exhibit slightly different homing patterns. Mesenteric lymph node cells tend to localize in the gut, while peripheral lymph node cells tend to home toward peripheral lymph nodes [9]. This difference in homing patterns has been greatly amplified by employing activated lymphoblasts rather than unfractionated or resting lymphocytes. Griscelli and coworkers [10] first demonstrated that radiolabeled blast cells obtained from rat mesenteric lymph nodes accumulate after adoptive transfer in highest number in the gut wall and the mesenteric lymph nodes. By contrast, peripheral node lymphoblasts selectively localize in peripheral lymph nodes at the expense of the gut and the mesenteric nodes. Rose [11] has determined that T lymphoblasts generated within peripheral lymph nodes will preferentially infiltrate skin when given the choice of inflammatory sites simultaneously present in skin and gut. As expected, mesenteric T lymphoblasts display the opposite homing pattern. Asherson [12] and Tigelaar (personal communication) have employed oxazolone-stimulated peripheral lymph node blast cells to compare their ability to enter sites of allergic contact dermatitis. Blast cells enter these sites with much greater facility than do unstimulated or small lymphocytes. To determine whether their attraction to the skin test site was related in any way to their ability to recognize haptenic antigen, Tigelaar has recently produced alloantigen-activated T-cell blasts and examined their migratory potential in recipients in whom cutaneous inflammation was induced with hapten. The results clearly indicate that the migratory properties of T lymphoblasts are independent of the antigen specificity of the activated cells. T lymphoblasts activated in peripheral lymph nodes of F 1 hybrids demonstrate a distinct proclivity for returning to peripheral lymph nodes of syngeneic mice, and they preferentially infiltrate cutaneous, as opposed to gastrointestinal, sites of hapten-induced inflammation. Finally, Liden [13] has accumulated evidence to suggest that the localization of T cells in the skin leads to preferential infiltration by these cells into the epidermis itself. Thus, once localized within the dermis, sub-populations of T cells display a propensity to migrate across the dermal-epidermal junction and infiltrate the epidermis. While this mass of evidence remains largely circumstantial, it is impressive and strongly supports the contention that subpopulations of T lymphocytes exhibit epidermotropism. THE IMMUNOLOGIC ENDOWMENT OF SKIN The ingredients thought to be necessary for successful induction of immune response to antigen include immunocompetent lymphocytes, antigen-presenting cells, and molecules with the capacity to influence the activation of lymphocytes. The next evidence cited indicates that these essential ingredients exist not only within skin, but within the epidermis itself. For the thesis being developed, it is important to know whether T cells within the skin possess immunocompetence in situ. This is a difficult point to document because of technical considerations. One experimental approach to this issue has been to exploit alloimmune reactivities by grafting skin ortho-topically from an inbred laboratory animal (chicken, in the ease of Solomon [14]; mice, in the case of Barker and Billingham [15]) to its appropriate F 1 hybrid. Solomon was able to show that contrary to the apparent rules of transplantation genetics, such grafts would be rejected. With appropriate controls, he was able to conclude that the rejection process which the F 1 hybrid is genetically incapable of mounting in its own right stemmed from the vigor of the local response that resident lymphocytes of the graft made against the histocompatibility antigens of the recipient. Once the process was initiated, nonspecific inflammatory forces procured the destruction of the graft. A perhaps more convincing demonstration of the immunocompetence of lymphocytes present within normal skin was provided by the experiments of Barker and Billingham. Parentalstrain skin was grafted orthotopically to F 1 hybrid mice, whose draining regional lymph nodes were extirpated and weighed 7 days later. Specific nodal hypertrophy developed under circumstances in which the only explanation could be that immunocompetent cells present within the graft traveled to the host s draining lymph nodes and initiated a local graft-versus-host reaction. These findings confirm that immunocompetent lymphocytes exist within skin under physiologic conditions and in sufficient number to cause observable reactions to appropriate antigens. Much has been learned over the past decade about the immunologic properties of epidermal Langerhans cells [16]. These cells can be identified with the light microscope by aid of cell-surface ATPase staining as well as with immunofluorescence markers; the cell is uniquely distinguished under the electron microscope by the presence within its cytoplasm of unique granules described originally by Birbeck. While the origin of Langerhans cells was shrouded in mystery for many years, it now seems fairly clear that this cell type is derived from a progenitor that resides within the marrow. In histologic preparations, Langerhans cells assume a suprabasilar location within the epidermis, where they form a horizontally disposed network of dendritic cells. In our laboratories we have quantified the distribution and density of Langerhans cells in various types of skin. We have learned that there is a relatively homogeneous distribution of Langerhans cells throughout the body wall skin of most areas. However, the cornea of normal eyes is devoid of Langerhans cells, except at its limbic attachment, and the density of Langerhans cells in the scaled portions of murine tail skin is exceedingly low. In addition, the density of Langerhans cells in the epithelium of the hamster cheek pouch is approximately 20 percent of that found in normal hamster skin [17]. Langerhans cells display on their surfaces receptors for the Fc component of immunoglobulin and for the C3b component of complement. These cells also express an extremely high density of surface markers encoded by the I region of the mouse major histocompatibility complex (H-2) and the D region of HLA. Over the past several years, a new lineage of leukocytes has been defined by Steinman and Nuzzenzweig [18] (using mice) and corroborated by Mason and his colleagues [19] (using rats): Dendritic cells, of which Langerhans cells are very likely to be a subset, are pervasively located throughout (apparently) all organs of the body, chiefly in the connective-tissue stroma. All dendritic cells share several important properties: (1) they contain high concentrations of surface Ia antigens; (2) they stimulate exceedingly well in mixed lymphocyte reactions, and (3) they present nominal antigens to primed T lymphocytes. Langerhans cells display all these properties, suggesting that a physiologic role for these cells in skin might be antigen presentation. It was therefore gratifying, in our experiments with cutaneous surfaces depleted of Langerhans cells, to find that the ability of these surfaces to promote the induction of contact hypersensitivity was very reduced [20]. Thus epidermal Langerhans cells, along with similar cells located within the dermis, provide skin with an extremely efficient cellular mechanism for the presentation and processing of antigen to immunocompetent lymphoid cells. One effective method we have employed to reduce the number of normal Langerhans cells in the epidermis of mouse skin is to irradiate 13s

3 STREILEIN Vol. 80, No. 6 Supplement with ultraviolet light B. When skins of C3H/HeJ and C57BL/6 mice are irradiated with UV-B and then painted with immunizing doses of DNFB, the animals fail to become sensitized. However, when comparable treatment protocols are administered to mice of the BALB/c strain, sensitization of normal intensity takes place. We have now examined numerous inbred mouse strains in this regard; the results are listed in Table I. The data indicate that, susceptibility of skin to UV-B in the induction of contact hypersensitivity is under genetic control. Since F 1 hybrids from suppressive and nonsuppressible parents are suppressible, the trait is apparently dominant. However, when two suppressible strains (B6 and C3H) are crossed, the F 1 hybrid is now resistant to the effects of UV-B, suggesting that gene complementation is at work and that more than one locus is involved. As the data in Table II indicate, analysis of suppressibility of segregant populations in mice reveals that at least two, and more likely three, distinct, unlinked genetic loci govern this effect. Preliminary studies with H-2 congenic mice strongly implicate this important chromosomal segment in the process. We are very encouraged by these experiments, because they offer hope of understanding not only the genetic basis of susceptibility to UV-B but also perhaps the genetic mechanisms that govern antigen recognition within skin. Finally, we have succeeded in drastically depleting the number of resident Langerhans cells in epidermis by stripping with repeated applications of cellophane tape. When skin from C3H mice has been stripped in this fashion and then painted with DNFB, the animals fail to develop contact hypersensitivity; by contrast, similar stripping of the skin of BALB/c mice fails to interfere with the induction of contact hypersensitivity. The difference in susceptibility to tape stripping in these two strains of mice resembles that seen when they are exposed to ultraviolet B radiation. We are currently entertaining the hypothesis that in C3H mice, Langerhans cells alone serve as the antigen presenting cells in the induction of contact hypersensitivity. By contrast, BALB/c mice possess a secondary, or fail-safe, antigen presenting system that can be brought into play when the Langerhans cell network has been disrupted. It will be important to determine whether these differences are genetically determined by the loci that we have identified as being responsible for the ultraviolet B effect. TABLE I. Effect of UV-B on induction of contact hypersensitivity Suppressible strains Nonsuppressible strains C57BL/6 C57BL/10 CSH/HeJ A/J BALB/c DBA/2 (B6 BALB/c)F 1 (B6 C3H)F 1 (Bfi A/J)F 1 (BALB/c C3H)F 1 TABLE II. Genetic basis of UV-B effect on induction of contact hypersensitivity Segregant population: (C57BL/6 A/J)F 1 A/J Resistant (470% positive control) 65% Intermediate (30 50% positive control) 18% Suppressible (o10% positive control) 18% N=40 In the induction of immune responses, regulatory molecules are released from antigen presenting cells as well as from activated lymphocytes. These factors further amplify, recruit, and modulate the developing response. In the skin, production of a molecule that resembles to a striking degree interleukin-1 has been documented and termed epidermal-cell-derived thymocyte-activating factor (ETAF) [21,22]. This locally secreted molecule is unable to stimulate mature T cells directly, but it can activate them in the presence of antigen. Surprisingly, the cellular origin of this molecule appears to be the keratinocyte rather than the Langerhans cell. We have thus identified within skin, and probably even within epidermis itself, the essential ingredients for effective recognition of nominal antigen. ANTIGEN RECOGNITION TAKES PLACE WITHIN SKIN Given the facts that certain subpopulations of lymphocytes have a predilection for entering the skin and that the skin contains cells capable of effective antigen presentation, including local secretion of immunomodulatory molecules, it remains to be determined whether effective antigen presentation actually takes place within the confines of the skin. Several lines of evidence bear on this important issue. In the late 1950s and during the 1960s, a common experimental maneuver in transplantation immunobiology was to inoculate monodisperse suspensions of lymphocytes into the skin and observe the appearance of a local inflammatory reaction. The so-called normal lymphocyte transfer reaction, first examined by Brent and Medawar [23], was achieved by inoculating parental-strain lymphoid cells intradermally into the body wall of F 1 hybrid guinea pigs. Local delayed inflammatory reactions regularly ensued, indicating that a local graft-versus-host reaction had taken place. At its most elementary level, this result documents that recognition of antigen (in this case, alloantigens displayed on constitutive dermal cells) was achieved in situ by the inoculated lymphocytes. Moreover, the elicitation of a delayed inflammatory reaction signifies that an effector response was also generated; that is, antigen recognition by lymphocytes proceeded to that stage in the efferent limb of the immune response where recruitment of nonspecific host cells was achieved, and a typical inflammatory reaction ensued. This approach was refined by Ramseier and Streilein, who demonstrated that the skin of lethally irradiated Syrian hamsters could serve as a suitable milieu in which alloreactive lymphoid-cell mixtures could be placed and generate local immune reactions [24]. To test this idea more directly, a series of painstaking and elegant experiments by Macher and Chase [25] gave the first indication that meaningful antigen recognition leading to the expression of contact hypersensitivity takes place at the site at which hapten is injected. Using DNCB injected into the ears of guinea pigs, these investigators discovered that the site of application had to remain intact during the first 24 hr after skin painting. Excision of the ear prior to the end of the first 24-hr interval failed to achieve sensitization. These investigators reasoned that this length of time was probably required for lymphocytes, trafficking through the haptenated skin, to recognize hapten and thereby become activated a process termed peripheral sensitization. The approach we have taken to answering this important question has been to make use of the observation that hapten-derivatized skin as a graft can be used as a source of immunogen to induce contact hypersensitivity. In typical experiments, shaved abdominal-wall skin is painted with an immunizing dose of DNFB; grafts are then prepared and placed orthotopically on recipient mice. Six days later the ears of the recipient mice are challenged with dilute DNFB to determine whether the animals have developed contact hypersensitivity. Using this approach, we have learned that very much smaller doses of hapten are required to induce systemic sensitization compared with conventional immunizing protocols. Thus BALB/c mice can be rendered sensitized to DNFB by grafting them with syngeneic; skin that has been painted with as little as 20 ml of 0.02% DNFB approximately 60 times less than the amount applied during conventional sensitizing maneuvers. Using this limiting dose of hapten applied to skin, we then 14s

4 June 1983 SKIN-ASSOCIATED LYMPHOID TISSUES TABLE III. Induction of DNFB contact hypersensitivity in BALB/c mice with syngeneic, semiallogeneic, and allogeneic derivatized skin Peak ear swelling Experimental pane N ( 10 4 in±sem) DNFB-BALB/c skin 4 24±2 DNFB-(BALB/c B6)F 1 skin 4 16±2 DNFB-C57BL/6skin 4 7±2 Positive control 4 32±4 Negative control 4 6±1 TABLE IV. Capacity to reimmunize BALB/c mice with DNFB following initial exposure with derivatized syngeneic, semiallogeneic, and allogeneic skin Peak ear swelling Experimental panel N ( 10 4 in ± SEM) DNFB-BALB/c skin 4 82±4 DNFB-(BALB/c B6)F 1 skin 4 76±5 DNFB-C57BL/6 skin 4 31±2 Positive control 4 32±2 Previous negative control 4 37±1 New negative control 4 4±1 devised an experiment in which recipient mice received haptenderivatized skin from syngeneic, semiallogeneic, and fully allogeneic donors (see Table III). When the recipient mice were ear-challenged subsequently, only those animals which received hapten-derivatized skin bearing the host s own H-2 alloantigens displayed significant ear swelling. Recipients of haptenated H-2 incompatible skin grafts developed ear-swelling responses indistinguishable from negative controls. This observation confirms that at limiting doses of hapten, antigen recognition takes place within the skin graft itself. As a consequence, effector T cells are restricted in their ability to express or recognize the hapten by H-2 determinants expressed on the immunizing graft. In this instance, since the H-2 determinants at the ear-challenge site are of recipient type, only derivatized grafts expressing recipient H-2 determinants evoked DTH effector cells able to produce lesions at the challenge site. In a subsequent experiment, mice grafted in this manner were subjected to a reimmunization procedure (see Table IV). Animals that first experienced hapten in the form of derivatized syngeneic or semiallogeneic grafts gave exaggerated responses indicative of a previously immunized state. By contrast, animals first grafted with haptenated allogeneic skin responded to the second immunization regimen by expressing contact hypersensitivity similar in intensity to that of the primary positive controls. This result indicates that at limiting doses of hapten, presentation on H-2 allogeneic skin is an immunologically null event. Since these animals displayed no evidence of unresponsiveness, we would argue that little, if any, antigen reprocessing takes place on the recipients own antigen presenting cells. To our minds, we regard this as the strongest evidence to date that the phenomenon of peripheral sensitization occurs; moreover, it strongly suggests that the entire immunity sequence from antigen presentation through effectation of the response takes place within the cutaneous environment itself. ANTIGEN THAT ESCAPES INTRACUTANEOUS PRESENTATION INDUCES SYSTEMIC UNRESPONSIVENESS It is axiomatic among immunologists that some, but not all, exposures to antigenic substances lead to specific sensitization. When translated into experimental systems in laboratory rodents, this idea means that within a certain dosage range, contact hypersensitivity can be induced by skin painting with reactive haptens. When we elected to use ultraviolet radiation as a means of reducing epidermal Langerhans cells, we did so with the full expectation that we would impair the induction of contact hypersensitivity. And that is precisely what we found. However, while the thought had crossed our minds, we were still surprised to discover that a dose of DNFB that is immunogenic when applied to normal murine bodywall skin proves to be tolerogenic when applied to skin depleted of (or naturally deficient in) normal Langerhans cells [20]. This observation has profoundly altered the way we now consider induction of contact sensitivity through cutaneous surfaces. In the first place, it raises the possibility that doses of hapten that had heretofore been regarded assubimmunogenic might instead be tolerogenic. In the second place, it forces us to abandon the naive view that application of doses of hapten (and other types of eutaneously applied antigens) that fail to evoke specific sensitization are immunologically null events. Instead, it would appear that antigens that gain access to the body other than through the skin have a propensity for inducing unresponsiveness rather than sensitization. From this vantage point, the cutaneous surface may now be regarded as having been devised immunologically to prevent this occurrence by tipping the balance of the immune response in favor of sensitization. At present we and others are in the process of examining the cellular and molecular bases of the unresponsive state induced by painting hapten on skin that has been exposed to low-dose ultraviolet radiation. Experiments conducted chiefly by Dr. Craig Elmets in our laboratory have indicated the following attributes of the unresponsiveness achieved by painting DNFB on skin pretreated with ultraviolet light: (1) unresponsiveness is specific for the hapten in question; (2) unresponsiveness is achieved only if hapten is applied to the site of irradiation; application of hapten to an unirradiated site on an irradiated animal achieves normal sensitization; (3) the unresponsive state can be transferred adoptively to naive syngeneic animals by infusions of spleen and lymph node cells from animals that have been rendered unresponsive; (4) the responsible cell type in the adoptive transfer inocula is a T cell that is capable of inhibiting the afferent limb (induction) of the contact-hyper-sensitivity response; and (5) the T suppressor cells are unable to interfere with the expression of contact hypersensitivity and therefore do not work on the efferent limb. It is important to note that the functional attributes and surface phenotype of the suppressor cell induced by low-dose UV-B and DNFB skin painting resembles to a remarkable degree the suppressor cells generated when DNBS is inoculated into mice by the intravenous route. The similarities of these two regimens for the induction of unresponsiveness, combined with the realization that animals that are susceptible to UV-B irradiation effects do not appear to have a fail-safe secondary mechanism for processing of epicutaneously applied antigen, suggests that if epicutaneously applied haptens fail to meet resident Langerhans cells for antigen presentation, they gain access directly to the vascular tree and induce systemic unresponsiveness. It comes as no surprise that genes within the major histocompatibility complex might be expected to play a role in guiding the sensitivity or unresponsiveness in this system. SALT: A UNIQUE IMMUNE SURVEILLANCE SYSTEM FOR THE SKIN The data heretofore presented have been selected to make a strong argument in favor of the hypothesis that skin-associated lymphoid tissues (SALT) exist as an integrated system of immune surveillance designed uniquely for the skin. SALT is comprised of (1) a specialized set of antigen presenting cells within the epidermis, Langerhans cells; (2) distinctive populations of recirculating T lymphocytes that 15s

5 STREILEIN Vol. 80, No. 6 Supplement preferentially infiltrate the skin, especially the epidermis; (3) keratinocytes that provide an anatomically distinct environment for these lymphoreticular cells and secrete into that environment immunoregulatory molecules that can profoundly affect the consequences of immune recognition and differentiation; and (4) a set of draining peripheral lymph nodes, integrating this multicellular system, that contain, along with the dermis, blood vessels with endothelial cells whose surfaces capture lymphocytes passing through the blood. In an anatomic sense, skin is uniquely situated, and as a consequence, it faces a unique spectrum of antigenic demands. The skin is bombarded daily with the potentially damaging effects of ultraviolet radiation, raising the risk of neoplastic transformation of cells, especially within the epidermis. Neoplastic transformation, especially that induced by ultraviolet light, is associated with expression of unique neoantigens on the surfaces of the neoplastic cells. It is within this context that a formulation of SALT takes on special meaning. It is proposed that SALT represents the physiologic mechanism created to deal with this special demand placed on skin. The scenario by which SALT operates might be stated as follows: Neoplastic transformation of (for example) a keratinocyte results in the expression of surface neoantigen(s). When transferred to the ubiquitous Langerhans cells, the neoantigen is processed and presented in immunogenic form. One of two subsequent events may then transpire: either (1) peripatetic T lymphocytes with predetermined affinity for the epidermis and with immunologic Specificity for the neoantigen migrate into the epidermis, recognize the neoantigen presented on Langerhans cells, transform in situ into effector cells, and directly destroy the adjacent malignant cells, or (2) the neoantigen-bearing Langerhans cell drops down into the dermis, flows with the draining lymph into the regional lymph node, settles into the cortex, and forms a nidus for the attraction of antigenspecific lymphocytes that are activated within the node, proliferate, and then disseminate systemically to return predominantly to the skin in order to effect, destruction of the neoplastic keratinocytes. We are beginning to collect relevant data supporting the second of these options. However, to the present, essentially no information exists that indicates that T cells can be activated directly in situ, can destroy the offending cells, and yet never send this signal to be amplified and disseminated systematically. It is conceivable that this may be the fundamental mechanism by which SALT works. Selecting ultraviolet radiation as the oncogenic agent in this context does not imply an exclusive role for this agent. Oncogenic viruses as well as other carcinogenic agents could act. similarly to induce cutaneous neoplasia. Ultraviolet light, however, has other effects on the skin that make it particularly appropriate in this consideration. Ultraviolet light induces melanization of skin and alters remarkably the functional properties of Langerhans cells with regard to antigen presentation. We would presume that under normal circumstances the immediate effects of ultraviolet, light (that impair antigen presentation by Langerhans cells) prove to be transient. This is so because activation of melanogenesis by ultraviolet light rapidly provides the epidermis with a protective photoshield that allows Langerhans cells to restore their immune functional properties even in the face of continued exposure to ultraviolet light. Failure of SALT could occur through numerous pathways. Inadequate melanization would permit cutaneous cells, including Langerhans cells, to be exposed chronically to the deleterious effects of ultraviolet light; in the case of Langerhans cells, UV-induced inability to process neoantigens effectively can lead to the establishment of specific unresponsiveness rather than specific sensitization. In this manner, chronic impairment of the presentation of neoantigens by Langerhans cells would promote, rather than prevent, the escape of neoplastic cells from detection by the immune system. Reduction in the numbers and varieties of antigen-specific recirculating lymphocytes, as might, be achieved by chronic immunosuppression, would similarly disrupt SALT, robbing the organism of the antigen-specific and tissue-destructive components of the surveillance mechanism. REFERENCES 1. Thomas L: Reactions to homologous tissue antigens in relation to hypersensitivity, Cellular and Humoral Aspects to the Hypersensitivity State. Edited by HS Lawrence. New York, Hoeber-Harper, 1959, p Burnet FM: Immunologic Surveillance. Oxford, Pergamon Press, Kinlen LJ, Sheil AGR, Peto J, and Doll R: Collaborative United Kingdom- Australasian study of cancer in patients treated with immunosuppressive drugs. Br Med J 2: , Mitchison NA, Kinlen LJ: Present concepts on immune surveillance. Prog Immunol 4: , Streilein.JW: Lymphocyte traffic, T-cell malignancies and the skin. J Invest Dermatol 71: , Edelson RL: Cutaneous T cell lymphoma: Mycosis fungoides, Sézary syndrome, and other variants. J Am Acad Dermatol 2:89 106, Janossy G, Tidman N, Selby WS, et al: Human T lymphocytes of inducer and suppressor tvpe occupy different microenvironmenls. Nature 288:81 84, Janossy G, Tidman N, Papagerogiou ES, et al: Distribution of T lymphocyte subsets in the human bone marrow and thymus: An analysis with monoclonal antibodies. J Immunol 126: , McWilliams M, Philips-Quagliata JM, Lamm ME: Characteristici of mesenteric lymph node cells homing to gut associated lymphoid tissue in syngeneic mice. J Immunol 115:54 58, Griscelli C, Vasalli P, McCluskey RT: The distribution of large dividing lymph node cells in syngeneic recipients after intravenous injection. J Exp Med 130: , Rose ML, Parrot DMV, Bruce RG: Divergent migration of mesenteric and peripheral immunoblasts to sites of inflammation in the mouse. Cell Immunol 27:36 46, Asherson GL, Allwood GG, Mayhew B: Movement of T blasts in the draining lymph nodes to sites of inflammation. Immunology 25: , Liden S: The mononuclear-cell infiltrate in allergic contact dermatitis: 3. Selective accumulation of cells from lvmph nodes. Acta Pathol Microbiol Scand 70: , Solomon JB: Ontogenesis of immunological competence of lymphocytes in chicken skin. Transplantation 1: , Barker CF, Billingham RE: Immunologically competent passenger cells in mouse skin. Transplantation 14: , Special issue on dendritic and lymphocytic cells in the epidermis, 29th Symposium on the Biologv of Skin. J Invest Dermatol 75:1 127, Bergstresser PR, Fletcher CR, Streilein JW: Surface densities of Langerhans cells in relation to rodent epidermal sites with special immunologic properties. J Invest Dermatol 74:77 84, Steinman RM, Nuzzenzweig MC: Dendritic cells, features and functions. Immunol Rev 53: , Mason DW, Pugh CW, Webb M: The rat mixed lymphocyte reaction: Roles of a dendritic cell in intestinal lymph and T-cell subsets defined by monoclonal antibodies. Immunology 44:75 87, Toews GB, Bergstresser PR, Streilein JW: Epidermal Langerhans cell density determines whether contact hypersensitivity or unresponsiveness follows skin painting with DNFB. J Immunol 124: , Luger TA, Stalder BM, Katz SI, Oppenheim JJ: Epidermal cell (keratinocyte)-derived thymoevte-activating factor (ETAF). J Immunol 127: , Sauder DN, Carter D, Katz SI, Oppenheim JJ: Epidermal cell production of thymocyte activating factor. J Invest Dermatol 79:34 39, Brent L, Medawar PB: Nature of the normal lymphocyte transfer reaction. Nature 204:90 91, Ranvseier H, Streilein JW: Homograft sensitivity reaction in irradiated hamsters. Lancet 1: , Macher E, Chase MW: Studies on the sensitization of animals with simple chemical compounds: XII. Influence of excision of allergenic depots on onset of delayed hypersensitivity and tolerance. J Exp Med 129: , s