Both the UVB and UVA wavebands within sunlight are immunosuppressive. This article reviews the relationship between wavebands and dose in UV-induced immunosuppression mainly concentrating on responses in humans. It also contrasts the effects of UVB and UVA on cellular changes involved in immunosuppression. Over physiological sunlight doses to which humans can be exposed during routine daily living or recreational pursuits, both UVA and UVB suppress immunity. While there is a linear dose relationship with UVB commencing at doses less than half of what is required to cause sunburn, UVA has a bell-shaped dose response over the range to which humans can be realistically exposed. At doses too low for either waveband to be suppressive, interactions between UVA and UVB augment each other, enabling immunosuppression to occur. At doses beyond where UVA is immunosuppressive, it still contributes to sunlight-induced immunosuppression via this interaction with UVB. While there is little research comparing the mechanisms by which UVB, UVA and their interactions can cause immunosuppression, it is likely that different chromophores and early molecular events are involved. There is evidence that both wavebands disrupt antigen presentation and effect T cell responses. Different individuals are likely to have different immunomodulatory responses to sunlight.
Professor Margaret Kripke instigated the entirely new field of photoimmunology when she showed that UV radiation suppresses the ability of the immune system to reject skin cancers in mice. She made a large number of pivotal observations in this field, including that UV immunosuppression is important for enabling primary skin tumors to develop. Since then a large number of major advances have been made. It has been established that sunlight has two major biological effects responsible for causing skin cancer in humans—immunosuppression and gene modulations that alter cell growth and homeostasis. Some aspects of the mechanisms underlying photoimmunology have been identified and it is now recognized that UVC, UVB and UVA are immunomodulatory in humans and many animal systems.
Tumor immunity, contact and delayed hypersensitivity are suppressed by UV
Kripke’s early work in tumor models showed that chronic UV exposure enhances the outgrowth of primary skin cancer (1). In addition, she discovered that acute UV irradiation permits the growth of transplanted skin tumor lines, including squamous cell carcinoma (SCC) and melanoma, which are normally immunologically rejected in immunocompetent, unirradiated mice (2,3). Kripke thus drew the first parallel that poor immunity caused by UV and skin tumor risk are related. This can be demonstrated in inbred mouse strains, as mice that are genetically susceptible to UV-induced immunosuppression are prone to develop tumors (4). Furthermore, photocarcinogenesis can be prevented when UV-induced immunosuppression is inhibited (5). Despite the historical and clinical importance of tumor models, insights into the molecular and cellular mechanisms involved have mostly been achieved through the use of nontumor systems.
Kripke et al. popularized the use of contact hypersensitivity (CHS) and delayed type hypersensitivity (DTH) models in photoimmunology as a tool for investigating further the photobiological and immunological effects of UV (6,7). These experimental systems are amendable to short term or acute UV irradiation regimes, and have provided many clues as to how UV modulates the immune system. This approach has also been extended to studies in humans.
CHS is a skin immune response to contact allergens, such as nickel. Individuals that are sensitive to contact allergens due to a previous encounter, develop a rapid immune reaction in the skin when the contact allergen is reapplied epicutaneously. The CHS response that ensues is characterized by erythema, swelling and an infiltration of immune cells (8). DTH responses have a similar clinical presentation compared to CHS; however, the immune response here is directed against antigens that have been intradermally or subcutaneously injected into skin. The Mantoux test for BCG immunization is a classic example of a DTH reaction. Both CHS and DTH are driven by cell-mediated immunity; however, it is thought that CD8+ T cells primarily mediate CHS whilst DTH are modulated by CD4+ T cells (8). Inhibition of CHS or DTH by UV is typically demonstrated by a decrease in tissue swelling or erythema as a result of a general reduction in the level of the immune response following UV exposure and antigen challenge. This suppression can be a “local” effect, i.e. when the sites of UV irradiation and antigen application are the same, or via a “systemic” effect, i.e. when the sites of UV irradiation and antigen application are distal to one another.
UVB suppresses immunity in humans and mice
UVB (290–320 nm) accounts for about 5% of terrestrial UV radiation, with UVA (320–400 nm) making up the remainder. The exact percentage of UVB is dependent on time of day, season, latitude and atmospheric conditions so that the terrestrial UV spectrum is within a variable range, rather than being absolute. However, the amount of UVB is always considerably lower than UVA.
An action spectrum for systemic suppression of CHS in mice found a broad peak of maximum immunosuppression at 260–270 nm, wavelengths lower than found in sunlight, with a shoulder at 280–290 nm, just below the UVB band and then a steady decline to 3% of the maximum at 320 nm. Hence, low wavelength UVB was found to be immunosuppressive, but the peak lies within the UVC band (9). This study did not extend beyond 320 nm and, therefore, did not investigate UVA. However, this was the first study to suggest that urocanic acid (UCA) is a chromophore for immunosuppression, based on the similarity of its absorption curve with the action spectrum for systemic immunosuppression. The role of UCA has been confirmed in many subsequent studies using a range of methodologies in various mouse strains (10). A similar action spectrum has been shown for local immunosuppression in mice between 260 and 300 nm. This study also did not extend into the UVA waveband (11).
UVB was first shown to be immunosuppressive in the pioneering work by Dr. Margaret Kripke in murine studies. She showed that UVB immunosuppression enables the outgrowth of skin tumors that are highly antigenic, implying that many skin tumors would be immunologically rejected if UV had not suppressed immunity (2). UVB is also immunosuppressive in man. It is necessary to have details of the spectra used in order to compare UVB doses in different experiments. This is because, as discussed above, low wavelength UVB is much more effective at causing immunosuppression than higher wavelength UVB, with 320 nm having little effect on the immune system. Hence similar doses, but with different spectra can have different biological effects on the immune system. To partially compensate for this, UVB doses are often quoted as multiples of the minimum erythemal dose (MED). However, this is not a perfect way to overcome this issue, as the wavelength dependency for erythema and immunosuppression are likely to differ.
Exposure to the suberythemal dose of 144 mJ cm−2 UVB daily for four consecutive days caused local immunosuppression (12). A single exposure to 4 MED of UVB or repetitive daily exposures to 0.75 or 2 MED for four consecutive days, all suppressed the induction of CHS in humans when the antigen was applied locally to the UV-irradiated skin. The single 4 MED, but not the lower exposures, also inhibited the systemic induction of immunity (13). UVB also suppresses the reactivation of memory immunity to nickel in humans. Daily irradiations with the suberythemal dose of 144 mJ cm−2 UVB caused increasing levels of immunosuppression with increasing daily irradiations (14). Increasing doses of single exposures to UVB also cause greater levels of immunosuppression in humans (15) with the suppression being discernible within 24 h of irradiation and lasting for at least 3 days. Therefore, UVB suppresses the induction of local and systemic immunity, as well as reactivation of memory immunity in humans over a large dose range, from 0.5 to 4 MED.
UVA suppresses immunity in humans and mice
UVA is highly immunosuppressive in humans and mice. As UVA has little effect on MED, it is not realistic to relate UVA doses to MED. However, it is sometimes useful to indicate the UVA dose in relation to the amount of solar-simulated UV (ssUV, mixture of UVB and UVA designed to mimic or simulate sunlight UV spectrum) that would cause sunburn, as the UVB portion of the ssUV would cause sunburn and this can help put the dose into biological context. Irradiation with 1900 mJ cm−2 UVA has been shown to suppress reactivation of memory immunity to nickel in humans (14). While this dose appears higher than the amount of UVB required to cause immunosuppression, it needs to be regarded within the context that there is about 20 times more UVA than UVB in sunlight. This dose is approximately the amount of UVA that would be found in 0.5 MED of sunlight. Interestingly, UVA immunosuppression increased with larger numbers of exposures from 1 to 3 on consecutive days. However, continuing irradiation for 5 days was not immunosuppressive, probably due to the development of protective mechanisms either due to the increased time of repetitive irradiations, or the higher cumulative doses. Similar observations can be seen in mice despite methodological differences. In mice, UVA has been shown to suppress the induction of local immunosuppression (16) and the reactivation of memory immunity (17). Systemic suppression of the induction of primary CHS occurred with three daily exposures to the relatively low UVA dose of 1680 mJ cm−2, while the higher dose of 3360 mJ cm−2 was not suppressive (18). This bell-shaped dose–response curve with the UVA component of about 0.5 MED of sunlight, but not higher doses being immunosuppressive, also occurs in a DTH response to ovalbumin in mice (19). Therefore, while UVB induces a linear dose response with increasing doses leading to greater levels of immunosuppression, this does not appear to be the case with UVA. These higher doses of UVA are not inert, but in the case of co-irradiation with UVB, interact with UVB to cause immunosuppression, as will be discussed later. More recently, using narrow wavebands of UVA to suppress immunity in humans, we have shown the same dose response, immunosuppression with low but not higher UVA doses (D.L. Damian, Y. Renwick, T.A. Phan and G.M. Halliday, unpublished). Therefore, in a variety of models in humans and mice, UVA has been shown to be immunosuppressive, but only over a particular dose range with higher doses of UVA not causing immunosuppression. The dose range over which UVA is immunosuppressive is within the range that can be achieved by natural sunlight exposure, and is therefore likely to be important for human responses to sunlight. It is unclear why UVA causes this bell-shaped dose–response curve. Higher doses of UVA may switch on protective mechanisms, or destroy the unknown UVA chromophore.
UVA has also been shown to suppress recall immunity to the multitest kit Merieux, which is a DTH test to seven antigens to which humans are commonly exposed. This test requires exposure to a large area of human skin and tends to require larger UV doses than other procedures for monitoring photoimmunosuppression in humans. This may be related to the large area of skin irradiation causing systemic immunosuppression (20). UVA has been shown to suppress this response in humans, both locally and systemically (21).
Additional evidence that UVA is immunosuppressive comes from studies using sunscreens. Several studies into different models of UV-induced immunosuppression in mice have shown that sunscreens need to provide protection from both UVB and UVA in order to inhibit ssUV-induced immunosuppression (17,22). Similarly in humans, a number of studies have shown that good protection of the immune system can only be provided by broad-spectrum sunscreens that absorb or reflect both UVB and UVA (20,23–28). By comparing a large number of commercial sunscreens for protection from ssUV-induced immunosuppression, it was shown that this significantly correlated with the UVA protective capacity of the sunscreens (29). Considering all data, it is clear that UVA causes immunosuppression in humans; however, the wavelength dependency and chromophore which absorbs UVA remains to be determined in order to understand this response.
Immunomodulation due to interactive effects between UVB and UVA
UV containing a mixture of about 90% UVA and 10% UVB is immunosuppressive in man (30). In this study, Hersey and colleagues irradiated human volunteers in a commercial solarium for 12 × 0.5 h exposures on consecutive days, which reduced the induction of CHS to a hapten. Doses of ssUV as low as a single exposure to 800 mJ cm−2 suppressed the induction of immunity at the irradiated site in humans (31). A progressively increasing ssUV dose regime of 10 daily exposures to an average of 1.45 MED per day has also been shown to suppress recall DTH to Merieux (21). This clearly shows that UV, which mimics the UV portion of the solar spectrum is immunosuppressive in humans.
In a comparison of ssUV with UVB and UVA, it was noted that ssUV-induced immunosuppression was greater than could be accounted for by the UVB component of ssUV, although different groups of volunteers were irradiated with single doses of ssUV and UVB (14). This suggested that the level of sunlight-induced immunosuppression is greater than what could have been caused by the UVB waveband. In subsequent more comprehensive studies, it has been clearly shown that interactions between UVB and UVA make ssUV more immunosuppressive than either of the component wavebands alone in humans (15). While both UVB and UVA could suppress memory immunity to nickel in humans, dose responses showed that ssUV induced a greater degree of immunosuppression than the additive effects of these wavebands. Sunlight was even suppressive at doses below the immunosuppressive threshold of the UVB or UVA wavebands. We have recently made similar observations for suppression of DTH in humans (D.L. Damian, C. Patterson and G.M. Halliday, unpublished). Interestingly, this cooperation between the mechanisms activated by UVB and UVA required 72 h to augment immunosuppression, while UVB alone was suppressive within 24 h and UVA within 48 h. There appear to be interactive effects between the molecular changes caused by UVB and UVA, which require some time to result in suppressed reactions to antigens in the skin.
Molecular cross-talk between UVB and UVA signaling has been shown to cause a level of response larger than the additive effects of UVA and UVB (32). In cultured human keratinocytes, the signal transduction molecule ERK1/2 was only transiently activated to a small extent by 30 000 mJ cm−2 UVA while 10 mJ cm−2 UVB caused a more pronounced activation. However, p38MAPK and JNK1/2 were only slightly activated by either waveband. In contrast, irradiation with both UVA and UVB did not activate ERK1/2 (it was activated by UVB alone) but enhanced activation of p38MAPK and JNK1/2 (33). Therefore, the signaling cascade activated by a mixture of UVB and UVA is different to the additive effects of these wavebands indicating an interactive effect, so that the response to this combination differs to either waveband alone. In these experiments, the UVA:UVB ratio was considerably higher than that found in sunlight, with a lower UVB and higher UVA dose than shown by Poon et al. (15) to interact making sunlight more immunosuppressive than the additive effects of each waveband at low sunlight doses. However, these results show cross-talk at the molecular level in humans between UVA and UVB responses, supporting studies that interactions between these wavebands could have a different immunomodulatory effect to either waveband alone. In a further study in which humans were irradiated with ssUV for four consecutive days with and without additional UVA (17 800 mJ cm−2), interactive effects were observed (34). These doses of UVA were too high to be immunosuppressive in their own right (see above section, UVA has a bell-shaped dose–response curve for immunosuppression), but greatly augmented ssUV-induced immunosuppression. This provides further evidence for molecular cross-talk between these wavebands leading to a different immunological outcome from the additive effects of the two wavebands. It also shows that this interactive effect of UVA occurs not only with single doses of UV, but also with multiple exposures. UVA augmentation of UVB-induced immunosuppression can occur not only with doses of UVA that are too low to be suppressive, but also with doses of UVA that are too high to be autonomously immunosuppressive.
Interactive effects between UVA and UVB have also been described to cause the opposite immunological outcome, immunoprotection. This was first shown in mice by Reeve et al. (35). Immunosuppression induced by 700 mJ cm−2 UVB was found to be prevented by 38 700 mJ cm−2 UVA, which was not immunosuppressive in its own right. This high dose of UVA is equivalent to the amount found in 10 MED of sunlight in humans and is consistent with studies discussed above that these high doses of UVA are not independently immunosuppressive. Subsequent studies in mice showed that while lower doses of UVA are immunosuppressive, higher doses protect from UVB immunosuppression (18). The study described above by Kuchel et al., however, did not find UVA doses within this range to protect from UVB-induced suppression of reactivation of memory immunity to nickel in humans. A number of studies using a variety of different conditions in humans have found linear dose responses for ssUV-induced immunosuppression (24,27–29,31) showing that higher doses of UVA do not prevent UVB induced-immunosuppression in humans over the dose ranges tested. However, the doses of ssUV were limited by UVB-induced sunburn. An exceedingly high dose of 251 600 mJ cm−2 UVA (wavelengths above 340 nm) has been found to protect humans from 225 mJ cm−2 UVB-induced induction of immunosuppression to diphenylcyclopropenone (36). The UVA dose in this study was 14-fold higher than that found by Kuchel et al. to augment ssUV-induced immunosuppression in humans, and had a UVA:UVB ratio of 1:100, which is about 55 times greater than is found in sunlight.
In summary, there are four different waveband patterns of sunlight-induced immunomodulation. Both UVB and UVA are autonomously immunosuppressive over a dose range to which humans can be exposed by natural sunlight. There are also interactions between UVA and UVB, probably resulting from cross-talk between the molecular events triggered by these wavebands. These interactions can either enhance immunosuppression at doses to which humans can be exposed by natural sunlight, or UVA can inhibit UVB immunosuppression at doses far higher than humans can receive from sunlight. Interestingly, the immunoprotective interactive effect of UVA occurs at much lower doses in mice than in humans. These enhancing and inhibiting interactions probably occur via different pathways based on studies in genetically different mouse strains. Like humans, C57BL/6 mice can respond to UVA by displaying both immunosuppression and immunoprotection, while BALB/c mice do not respond to either of these UVA responses (18), but UVA protects from UVB without causing immunosuppression in SKH:HR-1 mice (37).
While both UVB and UVA, in addition to interactive effects, suppress CHS and DTH, the wavebands that suppress tumor immunity have received much less research attention. Direct investigations into UV-induced suppression of tumor immunity have been limited to animal models for ethical reasons.
Direct evidence that immunosuppression enhances growth of skin cancer in humans comes from studies in patients with HIV infection, a long history of phototherapy and pharmacologically induced immunosuppression to enhance survival of organ transplants. Patients immunosuppressed by infection with HIV can develop rapidly growing cutaneous SCC at a young age, with a high risk of local recurrence and metastasis (38). Psoralens combined with UVA (PUVA) phototherapy is both mutagenic and immunosuppressive. Recent work has demonstrated that activation of the platelet-activating factor (PAF) pathway is necessary for PUVA-induced immune suppression (39). Long-term high-dose exposure to PUVA is consistently observed to significantly increase the risk of SCC (40). Skin cancers occur more frequently in organ transplant recipients than in the general population. They often develop multiple skin cancers with a higher rate of local recurrence and a greater propensity to invade locally and metastasize. Restoring immunosurveillance, by reducing or stopping immunosuppressive therapy, can be useful for treating patients with highly aggressive skin cancer (41,42). Patients with a positive history of skin cancer are more sensitive than subjects without a history of skin cancer to UVB-induced suppression of CHS responses (12). Humans were irradiated with 144 mJ cm−2 UVB per day for four consecutive days, a dose that suppressed immunity in 40% of controls but 92% of patients with a history of skin cancer. Thus, the immune response inhibits development of many potential skin cancers in humans and UV is particularly suppressive in patients with a history of skin cancer, suggesting that UV suppression of tumor immunity is likely to be clinically important in humans.
A large number of experiments show that UV-induced immunosuppression enables the outgrowth of skin tumors in mice, and many of these experiments have come from Kripke’s research group. UVB-induced immunosuppression enables the outgrowth of transplanted epithelial skin cancers and melanomas in mice (3,43,44). Specific T cells activated in UVB-irradiated mice by antigen exposure can transfer suppression to normal recipients, inhibiting tumor immunity and, therefore, enabling UV-induced skin tumors to grow (45).
However, there are few studies into the wavelength dependency of UV suppression of tumor immunity. In one study, exposure of mice to five daily doses of 1750 mJ cm−2 ssUV per day prior to injection of a tumor cell line enhanced skin tumor growth (46). Therefore, the UV wavelengths found within sunlight, not just UVB, are able to suppress immune-mediated rejection of skin tumors. Whether UVA contributes to this, or interacts with UVB to enhance suppression of antitumor immunity has not been investigated.
Multiple pathways are involved in UV-induced immunosuppression
Many rigorous and varied efforts have been made to fully elucidate the mechanism of UV-induced immune suppression in view of the fact that, if it can be counteracted or reversed, skin tumor development may be delayed or halted. Although the exact molecular and cellular factors involved are only just coming into light, it is likely that UV initiates numerous pathways that all play a function in inhibiting immune responses. Molecular and biochemical changes within the skin signal the first effects of UV, which eventually leads to inhibitory effects on the cellular lymphoid system. These linked pathways will be discussed within the context of UVB and UVA wavebands.
Cellular events involved in UVB-induced immunosuppression
UVB chromophores in skin. As UVB penetrates into the epidermis, the energy from the photons is absorbed by chromophores in the skin. Studies by Kripke were the first to suggest that DNA may be the chromophore that initiates UVB-induced immunosuppression (47). UVB damages DNA by generating cyclobutane pyrimidine dimers (CPD). In marsupial Monodelphus domestica mammals, this DNA damage is repaired by a photolyase that is photoreactivated by visible light. Visible light reversed UVB-induced DNA damage and immune suppression. Application of liposomes containing the DNA repair enzyme T4 endonuclease V (T4N5), which initiates removal of CPDs from photodamaged DNA, can protect against local and systemic UVB-induced immune suppression of CHS and DTH in mice (48). Photolyase treatment and T4N5 liposomes are also effective in humans by inhibiting UVB and ssUV-induced immunosuppression (49,50).
Tryptophan is a second chromophore that is present in the cytoplasm and is independent of UV effects within the nucleus. Recently, it has been shown that photoproducts of tryptophan bind to the cytosolic arylhydrocarbon receptor (AhR). Activation of AhR leads to two secondary signals being delivered to the nucleus and to the cell membrane (51). The consequences of this signaling pathway are likely to be related to the activation of the epidermal growth factor receptor and to the MAPK pathway that is known to be important for skin carcinogenesis and UVB-induced inflammation (52,53).
De Fabo and Noonan contributed another unique photoreceptor, trans-UCA, when they discovered that prior to UVB irradiation, tape stripping mouse ears to remove the stratum corneum of the epidermis prevented suppressed CHS responses (9). Trans-UCA is a natural component of the stratum corneum at high concentrations in mice and humans. UV induces a photoisomeric transition of trans-UCA to cis-UCA that can be detected in elevated proportions in the skin and urine of humans irradiated with UVB (54). Blocking cis-UCA with a specific monoclonal antibody can reverse DTH suppression by UVB (55) and treatment with cis-UCA in the absence of UVB can mimic the immunosuppressive effects of UVB on DTH in a dose-dependent manner (56). Moreover, treatment of mice with anti-cis-UCA antibodies during chronic UVB irradiation reduced tumor incidence compared to control mice given an irrelevant antibody. Thus, cis-UCA may be involved in UV skin carcinogenesis (57). Interestingly, cis-UCA has been shown to bind to the serotonin receptor, 5-HT2A, which is found on dendritic cells, T cells, B cells and peripheral nerves. Walterscheid et al. present a convincing case arguing that cis-UCA may be modulating the immune system during UVB-induced immunosuppression by binding to 5-HT2A on these cell types (58).
UVB-induced mediators. UVB irradiation triggers the production of various immunomodulatory mediators within the skin, which contribute toward UVB-induced immunosuppression. Examples of these molecules include the enzyme, cyclooxygenase-2 (COX-2), which arises from activation of the MAPK pathway by UVB (59). Skin carcinogenesis is associated with UVB-induced inflammation, and enhanced levels of COX-2 are a common feature of nonmelanoma skin neoplasms (60). In mice, inhibition of COX-2 can reduce the incidence of tumor formation (61), as well as UVB-induced systemic immunosuppression (62). COX-2 is the rate-limiting enzyme that catalyzes prostaglandin production from arachidonic acid. Topical application of prostaglandin E2 (PGE2) onto unirradiated mouse skin can inhibit CHS responses similar to UVB, suggesting that PGE2 is involved in UVB-induced immune suppression (63). UVB-irradiated keratinocytes are the main producers of PGE2 and they can be stimulated to secrete this molecule by inflammatory phospholipids like PAF. In providing one of the primary signals in the UVB response, PAF is critical to immune suppression and this can be seen in investigations wherein delivery of PAF receptor antagonists can prevent suppression of DTH in UVB-irradiated mice (64). PGE2 promotes immune suppression by stimulating a cytokine cascade that leads to the circulation of antiinflammatory IL-4 and IL-10 (65).
As a byproduct of degranulated mast cells, histamine can also stimulate PGE2 production in keratinocytes and macrophages. Histamine is an important mediator of UVB-induced immune suppression, but to date the mechanism for this, besides its role in enhancing PGE2 production, is unclear. Mice deficient in mast cells are resistant to immunosuppression by UVB (66) and increased mast cell density has been proposed to be a predisposing factor for basal cell carcinoma development in humans (67). Binding of cis-UCA to peripheral nerves may stimulate the release of neuropeptides known to cause mast cell degranulation (58,68), for example calcitonin gene-related peptide (CGRP) (69). As further evidence to this, mice depleted of neuropeptides by capsaicin cannot be immune suppressed by cis-UCA (68), and topical treatment on human skin reduces UVB suppression of DTH (70). Chronic UVB irradiation causes increased production of CGRP in skin and antagonists for the CGRP receptor can reverse UV-induced immunosuppression in mouse CHS models (71). As inhibition of neuropeptide action can decrease mast cell degranulation, the primary mode to explain the immunosuppressive activities of neuropeptides has been correlated with its modulatory effects on mast cells to release tumor necrosis factor (TNF) (72).
UVB causes aberrant antigen presenting cell (APC) behavior and migration Since the study by Yoshikawa and Streilein that first described the similarities between the effects of TNF administration and UVB with regard to CHS inhibition (73), TNF has traditionally been considered to be an important cytokine involved in the mechanism of UVB-induced immunosuppression. Injured keratinocytes, degranulated mast cells and activated macrophages all produce TNF in response to UVB. UVB, through the action of TNF, induces Langerhans cell (LC) activation and migration out of the skin and into draining lymph nodes (74,75). Loss of LC in this manner in UVB-irradiated skin limits the capacity for antigen processing and presentation. Local UVB irradiation perturbs APC by reducing their expression of co-stimulatory molecules and profile of cytokine secretion necessary to stimulate the generation of antigen-specific Th1 cells (76,77). As CHS, DTH and tumor immunity are primarily mediated by Th1-type cells, a failure to adequately stimulate the respective effector cells would contribute toward UVB-induced immune suppression. Similarly, disrupted APC–T cell interactions are involved in systemic UVB-induced immune suppression (78,79). UVB activates B cells in draining lymph nodes that suppress dendritic cell antigen presentation to T cells (80).
DNA damage in the form of CPDs sustained by LC, which migrate from the skin during local UVB exposure, can be repaired by photolyase treatment. This restores normal APC function (81). Remarkably, the cytokine, IL-12, which is typically associated with driving Th1-type immune responses, can promote repair of keratinocyte DNA via a nucleotide-excision repair mechanism (82). Injection of IL-12 before UVB irradiation reduced the number of CPD+ LC accumulating in skin draining lymph nodes, suggesting that IL-12 prevents UVB-induced DNA damage in these cells. In this way, mice administered IL-12 exhibit CHS responses similar to unirradiated mice (83), whilst mice deficient in IL-12 show a higher prevalence to develop malignant skin tumors than wildtype mice (84).
UVB generates suppressor cells. Protection from UVB-induced DNA damage by IL-12 also prevented UVB-induced tolerance, which is a hallmark feature of UVB-induced immunosuppression (83). Exposure to UVB not only inhibits the primary immune response, but several weeks following UVB exposure, prolonged nonresponsiveness or tolerance can occur when antigen is rechallenged at an unirradiated site (85). Kripke’s original work eluded to a master suppressor cell type orchestrating antigen-specific UVB-induced immune suppression and skin tumor development (1,86,87). Several candidate UVB-induced suppressor cells have since been proposed including suppressor natural killer (NKT) (88), B (80) and CD4+ T cells (89). The existence of these cells is frequently demonstrated by adoptive transfer studies, wherein transfer of lymphoid cells from UVB-irradiated mice can induce antigen-specific nonresponsiveness in unirradiated mice when challenged with the same antigen. Suppressor cells are generated within the first few days following UVB (90) and as originally hypothesized by Kripke’s research group (7), they arise either as a direct result of tolerogenic or impaired APC (91), or as a consequence of the cytokine environment within lymphoid organs draining UVB-irradiated skin that promotes the activation of suppressor cells (80,92). Due to their limited capacity to migrate (89), suppressor cells localized within lymphoid organs impart suppression through the production of immunosuppressive cytokines, namely IL-4 and IL-10, that are potent inhibitors of cell-mediated immune responses (80,88,93,94). In effect, tumor immunity is hindered by the presence of UVB-induced suppressor cells, as shown by the example that lethally X-ray-irradiated mice are unable to reject regressor tumors if they are reconstituted with splenic NKT cells derived from UVB-irradiated mice (88).
UVB inhibits generation of effector and memory T cells. Immune responses to contact allergens or tumor antigens are predominantly mediated by antigen-specific T cells that are initially activated in lymphoid organs. Inhibited CHS responses denoted in UVB-irradiated mice are partially attributed to a reduced migration of T cells into unirradiated challenged sites. This observation is paralleled by a decreased generation of effector T cells within lymph nodes draining the site of antigen sensitization and UVB irradiation (S. Rana, L.M. MacDonald, S.N. Byrne and G.M. Halliday, unpublished). Factors such as poor APC activity in the presence of UVB-induced suppressor cells may impede the normal activation of T cells against their cognate antigen, or this UVB-induced reduction in the number of effector T cells could occur by a different mechanism. Inhibitory signaling through CTLA-4, a molecule expressed on activated T cells and suppressor cells, could also be limiting T cell reactivity in lymph nodes. In UVB-irradiated mice, blockade of CTLA-4 signaling via anti-CTLA-4 antibodies promoted the rejection of tumors that are usually progressive in these mice. In addition, it was found that in vitro anti-CTLA-4 antibodies blocked the suppressive activity of CD4+ CD25+ regulatory T cells on T cell proliferation. Thus, anti-CTLA-4 antibody treatment may enhance antitumor immunity in UVB-irradiated mice by inhibiting the activity of UVB-induced suppressor T cells on effector T cells (95). Interference by UVB during primary T cell responses has a corresponding effect on memory T cell development. Mice that are exposed to UVB prior to the application of antigen, fail to develop peripheral memory T cells unlike unirradiated mice (S. Rana, L.M. MacDonald, S.N. Byrne and G.M. Halliday, unpublished).
Thus, UVB ultimately suppresses the immune system by inducing the production of immunosuppressive mediators, by damaging and triggering the premature migration of APC required to stimulate antigen-specific immune responses, by inducing the generation of suppressor cells and by inhibiting the activation of effector and memory T cells. These events are likely to be linked. As a result, the immune cells important in cell-mediated immune responses typical of CHS, DTH and tumor immunity are not able to respond appropriately and are prevented from providing long lasting immunity.
Cellular events involved in UVA-induced immunosuppression
In contrasting the effects of UVB and UVA on cellular events leading to immunosuppression, it appears likely that differences should emerge; however, UVA has received little research attention. Some of the mechanisms implicated in UVA-induced immunosuppression, such as increased COX-2 activity (96), are common to those already described for UVB. UVA has an important role in skin cancer development as evidenced by the finding that the basal layer of human skin tumors contains more UVA fingerprint mutations than UVB fingerprint mutations (97). The mechanisms related to UVA-induced immunosuppression will be discussed.
UVA-induced mediators. Reactive oxygen species (ROS) are generated from chromophores, which primarily absorb wavelengths in the UVA spectrum (98). Although the identity of the UVA chromophore in skin is unknown, molecules like porphyrins have been proposed (99). Production of ROS alters the redox equilibrium, which is usually held in balance by antioxidant enzymes and free radical scavengers such as superoxide dismutase, catalase, glutathione peroxidase, vitamin C and vitamin E. UV, however, actively depletes antioxidants, such that ROS activity continues unchecked (100). ROS like superoxide anion, singlet oxygen and hydrogen peroxide target proteins, lipids and DNA. Loss of sulphydryl groups in proteins can inactivate enzymes and increase proteolysis, whilst lipid peroxidation can cause permanent cell membrane damage. Oxidative-induced damage of DNA, in particular to the base guanidine, causes strand breaks and DNA-protein cross-links. The dangers associated with UVA and DNA damage are underlined in a study by O’Donovan et al. that examined the interactions of UVA with an immunosuppressive drug, azathioprine, commonly administered to transplant patients. Azathioprine acts as a UVA chromophore within cells, and in vitro treatment of cells with azathioprine and UVA stimulated the increased production of ROS. Based on these findings, O’Donovan et al. speculated whether UVA from normal sunlight and azathioprine treatment is linked to the high incidence rate of SCC exhibited by transplant recipients (101). There is much interest today to protect the skin from photocarcinogenesis by treatment with antioxidants like carotenoids; however, the successes of these numerous studies in mice are not always translatable to humans (102).
UVA also upregulates the activity of nitric oxide synthase, an enzyme that catalyzes nitric oxide (NO) production. Interaction of NO with ROS leads to the increased generation of reactive nitrogen species (RNS) like peroxynitrite. These have damaging properties on cellular structures similar to ROS. Treatment of nickel-allergic individuals with the NO synthase inhibitor, l-NMMA, reverses the immunosuppressive effects of ssUV on a nickel CHS response, thereby supporting the notion that NO mediates immune suppression of UVA (103). UVA immunosuppression is likely to involve different chromophores to UVB and is likely to be mediated via ROS, but this requires additional studies.
UVA causes aberrant APC behavior and migration. Similar to UVB, UVA irradiation causes LC depletion from skin (16), which appears to be triggered by ROS and RNS. Inhibition of ROS and RNS by l-NMMA (103,104) or topical α-tocopherol (vitamin E) prevented UVA-induced suppression of CHS and LC loss (105). Moreover, purified LC or epidermal cells exposed to UVA in vitro demonstrated poor APC activity in a mixed epidermal cell-lymphocyte reaction, which was partially recovered when cells were incubated with vitamin E (106) or glutathione (107). Thus, immunosuppression caused by UVA is closely associated with ROS and RNS-mediated damage to LC.
UVA inhibits T cell activation and generates suppressor cells. Photooxidative damage incurred by migrating LC could interfere with their ability to initiate immune reactions in draining lymph nodes in vivo. Iwai et al. observed that UVA irradiation reduced lymph node cell proliferation, which was reversed when glutathione was applied onto the skin during irradiation (107). In addition, we have found that UVA irradiation impaired the development of peripheral memory T cells in skin (S. Rana, L.M. MacDonald and G.M. Halliday, unpublished). Therefore, it is possible that UVA-damaged LC unable to adequately stimulate T cells would cause the reduced activation and proliferation of effector T cells, and the consequent inhibition of memory T cells.
Like UVB, UVA can also induce the generation of suppressor cells that can inhibit T cell immune responses in lymphoid organs. Chronic exposure to UVA for 4 weeks prior to antigen application reduced local primary and secondary CHS (16), whilst UVA irradiation beginning after immunization but lasting for 9 days postimmunization also reduced DTH in mice (108). In these studies, tolerance was transferred into naïve mice via the adoptive transfer of splenic cells derived from UVA-irradiated donor mice. This indicates that chronic or prolonged UVA irradiation stimulates the generation of suppressor cells. The induction of UVA-induced suppressor cells, however, appears to be time or dose dependent. Acute UVA irradiation for 3 consecutive days and examination 3 days later did not cause an activation of suppressor B cells in draining lymph nodes, unlike UVB (109). Similarly, Nghiem et al. also reported that significant UVA-induced reduced DTH was only achieved when mice were exposed to UVA for at least 7 days postimmunization (108). Short-term irradiation with UVA also failed to induce tolerance in mice suggesting that suppressor cells were not generated to maintain nonresponsiveness to antigens (19). Clearly, further studies are required to elucidate the apparent dose dependency of UVA-induced suppressor cells in order to understand the involvement of these cells during UVA-induced immunosuppression. The mechanisms behind the formation of these cells are also likely to be different from those that stimulate the generation of UVB-induced suppressor cells.
Margaret Kripke initiated an intellectually challenging and medically important area of research. We cannot help wonder whether she realized how complex a discipline she had set a large number of scientists to pursue. Understanding the wavelength and cellular events involved in different immune responses to UV alone is a large task, which is not completely understood.
UVB and UVA suppress local, systemic, primary and memory immunity in humans and animal models. They suppress CHS to epicutaneously applied antigen, and DTH to intradermal antigens. To put UV dose and wavelength responses into perspective, the CHS data in humans summarized in this review have been converted to time of sunlight exposure. Natural sunlight in early spring in Sydney, Australia, at midday (0.64 mW cm−2, composed of 0.04 mW cm−2 UVB and 0.60 mW cm−2 UVA) is used for this purpose, and an MED is taken as 2800 mJ cm−2 as we find this to be an average MED in Caucasian Australians. This relationship of immunosuppression with time of sunlight exposure (Fig. 1), while based on current data described in this review, is likely to be dependent on a number of variables such as type of immune response, skin type and undefined genetic factors. Many of these are inadequately understood and, therefore, this is a guide only to attempt to put these issues into perspective.
UVB exposures from about 0.5 to 4 MED, or 40 min to 5 h midday spring sunlight cause a linear dose-related immunosuppression. The effect of higher doses is unknown and cannot be determined due to the limitation of inducing unacceptably high levels of sunburn in volunteers. UVA exposures of about 0.5 to less than 4 times the amount found in 1 MED of sunlight (40 min to 5 h) are suppressive in humans while higher doses cease to be independently immunosuppressive. Higher doses of UVA do however contribute to sunlight-induced immunosuppression by interacting with UVB. Doses of ssUV from about 0.3 MED (22 min) are suppressive up to at least 17 800 mJ cm−2 UVA because of interactions between UVB and UVA. Thus, UVA in midday spring sunlight exposures of about 20 min up to 8 h interact with UVB to enhance immunosuppression. Increased exposure of 251 600 mJ cm−2 UVA, requiring 110 h exposure to midday spring sunlight, however, protects from UVB immunosuppression. Thus, UVB and UVA within the dose range to which humans could be exposed for either normal daily, work or recreational activities all suppress immunity. However, it is the interactions between UVB and UVA, which are most effectively suppressive.
However, there are a considerable number of gaps in our knowledge. The wavelength dependency for UVA immunosuppression is unknown, and this will be linked to the identification of the chromophore for UVA immunosuppression. There are likely to be substantial differences between wavelength and dose-related effects on different types of immune suppression. There is little information available on the differences in mechanisms responsible for UVB- and UVA-induced immunosuppression, which may also be related to differences in the chromophores involved and to the signaling pathways that are triggered by UVB and UVA upon photon absorbance. For example, the UVB response has nucleic, cytoplasmic and extracellular signaling components due to the presence of UVB chromophores within these sites in the skin. Signaling from these known chromophores all contribute toward UVB-induced immunosuppression. In comparison, the immunosuppressive effects of UVA are likely to be limited to cytoplasmic chromophores as the generation of ROS and RNS by skin cells plays a significant role in UVA-induced immunosuppression. Despite the apparent differences in the early UVB and UVA response, it is interesting to note that some of the processes leading from these signaling pathways appear to converge. UVB and UVA have similar downstream effects on the cellular immune system, such as its modulation of LC, induction of suppressor cells and inhibition of effector and memory T cells. However, the exact mechanisms initiated by UVB, and in particular UVA, that bring about these changes need further definition. Moreover, UVB and UVA interact to make sunlight more potent than either waveband alone, and it is possible that synergistic interactions taking place between the signaling pathways would drive the enhanced immunosuppressive effects shown by ssUV compared to UVB or UVA. In addition, sunlight-induced immunosuppression may have other unknown players that may be independent of those already acting in UVB- or UVA-induced immunosuppression. Therefore, the mechanism of this UVB and UVA interaction needs to be determined if we are to fully appreciate the effects of solar radiation on the immune system.
Most of our current research has utilized CHS or DTH models in mice or humans; however, tumor immunity has received much less attention. Tumor immunity must be more complex as the tumor antigens change during tumor development and emerge on a small population of cells that increase in number, until at some stage there would be a sufficiently large quantity of tumor antigen for the immune system to take notice. This is more complex than experimental models of exposure to antigen during CHS or DTH where at a defined time point there is contact with immunogenic quantities of antigen. Whether the same wavelength dependencies occur for tumor immunity as have been described for CHS and DTH is unknown. It is not known whether UVA contributes to sunlight suppression of tumor immunity as this has not been investigated. Small variations in the magnitude of the immune reaction may have large consequences on immune rejection of skin tumors; however, it is not known how much sunlight is required to be clinically damaging to antitumor immunity in humans. Therefore, despite there being a large amount known about UV immunosuppression, large and clinically important questions remain to be resolved.
Sunlight has some health benefits while at the same time causing damage. UVB, but not UVA, induces production of vitamin D (110), which is photoprotective, reducing sunlight-induced damage to DNA (111). This steroid hormone also has other health benefits including muscle and bone function and there is some recent evidence that it may be beneficial for some types of cancers including prostate, colorectal, breast cancer and, paradoxically, melanoma (112). This beneficial effect of UVB is due to low levels of exposure as vitamin D production saturates with increasing levels of exposure, and higher levels cause photodegradation. Therefore, further levels of exposure do not lead to higher levels of vitamin D. It is difficult to estimate a UVB dose that produces maximum vitamin D production in human skin; however, about 0.33 of an MED of sunlight exposure has been recommended for adequate vitamin D production (113). Thus, using the example of midday spring sunlight described above, 24 min of exposure would produce this amount of vitamin D (Fig. 2). At higher doses of about 1 MED (1.2 h exposure to spring sunlight), UVB initiates apoptosis in cells, resulting in sunburn cell formation and removal of heavily damaged cells. This is another photoprotective mechanism, which reduces the level of cells in the skin with severe genetic damage. UVA contributes little to this effect. In contrast to this, both UVB and UVA cause genetic damage. Low doses of both UVB and UVA are able to cause immunosuppression, and these wavebands interact to make sunlight-induced immunosuppression greater than that achieved by both wavebands alone.
While UVB has some beneficial effects (vitamin D and apoptosis), UVA does not have any known health effects. All of this creates a difficult issue for advice to the public with regard to the benefits versus the dangers of sun exposure and also how to protect from sun exposure. A small amount of UVB is likely to be beneficial; however, even these low UVB doses cause damage, and the point at which the dangers outweigh the benefits is an unknown and hotly debated issue, but it will be at a low level of exposure. Contrary to this, UVA causes damage without any health benefits even at low levels of exposure. Hence, we need to be devising ways of providing complete protection from UVA while enabling vitamin D production and not inhibiting UVB-induced apoptosis. How this could be achieved is unknown.