Explaining a possible protective role of polymorphous light eruption against skin cancer

UV radiation in sunlight, especially short wavelength (<315 nm) UV-B radiation, is absorbed by a great variety of organic molecules, such as lipids, proteins and DNA. Upon absorption these molecules can be altered and damaged, or can initiate further photochemical reactions with other molecules. Our skin is regularly subjected to these alterations and is necessarily well adapted. With the marked exception of the formation of vitamin D3, most of the UV-induced alterations are detrimental and need to be counteracted or eliminated. Damage that acutely hampers cell function, such as damage in active genes, needs to be repaired with high priority. Minor damage can be tolerated but can become critical upon accumulation. To keep the exposed epidermis vital and functional it is continually renewed every 3–4 weeks.

Unfortunately, the small quantity of DNA is a disproportionally strong UV-B absorber in the cell, but the exposed epidermal cells are equipped with impressively efficient repair systems, among which the high-fidelity cut-and-paste replacement of stretches of damaged DNA by nucleotide excision repair (NER) stands out as the most important. A defect in NER, as in xeroderma pigmentosum (XP), results in an exquisite UV sensitivity of the skin and a dramatically increased risk of skin cancer. UV radiation induces a specific type of DNA damage (cyclobutane pyrimidine dimers and 6–4 photoproducts between neighboring pyrimidines in the same DNA strand), which cause specific gene mutations (cytosine to thymine) at dipyrimidine sites. Such mutations predominate in the p53 tumor suppressor genes of skin carcinomas from XP patients, but also from NER-proficient patients. Hence, there would appear to be a direct causal link from UV-induced DNA damage to skin carcinomas [1].

However, classic mouse experiments by Margaret Kripke demonstrated that UV-induced skin tumors are highly immunogenic: tumor implants in a syngeneic host are rejected unless the host is exposed to subcarcinogenic but sufficient doses of UV-B radiation [2]. UV irradiation suppressed tumor immunity and induced a UV tumor-specific tolerance transferable with ‘T suppressor cells’, now dubbed ‘regulatory T cells’ or ‘Tregs’ [3]. Likewise, UV irradiation was found to suppress sensitization of delayed-type and contact hypersensitivity to a skin allergen, and to induce Treg-mediated tolerance toward the allergen. Remarkably, XP mice are also exquisitely sensitive to the UV-induced immunosuppression, indicating that DNA is a UV target for this suppression, and that the increased sensitivity to immunosuppression contributes to the enhanced UV carcinogenesis in XP mice and patients [4]. However, next to DNA, UV also targets other cell components that may cause immunosuppression: urocanic acid (isomerisation from trans to cis) and cell membranes (generating reactive oxygen species), but also delayed the synthesis of vitamin D3 and its hormonal metabolite 1,25-dihydroxyvitamin D3 [5].

In 1990, the group of the late Wayne Streilein reported that a low dose of UV radiation also suppressed sensitization to contact allergens in human volunteers, and that among people who had skin carcinomas removed, a higher percentage (92 vs 40%) were susceptible to UV-induced immunosuppression, of whom 45% acquired an allergen-specific tolerance, whereas none of the controls developed a tolerance [6]. This indicated that UV-induced immunosuppression and tolerance are risk factors for skin carcinomas in humans. UV-induced immunosuppression would appear to be an adverse response, more so as it can potentially aggravate certain infections [7]. However, later studies showed that almost all volunteers displayed the UV-induced immunosuppression with sufficient UV exposure matched to their sunburn sensitivity [8,9]. This finding strongly indicates that UV-induced immunosuppression is a generally sound physiological response, an unanticipated adaptation of the skin to UV exposures. It has to be noted further that this suppression of acquired cellular immunity by UV exposure is accompanied by a boost of innate immunity as the rapidly acting first line of defense against pathogens [10], that is, UV exposure causes an apparent shift from antigen-specific acquired cellular immunity to nonspecific innate immunity. An inadequate suppression of cellular immunity might allow some illicit pathological immune reaction to develop against the UV-exposed skin, for example, against UV-altered molecules forming ‘neoantigens’. This potential reaction to neoantigens fits a hypothesis regarding the pathogenesis of the idiopathic photodermatosis polymorphic light eruption (PLE). The UV-induced immunosuppression would, therefore, prevent us from developing a ‘sun allergy’. Thus, the skin appears to perform a fine balancing act in mounting a cellular immune reaction against infections and tumors when needed, but also in downregulating such immune reactions to prevent repetitive allergic skin reactions upon exposure to the sun. This balance seems to waver often, as demonstrated in a multicentered collaborative pan-European study [11], in which it was found that approximately one in five people with indoor professions had experienced PLE-like skin reactions at some point, but only a small minority developed symptoms serious enough to visit a doctor.

Studying the skin reactions of PLE patients to acute UV over-exposures (six-times the threshold dose for a minimal erythema) using immunohistochemistry showed diminished migratory responses of Langerhans’ cells and neutrophils [12,13]. The original idea that cytokines instrumental for these migratory responses and ensuing tolerance would be weaker in PLE was not confirmed, instead the PLE skin showed an IL-1/IL1-receptor a ratio which indicated a basic bias toward inflammatory reactions [14], which indicates an unbalance in the skin reaction, more aberrant in the proinflammatory arm than in the suppressive arm. Two British groups provided the important functional evidence that PLE patients indeed show a diminished UV-induced suppression of sensitization to a contact allergen with physiological UV dosages around the threshold dose for an erythema [15,16], very much in contrast to the earlier mentioned group of people who had skin carcinomas removed.

After pointing out this difference in UV immunosuppressive responses between former skin cancer patients and PLE patients in a lecture, an attentive student will promptly ask whether PLE patients run a lower risk of skin cancer. The answer, until recently, was that we had no data regarding this suggestion. Recently, the group of Gillian Murphy reported on a study that attempted to fill in this gap [17] but, as with most observational epidemiologic studies, there are inherent uncertainties about the causality in the associations that were found. The paper by Murphy and coworkers reported on two case–control studies: one in which patients treated for skin cancers (n = 214 and 210 controls) were investigated on having experienced features of PLE, and one in which PLE patients (n = 100 and 155 controls) were investigated on having had skin cancers. They found a significantly lower prevalence of PLE among skin cancer patients than in controls of 7.5 versus 21.4% (p < 0.001) and a nonsignificant indication, because of low numbers, that the prevalence of skin cancer is lower among PLE patients (4 vs 7.1%). This is clearly supportive of the premise that there should be an inverse relationship between the risks of skin cancer and PLE because of opposite tendencies in UV-induced immunosuppression.

However, behavioral aspects instead of a true physiologic mechanism may also underlie the observed differences: sun avoidance in PLE patients, in order to avoid experiencing the uncomfortable rash, and sun seeking behavior among skin cancer patients, which might increase occurrence of PLE or conversely could, through the many hours of sun exposure linked to non-melanoma skin cancer, ‘harden’ the skin against PLE expression. The authors assessed sun exposures in three levels (low: both work and leisure indoors; medium: either work indoors and leisure outdoors or work outdoors and leisure indoors; and high: both work and leisure outdoors). Based on their data we found that significantly fewer PLE patients were in the high exposure category (3 vs 18%; p < 0.0005), and significantly fewer skin cancer patients in the low exposure category (6 vs 12%; p = 0.037). However, in multifactorial analysis they found that skin cancer history itself is significantly predictive of a lower PLE risk, whereas sun exposure level is not. This would indeed suggest that some inherent personal factor associated with skin cancer risk lowers the risk of PLE. The reverse relationship, that is, a lower risk of skin cancer in PLE patients, could not be established. In support of these findings, Francois Aubin and collaborators found a lower prevalence of PLE among renal transplant patients (2% instead of the expected 17.5% in controls; p < 0.001) who are known to run a much elevated risk of skin carcinomas owing to the immunosuppressive medication they take [18]. They argue that a lower sun exposure of renal transplant patients as an explanation of the low prevalence of PLE is unlikely as only a small percentage complies with the advice to take sun protective measures. They conclude that the data support the involvement of the immune response in both skin carcinomas and PLE. Immunosuppressive drugs are sometimes used to treat severe PLE [19], and the immunosuppressive medication given to the transplant patients may be suppressing PLE expression. The work of Murphy’s group does, however, strongly support a potential reciprocal relationship between PLE and skin carcinomas. Taken together with the experimental work in PLE patients described earlier in this editorial, we suspect that UV-induced immunosuppression may provide the link.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.