The importance of considering skin diseases from a temporal perspective: Special emphasis on the effects of corticosteroids and virally induced diseases

The efficacy and safety of medical therapies assessed on the basis of short‐term outcomes are often considered to inform the long‐term outcomes. Even in corticosteroid therapy, either systemic or topical, which is most frequently used as a first‐line therapy to control inflammation, few studies have reported the risk/benefit ratio for long‐term outcomes. Thus, a temporal perspective should be added to our conventional understanding of the effectiveness of medical therapies. In this review, we initially describe the importance of considering the efficacy of corticosteroid therapy based not only on short‐term outcomes, but also on long‐term outcomes of a certain type of severe drug eruption, drug‐induced hypersensitivity syndrome/drug reaction with eosinophilia and systemic symptoms (DiHS/DRESS). Our view is exemplified by our recent analysis of previously reported cases of DiHS/DRESS, revealing that aggressive treatment such as pulsed corticosteroids and intravenous immunoglobulin, although beneficial in the short term, paradoxically has deleterious effects such as autoimmune sequelae in the long term. Thus, measuring the therapeutic efficacy from only a short‐term perspective is insufficient and follow‐up examination of these patients for at least three years is required even after complete resolution is observed. Various cutaneous diseases refractory to conventional therapies that occur after resolution of herpes zoster also require a temporal perspective to understand how cutaneous diseases develop at a certain site. We describe the clinical usefulness of considering the dual actions of corticosteroids, antiinflammatory, and immunostimulatory, depending on the situations in which they are used.

long-term outcomes. Thus, a temporal perspective should be added to our conventional understanding of the effectiveness of medical therapies. In this review, we initially describe the importance of considering the efficacy of corticosteroid therapy based not only on short-term outcomes, but also on long-term outcomes of a certain type of severe drug eruption, drug-induced hypersensitivity syndrome/drug reaction with eosinophilia and systemic symptoms (DiHS/DRESS). Our view is exemplified by our recent analysis of previously reported cases of DiHS/DRESS, revealing that aggressive treatment such as pulsed corticosteroids and intravenous immunoglobulin, although beneficial in the short term, paradoxically has deleterious effects such as autoimmune sequelae in the long term. Thus, measuring the therapeutic efficacy from only a short-term perspective is insufficient and follow-up examination of these patients for at least three years is required even after complete resolution is observed. Various cutaneous diseases refractory to conventional therapies that occur after resolution of herpes zoster also require a temporal perspective to understand how cutaneous diseases develop at a certain site. We describe the clinical usefulness of considering the dual actions of corticosteroids, antiinflammatory, and immunostimulatory, depending on the situations in which they are used.

| INTRODUCTION
Corticosteroids, either topical or systemic, are most frequently used as a first-line therapy to control inflammation in patients with autoimmune diseases and cutaneous inflammatory diseases. The safety of corticosteroid use over long periods, however, is always a concern, although previous long-term, controlled studies suggest that long-term adverse effects, such as osteonecrosis, are uncommon when the corticosteroid dosage is low. 1,2 Thus, it is important to investigate whether the efficacy and safety in the short term necessarily assure the safety and efficacy in the long term. Even in corticosteroid therapy, however, few studies have demonstrated the risk/ benefit ratio in the long term; an excellent risk-benefit ratio of a certain therapy in the short term may not be necessarily favorable in the long term. It is also difficult to evaluate the effectiveness of medical therapy because of the effect of the disease duration before therapy is started; if treated early, many patients might have a life expectancy similar to that in a healthy control population, but if treatment is delayed, patients with late-stage disease might have a significantly greater risk of morbidity and mortality. Thus, a temporal perspective should be added to our conventional understanding of the effectiveness of medical therapy.
In addition, considering skin diseases as dynamic processes in which phenotypic manifestations evolve in a continuous manner rather than as a distinct disease onset is important in terms of understanding the pathogenesis or management. Indeed, careful follow-up of patients for years may reveal the transformation of an immature form of disease into a more classic or typical form of disease; typical examples of such transitions are often reported and include conversion from granuloma annulare to sarcoidosis and from discoid lupus erythematosus to systemic lupus erythematosus. In this review, we focus on the importance of considering skin disease as a continuous spectrum rather than as a binary process (ie, present or absent). In support of this possibility, we recently demonstrated that a variety of inflammatory diseases such as lichen planus (LP) and lichen amyloidosis (LA) develops in previously healed herpes zoster (HZ) lesions (manuscript submitted). 3 [NOTE: Rather than "manuscript submitted," please report the "authors name(s), personal communication."] We further describe how autoimmune sequelae or autoantibodies (autoAbs) arise long after the clinical resolution of severe drug eruptions and then trigger events that drive the disease process forward.

DRUG ERUPTION S
Severe drug eruptions encompass two distinct clinical entities, the most common being drug-induced hypersensitivity syndrome (DiHS)/ drug reaction with eosinophilia and systemic symptoms (DRESS).
Another is Stevens-Johnson syndrome (SJS)/toxic epidermal necrolysis (TEN). On the basis of clinical comparisons as well as immunologic studies of the two diseases, these two forms of severe drug eruptions likely have distinct pathogenic mechanisms, 4,5 although some of the causative drugs are common. DiHS/DRESS is characterized by sequential reactivations of herpesviruses, probably due to functional defects in regulatory T cells (Tregs). 6 Several autoimmune diseases and the generation of autoAbs concurrently or sequentially occur as sequelae of DiHS/DRESS. [7][8][9] Inflammation and subsequent damage to the skin and liver in patients with DiHS/DRESS are thought to lead to autoAbs generation and the onset of autoimmune disease. 9 Treatment is therefore aimed at suppressing the inflammation early and preventing the development of tissue destruction.
Although systemic corticosteroids are accepted as the gold standard treatment for ameliorating the clinical symptoms of DiHS/DRESS, it is unclear whether aggressive therapies such as pulsed corticosteroids or high doses of intravenous immunoglobulin (IVIG) would be more effective than the usual dosage of prednisolone (40-60 mg/d) for reducing symptoms and damage. Because autoimmune sequelae such as autoimmune thyroiditis, type 1 diabetes mellitus, and lupus erythematosus 10-12 may develop long after clinical resolution of DiHS/DRESS, the effect of systemic corticosteroid treatment for DiHS/DRESS should be evaluated not only for short-term outcomes, but also for long-term outcomes. In this regard, we investigated whether systemic corticosteroid treatment during the acute stage could be efficacious for both short-term and long-term outcomes. In our retrospective analysis, 13 patients with DiHS/DRESS were divided into two groups depending on the use of systemic corticosteroids during the first six-month period. During this period, various clinical symptoms due to reactivation of herpesviruses and other bacterial infections occurred more frequently in the corticosteroid-treated group than in the noncorticosteroid-treated group; in the corticosteroid-treated group, the findings were thought to relate to the reduction or withdrawal of corticosteroids. Flare-ups of the original symptoms of the disease also occurred more frequently in the corticosteroid-treated group upon reduction or withdrawal of the corticosteroid dose. Marked deterioration of various clinical symptoms is observed following accidental discontinuation or rapid tapering of corticosteroids. 13,14 We recommend that corticosteroids started at a dose of 40-60 mg/d be tapered over a period of six to eight weeks to prevent the relapse of various clinical symptoms and therefore continued for two to three months. Corticosteroid doses should be reduced gradually (10 mg/d for about two weeks), even upon clinical resolution of the disease manifestations. A reasonable explanation for this is that DiHS/DRESS represents another manifestation of immune reconstitution inflammatory syndrome (IRIS), which is observed with a broad-spectrum of immunosuppressive therapyrelated opportunistic infectious diseases. 14,15 According to the con- long-term outcomes, contrary to our initial expectation. In this case, would administration of pulsed corticosteroids or IVIG, which is effective for treating SJS/TEN, also be effective for treating DiHS/ DRESS? We are concerned that these aggressive therapies may have adverse consequences in either the short or long term, because these therapies are intrinsically associated with a process that requires either rapid or great reduction or discontinuation of the prednisolone immediately after therapy.
In this regard, we recently searched for published DiHS/ DRESS cases with autoimmune sequelae. Most, if not all, reported cases with autoimmune sequelae had clinical features consistent with type III polyglandular autoimmune syndrome, which is characterized by the coexistence of at least two glandular autoimmune diseases. 16 Most importantly, pulsed corticosteroids or IVIG was used to treat DiHS/DRESS during the acute stage in four of the five reported DiHS/DRESS cases that eventually developed type III polyglandular autoimmune syndrome.
Although the current literature recommends aggressive treatment with systemic corticosteroids followed by a slow taper or IVIG therapy, autoimmune sequelae developed more than six months after successful resolution of the initial clinical symptoms with these therapies in the previously reported cases. 7,10,17 These results indicate that treatment outcomes in DiHS/DRESS should be assessed with long-term follow-ups of at least six months after clinical resolution.
In contrast to the short-term outcome, in which noncorticosteroid treatment was thought to be more beneficial than corticosteroid treatment, our data on the effect of systemic corticosteroids on the long-term outcomes of DiHS/DRESS suggest the opposite of what would be expected from the results of the short-term outcomes. Autoimmune disease developed more frequently in the noncorticosteroid-treated group than in the corticosteroid-treated group. 9 Considering that inflammation associated with severe tissue damage is thought to be an underlying or preceding event for the subsequent development of autoimmune diseases, the severe tissue damage associated with Treg dysfunction observed in the acute to resolution stages (days 11-36) of DiHS/DRESS may increase the risk for developing autoimmune disease. We then considered whether the generation of autoAbs against plakin family protein was related to tissue damage in the acute stage and if the subsequent generation of autoAbs could be prevented by the use of systemic corticosteroids. We found that severe liver damage during the acute stage of DiHS/DRESS was associated with subsequent generation of auto-Abs to the plakin family and that the autoAb generation could be DRESS at the acute stage should no longer be loosely referred to as cured, because our data strongly support the need for improved longterm management for these patients, especially for these serious complications that occur after one to two year disease-free intervals.

DURING AN EPISODE OF HE RPES ZOSTER
The usual clinical features of herpes zoster (HZ) are readily recognized. In most cases, the characteristic features of HZ rash establish the clinical diagnosis. Varicella zoster virus (VZV), however, also has an atypical presentation. These lesions encompass folliculitis, 20 syringitis, verruciform lesions, 21 nodular lesions, 22 intracranial hemorrhage, 23 and acute abdomen. 24 Involvement of the hair follicles or eccrine epithelium by VZV is an infrequently described histologic pattern mostly reported in HIV-infected patients and rarely reported even in immunocompetent patients. In these lesions, VZV DNA and VZV-specific antigens such as VZV major envelope glycoprotein (gE) are detected in follicular keratinocytes, the epithelium of eccrine glands, or the gastric epithelium, which may be associated with cytopathic changes with intranuclear inclusions. The diagnosis of HZ can be also confirmed by polymerase chain reaction (PCR) analysis for VZV in the cerebrospinal fluid. Pruritus is also reported in many patients with HZ, not only during the herpetic period, but also in the pre-or postherpetic period. VZV gE can be detected as early as two to three days after the onset of HZ, and expression persists for at least several months in HZ lesions, whereas VZV DNA disappears much earlier than VZV gE antigens. 3 Because VZV gE is an essential late protein expressed on the cell membrane during lytic infection and is a predominant component of the virion envelope, 24,25 immunohistochemical detection of VZV gE in these lesions, either cutaneous or gastrointestinal, suggests VZV involvement in the development of these lesions.
Severe abdominal pain may occur as a serious manifestation of HZ, known as abdominal zoster, 26,27 particularly in an immunosuppressed host. 27 In these patients, the diagnosis of abdominal zoster is not usually considered until the typical vesicular eruptions begin to appear. Because the typical vesicular eruptions could be entirely absent in some patients with abdominal zoster, 28 however, the diagnosis in the setting of the absence of cutaneous lesions (zoster sine herpete) can only be made by PCR analysis of blood or saliva and biopsy or autopsy samples. Indeed, autopsy studies reveal high frequencies of abdominal involvement of VZV. 27 In these cases, herpetic lesions on the serosal and mucosal gut wall were observed at laparotomy. In this regard, our previous immunohistochemical study clearly demonstrated that detection of VZV antigens in the eccrine epithelium can be used as a highly reliable and useful clue to the diagnosis of HZ with unusual manifestations and inflammatory dermatoses induced by VZV infection. 3 Immunoglobulin A vasculitis (IAV) may develop before or after the appearance of varicella or zoster rash, 28,29 and therefore, we hypothesized that IAV associated with gastrointestinal symptoms results from VZV reactivation. Consistent with this hypothesis, there are similarities in the nature of the abdominal pain and endoscopic findings in gastrointestinal lesions between patients with IAV and gastric or enteric zoster. 30 We therefore investigated whether VZV gE antigens could be detected in both cutaneous and gastrointestinal lesions, and whether a significant increase in VZV IgG titers could be specifically detected in IAV patients with severe abdominal pain. 31,32 We analyzed consecutive IAV patients with severe abdominal mani- The term "postherpetic isotopic response" or "Wolf's isotopic response" encompasses a wide spectrum of clinical entities, ranging from LP to sarcoid granuloma. [33][34][35] Among them, granulomatous reactions ranging from granuloma annulare to granulomatous vasculitis at sites of resolved HZ lesions are well described. 34,36 These granulomatous reactions can occur immediately after the resolution of vesicular lesions, or at various times after the HZ. Earlier studies using PCR, however, failed to identify VZV DNA in the granulomatous reactions arising between one month and up to four years after the resolution of HZ. 37 These results may indicate that persistent expression of VZV gE antigens in the resolved HZ lesions could induce VZV-directed immune responses.
In addition to granulomatous reactions, many cutaneous disorders occur within healed HZ lesions, including LP, lichen simplex chronicus, 38 pseudolymphoma, 39 psoriasis, 40 lichenoid chronic graftvs-host disease, 41 and lymphoma. 42 The term "postherpetic isotopic response" or "Wolf's isotopic response" describes the occurrence of a new, unrelated disease appearing at the same location as previously healed HZ or, rarely, herpes simplex lesions. At present, however, there are no satisfying explanations for how these cutaneous disorders develop at the healed herpetic lesions. Recent reports on herpetic syringitis suggest the involvement of sweat glands in VZV infection. 43 Although localized unilateral hypohidrosis with a dermatomal or segmental distribution is suggestive of VZV involvement, a relationship between localized hypohidrosis and VZV infection has not yet been demonstrated, probably due to the low frequency of detection of VZVDNA in the lesions. Nevertheless, in view of our recent observation that long-lasting hypohidrosis localizing to the involved dermatome in healed HZ lesions is relatively common, 44 hypohidrosis after HZ might contribute to the onset of these secondary isotopic diseases at the involved dermatomes. We recently experienced two patients who developed LP and LA, respectively, in healed HZ lesions, prompting us to question whether hypohidrosis after VZV reactivation could be involved in the development of these isotopic responses. Because these lesions were characterized by dry skin and thought to be induced by postherpetic hypohidrosis, we examined sweating responses to thermal stimulation in these lesions using the impression mold technique, which allows for accurate quantification of the activity of each sweat gland/duct to produce and deliver sweat. 45  sweating responses to thermal stimulation is a relatively common sequel of HZ but most of these cases may go unrecognized due to the absence of apparent clinical symptoms. Unless a history of HZ is sought in patients characterized with isotopic responses in a dermatomal distribution, the lesions would remain idiopathic.

IMMUNE RESPONSES
Although the actions of corticosteroids are generally thought to be immunosuppressive, in some cases their actions may be immunostimulatory. Evidence is accumulating that stress-induced enhancement of immune function could be mediated by the action of corticosteroids in the acute phase of a stress response. Thus, the effect of corticosteroids may differ depending on the status of the activation of immune cells and tissue microenvironment in which the immune cells are activated. In addition, there may be confusion between a direct drug effect and an unanticipated consequence of a decrease in the dose of corticosteroids. As described above, the development of IRIS can also occur in non-HIV patients receiving immunosuppressive agents, such as prednisolone and tumor necrosis factor -α inhibitors, upon their reduction and withdrawal. 51 Based on these considerations, we postulate that withdrawal of corticosteroids could potentiate immune responses against tumors when applied before antitumor therapy (Figure 2).
Despite the initial enthusiasm for topical imiquimod (IMQ) as a novel treatment modality for malignant neoplasms, a recent review article concluded that its efficacy is limited to a given disease setting. 51 Thus, an adjunct therapy that can improve the antitumor efficacy of IMQ is needed. We initially hypothesized that decreasing the recruitment of Treg cells at the tumor site prior to IMQ therapy would have therapeutic value, because our previous unpublished preliminary study on IMQ monotherapy showed that Treg cells occur in significantly lower frequencies in dermal infiltrates of Bowen's disease lesions before treatment in patients who eventually exhibited a complete response compared to patients with a partial response.
Starting IMQ therapy at a time when the frequency of Treg cells is lowest could be essential for achieving more robust immune responses to tumor cells. We conducted an open-label, nonrandomized study to investigate whether sequential therapy with topical corticosteroids and IMQ could produce a higher clearance rate for Bowen's disease than IMQ monotherapy. 52 Patients with Bowen's disease were assigned to either IMQ monotherapy or sequential therapy. A complete response occurred at 8 weeks in all patients receiving the sequential therapy, but only in some patients in the IMQ-treated group (37.5%). Our immunohistochemical analysis showed that sequential therapy greatly improved the antitumor efficacy of IMQ by excluding tumor-associated Tregs from the lesions before starting IMQ therapy. We recommend using topical corticosteroids before IMQ therapy to provide marked pain relief and an excellent subsequent response to IMQ, because prior therapy with topical corticosteroids before starting IMQ therapy reduced adverse SHIOHARA AND MIZUKAWA | 89 reactions such as local skin irritation at the application site and none of the patients discontinued the therapy because of such adverse events. Particularly in the treatment of large or very sensitive areas such as the genitalia, topical corticosteroids may have a strong effect not only for enhancing the treatment efficacy of IMQ, but also for reducing pain. Our immunohistochemical studies clearly demonstrated that Treg cells were more profoundly deleted from tumor lesions after a two-week treatment with topical corticosteroids, whereas CD8 + T cells were recruited to the lesions more rapidly than Treg cells after starting IMQ therapy (Figure 3). Such time-dependent changes in the pattern of T-cell recruitment in sequential therapy were reflected by the dramatic increase in the ratio of CD8 + T cells to Treg cells at two to four weeks after starting IMQ therapy. 52 Thus, the use of topical corticosteroids, when induced before antitumor therapy, enhances the antitumor effects of IMQ on the one hand, but may aggravate the consequences of contact sensitization on the other hand. It is therefore necessary to examine whether frequent withdrawal of topical corticosteroids paradoxically aggravates atopic dermatitis lesions in patients with atopic dermatitis where the immune system has already been Th2-primed by allergens. If so, topical corticosteroids may have both antiinflammatory and immunopathologic effects, depending on the situation in which topical corticosteroids are used.

| CONCLUSION
Anticytokine therapies have recently received special attention as treatment modalities for allergic diseases. Detailed characterization of the timeline of the immunologic events occurring in patients treated with these anticytokine therapies, however, is necessary.
Although IL-4/IL-13 blockade is considered most effective in ameliorating atopic inflammation, it remains unknown whether these anticytokine therapies could have beneficial long-term outcomes when used over a prolonged period of time. It is not clear whether longterm inhibition of the activity of these cytokines will be safe in F I G U R E 2 Difference in the antitumor efficacy between IMQ monotherapy and sequential therapy with topical corticosteroids and IMQ humans. These previous analyses were largely performed in mouse models, and thus more patient-based studies are required to address these important issues given the differences in skin physiology between human and mice.