SEARCH

SEARCH BY CITATION

Keywords:

  • Koebner phenomenon;
  • vitiligo;
  • topical immunomodulators;
  • treatment;
  • experimental model

Summary

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

In this study, we developed an in vivo vitiligo induction model to explore the underlying mechanisms leading to Koebner’s phenomenon and to evaluate the efficacy of therapeutic strategies. The model consisted of 12 pigmented test regions on the back of generalized vitiligo patients that were exposed to three Koebner induction methods: cryotherapy, 755 nm laser therapy, and epidermal abrasion. In addition, four cream treatments (pimecrolimus, tacrolimus, steroid and placebo) were randomly applied. Koebnerization was efficiently induced by all three induction methods. In general, cryotherapy was the best method of Koebner induction, followed by 755 nm laser therapy and epidermal abrasion. Reproducible results were obtained, which showed enhanced depigmented surface areas and higher amounts of T lymphocytes in placebo-treated test zones compared to active treated areas. Tacrolimus and local steroids were better inhibitors of Koebner’s process (P < 0.05) compared to pimecrolimus. Our in vivo vitiligo induction model is very informative to investigate vitiligo induction and to determine the efficacy of topical treatments in vitiligo. This proof of concept confirms the efficient comparison of head-to-head therapeutic strategies intra-individually in a standardized, specific and better timed way.


Significance

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

Koebner’s phenomenon in vitiligo is a well-known, although scarcely investigated, process. Nonetheless, the possibility to actively induce koebnerization on a localized area makes it suitable and interesting for experimental models. In this pilot study, we describe the proof of concept of a new experimental vitiligo induction model, which allows both to investigate the pathophysiology of traumatic induced depigmentation and to compare adequately the efficacy of treatments in a standardized and reproducible way. We believe this is an excellent model to unravel currently remaining questions in vitiligo.

Introduction

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

Vitiligo is a common acquired pigmentary disorder, affecting approximately 1% of the world population. A significant effect on the quality of life has been documented in previous studies (Bilgiç et al., 2011; Kostopoulou et al., 2009). Commonly used conventional treatments include UV therapy (UVB/PUVA), local steroids, calcipotriol and more recently topical immunomodulators (tacrolimus and pimecrolimus). The efficacy of the treatments is difficult to evaluate as the responses are often limited and standardized measures are lacking in the available clinical trials (Whitton et al., 2010). Moreover, active vitiligo lesions are clinically undistinguishable from established lesions, and it would require a biopsy to demonstrate the inflammatory infiltrate.

The exact origin of vitiligo is unknown, although there is accumulating evidence for a major etiologic role of melanocyte-specific cytotoxic T cells in progressive vitiligo (van den Boorn et al., 2009; Mandelcorn-Monson et al., 2003). Both helper and cytotoxic T cells from progressing margins generate predominantly type 1 cytokines, namely IFN-γ and TNF-α, suggesting that vitiligo is a T helper 1 (Th1)-mediated disease. This theory is supported by the fact that various effective treatment options in vitiligo have an immunosuppressive effect on the activation and maturation of T cells (e.g. local steroids and topical immunomodulators) (Ahmed et al., 2001; Paul and Seder, 1994). The finding of melanocyte-specific T cells in the blood and perilesional skin of vitiligo patients has been a major breakthrough (Mandelcorn-Monson et al., 2003). Furthermore, also genetic evidence has been provided for immune-based melanocyte elimination in vitiligo, as nearly all of the susceptibility genes encode components of the immune system [e.g. major histocompatibility complex (MHC), protein tyrosinase phosphatase, non-receptor type 22, NACHT leucin-rich-repeat protein-1] (Jin et al., 2011). However, the extent of the role of immune cells in the pathophysiology of vitiligo and the Koebner process may be best evaluated by assessing the objective preventive effect of anti-inflammatory treatments. Besides the autoimmune theory, other etiopathogenetic mechanisms have been proposed such as defective melanocyte adhesion and a deficient production of melanocyte growth factors (e.g. stem cell growth factor and basic fibroblast growth factor) (van Geel et al., 2011).

The hallmark of organ-specific autoimmune diseases is the breakage of tolerance to self-antigens leading to tissue destruction. However, the underlying mechanisms for this phenomenon in vitiligo remain to be investigated. Based on clinical observations in vitiligo, environmental initiators of breaking tolerance can be caused by tissue damage (Koebner’s phenomenon) (Lee and Modlin, 2005). Minor trauma alone is well known to cause new skin lesions in vitiligo patients. Furthermore, the typical common distribution pattern of generalized vitiligo lesions on pressure and friction points of the skin suggests that mechanical stress is a precipitating factor. Koebner’s phenomenon is evident in a high percentage of vitiligo patients, and some studies propose that vitiligo always develops after some type of koebnerization (Gauthier, 1995; Sweet, 1978). Further support for a crucial role of Koebner’s phenomenon in vitiligo is provided in an animal model. It was demonstrated that local inflammation at the site of self-antigen endogenous expression was required to break tolerance against a melanocyte self-antigen (murine tyrosinase-related protein 2) (Lane et al., 2004). Presently, the mechanism by which Koebner’s phenomenon promotes autoimmune destruction of melanocytes in vitiligo is not known. As demonstrated in our previous review on Koebner’s phenomenon, trials with respect to this issue are still missing (van Geel et al., 2011). Insight into this process may have major implications not only for the pathogenesis of vitiligo but also for treatment modalities of other autoimmune diseases or by providing a road map for melanoma therapy.

In this study, we introduce a new vitiligo induction model (experimentally induced Koebner phenomenon) to investigate the Koebner process and evaluate the efficacy of therapies. The clinical rational for this model is 3-fold: (i) it will lead to valuable research material obtained during the early dynamic and active stage of the Koebner process that may mimic the development of vitiligo in general; (ii) it enables intra-individual, head-to-head comparison of different topical treatments in vitiligo and will therefore help to guide the clinical practice; and (iii) it provides a standardized, efficient, and very informative method for vitiligo investigations. The primary purpose of this prospective study was to determine the usefulness of this in vivo model. The secondary aim was to investigate the clinical, histopathological, and immunological process of vitiligo induction (experimental KP) and to evaluate the therapeutic value of cream treatments and their effect on the immune balance.

Results

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

Patients

In total, five patients met the inclusion criteria of our study (Figure 1); however, two patients declined to be enrolled because of logistic reasons. All remaining three patients completed the intervention and the follow-up as planned. All participants were included in the main analysis. Age-eligible participants were recruited among patients attending our clinic for a visit at the time of the inclusion period. The characteristics of the participants who were actually included are summarized in Table 1.

image

Figure 1.  Design of the study.

Download figure to PowerPoint

Table 1.   Baseline demographic and clinical characteristics
 Patient 1Patient 2Patient 3
  1. *BSA, % body surface area.

Age (yr)484728
Sex (F/M)FFF
Ethnic originSurinameseCaucasianNorth African
Skin type524
Total extent depigmentations (BSA)*60%70%65%

Clinical evaluation of the in vivo vitiligo induction model

Twelve square test areas were allocated to the back of each patient where three induction methods were used and four treatments could be applied for each induction method. As such, the efficacy of induction methods and treatments was evaluated by comparing the depigmented surface area between the test sites. Koebnerization was efficiently induced in all three patients resulting in depigmented areas, which were observed both clinically and by Wood’s lamp investigation (Figure 2). At the last time point (7–11 months), the median percentage of depigmented surface area at placebo (petrolatum)-treated sites was 55.2% [interquartile range (IQR): 40.4–77.3%]. This represents the natural evolution after the Koebner induction methods, although some minor influence of hydratation might be present.

image

Figure 2.  Clinical pictures of patient 1 showing baseline situation, cream application, and depigmentation at the koebnerized test areas over time (1 month and 7 months after vitiligo induction).

Download figure to PowerPoint

Evaluating all three patients, cryotherapy, 755 nm laser, or epidermal abraded koebnerized test regions showed no significant differences in elicited depigmented surface area. Similar reactions to the Koebner induction methods were observed in patients 1 and 2 (Figure 3). Cryotherapy was the most effective vitiligo induction method, followed by 755 nm laser therapy whereas less depigmentation was achieved at epidermal abraded test regions. However, in patient 3, cryotherapy was significantly less effective in inducing depigmentation (P = 0.023), while 755 nm laser therapy was more effective and epidermal abraded test areas showed even significantly larger depigmented surface areas compared to the other patients (P = 0.009) (Figure 3).

image

Figure 3.  Mean depigmented surface area observed in cryotherapy (A), 755 nm laser-treated (B), and epidermal abraded (C) test areas per patient at time point 1 (1 month) and time point 2 (7–11 months).

Download figure to PowerPoint

Clinical evaluation of the efficacy of therapeutic intervention

Active treatment with corticosteroids, tacrolimus, and pimecrolimus all had a clear capacity to prevent depigmentation compared to placebo. The percentage of depigmented surface area was higher at all placebo-treated sites (n = 9/9) compared to all therapeutic test regions at the first time point (P < 0.05) and at the last time point (P < 0.001) (Figure 4). At the last time point, more than a 2-fold [59.3% (IQR: 38.0–84.1%)] smaller depigmented surface area was observed in active treated lesions compared to placebo.

image

Figure 4.  Comparison of the efficacy of corticosteroids, tacrolimus, and pimecrolimus to prevent depigmentation in all test areas: (A) all time points/overall (B) at the first time point (1 month) and (C) at the last time point (7–11 months) of evaluation (mean ± 1 SD). Overall (A) tacrolimus and local steroids were significant better inhibitors of the Koebner process compared to pimecrolimus (P < 0.01).

Download figure to PowerPoint

Tacrolimus was a significantly better inhibitor of the Koebner process compared to pimecrolimus in all test areas (n = 9/9 lesions) (P = 0.003), while corticosteroid treatment also performed better than pimecrolimus (P = 0.058). Tacrolimus-treated test areas had a 84.3% (IQR: 59.4–94.1%) less extensive depigmented surface area compared to placebo, corticosteroids 70.6% (IQR: 38.7–92.0%), and pimecrolimus 42.8% (IQR: 37.0–59.5%). No significant differences in responses were observed between local steroids and tacrolimus.

Reproducibility of the model

Intra-individual reproducibility of the model was confirmed by a significant correlation of the percentage of depigmented surface area and the applied treatments in patients 1 (P < 0.001, ρ = 0.615) and 2 (P = 0.010, ρ = 0.526) (adjusted for time point). Significance between the percentage of depigmented surface area and the applied treatments was not reached in patient 3 (P = 0.081) mainly because of the reduced ability of cryotherapy to induce depigmentation in this patient. If cryotherapy was not taken into account, a significant correlation was also observed in this patient (P = 0.025, ρ = 0.575).

Interindividual reproducibility of the model was investigated by comparing the percentage of depigmented surface area according to the treatment in the different patients. In all patients, placebo-treated lesions had the most extended depigmented surface area for the three induction methods (n = 9/9 lesions), whereas tacrolimus was in all cases the best preventive factor (n = 9/9 lesions). So, reproducible results were achieved for all treatment modalities, both intra- and interindividually.

Pathological evaluation of the model

Biopsies (n = 17) were taken at day 3, day 10, and day 30 in several cryotherapy or epidermal abraded test zones. Immunohistochemistry revealed the presence of an inflammatory infiltrate mainly consisting of CD4+ and CD8+ T cells, in some cases clearly migrating toward the basal layer (Figure 5). CD4+ and CD8+ T cells were enhanced over time until day 10 at placebo-treated test sites (P = 0.010 and P = 0.048, respectively) and decreased again at day 30 (Figure 6).

image

Figure 5.  Comparison of histological pictures (day 10) between placebo and active treatment.

Download figure to PowerPoint

image

Figure 6.  Bars (mean ± 1 SD) showing the number of CD4 and CD8 T cells after koebnerization in placebo-treated test zones over time (A, B) and compared between placebo-, corticosteroid-, and tacrolimus-treated test zones (C, D).

Download figure to PowerPoint

The number of CD1a+ Langerhans cells decreased after koebnerization, which may be due to the direct cytotoxic effect of cryotherapy on epidermal Langerhans cells (data not shown). The absolute number of Foxp3+ cells showed a similar pattern as CD4 staining with an increase over time until day 10 followed by a decrease at day 30. However, the ratio between the number of Foxp3 cells and CD4+ T cells (median: 5.4%; IQR: 1.8–18.0%) decreased after Koebner induction compared to baseline (P = 0.012). After cryotherapy, melanocytes displayed an abnormal structure on Mart-1 staining with an enlarged cytoplasm, long clubbed dendrites, and remnants of melanocytes staining for MelanA throughout the basal layer (Figure 7). Interestingly, similar structural abnormalities have been described in unstable vitiligo and have been linked to a deficient adhesion capacity (Gauthier et al., 2003; Kumar et al., 2011).

image

Figure 7.  Remarkable MelanA staining 10 days after koebnerization using cryotherapy.

Download figure to PowerPoint

Pathological evaluation of the therapeutic interventions on the Koebner process

In all three patients, CD4+ and CD8+ T lymphocytes were significantly reduced in active treated lesions compared to placebo (P < 0.001 and P < 0.001) (Figures 5 and 6). Three days after koebnerization, a mean number of 261 CD4+ T cells/high power field (HPF) and 67 CD8+ T cells/HPF was found in placebo-treated lesions compared to 108 CD4+ T cells/HPF and 24 CD8+ T cells/HPF in active treated test zones. No clear difference was observed between tacrolimus (mean CD4 T cells: 108 cells/HPF; mean CD8 T cells: 23 cells/HPF) and corticosteroid treatment (mean CD4 T cells: 147 cells/HPF; mean CD8 T cells: 25 cells/HPF). As such, these results are in agreement with the clinical observations showing a decreased depigmented surface area in active treated test zones. The number of CD1a cells and Foxp3 cells did not differ between the applied treatments (data not shown).

Immunomonitoring of induction and treatment

In one patient (patient 1), biopsies were taken at baseline and at day 10 after cryotherapy in a placebo- and tacrolimus-treated test zone for flow cytometry analysis. In the placebo-treated test region at day 10, the isolated lesional T cells displayed a decrease in CD4/CD8 ratio compared to baseline with a concomitant increase in Mart-1- and Gp100280–288-specific CD8+ T cells (Table 2). At day 10, the lymphocytes in the tacrolimus-treated area had an increased CD4/CD8 ratio and lower presence of Gp100280–288- or Mart-126–35-specific lymphocytes, as compared to the placebo-treated test area (Table 2).

Table 2.   Flow cytometric analysis of the lesional T cell infiltrate at baseline and after koebnerization in a placebo- and tacrolimus-treated lesion
 Percentage of CD3+ cells
 Percentage tetramer-positive CD8+ cells
CD4CD8CD4/CD8Mart-126–35 CD8 T cellsGp100209–217 CD8 T cellsGp100280–288 CD8 T cellsTyr369–377 CD8 T cellsFlu58–66 CD8 T cells
PBMCs25.851.70.500.110.010.490.0120.00
Lesional Day 130.544.40.690.130.180.600.0010.04
Lesional D10: Tacrolimus treated48.021.32.250.150.110.480.0040.06
Lesional D10: placebo treated23.059.00.390.290.121.080.0020.03

Adverse events

All patients completed the trial without a serious side effect. However, in one patient, two small hypertrophic scars appeared at two biopsy locations. They were subsequently successfully treated with intra-lesional steroid (triamcinolone acetonide).

Discussion

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

In this article, we introduce a reproducible model to investigate the process of koebnerization and compare the efficacy of therapeutic agents for vitiligo. This pilot study involved the application of several topical creams on the back of three patients after the induction of vitiligo using three different koebnerization methods. This vitiligo induction model is easy to reproduce on a standardized way and allows objective evaluation of outcome measures (digital image analysis system). Koebnerization was efficiently induced in all three patients [median depigmentation at placebo treated sites: 55.2% (IQR: 40.4–77.3%)], although the capacity of the induction method differed individually. Nonetheless, the applied treatments were highly effective [median inhibition of depigmentation at active treated sites: 59.3% (IQR: 38.0–84.1%)], and these results were reproducible both inter- and intra-individually. The amount of CD4+ and CD8+ T cells increased markedly after koebnerization and could be partly prevented by topical anti-inflammatory therapy. This implicates that mometasone furoate, pimecrolimus, and tacrolimus all resulted in a good clinical response in this experimental model, which reinforces the use of these treatments especially in the active phase of vitiligo. Furthermore, these remarkable clinical responses confirm the key role of an inflammatory-mediated destruction of melanocytes in the Koebner process. As treatments in this study were only applied during the first week, these results stress the importance of an early intervention during the active inflammatory phase, which seemed to peak at 10 days after induction. The marked preventive effects of topical anti-inflammatory creams observed here contrast with the often disappointing clinical results in daily practice. Our results suggest that the timing of therapy during the early inflammatory phase may be critical to obtain a good clinical response that could prevent development of new depigmentations.

Future studies using this model may provide new essential information on the Koebner process in vitiligo. The production of proinflammatory cytokines and chemokines might facilitate lymphocyte homing to the skin and migration toward melanocytes. The inflammatory response might also expose melanocytes by upregulating their MHC expression, which is normally low or absent. The clarification of this immunological process in vitiligo is largely obstructed by the fact that immune infiltrates most often dissipate, once the lesions are clinically visible, which was confirmed in our model by a spontaneous decrease of T cells at day 30. By introducing this vitiligo induction model, this crucial practical limitation can be overcome.

Furthermore, as the role of topical cream treatment in vitiligo is controversial and its evaluation is still a difficult task, this model could be helpful in clarifying this issue. We realize that this study is mainly focusing on the effect on actively spreading lesions. Nonetheless, our model can also have implications for the design of further studies on therapeutic interventions for the maintenance of repigmentation.

So far, several studies have reported improvement with local steroids, pimecrolimus and tacrolimus, especially on head and neck regions or when combined with UV therapy or sun exposure, whereas others have shown only limited improvement using these products (Choi et al., 2008; Lotti et al., 2008). In this pilot study, we found that pimecrolimus was less efficient in preventing the development of depigmentations than tacrolimus and corticosteroids.

A limitation of this study is that the intervention was only performed in three patients with extensive generalized vitiligo patients. So, we cannot conclude that all patients with generalized vitiligo will react similar on vitiligo induction and cream treatments. Nonetheless, these patients were from different origin, implicating that this model is likely to be useful for all skin types.

In summary, as the management of vitiligo is still often frustrating for both the patient and physician and the effect of vitiligo on the quality of life is significant, further randomized controlled trials are warranted to improve the management of this recalcitrant disorder. Our study shows that all treatments are well effective in preventing the Koebner process in vitiligo both in reducing depigmentation clinically and in decreasing the inflammatory infiltrate on histology. This pilot study revealed that an early and short treatment is sufficient to reduce depigmentation resulting from trauma and identified pimecrolimus as a less efficient therapeutic agent, as compared to tacrolimus and corticosteroids. Future research with this experimental setup may lead to new insights into both the pathophysiology and therapy of vitiligo.

Methods

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

This is a monocenter double-blind placebo-controlled trial conducted at the department of Dermatology, Ghent University Hospital. This pilot study took place from March 2010 to April 2011, while recruitment of patients was limited to the first 5 months. The study was approved by our Ethics committee, and all patients gave a written informed consent. Eligible participants were all adults aged 18 or over with extensive generalized vitiligo (body surface area involvement of vitiligo >60%), who were seeking a total depigmentation of the skin and met the eligibility criteria for depigmenting therapy according to the British guidelines (Gawkrodger et al., 2008). Exclusion criteria included pregnancy, lactation, and use of any topical treatment or UV exposure on the back during the study.

Setup of the in vivo vitiligo induction model

Traumatic factors have long been recognized as triggering factors for depigmentation in vitiligo patients. We chose three different induction methods: cryotherapy, pigment laser, and epidermal abrasion (Figure 1). The treatment phase consisted of a topical treatment applied at baseline (day 0), day 3, and day 6. As hydratation may exert also a certain protective effect, petrolatum ointment was used as a placebo control. Twelve test regions (2.5 × 2.5 cm) on the back of each patient were exposed to three different Koebner induction methods: (i) cryotherapy (induces mainly inflammation), (ii) alexandrite (755 nm) laser therapy (Candela Alex TriVantage; Candela Corporation, Wayland, MA, USA) (induces mainly melanocyte destruction), and (iii) epidermal abrasion using CO2 laser (Coherent UltraPulse; Parallas Technology Inc, Waltham, MA, USA) or Stiefel-curette (induces epidermal wounding). Cryotherapy was performed by gently rolling three times over the test area with a cotton tip that was dipped into liquid nitrogen. Alexandrite laser treatment was carried out using a 3 mm spot size; 10 J/cm2 was used for all skin types, except skin type 5 that required 6.5 J/cm2 to achieve a similar skin reaction. CO2 laser induction was performed using a 3-mm hand piece, single pass at 300 mJ. In addition, four cream treatments were randomly applied on these injured test regions (at baseline, day 3, and day 6): (i) pimecrolimus 1% cream (Elidel, Novartis, Vilvoorde, Belgium); (ii) tacrolimus 0.1% ointment (Protopic, Fujisawa, Leuven, Belgium); (iii) local steroid (Mometasone furoate 1 mg/g; Elocom, Schering-Plough, Brussel, Belgium); and (vi) placebo in the form of petrolatum ointment (Vaseline; 100% petroleum jelly, Unilever, Brussel, Belgium) to exclude bias resulting from hydratation. Two pea-sized amounts of cream were used in all the test zones, which were subsequently covered by ordinary dressing (Cosmopor E; Hartmann, Heidenheim, Germany).

Clinical evaluation of test regions was performed at baseline, after 1 month (time point 1), and a last evaluation after 7−11 months (time point 2). Histological evaluation was performed at baseline and at days 3, 10, 30, and 60. Photographs were taken with and without Wood’s light using a digital camera (Nikon, Brussels, Belgium) at baseline and throughout the study period. Patients were encouraged to contact our department if they had adverse effects (such as pain or itching at test sites). Side effects were also checked during each visit.

Randomization procedure

Each test region per induction method was assigned following a randomization procedure to one of the four treatments, sequentially numbered. The allocation sequence was concealed from the investigator (RS) who performed the semiautomatic digital surface analyses of skin pigmentation. The randomization code was broken after the enrolled participants completed all baseline assessments and all primary end points were available.

Besides the physician who performed the intervention, patients and other investigators were kept blinded to the allocation. Blinding of patients was assured by the location of the test lesions (back) and the additional skin protection with white bandages after each cream application. Patients were not allowed to remove the bandages or apply the creams by themselves.

Evaluation of depigmentation

Digital clinical pictures were taken with and without Wood’s light. The depigmentation area in the different test zones was determined using image j software (NIH, Bethesda, MD, USA). The depigmented zone was evaluated semiautomatically by a blinded investigator (RS) using a binary threshold filter (Multicolor Threshold filter plugin, image j Software). The depigmented area was then calculated as a percentage of the total test area. The preventive capacity of active treatment was determined by comparing the depigmented area in test zones with active treatment to the depigmented area in the corresponding placebo-treated test zones [(1−(% depigmented area in active treated test zone/% depigmented area in placebo-treated test zone))×100]. Data are expressed as median (IQR), and significance was set at P < 0.05.

Immunohistochemistry

Biopsies were taken at baseline and at day 3, 10, and 30 after cryotherapy or epidermal abrasion. The paraffin-embedded tissue was cut into 4-micrometer sections and stained with the appropriate antibodies in a DAKO Autostainer (Dako, Meerbeke, Belgium) after heat-mediated antigen retrieval (PT Link; Dako, Meerbeke, Belgium). To evaluate the inflammatory infiltrate, staining with antibodies against CD1a, CD3, CD4, CD8 (Dako, Meerbeke, Belgium), and Foxp3 (eBioscience, Vienna, Austria) was performed. The presence of melanocytes was evaluated by MelanA/Mart-1 staining (Dako, Meerbeke, Belgium). For all antibodies, an incubation time of 30 min was used except for Foxp3 that required an incubation time of 1 h. The staining was visualized by the Envision Flex system (Dako, Meerbeke, Belgium) except for Foxp3 where the classical avidin–biotin–peroxidase method was carried out. 3-Amino-9-ethylcarbazole was used as chromogen, and counterstaining of the nuclei was carried out with hematoxylin.

HLA phenotyping and tetramer staining

The HLA phenotype of the patients was determined by flow cytometry using HLA-A1 (BD Biosciences, Erembodegem, Belgium), HLA-A2 (BD Biosciences, Erembodegem, Belgium and OneLambda, Canoga Park, CA, USA), and HLA-A3 antibodies (BD Biosciences, Erembodegem, Belgium). In one HLA-A2 patient, three biopsies were taken: one at baseline and after 10 days in a tacrolimus- and placebo-treated test area. T cells were isolated using the lymphocyte skin isolation method according to Clark et al. (2006). T cells were expanded with anti-CD3/CD28 dynabeads (Invitrogen, Merelbeke, Belgium) during 2 weeks. Flow cytometry was carried out after staining with an anti-CD8 antibody (BD Biosciences, Erembodegem, Belgium) and HLA-A2 tetramers containing Mart-126–35, Gp100280–288, Gp100209–127, tyrosinase369–377, or the control antigen influenza virus58–66. A minimum of 200.000 events were recorded using a FACS Canto II flow cytometer (BD Biosciences). Analysis was performed with flowjo 7.6.2 software (Treestar Inc., San Carlos, CA, USA).

Statistical methods

Results were primarily assessed by measuring the differences in pigmentation and number of stained cells/HPF in each test region comparing the different treatments, induction methods, and time points. Three HPFs (×250) per test region were counted. The data were analyzed using Mann–Whitney U analysis and Kruskal–Wallis test. Reliability and reproducibility were assessed by partial correlation analysis. All statistical analyses were performed using spss 17.0 software (SPSS Inc, Chicago, IL, USA).

Acknowledgements

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References

The authors thank Peter Batsleer, Department of Dermatology, Ghent University Hospital, Ghent, Belgium, for providing technical assistance with digital image analysis. This research was supported by a research grant to R. Speeckaert from the Research Foundation (No. BOF10/doc/403), Ghent University, and to Nanja van Geel from the Scientific Research Foundation-Flanders (FWO Senior Clinical Investigator).

References

  1. Top of page
  2. Summary
  3. Significance
  4. Introduction
  5. Results
  6. Discussion
  7. Methods
  8. Acknowledgements
  9. References