The Journal of Allergy and Clinical Immunology
Volume 119, Issue 6 , Pages 1470-1480, June 2007

Chemokine responses distinguish chemical-induced allergic from irritant skin inflammation: Memory T cells make the difference

  • Stephan Meller, MD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • These authors contributed equally to this work.
    • ,
  • Antti I. Lauerma, MD, PhD

      Affiliations

    • Control of Hypersensitivity Diseases, Finnish Institute of Occupational Health, Helsinki
    • These authors contributed equally to this work.
    • ,
  • Frank Michael Kopp, PhD

      Affiliations

    • Institute of Environmental Research, Heinrich-Heine University, Düsseldorf
    • These authors contributed equally to this work.
    • ,
  • Franziska Winterberg, PhD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • These authors contributed equally to this work.
    • ,
  • Minna Anthoni, PhD

      Affiliations

    • Control of Hypersensitivity Diseases, Finnish Institute of Occupational Health, Helsinki
    • ,
  • Anja Müller, MD

      Affiliations

    • Department of Radiation Oncology, Heinrich-Heine University, Düsseldorf
    • ,
  • Michael Gombert, PhD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • ,
  • Anna Haahtela, PhD

      Affiliations

    • Skin and Allergy Hospital, Helsinki University Central Hospital
    • ,
  • Harri Alenius, PhD

      Affiliations

    • Centre of Excellence in Immunotoxicology, Finnish Institute of Occupational Health, Helsinki
    • ,
  • Juliane Rieker, MD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • ,
  • Marie-Caroline Dieu-Nosjean, PhD

      Affiliations

    • Institut National de la Santé et de la Recherche Médicale Unité 255, Laboratoire d'Immunologie Cellulaire et Clinique, Centre de Recherches Biomédicales des Cordeliers, Paris
    • ,
  • Robert Christof Kubitza

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • ,
  • Ernst Gleichmann, MD

      Affiliations

    • Institute of Environmental Research, Heinrich-Heine University, Düsseldorf
    • ,
  • Thomas Ruzicka, MD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • ,
  • Albert Zlotnik, PhD

      Affiliations

    • Neurocrine Bioscience, San Diego
    • ,
  • Bernhard Homey, MD

      Affiliations

    • From the Department of Dermatology, Heinrich-Heine-University, Düsseldorf
    • Corresponding Author InformationReprint requests: Bernhard Homey, MD, Department of Dermatology, Heinrich-Heine-University, Moorenstr. 5, D-40225 Düsseldorf, Germany.

Received 9 August 2006; received in revised form 22 November 2006; accepted 12 December 2006. published online 10 March 2007.

Düsseldorf, Germany, Helsinki, Finland, Paris, France, and San Diego, Calif

Article Outline

Background

As clinical and histological features of allergic and irritant contact dermatitis share common characteristics, the differentiation between them in the preclinical and clinical evaluations of chemicals remains difficult.

Objective

To identify the differences in the underlying immunological mechanisms of chemical-induced allergic or irritant skin responses.

Methods

We systematically studied the involvement of chemokines in both diseases by quantitative real-time polymerase chain reaction in mice and humans. The cellular origin of relevant chemokines and receptors was determined using immunohistochemistry; functional relevance was demonstrated in vitro by transwell chemotaxis and in vivo by adoptive transfer experiments using a model of hapten-induced murine contact hypersensitivity.

Results

Independent of overall skin inflammation, chemical-induced allergic and irritant skin responses showed distinct molecular expression profiles. In particular, chemokine genes predominantly regulated by T-cell effector cytokines demonstrated differential upregulation in hapten-specific skin inflammation. Notably, the expression of CXCR3 ligands, such as CXCL9 (Mig) and CXCL10 (IP-10), was upregulated in chemical-induced allergic skin responses when compared with irritant skin responses. Furthermore, we showed that inflammatory chemokines such as CXCL10 prime leukocytes to respond to CXCL12 (SDF-1), increasing their recruitment both in vitro and in vivo.

Conclusion

We provide important insights into the molecular basis of chemical-induced allergic and irritant contact dermatitis, identify novel markers suitable for their differentiation, and demonstrate the cooperation of inflammatory and homeostatic chemokines in the recruitment of pathogenic leukocyte subsets.

Clinical implications

Molecular differences between both diseases represent the basis for new approaches to diagnostics and therapy.

Key words: Allergy, chemokines, chemotaxis, inflammation, cell trafficking, irritancy, memory T cells, skin, TH1/TH2 cells

Abbreviations used: DC, Dendritic cell, DNFB, Dinitrofluorobenzene, DNP, Dinitrophenyl, MCP, Monocyte chemotactic protein, MIP, Macrophage inflammatory protein, NiSO4, Nickel sulphate, SLS, Sodium lauryl sulfate

 

Exposure to chemicals through skin or mucosa causes an increasing number of allergic and irritant responses, leading to significant socioeconomical problems.1 Although irritant skin inflammation ideally represents a decrescendo reaction and allergic skin inflammation a crescendo reaction, the clinical features of both are comparable.2, 3 Nevertheless, the underlying pathophysiological mechanisms are thought to be substantially different.4 In contrast to direct tissue damage induced by chemical irritants, which subsequently leads to local inflammation, the induction of allergic responses by chemical allergens is characterized by induction of antigen-specific effector and memory T cells.5

The recruitment of distinct leukocyte subsets to sites of inflammation has been shown to be critically regulated by chemokines and their receptors.6, 7 Chemokines have recently attracted considerable attention because this superfamily of chemoattractive proteins is thought to be among the first complete protein superfamilies characterized at the molecular level.8, 9 In the current study, we sought to identify all relevant members of this protein superfamily involved in chemical-induced allergic or irritant skin inflammation, obtain new insights into their underlying pathomechanisms, and provide perspectives for diagnosis, prevention, and treatment of chemical-induced skin inflammation.

Back to Article Outline

Methods 

Patients 

For the induction of chemical-induced allergic skin inflammation, nickel sulphate (NiSO4) at 5% in petrolatum (Epikon Ltd, Helsinki, Finland) was applied in small Finn Chambers (Epitest Ltd, Tuusula, Finland) on nonlesional back skin of nickel-sensitized patients (n = 8). For the induction of chemical-induced irritant skin inflammation, the chemical irritant sodium lauryl sulphate (SLS) (Merck KGaA, Darmstadt, Germany) at 1% in water was applied in large Finn Chambers on nonlesional back skin of nickel-sensitized volunteers (n = 8). Subsequently, 6-mm punch biopsies were obtained 2, 6, and 48 hours after chemical exposure. Additionally, a control biopsy was obtained from untreated nonlesional skin of the back. The samples were snap-frozen in liquid nitrogen and stored at −70°C. The clinical score of patch test lesions was evaluated according to recommendations of the International Contact Dermatitis Research Group.10 The degree of inflammation was histopathologically analyzed and scored by 2 independent investigators (see Table E1 in this article's Online Repository at www.jacionline.org). These studies were approved by the local ethics committee (Helsinki-Uusimaa Hospital District Ethics Committee).

Quantitative real-time polymerase chain reaction (PCR) analysis 

Skin biopsies were homogenized in liquid nitrogen using a Mikro-Dismembrator U (Braun Biotech, Lancaster, Pa), ribonucleic acid (RNA) was extracted, and subsequently, complementary deoxyribonucleic acid (cDNA) was subjected to quantitative real-time PCR analyses (TaqMan; Roche Molecular Systems, Inc., Pleasanton, Calif) of chemokine expression as described previously.11

Mouse models 

For the induction of chemical-induced allergic skin inflammation, 8- to 12-week-old female BALB/c mice were topically treated with 100 μL 0.5% dinitrofluorobenzene (DNFB) on their shaved abdomen. Six days later, hapten-specific skin inflammation was elicited after painting of both ears with a non–irritant dose of 0.2% DNFB. For chemical-induced irritant skin inflammation, mice were topically treated with the standard irritant croton oil (2%) or an irritant dose of the hapten DNFB (0.5%) on both ears. Vehicle-treated (DAE 433, dimethylacetamid 40%, acetone 30%, ethanol 30%) animals served as controls. Ear-swelling responses were monitored 6, 12, 24, and 72 hours after chemical exposure with a spring-loaded micrometer (Oditest; Kroeplin Laengenmesstechnik, Schluechtern, Germany). Subsequently, mice were sacrificed and epidermal sheets of the ears as well as draining auricular lymph nodes were obtained for further analyses as previously described.7 As ear-swelling responses within the groups showed homogeneous distribution and small standard deviations (Fig 1, A), epidermal sheets (10/group) and local draining lymph nodes (10/group) were pooled and subjected to RNA extraction and subsequent quantitative real-time PCR analyses.

  • View full-size image.
  • Fig 1. 

    CXCL9 and CXCL10 are selectively upregulated in hapten-specific but not in irritant contact dermatitis. A, Ear-swelling responses (mm × 10−2) after epicutaneous application of either vehicle, the standard irritant croton oil or the hapten DNFB. Quantitative real-time PCR analysis of CXCL9 (B), CXCL10 (C), and CCL20 (D) at indicated time points in epidermal sheets. E, Ear-swelling responses (mm × 10−2) of nickel-sensitized mice 48 hours after intradermal challenge with NiCl2 or saline. Quantitative real-time PCR analysis of CXCL9 (F) and CXCL10 (G) in epidermal sheets of nickel-specific allergic contact dermatitis and controls. Values are expressed as femtograms of target gene in 12.5 ng of total cDNA.

For a mouse model of nickel-induced contact hypersensitivity, a modification of the mouse ear-swelling test was used, as described by van Hoogstraten et al.12 For sensitization, C57BL/6J mice received intradermal injections consisting of either 50 μL 10 mmol/L NiCl2 in 0.9% saline or a mixture of 10 mmol/L NiCl2/1% H2O2 in saline (H2O2 serving as adjuvant), into both flanks.13 Ten days later, mice were challenged for recall by injecting either 10 mmol/L NiCl2 in 50 μL of saline or saline alone into the pinna of each ear. Forty-eight hours after challenge, delayed-type hypersensitivity reactions were determined by measuring the increment in ear thickness compared with the prechallenge value; measurements were performed with a micrometer (S5010 gauge; Kroeplin Laengenmesstechnik). Data shown represent the mean ear-swelling response (10−2 mm + standard error of the mean) of groups comprised of 5 mice. Studies were approved by the local committees on animal welfare.

Adoptive transfer experiments 

BALB/c mice were treated with 0.5% DNFB on the dorsal and ventral surfaces of both ears. Six to 10 days later, draining auricular lymph nodes and spleens were collected and pooled, and single-cell suspensions were generated. Cells were stimulated with either 1000-ng/mL CXCL10 (R&D Systems, Abingdon, United Kingdom [UK]) in phosphate-buffered solution (PBS) for 20 minutes or left untreated (PBS only). Subsequently, 5 × 107 to 1 × 108 resting or CXCL10-stimulated cells were injected intravenously into the tail vein of naïve recipient mice (5-10 mice per group, n = 41), which had been treated with .25% DNFB on both ears 1 hour before injection. Ear swelling was monitored before and 24, 48, or 72 hours after challenge with relevant hapten as described above.

Histology and immunohistochemistry 

For immunohistochemical analyses of chemokine expression in NiSO4 or SLS patch test lesions, skin sections were fixed with 3% acetone and preprocessed with H2O2 followed by avidin and biotin blocking (VECTOR Blocking Kit; Vector Laboratories, Burlingame, Calif). Sections were stained with antibodies against CXCL9/MIG, CXCL10/IP-10 (both goat IgG; R&D Systems, Minneapolis, Minn) or CXCL12/SDF-1α (K15C; IgG2a; Laboratoire de Pathogénie Virale Moléculaire, Institute Pasteur, Paris, France) at 4°C. The staining was detected with an ABC-Kit (VECTOR Vectastain ABC-Kit, Vector Laboratories) and an AEC-Kit (Vector Laboratories). Sections were counterstained with hematoxylin.

Chemotaxis assays 

Transwell chemotaxis assays were performed as previously described.14 T cells were purified from human blood samples and subsequently incubated with the indicated concentrations of recombinant hCXCL10 and hCXCL12 (R&D Systems) in the bottom chamber for 3 hours in a 5% CO2 environment at 37°C. The number of migrated cells was determined by flow cytometry using monoclonal antibody specific for human CD8, CD4, and CLA, respectively (BD PharMingen, San Diego, Calif). To determine the absolute number of migrated cells, a known number of 15 μm microsphere beads (Bangs Laboratories, Inc., Fishers, Ind) was added to each sample before analysis. The relation of cells to beads of the starting population is defined as 100% migration. Based on that, we calculated the percentage of migration in each sample. For each set of experimental conditions, at least 3 separate experiments were performed.

Cell culture of structural cells and dendritic cell subsets 

Langerhans-type and interstitial-type dendritic cells (DCs) were generated from isolated human monocytes and, subsequently, stimulated with either lipopolysaccharide (25 ng/mL; Sigma Aldrich GmbH, Steinheim, Germany) or CD40L for 6 or 24 hours as described previously.15

Human primary epidermal keratinocytes, dermal fibroblasts, and dermal microvascular endothelial cells were cultured in keratinocyte (KGM-2), fibroblast (FGM-2), or endothelial cell (EGM-2) growth medium (Clonetics; Cambrex Corporation, East Rutherford, NJ), as described previously.11 Cells were treated with either TNF-α plus IL-1β for 18 hours or left untreated.

Human allergen-specific T-cell clones 

For generation of antigen-specific T-cell lines, T cells were purified from peripheral blood mononuclear cells of dinitrochlorobenzene-sensitized volunteers and cultured at a density of 1.5 × 106 cells/well in complete RPMI 1640 medium containing 10 μg/mL dinitrophenyl (DNP; Merck & Co, Inc, Whitehouse Station, NJ) antigen. On day 6, human rIL-2 (Sigma Corporation, St Louis, Mo) was added at 50 IU/mL and on day 9 optimized to 25 IU/mL. The established T-cell lines were expanded, and T-cell lines with highest stimulation indices to DNP were cloned by the limiting dilution method. When clones were established, the most vigorous clones were stimulated with DNP. A total of 80 clones were found to be DNP-specific and subsequently expanded in culture. For proliferation tests, the clone cells were stimulated with 10 μg/mL DNP, as described above, or with 200 μg/mL phytohemagglutinin (PHA; Sigma Corporation). Cells obtained from proliferation tests were homogenized in Trizol (Gibco BRL, Paisley, UK), RNA was extracted, and quantitative real-time PCR analysis was performed as described above.

Statistical analysis 

To calculate statistical significance, the F test was used to analyze variances, and subsequently, the Student t test was performed using SPSS software (version 11.0, 2000; SPSS, Chicago, Ill). Differences with P values less than .05 were considered as statistically significant. To identify differentially expressed genes, the results of quantitative real-time PCR analyses of skin specimens of NiSO4 patch test patients were compared with data obtained from SLS patch test patients using centroid analysis. Shrunken centroid analysis was performed as previously described.16

Back to Article Outline

Results 

Chemokine responses in murine models for chemical-induced allergic and irritant contact dermatitis 

To investigate the molecular differences between chemical-induced allergic and irritant skin inflammation, mice were either sensitized and challenged with the standard hapten DNFB or exposed to the standard chemical irritant croton oil or an irritant dose of DNFB. After chemical exposure, ear swelling as well as molecular responses were determined within the skin and draining lymph nodes (Fig 1, A-D).

Although maximum ear-swelling responses in chemical-induced irritant skin inflammation did not significantly differ from those observed in allergic skin inflammation (Fig 1, A), profound differences in chemokine gene expression were observed at the molecular level (Fig 1, B-D). The CXCR3 ligands CXCL9 and CXCL10, which were absent in irritant-induced skin inflammation, were selectively induced during hapten-specific immune responses. Peak expression occurred 24 hours after hapten exposure (Fig 1, B and C). In contrast to haptens, chemical irritants rapidly induced another set of human allergen-specific T-cell clones CCL1 (I-309), CCL2 (monocyte chemotactic protein [MCP]-1), CCL3 (macrophage inflammatory protein [MIP]-1α), CCL4 (MIP-1β, CCL7/MCP-3, CCL10/MIP-1γ), CCL11 (Eotaxin), CCL17 (thymus and activation-regulated chemokine [TARC]), CCL19 (MIP-3β), and CCL22 (macrophage-derived chemokine [MDC], data not shown) including CCL20 (MIP-3α) (Fig 1, D). The level of gene expression directly correlated with the irritant-induced ear-swelling response (Fig 1, A and D).

Surprisingly, no significant differences in chemokine gene expression were observed in draining lymph nodes during hapten-specific or irritant skin inflammation (data not shown). To validate whether results obtained during contact hypersensitivity responses to DNFB are also relevant for other haptens as well, chemokine gene expression was also analyzed in a mouse model for nickel contact allergy. Similarly to DNFB-induced skin inflammation, CXCL9 and CXCL10 were markedly induced in the ears after exposure to NiCl2, provided mice were previously sensitized with NiCl2/H2O2 (Fig 1, E-G).

Chemokine responses in human models for chemical-induced allergic and irritant contact dermatitis 

Quantitative real-time PCR analyses of human skin specimens of patients suffering from nickel allergy obtained before as well as after either NiSO4 or SLS patch testing demonstrated that the CXCR3 ligands CXCL9, CXCL10, and CXCL11 were time-dependently induced in 7 out of 8 nickel-sensitized patients upon NiSO4-exposure (Fig 2, A-C). Peak expression was observed 48 hours after hapten exposure (Fig 2, A-C). In contrast, epicutaneous treatment with the chemical irritant SLS failed to induce CXCR3 ligands within the skin of nickel-sensitized patients (Fig 2, A-C). Notably, no significant differences in the clinical and histological scores between NiSO4 patch–tested and SLS patch–tested collectives were observed (see Table E1 in this article's Online Repository at www.jacionline.org), indicating that differences in gene expression were a result of different molecular pathomechanisms rather than differences in the degree of skin inflammation.

  • View full-size image.
  • Fig 2. 

    Quantitative real-time PCR analysis of CXCL9 (A), CXCL10 (B), CXCL11 (C), and CCL20 (D) expression in skin specimens obtained from NiSO4 and SLS patch tests of nickel-sensitized patients (n = 8) at 0, 2, 6, and 48 hours after chemical exposure, respectively. Values are expressed as femtograms of target gene in 25 ng of total cDNA.

In contrast to CXCR3 ligands, the inflammatory chemokine CCL20 did not discriminate between allergic or irritant contact dermatitis. It was induced in both NiSO4 and SLS patch test lesions and followed the course of overall skin inflammation (Fig 2, D). During irritant skin inflammation, CCL20 was induced at earlier time points and in some patients showed peak expression 6 hours after irritant exposure. In contrast, nickel-sensitized patients uniformly demonstrated a maximum of CCL20 expression 48 hours after relevant hapten treatment (Fig 2, D).

Shrunken centroid analyses of chemokine expression in patch test lesions 48 hours after either NiSO4 or SLS exposure demonstrated that IFN-γ-regulated, TH1-associated CXCR3 ligands are among the 10 most differentially expressed chemokine genes discriminating between chemical-induced allergic and irritant skin inflammation (Fig 3). In particular, CXCL10 was the most differentially expressed gene in the 2 collectives. Moreover, the TH2-associated chemokines CCL11, CCL17, and CCL22 were markedly induced in allergic contact dermatitis and ranked among the 10 most differentially regulated chemokines as well. In addition, CCL7, CCL8, and CCL13, all belonging to the MCP subfamily of chemokines, showed increased RNA levels in allergic contact dermatitis and are among the 5 most differentially regulated chemokines (Fig 3). To date, no mouse analog could be identified for CCL13; therefore, this gene did not appear in our comprehensive mouse analyses.9 Furthermore, centroid analyses identified CCL2 as a chemokine associated with irritant but not with allergic skin inflammation (Fig 3). Similarly to our previous observations obtained from the mouse model of allergic and irritant contact dermatitis, CCL1, CCL20, and CCL21 were uniformly expressed in both NiSO4-induced and SLS-induced skin lesions (Fig 3).

  • View full-size image.
  • Fig 3. 

    Shrunken centroids ranking chemokine expression in human skin specimens of NiSO4 or SLS patch tests of Ni-sensitized patients (n = 8). Horizontal values represent log ratios of gene expression.

Subsequently, we compared mouse and human models for hapten-specific skin inflammation at the molecular level. Therefore, gene expression profiles of a set of homeostatic, inflammatory, TH2- and TH1-associated chemokines in NiSO4 patch test lesions, NiCl2-specific skin inflammation, and DNFB-specific skin inflammation were compared by quantitative real-time PCR (see Fig E1, A-C in this article's Online Repository at www.jacionline.org). Overall, similar gene expression profiles were induced in both mouse and human models for hapten-specific skin inflammation, indicating that murine hapten-induced contact hypersensitivity represents a valid model for chemical-induced allergic contact dermatitis in man (see Fig E1, A-C in this article's Online Repository at www.jacionline.org). Comparing the expression of TH1- and TH2-associated chemokines in hapten-specific skin inflammation directed against different haptens, DNFB-specific immune responses demonstrated a more pronounced shift to a type 1 response and a markedly elevated inflammatory chemokine profile (see Fig E1, C in this article's Online Repository at www.jacionline.org), in comparison with nickel-induced skin inflammation (see Fig E1, A and B in this article's Online Repository at www.jacionline.org), demonstrating that the outcome of hapten-specific responses is closely related to the nature of the hapten.

Immunohistochemical analyses in NiSO4 and SLS patch test lesions 

Our next set of experiments aimed at the identification of the cellular origin of CXCL9 and CXCL10 protein expression in chemical-induced allergic and irritant skin inflammation. Immunohistochemical analyses of human skin specimens obtained before as well as 6 or 48 hours after either NiSO4 or SLS exposure demonstrated that CXCL9 and CXCL10 were distinctly expressed in allergic versus irritant skin inflammation (Fig 4). Within normal skin, CXCL9 showed constitutive expression in dermal cells possessing dendritic morphology (Fig 4, A and D). Notably, a marked induction of CXCL9 protein in dermal as well as in epidermal compartments was observed during the development of hapten-specific skin inflammation in patients sensitized to NiSO4 (Fig 4, B and C). In contrast, CXCL9 was not induced in chemical-induced irritant skin inflammation in SLS-exposed individuals (Fig 4, C-F).

Although CXCL10 protein was absent in untreated skin (Fig 4, G and I) or in skin specimens obtained 6 hours after chemical exposure (Fig 4, H and K), marked expression of this CXCR3 ligand was selectively induced in the skin of nickel-sensitized patients 48 hours after relevant hapten exposure (Fig 4, I), and this expression could predominantly be attributed to keratinocytes of basal and suprabasal layers of the epidermis (Fig 4, I). On the contrary, treatment of nickel-sensitized patients' (n = 8) skin with the chemical irritant SLS did not induce CXCL10 protein (Fig 4, J-L), confirming previous quantitative real-time PCR analyses (Fig 2). In contrast to the CXCR3 ligands CXCL9 and CXCL10, there were no differences observed in the expression of the homeostatic chemokine CXCL12 between skin specimens of NiSO4- and SLS-treated individuals (Fig 4, M-R).

Regulation of disease-associated chemokines in cellular constituents of the skin 

To obtain insights into the regulation of disease-associated chemokines, we stimulated structural cells of the skin, that is, primary epidermal keratinocytes, dermal fibroblasts, and dermal endothelial cells, with proinflammatory cytokines, represented by TNF-α/IL-1β or the T-cell-derived effector cytokines IFN-γ or IL-4. Quantitative real-time PCR analyses indicated that the CXCR3 ligands CXCL9, CXCL10, and CXCL11 were predominantly regulated by the type 1 cytokine IFN-γ, and keratinocytes as well as dermal endothelial cells were identified as the most abundant sources for these chemokines (Fig 5, A-C). In contrast, chemokines showing no discriminative expression between chemical-induced allergic or irritant skin inflammation, such as CCL20, were induced by primary proinflammatory cytokines, such as TNF-α and IL-1β (Fig 5, D).

  • View full-size image.
  • Fig 5. 

    Quantitative real-time PCR analysis of CXCL9, CXCL10, CXCL11 , and CCL20 expression in resting and activated structural cells of the skin (A-D: representative results from single donors) as well as Langerhans-type and interstitial-type dendritic cells (E-H; mean + SD of 2 independent donors). Quantitative real-time PCR analysis of CD25 (I) and IFN-γ (J) expression in DNP-specific human T-cell clones stimulated with PHA or with their relevant antigen DNP. Results are representative of 2 individual experiments with 2 different donors. Values are expressed as femtograms of target gene in 12.5 ng of total cDNA.

Next to structural cells of the skin, resident or skin-infiltrating DCs are an abundant source of chemokine production (Fig 4). Quantitiative real-time PCR analyses of chemokine messenger RNA expression in resting and activated Langerhans-type or interstitial-type DCs show that activated DCs can produce 10- to 20-fold more CXCL9, CXCL10, and CXCL11 when compared with activated structural cells of the skin (for example, keratinocytes and endothelial cells). Conversely, activated dermal endothelial cells, fibroblasts, and keratinocytes expressed approximately 10 to 30 times higher levels of CCL20 transcripts when compared with different dendritic cell subsets (Fig 5, A-H).

Considering the influence of cytokines on the induction of chemokine profiles, we investigated the effect of Ag-specific and nonspecific priming of T cells on their cytokine expression. Looking at the activation of hapten-specific T-cell clones, we observed that Ag-specific stimulation of DNP-specific lymphocytes resulted in a stronger induction of IFN-γ in comparison with potent nonspecific stimuli such as PHA. Differences in IFN-γ expression seemed to be independent of the overall activation status of the stimulated T cells, because CD25 was expressed at comparable levels after either PHA or hapten-specific stimulation of T-cell clones (Fig 5, I and J).

Cooperation of homeostatic and inflammatory chemokines during the recruitment of memory T cells into the skin 

Comprehensive analyses of chemokine expression in mouse and human skin samples during chemical-induced allergic and irritant skin responses demonstrated the presence of a complex temporo-spatial network of both homeostatic and inflammatory chemokines. In vitro, transwell chemotaxis experiments showed that CD4+ or CD8+ skin-homing CLA+ memory T cells showed modest but dose-dependent migration responses to CXCL10 (Fig 6, A and B). In the presence of a suboptimal dose of CXCL12 (10 ng/mL), CXCL10 gradients induced profoundly increased migratory responses of skin-homing memory T cells (Fig 6, A and B). This synergistic effect resulted in the recruitment of approximately 60% to 85% of CLA+ skin-homing memory T cells. In a series of at least 3 independent experiments, especially CD8+ memory T cells showed an enhanced responsiveness to combined gradients of CXCL10 and CXCL12 (Fig 6, B). In contrast, the combination of CXCL12 with another homeostatic chemokine, such as the skin-specific CC chemokine CCL27, did not result in a comparable synergistic response (Fig 6, C and D). Furthermore, pre-incubation of circulating T cells with CXCL10 primed skin-homing memory T cells responds to the subsequent presentation of CXCL12 gradients in synergistical fashion (Fig 6, E and F), comparable with cells receiving a treatment with both chemokines at the same time. However, pretreatment with CXCL12 vice versa did not enhance the migratory response of memory T cells toward CXCL10 gradients (Fig 6, E and F). Taken together, these results demonstrate that inflammatory chemokines, for example, CXCL10, enhance the responsiveness of skin-homing T cells to homeostatic chemokines, in our case CXCL12, and may amplify the recruitment of pathogenically relevant T-cell subpopulations to sites of allergic inflammation.

  • View full-size image.
  • Fig 6. 

    Migratory response toward CXCL10 of CD4+CLA+ T cells (A) or CD8+CLA+ T cells (B) in the presence or absence of a suboptimal dose of CXCL12 (10 ng/mL). Dose response of CD4+CLA+ T cells (C) or CD8+CLA+ T cells (D) to CCL27 in the presence or absence of a suboptimal dose of CXCL12 (10 ng/mL). Migratory response of CD4+CLA+ T cells (E) or CD8+CLA+ T cells (F) to a suboptimal dose of CXCL12 after priming with CXCL10. Results of triplicate experiments are shown. G, Ear-swelling responses after adoptive transfer of spleen and lymph node cells from DNFB-sensitized mice. Before adoptive transfer cells were exposed to 1 μg/mL CXCL10 or saline. Ear-swelling response was measured 24, 48, and 72 hours after adoptive transfer. For statistical analysis, a Student t test was performed. ∗P < .05; ∗∗∗P < .005.

CXCL10 “primes” leukocytes to extravasate at sites of hapten exposure 

To investigate the in vivo relevance of CXCL10 priming for leukocyte chemotaxis, lymph node, and spleen cells of sensitized mice were pooled and treated ex vivo with either CXCL10 or PBS alone and then adoptively transferred into naïve recipient mice. One hour before cell transfer, recipients were challenged with the relevant hapten DNFB on both ears, and ear swelling was monitored 24, 48, and 72 hours later. In contrast to mice receiving PBS-treated leukocytes, mice receiving CXCL10-primed lymph node and spleen cells showed significantly (P < .05) increased skin inflammation 24, 48, and 72 hours after hapten challenge and cell transfer (Fig 6, G). These findings suggest that chemokine-induced “priming” of relevant leukocyte subpopulations contributes to Ag/hapten-specific skin inflammation in vivo.

Back to Article Outline

Discussion 

Members of the chemokine superfamily play a pivotal role in regulating the migration of leukocytes in vitro and in controlling leukocyte trafficking in vivo.17, 18 In this study, we provide the first comprehensive analysis of chemokines and their receptors in allergic and irritant contact dermatitis in mice and humans. Taking a global view, we identified a distinct set of chemokines that distinguishes allergic from irritant contact dermatitis.

Among the 5 most differentially regulated genes, we identified chemokines induced by products of effector memory T cells displaying either a TH1- or a TH2-phenotype. In contrast, genes that do not show differential expression between allergic and irritant contact dermatitis are mainly homeostatic chemokines or inflammatory chemokines regulated by primary proinflammatory cytokines, that is, TNF-α and IL-1β.

Notably, CXCL10 ranked at the top of differentially regulated chemokines in allergen-induced contact dermatitis. Moreover, CXCL9 was among the 5 and CXCL11 among the 10 most highly regulated chemokines in allergic skin inflammation. All 3 of them, CXCL9, CXCL10, and CXCL11, bind to their shared receptor CXCR3, are predominantly regulated by the type 1 effector cytokine IFN-γ, and are therefore TH1-associated.19, 20 In good agreement with quantitative real-time PCR analyses, immunohistochemical analyses depicted strong expression of CXCL9 and CXCL10 protein in hapten-specific but not in chemical-induced irritant skin inflammation of nickel-sensitized patients. In contrast to previous studies, providing only limited data based on immunohistochemistry or in situ hybridization,21, 22 the current study demonstrates the sequential and spatial expression of a distinct set of chemokines in irritant or allergic contact dermatitis in humans. Corresponding results were observed in murine models of chemical-induced allergic inflammation showing differential expression of CXCL9 and CXCL10. Thus, our results demonstrate that CXCR3 ligands are selectively induced during hapten-specific but not irritant-induced skin inflammation. This expression profile is likely to be attributed to the presence of activated type 1 memory T cells. The important role of CXCR3+ memory T cells in the pathogenesis of hapten-specific skin inflammation was documented by findings of Moed et al,23 demonstrating that CXCR3+ CD4+CLA+ memory T cells were essential for nickel-specific T-cell proliferation and production of cytokines in vitro. Depletion of this T-cell subset resulted in a significant decrease of hapten-specific responses. Moreover, the T-cell–derived effector cytokine IFN-γ, which is a signature cytokine of type 1 responses, regulates the expression of top-ranked CXCR3 ligands, that is, CXCL9, CXCL10, and CXCL11, both in vitro and in vivo.24, 25, 26 The pivotal role of IFN-γ in TH1-induced delayed-type hypersensitivity, such as allergic contact dermatitis, is supported by findings of Wakabayashi et al,27 showing that IFN-γ knockout mice failed to elicit contact hypersensitivity responses. Beyond that, experiments with CXCL10−/− mice, described by Dufour et al,28 demonstrated the relevance of inflammatory chemokines in hapten-induced allergic skin reactions. In the current study, we showed that upon stimulation with IFN-γ, keratinocytes as well as dermal endothelial cells produce large amounts of CXCL9 and CXCL10 transcripts. Taken together, we conclude that hapten-specific stimulation of skin-infiltrating CXCR3+ memory T cells in vivo may lead to the production and release of large amounts of the effector cytokines, including IFN-γ, perpetuating and enhancing the production of chemokines by surrounding resident structural cells and DCs, leading to the attraction of increasing numbers of effector cells.

In addition to chemokines predominantly regulated by TH1 cytokines, our centroid analyses also showed TH2-associated chemokines, such as CCL11, CCL17, and CCL22, among differentially regulated genes in hapten-specific skin inflammation. Similar to CCL11, the expression of CCL17 and CCL22 is regulated by type 2 effector cytokines such as IL-4. CCL11 was shown to be involved in the recruitment of eosinophils, a leukocyte subset frequently observed in allergic contact dermatitis; CCL17 and CCL22 preferentially attract CCR4-bearing TH2 cells.29, 30, 31 These observations suggest the existence of a TH2-associated amplification loop, analogous to the TH1-associated described above, through the production of CCL17 and CCL22 by hapten/allergen-loaded Ag-presenting cells, the subsequent recruitment of hapten-specific TH2 cells releasing IL-4 and in turn enhancing TH2-associated chemokine production.32, 33

Murine models of chemical-induced allergic or irritant skin inflammation demonstrated corresponding results to the human model. Moreover, comparison of DNFB-induced skin inflammation with nickel-allergy in mice or humans showed that in addition to a strong TH1-mediated component indicated by the induction of IFN-γ–inducible chemokines, such as CXCL9 and CXCL10, nickel-induced allergic skin inflammation had also considerable TH2 aspects, indicated by the upregulation of IL-4-inducible chemokines, including CCL11, CCL17, and CCL22, supporting the presence of both activated TH1 as well as TH2 cells at sites of allergic skin inflammation. These findings are consistent with the expression of IL-4 demonstrated in nickel-specific T-cell clones derived from the skin of patients with nickel allergies.34 Although this study focused on few standard irritants and allergens, the basic congruence at the molecular level between human and murine models for chemical-induced allergic or irritant skin inflammation indicates that these mouse models are valid experimental systems for chemical-induced skin inflammation in humans.

In addition to differentially regulated TH1-associated and TH2-associated chemokines, we identified inflammatory, nondiscriminative chemokines such as CCL20 that do not show significant differences between allergic or irritant contact dermatitis. In vivo, the induction of CCL20 is dependent on the overall inflammatory response in allergic as well as in irritant contact dermatitis. In vitro, this inflammatory chemokine is predominantly regulated by primary proinflammatory cytokines such as TNF-α and IL-1β, and attracts CCR6+ T cells and immature DCs into sites of skin inflammation.35, 36 The important role of primary proinflammatory cytokines is reflected by abrogation of inflammatory skin reactions such as allergic and irritant contact dermatitis by anti-TNF-α antibodies.37 Interestingly, although levels of nondiscriminative chemokines, especially CCL20, increased rapidly (2-12 hours) after challenge in both human and murine models, discriminative and allergy-associated chemokines, including CXCL9, CXCL10, and CXCL11, achieved their maximum at later time points (24-48 hours), reflecting qualitative as well as time-dependent differences in the orchestration of chemokine production during irritant and allergic contact dermatitis.

Hence, IL-1–driven and TNF-α–driven innate immune responses play a role during early phases of both allergic and irritant contact dermatitis. During later phases of irritant responses, skin inflammation still is critically dependent on innate responses; however, during allergic contact dermatitis, adaptive immune responses take over and amplify skin inflammation.

Next to inflammatory chemokines, we identified homeostatic chemokines such as CXCL14 and CXCL12 as nondiscriminative genes. We show that the CXCR4 ligand CXCL12 is preferentially expressed by endothelial cells as well as by epidermal and dermal cells with dendritic morphology in either type of contact dermatitis. Interestingly, in hapten-specific skin inflammation, the TH1-associated CXCR3 ligands CXCL9 and CXCL10 are coexpressed at similar anatomical locations with the homeostatic chemokine CXCL12. Recently, Vanbervliet et al38 demonstrated that the migration of plasmacytoid DCs is regulated by the synergistic action of inducible chemokines binding to CXCR3 and homoeostatic chemokines such as CXCL12. Hence, it is conceivable that plasmacytoid DCs may accumulate in allergic contact dermatitis but not or to a lesser degree in irritant contact dermatitis. In fact, Bangert et al39 recently observed increasing numbers of plasmacytoid DCs in allergic contact dermatitis versus normal skin. These data are in line with a previous study by Wollenberg et al40 demonstrating the accumulation of plasmacytoid DCs in contact dermatitis.

In the current study, we extend the findings of Vanbervliet et al showing that the homeostatic chemokine CXCL12 and the inflammatory CXCR3 ligands CXCL9 and CXCL10 cooperate synergistically in the recruitment of both CD4+CLA+ and CD8+CLA+ skin-homing memory T cells. In this regard, CD8+CLA+ memory T cells showed enhanced responsiveness after stimulation with CXCL10 and CXCL12 compared with CD4+CLA+ T cells. In vivo, CXCR3 and CXCR4 ligands are perfectly positioned to collaborate in allergic skin inflammation, as they are expressed in close proximity to each other in the dermis and epidermis of lesional skin. The constitutive expression of CXCL12 within the skin may facilitate the homeostatic recruitment of low numbers of lymphocytes. Upon hapten exposure, inflammatory chemokines generated within the tissue may cooperate with constitutive chemokines by exerting a synergistic effect. This cooperation may reduce the threshold of sensitivity for homeostatic chemokines such as CXCL12 and may have its physiological relevance in the amplification of lymphocyte migration, further boosting the development of skin inflammation. In fact, we have uncovered that T cells primed with the inflammatory chemokine CXCL10 show enhanced migration to homeostatic chemokines in vitro. Remarkably, this effect was independent of simultaneous presence of both homeostatic and inflammatory chemokines. Furthermore, T cells primed with CXCL10 not only exert such an effect in vitro but also cause stronger inflammatory reactions in mice, in vivo, as shown in our adoptive transfer experiments. The molecular mechanism of chemokine cooperation remains elusive. Conceivably, the synergy between chemokines may be caused by heterodimerization of receptors.41 Furthermore, intracellular cooperation in signal transduction by receptors for inflammatory and homeostatic chemokines may reduce the threshold of sensitivity for homeostatic chemokines.

Only a small fraction of memory T cells enter the healthy skin during steady-state conditions. After allergen exposure, antigen-specific memory T cells are activated and produce effector cytokines. However, expression levels of T-cell–derived effector cytokines within peripheral tissues such as skin are very low. Hence, detection of relevant amounts is difficult. Our data indicate that effector cytokines released by a small number of activated hapten-specific memory T cells stimulate gene expression of a large number of surrounding resident cells, leading to the production of a discriminative chemokine signature. In contrast, the absence of antigen-specific T-cell activation in irritant skin responses results in only negligible amounts of T-cell–derived effector cytokines.

Taken together, findings of the current study contribute to the understanding of the molecular pathomechanisms involved in chemical-induced allergic and irritant contact dermatitis and identify candidates to discriminate allergic from irritant skin reactions.

Back to Article Outline

 

We thank Dr Ali Amara, PhD, Laboratoire de Pathogénie Virale Moléculaire, Institute Pasteur, France, for providing antibodies against CXCL12/SDF-1α.

Back to Article Outline

Appendix. Supplementary data 

Download file

download text

Online Repository.

Back to Article Outline

References 

  1. Chew AL, Maibach HI. Occupational issues of irritant contact dermatitis. Int Arch Occup Environ Health. 2003;76:339–346
  2. Levin CY, Maibach HI. Irritant contact dermatitis: is there an immunologic component?. Int Immunopharmacol. 2002;2:183–189
  3. Chew AL. Biologic similarities of irritation and allergic contact dermatitis. In:  Chew AL,  Maibach HI editor. Irritant dermatitis. Berlin: Springer; 2006;
  4. Corsini E, Galli CL. Epidermal cytokines in experimental contact dermatitis. Toxicology. 2000;142:203–211
  5. Saint-Mezard P, Rosieres A, Krasteva M, Berard F, Dubois B, Kaiserlian D, et al. Allergic contact dermatitis. Eur J Dermatol. 2004;14:284–295
  6. Muller A, Homey B, Soto H, Ge N, Catron D, Buchanan ME, et al. Involvement of chemokine receptors in breast cancer metastasis. Nature. 2001;410:50–56
  7. Homey B, Alenius H, Muller A, Soto H, Bowman EP, Yuan W, et al. CCL27-CCR10 interactions regulate T cell-mediated skin inflammation. Nat Med. 2002;8:157–165
  8. Baggiolini M. Chemokines and leukocyte traffic. Nature. 1998;392:565–568
  9. Zlotnik A, Yoshie O. Chemokines: a new classification system and their role in immunity. Immunity. 2000;12:121–127
  10. Fregert S. Manual of contact dermatitis. 2nd ed.. Copenhagen, Denmark: Munksgaard; 1981;
  11. Homey B, Wang W, Soto H, Buchanan ME, Wiesenborn A, Catron D, et al. Cutting edge: the orphan chemokine receptor G protein-coupled receptor-2 (GPR-2, CCR10) binds the skin-associated chemokine CCL27 (CTACK/ALP/ILC). J Immunol. 2000;164:3465–3470
  12. Van Hoogstraten IM, Boos C, Boden D, Von Blomberg ME, Scheper RJ, Kraal G. Oral induction of tolerance to nickel sensitization in mice. J Invest Dermatol. 1993;101:26–31
  13. Artik S, von Vultee C, Gleichmann E, Schwarz T, Griem P. Nickel allergy in mice: enhanced sensitization capacity of nickel at higher oxidation states. J Immunol. 1999;163:1143–1152
  14. Morales J, Homey B, Vicari AP, Hudak S, Oldham E, Hedrick J, et al. CTACK, a skin-associated chemokine that preferentially attracts skin-homing memory T cells. Proc Natl Acad Sci U S A. 1999;96:14470–14475
  15. Gombert M, Dieu-Nosjean MC, Winterberg F, Bunemann E, Kubitza RC, Da Cunha L, et al. CCL1-CCR8 interactions: an axis mediating the recruitment of T cells and Langerhans-type dendritic cells to sites of atopic skin inflammation. J Immunol. 2005;174:5082–5091
  16. Tibshirani R, Hastie T, Narasimhan B, Chu G. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proc Natl Acad Sci U S A. 2002;99:6567–6572
  17. Zlotnik A, Morales J, Hedrick JA. Recent advances in chemokines and chemokine receptors. Crit Rev Immunol. 1999;19:1–47
  18. Homey B, Zlotnik A. Chemokines in allergy. Curr Opin Immunol. 1999;11:626–634
  19. Kim CH, Rott L, Kunkel EJ, Genovese MC, Andrew DP, Wu L, et al. Rules of chemokine receptor association with T cell polarization in vivo. J Clin Invest. 2001;108:1331–1339
  20. Sallusto F, Lenig D, Mackay CR, Lanzavecchia A. Flexible programs of chemokine receptor expression on human polarized T helper 1 and 2 lymphocytes. J Exp Med. 1998;187:875–883
  21. Flier J, Boorsma DM, Bruynzeel DP, Van Beek PJ, Stoof TJ, Scheper RJ, et al. The CXCR3 activating chemokines IP-10, Mig, and IP-9 are expressed in allergic but not in irritant patch test reactions. J Invest Dermatol. 1999;113:574–578
  22. Goebeler M, Trautmann A, Voss A, Brocker EV, Toksoy A, Gillitzer R. Differential and sequential expression of multiple chemokines during elicitation of allergic contact hypersensitivity. Am J Pathol. 2001;158:431–440
  23. Moed H, Boorsma DM, Stoof TJ, von Blomberg BM, Bruynzeel DP, Scheper RJ, et al. Nickel-responding T cells are CD4+ CLA+ CD45RO+ and express chemokine receptors CXCR3, CCR4 and CCR10. Br J Dermatol. 2004;151:32–41
  24. Cole KE, Strick CA, Paradis TJ, Ogborne KT, Loetscher M, Gladue RP, et al. Interferon-inducible T cell alpha chemoattractant (I-TAC): a novel non-ELR CXC chemokine with potent activity on activated T cells through selective high affinity binding to CXCR3. J Exp Med. 1998;187:2009–2021
  25. Farber JM. Mig and IP-10: CXC chemokines that target lymphocytes. J Leukoc Biol. 1997;61:246–257
  26. Abbas AK, Murphy KM, Sher A. Functional diversity of helper T lymphocytes. Nature. 1996;383:787–793
  27. Wakabayashi T, Hu DL, Tagawa Y, Sekikawa K, Iwakura Y, Hanada K, et al. IFN-gamma and TNF-alpha are involved in urushiol-induced contact hypersensitivity in mice. Immunol Cell Biol. 2005;83:18–24
  28. Dufour JH, Dziejman M, Liu MT, Leung JH, Lane TE, Luster AD. IFN-gamma-inducible protein 10 (IP-10; CXCL10)-deficient mice reveal a role for IP-10 in effector T cell generation and trafficking. J Immunol. 2002;168:3195–3204
  29. Lehto M, Koivuluhta M, Wang G, Amghaiab I, Majuri ML, Savolainen K, et al. Epicutaneous natural rubber latex sensitization induces T helper 2-type dermatitis and strong prohevein-specific IgE response. J Invest Dermatol. 2003;120:633–640
  30. Imai T, Nagira M, Takagi S, Kakizaki M, Nishimura M, Wang J, et al. Selective recruitment of CCR4-bearing Th2 cells toward antigen-presenting cells by the CC chemokines thymus and activation-regulated chemokine and macrophage-derived chemokine. Int Immunol. 1999;11:81–88
  31. Bonecchi R, Sozzani S, Stine JT, Luini W, D'Amico G, Allavena P, et al. Divergent effects of interleukin-4 and interferon-gamma on macrophage-derived chemokine production:an amplification circuit of polarized T helper 2 responses. Blood. 1998;92:2668–2671
  32. Xiao T, Fujita H, Saeki H, Mitsui H, Sugaya M, Tada Y, et al. Thymus and activation-regulated chemokine (TARC/CCL17) produced by mouse epidermal Langerhans cells is upregulated by TNF-alpha and IL-4 and downregulated by IFN-gamma. Cytokine. 2003;23:126–132
  33. Mantovani A, Gray PA, Van Damme J, Sozzani S. Macrophage-derived chemokine (MDC). J Leukoc Biol. 2000;68:400–404
  34. Werfel T, Hentschel M, Kapp A, Renz H. Dichotomy of blood- and skin-derived IL-4-producing allergen-specific T cells and restricted V beta repertoire in nickel-mediated contact dermatitis. J Immunol. 1997;158:2500–2505
  35. Liao F, Rabin RL, Smith CS, Sharma G, Nutman TB, Farber JM. CC-chemokine receptor 6 is expressed on diverse memory subsets of T cells and determines responsiveness to macrophage inflammatory protein 3 alpha. J Immunol. 1999;162:186–194
  36. Nakayama T, Fujisawa R, Yamada H, Horikawa T, Kawasaki H, Hieshima K, et al. Inducible expression of a CC chemokine liver- and activation-regulated chemokine (LARC)/macrophage inflammatory protein (MIP)-3 alpha/CCL20 by epidermal keratinocytes and its role in atopic dermatitis. Int Immunol. 2001;13:95–103
  37. Piguet PF, Grau GE, Hauser C, Vassalli P. Tumor necrosis factor is a critical mediator in hapten induced irritant and contact hypersensitivity reactions. J Exp Med. 1991;173:673–679
  38. Vanbervliet B, Bendriss-Vermare N, Massacrier C, Homey B, de Bouteiller O, Briere F, et al. The inducible CXCR3 ligands control plasmacytoid dendritic cell responsiveness to the constitutive chemokine stromal cell-derived factor 1 (SDF-1)/CXCL12. J Exp Med. 2003;198:823–830
  39. Bangert C, Friedl J, Stary G, Stingl G, Kopp T. Immunopathologic features of allergic contact dermatitis in humans: participation of plasmacytoid dendritic cells in the pathogenesis of the disease?. J Invest Dermatol. 2003;121:1409–1418
  40. Wollenberg A, Wagner M, Gunther S, Towarowski A, Tuma E, Moderer M, et al. Plasmacytoid dendritic cells: a new cutaneous dendritic cell subset with distinct role in inflammatory skin diseases. J Invest Dermatol. 2002;119:1096–1102
  41. Sebastiani S, Danelon G, Gerber B, Uguccioni M. CCL22-induced responses are powerfully enhanced by synergy inducing chemokines via CCR4: evidence for the involvement of first beta-strand of chemokine. Eur J Immunol. 2005;35:746–756

 Supported by Contract QLK4-CT-2001-00366 Chemokine-Atopy from the European Commission (to A.I.L., B.H., H.A.) and by Grant SFB503/C9 from the German Research Foundation (to B.H.).Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest.

PII: S0091-6749(07)00115-7

doi:10.1016/j.jaci.2006.12.654

The Journal of Allergy and Clinical Immunology
Volume 119, Issue 6 , Pages 1470-1480, June 2007