The Journal of Allergy and Clinical Immunology
Volume 124, Issue 6 , Pages 1235-1244.e58, December 2009

Broad defects in epidermal cornification in atopic dermatitis identified through genomic analysis

  • Emma Guttman-Yassky, MD, MSc

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • Department of Dermatology, Weill-Cornell Medical College, Cornell University, New York, NY
    • These authors contributed equally to this work.
    • ,
  • Mayte Suárez-Fariñas, PhD

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • These authors contributed equally to this work.
    • ,
  • Andrea Chiricozzi, MD

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • Department of Dermatology, University of Rome “Tor Vergata”, Rome, Italy
    • ,
  • Kristine E. Nograles, MD

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • ,
  • Avner Shemer, MD

      Affiliations

    • Tel-Hashomer Hospital and Tel-Aviv University, Tel-Aviv, Israel
    • ,
  • Judilyn Fuentes-Duculan, MD

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • ,
  • Irma Cardinale, MSc

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • ,
  • Peng Lin, BSc

      Affiliations

    • Department of Genetics, Washington University School of Medicine, St Louis, Mo
    • ,
  • Reuven Bergman, MD

      Affiliations

    • Rambam Medical Center and Bruce Rappaport Faculty of Medicine, Technion, Haifa, Israel
    • ,
  • Anne M. Bowcock, PhD

      Affiliations

    • Department of Genetics, Washington University School of Medicine, St Louis, Mo
    • ,
  • James G. Krueger, MD, PhD

      Affiliations

    • Laboratory for Investigative Dermatology, Rockefeller University, New York, NY
    • Corresponding Author InformationReprint requests: James G. Krueger, MD, PhD, Laboratory for Investigative Dermatology, Rockefeller University, 1230 York Avenue, New York, NY 10065.

Received 4 August 2009; received in revised form 15 September 2009; accepted 17 September 2009.

Article Outline

Background

Psoriasis and atopic dermatitis (AD) are common, complex inflammatory skin diseases. Both diseases display immune infiltrates in lesions and epidermal growth/differentiation alterations associated with a defective skin barrier. An incomplete understanding of differences between these diseases makes it difficult to compare human disease pathology to animal disease models.

Objective

To characterize differences between these diseases in expression of genes related to epidermal growth/differentiation and inflammatory circuits.

Methods

We performed genomic profiling of mRNA in chronic psoriasis (n = 15) and AD (n = 18) skin lesions compared with normal human skin (n = 15).

Results

As expected, clear disease classifications could be constructed on the basis of expected immune polarity (TH1, TH2, TH17) differences. However, even more striking differences were identified in epidermal differentiation programs that could be used for precise disease classifications. Although both psoriasis and AD skin lesions displayed regenerative epidermal hyperplasia, which is a general alteration in epidermal growth, keratinocyte terminal differentiation was differentially polarized. In AD, we found selective defects in expression of multiple genes encoding the cornified envelope, with the largest alteration in loricrin (expressed at 2% of the level of normal skin). At the ultrastructural level, the cornified envelope in AD was broadly defective with highly decreased compaction of corneocytes and reduced intercellular lipids. Hence, the entire keratinocyte terminal differentiation program (cytoplasmic compaction, cornification, and lipid release) is defective in AD, potentially underlying the immune differences.

Conclusion

Our study shows that although alterations in barrier responses exist in both diseases, epidermal differentiation is differentially polarized, with major implications for primary disease pathogenesis.

Key words: Atopic dermatitis, psoriasis, cornified envelope, terminal differentiation, altered barrier

Abbreviations used: AD, Atopic dermatitis, CDSN, Corneodesmosin, CE, Cornified envelope, EDC, Epidermal differentiation complex, ES, Enrichment score, FC, Fold change, FDR, Fold discovery rate, FLG, Filaggrin, GSEA, Gene Set Enrichment Analysis, IVL, Involucrin, LCE, Late cornified envelope, LOR, Loricrin, SC, Stratum corneum, SPRR, Small proline rich protein

 

Psoriasis and atopic dermatitis (AD) are common inflammatory skin diseases often compared to determine common versus unique cellular and molecular characteristics.1, 2 These diseases share important features, including (1) large infiltrates of T cells and inflammatory dendritic cells in skin lesions; (2) immune activation with upregulated expression of many cytokines, chemokines, and inflammatory molecules2, 3; (3) marked epidermal hyperplasia in chronic diseased skin3, 4; and (4) defective barrier function with increased transepidermal water loss, which reflects underlying alterations in keratinocyte differentiation.5, 6, 7 Beyond these broad similarities, there are also key differences between psoriasis and atopic dermatitis. A major focus of research has been on immune polarization differences between these diseases. Whereas psoriasis is considered a TH1 and TH17 disease, AD is predominantly a TH2-polarized disease, with a possible TH1 shift in the chronic phase.3, 8, 9, 10 More recently, it has been found that immune-related cytokines can have a major impact on growth and differentiation of epidermal keratinocytes. For example, IL-22 and other IL-20 family cytokines are major mitogens for epidermal keratinocytes that also inhibit terminal differentiation of these cells.11, 12, 13 Hence, pathologic changes in keratinocytes, many of which are similar in psoriasis and AD, may be reactive changes to inflammation. However, the similarities and differences in altered epidermal differentiation (resulting from either an immune-stimulated effect or underlying genetics) between these diseases are not yet fully understood.

An intrinsic keratinocyte differentiation defect involving the expression of filaggrin (FLG) has recently been suggested in AD.14, 15, 16 Although loss of function mutations in the gene encoding FLG were proposed in AD, FLG mutations alone do not provoke AD, because the mutations were observed only in a group of patients from certain ethnic groups (as many as 20% in all patients with AD and 50% in patients with severe AD). Similar mutations were also reported in ichthyosis vulgaris and in approximately 9% of the European population, without concomitant inflammation.14, 17 Furthermore, most patients with AD demonstrate a residual granular layer and orthokeratotic keratinocyte maturation, implying a residual function of FLG in AD.3, 18 Another argument that defective epidermal barrier is not restricted to the FLG defect stems from analysis of lipids synthesized during the cornification process. Elias and others7, 19, 20 demonstrated that AD is associated with distinctive abnormalities in lipid composition in the stratum corneum (SC; ie, decreased total lipids and ceramides) that correlate with defective barrier function. Inappropriate barrier response was also implicated recently in psoriasis pathogenesis with identification of a deletion of late cornified envelope (LCE) LCE3B and LCE3C genes.21

A number of animal models of skin inflammation have been created by modifying keratinocyte differentiation genes, suggesting that structural epidermal defects have the potential of contributing to psoriasis and AD pathogenesis.22, 23, 24, 25, 26, 27 Development of mouse models for inflammatory skin diseases that accurately represent the disease compared with normal mouse skin mandates a good molecular fingerprinting of the disease phenotype compared with normal human skin.

In this study, we describe gene (mRNA) expression profiles of psoriasis and AD by using normal skin as a reference. This has allowed us to classify disease-related alterations that are unique to each disease as well as those shared by both diseases. Whereas many expected alterations in keratinocyte terminal differentiation were detected in psoriasis,28, 29 we found many new, as well as extremely broad, differences in terminal differentiation in AD skin lesions. Although psoriasis and AD share epidermal hyperplasia and regenerative epidermal growth, terminal differentiation is accelerated in psoriasis but broadly suppressed or defective in AD. We show that psoriasis and AD can be discriminated on the basis of differences in expression of terminal differentiation genes, much like disease distinctions made on the basis of immune-response phenotypes. These data provide new insights into AD pathogenesis and a means to validate disease models.

Back to Article Outline

Methods 

Skin samples 

Skin biopsies were collected from 18 patients with chronic AD (lesional skin only; 12 males, 6 females; age, 17-66 years; median, 37 years), 15 patients with psoriasis (11 males, 4 females; age, 28-59 years; median, 48 years), and 15 healthy volunteers (7 males, 8 females; age, 24-69 years; median, 41 years) under a Rockefeller University Institutional Review Board–approved protocol as previously described.3 Patients with moderate to severe psoriasis (involvement of >10% body surface area) and with an acute exacerbation of chronic AD (Scoring Atopic Dermatitis between 20 and 70, all with elevated IgE) that did not receive any therapy for >4 weeks were included. Biopsies were frozen in optimal cutting temperature medium for immunohistochemistry and liquid nitrogen for RNA extraction.

Immunohistochemistry 

Cryostat tissue sections of all patients with AD and psoriasis were stained with hematoxylin (Fisher, Fair Lawn, NJ) and eosin (Shandon, Pittsburgh, Pa), and purified mouse antihuman mAbs were used (for details, see this article's Methods in the Online Repository at www.jacionline.org).

Electron microscopy 

Punch biopsies of 6 mm were obtained from 4 healthy volunteers and lesional skin of 4 patients with chronic AD. After rinsing in cocodylate buffer, samples were fixed in half-strength Karnovsky fixative, processed routinely through reduced 1% osmium tetroxide, and dehydrated in graded ethanols, followed by embeddement in an Epon-epoxy mixture (Electron Microscopy Sciences, Fort Washington, Pa) as previously described30, 31, 32 (see this article's Methods in the Online Repository).

Immunofluorescence localization of neutral lipids 

Fresh cryostat sections were stained with Nile red solution (500 μg/mL) in acetone, and images were acquired by using appropriate filters of Zeiss Axioplan 2 (LSM 510) (Carl Zeiss Microimaging, Thornwood, NY) as described30 (see this article's Methods in the Online Repository).

Sample preparation for real-time PCR and gene chip analysis 

The microarrays used for this study were the U133A 2 and U133 plus 2 GeneChip probe arrays (Affymetrix Inc, Santa Clara, Calif) as previously described10, 33 (see this article's Methods in the Online Repository).

Real-time PCR analysis 

The primers and probes for the TaqMan real-time PCR assays were generated with the Primer Express algorithm, version 1.5, by using published genetic sequences (NCBI-PubMed) for each gene.

The real-time PCR reaction was performed by using EZ PCR Core Reagents (Applied Biosystems, Foster City, Calif) as previously reported3 and for LCE genes as described by Jackson et al.34

See this article's Methods in the Online Repository detailed information.

Statistical analyses 

CEL files quality control was assessed by Harshlight35 and affyQCReport packages from R/Bioconductor. Expression measures were obtained by using robust multi-array average with GC background correction (GCRMA). Differences among groups were determined by moderated t test,36 and P values were adjusted by using the Benjamini-Hochberg procedure. Classic clustering algorithms were applied. Gene-based classifiers were obtained by using the nearest shrunken centroid method of Tibshirani et al.37 The relevance of terminal differentiation gene sets in comparisons of interests was evaluated by Gene Set Enrichment Analysis (GSEA).38 See this article's Methods in the Online Repository for detailed statistical analyses.

Back to Article Outline

Results 

Disease-defining differences in gene expression 

We used Affymetrix U133 A2.0 (for psoriasis) and U133 plus 2 (for normal and AD) arrays to quantify expression of transcripts in lesional skin of AD and psoriasis. All 3 conditions were compared by using genes common to both array series, with extra normalization steps across array series to reduce batch bias. Using normal skin as the reference, we found 1278 probe-sets (1073 unique genes using ENTREZ identifiers) with significantly increased expression in psoriasis and 720 probe-sets (640 genes) with significantly increased expression in AD skin lesions (using criteria of fold change [FC] ≥3, fold discovery rate [FDR] <0.05). Conversely, 1474 probe-sets (1264 genes) were decreased in psoriasis, and 939 probe-sets (788 genes) were decreased in AD skin lesions (see this article's Fig E1 in the Online Repository at www.jacionline.org). Some genes with differential expression are known to be commonly upregulated or downregulated in psoriasis and AD, including the markers of epidermal hyperplasia K16 and Ki67 (MKI67) mRNA. These are increased by >10-fold in both diseases (see this article's Table E1 and Fig E2 in the Online Repository at www.jacionline.org). Numerous cell-cycle genes associated with increased cell growth are commonly upregulated in psoriasis and AD, as expected from epidermal hyperplasia that characterizes disease pathology in both conditions (Table E1; Fig E2).

However, the majority of disease-related gene expression differences were distinct when the 2 diseases were compared. In AD, more than 50% of genes upregulated or downregulated compared with normal skin did not have cognate changes in psoriasis, whereas an even larger fraction of psoriasis-related genes were not commonly altered in AD (Fig E1; Table E1). Fig 1 shows the top 25 genes exclusively upregulated or downregulated in only 1 disease. Many of the genes shown in Fig 1 relate to inflammation or epidermal growth/differentiation, potentially suggesting key differences in inflammation and epidermal proliferation/differentiation between these inflammatory skin diseases. Although FLG expression was approximately 5 FC lower in lesional AD than normal skin, we were struck by the magnitude of reduced expression of other genes associated with terminal differentiation in AD compared with normal skin (7 genes; P < .006) and psoriasis (18 genes; P < 10−10). Differentially regulated genes included corneodesmosin (CDSN), LCE2B, and loricrin (LOR). CDSN, LCE2B, and LOR were decreased between 10-fold and 100-fold in AD versus normal skin (Table E1; Fig 1). We thus focused our analysis on the hypothesis that AD lesions have broad, disease-specific defects in keratinocyte terminal differentiation.

Deeper study of epidermal differentiation 

To examine further the extent to which genes of cornified envelope (CE) and epidermal differentiation complex (EDC; collectively referred to as terminal differentiation) were selectively altered in AD versus psoriasis, we explored all genes associated with these pathways in psoriasis, AD, and normal skin (Fig 2). From Affymetrix data, selective decreases in CE genes, including psoriasis 1 candidate 2 (PSORS1C2) and CDSN (part of the MHC complex on chromosome 6p21.3 and harboring the psoriasis susceptibility locus 1 variant [PSORS1]; Fig 2, A) and EDC genes (Fig 2, A), such as LCE2B, LOR, involucrin (IVL), and small proline rich proteins (SPRRs), were found in AD compared with both normal and psoriatic skin (Fig 2, A). Several LCE genes were highly downregulated in AD versus normal skin by using U133 Plus 2.0 arrays (see this article's Table E2 in the Online Repository at www.jacionline.org). However, because the U133A 2.0 array series has poor representation of LCE genes, expression levels of multiple LCE genes was performed by quantitative real-time PCR (Fig 2, B). A striking downregulation (to an almost undetectable level) was observed in expression of class 1 and 2 LCE genes in lesional AD compared with both normal and psoriatic skin (Fig 2, B). SPRR2C and CDSN were significantly downregulated compared with psoriasis, and CDSN was significantly downregulated compared with normal skin (Fig 2, B). We further tested the hypothesis of downregulation of terminal differentiation genes in AD versus upregulation in psoriasis compared with normal skin by GSEA, which considered gene sets rather than gene-by-gene analysis, and for each gene set, the enrichment score (ES) was calculated.38 Indeed, a negative and highly significant enrichment score (ES, –2) was obtained in AD, in contrast with a positive score (ES, 1.9) in psoriasis (P < .001 for both comparisons), supporting our hypothesis.

  • View full-size image.
  • View full-size image.
  • Fig 2. 

    Genomic expression of terminal differentiation (CE and EDC) genes. A, Downregulation of terminal differentiation genes in AD compared with normal and psoriasis (PS) skin. FC by microarrays in log2 scale, ∗FDR <0.05. B, Real-time PCR analysis showing significant downregulation of LCE1, LCE2, CDSN, and SPRR2C in AD; ∗P <.05; ∗∗P <.01; ∗∗∗P <.001. C, Terminal differentiation class prediction, discriminating AD and PS. Horizontal bars represent centroids for the AD and PS groups (upregulated genes [vs normal] on the right and downregulated on the left). Norm, Normal.

New disease classification based on expression of terminal differentiation genes 

Currently, the best understood differences in gene expression that might be used to distinguish psoriasis from AD relate to distinct TH1, TH2, and TH17 gene products and chemokines.1, 3, 10, 39 The data presented in this article's Fig E3 in the Online Repository at www.jacionline.org illustrate that gene expression measured by gene array (Fig E3, A) or real-time PCR (Fig E3, B) techniques can be used to classify these diseases with high certainty. An “immune-signature” class prediction through a discriminant analysis provided a set of 10 immune-related genes that could accurately classify all of the cases in our array data as psoriasis or AD (Fig E3, C). Because of the striking differences in expression of terminal differentiation genes between psoriasis and AD, supported by the highly significant GSEA results, we attempted a new disease classification scheme based solely on these genes. Using the method of Tibshirani et al,37 we obtained a class prediction with 13 genes expressed in opposite polarities in AD versus psoriasis relative to overall centroids (Fig 2, C). This classifier was able to determine correctly (with 0 error) the nature of the disease (AD/psoriasis) in the array data of all study participants. The horizontal bars in Fig 2, C, represent centroid expression values of the individual 13 genes comprising the disease classifier, with higher expression to the right and lower expression to the left, compared with the overall centroid.

Compromised integrity of epidermal barrier in AD 

To determine whether the observed gene expression differences might translate into altered epidermal structure or function, we examined the in vivo formation of the terminal differentiation structures using a combination of electron microscopy and immunofluorescence or immunohistochemistry staining. This approach was used because the cross-linking of CE precursors into the CE makes for an insoluble protein network in which quantification of levels of individual proteins is not ascertainable by conventional techniques of protein chemistry.

The integrity of CE in lesional AD versus normal skin was assessed by electron microscopy. Compared with normal skin (Fig 3, A and E), in all studied patients with AD there was increased thickness of corneocytes and less cytoplasmic compaction (Fig 3, B and F; mean thickness of a normal corneocyte was 1.09 ± 0.38 μm, compared with 2.15 ± 0.45 μm in AD corneocytes; Fig 3, A, E, and B, F, respectively). Unlike normal skin (Fig 3, A), in which CE is clearly visualized as a continuous distinguishable structure, in AD skin there is general thinning, disruption, and even a virtual absence of CE in certain areas (Fig 3, B-D). A degradation of corneodesmosome structure can also be visualized in lesional AD compared with normal skin (Fig 3, B-D). These findings suggest a disturbed cohesion pattern of AD corneocytes. For better visualization of SC intercellular lipids, Nile red immunofluorescence staining was used (Fig 3, G and H). In contrast with normal skin (Fig 3, G), in which distinct layers of extracellular lipids exist between thin compacted corneocytes, in AD we observed a highly decreased compaction of corneocytes and highly reduced intercellular lipids (Fig 3, H).

  • View full-size image.
  • Fig 3. 

    Compromised epidermal barrier integrity in AD versus normal skin. SC of a healthy volunteer (A and E) and a representative patient with AD patient (B-D, F). Nile red fluorescence in normal (G) and AD SC (H). CE (white arrows), corneodesmosomes (black arrows), lamellar body (white asterisk), and keratin filaments (white triangles). Disruption of CE, degradion of corneodesmosomes (B-D), increased corneocyte thickness, and decreased compaction of keratin filaments in AD (F) compared with normal skin (E). Decreased corneocyte compaction and reduced intercellular lipids in AD (H) versus normal skin (G). Scale bars: A-D, 500 nm; E and F, 5 um; G and H, × 40 magnification, zoom × 9.7.

Major differences in epidermal maturation and differentiation between AD and psoriasis 

Expression of proteins associated with terminal differentiation was studied by immunohistochemistry. In normal skin, as expected as part of homeostatic growth,40 there is a coordinated pattern of expression of terminal differentiation proteins, confined to the granular layer (Fig 4, A-E). Psoriasis, similar to normal wounds, displays regenerative maturation,40 with accelerated growth and maturation of keratinocytes. Within this growth program that is marked by acanthosis and keratin 16 expression, many “granular” layer products, such as FLG, IVL, CDSN, and LOR, are prematurely synthesized by keratinocytes at the midspinous layer level and continue to be expressed in the granular layer and often into the SC (Fig 4, A-E). The epidermis of chronic AD also shows regenerative growth on the basis of marked acanthosis and K16 expression (Fig 4, A). However, in AD, the coordinated expression of terminal differentiation proteins is largely delayed, occurring mostly in the granular layer, and for some proteins, such as FLG and LOR, as a discontinuous band (Fig 4, B-E). Importantly, both normal skin and psoriasis retain expression of CDSN and LOR in the SC as part of normal CE formation (Fig 4, D and E). However, in AD, there is a relative absence of CDSN and LOR in corneocytes above the granular layer, suggesting abnormal formation or retention of these terminal differentiation structures in corneocytes (Fig 4, D and E).

  • View full-size image.
  • Fig 4. 

    Characterization of major terminal differentiation proteins in normal skin, psoriasis (PS), and AD by immunohistochemistry. Skin sections were stained for hyperplasia (K16) (A) and the following: FLG (B), IVL (C), CDSN (D), and LOR (E). Although PS and AD showed similar acanthosis (A), delayed expression and abnormal formation of differentiation proteins characterized AD.

Back to Article Outline

Discussion 

This study represents the most comprehensive genomic comparison of chronic psoriasis and AD skin lesions compared with normal skin. Using this approach, we measured many more gene differences than smaller previous studies with more limited comparisons (summarized in this article's Table E3 in the Online Repository at www.jacionline.org). Our study has identified broad alterations in the expression of epidermal cornification genes in patients with AD.3, 4, 8, 9, 10, 41, 42, 43, 44 Collectively, past studies convey a confusing view of epidermal differentiation in AD, with identification of small sets of genes inconsistently increased or decreased in skin lesions.3, 4, 8, 9, 10, 41, 42, 43, 44

The current study identifies new and unexpected differences in “pathologic” epidermal hyperplasia between psoriasis and AD. The epidermal reaction in psoriasis bears strong similarity to regenerative maturation, the alternate growth/differentiation program of the epidermis that is triggered in response to skin injury or wounding.40 In psoriasis, regenerative epidermal hyperplasia is driven by underlying immune activation, either as a direct response to IL-20 family cytokines (which induce hyperplasia and inhibit keratinocyte terminal differentiation)11, 12, 45 or as an indirect response to immune-mediated injury to keratinocytes.46 The epidermal reaction in psoriasis is largely restored to normal or homeostatic growth pathway with selective immune suppression.47 Hence, one might hypothesize that similar regenerative epidermal responses should occur in the presence of “generalized” cellular immune activation, such as in diseases with similar inflammatory infiltrates. In fact, psoriasis and AD share features of dense T cells and dendritic cell infiltrates3 as well as overexpression of IL-22 in skin lesions.48 In conceptual terms, both diseases could be considered to have an altered epidermal barrier, and on the basis of similar epidermal hyperplasia and KRT 16 expression, both have regenerative maturation responses in the epidermis. It is thus surprising that expression of epidermal differentiation genes is vastly different between these “regenerative” processes.

The lower expression of a large set of keratinocyte terminal differentiation genes in AD (extending far beyond the FLG gene) translates to very late onset of terminal differentiation in granular keratinocytes, as assessed by protein expression/immunohistochemistry. In contrast, terminal differentiation of keratinocytes in psoriasis begins in the midspinous layer, which is the expected regenerative response (Fig 2, Fig 4). Both diseases also show defective secretion or retention of intracellular lipids in SC, which could represent a common barrier abnormality.49 These data do not solve the fundamental dispute in AD of whether primary pathogenesis is “inside out” or immune-driven versus “outside-in” or driven by a defective barrier.7, 50, 51 However, our data do suggest some new possibilities for interaction of pathologic immunity with diseases of defective skin barriers. One must now ask whether specific immune cytokines or effector immune functions can have vastly different effects on the process of keratinocyte terminal differentiation within a regenerative growth program. Because there is convincing evidence for a genetic linkage between the barrier dysfunction in AD and a locus on chromosome 1q21 containing the EDC,14 and initial evidence of genetic variations within the MHC52, 53 and also the EDC54, 55, 56, 57, 58, 59 that might contribute to psoriasis susceptibility, the differences in epidermal barrier defects between these diseases may be crucial in their pathogenesis. Although AD and psoriasis are differentially immune polarized diseases, with psoriasis considered a TH1 and TH17 disease, and AD a mostly TH2 disease,10 the differences in barrier polarity between these 2 diseases are equally or more impressive. A possible link between the immune and barrier defects was recently suggested for both diseases.12, 48, 60, 61 In AD, IL-4 and IL-13 TH2 cytokines are expressed at higher levels and can modify some keratinocyte responses, including inhibiting production of terminal differentiation60 and/or antimicrobial proteins.61 In contrast, in psoriasis, IFN-γ and IL-17 are expressed at higher levels, each inducing distinct gene expression programs and having synergistic interactions with IL-22 for regulation of terminal differentiation genes, such as S100 family proteins.12 Conversely, the relatively defective differentiation of corneocytes in AD may allow vastly different penetration of epicutaneous antigens and thus differential triggering of underlying immune responses.49 From this standpoint, activation of TH2 immunity is selectively triggered by Langerhans cells,62 and one could hypothesize that greater access of surface antigens to Langerhans cells in AD might result in TH2 activation, which is a hallmark of this disease.

The ability to pursue some of these questions and test specifics of immune interactions with keratinocyte differentiation programs will likely depend on animal models of inflammatory skin diseases. In this context, and especially for epidermal responses to TH1 versus TH2 versus TH17 immune reactions, it will be important to determine how faithfully the model systems reflect the actual pathology of AD and psoriasis in human skin. With many emerging mouse models that try to simulate the characteristics of human AD and psoriasis diseases,22, 23, 24, 25, 26, 27, 63 we need a unique set of genes that can be used as disease identifiers and help produce mouse models that will accurately represent the disease phenotype in human skin. Because epidermal hyperplasia and immune infiltrates are very similar in chronic AD and psoriasis vulgaris,3, 4 it is not clear that distinctions can be drawn by examining epidermal hyperplasia or immunophenotyping leukocytes in model systems. Instead, model systems should be examined for polarity of T-cell responses and also for the set of differentiation genes that are completely different in psoriasis and AD. Previous mouse models of both AD and psoriasis failed fully to reproduce disease characteristics in human beings, leading to mostly a psoriasiform hyperplasia that cannot differentiate between these diseases. We think that the disease classification dilemma in model systems is well illustrated by a targeted deletion of the murine CDSN gene, showing a “psoriasis-like” phenotype that might be impossible to differentiate from an “AD-like phenotype.”64 Similarly, epidermal deletion of JunB was found to produce a “psoriasis-like” phenotype in mice,25 whereas JunB was shown to be upregulated in human psoriatic lesions.65 Only more detailed genomic or cellular fingerprinting can determine whether this and many other models—knock-in, knock-out, immune-mediated, or spontaneous mutations—are more reflective of psoriasis, AD, or unrelated disease processes. The differential gene expression maps presented here probably represent the best current basis for drawing these distinctions. Our data provide a useful tool of a restricted number of genes able to identify AD and psoriasis reliably, compared with normal skin.

Clinical implications

Targeted treatment of atopic dermatitis will need to be directed either to disease-specific epidermal defects or to underlying (polar) cellular immune circuits that drive distinct hyperproliferative epidermal responses.

Back to Article Outline

 

We thank Shenghui Duan for her help with real-time PCR analysis for LCE genes.

Back to Article Outline

Methods 

Skin samples 

Skin biopsies were collected from 18 patients with chronic AD (lesional skin only; 12 males, 6 females; age, 17-66 years; median, 37 years), 15 patients with psoriasis (lesional and nonlesional skin; 11 males, 4 females; age, 28-59 years; median, 48 years), and 15 healthy volunteers (7 males, 8 females; age, 24-69 years; median, 41 years), under a Rockefeller University Institutional Review Board–approved protocol as previously described.E1 Patients with moderate to severe psoriasis (involvement of >10% body surface area) and with an acute exacerbation of chronic AD (Scoring Atopic Dermatitis between 20 and 70, all with elevated IgE) that did not receive any therapy for >4 weeks were included. Diagnoses were confirmed histologically, and there were no cases of diagnostic discordance. Biopsies were frozen in optimal cutting temperature medium for immunohistochemistry and liquid nitrogen for RNA extraction.

Immunohistochemistry 

Cryostat tissue sections of all patients with AD and psoriasis were stained with hematoxylin (Fisher, Fair Lawn, NJ) and eosin (Shandon, Pittsburgh, Pa). For immunohistochemistry, purified mouse antihuman mAbs were used to the following: CDSN (Abcam, Cambridge, Mass), FLG (Acris Antibodies GmbH, Hiddenhausen, Germany), IVL (GeneTex, Irvine, Calif), and K16 (Clone 8.12; Sigma, St Louis, Mo); purified polyclonal rabbit antibody to LOR was used (Abcam). Biotin-labeled horse antimouse antibody (Vector Laboratories, Burlingame, Calif) was amplified with avidin-biotin complex (Vector Laboratories) and developed with chromogen 3-amino-9-ethylcarbazole (Sigma Aldrich, St Louis, Mo).

Electron microscopy 

Punch biopsies of 6 mm were obtained from normal appearing non–sun-exposed skin of 4 healthy volunteers and lesional skin of 4 patients with chronic AD. After rinsing in cocodylate, buffer samples were fixed in half-strength Karnovsky fixative and processed routinely through reduced 1% osmium tetroxide. The tissues were dehydrated in graded ethanols, followed by embeddement in an Epon-epoxy mixture (Electron Microscopy Sciences, Fort Washington, Pa), as previously described.E2, E3, E4 Ultrathin sections were cut after further contrasting with 2% uranyl acetate and 1% citric acid solution and observed under an electron microscope (FEI Tecnai G2 12, FEI, Hillsboro, Ore) with a digital camera imaging system (Gatan 895 Ultrascan 4000, Walendale, Wash).

Immunofluorescence localization of neutral lipids 

A stock solution of Nile red (500 μg/mL) in acetone was prepared, stored at –20°C, and protected from light. Fresh staining solution of Nile red was prepared as previously described,E2 and fresh frozen cryostat sections were stained and examined after 10 minutes at room temperature in the dark.E2 Images were acquired by using appropriate filters of a Zeiss Axioplan 2 (LSM 510) (Carl Zeiss Microimaging, Thornwood, NY) laser scanning confocal microscope fitted with C-Apochromat 40x/1.2 W objective lens and a krypton/argon laser at a 568-nm excitation frequency. The zoom image is ×9.7 the image taken at ×40 magnification.

Sample preparation for real-time PCR and gene chip analysis 

The microarrays used for this study were the U133A 2.0 for psoriasis skin samples and U133 plus 2 GeneChip probe arrays for AD and normal skin samples (Affymetrix Inc, Santa Clara, Calif). The U133A 2.0 array includes more than 22,000 probe sets to analyze expression level of more than 18,400 transcripts, including 14,500 well characterized genes. The U133 Plus 2.0 Array includes more than 54,000 probe sets to assess expression of more than 47,000 transcripts, including approximately 38,500 genes and unigenes. The labeled target was fragmented and hybridized to probe arrays as previously described.E5, E6 Briefly, total RNA was extracted from tissues frozen in liquid nitrogen using the RNeasy Mini Kit (Qiagen, Valencia, Calif). DNA was removed by on-column DNAse digestion by the Quiagen RNAse-free DNAse Set. Total RNA (∼2 ug) was reverse-transcribed, amplified, and labeled as previously reported.E5, E6 mRNA was isolated and converted to double-strand cDNA and then to biotinylated cRNA (BioArray High Yield RNA Transcription Labeling Kit; Enzo Biochem, Inc, Farmingdale, NY). After fragmentation and quality confirmation with Affymetrix Test-3 Array, 15 μg of the biotinylated cRNA was hybridized to Affymetrix Human Genome U133A2 (21,722 probe sets) or U133 plus 2 GeneChips (54,120 probe sets; Affymetrix, Inc). The chips were washed, stained with streptavidin-phycoerythrin, and scanned (HP GeneArray Scanner; Hewlett-Packard Company, Palo Alto, Calif). On each chip, the human housekeeping genes β-actin and glyceraldehyde-3-phosphate dehydrogenase GAPDH served as controls. Suite 5.0 software normalized the expression level values using these controls. Chips with 3′ to 5′ ratios for GAPDH less than 3 and scaling factor within 3-fold of each other were compared for the study.

Real-time PCR analysis 

The primers and probes for the TaqMan real-time PCR assays were generated with the Primer Express algorithm, version 1.5, using published genetic sequences (NCBI-PubMed) for each gene. Primer sequences were as follows: IFN-γ forward, GGTCTGTGAAGAGCCGTTGTC; IFN-γ reverse, TGTTCCACTTTTCCTGGATTGTC; IFN-γ probe 6-FAM- CCAGCAACAGTTCCAGGCATGCA-; iNOS forward, CCTCAAGTCTTATTTCCTCAACGTT; iNOS reverse, CCGATCAATCCAGGGTGCTA; inducible nitric oxide synthase (iNOS) probe 6-FAM-CCCCATCAAGCCCTTTACTTGACCTCC-TAMRA (Gene Bank Accession Number AF068236); IL-10 forward, GGCGCTGTCATCGATTTCTT; IL-10 reverse, TGGAGCTTATTAAAGGCATTCTTCA; IL-10 probe 6-FAM AAACAAGAGCAAGGCCGTGGAGCAG; IL-8 forward, GCTGGCCGTGGCTCTCT; IL-8 reverse, TTTAGCACTCCTTGGCAAAACTG; IL-8 probe 6-FAM CCTTCCTGATTTCTGCAGCTCTGTGTGA; CCL18 (PARC) forward, CCAGCTCACTCTGACCACTTCTC; CCL18 reverse, GGTGCAGACGAGGACAAGGA; CCL18 probe 6-FAM CTGCCCAGCATCATGAAGGGCCT. The primers and probes for IL-17A (Assay ID Hs00174383), IL-17F (Assay ID Hs00369400), Elafin/peptidase inhibitor 3, skin derived (SKALP; Assay ID Hs00160066), DEFB4/hBD2 (Assay ID Hs00823638), p19 (Assay ID Hs00372324), IL-1β (Assay ID Hs00174097), IFN-γ (Assay ID Hs99999041), signal transducer and activator of transcription 1 (Assay ID Hs01014002), TSLPR (Assay ID Hs00845692), IL-13 (assay ID Hs99999038), IL-15 (Assay ID Hs00542562), CCL17 (Assay ID Hs00171074), CDSN (Assay ID Hs00169911), and SPRR2C (Assay ID Hs00272438) were designed by Applied Biosystems (Foster City, Calif).

The real-time PCR reaction was performed by using EZ PCR Core Reagents (Applied Biosystems) according to the manufacturer's directions and as previously reported.E1 The human acidic ribosomal protein (hARP) gene, a housekeeping gene, was used to normalize each sample and each gene: HARP forward, CGCTGCTGAACATGCTCAA; HARP reverse, TGTCGAACACCTGCTGGATG; HARP probe 6-FAM-TCCCCCTTCTCCTTTGGGCTGG-TAMRA (Gene Bank Accession Number NM-001002). The data were analyzed and quantified by the software provided with the Applied Biosystems PRISM 7700 (Sequence Detection Systems, version 1.7). Quantitative real-time PCR was performed with routine methods by using primers for LCE genes described by Jackson et al.E7 Levels of PCR products were evaluated in real time with an ABI Prism 7000 Sequence Detection System. For each sample/transcript, expression levels were normalized by comparison with endogenous 18S rRNA transcripts. Levels of each transcript that are provided are expressed relative to the low abundance transcript LCE2C in individual AL17. Real-time PCR was used to quantify expression of cytokine mRNA levels, because most of these mRNAs are expressed at a relatively low level and are not accurately detected on Affymetrix arrays.

Statistical analyses 

General analysis strategy of 2 diseases 

We used Affymetrix U133 A2.0 (for psoriasis) and U133 plus 2 (for normal and AD) arrays to quantify expression of transcripts in lesional skin of AD and psoriasis. All 3 conditions were compared by using a similar number of genes existing on U133 A2.0 arrays. We also quantified transcripts in normal human skin so that we could determine whether genes differentially expressed in these inflammatory diseases were increased, decreased, or unchanged from normal values. The advantage to our approach is that it permits a better understanding of whether disease-related expression differences are abnormal with respect to normal biology.

Gene array analysis. 

Quality control, preprocessing, and filtering 

GeneChip CEL files were scrutinized for spatial artifacts by using the Harshlight package (http://asterion.rockefeller.edu/Harshlight/index2.html).E8 Classic microarray quality control was obtained by using the affyQCReport package from Bioconductor. For obtaining expression values, CEL file intensity values were preprocessed by GCRMA algorithm for samples in each chip series separately. For proper analysis of all conditions, because different chips were used for psoriasis (U133 A2.0) and AD and normal samples (U133 Plus 2.0), only U133 A2.0 probe sets were used. The scope of the analysis was reduced to genes that appeared in both chips, and extra normalization steps across both series were taken to reduce batch bias. The expression values were filtered to eliminate probe sets with low variation/intensity. Probe sets with expression values >3 in at least 1 sample and SD >0.1 were kept for further analysis.

Differential expression criteria 

A linear model with condition (AD, psoriasis, and normal) as a factor was used to model gene expression. Model fitting and hypothesis testing were conducted by using the Bioconductor limmapackage (http://www.bioconductor.org/).E9 For each gene, a moderated t test was used to assess the significance of change between the variable groups (AD vs normal, psoriasis vs normal, and AD vs psoriasis). P values were adjusted for multiple hypotheses testing, controlling the FDR by using the Benjamini-Hochberg procedure. Genes with similar patterns of expression were grouped together as hierarchical clusters and presented as color-coded heatmaps. Single linkage was used as an agglomeration method with Euclidean distance.

Discriminant analysis 

We used a discriminant analysis method that performs sample classification from gene expression data via the nearest shrunken centroid method of Tibshirani and Efron.E10 This method reduces the effect of noisy genes and performs automatic gene selection, thus increasing the classifier accuracy. Predictive performance was assessed by a 9-fold validation. Under these settings, the value of the classifier threshold (which determines the number of genes included in the final classifier) was based on misclassification error in cross-validation and the resulting FDR for genes included in the classifier. Analysis was performed by using the package pamr under the programming language R (http://www.R-project.org/).

GSEA 

To determine whether the set of terminal differentiation genes are differentially expressed in AD and PS compared with normal skin, we used GSEA as previously described.E11 The GSEA method uses a general framework in which any collection of gene sets (eg, Kyoto Encyclopedia of Genes and Genomes [KEGG]) can be mined for correlations with a specific phenotype (in this case, disease vs normal). For each gene set, the ES is calculated by (1) ranking the available probe-sets with respect to phenotype (ie, disease/normal fold change) and (2) using the Kolmogorov-Smirnov statistics to measure the proximity of the gene set to the top of the list. A high and positive ES indicates that the gene set is upregulated in the disease versus the normal phenotype. The significance (P value) of the observed ES is calculated by using 1000 simulations. The software GSEA (available at http://www.broadinstitute.org/gsea/) was used for this analysis.

Back to Article Outline

Real-time PCR analysis 

Statistical comparisons of mRNA expression level were performed among lesional AD, psoriasis, and normal skin. A 2-sided Student t test on the log2 transformed data was used, and results with a significance level of at least .05 were considered significant.

Back to Article Outline

Fig E1. 

  • View full-size image.

  • Differentially and commonly upregulated (red) and downregulated (green) regulated genes in AD and psoriasis (PS) versus normal skin, by arrays. Overall many more genes were upregulated and downregulated in psoriasis than in AD, with 322 and 441 genes commonly upregulated and downregulated, respectively, in both diseases. Criteria of FC ≥3 and FDR <0.05 applied.

Back to Article Outline

Fig E2. 

  • View full-size image.

  • Genomic expression similarities between AD and psoriasis (PS) compared with normal skin. Heat maps representing the top 25 upregulated genes (A) and downregulated genes (B) in both AD and PS compared with normal skin. FC values represent AD and PS versus normal skin. All FDR values were highly significant (P < .001).

Back to Article Outline

Fig E3. 

  • View full-size image.
  • View full-size image.

  • Genomic expression differences of TH1, TH2, and TH17 immune pathways in AD, psoriasis (PS), and normal skin by microarray (A) and real-time PCR (B). Whereas PS deviates toward TH1/TH17, AD polarizes toward TH2. C, “Immune-signature” class prediction classifying these diseases. Horizontal bars represent centroids for AD and PS groups (upregulated genes (vs normal) on the right, downregulated on the left). Norm, Normal.

Back to Article Outline

Table E1. 

Genomic expression differences and similarities between AD and psoriasis compared with normal skin. Criteria of FC ≥3 and FDR <0.05 were used for microarray analysis.

Back to Article Outline

Table E2. 

Markedly downregulated expression of several class 1 and 2 LCE genes in AD compared with normal skin on U133 Plus 2.0 microarrays.
ProbeSymbolAD vs normalFDR-LS-normal
1559224_atLCE1B−4.411.18E-08
1559226_x_atLCE1E−4.615.55E-09
1560531_atLCE2B−6.394.28E-07
207710_atLCE3D2.68∗3.10E-07

FC obtained by microarray data presented in log2 scale. LS, Lesional.

∗FDR < 0.05. LS, Lesional.

Back to Article Outline

Table E3. 

Summary of previously published array-based studies comparing AD, psoriasis, and normal skin, or combinations of 2 of these conditions
AD vs PsoCriteriaReference
1Array comparison largely AD vs Pso, genes presented by signal intensity and not by fold change, only few significant genes.6 Patients with AD and 7 with Pso; 4 normal used for real-time PCR, only few genes with significant P values. Signal intensity >2 FC, P < .05. Same numbers as the next one.E12
2Array comparison is AD vs Pso, without expression values of each disease compared to normal skin.6 Patients with AD and 7 with Pso. Signal intensity > 2 FC, P < .05. 62 Genes were upregulated in Pso vs AD and 18 genes only found significantly increased in AD vs Pso.E13
3Array comparison of epidermal AD and psoriasis genes, some comparison with normal for selected antimicrobials obtained by real-time PCR and IHC.16 Patients with chronic AD and 20 with Pso; 11 normal used for real-time PCR and IHC only, FC > 2, P < .1. 183 Genes were differentially expressed, but study focused on antimicrobial genes—18 genes.E14
4Sage analysis and PIQOR microarray Genes previously characterized as psoriasis genes (like CXCL9, K16, MX1) were described as “specifically found in LP.”20 Patients with lichen planus, Pso, AD, normal (not clear how many in each group), P < .01, only 56 genes were identified as differentially expressed between AD and Pso, and 60 were found as “unique” to LP.E15
AD vs normalCriteriaReference
1Upregulation of antimicrobials (S100A8, S100A9, elafin, hBD2, and several EDC genes (involucrin), and downregulation of others (filaggrin, loricrin).P < .05, 17 patients with AD and 4 normal. 10 Genes only showed a FC >5.E16
2No comparison to Pso, small patient population.6 AD and 5 normal, FC >2, P < .01. Identified 169 increased in AD vs normal, and 212 decreased vs normal.E17
3Upregulated inflammatory genes and downregulated lipid homeostasis genes in AD. Limited expression of cornification genes on arrays (FLG and LOR not expressed). Differential expression of CDSN and TGM1 and TGM3 in AD vs normal. No comparison to Pso. Small patient population. Also evaluated patch tested skin sensitized to Malassezia sympodialis.7 Patients with AD and 4 normal. 2194 Genes were upregulated in AD vs normal skin. P < .05. Spotted cDNA arrays used. FCH not described.E18
Pso vs normalCriteriaReference
1The second largest array study comparing Pso to normal and identifying major T-cell and dendritic cell genes.16 Patients with Pso and 8 normal, FC >1.2, P < .05. 710 Genes were differentially expressed between Pso and normal.E19
2The largest array study comparing lesional and nonlesional psoriasis and normal skin.26 Patients with Pso and 21 normal, FC >2, P < .05. 706 Genes were upregulated and 1046 were downregulated vs normal skin.E20

Few studies used normal skin as a reference when comparing AD and psoriasis.

IHC, Immunohistochemistry; Pso, psoriasis.

Back to Article Outline

References 

  1. Leung DY, Bieber T. Atopic dermatitis. Lancet. 2003;361:151–160
  2. Wilsmann-Theis D, Hagemann T, Jordan J, Bieber T, Novak N. Facing psoriasis and atopic dermatitis: are there more similarities or more differences?. Eur J Dermatol. 2008;18:172–180
  3. Guttman-Yassky E, Lowes MA, Fuentes-Duculan J, Whynot J, Novitskaya I, Cardinale I, et al. Major differences in inflammatory dendritic cells and their products distinguish atopic dermatitis from psoriasis. J Allergy Clin Immunol. 2007;119:1210–1217
  4. de Jongh GJ, Zeeuwen PL, Kucharekova M, Pfundt R, van der Valk PG, Blokx W, et al. High expression levels of keratinocyte antimicrobial proteins in psoriasis compared with atopic dermatitis. J Invest Dermatol. 2005;125:1163–1173
  5. Fartasch M, Ponec M. Improved barrier structure formation in air-exposed human keratinocyte culture systems. J Invest Dermatol. 1994;102:366–374
  6. Fartasch M. Epidermal barrier in disorders of the skin. Microsc Res Tech. 1997;38:361–372
  7. Elias PM, Hatano Y, Williams ML. Basis for the barrier abnormality in atopic dermatitis: outside-inside-outside pathogenic mechanisms. J Allergy Clin Immunol. 2008;121:1337–1343
  8. Nomura I, Gao B, Boguniewicz M, Darst MA, Travers JB, Leung DY. Distinct patterns of gene expression in the skin lesions of atopic dermatitis and psoriasis: a gene microarray analysis. J Allergy Clin Immunol. 2003;112:1195–1202
  9. Nomura I, Goleva E, Howell MD, Hamid QA, Ong PY, Hall CF, et al. Cytokine milieu of atopic dermatitis, as compared to psoriasis, skin prevents induction of innate immune response genes. J Immunol. 2003;171:3262–3269
  10. Guttman-Yassky E, Lowes MA, Fuentes-Duculan J, Zaba LC, Cardinale I, Nograles KE, et al. Low expression of the IL-23/Th17 pathway in atopic dermatitis compared to psoriasis. J Immunol. 2008;181:7420–7427
  11. Boniface K, Bernard FX, Garcia M, Gurney AL, Lecron JC, Morel F. IL-22 inhibits epidermal differentiation and induces proinflammatory gene expression and migration of human keratinocytes. J Immunol. 2005;174:3695–3702
  12. Nograles KE, Zaba LC, Guttman-Yassky E, Fuentes-Duculan J, Suarez-Farinas M, Cardinale I, et al. Th17 cytokines interleukin (IL)-17 and IL-22 modulate distinct inflammatory and keratinocyte-response pathways. Br J Dermatol. 2008;159:1092–1102
  13. Ma HL, Liang S, Li J, Napierata L, Brown T, Benoit S, et al. IL-22 is required for Th17 cell-mediated pathology in a mouse model of psoriasis-like skin inflammation. J Clin Invest. 2008;118:597–607
  14. Palmer CN, Irvine AD, Terron-Kwiatkowski A, Zhao Y, Liao H, Lee SP, et al. Common loss-of-function variants of the epidermal barrier protein filaggrin are a major predisposing factor for atopic dermatitis. Nat Genet. 2006;38:441–446
  15. O'Regan GM, Sandilands A, McLean WH, Irvine AD. Filaggrin in atopic dermatitis. J Allergy Clin Immunol. 2008;122:689–693
  16. McGrath JA, Uitto J. The filaggrin story: novel insights into skin-barrier function and disease. Trends Mol Med. 2008;14:20–27
  17. Sandilands A, Terron-Kwiatkowski A, Hull PR, O'Regan GM, Clayton TH, Watson RM, et al. Comprehensive analysis of the gene encoding filaggrin uncovers prevalent and rare mutations in ichthyosis vulgaris and atopic eczema. Nat Genet. 2007;39:650–654
  18. Lee JH, Lee SJ, Kim D, Bang D. The effect of wet-wrap dressing on epidermal barrier in patients with atopic dermatitis. J Eur Acad Dermatol Venereol. 2007;21:1360–1368
  19. Elias PM. Barrier repair trumps immunology in the pathogenesis and therapy of atopic dermatitis. Drug Discov Today Dis Mech. 2008;5:e33–e38
  20. Proksch E, Jensen JM, Elias PM. Skin lipids and epidermal differentiation in atopic dermatitis. Clin Dermatol. 2003;21:134–144
  21. de Cid R, Riveira-Munoz E, Zeeuwen PL, Robarge J, Liao W, Dannhauser EN, et al. Deletion of the late cornified envelope LCE3B and LCE3C genes as a susceptibility factor for psoriasis. Nat Genet. 2009;41:211–215
  22. Scharschmidt TC, Segre JA. Modeling atopic dermatitis with increasingly complex mouse models. J Invest Dermatol. 2008;128:1061–1064
  23. Zheng T, Zhu Z. Lessons from murine models of atopic dermatitis. Curr Allergy Asthma Rep. 2005;5:291–297
  24. Gudjonsson JE, Johnston A, Dyson M, Valdimarsson H, Elder JT. Mouse models of psoriasis. J Invest Dermatol. 2007;127:1292–1308
  25. Zenz R, Eferl R, Kenner L, Florin L, Hummerich L, Mehic D, et al. Psoriasis-like skin disease and arthritis caused by inducible epidermal deletion of Jun proteins. Nature. 2005;437:369–375
  26. He R, Oyoshi MK, Jin H, Geha RS. Epicutaneous antigen exposure induces a Th17 response that drives airway inflammation after inhalation challenge. Proc Natl Acad Sci U S A. 2007;104:15817–15822
  27. Kess D, Lindqvist AK, Peters T, Wang H, Zamek J, Nischt R, et al. Identification of susceptibility loci for skin disease in a murine psoriasis model. J Immunol. 2006;177:4612–4619
  28. Zhou X, Krueger JG, Kao MC, Lee E, Du F, Menter A, et al. Novel mechanisms of T-cell and dendritic cell activation revealed by profiling of psoriasis on the 63,100-element oligonucleotide array. Physiol Genomics. 2003;13:69–78
  29. Yao Y, Richman L, Morehouse C, de los Reyes M, Higgs BW, Boutrin A, et al. Type I interferon: potential therapeutic target for psoriasis?. PLoS ONE. 2008;3:e2737
  30. Ghadially R, Brown BE, Sequeira-Martin SM, Feingold KR, Elias PM. The aged epidermal permeability barrier: structural, functional, and lipid biochemical abnormalities in humans and a senescent murine model. J Clin Invest. 1995;95:2281–2290
  31. Hou SY, Mitra AK, White SH, Menon GK, Ghadially R, Elias PM. Membrane structures in normal and essential fatty acid-deficient stratum corneum: characterization by ruthenium tetroxide staining and x-ray diffraction. J Invest Dermatol. 1991;96:215–223
  32. McNutt NS, Crain WR. Quantitative electron microscopic comparison of lymphocyte nuclear contours in mycosis fungoides and in benign infiltrates in skin. Cancer. 1981;47:698–709
  33. Haider AS, Peters SB, Kaporis H, Cardinale I, Fei J, Ott J, et al. Genomic analysis defines a cancer-specific gene expression signature for human squamous cell carcinoma and distinguishes malignant hyperproliferation from benign hyperplasia. J Invest Dermatol. 2006;126:869–881
  34. Jackson B, Tilli CM, Hardman MJ, Avilion AA, MacLeod MC, Ashcroft GS, et al. Late cornified envelope family in differentiating epithelia—response to calcium and ultraviolet irradiation. J Invest Dermatol. 2005;124:1062–1070
  35. Suarez-Farinas M, Pellegrino M, Wittkowski KM, Magnasco MO. Harshlight: a “corrective make-up” program for microarray chips. BMC Bioinformatics. 2005;6:294
  36. Smyth GK. Linear models and empirical bayes methods for assessing differential expression in microarray experiments. Stat Appl Genet Mol Biol. 2004;3:Article3
  37. 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
  38. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–15550
  39. Ong PY, Leung DY. Immune dysregulation in atopic dermatitis. Curr Allergy Asthma Rep. 2006;6:384–389
  40. Mansbridge JN, Knapp AM, Strefling AM. Evidence for an alternative pathway of keratinocyte maturation in psoriasis from an antigen found in psoriatic but not normal epidermis. J Invest Dermatol. 1984;83:296–301
  41. Wenzel J, Peters B, Zahn S, Birth M, Hofmann K, Kusters D, et al. Gene expression profiling of lichen planus reflects CXCL9+-mediated inflammation and distinguishes this disease from atopic dermatitis and psoriasis. J Invest Dermatol. 2008;128:67–78
  42. Plager DA, Leontovich AA, Henke SA, Davis MD, McEvoy MT, Sciallis GF, et al. Early cutaneous gene transcription changes in adult atopic dermatitis and potential clinical implications. Exp Dermatol. 2007;16:28–36
  43. Sugiura H, Ebise H, Tazawa T, Tanaka K, Sugiura Y, Uehara M, et al. Large-scale DNA microarray analysis of atopic skin lesions shows overexpression of an epidermal differentiation gene cluster in the alternative pathway and lack of protective gene expression in the cornified envelope. Br J Dermatol. 2005;152:146–149
  44. Sääf AM, Tengvall-Linder M, Chang HY, Adler AS, Wahlgren CF, Scheynius A, et al. Global expression profiling in atopic eczema reveals reciprocal expression of inflammatory and lipid genes. PLoS One. 2008;3:e4017
  45. Sa SM, Valdez PA, Wu J, Jung K, Zhong F, Hall L, et al. The effects of IL-20 subfamily cytokines on reconstituted human epidermis suggest potential roles in cutaneous innate defense and pathogenic adaptive immunity in psoriasis. J Immunol. 2007;178:2229–2240
  46. Nickoloff BJ, Bonish BK, Marble DJ, Schriedel KA, DiPietro LA, Gordon KB, et al. Lessons learned from psoriatic plaques concerning mechanisms of tissue repair, remodeling, and inflammation. J Investig Dermatol Symp Proc. 2006;11:16–29
  47. Zaba LC, Cardinale I, Gilleaudeau P, Sullivan-Whalen M, Suarez-Farinas M, Fuentes-Duculan J, et al. Amelioration of epidermal hyperplasia by TNF inhibition is associated with reduced Th17 responses. J Exp Med. 2007;204:3183–3194
  48. Nograles KE, Zaba LC, Shemer A, Fuentes-Duculan J, Cardinale I, Kikuchi T, et al. IL-22-producing “T22” T cells account for upregulated IL-22 in atopic dermatitis despite reduced IL-17-producing TH17 T cells. J Allergy Clin Immunol. 2009;123:1244–1252
  49. Proksch E, Brandner JM, Jensen JM. The skin: an indispensable barrier. Exp Dermatol. 2008;17:1063–1072
  50. Madison KC. Barrier function of the skin: “la raison d'etre” of the epidermis. J Invest Dermatol. 2003;121:231–241
  51. Elias PM, Steinhoff M. “Outside-to-inside” (and now back to “outside”) pathogenic mechanisms in atopic dermatitis. J Invest Dermatol. 2008;128:1067–1070
  52. Nair RP, Stuart PE, Nistor I, Hiremagalore R, Chia NV, Jenisch S, et al. Sequence and haplotype analysis supports HLA-C as the psoriasis susceptibility 1 gene. Am J Hum Genet. 2006;78:827–851
  53. Bowcock AM, Cookson WO. The genetics of psoriasis, psoriatic arthritis and atopic dermatitis. Hum Mol Genet. 2004;13 Spec No 1:R43–R55
  54. Capon F, Semprini S, Chimenti S, Fabrizi G, Zambruno G, Murgia S, et al. Fine mapping of the PSORS4 psoriasis susceptibility region on chromosome 1q21. J Invest Dermatol. 2001;116:728–730
  55. Giardina E, Sinibaldi C, Chini L, Moschese V, Marulli G, Provini A, et al. Co-localization of susceptibility loci for psoriasis (PSORS4) and atopic dermatitis (ATOD2) on human chromosome 1q21. Hum Hered. 2006;61:229–236
  56. Giardina E, Capon F, De Rosa MC, Mango R, Zambruno G, Orecchia A, et al. Characterization of the loricrin (LOR) gene as a positional candidate for the PSORS4 psoriasis susceptibility locus. Ann Hum Genet. 2004;68:639–645
  57. Semprini S, Capon F, Tacconelli A, Giardina E, Orecchia A, Mingarelli R, et al. Evidence for differential S100 gene over-expression in psoriatic patients from genetically heterogeneous pedigrees. Hum Genet. 2002;111:310–313
  58. Semprini S, Capon F, Bovolenta S, Bruscia E, Pizzuti A, Fabrizi G, et al. Genomic structure, promoter characterisation and mutational analysis of the S100A7 gene: exclusion of a candidate for familial psoriasis susceptibility. Hum Genet. 1999;104:130–134
  59. Sun C, Mathur P, Dupuis J, Tizard R, Ticho B, Crowell T, et al. Peptidoglycan recognition proteins Pglyrp3 and Pglyrp4 are encoded from the epidermal differentiation complex and are candidate genes for the Psors4 locus on chromosome 1q21. Hum Genet. 2006;119:113–125
  60. Kim BE, Leung DY, Boguniewicz M, Howell MD. Loricrin and involucrin expression is down-regulated by Th2 cytokines through STAT-6. Clin Immunol. 2008;126:332–337
  61. Howell MD, Fairchild HR, Kim BE, Bin L, Boguniewicz M, Redzic JS, et al. Th2 cytokines act on S100/A11 to downregulate keratinocyte differentiation. J Invest Dermatol. 2008;128:2248–2258
  62. Klechevsky E, Morita R, Liu M, Cao Y, Coquery S, Thompson-Snipes L, et al. Functional specializations of human epidermal Langerhans cells and CD14+ dermal dendritic cells. Immunity. 2008;29:497–510
  63. Fallon PG, Sasaki T, Sandilands A, Campbell LE, Saunders SP, Mangan NE, et al. A homozygous frameshift mutation in the mouse Flg gene facilitates enhanced percutaneous allergen priming. Nat Genet. 2009;41:602–608
  64. Matsumoto M, Zhou Y, Matsuo S, Nakanishi H, Hirose K, Oura H, et al. Targeted deletion of the murine corneodesmosin gene delineates its essential role in skin and hair physiology. Proc Natl Acad Sci U S A. 2008;105:6720–6724
  65. Haider AS, Duculan J, Whynot JA, Krueger JG. Increased JunB mRNA and protein expression in psoriasis vulgaris lesions. J Invest Dermatol. 2006;126:912–914

Back to Article Outline

References 

  1. Guttman-Yassky E, Lowes MA, Fuentes-Duculan J, Whynot J, Novitskaya I, Cardinale I, et al. Major differences in inflammatory dendritic cells and their products distinguish atopic dermatitis from psoriasis. J Allergy Clin Immunol. 2007;119:1210–1217
  2. Ghadially R, Brown BE, Sequeira-Martin SM, Feingold KR, Elias PM. The aged epidermal permeability barrier: structural, functional, and lipid biochemical abnormalities in humans and a senescent murine model. J Clin Invest. 1995;95:2281–2290
  3. Hou SY, Mitra AK, White SH, Menon GK, Ghadially R, Elias PM. Membrane structures in normal and essential fatty acid-deficient stratum corneum: characterization by ruthenium tetroxide staining and x-ray diffraction. J Invest Dermatol. 1991;96:215–223
  4. McNutt NS, Crain WR. Quantitative electron microscopic comparison of lymphocyte nuclear contours in mycosis fungoides and in benign infiltrates in skin. Cancer. 1981;47:698–709
  5. Haider AS, Peters SB, Kaporis H, Cardinale I, Fei J, Ott J, et al. Genomic analysis defines a cancer-specific gene expression signature for human squamous cell carcinoma and distinguishes malignant hyperproliferation from benign hyperplasia. J Invest Dermatol. 2006;126:869–881
  6. Guttman-Yassky E, Lowes MA, Fuentes-Duculan J, Zaba LC, Cardinale I, Nograles KE, et al. Low expression of the IL-23/Th17 pathway in atopic dermatitis compared to psoriasis. J Immunol. 2008;181:7420–7427
  7. Jackson B, Tilli CM, Hardman MJ, Avilion AA, MacLeod MC, Ashcroft GS, et al. Late cornified envelope family in differentiating epithelia—response to calcium and ultraviolet irradiation. J Invest Dermatol. 2005;124:1062–1070
  8. Suarez-Farinas M, Pellegrino M, Wittkowski KM, Magnasco MO. Harshlight: a “corrective make-up” program for microarray chips. BMC Bioinformatics. 2005;6:294
  9. Smyth GK. Limma: linear models for microarray data. In: Gentleman R VC, Huber W, Irizarry R, Dudoit S, editors. Bioinformatics and computational biology solutions using R and Bioconductor. New York: Springer; 2005. p. 397–420
  10. Tibshirani RJ, Efron B. Pre-validation and inference in microarrays. Stat Appl Genet Mol Biol. 2002;1:Article1
  11. Subramanian A, Tamayo P, Mootha VK, Mukherjee S, Ebert BL, Gillette MA, et al. Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A. 2005;102:15545–15550
  12. Nomura I, Goleva E, Howell MD, Hamid QA, Ong PY, Hall CF, et al. Cytokine milieu of atopic dermatitis, as compared to psoriasis, skin prevents induction of innate immune response genes. J Immunol. 2003;171:3262–3269
  13. Nomura I, Gao B, Boguniewicz M, Darst MA, Travers JB, Leung DY. Distinct patterns of gene expression in the skin lesions of atopic dermatitis and psoriasis: a gene microarray analysis. J Allergy Clin Immunol. 2003;112:1195–1202
  14. de Jongh GJ, Zeeuwen PL, Kucharekova M, Pfundt R, van der Valk PG, Blokx W, et al. High expression levels of keratinocyte antimicrobial proteins in psoriasis compared with atopic dermatitis. J Invest Dermatol. 2005;125:1163–1173
  15. Wenzel J, Peters B, Zahn S, Birth M, Hofmann K, Kusters D, et al. Gene expression profiling of lichen planus reflects CXCL9+-mediated inflammation and distinguishes this disease from atopic dermatitis and psoriasis. J Invest Dermatol. 2008;128:67–78
  16. Sugiura H, Ebise H, Tazawa T, Tanaka K, Sugiura Y, Uehara M, et al. Large-scale DNA microarray analysis of atopic skin lesions shows overexpression of an epidermal differentiation gene cluster in the alternative pathway and lack of protective gene expression in the cornified envelope. Br J Dermatol. 2005;152:146–149
  17. Plager DA, Leontovich AA, Henke SA, Davis MD, McEvoy MT, Sciallis GF, et al. Early cutaneous gene transcription changes in adult atopic dermatitis and potential clinical implications. Exp Dermatol. 2007;16:28–36
  18. Sääf AM, Tengvall-Linder M, Chang HY, Adler AS, Wahlgren CF, Scheynius A, et al. Global expression profiling in atopic eczema reveals reciprocal expression of inflammatory and lipid genes. PLoS One. 2008;3:e4017
  19. Zhou X, Krueger JG, Kao MC, Lee E, Du F, Menter A, et al. Novel mechanisms of T-cell and dendritic cell activation revealed by profiling of psoriasis on the 63,100-element oligonucleotide array. Physiol Genomics. 2003;13:69–78
  20. Yao Y, Richman L, Morehouse C, et al. PLoS One 2008 Jul 16;3(7):e2737.

 Supported by grant number 5UL1RR024143-02 from the National Center for Research Resources, a component of the National Institutes of Health, and the National Institutes of Health Roadmap for Medical Research.

 Disclosure of potential conflict of interest: J. G. Krueger has received research support from Centocor, Amgen, and Wyeth. The rest of the authors have declared that they have no conflict of interest.

PII: S0091-6749(09)01432-8

doi:10.1016/j.jaci.2009.09.031

The Journal of Allergy and Clinical Immunology
Volume 124, Issue 6 , Pages 1235-1244.e58, December 2009

Article Tools

* Image Usage & Resolution