Advertisement

Baricitinib blocks cytokine mediated downregulation of PAD1 in human keratinocytes: a possible molecular link to the effects of JAK-inhibitors in atopic dermatitis

Open AccessPublished:January 10, 2023DOI:https://doi.org/10.1016/j.jid.2022.12.012

      Abbreviations:

      AD (atopic dermatitis), FLG (filaggrin gene), JAK (Janus Kinase), KC (keratinocyte), NHEK (normal human epidermal keratinocyte), PAD (peptidylarginine deiminase)
      TO THE EDITOR
      Atopic Dermatitis (AD) is one of the most common chronic inflammatory skin conditions in the developed world. It has a complex aetiology with genetic, immunological and environmental factors causing skin barrier disruption and immune dysregulation (
      • Elias P.M.
      • Steinhoff M.
      Outside-to-inside" (and now back to "outside") pathogenic mechanisms in atopic dermatitis.
      ). Available immune modulating treatments for AD have proven the importance of Th2 immune activation/dysregulation in AD pathogenesis. In contrast, loss-of-function mutations in the gene encoding filaggrin (FLG) is the strongest genetic pre-disposing factor for development of AD, emphasizing the involvement of skin barrier dysregulation in AD pathogenesis. However, FLG mutations are neither crucial nor sufficient to cause AD and additional factors involved in skin barrier dysregulation have been proposed.
      The process of epidermal differentiation is characterized by a series of molecular changes and involves post-translational modifications (PTMs) of proteins. One of them is protein deimination i.e., conversion of peptidyl-arginine into peptidyl-citrulline by peptidylarginine deiminases (PADs). Human epidermis expresses three isotypes of PAD enzymes, PAD1, PAD2, and PAD3, each being localized in a specific layer of skin (
      • Nachat R.
      • Mechin M.C.
      • Takahara H.
      • Chavanas S.
      • Charveron M.
      • Serre G.
      • et al.
      Peptidylarginine deiminase isoforms 1-3 are expressed in the epidermis and involved in the deimination of K1 and filaggrin.
      ). Even though deimination is critical for skin homeostasis the precise role of PADs in skin diseases is poorly studied. We recently showed decreased levels of PAD1 in lesional psoriatic skin (
      • Padhi A.
      • Srivastava A.
      • Ramesh A.
      • Ehrstrom M.
      • Simon M.
      • Sonkoly E.
      • et al.
      IL-22 Downregulates Peptidylarginine Deiminase-1 in Human Keratinocytes: Adding Another Piece to the IL-22 Puzzle in Epidermal Barrier Formation.
      ), while others have shown the prevalence of PADI3 mutations in Central Centrifugal Cicatricial Alopecia patients (
      • Malki L.
      • Sarig O.
      • Romano M.T.
      • Mechin M.C.
      • Peled A.
      • Pavlovsky M.
      • et al.
      Variant PADI3 in Central Centrifugal Cicatricial Alopecia.
      ). One previous study in AD, reported decreased overall deimination and more specifically of filaggrin in lesional skin, but did not explore the expression of PADs (
      • Winget J.M.
      • Finlay D.
      • Mills K.J.
      • Huggins T.
      • Bascom C.
      • Isfort R.J.
      • et al.
      Quantitative Proteomic Analysis of Stratum Corneum Dysfunction in Adult Chronic Atopic Dermatitis.
      ). Hence, we sought to determine the expression of PADs in biopsies from AD patients (Figure S1a, Table S1 and Supplementary Text). All tissue samples were collected according to the principles of the Declaration of Helsinki and approved by the institutional review board of the ethics committee of the Karolinska Institutet, the regional ethics committee of Stockholm. Written, informed consent was obtained from all patients and healthy controls before enrolment.
      First, we stained skin sections with anti-PAD1,2,3 antibody and found a significant decrease in PAD1 expression in lesional AD skin as compared to skin from healthy individuals (Figure 1a). In contrast to our findings in psoriasis we observed a pattern of decreased PAD1 expression in non-lesional AD skin, though statistically non-significant (Figure 1b, p = 0.0685). This could be because the non-lesional skin in AD also presents with some epidermal defects (

      Suarez-Farinas M, Tintle SJ, Shemer A, Chiricozzi A, Nograles K, Cardinale I, et al. Nonlesional atopic dermatitis skin is characterized by broad terminal differentiation defects and variable immune abnormalities. J Allergy Clin Immunol 2011;127(4):954-964 e1-4.

      ). These findings are consistent with RNA-seq tape strip profiling of AD skin, which showed a similar reduction in PADI1 expression in lesional skin relative to healthy skin (
      • He H.
      • Bissonnette R.
      • Wu J.
      • Diaz A.
      • Saint-Cyr Proulx E.
      • Maari C.
      • et al.
      Tape strips detect distinct immune and barrier profiles in atopic dermatitis and psoriasis.
      ). However, we found no significant differences in PAD2 and PAD3 expression between the groups, which is consistent with our findings in psoriasis (Figure S1b, S1c and S1d). Of the four known PAD targets in epidermis, deimination of filaggrin is the most important for maintaining epidermal barrier function. Since we observed a decrease in PAD1, it was important to assess the deimination status of filaggrin. However, lesional AD skin is associated with decreased filaggrin expression (

      Howell MD, Kim BE, Gao P, Grant AV, Boguniewicz M, Debenedetto A, et al. Cytokine modulation of atopic dermatitis filaggrin skin expression. J Allergy Clin Immunol 2007;120(1):150-155.6

      ). Thus, it was necessary to determine the levels of deiminated filaggrin relative to its non-deiminated form within different groups. Using antibodies specific against non-deiminated and deiminated filaggrin, we found significant reduction in deimination of filaggrin in lesional as compared to non-lesional AD and healthy skin (Figure 1c and 1d). Next, we stratified included patients based on EASI (Eczema Area and Severity Index) score (mild, moderate, and severe) and FLG mutation (WT and ΔFLG). The observed decrease in PAD1 expression was independent of the degree of severity (Figure 1e) and FLG status (Figure S1e). Skin sections from all groups stained negative for the isotype control antibody (Figure S1f). Furthermore, using Spearman’s correlation we observed a linear association between PAD1 levels and filaggrin deimination, suggesting that the reduction in filaggrin deimination in lesional AD skin could be attributed to a decrease in PAD1 levels (Figure 1f). However further mechanistic studies are required to confirm this. A few of the enrolled patients were receiving Methotrexate (MTX) at the time of inclusion, but this didn’t constitute an exclusion criterion, since we observed that MTX treatment did not affect the expression of PADI1 in keratinocytes (Figure S1g).
      Figure thumbnail gr1
      Figure 1Decreased PAD1 expression and filaggrin deimination in lesional AD skin. (a) Representative images of chromogenic staining pattern for PAD1. Brown DAB staining shows the presence of PADs (black arrowheads). Scale bar = 100μm. (b) Quantification of DAB staining of sections of skin from healthy controls (n=10), and lesional (n=27) and non lesional skin (n=25) of AD patients with each dot on the graph representing individual patient. Because the stainings did not met the QuPath analysis threshold, non-lesional skin of 2 individuals was not represented. (c) Representative images of fluorescent staining patterns for non-deiminated and deiminated filaggrin. White arrowheads show absence of deiminated filaggrin. Dashed line indicates border between dermis and epidermis. Scale bar = 100μm. (d) Quantification of filaggrin staining using Image J to determine the ratio between deiminated and non-deiminated form in sections of skin from healthy controls (n=8), and lesional (n=27) or non lesional skin (n=20) of AD patients with each dot on the graph representing an individual patient. Stainings with high background on no signal were excluded from representation. (e) PAD1 staining quantification in patients subgrouped according to mild (n=10), moderate (n=8) or severe (n=9) AD as compared to healthy controls (n=10), with each dot on the graph representing an individual patient. (f) Correlation between the number of PAD1+ cells and the ratio of deiminated/non-deiminated filaggrin. Spearman’s correlation between the two parameters across all included samples is shown, with R representing correlation coefficient. 95% confidence interval is indicated in grey. All comparisons were performed using Kruskal−Wallis test and two-stage Benjamini, Krieger, and Yekutieli test. Bars represent median. *P < 0.05; **P < 0.01; ****P < 0.0001. AD, atopic dermatitis; DAB, 3,3′-Diaminobenzidine; FLG, filaggrin gene; ns, nonsignificant; PAD1, Peptidyl arginine deiminase 1.
      In several skin disorders including AD, the inflammatory environment regulates barrier dysfunction. AD skin is characterized by the overexpression of Th2 (IL-4, IL-13, IL-5, IL-31) cytokines and IL-22, which are known to mediate inhibition of keratinocyte differentiation (
      • Mitamura Y.
      • Nunomura S.
      • Nanri Y.
      • Ogawa M.
      • Yoshihara T.
      • Masuoka M.
      • et al.
      The IL-13/periostin/IL-24 pathway causes epidermal barrier dysfunction in allergic skin inflammation.
      ). Thus, we assessed the effects of AD relevant cytokines on PADI1 expression using differentiated normal human epidermal keratinocytes (NHEKs) (Figure 2a). Amongst the Th2 cytokines, IL-4 alone or in combination with IL-13 decreased the expression of PADI1 in keratinocytes (Figure 2b). In line with our previous data, IL-22 treatment caused a reduction in PADI1 levels. To mimic the cytokine milieu in AD skin, we stimulated keratinocytes with a combination of IL-4, IL-13, and IL-22, which resulted in further decrease in PADI1 levels as compared to IL-4 and IL-13 together (Figure 2b). Treating keratinocytes with increasing concentration of cytokines did not further lower the PADI1 levels (Figure 2c). Next, we determined whether we could block this effect. Recently Janus Kinase (JAK) inhibitors have been introduced as the next generation of targeted AD therapy (
      • Chovatiya R.
      • Paller A.S.
      JAK inhibitors in the treatment of atopic dermatitis.
      ). To this end, keratinocytes were pre-treated with Baricitinib (JAK1/2) or vehicle control dimethylsulfoxide (DMSO) followed by treatment with cytokines. Notably, Baricitinib was found to induce PAD1I expression (Figure S1h) and pre-treatment of keratinocytes with baricitinib inhibited the dramatic decrease in PADI1 expression caused by the cytokine combination of IL-4, IL-13, and IL-22 (Figure 2d). The JAK signalling pathway is activated by these cytokines to propagate inflammation in keratinocytes (
      • Bao L.
      • Zhang H.
      • Chan L.S.
      The involvement of the JAK-STAT signaling pathway in chronic inflammatory skin disease atopic dermatitis.
      ,
      • Welsch K.
      • Holstein J.
      • Laurence A.
      • Ghoreschi K.
      Targeting JAK/STAT signalling in inflammatory skin diseases with small molecule inhibitors.
      ).
      Figure thumbnail gr2
      Figure 2Pre-treatment with Baricitinib blocks the inhibitory effect of Th2 cytokines and IL-22 on PADI1 expression in keratinocytes. NHEKs isolated from skin of three healthy donors were cultured and differentiated in the presence of calcium chloride for 72 hours. The cells were pre-treated with Baricitinib (20μM) or DMSO equivalent (0.144%) to the volume used to dissolve Baricitinib, followed by cytokine stimulation (20ng/ml) for 72 hours. (a) Experimental layout. (b-d) Total RNA was isolated, converted to cDNA and relative expression of PADI1 was determined by qRT-PCR in keratinocytes stimulated with different cytokines alone or in combination (b), increasing concentrations (c) and in the absence or presence of Baricitinib (d). Dashed line represents PADI1 level in NHEKs treated with DMSO control. All comparisons were performed using ordinary one-way analysis of variance (ANOVA) with Holm-Šidák multiple comparison test. Data is represented as mean±SD (n=3). *P < 0.05; **P < 0.01; ***P < 0.001. DMSO, dimethyl sulfoxide; NHEK, normal human epidermal keratinocytes; ns, nonsignificant; PAD, peptidyl arginine deiminase; qRT-PCR, quantitative real-time PCR; μM, micromolar.
      These results show that Th2 cytokines suppress PAD1 expression in human epidermal keratinocytes and that baricitinib can block this effect, which could have therapeutic implications. We found that the levels of PAD1 and filaggrin deimination were reduced in AD skin, similar to our findings in psoriasis, suggesting impaired deimination in the event of skin inflammation. However, further mechanistic studies are necessary to better understand the role of PADs and protein deimination in pathogenesis of AD.

      ACKNOWLEDGMENTS

      We express our gratitude to all the patients and healthy donors who took part in the study. We would like to thank Dr Michel Simon (Toulouse Institute for Infectious and Inflammatory Diseases, University of Toulouse, Toulouse, France) for providing antibodies against native and deiminated filaggrin. We thank Prof. Liv Eidsmo (Karolinska Institutet, Sweden and Leo Foundation Skin Immunology Research Center, University of Copenhagen, Denmark) for her technical support. We are grateful to the research nurses Maria Lundqvist, Helena Griehsel and Inger Edén at Department of Dermatology, Karolinska University Hospital for clinical sample collection. This work was supported by Hudfonden, Psoriasisfonden, Gösta A Karlssons 60-års fond, Magnus Bergvalls stiftelse, and Åke Wibergs stiftelse. JL was supported by Region Stockholm (clinical postdoctoral appointment). The graphical abstracts were created with BioRender.com.

      SUPPLEMENTARY METHODS

      Patients and biopsies
      Punch biopsies (4mm) were taken from active infiltrated plaques (lesional) and non lesional skin at least 5 cm from active plaque without obvious active Atopic Dermatitis (AD) at the Department of Dermatology, Karolinska University Hospital (Stockholm, Sweden). 8 out of the 28 patients at inclusion had been on systemic treatment with MTX, whilst others were not forced to stop taking topical treatments with moisturizer or glucocorticoids. Patients were subgrouped into either FLG wildtype (WT) or FLG heterozygous (ΔFLG) determined by genotype data. All enrolled patients were examined by a dermatologist, and Eczema Area and Severity Index (EASI) score was determined at the time of biopsy collection. Patients with a EASI score ranging between 0-7 were considered having mild disease, 7-21 as moderate and more than 21 as severe. Punch biopsies (6mm) were obtained from healthy individuals (n=10) undergoing abdominoplasty through the plastic surgery department and cryopreserved at -800C until analysis.
      FLG Genotyping
      Peripheral blood samples were collected, and genomic DNA was extracted by standard procedures. We genotyped the three most common filaggrin gene (FLG) mutations in the European population (R501X, R2447X and 2282Del4) (Sandilands et al., 2007). Real-Time PCR was performed in QuantStudio 7 (ThermoFisher Scientific) using allele specific Taqman MGB probes labelled with fluorescent dyes FAM and VIC (Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s protocols. Allelic discrimination was made with the QuantStudioTM Real-Time PCR Software (Applied Biosystems).
      Immunohistochemistry
      Stored biopsies were sectioned (7μm) at Biomedicum histological core facility (Histocore, Karolinska Institutet, Stockholm, Sweden) using Cryostar NX70 (Thermo Fisher Scientific, Waltham, MA, USA). The sections were fixed in precooled acetone (Sigma Aldrich, St Louis, MO, USA) and washed three times with filtered PBS. The periphery of individual sections was marked with a PAP pen (Vector Laboratories, Burlingame, CA, USA). Blocking steps included blocking endogenous peroxidase activity with BLOXALL Endogenous Blocking solution (Vector Laboratories), protein blocking with 10% goat serum in 1% BSA and biotin blocking with Avidin/Biotin blocking kit (Vector Laboratories). Sections were incubated with anti-PAD1 (Sigma, HPA062294), anti-PAD2 (Proteintech, 12110-1-AP), anti-PAD3(Novus, NBP1-92240) or normal goat IgG isotype control (R&D systems, Ab-108-C) antibody overnight at 40C and washed thrice with PBS followed by addition of biotinylated secondary antibody (goat anti-rabbit IgG antibody (H+L), Vector laboratories). Next, the sections were stained with Vectastain ABC HRP kit-peroxidase reagent (Vector laboratories), followed by PBS wash and addition of ImmPACT DAB substrate (Vector laboratories). The sections were stained with a Hematoxylin nuclear counterstain (Vector laboratories), washed with running water, and mounted using Vectamount AQ mounting medium (Vector laboratories). Stained sections were visualized and imaged with a Leica DMRXA light microscope (Leica microsystems, Wetzlar, Germany). Images were analysed for appearance of brown DAB precipitates and quantified using an open source bioimage analysis software QuPath (Bankhead et al., 2017). Firstly, training annotations were created to aid in the classification of positive and negative cells. Next borders of the regions of interest (ROI) were demarcated with an expansion of the epidermal ROI. Using the colour deconvolution feature, the stains were separated digitally, and an appropriate threshold was set for the DAB channel. All cells with clear haematoxylin and DAB staining were marked as positive and cell count was determined. The cell count was normalized to the epidermal area to determine the number positive cells per μm2 tissue.
      Immunofluorescence
      Sectioned biopsies (7μm) from healthy individuals and AD patients were allowed to come to room temperature and fixed with pre-cooled acetone. The sections were blocked in a sequential manner with Image-iT™ FX Signal Enhancer (ThermoFisher Scientific), Background Buster (Innovex Biosciences, Richmond, USA), and 10% goat serum in 1% BSA with intermittent washing in PBS supplemented with 0.1% saponin (Sigma Aldrich) to reduce nonspecific dyes to tissues and quenching background fluorescence. The sections were incubated with anti-filaggrin (AHF3, FOVA711C) or anti-deiminated filaggrin (AHF7, FOVA311C) primary antibodies overnight at 40C (Simon et al., 1995). A second blocking step with 1% goat serum in PBS with 0.1% saponin for 30mins was performed. Following this, the sections were stained with goat anti-mouse IgG (H+L) Alexa Flour 594 (ThermoFisher Scientific) secondary antibody for 40mins at room temperature in dark. The sections were mounted on slides with an antifade reagent containing DAPI (ThermoFisher Scientific) to stain nucleus and imaged under a Nikon A1R confocal microscope (Nikon, Tokyon, Japan). Quantification of images was done using Image J software as described previously (
      • Padhi A.
      • Srivastava A.
      • Ramesh A.
      • Ehrstrom M.
      • Simon M.
      • Sonkoly E.
      • et al.
      IL-22 Downregulates Peptidylarginine Deiminase-1 in Human Keratinocytes: Adding Another Piece to the IL-22 Puzzle in Epidermal Barrier Formation.
      ).
      Keratinocyte culture and stimulations
      NHEKs isolated from three healthy adult donors (ThermoFisher Scientific) were cultured in Epilife medium supplemented with human keratinocyte growth supplement. The cells were trypsinized and seeded on tissue culture plates (Sarstedt, Helsingborg, Sweden) in basal medium without any supplements. The cells were allowed to differentiate in the presence of 1.5mM calcium chloride (Sigma Aldrich) for 3days. Differentiated cells were pre-treated with 5μM, 10μM or 20μM Baricitinib (MedChem express, NJ, USA) or equivalent amount of DMSO (Sigma Aldrich), equivalent to the concentration used to dissolve the highest dose of Baricitinib for 24hours, followed by cytokine (20ng/ml) (IL-5, IL-31, IL-4, IL-13 and IL-22) (R&D Systems) stimulations for 72hours.
      RNA extraction and quantitative real time PCR
      Total RNA from cultured cells was isolated using Bioline isolate II RNA kit (Meridian Bioscience, Cincinatti, OH, USA). RNA was quantified using a nanodrop spectrophotometer (Thermo Fisher Scientific) and 1μg of RNA from each condition was converted to cDNA using Bioline SensiFAST cDNA synthesis kit (Meridian Biosceince). The diluted cDNA (1:2) was used as a template, and gene expression was determined with PADI1 specific primers (FP- 5` TGCAGCTGTCCCTGAAGATG 3`, RP- 5` CACCCTTGGGCACATCACTGT 3`). All reactions were performed in volume of 20μl with SYBR green master mix (ThermoFisher Scientific) in a CFX96 PCR instrument (BioRad, Hercules, CA, USA). Data presented for each sample was relative to expression of 18s housekeeping gene calculated using 2-ΔΔcycle threshold method. Gene expression of all control samples was normalized to 1.
      Statistical analysis
      Statistical analysis was performed using PRISM (version 9, GraphPad software, La Jolla, CA, USA). The tests used for individual experiments are mentioned in respective figure legends.
      SUPPLEMENTARY REFERENCES
      Bankhead P, Loughrey MB, Fernandez JA, Dombrowski Y, McArt DG, Dunne PD, et al. QuPath: Open source software for digital pathology image analysis. Sci Rep 2017;7(1):16878.
      Padhi A, Srivastava A, Ramesh A, Ehrstrom M, Simon M, Sonkoly E, et al. IL-22 Downregulates Peptidylarginine Deiminase-1 in Human Keratinocytes: Adding Another Piece to the IL-22 Puzzle in Epidermal Barrier Formation. J Invest Dermatol 2022;142(2):333-42 e6.
      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(5):650-4.
      Simon M, Sebbag M, Haftek M, Vincent C, Girbal-Neuhauser E, Rakotoarivony J, et al. Monoclonal antibodies to human epidermal filaggrin, some not recognizing profilaggrin. J Invest Dermatol 1995;105(3):432-7.
      SUPPLEMENTARY FIGURE LEGEND
      Figure thumbnail fx1
      Supplementary Figure S1PAD2 and PAD3 expression in lesional AD skin are unaltered, and effect of Methotrexate and Baricitinib on expression of PADI1 expression in NHEKs. (a) Experimental workflow of the study. Biopsies were obtained from healthy individuals and lesional or non-lesional skin of AD patients and sectioned for IHC and immunofluorescence. (b) Representative images of chromogenic staining patterns for PAD2 and PAD3. Brown DAB staining shows the presence of PADs (black arrowheads) Scale bar = 100μm. Quantification of DAB staining of sections for PAD2 (c) and PAD3 (d). (e) PAD1 staining quantification in patients subgrouped according to FLG status as compared to healthy controls. Cryosections from healthy, AD non-lesional and AD lesional were stained with normal goat IgG isotype control and a corresponding biotinylated secondary antibody, followed by DAB detection. (f) Representative images for isotype control staining are shown here. Scale bar=50μm. (g) NHEKs isolated from three healthy donors were cultured and differentiated in presence of the calcium chloride for 72 hours. The cells were treated with increasing doses of Methotrexate (MTX) or DMSO equivalent to the volume used to dissolve the highest dose of MTX. Total RNA was isolated, converted to cDNA and relative expression of PADI1 was determined by qRT-PCR. (h) Differentiated NHEKs were treated with increasing doses of Baricitinib (5μM, 10μM or 20μM) or equivalent volumes of DMSO for 24hours and PADI1 expression was determined by qRT-PCR. All comparisons were performed using Kruskall−Wallis test and two-stage Benjamini, Krieger, and Yekutieli test. Bars represent median. *P<0.05; **P<0.01; ****P<0.0001. AD, atopic dermatitis; DAB, 3,3′-Diaminobenzidine; FLG, Filaggrin; PAD, peptidylarginine deiminase; ns, nonsignificant.
      SUPPLEMENTARY TABLE S1PATIENT INFORMATION
      Patient numberFilaggrin StatusEASISexAge at Biopsy collectionTreatment
      P1WT3M37MTX
      P2ΔFLG5M59MTX
      P3ΔFLG24M46MTX
      P4WT5F49MTX
      P5WT36M22No systemic tx
      P6ΔFLG1M48MTX
      P7WT30M81No systemic tx
      P8WT42M54No systemic tx
      P9WT7F42No systemic tx
      P10WT5M37MTX
      P11ΔFLG4M28MTX
      P12WT4M51MTX
      P13WT8M26No systemic tx
      P14WT46M35No systemic tx
      P15ΔFLG22M45No systemic tx
      P16WT16M64No systemic tx
      P17WT9M58No systemic tx
      P18WT17M23No systemic tx
      P19WT11F55No systemic tx
      P20WT13F55No systemic tx
      P21WT8F57No systemic tx
      P22WT15F64No systemic tx
      P23ΔFLG3F36No systemic tx
      P24ΔFLG24M24No systemic tx
      P25WT31F71No systemic tx
      P26WT37,4F26No systemic tx
      P27ΔFLG<7F27No systemic tx
      P28WT31M84No systemic tx
      Abbreviations: WT, Wild type; ΔFLG, Heterozygous for FLG; F, Female; M, Male; tx, treatment.

      REFERENCES

        • Bao L.
        • Zhang H.
        • Chan L.S.
        The involvement of the JAK-STAT signaling pathway in chronic inflammatory skin disease atopic dermatitis.
        JAKSTAT. 2013; 2e24137
        • Chovatiya R.
        • Paller A.S.
        JAK inhibitors in the treatment of atopic dermatitis.
        J Allergy Clin Immunol. 2021; 148: 927-940
        • Elias P.M.
        • Steinhoff M.
        Outside-to-inside" (and now back to "outside") pathogenic mechanisms in atopic dermatitis.
        J Invest Dermatol. 2008; 128: 1067-1070
        • He H.
        • Bissonnette R.
        • Wu J.
        • Diaz A.
        • Saint-Cyr Proulx E.
        • Maari C.
        • et al.
        Tape strips detect distinct immune and barrier profiles in atopic dermatitis and psoriasis.
        J Allergy Clin Immunol. 2021; 147: 199-212
      1. Howell MD, Kim BE, Gao P, Grant AV, Boguniewicz M, Debenedetto A, et al. Cytokine modulation of atopic dermatitis filaggrin skin expression. J Allergy Clin Immunol 2007;120(1):150-155.6

        • Malki L.
        • Sarig O.
        • Romano M.T.
        • Mechin M.C.
        • Peled A.
        • Pavlovsky M.
        • et al.
        Variant PADI3 in Central Centrifugal Cicatricial Alopecia.
        N Engl J Med. 2019; 380: 833-841
        • Mitamura Y.
        • Nunomura S.
        • Nanri Y.
        • Ogawa M.
        • Yoshihara T.
        • Masuoka M.
        • et al.
        The IL-13/periostin/IL-24 pathway causes epidermal barrier dysfunction in allergic skin inflammation.
        Allergy. 2018; 73: 1881-1891
        • Nachat R.
        • Mechin M.C.
        • Takahara H.
        • Chavanas S.
        • Charveron M.
        • Serre G.
        • et al.
        Peptidylarginine deiminase isoforms 1-3 are expressed in the epidermis and involved in the deimination of K1 and filaggrin.
        J Invest Dermatol. 2005; 124: 384-393
        • Padhi A.
        • Srivastava A.
        • Ramesh A.
        • Ehrstrom M.
        • Simon M.
        • Sonkoly E.
        • et al.
        IL-22 Downregulates Peptidylarginine Deiminase-1 in Human Keratinocytes: Adding Another Piece to the IL-22 Puzzle in Epidermal Barrier Formation.
        J Invest Dermatol. 2022; 142: 333-342 e6
      2. Suarez-Farinas M, Tintle SJ, Shemer A, Chiricozzi A, Nograles K, Cardinale I, et al. Nonlesional atopic dermatitis skin is characterized by broad terminal differentiation defects and variable immune abnormalities. J Allergy Clin Immunol 2011;127(4):954-964 e1-4.

        • Welsch K.
        • Holstein J.
        • Laurence A.
        • Ghoreschi K.
        Targeting JAK/STAT signalling in inflammatory skin diseases with small molecule inhibitors.
        Eur J Immunol. 2017; 47: 1096-1107
        • Winget J.M.
        • Finlay D.
        • Mills K.J.
        • Huggins T.
        • Bascom C.
        • Isfort R.J.
        • et al.
        Quantitative Proteomic Analysis of Stratum Corneum Dysfunction in Adult Chronic Atopic Dermatitis.
        J Invest Dermatol. 2016; 136: 1732-1735