MicroRNA-146a alleviates chronic skin inflammation in atopic dermatitis through suppression of innate immune responses in keratinocytes

MicroRNA-146a alleviates chronic skin inflammation in atopic dermatitis through suppression of innate immune responses in keratinocytes € ckert, BSc,a Ana Rebane, PhD,a,b Toomas Runnel, MSc,a,c Alar Aab, MSc,a,b Julia Maslovskaja, MSc,a,b Beate Ru a d a,b a €rner, MSc, Angela Treis, MSc, Maire Pihlap, BSc,b Maya Zimmermann, PhD, Mario Plaas, PhD, Jaanika Ka b b Uku Haljasorg, MSc, Helen Hermann, BSc, Nikoletta Nagy, MD, PhD,e,f Lajos Kemeny, MD,e,f Triin Erm, MD,g € lli Kingo, MD, PhD,h,i Mei Li, PhD,j Mark P. Boldin, PhD,k and Cezmi A. Akdis, MDa Ku Davos, Switzerland, Tartu, Estonia, Szeged, Hungary, Illkirch, France, and Duarte, Calif Background: Chronic skin inflammation in atopic dermatitis (AD) is associated with elevated expression of proinflammatory genes and activation of innate immune responses in keratinocytes. microRNAs (miRNAs) are short, single-stranded RNA molecules that silence genes via the degradation of target mRNAs or inhibition of translation. Objective: The aim of this study was to investigate the role of miR-146a in skin inflammation in AD. Methods: RNA and protein expression was analyzed using miRNA and mRNA arrays, RT-quantitative PCR, Western blotting, and immunonohistochemistry. Transfection of miR146a precursors and inhibitors into human primary

From athe Swiss Institute of Allergy and Asthma Research (SIAF), University of Z€urich, Davos; bthe Institute of Biomedicine and Translational Medicine, cthe Institute of Molecular and Cellular Biology, and dthe Transgenic Technology Core Laboratory, University of Tartu, Tartu; ethe Department of Dermatology and Allergology, University of Szeged, Szeged; fDermatological Research Group of the Hungarian Academy of Sciences, Szeged; gthe Department of Pathology, Tartu University Hospital, Tartu; hthe Department of Dermatology and Venereology, University of Tartu, Tartu; ithe Dermatology Clinic, Tartu University Hospital, Tartu; jInstitut de Genetique et de Biologie Moleculaire et Cellulaire, Centre National de la Recherche Scientifique/Institut National de la Sante et de la Recherche Medicale/Universite de Strasbourg, Illkirch; and kthe Department of Molecular and Cellular Biology, Beckman Research Institute, City of Hope, Duarte. This work was supported by the Swiss National Science Foundation (grant no. 32-132899 and grant no. 32-112306), the Christine K€uhne-Center for Allergy Research and Education, Davos, Switzerland (CK-CARE), Swiss-Polish contribution, the Estonian Science Foundation (grant no. ESF8350 and grant no. ESF7437), European Regional Fund with Archimedes Foundation, European Union structural assistance grant (grant no. SARMP12219T), institutional research grant (grant no. IUT2-2), and personal research grants (grant no. PUT214 and grant no. PUT177) from the Estonian Research Council. A. Rebane was supported by fellowships from the SCIEX Program NMS-CH and ESTBIOREG. Disclosure of potential conflict of interest: M. Zimmermann has received research support from SNF. M. P. Boldin is employed by City of Hope, has received research support from the Tim Nesvig Lymphoma Foundation, and has stock/stock options in Regulus Therapeutics. C. A. Akdis has received research support from Novartis, PREDICTA, the Swiss National Science Foundation, Mechanisms of the Development of ALLergy, the Christine K€ uhne Center for Allergy Research and Education (CKCARE), and Swiss-Polish Research Cooperation; has received consultancy fees from Actelion, Aventis, Stallergenes, Allergopharma, Circasia, and Anergis; and is employed by the Swiss Institute of Allergy and Asthma Research. The rest of the authors declare that they have no relevant conflicts of interest. Received for publication January 27, 2014; revised April 24, 2014; accepted for publication May 6, 2014. Available online July 2, 2014. Corresponding author: Ana Rebane, PhD, Institute of Biomedicine and Translational Medicine, University of Tartu, Ravila 19, 50411 Tartu, Estonia. E-mail: ana. [email protected]. 0091-6749/$36.00 Ó 2014 American Academy of Allergy, Asthma & Immunology http://dx.doi.org/10.1016/j.jaci.2014.05.022

836

keratinocytes, luciferase assays, and MC903-dependent mouse model of AD were used to study miR-146a function. Results: We show that miR-146a expression is increased in keratinocytes and chronic lesional skin of patients with AD. miR-146a inhibited the expression of numerous proinflammatory factors, including IFN-g–inducible and ADassociated genes CCL5, CCL8, and ubiquitin D (UBD) in human primary keratinocytes stimulated with IFN-g, TNF-a, or IL-1b. In a mouse model of AD, miR-146a–deficient mice developed stronger inflammation characterized by increased accumulation of infiltrating cells in the dermis, elevated expression of IFN-g, CCL5, CCL8, and UBD in the skin, and IFN-g, IL-1b, and UBD in draining lymph nodes. Both tissue culture and in vivo experiments in mice demonstrated that miR-146a–mediated suppression in allergic skin inflammation partially occurs through direct targeting of upstream nuclear factor kappa B signal transducers caspase recruitment domain-containing protein 10 and IL-1 receptor–associated kinase 1. In addition, human CCL5 was determined as a novel, direct target of miR-146a. Conclusion: Our data demonstrate that miR-146a controls nuclear factor kappa B–dependent inflammatory responses in keratinocytes and chronic skin inflammation in AD. (J Allergy Clin Immunol 2014;134:836-47.) Key words: Allergy, noncoding RNA, atopic eczema, gene therapy

Atopic dermatitis (AD) is a common chronic inflammatory skin disease that affects 10% to 20% of children and persists throughout adulthood in 1% to 3% of the cases. Recurrent exacerbations, activation of the immune system, expansion of TH2 cells, increased IgE serum levels, production of thymic stromal lymphopoietin (TSLP) by keratinocytes in the early phase, and hyperplasia and keratinocyte apoptosis in the chronic phase represent cardinal features of AD.1-5 Skin inflammation during an exacerbation can persist for several weeks and is nearly exclusively complicated with secondary infections, including Staphylococcus aureus (S aureus), that contribute to the activation of innate immune responses and the nuclear factor kappa B (NF-kB) pathway in keratinocytes.6-9 Recent studies describe novel mediators of NF-kB activation, such as caspase recruitment domain-containing protein 10 (CARD10; also known as CARD recruited membrane associated protein 3) and ubiquitin D (UBD, also known as ubiquitin-like protein FAT10). CARD10 is activated via G protein–coupled receptors and forms a complex with mucosa-associated lymphoid

REBANE ET AL 837

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

Abbreviations used AD: Atopic dermatitis CARD10: Caspase recruitment domain-containing protein 10 IRAK1: IL-1 receptor–associated kinase 1 LN: Lymph node miRNA: microRNA NF-kB: Nuclear factor kappa B siRNA: Small-interfering RNA STAT: Signal transducer and activator of transcription TSLP: Thymic stromal lymphopoietin TLR: Toll-like receptor UBD: Ubiquitin D 39UTR: 39 untranslated region WT: Wild type

tissue lymphoma translocation protein 1, which mediates the degradation of inhibitor of kB kinase, and thereby activates NF-kB.10,11 UBD plays a role in TNF-a–dependent NF-kB activation by targeting the inhibitor of kB kinase for degradation in a ubiquitin-independent mode.12,13 The chronic phase of skin inflammation in AD is affected by IL-22, mainly produced by TH22 cells,1,14,15 and TH1-type cytokine IFN-g.16-19 Although IL-22 has been associated with tissue regeneration and hyperplasia,1,14,15,20 IFN-g is the main factor responsible for the apoptosis of keratinocytes and intensifies the inflammation through the upregulation of numerous proinflammatory genes in the skin.16-19 Two characteristic chemokines that are induced by IFN-g in keratinocytes and are upregulated in the skin of patients with AD are CCL5 (also known as RANTES) and CCL8 (also known as MCP-2).18,21 CCL5 has been shown to attract T cells, macrophages, and eosinophils.22 CCL5 promoter polymorphisms and enhanced levels have been shown to be associated with AD.23-25 CCL8 mediates the recruitment of IL-5–enriched TH2 cells in the inflamed skin in a mouse model of chronic AD.26 Regulation of inflammatory responses in diseases is mediated by coordinated control of gene expression, which is also modulated by microRNAs (miRNAs). miRNAs silence genes by triggering the degradation of target mRNAs and inhibition of translation.27,28 Several miRNAs, including miR-146a, function in the regulation of various inflammatory processes.29-31 Expression of miR-146a is dependent on NF-kB, and it has been shown to suppress the NF-kB signaling pathway through direct targeting of IL-1 receptor–associated kinase 1 (IRAK1), TNF receptor– associated factor 6,32 v-rel avian reticuloendotheliosis viral oncogene homolog B (RELB),33 and CARD10.34 miR-146a– deficient mice spontaneously develop autoimmunity later in life because of aberrant activation of NF-kB in T cells and signal transducer and activator of transcription 1 (STAT1) in regulatory T cells.35,36 In lung alveolar epithelial cells, miRNA-146a negatively regulates the release of IL-8 and CCL5 in an IL-1b signaling-independent manner.37 The function of miR-146a in the normal skin and keratinocytes and in AD has not been previously explored. Here, we demonstrate that the level of miR-146a is increased in keratinocytes from patients with AD and during skin inflammation in AD and that this elevated expression, in turn, controls chronic inflammatory processes that are triggered by IFN-g and activation of NF-kB in keratinocytes.

FIG 1. miR-146a expression in the skin and keratinocytes. A, Heatmap of miRNAs that are differentially expressed in keratinocytes from patients with AD (AD1-AD3) and healthy individuals (H1-H3). Log2 values of miRNA expression signals are mean-centered for each gene separately. Color scale from blue (lower) to red (higher) represents deviation from the mean (black). B, The relative miR-146a expression in lesional (L) and nonlesional (NL) skin of patients with AD compared with the mean of healthy (H) controls. C and D, miR-146a expression in keratinocytes stimulated or costimulated with indicated cytokines, the TLR4 ligand LPS, or the TLR2 ligand Pam3CSK4 (Pam3C) for 48 hours, or left unstimulated (us). Data on C are presented in log10 scale. Statistical analysis was performed relative to us (*) or single treatments (#). cont, Control; HKSA, heat-killed S aureus.

METHODS Patients The study was approved by the Ethical Review Committees on Human Research of the University of Tartu and the University of Szeged. All participants signed a written informed consent form. Detailed description of

838 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

FIG 2. miR-146a inhibits the expression of NF-kB–dependent genes in primary keratinocytes (KCs). A-F, KCs were transfected either with control (cont) or pre-miR-146a (miR-146a) for 24 hours and then stimulated with IFN-g or TNF-a for 48 hours or left unstimulated (us). A and D, Heatmaps of the predicted direct targets (A) and miR-146a–affected genes from the selected signaling pathways (D). Color scale was determined as in

REBANE ET AL 839

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

FIG 3. IFN-g–inducible miR-146a–affected genes are dysregulated in AD. A and B, Heatmaps of miR-146a– affected genes that are differentially expressed in the skin of patients with AD (A) and the genes most highly upregulated by IFN-g and strongly suppressed by miR-146a in keratinocytes (B). Color scale was determined as in Fig 1, A. C, Keratinocytes were stimulated with the indicated cytokines, LPS, Pam3CSK4 (Pam3C), and heat-killed S aureus (HKSA) for 48 hours or left unstimulated (us). Statistical analysis was performed relative to us (n 5 4). C and D, The relative mRNA expression levels are presented in log10 scale relative to us control (cont). E, The WT and mutated miR-146a binding site in the human CCL5 39UTR and the corresponding region of mouse CCL5. F, The relative firefly luciferase (LUC) activity is compared with that of the cont (51), (n 5 8). L, Lesional; NL, nonlesional.

the study participants is provided in the Methods section in this article’s Online Repository at www.jacionline.org.

Gene expression analysis and tissue culture experiments For functional studies, pooled, normal human epidermal keratinocytes (Promocell, Heidelberg, Germany) were used. miRNA and mRNA profiling was performed using Illumina miRNA Universal-16 and HumanHT-12

=

Expression BeadChips. The array data are available at ArrayExpress as E-MTAB-1740 and E-MTAB-1739. The cytokines and chemokines in keratinocyte cell culture supernatants were measured using the Human Cytokine 27-plex assay according to the manufacturer’s instructions (BioRad Laboratories, Hercules, Calif). In Western blots, the mean intensities of each protein level were normalized to glyceraldehyde 3-phosphate dehydrogenase. One representative blot of 3 for each studied protein is shown. For detailed description of total RNA isolation, cDNA synthesis, RT-quantitative PCR, array and pathway analysis, transfections, stimulations, Western blot

Fig 1, A. The verified direct targets IRAK1 and CARD10 are marked with an asterisk. B, Relative mRNA expression for CCL5 is presented in log10 scale. C, Chemokine levels in supernatants of keratinocyte cultures (n 5 2). E and F, Western blot analysis of whole keratinocyte extracts; P-p65 phosphorylated at Ser468 and P-STAT1 at Tyr701. The mean intensities of each protein level were compared with us cont or IFN-g–stimulated cont (STAT1). GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

840 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

FIG 4. CARD10 is important for the expression of miR146a-affected proinflammatory factors. A and B, Keratinocytes were transfected with either control (cont) siRNA or siRNAs silencing IRAK1 or CARD10 or both for 24 hours and then stimulated with IFN-g for 48 hours or left unstimulated (us). A, The relative mRNA expression is presented in log10 scale. B, Quantification of chemokines and cytokines in keratinocyte supernatants (n 5 2).

analysis, and luciferase assays, see the Methods section in this article’s Online Repository.

Mouse experiments Animal experiments were approved by the Animal Ethics Committee of the University of Tartu. The generation of miR-146a2/2 mice in the C57BL/6J background was described earlier.36 For the MC903-dependent mouse model of AD, histopathology, immunohistochemistry, and laser capture microdissection, see the Methods section in this article’s Online Repository.

Statistics Statistical analysis was performed using the unpaired Student t test. The results were considered significant at P < .05 (* or #) and highly significant at P < .01 (** or ##). Data are presented as scatter blots, with line indicating the mean 6 SEM, bar blots indicating the mean 6 SEM, or box blots with whiskers showing minimum and maximum.

RESULTS miR-146a expression is elevated in keratinocytes and skin from patients with AD To study the role of miRNAs in tissue responses in AD, we performed miRNA expression profiling of cultured primary keratinocytes from healthy individuals and patients with AD. Twelve miRNAs were differentially expressed in keratinocytes from patients with AD (Fig 1, A; see Table E1 in this article’s Online Repository at www.jacionline.org). In this miRNA set,

miR-146a was the second most strongly upregulated miRNA after miR-10b, however, with 10-fold higher expression level than that of miR-10b (Table E1). Because miR-146a had been shown to have immune regulatory functions in other cell types, we focused on miR-146a and confirmed its increased expression in nonlesional and chronic lesional skin of patients with AD (Fig 1, B). To identify AD-related factors and proinflammatory cytokines that affect miR-146a expression, we stimulated normal human primary keratinocytes with different cytokines, heat-killed S aureus, or Toll-like receptor (TLR) ligands. miR-146a was strongly upregulated by IL-1b, TNF-a, and IL-17A, with a synergistic effect between the 3 cytokines. IFN-g did not induce miR146a expression, suppressed the effect of TNF-a, and did not affect the activation by IL-1b and IL-17A (Fig 1, C). Heat-killed S aureus, IL-13, TLR2, and TLR4 ligands slightly upregulated and IL-4 and IL-22 did not affect miR-146a expression (Fig 1, D).

Overexpression of miR-146a inhibits NF-kB–dependent genes in human primary keratinocytes To study miR-146a function in keratinocytes, we overexpressed miR-146a in unstimulated and IFN-g– or TNF-a–stimulated keratinocytes and performed mRNA array analysis. TNF-a was used as a strong stimulant of miR-146a (Fig 1, C) and IFN-g as a hallmark cytokine of chronic skin inflammation in AD known to cause global changes in gene expression in keratinocytes.18 The

REBANE ET AL 841

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

FIG 5. Inhibition of miR-146a leads to the upregulation of downstream proinflammatory proteins. A-D, Keratinocytes were transfected with control (cont) LNA or miR-146a inhibitor (LNA-146a) for 24 hours and then stimulated with TNF-a or IL-1b for 48 hours or left unstimulated (us). B, The mean intensities of protein levels are compared with those of us cont. A and C, Relative mRNA expression. D, Quantification of secreted chemokines and cytokines (n 5 6). GAPDH, Glyceraldehyde 3-phosphate dehydrogenase; LNA, locked nucleic acid.

transfection of miR-146a resulted in the downregulation of 102, 37, and 410 genes in unstimulated, TNF-a–stimulated, and IFNg–stimulated keratinocytes, respectively (see Fig E1, A-C, in this article’s Online Repository at www.jacionline.org). According to Targetscan 6.2,38,39 21 predicted targets were suppressed by miR-146a in both IFN-g–stimulated and unstimulated conditions, including the previously identified miR-146a direct targets IRAK132 and CARD1034 (Fig 2, A). Suppression of CARD10, IRAK1, CCL5, and IL-8 was confirmed by RT-quantitative PCR (Fig 2, B). Reduced amounts of CCL5 and CXCL10 in IFNg–stimulated, IL-6 in TNF-a– and IFN-g–stimulated, and IL-8 in all conditions were detected in supernatants of keratinocytes on overexpression of miR-146a (Fig 2, C). Because miR-146a downregulated a large number of genes in IFN-g–stimulated cells, we next performed pathway analysis, which revealed that genes from several immune system–related pathways, including

NF-kB signaling, chemokine signaling (Fig 2, D; see Table E2 in this article’s Online Repository at www.jacionline.org), and cytokine-cytokine receptor interaction pathways (Table E2) were enriched among the downregulated genes. The effect of miR-146a overexpression was similar in TNF-a–stimulated keratinocytes (Fig 2, D). There was no significant overlap between the set of genes affected by miR-146a and genes from the janus kinase-STAT pathway, which is a primary pathway activated by IFN-g40 (Table E2). Consistent with the pathway analysis, reduced levels of phosphorylated and active NF-kB subunit p65 and total p65 were detected in miR-146a–transfected cells in IFN-g–stimulated and unstimulated cells. The NF-kB domain p50 was expressed at a higher level in IFN-g–stimulated cells, and overexpression of miR-146a slightly reduced its levels. The expression of the STAT1 protein was slightly enhanced in response to IFN-g, and miR-146a decreased its levels by

842 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

FIG 6. The relative expression of mouse miR-146a and its targets in the epidermis and the dermis. A, A representative view of the LCM before (left panel) and after (right panel) capture (bar 5 100 mm). B-D, Statistical analysis was performed relative to WT epi. E, Protein levels are compared with WT epi. der, Dermis; epi, epidermis; hf, hair follicle; LCM, laser capture microdissection.

approximately 20%. Phosphorylated, active STAT1 appeared only in response to IFN-g and was not significantly different in the presence of miR-146a (Fig 2, E). Protein levels of the direct targets CARD10 and IRAK1 were suppressed by miR-146a in both conditions (Fig 2, F).

Proinflammatory genes that are suppressed by miR-146a are dysregulated in AD We then investigated whether genes suppressed by miR-146a are differentially expressed in the skin of patients with AD. Comparison of the set of genes that are dysregulated in the skin of patients with AD (82 genes)18 with the list of genes that are suppressed by miR-146a in IFN-g–stimulated keratinocytes (410 genes) revealed 7 overlapping genes (Fig 3, A). Interestingly, 3 AD-related genes, CCL5, CCL8, and UBD, were induced more than 10-fold by IFN-g and suppressed by more than 50% by miR146a (Fig 3, B). The other tested proinflammatory cytokines and TLR ligands had only a moderate effect on the expression of CCL5, CCL8, and UBD in human primary keratinocytes (Fig 3, C). CCL8 and UBD were strongly suppressed by miR-146a (66.1% and 73.4%, respectively) in IFN-g–stimulated keratinocytes (Fig 3, D), as was shown for CCL5 (Fig 2, B). Further

analysis using Targetscan38 detected that of these 3 genes, the 39 untranslated region (39UTR) of the human CCL5 mRNA contains a potential, nonconserved binding site for miR-146a (Table E3 and Fig 3, E). To study whether this predicted binding site is functional, we inserted the corresponding 39UTRs of human CCL5, the previously confirmed direct targets IRAK132 and CARD10,34 and the CCL5 39UTR containing a mutation in the miR-146a binding site (Fig 3, E) into the luciferase reporter vector and performed luciferase assays. As presented in Fig 3, F, the transfection of miR-146a resulted in suppression of the luciferase activity of the reporters containing regions of the 39UTRs of human IRAK1, CARD10, and CCL5 but not the construct containing the mutated CCL5 39UTR (Fig 3, F), confirming that human CCL5 is a novel, direct target for miR-146a.

CARD10 is important for the expression of miR146asuppressed proinflammatory factors in keratinocytes Because the miR-146a–regulated genes CARD10 and IRAK1 are upstream elements in the NF-kB pathway, we used a smallinterfering RNA (siRNA)-silencing approach to test whether they are important for the expression of downstream genes suppressed

REBANE ET AL 843

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

FIG 7. Increased expression of IFN-g2inducible proinflammatory cytokines in inflamed skin of miR-146a– deficient mice. A-F, WT and miR-146a2/2 (146a2/2) mice were topically treated with MC903 or ethanol (control [cont]). A, Representative of hematoxylin-eosin-(HE)-stained ear sections (bar 5 100 mm). Arrows point to infiltrating cells in the cross-section and in the corresponding inserts. B, Epidermal thickness measurements (n 5 80). C, Quantification of infiltrating cells in the dermis (n 5 80). D and F, The relative expression of miR-146a and mRNA levels compared with the WT cont. E, Western blot analysis of tissue extracts. The mean intensities of protein levels are compared with WT cont. GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

by miR-146a. IRAK1 and CARD10 were efficiently inhibited by specific siRNAs (see Fig E2 in this article’s Online Repository at www.jacionline.org). Silencing of CARD10 resulted in strong downregulation of the CCL5 mRNA (71.0%) and protein (58.1%) in IFN-g–stimulated keratinocytes. Similar to CCL5, the expression of UBD mRNA was downregulated by siRNAs against CARD10 (71.4%), but not by siRNA against IRAK1. The expression of the IFN-g-independent chemokine, IL-8, was synergistically suppressed by siRNAs against IRAK1 and

CARD10 in unstimulated and IFN-g–stimulated keratinocytes (Fig 4). The siRNAs did not significantly affect the CCL8 mRNA (data not shown) and IL-6 protein (Fig 4, B) levels.

Inhibition of miR-146a results in the upregulation of proinflammatory factors in keratinocytes To study whether endogenous levels of miR-146a are sufficient for its suppressor function in keratinocytes, we used locked

844 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

IL-1β

IFN- γ *

**

8 6 4 2

6 Relative expression

*

4 2 0

co nt M C 90 3 co nt M C 90 3

co nt M C 90 3 co nt M C 90 3

0

UBD

IL- 4 **

150 100

50

wt 146a-/-

20 10

co nt M C 90 3 co nt M C 90 3

co nt M C 90 3 co nt M C 90 3

CCL8

M

3

nt

0

nt

0

co

5

co nt M C 90 3 co nt M C 90 3

10

wt 146a-/-

90

20

**

co

30

*

10

90 3

**

C

*

Relative expression

CCL5 Relative expression

**

0

0

40

*

** 30

C

Relative expression

*

Relative expression

40

200

50

wt 146-/-

M

Relative expression

10

FIG 8. Increased inflammatory response in the LNs of miR-146a–deficient mice. WT and miR-146a2/2 (146a2/2) mice were topically treated with ethanol (control [cont]) or MC903. The relative mRNA expression levels of proinflammatory cytokines and chemokines in LNs are shown compared with the WT cont.

nucleic acid–based miRNA inhibitors and stimulations with TNF-a and IL-1b. The transfection of the miR-146a inhibitor was efficient (see Fig E3 in this article’s Online Repository at www.jacionline.org) and resulted in upregulation or a tendency for upregulation of the direct targets IRAK1 and CARD10 at the mRNA and protein levels (Fig 5, A and B). The mRNA and protein levels of CCL5 and IL-8 showed a significant increase, except for the CCL5 protein in the presence of IL-1b, in all conditions. The levels of secreted IL-6 and CXCL10 were increased on miR-146a inhibition in all conditions (Fig 5, C and D). CCL8 and UBD were not significantly induced by TNF-a and IL-1b (data not shown).

miR-146a, IRAK1, and CARD10 are expressed in the epidermis of the mouse skin We next analyzed the expression of miR-146a and its targets in the skin of wild-type (WT) and miR-146a2/2 mice. The expression of miR-146a was detected in the epidermis, areas around hair

follicles, and connective tissue in the dermis using laser capture microdissection (Fig 6, A). Slightly higher expression of miR-146a was observed in the dermis than in the epidermis, when the epidermis was separated using dispase treatment (Fig 6, B). The miR-146a direct targets CARD10 and IRAK1 had significantly higher expression in the epidermis and slightly elevated protein levels in miR-146a2/2 mice than in WT mice. No difference was observed in the mRNA expression levels of CARD10, IRAK1, and the downstream indirect targets CCL5, UBD, and CXCL2 (mouse homologue of IL-8) in the epidermis of WT mice compared with that of miR-146a2/2 mice (Fig 6, C and D). The expression of CCL8 mRNA was not detectable (data not shown).

miR-146a controls AD-like skin inflammation in mice To investigate the role of miR-146a during skin inflammation, we used MC903-dependent mouse model of AD that involves repeated topical treatments of mouse ears with the vitamin D3 analog MC903, which leads to increased TSLP production and strong TH2-type inflammation in the skin.41 Histologic analysis showed that on application of MC903, WT and miR-146a2/2 mice developed severe skin inflammation with heavy infiltrations in the dermis, strong hyperplasia, and thickening of the epidermis (Fig 7, A). The epidermal thickness of WT and miR-146a–deficient mice was enhanced to a similar extent (Fig 7, B). Increased numbers of skin-infiltrating cells were present in MC903-treated, miR-146a–deficient mice than in WT mice (Fig 7, C). The expression of miR-146a itself was enhanced in the skin of MC903treated WT mice, and this correlated with the downregulation of the mRNA and protein levels of miR-146a direct targets CARD10 and IRAK1. In the resting condition, a slight upregulation of the CARD10 and IRAK1 protein levels was observed in miR-146a2/2 mice than in WT mice without a significant difference in the mRNA levels (Fig 7, D and E). Consistent with the results in human keratinocytes, the mRNA levels of the IFN-g and AD-related genes CCL5, CCL8, and UBD were enhanced in the skin of MC903-treated miR-146a2/2 mice compared with WT mice (Fig 7, E and F). The TH2-type cytokines TSLP and IL-4 were equally strongly induced by MC903 in WT and miR146a2/2 mice (see Fig E4, A and B, in this article’s Online Repository at www.jacionline.org). Of the other tested cytokines and chemokines, eosinophil-chemoattractant CCL24 (also known as Eotaxin-2) was significantly elevated in MC903-treated miR-146a2/2 mice. IL-6, IL-1b, and CXCL2 were elevated in the skin of MC903-treated mice, but no difference between WT and 146a-deficient mice was observed (Fig E4, C). The mRNAs of IL-22 and IL-17A were not detectable in all conditions. The number of CD3-expressing T cells was enhanced in response to MC903 in both WT and miR-146a2/2 mice (Fig E4, D). In line with the results of the skin, increased levels of IFN-g, IL-1b, and UBD were detected in draining lymph nodes (LNs) of miR-146a–deficient mice compared with WT mice on MC903 treatment. Enhanced expression levels of CCL5, CCL8, and IL-4 were detected in draining LNs in both WT and miR-146a– deficient mice (Fig 8). DISCUSSION Information regarding the role of miRNAs in allergic diseases is limited.42,43 The present study demonstrates an essential immune

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

regulatory function of miR-146a in the skin. miR-146a is upregulated in the keratinocytes and skin of patients with AD and inhibits the expression of numerous proinflammatory factors, including IFN-g–inducible and AD-associated genes CCL5, CCL8, and UBD in human primary keratinocytes and in vivo in mouse model of AD. Both siRNA silencing and in vivo mouse experiments suggest that miR-146a–mediated regulation occurs partially through direct targeting of IRAK1, CARD10, and CCL5 in keratinocytes. Our data together show that the expression of miR-146a is increased during skin inflammation in AD and this, in turn, helps to control chronic inflammatory processes triggered by IFN-g and activation of NF-kB in keratinocytes (see Fig E5 in this article’s Online Repository at www.jacionline.org). Chronic skin inflammation in AD is affected by numerous factors that besides TH2-type cytokines include secondary infections and activation of keratinocytes by IL-22 and IFN-g.1,7,17,18,20,44 Remarkably, IFN-g induces expression changes of thousands of genes in keratinocytes, while only 39 genes are affected by IL-22.45 To study the function of miR146a in keratinocytes, we used stimulations with IFN-g and TNF-a. miR-146a downregulated 410 genes in IFN-g–stimulated keratinocytes, pointing to a strong anti-inflammatory function of miR-146a in chronic AD. Only 37 genes were inhibited in the presence of TNF-a, although a similar set of genes tended to be suppressed. It is possible that the increased endogenous level of miR-146a is sufficient for the suppression in TNF-a–stimulated keratinocytes. Keratinocytes express numerous pattern-recognition receptors, the stimulation of which leads to the activation of the NF-kB pathway. The enhanced expression of NF-kB–dependent miR146a32 in keratinocytes and nonlesional skin of patients with AD could be due to epigenetic changes in the miR-146a promoter or could represent a different activation state of keratinocytes of patients with AD. miR-146a is also upregulated in the skin of patients with psoriasis,46 which indicates that elevated expression of miR-146a is a general characteristic of inflammatory responses in the skin. Enhancement of miR-146a expression by NF-kB appears to have a feedback loop effect, because miR-146a targets several modulators of the NF-kB pathway, including IRAK1, TRAF6, RELB, and CARD10, in different cell types.32-34 In the present study, we observed suppression of CARD10 and IRAK1 by miR-146a, whereas silencing of CARD10, but not of IRAK1, led to robust downregulation of the AD-related genes CCL5 and UBD. Interestingly, human CCL5 seems to be particularly well controlled by miR-146a. First, human CCL5 is a direct target of miR-146a, as we demonstrated using luciferase assays, and second, it is suppressed as soon as miR-146a targets CARD10. The targeting of CARD10 and IRAK1 seems to be crucial, because other miR-146a–affected mRNAs, such as UBD, CCL8, and IL-8, do not contain potential binding sites for miR-146a. As we demonstrated using locked nucleic acid inhibitors, endogenous levels of miR-146a were sufficient for its function in keratinocytes. Previously, it was shown that autoimmunity in miR-146a2/2 mice is partially caused by dysregulated IFN-g production due to the overactivation of STAT1 in regulatory T cells, which renders them incapable of suppressing TH1-type responses.35 In addition, miR-146a has been shown to suppress STAT1 in HEK293 cells.47 Although we observed slight downregulation of the total STAT1 protein level by miR-146a in IFNg–stimulated keratinocytes, there was no difference in the

REBANE ET AL 845

phosphorylated STAT1 level or significant influences on janus kinase-STAT pathway genes. IFN-g plays an important role in the chronic phase of allergic inflammation.16,18,44 Although fewer IFN-g–producing CD41 cells are present in AD than in psoriatic skin, the number of IFN-g–producing CD81 cytotoxic T cells is similar in these 2 diseases.48,49 Because CCL5 and CCL8, which attract T cells, macrophages, and eosinophils,22,26 are strongly induced by IFN-g in keratinocytes, they could provide a link into how the activation of keratinocytes by IFN-g contributes to the development of chronic inflammation. Interestingly, our data suggest that CARD10 and UBD are novel factors that link the IFN-g and NF-kB pathways. CARD10 is activated by angiotensin II and lysophosphatidic acid via G protein–coupled receptors,11,50,51 and UBD has been shown to be important for TNF-a–dependent NF-kB activation in renal tubular epithelial cells and mouse lymphocytes.52 No effect of IFN-g on UBD and CARD10 has been described. Although miR-146a was expressed both in the epidermis and in the dermis, its direct targets IRAK1 and CARD10 were expressed higher in the epidermis, suggesting an important function for them in the regulation of the inflammatory responses in keratinocytes. Previously, it was reported that the protein, but not mRNA, expression of the miR-146a direct targets IRAK1 and TRAF6 was elevated in bone marrow–derived macrophages and B cells in miR-146a2/2 mice.36 Similarly, we observed slightly higher protein levels of IRAK1 and CARD10 in miR-146a2/2 mice in resting conditions; however, there was no difference in their mRNA expression levels. These and previously published results indicate that the effect of miR-146a is not always evident on the mRNA level. To study how miR-146a affects AD-related inflammation in vivo, we used MC903-dependent mouse model of AD in which ablation of miRNA-processing factor Dicer-1 resulted in the aggravation of AD.53 We observed increased mRNA expression of IFN-g, CCL5, CCL8, and UBD in the skin of MC903treated miR-146a2/2 mice and reduced mRNA and protein levels of miR-146a direct targets IRAK1 and CARD10 in MC903treated WT mice. In addition, IFN-g and IL-1b were upregulated in draining LNs. These results demonstrate that miR-146a has antiinflammatory function in the skin, which involves targeting of the epidermal NF-kB pathway-activating elements CARD10 and IRAK1 and suppression of the downstream IFN-g– inducible, AD-related genes CCL5, CCL8, and UBD. In line with gene expression changes, we observed an increased number of skin-infiltrating cells in MC903-treated, miR-146a–deficient mice. Anti-CD3 staining of frozen skin sections from MC903treated mice demonstrated increased accumulation of CD3 positive cells, particularly in the area of the dermis and epidermis connection both in WT and in miR-146a2/2 mice, suggesting that most of the infiltrating cells were other than T cells. This finding opens a window for future studies to further characterize the types of cells and their localization in tissues affected by miR-146a. In mammals, the miR-146 family consists of 2 members, miR146a and miR-146b, which are encoded by 2 independent genes located on chromosomes 5 and 10, respectively. Mature miR-146a and miR-146b differ from each other in 2 nucleotides, and they most likely target a similar set of genes because of the same seed sequence.32,54 Although miR-146a function has been intensively studied in immune cells and in relation to many diseases, the function and regulation of miR-146b is less described. Recent studies

846 REBANE ET AL

note the importance of miR-146b in different cell types.55,56 In monocytes, miR-146b is upregulated via an IL-10–mediated STAT3-dependent loop on LPS stimulation and directly targets TLR4.55 We did not observe a difference in miR-146b expression in keratinocytes from patients with AD. Future studies are needed to explore functions of miR-146b in the skin. We observed increased expression of the eosinophil chemoattractant CCL24 in the skin; however, there was no difference in TSLP and IL-4 expressions in the skin and LNs. It is possible that miR-146a suppresses TH2-type immune responses on antigenspecific stimulation because miR-146a has been shown to downregulate T-cell receptor–driven NF-kB activation.57 A limited number of studies have been published on miRNA functions in inflammatory skin diseases. miR-155 is overexpressed in the skin of patients with AD and apparently affects the development of AD by downregulating cytotoxic T lymphocyte– associated antigen CTLA-4.58 miR-125b is enhanced in psoriatic skin and functions as a suppressor of keratinocyte proliferation via direct targeting of fibroblast growth factor receptor 2.59 In the present study, we demonstrated an important role of miR146a in the suppression of innate immune responses in keratinocytes during chronic skin inflammation in AD. Our results suggest that overexpression of miR-146a may have potential for the treatment of inflammatory skin diseases. Uku H€amarik (Institute of Biomedicine and Translational Medicine, University of Tartu) is acknowledged for technical help during mouse experiments. Kristi Abram and Maire Karelson (Dermatology Clinic, Tartu University Hospital) are thanked for their collection of skin biopsies. Konstatin Tretjakov (Institute of Computer Sciences, University of Tartu) is acknowledged for help with array data analysis. P€art Peterson (Institute of Biomedicine and Translational Medicine, University of Tartu) and Zsolt Komlosi (SIAF, Davos) are greatly thanked for helpful discussions.

Clinical implications: Our results demonstrate that miR-146a inhibits innate immune responses in keratinocytes and suggest that overexpression of miR-146a may be a potential treatment of chronic inflammatory conditions in the skin.

REFERENCES 1. Eyerich S, Eyerich K, Pennino D, Carbone T, Nasorri F, Pallotta S, et al. Th22 cells represent a distinct human T cell subset involved in epidermal immunity and remodeling. J Clin Invest 2009;119:3573-85. 2. Bieber T. Atopic dermatitis. Ann Dermatol 2010;22:125-37. 3. Boguniewicz M, Leung DY. Atopic dermatitis: a disease of altered skin barrier and immune dysregulation. Immunol Rev 2011;242:233-46. 4. Soumelis V, Reche PA, Kanzler H, Yuan W, Edward G, Homey B, et al. Human epithelial cells trigger dendritic cell mediated allergic inflammation by producing TSLP. Nat Immunol 2002;3:673-80. 5. Watanabe N, Hanabuchi S, Soumelis V, Yuan W, Ho S, de Waal Malefyt R, et al. Human thymic stromal lymphopoietin promotes dendritic cell-mediated CD41 T cell homeostatic expansion. Nat Immunol 2004;5:426-34. 6. Medzhitov R. Inflammation 2010: new adventures of an old flame. Cell 2010; 140:771-6. 7. Guttman-Yassky E, Nograles KE, Krueger JG. Contrasting pathogenesis of atopic dermatitis and psoriasis–part II: immune cell subsets and therapeutic concepts. J Allergy Clin Immunol 2011;127:1420-32. 8. Kuo IH, Yoshida T, De Benedetto A, Beck LA. The cutaneous innate immune response in patients with atopic dermatitis. J Allergy Clin Immunol 2013;131: 266-78. 9. Classen A, Kalali BN, Schnopp C, Andres C, Aguilar-Pimentel JA, Ring J, et al. TNF receptor I on human keratinocytes is a binding partner for staphylococcal protein A resulting in the activation of NF kappa B, AP-1, and downstream gene transcription. Exp Dermatol 2011;20:48-52.

J ALLERGY CLIN IMMUNOL OCTOBER 2014

10. Grabiner BC, Blonska M, Lin PC, You Y, Wang D, Sun J, et al. CARMA3 deficiency abrogates G protein-coupled receptor-induced NF-{kappa}B activation. Genes Dev 2007;21:984-96. 11. Mahanivong C, Chen HM, Yee SW, Pan ZK, Dong Z, Huang S. Protein kinase C alpha-CARMA3 signaling axis links Ras to NF-kappa B for lysophosphatidic acid-induced urokinase plasminogen activator expression in ovarian cancer cells. Oncogene 2008;27:1273-80. 12. Gong P, Canaan A, Wang B, Leventhal J, Snyder A, Nair V, et al. The ubiquitinlike protein FAT10 mediates NF-kappaB activation. J Am Soc Nephrol 2010;21: 316-26. 13. Hipp MS, Kalveram B, Raasi S, Groettrup M, Schmidtke G. FAT10, a ubiquitinindependent signal for proteasomal degradation. Mol Cell Biol 2005;25:3483-91. 14. Gittler JK, Shemer A, Suarez-Farinas M, Fuentes-Duculan J, Gulewicz KJ, Wang CQ, et al. Progressive activation of T(H)2/T(H)22 cytokines and selective epidermal proteins characterizes acute and chronic atopic dermatitis. J Allergy Clin Immunol 2012;130:1344-54. 15. 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;123:1244-52.e2. 16. Trautmann A, Altznauer F, Akdis M, Simon HU, Disch R, Brocker EB, et al. The differential fate of cadherins during T-cell-induced keratinocyte apoptosis leads to spongiosis in eczematous dermatitis. J Invest Dermatol 2001;117: 927-34. 17. Klunker S, Trautmann A, Akdis M, Verhagen J, Schmid-Grendelmeier P, Blaser K, et al. A second step of chemotaxis after transendothelial migration: keratinocytes undergoing apoptosis release IFN-gamma-inducible protein 10, monokine induced by IFN-gamma, and IFN-gamma-inducible alpha-chemoattractant for T cell chemotaxis toward epidermis in atopic dermatitis. J Immunol 2003;171: 1078-84. 18. Rebane A, Zimmermann M, Aab A, Baurecht H, Koreck A, Karelson M, et al. Mechanisms of IFN-gamma-induced apoptosis of human skin keratinocytes in patients with atopic dermatitis. J Allergy Clin Immunol 2012;129:1297-306. 19. Trautmann A, Akdis M, Kleemann D, Altznauer F, Simon HU, Graeve T, et al. T cell-mediated Fas-induced keratinocyte apoptosis plays a key pathogenetic role in eczematous dermatitis. J Clin Invest 2000;106:25-35. 20. Akdis M, Palomares O, van de Veen W, van Splunter M, Akdis CA. TH17 and TH22 cells: a confusion of antimicrobial response with tissue inflammation versus protection. J Allergy Clin Immunol 2012;129:1438-49; quiz 50-1. 21. Fujita H, Shemer A, Suarez-Farinas M, Johnson-Huang LM, Tintle S, Cardinale I, et al. Lesional dendritic cells in patients with chronic atopic dermatitis and psoriasis exhibit parallel ability to activate T-cell subsets. J Allergy Clin Immunol 2011;128:574-82, e1-12. 22. Ying S, Taborda-Barata L, Meng Q, Humbert M, Kay AB. The kinetics of allergen-induced transcription of messenger RNA for monocyte chemotactic protein-3 and RANTES in the skin of human atopic subjects: relationship to eosinophil, T cell, and macrophage recruitment. J Exp Med 1995;181: 2153-9. 23. Bai B, Tanaka K, Tazawa T, Yamamoto N, Sugiura H. Association between RANTES promoter polymorphism-401A and enhanced RANTES production in atopic dermatitis patients. J Dermatol Sci 2005;39:189-91. 24. Tanaka K, Roberts MH, Yamamoto N, Sugiura H, Uehara M, Hopkin JM. Upregulating promoter polymorphisms of RANTES relate to atopic dermatitis. Int J Immunogenet 2006;33:423-8. 25. Nickel RG, Casolaro V, Wahn U, Beyer K, Barnes KC, Plunkett BS, et al. Atopic dermatitis is associated with a functional mutation in the promoter of the C-C chemokine RANTES. J Immunol 2000;164:1612-6. 26. Islam SA, Chang DS, Colvin RA, Byrne MH, McCully ML, Moser B, et al. Mouse CCL8, a CCR8 agonist, promotes atopic dermatitis by recruiting IL-51 T(H)2 cells. Nat Immunol 2011;12:167-77. 27. Makeyev EV, Maniatis T. Multilevel regulation of gene expression by microRNAs. Science 2008;319:1789-90. 28. Djuranovic S, Nahvi A, Green R. A parsimonious model for gene regulation by miRNAs. Science 2011;331:550-3. 29. O’Connell RM, Rao DS, Baltimore D. microRNA regulation of inflammatory responses. Annu Rev Immunol 2012;30:295-312. 30. Boldin MP, Baltimore D. MicroRNAs, new effectors and regulators of NF-kappaB. Immunol Rev 2012;246:205-20. 31. O’Neill LA, Sheedy FJ, McCoy CE. MicroRNAs: the fine-tuners of Toll-like receptor signalling. Nat Rev Immunol 2011;11:163-75. 32. Taganov KD, Boldin MP, Chang KJ, Baltimore D. NF-kappaB-dependent induction of microRNA miR-146, an inhibitor targeted to signaling proteins of innate immune responses. Proc Natl Acad Sci U S A 2006;103:12481-6.

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

33. Etzrodt M, Cortez-Retamozo V, Newton A, Zhao J, Ng A, Wildgruber M, et al. Regulation of monocyte functional heterogeneity by miR-146a and Relb. Cell Rep 2012;1:317-24. 34. Crone SG, Jacobsen A, Federspiel B, Bardram L, Krogh A, Lund AH, et al. microRNA-146a inhibits G protein-coupled receptor-mediated activation of NF-kappaB by targeting CARD10 and COPS8 in gastric cancer. Mol Cancer 2012;11:71. 35. Lu LF, Boldin MP, Chaudhry A, Lin LL, Taganov KD, Hanada T, et al. Function of miR-146a in controlling Treg cell-mediated regulation of Th1 responses. Cell 2010;142:914-29. 36. Boldin MP, Taganov KD, Rao DS, Yang L, Zhao JL, Kalwani M, et al. miR-146a is a significant brake on autoimmunity, myeloproliferation, and cancer in mice. J Exp Med 2011;208:1189-201. 37. Perry MM, Moschos SA, Williams AE, Shepherd NJ, Larner-Svensson HM, Lindsay MA. Rapid changes in microRNA-146a expression negatively regulate the IL-1beta-induced inflammatory response in human lung alveolar epithelial cells. J Immunol 2008;180:5689-98. 38. Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res 2009;19:92-105. 39. Garcia DM, Baek D, Shin C, Bell GW, Grimson A, Bartel DP. Weak seed-pairing stability and high target-site abundance decrease the proficiency of lsy-6 and other microRNAs. Nat Struct Mol Biol 2011;18:1139-46. 40. Platanias LC. Mechanisms of type-I- and type-II-interferon-mediated signalling. Nat Rev Immunol 2005;5:375-86. 41. Li M, Hener P, Zhang Z, Kato S, Metzger D, Chambon P. Topical vitamin D3 and low-calcemic analogs induce thymic stromal lymphopoietin in mouse keratinocytes and trigger an atopic dermatitis. Proc Natl Acad Sci U S A 2006;103:11736-41. 42. Lu TX, Rothenberg ME. Diagnostic, functional, and therapeutic roles of microRNA in allergic diseases. J Allergy Clin Immunol 2013;132:3-13, quiz 4. 43. Rebane A, Akdis CA. MicroRNAs: essential players in the regulation of inflammation. J Allergy Clin Immunol 2013;132:15-26. 44. Soyka MB, Wawrzyniak P, Eiwegger T, Holzmann D, Treis A, Wanke K, et al. Defective epithelial barrier in chronic rhinosinusitis: the regulation of tight junctions by IFN-gamma and IL-4. J Allergy Clin Immunol 2012;130:1087-96.e10. 45. 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; 159:1092-102. 46. Sonkoly E, Wei T, Janson PC, Saaf A, Lundeberg L, Tengvall-Linder M, et al. MicroRNAs: novel regulators involved in the pathogenesis of psoriasis? PLoS One 2007;2:e610.

REBANE ET AL 847

47. Tang Y, Luo X, Cui H, Ni X, Yuan M, Guo Y, et al. MicroRNA-146A contributes to abnormal activation of the type I interferon pathway in human lupus by targeting the key signaling proteins. Arthritis Rheum 2009;60:1065-75. 48. Hamid Q, Boguniewicz M, Leung DY. Differential in situ cytokine gene expression in acute versus chronic atopic dermatitis. J Clin Invest 1994;94:870-6. 49. Hijnen D, Knol EF, Gent YY, Giovannone B, Beijn SJ, Kupper TS, et al. CD8(1) T cells in the lesional skin of atopic dermatitis and psoriasis patients are an important source of IFN-gamma, IL-13, IL-17, and IL-22. J Invest Dermatol 2013;133: 973-9. 50. McAllister-Lucas LM, Ruland J, Siu K, Jin X, Gu S, Kim DS, et al. CARMA3/ Bcl10/MALT1-dependent NF-kappaB activation mediates angiotensin IIresponsive inflammatory signaling in nonimmune cells. Proc Natl Acad Sci U S A 2007;104:139-44. 51. Medoff BD, Landry AL, Wittbold KA, Sandall BP, Derby MC, Cao Z, et al. CARMA3 mediates lysophosphatidic acid-stimulated cytokine secretion by bronchial epithelial cells. Am J Respir Cell Mol Biol 2009;40:286-94. 52. Canaan A, Yu X, Booth CJ, Lian J, Lazar I, Gamfi SL, et al. FAT10/diubiquitinlike protein-deficient mice exhibit minimal phenotypic differences. Mol Cell Biol 2006;26:5180-9. 53. Hener P, Friedmann L, Metzger D, Chambon P, Li M. Aggravated TSLP-induced atopic dermatitis in mice lacking Dicer in adult skin keratinocytes. J Invest Dermatol 2011;131:2324-7. 54. Griffiths-Jones S. miRBase: the microRNA sequence database. Methods Mol Biol 2006;342:129-38. 55. Curtale G, Mirolo M, Renzi TA, Rossato M, Bazzoni F, Locati M. Negative regulation of Toll-like receptor 4 signaling by IL-10-dependent microRNA-146b. Proc Natl Acad Sci U S A 2013;110:11499-504. 56. Ahn J, Lee H, Jung CH, Jeon TI, Ha TY. MicroRNA-146b promotes adipogenesis by suppressing the SIRT1-FOXO1 cascade. EMBO Mol Med 2013;5: 1602-12. 57. Yang L, Boldin MP, Yu Y, Liu CS, Ea CK, Ramakrishnan P, et al. miR-146a controls the resolution of T cell responses in mice. J Exp Med 2012;209: 1655-70. 58. Sonkoly E, Janson P, Majuri ML, Savinko T, Fyhrquist N, Eidsmo L, et al. MiR155 is overexpressed in patients with atopic dermatitis and modulates T-cell proliferative responses by targeting cytotoxic T lymphocyte-associated antigen 4. J Allergy Clin Immunol 2010;126:581-9, e1-20. 59. Xu N, Brodin P, Wei T, Meisgen F, Eidsmo L, Nagy N, et al. MiR-125b, a microRNA downregulated in psoriasis, modulates keratinocyte proliferation by targeting FGFR2. J Invest Dermatol 2011;131:1521-9.

847.e1 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

METHODS Patients, human primary keratinocytes, and skin biopsies This study was approved by the Ethical Review Committees on Human Research of the University of Tartu and the University of Szeged. All participants signed a written informed consent form. Primary keratinocytes (all passage 4) from the nonlesional skin of 3 patients with AD and 3 healthy individuals undergoing plastic surgery were used in miRNA profiling (Fig 1, A). Patients and the preparation of the keratinocytes have been described earlier.E1,E2 In Fig 1, B, skin biopsies (diameter, 4 mm) from 9 patients with chronic AD (5 women, 4 men, age, 18-42 years) and 9 healthy subjects (4 men, 5 women, age 19-43 years) were used. All patients had experienced 6to 14-day-long severe exacerbation of the disease. None of the patients had been treated with systemic antihistamines or topical corticosteroids for at least 1 week before inclusion in the study.

Transfections and stimulations

miRBase 12.0E3 and Illumina HumanHT-12 Expression BeadChips, respectively). miRNA and mRNA Illumina arrays were carried out at the Core Facility at the Department of Biotechnology, University of Tartu, using 500 ng of total RNA per each sample. The data were analyzed with GenomeStudio 2011.1 Gene Expression Module using average normalization for miRNA data and Illumina’s custom rank invariant method for mRNA arrays. Genes were considered to be expressed at detection P < .05 and differentially expressed at P < .05. Detection P value is a parameter that estimates whether a particular gene is expressed with comparing signal intensities of the genes with the controls and noise of the same array. Further analyses and visualizations were performed using Microsoft Excel and Multi Experiment Viewer 4.6.2. Pathway analysis was performed with g:Profiler (http://biit.cs.ut.ee/ gprofiler), which retrieves the most significant gene ontology groups and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways and enables one to estimate the significance of the search results by calculating enrichment P value using Fisher 1-tailed test.E4

Isolation of RNA, cDNA synthesis, and RT-quantitative PCR

For functional studies, pooled, normal human epidermal keratinocytes (Promocell) were used. Keratinocytes were stimulated for 48 hours with the following concentrations of cytokines: TNF-a, 25 ng/mL; IL-1b, 10 ng/mL; IL-4, 40 ng/mL; IFN-g, 10 ng/mL; IL-17A, 10 ng/mL; IL-22, 10 ng/mL; IL-13, 50 ng/mL (R&D Systems, Minneapolis, Minn); LPS, 10 ng/mL; Pam3CSK4, 2.5 ng/mL; and heat killed S aureus, 106 cells/mL (Invivogen, San Diego, Calif). Transfections were carried out in 24-well plates using 0.8 mL of siPORT NeoFX (Life Technologies, Grand Island, NY) and 5 3 104 keratinocytes in 0.6 mL growth medium, which was for keratinocytes Keratinocyte-SFM and for HEK293 cells Dulbecco modified Eagle medium (both from Invitrogen, Grand Island, NY). After 24 hours, keratinocytes were stimulated with cytokines as indicated for 48 hours. Transfections in keratinocytes were performed at 30 nM of pre-miRNAs (Pre-miR Precursors Molecules hsa-miR-146a and negative control #1, Life Technologies) and miRCURY LNA inhibitors (microRNA Power inhibitor hsa-miR-146a and negative control A, Exiqon, Vedbaek, Denmark). Silencer Select Validated IRAK1 (s323), CARD10 (s26577), and negative control #1 siRNAs (Life Technologies) were used at 50 nM concentration.

Total RNA was extracted using miRNAeasy Mini Kit (Qiagen, Valencia, Calif). Homogenization of the human and mouse skin, epidermis, and dermis samples and mouse LNs was done by using Precellys 24 (Precellys, Southampton, United Kingdom) homogenizer or gentleMACS diccociator (Miltenyi Biotec, Heidelberg, Germany). cDNA was synthesized from 200 to 900 ng of total RNA using oligo-dT and reagents from Thermo Scientific. RNA concentration and quality were assessed with NanoDrop ND-1000 and Agilent 2100 Bioanalyzer. SYBR/Rox master mix (Bio-Rad Laboratories or Solis BioDyne, Tartu, Estonia) and ABI Prism 7900 were used for quantitative PCR (qPCR). The relative gene expression levels were normalized to mouse or human EF1A levels and calculated using the comparative Ct (DDCt) method (Life Technologies). The mean level of control experiments or control group was equalized to 1. miRNA qPCR was carried out by using TaqMan microRNA Assays (Life Technologies) and were normalized to Let-7a. All PCR primers were designed with the assistance of Primer 3 and were ordered from Microsynth (Balgach, Switzerland).

Array and pathway analysis

RT-qPCR primers

miRNA and mRNA profiling was carried out using Illumina miRNA Universal-16 BeadChips (probes for 858 human mature miRNAs based on

All primers were designed with the assistance of Primer 3 and were ordered from Microsynth.

Gene

CARD10 CCL8 CCL5 EF1Aex7-8 IRAK1 IL8 hUBD mCARD10 mCCL24 mCCL5 mCCL8 mCXCL2 mEF1A mIFNg mIL1b mIL4 mIL6 mIRAK1 mTSLP mUBD mLor

Forward

Reverse

AGTGTGCCCAGCGGAAAGCC AGGGACTTGCTCAGCCAGATTCAGT AGTCGTCTTTGTCACCCGAAA CCACCTTTGGGTCGCTTTGCTGT CACCTTCAGCTTTGGGGTGGTAGTG GCAGCTCTGTGTGAAGGTGCAGTT CAGAGATGGCTCCCAATGCT GCGAGGTCTACCCCATTGTC TCCCAAGGCAGGGGTCATCTTCA GCCTCACCATATGGCTCGGACAC GGGCCAGATAAGGCTCCAGTCACC TGAACAAAGGCAAGGCTAACTG CGGCAGTCGCCTTGGACGTT TGCCAAGTTTGAGGTCAACAACCCA AGCTTCCTTGTGCAAGTGTCT CAGGAGAAGGGACGCCATGCAC ACTTCACAAGTCGGAGGCTT TGTGAGGACACAAGGTGCAA ATCGAGGACTGTGAGAGCAAGCCAG TTCTGTCCGCACCTGTGTTG TCCCTGGTGCTTCAGGGTAAC

GATGGCCCGGATCCTGCTGC TCTTGTGTAGCTCTCCAGCCTCTGG TCTCCATCCTAGCTCATCTCCAA TGCCAGCTCCAGCAGCCTTCTT CCAGCCTCCTCAGCCTCCTCT TTCTGTGTTGGCGCAGTGTGGTC CGCTGTCATATGGGTTGGCA CAACAGGCCCCTGATCTCAC TGCCTCTGAACCCACAGCAGC TTGACGTGGGCACGAGGCAG TCAGCACCCGAAGGGGGATCT CAGGTACGATCCAGGCTTCC CGGTGGTTTTCACAACACCTGCGT ACAGCTGGTGGACCACTCGGA TGGGGTCCGTCAACTTCAAA GCGAAGCACCTTGGAAGCCCT TGCCATTGCACAACTCTTTTCTC TAGGCTGGGTGCTTTTCAGG GTGAAGGGCAGCCAGGGATAGGA GAGACCTTGGTTTGGGACCT TCTTTCCACAACCCACAGGAG

REBANE ET AL 847.e2

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

Western blot analysis

Laser capture microdissection

Tissue samples were homogenized using gentleMACS in Luciferase cell lysis buffer and then boiled in 13 Laemmli buffer for either 5 minutes (epidermis and dermis samples) or 60 minutes (formaldehyde cross-linked frozen samples) at 1008C before the loading to the gel. A total of 5 mg of total protein was loaded in each line of SDS PAGE gel. Antibodies were diluted in Tris-buffered saline containing 0.05% Tween 20 and 5% milk or 2.5% BSA (antimouse CCL5). Primary antibodies were used at the following dilutions: mouse monoclonal anti-IRAK1 (clone 3F7, ab119289, Abcam, Cambridge, United Kingdom) 1:1000, mouse monoclonal anti-human glyceraldehyde 3-phosphate dehydrogenase (clone 6C5) 1:5000, goat polyclonal antihuman CARD10 (Sc-32233) 1:1000, goat polyclonal anti-mouse CCL5 (sc-1410, cloneC-19) 1:1000, rabbit polyclonal anti-mouse IRAK-1 (sc-7883, H-273) 1:1000 (all from Santa Cruz Biotech, Santa Cruz, Calif); rabbit mAbs recognizing human STAT1 (clone 42H3) and phosoho-STAT1 (Tyr701, clone 58D6) (both from Cell Signaling Technology, Danvers, Mass) 1:1000, mouse monoclonal anti-human NFKB1 (clone 5D10, Thermo Scientific) 1:1000, and antiphospho-NF-kB p65 (Ser468) (Cell Signaling) 1:1000. Anti-rabbit, -goat and -mouse IgG horseradish peroxidase– conjugated antibodies from Cell Signaling Technology were used as secondary antibodies. Signals were detected with the enhanced chemiluminescence plus substrate (Thermo Scientific, Waltham, Mass) and captured with LAS-1000 (FUJIFILM Europe GmbH, Duesseldorf, Germany) or ImageQant LAS gel (GE Healthcare, Fairfield, Conn) systems. Western blots were quantified using ImageJ software. The mean intensity of the background value was subtracted from the mean intensity of each protein band followed by normalization against the mean values of glyceraldehyde 3-phosphate dehydrogenase.

Pieces of the skin (5 3 5 mm) from trimmed mouse back were embedded in Tissue-Tek (Thermo Scientific) and quick-frozen using the mixture of ethanol and dry ice. For laser capture microdissection, 10-mm tissue sections were cut, fixed in methanol for 10 minutes, stained with hematoxylin, and stored at 2208C. Laser capture microdissection was performed with Zeiss PALM MicroBeam. RNA from microdissected tissue was purified using miRNAeasy Mini Kit (Qiagen). A total of 0.8 mg of bacteriophage MS2 RNA (Roche Applied Science Switzerland) was added as carrier RNA per each sample.

Luciferase assay 39UTR fragments of CCL5, CARD10, and IRAK1 were cloned into the pmirGLO Dual-Luciferase miRNA Target Expression Vector (Promega, Madison, Wis) using the PCR primers containing NheI and SalI sites. The following primers were used: CARD10 39UTR FOR 59-ATTGC TAGCGGCCTCTCACCCCCAAGGCT-39; CARD10 39UTR REV 59ATTGTCGACCACGCGGTAACGGGCTCAGG-39; IRAK1 39UTR FOR 59-ATTGCTAGCT GTGTTCACCTGGGCAGA TCC-39; IRAK1 39UTR REV 59-ATTGTCGACTCCCAACATGCGCCAGCC-39; CCL5 39UTR FOR 59-ATTGCTAGCAA CTAA AGCCTAGAAGAGCTT-39; CCL5 39UTR REV 59-ATTGTCGACGGT TGC ATT GAG AAC TTT AAT G-39. For the generation of CCL5 39UTR fragment with mutated miR146a binding site, CCL5 39UTR FOR and CCL5mut 39UTR REV 59 GTCGACGGTTGCATTGACTACTTTAATGGTAAGCCGA 39 primers were used. Transfections were carried out in 24-well plates using 0.8 mL of siPORT NeoFX (Life Technologies), 30 nM of pre-miRNAs (Pre-miR Precursors Molecules hsa-miR-146a or negative control #1, Life Technologies), 50 ng of the reporter plasmid, and 5 3 104 human embryonic kidney epithelial cells HEK293 in 0.6 mL medium for 24 hours. Firefly and renilla luciferase activities were measured using Promega dual luciferase assay, and firefly luciferase activities were normalized to the values of the renilla luciferase.

Mouse AD model

miR-146a2/2 mice in C57BL/6J background and C57BL/6J (B6) WT mice were purchased from the Jackson Laboratory (Bar Harbor, Me). Generation of miR-146a2/2 mice has been described earlier.E5 The mice were maintained and bred in the animal facility at the Institute of Molecular and Cell Biology, University of Tartu, in accordance with the institute’s regulations. Nine- to 10week-old female mice were used for the experiments. Animal experiments were approved by the Animal Ethics Committee of the University of Tartu. MC903 (calcipotriol hydrate, C4369, Sigma-Aldrich, St Louis, Mo) was dissolved in EtOH and topically applied on mouse ears (1 nmol per ear) on every other day 10 times starting at day 0. As vehicle control, the same amount of EtOH was applied. Analyses of the ear biopsies and LNs were performed on day 19.

Separation of the epidermis and the dermis For separation of the epidermis from the dermis, 1 3 1 cm pieces of the skin from trimmed mouse back were incubated for 15 to 18 hours at 48C in Keratinocyte-SFM Medium (Invitrogen) in the presence of 5 mg/mL of Dispase II (Life Technologies).

Histopathology and immunohistochemistry Ear biopsies were fixed in 4% paraformaldehyde for 24 hours at 48C and embedded in paraffin. Five-micrometer sections were stained with hematoxylin/eosin. For immunohistochemistry, heat-induced epitope retrieval method using Ph 9.0 Target Retrieval Solution (Dako Denmark) was applied. DAKO REAL peroxidase blocking was used for 7 to 15 minutes to block endogenous peroxidase. For detection, rabbit polyclonal anti-mouse TSLP (ab115700, Abcam) at 1:200 dilution, EnVision-labeled polymer containing horseradish peroxidase–conjugated antibodies for rabbit IgG antibodies, and DAKO DAB chromogen were used. Zeiss Axiovert 405 M microscope and Axiovision Rel. 4.8 software were used to capture images and to perform measurements. For epidermal thickness, the distance between the epithermal-dermal junction and the epidermal-stratum corneum junction was measured from 80 different randomly selected places for each group. The infiltrating cells were counted from 25 random dermal fields per each group with an area of 2000 to 8000 mm2.

Immunofluorescence Immunofluorescence was performed essentially as described before. Cryosections (10 mm) were fixed for 7 minutes with 4% formaldehyde in PBS followed by the permeabilization/blocking step in 10% donkey serum, 1% BSA, and 0.2% Triton X100 in PBS, pH 7.4, for 30 minutes. Specimens were further incubated for 45 minutes with rat anti-mouse CD3 (clone 17A2, BD Bioscience) with dilution 1:200. Donkey anti-goat IgG Alexa Fluor546 1:4000 was used as secondary antibody. Before and after antibody incubations, specimens were washed for 3 3 5 minutes with 0.05% Tween in PBS. Slides were mounted with VectaShield mounting medium containing 49,6-diamidino-2-phenylindol, dihydrochloride (Vector Laboratories, Burlingame, Calif) and analyzed with a Leica Fluorescence microscope (Leica Microsystems, Wetzlar, Germany).

RESULTS Overexpression of miR-146a inhibits proinflammatory factors in IFN-g– and TNF-a– stimulated human primary keratinocytes As demonstrated previously,E6 1767 genes were upregulated and 1021 downregulated in IFN-g–treated cells. Only 233 genes were determined to be upregulated and 61 genes downregulated in TNF-a–stimulated keratinocytes (data not shown). In unstimulated keratinocytes, the transfection of miR-146a resulted in the downregulation of 102 genes and the upregulation of 33 genes; and in IFN-g–stimulated cells, 410 genes were downregulated and 175 were upregulated by miR-146a (see Fig E1, A). Only 37 genes were downregulated and 12 genes were upregulated in response to miR-146a overexpression in TNF-a–stimulated keratinocytes, which showed higher levels of endogenous miR-146a (data not shown). According to Targetscan 6.2E7,E8 prediction and the array analysis, miR-146a had 1019 direct targets expressed

847.e3 REBANE ET AL

in keratinocytes (Fig E1, C). In unstimulated conditions, 39.2% of the downregulated genes (40 of 102) and 18% of the genes in IFN-g–stimulated keratinocytes (74 of 410) were predicted to be direct targets of miR-146a, suggesting that miR-146a indirectly targets many genes in IFN-g–stimulated cells (Fig E1, C). Although the downregulating effect of miR-146a was statistically significant only in the case of 37 genes in TNF-a–stimulated keratinocytes, there was a tendency for the suppression of the same set of genes that were suppressed in IFN-g–stimulated keratinocytes. In conclusion, these data together demonstrate that miR-146a suppresses the expression of numerous proinflammatory factors, including NF-kB signal transducers and direct miR-146a targets IRAK1 and CARD10 in unstimulated, TNF-a–, and IFNg–stimulated keratinocytes.

J ALLERGY CLIN IMMUNOL OCTOBER 2014

E2.

E3. E4.

E5.

E6.

E7. REFERENCES E1. Meyer N, Zimmermann M, Burgler S, Bassin C, Woehrl S, Moritz K, et al. IL-32 is expressed by human primary keratinocytes and modulates

E8.

keratinocyte apoptosis in atopic dermatitis. J Allergy Clin Immunol 2010;125: 858-65.e10. Zimmermann M, Koreck A, Meyer N, Basinski T, Meiler F, Simone B, et al. TNFlike weak inducer of apoptosis (TWEAK) and TNF-alpha cooperate in the induction of keratinocyte apoptosis. J Allergy Clin Immunol 2011;127:200-7, 207. e1-10. Griffiths-Jones S, Saini HK, van Dongen S, Enright AJ. miRBase: tools for microRNA genomics. Nucleic Acids Res 2008;36:D154-8. Reimand J, Kull M, Peterson H, Hansen J, Vilo J. g:Profiler–a web-based toolset for functional profiling of gene lists from large-scale experiments. Nucleic Acids Res 2007;35:W193-200. Boldin MP, Taganov KD, Rao DS, Yang L, Zhao JL, Kalwani M, et al. miR-146a is a significant brake on autoimmunity, myeloproliferation, and cancer in mice. J Exp Med 2011;208:1189-201. 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: 1092-102. Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res 2009;19:92-105. Garcia DM, Baek D, Shin C, Bell GW, Grimson A, Bartel DP. Weak seed-pairing stability and high target-site abundance decrease the proficiency of lsy-6 and other microRNAs. Nat Struct Mol Biol 2011;18:1139-46.

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

REBANE ET AL 847.e4

FIG E1. miR-146a inhibits the expression of proinflammatory factors in TNF-a– and IFN-g–stimulated keratinocytes. A-C, Keratinocytes were transfected either with control (cont) or with pre-miR-146a (miR146a) for 24 hours and then stimulated with IFN-g or TNF-a for 48 hours or left unstimulated (us). B, Scatter blots are shown in log10 scale. Black dots designate differentially expressed genes, red lines indicate a fold difference of 2.0, and r2 is the correlation coefficient. C, The Venn diagram represents the overlap of predicted miR-146a direct targets expressed in keratinocytes and miR-146a–suppressed genes. Relative expression levels of miR-146a are shown in log10 scale.

847.e5 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

FIG E2. Silencing of CARD10 and IRAK1 in human primary keratinocytes. Keratinocytes were transfected either with control (cont) siRNA or with siRNAs silencing IRAK1 or CARD10 or both for 24 hours and then stimulated with IFN-g for 48 hours or left unstimulated (us). A, Data represent mean 6 SEM. Statistical analysis was performed relative to cont. B, Western blot analysis of CARD10 and IRAK1 in primary keratinocytes transfected with indicated siRNAs. Mean intensity of each protein level was normalized to the GAPDH and to cont-transfected and us cells. GAPDH, Glyceraldehyde 3-phosphate dehydrogenase.

REBANE ET AL 847.e6

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

miR-146a 8

Relative expression

**

**

* 6

us IL-1β TNF-α

4 2

6a

co nt A14 LN

co nt A14 6a LN

LN

co

nt A14 6a

0

FIG E3. LNA-based inhibition of miR-146a in primary keratinocytes. Keratinocytes were transfected either with control (cont) LNA inhibitor or with miR-146a inhibitor (LNA-146a) for 24 hours and then stimulated with TNF-a or IL-1b for 48 hours or left unstimulated (us). LNA, Locked nucleic acid.

847.e7 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

FIG E4. The expression of proinflammatory cytokines and chemokines in the skin. WT and miR-146a2/2 (146a2/2) mice were topically treated with ethanol (control [cont]) or MC903 10 times on every other day. Skin samples were collected 24 hours after the last treatment. A, Immunohistochemical staining of TSLP on ear sections (bar 5 100 mm). Relative mRNA expression of TH2-type (B) and proinflammatory cytokines and chemokines (C). D, Immunofluorescence staining of CD3 on ear sections (bar 5 100 mm). DAPI, 4’-6Diamidino-2-phenylindole, dihydrochloride.

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

FIG E5. The function of miR-146a in keratinocytes and AD. During chronic skin inflammation in AD, the expression of miR-146a is increased in response to NF-kB activation in keratinocytes. This elevated expression of miR-146a, in turn, contributes to the suppression of proinflammatory factors, including CCL5, UBD, and CCL8, which can be activated by the stimulation of IFN-g and to a less extent by other AD-related factors. The suppression of IL-8 by miR-146a occurs through the targeting of NF-kB signaling transducers IRAK1 or CARD10, and the repression of UBD and CCL5 occurs through the targeting of CARD10. In human cells, CCL5 is a direct target of miR-146a. Suppression of IRAK1 and CARD10 does not affect the repression of CCL8 by miR-146a, which indicates that miR-146a has more direct targets in keratinocytes. It is not known how IFN-g activates CARD10 and NF-kB in keratinocytes. The increased AD-related expression of endogenous miR-146a in keratinocytes helps to balance the chronic inflammation in the skin.

REBANE ET AL 847.e8

847.e9 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

TABLE E1. The list of the genes differentially expressed in AD KCs* ILMN_GENE

hsa-miR-10b hsa-miR-146a hsa-miR-10a* hsa-miR-216a hsa-miR-10a hsa-miR-454* hsa-miR-29b-1* hsa-miR-92a-1* hsa-miR-99a* hsa-miR-30a hsa-miR-34a* hsa-miR-34c-5p

KC H average signal

KC H array SD

KC AD average signal

KC AD array SD

KC AD difference score

FD

34.1 498.4 334.5 95.8 3337.6 223.4 7678.7 757.8 260.2 10277.3 484.8 512.2

29.347 177.533 174.946 6.974 738.624 88.831 3416.671 169.899 35.403 2376.951 121.184 116.73

273 2459 972.4 260.8 8167.1 517.7 14535.9 1370.1 145.9 5416 243.1 251.2

38.728 1182.625 174.575 70.976 1117.816 115.96 2679.916 211.683 48.893 1444.568 97.456 95.271

343.752 18.155 55.496 37.096 124.694 29.936 16.855 41.233 220.459 221.476 215.418 220.231

8.01 4.93 2.91 2.72 2.45 2.32 1.89 1.81 0.56 0.53 0.50 0.49

FD, Fold difference; KC, keratinocytes. *Average expression signals of miRNAs with differential P < .05 (differential score >13 or <213) and with FD >1.8 or <0.55. The average signals and array SD values are calculated on the basis of results of 3 different donors from both healthy individuals (H) and patients with AD.

REBANE ET AL 847.e10

J ALLERGY CLIN IMMUNOL VOLUME 134, NUMBER 4

TABLE E2. Pathway analysis of genes suppressed by miR-146a in unstimulated keratinocytes* No. of genes in functional group

No. of genes in input list

No. of overlapped genes

Functional group

119

317

11

KEGG:04064

NF-kB signaling pathway

CXCL2, ICAM1, IL1B, IL8, IRAK1, NFKB1, NFKBIA, PLAU, RELB, TNFAIP3

2.31 3 1022

66

317

7

KEGG:04623

Cytosolic DNA-sensing pathway

2.46 3 1023

196

317

14

KEGG:04062

Chemokine signaling pathway

CCL5, IL1B, IL6, IRF3, NFKB1, NFKBIA, TMEM173 BCAR1, CCL20, CCL5, CCL8, CXCL1, CXCL2, CXCL5, IL8, NFKB1, NFKBIA, PPBP, PXN, RAC2, STAT2

4.88 3 1022

291

317

15

KEGG:04060

Cytokine-cytokine receptor interaction

3.55 3 1024

65

317

9

KEGG:04621

NOD-like receptor signaling pathway

2.85 1022

49

317

6

KEGG:00480

Glutathione metabolism

160

317

8

KEGG:04630

Janus kinase-STAT signaling pathway

P value 23

1.68 3 10

1.00

Functional group name

*The P value from Fisher exact test showing the significance of the overlap between the target list and the indicated functional group.

Overlapped genes

BMP2, CCL20, CCL5, CCL8, CLCF1, CSF2, CXCL1, CXCL2, CXCL5, IL1B, IL1R2, IL6, IL8, PPBP, TNFRSF1B CCL5, CXCL1, CXCL2, IL1B, IL6, IL8, NFKB1, NFKBIA, TNFAIP3 GSS, GSTO1, GSTT1, MGST3, ODC1, PGD CCND3, CLCF1, CSF2, IL13RA2, IL6,SOCS1, STAT2, STAT5A

847.e11 REBANE ET AL

J ALLERGY CLIN IMMUNOL OCTOBER 2014

TABLE E3. Predicted miR-146a binding sites according to Targetscan 6.240 Conserved sites* Target gene

IRAK1 CARD10 CCL5

Poorly conserved sites

Total

8mer

7mer-m8

7mer-1A

Total

8mer

7mer-m8

7mer-1A

Total context scorey

2 1 0

2 1 0

0 0 0

0 0 0

0 1 1

0 0 1

0 0 0

0 1 0

20.70 20.26 20.40

*Conserved 5 conserved across most mammals, but usually not beyond placental.  The context1 score for a specific site is the sum of the contribution of 6 features (site type, 39 pairing, local AU, position, target-site abundance, and seed-pairing stability contributions) and is calculated as in Li et al.41