MicroRNA-31 Is Overexpressed in Psoriasis and Modulates Inflammatory Cytokine and Chemokine Production in Keratinocytes via Targeting Serine/Threonine Kinase 40 This information is current as of September 24, 2021. Ning Xu, Florian Meisgen, Lynn M. Butler, Gangwen Han, Xiao-Jing Wang, Cecilia Söderberg-Nauclér, Mona Ståhle, Andor Pivarcsi and Enikö Sonkoly J Immunol 2013; 190:678-688; Prepublished online 10 December 2012; Downloaded from doi: 10.4049/jimmunol.1202695 http://www.jimmunol.org/content/190/2/678 http://www.jimmunol.org/ Supplementary http://www.jimmunol.org/content/suppl/2012/12/10/jimmunol.120269 Material 5.DC1 References This article cites 46 articles, 7 of which you can access for free at: http://www.jimmunol.org/content/190/2/678.full#ref-list-1

Why The JI? Submit online. by guest on September 24, 2021

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2013 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

MicroRNA-31 Is Overexpressed in Psoriasis and Modulates Inflammatory Cytokine and Chemokine Production in Keratinocytes via Targeting Serine/Threonine Kinase 40

Ning Xu,* Florian Meisgen,* Lynn M. Butler,† Gangwen Han,‡ Xiao-Jing Wang,‡ Cecilia So¨derberg-Naucle´r,† Mona Sta˚hle,* Andor Pivarcsi,* and Eniko¨ Sonkoly*

Psoriasis is characterized by a specific microRNA expression profile, distinct from that of healthy skin. MiR-31 is one of the most highly overexpressed microRNAs in psoriasis skin; however, its biological role in the disease has not been studied. In this study, we show that miR-31 is markedly overexpressed in psoriasis keratinocytes. Specific inhibition of miR-31 suppressed NF-kB–driven promoter luciferase activity and the basal and TNF-a–induced production of IL-1b, CXCL1/growth-related oncogene-a, CXCL5/

epithelial-derived neutrophil-activating peptide 78, and CXCL8/IL-8 in human primary keratinocytes. Moreover, interference with Downloaded from endogenous miR-31 decreased the ability of keratinocytes to activate endothelial cells and attract leukocytes. By microarray expression profiling, we identified genes regulated by miR-31 in keratinocytes. Among these genes, we identified serine/threonine kinase 40 (STK40), a negative regulator of NF-kB signaling, as a direct target for miR-31. Silencing of STK40 rescued the suppressive effect of miR-31 inhibition on cytokine/chemokine expression, indicating that miR-31 regulates cytokine/chemokine expression via targeting STK40 in keratinocytes. Finally, we demonstrated that TGF-b1, a cytokine highly expressed in psoriasis epidermis, upregulated miR-31 expression in keratinocytes in vitro and in vivo. Collectively, our findings suggest that overexpres- http://www.jimmunol.org/ sion of miR-31 contributes to skin inflammation in psoriasis lesions by regulating the production of inflammatory mediators and leukocyte chemotaxis to the skin. Our data indicate that inhibition of miR-31 may be a potential therapeutic option in psoriasis. The Journal of Immunology, 2013, 190: 678–688.

soriasis is a common chronic inflammatory skin disease, epidermis (1). There is a close interdependence between kerati- which affects 2–3% of the population. It is a lifelong nocytes and immune cells in psoriatic skin: the cytokines and P disease with spontaneous remissions and exacerbations chemokines secreted by keratinocytes, such as IL-1b,TNF-a, that are severely detrimental to the patients’ quality of life (1). CXCL1/growth-related oncogene-a, CXCL5/epithelial-derived Psoriasis skin lesions are typically characterized by keratinocyte neutrophil-activating peptide 78, and CXCL8/IL-8 activate and by guest on September 24, 2021 hyperproliferation and aberrant differentiation, increased vascu- attract immune cells to migrate into epidermis and dermis; immune larity in the dermis, and infiltration of inflammatory cells, such as cell–derived cytokines, in turn, act on keratinocytes to increase the macrophages, neutrophils, and lymphocytes into the dermis and expression of inflammatory genes, promote keratinocyte prolifer- ation, and impair keratinocyte differentiation (reviewed in Ref. 1). microRNAs (miRNAs) are ∼ 22nt long single-stranded noncoding *Molecular Dermatology Research Group, Unit of Dermatology and Venereology, Department of Medicine, Karolinska Institute, SE-17176 Stockholm, Sweden; RNAs that mediate posttranscriptional silencing by binding with †Cell and Molecular Immunology Group, Department of Medicine, Karolinska In- ‡ partial complementarity to the 39-untranslated region (39-UTR) of stitute, SE-17176 Stockholm, Sweden; and Department of Pathology, University of ∼ Colorado Denver, Aurora, CO 80045 target mRNA (2). miRNAs are estimated to regulate 60% of all protein-coding genes in humans and have been shown to participate Received for publication September 26, 2012. Accepted for publication November 5, 2012. in the regulation of almost every cellular process investigated to date This work was supported by the Swedish Research Council (Vetenskapsra˚det), the (3). We and others have previously identified a distinct miRNA Swedish Society of Medicine (Svenska La¨karesa¨llskapet), the Swedish Psoriasis expression profile in psoriasis skin compared with healthy skin (4– Association (Psoriasisfo¨rbundet), the European Skin Research Foundation, the 8). Several of these deregulated miRNAs have been shown to act on Welander and Finsens Foundation, the Tore Nilssons Foundation, the Lars Hierta Memorial Foundation, the Sigurt and Elsa Golje Memorial Foundation, the Centre of cellular processes crucial for psoriasis. For example, miR-203 (7), Excellence for Research on Inflammation and Cardiovascular Disease, the Karolinska miR-125b (9), miR-424 (4), and miR-99a (6) regulate keratinocyte Institute, and the Stockholm County Council. N.X. was supported by the Swedish Society for Medical Research. A.P. was supported by the LEO Pharma Research proliferation and differentiation, whereas miR-21 suppresses T cell Foundation. E.S. was supported by the Stockholm County Council. apoptosis (10). Although being in the infancy, these studies reveal The datasets presented in this article have been submitted to the National Center for important roles for miRNAs in the biology of psoriasis. Biotechnology Information’s Gene Expression Omnibus (http://www.ncbi.nlm.nih. In this study, we identify a function for miR-31 in the context of gov/geo/) under accession number GSE41905. psoriasis. We show that specific inhibition of miR-31, a miRNA Address correspondence and reprint requests to Dr. Ning Xu, Molecular Dermatology overexpressed in psoriasis keratinocytes, suppresses NF-kBsig- Research Group, Centre for Molecular Medicine, L8:02, Department of Medicine, Karolinska Institute, SE-17176 Stockholm, Sweden. E-mail address: [email protected] naling and the production of IL-1b, CXCL1, CXCL5, and CXCL8/ The online version of this article contains supplemental material. IL-8. Serine/threonine kinase 40 (STK40), a negative modulator of Abbreviations used in this article: GSEA, gene set enrichment analysis; K5, keratin 5; NF-kB signaling, was identified as a direct target for miR-31. LNA, locked nucleic acid; miRNA, microRNA; qRT-PCR, quantitative real-time Furthermore, we identify TGF-b1, a cytokine highly expressed in PCR; siRNA, small interfering RNA; STK40, serine/threonine kinase 40; Treg, psoriasis epidermis, as a potent regulator of miR-31 expression in regulatory T cell; 39-UTR, 39-untranslated region; WT, wild-type. keratinocytes in vitro and in vivo. Taken together, our results Copyright Ó 2013 by The American Association of Immunologists, Inc. 0022-1767/13/$16.00 suggest that suppressing miR-31 in psoriasis skin may alleviate www.jimmunol.org/cgi/doi/10.4049/jimmunol.1202695 The Journal of Immunology 679 inflammation by interfering with the cross-talk between kerati- 2100 Bioanalyzer and Nanodrop ND-1000. One hundred nanograms of nocytes and immune cells. total RNA was used to prepare cDNA following the Affymetrix 39IVT Express Kit labeling protocol. Standardized array processing procedures recommended by Affymetrix including hybridization, fluidics processing, Materials and Methods and scanning were used. Genes showing at least 1.2-fold regulation, and Patients p , 0.05 was considered to be differentially expressed. Gene set enrich- ment analysis (GSEA) was performed using public software from Broad Four-millimeter punch biopsies were taken, after informed consent, from Institute (14, 15). The data discussed in this publication have been de- nonlesional (n = 20) and lesional skin (n = 43) of patients with moderate or posited in the National Center for Biotechnology Information’s Gene severe chronic plaque psoriasis and from noninflamed, nonirritated skin of Expression Omnibus (16) and are accessible through GEO Series ac- healthy individuals (n = 35). The psoriasis patients had not received sys- cession number GSE41905 (http://www.ncbi.nlm.nih.gov/geo/query/acc. temic immunosuppressive treatment or psoralen1ultraviolet A/solarium/ cgi?acc=GSE41905). UV for at least 1 mo and topical therapy for at least 2 wk before skin biopsy. ELISA RNA extraction and quantitative real-time PCR The culture medium was collected from transfected keratinocytes at the Total RNA was extracted from tissues and cells using miRNeasy Mini kit indicated time points and stored at 280˚C. The protein levels of CXCL1, (Qiagen). Skin biopsies from human and K5.TGF-b1 mice (11, 12) were CXCL5, and CXCL8/IL-8 were analyzed by ELISA (R&D Systems) fol- homogenized in liquid nitrogen using a Mikro-Dismembrator U (Braun lowing the manufacturer’s instructions. Biotech) prior to RNA extraction. Epidermal cells were isolated from punch skin biopsies by dispase treatment and trypsin digestion as described Leukocyte chemotaxis in Ref. 13. CD45-negative cells were subsequently isolated using MACS separation columns (Miltenyi Biotec), according to the manufacturer’s Human leukocytes were isolated from 0.2% EDTA anticoagulated whole instructions. Quantification of miR-31 by TaqMan Real-Time PCR was blood collected by venipuncture from healthy donors. Erythrocytes were Downloaded from performed as described previously (7). Its expression was normalized removed using dextran sedimentation (1:1 mixture of blood: 6% dextran/0.9% between different samples based on the values of U48 RNA expression in NaCl), followed by one or two rounds of hypotonic lysis using ddH2O. human and snoRNA 251 in mouse. The primary miRNA transcripts were Purified leukocytes were suspended in EpiLife serum-free keratinocyte 5 quantified by TaqMan Pri-miRNA assay (Hs03302684_pri; Applied Bio- growth medium, and 3 3 10 cells were added to the inner chamber of systems) following the manufacturer’s instructions. To quantify mRNAs, a BD Falcon Cell Culture Insert. Leukocytes migrate through a 3-mm 500 ng total RNA was reverse transcribed using the RevertAid First Strand porosity polyethylene terephthalate membrane toward the outer chamber cDNA Synthesis Kit (Fermentas). The mRNAs of IL-1b, CXCL1, CXCL5, containing culture medium from keratinocytes, which were transfected CXCL8/IL-8, STK40, ICAM-1, VCAM-1, and E-selectin were quantified with anti–miR-31 or anti–miR-Ctrl for 72 h. After incubation for 3 h at 37˚C http://www.jimmunol.org/ by TaqMan gene expression assays (Applied Biosystems). Target gene in 5% CO2, the migrating cells in the medium of the outer chamber were expression was normalized based on the expression of the internal positive quantified by CyQUANT GR dye (Life Technologies) staining. control 18S RNA: 59-CGGCTACCACATCCAAGGAA-39 (forward), 59- Endothelial cell activation GCTGGAATTACCGCGGCT-39 (reverse), and 59-FAM-TGCTGGCAC- CAGACTTGCCCTC-TAMRA-39 (probe). HUVECs were isolated as previously described (17) and maintained in Medium 199 (Invitrogen) containing 20% FCS, 28 mg/ml gentamicin, 2.5 In situ hybridization mg/ml amphotericin B, 1 ng/ml epidermal growth factor, and 1 mg/ml In situ hybridization was performed on frozen sections (10 mm in thickness) hydrocortisone (all from Sigma-Aldrich). Second-passage HUVECs were of skin biopsy specimens as described previously (9). Briefly, after incu- treated with keratinocytes culture medium for 4 h and then harvested. bated in acetylation solution (0.06 M HCl, 1.3% trietanolamin, and 0.6% Luciferase reporter assays by guest on September 24, 2021 acetic anhydride in diethyl pyrocarbonate-treated water) for 10 min at room temperature, sections were incubated in permeabilization buffer (1% Renilla luciferase reporter plasmids containing synthetic sequence repeats of Triton X-100) for 30 min at room temperature, washed, and prehybri- that are fully complementary to miR-31 (miR-31 sensor) or 39-UTR of the dized for 1 h at 50˚C. Hybridization with digoxigenin-labeled miRCURY STK40 gene-cloned downstream of the reporter gene and empty luciferase locked nucleic acid (LNA) probes (Exiqon) was performed over night at vector were obtained from SwitchGear Genomics. The mutations were gen- 50˚C. Slides were then washed four times with 23 SSC buffer followed by erated with the predicted target site of STK40 39-UTR using the QuickChange one time with 0.13 SSC buffer at 67˚C. The probe binding was detected XL site-directed mutagenesis kit (Stratagene), according to the manu- by incubating the sections with alkaline phophatase–conjugated sheep anti- facturer’s instructions. NF-kB reporter plasmid pGL4.32[luc2P/NF-kB- digoxigenin Fab fragments (1:2500 [Roche]) for 1 h at room temperature. RE/Hygro] Vector was obtained from Promega, which contains five copies Sections were visualized by adding BM purple alkaline phophatase sub- of an NF-kB response element that drives transcription of the luciferase re- strate (Roche), according to the manufacturer’s instructions. porter gene luc2P (Photinus pyralis). Human primary keratinocytes growing in 24-well plates were cotransfected with the luciferase reporters (25 ng/ml) Cells culture and treatments together with 10 nM anti–miR-31 or anti–miR-Ctrl using FuGENE HD Human adult epidermal keratinocytes used in miR-31 function study were transfection reagent (Promega). Luciferase activity was analyzed 24 h purchased from Cascade Biologics and cultured in EpiLife serum-free posttransfection using LightSwitch Luciferase Assay reagent (SwitchGear) keratinocyte growth medium including Human Keratinocyte Growth or Dual-Luciferase Reporter Assay System (Promega). Supplement at a final Ca2+ concentration of 0.06 mM (Cascade Biologics) Immunohistochemistry at 37˚C in 5% CO2. To study the biological effects of miR-31 on kerati- nocytes, third passage keratinocytes at 30–50% confluence were trans- STK40 protein expression was analyzed in both frozen and formalin-fixed fected with 10 nM miRIDIAN miR-31 hairpin inhibitor or miRNA hairpin paraffin embedded skin sections (7 mm in thickness) using rabbit anti- inhibitor negative control number 1 (ThermoFisher Scientific), 30 nM human STK40 Ab (1:200) (Sigma-Aldrich) and the avidin-biotin-peroxidase Silencer select predesigned small interfering RNA (siRNA) for STK40 complex staining system (Vector Laboratories) following the manu- (siRNA ID: s38327), or siRNA-negative control number 1 (Ambion) using facturer’s instructions. Replacement of the primary Ab with rabbit Ig Lipofectamine 2000 (Invitrogen). Keratinocytes were treated with TNF-a fraction (DakoCytomation) in the staining process was used as negative (50 ng/ml; Immunotools) or TGF-b1 (3 ng/ml; R&D Systems) at the in- control (Supplemental Fig. 3). dicated time points. Three-dimensional epidermal equivalents were pur- Statistics chased from MatTek. TGF-b1 (10 ng/ml; R&D systems) was added into the culture medium. Seventy-two hours later, the total RNA was extracted Statistical significance for experiments was determined by Mann–Whitney with miRNeasy Mini kit (Qiagen). U test or Student t test. Correlation between the expression of different genes in the same samples was made using Pearson’s correlation test on log- Gene expression microarray transformed data. p , 0.05 was considered to be statistically significant. Expression profiling of primary human keratinocytes transfected with 10 nM Study approval anti–miR-31 or anti–miR-Ctrl for 48 h (in triplicates) was performed using Affymetrix Genechip system at the Microarray core facility of Karolinska The clinical materials were taken after patients’ consent and the study was Institute. In brief, total RNA was extracted using the miRNeasy Mini Kit approved by the Stockholm Regional Ethics Committee and conducted (Qiagen), and RNA quality and quantity were determined using Agilent according to the Declaration of Helsinki’s principles. All procedures in- 680 miR-31 IN PSORIASIS volving mice were approved by Institutional Animal Care and Use Com- expression of mature miR-31 in these skin biopsies, suggesting that mittee at Colorado School of Public Health (University of Colorado the upregulation of miR-31 in psoriasis skin may occur at the Denver, Aurora, CO). transcriptional level, rather than during processing of the primary transcript. Results To investigate which cell type(s) in the skin expresses miR-31, miR-31 is upregulated in psoriasis keratinocytes we surveyed miR-31 expression in a panel of isolated primary Earlier expression profiling data from our and other groups showed human cell types. qRT-PCR results showed that miR-31 was mainly that miR-31 is one of the miRNAs overexpressed in psoriasis skin expressed by keratinocytes, fibroblasts and melanocytes isolated (5, 7, 8). To confirm these data, we measured mature miR-31 from the skin (Supplemental Fig. 1). To identify which cell type(s) expression in skin biopsies from healthy donors (n = 14), nonle- in the skin are primarily responsible for the increased expression sional (n = 10), and lesional skin (n = 25) from psoriasis patients of miR-31 in psoriasis, we performed in situ hybridization on skin using quantitative real-time PCR (qRT-PCR) (Fig. 1A). miR-31 sections from healthy individuals (n = 11), nonlesional (n = 6), and expression was found to be dramatically increased in psoriasis lesional skin (n = 11) from psoriasis patients using miR-31–specific lesional skin compared with healthy skin (33-fold change; p =4.13 LNA-modified probes (Fig. 1C). The expression of miR-31 was 2 10 7) and compared with psoriasis nonlesional skin (14-fold change; low in healthy skin and restricted to the basal cell layer of the 2 p =8.13 10 7). In addition, we found that the expression of pri- epidermis. In contrast, miR-31 expression was higher in psoriasis mary miR-31 transcript (pri-miR-31), from which the mature miR- nonlesional skin and appeared at both basal and suprabasal cell 31 is processed, was also upregulated in psoriasis lesional skin (n = layers. In lesional skin from the same psoriasis patients, miR-31

9) compared with healthy skin (n = 7) (Fig. 1B). The expression of expression was further upregulated and mainly detected in the Downloaded from pri-miR-31 correlated positively (R = 0.7659; p = 0.0014) with the suprabasal layers. http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 1. The expression of miR-31 in healthy and psoriasis skin. (A) MiR-31 expression was analyzed in healthy (n = 14), psoriasis nonlesional (n =10),and lesional skin samples (n = 25) using qRT-PCR. Results for individual patients and mean are shown. (B) Expression of primary miR-31 transcript (pri-miR-31) was analyzed in healthy (n = 7) and psoriasis lesional skin (n = 9) by qRT-PCR. (C) In situ hybridization was performed on healthy skin (upper panel, n = 11), psoriasis nonlesional skin (middle panel, n = 6), and lesional skin sections (lower panel, n = 11) using miR-31–specific LNA probe (left panel) or scrambled probe (right panel). Blue-purple color indicates miR-31 expression. Scale bar, 50 mm. (D) miR-31 expression was analyzed on RNA from epidermal keratinocytes (CD45neg) isolated from healthy skin (n = 10), psoriasis nonlesional (n =4),andlesionalskin(n = 7) using qRT- PCR. **p , 0.01, ***p , 0.001; Mann–Whitney U test. The Journal of Immunology 681

To quantify the change of miR-31 level in epidermal cells, we the expression of the host gene of miR-31, encoding a long measured miR-31 expression in sorted epidermal CD45 (common noncoding RNA, LOC554202 (18), was not affected by miR-31 leukocyte Ag)-negative cells from healthy (n =10),psoriasis inhibition, neither that of the adjacent IFN ε (IFNE1) gene nonlesional (n = 4), and lesional skin (n = 7) (Fig. 1D). In line with (Supplemental Fig. 2C). the results of in situ hybridization, qRT-PCR analysis revealed Interestingly, we found that several mediators of key importance a 13-fold (p = 0.0001) higher miR-31 expression in CD45neg cells in psoriasis pathogenesis, such as IL-1b,CXCL1,CXCL5,and sorted from psoriasis lesional skin compared with those obtained CXCL8/IL-8, were downregulated by anti–miR-31 in keratinocytes from healthy skin and a 9-fold (p = 0.0061) increase compared (Supplemental Table I), and this was validated by qRT-PCR on with psoriasis nonlesional skin. Collectively, our results demon- keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl for strate that miR-31 is upregulated in keratinocytes in psoriasis skin 24, 48, 72, and 96 h (Fig. 2A). Consistently, the amount of these lesions. chemokines secreted into the culture medium was decreased by anti–miR-31, shown by ELISA (Fig. 2B). Notably, IL-1b was not miR-31 regulates the production of inflammatory mediators in detectable in the culture medium, because the cultured human ker- keratinocytes atinocytes produce but do not process the IL-1b precursor into its To study the biological role of miR-31 in keratinocytes, we biologically active form because of the lack of IL-1 convertase (19). transfected miR-31 hairpin inhibitor (anti–miR-31) into primary Next, we aimed to test whether inhibition of miR-31 can effi- human keratinocytes to inhibit endogenous miR-31. Inhibition of ciently suppress the production of IL-1b, CXCL1, CXCL5, and miR-31 was confirmed by qRT-PCR analysis of miR-31 expres- CXCL8/IL-8 under inflammatory conditions. To this end, we mea- sion and by luciferase assay using a synthetic miR-31-target as sured the effect of miR-31 inhibition on mRNA and protein levels Downloaded from sensor (Supplemental Fig. 2A, 2B). We performed a global tran- of these cytokine/chemokines in keratinocytes treated with the scriptome analysis of keratinocytes upon suppression of endoge- proinflammatory cytokine TNF-a (Fig. 2C, 2D). Consistent with nous miR-31 using Affymetrix arrays, which identified 234 genes earlier reports, TNF-a induced the expression of IL-1b, CXCL1, that were significantly changed (Supplemental Table I). Of note, CXCL5, and CXCL8/IL-8. Inhibition of endogenous miR-31 http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 2. Inhibition of miR-31 suppresses the production of cytokines/chemokines by keratinocytes. The expression and secretion of CXCL1, CXCL8, CXCL5, and IL-1b were analyzed in keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl for 24–96 h by qRT-PCR (A) and ELISA (B) or analyzed in keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl for 48 h and then treated with TNF-a for 6 h by qRT-PCR (C) and ELISA (D). *p , 0.05, **p , 0.01, ***p , 0.001; Student t test. 682 miR-31 IN PSORIASIS suppressed TNF-a–induced cytokine/chemokine production both miR-31 regulates the NF-kB pathway in keratinocytes at mRNA (Fig. 2C) and protein level (Fig. 2D). The NF-kB pathway is one of the signaling pathways commonly miR-31 regulates the endothelial cell–activating and regulating the expression of IL-1b, CXCL1, CXCL5, and CXCL8/ leukocyte-attracting capacity of keratinocytes IL-8. We therefore investigated whether other target genes of the NF-kB pathway are also regulated by miR-31 in keratinocytes. To Because CXCL1, CXCL5, and CXCL8 have the ability to activate answer this question, we evaluated the enrichment for known endothelial cells and recruit polymorphonuclear leukocytes into target genes of the NF-kB pathway (summarized on the Web site: inflamed tissues (20), we examined whether the capacity of kera- http://bioinfo.lifl.fr/NF-KB/) in our microarray data of keratino- tinocytes to attract leukocytes was affected by miR-31. The first cytes transfected with anti–miR-31 or anti–miR-Ctrl (Fig. 4A). step of recruitment of leukocytes to the skin is the attachment of Results of the GSEA (14, 15) revealed that NF-kB pathway target circulating cells to vascular endothelial cells, which is mediated genes were significantly enriched among the genes downregulated by cell adhesion molecules, such as ICAM-1, VCAM-1, and E- by anti–miR-31, and a negative enrichment score curve was gen- selectin (21). It has been shown that the expression of these cell erated by GSEA (p , 0.001). These data indicated that inhibition adhesion molecules is increased on dermal vessels of psoriatic skin of miR-31 might affect the activity of the NF-kB signal trans- (22–25). To test the effect of miR-31 on endothelial cell-activating duction pathway. capacity of keratinocytes, we measured the expressions of ICAM- To test this hypothesis, we determined the effect of miR-31 1, VCAM-1, and E-selectin in primary HUVECs incubated with inhibitor on NF-kB–dependent promoter luciferase reporter gene culture medium from miR-31 inhibitor–treated keratinocytes. qRT- activity in human primary keratinocytes (Fig. 4B). Results of lu- PCR results demonstrated that inhibition of endogenous miR-31 in Downloaded from ciferase assays demonstrated that inhibition of miR-31 suppressed keratinocytes decreased their capacity to induce the expression of both the basal and TNF-a–induced NF-kB–dependent promoter these cell adhesion molecules in endothelial cells (Fig. 3A). luciferase activity, indicating that miR-31 regulates the activity of Next, we considered whether altered expression of miR-31 in NF-kB pathway in keratinocytes. keratinocytes could affect their capacity to attract leukocytes. To address this question, we performed migration assays with pe- miR-31 targets STK40, a negative regulator of the NF-kB

ripheral blood leukocytes using conditioned supernatant from miR- pathway http://www.jimmunol.org/ 31 inhibitor–treated keratinocytes (Fig. 3B). The results revealed Next, we aimed to identify the molecular mechanism by which that supernatant from keratinocytes with inhibited miR-31 ex- miR-31 modulates the NF-kB pathway. miRNAs exert biological pression attracted less leukocytes compared with medium from functions by regulating their target genes. To predict target genes control treated cells. Taken together, these data show that inhibi- of miR-31, we used three different public algorithms, TargetScan tion of miR-31 in keratinocytes results in decreased endothelial cell activation and leukocyte chemotaxis. by guest on September 24, 2021

FIGURE 4. miR-31 regulates the NF-kB pathway in keratinocytes. (A) Microarray analysis was performed in independent biological triplicates for primary human keratinocytes transfected with either anti–miR-31 or anti– FIGURE 3. miR-31 regulates the endothelial cell–activating and leuko- miR-Ctrl. Genes represented in the profile data set were ranked by fold cyte-attracting capacity of keratinocytes. (A) HUVECs were treated with change (anti–miR-31/anti–miR-Ctrl). GSEA evaluated enrichment within the medium from cultured keratinocytes transfected with anti–miR-31 or anti– profile data set for the reported target genes of NF-kB signaling pathway. miR-Ctrl, and the expression of endothelial cell activation markers (ICAM-1, Vertical bars along the x axis of the GSEA plot denote the positions of NF-kB VCAM-1, and E-selectin) was analyzed by qRT-PCR. (B) Human leukocytes target genes within the ranked list. NES, normalized enrichment score. (B) chemotaxis toward the unused medium (med) or the medium from cultured Keratinocytes were transfected with NF-kB luciferase reporter plasmid keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl for 54 h and then 48 h after the transfection of anti–miR-31 or anti–miR-Ctrl. Eighteen hours treated with or not with TNF-a for 18 h. The migrating cells were quantified later, the cells were treated with or not with TNF-a for 5 h, and luciferase by CyQUANT GR dye staining. *p , 0.05, **p , 0.01; Student t test. activity was measured. **p , 0.01, ***p , 0.001; Student t test. The Journal of Immunology 683

(26), miRanda (27), and PicTar (28). Within the ranked gene list The gene showing highest up-regulation upon miR-31 inhibition from microarray analysis of keratinocytes transfected with either was STK40 (Supplemental Table I), which had been identified as anti–miR-31 or anti–miR-Ctrl, the enrichment of these predicted a suppressor of NF-kB–mediated transcription (29). The regulation miR-31 targets was evaluated with GSEA (Fig. 5A). Most pre- of STK40 expression by miR-31 has been previously observed in dicted target genes were found to be upregulated upon inhibition ovarian cancer cells (30). STK40 contains two conserved potential of miR-31, and GSEA generated positive enrichment score curves, binding sites for miR-31 in its 39-UTR (Fig. 5B), and the dere- irrespective of which prediction algorithm was used, indicating pression of STK40 transcript upon inhibition of miR-31 in human that the microarray analysis was specific and sensitive for detecting primary keratinocytes was confirmed by qRT-PCR analysis (Fig. direct target genes for miR-31. 5C). To determine whether STK40 is a bona fide target of miR-31, Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 5. MiR-31 targets STK40, a negative regulator of the NF-kB signaling pathway. (A) GSEA evaluated the enrichment of predicted miR-31 targets, as determined by the given algorithms (TargetScan, miRanda, and PicTar), within the microarray profiling data set of keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl. (B) Nucleotide resolution of the predicted miR-31 binding site in 39-UTR of STK40 mRNA: seed sequence (green letters); and mutated miR-31 binding sites (red letters). (C) The miR-31 mediated regulation of STK40 was verified by qRT-PCR in keratinocytes transfected with anti–miR-31 or anti–miR-Ctrl for 24–96 h. (D) Keratinocytes were transfected with luciferase reporter plasmid containing WT or mutant (Mut) STK40 39- UTR or empty vector (Vector) together with anti–miR-31 or anti–miR-Ctrl. **p , 0.01, ***p , 0.001; Student t test. (E) Human skin from the same donors was used for the detection of miR-31 by in situ hybridization (left panel) and STK40 expression by immunohistochemistry (right panel). Red-brown color indicates STK40 expression. Scale bar, 50mm. 684 miR-31 IN PSORIASIS we performed 39-UTR luciferase reporter assays with luciferase signal of STK40 was detected also in the basal and spinous layers reporter gene constructs containing the full-length 39-UTR of of the epidermis. In psoriasis nonlesional skin, the strong signal of STK40 mRNA in human primary keratinocytes (Fig. 5D). The STK40 in granular layers was similar as in healthy skin, whereas wild-type (WT) STK40 39-UTR luciferase activity was increased the expression of STK40 in basal and spinous layers was slightly 8-fold (p = 0.004) by anti–miR-31 in comparison with anti–miR- decreased. In psoriasis lesional skin, the STK40 signal was absent Ctrl. Mutation of one of the predicted target sites (Mut1) markedly in the lower spinous layers but strongly expressed in the upper decreased luciferase activity upon miR-31 inhibition, although the spinous layers, which is opposite to the pattern of miR-31 ex- luciferase activity was still affected by anti–miR-31. Mutation of pression in psoriasis lesional skin. No STK40-positive leukocytes target site #2 (Mut 2) or both target sites (Mut1+2), however, were observed in the dermis. The reciprocal expression of miR-31 completely abolished the effect of miR-31 inhibition on reporter and STK40 further supports that STK40 is regulated by miR-31 in gene expression. These data demonstrate that miR-31 directly psoriasis skin. regulates STK40 expression through binding to the two predicted target sites in its 39-UTR. miR-31 regulates cytokine/chemokine expression via targeting Next, we analyzed the expression of STK40 in healthy (n = 11), STK40 psoriasis nonlesional (n = 9), and lesional skin (n = 11) by im- To determine whether the observed effects of miR-31 on kerati- munohistochemistry. miR-31 expression in the same donors was nocyte cytokine/chemokine production are, at least partially, me- detected by in situ hybridization (Fig. 5E, Supplemental Fig. 3). diated through STK40, we analyzed the effects of silencing of STK40 expression was mainly detected in the cytoplasm and the STK40 expression by siRNA in keratinocytes (Fig. 6). STK40 plasma membrane of epidermal keratinocytes. In healthy skin, expression in keratinocytes was significantly increased by anti– Downloaded from STK40 was strongly expressed in the granular layers and a weaker miR-31, which was prevented by simultaneous transfection with http://www.jimmunol.org/ by guest on September 24, 2021

FIGURE 6. MiR-31 regulates cytokine/che- mokine expression via targeting STK40. Kera- tinocytes were cotransfected with siRNA and anti-miR for 48 h and then treated with TNF-a for 6 h. STK40 expression was detected by qRT-PCR (A). The expression and secretion of IL-1b, CXCL1, CXCL5, and CXCL8/IL-8 were analyzed by qRT-PCR (B) and ELISA (C). (D) Keratinocytes were transfected with an NF-kB reporter plasmid 48 h after the transfection of anti-miR and siRNA. Eighteen hours later, the cells were treated with TNF-a for 5 h, and the luciferase activity was measured. *p , 0.05, **p , 0.01, ***p , 0.001; Student t test. The Journal of Immunology 685

STK40 siRNAs (Fig. 6A). In line with our previous data, the basal a multilayered artificial epidermis built with human primary ker- and TNF-a–induced expression of IL-1b, CXCL1, CXCL5, and atinocytes on an air-liquid interphase (Fig. 7A). CXCL8/IL-8 was decreased by anti–miR-31 in keratinocytes. To investigate whether miR-31 is regulated by TGF-b1invivo, However, silencing of STK40 rescued the suppressive effect of we measured miR-31 expression in the skin of K5.TGF-b1 trans- anti–miR-31 on IL-1b, CXCL1, CXCL5, CXCL8/IL-8, which was genic mice, which overexpress this cytokine in an epidermis- shown both at mRNA level by qRT-PCR (Fig. 6B), and at protein specific manner and develop a psoriasis-like skin disorder (11, 12). level by ELISA (Fig. 6C). These data indicate that the effect of qRT-PCR results showed that the expression of miR-31 was in- miR-31 on the production of inflammatory mediators requires creased 35-fold (p = 0.01) in the skin of K5.TGF-b1 mice com- STK40. In line with these results, knockdown of STK40 expres- pared with WT littermates (Fig. 7B). Treatment with etanercept, sion rescued the suppressive effect of anti–miR-31 on NF-kB– which blocks TNF-a activity and is one of the current therapies dependent promoter luciferase activty (Fig. 6D). These data in- for psoriasis, efficiently alleviated the psoriasis phenotype of K5. dicate that the regulation of keratinocyte cytokine/chemokine TGF-b1 mice (11) and resulted in a 4-fold (p = 0.005) reduction production by miR-31 is, at least partially, mediated through of miR-31 expression compared with the K5.TGF-b1 mice treated STK40. with normal saline (Fig. 7B). In line with this, in situ hybridization showed that miR-31 was upregulated in the hyperplastic epidermis b miR-31 is upregulated by TGF- 1 in keratinocytes both in vitro of K5.TGF-b1 mice compared with WT littermates, whereas and in vivo treatment with etanercept reduced epidermal hyperplasia and To understand the mechanism of miR-31 overexpression in pso- miR-31 expression (Fig. 7C). Taken together, our results suggest riasis keratinocytes, we systematically surveyed the effect of that increased levels of TGF-b1 in psoriatic skin may contribute to Downloaded from cytokines, growth factors, and cell differentiation on miR-31 ex- the upregulation of miR-31 in psoriasis. pression in keratinocytes. Cells were treated with cytokines rele- vant for psoriasis pathology (TNF-a, IL-22, TGF-b1, IL-6, IL-20, Discussion IFN-g, and GM-CSF), growth factors (epidermal growth factor In this study, we show that miR-31 is upregulated in psoriasis and and keratinocyte growth factor) and keratinocyte differentiation- identify it as a novel regulator of NF-kB activity. We demonstrate driving factors (1.5 mM CaCl2,12-O-tetradecanoylphorbol-13- that inhibition of endogenous miR-31 in keratinocytes suppresses http://www.jimmunol.org/ acetate, and cell confluence) and miR-31 expression was ana- the production of inflammatory mediators and the capability of lyzed by qRT-PCR. miR-31 was not significantly changed by keratinocytes to activate endothelial cells and to attract leuko- these factors (data not shown) with the exception of TGF-b1, cytes. Moreover, we identify STK40, a negative regulator of the a cytokine previously reported to be upregulated in the epidermis NF-kB pathway, as a novel target for miR-31 in keratinocytes and and serum of psoriasis patients (31). TGF-b1 significantly induced demonstrate that silencing of STK40 can rescue the effect of miR- miR-31 expression in keratinocyte cultures (Fig. 7A). Further- 31 on inflammatory mediators. Finally, we identify TGF-b1as more, upregulation of miR-31 by TGF-b1 was also observed in a regulator of miR-31 expression in vitro and in vivo. Our results three-dimensional reconstructed epidermal equivalents, which is thus suggest a model in which TGF-b1 induces miR-31 in pso- by guest on September 24, 2021

FIGURE 7. miR-31 is upregulated by TGF-b1 in keratinocytes in vitro and in vivo. (A) miR-31 expression was analyzed in keratinocytes in monolayer cell culture (two-dimensional) and three-dimensional epidermal equivalents treated with TGF-b1 (3 or 10 ng/ml, respectively) for 24 or 72 h. (B) miR-31 expression was analyzed in skins of WT mice (n = 3) and K5.TGF-b1 transgenic mice treated with normal saline (control, n = 3) or etanercept (n =3)by qRT-PCR. *p , 0.05, **p , 0.01, ***p , 0.001; Student t test. (C) In situ hybridization was performed on skin from WT mice (left panel) and K5.TGF-b1 transgenic mice treated with normal saline (middle panel) or Etanercept (right panel) using an miR-31–specific LNA probe (upper panel) or scrambled probe (middle panel). Blue-purple color indicates miR-31 expression. H&E staining (bottom panel) was performed for the same biopsies. Scale bar, 50 mm. 686 miR-31 IN PSORIASIS riasis keratinocytes that leads to increased NF-kB activity partially (reviewed in Ref. 1). In accordance with known functions of through suppression of STK40. In turn, inflammatory cytokine/ CXCL1, CXCL5, and CXCL8/IL-8, inhibition of endogenous chemokines are induced, which contributes to endothelial cell miR-31 in keratinocytes impaired the capability of conditioned activation, leukocyte attraction and clinically, skin inflammation supernatant to activate endothelial cells or to attract leukocytes, (Fig. 8). suggesting that miR-31 may be involved in the cross-talk between miR-31 is widely expressed and plays diverse roles in different keratinocytes and immune cells in psoriasis. Because miR-31 tissue and cell types. This miRNA has been shown to positively expression is not confined to keratinocytes, its deregulation in regulate corneal epithelial glycogen metabolism (32), to inhibit other cell types may also contribute to psoriatic skin inflammation. fibrogenesis and pulmonary fibrosis (33), to regulate lymphatic It was previously shown that low levels of miR-31 in human vascular lineage–specific differentiation (34), and to reduce TNF- natural regulatory T cells (Tregs) contributes to high expression of induced expression of E-selectin on human endothelial cells as FOXP3, which is a master transcription factor crucial for Treg a feedback control of inflammation (35). In pathological con- function and identified as the direct target of miR-31 (40). Pso- ditions, miR-31 has been mainly studied in cancer and was riasis has been associated with impaired immune suppressive ca- identified as a critical and pleiotropical regulator of tumor me- pacity of Tregs (41) and their enhanced propensity to differentiate tastasis and growth (reviewed in Ref. 36). In mouse skin, miR-31 into inflammatory IL-17A–producing cells (42). Thus, it would be has been shown to control hair cycle–associated gene expression interesting to further investigate the effect of miR-31 on the programs (37). However, its role in skin diseases and in particular, function and differentiation of Tregs in the context of psoriasis. psoriasis, was unknown. We found that inhibition of miR-31 in Our findings show that miR-31 regulates the activity of NF-kB keratinocytes decrease the production of IL-1b, CXCL1, CXCL5, pathway in keratinocytes, which is a key signaling pathway reg- Downloaded from and CXCL8, suggesting that miR-31 has an immunomodulatory ulating cytokine/chemokine expression (43) and which is activated function in keratinocytes. Keratinocytes in psoriasis lesions in psoriasis lesional skin (1, 44, 45). Importantly, activation of NF- overexpress several proinflammatory cytokines and chemokines, kB in both keratinocytes and lymphocytes is required to develop which contribute to the maintenance of inflammation in psoriasis the psoriasis-like phenotype in transgenic mice, indicating the skin lesions and represent a crucial element in psoriasis patho- importance of this pathway in psoriasis (46). In this study, we

genesis (1). IL-1b is a proinflammatory cytokine, which is es- demonstrate that in keratinocytes miR-31 directly targets STK40, http://www.jimmunol.org/ sential for Th17 differentiation (38) and potentiates immune cell which was previously shown to inhibit TNF-induced NF-kB ac- activation by regulating T cell–targeting chemokine production in tivation (29). Interestingly, the regulation of STK40 expression by keratinocytes (39). Keratinocyte-derived CXCL1, CXCL5, and miR-31 was also observed in ovarian cancer cells (30). In the skin, CXCL8/IL-8 can potently activate endothelial cells and are che- STK40 showed a reciprocal expression pattern with miR-31, moattractants for multiple subsets of leukocytes, especially neu- which further supports a functional miRNA:mRNA interaction trophil granulocytes, by binding to the cognate receptors CXCR1 in vivo. Derepression of STK40 through miR-31 may, at least and CXCR2 (20). These chemokines play prominent roles in re- partially, explain the reduced activity of NF-kB signaling by anti– cruitment and retention of inflammatory cells into psoriasis skin, miR-31 in keratinocytes. miR-31 was recently shown to down- as well as regulation of the formation of new blood vessels regulate noncanonical pathway of NF-kB activation via targeting by guest on September 24, 2021 NF-kB–inducing kinase in leukemic T cells (47). On the contrary, the expression of NF-kB–inducing kinase was not significantly altered by miR-31 inhibition in human primary keratinocytes, as shown in our microarray profiling data. The seemingly paradoxi- cal findings with regard to the regulatory role of miR-31 in the NF-kB signaling pathway indicate that the biological function of miR-31 is highly cell-type dependent. By a systematic screen using cytokines, growth factors, and factors modulating cell differentiation, we identified TGF-b1as a potent regulator of miR-31 expression in primary keratinocytes, in reconstructed human epidermal equivalents, and in a transgenic mouse model. TGF-b1 is a cytokine with increased levels in the epidermis and serum of psoriasis patients (9, 29, 30), and its in- volvement in psoriasis pathogenesis is supported by the finding that mice overexpressing TGF-b1 driven by a keratin 5 (K5) promoter in keratinocytes (K5.TGF-b1), exhibit a psoriasis-like phenotype (11, 12). The infiltration of inflammatory cells ob- served in K5.TGF-b1 mice may partially be due to the upregu- lation of miR-31 by TGF-b1 in keratinocytes, leading to increased production of chemoattractants. Importantly, TGF-b1, along with IL-1b, is critical for the differentiation of Th17 cells (38), which play essential pathogenic role in psoriasis. TGF-b1–induced miR- FIGURE 8. Proposed mechanism by which miR-31 modulates IL-1b, 31 can lead to increased IL-1b levels and thus can contribute to CXCL1, CXCL5, and CXCL8/IL-8 production by keratinocytes in psori- the amplification of Th17-type inflammation in psoriasis. Notably, asis skin. TGF-b1, which is highly expressed in psoriatic skin, upregulates we found that in keratinocytes miR-31 was not induced by TNF-a, the expression of miR-31 in keratinocytes. miR-31 directly suppresses STK40, a suppressor of NF-kB signaling activation induced by TNF-a. although this cytokine has been shown to upregulate miR-31 ex- The activation of NF-kB signaling contributes to the elevated expression pression in endothelial cells (27), further underlining that miR-31 and secretion of IL-1b, CXCL1, CXCL5, and CXCL8/IL-8, which pro- is regulated in a cell-type dependent manner. mote vascular endothelial cell activation and attract leukocytes via che- Taken together, we show that miR-31 is overexpressed in pso- motaxis into the skin. riasis keratinocytes and identify STK40, a negative regulator of The Journal of Immunology 687

NF-kB as a direct target. We further show that the expression of 19. Mizutani, H., R. Black, and T. S. Kupper. 1991. Human keratinocytes produce but do not process pro-interleukin-1 (IL-1)b: different strategies of IL-1 pro- miR-31 in keratinocytes is regulated by the important psoriasis- duction and processing in monocytes and keratinocytes. J. Clin. Invest. 87: associated cytokine, TGF-b1. Overexpression of miR-31 in pso- 1066–1071. riasis keratinocytes may contribute to skin inflammation by en- 20. Murphy, K. 2012. Janeway’s Immunobiology. Garland Science, New York. 21. Butcher, E. C. 1991. Leukocyte-endothelial cell recognition: three (or more) hancing leukocyte migration into the skin. These findings suggest steps to specificity and diversity. Cell 67: 1033–1036. that targeting miR-31 in psoriasis skin may alleviate inflammation 22. de Boer, O. J., I. M. Wakelkamp, S. T. Pals, N. Claessen, J. D. Bos, and by reducing the activity of NF-kB signaling and interfering with P. K. Das. 1994. Increased expression of adhesion receptors in both lesional and non-lesional psoriatic skin. Arch. Dermatol. Res. 286: 304–311. the cross-talk between keratinocytes and immune cells. 23. Groves, R. W., M. H. Allen, J. N. Barker, D. O. Haskard, and D. M. MacDonald. 1991. Endothelial leucocyte adhesion molecule-1 (ELAM-1) expression in cu- taneous inflammation. Br. J. Dermatol. 124: 117–123. Acknowledgments 24. Horrocks, C., J. I. Duncan, A. M. Oliver, and A. W. Thomson. 1991. Adhesion We thank Anna-Lena Kastman for excellent technical support as well as molecule expression in psoriatic skin lesions and the influence of cyclosporin A. Maria Lundqvist and Helena Griehsel for help in collecting patient samples. Clin. Exp. Immunol. 84: 157–162. We also thank the Microarray Core Facility at Novum, Bioinformatics and 25. Wakita, H., and M. Takigawa. 1994. E-selectin and vascular cell adhesion molecule-1 are critical for initial trafficking of helper-inducer/memory T cells in Expression Analysis, which is supported by the board of research at Kar- psoriatic plaques. Arch. Dermatol. 130: 457–463. olinska Institute and the research committee at the Karolinska Hospital. 26. Lewis, B. P., C. B. Burge, and D. P. Bartel. 2005. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120: 15–20. Disclosures 27. John, B., A. J. Enright, A. Aravin, T. Tuschl, C. Sander, and D. S. Marks. 2004. The authors have no financial conflicts of interest. Human MicroRNA targets. PLoS Biol. 2: e363. 28. Krek, A., D. Gru¨n, M. N. Poy, R. Wolf, L. Rosenberg, E. J. Epstein, P. MacMenamin, I. da Piedade, K. C. Gunsalus, M. Stoffel, and N. Rajewsky. Downloaded from 2005. Combinatorial microRNA target predictions. Nat. Genet. 37: 495–500. References 29. Huang, J., L. Teng, T. Liu, L. Li, D. Chen, F. Li, L. G. Xu, Z. Zhai, and 1. Lowes, M. A., A. M. Bowcock, and J. G. Krueger. 2007. Pathogenesis and H. B. Shu. 2003. Identification of a novel serine/threonine kinase that inhibits therapy of psoriasis. Nature 445: 866–873. TNF-induced NF-kB activation and p53-induced transcription. Biochem. Bio- 2. Ambros, V., R. C. Lee, A. Lavanway, P. T. Williams, and D. Jewell. 2003. phys. Res. Commun. 309: 774–778. MicroRNAs and other tiny endogenous RNAs in C. elegans. Curr. Biol. 13: 807– 30. Creighton, C. J., M. D. Fountain, Z. Yu, A. K. Nagaraja, H. Zhu, M. Khan, 818. E. Olokpa, A. Zariff, P. H. Gunaratne, M. M. Matzuk, and M. L. Anderson. 2010.

3. Friedman, R. C., K. K. Farh, C. B. Burge, and D. P. Bartel. 2009. Most mam- Molecular profiling uncovers a p53-associated role for microRNA-31 in inhib- http://www.jimmunol.org/ malian mRNAs are conserved targets of microRNAs. Genome Res. 19: 92–105. iting the proliferation of serous ovarian carcinomas and other cancers. Cancer 4. Ichihara, A., M. Jinnin, K. Yamane, A. Fujisawa, K. Sakai, S. Masuguchi, Res. 70: 1906–1915. S. Fukushima, K. Maruo, and H. Ihn. 2011. microRNA-mediated keratinocyte 31. Flisiak, I., B. Chodynicka, P. Porebski, and R. Flisiak. 2002. Association be- hyperproliferation in psoriasis vulgaris. Br. J. Dermatol. 165: 1003–1010. tween psoriasis severity and transforming growth factor b(1) and b(2) in plasma 5. Joyce, C. E., X. Zhou, J. Xia, C. Ryan, B. Thrash, A. Menter, W. Zhang, and and scales from psoriatic lesions. Cytokine 19: 121–125. A. M. Bowcock. 2011. Deep sequencing of small RNAs from human skin reveals 32. Peng, H., R. B. Hamanaka, J. Katsnelson, L. L. Hao, W. Yang, N. S. Chandel, major alterations in the psoriasis miRNAome. Hum. Mol. Genet. 20: 4025–4040. and R. M. Lavker. 2012. MicroRNA-31 targets FIH-1 to positively regulate 6. Lerman, G., C. Avivi, C. Mardoukh, A. Barzilai, A. Tessone, B. Gradus, corneal epithelial glycogen metabolism. FASEB J. 26: 3140–3147. F. Pavlotsky, I. Barshack, S. Polak-Charcon, A. Orenstein, et al. 2011. MiRNA 33. Yang, S., N. Xie, H. Cui, S. Banerjee, E. Abraham, V. J. Thannickal, and G. Liu. expression in psoriatic skin: reciprocal regulation of hsa-miR-99a and IGF-1R. 2012. miR-31 is a negative regulator of fibrogenesis and pulmonary fibrosis. PLoS ONE 6: e20916. FASEB J. 26: 3790–3799.

7. Sonkoly, E., T. Wei, P. C. Janson, A. Sa¨a¨f, L. Lundeberg, M. Tengvall-Linder, 34. Pedrioli, D. M., T. Karpanen, V. Dabouras, G. Jurisic, G. van de Hoek, by guest on September 24, 2021 G. Norstedt, H. Alenius, B. Homey, A. Scheynius, et al. 2007. MicroRNAs: J. W. Shin, D. Marino, R. E. Ka¨lin, S. Leidel, P. Cinelli, et al. 2010. miR-31 novel regulators involved in the pathogenesis of psoriasis? PLoS ONE 2: e610. functions as a negative regulator of lymphatic vascular lineage-specific differ- 8. Zibert, J. R., M. B. Løvendorf, T. Litman, J. Olsen, B. Kaczkowski, and L. Skov. entiation in vitro and vascular development in vivo. Mol. Cell. Biol. 30: 3620– 2010. MicroRNAs and potential target interactions in psoriasis. J. Dermatol. Sci. 3634. 58: 177–185. 35. Sua´rez, Y., C. Wang, T. D. Manes, and J. S. Pober. 2010. Cutting edge: TNF- 9. Xu, N., P. Brodin, T. Wei, F. Meisgen, L. Eidsmo, N. Nagy, L. Kemeny, induced microRNAs regulate TNF-induced expression of E-selectin and inter- M. Sta˚hle, E. Sonkoly, and A. Pivarcsi. 2011. MiR-125b, a microRNA down- cellular adhesion molecule-1 on human endothelial cells—feedback control of regulated in psoriasis, modulates keratinocyte proliferation by targeting FGFR2. inflammation. J. Immunol. 184: 21–25. J. Invest. Dermatol. 131: 1521–1529. 36. Valastyan, S., and R. A. Weinberg. 2010. miR-31: a crucial overseer of tumor 10. Meisgen, F., N. Xu, T. Wei, P. C. Janson, S. Obad, O. Broom, N. Nagy, metastasis and other emerging roles. Cell Cycle 9: 2124–2129. S. Kauppinen, L. Keme´ny, M. Sta˚hle, et al. 2012. MiR-21 is up-regulated in 37. Mardaryev, A. N., M. I. Ahmed, N. V. Vlahov, M. Y. Fessing, J. H. Gill, psoriasis and suppresses T cell apoptosis. Exp. Dermatol. 21: 312–314. A. A. Sharov, and N. V. Botchkareva. 2010. Micro-RNA-31 controls hair cycle- 11. Han, G., C. A. Williams, K. Salter, P. J. Garl, A. G. Li, and X. J. Wang. 2010. A associated changes in gene expression programs of the skin and hair follicle. role for TGFb signaling in the pathogenesis of psoriasis. J. Invest. Dermatol. FASEB J. 24: 3869–3881. 130: 371–377. 38. Volpe, E., N. Servant, R. Zollinger, S. I. Bogiatzi, P. Hupe´, E. Barillot, and 12. Li, A. G., D. Wang, X. H. Feng, and X. J. Wang. 2004. Latent TGFb1 over- V. Soumelis. 2008. A critical function for transforming growth factor-b, inter- expression in keratinocytes results in a severe psoriasis-like skin disorder. EMBO leukin 23 and proinflammatory cytokines in driving and modulating human T J. 23: 1770–1781. (H)-17 responses. Nat. Immunol. 9: 650–657. 13. Pivarcsi, A., L. Bodai, B. Re´thi, A. Kenderessy-Szabo´, A. Koreck, M. Sze´ll, 39. Sanmiguel, J. C., F. Olaru, J. Li, E. Mohr, and L. E. Jensen. 2009. Interleukin-1 Z. Beer, Z. Bata-Cso¨rgoo, M. Mago´csi, E. Rajnavo¨lgyi, et al. 2003. Expression regulates keratinocyte expression of T cell targeting chemokines through and function of Toll-like receptors 2 and 4 in human keratinocytes. Int. Immunol. interleukin-1 receptor associated kinase-1 (IRAK1) dependent and independent 15: 721–730. pathways. Cell. Signal. 21: 685–694. 14. Mootha, V. K., C. M. Lindgren, K. F. Eriksson, A. Subramanian, S. Sihag, 40. Rouas, R., H. Fayyad-Kazan, N. El Zein, P. Lewalle, F. Rothe´, A. Simion, J. Lehar, P. Puigserver, E. Carlsson, M. Ridderstra˚le, E. Laurila, et al. 2003. H. Akl, M. Mourtada, M. El Rifai, A. Burny, et al. 2009. Human natural Treg PGC-1a‑responsive genes involved in oxidative phosphorylation are coordi- microRNA signature: role of microRNA-31 and microRNA-21 in FOXP3 ex- nately downregulated in human diabetes. Nat. Genet. 34: 267–273. pression. Eur. J. Immunol. 39: 1608–1618. 15. Subramanian, A., P. Tamayo, V. K. Mootha, S. Mukherjee, B. L. Ebert, 41. Sugiyama, H., R. Gyulai, E. Toichi, E. Garaczi, S. Shimada, S. R. Stevens, M. A. Gillette, A. Paulovich, S. L. Pomeroy, T. R. Golub, E. S. Lander, and T. S. McCormick, and K. D. Cooper. 2005. Dysfunctional blood and target tissue J. P. Mesirov. 2005. Gene set enrichment analysis: a knowledge-based approach CD4+CD25high regulatory T cells in psoriasis: mechanism underlying unre- for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA strained pathogenic effector T cell proliferation. J. Immunol. 174: 164–173. 102: 15545–15550. 42. Bovenschen, H. J., P. C. van de Kerkhof, P. E. van Erp, R. Woestenenk, 16. Edgar, R., M. Domrachev, and A. E. Lash. 2002. Gene Expression Omnibus: I. Joosten, and H. J. Koenen. 2011. Foxp3+ regulatory T cells of psoriasis NCBI gene expression and hybridization array data repository. Nucleic Acids patients easily differentiate into IL-17A‑producing cells and are found in lesional Res. 30: 207–210. skin. J. Invest. Dermatol. 131: 1853–1860. 17. Butler, L. M., G. E. Rainger, and G. B. Nash. 2009. A role for the endothelial 43. Pasparakis, M. 2009. Regulation of tissue homeostasis by NF-kB signalling: glycosaminoglycan hyaluronan in neutrophil recruitment by endothelial cells implications for inflammatory diseases. Nat. Rev. Immunol. 9: 778–788. cultured for prolonged periods. Exp. Cell Res. 315: 3433–3441. 44. Johansen, C., E. Flindt, K. Kragballe, J. Henningsen, M. Westergaard, 18. Augoff, K., B. McCue, E. F. Plow, and K. Sossey-Alaoui. 2012. miR-31 and its K. Kristiansen, and L. Iversen. 2005. Inverse regulation of the nuclear factor-kB host gene lncRNA LOC554202 are regulated by promoter hypermethylation in binding to the p53 and interleukin-8 kB response elements in lesional psoriatic triple-negative breast cancer. Mol. Cancer 11: 5. skin. J. Invest. Dermatol. 124: 1284–1292. 688 miR-31 IN PSORIASIS

45. Lizzul, P. F., A. Aphale, R. Malaviya, Y. Sun, S. Masud, V. Dombrovskiy, and Crosstalk between keratinocytes and adaptive immune cells in an IkBa protein- A. B. Gottlieb. 2005. Differential expression of phosphorylated NF-kB/RelA in mediated inflammatory disease of the skin. Immunity 27: 296–307. normal and psoriatic epidermis and downregulation of NF-kB in response to 47. Yamagishi, M., K. Nakano, A. Miyake, T. Yamochi, Y. Kagami, A. Tsutsumi, treatment with etanercept. J. Invest. Dermatol. 124: 1275–1283. Y. Matsuda, A. Sato-Otsubo, S. Muto, A. Utsunomiya, et al. 2012. Polycomb- 46. Rebholz, B., I. Haase, B. Eckelt, S. Paxian, M. J. Flaig, K. Ghoreschi, mediated loss of miR-31 activates NIK-dependent NF-kB pathway in adult S. A. Nedospasov, R. Mailhammer, S. Debey-Pascher, J. L. Schultze, et al. 2007. T cell leukemia and other cancers. Cancer Cell 21: 121–135. Downloaded from http://www.jimmunol.org/ by guest on September 24, 2021 DATA SUPPLEMENT:

Figure S1. Differential expression of miR-31 in isolated primary human cell types. The expression of miR-31 was analyzed in primary adult keratinocytes, dermal fibroblasts, melanocytes, monocyte-derived dendritic cells (MDDCs), polymorphonuclear leukocytes (PMN), granulocytes, eosinophils, CD69+ cells, CD19+ cells, NK cells (CD56+ cells), naïve T cells (CD4+CD25high cells), CD8+ cells, CD4+ cells and mast cells using qRT-PCR. Primary human keratinocytes, dermal fibroblasts and melanocytes were isolated from healthy skin and cultured. Monocytes were isolated from peripheral blood mononuclear cells (PBMCs) from healthy blood donors (Karolinska University Hospital Blood Bank, Stockholm, Sweden) using MACS separation. Immature monocyte-derived dendritic cells (MDDCs) were generated by culturing separated monocytes in the presence of GM-CSF (550 IU/ml) and IL-4 (800 IU/ml) (Biosource International) for 6 days. CD4+, CD8+, CD4+CD25high, CD56+, CD19+, and CD69+ cells were isolated from PBMCs of healthy blood donors by FACS sorting using Becton Dickinson FACSAria cell sorting system. Granulocytes and eosinophils were FACS sorted from whole blood following RBC lysis.

Figure S2. Ectopic inhibition of miR-31 in human primary keratinocytes. (A) MiR-31 expression was detected in keratinocytes transfected with miR-31 inhibitor (anti-miR-31) or miRNA inhibitor control (anti-miR-Ctrl) for 24-96 hours by qRT-PCR. (B) The luciferase plasmids containing the synthesized sequence complementary to miR-31 (miR-31 sensor) or empty luciferase vector (vector) were co-transfected with anti-miR-31 or anti-miR-Ctrl into keratinocytes and the luciferase activity was measured 24 hours later. ***P < 0.001. Student’s t test. (C) The expression data of miR-31 host gene LOC554202 and the adjacent interferon epsilon (IFNE1) gene were extracted from the global transcriptome analysis of keratinocytes upon suppression of endogenous miR-31 using Affymetrix arrays.

Figure S3. The expression of STK40 in healthy and psoriasis skin. STK40 expression was analyzed in healthy, psoriasis non-lesional and lesional skin samples by immunohistochemistry. Replacement of the primary antibody with rabbit IgG in the staining process was used as negative control. Red-brown colour indicates STK40 expression. Bar = 50 m.

Supplemental table S1: miR-31 regulated transcripts The predicted targets of miR-31 were labeled in red. The NF-b target genes (summarized on the website http://bioinfo.lifl.fr/NF-KB/) were labeled in green. anti-miR-31 anti-miR-Ctrl Fold Gene Assignment miRNA target prediction change Sample Sample Sample Sample Sample Sample (anti-miR- p-value Mean Mean Public id Symbol Name TargetScan miRanda PicTar 1 2 3 1 2 3 31 vs anti- miR-Ctrl) 419 398 403 406 159 158 142 153 2,7 0,000 NM_032017 STK40 serine/threonine kinase 40 Yes 121 147 101 123 46 67 52 55 2,2 0,010 NM_004942 DEFB4A defensin, beta 4A 881 706 730 772 403 502 379 428 1,8 0,007 NM_178435 LCE3E late cornified envelope 3E keratinocyte differentiation- 1802 1614 1133 1516 726 910 976 870 1,7 0,039 NM_207392 KRTDAP associated protein 627 557 728 637 418 418 298 378 1,7 0,015 NM_178428 LCE2A late cornified envelope 2A

matrix metallopeptidase 9 (gelatinase B, 92kDa gelatinase, 92kDa type IV 236 175 222 211 128 114 140 127 1,7 0,014 NM_004994 MMP9 collagenase)

potassium channel tetramerisation 400 396 405 401 254 267 215 246 1,6 0,001 NM_173562 KCTD20 domain containing 20 UDP-N-acetyl-alpha-D- galactosamine:polypeptide N- acetylgalactosaminyltransferase 5 185 188 200 191 125 126 115 122 1,6 0,000 NM_014568 GALNT5 (GalNAc-T5) 2094 2332 2653 2360 1743 1430 1375 1516 1,6 0,013 NM_000421 KRT10 keratin 10

103 78 81 87 56 62 55 58 1,5 0,024 NM_001124758 SPNS2 spinster homolog 2 (Drosophila) 197 161 195 184 118 145 108 124 1,5 0,020 NM_014365 HSPB8 heat shock 22kDa protein 8

neural precursor cell expressed, 729 828 724 761 529 569 438 512 1,5 0,008 NM_006154 NEDD4 developmentally down-regulated 4 188 180 187 185 133 133 117 128 1,4 0,001 NM_014897 ZNF652 zinc finger protein 652 Yes

250 325 279 285 218 204 174 198 1,4 0,028 NM_007207 DUSP10 dual specificity phosphatase 10

63 77 74 71 49 60 40 50 1,4 0,040 NM_006475 POSTN periostin, osteoblast specific factor 1986 1725 1642 1784 1141 1404 1198 1248 1,4 0,015 NM_019060 CRCT1 cysteine-rich C-terminal 1 RNA binding motif, single stranded Yes 918 1080 1068 1022 832 701 616 716 1,4 0,020 NM_016836 RBMS1 interacting protein 1

771 893 840 835 488 696 576 587 1,4 0,024 NM_002963 S100A7 S100 calcium binding protein A7

1 sema domain, immunoglobulin domain (Ig), short basic domain, 1138 1440 1077 1218 785 914 878 859 1,4 0,039 NM_006379 SEMA3C secreted, (semaphorin) 3C

non-specific cytotoxic cell receptor 608 606 659 624 503 438 386 442 1,4 0,009 NM_001001414 NCCRP1 protein 1 homolog (zebrafish) 152 142 177 157 131 101 108 114 1,4 0,035 NM_001170880 GPR137 G protein-coupled receptor 137 1310 1041 1260 1204 867 837 908 871 1,4 0,017 NM_001423 EMP1 epithelial membrane protein 1 Yes 364 458 353 392 268 294 288 284 1,4 0,034 NM_018368 LMBRD1 LMBR1 domain containing 1 97 99 125 107 82 68 82 78 1,4 0,044 NM_015265 SATB2 SATB homeobox 2 Yes Yes Yes

nudix (nucleoside diphosphate linked 60 63 53 59 50 40 37 43 1,4 0,031 NM_198949 NUDT1 moiety X)-type motif 1 147 174 159 160 123 128 102 117 1,4 0,019 NM_173084 TRIM59 tripartite motif-containing 59 Yes 243 266 265 258 197 182 193 190 1,4 0,002 AK296954 ZNF852 zinc finger protein 852 46 43 41 43 32 36 28 32 1,4 0,017 NM_183242 BTBD8 BTB (POZ) domain containing 8 61 65 53 60 46 47 41 45 1,3 0,016 NM_020880 ZNF530 zinc finger protein 530 132 158 157 149 120 110 106 112 1,3 0,017 NM_017523 XAF1 XIAP associated factor 1 quaking homolog, KH domain RNA Yes 581 769 672 674 511 530 478 507 1,3 0,041 NM_206855 QKI binding (mouse) v-maf musculoaponeurotic fibrosarcoma oncogene homolog 79 69 72 73 61 60 45 55 1,3 0,037 NM_001031804 MAF (avian)

89 102 106 99 81 65 78 75 1,3 0,027 NM_001135179 ZDHHC3 zinc finger, DHHC-type containing 3

37 41 43 40 32 31 28 30 1,3 0,012 NM_001037132 NRCAM neuronal cell adhesion molecule

94 76 73 81 59 66 58 61 1,3 0,048 NM_052853 ADCK2 aarF domain containing kinase 2

similar to potassium channel 169 187 172 176 124 134 142 134 1,3 0,006 XM_001724546 LOC647013 tetramerisation domain containing 9

57 46 48 51 43 35 37 38 1,3 0,037 NM_144584 C1orf59 chromosome 1 open reading frame 59 sodium channel, voltage-gated, type 53 40 47 47 35 35 37 36 1,3 0,045 NM_018400 SCN3B III, beta

304 323 319 316 252 253 219 241 1,3 0,004 NM_080660 ZC3HAV1L zinc finger CCCH-type, antiviral 1-like 163 192 175 176 126 158 120 135 1,3 0,047 NM_033414 ZNF622 zinc finger protein 622 1125 1221 1119 1155 861 1008 786 885 1,3 0,021 NM_001017418 SPRR2B small proline-rich protein 2B protein tyrosine phosphatase-like A 550 654 587 597 453 502 421 459 1,3 0,023 NM_001010915 PTPLAD2 domain containing 2

2 514 477 449 480 340 384 385 370 1,3 0,010 NM_173595 ANKRD52 ankyrin repeat domain 52 Yes KN motif and ankyrin repeat domains Yes Yes 451 396 437 428 316 346 329 331 1,3 0,006 NM_153186 KANK1 1 coatomer protein complex, subunit 228 204 232 221 179 184 152 172 1,3 0,021 NM_016429 COPZ2 zeta 2

358 361 317 345 285 261 258 268 1,3 0,009 NM_007157 ZXDB zinc finger, X-linked, duplicated B 302 340 335 326 252 283 223 253 1,3 0,025 NM_173080 SPRR4 small proline-rich protein 4 65 63 72 66 47 57 50 52 1,3 0,021 NM_020431 TMEM63C transmembrane protein 63C 232 247 195 225 163 184 178 175 1,3 0,041 NM_024548 CEP97 centrosomal protein 97kDa 2400 2610 2346 2452 1931 1891 1919 1914 1,3 0,003 NM_019895 CLDND1 claudin domain containing 1 43 39 39 41 32 32 32 32 1,3 0,003 ENST00000426433 CCNYL2 cyclin Y-like 2 chromosome 6 open reading frame 126 136 121 128 105 101 93 100 1,3 0,008 NM_001042493 C6orf162 162 499 542 437 493 381 396 380 386 1,3 0,026 NM_001017418 SPRR2B small proline-rich protein 2B 68 64 54 62 48 48 51 49 1,3 0,030 NR_026743 tAKR aldo-keto reductase, truncated

306 350 297 318 229 268 254 250 1,3 0,028 NM_014506 TOR1B torsin family 1, member B (torsin B)

Yes 226 220 216 221 179 192 151 174 1,3 0,021 NM_033390 ZC3H12C zinc finger CCCH-type containing 12C 75 73 80 76 64 60 55 60 1,3 0,009 BC039496 LOC388906 hypothetical protein LOC388906 67 69 62 66 54 54 48 52 1,3 0,010 NM_018071 FLJ10357 protein SOLO

similar to microtubule-associated 203 241 196 213 180 158 168 169 1,3 0,044 XM_373277 LOC392288 proteins 1A/1B light chain 3 41 48 44 44 34 33 37 35 1,3 0,020 NM_020855 ZNF492 zinc finger protein 492 251 285 251 262 209 199 217 208 1,3 0,012 NM_006963 ZNF22 zinc finger protein 22 (KOX 15) 327 347 359 344 268 300 253 274 1,3 0,014 NM_020801 ARRDC3 arrestin domain containing 3 Yes protein phosphatase, Mg2+/Mn2+ 316 339 315 323 260 283 228 257 1,3 0,021 NM_177968 PPM1B dependent, 1B adaptor-related protein complex 4, 155 169 138 154 125 111 132 123 1,3 0,047 NM_004722 AP4M1 mu 1 subunit solute carrier family 24 (sodium/potassium/calcium 161 165 183 170 144 121 139 135 1,3 0,023 NM_024959 SLC24A6 exchanger), member 6 287 278 294 286 224 237 225 229 1,3 0,001 NM_018202 TMEM57 transmembrane protein 57 85 93 76 85 72 67 64 68 1,2 0,033 NM_058230 ZNF354B zinc finger protein 354B

112 123 139 125 106 102 92 100 1,2 0,050 NM_016134 PGCP plasma glutamate carboxypeptidase

107 123 110 113 88 101 84 91 1,2 0,033 NM_144611 CYB5D2 cytochrome b5 domain containing 2

3 protein tyrosine phosphatase-like (proline instead of catalytic arginine), 235 215 224 225 179 171 191 180 1,2 0,006 NM_014241 PTPLA member A 68 65 66 66 58 56 46 53 1,2 0,023 NM_033512 TSPYL5 TSPY-like 5

45 48 50 48 35 40 41 39 1,2 0,018 NM_030907 C1orf89 chromosome 1 open reading frame 89 tumor protein p53 inducible nuclear 134 119 131 128 112 97 101 103 1,2 0,021 NM_021202 TP53INP2 protein 2 977 1053 998 1009 780 789 871 813 1,2 0,006 NM_005470 ABI1 abl-interactor 1 poly (ADP-ribose) polymerase family, 58 54 47 53 40 45 43 43 1,2 0,039 NM_001113523 PARP15 member 15 steroid sulfatase (microsomal), 199 206 167 191 148 160 154 154 1,2 0,045 NM_000351 STS isozyme S

423 521 457 467 381 361 398 380 1,2 0,048 NM_003272 GPR137B G protein-coupled receptor 137B 204 227 216 216 188 177 163 176 1,2 0,015 NM_015691 WWC3 WWC family member 3

integrin, alpha V (vitronectin receptor, Yes 1801 1889 1647 1779 1347 1470 1532 1450 1,2 0,021 NM_002210 ITGAV alpha polypeptide, antigen CD51) 1496 1568 1474 1513 1225 1272 1206 1234 1,2 0,001 NM_005109 OXSR1 oxidative-stress responsive 1 Yes

peptidase domain containing 114 107 124 115 94 97 92 94 1,2 0,016 NM_015430 PAMR1 associated with muscle regeneration 1

273 242 270 261 211 234 197 214 1,2 0,031 NM_021260 ZFYVE1 zinc finger, FYVE domain containing 1 201 197 187 195 156 167 156 160 1,2 0,003 NM_033104 STON2 stonin 2 pleiotropic regulator 1 (PRL1 963 1129 957 1016 815 847 839 834 1,2 0,033 NM_002669 PLRG1 homolog, Arabidopsis) 180 180 178 180 142 165 135 147 1,2 0,023 BX641032 WEE1 WEE1 homolog (S. pombe) 2476 2403 2714 2531 2031 2284 1926 2080 1,2 0,034 NM_006472 TXNIP thioredoxin interacting protein 49 47 52 49 45 41 36 41 1,2 0,035 NM_138368 DKFZp761E198 DKFZp761E198 protein

567 619 534 573 484 495 436 472 1,2 0,029 NM_001079537 TRAPPC6B trafficking protein particle complex 6B 445 380 417 414 350 331 341 341 1,2 0,020 NM_173872 CLCN3 chloride channel 3 Yes 926 900 792 872 662 757 736 718 1,2 0,037 NM_024594 PANK3 pantothenate kinase 3

145 148 134 143 119 124 111 118 1,2 0,011 NM_005908 MANBA mannosidase, beta A, lysosomal 47 42 43 44 40 34 37 37 1,2 0,029 NM_020546 ADCY2 adenylate cyclase 2 (brain)

sphingomyelin phosphodiesterase 1, 112 101 105 106 81 85 97 88 1,2 0,036 NM_000543 SMPD1 acid lysosomal 482 516 488 495 428 418 387 411 1,2 0,007 NM_005333 HCCS holocytochrome c synthase

4 Yes 557 603 544 568 453 473 490 472 1,2 0,010 NM_006178 NSF N-ethylmaleimide-sensitive factor 166 139 155 153 121 128 134 127 1,2 0,041 NM_145115 ZNF498 zinc finger protein 498 Yes 45 45 45 45 39 33 40 38 1,2 0,028 NM_001171979 ZNF829 zinc finger protein 829

transglutaminase 2 (C polypeptide, protein-glutamine-gamma- 1070 925 1066 1020 1282 1206 1187 1225 -1,2 0,021 NM_004613 TGM2 glutamyltransferase)

Yes 85 81 95 87 106 105 103 104 -1,2 0,014 NM_207386 SHISA6 shisa homolog 6 (Xenopus laevis) 137 140 130 136 155 167 167 163 -1,2 0,005 NM_022371 TOR3A torsin family 3, member A 91 83 98 90 107 104 115 109 -1,2 0,029 BC019830 PRO2012 hypothetical protein PRO2012 48 41 38 42 50 51 52 51 -1,2 0,037 NM_001122764 PPOX protoporphyrinogen oxidase

solute carrier family 13 (sodium- dependent dicarboxylate transporter), 78 77 86 81 104 94 93 97 -1,2 0,022 NM_022829 SLC13A3 member 3 994 1030 966 996 1192 1279 1136 1202 -1,2 0,011 NM_013442 STOML2 stomatin (EPB72)-like 2 53 46 55 51 63 66 58 62 -1,2 0,042 NM_024741 ZNF408 zinc finger protein 408

34 34 34 34 42 38 43 41 -1,2 0,006 ENST00000327581 NT5DC4 5'-nucleotidase domain containing 4 fibroblast growth factor binding 2548 2471 2547 2522 2910 3326 2929 3055 -1,2 0,018 NM_005130 FGFBP1 protein 1 579 505 608 564 663 675 712 683 -1,2 0,024 NM_004687 MTMR4 myotubularin related protein 4

96 77 89 88 102 109 108 106 -1,2 0,035 NM_032387 WNK4 WNK lysine deficient protein kinase 4 691 611 718 673 826 821 804 817 -1,2 0,012 NR_027244 LOC151009 hypothetical LOC151009 chromosome 9 open reading frame 165 143 151 153 190 171 198 186 -1,2 0,034 BC002613 C9orf142 142 serine peptidase inhibitor, Kunitz 34 42 44 40 50 49 47 49 -1,2 0,050 NM_006652 SPINT3 type, 3 103 104 107 105 126 136 120 127 -1,2 0,007 NM_053043 RBM33 RNA binding motif protein 33 DnaJ (Hsp40) homolog, subfamily C, 178 159 186 174 218 215 205 213 -1,2 0,012 NM_015190 DNAJC9 member 9

ubiquitin carboxyl-terminal esterase 135 149 131 138 162 167 179 169 -1,2 0,015 NM_004181 UCHL1 L1 (ubiquitin thiolesterase) 70 78 83 77 100 94 88 94 -1,2 0,029 NM_153021 PLB1 phospholipase B1 123 140 133 132 174 144 166 161 -1,2 0,045 NM_032776 JMJD1C jumonji domain containing 1C growth factor, augmenter of liver 56 50 55 54 70 64 63 66 -1,2 0,012 NM_005262 GFER regeneration

5 UDP-GlcNAc:betaGal beta-1,3-N- 81 72 84 79 99 94 96 96 -1,2 0,009 NM_030765 B3GNT4 acetylglucosaminyltransferase 4 29 30 34 31 37 41 37 38 -1,2 0,020 NM_032133 MYCBPAP MYCBP associated protein 40 33 34 36 40 45 46 44 -1,2 0,045 NR_029495 MIR23A microRNA 23a 39 36 34 36 43 45 45 44 -1,2 0,005 NM_182645 VGLL2 vestigial like 2 (Drosophila)

804 780 706 763 922 963 925 937 -1,2 0,006 NR_003329 SNORD116-14 small nucleolar RNA, C/D box 116-14 ubiquitin-conjugating enzyme E2T 284 243 254 260 299 325 334 319 -1,2 0,022 NM_014176 UBE2T (putative)

254 204 247 235 281 280 305 289 -1,2 0,038 NM_016504 MRPL27 mitochondrial ribosomal protein L27 387 394 452 411 508 529 481 506 -1,2 0,019 NM_002064 GLRX glutaredoxin (thioltransferase) ral guanine nucleotide dissociation 151 126 122 133 156 160 176 164 -1,2 0,049 NM_006266 RALGDS stimulator

781 821 992 865 1118 1040 1040 1066 -1,2 0,045 NM_018571 STRADB STE20-related kinase adaptor beta 102 81 89 91 117 109 110 112 -1,2 0,032 NM_139280 ORMDL3 ORM1-like 3 (S. cerevisiae)

ubiquinol-cytochrome c reductase, 1317 1056 1249 1208 1541 1366 1563 1490 -1,2 0,048 NM_001003684 UQCR10 complex III subunit X chromosome 19 open reading frame 116 106 99 107 138 129 128 132 -1,2 0,013 NM_199249 C19orf48 48 86 86 79 84 100 110 99 103 -1,2 0,010 NM_025074 FRAS1 Fraser syndrome 1 zinc finger, CCHC domain containing 100 89 108 99 134 114 119 122 -1,2 0,045 NM_017665 ZCCHC10 10 337 319 337 331 379 452 398 410 -1,2 0,026 NM_016093 RPL26L1 ribosomal protein L26-like 1

egf-like module containing, mucin- 176 151 190 172 205 214 221 213 -1,2 0,028 NM_013447 EMR2 like, hormone receptor-like 2 non-SMC element 1 homolog (S. 382 375 447 401 511 532 448 497 -1,2 0,049 NM_145080 NSMCE1 cerevisiae) 35 39 45 39 51 49 47 49 -1,2 0,036 NM_001160042 IQCC IQ motif containing C PrdX deacylase domain containing 1, 41 39 39 39 55 44 48 49 -1,2 0,037 NR_027258 PRDXDD1P pseudogene 46 50 44 46 59 55 60 58 -1,2 0,011 NM_024680 E2F8 E2F transcription factor 8 NIMA (never in mitosis gene a)- 344 289 355 329 399 432 401 411 -1,2 0,024 NM_001145001 NEK6 related kinase 6 1078 960 1179 1072 1384 1395 1241 1340 -1,2 0,029 AK095678 LOC151009 hypothetical LOC151009 von Willebrand factor A domain 50 51 60 54 65 64 73 67 -1,3 0,034 NM_138345 VWA5B2 containing 5B2 43 37 36 39 51 43 51 48 -1,3 0,049 NM_001451 FOXF1 forkhead box F1

6 achaete-scute complex homolog 1 44 40 40 41 52 50 54 52 -1,3 0,004 NM_004316 ASCL1 (Drosophila) 170 141 156 156 196 204 187 196 -1,3 0,015 NM_030758 OSBP2 oxysterol binding protein 2 Yes Yes

asparagine-linked glycosylation 10, alpha-1,2-glucosyltransferase 127 154 145 142 159 189 190 179 -1,3 0,046 NM_032834 ALG10 homolog (S. pombe)

273 257 262 264 330 353 316 333 -1,3 0,005 NM_025079 ZC3H12A zinc finger CCCH-type containing 12A

43 47 45 45 59 57 55 57 -1,3 0,002 NM_015989 CSAD cysteine sulfinic acid decarboxylase

116 127 120 121 160 164 135 153 -1,3 0,028 NR_026825 RPSAP52 ribosomal protein SA pseudogene 52 family with sequence similarity 43, 159 143 175 159 215 198 189 201 -1,3 0,026 NM_153690 FAM43A member A 630 629 761 673 919 873 763 851 -1,3 0,049 NM_003521 HIST1H2BM histone cluster 1, H2bm 2041 1682 1880 1868 2262 2475 2355 2364 -1,3 0,015 NR_030754 MIR622 microRNA 622 1592 1923 1583 1699 2165 2213 2081 2153 -1,3 0,018 NM_004060 CCNG1 cyclin G1 olfactory receptor, family 2, subfamily 31 29 32 31 40 36 41 39 -1,3 0,009 NM_007160 OR2H2 H, member 2 olfactory receptor, family 2, subfamily 31 29 32 31 40 36 41 39 -1,3 0,009 NM_007160 OR2H2 H, member 2 similar to high-mobility group 347 286 328 321 405 433 385 408 -1,3 0,019 XM_936840 LOC646853 nucleosomal binding domain 2 44 42 44 43 59 54 53 55 -1,3 0,004 AK292677 LOC494150 prohibitin pseudogene chromosome 15 open reading frame 561 523 571 552 702 776 633 704 -1,3 0,026 NM_033286 C15orf23 23 family with sequence similarity 86, 166 138 159 155 195 211 186 197 -1,3 0,019 NM_201400 FAM86A member A 1577 1388 1527 1498 1879 1983 1875 1912 -1,3 0,003 NM_000224 KRT18 keratin 18

76 66 76 73 96 87 97 93 -1,3 0,010 NM_198180 QRFP pyroglutamylated RFamide peptide enoyl CoA hydratase domain 86 68 66 73 97 94 91 94 -1,3 0,032 NM_024693 ECHDC3 containing 3

44 34 40 39 49 50 53 51 -1,3 0,023 NM_024758 AGMAT agmatine ureohydrolase (agmatinase) dimethylarginine 300 345 365 337 484 417 394 432 -1,3 0,046 NM_012137 DDAH1 dimethylaminohydrolase 1 aldehyde dehydrogenase 1 family, 352 305 351 336 420 475 404 433 -1,3 0,022 NM_000693 ALDH1A3 member A3 181 178 176 179 205 247 238 230 -1,3 0,017 NM_153840 GPR110 G protein-coupled receptor 110 36 30 27 31 39 39 41 40 -1,3 0,030 NM_002666 PLIN1 perilipin 1

7 hydroxysteroid (17-beta) 126 96 111 111 136 150 142 143 -1,3 0,029 NM_014234 HSD17B8 dehydrogenase 8

1179 1189 1119 1162 1395 1708 1401 1501 -1,3 0,033 NR_026911 RPL21P28 ribosomal protein L21 pseudogene 28 thioredoxin-related transmembrane 1195 1107 1081 1128 1573 1537 1283 1464 -1,3 0,026 NM_015959 TMX2 protein 2 formin homology 2 domain containing 95 107 112 105 148 136 126 136 -1,3 0,017 NM_025135 FHOD3 3 sodium channel, voltage-gated, type 77 76 80 78 96 102 107 102 -1,3 0,002 NM_174934 SCN4B IV, beta 224 169 199 198 246 246 283 258 -1,3 0,039 NM_002318 LOXL2 lysyl oxidase-like 2

86 75 92 84 118 100 115 111 -1,3 0,025 NM_006169 NNMT nicotinamide N-methyltransferase

93 82 87 87 119 125 100 115 -1,3 0,027 NM_013332 C7orf68 chromosome 7 open reading frame 68 1376 1330 1338 1348 1647 2024 1651 1774 -1,3 0,028 NM_000982 RPL21 ribosomal protein L21 72 68 74 71 95 90 97 94 -1,3 0,001 NM_015894 STMN3 stathmin-like 3

1039 850 1079 989 1374 1238 1298 1304 -1,3 0,018 NR_003334 SNORD116-20 small nucleolar RNA, C/D box 116-20

Yes 478 369 457 435 549 570 600 573 -1,3 0,019 NM_000620 NOS1/NOS1 nitric oxide synthase 1 (neuronal) 74 69 81 75 88 102 104 98 -1,3 0,019 NM_006850 IL24 interleukin 24 1436 1410 1351 1399 1685 2104 1745 1844 -1,3 0,029 NM_000982 RPL21 ribosomal protein L21 1281 1238 1246 1255 1538 1892 1539 1656 -1,3 0,028 NM_000982 RPL21 ribosomal protein L21 nuclear prelamin A recognition factor- 76 53 66 65 83 91 85 86 -1,3 0,043 NM_022493 NARFL like 36 33 34 34 43 43 51 46 -1,3 0,019 AK093796 ZXDC ZXD family zinc finger C chromosome 11 open reading frame Yes 104 117 123 115 141 144 171 152 -1,3 0,028 NM_012194 C11orf41 41 NOTCH-regulated ankyrin repeat 98 101 91 97 130 125 131 128 -1,3 0,001 NM_001004354 NRARP protein 137 102 106 115 147 155 157 153 -1,3 0,030 BC052611 PXN paxillin SH3 domain and tetratricopeptide 116 108 119 114 150 148 159 152 -1,3 0,001 NM_024577 SH3TC2 repeats 2 CDC28 protein kinase regulatory 906 913 1110 977 1363 1413 1120 1299 -1,3 0,046 NM_001827 CKS2 subunit 2 cytochrome P450, family 27, 293 308 333 312 421 428 396 415 -1,3 0,002 NM_000785 CYP27B1 subfamily B, polypeptide 1 nuclear receptor subfamily 2, group F, 54 51 47 51 66 62 75 68 -1,3 0,020 NM_005654 NR2F1 member 1 glucosaminyl (N-acetyl) transferase 1, 123 116 157 132 179 180 171 177 -1,3 0,026 NM_001490 GCNT1 core 2

8 nuclear factor of kappa light polypeptide gene enhancer in B-cells 469 429 476 458 615 695 531 614 -1,3 0,035 NM_031419 NFKBI inhibitor, zeta

309 281 350 313 419 467 374 420 -1,3 0,033 NM_153212 GJB4 gap junction protein, beta 4, 30.3kDa 62 51 55 56 70 81 74 75 -1,3 0,016 NM_006716 DBF4 DBF4 homolog (S. cerevisiae) lymphocyte antigen 6 complex, locus 311 239 289 280 364 379 386 376 -1,3 0,012 NM_003695 LY6D D 414 401 455 423 573 602 534 569 -1,3 0,005 NM_004633 IL1R2 interleukin 1 receptor, type II

1418 1281 1496 1398 1963 1857 1836 1885 -1,3 0,003 NR_003332 SNORD116-17 small nucleolar RNA, C/D box 116-17

1418 1281 1496 1398 1963 1857 1836 1885 -1,3 0,003 NR_003332 SNORD116-17 small nucleolar RNA, C/D box 116-17 FBJ murine osteosarcoma viral 106 74 93 91 119 118 133 123 -1,4 0,037 NM_005252 FOS oncogene homolog

317 338 418 358 497 465 499 487 -1,4 0,017 NM_005266 GJA5 gap junction protein, alpha 5, 40kDa 1102 1090 963 1052 1269 1690 1348 1436 -1,4 0,048 NM_000982 RPL21 ribosomal protein L21

apolipoprotein B mRNA editing 42 44 53 46 74 61 57 64 -1,4 0,048 NM_021822 APOBEC3G enzyme, catalytic polypeptide-like 3G ATPase family, AAA domain containing 251 187 201 213 259 293 329 294 -1,4 0,045 NM_031921 ATAD3B 3B anoctamin 1, calcium activated 201 159 203 188 279 244 256 259 -1,4 0,015 NM_018043 ANO1 chloride channel 28 29 38 31 45 45 40 43 -1,4 0,027 NM_001104587 SLFN11 schlafen family member 11 gamma-aminobutyric acid (GABA) A 61 65 65 64 100 91 74 88 -1,4 0,033 BC099705 GABRB2 receptor, beta 2 ADAM metallopeptidase domain 19 306 270 334 303 391 434 437 421 -1,4 0,008 NM_033274 ADAM19 (meltrin beta) 782 633 624 680 822 961 1053 945 -1,4 0,035 NM_003524 HIST1H2BH histone cluster 1, H2bh

87 79 88 84 121 101 132 118 -1,4 0,023 NM_021939 FKBP10 FK506 binding protein 10, 65 kDa

71 52 62 62 90 74 95 86 -1,4 0,040 NM_003812 ADAM23 ADAM metallopeptidase domain 23 1420 1488 1426 1444 2100 2073 1931 2034 -1,4 0,000 NM_000576 IL1B interleukin 1, beta chromosome 21 open reading frame 185 147 246 193 273 270 277 273 -1,4 0,049 AF391113 C21orf70 70 35 31 37 34 53 47 45 48 -1,4 0,007 NM_001131055 HRH2 histamine receptor H2 63 51 51 55 66 87 81 78 -1,4 0,034 NM_001040715 KIAA0895L KIAA0895-like 38 31 45 38 57 50 55 54 -1,4 0,023 NM_032726 PLCD4 phospholipase C, delta 4

9 chemokine (C-X-C motif) ligand 1 (melanoma growth stimulating 185 188 212 195 300 255 286 280 -1,4 0,006 NM_001511 CXCL1 activity, alpha)

40 34 50 41 63 57 59 60 -1,4 0,019 NM_201252 AKR7L aldo-keto reductase family 7-like

small nuclear ribonucleoprotein 43 37 45 42 62 51 69 61 -1,5 0,031 NR_024489 LOC100129534 polypeptide N pseudogene

39 41 39 40 60 57 61 59 -1,5 0,000 NM_002994 CXCL5 chemokine (C-X-C motif) ligand 5

solute carrier organic anion 78 77 96 83 138 131 106 125 -1,5 0,022 NM_016354 SLCO4A1 transporter family, member 4A1

42 44 58 48 68 74 73 72 -1,5 0,012 NR_003336 SNORD116-22 small nucleolar RNA, C/D box 116-22 38 29 26 31 43 47 51 47 -1,5 0,017 NR_029422 RNU12 RNA, U12 small nuclear 121 160 138 140 213 261 182 219 -1,6 0,037 NM_001009931 HRNR hornerin 180 132 204 172 288 303 220 270 -1,6 0,041 NM_015973 GAL galanin prepropeptide 637 603 687 642 994 1041 1063 1033 -1,6 0,000 NM_139314 ANGPTL4 angiopoietin-like 4 76 58 56 63 93 108 108 103 -1,6 0,008 NR_004407 RNU11 RNA, U11 small nuclear

60 92 83 78 144 135 109 129 -1,7 0,023 NM_014439 IL1F7 interleukin 1 family, member 7 (zeta) 69 46 71 62 99 117 95 104 -1,7 0,018 NM_001014342 FLG2 filaggrin family member 2 69 70 68 69 124 130 93 116 -1,7 0,014 NM_003914 CCNA1 cyclin A1 ATPase, H+/K+ transporting, 78 82 106 89 159 178 121 153 -1,7 0,028 NM_001676 ATP12A nongastric, alpha polypeptide lipid phosphate phosphatase-related 35 40 51 42 79 74 75 76 -1,8 0,002 NM_001037317 LPPR5 protein type 5 sodium channel, nonvoltage-gated 1, 58 45 49 51 94 111 83 96 -1,9 0,007 NM_001039 SCNN1G gamma 346 301 438 362 718 665 728 704 -1,9 0,002 NM_000930 PLAT plasminogen activator, tissue

10