Microrna-302D Targets IRF9 to Regulate the IFN-Induced Gene Expression in SLE

Journal of Autoimmunity xxx (2017) 1e7

Contents lists available at ScienceDirect

Journal of Autoimmunity

journal homepage: www.elsevier.com/locate/jautimm

Short communication MicroRNA-302d targets IRF9 to regulate the IFN-induced gene expression in SLE

Siobhan Smith a, Thilini Fernando b, c, Pei Wen Wu b, c, Jane Seo b, c, Joan Ní Gabhann a, Olga Piskareva a, Eoghan McCarthy d, Donough Howard d, Paul O'Connell e, Richard Conway e, Phil Gallagher e, Eamonn Molloy e, Raymond L. Stallings a, Grainne Kearns c, Lindsy Forbess b, Mariko Ishimori b, Swamy Venuturupalli b, * Daniel Wallace b, Michael Weisman b, Caroline A. Jefferies a, b, c, a Department of Molecular and Cellular Therapeutics, Royal College of Surgeons in Ireland, Dublin 2, Ireland b Division of Rheumatology, Department of Medicine, Cedars-Sinai Medical Centre, 8700 Beverly Blvd, Los Angeles, CA 90048, USA c Department of Biomedical Sciences, Cedars-Sinai Medical Centre, 8700 Beverly Blvd, Los Angeles, CA 90048, USA d Department of Rheumatology, Beaumont Hospital, Dublin 9, Ireland e Department of Rheumatology, St. Vincent's University Hospital, Dublin 4, Ireland article info abstract

Article history: Systemic lupus erythematosus (SLE) is a complex disease targeting multiple organs as a result of over- Received 31 January 2017 activation of the type I interferon (IFN) system, a feature currently being targeted by multiple biologic Received in revised form therapies against IFN-a. We have identified an estrogen-regulated microRNA, miR-302d, whose 2 March 2017 expression is decreased in SLE patient monocytes and identify its target as interferon regulatory factor Accepted 6 March 2017 (IRF)-9, a critical component of the transcriptional complex that regulates expression of interferon- Available online xxx stimulated genes (ISGs). In keeping with the reduced expression of miR-302d in SLE patient monocytes, IRF9 levels were increased, as was expression of a number of ISGs including MX1 and OAS1. In vivo Keywords: fl SLE evaluation revealed that miR-302d protects against pristane-induced in ammation in mice by targeting MicroRNA IRF9 and hence ISG expression. Importantly, patients with enhanced disease activity have markedly IFN signalling reduced expression of miR-302d and enhanced IRF9 and ISG expression, with miR-302d negatively IFN-stimulated genes correlating with IFN score. Together these findings identify miR-302d as a key regulator of type I IFN driven gene expression via its ability to target IRF9 and regulate ISG expression, underscoring the importance of non-coding RNA in regulating the IFN pathway in SLE. © 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction a signalling cascade that results in the activation of two cytoplasmic kinases JAK1 and TYK2, which subsequently phosphorylate the Systemic lupus erythematosus (SLE) is a chronic autoimmune associated transcription factors STAT1 and STAT2. Once phosphor- disease characterised by a wide variety of immunological defects ylated, STAT1 and STAT2 dimerize and interact with IRF9 to form including enhanced expression of type I interferon (IFN)-stimu- the transcriptionally active complex, ISGF3, which binds to IFN- lated genes (ISGs) as a result of increased circulating levels of the stimulated response elements (ISRE) in the promoter region of antiviral cytokines IFN-alpha and IFN-beta (IFN-a and -b) [1]. IFN-inducible genes (reviewed in Ref. [3]). Intraperitoneal injection Elevated IFN-a correlates with disease severity, ﬂare and tissue of the hydrocarbon oil 2,6,10,14-tetramethylpentadecane (TMPD; involvement (speciﬁcally in the skin, kidney and central nervous also known as pristane) in mouse strains including C57BL/6 or Balb/ system) [2]. Binding of IFN-a/b to the IFN receptor (IFNAR) activates c mice has been used to induce a TLR7-dependent induction of type I IFN and sustained expression of interferon stimulated genes as early as 2 weeks post-pristane injection [4,5]. Although dendritic * Corresponding author. Division of Rheumatology, Department of Medicine and cells are the major producers of IFN, it is Ly6Chi monocytes which Department of Biomedical Sciences, Cedars-Sinai Medical Centre, 8700 Beverly are the predominant source of IFN-a and -b in the pristane model of Blvd, Los Angeles, CA 90048, USA. E-mail address: [email protected] (C.A. Jefferies). lupus, suggesting monocytes to be of importance in the http://dx.doi.org/10.1016/j.jaut.2017.03.003 0896-8411/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: S. Smith, et al., MicroRNA-302d targets IRF9 to regulate the IFN-induced gene expression in SLE, Journal of Autoimmunity (2017), http://dx.doi.org/10.1016/j.jaut.2017.03.003 2 S. Smith et al. / Journal of Autoimmunity xxx (2017) 1e7 interferonopathy observed in SLE [6e8]. Both IRF9 and IRF5 have 2.2. Isolation of PBMCs and cellular subsets independently been shown to be required for IFN-driven responses in pristane-induced ISG expression and autoimmunity [5,9]. Peripheral blood mononuclear cells (PBMCs) were separated With 90% of SLE patients being female, enhanced signalling from whole blood by density-gradient centrifugation with Ficoll- þ through estrogen receptor-alpha (ER-a) has been proposed to help Paque Plus (GE Healthcare). CD14 monocytes were purified from þ explain the strong female prevalence of SLE [10], with many studies fresh PBMCs by positive selection using magnetic anti-CD14 beads in both mouse models of lupus and ex vivo in SLE patient cells (Miltenyi Biotec and STEMCELL Technologies) according to manu- supporting this [11e16]. In addition, we have recently reported that facturers' protocol. levels of ER-a (ESR1) expression are elevated in SLE patient monocytes, with enhanced expression corresponding with 2.3. Cell culture increased sensitivity of SLE patient monocytes to estrogen stimulation in driving the expression of IL-23 and IFN-b, for example [11]. Purified monocytes were cultured in phenol redefree RPMI Cross-talk between estrogen-driven signalling and the IFN system 1640 medium supplemented with 10% charcoalestripped foetal is evidenced by the ability of IFN-a to drive ESR1 upregulation in calf serum and 100 mg/ml of penicillin/streptomycin. Sterile splenic mouse cells, with enhanced levels of ISGs and ESR1-driven ethanol-soluble 17b-estradiol and MPP (Sigma) were prepared genes observed in SLE-prone (NZB NZW) F1 female mice fresh for each experiment. Cells were stimulated for 6 h with sterile compared to their male counterparts [17]. water-soluble type I IFN (Immunotools) at 1000 IU/ml. Regarding the potential that estrogen-regulated microRNAs (miRs) are involved in either SLE development or pathogenesis, 2.4. MicroRNA mimic and inhibitor transfection female lupus prone-NZB/W (F1) mice have increased expression of lupus-linked miRNAs including miR-32, miR-155, miR-127 and the MicroRNA oligonucleotides were obtained from Dharmacon. miR-182-96-183 cluster compared to their male counterparts [18]. Negative controls were based on the sequences of miRNA in Cae- Moreover, X-linked miRNAs - miR-503, miR-188-3p, miR-421 and norhabditis elegans (cel-miR-67). Reverse transfection of human þ miR-98 - were reported to be overexpressed in CD4 T cells from primary monocytes was performed using Metafectene SI trans- female SLE patients compared to male patients [19], potentially fection reagent according to the manufacturers protocol. Cells were contributing to the sex differences observed in SLE development. To harvested for analysis 48e72 h after transfection. date however no association between estrogen-regulated micro- RNAs in SLE and IFN-driven gene expression has been shown. 2.5. Real-time quantitative polymerase chain reaction (RT-qPCR) In this study, we have identified miR-302d as an estrogen- analysis regulated miRNA which is downregulated in the immune cells of patients with SLE. Bioinformatic analysis identified interferon RNA was extracted from cell cultures using TRIzol reagent regulatory factor 9 (IRF9), a transcription factor which plays a key (Sigma) according to manufacturer's protocol. Expression of >800 ® role in promoting IFN-a-mediated gene expression, as a target miRNAs was assessed by Nanostring Technologies using nCounter gene. We validated the miR-302d binding site in the 30 untranslated miRNA Expression Assays. To validate the results, miR-302d region (UTR) of IRF9 and demonstrated augmentation of IRF9 levels expression was analyzed using the miScript primer assay with transfection of a miR-302d mimic, accompanied by enhanced (MS00003920 Qiagen) and miScript SYBR Green PCR kit (Qiagen), ISG expression. We demonstrate that in vivo administration of miR- with normalisation to the U6 small nuclear RNA (U6 snRNA). RNA 302d decreases IRF9 levels both in the peritoneum and systemically was reverse transcribed using the Tetro cDNA synthesis kit (Med- and that this results in reduced expression of MX1 and ISG15. ical Supply Company) according to the manufacturer's recom- Translating this to human SLE, we have found that not only is IRF9 mendations. Expression of mRNA was determined using expression elevated in SLE monocytes, but that its expression is appropriate primers (Supplementary Table 2) with the Sensifast associated with enhanced levels of ISGs at baseline in SLE patient Real-Time PCR Kit (Medical Supply Company) according to the peripheral blood mononuclear cells (PBMCs). Importantly, reduced manufacturer's recommendations. Data were analyzed using the miR-302d expression in patient PBMCs is associated with more DDCt comparative quantification method following normalisation active disease, underscoring the importance of miR-302d expres- to 18sRNA [21]. sion in maintaining a balance in the IFN system via its ability to regulate IRF9 expression. 2.6. Luciferase reporter assay

The putative miR-302d target sequence in the 30UTR of human 2. Methods IRF9 (IRF9-UTR) was cloned into the psiCHECK-2 vector (Promega) downstream of the Renilla luciferase reporter gene. The GeneTailor 2.1. Patient samples Site-Directed Mutagenesis kit (Thermo Fischer Scientiﬁc) was used to generate base deletions (IRF9-DEL) at the putative miR-302d All SLE patients (as per ACR diagnostic criteria) were recruited binding site in the IRF9-UTR plasmid with the following primers: from Beaumont Hospital, Dublin, St. James Hospital, Dublin, St. forward, 50-ACTTTATATTTTCCTCTTAGATATTCACTAAGG-3’; and Vincent's University Hospital, Dublin and Cedars-Sinai Medical reverse, 50-CTAAGAGGAAAATATAAAGTTGGGAGGTCAG-3’. All con- Centre, CA, USA. The Systemic Lupus Erythematosus Disease Ac- structs were sequenced, and were prepared with the use of an tivity Index (SLEDAI) score was determined for each patient at the EndoFree Plasmid Maxi kit (Qiagen). HEK293T cells were seeded at time of the blood draw. For calculation of the IFN score, the gene 1 105 cells/well in a 96-well plate 1 day pretransfection and then expression data for MX1, IFIT3, and OAS1 were used, as described transfected with a mixture of 50 ng IRF9-UTR or IRF9-DEL luciferase previously [20]. Age- and sex-matched healthy donors who had no reporter vector and 50 nM of either a negative control/miR-302d history of autoimmune diseases or treatment with immunosup- mimic. The cells were harvested 24 h later, and luciferase activity pressive agents were included. All participants provided informed was assessed using a Dual Luciferase Reporter Assay System written consent and the study received prior approval from the (Promega). Fireﬂy luciferase was used to normalise the Renilla relevant institutional ethics review boards. luciferase.

2.7. Western blotting 3. Results

Cells were lysed in loading buffer containing 63 mM Tris HCl (pH 3.1. Estrogen-regulated miR-302d is altered in SLE patient 6.8), 10% glycerol, 2% SDS, and 0.0025% bromophenol blue. Equal monocytes quantities of whole cell lysates were resolved by electrophoresis on a denaturing SDSepolyacrylamide gel and transferred to a nitro- To identify novel estrogen-regulated miRNAs in human mono- cellulose membrane followed by immunoblotting for anti-GAPDH cytes, we first performed an nCounter Human miRNA Assay (Santa Cruz Biotechnology) and antieIRF9 (Abcam). (Nanostring Technologies) on pooled samples derived from 5 healthy control monocytes either untreated or treated with 17b- estradiol (10 7 M). Physiological estradiol levels vary widely depending on age and stage in menstrual cycle and range between 2.8. Mice 10 11-10 8 M, whilst ex vivo stimulation of primary and cultured 9 7 C57BL/6J (B6) mice were purchased from The Jackson Labo- cells routinely use between 10 -10 M estradiol [22,23]. In our 7 ratory. Mice were maintained in the animal facility at the studies a concentration of 10 M17b-estradiol had previously Department of Comparative Medicine (Davis Building) in Cedars- been shown to induce responses in primary monocytes [11] and for fi Sinai Medical Centre, CA, USA. All animals were 6e8 weeks old at this study was determined to induce a signi cant decrease in miR- the time of experimentation. A lupus-like phenotype was induced 21 levels, a published estrogen-regulated miRNA [24] by intraperitoneal administration of 0.5 ml of pristane (Sigma). (Supplementary Fig. 1). Expression of 8 miRNAs was decreased The miR-302d and negative control mimics were purchased from (miR-720, miR-4454, miR-302d-3p, miR-576-5p, miR-1183, miR- Dharmacon. Invivofectamine 3.0 in vivo transfection reagent was 219-5p, miR-761 and miR-548ad), and 2 miRNAs increased (miR- purchased from Thermo Fisher Scientific. The transfection pro- 29a-3p and miR-1246) following estrogen stimulation tocol was based on manufacturer's instructions. Briefly, 100 mlof (Supplementary Table 1 A and B, respectively). qPCR validation of 1.2 mg/ml miRNA solution was prepared by adding 50 mlof individual miRNAs in the individual samples demonstrated a sig- fi complexation buffer to 50 ml of 2.4 mg/ml solution of miR-302d ni cant decrease of miR-302d-3p (miR-302d) alone or negative control. This solution was immediately added to (Supplementary Fig. 2). In line with hyperactive estrogen signalling 100 ml of Invivofectamine 3.0 transfection reagent which was pathway previously demonstrated in SLE patient monocytes [11], fi then mixed well to ensure complete complexation. The mixture reduced expression of miR-302d levels were identi ed in mono- was incubated at 50 C for 30min. The complex was then diluted cytes isolated from patients with SLE (Fig. 1A), with estrogen- 6-fold with sterile PBS and mixed well. For each injection, we receptor alpha (ER-a) inhibition (using methyl-piperidino- used 200 ml of working solution (10 nmol of the miR-302d mimic/ pyrazole dihydrochloride (MPP)) reversing 17b-estradiol-medi- negative control mimic). ated miR-302d decrease, suggesting that the effect of estrogen on miR-302d levels is mediated by ER-a (Fig. 1B).

2.9. Prime flow/flow cytometry 3.2. MiR-302d directly targets IRF9 whose expression is increased in SLE monocytes Peritoneal cells were harvested by peritoneal lavage with 10 ml of sterile, ice cold PBS. Cell viability was assessed by Fixable To investigate the biological significance of reduced miR-302d Viability dyes eFluor 506 (eBioscience). Cell populations were levels in SLE, bioinformatic analysis was employed to identify po- determined by flow cytometry following staining using CD45, tential gene targets. Three separate miRNA target prediction data- CD11b, Ly6G, Ly6C, CD19, CD4, CD8, and CD3. Intracellular mRNA bases each predicted IRF9 as a candidate target of miR-302d detection was performed using Prime FlowRNA assay kit (Affyme- (Supplemental Fig. 3A). Further analysis revealed a potential miR- trix) following manufacturer's protocols. In short, samples were 302d recognition site in the IRF9 3'untranslated region (UTR) to incubated for 2 h with target probes at 40 C. Target probe hy- be evolutionarily conserved, as shown in Supplemental Fig. 3B.To bridization was performed using type 1 (APC) probes for IRF9 confirm IRF9 as a direct target of miR-302d, a 300 base pair frag- 0 mRNA and type 6 (APC-Cy7) probes for the normalizer mRNA ment of the 3 UTR of IRF9 containing the putative miR-302d 0 (mouse ACTB). All samples were then incubated with the Pre- binding site (WT 3 UTR-IRF9) and its deletion mutant (Mut 0 Amplification (PreAmp) reagent for 2 h and the Amplification re- 3 UTR-IRF9) (Fig. 1C) were cloned into the psiCHECK-2 vector agent for an additional 2 h at 40 C. After signal amplification, downstream of the Renilla luciferase reporter gene. Cotransfection samples were incubated with label probes at 40 C for 1 h. Cell with miR-302d mimic significantly decreased luciferase activity of 0 staining were then analyzed on BD Biosciences LSRII, and data WT 3 -UTR-IRF9 reporter but as expected had no effect on the Mut 0 analysis was performed by using FlowJo_V10 (Tree Star). 3 -UTR-IRF9 (Fig. 1C). Transfection of primary human monocytes with a miR-302d mimic or antagomiR resulted in either repression or augmentation of IRF9 at both the mRNA (Fig. 1D) and protein 2.10. Statistical analysis level (Fig. 1E). In keeping with IRF9 as a target of miR-302d, IRF9 mRNA levels were elevated in SLE monocytes compared to healthy Data were analyzed using Prism 4 software, version 4.03 controls (Fig. 1F). (GraphPad Software). The nonparametric Mann-Whitney test used to compare gene expression between two groups, and the 3.3. MiR-302d and IRF9 modulate type I IFN signalling in human Spearman rank correlation coefficient was used to assess correla- monocytes tion between non-parametric data whilst an unpaired t-test was used to compare reporter gene activity. An unpaired t-test was used To explore the association between miR-302d and its ability to to analyse murine data. Statistical significance was defined as a p- target IRF9 expression in SLE monocytes, we evaluated the effect of value of <0.05. a miR-302d mimic on ISG expression, assessing ISRE-driven

Fig. 1. Altered expression of miR-302d and its target gene IRF9 regulates ISG expression. (A) miR-302d expression in monocytes isolated from SLE and control subjects (n ¼ 16). Each data point represents a single subject; horizontal lines and error bars show the mean ± SD, *** ¼ P 0.001. (B) miR-302d expression was determined by qPCR in monocytes pre- treated with 10 6 M MPP or ethanol (EtOH) vehicle control for 24 h followed by 17b-estradiol (E2) treatment (10 7 M) for an additional 6 h. Values are the mean ± SD of 3 samples, ** ¼ P 0.01. (C) Luciferase activity of WT-IRF9 or MUT-IRF9 plasmids following miR-302d transfection. Values are the mean ± SD of 3 samples, * ¼ P 0.05. (D) qPCR analysis of IRF9 expression in monocytes following miR-302d mimic/antagomiR transfection for 72 h. Values are the mean ± SD of 3 samples, * ¼ P 0.05 and ** ¼ P 0.01. (E) Western blot analysis of IRF9 in miR-302d mimic/antagomiR transfected monocytes after 72 h. Results are representative of 3 samples, with GAPDH used as a loading control. (F) qPCR analysis of IRF9 expression in monocytes isolated from SLE and control subjects (n ¼ 15). * ¼ P 0.05. (G) ISRE luciferase reporter gene activity in 293T cells transfected with miR-302d mimic/ antagomiR, followed by incubation with type I IFN for 6 h. Values are the mean and SEM, * ¼ P 0.05. (H and I) qPCR analysis of OAS1, Mx1 and IRF1 expression in miR-302d mimic (H)/antagomiR (I) transfected monocytes after 6 h of treatment with IFN-a (1000 IU/ml). Values are the mean ± SD of 3 samples, * ¼ P 0.05. reporter gene activity initially. Cotransfection of the ISRE reporter protecting against IFN-driven inflammation in vivo, using the with miR-302d mimic or inhibitor increased and decreased IFN-a- pristine-driven model of inducible SLE. miR-302d mimic or nega- driven luciferase activity, respectively (Fig. 1G). To further demon- tive control were complexed with Invivofectamene 3.0™ and strate modulation of type I IFN signalling by miR-302d, human injected i.p. into mice followed by 500 ml of pristane 24 h later. As monocytes were transfected with a miR-302d mimic or antagomiR, expected, miR-302d levels were significantly elevated in the peri- prior to stimulation with IFN-a (1000 IU/ml). Overexpression of toneum and systemically (Supplementary Figs. 4A and B, respec- miR-302d reduced expression of two representative ISGs, MX1 and tively). Pristane treatment induced increased expression of IRF9, OAS1 (Fig. 1H). In contrast, silencing of endogenous miR-302d led to MX1 and ISG15 in cells from the peritoneal cavity, blood, lung and increased activation of the type I IFN pathway and ISG expression kidney (Fig. 2A), in keeping with the role of IFN-a in driving (Fig. 1I). Activation by type I IFN can also lead to formation of STAT inflammation in this model. In the peritoneal lavage, blood and dimers which initiate gene transcription by binding to IFN gamma- lungs treatment with miR-302d mimic significantly reduced activated site (GAS) elements within gene promoters. Notably, miR- pristane-induced expression of IRF9 (p ¼ 0.01, 0.0037, 0.039, 302d failed to affect levels of the GAS-associated gene IRF1 respectively), MX1 (p ¼ 0.0002, 0.0002, 0.0025, respectively), and demonstrating the specificity of miR-302d activity on IRF9- ISG15 (p ¼ 0.0298, 0.0114, 0.0305, respectively) (Fig. 2BeD) whereas mediated activation of ISGs (Fig. 1H and I). Thus, miR-302d regu- decreases in IRF9 (p ¼ 0.0289) and MX1 (p ¼ 0.036) but not ISG15 lates expression of IRF9 through direct interaction with its 30UTR, were observed in the kidney (Fig. 2D). Expression of IRF1 (an IFN-g- thereby negatively regulating expression of IFN-inducible genes. driven gene) was not altered in cells obtained from peritoneal lavage (Fig. 2B, not significant, p ¼ 0.672). Interestingly, whilst a trend towards reduction of IRF1 expression in the blood was 3.4. In vivo administration of miR-302d reduces ISG expression in observed, it was not found to be statistically significant (Fig. 2C, not pristane driven IFN-induction significant, p ¼ 0.1113). Taken together these observations demonstrate the specificity of miR-302d for IRF9 and ISGF3- We next examined the biological relevance of miR-302d in

Fig. 2. In vivo administration of miR-302d attenuates IRF9-driven ISG expression. C57BL/6 female mice received a single intraperitoneal (IP) injection (10 nmol) of miR-302d mimic or negative control mimic (n ¼ 6) 1 day before an i.p. injection of 500 ml of pristane followed by sacrifice at 7 days for analysis. (A) Heatmap displaying relative expression of a panel of ISGs in PBS and pristane treated groups (n ¼ 6). (B) IRF9, MX1, ISG15 and IRF1 expression in the different treatment groups was analyzed in the peritoneal lavage by qPCR, p ¼ 0.01, 0.0002, 0.0298, and 0.67 (ns), respectively. (C) IRF9, MX1, ISG15 and IRF1 expression in the different treatment groups was analyzed in the blood by qPCR, p ¼ 0.0037, 0.0002, 0.0114, and 0.1113 (ns), respectively. (D) IRF9, MX1, and ISG15 expression in the different treatment groups was analyzed in the kidneys (top panel, p ¼ 0.0289, 0.036 and 0.7374 (ns), þ respectively) and lungs (bottom panel, p ¼ 0.039, 0.0025, 0.0305, respectively) by qPCR. (E) Flow cytometric analysis of numbers of Ly6C monocytes (left panel) and neutrophils (right panel)) in the peritoneum. In all cases, p values between different time points were by unpaired t-test * ¼ P 0.05, ** ¼ P 0.01 and *** ¼ P 0.001. mediated gene expression and indicate that the effects of miR-302d regulating the type I IFN pathway through the targeting IRF9. Pa- on the type I IFN pathway may result in inhibition of additional tient expression analysis revealed miR-302d negatively correlated inflammatory pathways in the periphery. with IFN score in our SLE patient cohort, further suggesting a reg- Given the association between elevated IFN-inducible gene ulatory interaction (Supplementary Fig. 5). As the IFN gene expression and accumulation of infiltrating cells in the peritoneal expression signature has been found to correlate with disease ac- lavage in pristane model of lupus [25], the effect of miR-302d tivity in SLE, we next explored whether miR-302d was associated administration on cellular recruitment was also assessed. Signifi- with an altered disease activity (SLEDAI) score in SLE patients. cant reduction in the percentage of neutrophils (Fig. 2E, right panel) Reduced miR-302d levels and increased IRF9 levels were identified recruited to the peritoneum was observed following miR-302d in- in SLE patients with active disease (SLEDAI score >4), respectively jection, with Ly6Chi monocyte recruitment also reduced although (Fig. 3A and B). In keeping with the role for IRF9 as a critical þ the effect was not significant (Fig. 2E, left panel). Neither CD8 T cell regulator of IFN stimulated gene expression, we observed enhanced or B cell recruitment were affected (Supplementary Fig. 4C), expression of two signature ISGs (MX1 and OAS1) in SLE patient þ whereas CD4 T cell numbers were slightly enhanced cells, compared to controls (Fig. 3C and D) with higher levels of (Supplementary Fig. 4D). Analysis of IRF9 expression by in-situ expression observed in patients with active disease (Fig. 3E and F). hybridization in immune cells following miR-302d administration Interestingly, a search of the GEO Profiles database [26,27] revealed in vivo showed reduced IRF9 expression in peritoneal neutrophils that IRF9 levels are elevated in SLE patient peripheral blood and monocytes (Supplementary Figs. 4E and F, respectively). IRF9 mononuclear cells (PBMCs) (GSE78193 [28]) and that elevated IRF9 expression T and B cell populations was unaffected (Supplementary expression correlates positively with expression of ISGs such as Figs. 4G and H, respectively). Thus in vivo targeting of IRF9 via ISG15 and IRF7 (Fig. 3G and H, respectively). In addition, analysis of transfection with a miR-302d mimic resulted in reduced ISG a separate cohort of patients revealed that ISG expression associ- expression via targeting IRF9 in neutrophils and monocytes. Our ates strongly with IRF9 expression as demonstrated in Fig. 3I. We data indicates that miR-302d protects against IFN-driven gene next investigated whether manipulation of miR-302d levels in SLE expression by targeting IRF9 both ex vivo in SLE monocytes and monocytes could attenuate both IRF9 expression and subsequent in vivo in the pristane model of SLE. ISG induction. Addition of a miR-302d mimic significantly reduced expression of IRF9 and the two representative ISGs Mx1 and OAS1 (Fig. 3J). Taken together our data demonstrates that reduced 3.5. MiR-302d and IRF9 expression levels correlate with SLE disease expression of miR-302d in SLE monocytes results in enhanced activity and ISG expression in SLE monocytes expression of IRF9 which in turn positively contributes to the IFN gene signature observed in patients. Our findings thus have The results described thus far imply a role for miR-302d in

Fig. 3. Increased IRF9 levels are associated with enhanced basal expression of ISGs in SLE patients. (A and B) Levels of miR-302d and IRF9 in SLE patient cells in inactive and active (SLEDAI score >4) disease course as determined by qPCR (n ¼ 42); (C and D) Levels of MX1 and OAS1 expression in SLE patient monocytes (n ¼ 8) versus healthy control subjects (n ¼ 12) as determined by qPCR; (E and F) Levels of MX1 and OAS1 expression in SLE patient cells both in inactive and active disease course as determined by qPCR (n ¼ 35). For A-F, each data point represents a single subject; horizontal lines and error bars show the mean ± SD. * ¼ P 0.05. (G and H) Correlation between IRF9 and IRF7 (G) and ISG15 (H) levels in SLE patient PBMCs. (I) Heatmap demonstrating relative expression of a panel of ISGs in SLE patients (n ¼ 79) versus healthy controls (n ¼ 10); (J) IRF9, MX1 and OAS1 expression in monocytes isolated from SLE patients (N ¼ 3) following negative control/miR-30d mimic transfection for 48 h as determined by qPCR. Values are the mean ± SD of 3 samples, * ¼ P 0.05 and *** ¼ P 0.001. important implications for our understanding of the molecular target for miR-302d. It has been demonstrated that 50% of SLE regulation of IFN gene expression in SLE. patients have dysregulated expression of ISGs (interferon stimulated genes), a phenotype which is strongly associated with more severe manifestation of this condition [29]. 4. Discussion A role for microRNAs in regulating the type I IFN pathway in SLE has become increasingly apparent, with miR-146a, for example Abnormal miRNA expression patterns in hormone-dependent been implicated at various stages of type I IFN signalling including conditions such as breast cancer have been reported to be manipulation of IRF5 and STAT1 levels. To the best of our knowl- induced by estrogen. Similarly, a growing number of miRNA have edge, this is the first study to investigate a role for miR-302d in fi been recently identi ed as estrogen-regulated including miR-146a autoimmunity and its contribution to dysregulated IFN signalling in [14] and miR-155, [15] both of which have links to SLE. Here, we SLE. In the present study, miR-302d levels were found to negatively fi establish identi ed that levels of the estrogen-regulated miRNA correlate with SLE patient IFN scores suggesting that the under- fi miR-302d are signi cantly reduced in SLE patients, in particular in expression of miR-302d observed in SLE contributes to the alter- those with a high IFN score and increased disease activity. Reduced ations observed in the type I IFN pathway observed in these pa- plasma levels of miR-302d have previously been linked to the tients. The ability of miR-302d to reduce activation of the IFN autoimmune condition experimental autoimmune encephalomy- pathway and expression and of the selected IFN-stimulated genes elitis (EAE), [16] highlighting the potential that miR-302d may be downstream of the IFNAR suggests modulation of miR-302d ac- an important autoimmune regulator. tivity could be useful as a strategy to directly inhibit the IFN MiR-302d is a member of the miR-302 family of miRNAs, a pathway in SLE. Antagonism of miR-155 and miR-21 in mouse family which have been demonstrated to target a variety of genes models of SLE have both been shown to ameliorate inflammation including interleukin-1 receptor-associated kinase 4 (IRAK4) [17], and pathology [18 19]. Consistent with findings demonstrating altered expression of which can lead to dysregulation of key cellular deletion of the IFNAR results in reduced pristane-induced ISG in- responses. In addition to these previously established miR-302 duction [30], our data shows that miR-302d-mediated down- fi targets, we identi ed and validated IRF9 - a key component of regulation of IRF9 decreases ISG expression in both in the ISGF3 complex, a transcriptional complex of IRF9, STAT2 and STAT1 peritoneum and systemically. which promotes transcription of IFN-inducible genes - as a novel

The IFN pathway has emerged as a major contributor to SLE altered IgG class switching, Eur. J. Immunol. 42 (6) (2012) 1477e1487. pathogenesis and a vast array of therapeutics have been developed [10] A. Inui, H. Ogasawara, T. Naito, I. Sekigawa, Y. Takasaki, Y. Hayashida, K. Takamori, H. Ogawa, Estrogen receptor expression by peripheral blood to dampen activity of this pathway. Of those developed and mononuclear cells of patients with systemic lupus erythematosus, Clin. participating in clinical trials, directly targeting the IFNAR receptor Rheumatol. 26 (10) (2007) 1675e1678. itself with a mAb has yielded the most promising results and is [11] S. Smith, J.N. Gabhann, E. McCarthy, B. Coffey, R. Mahony, J.C. Byrne, K. Stacey, E. Ball, A. Bell, G. Cunnane, et al., Estrogen receptor alpha regulates TRIM21 most advanced, currently in Phase III trials. Overall our study expression, contributing to dysregulated cytokine production in systemic demonstrates that underexpression of miR-302d contributes to the lupus erythematosus, Arthritis Rheumatism 66 (1) (2014) 163e172. IFN signature that is characteristic of SLE via its ability to target IRF9 [12] V. Rider, X. Li, G. Peterson, J. Dawson, B.F. Kimler, N.I. Abdou, Differential expression of estrogen receptors in women with systemic lupus erythema- and hence modulate ISGF3-dependent gene expression, and sug- tosus, J. Rheumatol. 33 (6) (2006) 1093e1101. gests modulation of miR-302d levels may be protective in SLE as [13] M.A. Cunningham, O.S. Naga, J.G. Eudaly, J.L. Scott, G.S. Gilkeson, Estrogen demonstrated by its ability to protect against pristane-induced receptor alpha modulates Toll-like receptor signaling in murine lupus, Clin. fl Immunol. 144 (1) (2012) 1e12. in ammation. Our data indicates that inhibition of the type I IFN [14] J.L. Svenson, J. EuDaly, P. Ruiz, K.S. Korach, G.S. Gilkeson, Impact of estrogen pathway through manipulation of miR-302d levels could poten- receptor deficiency on disease expression in the NZM2410 lupus prone tially be beneficial in SLE patients who present with clinical disease mouse, Clin. Immunol. 128 (2) (2008) 259e268. driven by IFN dysregulation. [15] E. Kassi, P. Moutsatsou, Estrogen receptor signaling and its relationship to cytokines in systemic lupus erythematosus, J. Biomed. Biotechnol. (2010) 317452. Financial support [16] M.A.L. Johansson, B. Moller, T. Smedby, S. Rantapaa-Dahlqvist, Oestrogen receptor a gene polymorphisms in systemic lupus erythematosus, Ann. Rheum. Dis. 64 (2005) 1611e1617. Funding for this study was provided by the Lupus Research [17] R. Panchanathan, H. Shen, X. Zhang, S.M. Ho, D. Choubey, Mutually positive Alliance (ALR/TIL/332436) and the Irish Research Council [GOIPD/ regulatory feedback loop between interferons and estrogen receptor-alpha in 2014/512] who had no role in study design, data collection and mice: implications for sex bias in autoimmunity, PLoS One. 5 (5) (2010) e10868. analysis, decision to publish, or preparation of the manuscript. [18] R. Dai, S. McReynolds, T. LeRoith, B. Heid, Z. Liang, S.A. Ahmed, Sex differences in the expression of lupus-associated miRNAs in splenocytes from lupus- Conflict of interest statement prone NZB/W(F1) mice, Biol. Sex Differ. 4 (2013) 19. [19] A. Hewagama, G. Gorelik, D. Patel, P. Liyanarachchi, W. Joseph McCune, E. Somers, T. Gonzalez-Rivera, C. The Michigan Lupus, F. Strickland, The authors have no financial conflict of interest with the work B. Richardson, Overexpression of X-Linked genes in T cells from women with and results presented herein. lupus, J. Autoimmun. 41 (2013) 60e71. [20] Q. Fu, X. Chen, H. Cui, Y. Guo, J. Chen, N. Shen, C. Bao, Association of elevated transcript levels of interferon-inducible chemokines with disease activity and Appendix A. Supplementary data organ damage in systemic lupus erythematosus patients, Arthritis Res. Ther. 10 (5) (2008). R112-R. Supplementary data related to this article can be found at http:// [21] K.J. Livak, T.D. Schmittgen, Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method, Methods 25 dx.doi.org/10.1016/j.jaut.2017.03.003. (4) (2001) 402e408. [22] A.J. Murphy, P.M. Guyre, C.R. Wira, P.A. Pioli, Estradiol regulates expression of References estrogen receptor ERalpha46 in human macrophages, PLoS One. 4 (5) (2009) e5539. [23] X. Xu, T.D. Veenstra, Concentration of endogenous estrogens and estrogen [1] L. Ronnblom, The importance of the type I interferon system in autoimmunity, metabolites in the NCI-60 human tumor cell lines, Genome Med. 4 (4) (2012) e Clin. Exp. Rheumatol. 34 (2016) 21 24, 4 Suppl 98. 31. [2] M.K. Crow, Type I interferon in the pathogenesis of lupus, J. Immunol. 192 (12) [24] N.S. Wickramasinghe, T.T. Manavalan, S.M. Dougherty, K.A. Riggs, Y. Li, e (2014) 5459 5468. C.M. Klinge, Estradiol downregulates miR-21 expression and increases miR-21 [3] J.J. O'Shea, D.M. Schwartz, A.V. Villarino, M. Gadina, I.B. McInnes, A. Laurence, target gene expression in MCF-7 breast cancer cells, Nucleic Acids Res. 37 (8) The JAK-STAT pathway: impact on human disease and therapeutic interven- (2009) 2584e2595. e tion, Annu. Rev. Med. 66 (2015) 311 328. [25] P.Y. Lee, J.S. Weinstein, D.C. Nacionales, P.O. Scumpia, Y. Li, E. Butfiloski, N. van [4] P.Y. Lee, Y. Kumagai, Y. Li, O. Takeuchi, H. Yoshida, J. Weinstein, E.S. Kellner, Rooijen, L. Moldawer, M. Satoh, W.H. Reeves, A novel type I interferon- D. Nacionales, T. Barker, K. Kelly-Scumpia, et al., TLR7-dependent and producing cell subset in murine lupus, J. Immunol. Baltim. Md 1950) 180 FcgammaR-independent production of type I interferon in experimental (7) (2008) 5101e5108. e mouse lupus, J. Exp. Med. 205 (13) (2008) 2995 3006. [26] R. Edgar, T. Barrett, NCBI GEO standards and services for microarray data, Nat. [5] D.L. Thibault, A.D. Chu, K.L. Graham, I. Balboni, L.Y. Lee, C. Kohlmoos, Biotechnol. 24 (12) (2006) 1471e1472. A. Landrigan, J.P. Higgins, R. Tibshirani, P.J. Utz, IRF9 and STAT1 are required [27] T. Barrett, R. Edgar, Mining microarray data at NCBI's gene expression for IgG autoantibody production and B cell expression of TLR7 in mice, J. Clin. omnibus (GEO)*, Methods Mol. Biol. 338 (2006) 175e190. e Invest. 118 (4) (2008) 1417 1426. [28] P. Chen, T. Vu, A. Narayanan, W. Sohn, J. Wang, M. Boedigheimer, fi [6] P.Y. Lee, J.S. Weinstein, D.C. Nacionales, P.O. Scumpia, Y. Li, E. But loski, N. van A.A. Welcher, B. Sullivan, D.A. Martin, J.J. Ruixo, et al., Pharmacokinetic and Rooijen, L. Moldawer, M. Satoh, W.H. Reeves, A novel type I IFN-producing cell pharmacodynamic relationship of AMG 811, an anti-IFN-gamma IgG1 e subset in murine lupus, J. Immunol. 180 (7) (2008) 5101 5108. monoclonal antibody, in patients with systemic lupus erythematosus, Pharm. [7] P.Y. Lee, Y. Li, Y. Kumagai, Y. Xu, J.S. Weinstein, E.S. Kellner, D.C. Nacionales, Res. 32 (2) (2015) 640e653. fi E.J. But loski, N. van Rooijen, S. Akira, et al., Type I interferon modulates [29] E.C. Baechler, F.M. Batliwalla, G. Karypis, P.M. Gaffney, W.A. Ortmann, fl monocyte recruitment and maturation in chronic in ammation, Am. J. Pathol. K.J. Espe, K.B. Shark, W.J. Grande, K.M. Hughes, V. Kapur, et al., Interferon- e 175 (5) (2009) 2023 2033. inducible gene expression signature in peripheral blood cells of patients with [8] W.H. Reeves, P.Y. Lee, J.S. Weinstein, M. Satoh, L. Lu, Induction of autoim- severe lupus, Proc. Natl. Acad. Sci. 100 (5) (2003) 2610e2615. munity by pristane and other naturally occurring hydrocarbons, Trends [30] D. Braun, P. Geraldes, J. Demengeot, I. Type, Interferon controls the onset and e Immunol. 30 (9) (2009) 455 464. severity of autoimmune manifestations in lpr mice, J. Autoimmun. 20 (1) fi [9] D. Feng, L. Yang, X. Bi, R.C. Stone, P. Patel, B.J. Barnes, Irf5-de cient mice are (2003) 15e25. protected from pristane-induced lupus via increased Th2 cytokines and