ava i l ab l e a t www.sc i enced i r ec t . com

C l i n i ca l Immuno logy

www.e l sev i e r . com/ l oca te /yc l im

Clinical Immunology (2009) 133, 428–436

Estradiol targets T cell signaling pathways in humansystemic lupusEmily Walters a,1, Virginia Rider a,⁎,1, Nabih I. Abdou b, Cindy Greenwell b,Stan Svojanovsky c, Peter Smith c, Bruce F. Kimler d

a Department of Biology, Pittsburg State University, Pittsburg, KS 66762, USAb Center for Rheumatic Disease, Allergy-Immunology, Kansas City, MO 64111, USAc Kansas Intellectual and Developmental Disabilities Research Center, The University of Kansas Medical Center, Kansas City,KS 66160, USAd Department of Radiation Oncology, The University of Kansas Medical Center, Kansas City, KS 66160, USA

Received 13 July 2009; accepted with revision 8 September 2009Available online 30 September 2009

⁎ Corresponding author. Fax: +1 620E-mail address: vrider@pittstate.e

1 Contributed equally to the content

1521-6616/$ – see front matter © 200doi:10.1016/j.clim.2009.09.002

KEYWORDSSLE;Estradiol;Interferon-α;T cell signaling;Microarray

Abstract The major risk factor for developing systemic lupus erythematosus (SLE) is beingfemale. The present study utilized gene profiles of activated T cells from females with SLE andhealthy controls to identify signaling pathways uniquely regulated by estradiol that couldcontribute to SLE pathogenesis. Selected downstream pathway genes (+/− estradiol) weremeasured by real time polymerase chain amplification. Estradiol uniquely upregulated sixpathways in SLE T cells that control T cell function including interferon-α signaling. Measurementof interferon-α pathway target gene expression revealed significant differences (p= 0.043) in

DRIP150 (+/− estradiol) in SLE T cell samples while IFIT1 expression was bimodal and correlatedmoderately (r= 0.55) with disease activity. The results indicate that estradiol alters signalingpathways in activated SLE T cells that control T cell function. Differential expression oftranscriptional coactivators could influence estrogen-dependent gene regulation in T cellsignaling and contribute to SLE onset and disease pathogenesis.© 2009 Elsevier Inc. All rights reserved.

235 4194.du (V. Rider).of this paper.

9 Elsevier Inc. All rights reser

Introduction

Systemic lupus erythematosus (SLE) is an autoimmunedisease that affects several organ systems in the body [1].SLE is characterized by the production of pathogenicautoantibodies [2] resulting in part from abnormal interac-tions between T and B cell signaling [3]. Genetic suscepti-

ved

bility and environmental triggers contribute to SLE diseaseonset [4]. However, the greatest risk factor for developingSLE is female gender [5]. Sex hormones contribute to thisgender bias in SLE but the mechanisms involved have notbeen fully elucidated [6-8].

Estradiol is a female sex hormone that binds to specificestrogen receptors (ER) and alters the transcription rates oftarget genes [9]. ERs function as transcription factorsthrough subdomains that serve as docking surfaces forother essential transcriptional regulatory proteins known ascoactivators and corepressors [10]. Nuclear receptor coacti-vators enhance transcription [11] while nuclear corepressors

.

429Human systematic lupus

suppress transcription [12]. The ligand activated ER interactswith the p160 coactivators to open chromatin at thepromoter region of target genes [13]. The DRIP coactivators,a second class of coactivator proteins, share subunits withthe Mediator complex, suggesting these coactivators operateby recruiting RNA polymerase II to the promoter of estradiolregulated genes [14].

The molecular basis underlying the deregulation of T andB cells in SLE appears to be global in nature since multiplegenes are inappropriately regulated [4,15]. For example,measurement of 160 variants in the sera of SLE patientsusing a high-throughput protein microarray revealed 30abnormally regulated proteins including cytokines, chemo-kines, growth factors and soluble receptors in SLE patientscompared with control samples [16]. Gene profiling ofheterogeneous populations of peripheral blood mononuclearcells revealed genes within molecular pathways that areconsistently altered in SLE and may contribute to thedevelopment of autoimmunity [17-19]. The overexpression ofthe type I interferons (IFNs) and, in particular the IFN-αsubtype, is implicated in the initiation and development ofSLE. Secretion of IFN-α normally declines within hours afterinitial viral infection, but in susceptible individuals it can besustained because genes that control INF-α or downstreamtargets such as IFN regulatory factor (IRF5) and signaltransducer and activator of transcription (STAT4) remainelevated [20,21]. The IFN signature refers to upregulatedgenes in the interferon pathway that correlate with diseaseactivity in SLE. The IFN-α signaturemay also bealtered in otherautoimmune disorders including Sjögren's syndrome, demato-myosisits, psoriasis and in a small number of RA patients [22].

Recent completion of a clinical trial showed that monthlyadministration of the ER antagonist, Faslodex, for 1 yearimproved disease activity in SLE patients [23]. Calcineurinand CD154 expression were downregulated in the Faslodexarm of the study. This finding is consistent with our previousreports showing calcineurin and CD154 genes were upregu-lated by estradiol in SLE but not in normal T cells [24,25]. Togain additional insight into the molecular basis of estradiolaction in SLE T cells, the present study utilized gene profilingand compared estradiol effects in SLE and normal T cells. Theresults suggest that estradiol alters signaling pathways in SLET cells that are associated with disease onset and progres-sion. One of the pathways significantly altered by estradiol inSLE T cells is interferon-α. While several genes within theinterferon-α pathway appeared to be upregulated, expres-sion of vitamin D receptor interacting protein (DRIP150) is ofparticular interest because it provides preliminary evidencethat deregulated cofactor expression could modulate theresponse to estradiol in target cells and may underlie thesensitivity to estradiol in some SLE T cells.

Patients and methods

Study subjects

This study was approved by the St. Luke's Hospital Institu-tional Review Board and all subjects provided writteninformed consent prior to participation. Thirteen femaleSLE patients between the ages of 24 and 46 years with amedian age of 42 years were enrolled in the study. The SLE

patients fulfilled the American College of Rheumatologycriteria for SLE [26] and were diagnosed with moderatedisease activity at the time of blood draw (SLEDAI score 2–18,median 6). Patients enrolled in the study exhibited symmet-rical polyarticular non-erosive arthritis (n=8), lupus malar ordiscoid rash (n= 8), Raynaud's syndrome (n=1), serositis(n=3), and renal disease (n= 6). Nine of the patients weretaking prednisone, three patients were taking mycopheno-late mofetil, three patients were taking azathioprine andseven patients were taking hydroxychloroquine. Duration oflupus in the patients ranged from 3 to 20 years (medianduration, 8 years). Eight of the patients were Caucasian, fourwere African American and one was of Asian descent. Tenhealthy control females were enrolled in the study. The ageof the control females ranged from twenty-one to 51 yearswith a median age of 37. To control for the effects ofmedications, five female patients with rheumatoid arthritis(RA) were enrolled in the study. The age of the RA patientsranged from 28 to 48 years with a median age of 42. Two ofthe RA patients were taking prednisone, one was takingmethotrexate and onewas taking azathioprine. Two of the RApatients were not taking medications. Participants hadregular menstrual cycles and none were taking hormonereplacement therapy, oral contraceptives, or had a history ofother collagen vascular diseases.

Collection of T cell enriched peripheral bloodmononuclear cells

T cell enriched mononuclear cells were separated from bloodsamples (∼90 ml) by density gradient (Histopaque, Sigma, St.Louis, MO, USA). The lymphocytes were removed and washedtwice in serum-free medium (RPMI 1640, Fisher Scientific,Hanover Park, IL, USA) and residual red blood cells were lysed(H-Lyse buffer, R and D Systems, Minneapolis, MN, USA). Tcells were purified by negative selection through T cellisolation columns (Human T Cell Enrichment Columns, R and DSystems). T cells were cultured overnight (18 h) at 37 °C under5% CO2 in serum-free medium (Hybridoma, Sigma, St. Louis,MO, USA) supplemented with L-glutamine (200 mM). T cellswere activated after 18 h of culture for 4 h with phorbol 12myristate 13-acetate (PMA, Sigma, 10 ng/ml) and ionomycin(Sigma, 0.5 μg/ml). To test the effect of estradiol on activatedT cells, estradiol-17β (10−7 M) was added (or not) to half of thereplicate cultures for the entire culture period. We havepreviously shown that calcineurin and CD154 expression areupregulated in SLE T cells in response to estradiol at this doseand time as described in detail elsewhere [24,25].

RNA isolation

T cell RNA was isolated using the TRIzol reagent (Invitrogen,Carlsbad, CA) and Phase Lock Heavy Gels (Eppendorf, FisherScientific). Total RNA was purified from T cells and treatedwith DNase I according to the manufacturer's protocol (DNA-free, Ambion, Austin, TX).

Microarray analysis

Gene profiling was carried out at the Kansas UniversitySchool of Medicine Microarray Facility. The concentration

430 E. Walters et al.

and purity of total RNA was assessed with an AgilentBioanalyzer and samples with RIN scores above 7.0 wereused for complimentary (cRNA) RNA synthesis. BiotinylatedcRNA was hybridized to high density Affymetrix humanGeneChips HG-U133_Plus_2, which contained 54,675 probesets. The chips were scanned and analyzed using MAS5 typeof data analysis with Affymetrix and Gene spring GX 7.3.1(Agilent Technologies) software, where the data from themicroarray chips were subsequently normalized per gene andchip. The absence, presence, or marginal status of specificgene probes was based on the detection p-value (Student's ttest). Detection p-value under 5% (b0.05) determines theprobe as present. In cases where the detection p-value wasbetween 5% and 6.5% (0.05–0.065) the detection interpre-tation was marginal lowering the significance level to lessthan 95%. If the detection value exceeded 6.5% (N0.065), theprobes were designated as absent, since the significance ofthe detection was less than 94%. Signal intensities of genespresent in estradiol-treated samples were compared to thenon-treated samples in order to generate a value for fold-change.

Pathway analysis

Cell signaling pathways were identified using the IngenuityPathways Analysis (IPA, Ingenuity Systems, Redwood City,CA) library of canonical pathways that were most significantto the data sets. Genes from the data sets that met the foldchange cutoff in all SLE patients but not in control femalesand, in all control females but not in SLE patients werepositioned in the appropriate canonical pathways using theIPA Knowledge Base. Fischer's exact test was used tocalculate a p-value determining the probability that theassociation between the genes in the data set and thecanonical pathway were explained by chance alone.

Real-time polymerase chain amplification

Selected target genes within upregulated pathways wereindependently investigated by examining expression levelsusing real-time PCR. Total T cell RNA was digested usingDNase 1 and cDNA was synthesized from 4 μg of the resultingRNA using a High Capacity cDNA kit (Applied Biosystems,Foster City, CA). Real-time PCR (Step-one, Applied Biosys-tems) was carried out according to the manufacturer'sprotocol. The templates were quantified using a Taqmanprobe and specific gene primers for vitamin D receptorinteracting protein [DRIP150, (also known as mediatorcomplex subunit 14), Hs00188481, Applied Biosystems] andinterferon induced with tetratricopeptide repeats 1 (IFIT1,Hs01911452, Applied Biosystems). A Taqman probe andglyceraldehyde 3-phosphate dehydrogenase (GAPDH,Hs99999905, Applied Biosystems) specific gene primerswere used for the internal control. In each cycle, fluorescentsignals for the target gene and GAPDH were collected fromtriplicate samples. The value of Ct (target Ct minus GAPDHCt), which was inversely correlated with the number oftarget mRNA copies, was calculated and compared with apooled T cell sample that was used as a positive control forall of the assays. Samples without template were included oneach plate as a negative control. The fold change in

expression levels was calculated by dividing the sample Ctvalues obtained from T cells cultured with and withoutestradiol from the same individual.

Statistical analysis

Because the entire content of a blood draw was required ofan individual assay, samples from each subject enrolled inthis study were not tested in all assays. After mediannormalization for each chip, the Student's t test was used tocompare differences in T cell gene expression without andwith estradiol. The significance of the association between adata set and the canonical pathway (IPA, www.ingenuity.com) was measured by: (i) a ratio of the number of genesfrom the data set that mapped to the pathway divided by thetotal number of genes that map to the canonical pathway;(ii) Fisher's exact test to calculate a p-value determining theprobability that the association between the genes in thedataset and the canonical pathway was explained by chancealone. A nonparametric Mann–Whitney U test was used tocompare differences in DRIP150 and IFIT1 expression. Linearregression analysis correlated SLEDAI scores with IFIT1 andDRIP150 response to estradiol. A p valueb0.05 was consid-ered statistically significant.

Results

Gene profiling reveals differential expression ofestrogen responsive genes between SLE and normalT cells

Microarray analysis was used to identify specific genes thatwere differentially regulated in samples of peripheral T cellspurified and activated from SLE patients (n=6) and controlfemales (n=5). Of the six patients used for the microarraystudy, four were Caucasians and two were African American.The median SLEDAI at the time of blood collection was 6.8.Normalized probe values (see Patients and methods) wereobtained from profiled T cells and the data were used togenerate fold changes in gene expression. Since we expectedsome heterogeneity among human samples we arbitrarilychose to include fold change values≥1 for all upregulatedgenes and ≤1 for all downregulated genes. Four gene listswere assembled using GeneSpring software (data notshown). The first list contained genes that were upregulatedby estradiol in all SLE T cell samples but not in all control Tcell samples. The second list contained genes that weredownregulated by estradiol in all SLE T cell samples but notin all control T cell samples. The third list contained genesthat were upregulated by estradiol in all control T cellsamples but not in all SLE T cell samples. The fourth listcontained genes that were downregulated by estradiol in allcontrol T cell samples but not in all SLE T cell samples.

Signal transduction pathways are altered byestradiol in SLE T cells

As expected, the expression of a large number of genes wasaltered in SLE and normal T cells by estradiol (data can beaccessed http://bioinformatics.kumc.edu/mdms/login.

Table 2 Estradiol uniquely downregulates signaling pathwaysin activated SLE T cells but not in normal T cells.

Pathway SLE T cellsp-value

Control T cellsp-value

PPARα/RXRγ activation 0.0006 NSG-protein coupled receptorsignaling

0.0017 NS

Hepatic fibrosis/hepaticstellate cell activation

0.0019 NS

cAMP-mediated signaling 0.0024 NSNitrogen metabolism 0.0058 NSPropanoate metabolism 0.0102 NSCircadian rhythm signaling 0.0120 NSAxonal guidance signaling 0.0174 NSKeratan sulfate biosynthesis 0.0234 NSCalcium signaling 0.0275 NSFatty acid metabolism 0.0282 NSValine, leucine and isoleucinedegradation

0.0309 NS

Cardiac β2-adrenergic signaling 0.0324 NSβ-alanine metabolism 0.0417 NSPhospholipid degradation 0.0417 NSSynaptic long term potentiation 0.0437 NSLPS/IL-1 mediated inhibition ofRXR function

0.0437 NS

TGF-β signaling 0.0447 NSIL-2 signaling 0.0457 NSAmyloid processing 0.0457 NS

Gene lists that were generated from microarray data were

431Human systematic lupus

php). In order to obtain a more focused view of how thesegenes might alter SLE T cell function, we used the IPA libraryto investigate whether estradiol altered signaling pathwaysin all of the T cell samples from SLE patients versus pathwaysin all of the control T cell samples. In general, estradioluniquely downregulated more signaling pathways in both SLEand control T cells (20 down versus 16 up in SLE T cellsamples; 28 down versus 13 up in control T cell samples,Tables 1–4). Three pathways, including B cell receptorsignaling, cardiac β2-adrenergic signaling and IL-4 signaling,were upregulated in both SLE and control T cell samples butthe genes involved were different between SLE and control Tcells. These pathways are therefore not included as uniquelyregulated in SLE or normal T cells.

Sixteen signaling pathways were uniquely upregulated inall of the SLE T cell samples (Table 1). Eight of thesepathways are involved in cellular metabolism while sixpathways control various aspects of T cell function. Five ofthe six pathways involved with T cell function are known tobe altered in autoimmune diseases. Twenty pathways wereuniquely downregulated by estradiol in all of the SLE T cellsamples (Table 2). The downregulated responses includedmetabolic pathways, cell signaling pathways and pathwaysthat are crucial for T cell function including PPARαactivation, calcium signaling, cAMP-mediated signaling,and IL-2 signaling.

Estradiol increased gene expression of thirteen pathwaysin all of the control but not the SLE T cell samples (Table 3).The antigen presentation pathway was the only pathwayidentified that is known to play a role in T cell function. The

Table 1 Estradiol uniquely upregulates signaling pathwaysin activated SLE T cells but not in normal T cells.

Pathway SLE T cellsp-value

Control T cellsp-value

Pentose phosphate pathway 0.0002 NSInositol phosphate metabolism 0.0030 NSSulfur metabolism 0.0037 NSPI3K/AKT signaling 0.0041 NSInterferon signaling 0.0047 NST cell receptor signaling 0.0062 NSGlucocorticoid receptorsignaling

0.0069 NS

GM-CSF signaling 0.0071 NSInsulin receptor signaling 0.0141 NSβ-alanine metabolism 0.0155 NSPropanoate metabolism 0.0170 NSValine, leucine and isoleucinedegradation

0.0170 NS

Pantothenate and CoAbiosynthesis

0.0191 NS

Calcium signaling 0.0275 NSSynaptic long termpotentiation

0.0380 NS

SAPK/JNK signaling 0.0407 NSHuntington's disease signaling 0.0407 NS

Gene lists that were generated from microarray data wereanalyzed for pathways uniquely upregulated in all SLE but not incontrol T cells. List shown is of pathways in the SLE T cells whichmet the criteria for significance (p≤0.05).

analyzed for pathways uniquely downregulated in all SLE T cellsand not in the control T cells. List shown is of pathways whichmet the criteria for significance (p≤0.05) in SLE T cells.

other types of signaling pathways regulated by estradiol inthe control T cell samples involved cell signaling, motility,and metabolism. Twenty eight pathways were significantlydownregulated by estradiol in the control T cell samples(Table 4). It is interesting to note that five of those pathwaysdownregulated in the control but not in the SLE T cellsamples, regulate processes that are defective in SLE T cells.

The interferon-α pathway is upregulated inresponse to estradiol in SLE T cells

Type I interferons are suggested to be key modulators in thepathogenesis of SLE [27]. The microarray data showedestradiol upregulated the interferon-α pathway in SLE (p=0.005) and not in control T cell samples (Table 1). We furtherinvestigated interferon-α signaling in order to identify thosegenes that were contributing to the upregulated response.Normalized values from the microarray data showed seven ofthe genes including the Janus kinase 2 (JAK2), proteintyrosine phosphatase, non-receptor type 11 (PTPN11),vitamin D receptor interacting protein (DRIP150), 2′, 5′-oligoadenylate synthetase 1 (OAS1), the SEC14-like 2(Saccharomyces cerevisiae) (sec14L2), the suppressor ofcytokine signaling 1 (SOCS1), and interferon induced withtetratricopeptide repeats 1 (IFIT1) showed mean fold-values

Table 3 Estradiol uniquely upregulates signaling pathwaysin activated normal T cells but not in SLE T cells.

Pathway Control T cellsp-value

SLE T cellsp-value

Axonal guidance signaling 0.0001 NSEphrin receptor signaling 0.0011 NSRegulation of actin-basedmotility by rho

0.0021 NS

Valine, leucine andisoleucine degradation

0.0043 NS

Antigen presentationpathway

0.0102 NS

Sonic hedgehog signaling 0.0102 NSBile acid biosynthesis 0.0110 NSSerotonin receptor signaling 0.0123 NSN-glycan biosynthesis 0.0141 NSXenobiotic metabolismsignaling

0.0191 NS

FXR/RXR activation 0.0372 NSβ-alanine metabolism 0.0407 NSAryl hydrocarbon receptorsignaling

0.0427 NS

Gene lists that were generated from microarray data wereanalyzed for pathways uniquely upregulated in all control but notin SLE T cells. List shown is of pathways which met the criteria forsignificance (p≤0.05) in control T cells.

Table 4 Estradiol uniquely downregulates signaling pathwaysin activated control T cells but not in SLE T cells.

Pathway Control T cellsp-value

SLE T cellsp-value

IGF-1 signaling 0.0002 NSInsulin receptor signaling 0.0002 NSPI3K/AKT signaling 0.0009 NSNatural killer cell signaling 0.0016 NSFc epsilon RI signaling 0.0017 NSMitochondrial dysfunction 0.0025 NSCeramide signaling 0.0030 NSEstrogen receptor signaling 0.0039 NSNRF2-mediated oxidativestress response

0.0044 NS

Synaptic long term depression 0.0056 NSGlucocorticoid receptorsignaling

0.0058 NS

T cell receptor signaling 0.0091 NSXenobiotic metabolismsignaling

0.0100 NS

JAK/stat signaling 0.0117 NSOxidative phosphorylation 0.0126 NSVDR/RXR activation 0.0138 NSIntegrin signaling 0.0145 NSProtein ubiquitinationpathway

0.0195 NS

Neuregulin signaling 0.0200 NSAmyotrophic lateral sclerosissignaling

0.0234 NS

Death receptor signaling 0.0251 NSSAPK/JNK signaling 0.0282 NSUbiquinone biosynthesis 0.0347 NSERK/MAPK signaling 0.0398 NSFXR/RXR activation 0.0407 NSIL-4 signaling 0.0417 NSNF-κB signaling 0.0427 NSApoptosis signaling 0.0437 NS

Gene lists that were generated from microarray data wereanalyzed for pathways uniquely downregulated in all normal andnot in the control T cells. List shown is of pathways which met thecriteria for significance (p≤0.05) in control T cells.

432 E. Walters et al.

greater than 1.5 in response to estradiol across the SLE T cellsamples (Fig. 1).

Real-time PCR allowed us to investigate pathway genes inT cells from additional SLE patients as well as T cells frompatients with RA to control for the effects of medications.We measured expression levels of two genes in theinterferon-α pathway in SLE T cell samples from five of thesix SLE patients and five controls who donated blood formicroarray study. In addition, we added five additional SLE Tcell samples, five additional control T cell samples and Tcells from 5 patients with RA. First, we measured the IFIT1gene product that showed a modest mean-fold increase inresponse to estradiol in the SLE T cell samples (see Fig. 1).The IFIT1 gene was chosen in spite of this modest increasebecause it codes for a protein that has been suggested tofunction in the pathogenesis of SLE [28]. Second was theDRIP150 gene since this gene product functions as a cofactorfor ER mediated transcription [29] and could be responsiblefor the increased sensitivity of SLE T cells to estradiol.

As measured by real-time PCR, the estradiol-inducedchange in IFIT1 expression was not significantly different (p=0.63, Mann–WhitneyU test) between the SLE and control T cellsamples. The mean fold change between SLE T cell samplescultured without and with estradiol from the same person was2.1 (n= 10), while the mean fold change for the control T cellsamples was 1.18 (n= 10). The estradiol response in the SLET cells was bimodal in that five of the T cell samples respondedstrongly to estradiol (Table 5). The lowest and highestresponders were African American suggesting that the bimodalresponse was not based on race. The five high responders hadrenal involvement in their lupus. IFIT1 expression was lower infive of the SLE patients in response to estradiol. Two of thosepatients, both African American, had renal involvement in

their disease. Changes in IFIT1 expression showed a moderatecorrelation (r= 0.55) with the patient's SLEDAI scores.

Relative DRIP150 expression (+/− estradiol) was signifi-cantly different in SLE T cells compared with the control Tcells (p= 0.043, 2-sided Mann–Whitney U test). The meanfold change in DRIP150 expression for the SLE T cells was 6.8(n=10). The median fold change in DRIP150 expressionamong the SLE patient's T cells was 2.6. Among the top fiveresponders, the mean fold change for DRIP150 expressionwas 8.4 while in the five lower responders the mean foldincrease in expression was 1.1. T cell samples from six of theten SLE patients showed a greater than two fold increase inDRIP150 and those six patients had renal involvement in theirlupus. The T cell samples from two SLE patients showed adecline in DRIP150 expression compared with the same T cellsamples cultured without estradiol. Both of those patientsinitially presented with arthritis and one exhibited Raynaud'ssyndrome. Neither of those patients exhibited renal

Figure 1 Expression of interferon-related genes in the activat-ed T cells from SLE patients (n=6) increased in response toestradiol but not in the activated T cells from healthy donors(n=5). Activated T cells were cultured without and with estradiol.Fold change (FC) values were calculated from gene profiling dataas a ratio of expression levels with estradiol compared toexpression levels without estradiol from the same individuals.The mean FC for thirteen genes representing those associatedwith the interferon-α pathway are displayed (1–13). Dark bars(black) are mean FC values from all T cell samples obtained fromcontrol females. Light bars (gray) are mean FC values from all Tcell samples obtained from female SLE patients. The horizontallines represent mean FC values for each group, 1.0 for the controlT cells and 1.6 for the SLE T cells. 1= DRIP150; 2= IFIT1; 3=IFNAR1; 4= INFAR2; 5= JAK1; 6= JAK2; 7= OAS1; 8= PTPN11; 9=PTPN2; 10= SEC14; 11= SOCS1; 12= STAT1; 13= STAT2.

433Human systematic lupus

involvement in their disease. DRIP150 expression in thecontrol and RA T cell samples showed a mean fold differenceof 1 (n=10) and 0.85 (n=5), respectively. The values for theRA T cell samples were closer to the controls than to the SLET cell samples indicating that altered DRIP150 expression is adisease effect rather than a response to medications.

Table 5 Estradiol differentially regulates expression of genes in

SLE DRIP150 FC IFIT1 FC Controls DRIP15

SLE 1 1.11 0.80 CTRL 1 1.06SLE 2 0.97 CTRL 2 1.02SLE 3 2.07 2.78 CTRL 3 1.71SLE 4 16.57 0.01 CTRL 4 0.62SLE 5 18.89 0.20 CTRL 5 1.25SLE 6 1.08 2.22 CTRL 6 0.63SLE 7 15.27 ND CTRL 7 0.74SLE 8 0.34 0.59 CTRL 8 1.80SLE 9 3.22 1.93 CTRL 9 0.97SLE 10 8.03 0.88 CTRL 10 1.06SLE 11 ND 5.15SLE 12 ND 6.60Mean FC 6.75 2.12 Mean FC 1.09Median FC 2.65 1.41 Median FC 1.04Std. Dev. 7.38 2.20 Std. Dev. 0.41

Data shown are the fold change (FC) in target gene expression from thedone.

Discussion

Previous results from our laboratory showed SLE T cellsexhibit an increased sensitivity to estradiol [6,7]. Adminis-tration of the ER antagonist Faslodex reduced diseaseactivity in some SLE patients [23]. We have extended thosefindings and used gene profiling to gain greater insight intosignaling pathways that are stimulated by estradiol and mayalter T cell function. We selected pathways that wereuniquely controlled by estradiol in all six female SLE and allfive control T cell samples obtained from healthy females. Itis important to note that such an approach will not detectaltered pathways in subsets of patients. Our rationale was toselect those pathways differentially regulated by estradiol inall T cell samples in order to identify signaling pathways thatwere targets in the majority of SLE patient's T cells.Confirming the validity of this approach was the identifica-tion of the T cell receptor signaling pathway as an estrogentarget shown in earlier studies [24] and known to be alteredin SLE T cells [15].

Estradiol increases signaling through PI3K/AKT, T cellreceptor, interferon, GM-CSF, Calcium and SAPK/JNK path-ways (Table 1) that are normally associated with regulating Tcell function. Estradiol increases expression of the PI3K/AKTpathway in SLE and not in control T cells. In normal T cells,activation of the PI3K/AKT pathway stimulates cell survival,glucose metabolism and inducible transcription [30]. It isimportant to note that alteration in the PI3K/AKT pathway issufficient to promote autoimmunity [15]. Estradiol enhancesthe T cell receptor signaling pathway which we [24,25], andothers [3,4,31] have shown is altered in human SLE T cells.Estradiol increases GM-CSF signaling uniquely in activated SLET cells (Table 1). GM-CSF is a critical hematopoietic growthfactor that affects circulating leukocytes and is produced byactivated T cells [32]. GM-CSF can increase antigen-inducedimmune responses and alter the Th1/Th2 cytokine balance inT cells. Estradiol also augmented the stress-activated proteinkinase (SAPK/JNK) pathway in SLE T cells. JNK1 and JNK2

volved in the interferon-α pathway in SLE T cells.

0FC IFIT1 FC Rheumatoidarthritis

DRIP150 FC IFIT1 FC

1.19 RA 1 1.18 0.931.60 RA 2 0.90 1.741.09 RA 3 1.22 2.920.05 RA 4 0.46 0.930.72 RA 5 0.49 0.890.741.252.342.080.74

1.18 Mean FC 0.85 1.481.14 Median FC 0.90 0.930.68 Std. Dev. 0.36 0.88

same T cell sample cultured without and with estradiol. ND= not

434 E. Walters et al.

stimulate the release of proinflammatory cytokines inactivated T cells and modulate the differentiation of Th1and Th2 cells [33].

Calcium signaling is dysregulated in SLE T cells by severalmechanisms [34–39]. Engagement of the T cell receptor leadsto a transient release of calcium that occurs earlier and is ofgreater magnitude in SLE compared with normal T cells [36].By contrast, long-term changes in calcium signaling isdiminished in SLE T cells and may underlie abnormal T cellactivation and/or the inability of SLE T cells to produce IL-2[37,38]. It is interesting to note that estradiol both stimulatesand reduces calcium signaling in SLE T cells (compare Tables 1and 2). This finding suggests that estradiol plays a role indifferentially affecting short term versus sustained calciumsignaling in SLE T cells. Some clues regarding the molecularmechanisms involved in this apparent paradox are suggestedby increased signaling through the pentose phosphate path-way in SLE T cells, which is increased by estradiol (Table 1).Previous studies have shown that the mitochondrial mem-brane in SLE T cells is hyperpolarized owing to reductions inATP and glutathione [38,39]. Mitochondrial hyperpolariza-tion and persistent ATP depletion is a critical checkpointthat predispose SLE T cells to undergo cell death by necrosisrather than apoptosis. The mitochondrial transmembranepotential is regulated in part by the supply of NADPH producedby the pentose phosphate pathway [39]. Since abnormal T cellactivation and cell death underpin the pathology of SLE it isimportant to clarify the estrogen-dependent steps thatcontribute to alterations in this metabolic pathway and maychange calcium signaling in SLE T cells.

Of the twenty signaling pathways that were uniquelydownregulated by estradiol (Table 2), the decrease in IL-2signaling is a central component of defective SLE T cellsignaling [40,41]. Local increases in nitric oxide at the site ofT cell receptor engagement reduce IL-2 production [42].Nitric oxide is regulated by NADPH and, as revealed by ourstudy, the pentose phosphate pathway is altered by estradiolin SLE T cells. Alternatively, IL-2 is regulated by a number oftranscription factors including the cAMP response element-binding protein (CREB). Of relevance for IL-2 regulation in SLET cells is the report from Katsiari et al. [43] showing thatprotein phosphatase 2A expression is higher in SLE thannormal T cells leading to reduced transcription of IL-2.Protein kinase A (PKA) is a major target of cAMP activation.PKA signaling is maintained by a balance between kinase andphosphatase activity and the pathway is deficient in SLE Tcells [44]. Although additional studies are required to identifythe targets for reduced cAMP signaling in response toestradiol, our data indicate that estradiol targets pathwaysknown to be markers for abnormal signaling in SLE T cells.

Estradiol uniquely downregulates 28 signal transductionpathways in activated control T cells (Table 4). Several ofthese pathways are known to be defective in SLE T cellsincluding cell receptor signaling [45], ERK/MAPK signaling[46,47], IL-4 [48], NF-κB [49] and apoptosis signaling [50].Although nothing is known about estradiol-dependent sig-naling in normal T cells, our results suggest that estradioldownregulates key signaling pathways in activated normal Tcells. Estrogen is considered a promoter of the immuneresponse [51]. Our results indicate that estradiol down-regulates numerous pathways in activated T cells andsuggests that failure to downregulate these same pathways

in SLE T cells could contribute to SLE T cell hyperactivation.The concept now requires further investigation because itsuggests downregulation is an important mechanism that isdefective in the SLE T cell response to estradiol.

In this report, we chose to further study the interferon-αpathway owing to its importance in SLE and other autoim-mune diseases [17,18,22]. Seven of thirteen genes identifiedin interferon-α signaling showed greater than a 1.5 mean-fold increase in the SLE T cell samples (Fig. 1). If regulationof this pathway is typical for estradiol targets, it suggeststhat estrogen-dependent upregulation of signaling will occurby increased expression of multiple genes within a pathwayrather than a large change in the magnitude of expression ofa single gene. We found that expression of IFIT1 in SLE T cellsamples was bimodal in that some patient's T cells respondedstrongly to estradiol while others were less responsive to thehormone. Changes in IFIT1 expression correlated moderatelywith SLE disease activity. Interestingly, T cells from patientswith renal disease showed the greatest sensitivity toestradiol. Other studies have found significant differencesin IFIT1 expression in SLE compared with control T cells[28,52] and increased expression correlated with renaldisease [53]. It is likely that our study underestimatesestradiol effects on IFIT1 expression since the T cells werecultured for 16 h in serum-free medium without addedinterferon-α.

The interferon stimulated gene factor 3 (ISGF3) is aheterotrimeric complex central to the regulation of IFN genetranscription [54,55]. Binding of IFN to cell surface receptorsactivates STAT1 and STAT2 proteins which complex withinterferon regulatory factor 9 (IRF9) forming the ISGF3activation complex. The ISGF3 complex enters the nucleuswhere it binds to specific DNA sequences and increases therates of transcription of interferon gene targets. While manyof the signaling events downstream of interferon signalingare known [22,27,54], the mechanisms involved in regulatingIFN-dependent gene transcription are less well understood.

Transcriptional coactivators are important modulators ofgene regulation. Coactivators regulate gene transcription bymodifying chromatin structure and enhancing promoteraccessibility. In addition, coactivators can recruit otherfactors to the promoter of responsive genes that enhance thetranscriptional response. DRIP150 is a member of a multi-protein complex that shares several subunits with themammalian Mediator complex [29,54]. Mediator complexesfunction as a bridge between distal transcription activatorsand RNA polymerase II. The Mediator complex is essential forthe regulation of NF-κB, SP1 and most nuclear hormonereceptor signaling pathways including the ER [13]. DRIP150interacts with both ER-α and ER-β [56]. ISGF3-mediatedtranscription is dependent on STAT2 interactions withDRIP150 [55]. Interestingly, one of the DRIP coactivators,TRAP220 showed preference for interacting with ER-β overER-α [56].

Our data indicate that DRIP150 expression is altered in SLEcompared with normal T cells in response to estradiol. Theseresults are the first to suggest that abnormal cofactorrecruitment to the promoter of genes involved in cellsignaling could lead to enhanced pathway activation anddisrupt normal T cell function. Over and or underexpressionof key signaling pathways is expected to contribute to thedevelopment of autoimmune disease. It may be that

435Human systematic lupus

abnormal coactivator expression recruits ERs to the promo-ters of genes that are not normally regulated by estradiol.Alternatively, inappropriate cofactor recruitment couldoccur by differential expression of ER subtypes such thatratio of ER-α/ER-β is altered in SLE T cells [57] leading toabnormal signal transduction. It is now clear that the timehas come to explore the role of coactivators and corepressorsas possible mechanisms underlying the deregulation ofnumerous genes in SLE T cells.

Acknowledgments

The research reported in this manuscript is funded by theNational Institutes of Health (AI49272 and P20RR0-16475 forthe INBRE Program of the National Center for ResearchResources). We are grateful to the patients who donatedblood for this study. We thank Phyllis Smotherman and ClarkBloomer for technical assistance. We are grateful to ourcolleagues at the KUMC Microarray Facility for their supportand guidance.

References

[1] D.T. Boumpas, B.J. Fessler, H.A. Austin 3rd, J.E. Balow, J.H.Klippel, M.D. Lockshin, Systemic lupus erythematosus:emerging concepts. Part 2: dermatologic and joint disease,the antiphospholipid antibody syndrome, pregnancy and hor-monal therapy, morbidity and mortality, and pathogenesis,Ann. Intern. Med. 123 (1995) 42–53.

[2] J. Zhang, A.M. Jacobi, T. Wang, R. Berlin, B.T. Volpe, B.Diamond, Polyreactive autoantibodies in systemic lupuserythematosus have pathogenic potential, J. Autoimmun.(2009) doi:10.1016/j.jaut.2009.03.011.

[3] K. Tenbrock, Y.T. Juang, V.C. Kyttaris, G.C. Tsokos, Alteredsignal transduction in SLE T cells, Rheumatology 46 (2007)1525–1530.

[4] V.C. Kyttaris, S. Krishnan, G.C. Tsokos, Systems biology insystemic lupus erythematosus: integrating genes, biology andimmune function, Autoimmunity 39 (2006) 705–709.

[5] R. Cervera, M.A. Khamashta, J. Font, G.D. Sebastiani, A. Gil,P. Lavilla, et al., Systemic lupus erythematosus: clinical andimmunologic patterns of disease expression in a cohort of 1,000patients. The European Working Party on Systemic LupusErythematosus, Medicine (Baltimore) 72 (1993) 113–124.

[6] V. Rider, N.I. Abdou, Gender differences in autoimmunity:molecular basis for estrogen effects in systemic lupus erythe-matosus, Int. Immunopharmacol. 1 (2001) 1009–1024.

[7] V. Rider, X. Li, N.I. Abdou, Hormonal influences in theexpression of systemic lupus erythematosus, in: G.C. Tsokos,P.C. Gordon, J.S. Smolen (Eds.), Systemic Lupus Erythematosus:A Companion to Rheumatology Elsevier, New York, NY, 2007,pp. 87–94.

[8] J.F. Cohen-Solal, V. Jeganathan, C.M. Grimaldi, E. Peeva, B.Diamond, Sex hormones and SLE: influencing the fate ofautoreactive B cells, Curr. Top. Microbiol. Immunol. 305(2006) 67–88.

[9] S. Khorasanizadeh, F. Rastinejad, Nuclear-receptor interac-tions on DNA-response elements, Trends Biochem. Sci. 26(2001) 384–390.

[10] B.W. O'Malley, Coregulators: from whence came these “mastergenes”, Mol. Endocrinol. 21 (2007) 1009–1013.

[11] I.M. Wolf, M.D. Heitzer, M. Grubisha, D.B. DeFranco, Coacti-vators and nuclear receptor transactivation, J. Cell. Biochem.104 (2008) 1580–1586.

[12] P. Ordentlich, M. Downes, R.M. Evans, Corepressors andnuclear hormone receptor function, Curr. Top. Microbiol.Immunol. 254 (2001) 101–116.

[13] D. Burakov, L.A. Crofts, C.P. Chang, L.P. Freedman, Reciprocalrecruitment of DRIP/mediator and p160 coactivator complexesin vivo by estrogen receptor, J. Biol. Chem. 277 (2002)14359–14362.

[14] N. Chiba, Z. Suldan, L.P. Freedman, J.D. Parvin, Binding ofliganded vitamin D receptor to the vitamin D receptorinteracting protein coactivator complex induces interactionwith RNA polymerase II holoenzyme, J. Biol. Chem. 275 (2000)10719–10722.

[15] P.S. Ohashi, T-cell signalling and autoimmunity: molecularmechanisms of disease, Nat. Rev., Immunol. 2 (2002) 427–438.

[16] J.W. Bauer, E.C. Baechler, M. Petri, F.M. Batliwalla, D.Crawford, W.A. Ortmann, et al., Elevated serum levels ofinterferon-regulated chemokines are biomarkers for activehuman systemic lupus erythematosus, PLoS Med. 3 (2006)2274–2284.

[17] E.C. Baechler, F.M. Batliwalla, G. Karypis, P.M. Gaffney, W.A.Ortmann, K.J. Espe, et al., Interferon-inducible gene expres-sion signature in peripheral blood cells of patients with severelupus, Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 2610–2615.

[18] M.K. Crow, K.A. Kirou, J. Wohlgemuth, Microarray analysis ofinterferon-regulated genes in SLE, Autoimmunity 36 (2003)481–490.

[19] M.K. Crow, K.A. Kirou, Interferon-alpha in systemic lupuserythematosus, Curr. Opin. Rheumatol. 16 (2004) 541–547.

[20] S.V. Kozyrev, M.E. Alarcon-Riquelme, The genetics and biologyof Irf5-mediated signaling in lupus, Autoimmunity 40 (2007)591–601.

[21] S.N. Kariuki, K.A. Kirou, E.J. MacDermott, L. Barillas-Arias, M.K.Crow, T.B. Niewold, Cutting edge: autoimmune disease riskvariant of STAT4 confers increased sensitivity to IFN-alpha inlupus patients in vivo, J. Immunol. 182 (2009) 34–38.

[22] E.C. Baechler, F.M. Batliwalla, A.M. Reed, E.J. Peterson, P.M.Gaffney, K.L. Moser, P.K. Gregersen, T.W. Behrens, Geneexpression profiling in human autoimmunity, Immunol.Rev. 210(2006) 120–137.

[23] N.I. Abdou, V. Rider, C. Greenwell, X. Li, B.F. Kimler,Fulvestrant (Faslodex), an estrogen selective receptordownregulator, in therapy of women with systemic lupuserythematosus. Clinical, serologic, bone density, and T cellactivation marker studies: a double-blind placebo-controlledtrial, J. Rheumatol. 35 (2008) 797–803.

[24] V. Rider, R.T. Foster, M. Evans, R. Suenaga, N.I. Abdou, Genderdifferences in autoimmune diseases: estrogen increases calci-neurin expression in systemic lupus erythematosus, Clin.Immunol. Immunopathol. 89 (1998) 71–80.

[25] V. Rider, S. Jones, M. Evans, H. Bassiri, Z. Afsar, N.I. Abdou,Estrogen increases CD40 ligand expression in T cells fromwomen with systemic lupus erythematosus, J. Rheumatol. 28(2001) 2644–2649.

[26] E.M. Tan, A.S. Cohen, J.F. Fries, A.T. Masi, D.J. McShane, N.F.Rothfield, et al., The 1982 revised criteria for the classificationof systemic lupus erythematosus, Arthritis Rheum. 25 (1982)1271–1277.

[27] M.K. Crow, Interferon pathway activation in systemic lupuserythematosus, Curr. Rheumatol. Rep. 7 (2005) 463–468.

[28] S. Ye, H. Pang, Y. Gu, J. Hua, X.G. Chen, C.D. Bao, et al.,Protein interaction for an interferon-inducible systemic lupusassociated gene, IFIT1, Rheumatology 42 (2003) 1155–1163.

[29] J. Lee, S. Safe, Coactivation of estrogen receptor alpha (ERalpha)/Sp1 by vitamin D receptor interacting protein 150(DRIP150), Arch. Biochem. Biophys. 461 (2007) 200–210.

[30] L.P. Kane, A. Weiss, The PI-3 kinase/Akt pathway and T cellactivation: pleiotropic pathways downstream of PIP3, Immunol.Rev. 192 (2003) 7–20.

436 E. Walters et al.

[31] J. Zikherman, A. Weiss, Antigen receptor signaling in therheumatic diseases. Arthritis Res, Ther. 11 (2009) 202, doi:10.1186/ar2528.

[32] Y. Shi, C.H. Liu, A.I. Roberts, J. Das, G. Xu, G. Ren, et al.,Granulocyte-macrophage colony-stimulating factor (GM-CSF)and T-cell responses: what we do and don't know, Cell Res. 16(2006) 126–133.

[33] M. Rincón, R.J. Davis, Regulation of the immune response bystress-activated protein kinases, Immunol. Rev. 228 (2009)212–224.

[34] D. Fernandez, E. Bonilla, N. Mirz, B. Niladn, A. Perl, Rapamycinreduces disease activity and normalizes T cell activation-induced calcium fluxing in patients with systemic lupuserythematosus, Arthritis Rheum. 54 (2006) 2983–2988.

[35] S. Feske, Calcium signaling in lymphocyte activation anddisease, Nature 7 (2007) 690–702.

[36] G.C. Tsokos, Calcium signaling in systemic lupus erythematosuslymphocytes and its therapeutic exploitation, Arthritis Rheum.58 (2008) 1216–1219.

[37] G. Nagy, A. Koncz, D. Fernandez, A. Perl, Nitric oxide,mitochondrial hyperpolarization and T-cell activation, FreeRadic. Biol. Med. (2007) 1625–1631.

[38] D. Fernandez, A. Perl, Metabolic control of T cell activation anddeath in SLE, Auotimmunity Rev. (2009) 184–189.

[39] A. Perl, P. Gergely Jr., A. Koncz, K. Banki, Mitochondrialhyperpolarization: a checkpoint of T cell life, death, andautoimmunity, Trends Immunol. 25 (2004) 360–367.

[40] J.C. Crispín, G.C. Tsokos, Transcriptional regulation of IL-2 inhealth and autoimmunity, Autoimmunity Rev. 8 (2009) 190–195.

[41] D. Gómez-Martín, M. Díaz-Zamudio, J.C. Crispín, J. Alcocer-Varela, Interleukin 2 and systemic lupus erythematosus:beyond the transcriptional regulatory net abnormalities,Autoimmun. Rev. 9 (2009) 34–39.

[42] S. Ibiza, V.M. Víctor, I. Boscá, A. Ortega, A. Urzainqui, J.E.O'Connor, et al., Endothelial nitric oxide synthase regulates Tcell receptor signaling at the immunological synapse, Immunity24 (2006) 753–765.

[43] C.G. Katsiari, V.C. Vasileios, Y.-T. Juang, G.C. Tsokos, Proteinphosphatase 2A is a negative regulator of IL-2 production inpatients with systemic lupus erythematosus, J. Clin. Invest 115(2005) 3193–3204.

[44] G.M. Kammer, Deficient protein kinase a in systemic lupuserythematosus: a disorder of T lymphocyte signal transduction,Ann. NY Acad. Sci. 968 (2002) 96–105.

[45] J. Zikherman, A. Weiss, Antigen receptor signaling in therheumatic diseases, Arthritis Res. Ther. 11 (2009) 202, doi:10.1186/ar2528.

[46] S. Gorjestani, V. Rider, B.F. Kimler, C. Greenwell, N.I. Abdou,Extracellular signal-regulated kinase 1/2 signalling in SLE Tcells is influenced by oestrogen and disease activity, Lupus 17(2008) 548–554.

[47] A.H. Sawalha, M. Jeffries, R. Webb, Q. Lu, G. Gorelik, D. Ray,et al., Defective T-cell ERK signaling induces interferon-regulated gene expression and overexpression of methylation-sensitive genes similar to lupus patients, Genes Immun. 9(2008) 368–378.

[48] R.R. Singh, IL-4 and many roads to lupus like autoimmunity,Clin. Immunol. 108 (2003) 73–79.

[49] K.D. Brown, E. Claudio, U. Siebenlist, The roles of the classicaland alternative nuclear factor-kappaB pathways: potentialimplications for autoimmunity and rheumatoid arthritis,Arthritis Res. Ther. 10 (2008) 212, doi:10.1186/ar2457.

[50] E. Maniati, P. Potter, N.J. Rogers, B.J. Morley, Control ofapoptosis in autoimmunity, J. Pathol. 214 (2008) 190–198.

[51] R.H. Straub, The complex role of estrogens in inflammation,Endocr. Rev. 28 (2007) 521–574.

[52] K.A. Kirou, C. Lee, S. George, K. Louca, I.G. Papagiannis, M.G.Peterson, et al., Coordinate overexpression of interferon-alpha-induced genes in systemic lupus erythematosus, ArthritisRheum. 50 (2004) 3958–3967.

[53] K.A. Kirou, C. Lee, S. George, K. Louca, M.G. Peterson, M.K.Crow, Activation of the interferon-alpha pathway identifies asubgroup of systemic lupus erythematosus patients withdistinct serologic features and active disease, ArthritisRheum. 52 (2005) 1491–1503.

[54] J.F. Lau, I. Nusinzon, D. Burakov, L.P. Freedman, C.M. Horvath,Role of metazoan mediator proteins in interferon-responsivetranscription, Mol. Cell Biol. 23 (2003) 620–628.

[55] J.Z. Chadick, F.J. Asturias, Structure of eukaryotic Mediatorcomplexes, Trends Biochem. Sci. 30 (2005) 264–271.

[56] A. Wärnmark, T. Almlöf, J. Leers, J.A. Gustafsson, E. Treuter,Differential recruitment of the mammalian mediator subunitTRAP220 by estrogen receptors ERalpha and ERbeta, J. Biol.Chem. 276 (2001) 23397–23404.

[57] V. Rider, X. Li, G. Peterson, J. Dawson, B.F. Kimler, N.I. Abdou,Differential expression of estrogen receptors in women withsystemic lupus erythematosus, J. Rheumatol. 33 (2006)1093–1101.