45. Pustaka Galon Dkk

Gene profiling reveals unknown enhancing andsuppressive actions of glucocorticoidson immune cells

JEROME GALON,*,‡,1,2 DENIS FRANCHIMONT,†,§,1 NAOKI HIROI,†,‡‡ GREGORY FREY,†

ANTJE BOETTNER,†† MONIKA EHRHART-BORNSTEIN,†,‡‡ JOHN J. O’SHEA,*GEORGE P. CHROUSOS,† AND STEFAN R. BORNSTEIN†,‡‡

*Lymphocyte Cell Biology Section, NIAMS, and †Pediatric and Reproductive Endocrinology Branch,NICHD, National Institutes of Health, Bethesda, Maryland 20892, USA; ‡INSERM U255, CurieInstitute, Paris, France; §Department of Gastroenterology, Erasme Hospital, Brussels, Belgium;and ††Department of Pediatrics, University of Leipzig, 04103 Leipzig, and ‡‡Department ofEndocrinology, University of Dusseldorf, 40225 Dusseldorf, Germany

ABSTRACT Glucocorticoids continue to be the ma-jor immunomodulatory agents used in clinical medicinetoday. However, their actions as anti-inflammatory andimmunosuppressive drugs are both beneficial and del-eterious. We analyzed the effect of glucocorticoids onthe gene expression profile of peripheral blood mono-nuclear cells from healthy donors. DNA microarrayanalysis combined with quantitative TaqMan PCR andflow cytometry revealed that glucocorticoids inducedthe expression of chemokine, cytokine, and comple-ment family members as well as of newly discoveredinnate immune-related genes, including scavenger andToll-like receptors. In contrast, glucocorticoids re-pressed the expression of adaptive immune-relatedgenes. Simultaneous inhibitory and stimulatory effectsof glucocorticoids were found on inflammatory Thelper subsets and apoptosis-related gene clusters. Incells activated by T cell receptor cross-linking, glucocor-ticoids down-regulated the expression of specific genesthat were previously up-regulated in resting cells, sug-gesting a potential new mechanism by which they exertpositive and negative effects. Considering the broadand continuously renewed interest in glucocorticoidtherapy, the profiles we describe here will be useful indesigning more specific and efficient treatment strate-gies.—Galon, J., Franchimont, D., Hiroi, N., Frey, G.,Boettner, A., Ehrhart-Bornstein, M., O’Shea, J. J.,Chrousos, G. P., Bornstein, S. R. Gene profiling revealsunknown enhancing and suppressive actions of glu-cocorticoids on immune cells. FASEB J. 16, 61–71(2002)

Key Words: DNA microarray � clinical medicine � immunesystem

Inflammation is the physiological homeostatic re-sponse elicited to shield our tissues from injuriousagents. The regulation of the immense array of cellularand molecular mediators involved in inflammation isharmonized by the immune and neuroendocrine sys-

tems (1, 2). Adrenal-derived glucocorticoids (GC) assistthe innate and adaptive immunity to eliminate toxins,tumor cells, and foreign pathogens to limit tissuedamage (1, 3, 4). The broad and triumphant clinicaluse of synthetic GC has improved the life of millions ofpatients with inflammatory/autoimmune diseases andgraft rejection during the past 50 years (5). Renewedinterest is surfacing within the field of GC therapy forpatients with severe sepsis, septic shock, and the acuterespiratory distress syndrome (ARDS), based on evi-dence from several clinical trials where prolonged GCtreatment resulted in significant improvement and adecreased mortality (6–9). However, GC may alsoworsen the course of such diseases and prime deleteri-ous inflammatory cascades, resulting in death (10).

A great deal of effort has been focused on trying tounderstand the cellular and molecular mechanisms ofaction of GC (5, 11). In vivo and in vitro studies havesuggested both enhancing and suppressive effects ofGC on the immune and inflammatory response, andeven more discrepancies have emerged (3, 12, 13). Acomprehensive analysis of gene regulation by GC hasnot yet been performed. Therefore, we analyzed theeffect of GC on the gene expression profile of humanperipheral blood mononuclear cells (PBMC) fromhealthy donors treated with dexamethasone (Dex). Toaccomplish this, we used the novel technique of DNAmicroarray chip analysis in combination with onlinePCR and flow cytometry. First, our study showed thatGC regulated numerous newly discovered genes thatplay critical roles in innate and adaptive immuneresponses. Second, the data revealed a role for GC notonly as immunosuppressants, but also as major immu-nopermissive and immuno-enhancing agents. Finally,we showed that GC could regulate some genes in

1 These authors contributed equally to the work.2 Correspondence: Laboratoire d’Immunlogie Cellulaire et

Clinique, Unite INSERM U255, Centre de Recherches Biomedi-cales des Cordeliers, 15 Rue de l’Ecole de Medecine, 75270 ParisCedex 06 France. E-mail: [email protected]

610892-6638/02/0016-0061 © FASEB

opposite directions, depending on the state of activa-tion of the target cells. This study supports a major andhighly complex regulatory role of GC in human im-mune functions.

MATERIALS AND METHODS

Antibodies and cells

Monoclonal mouse anti-human CD3 and anti-CD28, FITC- orPE-conjugated monoclonal anti-CD68, anti-TNFR1, anti-CD36, anti-CD163, anti-CD27, anti-CD49d, anti-CD39, anti-HLA-DR, anti-CD74, anti-TCR-��, anti-TCR-��, anti-CD127,and isotype-matched IgG controls were purchased fromPharMingen (San Diego, CA). PBMC from healthy donors

were isolated by Ficoll-Paque gradient centrifugation (Phar-macia, Piscataway, NJ), activated or not with coated anti-CD3/anti-CD28 antibodies at 10 �g/ml, and stimulated with dexa-methasone (10�7 M) for 18 h in RPMI 1640 containing 1%FCS. TaqMan primers are described in the online Table 1supplemental.

Microarrays

Peripheral blood mononuclear cells were prepared as de-scribed. RNA extraction was performed (RNAgents, Promega,Madison, WI) and mRNA was purified with oligoTex mRNAisolation columns (Qiagen, Valencia, CA). The cDNA madefrom 600 ng of purified mRNA from each sample was labeledwith the fluorescent dyes Cy5 and Cy3. The two cDNA probeswere mixed and simultaneously hybridized to a microarraycontaining 9182 human expressed genes (GEM microarray;

TABLE 1 Supplemental. Primer sequences, size and Genebank accession numbera

Name Primer sequences (5�-3�) Primer size Genebank no.

Indoramine-pyrrole S AGTCCGTGAGTTTGTCCTTTCAA 967–989 NM_002164AS TTTCACACAGGCGTCATAAGCT 1033–1012TM CCCGCAGGCCAGCATCACCT 991–1010

TLR4 S AGAGTTTCCTGCAATGGATCAAG 1927–1949 NM_003266AS TTATCTGAAGGTGTTGCACATTCC 2008–1985TM TTCGTTCAACTTCCACCAAGAGCTGCCT 1956–1983

CD44 S AAAAATGGTCGCTACAGCATCTC 140–162 NM_000610AS GGTGCTATTGAAAGCCTTGCA 205–185TM AGGTCAGCGGCCTCCGTCCG 164–183

CD163 S GGTCGCTCATCCCGTCAGT 3213–3231 NM_004244AS CGAAAATGGCCAACAGAACA 3282–3263TM CCCAAGGATCCCGACTGCAATAAAGGAT 3233–3260

Integrine alphal S TCTGGTTTTAAGTAGCAGCAATCAA 3261–3285 X68742AS GCAGCATTAACAGCAACAATCC 3391–3370TM TGCCCGGTAGCCCATCTTTGGATATTT 3305–3331

MARCO S GACAGCCCGTCCTTCTCCTT 268–287 NM_006770AS GACTTGCAGCCTCGATGCA 345–327TM AGCACACCCTGGAGAACACCTGGCT 297–321

THBS1 S CATCCGCAAAGTGACTGAAGAG 1002–1023 NM_003246AS GTACTGAACTCCGTTGTGATAGCATAG 1086–1060TM CCAATGAGCTGAGGCGGCCTCC 1037–1058

STAT1 S GTGGAAAGACAGCCCTGCAT 1151–1170 NM_007315AS ACTGGACCCCTGTCTTCAAGAC 1217–1196TM CAACGCACCCTCAGAGGCCGC 1173–1193

IRF4 S GCCCAGGTTCACAACTACATGAT 612–634 NM_002460AS TTTCCGGGTGTGGCTGAT 690–673TM ACCGAAGCTGGAGGGACTACGTCCC 646–670

Diubiquitin S ATGGCTCCCAATGCTTCCT 19–37 NM_006398AS TGGCATCAAAGGTCATTAAATCC 91–69TM CCTCTGTGTGCATGTCCGTTCCGA 39–62

PAI2 S GCATGGAGGACGCCTTCA 1055–1072 NM_002575AS ACAGGTCATTCCTCTCCGACAT 1120–1099TM CAAGGGACGGGCCAATTTCTCAGG 1074–1097

CD127 S GCATGTGACGCCCCTATTCT 1148–1167 XM_004013AS GGCCCATTCTTGCCACTCT 1215–1197TM TCCTCTTCCAGGTCCCTAGACTGCAGG 1169–1195

IL10 S ACGGCGCTGTCATCGATT 399–416 M57627AS TGGAGCTTATTAAAGGCATTCTTCA 479–455TM TGTGAAAACAAGAGCAAGGCCGTGGA 424–499

CCR2 S TGAGTAACTGTGAAAGCACCAGTCA 901–925 NM_000648AS GCAGTGAGTCATCCCAAGAGTCT 974–952TM CTGGACCAAGCCACGCAGGTGAC 927–949

a S: sense primer; AS: antisense primer; TM: TaqMan probe; Primer size: Base pairs; TLR4: Toll-like receptor 4; MARCO: macrophagereceptor with collagenouse structure; THBS1: thrombospondin 1; STAT1: signal transducer and activator of transcription 1; IRF4: interferonregulatory factor 4; PAI2: plasminogen activator inhibitor type2; CD127: interleukin 7 receptor; IL10: interleukin 10; CCR2: chemokine receptor 2.

62 Vol. 16 January 2002 GALON ET AL.The FASEB Journal

Incyte, St. Louis, MO). The microarray was scanned and theintensity of the fluorescence at each array element wasproportional to the expression level of that gene in thesample. The ratio of the two fluorescence intensities provideda quantitative measurement of the relative gene expressionlevel in the two cell samples. The detectable dynamic range ofindividual mRNA in a sample range from 2 to 2000 pg andreproducibility was assessed on a validation study (www.incyte.com/incyte science/technology/gem/reproducibility.shtml). Using specially designed DNA control elements and aknown RNA set added to each probe labeling reaction, allGEM microarray hybridizations are assessed for labeling,hybridization efficiency, and sensitivity. A 96-well plate isarrayed on each GEM microarray that contains a variety ofcontrol elements arrayed in quadruplicate. The elements arecomposed of DNA fragments derived from inter-ORF regionsin Saccharomyces cerevisiae. These fragments are all �1000 bp inlength and have minimal homology to any known gene, toensure that no cross-hybridization occurs with experimentalsamples. A control set of 14 RNA molecules is generated fromthese fragments using PCR with composite primers thatinclude a T7 promoter and a 30-base oligo-dT sequence. Asimple transcription reaction then generates artificial mRNAfragments that hybridize specifically to individual controlelements. These fragments are added to the two probe-labeling reactions in known amounts, together with theexperimental samples. A set of four fragments is added inincreasing concentrations in both the Cy3 and Cy5 reactionsto control for signal sensitivity. Six additional are added inknown ratios in the two fluorescent labeling reactions tocontrol for expected ratios. Four RNA fragments previouslyconverted to fluorescent DNA are either arrayed onto thecontrol plate or added to the experimental sample to controlfor overall scanning and hybridization signal, independent ofeither labeling reaction. Additional elements composed ofeither housekeeping genes or mixed cDNA libraries arearrayed in the last two rows of the control plate. Theseelements are hybridized by probe from the experimentalsample, and their signal intensity is used to measure experi-mental RNA quality. All signal intensities in the control plateas well as total GEM microarray hybridization signals arecollected. The values are then compared with a set ofminimum intensity criteria for each particular control. Be-cause the controls differentiate between processes in thelabeling and hybridization reaction, the quality of experimen-tal RNA can be measured independent of labeling or hybrid-ization efficiency. An absolute balanced differential expres-sion 1.4 was considered significant based on threeparameters: reproducibility of differential expression basedon a statistical study encompassing 70 differential hybridiza-tion, known genes regulated by GC under these conditions,and TaqMan PCR confirmation. Differential expression accu-racy and precision were determined by hybridizing samplesagainst themselves in replicates of 10. The expected differen-tial expression ratio for each gene being 1, experimentaldifferential expression (calculated from 100,000 data points)ranged from 0.99977 to 0.99998. Known and unknown GC-inducible genes were confirmed for their up- or down-regulation by GC using quantitative online PCR or flowcytometry. GemTools software (version 2.5.0) was used toanalyze the gene distribution, and genes were clustered basedon GemTools clustering software and published literature.

Online TaqMan PCR

RT-PCR experiments were carried out according to theThermoscript RT-PCR system kit for RT-PCR instructionsprovided by the manufacturer (Life Technologies, Gaithers-burg, MD). Total RNA (1 �g) of human lymphocyte were

reverse transcribed to complementary DNA (cDNA) by areaction containing 2 mM deoxynucleotide mix, 100 mMDTT, 40 units RNase inhibitor, 50 ng random primer, and 15units Thermoscript reverse transcriptase. The reaction wasrun 25°C for 10 min and 50°C for 50 min, heated at 85°C for5 min, then cooled to 4°C.

To quantify expression of IDO, TLR-4, CD44, CD163,integrin alpha1, MARCO, TSP-1, STAT1, PLA2, IRF-4,CD127, IL-10, CCR2, and diubiquitin, we applied the TaqManPCR using the 7700 Sequence Detector (Perkin Elmer Ap-plied Biosystems, Foster, CA). Reactions contained 1 Taq-Man Universal PCR Master Mix, 900 nM of forward andreverse primers and 200 nM for the TaqMan-probes (Table 1,supplemental). 18 S primers and probes were added at 50 nM.Thermal cycling proceeded with 50 cycles of 95°C for 15 s and60° C for 1 min. Input RNA amounts were calculated withrelative standard curves for both the mRNAs of interest and 18S.

Statistical analysis

Microarrays

The genes were classified into three categories: not regulated,up-regulated, and down-regulated. The statistical significanceof the correlation between the two DNA microarray experi-ments was determined using correlation matrix algorithm,Kendall correlation algorithm, or Spearman correlation algo-rithm. Similar results were found with all algorithms. Whenno fluorescent intensities were detected in one or bothexperiments, the corresponding genes were removed fromthe statistical analysis. Correlation analyses were performedby plotting differential gene expression of one experimenta-gainst the value from the other experiment. The correlationcoefficient (r) and the statistical significance (P) were deter-mined.

Taqman PCR

Input RNA amounts were calculated with relative standardcurves for all mRNA of interest and 18S RNA. Division bythe amount of 18S RNA in each sample corrected theamounts of input mRNAs. Specific mRNA transcript levelsafter Dex stimulation were expressed as percentage ofincrease vs. unstimulated cells. Data are presented asmean � se and analyzed by ANOVA T test for each genefrom individual healthy donors (n�6). The statistical sig-nificance (P) was determined and is represented for eachhistogram.

Flow cytometry

Flow cytometry (FACS) was previously described (14). Cellswere washed and incubated with conjugated antibodies andisotype-matched IgG-PE/and FITC-IgG as controls for 30 minat 4°C. Cells were washed three times with phosphate-buff-ered saline containing 0.5% bovine serum albumin andanalyzed by flow cytometry on a FACSCalibur flow cytometer(Becton Dickinson, San Jose, CA).

RESULTS

Global analysis of gene expression profile

We analyzed PBMC gene expression changes mediatedthrough GC (Fig. 1). In a DNA microarray containing

63GENE EXPRESSION PROFILES OF GLUCOCORTICOIDS ON IMMUNE CELLS

9182 human expressed genes, 9% were considereddown-regulated (845�41 genes) and 12% up-regulated(1125�135 genes) (Fig. 1a). The sensitivity and repro-ducibility of DNA microarrays are highlighted by thefact that appropriate changes of known GC-regulatedgenes were detected in separate DNA microarray ex-periments. Statistical significance of the correlationbetween the two DNA microarray experiments wasdetermined. When no fluorescent intensities were de-tected in one or both experiments, the correspondinggenes were removed from the statistical analysis. Oneexperiment showed a greater difference among gene

expression values. This is represented by the histograms(Fig. 1c). However, even with a broader differentialexpression, the genes were coregulated in the samemanner in both experiments. Correlation analyses wereperformed by plotting differential gene expression ofone experiment against the value from the other ex-periment. As shown in Fig. 1c, on average, we found acorrelation between data points from the two experi-ments. The correlation coefficient (r�0.85) confirmedthe good correlation between experiments and thecorrelation appears to be statistically significant(P 0.0001).

Using independent experimental samples (n�6),online PCR results and/or FACS analysis confirmedin all cases the direction and magnitude of the geneexpression changes observed (panels a and b inTables 1–3; data not shown). Global gene expressionanalysis, based on GemTools clustering software andpublished literature, enabled us to quantitate therelative contribution of GC to changes in genecluster expression. Of six major clusters analyzed,immune response-related genes and unknown genes(ESTs) were mostly regulated. The immune clusterwas divided into subclusters and subcategories (Fig.1b). The first group of immune genes regulated byGC included key anti-inflammatory and proinflam-matory factors (Table 1). Global analysis revealedsymmetrical up- or down-regulation within the in-flammation subcluster. In contrast, the subclusteringanalysis revealed a shift toward innate compared withadaptive recognition (Fig. 1b). Within the immunesystem, a second group involved genes related to theinnate immune response that constitutes the first lineof defense during an inflammatory reaction (Table2). Finally, we identified a third clusters of genesinvolved the adaptive immune response (Table 3).This group of genes specifically directs the appropri-ate type of cellular or humoral immune reactionaccording to the nature of the injurious agent.

GC as anti-inflammatory agents

Inflammation is an important homeostatic mecha-nism that limits the effects of injurious agents, buthas the potential for inducing host tissue damageand must be controlled by the organism. GC suppressinflammation by down-regulating the expression ofproinflammatory cytokines such as IL-1�, lympho-toxin-�, IL-1�, IL-8, IFN-�, and IFN-� (11). Indeed,the expression of these cytokines was decreased byGC (Table 1, panel a). A balance between the effectsof pro- and anti-inflammatory cytokines may deter-mine the outcome of the inflammatory response.Cytokines such as transforming growth factor (TGF)-�1–3 and IL-10 suppress the production of proin-flammatory mediators. We found that the expressionof TGF-�3, IL-10, and IL-10R was up-regulated by GC,supporting their anti-inflammatory action (Table 1).An important group of proinflammatory genes is thatof chemokines, small peptides that facilitate chemo-

Figure 1. Global gene expression analysis. Untreated andDex-treated PBMC were analyzed for gene expression usingDNA microarray. a) Clustering analysis. Total number ofgenes up- or down-regulated among 6 major clusters. b)Subclustering analysis. The immune cluster was divided intosubclusters and subcategories; percentage of GC-regulatedgenes within the immune cluster is presented. c) Correlationanalysis was performed by plotting differential gene expres-sion of one experiment against the value from the otherexperiment. Dot plot represents the correlation between datapoints from the two experiments and histograms representthe frequency of differential expression. Correlation coeffi-cient (r�0.85) between experiments (P 0.0001).


taxis of leukocytes from the circulation to the tissues.The prototypical chemoattractant IL-8 causes neutro-phils to degranulate and cause tissue damage. Theexpression of many chemokines, including IL-8, Mip-1�, Mip-3�, Mcp-2, Mcp-3, Mcp-4, TARC, andeotaxin, was down-regulated by dexamethasone treat-

ment (15, 16). In contrast, we found that I-309, IP-10,and fractalkine were up-regulated by GC (Table 2).Finally, IL-1RII, a decoy receptor limiting the delete-rious effects of IL-1, was among the genes moststrongly up-regulated by GC (17).

Thus, many proinflammatory ligands were down-

TABLE 1.


regulated whereas anti-inflammatory soluble mediatorswere mostly up-regulated, by GC, highlighting theirimmunosuppressive and protective role against inflam-mation.

GC induce proinflammatory mediators

A different pattern of regulation was observed forimmune-related membrane receptors and intracellularmediators. Among genes most strongly up-regulated byGC were some proinflammatory receptors IL-1RI, thegp130 subunit of the IL-6 type receptors, IL-8R, IFN�R,CSFRI, CFSRII, IFN�RI, IFN�RII, and TNFR familymembers (Table 1). Some of the intracellular media-tors from the IL-1/Toll signaling pathway, such asIRAK, were up-regulated by GC. In contrast, the anti-inflammatory mediator, IL-1Ra, a soluble receptor an-tagonist released during inflammation, was among thegenes most strongly down-regulated by GC (18). An-

other hallmark of inflammation is the influx of im-mune cells, mediated by chemokines and chemotacticfactors of the complement system (19). Chemokinereceptors or associated molecules, including Dez, anorphan G receptor-coupled protein, CCR1, CCR2, andCCR2-like, were up-regulated by GC. The activation ofcomplement forms the first line of defense and causesthe release of potent chemotactic complement factorsC3a, C4a, and C5a. The proinflammatory componentsof the classical pathway (C1q, C3, C5) and receptorsC3aR1, CR2, and C5aR1 were up-regulated. Our anal-ysis also revealed that C8, a terminal complementcomponent involved in the lytic activity of complement,was up-regulated by GC. In contrast, the two inhibitorstightly controlling the complement cascade, C1-INHand C4bp, were down-regulated.

The release of matrix metalloproteinases (MMPs)allows leukocytes to extravasate and penetrate tissues, akey event in inflammatory disease (20). They also act as

TABLE 2.


regulatory molecules by functioning in enzyme cas-cades and processing matrix proteins such as cytokines,growth factors, and adhesion molecules to generatefragments with biological effects. With the exception ofMMP-9, metalloproteinases were up-regulated by GC(Table 2). Other molecules promoting chemotaxis ofleukocytes to inflammatory sites, such as the throm-bospondins (TSPs) and their APC receptor (CD36), areabundantly expressed in chronically inflamed tissues.Thrombospondins are structurally related to MMPs and

regulate their functions (21, 22). The expression of TSP-1,2, 4 and CD36 was strongly up-regulated by GC (Table 1,panels a, b). In contrast, TSPs membrane receptor on Tcells (CD47) was down-regulated by GC (Table 2).

Therefore, proinflammatory mediators, such ascomplement, or cytokine and chemokine receptorswere up-regulated by GC whereas soluble inhibitorsof these pathways were down-regulated, supportingthe positive, enhancing role of GC in the inflamma-tory response.

TABLE 3.


GC support innate immunity and regulate specificgenes in opposite directions depending on theactivation state of the cells

A key aspect of the innate immune response is theability to discriminate between self and non-self mole-cules. This is achieved through pattern recognitionreceptors, which directly recognize molecular epitopesexpressed by microbes. Scavenger receptors (SRs) arealso involved in the innate immune response for theremoval of damaged tissue and debris, and recognize awide variety of pathogens. CD163 expression is down-regulated by proinflammatory and up-regulated byanti-inflammatory mediators and GC (23, 24). Here weshow that CD163, as well as eight other scavengerreceptors, were strongly up-regulated by GC (Table 2,panels a, b).

Finally, the IL-1R/Toll-like receptor (TLR) super-family has emerged as an expanding family of receptorswhose function is to respond rapidly to infection andinjury (25). TLR-4 and TLR-2 mediate the host re-sponse to gram-negative and gram-positive bacteria,respectively, and the finding that TLR-4 is importantfor responses to LPS may allow for novel means tointervene therapeutically during sepsis (25, 26). Threemembers of the Toll family were regulated by GC;TLR-4 and TLR-2 were up-regulated, and TLR-3 down-regulated (Table 2). That GC regulate this superfamilycould be critical for many aspects of inflammation andhost defense, since TLRs control innate immune re-sponses in vivo.

Besides the positive and negative role of GC onimmune gene clusters, we analyzed the potential differ-ential action of GC at the single gene level. Theregulation of genes implicated in innate and adaptiveimmune response in cells at different activation stageswas of particular interest. To mimic an antigenic re-sponse, we analyzed the action of GC on humanimmune cells activated by T cell receptor cross-linkingwith anti-CD3 and anti-CD28 antibodies. Among 20genes tested by online PCR under these differentconditions, 25% revealed a bidirectional action of GC.For instance, GC induced TLR-4 expression in theresting condition, yet after T cell receptor activation,GC decreased TLR-4 expression (n�7) (Fig. 2a, b).Indoleamine 2,3-dioxygenase (IDO), which partici-pates in innate and adaptive immune responses, is anovel antioxidant defense molecule. It prevents oxida-tive tissue damage and suppresses T cell-mediatedimmune responses during inflammation (27, 28). Theexpression of IDO was oppositely regulated by GC inresting and activated conditions (n�7) (Fig. 2c, d).

These results illustrate the complex bi-directionalregulation of a single gene by GC, depending on thestate of activation of the cells. This study reveals a majorpositive role of GC on the innate immune system, andsupports differential actions of GC on the innate andadaptive immune response during an antigen-mediatedimmune reaction.

GC inhibit adaptive immunity

Innate immunity is the first line of defense and func-tions to protect the host from infection while sloweradaptive (acquired) responses are developing. Co-stimulatory molecules such as CD40-ligand (CD40L)/CD40 are central regulators of the adaptive immuneresponse, and we found that GC down-regulated CD40(Table 3). CD40 regulates immune responses throughits ability to enhance the expression of B-7 costimula-tory molecules and through the increased productionof IL-12, a major Th1 cytokine, by monocytes anddendritic cells. Thus, GC appeared to have an impor-tant role in down-regulating CD4� T cell-dependentTh1 immune responses (29, 30). Furthermore, inter-feron regulatory factors (IRF) -1, -3, and -7 induced byTh1 cytokines were down-regulated by GC. Conversely,the expression of c-maf, the transcriptional regulatoryprotein involved in Th2 differentiation and IL-4 tran-scription, was up-regulated by GC. Thus, corticosteroidsappear to bias the immune response toward Th2 (31–33). Another T cell type that produces only transform-ing growth factor-� (TGF-�), but not IL-4 or IFN-�, hasbeen termed a Th3 cell. This type of cell has beensuggested to be important in mediating immune sup-pression in autoimmune disease. Six members of theTGF family [TGF-�, TGF-�3, TGF-�-induced (68 kDa)protein, TRAP-1, TGF-�RI, TGF-�RII, and TGF-�RIII]were up-regulated by GC (Table 1). Glucocorticoids arepotent immunosuppressants and induce apoptosis. Atthe same time, they up-regulate �50% of antiapoptotic-and proliferation-related molecules (Table 3 and datanot shown).

Figure 2. Quantitation of TLR-4 (a, b) and IDO (c, d) mRNAexpression by TaqMan PCR in resting (a, c) and CD3/CD28-activated (b, d) PBMC stimulated or not with the indicatedconcentrations of Dex. Data represent the mean � se (n�7).*P 0.05; **P 0.01; ***P 0.001.


Given the essential function of MHC-II molecules inT cell activation, the molecular mechanisms regulatingtheir complex expression profile obviously representan important parameter in the normal physiologicalcontrol of the immune response (34). Transcription ofMHC-II, CD74 (the invariant chain of the class II MHCmolecules, Ii), and MHC-II-DM genes is up-regulatedby the major CIITA complex, which is induced byintracellular signals (such as STAT1) generated byIFN-�. Products of the MHC-II-DM genes are essentialfor the peptide loading step and are modulated byMHC-II-DO, which enhances stringent MHC/peptideassociation through antigen receptor recognition (35).Among the adaptive immunity clusters, 17 genes encod-ing MHC molecules, Ii, MHC-II-DM and -DO, TCR-�,TCR-�, granzyme A and B, CD86, and STAT1 andSTAT5a were down-regulated by GC (Table 3, seepanels a, b). A remarkably consistent and coherentpicture emerged, where GC down-regulated positiveacting elements of the MHC-II expression and allfunctional units necessary for optimal antigen recogni-tion. Therefore, these data support a role of GC ininhibiting adaptive immunity.

DISCUSSION

The ubiquitous distribution of the glucocorticoid re-ceptor renders cell differentiation crucial in discrimi-nating the effects of GC on different physiologicalsystems. Our analysis was carried out on human PBMCand represents a large panel of a specific set of genesthat participate in the immune and inflammatory re-sponse. This panoramic study of gene regulation re-veals that GC regulate numerous newly discoveredgenes and control the expression of gene families, suchas SRs, TLRs, TSPs, complement, chemokine, TNF,TGF-�, and MHC families. It also elucidates new mech-anisms of GC action by unveiling unexpected regulatedgenes or gene (super) families and shows how GC mayfacilitate one biological signal while counteracting itsinhibitor. This is well documented with the selectivemodulation of the complement cascade activation andinhibition and the regulation of cytokines/chemokinesand their receptors. Finally, it is noteworthy that withinthe same gene family, some genes are stimulated andothers repressed, increasing the complexity of GCaction. Potential sources of discordance between RNAand protein levels include translational control andaltered protein stability, and protein analysis may un-cover post-translational modifications that are not pre-dictable at the RNA level. Therefore, we analyzedprotein expression by FACS analysis; for all membranereceptors tested, a perfect correlation between RNAand protein expression was found. Thus, these resultsillustrate the broad and coherent regulation of theimmune system by GC.

Glucocorticoids clearly have immunosuppressive ac-tions. In fact, anti-inflammatory genes such as IL-10,TGF-�, and IL-1Ra, whose deletion in mice results in

increased susceptibility to inflammatory disease (36,37), are strongly increased by GC. However, a morecomprehensive genetic dissection of the inflammatoryand immune response shows a more selective andcomplex picture of the actions of GC. First, at the genecluster level, our study shows a bidirectional action ofGC, which are both immunostimulatory and immuno-suppressive at the same time even for the inflammationcluster (12). They seemed to prime and enhance theinnate immune response while repressing part of theadaptive immune response in a resting state. Thissuggests that GC help clear antigens by stimulating celltrafficking, scavenger systems, and matrix metallopro-teinases while they halt cellular immune responses byinhibiting antigen presentation and T cell activation.Second, at the single gene level, we found an oppositeregulation of molecules implicated in the inflammatoryand immune responses by GC, depending on the stateof activation of the cells. This effect may underline thatthe signaling through the glucocorticoid receptor sys-tem may be mediated by a different set of transcriptionfactors, coactivators, and corepressors, allowing a switchin promoter activity depending on the state of cellularactivation (38). Therefore, GC may drive the inflamma-tory and immune response in a context-dependentfashion, which could be related to the state of disease(13, 39). Indeed, the complete understanding of thecomplex GC-mediated gene regulation may shed lighton the variable responses to GC, in patients withGC-sensitive or -resistant inflammatory and autoim-mune diseases (33, 40).

Glucocorticoids increase susceptibility to intracellu-lar and opportunistic infections, which appears toresult from the profound immune suppression of anti-gen cell presentation and adaptive immune responseinduced by these hormones. At the same time, GCappear highly beneficial in the presence of an over-whelming systemic inflammation when administered ina sustained fashion throughout the course of the dis-ease. This is observed in sepsis, septic shock, ARDS,and, less frequently, in meningococcal meningitis (6, 9,41, 42). However, their delayed and time-limited usecan be devoid of beneficial effects and even disastrous(10). This is clearly documented by the enhancedTLR-4 expression by GC, which may be progressivelylost in the later phase of the disease, reflecting the clearpermissive action of GC early on the innate immuneresponse. Depending on the state of activation of thecells, the differential GC actions underscore that thetiming of GC administration is crucial and may explainsome of the harmful effects of GC.

Therefore, clinical experience, together with a globalmolecular approach of gene regulation, may reveal theefficient timing of GC treatment and will allow a betterselection of patient groups that are likely to producethe most benefit. Although it is possible that peripheralblood mononuclear cells from patients with differentdiseases may respond differently from healthy donors,microarray screening techniques in combination withquantitative RNA and protein assays will enable testing


the profile of new steroids in the search for discrimi-nating between desired and adverse effects. Thus,synthetic steroids could be designed to act more spe-cifically on certain components of the immune re-sponse, selectively enhancing or suppressing such aresponse.

REFERENCES

1. Chrousos, G. P. (1995) The hypothalamic-pituitary-adrenal axisand immune-mediated inflammation. N. Engl. J. Med. 332,1351–1362

2. Wilder, R. L. (1995) Neuroendocrine-immune system interac-tions and autoimmunity. Annu. Rev. Immunol. 13, 307–338

3. Dhabhar, F. S., and McEwen, B. S. (1999) Enhancing versussuppressive effects of stress hormones on skin immune function.Proc. Natl. Acad. Sci. USA 96, 1059–1064

4. Sapolsky, R. M., Romero, L. M., and Munck, A. U. (2000) Howdo glucocorticoids influence stress responses? Integrating per-missive, suppressive, stimulatory, and preparative actions. En-docr. Rev. 21, 55–89

5. Boumpas, D. T., Chrousos, G. P., Wilder, R. L., Cupps, T. R., andBalow, J. E. (1993) Glucocorticoid therapy for immune-medi-ated diseases: basic and clinical correlates. Ann. Intern. Med. 119,1198–1208

6. Bollaert, P. E., Charpentier, C., Levy, B., Debouverie, M.,Audibert, G., and Larcan, A. (1998) Reversal of late septic shockwith supraphysiologic doses of hydrocortisone. Crit. Care Med.26, 645–650

7. Bronicki, R. A., Backer, C. L., Baden, H. P., Mavroudis, C.,Crawford, S. E., and Green, T. P. (2000) Dexamethasonereduces the inflammatory response to cardiopulmonary bypassin children. Ann. Thorac. Surg. 69, 1490–1495

8. Meduri, G. U., Headley, A. S., Golden, E., Carson, S. J.,Umberger, R. A., Kelso, T., and Tolley, E. A. (1998) Effect ofprolonged methylprednisolone therapy in unresolving acuterespiratory distress syndrome: a randomized controlled trial.J. Am. Med. Assoc. 280, 159–165

9. Meduri, G. U., and Kanangat, S. (1998) Glucocorticoid treat-ment of sepsis and acute respiratory distress syndrome: time fora critical reappraisal. Crit. Care Med. 26, 630–633

10. Cronin, L., Cook, D. J., Carlet, J., Heyland, D. K., King, D.,Lansang, M. A., and Fisher, C. J. (1995) Corticosteroid treat-ment for sepsis: a critical appraisal and meta-analysis of theliterature. Crit. Care Med. 23, 1430–1439

11. Ashwell, J. D., Lu, F. W., and Vacchio, M. S. (2000) Glucocorti-coids in T cell development and function. Annu. Rev. Immu-nol.18, 309–345

12. Wilckens, T., and De Rijk, R. (1997) Glucocorticoids andimmune function: unknown dimensions and new frontiers.Immunol. Today 18, 418–424

13. Franchimont, D., Martens, H., Hagelstein, M. T., Louis, E.,Dewe, W., Chrousos, G. P., Belaiche, J., and Geenen, V. (1999)Tumor necrosis factor alpha decreases, and interleukin-10 in-creases, the sensitivity of human monocytes to dexamethasone:potential regulation of the glucocorticoid receptor. J. Clin.Endocrinol. Metab. 84, 2834–2839

14. McDermott, M. F., Aksentijevich, I., Galon, J., McDermott,E. M., Ogunkolade, B. W., Centola, M., Mansfield, E., Gadina,M., Karenko, L., Pettersson, T., McCarthy, J., Frucht, D. M.,Aringer, M., Torosyan, Y., Teppo, A. M., Wilson, M., Karaarslan,H. M., Wan, Y., Todd, I., Wood, G., Schlimgen, R., Kumara-jeewa, T. R., Cooper, S. M., Vella, J. P., Kastner, D. L., et al.(1999) Germline mutations in the extracellular domains of the55 kDa TNF receptor, TNFR1, define a family of dominantlyinherited autoinflammatory syndromes. Cell 97, 133–144

15. Jahnsen, F. L., Haye, R., Gran, E., Brandtzaeg, P., and Johansen,F. E. (1999) Glucocorticosteroids inhibit mRNA expression foreotaxin, eotaxin-2, and monocyte-chemotactic protein-4 in hu-man airway inflammation with eosinophilia. J. Immunol. 163,1545–1551

16. Pype, J. L., Dupont, L. J., Menten, P., Van Coillie, E., Opdenak-ker, G., Van Damme, J., Chung, K. F., Demedts, M. G., and

Verleden, G. M. (1999) Expression of monocyte chemotacticprotein (MCP)-1, MCP-2, and MCP-3 by human airway smooth-muscle cells. Modulation by corticosteroids and T-helper 2cytokines. Am. J. Resp. Cell. Mol. Biol. 21, 528–536

17. Re, F., Muzio, M., De Rossi, M., Polentarutti, N., Giri, J. G.,Mantovani, A., and Colotta, F. (1994) The type II ‘receptor’ asa decoy target for interleukin 1 in polymorphonuclear leuko-cytes: characterization of induction by dexamethasone andligand binding properties of the released decoy receptor. J. Exp.Med. 179, 739–743

18. Arzt, E., Sauer, J., Pollmacher, T., Labeur, M., Holsboer, F.,Reul, J. M., and Stalla, G. K. (1994) Glucocorticoids suppressinterleukin-1 receptor antagonist synthesis following inductionby endotoxin. Endocrinology 134, 672–677

19. Carroll, M. C. (1998) The role of complement and complementreceptors in induction and regulation of immunity. Annu. Rev.Immunol.16, 545–568

20. Opdenakker, G., and Van Damme, J. (1992) Cytokines andproteases in invasive processes: molecular similarities betweeninflammation and cancer. Cytokine 4, 251–258

21. Kuno, K., Kanada, N., Nakashima, E., Fujiki, F., Ichimura, F.,and Matsushima, K. (1997) Molecular cloning of a gene encod-ing a new type of metalloproteinase-disintegrin family proteinwith thrombospondin motifs as an inflammation associatedgene. J. Biol. Chem. 272, 556–562

22. Chen, H., Herndon, M. E., and Lawler, J. (2000) The cellbiology of thrombospondin-1. Matrix Biol. 19, 597–614

23. Hogger, P., Dreier, J., Droste, A., Buck, F., and Sorg, C. (1998)Identification of the integral membrane protein RM3/1 onhuman monocytes as a glucocorticoid-inducible member of thescavenger receptor cysteine-rich family (CD163). J. Immunol.161, 1883–1890

24. Buechler, C., Ritter, M., Orso, E., Langmann, T., Klucken, J.,and Schmitz, G. (2000) Regulation of scavenger receptor CD163expression in human monocytes and macrophages by pro- andantiinflammatory stimuli. J. Leukoc. Biol. 67, 97–103

25. Aderem, A., and Ulevitch, R. J. (2000) Toll-like receptors in theinduction of the innate immune response. Nature (London) 406,782–787

26. Krutzik, S. R., Sieling, P. A., and Modlin, R. L. (2001) The roleof Toll-like receptors in host defense against microbial infec-tion. Curr. Opin. Immunol. 13, 104–108

27. Mellor, A. L., and Munn, D. H. (1999) Tryptophan catabolismand T-cell tolerance: immunosuppression by starvation? Immu-nol. Today 20, 469–473

28. Mellor, A. L., Sivakumar, J., Chandler, P., Smith, K., Molina, H.,Mao, D., and Munn, D. H. (2001) Prevention of T cell-drivencomplement activation and inflammation by tryptophan catab-olism during pregnancy. Nat. Immunol. 2, 64–68

29. Franchimont, D., Galon, J., Gadina, M., Visconti, R., Zhou, Y.,Aringer, M., Frucht, D. M., Chrousos, G. P., and O’Shea, J. J.(2000) Inhibition of Th1 immune response by glucocorticoids:dexamethasone selectively inhibits IL-12-induced Stat4 phos-phorylation in T lymphocytes. J. Immunol. 164, 1768–1774

30. Blotta, M. H., DeKruyff, R. H., and Umetsu, D. T. (1997)Corticosteroids inhibit IL-12 production in human monocytesand enhance their capacity to induce IL-4 synthesis in CD4�lymphocytes. J. Immunol. 158, 5589–5595

31. Wu, C. Y., Sarfati, M., Heusser, C., Fournier, S., Rubio-Trujillo,M., Peleman, R., and Delespesse, G. (1991) Glucocorticoidsincrease the synthesis of immunoglobulin E by interleukin4-stimulated human lymphocytes. J. Clin. Invest. 87, 870–877

32. Dozmorov, I. M., and Miller, R. A. (1998) Generation ofantigen-specific Th2 cells from unprimed mice in vitro: effectsof dexamethasone and anti-IL-10 antibody. J. Immunol. 160,2700–2705

33. Lane, S. J., Adcock, I. M., Richards, D., Hawrylowicz, C., Barnes,P. J., and Lee, T. H. (1998) Corticosteroid-resistant bronchialasthma is associated with increased c-fos expression in mono-cytes and T lymphocytes. J. Clin. Invest. 102, 2156–2164

34. McDevitt, H. O. (2000) Discovering the role of the majorhistocompatibility complex in the immune response. Annu. Rev.Immunol.18, 1–17

35. Alfonso, C., and Karlsson, L. (2000) Nonclassical MHC class IImolecules. Annu. Rev. Immunol.18, 113–142

36. Shull, M. M., Ormsby, I., Kier, A. B., Pawlowski, S., Diebold, R. J.,Yin, M., Allen, R., Sidman, C., Proetzel, G., Calvin, D., et al.


(1992) Targeted disruption of the mouse transforming growthfactor-beta 1 gene results in multifocal inflammatory disease.Nature (London) 359, 693–699

37. Kuhn, R., Lohler, J., Rennick, D., Rajewsky, K., and Muller, W.(1993) Interleukin-10-deficient mice develop chronic enteroco-litis. Cell 75, 263–274

38. McKenna, N. J., Lanz, R. B., and O’Malley, B. W. (1999) Nuclearreceptor coregulators: cellular and molecular biology. Endocr.Rev. 20, 321–344

39. Leung, D. Y., Martin, R. J., Szefler, S. J., Sher, E. R., Ying, S., Kay,A. B., and Hamid, Q. (1995) Dysregulation of interleukin 4,interleukin 5, and interferon gamma gene expression in steroid-resistant asthma. J. Exp. Med. 181, 33–40

40. Bamberger, C. M., Schulte, H. M., and Chrousos, G. P. (1996)Molecular determinants of glucocorticoid receptor function

and tissue sensitivity to glucocorticoids. Endocr. Rev. 17,245–261

41. McGowan, J. E., Chesney, P. J., Crossley, K. B., and LaForce,F. M. (1992) Guidelines for the use of systemic glucocorticoste-roids in the management of selected infections. Working Groupon Steroid Use. Antimicrobial Agents Committee, InfectiousDiseases Society of America. J. Infect. Dis. 165, 1–13

42. Odio, C. M., Faingezicht, I., Paris, M., Nassar, M., Baltodano, A.,Rogers, J., Saez-Llorens, X., Olsen, K. D., and McCracken, G. H.(1991) The beneficial effects of early dexamethasone adminis-tration in infants and children with bacterial meningitis. N. Engl.J. Med. 324, 1525–1531

Received for publication May 1, 2001.Revised for publication August 15, 2001.


Documents

45. Pustaka Galon Dkk