ORIGINAL ARTICLES

Molecular profiling and gene expression analysis incutaneous sarcoidosis: The role of interleukin-12,

interleukin-23, and the T-helper 17 pathway

Marc A. Judson, MD,a Richard M. Marchell, MD,a MaryAnn Mascelli, PhD,b Alexa Piantone, PharmD,b

Elliot S. Barnathan, MD,b Kevin J. Petty, MD,b Dion Chen, PhD,b Hongtao Fan, PhD,b Heidi Grund, RN,a

Keying Ma, PhD,b Fr�ed�eric Baribaud, PhD,b and Carrie Brodmerkel, PhDb

Charleston, South Carolina, and Malvern, Pennsylvania

From

Re

Ph

Dr Ju

C

ic

Fund

Jo

LL

Discl

Ba

sa

su

Ba

Jo

ha

Background: Cutaneous sarcoidosis (CS) skin provides relatively noninvasive access to granulomatoussarcoidosis tissue.

Objective: We sought to explore the role of the T-helper (Th)1 and Th17 pathways in sarcoidosis.

Methods: We used molecular profiling and gene expression analysis to analyze the Th1 and Th17pathways and other immune-mediated pathways in CS. Molecular profiles were obtained from sarcoidosisskin lesions (lesional skin [LS]), unaffected skin from patients with CS (non-LS), and the skin of healthycontrol subjects. Whole blood was collected to compare the molecular profile of sarcoidosis skin lesionsand whole blood.

Results: Twenty participants were enrolled: 15 with active CS and 5 healthy volunteers. Microarrayanalyses comparing non-LS and healthy volunteer skin with LS showed several thousand genes differ-entially expressed ($ 2-fold change false discovery rate, P\ .01). Targeted selections of genes associatedwith Th1 and Th17 phenotypes showed a strong Th1 profile of sarcoidosis and expression of interleukin(IL)-23 and IL-23R with limited expression of other Th17 pathway genes. IL-21 and signal transducer andactivator of transcription 3 (STAT3) were also dysregulated in skin and whole blood, providing additionalevidence for involvement of the IL-12 pathway and potential activation of the Th17 pathway.

Limitations:Measurements were made at a single point in time and may not identify mechanisms that maybe identified in patients followed up longitudinally.

Conclusion: These findings provide novel insight into the dysregulated pathways that may be involved inthe pathogenesis of sarcoidosis. ( J Am Acad Dermatol 2012;66:901-10.)

Key words: interferon-gamma; interleukin-12; interleukin-21; interleukin-23; lesional sarcoidosis; nonle-sional sarcoidosis.

the Medical University of South Carolinaa and Centocor

search and Development, a Division of Johnson & Johnson

armaceutical Research & Development, LLC, Malvern.b

dson is currently affiliated with Division of Pulmonary and

ritical Care Medicine, AMC Pulmonary and Critical Care Med-

ine, Albany, New York.

ed by Centocor Research and Development, a Division of

hnson & Johnson Pharmaceutical Research & Development,

C.

osure: Dr Mascelli is a salaried consultant and Drs Piantone,

rnathan, Petty, Chen, Fan, Ma, Baribaud, and Brodmerkel are

laried employees of Centocor Research and Development, a

bsidiary of Johnson & Johnson. Drs Barnathan, Petty,

ribaud, and Brodmerkel own stock and stock options in

hnson & Johnson. Drs Judson and Marchell, and Ms Grund

ve no conflicts of interest to declare.

Data from this study were presented in part at the American

Thoracic Society Annual Meeting, May 2009, San Diego, California,

and the Federation of Clinical Immunology Societies Annual

Meeting, June 2009, San Francisco, California.

Accepted for publication June 1, 2011.

Reprint requests: Marc A. Judson, MD, Division of Pulmonary and

Critical Care Medicine, AMC Pulmonary and Critical Care

Medicine, 47 New Scotland Ave, MC 91, Physicians Pavilion,

Fourth Floor, Albany, NY 12208. E-mail: judsonm@mail.amc.

edu.

Published online September 19, 2011.

0190-9622/$36.00

� 2011 by the American Academy of Dermatology, Inc.

doi:10.1016/j.jaad.2011.06.017

901

J AM ACAD DERMATOL

JUNE 2012902 Judson et al

Sarcoidosis is a multisystem granulomatous diseaseof unknown cause. As with known granulomatousdiseases, sarcoidosis is thought to result from contactof an external antigen with antigen-presentingcells. The antigen-presenting cells process andpresent the antigen via HLA class II molecules toT-cell receptors attached to T lymphocytes.1 Once

CAPSULE SUMMARY

d Genes involved in the T-helper 1 andT-helper 17 pathways were up-regulatedin the skin lesions of patients withsarcoidosis.

d These genes were differentially up-regulated in sarcoidosis skin lesionscompared with nonlesional skin of thesepatients and the skin of healthy controlsubjects.

d Gene profiling in skin and bloodsuggests there is a potential fordetermining whether therapy thatblocks particular pathways will be usefulin assessing the sarcoidosis treatmentresponse.

this interaction occurs, theantigen-presenting cell andT lymphocytes secrete vari-ous cytokines of the T-helper(Th)1 class, resulting in theformation of granulomas.Cytokines such as tumor ne-crosis factor (TNF)-alpha andinterleukin (IL)-12 are impor-tant Th1 cytokines that arereleased during the formationof sarcoid granulomas.2-7

TNF antagonists have dem-onstrated some success in thetreatment of sarcoidosis, in-cluding cutaneous forms.8

Although IL-12 has been re-ported to be up-regulated inpulmonary sarcoidosis, itsrole in the development ofskin sarcoidosis has not been

elucidated.

Evaluating the role of IL-12 has been complicatedby the identification of IL-23, a heterodimeric cyto-kine comprising the common p40 subunit of IL-12but binding a specific p19 subunit. IL-23 is a keycytokine in the Th17 pathway.9 Although IL-12 hasbeen linked to sarcoidosis, the role of IL-23 or theTh17 pathway, including IL-17, has not been exam-ined in this disease.

Much of the information concerning cytokinenetworks involved in sarcoidosis is derived fromsamples obtained from the lung or periphery (ie,serum, plasma, peripheral blood).10 Lung sampleshave predominantly been obtained by bronchoal-veolar lavage. Because pulmonary sarcoidosis is apatchy disease, such samples contain material frominvolved and uninvolved lung segments.11 In addi-tion, the assessment of pulmonary activity in sar-coidosis is problematic.12 Biopsy specimen ofsarcoidosis skin lesions provides a sample withactive granulomatous inflammation. In addition,the activity of sarcoidosis skin lesions is readilyassessable. Furthermore, biopsy specimen of unaf-fected skin in a patient with cutaneous sarcoidosis(CS) allows each patient to serve as his/her owncontrol. These facts suggest that analysis of granu-lomas from sarcoidosis skin lesions may reflect

active mechanisms of disease more accurately thanfrom sarcoidosis lung lesions. To this end, weassessed the expression of IL-12, IL-23, and theTh17 pathway in sarcoidosis skin lesions by com-paring their expression in lesions versus areas ofuninvolved skin of patients with sarcoidosis, theskin of healthy volunteers, and in whole blood from

these patients.

METHODSThis study was approved

by the Medical University ofSouth Carolina InstitutionalReview Board. Patients aged18 to 80 years were eligiblefor this study if they hadbiopsy-confirmed sarcoido-sis; visual evidence of faciallupus pernio skin lesions oractive sarcoidosis skin le-sions over the remainder ofthe body; and a primary skinlesion for assessment (targetlesion) that could be photo-graphed. Control subjectswere healthy volunteersaged 18 to 80 years. Patientswith sarcoidosis and healthy

volunteers were excluded if they were pregnant; hada history of keloid formation or psoriasis; had activetuberculosis or nonsarcoidosis active granulomatousdisease; or had an underlying medical condition that,in the opinion of the investigators, would place themat undue risk or preclude the completion of follow-up measurements. Organ involvement in patientswith sarcoidosis was determined using an estab-lished protocol.13

All patients with sarcoidosis had the skin lesionsthat were biopsied and assessed using a standard-ized visual evaluation system (the SarcoidosisActivity and Severity Index scoring system).14 Inpatients given the diagnosis of lupus pernio, a skinbiopsy (two adjacent 3 mm) was performed on afacial lesion. Patients with sarcoidosis skin lesionson areas other than the face had a skin biopsy (twoadjacent 3 mm) performed on lesional skin (LS) andtwo additional adjacent biopsies performed in anarea of non-LS (NLS). Whenever possible, the NLSbiopsy was performed at a site contralateral to thatof the LS biopsy. Biopsies for healthy volunteer skin(two adjacent 3 mm) were performed in the axillato minimize the cosmetic impact. Absorbable su-tures were used for all biopsies. A total of 12 mL ofblood was phlebotomized from each participant.Participants were contacted by telephone 7 to 10

Abbreviations used:

cDNA: complementary DNACS: cutaneous sarcoidosisCXCL: chemokine (C-X-C motif) ligandFDR: false discovery rateIFN: interferonIL: interleukinLS: lesional skinNLS: nonlesional skinPCR: polymerase chain reactionRRT: real-time reverse transcriptionSTAT: signal transducer and activator of

transcriptionTGF: transforming growth factorTh: T-helperTNF: tumor necrosis factor

J AM ACAD DERMATOL

VOLUME 66, NUMBER 6Judson et al 903

days after biopsy to ensure that there were nobiopsy-related complications. Participants were in-structed to return to the clinical investigator if anycomplications arose.

Histologic analysesSkin biopsy specimens were frozen in optical

coherence tomography and stored at e808C beforeprocessing. Samples were cut and stained withhematoxylin and eosin using standard stainingprocedures.

RNA extraction from skin and whole bloodSkin biopsy specimens were collected and frozen.

RNA was extracted with RNeasy fibrous tissue minikit according to the manufacturer’s instructions(Qiagen Inc, Valencia, CA). RNA of sufficient qualityfor microarray analyses was obtained from 14 LS, 10NLS, and 5 healthy volunteer skin biopsy samples.Blood samples were collected and frozen inPAXgene blood RNA tubes (BD, Franklin Lakes,NJ). RNA was extracted using the PAXgene bloodRNA kit according to manufacturer’s instructions(Qiagen Inc). RNA from whole blood of sufficientquality for microarray analyses was obtained from 15CS and 5 healthy volunteer samples.

Microarray and microarray data analysesComplementary DNA (cDNA) was generated us-

ing the WT-Ovation Pico RNA amplification system(NuGEN, San Carlos, CA). cDNAwas normalized to aconcentration of 140 ng/�L, then fragmented andlabeled for hybridization using the FL-Ovation cDNAbiotin module (NuGEN). A total of 1.8 �g of labeledcDNA was added to hybridization buffer and 210 �Lwas hybridized to U133 (Affymetrix, Santa Clara,CA).

Plus2.0 microarrays (Affymetrix) were generatedfollowing manufacturer’s protocols. Microarrays

were hybridized for 16 hours, washed and stainedusing a FS450 wash station (Affymetrix), andscanned with a GCS-3000 scanner with autoloader(Affymetrix).

All data were quantile-transformed before dataanalysis. A t test was performed to compare NLS withLS, healthy volunteer skin with NLS, healthy volun-teer skin with LS, and healthy volunteer whole bloodwith whole blood from patients with CS. Falsediscovery rate (FDR) was used to adjust for multiplecomparisons. Significance was assigned for genesgreater than 2-fold differentially expressed in thecomparator group with a t test P value of less than .05and an FDR of less than .01. These data have beenplaced in Gene Expression Omnibus and can befound at http://www.ncbi.nlm.nih.gov/geo/.

Real-time reverse transcription andpolymerase chain reaction

The primers and probes used were based on themanufacturer’s manual (Applied Biosystems, FosterCity, CA). The expression of each gene was quanti-fied using 20 ng of cDNA. The RRT-PCR reactioncontained cDNA, 1X primers and probe mix, and 1XABI TaqMan universal polymerase chain reaction(PCR) master mix without ampErase UNG (AppliedBiosystems). Forty cycles (958C for 15 seconds and608C for 1 minute) were performed on an ABI7900HT unit (Applied Biosystems). Data were nor-malized to glyceraldehyde phosphate dehydrogen-ase expression.

The RRT-PCR from healthy volunteer skin, NLS,and LS were compared using the medians for eachgene, and the significance was assessed using theanalysis of variance Kruskal-Wallis test. Whole bloodRRT-PCR data comparing median healthy volunteerwith CS gene expression data were tested for signif-icance using the Mann-Whitney U test. For both tests,a P value of less than .05 was considered significant.

Pathway analysesNetwork classification and functional analyses

were generated using Ingenuity pathways analysis(Ingenuity Systems, Redwood City, CA). The func-tional analysis identified the biological functions thatwere most significant to the data set. Genes from thedata set that met the cutoff of greater than or equal to2-fold change with an FDR less than .01 and wereassociated with biological functions in the IngenuitySystems pathways knowledge base were consideredfor the analysis.

Serum protein analysesAn 89-protein multiplex panel (Human Map 1.6,

Rules Based Medicine, Austin, TX) was used to

Table I. Demographics and organ involvement

Healthy

volunteers

(n = 5)

Patients with

sarcoidosis

(n = 15)

Age, yMean 6 SD 36.6 6 9.9 42.9 6 9.2Median (range) 34 (27-53) 47 (27-54)

Race, n (%)White 3 (60) 3 (20)Black 2 (40) 12 (80)

Gender, n (%)Female 0 (0) 11 (73)Male 5 (100) 4 (27)

Lungs,* n (%)None 3 (20)Definite 12 (80)

Skin,* n (%)None 0 (0)Probable 0 (0)Definite 15 (100)

Any extracutaneous organinvolvement,* n (%)

Definite 14 (93)

*Organ involvement as described in Judson et al.13

Table II. Biopsy site and target lesion assessment

Healthy

volunteers

(n = 5)

Patients with

sarcoidosis

(n = 15)

Lesional skin site characteristicsErythema score, n (%)Slight 2 (13)Moderate 11 (73)Severe 2 (13)Very severe 0 (0)

Induration score, n (%)Slight 1 (7)Moderate 4 (27)Severe 3 (20)Very severe 7 (47)

Area of involvement (cm2)Mean 6 SD 3.8 6 5.7Median (range) 1.8 (0.2-19.0)

Sarcoidosis Activity and Severity Index scoring system was used

for skin lesions.14

Table III. Significantly dysregulated genes inlesional skin

Gene Fold difference NLS vs LS P value

IL-12p40 125.9 # .01IL-12RB1 12.3 # .01IL-12RB2 3.1 # .01IL-23R 5.7 # .05IFN-g 7.0 # .01STAT1 3.2 # .01STAT4 2.4 # .05IL-21 170.6 # .01TNF-alpha 2.8 # .01IL-13Ra2 0.2 # .01

Real-time reverse transcription-polymerase chain reaction data

were compared using median expression of each gene from

normal-appearing skin (healthy volunteer skin) and NLS versus

median expression of each gene in LS. P value was derived using

analysis of variance Kruskal-Wallis test.

IFN, Interferon; IL, interleukin; LS, lesional skin; NLS, nonlesional

skin; STAT, signal transducer and activator of transcription; TNF,

tumor necrosis factor.

J AM ACAD DERMATOL

JUNE 2012904 Judson et al

profile differential serum protein expression fromhealthy volunteers (n = 5) and patients with sarcoid-osis (n = 15). The complete list of analytes in theHuman Map 1.6 can be found at http://www.rulesbasedmedicine.com/products-services/humanmap-services/humanmap/. The data were log-transformed before analyses to achieve an approx-imately normal distribution. For IL-17 (MesoscaleDiscovery, Gaithersburg, MD), IL-21 (Biovendor,Candler, NC), and IL-23 (eBiosciences, San Diego,CA), single analyte assays were used according tomanufacturer’s instructions. The lower limit of de-tection of these assays was 1.22, 78, and 78 pg/mL,respectively. Student t test with multiple testingadjustment (FDR \ .05) was performed to identifymarkers significantly different (P \ .05) betweenhealthy volunteers and patients with CS.

RESULTSParticipant demographics

Fifteen patients with CS were enrolled. Twopatients had lupus pernio, and the remaining 13patients had active sarcoidosis skin lesions in non-facial areas. Fourteen of the 15 patients had morethan one organ involved, of which 11 patients had3 or more organs involved. All patients with CShad confirmatory biopsy specimens and thereforewere classified as definite cases of CS.13 Fivehealthy volunteers of similar age and ethnicityto the patients with sarcoidosis were enrolled(Table I).

Two patients with CS were not receivingsarcoidosis-related medication, one was receivinghydroxychloroquine, and the remaining 12 werereceiving prednisone: 6 on monotherapy and 6 oncombination therapy (including 4 patients takingprednisone plus methotrexate, one taking predni-sone with hydroxychloroquine, and one takingprednisone with clobetasol). The characteristics ofthe skin lesion biopsy specimens of the patientswith CS are described in Table II.

Fig 1. Real-time reverse transcription-polymerase chain reaction of interleukin (IL)-12p40,tumor necrosis factor (TNF )-alpha, interferon (IFN )-g, IL-23a, and IL-21 in healthy volunteerskin (HVS ), nonlesional skin (NLS ), and lesional skin (LS ). Each symbol represents oneparticipant. The solid line represents the median. Statistical comparisons were made betweenthe median for each group using analysis of variance Kruskal-Wallis test. Significant compar-isons (P\ .05) are indicated. GAPDH, Glyceraldehyde phosphate dehydrogenase.

J AM ACAD DERMATOL

VOLUME 66, NUMBER 6Judson et al 905

IL-12, IL-23, and Th17 pathway genes in CSskin versus NLS

Seventeen genes, including IL-12, IL-23, TNF-alpha, and Th17 pathway genes, were analyzed byRRT-PCR. Of the 17 selected genes, 10 were signif-icantly dysregulated in LS (Table III). IL-12p40 andthe Th17 cytokine, IL-21, were more than 100-foldincreased in LS compared with NLS. IL-23p19,IL-12p35, signal transducer and activator of tran-scription (STAT) 3, IL-17A, IL-17F, IL-20, IL-22, andIL-27 were not significantly dysregulated; although

IL-23p19 demonstrated a trend toward significancewith 8 of 12 LS samples expressing the gene.Individual expression levels for selected genes areshown in Fig 1.

IL-12, IL-23, and Th17 pathway genes in wholeblood: patients with sarcoidosis versus healthyvolunteers

Of the 17 genes assessed in skin by RRT-PCR,STAT1, STAT3, IL-12Rb1, and TNF-alpha were signif-icantly differentially expressed in the blood of

Fig 2. Real-time reverse transcription-polymerase chain reaction of whole blood messengerRNA expression of signal transducer and activator of transcription (STAT) 1, STAT3, interleukin(IL)-12RB1, and tumor necrosis factor (TNF )-alpha comparing healthy volunteer (HV) withcutaneous sarcoidosis (CS ). Each symbol represents one participant. The solid line representsthe median. Statistical comparisons were made between the median for each group usingMann-Whitney U test. Significant comparisons (P\.05) are indicated.GAPDH, Glyceraldehydephosphate dehydrogenase.

J AM ACAD DERMATOL

JUNE 2012906 Judson et al

patients with sarcoidosis (Fig 2). STAT3 was the onlygene significantly differentially expressed in bloodbut not in skin.

Transcriptomic analysis of skin biopsyspecimens

The number of genes significantly differentiallyexpressed was 334 between healthy volunteer skinand NLS, 2543 between healthy volunteer skin andLS, and 2059 between NLS and LS (Fig 3, A). Fig 3, B,shows the relative intensity of the genes signifi-cantly differentially expressed by significant probeset for all skin biopsy specimens, including a clearseparation of healthy volunteer skin and NLS fromLS. Two thirds of the dysregulated transcripts wereoverexpressed in LS. By magnitude, the most highlyup-regulated known genes were products of acti-vated macrophages and inflammatory-related pro-teins (Table IV). The most down-regulated knowngenes belonged to families of reductases and dehy-drogenases (Table IV). Pathway analyses indicateinterferon (IFN) signaling as a significantly dysregu-lated canonical pathway with genes such as janus-

activated kinase 2, interferon-regulatory factor-1,and MX1 being highly up-regulated along withSTAT1, which was confirmed by RRT-PCR.Assessment of Th17 pathway genes dysregulatedin the microarray analysis include transforminggrowth factor (TGF) b, which was significantly up-regulated, as were IL-17R, chemokine (C-X-C motif)ligand (CXCL) 1, CXCL10, tissue inhibitor of metal-loproteinases 1, myeloid cell nuclear differentiationantigen, and IL-26.

Whole blood and serum profiling ofsarcoidosis

The number of transcripts differentially expressedin blood of patients with CS compared with healthyvolunteers was 411. Approximately one third of the411 transcripts were up-regulated. The most signif-icant canonical pathway affected in the blood profileas identified by pathway analysis was IFN-g signal-ing, which was also one of the top canonical path-ways in LS, suggesting a whole-blood correlation tothe disease site. A key regulator of Th17 pathway

Fig 3. A, Venn diagram of biopsy specimen microarray results showing number of genesdifferentially expressed between healthy volunteer skin (HVS ) and nonlesional skin (NLS )(red ), HVS and lesional skin (LS ) (blue), and NLS and LS ( green). B, Heat map of differentialgene expression in LS (blue) and NLS ( purple) from patients with cutaneous sarcoidosis andHVS ( green). All significantly dysregulated transcripts have fold change cut-off of 2-fold andfalse discovery rate of P\ .01.

J AM ACAD DERMATOL

VOLUME 66, NUMBER 6Judson et al 907

genes, STAT3, was also significantly up-regulated inwhole blood.

Of the 90 analytes tested by multianalyte profilingand the 3 single analyte assays (IL-17, IL-21, andIL-23), none were significantly elevated at P less than.05 with an FDR adjustment for multiple compari-sons. However, 17 markers were significantly ele-vated based on P less than .05 using a Student t testwith no multiple testing adjustment (Table V; avail-able at http://www.jaad.org). IL-21 was detected inthe serum of 3 of 15 patients with CS (data notshown).

DISCUSSIONOur results detail the novel evaluation of IL-23

and Th17-related genes and transcriptomic profilingin CS. We found that IL-12p40 and the IFN-gpathway were highly up-regulated in the skin ofpatients with CS. IL-23 was also up-regulated in twothirds of patients, with no significant expression ofIL-17 but with highly significant expression of IL-21,

a Th17 pathway gene. To our knowledge, this is thefirst whole genome molecular profiling study com-pleted in CS. The tissue level transcriptomic datademonstrate that the molecular profile of CS issimilar to that reported for pulmonary sarcoido-sis.15,16 As such, although the initiating factor of thedisease resulting in involvement of a given organmay differ, there is commonality in the biologicalpathways that are dysregulated in systemic sarcoid-osis. A molecular disease profile was also measur-able in whole blood, with the IFN-g pathwaysignificantly dysregulated, which independentlyconfirms recent data on the molecular profile ofsarcoidosis in whole blood.16 In addition, we de-scribe significant up-regulation of STAT3, which hasbeen implicated in the generation of the Th17 celllineage.17

IL-23 and TGFb play a role in the proliferation anddifferentiation of Th17 cells, which produce IL-17A,IL-17F, IL-21, and IL-22. Here we describe theexpression of IL-23, TGFb, IL-17R, and IL-21 but

Table IV. Twenty most highly up- or down-regulated known genes in lesional skin versushealthy volunteer skin

Gene Function

Fold

change

Up-regulatedATP6VOD2 ATPase 50.1PLA2G2D Phospholipase 45.9CHIT1 Chitinase 43.7TNIP3 Cytokine regulation 34.5DCSTAMP Dendritic cell transmembrane 34.5C2 Complement cascade 32.6MMP12 Extracellular matrix degradation 30.2LYZ Lysozyme 25.3CLECSF9 Sugar binding 24.5NKIR Inhibitory receptor 23.9

Down-regulatedSLC26A3 Transporter 0.003UGT2B28 Glucuronosyltransferase 0.003DHRS2 Dehydrogenase/reductase 0.004AKR1D1 Aldo-keto reductase 0.007BUCS1 Catalysis 0.014HMGCS2 HMGCoA synthase 0.015TAT Tyrosine transaminase 0.019SAH Fatty acid ligase 0.038HIST1H2Bg Histone cluster gene 0.041MUC7 Protein binding 0.042

ATP6VOD2, ATPase H+ transporting lysosomal 38kDa VO subunit

d2; PLA2G2D, phospholipase A2; CHIT1, chitinase 1; TNIP3, tumor

necrosis factor-alpha induced protein interacting protein 3;

DCSTAMP, dendritic cell specific transmembrane protein; C2,

complement component 2; MMP12, matrix metalloproteinase 12;

LYZ, lysozyme; CLECSF9, C type lectin domain family 4, member E;

NKIR, NK inhibitory receptor precursor; SLC26A3, solute carrier

family 26, member 3; UGT2B28, UDP glucuronosyltransferase 2

family, polypeptide B28; DHRS2, dehydrogenase reductase,

member 2; AKR1D1, aldo-keto reductase family 1, member D1;

BUCS1, butyryl coenzyme A synthetase 1; HMGCS2, 3-hydroxy-

3-methylglutaryl-Coenzyme A synthase 2; TAT, tyrosine

aminotransferase; SAH, SA hypertension-associated homologue;

HIST1H2Bg, histone 1; MUC7, mucin 7.

J AM ACAD DERMATOL

JUNE 2012908 Judson et al

did not detect expression of IL-17A, IL-17F, or IL-22.Although IL-17 is a key member of the Th17 path-way, there are data to suggest that IL-2118 and IL-2319

can act independently of IL-17, which could explainthe lack of IL-17 expression with the presence ofother Th17 pathway molecules. Interestingly, it hasbeen reported that sarcoidosis is characterized by thepresence of immature dendritic cells in the lung andskin,20 which may in part be mediated by overpro-duction of IL-21.

Whole blood molecular profiling has been usedto develop disease signatures that may be useful inpredicting the response to a therapy.21 Recently,two studies of whole genome molecular profiling ofsarcoidosis lung biopsy specimens compared with

normal lung tissue demonstrated that the IFN-gpathway was significantly dysregulated.15,16

Interestingly, many of the genes these authorsfound up-regulated in lung sarcoidosis showedsignificant up-regulation in the LS of our patients,including STAT1, a disintegrin and metalloprotei-nase domain-like protein decysin 1, IL-15, matrixmetalloproteinase 12, interferon-regulatory factor 1,and CXCL9. Classification of the dysregulated genesinto functional networks was also comparable andrevealed that significant biologic functions dysregu-lated in LS compared with healthy volunteer skin orNLS were related to the cellular and humoralimmune response; cell growth and proliferation,particularly Th cells; and immunologic and inflam-matory disease-related genes. These data supportthe hypothesis that mechanisms of granuloma for-mation in sarcoidosis are similar in different affectedtissues.

We also assessed whole blood and serum todetermine whether these samples could be used asa surrogate for inflammatory events occurring in thetarget organ, thereby avoiding more invasive tissuesampling. We confirmed previous data16 that dysre-gulation of the IFN-g pathway in sarcoidosis targetorgans can be measured by whole blood geneexpression in these patients. In addition, we reportthat STAT3 is significantly expressed in wholeblood. STAT3 has been linked to the generation ofTh17 cells and regulation of IL-21,22 and its over-expression peripherally in sarcoidosis indicates ac-tivity of the Th17 pathway. In addition, there wasdysregulation of cell death genes in the peripherythat were not observed in the skin, such as caspases1 and 4 along with FAS. The activation of theseapoptotic genes may result from an attempt tocontrol the general activation of immune responsepathways.

In general, we did not find a consistent pattern ofup-regulation of proteins in the serum of patientswith CS, including the IL-12p40 and Th17 pathway-related proteins. The lack of expression of IL-17 inserum is consistent with the lack of expression in theskin and whole blood RNA. That we did not findlevels of IL-23, a cytokine that acts locally, in serum isnot an unexpected finding as systemic levels of thiscytokine have not been reliably reported.

This study has potential limitations. First, mea-surements were made at a single point in time at thelevel of gene expression in the tissue, which may notallow for identification of dysregulated pathwaysthat would be better elucidated at the protein leveland/or via longitudinal profiling of patients. As acorollary, confirmation that a pathway is importantin the development of a disease requires therapy that

J AM ACAD DERMATOL

VOLUME 66, NUMBER 6Judson et al 909

blocks this pathway to be effective. The use ofmedications by the majority of the patients withsarcoidosis in this study may have influenced theconclusions. Because the number of patients receiv-ing and not receiving antisarcoidosis medications atthe time of this study was small, we found noobvious effect related to antisarcoidosis therapy.Finally, the fact that only a small number of healthyvolunteers participated as control subjects and thatthey were not well matched in terms of gender orrace with the patients with sarcoidosis could havebiased some of our interpretations. However, therewere limited associations of genes with gender, age,or race in the analysis, and each patient withsarcoidosis served as his or her own control in thattheir NLS was used for comparison. The overallpatterns demonstrated in LS, NLS, and skin fromhealthy volunteers suggest that the inflammatorypathways demonstrated in nonlesional sarcoidosisskin represented amiddle ground betweenwhat wasfound in lesional sarcoidosis skin and healthy vol-unteer skin. This suggests that our healthy volunteerswere an adequate negative control group. Follow-upstudies could be conducted to understand the con-tribution of the various cell types found in LS to thegene expression disease profile.

In summary, our data confirm that IL-12 and theIFN pathway are up-regulated in CS and demonstratenovel findings on the up-regulation of IL-23 andIL-21 in CS. The differential expression of the Th17pathway genes IL-23 and IL-21 but not IL-17 that wedemonstrated differentiates sarcoidosis from severalother immune-mediated diseases such as psoriasisand Crohn’s disease, where up-regulation of IL-17and Th17 pathway genes have been reported.23,24 Ananti-IL-12p40 molecule is currently being tested forthe treatment of chronic sarcoidosis in a proof-of-concept, phase II clinical trial (http://clinicaltrials.gov/ct2/show/NCT00955279), which assesses thecontribution of these pathways to the pathogenesisof sarcoidosis, Th1 via inhibition of IL-12, and Th17via inhibition of IL-23.

The authors thank Jennifer Han and Robert Achenbachof Janssen Services, LLC, for assistance with editing andpreparing the manuscript for publication.

REFERENCES

1. Newman LS, Rose CS, Maier LA. Sarcoidosis. N Engl J Med

1997;336:1224-34.

2. Moller DR, Forman JD, Liu MC, Noble PW, Greenlee BM, Vyas P,

et al. Enhanced expression of IL-12 associated with Th1

cytokine profiles in active pulmonary sarcoidosis. J Immunol

1996;156:4952-60.

3. Shigehara K, Shijubo N, Ohmichi M, Kamiguchi K, Takahashi R,

Morita-Ichimura S, et al. Increased circulating interleukin-12

(IL-12) p40 in pulmonary sarcoidosis. Clin Exp Immunol 2003;

132:152-7.

4. Shigehara K, Shijubo N, Ohmichi M, Kon S, Shibuya Y,

Takahashi R, et al. Enhanced mRNA expression of Th1 cyto-

kines and IL-12 in active pulmonary sarcoidosis. Sarcoidosis

Vasc Diffuse Lung Dis 2000;17:151-7.

5. Shigehara K, Shijubo N, Ohmichi M, Takahashi R, Kon S,

Okamura H, et al. IL-12 and IL-18 are increased and stimulate

IFN-gamma production in sarcoid lungs. J Immunol 2001;166:

642-9.

6. Taha RA, Minshall EM, Olivenstein R, Ihaku D, Wallaert B,

Tsicopoulos A, et al. Increased expression of IL-12 receptor

mRNA in active pulmonary tuberculosis and sarcoidosis. Am J

Respir Crit Care Med 1999;160:1119-23.

7. Ziegenhagen MW, Muller-Quernheim J. The cytokine network

in sarcoidosis and its clinical relevance. J Intern Med 2003;253:

18-30.

8. Stagaki E, Mountford WK, Lackland DT, Judson MA. The

treatment of lupus pernio: results of 116 treatment courses

in 54 patients. Chest 2009;135:468-76.

9. Wilson NJ, Boniface K, Chan JR, McKenzie BS, Blumenschein

WM, Mattson JD, et al. Development, cytokine profile and

function of human interleukin 17-producing helper T cells. Nat

Immunol 2007;8:950-7.

10. Bargagli E, Mazzi A, Rottoli P. Markers of inflammation in

sarcoidosis: blood, urine, BAL, sputum, and exhaled gas. Clin

Chest Med 2008;29:445-58, viii.

11. Reynolds HY. Bronchoalveolar lavage and other methods to

define the human respiratory tract milieu in health and

disease. Lung 2011;189:87-99.

12. Baughman RP, Culver DA, Judson MA. A concise review of

pulmonary sarcoidosis. Am J Respir Crit Care Med 2011;183:

573-81.

13. Judson MA, Baughman RP, Teirstein AS, Terrin ML, Yeager H Jr.

Defining organ involvement in sarcoidosis: the ACCESS pro-

posed instrument. ACCESS Research Group. A case control

etiologic study of sarcoidosis. Sarcoidosis Vasc Diffuse Lung

Dis 1999;16:75-86.

14. Baughman RP, Judson MA, Teirstein A, Lower EE, Lo K,

Schlenker-Herceg R, et al. Chronic facial sarcoidosis including

lupus pernio: clinical description and proposed scoring sys-

tems. Am J Clin Dermatol 2008;9:155-61.

15. Crouser ED, Culver DA, Knox KS, Julian MW, Shao G,

Abraham S, et al. Gene expression profiling identifies

MMP-12 and ADAMDEC1 as potential pathogenic mediators

of pulmonary sarcoidosis. Am J Respir Crit Care Med 2009;

179:929-38.

16. Rosenbaum JT, Pasadhika S, Crouser ED, Choi D, Harrington

CA, Lewis JA, et al. Hypothesis: sarcoidosis is a STAT1-medi-

ated disease. Clin Immunol 2009;132:174-83.

17. Ma CS, Chew GY, Simpson N, Priyadarshi A, Wong M,

Grimbacher B, et al. Deficiency of Th17 cells in hyper IgE

syndrome due to mutations in STAT3. J Exp Med 2008;205:

1551-7.

18. Nurieva RI, Chung Y, Hwang D, Yang XO, Kang HS, Ma L, et al.

Generation of T follicular helper cells is mediated by

interleukin-21 but independent of T helper 1, 2, or 17 cell

lineages. Immunity 2008;29:138-49.

19. Volpe E, Touzot M, Servant N, Marloie-Provost MA, Hupe P,

Barillot E, et al. Multiparametric analysis of cytokine-driven

human Th17 differentiation reveals a differential regulation of

IL-17 and IL-22 production. Blood 2009;114:3610-4.

20. Zaba LC, Smith GP, Sanchez M, Prystowsky SD. Dendritic cells

in the pathogenesis of sarcoidosis. Am J Respir Cell Mol Biol

2010;42:32-9.

J AM ACAD DERMATOL

JUNE 2012910 Judson et al

21. Chaussabel D, Quinn C, Shen J, Patel P, Glaser C, Baldwin N,

et al. A modular analysis framework for blood genomics

studies: application to systemic lupus erythematosus. Immu-

nity 2008;29:150-64.

22. Yang XO, Panopoulos AD, Nurieva R, Chang SH, Wang D,

Watowich SS, et al. STAT3 regulates cytokine-mediated gen-

eration of inflammatory helper T cells. J Biol Chem 2007;282:

9358-63.

23. Guttman-Yassky E, Lowes MA, Fuentes-Duculan J, Zaba LC,

Cardinale I, Nograles KE, et al. Low expression of the

IL-23/Th17 pathway in atopic dermatitis compared to psori-

asis. J Immunol 2008;181:7420-7.

24. Noble CL, Abbas AR, Lees CW, Cornelius J, Toy K, Modrusan Z,

et al. Characterization of intestinal gene expression profiles in

Crohn’s disease by genome-wide microarray analysis. Inflamm

Bowel Dis 2010;16:1717-28.

Table V. Serum analytes from healthy volunteers and patients with cutaneous sarcoidosis (log-transformeddata)

Biomarker

Healthy volunteers (n = 5)

Mean 6 SD (range)

Patients with sarcoidosis (n = 15)

Mean 6 SD (range)

a1-Antitrypsin 2.010 6 0.272 (1.720-2.440) 2.154 6 0.456 (1.360-3.230)Complement 3 0.853 6 0.120 (0.682-0.976) 1.116 6 0.187 (0.764-1.380)Cancer antigen 125 9.834 6 5.395 (2.110-15.000) 36.241 6 20.311 (7.120-81.500)Cancer antigen 199 3.460 6 5.387 (0.123-12.500) 19.821 6 31.311 (0.123-93.100)Calcitonin 14.516 6 25.706 (3.020-60.500) 3.095 6 0.508 (2.340-4.820)CD40 0.635 6 0.180 (0.413-0.907) 1.583 6 1.372 (0.560-5.890)CD40 ligand 0.658 6 0.466 (0.110-1.260) 0.977 6 0.488 (0.010-1.690)Carcinoembryonic antigen 1.261 6 0.529 (0.397-1.730) 1.912 6 1.236 (0.492-5.080)Creatine kinase MB 0.882 6 0.853 (0.105-2.130) 0.403 6 0.426 (0.059-1.750)C-reactive protein 1.954 6 3.204 (0.261-7.670) 12.263 6 10.616 (0.581-32.600)Epidermal growth factor 77.515 6 83.326 (3.675-218.00) 73.535 6 50.015 (9.020-173.00)Adiponectin 3.742 6 1.983 (1.890-6.210) 2.698 6 1.658 (0.574-6.050)Epithelial neutrophil-activating protein 78 7.223 6 8.671 (0.786-19.900) 11.255 6 6.810 (1.370-22.500)Endothelin 1 7.872 6 9.575 (3.590-25.000) 6.010 6 4.550 (3.590-16.900)Extracellular newly identified receptor foradvanced glycation end product

33.280 6 11.979 (23.400-51.700) 40.533 6 26.281 (10.000-86.300)

Eotaxin 49.800 6 34.966 (12.200-104.00) 63.814 6 39.848 (8.710-164.00)Erythropoietin 83.000 6 0.000 (83.000-83.000) 83.000 6 0.000 (83.000-83.000)Fatty acid binding protein 2.132 6 1.739 (0.930-5.210) 1.482 6 0.820 (0.530-4.200)Factor VII 445.800 6 139.709 (273.00-586.00) 425.867 6 126.140 (257.00-729.00)Ferritin 157.180 6 86.973 (54.900-289.00) 317.820 6 647.698 (16.900-2580.0)Fibroblast growth factor basic 69.340 6 30.102 (49.000-116.00) 79.567 6 75.390 (28.600-291.00)Fibrinogen 0.005 6 0.001 (0.005-0.008) 0.009 6 0.007 (0.005-0.027)a2-Macroglobulin 0.563 6 0.255 (0.282-0.973) 0.463 6 0.143 (0.266-0.737)Granulocyte colony-stimulating factor 6.896 6 4.327 (2.500-13.800) 14.536 6 6.166 (6.460-24.800)Growth hormone 0.098 6 0.050 (0.061-0.179) 1.044 6 1.087 (0.067-3.610)Granulocyte-macrophage colony-stimulatingfactor

18.414 6 14.088 (2.570-28.700) 22.595 6 10.539 (3.400-28.700)

Glutathione S-transferase 0.744 6 0.126 (0.635-0.925) 0.714 6 0.189 (0.451-1.140)Haptoglobin 1.087 6 0.422 (0.565-1.440) 1.673 6 1.328 (0.177-5.060)Intercellular adhesion molecule-1 62.220 6 21.787 (27.500-77.700) 97.766 6 74.596 (0.939-295.00)Interferon-g 2.856 6 1.243 (2.300-5.080) 3.446 6 2.682 (2.300-12.600)Immunoglobulin A 2.420 6 1.512 (1.460-5.070) 2.731 6 1.971 (0.008-7.220)Immunoglobulin E 7.512 6 0.802 (7.000-8.830) 266.014 6 609.689 (1.380-1780.0)Insulin-like growth factor-1 21.420 6 10.040 (13.100-36.900) 18.220 6 22.906 (1.320-90.800)Alpha fetoprotein 1.431 6 0.815 (0.868-2.840) 2.930 6 1.463 (0.712-5.190)Immunoglobulin M 1.197 6 0.510 (0.636-1.930) 1.140 6 0.770 (0.053-3.350)Interleukin-10 4.322 6 1.564 (2.210-5.870) 7.778 6 2.799 (3.380-13.200)Interleukin-12p40 0.610 6 0.000 (0.610-0.610) 0.542 6 0.137 (0.177-0.610)Interleukin-12p70 33.520 6 8.973 (23.800-44.300) 30.540 6 6.183 (21.900-40.600)Interleukin-13 48.120 6 6.769 (39.700-55.400) 46.920 6 11.283 (34.300-74.800)Interleukin-15 0.351 6 0.168 (0.236-0.640) 0.412 6 0.198 (0.131-0.966)Interleukin-16 328.000 6 144.300 (202.00-566.00) 384.667 6 157.983 (152.00-651.00)Interleukin-18 186.440 6 96.692 (66.200-313.00) 591.467 6 661.000 (2740.0-159.00)Interleukin-1a 0.001 6 0.000 (0.001-0.001) 0.001 6 0.000 (0.001-0.003)Apolipoprotein A1 0.298 6 0.149 (0.160-0.538) 0.358 6 0.151 (0.203-0.799)Interleukin-1b 0.527 6 0.231 (0.284-0.858) 0.870 6 0.337 (0.417-1.450)Interleukin-1 receptor antagonist 73.380 6 37.527 (28.500-126.00) 200.880 6 151.342 (7.500-669.00)Interleukin-2 30.000 6 0.000 (30.000-30.000) 30.000 6 0.000 (30.000-30.000)Interleukin-3 0.057 6 0.027 (0.030-0.087) 0.078 6 0.022 (0.027-0.122)Interleukin-4 52.000 6 0.000 (52.000-52.000) 52.000 6 0.000 (52.000-52.000)Interleukin-5 16.250 6 0.000 (16.250-16.250) 15.316 6 3.617 (2.240-16.250)Interleukin-6 4.296 6 2.486 (1.190-6.100) 11.125 6 27.137 (1.190-109.00)

Continued

J AM ACAD DERMATOL

VOLUME 66, NUMBER 6Judson et al 910.e1

Table V. Cont’d

Biomarker

Healthy volunteers (n = 5)

Mean 6 SD (range)

Patients with sarcoidosis (n = 15)

Mean 6 SD (range)

Interleukin-7 72.880 6 14.684 (53.900-93.100) 67.187 6 17.388 (40.700-93.100)Interleukin-8 8.678 6 3.134 (3.390-11.100) 25.282 6 24.775 (2.300-96.000)Apolipoprotein CIII 90.780 6 42.110 (51.100-159.00) 151.400 6 172.896 (34.000-748.00)Insulin 7.422 6 4.664 (2.470-12.400) 8.260 6 7.981 (0.430-23.400)Leptin 3.870 6 3.386 (0.490-9.060) 32.597 6 26.853 (0.227-105.00)Lipoprotein A 109.350 6 51.831 (72.700-146.00) 149.640 6 117.428 (23.100-419.00)Lymphotactin 0.189 6 0.000 (0.189-0.189) 0.189 6 0.000 (0.189-0.189)Monocyte chemotactic protein 1 89.860 6 23.325 (70.000-130.00) 75.113 6 56.529 (27.000-232.00)Monocyte-derived chemokine 320.000 6 69.430 (210.00-379.00) 642.600 6 454.852 (224.00-2070.0)Monocyte inflammatory protein 1a 24.160 6 6.539 (13.800-31.200) 35.555 6 20.473 (9.620-88.900)Monocyte inflammatory protein 1b 187.400 6 107.043 (66.000-355.00) 363.733 6 196.741 (113.00-716.00)Matrix metalloproteinase 2 63.780 6 25.089 (18.900-75.000) 61.227 6 24.954 (15.300-82.400)Matrix metalloproteinase 3 9.996 6 2.347 (7.510-13.200) 9.324 6 9.315 (1.790-32.100)Apolipoprotein H 210.400 6 24.399 (180.00-247.00) 203.733 6 42.296 (140.00-272.00)Matrix metalloproteinase 9 21.760 6 7.066 (18.600-34.400) 23.300 6 15.395 (10.100-61.200)Myeloperoxidase 485.600 6 331.204 (197.00-940.00) 1141.07 6 913.585 (170.00-3570.0)Myoglobin 37.156 6 49.426 (7.090-124.00) 11.827 6 7.691 (4.510-28.400)Plasminogen activator inhibitor 1 204.520 6 89.880 (98.600-330.00) 296.333 6 91.503 (193.00-529.00)Prostatic acid phosphatase 0.291 6 0.083 (0.201-0.384) 0.379 6 0.125 (0.141-0.571)Pregnancy-associated plasma protein A 0.019 6 0.000 (0.019-0.019) 0.018 6 0.002 (0.012-0.019)Prostate specific antigen free 0.386 6 0.154 (0.208-0.561) 0.238 6 0.583 (0.012-2.190)Regulated on activation, normal T expressedand secreted

34.060 6 22.611 (14.500-73.000) 41.326 6 25.178 (9.190-88.300)

Serum amyloid P 22.640 6 5.766 (14.100-28.300) 23.647 6 5.884 (16.500-39.000)Stem cell factor 302.880 6 238.681 (65.400-649.00) 415.933 6 193.389 (179.00-799.00)b2-Microglobulin 1.460 6 0.437 (0.938-2.020) 2.903 6 1.840 (1.330-8.350)Serum glutamic oxaloacetic transaminase(or aspartate aminotransferase)

15.982 6 5.670 (7.410-23.000) 18.467 6 5.730 (10.000-29.600)

Sex hormone-binding globulin 48.520 6 18.134 (32.600-73.400) 49.990 6 36.449 (7.250-140.00)Thyroxine binding globulin 57.780 6 7.397 (46.400-66.400) 68.680 6 17.773 (43.500-107.00)Tissue factor 0.372 6 0.109 (0.177-0.421) 0.383 6 0.083 (0.177-0.470)Tissue inhibitor of metalloproteinase 1 106.520 6 34.494 (72.300-154.00) 150.067 6 68.802 (85.000-374.00)Tumor necrosis factor receptor II 3.770 6 1.221 (2.850-5.840) 8.717 6 5.843 (2.500-23.800)Tumor necrosis factor-alpha 2.268 6 0.454 (1.995-3.070) 8.678 6 8.573 (0.812-29.300)Tumor necrosis factor-beta 18.766 6 9.020 (22.800-2.630) 16.555 6 9.306 (2.070-22.800)Thrombopoietin 2.260 6 0.564 (1.350-2.830) 2.315 6 0.506 (1.490-3.150)Thyroid-stimulating hormone 1.261 6 0.568 (0.386-1.660) 1.190 6 0.455 (0.306-1.930)Brain-derived neurotrophic factor 20.960 6 6.569 (13.700-31.200) 22.987 6 9.527 (10.400-42.600)Vascular cell adhesion molecule 1 421.200 6 151.397 (224.00-575.00) 675.133 6 281.835 (327.00-1230.0)Vascular endothelial growth factor 434.400 6 248.652 (227.00-721.00) 639.867 6 431.447 (217.00-1870.0)von Willebrand factor 56.440 6 17.940 (35.500-78.500) 64.347 6 32.254 (27.900-146.00)

J AM ACAD DERMATOL

JUNE 2012910.e2 Judson et al