Integrative biology identifies shared transcriptional networks in CKD

BASIC RESEARCH www.jasn.org

Integrative Biology Identifies Shared TranscriptionalNetworks in CKD

Sebastian Martini,*†‡ Viji Nair,*†‡ Benjamin J. Keller,§ Felix Eichinger,*†‡ Jennifer J. Hawkins,*†‡

Ann Randolph,*†‡ Carsten A. Böger,| Crystal A. Gadegbeku,¶ Caroline S. Fox,**††

Clemens D. Cohen,‡‡ Matthias Kretzler,*†‡ the European Renal cDNA Bank,C-PROBE Cohort, and CKDGen Consortium

Departments of *Internal Medicine, †Nephrology, and ‡Computational Medicine and Bioinformatics, University ofMichigan, Ann Arbor, Michigan; §Department of Computer Science, Eastern Michigan University, Ypsilanti, Michigan;|Department of Internal Medicine II, University Hospital Regensburg, Regensburg, Germany; ¶Department ofMedicine, Section of Nephrology and Kidney Transplantation, Temple University School of Medicine,Philadelphia, Pennsylvania; **Division of Intramural Research and Laboratory for Population and Metabolic Health,National Heart, Lung, and Blood Institute, Framingham, Massachusetts; ††Department of Endocrinology, Brighamand Women’s Hospital, Boston, Massachusetts; and ‡‡Institute of Physiology, University of Zürich, Zürich,Switzerland

ABSTRACTApreviousmeta-analysis of genome-wide association data by theCohorts for Heart andAging Research inGenomic Epidemiology and CKDGen consortia identified 16 loci associated with eGFR. To define howeach of these single-nucleotide polymorphisms (SNPs) could affect renal function, we integrated GFR-associated loci with regulatory pathways, producing a molecular map of CKD. In kidney biopsy specimensfrom 157 European subjects representing nine different CKDs, renal transcript levels for 18 genes inproximity to the SNPs significantly correlated with GFR. These 18 genes were mapped into their biologiccontext by testing coregulated transcripts for enriched pathways. A network of 97 pathways linked byshared genes was constructed and characterized. Of these pathways, 56 pathways were reportedpreviously to be associatedwith CKD; 41 pathways without prior associationwith CKDwere ranked on thebasis of the number of candidate genes connected to the respective pathways. All pathways aggregatedinto a network of two main clusters comprising inflammation- and metabolism-related pathways, with theNRF2-mediated oxidative stress response pathway serving as the hub between the two clusters. In all, 78pathways and 95% of the connections among those pathways were verified in an independent NorthAmerican biopsy cohort. Disease-specific analyses showed that most pathways are shared between setsof three diseases, with closest interconnection between lupus nephritis, IgA nephritis, and diabetic ne-phropathy. Taken together, the network integrates candidate genes from genome-wide association stud-ies into their functional context, revealing interactions and defining established and novel biologicmechanisms of renal impairment in renal diseases.

J Am Soc Nephrol 25: 2559–2572, 2014. doi: 10.1681/ASN.2013080906

CKD affects .13% of the Unites States populationand is a major contributor to cardiovascular morbid-ity and mortality.1,2 Defining the pathophysiology ofCKD is critical to identify predictors of the diseasecourse and therapeutic targets. Diverse mechanismshave been linked with the development and progres-sion of CKD using experimental models and renaltissues based expression studies.3–6

Received August 28, 2013. Accepted April 30, 2014.

Published online ahead of print. Publication date available atwww.jasn.org.

Correspondence: Dr. Matthias Kretzler, University of Michigan,Department of Internal Medicine/Division of Nephrology, 1150West Medical Center Drive, 1560 MSRB II, Ann Arbor, MI 48109-0676. Email: [email protected]

Copyright © 2014 by the American Society of Nephrology

J Am Soc Nephrol 25: 2559–2572, 2014 ISSN : 1046-6673/2511-2559 2559

The CKDGen and CHARGE consortia were able to usegenome-wide association studies (GWASs) to identify preex-isting genetic risk factors for renal function decline.7,8 How-ever, GWASs, just like renal tissue gene expression profilingstudies, only capture one aspect of the underlying pathophys-iology of CKD. A combined effort, linking the knowledge ofgenetic and transcriptomic alterations with clinical phenotypeinformation in CKD, is now feasible and offers the opportu-nity for an integrated understanding of the intrarenal driversof CKD.

This study used a sequential strategy to construct andinterpret a network of CKD-associated pathways that com-bines distinct but complementary sources of data: GWAScandidate genes, renal biopsy-derived transcriptional profileswith matching clinical information, and literature-derivedknowledge of molecular pathways.

The CKDGen GWAS identified genetic loci associated witheGFR. Intrarenal gene expression levels generated from anindependent population of subjects were tested for correlationwith eGFR. Genes with evidence for eGFR association fromthose two independent lines of evidence (genetic and tran-scriptomic) were evaluated for their molecular functionalcontext in CKD using a coexpression strategy. For each gene inthe intersection, mRNAs with stringent coexpression in therenal tissues of subjects with CKD were identified. Thefunctional context of coexpressed gene sets was defined usingprior biologic knowledge derived from comprehensive path-way databases, linking each gene from GWAS through geneexpression and eGFR correlation to a set of molecular path-ways. Connections between these pathways indicate that theyshare at least one transcript correlated with one or moreGWAS-derived candidate genes, allowing the construction of anetworkof interacting pathways inCKD. For each renal diseaseincluded in the CKD dataset, the disease-specific interplay ofCKD-related pathways was identified and used to define atranscript-based similarity matrix between the glomerulardiseases studied.

The combinationof genetic,molecular, and clinical datasetsin individuals with CKD allows for defining renal diseases onthe basis of their shared and specific molecular mechanism.Identification of CKD pathways and their interplay providesstarting points for experimental studies to define the biologicmechanisms underlying chronic renal failure.

RESULTS

eGFR CorrelationA meta-analysis by the CKDGen and CHARGE consortiaidentified 16 loci associated with renal function.7 A follow-upstudy identified 13 additional loci for renal function and CKDin the same regions.8 Forty candidate genes were located inproximity (660 kb) of 16 single-nucleotide polymorphisms(SNPs) and evaluated fordifferential regulation inCKD(Table1).Of them, 29 transcripts were found to be expressed above

background in transcriptional expression profiles of renal bi-opsies of 157 subjects (Figure 1, step 1). These biopsies werefrom subjects diagnosed as having one of nine different chronicrenal diseases (FSGS, membranous GN [MGN], minimalchange disease [MCD], diabetic nephropathy [DN], hyperten-sive nephropathy [HTN], IgA nephritis [IgAN], lupus nephri-tis [LN], and thin-membrane disease) or histologically unaf-fected parts of tumor nephrectomies. Renal biopsies from 10living kidney donors (LDs) served as controls to test for dis-ease-specific regulation. Disease cohort sizes ranged from 4 to30 patients, with LN (30 patients) and IgAN (24 patients) beingthe largest subcohorts. The mean age at the time of biopsy was46617 years (mean6SD), sex ratio was 90:67 (men/women),and the average eGFR was 70636 ml/min per 1.73 m2,covering a range from 44 to 101 ml/min per 1.73 m2 (Table2). The validation cohort, although different in disease compo-sition, had comparable clinical characteristics (Table 2).

Tubulointerstitial and glomerular gene expression profileswere used to compute the correlation of log eGFRwith the log-transformed steady-state expression levels of 29 candidateswithin each subject (Figure 1, step 2A). In the tubulointerstitialcompartment, 18 of 29 candidate genes were found to signif-icantly correlate with renal function:VEGFA,ANXA9,NAT8B,SLC34A1, TFDP2, ACSM5, SLC7A9, LASS2, FBXO22,UMOD,PIP5K1B, NAT8, GP2, DAB2, ALMS1, LMAN2, PRUNE, andF12 (false discovery rate [FDR]#0.01, |r|$0.25) (Figure 2A).Tubulointerstitial gene expression profiles for these 18 tran-scripts across all subjects with CKDs versus LD controls areshown in Table 1. Four of these transcripts also passed thesignificance filter in the glomerular compartment (VEGFA,ANXA9, NAT8B, and SLC34A1), with only DACH1 specificto the glomerular compartment. Therefore, CKDGen candi-date genes were enriched for eGFR-correlated genes in thetubulointerstitial compartment comparedwith a randomgeneset (z-score for enrichment compared with random datasets:3.86), and additional analysis focused on the tubulointerstitialgene expression datasets. The directionality of correla-tion of transcript levels with eGFR was conserved across alldiseases.

Gene Coexpression PathwaysThe 18 CKDGen-associated and eGFR-correlated candidategenes were evaluated for their functional context using acoexpression strategy (Figure 1, step 2B). For 14 of 18 candi-date genes, the following number of transcripts correlatedwith the expression of the candidate gene (FDR#0.01,|r|.0.5), providing a basis for the detection of enriched path-ways among coexpressed genes: SLC7A9 (1078), VEGFA(988), ACSM5 (925), SLC34A1 (839), NAT8B (811), ANXA9(690), LASS2 (519), DAB2 (365), NAT8 (283), GP2 (174),TFDP2 (118), UMOD (83), LMAN2 (57), and F12 (22). Theresulting 14 coexpressed gene sets show significant enrichmentof 147 unique canonical pathways. Of these pathways, 97 path-ways were identified in at least two candidate-correlatedgene sets.

2560 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

Association of Pathway Network withCKDMultiple lines of evidence implicate thepathways constructed from CKDGen can-didates in CKD.

(1) Fifty-six of ninety-seven pathways areknown to be associated with CKD.Supplemental Table 1 shows these path-ways, an exemplar reference, and a briefsummary of the proposed mechanismof CKD involvement of that pathwayor group of pathways. The 41 pathwaysnot previously associated with CKDare shown in Supplemental Table 2ranked on the basis of the number ofcandidate genes connected to the re-spective pathways.

(2) Because groups of genes can contributeto several disease-relevant pathways, wecreated a network of all 97 pathwaysto investigate the inter-relationshipsamong pathways in the context ofCKD. In this network, each pathwayis represented by a single node, andgenes shared between pathways form theedges (connections betweennodes). Thepathways cluster in a bowtie structureconnecting two subclusters with CKDrelevance: inflammation-related path-ways (e.g., NRF2-mediated oxidativestress response, Cdc42 signaling, andNF-kB signaling) and metabolic path-ways (e.g., fatty acid metabolism, gly-colysis/gluconeogenesis, and tyrosinemetabolism) (Figure 3). Transcripts fromthese subclusters show differential ex-pression consistent with independentstudies where the majority of genes en-riched in inflammatory pathways haveincreased expression in CKD,9 whereasmetabolism pathway genes are mainlyrepressed.10–12 The two subclusters areinterconnected by pathways with mul-tiple connecting genes. The pathwayswith the largest number of neighbors arexenobiotic signaling (53 connections),arylhydrocarbon-receptor signaling (45connections), and NRF2-mediated oxi-dative stress response (44 connections)pathways.

(3) The majority of genes shared amongpairs of pathways (281 of 309 genes

Table 1. List of 16 SNPs associated with CKD originally identified by Köttgenet al.7,8 and candidate genes (Gene Symbol and Entrez GeneID) within a 60-kbrange, gene expression correlation with eGFR within the original cohort (n=157)for candidate genes correlated with eGFR (FDR,0.01), and log2 fold change ofcandidate genes in all patients with CKDs compared with LD controls (FDR wasNS)

SNP/Genes within660 kb (Gene Symbol)

Entrez GeneIDeGFR Correlationr Original Cohort

(FDR,0.01)

log2 Fold Change AllCKD Versus LD (FDR)

rs17319721SHROOM3 57619CCDC158 339965

rs10109414STC1 6781

rs12917707ACSM5 54988 0.44 20.18 (0.1)GP2 2813 0.32 20.11 (0.11)PDILT 204474UMOD 7369 0.37 0.14 (0.29)

rs267734LASS2 29956 0.39 20.08 (0.17)ANXA9 8416 0.61 20.04 (0.28)SETDB1 9869FAM63A 55793BNIPL 149428PRUNE 58497 0.27 0.14 (0.03)

rs1260326GCKR 2646FNDC4 64838IFT172 26160

rs13538ALMS1 7840 20.38 0.11 (0.06)NAT8 9027 0.29 0.11 (NS)NAT8B 51471 0.57 0.41 (0.16)

rs347685TFDP2 7029 0.45 20.03 (0.27)

rs11959928DAB2 1601 0.28 0.13 (0.34)C9 735

rs6420094SLC34A1 6569 0.55 20.42 (0.06)F12 2161 20.43 0.22 (0.07)GRK6 2870LMAN2 10960 20.38 0.18 (0.01)PFN3 345456PRR7 80758RGS14 10636

rs881858VEGFA 7422 0.59 20.33 (0.01)

rs7805747PRKAG2 51422

rs4744712PIP5K1B 8395 0.3 20.14 (0.1)FAM122A 116224

rs653178ATXN2 6311

rs626277DACH1 1602

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2561

www.jasn.org BASIC RESEARCH

forming edges in the network) displays an expressionpattern that correlates with log eGFR (0.26#|r|#0.63,FDR,0.01). Protein products of the connecting genesalso form a protein–protein interaction network with ahigher connectivity than randomized versions of the samenetwork (clustering coefficient=0.27 versus 0.09, P=0.01),which suggests a strong functional interconnection of theencoded proteins (which has previously been shown forSaccharomyces cerevisiae)13 (Supplemental Figure 1).

The pathway network relationships were tested in an indepen-dent biopsy gene expression dataset obtained from the ClinicalPhenotyping Resource and Biobank Core (C-PROBE) cohort, aNorth American CKD cohort with different environmentalexposures and ethnic backgrounds but similar range of renalimpairment (Table 2). Using the same strategy, 78 of 97 pathways(80%) from the original cohort were identified in the independentdataset. The validation study found 127 pathways connected by341 genes shared among pathways compared with 313 genesshared among pathways in the original network. Among the rep-licated pathways, 95% of the genes connecting sets of pathwayswere retained in the independent cohort. The 14 candidate genesthat went into additional analysis in the original cohort preservedtheir directionality of eGFR correlation in the independent valida-tion cohort. Pearson correlation for genesALMS1,ANXA9,LASS2,NAT8B, andVEGFA achieved statistical significance in this cohort.

CKD Pathway Networks: Shared and Disease-SpecificFeaturesThe pathway network shows that different renal diseases shareeGFR-associated pathways. By analyzing differentially regu-lated genes between the subjects with CKD and tissue from thehealthy kidney donors, enriched pathways were detected foreach disease. Additional subtle relationships among differentglomerular diseases leading to CKD can be identified. Thedisease–pathway bipartite network (Figure 4) and one-modeprojection disease network (Figure 5) compress the informa-tion shown in Figure 3 to reveal disease–pathway relation-ships. Using this approach, the following elements of thedisease–pathway network can be displayed. Although none

of the pathways are shared among allseven CKDs, the cytotoxic T lymphocyte-mediated apoptosis pathway is sharedamong FSGS, DN, MGN, IgAN, and LN(Supplemental Figure 2, A–G), and theNRF2-mediated oxidative stress responsepathway is shared among DN, HTN,MCD, and MGN. Three diseases showunique pathways (DN: caveolar-mediatedendocytosis signaling, crosstalk betweendendritic cells and natural killer cells, leuko-cyte extravasation signaling, mitochondrialdysfunction, phospholipase C signaling,and a-adrenergic signaling; MCD: glioblas-toma multiforme signaling; MGN: xenobi-

otic metabolism signaling), whereas 12 pathways are sharedamong various sets of three diseases. The most pathways fromthe CKD pathway network are enriched among differentially reg-ulated genes inDN(31) andLN(22),whereas the fewest pathwayswere found for FSGS (5) andHTN (1). The strongest relationshipamong diseases as defined by Jaccard coefficient (defined as theratio of the number of pathways shared by a set of diseases to thenumber of pathways associated with all diseases in the set) in-volves LNwith both IgANandDN (Table 3). These three diseasesalso share the most pathways of any set of three diseases.

DISCUSSION

An Integrative Approach to GWAS Data AnalysesGWASs have identified several genetic loci with highly signif-icant associations with CKD and eGFR. These genetic associ-ations have yet to be translated into a molecular understandingof disease susceptibility (i.e., how the SNPs affect renal functionmechanistically). SNPs in complex diseases have been shown tobe able to influence protein function14 andmRNA expression.15

Various approaches have yielded insights into molecular mech-anisms of CKD (reviews in refs. 16–18), but comprehensivefunctional integration of genetic CKD risk loci with the biologyof CKD is still lacking.

Our study introduces an approach to extend GWAS resultsby transcriptomic analyses to uncover the rich informationhidden in CKDGWAS. Several types of data were integrated toidentify a hierarchy relating candidate genes to coexpressedtranscripts to function in relationship to CKD (overview inFigure 1).19 The validity of this unbiased approach is suppor-ted by the identification of pathways with known CKD asso-ciations, providing a rationale to further evaluate the knownand novel pathways as targets for therapeutic intervention.

The CKD Pathway Network: Crosstalk among MultipleMolecular MechanismsThe concept of a single molecular pathway driving a diseaseprocess is aconstruct to facilitateourunderstandingofbiology. Incontrast to this concept, comprehensive studies in model

Table 1. Continued

SNP/Genes within660 kb (Gene Symbol)

Entrez GeneIDeGFR Correlationr Original Cohort

(FDR,0.01)

log2 Fold Change AllCKD Versus LD (FDR)

rs1394125UBE2Q2 92912FBXO22 26263 0.27 20.04 (0.23)

rs12460876TDRD12 91646CCDC123 84902SLC7A9 11136 0.39 20.18 (0.12)

Association between eGFR and gene expression levels of 18 candidate genes analyzed further remainssignificant after correction of age and sex.

2562 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

organisms show genes and pathways in dense interrelationships,with multiple genes mapping onto multiple pathways.20 Weexploit this property in our data by performing a pathway–crosstalk analysis of the gene sets linked to eGFR. Graphicpresentation of the network of pathways reveals that themajorityof the CKD pathways aggregates in either an inflammation- or ametabolism-related cluster, corresponding to two major hall-marks of CKD.21 Genes from pathways in those clusters displaysteady-statemRNA expression patterns consistent with previousfindings in CKD (i.e., upregulation of inflammatory genes, suchas HLA isoforms, TLR1, TLR3, and NFKB1) (review in ref. 22).Multiple eGFR-correlated genes, including RAC1, RAC2,TNFRSF1B, IL1R1, and RXRA, establish links between thesetwo dominant CKD pathway groups.

The structure of the network identifies three pathways withthe most neighbors: the xenobiotic signaling pathway, thearylhydrocarbon-receptor signaling pathway, and the NRF2 ox-idative stress pathway. The xenobiotic signaling pathway wasoriginally identified as a major player in the elimination ofdrugs.23 Beyond this metabolic function, additional hypothesesfor this pathway in the progression of CKD can be postulated.Transcriptional mechanisms mediating the coordinated down-regulation of this pathway have been associated with disease

progression in inflammatory bowel disease,24 and similar mech-anisms could be involved in progression of CKD.25

The second most interconnected pathway is thearylhydrocarbon-receptor signaling pathway. Arylhydrocar-bon receptors are members of the basic helix-loop-helixPer-Arnt-Sim receptor superfamily that are closely related tothe xenobiotic signaling pathway, with receptor activation (i.e.,by statins) leading to upregulation of xenobiotic pathwaymembers.26,27 In CKD animal models, statins, through thearylhydrocarbon receptor, are involved in the elimination of ure-mic toxins and have been shown to decrease organ damage.27

The third most interconnected pathway, the NRF2 oxidativestress pathway, links to the relevance of oxidative stress in CKD,particularly in DN.28–30 In general, low concentrations of pro-oxidants have potentially protective effects, because they act assecondarymessengers of cell survival.31 Excessive oxidative stress,however, is highly prevalent in patients withCKD and believed tocontribute to more traditional cardiovascular risk factors in thispatient cohort.28 TargetingNRF2, therefore, is an attractive strat-egy, and an NRF2 agonist has been studied in phase II and IIItrials in DN CKD stages 3 of 4. Studies were halted because of anincreased cardiovascular mortality before conclusive evidence onthe effect on CKD could be generated.32,33

Figure 1. Integration of GFR-associated loci with regulatory pathways allowed for the generation of a molecular map of CKD. Schematicillustration of the strategy to identify CKD-associated pathways and their connections to molecular mechanisms of renal function.

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2563

www.jasn.org BASIC RESEARCH

The crosstalk between pathways indicates a plethora of con-nectionsbetweenmetabolism-andinflammation-relatedpathways.Becausemanygenes enriched inoneof these top three pathways arealso enriched in other pathways, they form major hubs in thenetwork and might be capable of affecting the entire regulatorynetwork balance, establishing intriguing therapeutic targets.

Additional evidence for the CKD association for genesconnecting pathways within the network is provided by theirrole as CKD candidate biomarkers (i.e., TIMP-1, FABP1, GGT1,and TNFRSF1B).6 These CKD coregulated genes also interacton a protein level (Supplemental Figure 1), supporting the con-cept of identifying shared functional concepts by coregulationanalysis. Only 1 of the original 14 CKDGen candidates,ACSM5,was also a pathway-connecting gene, highlighting the impor-tance of integrating additional molecular information fromCKD to unravel information linked to GWAS candidate genes.

Pathways Specific to Individual Kidney DiseasesThe bipartite disease–pathway network reveals relationships be-tween individual renal diseases and CKD pathways (Figure 4),highlighting a key finding of this study: renal diseases show asubstantial overlap in gene expression changes, irrespective ofthe initiating disease mechanism. The observation that the tu-bulointerstitial compartment shows robust correlation ofCKDGen-associated transcripts across renal diseases is consis-tent with the longstanding knowledge in nephropathology that

tubulointerstitial lesions correlate and predict renal functionmore robustly than glomerular alterations.34 The limited num-ber of pathways common among diseases sharing a clinical pat-tern (i.e., nephrotic syndrome) might also be explained by theanalysis of the tubulointerstitial versus glomerular compartmentof renal biopsies. Additional mechanisms and pathways (e.g., fordisease initiation and development)might bemore similar in theglomerular compartment among nephrotic diseases.

Pathways Not Previously Associated with KidneyDiseaseAlthough an in-depth analysis of 41 pathways not previouslyassociatedwithCKD is beyond the scope of this work, a ranked listof these pathways on the basis of the number of candidate genesconnected to the respective pathways is provided (SupplementalTable 2). This list could prioritize future research on mechanismsofCKD.Among41pathwaysnot previously found tobe associatedwith CKD is the hepatic fibrosis/hepatic stellate cell activationpathway, which might be activated in both CKD and hepatic fi-brosis. Although the pathway was originally described for tran-scripts associated with liver disease, core components, like theidentified transcripts CTGF, SMAD2, SMAD3, SMAD4, VEGFA,VEGFB, TGFBR1, TGFBR2, andMMP13, have also been found inrenalfibroticdisease states.35–37Elements of the renalfibrosis path-ways replicating in independent organ systems can be consideredcore elements of chronic fibrotic mechanism of human disease.

Table 2. Clinical characteristics of subjects per disease group analyzed by oligonucleotide array-based gene expressionprofiling and correlation analysis with clinical parameters

Condition N Sex (men/women) Age (yr) GFR (ml/min per 1.73 m2) Proteinuria (g/d)

Original cohortLD 10 9:1 46613 81616 ,0.2DN 17 12:5 58611 44625 3.162.7FSGS 16 7:9 46618 73638 4.462.7HTN 20 15:5 57612 44625 1.461.5IgAN 24 18:6 36615 76638 2.462.4MCD 12 8:4 36617 101634 6.765.8MGN 18 10:8 53619 89641 4.663.2LN 30 7:23 35613 64629 3.163.3Thin-membrane disease 6 4:2 46614 93629 0.560.6Tumor nephrectomy 4 0:4 6567 60610 ,0.2Total 157 90:67 46617 70636 3.163.3

Validation cohortFSGS 4 4:0 57613 51632 3.363.4IgAN 6 3:3 50614 60635 2.262.0Immuntactoid GN 1 1:0 71 40 0.5LN 7 1:6 34611 76630 4.662.4MCD 1 1:0 27 152 0MGN 14 6:8 46614 92637 5.064.0HTN 2 1:1 55614 47647 0.560.6HTN/DN 1 1:0 71 70 15Vasculitis 1 1:0 66 22 2.1Amyloidosis 4 2:2 53613 49619 2.262.0DN 2 1:1 50614 39613 6.061.7Total 43 22:21 48615 71638 3.963.6

GFR and proteinuria are displayed as average values6SD.

2564 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

Twenty-four of forty-one pathways with no previousassociation to CKD have been linked to specific renal diseases.As shown in Supplemental Figure 2, A–F between 9 and 18 ofthe pathways not previously implicated in CKD were enrichedamong transcripts differentially expressed in disease versuscontrols for specific diseases. Nine pathways (allograft rejec-tion signaling, antigen presentation pathway, cytotoxic Tlymphocyte-mediated apoptosis of target cells, graft-versus-hostdisease signaling, hepatic fibrosis/hepatic stellate cell activation,LPS/IL-1–mediated inhibition of RXR function,OX40 signalingpathway, starch and sucrose metabolism, and type 1 diabetesmellitus signaling) were enriched (P,0.05) among six of sevendiseases, supporting the notion that these pathways or severalmembers of these pathways do play a role in renal function declinefor specific diseases.

Limitations of the ApproachOur study analyzed genetic association and transcriptomicdata fromtwo independent cohorts byfiltering candidate genesfor significant correlationof theirmRNAexpressionwith eGFRin the European Renal cDNA Bank-Kroener-Fresenius BiopsyBank (ERCB) and C-PROBE cohorts. All study participants inthe ERCBcohortwere of European descent. Themajority (129;87%)of all renal disease samples for this cohortwas collected instudy centers in southern Germany and northern Italy; theremaining 18 samples were procured in France, Czech Re-public, Ireland, and Spain. No correction for ancestry wasperformed in our study. The association between geneexpression levels and eGFR remained significant after correct-ing for age and sex in the 18 CKDGen candidate genes studied.The study reported here remains agnostic to how a specific

Figure 2. Tubulointerstitial gene expression profiles for 18 CKD candidate genes correlate with log eGFR across all subjects with CKD.(A) Heat map of the correlation of logarithmic eGFR with mRNA expression for 29 CKDGen candidate genes (r values ranging from20.76to 0.79). Includes 157 subjects from seven individual CKDs. The CKD column displays the aggregate correlation across all diseases.Glomerulum and tubulointerstitium of renal biopsies were analyzed separately. Gene names in red indicate significance in the tubu-lointerstitial expression dataset (FDR#0.01, |r|$0.25). Blue underlining or blue font indicates significance in the glomerular expressiondataset. Red gene names with blue underlining indicate significant coregulation in both renal compartments. FG, FSGS; HT, hyper-tensive nephropathy; IA, IgA nephritis; LN, lupus nephritis; MC, minimal change disease; MN, membranous GN. (B) Exemplary graphshowing the correlation of VEGFA expression with corresponding eGFR data across 157 subjects and statistical assessment.

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2565

www.jasn.org BASIC RESEARCH

SNP is affecting the mRNA levels found to be correlated witheGFR, and substantial additional work integrating multipleinformation sources and approaches, like cis-eQTL studies,will be needed to address this question (a recent study in DNgenetics is in the work by Martini et al.38). With regard to thelimited number of samples in each disease subgroup, ourstudy analyzed the CKD cohort as a whole rather than facili-tating adjustments for aforementioned potential confounders.The analysis of pathways shared among specific disease sub-groups in the CKD cohort was not corrected for heterogeneityin sample size given the limited sample sizes in each diseasesubcohort. Diseases with larger sample sizes (i.e., LN) could,therefore, have a greater influence on the bipartite and themode projection networks presented in Figures 4 and 5.

Of the original 40 CKDGen candidate genes, 29 genes werepresent on the Affymetrix Chip, excluding the missing genesfrom our analysis. Newer transcriptomic analysis techniques,such as RNA sequencing, could fill these gaps in the future. It

should be noted that the decision to include correlatedtranscripts of candidate genes passing a cutoff of FDR,0.01and |r|.0.50 was operationally derived. With this cutoff, eachcandidate gene generated a list of correlated transcripts ofdifferent length and thereby, potentially biased the influenceof each candidate gene in identifying pathways. Finally, tran-scriptional data are a reflection of the regulatory mechanismactivated in a given tissue and not a reflection of the functionalstate of the encoded proteins. Complementing the transcrip-tional approach with genome-scale proteomic analysis will bean important next step.

The CKD Pathway Network as a Resource for the RenalResearch CommunityTo allow interrogation of the CKD pathway network andthe underlying datasets for individual genes and pathwaysas well as pathway interaction, the data generated in thiswork are provided in Supplemental Material linked to a

Figure 3. Integration of GFR-associated loci with transcriptional data allowed for the generation of a comprehensive pathway map ofCKD. Network graph of 97 CKD-associated pathway nodes connected by transcripts coregulated with CKD candidate genes. Nodesdisplay significantly (P,0.05) enriched pathways derived from two or more lists of coregulated candidate genes, and node size reflectsthe number of connections to other pathways (degree). A spring-embedded layout is applied, pulling together pathway nodes ac-cording to the number of shared genes among pathways. Unbiased cluster analysis of the CKD pathway network separates a subclusterof mainly inflammation-related pathways (red nodes; i.e., NRF2-mediated oxidative stress response, Cdc42 signaling, and NF-kBsignaling) and a subcluster of several metabolic pathways (green nodes; i.e., fatty acid metabolism, glycolysis/gluconeogenesis, andtyrosine metabolism). Subclusters were identified by AllegroMCODE, a Cytoscape plug-in, on the original CKD pathway network.

2566 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

searchable interface through the open-source, platform-independent Cytoscapesoftware (version 2.8.3). Using Cytoscape,all visualization and network analysistools described in Concise Methods can beused. Information on individual pathways,genes that are shared among sets of path-ways, and their eGFR correlation valuesfrom the entire cohort of 157 subjectscan be retrieved through the dynamic in-terface. Unprocessed gene expression data-sets are available in the Gene ExpressionOmnibus and have also been integratedinto the renal search engine Nephromine(www.nephromine.org) for additional sys-tems biology analysis by the renal researchcommunity.39

GWAS, intrarenal transcriptional profiles,andbiologic knowledge togetherwere able todefine a tight pathway–crosstalk activated

Figure 4. CKD associated pathways are shared between renal diseases. Bipartite network of the relationship between diseases andpathways, where a yellow node represents a disease, an open node represents a pathway, and an edge represents pathway associationwith a disease. Differentially regulated genes in disease versus control are mapped into pathways. For pathways, the node size reflectsthe number of diseases showing enrichment for these pathways genes, whereas the node size for diseases reflects the number ofpathway members regulated in disease versus control.

Figure 5. Disease-specific analyses shows a close interconnection between lupusnephritis, IgA nephritis, and diabetic nephropathy. One-mode projection of the re-lationships between diseases and their associated pathways, where nodes representthe individual diseases and edges represent the number of pathways shared betweenthem. Node size reflects the number of pathways shared with other diseases.

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2567

www.jasn.org BASIC RESEARCH

with impaired renal function. The activation of inflammatorysignaling cascades and the loss ofmetabolite functions provide awealth of information to be tested for causal relationships inadditional experimental studies. Targeting the key regulatoryhubs of the interlinked pathways will be a rational therapeuticapproach to affect the CKD network at multiple levels.

CONCISE METHODS

The goal of this study was to define CKD-related pathways and the

functional connections among them by integrating CKD candidate

gene information with transcriptomic and clinical data from 157

subjects with CKDs.

StrategyIn this study, previousobservations fromGWASs and transcriptomic

studies in CKDwere extended to integrate currently available GWAS

results into their functional context. Three distinct types of datawere

used: GWAS candidates, renal biopsy-derived transcriptional pro-

files, and clinical information. Combining these diverse data for all

CKDGen candidates allows the construction of a comprehensive

network of CKD pathways. Correlation analyses were first used to

test GWAS-derived CKD candidate genes for their association with

renal function in a kidney biopsy gene expression dataset (Figure 1,

step 2A). Pathway annotations were then used to identify functional

categories for transcripts correlated with GWAS-derived candidate

genes (Figure 1, step 2B). Relationships were captured among sig-

nificantly enriched pathways and displayed through a network in

which pathways are represented by single nodes and transcripts

shared by pathway pairs form edges (connections between nodes)

(Figure 1, steps 3 and 4). Individual pathways of this network were

then interrogated for known associations with CKD. A disease-

specific analysis defined the functional interplay of diseases on the basis

of enriched pathways among differentially expressed genes of each

glomerular disease studied (Figure 1, step 5). Genes forming the edges

of the pathway network were tested for their association with CKD as

well as their correlation with GFR (Figure 1, step 6).

Selecting Subjects for Human Renal BiopsyForty CKD candidate genes were derived from a published meta-

analysis that identified susceptibility loci for GFR.7 GWAS candidates

were linked with gene expression profiles from the ERCB. Renal tran-

scriptional expression profiles from renal biopsies of 157 subjects

Table 3. Relationships among diseases as defined by the Jaccard coefficient (defined as ratio of the number of sharedpathways for a set of diseases to the number of all pathways for these diseases)

Disease Count Diseases Disease and Disease Group-Specific PathwaysTotal SharedPathways

JaccardCoefficient

1 DN Caveolar-mediated endocytosis signaling, crosstalkbetween dendritic cells and natural killer cells, leukocyteextravasation signaling, mitochondrial dysfunction,phospholipase C signaling, a-adrenergic signaling

31 1.00

1 MCD Glioblastoma multiforme signaling 6 1.001 MGN Xenobiotic metabolism signaling 6 1.002 IgAN, LN Hepatic fibrosis hepatic stellate cell activation 15 0.652 DN, LN Altered T cell and B cell signaling in rheumatoid arthritis,

Cdc42 signaling, Fc-g-receptor–mediated phagocytosisin macrophages and monocytes, neuroprotective role ofTHOP1 in Alzheimer’s disease, Rac signaling

20 0.59

2 DN, IgAN B cell development 14 0.413 DN, IgAN, LN Allograft rejection signaling, autoimmune thyroid

disease signaling, CD28 signaling in T-helper cells,complement system, graft-versus-host disease signaling,type 1 diabetes mellitus signaling, fMLP signaling inneutrophils, iCOS/iCOSL signaling in T-helper cells

13 0.38

2 DN, FSGS Aryl hydrocarbon receptor signaling 5 0.162 DN, MCD Virus entry through endocytic pathways 4 0.123 DN, FSGS, LN Nur77 signaling in T lymphocytes 4 0.114 DN, FSGS, IgAN, LN Antigen presentation pathway, dendritic cell maturation 3 0.094 DN, IgAN, LN, MGN Calcium-induced T lymphocyte apoptosis,

OX40 signaling pathway3 0.09

3 DN, MCD, MGN 1433-mediated signaling 2 0.063 IgAN, LN, MCD PI3K signaling in B lymphocytes 1 0.043 DN, LN, MCD PXR/RXR activation 1 0.034 DN, HTN, MCD, MGN NRF2-mediated oxidative stress response 1 0.035 DN, FSGS, IgAN, LN, MGN Cytotoxic T lymphocyte-mediated

apoptosis of target cells1 0.03

2568 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

were selected representing the following nine forms of CKD: FSGS

(n=16), MCD (n=12), MGN (n=18), DN (n=17), IgAN (n=24), LN

(n=30), HN (n=20), thin-membrane disease (n=6), and tumor ne-

phrectomy (n=4) as well as controls (LD; n=10). Biopsy specimens

were obtained from patients after informed consent and with the

approval of the respective local ethics committees. Clinical and his-

tologic characteristics of the subjects are summarized in Table 2.

Manual microdissection of glomeruli and tubulointerstitial compart-

ments was performed as previously described.40

RNA Isolation, RNA Preparation, and Microarray DataAnalysesTotal RNA from the microdissected glomerular and tubulointerstitial

compartments was isolated, reverse transcribed, linearly amplified,

and hybridized on Affymetrix HG-U133A and HG-U133Plus2

microarrays.40,41 Affymetrix CEL files were normalized using the ro-

bust multi-array average method and annotated at EntrezGene level

(HumanEntrez Gene custom CDF annotation, version 10; http://

brainarray.mbni.med.umich.edu/Brainarray/default.asp), corrected

for batch effect,42 and log2 transformed through the GenePattern

pipeline (http://www.GenePattern.org). Underexpressed genes were

filtered by applying a cutoff of the median of the non-human

Affymetrix control probe sets plus two times their SD.43 Gene expres-

sion datasets are available as CEL files in the Gene Expression Om-

nibus (submission no. GSE47183). In addition, expression datasets

have been annotated with clinical information and uploaded into the

renal systems biology search engineNephromine39 to enable individual-

and system-level analysis.

Correlation to eGFRCandidate genes were prioritized by correlation of their expression

profiles with log eGFR. In analogy of the analysis performed by

CKDGen, eGFRwas estimated using theModification ofDiet in Renal

Disease equation. For 29 of 40 CKD candidate genes, renal gene

expression profiles above background were available; the other 11

geneswere not included on theAffymetrix 133A expression platform.

Pearson correlation of gene expression to log eGFR was calculated for

these 29 candidate genes across all eight diseases as well as the LD

controls todefine the linear relationshipsbetweengene expressionand

log eGFR. Correlationswith |r|$0.25 and FDR#0.01were considered

significant and calculated for both compartments, every disease in-

dividually, and all diseases together (Figure 2). Random association of

transcripts with log eGFR was tested over 10 random gene sets of size

identical to the CKD candidate gene set (n=40) generated using the

above threshold parameters and statistically assessed by z score.

Gene CoexpressionA genome-wide expression correlation analysis detected transcripts

with expression profiles that are similar to those of the CKDGen

candidate genes (Figure 1, step 2). The goal of this step was to link

specific transcripts by their coregulated gene set to defined molecular

functions in CKD. This concept is on the basis of the observation that

mRNA coexpression patterns can result from common upstream

transcriptional regulatory events that determine the transcriptional

response in the functional context of a disease.43,44 Pearson r

correlation of tubulointerstitial compartment expression of each can-

didate gene was computed with the CKD gene expression dataset.

Correlations with |r|$0.5 and FDR#0.01 were considered signifi-

cant.

Analysis of Canonical PathwaysCandidate-correlated transcripts were analyzed for significant

(P,0.05) enrichment of canonical pathways and functional group-

ings using the Ingenuity Pathway Analysis Software Suite. Pathway

analysis was conducted separately for each list of correlated tran-

scripts expanded from individual CKD candidate genes (Figure 1,

step 3). FBXO22, ALMS1, PRUNE, and PIP5K1B were not included,

because each had five or fewer correlated transcripts, limiting the

ability to identify associated molecular pathways.45

Pathway Network, Generation, and AnalysesA network of all pathways significantly enriched in at least two sets of

candidate gene-correlated transcripts was constructed (Figure 1, step 4).

In this network, a node represents a pathway, and an edge between two

pathways represents the occurrence of one or more transcripts both

correlated with two or more candidate genes and shared between these

two pathways. The network graph was visualized in Cytoscape (http://

www.cytoscape.org/) using a spring-embedded layout (Figure 3).

After the generation of the network, analysis focused on the

pathways of the CKD network, the connecting genes of this network,

and the network structure itself. First, the pathway network was

evaluated for pathways with known CKD associations using the

literature mining tool LitInspector (Genomatix) followed by verifi-

cation in PubMed. For each result, an exemplar citation as well as a

brief descriptionof the proposedmechanismof action inCKD is given

in Supplemental Table 1. Pathways that had not previously been im-

plicated in CKD were ranked in descending order according to the

number of original candidate genes that the pathway was associated

with in Supplemental Table 2.

Second, we tested significantly differentially regulated genes

between seven CKDs and LDs for their enrichment in any of 97

pathways in the network. This analysis identifies pathways likely to be

involved in disease-specific processes among all CKD-associated

pathways that are responsible for the development and progression

of a particular CKD.

Third, we focused on the connecting genes of the CKD pathway

network. These genes connect two or more pathways within the

network, because they are correlated with original CKD candidate

genes and also enriched in two or more pathways of the network. We

tested the hypothesis that transcripts connecting pathways in the

network play an important role in the pathogenesis of CKD. We

calculated their correlationwith log eGFR and testedwhether pathway

connecting transcripts were found among the CKD biomarkers as

described in the review by Fassett et al.6

Fourth, we constructed a protein–protein interaction network

using the Cytoscape MiMI plug-in from all 309 connecting genes

(settings: all databases, human)46 and assessed the functional con-

nections among themolecules of this network by comparing against a

randomized version of the network (random networks generated and

compared by Cytoscape Random Networks plug-in).

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2569

www.jasn.org BASIC RESEARCH

Fifth, we tested the hypothesis that biologically related pathways

aggregate in the pathway network with an unbiased cluster analysis of

the network using the Cytoscape plug-in Allegro mCode (default

settings; analysis on the complete pathway network). This cluster

analysis identifies strongly connected subnetworks, defining groups of

possibly functionally related pathways. Gene expression levels of

candidate-correlated transcripts that connect pathways within the

resulting two subclusters were then compared separately between

153 subjectswithCKDs and 32 healthy LDcontrols by SAManalysis

in Tigr MeV to identify differential regulation of genes with the

goal of detecting expression patterns reflecting established CKD

pathomechanism.

The CKD pathway network and its detailed pathway and eGFR

correlation information are available through a downloadable Cyto-

scape network file (Supplemental Material). Cytoscape is a multiplat-

form, open-source software, enabling researchers to scrutinize our

data for their molecule(s) and/or pathway(s) of interest.47

CKD Pathway ValidationPathwaynetworkswere validated in an independentCKDcohort of 43

subjects from the C-PROBE study cohort (Table 2). We followed the

same procedure starting with the initial set of candidate genes. The

overlap of pathways retrieved for each cohort was compared to assess

reproducibility of our approach.

ACKNOWLEDGMENTS

We acknowledge the great efforts of all contributors of the CKDGen

consortium. We also thank all participating centers of the European

Renal cDNA Bank-Kroener-Fresenius Biopsy Bank (ERCB-KFB) and

theClinical PhenotypingResource andBiobankCore (C-PROBE)and

their participants for their cooperation. The expert technical assis-

tanceofAnnaHenger is gratefully acknowledged.We thankChrystaC.

Lienczewski and Tennille S. Leak-Johnson for carefully proofreading

the manuscript.

This study was supported, in part, by O’Brien Renal Center Na-

tional Institutes of Health Grant P30-DK081943-01, National In-

stitutes of Health Grant R01-DK079912-01, and National Center for

Integrative Bioinformatics National Institutes of Health Grant U54-

DA021519 01A1 (to M.K.). C.A.B. was supported by Else Kröner-

Fresenius-Stiftung Grants Nr P48/08//A11/08 and 2012_A147.

Parts of this study were presented at the Keystone Meeting

“Complex Traits: Genomics and Computational Approaches” in

Beckenridge, CO on February 20–25, 2012.

The contribution of V.N. in this study is, in part, to fulfill the thesis

requirement at the Ludwig Maximilian University of Munich. M.K. is

the guarantor of this work and as such, had full access to all the data in

the study and takes responsibility for the integrity of the data and the

accuracy of the data analysis.

Activemembers of theERCB-KFBat the timeof the study:Clemens

David Cohen, Holger Schmid, Michael Fischereder, Lutz Weber,

Walter Samtleben,MatthiasKretzler, andDetlef Schlöndorff (Ludwig-

Maximilians University, Munich, Germany); Jean Daniel Sraer and

Pierre Ronco (Tenon Hospital, Paris, France); Maria Pia Rastaldi

and Giuseppe D’Amico (San Carlo Borromeo Hospital, Milan, Italy);

Peter Doran and Hugh Brady (University College Dublin, Dublin,

Ireland); Detlev Mönks and Christoph Wanner (University of

Würzburg, Germany); Andrew Rees (University of Aberdeen, UK);

Frank Strutz and Gerhard Anton Müller (University of Göttingen,

Germany); Peter Mertens and Jürgen Floege (Rheinisch-Westfälische

Technische Hochschule Aachen University, Germany); Norbert Braun

and Teut Risler (University of Tübingen, Germany); Loreto Gesualdo

and Francesco Paolo Schena (University of Bari, Italy); Jens Gerth and

Gunter Wolf (University of Jena, Germany); Rainer Oberbauer and

Dontscho Kerjaschki (University of Vienna, Austria); Bernhard Banas

and Bernhard Krämer (University of Regensburg, Germany); Moin

Saleem (University of Bristol, UK); Rudolf Wüthrich (University of

Zürich, Switzerland);HarmPeters andHans-HellmutNeumayer (Charite,

Berlin, Germany); Mohamed Daha (University of Leiden, Netherlands);

Katrin Ivens and Bernd Grabensee (University of Düsseldorf,

Germany); Francisco Mampaso (University of Madrid, Spain); Jun

Oh, Franz Schaefer, and Martin Zeier (University of Heidelberg,

Germany); Hermann-JosephGröne (GermanCancer ResearchCenter

[DKFZ], Heidelberg, Germany); Peter Gross (University of Dresden,

Germany); Giancarlo Tonolo (University of Sassari, Italy); Vladimir

Tesar (University of Prag,CzechRepublic);Harald Rupprecht (Klinikum

Bayreuth, Germany); Hermann Pavenstädt (University of Münster,

Germany); and Hans-Peter Marti (University of Bern, Switzerland).

DISCLOSURESNone.

REFERENCES

1. US Renal Data System: USRDS 2010 Annual Data Report: Atlas ofChronic Kidney Disease and End-Stage Renal Disease in the UnitedStates, Bethesda, MD, US Renal Data System, 2010

2. Covic A, Kothawala P, Bernal M, Robbins S, Chalian A, Goldsmith D:Systematic review of the evidence underlying the association betweenmineral metabolism disturbances and risk of all-cause mortality, car-diovascular mortality and cardiovascular events in chronic kidney dis-ease. Nephrol Dial Transplant 24: 1506–1523, 2009

3. Sharma S, Sirin Y, Susztak K: The story of Notch and chronic kidneydisease. Curr Opin Nephrol Hypertens 20: 56–61, 2011

4. Ly J, AlexanderM,Quaggin SE:A podocentric viewof nephrology.CurrOpin Nephrol Hypertens 13: 299–305, 2004

5. Zoja C, Abbate M, Remuzzi G: Progression of chronic kidney disease:Insights from animal models. Curr Opin Nephrol Hypertens 15: 250–257, 2006

6. Fassett RG, Venuthurupalli SK, Gobe GC, Coombes JS, Cooper MA,HoyWE: Biomarkers in chronic kidney disease: A review. Kidney Int 80:806–821, 2011

7. Köttgen A, Glazer NL, Dehghan A, Hwang SJ, Katz R, Li M, Yang Q,Gudnason V, Launer LJ, Harris TB, Smith AV, Arking DE, Astor BC,Boerwinkle E, Ehret GB, Ruczinski I, Scharpf RB, Chen YD, de Boer IH,Haritunians T, Lumley T, Sarnak M, Siscovick D, Benjamin EJ, Levy D,Upadhyay A, Aulchenko YS, Hofman A, Rivadeneira F, Uitterlinden AG,van Duijn CM, Chasman DI, Paré G, Ridker PM, Kao WH, Witteman JC,Coresh J, Shlipak MG, Fox CS: Multiple loci associated with indices ofrenal function and chronic kidney disease. Nat Genet 41: 712–717, 2009

2570 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org

8. Köttgen A, Pattaro C, Böger CA, Fuchsberger C, Olden M, Glazer NL,Parsa A, Gao X, Yang Q, Smith AV, O’Connell JR, Li M, Schmidt H,Tanaka T, Isaacs A, Ketkar S, Hwang SJ, Johnson AD, Dehghan A,Teumer A, Paré G, Atkinson EJ, Zeller T, Lohman K, Cornelis MC,Probst-Hensch NM, Kronenberg F, Tönjes A, Hayward C, Aspelund T,Eiriksdottir G, Launer LJ, Harris TB, Rampersaud E, Mitchell BD, ArkingDE, Boerwinkle E, Struchalin M, Cavalieri M, Singleton A, Giallauria F,Metter J, de Boer IH, Haritunians T, Lumley T, Siscovick D, Psaty BM,Zillikens MC, Oostra BA, Feitosa M, Province M, de Andrade M, TurnerST, Schillert A, Ziegler A, Wild PS, Schnabel RB, Wilde S, Munzel TF,Leak TS, Illig T, Klopp N,Meisinger C,Wichmann HE, KoenigW, ZgagaL, Zemunik T, Kolcic I, Minelli C, Hu FB, Johansson A, Igl W, Zaboli G,Wild SH, Wright AF, Campbell H, Ellinghaus D, Schreiber S, AulchenkoYS, Felix JF, Rivadeneira F, Uitterlinden AG, Hofman A, Imboden M,Nitsch D, Brandstätter A, Kollerits B, Kedenko L, Mägi R, Stumvoll M,Kovacs P, Boban M, Campbell S, Endlich K, Völzke H, Kroemer HK,NauckM, Völker U, PolasekO, Vitart V, Badola S, Parker AN, Ridker PM,Kardia SL, Blankenberg S, Liu Y, Curhan GC, Franke A, Rochat T,Paulweber B, Prokopenko I, Wang W, Gudnason V, Shuldiner AR,Coresh J, Schmidt R, Ferrucci L, Shlipak MG, van Duijn CM, Borecki I,Krämer BK, Rudan I, Gyllensten U, Wilson JF, Witteman JC, PramstallerPP, Rettig R, Hastie N, Chasman DI, KaoWH, Heid IM, Fox CS: New lociassociated with kidney function and chronic kidney disease.Nat Genet

42: 376–384, 20109. Rodríguez-Iturbe B, García García G: The role of tubulointerstitial in-

flammation in the progression of chronic renal failure. Nephron Clin

Pract 116: c81–c88, 201010. Leblond FA, Giroux L, Villeneuve JP, Pichette V: Decreased in vivo

metabolism of drugs in chronic renal failure. Drug Metab Dispos 28:1317–1320, 2000

11. Naud J, Michaud J, Beauchemin S, Hébert MJ, Roger M, Lefrancois S,Leblond FA, Pichette V: Effects of chronic renal failure on kidney drugtransporters and cytochrome P450 in rats. Drug Metab Dispos 39:1363–1369, 2011

12. Vaziri ND: Dyslipidemia of chronic renal failure: The nature, mecha-nisms, and potential consequences. Am J Physiol Renal Physiol 290:F262–F272, 2006

13. Song J, Singh M: How and when should interactome-derived clustersbe used to predict functional modules and protein function? Bio-

informatics 25: 3143–3150, 200914. Sauna ZE, Kimchi-Sarfaty C: Understanding the contribution of synon-

ymous mutations to human disease.Nat Rev Genet 12: 683–691, 201115. Lo HS, Wang Z, Hu Y, Yang HH, Gere S, Buetow KH, Lee MP: Allelic

variation in gene expression is common in the human genome. Ge-

nome Res 13: 1855–1862, 200316. Mohtat D, Susztak K: Fine tuning gene expression: The epigenome.

Semin Nephrol 30: 468–476, 201017. Perco P,Oberbauer R: Integrative analysis of -omics data and histologic

scoring in renal disease and transplantation: Renal histogenomics.Semin Nephrol 30: 520–530, 2010

18. Spasovski G, Ortiz A, Vanholder R, El Nahas M: Proteomics in chronickidney disease: The issues clinical nephrologists need an answer for.Proteomics Clin Appl 5: 233–240, 2011

19. Keller BJ, Martini S, Sedor JR, Kretzler M: A systems view of genetics inchronic kidney disease. Kidney Int 81: 14–21, 2012

20. Wang K, Li M, Bucan M: Pathway-based approaches for analysis of ge-nomewide association studies. Am J Hum Genet 81: 1278–1283, 2007

21. Mak RH, Cheung W, Cone RD, Marks DL: Mechanisms of disease: Cy-tokine and adipokine signaling in uremic cachexia. Nat Clin Pract

Nephrol 2: 527–534, 200622. Slee AD: Exploring metabolic dysfunction in chronic kidney disease.

Nutr Metab (Lond) 9: 36, 201223. Vondrácek J, Umannová L, Machala M: Interactions of the aryl hydro-

carbon receptor with inflammatory mediators: Beyond CYP1A regula-tion. Curr Drug Metab 12: 89–103, 2011

24. Reiff C, Delday M, Rucklidge G, Reid M, Duncan G, Wohlgemuth S,Hörmannsperger G, Loh G, Blaut M, Collie-Duguid E, Haller D, Kelly D:Balancing inflammatory, lipid, and xenobiotic signaling pathways byVSL#3, a biotherapeutic agent, in the treatment of inflammatory boweldisease. Inflamm Bowel Dis 15: 1721–1736, 2009

25. Handschin C, Podvinec M, Stöckli J, Hoffmann K, Meyer UA: Conser-vation of signaling pathways of xenobiotic-sensing orphan nuclear re-ceptors, chicken xenobiotic receptor, constitutive androstane receptor,and pregnane X receptor, from birds to humans. Mol Endocrinol 15:1571–1585, 2001

26. Schmidt JV, SuGH, Reddy JK, SimonMC, Bradfield CA: Characterizationof a murine Ahr null allele: Involvement of the Ah receptor in hepaticgrowth and development. ProcNatl Acad Sci U SA 93: 6731–6736, 1996

27. Suzuki T, Toyohara T, Akiyama Y, Takeuchi Y,Mishima E, Suzuki C, Ito S,Soga T, Abe T: Transcriptional regulation of organic anion transportingpolypeptide SLCO4C1 as a new therapeutic modality to preventchronic kidney disease. J Pharm Sci 100: 3696–3707, 2011

28. Del Vecchio L, Locatelli F, Carini M: What we know about oxidativestress in patients with chronic kidney disease on dialysis—clinical ef-fects, potential treatment, and prevention. Semin Dial 24: 56–64, 2011

29. GranataS,ZazaG,SimoneS,VillaniG,LatorreD,Pontrelli P,CarellaM,SchenaFP, Grandaliano G, Pertosa G: Mitochondrial dysregulation and oxidativestress in patients with chronic kidney disease. BMCGenomics 10: 388, 2009

30. Kuchta A, Pacanis A, Kortas-Stempak B, Cwikli�nska A, Ziętkiewicz M,Renke M, Rutkowski B: Estimation of oxidative stress markers in chronickidney disease. Kidney Blood Press Res 34: 12–19, 2011

31. Khan HY, Zubair H, Ullah MF, Ahmad A, Hadi SM: A prooxidantmechanism for the anticancer and chemopreventive properties of plantpolyphenols. Curr Drug Targets 13: 1738–1749, 2012

32. Pergola PE, Raskin P, Toto RD, Meyer CJ, Huff JW, Grossman EB,Krauth M, Ruiz S, Audhya P, Christ-Schmidt H, Wittes J, Warnock DG;BEAM Study Investigators: Bardoxolone methyl and kidney function inCKD with type 2 diabetes. N Engl J Med 365: 327–336, 2011

33. Zoja C, Benigni A, Remuzzi G: The Nrf2 pathway in the progression ofrenal disease. Nephrol Dial Transplant 29[Suppl 1]: S19–S24, 2014

34. BohleA,Mackensen-HaenS, vonGiseH,GrundKE,WehrmannM,BatzC,Bogenschütz O, Schmitt H, Nagy J, Müller C: The consequences oftubulo-interstitial changes for renal function in glomerulopathies. A mor-phometric and cytological analysis. Pathol Res Pract 186: 135–144, 1990

35. Samarakoon R, Dobberfuhl AD, Cooley C, Overstreet JM, Patel S,Goldschmeding R, Meldrum KK, Higgins PJ: Induction of renal fibroticgenes by TGF-b1 requires EGFR activation, p53 and reactive oxygenspecies. Cell Signal 25: 2198–2209, 2013

36. Berthier CC, Bethunaickan R, Gonzalez-Rivera T, Nair V, RamanujamM,Zhang W, Bottinger EP, Segerer S, Lindenmeyer M, Cohen CD,Davidson A, Kretzler M: Cross-species transcriptional network analysisdefines shared inflammatory responses in murine and human lupusnephritis. J Immunol 189: 988–1001, 2012

37. Hodgin JB,Nair V, ZhangH, RandolphA, Harris RC,Nelson RG,Weil EJ,Cavalcoli JD, Patel JM, Brosius FC 3rd, Kretzler M: Identification ofcross-species shared transcriptional networks of diabetic nephropathyin human and mouse glomeruli. Diabetes 62: 299–308, 2013

38. Martini S, Nair V, Patel SR, Eichinger F,NelsonRG,Weil EJ, PezzolesiMG,Krolewski AS, Randolph A, Keller BJ, Werner T, Kretzler M: From singlenucleotide polymorphism to transcriptional mechanism: A model forFRMD3 in diabetic nephropathy. Diabetes 62: 2605–2612, 2013

39. Martini S, Eichinger F, Nair V, Kretzler M: Defining human diabeticnephropathy on the molecular level: Integration of transcriptomicprofiles with biological knowledge. Rev Endocr Metab Disord 9: 267–274, 2008

40. Cohen CD, Frach K, Schlöndorff D, Kretzler M: Quantitative gene ex-pression analysis in renal biopsies: A novel protocol for a high-throughput multicenter application. Kidney Int 61: 133–140, 2002

41. Schmid H, Boucherot A, Yasuda Y, Henger A, Brunner B, Eichinger F,Nitsche A, Kiss E, BleichM, GröneHJ, Nelson PJ, Schlöndorff D, Cohen

J Am Soc Nephrol 25: 2559–2572, 2014 Transcriptional Networks in CKD 2571

www.jasn.org BASIC RESEARCH

CD, Kretzler M; European Renal cDNA Bank (ERCB) Consortium:Modular activation of nuclear factor-kappaB transcriptional pro-grams in human diabetic nephropathy. Diabetes 55: 2993–3003,2006

42. Johnson WE, Li C, Rabinovic A: Adjusting batch effects in microarray ex-pressiondata usingempirical Bayesmethods.Biostatistics8: 118–127, 2007

43. Thacker SG, Berthier CC, Mattinzoli D, Rastaldi MP, Kretzler M,Kaplan MJ: The detrimental effects of IFN-a on vasculogenesis inlupus are mediated by repression of IL-1 pathways: Potential role inatherogenesis and renal vascular rarefaction. J Immunol 185: 4457–4469, 2010

44. Fessele S, Maier H, Zischek C, Nelson PJ, Werner T: Regulatorycontext is a crucial part of gene function. Trends Genet 18: 60–63,2002

45. Huang W, Sherman BT, Lempicki RA: Bioinformatics enrichment tools:Paths toward the comprehensive functional analysis of large gene lists.Nucleic Acids Res 37: 1–13, 2009

46. Tarcea VG, Weymouth T, Ade A, Bookvich A, Gao J, Mahavisno V,Wright Z, Chapman A, Jayapandian M, Ozgür A, Tian Y, Cavalcoli J,Mirel B, Patel J, Radev D, Athey B, States D, Jagadish HV: Michiganmolecular interactions r2: From interacting proteins to pathways. Nu-cleic Acids Res 37: D642–D646, 2009

47. Cline MS, Smoot M, Cerami E, Kuchinsky A, Landys N, Workman C,Christmas R, Avila-Campilo I, CreechM, Gross B, Hanspers K, Isserlin R,Kelley R, Killcoyne S, Lotia S, Maere S, Morris J, Ono K, Pavlovic V, PicoAR, Vailaya A,Wang PL, Adler A, Conklin BR, Hood L, KuiperM, SanderC, Schmulevich I, Schwikowski B, Warner GJ, Ideker T, Bader GD: In-tegration of biological networks and gene expression data using Cy-toscape. Nat Protoc 2: 2366–2382, 2007

This article contains supplemental material online at http://jasn.asnjournals.org/lookup/suppl/doi:10.1681/ASN.2013080906/-/DCSupplemental.

2572 Journal of the American Society of Nephrology J Am Soc Nephrol 25: 2559–2572, 2014

BASIC RESEARCH www.jasn.org