ADPKD - Oxford Journals - Oxford University Press

Human Molecular Genetics, 2009, Vol. 18, No. 13 doi:10.1093/hmg/ddp165 Advance Access published on April 3, 2009

2328–2343

Systems biology of autosomal dominant polycystic kidney disease (ADPKD): computational identification of gene expression pathways and integrated regulatory networks Xuewen Song1, Valeria Di Giovanni2, Ning He1, Kairong Wang1, Alistair Ingram3, Norman D. Rosenblum2 and York Pei1, 1

Division of Nephrology, University Health Network and 2Program in Developmental and Stem Cell Biology, Hospital for Sick Children and 3Division of Nephrology, McMaster University, Hamilton, Ontario, Canada Received February 23, 2009; Revised and Accepted April 1, 2009

To elucidate the molecular pathways that modulate renal cyst growth in ADPKD, we performed global gene profiling on cysts of different size (50 ml, n = 3) and minimally cystic tissue (MCT, n = 5) from five PKD1 human polycystic kidneys using Affymetrix HG-U133 Plus 2.0 arrays. We used gene set enrichment analysis to identify overrepresented signaling pathways and key transcription factors (TFs) between cysts and MCT. We found down-regulation of kidney epithelial restricted genes (e.g. nephron segment-specific markers and cilia-associated cystic genes such as HNF1B, PKHD1, IFT88 and CYS1) in the renal cysts. On the other hand, PKD1 cysts displayed a rich profile of gene sets associated with renal development, mitogen-mediated proliferation, cell cycle progression, epithelial –mesenchymal transition, hypoxia, aging and immune/inflammatory responses. Notably, our data suggest that up-regulation of Wnt/betacatenin, pleiotropic growth factor/receptor tyrosine kinase (e.g. IGF/IGF1R, FGF/FGFR, EGF/EGFR, VEGF/ VEGFR), G-protein-coupled receptor (e.g. PTGER2) signaling was associated with renal cystic growth. By integrating these pathways with a number of dysregulated networks of TFs (e.g. SRF, MYC, E2F1, CREB1, LEF1, TCF7, HNF1B/ HNF1A and HNF4A), our data suggest that epithelial dedifferentiation accompanied by aberrant activation and cross-talk of specific signaling pathways may be required for PKD1 cyst growth and disease progression. Pharmacological modulation of some of these signaling pathways may provide a potential therapeutic strategy for ADPKD.

INTRODUCTION Autosomal dominant polycystic kidney disease (ADPKD) is the most common monogenic kidney disorder and is characterized by focal development of renal cysts in an age-dependent manner (1 – 4). Typically, only a few renal cysts are clinically detectable during the first three decades of life; however, by the fifth decade, tens of thousands of renal cysts of different sizes can be easily found in most patients (3,4). Progressive expansion of renal cysts during the clinical course of ADPKD leads to massive enlargement and distortion of the normal architecture of both kidneys, and ultimately, end-stage

renal disease (ESRD) in most patients by late middle age (1– 4). Mutations of PKD1 and PKD2 account for 85 and 15% of ADPKD, respectively, and together, they account for 5% of ESRD (1,2). Polycystin-1 (PC1) and polycystin-2 (PC2), the proteins encoded by PKD1 and PKD2, respectively, function as a macromolecular complex and regulate multiple signaling pathways to maintain the normal tubular structure and function (1,2). Micro-dissection studies have shown that renal cysts initially develop as tubular dilatations from any segment of the nephron—these dilated tubules eventually become cysts as a result of increased cellular proliferation and fluid secretion, and detach from their tubule of origin

*

To whom correspondence should be addressed at: Division of Nephrology, University Health Network, 8N838, 585 University Avenue, Toronto, Ontario, Canada M5G 2N2. Tel: þ1 4163404257; Fax: þ1 4163404999; Email: [email protected]

# The Author 2009. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

Human Molecular Genetics, 2009, Vol. 18, No. 13

after enlarging beyond a few millimeters in diameter (4). Recent studies have provided important insights into the molecular basis of cystogenesis in ADPKD (2,5). Monoclonal expansion of individual epithelial cells that have undergone a ‘second hit’ somatic mutation, resulting in biallelic inactivation of either PKD1 or PKD2, provides a major mechanism for focal cyst initiation (5), possibly through the loss of polycystin-mediated mechanosensory function in the primary cilium (6). Recent data from a large prospective observational study indicated that renal cysts in ADPKD expand exponentially with increasing age and patients with large polycystic kidneys are at increased risk for developing kidney failure (7). However, the key factors that modulate the growth of individual cysts in ADPKD remain incompletely understood. In this report, we used a systems biology approach to discover growth-modulating gene pathways in ADPKD. We found that human PKD1 cysts displayed loss of expression of kidney epithelial differentiation genes and up-regulation of developmental and mitogenic signaling pathways. The ligands, receptors and downstream target gene products of some of these signaling pathways represent potential therapeutic targets for ADPKD.

RESULTS Identifying samples with similar global gene expression profiles We performed cDNA microarrays to profile the expression of 47 400 transcripts and variants in 13 renal cysts of different sizes, 5 minimally cystic tissue (MCT) and 3 normal renal cortical samples. Using the top 200, 500, 1000 or 2000 most variable genes across all samples in an unsupervised hierarchical cluster analysis (8), we found all cyst samples consistently clustered as a single group, while the MCT and normal renal cortical samples clustered as a second group (Supplementary Material, Fig. S1). These results suggest that the gene expression pattern is very similar between renal cysts, albeit of different sizes, and between MCT and normal renal cortical tissue. For further statistical analysis, we used 13 cysts against 5 MCT samples. Due to small sample size (n = 3), we did not include the normal renal cortical samples for the analysis. Gene pathway analysis of PKD1 renal cysts We used gene set enrichment analysis (GSEA) (9), a novel algorithm for discovering differentially expressed gene pathways between two biological states, as the primary tool to identify gene pathways associated with renal cyst growth. We also performed significance analysis of microarrays (SAM) (10) and gene ontology (GO) analysis (11) to identify differentially expressed genes within the dysregulated gene sets. Of the 637 pathways tested, we found 212 (128 upand 84 down-regulated) pathways were dysregulated in the renal cysts (Supplementary Material, Table S1), although some pathways may be represented by multiple independent gene sets. The top 100 most up-regulated and 50 most downregulated gene sets in the renal cysts are shown in Tables 1 and 2. Similar to the gene expression results of the cpk mouse kidneys, a model of ARPKD (12), we found most

2329

(77/84) down-regulated genes sets in PKD1 renal cysts represent metabolic pathways (e.g. amino acid, fatty acid, urea cycle and ATP metabolism). In contrast, the analysis of up-regulated gene sets suggest that PKD1 renal cysts display a rich gene transcriptional profile for renal developmental processes, [Ca2+]i signaling, mitogen-mediated cell proliferation, cell cycle progression, aging, oxidative stress, genomic instability, hypoxic responses, epithelial– mesenchymal transition, angiogenesis and immune/inflammatory response. GO analysis of individual genes within the dysregulated pathways indicated that extracellular soluble signal modulators or pathway target genes, but not intracellular signal transducers, provided the major source of enrichment. Down-regulation of genes associated with specific nephron segments, renal cystic diseases and the primary cilium. Among the down-regulated pathways, we found two kidneyspecific gene sets (Table 2) containing the following marker genes of proximal tubule (ANPEP, CUBN, PCK1, FAH, SLC5A2, LRP2 and HNF1A), cortical thick ascending limb of Henle (UMOD, SLC12A1 and KCNJ1), distal convoluted tubule (SLC12A3, KLK1, KL and PVALB) and collecting ducts (AQP2, FXYD2 and SCNN1G) (13). By SAM [false discovery rate (FDR) 0.5%], we found the expression of these marker genes was greatly reduced or absent in the renal cysts compared with MCT (Fig. 1A). These data suggest that our PKD1 renal cysts have undergone de-differentiation and lost their nephron-segment specific transcriptional signatures. Another down-regulated gene set of interest was the hepatocyte nuclear factor-1b (HNF-1b) and its target genes (Fig. 1B). We found moderately reduced expression of HNF-1b, while 9 of its 15 target genes were dramatically reduced, in the renal cysts when compared with MCT. Notably, IFT88, PKHD1, TMEM27 and UMOD, which are genes associated with primary cilia and/or other renal cystic disorders, were decreased by 1.7, 15, 79 and 299-fold, respectively. On the other hand, we found no detectable NPHP1 expression in most of the cystic and control samples, while PKD2 expression was moderately increased in the renal cysts. The latter is consistent with the observation that mouse but not human PKD2 is regulated by HNF-1b (14 – 16). Additionally, SOCS3, a newly identified HNF-1b repressed gene, was greatly up-regulated in our renal cysts (17). There was no correlation between the expression levels of HNF-1b and its target genes and the cyst size (data not shown). By SAM, we found five additional genes associated with the primary cilium (down-regulated: DNAH5 and CYS1; up-regulated: PDGFRA, GLI2 and LGALS3) were differentially expressed in PKD1 renal cysts. Reactivation of signaling pathways for renal development and repair. We found multiple up-regulated gene sets in PKD1 renal cysts associated with kidney development and regeneration (18 – 20), including Wnt, Notch and BMP/TGFb/ Activin signaling (Table 1 and Fig. 2). Our previous data, which strongly implicated the activation of the canonical WNT/b-catenin pathway in PKD1 renal cysts, have been detailed in a recent publication and will not be reported here (21). Up-regulation of the sonic hedgehog (SHH) pathway has been recently reported to cause ciliopathy and renal

2330


Table 1. Top 100 most up-regulated gene sets in PKD1 renal cysts Up-regulated Proliferation (4) RIBOSOMAL_PROTEINS (n = 94) HSA03010_RIBOSOME (n = 62) SMITH_HTERT_UP (n = 128) TRANSLATION_FACTORS (n = 46) Mitogenic signaling pathways via RTKs, GPCRs and ECM/integrins (25) Growth factors/RTK signaling (13) IGF1_NIH3T3_UP (n = 35) SIG_INSULIN_RECEPTOR_PATHWAY_IN_CARDIAC_MYOCYTES (n = 51) SA_TRKA_RECEPTOR (n = 16) VEGFPATHWAY (n = 27) HDACPATHWAY (n = 29) ERK5PATHWAY (n = 17) NGFPATHWAY (n = 19) EGFPATHWAY (n = 27) IGF1PATHWAY (n = 20) EGF_HDMEC_UP (n = 41) IGF1MTORPATHWAY (n = 20) PDGFPATHWAY (n = 27) IGF1RPATHWAY (n = 15) ECM/integrins (6) HSA04510_FOCAL_ADHESION (n=190) HSA04512_ECM_RECEPTOR_INTERACTION (n = 81) ACTINYPATHWAY (n = 17) INTEGRINPATHWAY (n = 34) ECMPATHWAY (n = 21) RUIZ_TENASCIN_TARGETS (n =77) GPCR signaling (1) SIG_CHEMOTAXIS (n = 44) Others (5) ST_PHOSPHOINOSITIDE_3_KINASE_PATHWAY (n = 35) RASPATHWAY (n = 22) FCER1PATHWAY (n = 37) MAPKPATHWAY (n = 84) SIG_PIP3_SIGNALING_IN_CARDIAC_MYOCTES (n = 66) Cell cycle (8) IGLESIAS_E2FMINUS_UP (n = 147) RACCYCDPATHWAY (n = 22) HSA04110_CELL_CYCLE (n = 109) VERNELL_PRB_CLSTR1 (n = 67) SA_G1_AND_S_PHASES (n = 14) CELLCYCLEPATHWAY (n = 22) P27PATHWAY (n = 13) LEE_E2F1_UP (n = 60) Wnt pathway (7) ST_WNT_BETA_CATENIN_PATHWAY (n = 31) HSA04310_WNT_SIGNALING_PATHWAY (n = 141) WNT_SIGNALING (n = 58) WNT_TARGETS (n = 43) GSK3PATHWAY (n = 26) CTNNB1_ONCOGENIC_SIGNATURE (n = 84) PITX2PATHWAY (n = 16) Notch pathway (2) NGUYEN_KERATO_DN (n = 79) HSA04330_NOTCH_SIGNALING_PATHWAY (n = 43) BMP/TGFb/Activin pathway (10) TGFBETA_ALL_UP (n = 79) TGFBETA_EARLY_UP (n = 46) HSA04350_TGF_BETA_SIGNALING_PATHWAY (n = 82) TGFBETA_C3_UP (n = 12) TGFBETA_C1_UP (n = 17) ALKPATHWAY (n = 33) TGFBPATHWAY (n = 14) TGFBETA_LATE_UP (n = 33) TGFBETA_C2_UP (n = 17) TGFBETA_C4_UP (n = 11) Epithelial–mesenchymal transition (2)

NES

NOM P-value

FDR q-value

Ranking

3.29 3.21 2.05 1.63

,0.001 ,0.001 ,0.001 0.004

,0.001 ,0.001 0.002 0.055

1 2 13 85

2.14 1.99 1.73 1.7 1.67 1.66 1.63 1.63 1.62 1.62 1.62 1.62 1.61

,0.001 ,0.001 0.011 0.004 0.002 0.01 0.016 0.014 0.025 0.005 0.014 0.006 0.01

0.001 0.004 0.031 0.037 0.045 0.046 0.055 0.055 0.055 0.056 0.056 0.056 0.055

7 16 56 65 73 77 84 87 89 91 92 93 98

1.91 1.79 1.77 1.67 1.63 1.62

,0.001 ,0.001 0.006 0.015 0.022 0.005

0.009 0.024 0.024 0.045 0.055 0.055

26 40 49 74 83 90

1.62

0.009

0.055

88

1.79 1.75 1.71 1.69 1.68

0.008 0.004 0.007 ,0.001 ,0.001

0.024 0.025 0.035 0.041 0.043

41 55 62 67 71

2.05 2.02 1.86 1.83 1.68 1.65 1.61 1.6

,0.001 ,0.001 ,0.001 ,0.001 0.021 0.014 0.023 0.007

0.002 0.003 0.014 0.019 0.042 0.048 0.055 0.059

14 15 30 34 72 81 97 100

1.94 1.85 1.76 1.72 1.72 1.68 1.61

,0.001 ,0.001 ,0.001 0.006 0.012 ,0.001 0.022

0.007 0.015 0.025 0.033 0.033 0.043 0.055

21 32 51 60 61 69 95

2.11 2.05

,0.001 ,0.001

0.001 0.003

10 12

2.38 2.36 2.14 1.98 1.81 1.8 1.77 1.77 1.73 1.71

,0.001 ,0.001 ,0.001 ,0.001 0.002 0.004 0.004 0.004 0.004 0.012

,0.001 ,0.001 0.001 0.004 0.022 0.024 0.024 0.024 0.032 0.034

4 5 6 17 37 38 47 50 59 64 Continued


2331

Table 1. Continued Up-regulated

NES

NOM P-value

FDR q-value

Ranking

EMT_UP (n = 59) JECHLINGER_EMT_UP (n = 53) Hypoxia pathway (8) MENSE_HYPOXIA_UP (n = 107) HYPOXIA_FIBRO_UP (n = 20) HIFPATHWAY (n = 13) HIF1_TARGETS (n = 35) HYPOXIA_REG_UP (n = 38) HYPOXIA_REVIEW (n = 155) EPONFKBPATHWAY (n = 11) HYPOXIA_NORMAL_UP (n = 214) Aging (2) AGEING_KIDNEY_SPECIFIC_UP (n = 182) ZMPSTE24_KO_UP (n = 30) [Ca2+]i mediated signaling pathway (3) CALCINEURINPATHWAY (n = 18) NDKDYNAMINPATHWAY (n = 18) CDMACPATHWAY (n = 16) Genomic integrity (3) PARP_KO_UP (n = 84) DNMT1_KO_DN (n = 16) DNMT1_KO_UP (n = 70) Cytokine mediated JAK-STAT, MEK/ERK cascades (9) IL2RBPATHWAY (n = 34) TPOPATHWAY (n = 23) IL6PATHWAY (n = 20) IL3PATHWAY (n = 15) GLEEVECPATHWAY (n = 22) SIG_IL4RECEPTOR_IN_B_LYPHOCYTES (n = 27) IL22BPPATHWAY (n = 13) EPOPATHWAY (n = 19) IL10PATHWAY (n = 13) Immune and inflammatory response (3) 41BBPATHWAY (n = 18) NTHIPATHWAY (n = 21) PAR1PATHWAY (n = 19) MYC regulated genes (2) LEE_MYC_UP (n = 53) MYC_TARGETS (n = 40) Cell junctions (1) HSA04520_ADHERENS_JUNCTION (n = 74) p53 pathway (3) KANNAN_P53_UP (n = 35) PMLPATHWAY (n = 13) P53GENES_ALL (n = 16) Oxidative damage (1) ARENRF2PATHWAY (n = 14) Inhibition of matrix metalloproteinases (1) RECKPATHWAY (n = 9) Differentiation (1) ETSPATHWAY (n = 17) Nuclear receptor signaling (4) RARRXRPATHWAY (n = 15) CARM_ERPATHWAY (n = 26) GCRPATHWAY (n = 17) BREAST_CANCER_ESTROGEN_SIGNALING (n = 91) Wilms’ tumor signature (1) LI_FETAL_versus_WT_KIDNEY_DN (n = 158)

2.11 1.94

,0.001 ,0.001

0.001 0.007

9 22

2.41 1.96 1.91 1.81 1.76 1.73 1.69 1.65

,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.016 0.003

,0.001 0.006 0.009 0.022 0.025 0.031 0.041 0.047

3 20 25 36 52 58 66 80

1.96 1.79

,0.001 0.006

0.006 0.022

18 45

1.96 1.92 1.68

,0.001 0.004 0.018

0.006 0.009 0.042

19 23 68

2.12 1.9 1.79

,0.001 ,0.001 0.002

0.001 0.009 0.022

8 28 44

1.91 1.79 1.79 1.76 1.76 1.66 1.63 1.65 1.65

,0.001 ,0.001 0.004 0.006 0.01 0.01 0.012 0.014 0.019

0.009 0.023 0.023 0.025 0.025 0.047 0.055 0.048 0.048

27 43 42 54 53 76 86 79 78

1.67 1.64 1.62

0.006 0.016 0.019

0.045 0.049 0.056

75 82 94

1.92 1.68

,0.001 0.002

0.009 0.043

24 70

1.84

,0.001

0.017

33

1.82 1.8 1.6

0.002 0.002 0.016

0.019 0.024 0.058

35 39 99

1.78

0.006

0.024

46

1.73

0.01

0.031

57

1.71

0.02

0.035

63

2.08 1.89 1.85 1.61

,0.001 ,0.001 0.004 0.002

0.002 0.01 0.015 0.055

11 29 31 96

1.77

,0.001

0.023

48

NES represents the degree of enrichment of the gene set at the top or bottom of the ordered gene list. NOM P-value measures the significance of NES for a gene set by using permutation testing. The FDR is the estimated probability that a set with a given NES represents a false positive finding. ‘n’ is the number of genes in a specific gene set. The detailed description of each pathway can also be found on GSEA Molecular Signatures Database (MSigDB) website: http://www.broad.mit.edu/gsea/msigdb/ index.jsp.

2332


Table 2. Top 50 most down-regulated gene sets in PKD1 renal cysts Down-regulated Kidney specific genes (3) HUMAN_TISSUE_KIDNEY (n = 10) MOUSE_TISSUE_KIDNEY (n = 12) HNF1B TARGET GENES (n = 15) Aging (2) ZMPSTE24_KO_DN (n = 31) AGEING_KIDNEY_SPECIFIC_DN (n = 130) Wnt pathway (1) SANSOM_APC_4_DN (n = 67) Renin–angiotensin system (1) HSA04614_RENIN_ANGIOTENSIN_SYSTEM (n = 15) Major metabolic pathways (43) HSA00280_VALINE_LEUCINE_AND_ISOLEUCINE_DEGRADATION (n = 43) HSA00640_PROPANOATE_METABOLISM (n = 33) HSA00650_BUTANOATE_METABOLISM (n = 44) PROPANOATE_METABOLISM (n = 30) HSA00260_GLYCINE_SERINE_AND_THREONINE_METABOLISM (n = 43) HSA00190_OXIDATIVE_PHOSPHORYLATION (n = 113) HSA00310_LYSINE_DEGRADATION (n = 45) HSA00071_FATTY_ACID_METABOLISM (n = 45) ATP_SYNTHESIS (n = 20) HSA00252_ALANINE_AND_ASPARTATE_METABOLISM (n = 32) PHOTOSYNTHESIS (n = 21) OXIDATIVE_PHOSPHORYLATION (n = 57) VALINE_LEUCINE_AND_ISOLEUCINE_DEGRADATION (n = 35) HSA00620_PYRUVATE_METABOLISM (n = 42) HSA00220_UREA_CYCLE_AND_METABOLISM_OF_AMINO_GROUPS (n = 30) GLYCINE_SERINE_AND_THREONINE_METABOLISM (n = 34) HSA00410_BETA_ALANINE_METABOLISM (n = 25) HSA00020_CITRATE_CYCLE (n = 26) TRYPTOPHAN_METABOLISM (n = 51) ARGININE_AND_PROLINE_METABOLISM (n = 41) ETCPATHWAY (n = 9) BUTANOATE_METABOLISM (n = 26) PYRUVATE_METABOLISM (n = 26) GLYCOLYSIS_AND_GLUCONEOGENESIS (n = 42) HSA00330_ARGININE_AND_PROLINE_METABOLISM (n = 33) HSA00120_BILE_ACID_BIOSYNTHESIS (n = 37) HSA03320_PPAR_SIGNALING_PATHWAY (n = 66) CITRATE_CYCLE_TCA_CYCLE (n = 19) HSA00980_METABOLISM_OF_XENOBIOTICS_BY_CYTOCHROME_P450 (n = 59) BETA_ALANINE_METABOLISM (n = 26) CARBON_FIXATION (n = 20) HSA00710_CARBON_FIXATION (n = 23) HSA00903_LIMONENE_AND_PINENE_DEGRADATION (n = 27) HSA00051_FRUCTOSE_AND_MANNOSE_METABOLISM (n = 42) KREBS_TCA_CYCLE (n = 29) HSA00010_GLYCOLYSIS_AND_GLUCONEOGENESIS (n = 62) IRINOTECAN_PATHWAY_PHARMGKB (n = 9) HSA00480_GLUTATHIONE_METABOLISM (n = 36) LYSINE_DEGRADATION (n = 29) HSA00590_ARACHIDONIC_ACID_METABOLISM (n = 50) UREA_CYCLE_AND_METABOLISM_OF_AMINO_GROUPS (n = 17) GAMMA_HEXACHLOROCYCLOHEXANE_DEGRADATION (n = 26) ANDROGEN_AND_ESTROGEN_METABOLISM (n = 23)

NES

NOM P-value

FDR q-value

Ranking

22.26 22.08 22.09

,0.001 ,0.001 ,0.001

,0.001 0.001 0.001

18 36 35

22.19 22.17

,0.001 ,0.001

,0.001 ,0.001

22 26

22.18

,0.001

,0.001

23

21.97

,0.001

0.003

48

22.67 22.65 22.53 22.49 22.49 22.47 22.46 22.44 22.42 22.42 22.40 22.40 22.38 22.35 22.32 22.27 22.26 22.26 22.25 22.19 22.17 22.17 22.16 22.16 22.15 22.15 22.14 22.14 22.13 22.10 22.07 22.05 22.02 22.01 22.00 22.00 21.99 21.99 21.98 21.97 21.97 21.92 21.91

,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.002 ,0.001

,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.001 0.001 0.002 0.002 0.002 0.002 0.002 0.002 0.003 0.003 0.003 0.006 0.006

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 19 20 21 24 25 27 28 29 30 31 32 33 34 37 38 39 40 41 42 43 44 45 46 47 49 50

NES represents the degree of enrichment of the gene set at the top or bottom of the ordered gene list. NOM P-value measures the significance of NES for a gene set by using permutation testing. The FDR is the estimated probability that a set with a given NES represents a false positive finding. ‘n’ is the number of genes in a specific gene set. The detailed description of each pathway can also be found on GSEA Molecular Signatures Database (MSigDB) website: http://www.broad.mit.edu/gsea/msigdb/ index.jsp.

cystic disease (22). Although we found no definitive enrichment of this pathway (NOM P-value = 0.175), a number of its core signaling genes were up-regulated in the renal cysts, including PTCH1 (2.3-fold), GLI2 (4.6-fold) and multiple

SHH target genes of the WNT and BMP families. Notably, the membrane protein GAS1 (12-fold), a SHH regulator (23), was up-regulated in our renal cysts (Fig. 2). By combining SAM (FDR 0.5%) and GO analysis, we found


2333

Figure 1. (A) Loss of expression of tubular marker genes suggests that PKD1 renal cyst epithelial cells have undergone de-differentiation from the nephron segment-specific phenotype. (B) Dysregulation of genes associated with primary cilia (denoted by *) and HNF1b target genes in PKD1 renal cyst. All genes listed in the panels were differentially expressed between the cysts and MCT samples with a false discovery rate 0.5%. Up-regulated genes are shown in red, and down-regulated genes, in green. SC, small cysts; MC, medium cysts; LC, large cysts; MCT, minimally cystic tissue; KIDNEY, normal renal cortical tissue.

up-regulation of multiple developmental genes, including secreted modulators (SLIT2, SLIT3, NOG, WNT11, BMP2, BMP4, BMP7 and GREM1), transcription factors (TFs) (GLI2, HOXA11, HOXD11, FOXC1, FOXD1, SIX1, TCF21, WT1 and ZBTB16) and integrin receptor subunit (ITGA8, SPRY1)—many of these genes are essential for ureteric bud formation, outgrowth and branching during metanephric kidney development (18 – 20). On the other hand, up-regulation of gene sets for BMP/TGFb/activin signaling (n = 10) suggests epithelial – mesenchymal transition (EMT). Taken together, these data suggest re-activation of signaling pathways for renal development and repair in PKD1 renal cysts (19). Activation of mitogenic signaling pathways. We identified 32 up-regulated gene sets associated with mitogenic signaling

(24 – 36), including those associated with growth factor/RTK signaling (e.g. IGFs, FGFs, EGF, VEGF, PDGF and NGF; n = 14), GPCR signaling (n = 1), ECM/integrin signaling (n = 9) and downstream intracellular cascades of RTK/GPCR/ integrin activation (e.g. mTOR, PI3K, RAS/RAF; n = 8) (Table 1 and Fig. 3). Of note, we found EGFR was only moderately up-regulated in the renal cysts, while its ligands (EGF/ TGFA/NRG1) and other related receptors (ERBB3/ERBB4) were down-regulated. Although only one gene set for GPCR signaling was up-regulated in the renal cysts, we identified 26 up-regulated genes associated with GPCR (UPHAR GPCR database, http://www.iuphar-db.org/) by combined SAM and GO analysis (FDR0.5%). Of interest, PLA2G2A (30.4-fold), PTGS2 (COX2, 8.2-fold) and PTGER2 (3.7-fold) were all up-regulated, while PTGER3 (23.5-fold) was down-regulated. Increased [cAMP]i has been shown to

2334


Figure 2. Schematic summary of the aberrant activation of Wnt/b-catenin, sonic hedgehog (SHH), notch and BMP/TGFb/activin signaling pathways in PKD1 renal cysts. Up-regulated pathways/genes are shown in red, and down-regulated pathways/genes in green, with mean expression fold changes in brackets. Genes that were not differentially expressed are shown in black.

promote renal cystic epithelial proliferation in ADPKD (30). Our findings suggest that increased PGE2 and PTGER2 expression in the renal cyst may stimulate [cAMP]i production through an EP2 receptor-Gs mediated mechanism (31,32). By contrast, PTGER3, which inhibits [cAMP]i through an EP3 receptor-Gi mediated mechanism, was down-regulated in the renal cysts (32). Furthermore, we observed up-regulation of adenylate cyclases (ADCY2 and ADCY3) which may increase [cAMP]i production in the renal cysts. Alteration of [Ca2+]i homeostasis is a key pathogenic feature of ADPKD and may lead to increased [cAMP]i (1,2,30). We found multiple up-regulated gene sets for calcineurin/NFAT signaling (n = 4). Of interest, both NFATC1 (2.3-fold) and NFATC4 (1.3-fold) were up-regulated in the renal cysts, which can increase PTGS2 expression and promote cellular proliferation, migration and angiogenesis (33,34). Eight gene sets for ERK, mTOR and PI3K signaling cascades were up-regulated in the renal cysts (Table 1 and Fig. 3). For the ERK pathway, we found up-regulation of RRAS (3.3-fold), MRAS (1.6-fold), RAF1 (1.3-fold), ERK1 (1.4-fold) and several ERK target genes (e.g. RPS6KA2, MYC, FOS and SRF), and down-regulation of the RAF1

inhibitor, RKIP (-2.4-fold). In contrast, no expression of BRAF was detectable in most of our cyst and control samples (30). When considering the mTOR pathway, we found up-regulation of the catalytic subunit PIK3CA (1.5-fold) and regulatory subunit PIK3R1 (4.8-fold) of PI3K, AKT3 (2.1-fold), TSC2 (1.6-fold), RHEB (1.8-fold) and p70S6K (2.2-fold). On the other hand, we also found increased expression (2.4-fold) of PTEN, which is a negative feedback regulator of PI3K-AKT signaling. Activation of angiogenic and immune/inflammatory pathways. Multiple up-regulated gene sets in the renal cysts were associated with hypoxic/angiogenic (n = 10) pathways, and immune/ inflammatory (JAK-STAT (n = 11) and NF-kappa B (n = 3) signaling) responses (Fig. 4). Specifically, we found increased expression of angiogenic factors such as IL8 (12.9-fold), VEGF (2.6-fold), NRP2 (18.9-fold), ANGPT2 (5.3-fold) and multiple other regulators of angiogenesis; cytokines such as IL6 (20-fold), IL8 (12.9-fold), CXCL1 (9.3-fold) and IL17D (8.9-fold); and 11 component or regulatory genes of the complement system.


Figure 3. Schematic summary of aberrant activation of mitogenic (RTK, GPCR and integrin receptor) signaling pathways in PKD1 renal cysts. Up-regulated pathways/genes are shown in red, and downregulated pathways/genes in green, with mean expression fold changes in brackets. Genes that were not differentially expressed are shown in black.

2335

2336


Figure 4. Schematic summary of aberrant activation of cytokine, complement and angiogenic signaling pathways in PKD1 renal cysts. Up-regulated pathways/ genes are shown in red, and down-regulated pathways/genes in green, with mean expression fold changes in brackets. Genes that were not differentially expressed are shown in black.

Alteration of pathways for aging, oxidative stress and genome stability. Increased oxidative stress during aging can promote the somatic ‘second-hit’ mutations of an ADPKD gene (3,5), and the loss of PC1 has been recently shown to cause genomic instability (37). We found multiple dysregulated pathways for aging (n = 5), oxidative stress (n = 1), genome integrity (n = 3) and the p53 signaling (n = 3) (Supplementary Material, Table S1) in PKD1 renal cysts. In this context, it is interesting to note that the expression of Klotho (KL), an anti-aging protein that can down-regulate both IGF-1 and Wnt signaling (38– 40), was significantly decreased in our PKD1 renal cysts (by 6.7-fold compared with MCT and 18-fold compared normal renal cortical samples; FDR 0.5% by SAM). Concurrently, both IGF-1 and Wnt signaling pathways were up-regulated in PKD1 renal cysts (Table 1 and Figs 2 and 3). Transcriptional factor (TF) analysis of PKD1 renal cysts Computational analysis of promoters of co-expressed genes can provide additional evidence for the regulation of a gene set by specific TFs (41). Using GSEA, we searched for overrepresented TF promoter binding motifs among differentially expressed genes. Of 615 gene sets with shared TF binding motifs tested, we found 236 (225 up- and 11 down-regulated;

194 assigned to known TFs) were dysregulated in PKD1 renal cysts (Supplementary Material, Table S2). Table 3 lists all the dysregulated gene sets with shared TF binding sites for renal development, mitogen-mediated proliferation, cell cycle, epithelial – mesenchymal transition, angiogenesis and immune/ inflammatory response. Overall, there is excellent concordance of our results between the pathway and TF analysis. Of note, multiple gene sets could be assigned to the same TF and significant overlap of genes was observed between the gene sets belonging to the same TF family. Different degrees of overlap of genes were also seen between gene sets belonging to the different TF families since transcriptional regulation often results from combinatorial interaction between multiple TFs. We found that only 11 gene sets with shared TF binding motifs were down-regulated in the renal cysts and 7 of them were assigned to the HNF family of TFs. Specifically, four genes sets were assigned to HNF1A or HNF1B, but our in silico analysis could not differentiate between them because they share a high level of sequence identity for the DNA binding domains and recognize the same DNA sequence. Three other down-regulated gene sets were assigned to HNF4A, a direct target gene of HNF1B and a major activator of HNF1A. In contrast, we found multiple up-regulated gene sets with shared TF promoter binding motifs associated with mitogenic signaling (n = 22)


2337

Table 3. Dysregulated gene sets with shared TF binding sites for renal development, mitogen-mediated proliferation, cell cycle, epithelial– mesenchymal transition, angiogenesis and immune/inflammatory response (P 0.01 and FDR 0.25) Up-regulated gene sets (TRANSFAC matrices) Mitogenic signaling pathways (22) V$SRF_C, V$SRF_Q6, V$SRF_Q4, V$SRF_Q5_01 V$SRF_01, CCAWWNAAGG_V$SRF_Q4 (n = 6) V$ELK1_01 $RREB1_01 V$MYCMAX_03 V$MAX_01 V$EGR1_01 $EGR_Q6 V$NGFIC_01 V$CREB_01, V$CREB_02, V$CREB_Q2, V$CREB_Q4 (n = 4) V$P300_01 V$NFAT_Q6, V$NFAT_Q4_01 (n = 2) V$AP1_Q6 V$POU3F2_02 Cell cycle (28) V$E2F1_Q4_01, V$E2F1_Q6_01, V$E2F1_Q3, V$E2F1_Q3_01 V$E2F1_Q4, V$E2F1_Q6 (n = 6) V$E2F1DP1RB_01 V$E2F1DP1_01 V$E2F1DP2_01, SGCGSSAAA_V$E2F1DP2_01 (n = 2) V$E2F4DP1_01 V$E2F4DP2_01 V$E2F_Q4, V$E2F_Q6, V$E2F_02, V$E2F_03, V$E2F_Q3_01 V$E2F_Q2, V$E2F_Q4_01, V$E2F_Q6_01, V$E2F_Q3 (n = 9) V$E4F1_Q6 V$CDC5_01 V$FOXO1_01, V$FOXO1_02 (n = 2) V$FOXO4_01, V$FOXO4_02 (n = 2) V$FOXO3_01 Wnt signal transduction pathway (2) V$LEF1_Q6 V$TCF1P_Q6 BMP/TGFb signaling pathway (3) V$SMAD_Q6 V$SMAD4_Q6 V$SMAD3_Q6 Hypoxia pathway (2) V$HIF1_Q5, V$HIF1_Q3 (n = 2) Immune/inflammatory responses (13) V$AREB6_04 V$CEBPB_02, V$CEBPB_01 (n = 2) V$CEBPDELTA_Q6 V$CEBPGAMMA_Q6 V$E4BP4_01 V$CMYB_01 V$HEB_Q6 V$IRF7_01 V$EFC_Q6 V$CREL_01 V$NFKAPPAB65_01 V$IK1_01 Down-regulated gene sets (TRANSFAC matrices) V$HNF1_C, V$HNF1_01, V$HNF1_Q6 RGTTAMWNATT_V$HNF1_01(n = 4) V$HNF4_Q6, V$HNF4_0, V$HNF4_01_B (n = 3)

TF with fold change A: absent; NC: no change

Functions

SRF (2.6)

MEK/ERK cascade, induce FOS expression

ELK1 (NC) RREB1 (NC) MYC (8.5)/MAX (1.3) MAX (1.3) EGR1 (5.2) EGR1 (5.2)/EGR2 (5)/EGR3 (11.3) EGR4 (A) CREB1 (1.7) EP300 (2.2) NFATC4 (1.3) JUN (6.8) POU3F2 (A) E2F1 (A) E2F1 E2F1 E2F1 E2F4 E2F4 E2Fs

RAS/RAF-mediated cell differentiation Cell proliferation, differentiation and apoptosis Mitogenesis and differentiation cAMP-mediated signaling pathway cAMP and HIF1A-mediated pathway Ca2+ mediated signaling pathway

Prime regulator of cell cycle

(A)/TFDP1 (1.6)/RB1 (NC) (A)/TFDP1 (1.6) (A)/TFDP2 (21.6) (A)/TFDP1 (1.6) (A)/TFDP2 (21.6)

E4F1 (1.6) CDC5L (1.5) FOXO1 (2.1) FOXO4 (NC) FOXO3A (1.6)

Cell cycle G2/M progression Negative regulation of the cell cycle

LEF1 (1.6) TCF7 (NC)

Wnt signaling pathway

SMAD1-4 SMAD4 (1.5) SMAD3 (1.4)

BMP/TGFb Signaling Pathway

HIF1A (NC)

Homeostatic responses to hypoxia

ZEB1 (5.3) CEBPB (4.4) CEBPD (3.4) CEBPG (1.7) NFIL3 (3.9) MYB (A) TCF12 (1.3) IRF7 (NC) RFX1 (1.2) REL (NC) RELA (2.2) IKZF1 (NC) HNF1A (21.4) or HNF1B (21.7) HNF4A (25.4)

Tissue-specific TFs, regulates expression of kidney, liver genes Controls the expression of HNF1A

‘n’ is the number of dysregulated gene sets assigned to the same TF.

and cell cycle progression (n = 28). Of 225 up-regulated gene sets (Supplementary Material, Table S2), the top five gene sets all shared TF promoter binding motifs to serum response factor (SRF) and function in the Ras/MEK/ERK cascade. Consistent with our earlier results from the pathway analysis, we

also found four up-regulated gene sets with shared cAMP response element binding (CREB1) motif. On the other hand, enrichment of up-regulated gene sets with shared E2F family promoter binding motifs (n = 21) suggests activation of the Rb/E2F pathway, which promotes G1 to S phase pro-

2338


gression in the cell cycle (42). Consistent with our earlier results from the pathway analysis, the remaining up-regulated gene sets from Table 3 suggests activation of the signaling pathways for Wnt/b-catenin, BMP/TGFb, hypoxic/angiogenic and immune/inflammatory responses in PKD1 renal cysts. Validation of selected differentially expressed genes by qPCR Using real time RT – PCR in an expanded set of cyst and MCT samples derived from the same PKD1 kidneys, as well as normal renal cortical samples, we independently confirmed the microarray results on following genes of interest: KL (inhibitor of Wnt and IGF1R signaling); IGF1, IGF1R and EGFR (growth factor/RTK pathway); RAF1 and BRAF (RAS/RAF cascade); PTGS2, PLA2G2A, PTGER2 and PTGER3 (PGE2/PTGER2 pathway) (Fig. 5).

DISCUSSION In this study, we have generated a comprehensive map of the key signaling pathways and TFs associated with PKD1 renal cysts. We found a high level of complexity in the molecular pathobiology of human ADPKD: PKD1 renal cysts displayed signature transcriptional profiles characteristic of biological processes such as renal tubular dedifferentiation (downregulation of nephron segment-specific genes and reactivation of developmental pathways); cell proliferation and migration; tissue remodeling; angiogenesis; and activation of immune/ inflammatory cascades. At the same time, our data also support previous in vitro studies suggesting that aberrant activation of WNT b-catenin (21), EGF (2), IGF1 (26), ERK (26,30), mTOR (27), calcineurin/NFAT (33), AP1 (35) and JAK/STAT (36) may modulate renal cyst growth in ADPKD. The up-regulation of these pathways in renal cysts of different sizes suggests that they are associated with all stages of cyst expansion. An interesting observation of our study is the downregulation of genes associated with ciliary function or renal cystic diseases, including HNF1b (mature onset diabetes and renal cysts), UMOD (medullary cystic kidney disease type 2), PKHD1 (ARPKD), IFT88 (mouse orpk cystic disease) and CYS1 (mouse cpk cystic disease). HNF1b is a TF that regulates gene expression of kidney, liver, pancreas and other epithelial organs (14 – 16). Renal-specific inactivation of HNF1b in mice resulted in polycystic kidney disease and down-regulated several cystic disease genes localized to the primary cilia, including Umod, Pkhd1 and Ift88 (14). In vivo chromatin immunoprecipitation studies further demonstrated that HNF1b binds to putative cis-regulatory elements of the above target genes (14). Taken together, these data suggest a direct transcriptional hierarchy in which HNF1b regulates the other cystic genes although the biological significance of this finding remains unclear at this time. Our study suggests that HNF1b may be a down-stream target regulated by PKD1. PKD1 renal cysts displayed a rich network of up-regulated signaling pathways for mitogenic responses and focal adhesion, and their RTK, GPCR and integrin-mediated activation of intracellular cascades, including PLCs/DAG/

[Ca2+]i, MAPK (e.g. ERK, JUN, p38) and PI3K/AKT (1,2). Given the complexity, redundancy and cross-talk between these pathways, therapeutic intervention targeting a single mitogenic signaling pathway (e.g. IGF-1 and EGF) is unlikely to be effective. On the other hand, therapies targeting key points of convergence of intracellular signaling cascades may prove to be useful (Fig. 2). For example, ERK is an important regulator of cell proliferation and provides a key point of convergence that integrates mitogenic signals from IGF-1 (26), EGF (1) and cAMP (30) through the Ras/B-Raf/ MEK cascade. Treatment with B-Raf or MEK inhibitors may, therefore, be useful to retard renal cyst growth in ADPKD (25). Similarly, mTOR is another key regulator of cell proliferation and angiogenesis, and provides a major node for convergence of mitogenic signals from the Ras/Raf/ MEK and PI3K/AKT cascades (1,2). Notably, rapamycin (an mTOR inhibitor) treatment was highly effective in reducing renal cystic disease in two mouse models of PKD (27) and, in retrospective studies, found to reduce renal and liver cyst size in human ADPKD (27,43). Two randomized clinical trials are currently in progress to test the safety and efficacy of mTOR inhibitors in ADPKD (1). The loss of mechanosensory function of the polycystin complex in the primary cilia is associated with altered intracellular Ca2+ homeostasis and increased [cAMP]i in renal tubular epithelial cells (1,6). Increased [cAMP]i in turn promotes renal cystic epithelial proliferation (30). A recent study has shown that pharmacological antagonism of vasopressin V2 receptor (VPV2R), which lowers renal [cAMP]i, prevented renal cyst expansion in the Pkd22/tm1Som knock-out mice, an orthologous mouse model of ADPKD (44). Taken together, these data suggest that [cAMP]i in renal tubular epithelial cells is an important therapeutic target and provides the rationale for a current clinical trial in human ADPKD. However, VPV2R is expressed only in the collecting ducts, from whence most cysts arose in the Pkd22/tm1Som knock-out mice (44). In contrast, the tubular segmental origin of cysts from human ADPKD is unknown since most appear to have undergone dedifferentiation and lost their tubular segment markers. Thus, if most of them are derived from other segments of the nephron, therapeutic VPV2R antagonism may be of limited utility. In this context, our findings of up-regulation of PLA2G2A (30.4-fold), PTGS2 (COX2, 8.2-fold) and PTGER2 (3.7-fold) are interesting (Figs 3 and 5), as they suggest increased prostaglandin PGE2 synthesis and possibly an EP2 receptor-Gs coupled mechanism of [cAMP]i production in PKD1 renal cysts (31,32). Through the EP2 receptor, PGE2 may also trans-activate the canonical Wnt/b-catenin signaling pathway (45). Similarly, PGE2 can trans-activate EGFR and stimulate down-stream intracellular signaling cascades, including MEK/ERK and PI3K/AKT (46). Further studies are needed to characterize the cellular localization of PGE2/EP2 pathway and evaluate the role of PGE2/EP2 in modulating the above mitogenic signals in PKD1 renal cysts. Therapeutic COX2 inhibition may, therefore, provide an alternative or complementary approach to lowering [cAMP]i by VPV2R antagonism. We have recently reported that PKD1 renal cysts display up-regulation of canonical WNT/b-catenin signaling, which regulates two cellular processes: calcium-dependent cell


2339

Figure 5. Real-time PCR analysis of anti-aging hormone gene, KL, growth factor/RTK signaling pathway genes (IGF1, IGF1R and EGFR), MEK/ERK cascade genes (RAF1 and BRAF) and prostaglandin biosynthesis pathway genes (PLA2G2A, PTGS2, PTGER2 and PTGER3) in an expanded sample set (cyst = 38; MCT = 16). Corresponding expression values of these genes from four normal renal cortical tissue samples were also shown to illustrate the similarity to MCT (data expressed as mean + SD; one-way ANOVA with Tukey’s multiple comparison post-test).

2340


adhesion and T-cell factor/lymphoid-enhancing factor (TCF/ LEF)-dependent transcriptional activation (21). Compelling evidence supporting a direct role of WNT/b-catenin signaling in ADPKD pathogenesis has been provided by in vivo studies performed using transgenic mice harboring mutations that result in the accumulation of b-catenin (reviewed in 21). Targeted expression of a stabilized form of b-catenin to renal epithelial cells gives rise to animals with a cystic phenotype remarkably similar to ADPKD (47). We found that nuclear localization of the C-terminal tail (CTT) of PC-1 negatively regulates b-catenin-dependent induction of TCF-mediated transcription of target genes including c-myc and that the loss of the PC-1 CTT due to PKD1 mutations may lead to inappropriate activation of WNT/b-catenin signaling (21). In this regard, we found a marked reduction of Klotho (KL) expression (especially when compared with normal renal cortical tissue) and concomitant up-regulation of IGF1 and WNT signaling in PKD1 renal cysts interesting (Figs 2, 3 and 5). KL is a single-pass trans-membrane protein that functions as a co-receptor for FGF23, a bone-derived hormone that regulates phosphate re-absorption and vitamin D biosynthesis in the kidney (38). In addition, the extracellular domain of KL is shed and secreted—this circulating form of KL functions as an anti-aging hormone in rodents and has been shown to suppress IGF-1 and WNT signaling in vivo (39,40). Renal KL mRNA expression was greatly reduced in patients with chronic renal failure (48). On the other hand, experimental over-expression of KL ameliorated both immune- and nonimmune renal injury in rodent models and mitigated mitochondrial stress (49,50). Further studies are needed to assess whether therapeutic delivery (e.g. by adenoviral mediated gene transfer) of KL to the kidney may provide a novel means to down-regulate both IGF1 and WNT signaling and improve the renal cystic disease in ADPKD. Up-regulation of multiple hypoxic/angiogenic (n = 10) pathways was documented in our PKD1 renal cysts. Specifically, we found increased expression of angiogenic factors such as IL8 (12.9-fold), VEGF (2.6-fold), NRP2 (18.9-fold), ANGPT2 (5.3-fold) and multiple other regulators of angiogenesis. Although no increased expression of both VEGFR1 and VEGFR2 was detected, the expression of NRP2, a high-affinity kinase-deficient receptor for VEGF was increased 19-fold in the PKD1 renal cysts. Recent studies have suggested an important role of angiogenesis for renal cyst growth in ADPKD. Corrosion cast studies of human ADPKD kidneys have documented that a rich network of new blood vessels overlaying on individual cystic epithelia (51). Experimental VEGF receptor inhibition with ribozymes has been recently shown to reduce the renal cystic disease in the Cy/+ rat model (52). Thus, the therapeutic anti-angiogenic treatment may provide a novel approach to renal cyst growth. Further studies are needed to evaluate the safety and efficacy of direct or indirect (such as with an mTOR inhibitor) antagonism of VEGF or its receptor(s) in ADPKD. We found that the top five most up-regulated gene sets with shared TF promoter binding motifs were all assigned to SRF, a TF that regulates immediate early genes (e.g. FOS, EGR1 and JUNB) and cytoskeleton-related genes and is induced by MEK/ERK-dependent and Rho-dependent signaling, respectively (53,54). Indeed, our analysis of the individual genes

contributing to these enriched gene sets confirmed that both immediate early genes and cytoskeleton-related genes were up-regulated in the renal cysts. SRF plays an essential role in muscle and nervous system development, but little is known about its role in renal epithelial cells. The biological significance of this finding is presently unclear. However, these data are consistent with a prominent role of Ras/MEK/ ERK dysregulation in the pathobiology of ADPKD (25,26,30) and suggest that antagonism of this pathway may be a promising therapeutic approach. Finally, there are several limitations in our study: (i) given the correlative nature of our data they alone do not prove causality between the loss of polycystin function and alteration of a specific signaling pathway; (ii) both the renal cysts and minimally cystic (MCT) control samples contained a mixture of cell types; and (iii) we only evaluated transcriptional changes and due to the limited tissue availability, were not able to assess the levels or activity of specific proteins within a dysregulated signaling pathway. Nonetheless, by providing a detailed map of the key signaling pathways and TFs that are dysregulated in PKD1 renal cysts, our database will provide an invaluable resource for researchers interested in studying molecular pathobiology of ADPKD.

MATERIALS AND METHODS Renal cyst and control tissues Renal cysts of different sizes were obtained from five polycystic kidneys removed for medical reasons. The kidneys were kept on ice immediately after nephrectomy and throughout the tissue retrieval procedure in the pathology suite, where the renal capsule was stripped and individual cysts identified. The volume of each intact cyst was determined by withdrawing all the cyst fluid into an appropriately sized syringe via a 21-gauge needle. Small cysts (SC) were defined as less than 1 ml, medium cysts (MC) between 10 and 25 ml and large cysts (LC) greater than 50 ml. MCT, which might have contained a few microscopic cysts from the renal cortex, was obtained as PKD control tissue from the same kidneys. Additionally, non-cancerous renal cortical tissue from three nephrectomized kidneys with isolated renal cell carcinoma was used as normal control tissue. All tissues were dissected within 30 min of nephrectomy, washed in cold PBS, snap-frozen in liquid N2 and stored at 2808C. For the microarray study, 13 cysts (SC: each pooled from four different SC, n = 5; MC, n = 5, and LC, n = 3), five MCT and three normal renal cortical tissue samples were used. For validation of specific genes of interest by real-time RT – PCR (qPCR), an expanded sample set (SC: n = 16; MC, n = 19; LC, n = 3; MCT, n = 16) from the above PKD kidneys was used. All the study patients were shown to have PKD1 by DNA linkage or documentation of a pathogenic mutation identified through DNA sequencing by Athena DiagnosticsTM . Informed consent was obtained from all patients and the Institutional Review Board of the hospital where the nephrectomy was performed approved the research protocol used for this study.


RNA extraction Total RNA was extracted from each sample using Absolutely RNA RT – PCR Miniprep Kit (Stratagene) with an on-column DNA digestion step to minimize genomic DNA contamination. The integrity of the RNA was assessed using the RNA 6000 Nano Assay on 2100 Bioanalyzer (Agilent Technologies) to ensure that the ratio of 28S to 18S rRNA was at least 2. As negative control for potential DNA contamination, we routinely performed 45 cycles of PCR for beta-actin pseudogenes and did not detect any amplification products. Microarray analysis of cyst and control tissues cRNA samples were prepared according to the Affymetrix Two-Cycle Target Labeling Reagents for small sample (http://www.affymetrix.com/support/). In brief, the protocol involves two rounds of linear amplification in four stages: (i) synthesis of double stranded cDNA from 50 to 100 ng total RNA; (ii) first round amplification by in vitro transcription using the double-stranded cDNA to generate cRNA; (iii) synthesis of double-stranded cDNA from the cRNA produced in the first round amplification; and (iv) second round amplification by in vitro transcription using the double-stranded cDNA to produce labeled cRNA. After fragmentation into pieces of 50– 200 bases long, 15 mg of labeled cRNA sample was hybridized onto GeneChip Human Genome U133 Plus 2.0 Array (Affymetrix), using standard protocols at the Microarray and Gene Expression Facility of the Centre for Applied Genomics in the Hospital for Sick Children, Toronto. The U133 Plus 2.0 Array comprises 54 675 probe sets, representing over 47 400 transcripts and variants, including 38 500 well-substantiated human genes. Microarray quality control parameters were as follows: noise (RawQ) less than 5, background signal less than 100 (250 targeted intensity for array scaling), consistent number of genes detected as present across arrays, consistent scale factors and consistent detection of BioB and BioC hybridization spiked controls. Microarray data analysis Data preprocessing and normalization. Scanned raw data images were processed with GeneChip Operating Software (GCOS) 1.4. Probe set signal intensities were extracted and normalized by the robust multi-array average algorithm (55), which can be found in the R package affy that can be downloaded from the Bioconductor project website (http:// www.bioconductor.org). Microarray data are MIAME compliant and available in Gene Expression Omnibus (GEO, http:// www.ncbi.nlm.nih.gov/geo/) (ID: GSE7869). Hierarchical cluster analysis. We performed hierarchical cluster analysis based on gene expression signatures from the top 200, 500, 1000 or 2000 most variable genes (with the largest CV) across all samples to define the similarity of gene expression patterns between renal cysts of varying size, MCT and normal renal cortical tissue. The correlation visualization tool displays correlations between the samples. The colors of cells relate to Pearson’s correlation coefficient values, with deeper colors indicating higher positive (red) or

2341

negative (green) correlations. Hierarchical cluster analysis is a popular unsupervised method for arranging genes and samples according to underlying similarities in patterns of gene expression. Patterns in the data are discovered solely from the data itself without prior knowledge or grouping of the data. Two-dimensional hierarchical clustering sorts both samples and genes according to similarities and leads to a tree-structured dendrogram for easy viewing. Common crossing points represent similar sample characteristics as well as similarities with regard to the co-expression of distinct genes (8). Complete hierarchical clustering was carried out using the Cluster software v2.11 (http:// rana.lbl.gov/EisenSoftware.htm). The results were analyzed and visualized with the TreeView program v1.50 (http:// rana.lbl.gov/EisenSoftware.htm). Gene set enrichment analysis. Gene expression changes do not occur as independent events but in a highly coordinated and interdependent manner. In this study, we used GSEA (http:// www.broad.mit.edu/gsea/) as the primary tool to identify potential gene pathways and key TFs that may modulate renal cyst growth (9). This novel approach determines whether a defined set of genes displays a statistically significant, concordant difference between two biological states and is more powerful than strategies that focus on identifying individual differentially expressed genes (9). It begins by ranking the genes of the entire data set according to a test statistic (e.g. SAM and Signal2Noise) and calculates a normalized enrichment score (NES) that reflects the degree to which a particular gene set is overrepresented at the extreme (top or bottom) of the entire ranked gene list. Statistical significance of the nominal P-value of the NES is then estimated using an empirical phenotype-based permutation test procedure that preserves the complex correlation structure of the gene expression data. FDR is also calculated to estimate the probability that a set with a given NES represents a false positive finding. Before running GSEA, Affymetrix probe sets were collapsed to one gene level by using the maximum expression value of the probe set in each class (from 54 675 probe sets to 20 356 genes). Un-paired SAM scores (10) were used to rank the genes. Signal2Noise ranking within the GSEA program was also performed. For pathway analysis, we selected 637 gene sets from the GSEA C2 database (curated gene sets from BioCarta, KEGG, Signaling Pathway database, Signaling Gateway, Signal Transduction Knowledge Environment, Human Protein Reference database, GenMAPP, Sigma-Aldrich Pathways, Gene Arrays, BioScience Corp., Human Cancer Genome Anatomy Consortium database) and a customized gene set (for HNF1B target genes), which include well-studied metabolic and signaling pathways and published microarray data sets. For TF analysis, we used the GSEA C3 database with 615 genes sets containing shared and evolutionarily conserved TF binding motifs defined by the TRANSFAC database. The description of each gene sets can be found on the GSEA Molecular Signatures Database website: http://www.broad.mit.edu/gsea/ msigdb/index.jsp. The results shown here used the rank list generated by SAM; however, similar results were obtained for the top-ranked gene sets regardless of which

2342


ranking-ordering protocols were used. We defined overrepresented pathways and TF binding motifs by a NOM P-value 0.05 and 0.01, respectively; both with a FDR 0.25. Analysis of individual differentially expressed genes. We used Significance Analysis of Microarrays (SAM) and GO analysis as complementary tools to identify individual differentially expressed genes within a dysregulated gene set. We used SAM (http://www-stat.stanford.edu/tibs/SAM/) with a FDR0.5% to provide a conventional measure of statistical significance for individual differentially expressed genes between classes (10). GO enrichment analysis was performed by using the Database for Annotation, Visualization and Integration Discovery (DAVID) Knowledgebase (http:// david.abcc.ncifcrf.gov/) (11). Validation of microarray results by real time PCR. Total RNA was extracted from an expanded sample set derived from the same PKD1 kidneys: renal cysts (n = 38); MCT (n = 16). Additionally, we also included four normal renal cortical RNA samples. cDNA was generated using superscript II reverse transcriptase (Invitrogen). Real time RT – PCR (qPCR) was performed with the ABI PRISM 7900 Sequence Detection System, using Power SYBRw Green PCR Master Mix (Applied Biosystems). qPCR for Wnt signaling components, including CTNNB1, CCND1, GSK3B, MYC, AXIN2 and SFRP4 has been shown in another recently published paper (21). Primer sets for KL, IGF1, IGF1R, EGFR, RAF1, BRAF, PLA2G2A, PTGS2, PTGER2 and PTGER3 were designed to the exon sequences using the Primer Express software (Applied Biosystems). A standard curve of relative concentration was calculated for each reaction run using serial dilutions of human blood genomic DNA, and absolute transcript copy numbers were calculated using standard curves for each primer set (Supplementary Material, Table S3). We used the geNorm software (http://medgen.ugent.be/~jvdesomp/ genorm/) to identify the most stable housekeeping genes (56) and a normalization factor based on the geometric mean of three housekeeping genes EEFIA1, B2M and PPIA were calculated. The resulting normalized copy numbers were calculated by dividing the raw quantities by the normalization factor. qPCR results are expressed as mean + SD. GraphPad Prism 3.0 (GraphPad Software, San Diego, CA, USA) was used for statistical analysis of the results. Comparisons between cysts and control were done by one-way ANOVA with Tukey’s multiple comparison post test.

SUPPLEMENTARY MATERIAL Supplementary Material is available at HMG online.

ACKNOWLEDGEMENTS We thank Drs Joan Krepinsky and Anil Kapoor for his help in the retrieval of polycystic kidneys, all the PKD patients for donating their nephrectomized kidneys and Dr Lucy Osborne

for access to qPCR. We also thank Dr James Scholey for his critical review of this manuscript. Conflict of Interest statement. None declared.

FUNDING This work was supported by grants from the Kidney Foundation of Canada and Canadian Institutes of Health Research (MOP 77806) to Y.P., and from the Canadian Institutes of Health Research (MOP 14406) to N.D.R.

REFERENCES 1. Harris, P.C. and Torres, V.E. (2009) Polycystic kidney disease. Annu. Rev. Med., 60, 321 –337. 2. Ong, A.C. and Harris, P.C. (2005) Molecular pathogenesis of ADPKD: the polycystin complex gets complex. Kidney Int., 67, 1234–1247. 3. Paterson, A.D., Magistroni, R., He, N., Wang, K., Johnson, A., Fain, P.R., Dicks, E., Parfrey, P., St George-Hyslop, P. and Pei, Y. (2005) Progressive loss of renal function is an age-dependent heritable trait in type 1 autosomal dominant polycystic kidney disease. J. Am. Soc. Nephrol., 16, 755 –762. 4. Grantham, J.J., Geiser, J.L. and Evan, A.P. (1987) Cyst formation and growth in autosomal dominant polycystic kidney disease. Kidney Int., 31, 1145– 1152. 5. Pei, Y. (2001) A ‘two-hit’ model of cystogenesis in autosomal dominant polycystic kidney disease? Trends Mol. Med., 7, 151–156. 6. Nauli, S.M., Rossetti, S., Kolb, R.J., Alenghat, F.J., Consugar, M.B., Harris, P.C., Ingber, D.E., Loghman-Adham, M. and Zhou, J. (2006) Loss of polycystin-1 in human cyst-lining epithelia leads to ciliary dysfunction. J. Am. Soc. Nephrol., 17, 1015–1025. 7. Grantham, J.J., Torres, V.E., Chapman, A.B., Guay-Woodford, L.M., Bae, K.T., King, B.F. Jr, Wetzel, L.H., Baumgarten, D.A., Kenney, P.J., Harris, P.C. et al. (2006) Volume progression in polycystic kidney disease. N. Engl. J. Med., 354, 2122–2130. 8. Eisen, M.B., Spellman, P.T., Brown, P.O. and Botstein, D. (1998) Cluster analysis and display of genome-wide expression patterns. Proc. Natl Acad. Sci. USA, 95, 14863–14868. 9. Subramanian, A., Tamayo, P., Mootha, V.K., Mukherjee, S., Ebert, B.L., Gillette, M.A., Paulovich, A., Pomeroy, S.L., Golub, T.R., Lander, E.S. et al. (2005) Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA, 102, 15545–15550. 10. Tusher, V.G., Tibshirani, R. and Chu, G. (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl Acad. Sci. USA, 98, 5116– 5121. 11. Dennis, G. Jr, Sherman, B.T., Hosack, D.A., Yang, J., Gao, W., Lane, H.C. and Lempicki, R.A. (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biol., 4, 3 p. 12. Mrug, M., Zhou, J., Woo, Y., Cui, X., Szalai, A.J., Novak, J., Churchill, G.A. and Guay-Woodford, L.M. (2008) Overexpression of innate immune response genes in a model of recessive polycystic kidney disease. Kidney Int., 73, 63– 76. 13. Chabardes-Garonne, D., Mejean, A., Aude, J.C., Cheval, L., Di Stefano, A., Gaillard, M.C., Imbert-Teboul, M., Wittner, M., Balian, C., Anthouard, V. et al. (2003) A panoramic view of gene expression in the human kidney. Proc. Natl Acad. Sci. USA, 100, 13710–13715. 14. Gresh, L., Fischer, E., Reimann, A., Tanguy, M., Garbay, S., Shao, X., Hiesberger, T., Fiette, L., Igarashi, P., Yaniv, M. et al. (2004) A transcriptional network in polycystic kidney disease. EMBO J., 23, 1657– 1668. 15. Senkel, S., Lucas, B., Klein-Hitpass, L. and Ryffel, G.U. (2005) Identification of target genes of the transcription factor HNF1beta and HNF1alpha in a human embryonic kidney cell line. Biochem. Biophys. Acta, 1731, 179–190. 16. Rebouissou, S., Vasiliu, V., Thomas, C., Bellanne-Chantelot, C., Bui, H., Chretien, Y., Timsit, J., Rosty, C., Laurent-Puig, P., Chauveau, D. et al.


17.

18.

19. 20. 21.

22. 23. 24. 25.

26.

27.

28. 29. 30.

31.

32. 33.

34. 35.

36.

37.

(2005) Germline hepatocyte nuclear factor 1alpha and 1beta mutations in renal cell carcinomas. Hum. Mol. Genet., 14, 603 –614. Ma, Z., Gong, Y., Patel, V., Karner, C.M., Fischer, E., Hiesberger, T., Carroll, T.J., Pontoglio, M. and Igarashi, P. (2007) Mutations of HNF-1beta inhibit epithelial morphogenesis through dysregulation of SOCS-3. Proc. Natl Acad. Sci. USA, 104, 20386– 20391. Roxburgh, S.A., Murphy, M., Pollock, C.A. and Brazil, D.P. (2006) Recapitulation of embryological programmes in renal fibrosis—the importance of epithelial cell plasticity and developmental genes. Nephron Physiol., 103, 139–148. El-Nahas, A.M. (2003) Plasticity of kidney cells: role in kidney remodeling and scarring. Kidney Int., 64, 1553–1563. Gill, P.S. and Rosenblum, N.D. (2006) Control of murine kidney development by sonic hedgehog and its GLI effectors. Cell Cycle, 5, 1426– 1430. Lal, M., Song, X., Pluznick, J.L., Di Giovanni, V., Merrick, D.M., Rosenblum, N.D., Chauvet, V., Gottardi, C.J., Pei, Y. and Caplan, M.J. (2008) Polycystin-1 C-terminal tail associates with beta-catenin and inhibits canonical Wnt signaling. Hum. Mol. Genet., 17, 3105–3117. Hildebrandt, F., Attanasio, M. and Otto, E. (2009) Nephronophthisis: disease mechanisms of a ciliopathy. J. Am. Soc. Nephrol., 20, 23–35. Martinelli, D.C. and Fan, C.M. (2007) The role of Gas1 in embryonic development and its implications for human disease. Cell Cycle, 6, 2650– 2655. Schlessinger, J. (2000) Cell signaling by receptor tyrosine kinases. Cell, 103, 211–225. Omori, S., Hida, M., Fujita, H., Takahashi, H., Tanimura, S., Kohno, M. and Awazu, M. (2006) Extracellular signal-regulated kinase inhibition slows disease progression in mice with polycystic kidney disease. J. Am. Soc. Nephrol., 17, 1604–1614. Parker, E., Newby, L.J., Sharpe, C.C., Rossetti, S., Streets, A.J., Harris, P.C., O’Hare, M.J. and Ong, A.C. (2007) Hyperproliferation of PKD1 cystic cells is induced by insulin-like growth factor-1 activation of the Ras/Raf signalling system. Kidney Int., 72, 157– 165. Shillingford, J.M., Murcia, N.S., Larson, C.H., Low, S.H., Hedgepeth, R., Brown, N., Flask, C.A., Novick, A.C., Goldfarb, D.A., Kramer-Zucker, A. et al. (2006) The mTOR pathway is regulated by polycystin-1, and its inhibition reverses renal cystogenesis in polycystic kidney disease. Proc. Natl Acad. Sci. USA, 103, 5466– 5471. Rozengurt, E. (2007) Mitogenic signaling pathways induced by G protein-coupled receptors. J. Cell Physiol., 213, 589– 602. Schlaepfer, D.D. and Mitra, S.K. (2004) Multiple connections link FAK to cell motility and invasion. Curr. Opin. Genet. Dev., 14, 92– 101. Yamaguchi, T., Hempson, S.J., Reif, G.A., Hedge, A.M. and Wallace, D.P. (2006) Calcium restores a normal proliferation phenotype in human polycystic kidney disease epithelial cells. J. Am. Soc. Nephrol., 17, 178–187. Elberg, G., Elberg, D., Lewis, T.V., Guruswamy, S., Chen, L., Logan, C.J., Chan, M.D. and Turman, M.A. (2007) EP2 receptor mediates PGE2-induced cystogenesis of human renal epithelial cells. Am. J. Physiol. Renal Physiol., 293, F1622– F1632. Hao, C.M. and Breyer, M.D. (2008) Physiological regulation of prostaglandins in the kidney. Annu. Rev. Physiol., 70, 357–377. Puri, S., Magenheimer, B.S., Maser, R.L., Ryan, E.M., Zien, C.A., Walker, D.D., Wallace, D.P., Hempson, S.J. and Calvet, J.P. (2004) Polycystin-1 activates the calcineurin/NFAT (nuclear factor of activated T-cells) signaling pathway. J. Biol. Chem., 279, 55455– 55464. Buchholz, M. and Ellenrieder, V. (2007) An emerging role for Ca2+/ calcineurin/NFAT signaling in cancerogenesis. Cell Cycle, 6, 16–19. Le, N.H., van der Wal, A., van der Bent, P., Lantinga-van Leeuwen, I.S., Breuning, M.H., van Dam, H., de Heer, E. and Peters, D.J. (2005) Increased activity of activator protein-1 transcription factor components ATF2, c-Jun, and c-Fos in human and mouse autosomal dominant polycystic kidney disease. J. Am. Soc. Nephrol., 16, 2724– 2731. Bhunia, A.K., Piontek, K., Boletta, A., Liu, L., Qian, F., Xu, P.N., Germino, F.J. and Germino, G.G. (2002) PKD1 induces p21(waf1) and regulation of the cell cycle via direct activation of the JAK-STAT signaling pathway in a process requiring PKD2. Cell, 109, 157–168. Battini, L., Macip, S., Fedorova, E., Dikman, S., Somlo, S., Montagna, C. and Gusella, G.L. (2008) Loss of polycystin-1 causes centrosome amplification and genomic instability. Hum. Mol. Genet., 17, 2819–2833.

2343

38. Kuro-o, M. (2008) Klotho as a regulator of oxidative stress and senescence. Biol. Chem., 389, 233–241. 39. Kurosu, H., Yamamoto, M., Clark, J.D., Pastor, J.V., Nandi, A., Gurnani, P., McGuinness, O.P., Chikuda, H., Yamaguchi, M., Kawaguchi, H. et al. (2005) Suppression of aging in mice by the hormone Klotho. Science, 309, 1829–1833. 40. Liu, H., Fergusson, M.M., Castilho, R.M., Liu, J., Cao, L., Chen, J., Malide, D., Rovira, I.I., Schimel, D., Kuo, C.J. et al. (2007) Augmented Wnt signaling in a mammalian model of accelerated aging. Science, 317, 803– 806. 41. Elkon, R., Linhart, C., Sharan, R., Shamir, R. and Shiloh, Y. (2003) Genome-wide in silico identification of transcriptional regulators controlling the cell cycle in human cells. Genome Res., 13, 773 –780. 42. Bukanov, N.O., Smith, L.A., Klinger, K.W., Ledbetter, S.R. and Ibraghimov-Beskrovnaya, O. (2006) Long-lasting arrest of murine polycystic kidney disease with CDK inhibitor roscovitine. Nature, 444, 949– 952. 43. Qian, Q., Du, H., King, B.F., Kumar, S., Dean, P.G., Cosio, F.G. and Torres, V.E. (2008) Sirolimus reduces polycystic liver volume in ADPKD patients. J. Am. Soc. Nephrol., 19, 631–638. 44. Torres, V.E., Wang, X., Qian, Q., Somlo, S., Harris, P.C. and Gattone, V.H. 2nd (2004) Effective treatment of an orthologous model of autosomal dominant polycystic kidney disease. Nat. Med., 10, 363 –364. 45. Castellone, M.D., Teramoto, H. and Gutkind, J.S. (2006) Cyclooxygenase-2 and colorectal cancer chemoprevention: the beta-catenin connection. Cancer Res., 66, 11085– 11088. 46. Pai, R., Soreghan, B., Szabo, I.L., Pavelka, M., Baatar, D. and Tarnawski, A.S. (2002) Prostaglandin E2 transactivates EGF receptor: a novel mechanism for promoting colon cancer growth and gastrointestinal hypertrophy. Nat. Med., 8, 289–293. 47. Saadi-Kheddouci, S., Berrebi, D., Romagnolo, B., Cluzeaud, F., Peuchmaur, M., Kahn, A., Vandewalle, A. and Perret, C. (2001) Early development of polycystic kidney disease in transgenic mice expressing an activated mutant of the beta-catenin gene. Oncogene, 20, 5972–5981. 48. Koh, N., Fujimori, T., Nishiguchi, S., Tamori, A., Shiomi, S., Nakatani, T., Sugimura, K., Kishimoto, T., Kinoshita, S., Kuroki, T. et al. (2001) Severely reduced production of klotho in human chronic renal failure kidney. Biochem. Biophys. Res. Commun., 280, 1015–1020. 49. Haruna, Y., Kashihara, N., Satoh, M., Tomita, N., Namikoshi, T., Sasaki, T., Fujimori, T., Xie, P. and Kanwar, Y.S. (2007) Amelioration of progressive renal injury by genetic manipulation of Klotho gene. Proc. Natl Acad. Sci. USA, 104, 2331–2336. 50. Mitani, H., Ishizaka, N., Aizawa, T., Ohno, M., Usui, S., Suzuki, T., Amaki, T., Mori, I., Nakamura, Y., Sato, M. et al. (2002) In vivo klotho gene transfer ameliorates angiotensin II-induced renal damage. Hypertension, 39, 838 –843. 51. Wei, W., Popov, V., Walocha, J.A., Wen, J. and Bello-Reuss, E. (2006) Evidence of angiogenesis and microvascular regression in autosomal-dominant polycystic kidney disease kidneys: a corrosion cast study. Kidney Int., 70, 1261–1268. 52. Tao, Y., Kim, J., Yin, Y., Zafar, I., Falk, S., He, Z., Faubel, S., Schrier, R.W. and Edelstein, C.L. (2007) VEGF receptor inhibition slows the progression of polycystic kidney disease. Kidney Int., 72, 1358– 1366. 53. Chai, J. and Tarnawski, A.S. (2002) Serum response factor: discovery, biochemistry, biological roles and implications for tissue injury healing. J. Physiol. Pharmacol., 53, 147– 157. 54. Posern, G. and Treisman, R. (2006) Actin’ together: serum response factor, its cofactors and the link to signal transduction. Trends Cell Biol., 16, 588– 596. 55. Irizarry, R.A., Hobbs, B., Collin, F., Beazer-Barclay, Y.D., Antonellis, K.J., Scherf, U. and Speed, T.P. (2003) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics, 4, 249– 264. 56. Vandesompele, J., De Preter, K., Pattyn, F., Poppe, B., Van Roy, N., De Paepe, A. and Speleman, F. (2002) Accurate normalization of real-time quantitative RT–PCR data by geometric averaging of multiple internal control genes. Genome Biol., 3, RESEARCH0034.