- Email: [email protected]
Gene Expression Changes Induced by Unilateral Ureteral Obstruction in Mice
Gene Expression Changes Induced by Unilateral Ureteral Obstruction in Mice Bo Wu and James D. Brooks* From the Department of Urology, Stanford University School of Medicine, Stanford, California
Purpose: Loss of renal function is often the impetus for operative intervention in renal obstruction cases. Obstructive nephropathy is characterized by discrete morphological and physiological changes, including tubular dilatation, apoptosis and atrophy as well as interstitial cellular infiltration and progressive interstitial fibrosis. We hypothesized that gene expression alterations correlate with obstructive nephropathy and could serve as biomarkers for early intervention. Materials and Methods: C57BL/6 mice were subjected to unilateral ureteral obstruction or sham surgery at postnatal day 21. Kidneys were harvested 1, 2, 5 and 9 days postoperatively. RNA was extracted from kidneys and comprehensive gene expression profiling was performed with microarrays. IPA® pathway analysis software was used to analyze the biological function and gene networks of gene expression data. Results: Microarray analysis revealed more than 1,800 transcripts that were up-regulated or down-regulated during days 1 through 9 after obstruction, including many previously reported transcripts (FOS, CD44, CLU, SPP1 and EGF). Pathway analysis showed significant enrichment of transcripts in cell activation/ differentiation, immune/inflammatory responses, cell cycle, metabolic process and transport. Network analysis using IPA showed that transcriptional regulatory pathways involving CEBPB and HNF4A are involved in obstructive nephropathy. Conclusions: This data set provides a foundation for development of biomarkers for obstructive nephropathy. Key Words: kidney, ureteral obstruction, hydronephrosis, gene expression, mice UPPER urinary tract obstruction is a common condition in the adult and pediatric populations, although to our knowledge its true prevalence is unknown. Autopsy studies have identified hydronephrosis in 2% to 3% of adults and 2% of infants.1,2 Upper urinary tract obstruction can be partial or complete and transient, such as from a urinary calculus, or chronic. When encountered in any form, obstruction and hydronephrosis pose particular challenges to the clinician, who must decide
whether intervention is needed to relieve obstruction to protect renal function. Currently the decision to intervene is made based on imaging that shows worsening hydronephrosis or renal function loss, or serum chemistry studies that document altered renal function. However, these measures provide only an imprecise, crude or much delayed measure of renal compromise. Also, to our knowledge no marker or panel of markers has been developed for use in clinical practice.
0022-5347/12/1883-1033/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
http://dx.doi.org/10.1016/j.juro.2012.05.004 Vol. 188, 1033-1041, September 2012 RESEARCH, INC. Printed in U.S.A.
AND
Abbreviations and Acronyms CEBPB ⫽ CCAAT/enhancer binding protein  CUUO ⫽ complete unilateral ureteral obstruction EGF ⫽ epidermal growth factor HNF4A ⫽ hepatic nuclear factor 4␣ SAM ⫽ microarray significance analysis SPP1 ⫽ secreted phosphoprotein 1 Submitted for publication January 10, 2012. Study received Stanford University animal care and use committee approval. Supported by the Keith and Jan Hurlbut Research Fund. Supplementary material for this article can be obtained at http://jurology.com. * Correspondence: Department of Urology, Room S287, Stanford University School of Medicine, 300 Pasteur Dr., Stanford, California 943055118 (telephone: 650-498-4464; FAX: 650-7234200; e-mail: [email protected]).
www.jurology.com
1033
1034
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
We characterized comprehensive changes in gene expression in various contexts and used these data to provide mechanistic insights and define disease biomarkers.3–5 For single genes and proteins postCUUO expression changes occur relatively rapidly.6 Also, acute CUUO results in hydronephrosis and cortical loss within several days. In mice irreversible renal damage was documented within 3 days after obstruction.7 Since previous studies of gene expression changes after obstruction are limited in scope with only a few time points and were done in kidneys long after renal damage and hydronephrosis were established,6,8 we sought to understand transcriptional programs that are altered soon after obstruction. To identify early biomarkers of renal damage we performed gene expression profiling in mice during the first several days after ureteral obstruction. We also took advantage of contemporary bioinformatics approaches to gain functional insights into the data.
ated from 1 g total RNA in each reaction using a fluorescent linear amplification kit (Agilent Technologies) and 10.0 mM cyanine 3 or 5 labeled CTP. Labeled cRNAs were purified using the RNeasy Mini Kit and applied to the 44K Mouse Whole Genome Oligonucleotide Microarray (Agilent Technologies), which contains 41,534 independent oligonucleotides mapping to a total of 21,503 genes. Cyanine 5 labeled cRNA (825 ng) from each kidney sample was mixed with the same amount of cyanine 3 labeled cRNA generated from mouse Universal Reference RNA (Stratagene, La Jolla, California). Microarray hybridization, washing, scanning, data extraction and analysis were done as described in the manufacturer protocol.
Statistical Analysis
MATERIALS AND METHODS
Transcript levels for obstructed and sham operated control kidneys were collected from 2 to 4 animals on days 1, 2, 5 and 9 postoperatively. SAM 3.0 (www-stat.stanford.edu) was used to identify transcripts with significant differences in expression between obstructed and sham operated mice by comparing the 2 groups directly and at each time point.9 Since the 2 analyses largely overlapped, we used the transcript set from the comparison of all controls vs all obstructed animals with a false discovery rate of less than 2%.
CUUO Generation in Mice
Gene Ontology and Pathway Analysis
C57BL/6 mice (Charles River Laboratories, Wilmington, Massachusetts) were maintained and used in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council 2003). They received standard rodent chow (Zeigler Bros, Gardners, Pennsylvania) and had free access to water. On postnatal day 21 mice were anesthetized with ketamine and subjected to unilateral obstruction for 1, 2, 5 or 9 days (2 to 4 per time point). In this surgery the left ureter was ligated to generate an obstructed kidney. For each time point additional mice underwent laparotomy and dissection of the left ureter without ligation to serve as sham controls. The protocol was approved by the Stanford University animal care and use committee.
Kidney Collection and RNA Extraction After the mice were sacrificed the kidneys were rapidly excised and stored at – 80C. For RNA extraction kidneys were homogenized in TRIzol® reagent on ice. Total RNA was purified by RNeasy® Mini Kit purification according to manufacturer instructions. RNA concentration was determined by a NanoDrop® 1000 at 260 nm absorbance. All RNA samples had a 260:280 nm ratio above 1.9. RNA integrity was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, California). Only high quality RNA (28S/18S of 1.8 or greater) was used for further analysis.
Histology Samples fixed in 10% formalin were washed, dehydrated and embedded in paraffin. Sections (7 ) were collected and stained with hematoxylin and eosin.
Microarray Analysis Gene expression analysis was done using oligonucleotide arrays (part G4140-90050 v.5.7, Agilent Technologies). Briefly, fluorescently labeled cRNA targets were gener-
Transcripts that differed significantly between obstructed and sham operated mice were uploaded to the DAVID (Database for Annotation, Visualization and Integrated Discovery) Bioinformatics Database (http://david.abcc. ncifcrf.gov/tools.jsp) for functional annotation and classification based on gene ontology assignment. In the database, transcripts were assigned to predefined functional categories, such as immune response, or to subfunctional categories, such as chemotaxis. Tables 1 and 2 lists Benjamini test values. To discover gene interactions and correlation networks the transcripts were uploaded to IPA pathway analysis software. Each transcript was mapped to its corresponding gene and overlaid on a global molecular network developed from information in the Ingenuity® Knowledge Base. Significantly enriched networks and functions were identified by comparing obstruction induced gene expression changes to the total number of occurrences of these genes in all pathway annotations in the database using the right-tailed Fisher exact test. IPA calculates a score derived from a p value for each network that reflects the likelihood that the set of significant genes in a given network is enriched or the genes are found together due to random chance. A score of 2 indicates a 1% chance that the focus genes are together in a network due to chance.
RESULTS Hydronephrosis Generation and Mouse Histological Features CUUO in 21-day-old C57BL/6 mice was generated by suture ligation of the ureter. Obstruction length in kidneys increased with time and differed significantly from the contralateral side and from that in controls by day 5 after surgery (fig. 1, A). There was
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
Table 1. Function category of up-regulated genes in hydronephrotic kidney Function (term) Cell cycle/chromosome segregation: Cell cycle DNA metabolic process DNA packaging Cell stress response Organization: Extracellular matrix Pos cellular component regulation Cell growth regulation Microtubule based process Protein processing Immune ⫹ inflammatory responses: Immune response Immune response activation Chemotaxis Inflammatory response Wound response Defense response Adhesion: Cell Cell-cell Cell regulation Cell activation ⫹ differentiation: Proliferation regulation Activation Development: Vasculature Lung Hemopoietic or lymphoid organ Heart
No. Genes Fold Benjamini (%) p Value Enrichment Test
96 (12.73) 43 (5.70) 15 (1.99) 30 (3.98)
3.10E-32 2.70E-08 3.90E-05 0.00132
3.99 2.59 3.77 1.89
7E-29 4E-06 2E-03 4E-02
14 (1.86) 1.58E-04 15 (1.99) 3.06E-04
3.52 3.12
7E-03 1E-02
12 (1.59) 9.38E-04 28 (3.71) 6.24E-08 12 (1.59) 7.79E-04
3.31 3.37 3.39
3E-02 8E-06 3E-02
59 17 16 37 45 44
6.45E-15 1.96E-07 2.27E-05 5.22E-13 5.37E-12 5.59E-08
3.18 5.02 3.73 4.18 3.29 2.49
1E-12 2E-05 2E-03 1E-10 1E-09 7E-06
59 (7.82) 1.17E-11 21 (2.79) 0.00105 12 (1.59) 0.00112
2.67 2.26 3.24
2E-09 3E-02 3E-02
46 (6.10) 1.39E-06 27 (3.58) 4.46E-06
2.17 2.79
1E-04 4E-04
25 (3.32) 5.07E-05 14 (1.86) 4.08E-04 25 (3.32) 3.07E-04
2.54 3.20 2.26
3E-03 2E-02 1E-02
21 (2.79) 5.12E-04
2.39
2E-02
(7.82) (2.25) (2.12) (4.91) (5.97) (5.84)
no difference in kidney length between the control and contralateral nonobstructed kidneys from obstructed mice. Hematoxylin and eosin stained sections of obstructed kidneys revealed progressive tubular injury by postoperative day 5, as manifested by tubular dilatation, epithelial cell flattening and loss of brush border in proximal tubules (fig. 1, B). Peritubular inflammation, edema and fibrosis were also apparent histologically 5 and 9 days after obstruction, as evidenced by decreased tubular lumen size, loss of back to back tubular organization and infiltration of peritubular spaces with fibroblasts and inflammatory cells. There were no notable histological differences between obstructed and control kidneys at postoperative days 1 and 2 (fig. 1, B). Identifying Genes Modulated in Kidney After Ureteral Obstruction SAM was used to identify 857 transcripts significantly up-regulated in obstructed kidneys and 982 significantly down-regulated vs controls. Transcripts
1035
were grouped by hierarchical clustering analysis to show those with correlated patterns of expression changes with time (fig. 2, A). In many cases genes that were co-regulated shared similar functions. For example, many transcripts that were up-regulated soon after obstruction were implicated in cell activation/differentiation while genes up-regulated several days after obstruction were enriched for immune/ inflammatory responses and cell cycle regulation. Many down-regulated transcripts encoded for proteins important for metabolic processes and transport functions. Notably several transcripts known to be modulated in response to ureteral obstruction were represented in the data set, including FOS, CD44, clusterin, SPP1 and EGF (fig. 2, B).6,10–13 Significantly enriched gene ontology terms associated with each part of the dendrogram were identified using DAVID. For genes up-regulated in response to ureteral obstruction significant enrichment was seen for transcripts involved in cell cycle/chromosome segregation, organization, immune/inflammatory responses, cell adhesion, cell activation/differentiation and development (table 1). Enriched biological functions for down-regulated transcripts
Table 2. Function category of down-regulated genes in hydronephrotic kidney
Oxidation reduction Transport: Ion Transmembrane Cation Anion Carbohydrate Process: Coenzyme metabolic Fatty acid metabolic Organic acid catabolic Lipid biosynthetic Nitrogen compound biosynthetic Steroid metabolic Alditol metabolic Organic acid biosynthetic Cellular amino acid derivative metabolic Glycerol ether metabolic Neutral lipid metabolic Vitamin metabolic Organic ether metabolic Aspartate family amino acid metabolic Glycerol metabolic Pyruvate metabolic Heterocycle catabolic Carbohydrate stimulus response
No. Genes (%)
p Value
Fold Enrichment
Benjamini Test
85 (10.34)
1.57E-21
3.17
3E-18
146 58 95 45 11
(9.25) (7.06) (6.45) (2.68) (1.34)
5.68E-15 1.51E-14 5.49E-10 4.14E-08 1.15E-04
2.68 3.16 2.58 4.24 4.60
6E-12 1E-11 2E-07 9E-06 8E-03
36 29 16 29 30
(3.16) (3.53) (1.95) (3.53) (3.65)
3.80E-10 9.34E-10 5.92E-07 1.00E-05 1.08E-05
4.56 3.95 4.95 2.55 2.49
1E-07 2E-07 1E-04 1E-03 1E-03
20 8 18 28
(2.43) (0.97) (2.19) (2.19)
2.20E-05 3.92E-05 4.49E-05 4.49E-05
3.11 8.02 3.20 3.20
3E-03 4E-03 4E-03 4E-03
10 10 29 10 7
(1.22) (1.22) (1.46) (1.22) (0.85)
5.78E-05 5.78E-05 8.07E-05 9.84E-05 1.68E-04
5.57 5.57 4.36 5.22 7.98
5E-03 5E-03 6E-03 7E-03 1E-02
7 7 9 7
(0.85) (0.85) (1.09) (0.85)
2.19E-04 3.59E-04 4.11E-04 4.51E-04
7.63 7.02 4.91 6.75
1E-02 2E-02 2E-02 2E-02
1036
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
B A
Obstruction
Sham
Obstruction
Sham
P1
P1
P2 P2
P5
P5
P9 P9
Figure 1. Representative kidney morphology and histology after ureteral obstruction. A, kidneys from mice after left ureter was obstructed by ligation and after sham surgery (Sham). B, stained, formalin fixed, paraffin embedded samples of representative kidneys after ureteral obstruction. Sham operated control renal cortexes show mature glomeruli, normal tubules and few interstitial cells. Obstructed kidneys had progressive accumulation of architectural changes after obstruction. Cortex structure appeared relatively normal 1 and 2 days postoperatively but showed alterations by days 5 and 9 with narrowed glomerular spaces, fewer tubules and more interstitial cell proliferation with disrupted tubular architecture. P, postoperative day. H&E, reduced from ⫻20.
included those associated with oxidation-reduction, transport and metabolic process (table 2). To gain further insight into the regulatory networks and gene functions altered by CUUO the SAM list of transcripts was analyzed using IPA. When protein products of significant transcripts were grouped by subcellular location 124 of 574 up-regulated (22%) and 212 of 535 down-regulated (40%) gene products localized to the cytoplasm, 21% and 14% to the extracellular space, 31% and 14% to the nucleus and 21% and 25% to the plasma membrane, respectively. When grouped by molecular function, 7% of up-regulated and 5% of down-regulated genes were classified as cytokines, 13% and 31% as enzymes, 3% and 2% as growth factors, 1% and 3% as ion channels, 7% and 6% as kinases, 3% and 3% as peptidases, 2% and 2% as phosphatases, 14% and 9% as transcription regulators, 7% and 2% as transmembrane proteins, and 3% and 13% as transporters, respectively. Integrating gene function and cellular location data revealed that most up-regulated gene products localized to the nucleus and were transcription regulators or transmem-
brane proteins. In contrast, most down-regulated gene products localized to cytoplasm and functioned as enzymes, ion channels and transporters. Networks Involved in Renal Response to Ureteral Obstruction To assess the validity of our data and gain insight into the functional consequences of CUUO we used IPA to identify significantly enriched gene networks, cellular locations and gene functions. A total of 25 significantly enriched networks were generated based on a score of 3 or greater. The top network contained 34 transcripts, of which 33 were modulated significantly in response to UUO in our data set. Cognate proteins for 14 transcripts regulate or are regulated by CEBPB, of which the transcript level increased in response to obstruction (fig. 3, A). Five of the proteins related to CEBPB, including ERK, FOS, GLIPR2, CDKN2A and DGAT2, were previously implicated in ureteral obstruction or renal fibrosis. In the second network HNF4A occupied a central role. The IPA knowledge base associated HNF4A
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
A
1037
B Obstruction
Control
Obstruction
P1 P2 P5 P1 P1 P2 P5 P9 P9 P1 P1 P1 P2 P2 P2 P5 P5 P5 P9 P9 P9
P1 P2 P5 P1 P1 P2 P5 P9 P9 P1 P1 P1 P2 P2 P2 P5 P5 P5 P9 P9 P9
Control
Cell activation/ differentiation
FOS CD44 CLU SPP1
Immune/inflammatory response
EGF
1.00
0.67
0.33
0.00
0.33
0.67
Immune/inflammatory response
1.00
Organization
Cell adhesion/cell activation Immune/inflammatory response Cell activation/ differentiation Cell cycle
Immune/inflammatory response Metablic process
Transport Metablic process
Transport Metablic process
Transport Metablic process
Figure 2. Gene expression profiles of kidneys in CUUO mice and sham operated controls at postoperative days 1, 2, 5 and 9. A, heat map shows hierarchical clustering of transcript data on sham operated controls and CUUO mice. Black vertical lines indicate 5 major gene clusters enriched for gene ontology terms related to cell activation/differentiation, immune/inflammatory responses, cellular organization, cell adhesion and cell cycle in up-regulated genes, and 2 major clusters related to metabolic process and transport in down-regulated genes. Red areas represent increased expression relative to mean per gene across experiments. Green areas represent decreased expression. Color saturation corresponds to degree of expression modulation. B, CUUO resulted in significantly increased RNA expression of FOS, CD44, CLU and SPP1, and decreased EGF expression from day 1 to day 9 postoperatively.
with another 23 transcripts that we found were significantly modulated by CUUO (fig. 3, B). Previously 5 of these transcripts (AQP1, FHL1, CBR3, HMGA2 and PLA2G4A) were reported to be involved in hydronephrosis or fibrosis, suggesting that the CEBPB
and HNF4A signaling pathways have important roles in the renal response to CUUO. Also, 7 genes (CIITA, EZH2, histone H3, LGALS3, VIM, CDH16 and DAB2) that were in network 1 but not directly associated with CEBPB and 2 (HNF1A
1038
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
A
B
Figure 3. Top 2 networks of gene expression in kidney after CUUO. A, network 1. Genes or gene products (nodes) and putative biological relationships (arrows) between each gene product. Location of modulated gene cognate proteins are shown in extracellular space (top), cell membrane (light blue area), cytoplasm, nuclear membrane (pink area) and within nucleus (bottom). Transcriptional regulator CEBPB is at center of nuclear node. Red nodes indicate up-regulated transcripts. Green areas indicate down-regulation. Noncolored nodes represent genes added by IPA based on network algorithm that were not significantly altered by CUUO in microarray data. Node shapes correspond to gene product functional class. Orange lines indicate genes that interact directly with CEBPB. B, network 2. Transcriptional regulator HNF4A down-regulation was associated with modulation of many transcripts involved in normal renal tubular function, including several transmembrane transporters. Asterisk indicates gene products represented in more than 1 network.
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
and UMOD) in network 2 that did not directly interact with HNF4A were previously implicated in hydronephrosis or fibrosis. The remaining 23 networks identified by IPA as significantly modulated in response to UUO contained many transcripts and biological functions known to participate in renal damage. For instance, activation of the immune response is represented robustly in several networks. It includes transcripts involved in cellular movement, immune cell trafficking and inflammatory response. Transcripts and pathways associated with down-regulation of cell growth and proliferation were strongly represented, as were pathways associated with cell death. Also, transcripts functionally related to functions of terminally differentiated renal cells, such as molecule transport, energy production, lipid metabolism and small molecular biochemistry, were mostly down-regulated.
DISCUSSION CUUO produced significant changes in gene expression in the kidneys of mature mice. We identified more than 1,800 transcripts that were significantly altered in response to obstruction, a number significantly larger than in previous studies due to the number of samples analyzed and the number of transcripts assayed on the microarrays. Gene expression changes appeared to be time dependent, in that some transcript levels changed relative to control within 1 day after obstruction while others required several days to be significantly modulated. The number of transcripts induced by obstruction was comparable to the number suppressed. These data provide a comprehensive assessment of transcript changes induced by renal obstruction. They are a starting point to develop biomarkers correlating with renal damage in response to obstruction. Renal damage arising from obstruction depends on the degree (complete or partial) and the duration of obstruction. The renal response to obstruction also likely depends on when in life obstruction occurs, that is during gestation or in adulthood. In adult patients and animals complete obstruction can be tolerated for a short time with little apparent functional damage, provided that obstruction is relieved.7 A previous study showed that CUUO in C57BL/6 mice, which were used in our study, had irreversible renal damage with increased serum BUN after 3 days.7 In that context the gene expression changes that we observed in animals obstructed 5 or more days might reflect alterations in the kidney associated with fibrosis, permanent tubular loss and other aspects of permanent renal damage. Undoubtedly many but not all gene expression changes within the first 3 days after obstruction involve pathways that give rise to irreversible changes in
1039
the damaged kidney and could be reversed by early relief of obstruction. Those transcripts would be attractive candidate biomarkers of the renal damage induced by obstruction. Thus, understanding the biological functions of transcripts that were significantly modulated in response to obstruction is important to develop future biomarkers to predict the loss of renal function. Pathway analysis identified 25 gene networks that were significantly enriched in obstructed kidneys. The top 2 networks illustrate the potential usefulness of pathway analysis in gaining an understanding of the renal response to unilateral ureteral obstruction. The central gene in the top network was CEBPB. Through interactions with other transcriptional regulators14 CEBPB regulates genes involved in inflammation,15 muscle injury repair16 and cell proliferation.17 Genes that interact with CEBPB in the network were previously associated with hydronephrosis or fibrosis, including GLIPR-2, CDKN2A and DGAT2. GLIPR-2 up-regulation, and DGAT2 and CDKN2A down-regulation results in increased expression of genes associated with fibrosis, such as CTGF and various chemokine genes. HNF4A is a transcription factor in the nuclear hormone receptor family that is generally expressed in endoderm derived epithelial tissues and the developing metanephros.18 HNF4A expression negatively correlates with the expression of mesenchyma associated genes, eg ACTA1 and VIM.19 HNF4A knockdown using siRNA interfered with cellular organization development in the condensed mesenchyma of cultured embryonic kidneys.20 Network analysis showed that HNF4A down-regulation induced by UUO was associated with AQP1 and FHL down-regulation, and upregulation of CBR3, HMGA2 and PLA2G4A (fig. 3, B). Based on previous functional studies of these transcripts their modulation likely contributes to renal dysfunction and fibrosis.21–26 We were not the first group to use gene expression profiling to investigate the transcriptional changes induced by renal obstruction, although to our knowledge no group has comprehensively analyzed changes soon after obstruction. Silverstein et al identified 250 genes modulated in response to complete ureteral obstruction in 1-day-old rats after 4 and 12 days.6 Since the rat kidney undergoes significant development postnatally, gene expression changes in that study could differ significantly from those that we observed in adult mice. Regardless, significant overlaps were observed between that data set and ours, particularly in genes associated with the induction of the inflammatory response and stress response genes, including transforming growth factor-1, CLU, FOS, EGF, CEBPA/B, calponin and lumican. Seseke et al profiled expression changes in a rat congenital model of high grade partial renal obstruc-
1040
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
tion at days 32 to 35 postnatally.8 Microarrays were performed on pooled RNA samples from several rats and expression differences were seen in 18 transcripts. Transcripts significantly induced by obstruction included several growth factors altered in our data set (transforming growth factor-, IGF1 and IGFBP5/6) as well as several transport proteins that were down-regulated (SLC12A3 and 34A1). Cognate proteins for several genes found in our data set were also measured in the urine of patients after obstruction, including EGF,27 AMBP, CD44, CDH11, FN1, SERPINA1, SPP1, CD14, MMP9, S100A6,28 GGT,29 CXCL10 and L16, VCAM1 and COL4.30 Many transcripts in our data set could be developed as novel urinary markers of renal damage due to obstruction. We confirmed the modulation of previously identified transcripts in our study but this study is preliminary and has limitations. To our knowledge we report the largest data set after renal obstruction to date. However, more mice per time point and the inclusion of earlier time points would improve statistical power and open avenues for other analysis tools. Furthermore, acute CUUO in our model differs from clinical obstruction, which is often partial, chronic and gradual in onset. This difference could influence the number of transcripts and physiological response of the kidney. Another challenge in developing obstruction biomarkers arises when protein levels fail to correlate with transcript levels. While transcript levels of several proteins known to be modulated by obstruction were observed in our data set, additional study is
needed to test novel proteins predicted from our data set. Alternatively urinary transcript levels could be evaluated for renal damage markers, much as PCA3 has been validated as a urine based RNA biomarker for prostate cancer.
CONCLUSIONS Gene expression profiling of unilateral ureteral obstruction provides insights into the possible mechanisms underlying obstruction induced renal damage and is a potential source of renal damage biomarkers. Many genes related to cell activation/differentiation, immune/inflammatory responses, cell cycle regulation, metabolic processes and transport functions were identified that were associated with the renal damage induced by obstruction. Pathway analysis identified CEBPB and HNF4A as candidate regulators of transcripts and pathways involved in the response to kidney obstruction. Future study is necessary to verify whether CEBPB and HNF4A have important roles in early gene expression events associated with renal damage and fibrosis induced by obstruction. In addition, an ongoing study will investigate the possibility of detecting RNA species or their cognate proteins in urine after partial ureteral obstruction.
ACKNOWLEDGMENTS Rosie Nolley provided technical assistance. Microarray raw data are available at http://www.ncbi.nlm. nih.gov/geo/query/acc.cgi?acc ⫽ GSE36496 (Accession No. GSE36496).
REFERENCES 1. Lee RS, Cendron M, Kinnamon DD et al: Antenatal hydronephrosis as a predictor of postnatal outcome: a meta-analysis. Pediatrics 2006; 118: 586. 2. Siddiqui MM and McDougal WS: Urologic assessment of decreasing renal function. Med Clin North Am 2011; 95: 161. 3. Zhao H, Zongming M, Tibshirani R et al: Alteration of gene expression signatures of cortical differentiation and wound response in lethal clear cell renal cell carcinomas. PLoS One 2009; 4: e6039.
and partial androgen insensitivity syndrome. BMC Genomics 2007; 8: 376. 6. Silverstein DM, Travis BR, Thornhill BA et al: Altered expression of immune modulator and structural genes in neonatal unilateral ureteral obstruction. Kidney Int 2003; 64: 25. 7. Puri TS, Shakaib MI, Chang A et al: Chronic kidney disease induced in mice by reversible unilateral ureteral obstruction is dependent on genetic background. Am J Physiol Renal Physiol 2010; 298: F1024.
10. Florquin S and Rouschop KM: Reciprocal functions of hepatocyte growth factor and transforming growth factor-beta1 in the progression of renal diseases: a role for CD44? Kidney Int Suppl 2003; 86: S15. 11. Chevalier RL: Biomarkers of congenital obstructive nephropathy: past, present and future. J Urol 2004; 172: 852. 12. Kiley SC, Thornhill BA, Belyea BC et al: Epidermal growth factor potentiates renal cell death in hydronephrotic neonatal mice, but cell survival in rats. Kidney Int 2005; 68: 504.
4. Bhamre S, Sahoo D, Tibshirani R et al: Temporal changes in gene expression induced by sulforaphane in human prostate cancer cells. Prostate 2009; 69: 181.
8. Seseke F, Thelen P and Ringert RH: Characterization of an animal model of spontaneous congenital unilateral obstructive uropathy by cDNA microarray analysis. Eur Urol 2004; 45: 374.
13. Yoo KH, Thornhill BA, Forbes MS et al: Osteopontin regulates renal apoptosis and interstitial fibrosis in neonatal chronic unilateral ureteral obstruction. Kidney Int 2006; 70: 1735.
5. Holterhus PM, Deppe U, Werner R et al: Intrinsic androgen-dependent gene expression patterns revealed by comparison of genital fibroblasts from normal males and individuals with complete
9. Tusher VG, Tibshirani R and Chu G: Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001; 98: 5116.
14. Stein B and Baldwin AS Jr: Distinct mechanisms for regulation of the interleukin-8 gene involve synergism and cooperativity between C/EBP and NF-kappa B. Mol Cell Biol 1993; 13: 7191.
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
15. Poli V: The role of C/EBP isoforms in the control of inflammatory and native immunity functions. J Biol Chem 1998; 273: 29279. 16. Ruffell D, Mourkioti F, Gambardella A et al: A CREB-C/EBPbeta cascade induces M2 macrophage-specific gene expression and promotes muscle injury repair. Proc Natl Acad Sci U S A 2009; 106: 17475. 17. Wessells J, Yakar S and Johnson PF: Critical prosurvival roles for C/EBP beta and insulin-like growth factor I in macrophage tumor cells. Mol Cell Biol 2004; 24: 3238. 18. Kanazawa T, Konno A, Hashimoto Y et al: Expression of hepatocyte nuclear factor 4alpha in developing mice. Anat Histol Embryol 2009; 38: 34. 19. Kanazawa T, Ichii O, Otsuka S et al: Hepatocyte nuclear factor 4 alpha is associated with mesenchymal-epithelial transition in developing kidneys of C57BL/6 mice. J Vet Med Sci 2011; 73: 601. 20. Kanazawa T, Konno A, Hashimoto Y et al: Hepatocyte nuclear factor 4 alpha is related to survival
of the condensed mesenchyme in the developing mouse kidney. Dev Dyn 2010; 239: 1145. 21. Ma T, Yang B, Gillespie A et al: Severely impaired urinary concentrating ability in transgenic mice lacking aquaporin-1 water channels. J Biol Chem 1998; 273: 4296. 22. Topcu SO, Norregaard R, Pedersen M et al: Regulation of aquaporins and sodium transporter proteins in the solitary kidney in response to partial ureteral obstruction in neonatal rats. Urol Int 2011; 87: 94. 23. Tang T, Lai NC, Hammond HK et al: Adenylyl cyclase 6 deletion reduces left ventricular hypertrophy, dilation, dysfunction, and fibrosis in pressure-overloaded female mice. J Am Coll Cardiol 2010; 55: 1476. 24. Na GY, Seo SK, Lee SJ et al: Upregulation of the NNP-1 (novel nuclear protein-1, D21S2056E) gene in keloid tissue determined by cDNA microarray and in situ hybridization. Br J Dermatol 2004; 151: 1143.
1041
25. Pandit KV, Corcoran D, Yousef H et al: Inhibition and role of let-7d in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2010; 182: 220. 26. Nagase T, Uozumi N, Ishii S et al: A pivotal role of cytosolic phospholipase A(2) in bleomycin-induced pulmonary fibrosis. Nat Med 2002; 8: 480. 27. Bartoli F, Penza R, Aceto G et al: Urinary epidermal growth factor, monocyte chemotactic protein-1, and beta2-microglobulin in children with ureteropelvic junction obstruction. J Pediatr Surg 2011; 46: 530. 28. Mesrobian HG, Mitchell ME, See WA et al: Candidate urinary biomarker discovery in ureteropelvic junction obstruction: a proteomic approach. J Urol 2010; 184: 709. 29. Taha MA, Shokeir AA, Osman HG et al: Diagnosis of ureteropelvic junction obstruction in children: role of endothelin-1 in voided urine. Urology 2007; 69: 560. 30. Tesch GH: Review: serum and urine biomarkers of kidney disease: a pathophysiological perspective. Nephrology (Carlton) 2010; 15: 609.
Purpose: Loss of renal function is often the impetus for operative intervention in renal obstruction cases. Obstructive nephropathy is characterized by discrete morphological and physiological changes, including tubular dilatation, apoptosis and atrophy as well as interstitial cellular infiltration and progressive interstitial fibrosis. We hypothesized that gene expression alterations correlate with obstructive nephropathy and could serve as biomarkers for early intervention. Materials and Methods: C57BL/6 mice were subjected to unilateral ureteral obstruction or sham surgery at postnatal day 21. Kidneys were harvested 1, 2, 5 and 9 days postoperatively. RNA was extracted from kidneys and comprehensive gene expression profiling was performed with microarrays. IPA® pathway analysis software was used to analyze the biological function and gene networks of gene expression data. Results: Microarray analysis revealed more than 1,800 transcripts that were up-regulated or down-regulated during days 1 through 9 after obstruction, including many previously reported transcripts (FOS, CD44, CLU, SPP1 and EGF). Pathway analysis showed significant enrichment of transcripts in cell activation/ differentiation, immune/inflammatory responses, cell cycle, metabolic process and transport. Network analysis using IPA showed that transcriptional regulatory pathways involving CEBPB and HNF4A are involved in obstructive nephropathy. Conclusions: This data set provides a foundation for development of biomarkers for obstructive nephropathy. Key Words: kidney, ureteral obstruction, hydronephrosis, gene expression, mice UPPER urinary tract obstruction is a common condition in the adult and pediatric populations, although to our knowledge its true prevalence is unknown. Autopsy studies have identified hydronephrosis in 2% to 3% of adults and 2% of infants.1,2 Upper urinary tract obstruction can be partial or complete and transient, such as from a urinary calculus, or chronic. When encountered in any form, obstruction and hydronephrosis pose particular challenges to the clinician, who must decide
whether intervention is needed to relieve obstruction to protect renal function. Currently the decision to intervene is made based on imaging that shows worsening hydronephrosis or renal function loss, or serum chemistry studies that document altered renal function. However, these measures provide only an imprecise, crude or much delayed measure of renal compromise. Also, to our knowledge no marker or panel of markers has been developed for use in clinical practice.
0022-5347/12/1883-1033/0 THE JOURNAL OF UROLOGY® © 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION
http://dx.doi.org/10.1016/j.juro.2012.05.004 Vol. 188, 1033-1041, September 2012 RESEARCH, INC. Printed in U.S.A.
AND
Abbreviations and Acronyms CEBPB ⫽ CCAAT/enhancer binding protein  CUUO ⫽ complete unilateral ureteral obstruction EGF ⫽ epidermal growth factor HNF4A ⫽ hepatic nuclear factor 4␣ SAM ⫽ microarray significance analysis SPP1 ⫽ secreted phosphoprotein 1 Submitted for publication January 10, 2012. Study received Stanford University animal care and use committee approval. Supported by the Keith and Jan Hurlbut Research Fund. Supplementary material for this article can be obtained at http://jurology.com. * Correspondence: Department of Urology, Room S287, Stanford University School of Medicine, 300 Pasteur Dr., Stanford, California 943055118 (telephone: 650-498-4464; FAX: 650-7234200; e-mail: [email protected]).
www.jurology.com
1033
1034
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
We characterized comprehensive changes in gene expression in various contexts and used these data to provide mechanistic insights and define disease biomarkers.3–5 For single genes and proteins postCUUO expression changes occur relatively rapidly.6 Also, acute CUUO results in hydronephrosis and cortical loss within several days. In mice irreversible renal damage was documented within 3 days after obstruction.7 Since previous studies of gene expression changes after obstruction are limited in scope with only a few time points and were done in kidneys long after renal damage and hydronephrosis were established,6,8 we sought to understand transcriptional programs that are altered soon after obstruction. To identify early biomarkers of renal damage we performed gene expression profiling in mice during the first several days after ureteral obstruction. We also took advantage of contemporary bioinformatics approaches to gain functional insights into the data.
ated from 1 g total RNA in each reaction using a fluorescent linear amplification kit (Agilent Technologies) and 10.0 mM cyanine 3 or 5 labeled CTP. Labeled cRNAs were purified using the RNeasy Mini Kit and applied to the 44K Mouse Whole Genome Oligonucleotide Microarray (Agilent Technologies), which contains 41,534 independent oligonucleotides mapping to a total of 21,503 genes. Cyanine 5 labeled cRNA (825 ng) from each kidney sample was mixed with the same amount of cyanine 3 labeled cRNA generated from mouse Universal Reference RNA (Stratagene, La Jolla, California). Microarray hybridization, washing, scanning, data extraction and analysis were done as described in the manufacturer protocol.
Statistical Analysis
MATERIALS AND METHODS
Transcript levels for obstructed and sham operated control kidneys were collected from 2 to 4 animals on days 1, 2, 5 and 9 postoperatively. SAM 3.0 (www-stat.stanford.edu) was used to identify transcripts with significant differences in expression between obstructed and sham operated mice by comparing the 2 groups directly and at each time point.9 Since the 2 analyses largely overlapped, we used the transcript set from the comparison of all controls vs all obstructed animals with a false discovery rate of less than 2%.
CUUO Generation in Mice
Gene Ontology and Pathway Analysis
C57BL/6 mice (Charles River Laboratories, Wilmington, Massachusetts) were maintained and used in accordance with the Guide for the Care and Use of Laboratory Animals (National Research Council 2003). They received standard rodent chow (Zeigler Bros, Gardners, Pennsylvania) and had free access to water. On postnatal day 21 mice were anesthetized with ketamine and subjected to unilateral obstruction for 1, 2, 5 or 9 days (2 to 4 per time point). In this surgery the left ureter was ligated to generate an obstructed kidney. For each time point additional mice underwent laparotomy and dissection of the left ureter without ligation to serve as sham controls. The protocol was approved by the Stanford University animal care and use committee.
Kidney Collection and RNA Extraction After the mice were sacrificed the kidneys were rapidly excised and stored at – 80C. For RNA extraction kidneys were homogenized in TRIzol® reagent on ice. Total RNA was purified by RNeasy® Mini Kit purification according to manufacturer instructions. RNA concentration was determined by a NanoDrop® 1000 at 260 nm absorbance. All RNA samples had a 260:280 nm ratio above 1.9. RNA integrity was determined using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, California). Only high quality RNA (28S/18S of 1.8 or greater) was used for further analysis.
Histology Samples fixed in 10% formalin were washed, dehydrated and embedded in paraffin. Sections (7 ) were collected and stained with hematoxylin and eosin.
Microarray Analysis Gene expression analysis was done using oligonucleotide arrays (part G4140-90050 v.5.7, Agilent Technologies). Briefly, fluorescently labeled cRNA targets were gener-
Transcripts that differed significantly between obstructed and sham operated mice were uploaded to the DAVID (Database for Annotation, Visualization and Integrated Discovery) Bioinformatics Database (http://david.abcc. ncifcrf.gov/tools.jsp) for functional annotation and classification based on gene ontology assignment. In the database, transcripts were assigned to predefined functional categories, such as immune response, or to subfunctional categories, such as chemotaxis. Tables 1 and 2 lists Benjamini test values. To discover gene interactions and correlation networks the transcripts were uploaded to IPA pathway analysis software. Each transcript was mapped to its corresponding gene and overlaid on a global molecular network developed from information in the Ingenuity® Knowledge Base. Significantly enriched networks and functions were identified by comparing obstruction induced gene expression changes to the total number of occurrences of these genes in all pathway annotations in the database using the right-tailed Fisher exact test. IPA calculates a score derived from a p value for each network that reflects the likelihood that the set of significant genes in a given network is enriched or the genes are found together due to random chance. A score of 2 indicates a 1% chance that the focus genes are together in a network due to chance.
RESULTS Hydronephrosis Generation and Mouse Histological Features CUUO in 21-day-old C57BL/6 mice was generated by suture ligation of the ureter. Obstruction length in kidneys increased with time and differed significantly from the contralateral side and from that in controls by day 5 after surgery (fig. 1, A). There was
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
Table 1. Function category of up-regulated genes in hydronephrotic kidney Function (term) Cell cycle/chromosome segregation: Cell cycle DNA metabolic process DNA packaging Cell stress response Organization: Extracellular matrix Pos cellular component regulation Cell growth regulation Microtubule based process Protein processing Immune ⫹ inflammatory responses: Immune response Immune response activation Chemotaxis Inflammatory response Wound response Defense response Adhesion: Cell Cell-cell Cell regulation Cell activation ⫹ differentiation: Proliferation regulation Activation Development: Vasculature Lung Hemopoietic or lymphoid organ Heart
No. Genes Fold Benjamini (%) p Value Enrichment Test
96 (12.73) 43 (5.70) 15 (1.99) 30 (3.98)
3.10E-32 2.70E-08 3.90E-05 0.00132
3.99 2.59 3.77 1.89
7E-29 4E-06 2E-03 4E-02
14 (1.86) 1.58E-04 15 (1.99) 3.06E-04
3.52 3.12
7E-03 1E-02
12 (1.59) 9.38E-04 28 (3.71) 6.24E-08 12 (1.59) 7.79E-04
3.31 3.37 3.39
3E-02 8E-06 3E-02
59 17 16 37 45 44
6.45E-15 1.96E-07 2.27E-05 5.22E-13 5.37E-12 5.59E-08
3.18 5.02 3.73 4.18 3.29 2.49
1E-12 2E-05 2E-03 1E-10 1E-09 7E-06
59 (7.82) 1.17E-11 21 (2.79) 0.00105 12 (1.59) 0.00112
2.67 2.26 3.24
2E-09 3E-02 3E-02
46 (6.10) 1.39E-06 27 (3.58) 4.46E-06
2.17 2.79
1E-04 4E-04
25 (3.32) 5.07E-05 14 (1.86) 4.08E-04 25 (3.32) 3.07E-04
2.54 3.20 2.26
3E-03 2E-02 1E-02
21 (2.79) 5.12E-04
2.39
2E-02
(7.82) (2.25) (2.12) (4.91) (5.97) (5.84)
no difference in kidney length between the control and contralateral nonobstructed kidneys from obstructed mice. Hematoxylin and eosin stained sections of obstructed kidneys revealed progressive tubular injury by postoperative day 5, as manifested by tubular dilatation, epithelial cell flattening and loss of brush border in proximal tubules (fig. 1, B). Peritubular inflammation, edema and fibrosis were also apparent histologically 5 and 9 days after obstruction, as evidenced by decreased tubular lumen size, loss of back to back tubular organization and infiltration of peritubular spaces with fibroblasts and inflammatory cells. There were no notable histological differences between obstructed and control kidneys at postoperative days 1 and 2 (fig. 1, B). Identifying Genes Modulated in Kidney After Ureteral Obstruction SAM was used to identify 857 transcripts significantly up-regulated in obstructed kidneys and 982 significantly down-regulated vs controls. Transcripts
1035
were grouped by hierarchical clustering analysis to show those with correlated patterns of expression changes with time (fig. 2, A). In many cases genes that were co-regulated shared similar functions. For example, many transcripts that were up-regulated soon after obstruction were implicated in cell activation/differentiation while genes up-regulated several days after obstruction were enriched for immune/ inflammatory responses and cell cycle regulation. Many down-regulated transcripts encoded for proteins important for metabolic processes and transport functions. Notably several transcripts known to be modulated in response to ureteral obstruction were represented in the data set, including FOS, CD44, clusterin, SPP1 and EGF (fig. 2, B).6,10–13 Significantly enriched gene ontology terms associated with each part of the dendrogram were identified using DAVID. For genes up-regulated in response to ureteral obstruction significant enrichment was seen for transcripts involved in cell cycle/chromosome segregation, organization, immune/inflammatory responses, cell adhesion, cell activation/differentiation and development (table 1). Enriched biological functions for down-regulated transcripts
Table 2. Function category of down-regulated genes in hydronephrotic kidney
Oxidation reduction Transport: Ion Transmembrane Cation Anion Carbohydrate Process: Coenzyme metabolic Fatty acid metabolic Organic acid catabolic Lipid biosynthetic Nitrogen compound biosynthetic Steroid metabolic Alditol metabolic Organic acid biosynthetic Cellular amino acid derivative metabolic Glycerol ether metabolic Neutral lipid metabolic Vitamin metabolic Organic ether metabolic Aspartate family amino acid metabolic Glycerol metabolic Pyruvate metabolic Heterocycle catabolic Carbohydrate stimulus response
No. Genes (%)
p Value
Fold Enrichment
Benjamini Test
85 (10.34)
1.57E-21
3.17
3E-18
146 58 95 45 11
(9.25) (7.06) (6.45) (2.68) (1.34)
5.68E-15 1.51E-14 5.49E-10 4.14E-08 1.15E-04
2.68 3.16 2.58 4.24 4.60
6E-12 1E-11 2E-07 9E-06 8E-03
36 29 16 29 30
(3.16) (3.53) (1.95) (3.53) (3.65)
3.80E-10 9.34E-10 5.92E-07 1.00E-05 1.08E-05
4.56 3.95 4.95 2.55 2.49
1E-07 2E-07 1E-04 1E-03 1E-03
20 8 18 28
(2.43) (0.97) (2.19) (2.19)
2.20E-05 3.92E-05 4.49E-05 4.49E-05
3.11 8.02 3.20 3.20
3E-03 4E-03 4E-03 4E-03
10 10 29 10 7
(1.22) (1.22) (1.46) (1.22) (0.85)
5.78E-05 5.78E-05 8.07E-05 9.84E-05 1.68E-04
5.57 5.57 4.36 5.22 7.98
5E-03 5E-03 6E-03 7E-03 1E-02
7 7 9 7
(0.85) (0.85) (1.09) (0.85)
2.19E-04 3.59E-04 4.11E-04 4.51E-04
7.63 7.02 4.91 6.75
1E-02 2E-02 2E-02 2E-02
1036
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
B A
Obstruction
Sham
Obstruction
Sham
P1
P1
P2 P2
P5
P5
P9 P9
Figure 1. Representative kidney morphology and histology after ureteral obstruction. A, kidneys from mice after left ureter was obstructed by ligation and after sham surgery (Sham). B, stained, formalin fixed, paraffin embedded samples of representative kidneys after ureteral obstruction. Sham operated control renal cortexes show mature glomeruli, normal tubules and few interstitial cells. Obstructed kidneys had progressive accumulation of architectural changes after obstruction. Cortex structure appeared relatively normal 1 and 2 days postoperatively but showed alterations by days 5 and 9 with narrowed glomerular spaces, fewer tubules and more interstitial cell proliferation with disrupted tubular architecture. P, postoperative day. H&E, reduced from ⫻20.
included those associated with oxidation-reduction, transport and metabolic process (table 2). To gain further insight into the regulatory networks and gene functions altered by CUUO the SAM list of transcripts was analyzed using IPA. When protein products of significant transcripts were grouped by subcellular location 124 of 574 up-regulated (22%) and 212 of 535 down-regulated (40%) gene products localized to the cytoplasm, 21% and 14% to the extracellular space, 31% and 14% to the nucleus and 21% and 25% to the plasma membrane, respectively. When grouped by molecular function, 7% of up-regulated and 5% of down-regulated genes were classified as cytokines, 13% and 31% as enzymes, 3% and 2% as growth factors, 1% and 3% as ion channels, 7% and 6% as kinases, 3% and 3% as peptidases, 2% and 2% as phosphatases, 14% and 9% as transcription regulators, 7% and 2% as transmembrane proteins, and 3% and 13% as transporters, respectively. Integrating gene function and cellular location data revealed that most up-regulated gene products localized to the nucleus and were transcription regulators or transmem-
brane proteins. In contrast, most down-regulated gene products localized to cytoplasm and functioned as enzymes, ion channels and transporters. Networks Involved in Renal Response to Ureteral Obstruction To assess the validity of our data and gain insight into the functional consequences of CUUO we used IPA to identify significantly enriched gene networks, cellular locations and gene functions. A total of 25 significantly enriched networks were generated based on a score of 3 or greater. The top network contained 34 transcripts, of which 33 were modulated significantly in response to UUO in our data set. Cognate proteins for 14 transcripts regulate or are regulated by CEBPB, of which the transcript level increased in response to obstruction (fig. 3, A). Five of the proteins related to CEBPB, including ERK, FOS, GLIPR2, CDKN2A and DGAT2, were previously implicated in ureteral obstruction or renal fibrosis. In the second network HNF4A occupied a central role. The IPA knowledge base associated HNF4A
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
A
1037
B Obstruction
Control
Obstruction
P1 P2 P5 P1 P1 P2 P5 P9 P9 P1 P1 P1 P2 P2 P2 P5 P5 P5 P9 P9 P9
P1 P2 P5 P1 P1 P2 P5 P9 P9 P1 P1 P1 P2 P2 P2 P5 P5 P5 P9 P9 P9
Control
Cell activation/ differentiation
FOS CD44 CLU SPP1
Immune/inflammatory response
EGF
1.00
0.67
0.33
0.00
0.33
0.67
Immune/inflammatory response
1.00
Organization
Cell adhesion/cell activation Immune/inflammatory response Cell activation/ differentiation Cell cycle
Immune/inflammatory response Metablic process
Transport Metablic process
Transport Metablic process
Transport Metablic process
Figure 2. Gene expression profiles of kidneys in CUUO mice and sham operated controls at postoperative days 1, 2, 5 and 9. A, heat map shows hierarchical clustering of transcript data on sham operated controls and CUUO mice. Black vertical lines indicate 5 major gene clusters enriched for gene ontology terms related to cell activation/differentiation, immune/inflammatory responses, cellular organization, cell adhesion and cell cycle in up-regulated genes, and 2 major clusters related to metabolic process and transport in down-regulated genes. Red areas represent increased expression relative to mean per gene across experiments. Green areas represent decreased expression. Color saturation corresponds to degree of expression modulation. B, CUUO resulted in significantly increased RNA expression of FOS, CD44, CLU and SPP1, and decreased EGF expression from day 1 to day 9 postoperatively.
with another 23 transcripts that we found were significantly modulated by CUUO (fig. 3, B). Previously 5 of these transcripts (AQP1, FHL1, CBR3, HMGA2 and PLA2G4A) were reported to be involved in hydronephrosis or fibrosis, suggesting that the CEBPB
and HNF4A signaling pathways have important roles in the renal response to CUUO. Also, 7 genes (CIITA, EZH2, histone H3, LGALS3, VIM, CDH16 and DAB2) that were in network 1 but not directly associated with CEBPB and 2 (HNF1A
1038
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
A
B
Figure 3. Top 2 networks of gene expression in kidney after CUUO. A, network 1. Genes or gene products (nodes) and putative biological relationships (arrows) between each gene product. Location of modulated gene cognate proteins are shown in extracellular space (top), cell membrane (light blue area), cytoplasm, nuclear membrane (pink area) and within nucleus (bottom). Transcriptional regulator CEBPB is at center of nuclear node. Red nodes indicate up-regulated transcripts. Green areas indicate down-regulation. Noncolored nodes represent genes added by IPA based on network algorithm that were not significantly altered by CUUO in microarray data. Node shapes correspond to gene product functional class. Orange lines indicate genes that interact directly with CEBPB. B, network 2. Transcriptional regulator HNF4A down-regulation was associated with modulation of many transcripts involved in normal renal tubular function, including several transmembrane transporters. Asterisk indicates gene products represented in more than 1 network.
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
and UMOD) in network 2 that did not directly interact with HNF4A were previously implicated in hydronephrosis or fibrosis. The remaining 23 networks identified by IPA as significantly modulated in response to UUO contained many transcripts and biological functions known to participate in renal damage. For instance, activation of the immune response is represented robustly in several networks. It includes transcripts involved in cellular movement, immune cell trafficking and inflammatory response. Transcripts and pathways associated with down-regulation of cell growth and proliferation were strongly represented, as were pathways associated with cell death. Also, transcripts functionally related to functions of terminally differentiated renal cells, such as molecule transport, energy production, lipid metabolism and small molecular biochemistry, were mostly down-regulated.
DISCUSSION CUUO produced significant changes in gene expression in the kidneys of mature mice. We identified more than 1,800 transcripts that were significantly altered in response to obstruction, a number significantly larger than in previous studies due to the number of samples analyzed and the number of transcripts assayed on the microarrays. Gene expression changes appeared to be time dependent, in that some transcript levels changed relative to control within 1 day after obstruction while others required several days to be significantly modulated. The number of transcripts induced by obstruction was comparable to the number suppressed. These data provide a comprehensive assessment of transcript changes induced by renal obstruction. They are a starting point to develop biomarkers correlating with renal damage in response to obstruction. Renal damage arising from obstruction depends on the degree (complete or partial) and the duration of obstruction. The renal response to obstruction also likely depends on when in life obstruction occurs, that is during gestation or in adulthood. In adult patients and animals complete obstruction can be tolerated for a short time with little apparent functional damage, provided that obstruction is relieved.7 A previous study showed that CUUO in C57BL/6 mice, which were used in our study, had irreversible renal damage with increased serum BUN after 3 days.7 In that context the gene expression changes that we observed in animals obstructed 5 or more days might reflect alterations in the kidney associated with fibrosis, permanent tubular loss and other aspects of permanent renal damage. Undoubtedly many but not all gene expression changes within the first 3 days after obstruction involve pathways that give rise to irreversible changes in
1039
the damaged kidney and could be reversed by early relief of obstruction. Those transcripts would be attractive candidate biomarkers of the renal damage induced by obstruction. Thus, understanding the biological functions of transcripts that were significantly modulated in response to obstruction is important to develop future biomarkers to predict the loss of renal function. Pathway analysis identified 25 gene networks that were significantly enriched in obstructed kidneys. The top 2 networks illustrate the potential usefulness of pathway analysis in gaining an understanding of the renal response to unilateral ureteral obstruction. The central gene in the top network was CEBPB. Through interactions with other transcriptional regulators14 CEBPB regulates genes involved in inflammation,15 muscle injury repair16 and cell proliferation.17 Genes that interact with CEBPB in the network were previously associated with hydronephrosis or fibrosis, including GLIPR-2, CDKN2A and DGAT2. GLIPR-2 up-regulation, and DGAT2 and CDKN2A down-regulation results in increased expression of genes associated with fibrosis, such as CTGF and various chemokine genes. HNF4A is a transcription factor in the nuclear hormone receptor family that is generally expressed in endoderm derived epithelial tissues and the developing metanephros.18 HNF4A expression negatively correlates with the expression of mesenchyma associated genes, eg ACTA1 and VIM.19 HNF4A knockdown using siRNA interfered with cellular organization development in the condensed mesenchyma of cultured embryonic kidneys.20 Network analysis showed that HNF4A down-regulation induced by UUO was associated with AQP1 and FHL down-regulation, and upregulation of CBR3, HMGA2 and PLA2G4A (fig. 3, B). Based on previous functional studies of these transcripts their modulation likely contributes to renal dysfunction and fibrosis.21–26 We were not the first group to use gene expression profiling to investigate the transcriptional changes induced by renal obstruction, although to our knowledge no group has comprehensively analyzed changes soon after obstruction. Silverstein et al identified 250 genes modulated in response to complete ureteral obstruction in 1-day-old rats after 4 and 12 days.6 Since the rat kidney undergoes significant development postnatally, gene expression changes in that study could differ significantly from those that we observed in adult mice. Regardless, significant overlaps were observed between that data set and ours, particularly in genes associated with the induction of the inflammatory response and stress response genes, including transforming growth factor-1, CLU, FOS, EGF, CEBPA/B, calponin and lumican. Seseke et al profiled expression changes in a rat congenital model of high grade partial renal obstruc-
1040
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
tion at days 32 to 35 postnatally.8 Microarrays were performed on pooled RNA samples from several rats and expression differences were seen in 18 transcripts. Transcripts significantly induced by obstruction included several growth factors altered in our data set (transforming growth factor-, IGF1 and IGFBP5/6) as well as several transport proteins that were down-regulated (SLC12A3 and 34A1). Cognate proteins for several genes found in our data set were also measured in the urine of patients after obstruction, including EGF,27 AMBP, CD44, CDH11, FN1, SERPINA1, SPP1, CD14, MMP9, S100A6,28 GGT,29 CXCL10 and L16, VCAM1 and COL4.30 Many transcripts in our data set could be developed as novel urinary markers of renal damage due to obstruction. We confirmed the modulation of previously identified transcripts in our study but this study is preliminary and has limitations. To our knowledge we report the largest data set after renal obstruction to date. However, more mice per time point and the inclusion of earlier time points would improve statistical power and open avenues for other analysis tools. Furthermore, acute CUUO in our model differs from clinical obstruction, which is often partial, chronic and gradual in onset. This difference could influence the number of transcripts and physiological response of the kidney. Another challenge in developing obstruction biomarkers arises when protein levels fail to correlate with transcript levels. While transcript levels of several proteins known to be modulated by obstruction were observed in our data set, additional study is
needed to test novel proteins predicted from our data set. Alternatively urinary transcript levels could be evaluated for renal damage markers, much as PCA3 has been validated as a urine based RNA biomarker for prostate cancer.
CONCLUSIONS Gene expression profiling of unilateral ureteral obstruction provides insights into the possible mechanisms underlying obstruction induced renal damage and is a potential source of renal damage biomarkers. Many genes related to cell activation/differentiation, immune/inflammatory responses, cell cycle regulation, metabolic processes and transport functions were identified that were associated with the renal damage induced by obstruction. Pathway analysis identified CEBPB and HNF4A as candidate regulators of transcripts and pathways involved in the response to kidney obstruction. Future study is necessary to verify whether CEBPB and HNF4A have important roles in early gene expression events associated with renal damage and fibrosis induced by obstruction. In addition, an ongoing study will investigate the possibility of detecting RNA species or their cognate proteins in urine after partial ureteral obstruction.
ACKNOWLEDGMENTS Rosie Nolley provided technical assistance. Microarray raw data are available at http://www.ncbi.nlm. nih.gov/geo/query/acc.cgi?acc ⫽ GSE36496 (Accession No. GSE36496).
REFERENCES 1. Lee RS, Cendron M, Kinnamon DD et al: Antenatal hydronephrosis as a predictor of postnatal outcome: a meta-analysis. Pediatrics 2006; 118: 586. 2. Siddiqui MM and McDougal WS: Urologic assessment of decreasing renal function. Med Clin North Am 2011; 95: 161. 3. Zhao H, Zongming M, Tibshirani R et al: Alteration of gene expression signatures of cortical differentiation and wound response in lethal clear cell renal cell carcinomas. PLoS One 2009; 4: e6039.
and partial androgen insensitivity syndrome. BMC Genomics 2007; 8: 376. 6. Silverstein DM, Travis BR, Thornhill BA et al: Altered expression of immune modulator and structural genes in neonatal unilateral ureteral obstruction. Kidney Int 2003; 64: 25. 7. Puri TS, Shakaib MI, Chang A et al: Chronic kidney disease induced in mice by reversible unilateral ureteral obstruction is dependent on genetic background. Am J Physiol Renal Physiol 2010; 298: F1024.
10. Florquin S and Rouschop KM: Reciprocal functions of hepatocyte growth factor and transforming growth factor-beta1 in the progression of renal diseases: a role for CD44? Kidney Int Suppl 2003; 86: S15. 11. Chevalier RL: Biomarkers of congenital obstructive nephropathy: past, present and future. J Urol 2004; 172: 852. 12. Kiley SC, Thornhill BA, Belyea BC et al: Epidermal growth factor potentiates renal cell death in hydronephrotic neonatal mice, but cell survival in rats. Kidney Int 2005; 68: 504.
4. Bhamre S, Sahoo D, Tibshirani R et al: Temporal changes in gene expression induced by sulforaphane in human prostate cancer cells. Prostate 2009; 69: 181.
8. Seseke F, Thelen P and Ringert RH: Characterization of an animal model of spontaneous congenital unilateral obstructive uropathy by cDNA microarray analysis. Eur Urol 2004; 45: 374.
13. Yoo KH, Thornhill BA, Forbes MS et al: Osteopontin regulates renal apoptosis and interstitial fibrosis in neonatal chronic unilateral ureteral obstruction. Kidney Int 2006; 70: 1735.
5. Holterhus PM, Deppe U, Werner R et al: Intrinsic androgen-dependent gene expression patterns revealed by comparison of genital fibroblasts from normal males and individuals with complete
9. Tusher VG, Tibshirani R and Chu G: Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001; 98: 5116.
14. Stein B and Baldwin AS Jr: Distinct mechanisms for regulation of the interleukin-8 gene involve synergism and cooperativity between C/EBP and NF-kappa B. Mol Cell Biol 1993; 13: 7191.
GENE EXPRESSION CHANGES INDUCED BY UNILATERAL URETERAL OBSTRUCTION IN MICE
15. Poli V: The role of C/EBP isoforms in the control of inflammatory and native immunity functions. J Biol Chem 1998; 273: 29279. 16. Ruffell D, Mourkioti F, Gambardella A et al: A CREB-C/EBPbeta cascade induces M2 macrophage-specific gene expression and promotes muscle injury repair. Proc Natl Acad Sci U S A 2009; 106: 17475. 17. Wessells J, Yakar S and Johnson PF: Critical prosurvival roles for C/EBP beta and insulin-like growth factor I in macrophage tumor cells. Mol Cell Biol 2004; 24: 3238. 18. Kanazawa T, Konno A, Hashimoto Y et al: Expression of hepatocyte nuclear factor 4alpha in developing mice. Anat Histol Embryol 2009; 38: 34. 19. Kanazawa T, Ichii O, Otsuka S et al: Hepatocyte nuclear factor 4 alpha is associated with mesenchymal-epithelial transition in developing kidneys of C57BL/6 mice. J Vet Med Sci 2011; 73: 601. 20. Kanazawa T, Konno A, Hashimoto Y et al: Hepatocyte nuclear factor 4 alpha is related to survival
of the condensed mesenchyme in the developing mouse kidney. Dev Dyn 2010; 239: 1145. 21. Ma T, Yang B, Gillespie A et al: Severely impaired urinary concentrating ability in transgenic mice lacking aquaporin-1 water channels. J Biol Chem 1998; 273: 4296. 22. Topcu SO, Norregaard R, Pedersen M et al: Regulation of aquaporins and sodium transporter proteins in the solitary kidney in response to partial ureteral obstruction in neonatal rats. Urol Int 2011; 87: 94. 23. Tang T, Lai NC, Hammond HK et al: Adenylyl cyclase 6 deletion reduces left ventricular hypertrophy, dilation, dysfunction, and fibrosis in pressure-overloaded female mice. J Am Coll Cardiol 2010; 55: 1476. 24. Na GY, Seo SK, Lee SJ et al: Upregulation of the NNP-1 (novel nuclear protein-1, D21S2056E) gene in keloid tissue determined by cDNA microarray and in situ hybridization. Br J Dermatol 2004; 151: 1143.
1041
25. Pandit KV, Corcoran D, Yousef H et al: Inhibition and role of let-7d in idiopathic pulmonary fibrosis. Am J Respir Crit Care Med 2010; 182: 220. 26. Nagase T, Uozumi N, Ishii S et al: A pivotal role of cytosolic phospholipase A(2) in bleomycin-induced pulmonary fibrosis. Nat Med 2002; 8: 480. 27. Bartoli F, Penza R, Aceto G et al: Urinary epidermal growth factor, monocyte chemotactic protein-1, and beta2-microglobulin in children with ureteropelvic junction obstruction. J Pediatr Surg 2011; 46: 530. 28. Mesrobian HG, Mitchell ME, See WA et al: Candidate urinary biomarker discovery in ureteropelvic junction obstruction: a proteomic approach. J Urol 2010; 184: 709. 29. Taha MA, Shokeir AA, Osman HG et al: Diagnosis of ureteropelvic junction obstruction in children: role of endothelin-1 in voided urine. Urology 2007; 69: 560. 30. Tesch GH: Review: serum and urine biomarkers of kidney disease: a pathophysiological perspective. Nephrology (Carlton) 2010; 15: 609.