Global Gene Expression Profiling in PPAR-γ Agonist-Treated Kidneys in an Orthologous Rat Model of Human Autosomal Recessive Polycystic Kidney Disease

Kidneys are enlarged by aberrant proliferation of tubule epithelial cells leading to the formation of numerous cysts, nephron loss, and interstitial fibrosis in polycystic kidney disease (PKD). Pioglitazone (PIO), a PPAR-γ agonist, decreased cell proliferation, interstitial fibrosis, and inflammation, and ameliorated PKD progression in PCK rats (Am. J. Physiol.-Renal, 2011). To explore genetic mechanisms involved, changes in global gene expression were analyzed. By Gene Set Enrichment Analysis of 30655 genes, 13 of the top 20 downregulated gene ontology biological process gene sets and six of the top 20 curated gene set canonical pathways identified to be downregulated by PIOtreatment were related to cell cycle and proliferation, including EGF, PDGF and JNK pathways. Their relevant pathways were identified using the Kyoto Encyclopedia of Gene and Genomes database. Stearoyl-coenzyme A desaturase 1 is a key enzyme in fatty acid metabolism found in the top 5 genes downregulated by PIO treatment. Immunohistochemical analysis revealed that the gene product of this enzyme was highly expressed in PCK kidneys and decreased by PIO. These data show that PIO alters the expression of genes involved in cell cycle progression, cell proliferation, and fatty acid metabolism.


Introduction
Polycystic kidney diseases (PKD) are characterized by progressive enlargement of numerous fluid-filled cysts in both kidneys, often leading to chronic kidney disease (CKD). Autosomal dominant PKD (ADPKD) is one of the most common hereditary disorders in humans with an incidence of 1 : 500-1,000, caused by mutations in the PKD1 or PKD2 gene. Progressive kidney enlargement is due to aberrant proliferation of the cystic epithelia, together with an accumulation of fluid into the cyst cavities due to transepithelial chloride (Cl − ) and fluid secretion [1][2][3]. Autosomal recessive PKD (ARPKD) is known as a juveniletype cystic disease with an incidence of 1 : 20,000 [3]. Kidneys in ARPKD patients are characterized by cystic fusiform dilations of the collecting ducts accompanied by increased cell proliferation and fluid secretion, leading to massive kidney enlargement and renal failure occurring in the first few years after birth [4]. Increased cell proliferation, stimulated fluid secretion, and interstitial fibrosis are often observed in cystic liver disease in ARPKD as well [5].
Peroxisome proliferator-activated receptors (PPARs) belong to a nuclear receptor superfamily of ligand-activated transcription factors with subtypes α, β/δ, and γ. PPAR-γ is widely expressed in several organs including kidneys and known to be activated by fatty acids [6,7]. Antidiabetic agents, pioglitazone (PIO), troglitazone, ciglitazone, and rosiglitazone, are used to control blood sugar levels in patients with diabetes mellitus. These PPAR-γ agonists also have important roles in regulation of cell cycle, inhibition of fibrosis, infiltration and metastasis of cancer cells, and modulation of inflammatory cytokines.
Treatment with PIO improved survival and ameliorated cardiac defects and the degree of renal cystogenesis in embryos of Pkd1 −/− mice in a previous study [8]. In addition, long-term treatment of this agonist improved endothelial function by increasing production of nitric oxide in adult heterozygous Pkd1 +/− mice [8]. Another PPAR-γ agonist, rosiglitazone attenuated PKD progression and prolonged survival of Han: SPRD Cy rats [9]. In our recent study, daily treatment of PIO ameliorated polycystic kidney disease through inhibiting Raf/MEK/ERK and AKT/mTOR/S6 signaling cascades in the PCK rat, an orthologous model of human ARPKD [10]. These findings suggest that PPAR-γ agonists may have therapeutic value in ARPKD via altering several cellular signaling pathways. In the current study, we applied global gene expression profiling to explore novel cellular signaling pathways potentially related to the ameliorating effects of PIO in PCK rat kidneys.

PCK Rat and Study
Design. PCK rats were originally derived from a strain of Sprague-Dawley rats in Japan and descendants of this colony have been maintained at the Education and Research Center of Animal Models for Human Diseases, Fujita Health University. PCK rats and normal Sprague Dawley rats (+/+; Charles River Japan Inc., Kanagawa, Japan) were allowed free access to water and food throughout the study. Female PCK and +/+ rats, aged 4-20 weeks (n = 10 per gender) were randomly assigned to one of two groups: treatment with 10 mg/kg PIO (Takeda Pharmaceutical Company Limited, Osaka, Japan) or vehicle control (0.5% DMSO) by gavage every day as previously reported [10]. The protocol for the ethics and use of these animals was approved by the Animal Care and Use Committee at Fujita Health University.
At 20 weeks of age, rats were anesthetized with sodium pentobarbital (Schering-Plough Corp., Kenilworth, NJ), and the kidneys were removed rapidly, causing lethal exsanguination. Half of the left kidney was frozen in liquid nitrogen for RNA extraction. Half of the right kidney was immersed in 4% paraformaldehyde, embedded in paraffin, and sectioned for immunohistochemistry.

RNA Extraction.
RNA was extracted from kidneys of rats with or without PIO treatment using a monophasic solution of phenol/guanidine isothiocyanate and TRIzol reagent (Invitrogen Co., Carlsbad, CA, USA) in accordance with their manual, and the samples were incubated with RNase-free DNase I (Ambion, TX, USA). The quality and concentration of each sample was confirmed by spectrophotometry (NanoDrop ND-1000; Asahi glass Co. Ltd., Tokyo, Japan). Total RNA obtained from three females was pooled in each PIO-treated or control vehicle-treated (CONT) group in accordance with our previous report [11].

Microarrays.
DNA microarray experiments were performed essentially as described previously [11]. Briefly, 500 ng aliquots of total RNA obtained from kidneys of five rats were labeled using a Quick Amp Labeling Kit, one-color (Agilent Technologies, Inc., Santa Clara, CA, USA), according to the manufacturer's instructions. The pooled renal RNA of PIO-or vehicle-treated PCK rats were labeled with the Cy3-fluorescence dye. After determination of labeling efficiency, 1.65 μg aliquots of Cy3-labeled RNA were hybridized using the Gene Expression hybridization kit (Agilent Technologies) onto Rat Oligo Microarrays (Agilent Technologies, product no. G4130A) according to the manufacturer's hybridization protocol. The microarray slides were examined with an Agilent microarray scanner and software. Data analysis was performed with Agilent Feature Extraction software (version A.7.1.1).
Data from microarray experiments of PIO-or vehicletreated rats were analyzed independently. Primary microarray data are available from the Gene Expression Omnibus (GEO) (accession number GSE00000). Evaluation of signal intensity was divided into three classes, {0}: nondetected, {1}: weakly detected, and {2}: strongly detected transcription product. Gene ontology analysis of biological process (C5BP) and curated gene sets of canonical pathways (C2CP) were analyzed by importing the data into Gene Set Enrichment Analysis (GSEA version 2, the Broad Institute/Massachusetts Institute of technology, USA) [12].
Using the GeneSpring software, the changed probes were listed as "Log 2 ratio was over 1 (over 2-fold) or less than −1 (less than 1/2-fold) between PIO group and CONT group" and "the signal evaluation was {2} (strongly detected) in both groups". In the changed genes, Kyoto Encyclopedia of Gene and Genomes (KEGG) analysis was used [13].

Real-Time Reverse Transcriptase Polymerase Chain Reaction (RT-PCR)
. cDNA was produced from total RNA by reverse transcriptase using random hexamer primers (SuperScript II First Strand Synthesis System; Invitrogen Co., Carlsbad, CA, USA). To compare gene expression patterns of PCK kidneys with PIO or vehicle treatment, we selected a key enzyme in fatty acid metabolism, stearoyl-coenzyme A desaturase 1 (Scd1), and uncoupling protein 1 (Ucp1). Gene expression was detected by real-time RT-PCR (ABI 7300 real-time PCR system; Applied Biosystems, Foster City, CA, USA) using the Taq-Man reagent-based chemistry protocol. Glyceraldehyde-3phosphate dehydrogenase (GAPDH) as a housekeeping gene was used for data normalization. The probes of Scd1, UCP-1, and GAPDH were CCCACATGCTCCAAGAGATCTCCAG, CTCTTCAGGGAGAGAAACGCCTGCC, and AACCCAT-CACCATCTTCCAGGAGCG, respectively (TaqMan Gene Expression Assays; Applied Biosystems). Relative quantification of gene expression was compared to one in SD control vehicle-treated (CONT) kidneys (set to 1.0).

Statistical Analysis.
Results are expressed as the arithmetic mean ± standard error. Statistical comparisons between groups were performed by Student's t-test and twoway analysis of variance, and differences were considered to be significant at P < 0.05.

Identification of Differentially Expressed Genes by Expression Profiling.
Previous report indicates that PPAR-γ agonistic action decreases expression of endothelin receptor type A (EDNRA) [14], suggesting that EDNRA is one of the down-stream target gene of PPAR-γ agonists. In our current study, expression of Ednra was also downregulated in PIOtreated kidneys (Log2 ratio = −1.30). EDNRA expression is increased in human ADPKD, and overexpression of Ednra causes cyst formation in transgenic mouse kidneys [15]. Because not only EDNRA but also various genes  1(b)). These gene sets were related to catabolic and metabolic processes.
To examine the gene sets with the greatest changes, only 2,611 genes, which changed more than 1.25-fold in PIO-treated kidneys compared to vehicle-treated kidneys, were analyzed. 141 gene sets were formed from these 2,611 genes. 112 of those 141 gene sets were downregulated in PIO-treated kidneys compared with vehicle-treated kidneys of PCK rats. Of these, 6 gene sets were significantly different (P < 0.05, Table 2(a)). 4 of these 6 gene sets are related to cell cycle and cell proliferation (Table 2(a)). Common genes in these gene sets include G1/S or G2/M checkpoint related genes, breast cancer 2 (Brca2), cyclin-dependent kinase inhibitor 2B (Cdkn2b), CHK1 checkpoint homolog (Chek1), cell cycle checkpoint protein kinase Bub1 fragment (BUB1B), pololike kinase 1 (PLK1), and cyclin-dependent kinase inhibitor 1C (Cdkn1c) (Table 2(b)). Of the remaining 29 of the 141 gene sets that were upregulated in PIO-treated kidneys compared with vehicle-treated kidneys, only one, related to neurological system processes, was significantly elevated (P < 0.05) (Table 2(c)).
In curated gene sets of canonical pathways (C2CP), 257 were formed from the 11,809 genes detected. 201 of these 257 gene sets were downregulated in PIO-treated kidneys compared with vehicle-treated kidneys, of which 33 were significantly lower (P < 0.05). From the 20 downregulated C2CP gene sets with the highest significant differences (lowest P values), 6 gene sets were related to cell cycle and cell proliferation including c-Jun N-terminal kinase (JNK), epidermal growth factor (EGF), and platelet-derived growth factor (PDGF) pathways, and 3 gene sets were related to inflammatory signals including interleukin-1 receptor (IL1R) and interleukin-6 (IL6) pathways (Table 3(a)). One gene set, extracellular matrix (ECM) receptor interaction, also was in the top 20 downregulated in C2CP. On the other hand, 56 of 257 gene sets were upregulated in PIO-treated kidneys compared with vehicle-treated kidneys, of which 5 gene sets were significantly higher (P < 0.05, Table 3(b)). 3 of these 5 gene sets are related to glutamate, alanine, and aspartate metabolism.
GSEA is a computational method that determines whether an a priori defined set of genes shows statistically significant and concordant differences between two biological states and can detect important biological processes or canonical pathways by using the list rank information without using a threshold [12]. Among the 43,379 probes spotted on the microarray slide, 189 probes were significantly changed. From these 189 probes, 31 genes were identified by KEGG analysis. 23 of those 31 genes were downregulated in PIO-treated compared with vehicle-treated kidneys (Table 4(a)). Two key enzymes in fatty acid metabolism, stearoyl-coenzyme A desaturase 1 (Scd1) and uncoupling protein 1 (Ucp1), which are involved in PPAR signaling were in the top 15 genes downregulated by PIO treatment. On the other hand, 8 of the 31 genes were upregulated in PIO-treated kidneys compared with vehicle-treated kidneys (Table 4(b)).

Cellular Expression and Distribution of Scd1 in Rodent
Polycystic Kidneys. For Scd1 and Ucp1, in order to confirm the mRNA expression by DNA microarray screening above, real-time RT-PCR analysis was performed. The mRNA level of Scd1 in the kidney was increased in PCK rats compared to SD rats and was decreased by PIO treatment in PCK rats (Figure 1(a)). On the other hand, the mRNA level of Ucp1 was not significantly different between PCK and SD rats (data not shown). Scd1 is involved in cell proliferation via growth factors in some type of cancer cells [16][17][18]. To determine the cellular distribution of Scd1 in PCK and SD kidneys, immunohistochemistry was used. In normal SD kidneys, Scd1 was hardly detected. On the other hand, in untreated PCK kidneys, Scd1 was present in the cytoplasm of normalshaped tubule epithelia diffusely but not in growing cysts. With PIO treatment, the distribution of Scd1 decreased in those normal-shaped cells (Figures 1(b) and 1(c)). These findings suggest that Scd1 may relate to the onset of renal cyst formation originated from normal-shaped tubules.

Discussion
In our previous report, we demonstrated that PIO treatment in PCK rats inhibited renal Raf/MEK/ERK and AKT/mTOR/S6 activity and reduced proliferation of diseased renal cells [10]. In the current study, we analyzed DNA microarray using GSEA and KEGG pathway analysis in order to detect gene-based effects of PIO treatment [12,13]. The results of GSEA analysis of C5BP and C2CP are consistent with our previous findings, as a number of gene sets related to cell cycle and cell proliferation are downregulated in kidneys of PIO-treated PCK rats.
Both EGF and PDGF pathways were downregulated by PIO treatment (Table 3(a)). In PKD cystic epithelial cells, growth factors such as EGF and PDGF activate the Raf/MEK/ERK pathway via receptor binding and tyrosine kinase activation [19][20][21]. Therefore, PIO may ameliorate PKD in PCK rats by inhibiting cell proliferation through suppression of the activity of EGF and PDGF pathways. Further, in PKD patients, several reports show that cystic kidneys have significant levels of apoptosis [22,23]. The JNK pathway is known to have critical roles in cell apoptosis, and JNK is overexpressed in cystic epithelial cells in Pkd1 conditional knockout mice [23,24]. In the current study, the JNK MAPK pathway also was downregulated by PIO treatment. Therefore, PIO may have antiapoptotic effects via inactivation of the JNK pathway.
After searching each gene expression with significant change by PIO treatment, we then focused on Scd1 because it is known to stimulate cell proliferation in cancer cells through phosphorylation of AKT [16][17][18], one of the responsible kinases in cystic cell proliferation in PKD [10,36]. Immunohistochemical analysis demonstrated that Scd1 expression was increased in noncystic tubules in PCK kidneys, and PIOtreatment reduced its overexpression, suggesting that Scd1 may relate to the onset of cell proliferation in initial cyst formation through phosphorylation of AKT. In addition, activation of the cell cycle increases syntheses of phospholipids and cholesterol [37][38][39], and Scd1 controls the balance of saturated and monounsaturated fatty acids,  ). Expression levels were normalized to GAPDH. (b) Renal Scd1 distribution in vehicle-treated (CONT) or PIO-treated SD and PCK rats. Representative kidney sections from vehicle-treated (CONT) or PIO-treated SD and PCK rats were stained with an antibody to Scd1. Mouse IgG2b, κ isotype control antibody, did not show any reaction in the kidney. (c) Ratio of Scd1-positive cysts or noncystic tubules in kidney sections. Positivestained cysts or non-cystic tubules were counted in five random fields of kidney sections obtained from five rats in each group by a naive observer using a 20x objective. ( * * P < 0.01 PCK (CONT) versus PCK (PIO) in noncystic tubules in the kidney section).
regulating the composition of cholesterol esters and phospholipids in cell membrane structure [16]. Therefore, PIO may reduce cell proliferation by the downregulation of Scd1 gene expression not only through reducing AKT signaling activity but also through altering fatty acid synthesis. In abnormal cell proliferation in cancer, Scd1 expression is increased, and the cell proliferation is suppressed by treatment with PPAR-γ agonists, although the changes in Scd1 expression are not always consistent [16,40,41]. On the other hand, in diabetes mellitus with insulin resistance, adipose tissue or skeletal muscle Scd1 expression is decreased and increased by PPAR-γ agonists [42][43][44]. Therefore, the expression level of Scd1 and the effect of PPAR-γ agonists may depend on the disease and/or the state of cell proliferation.
Clinically, increased body weight, oedema, and urinary bladder tumors are concerned as possible side effects of PPAR-γ agonists. Although those phenomena were not observed in both genders of PCK rats in the current PIO treatment, the effect of longer term treatment with different doses will need to be studied carefully. Since ameliorative effects are reported in several animal models of PKD [8][9][10]45], PPAR-γ agonists are thought to be a potential candidate for therapeutic interventions in both ARPKD and ADPKD patients.

Conclusions
In the current study, PIO reduced PKD progression and altered the expression of renal genes involved in cell proliferation, cell cycle progression, and fatty acid metabolism in an orthologous rat model of human ARPKD. In addition to the previously demonstrated inhibition of Raf/MEK/ERK and AKT/mTOR/S6 signaling pathways by treatment of PCK rats with 10 mg/kg PIO for 16 weeks [10], suppression of cell proliferation may also be related to reductions in EGF, PDGF, and JNK pathways, cell cycle arrest related to Brca2, BUB1B, Cdkn1c, Cdkn2b, Chek1, and PLK1 genes, and alteration of fatty acid metabolism related to Scd1.