The expression profiling and ontology analysis of noncoding RNAs in peritoneal fibrosis induced by peritoneal dialysis fluid.

GENE-40380; No. of pages: 10; 4C: 3, 7 Gene xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Gene journal homepage: www.elsevier.com/locate/gene

Research paper

The expression profiling and ontology analysis of noncoding RNAs in peritoneal fibrosis induced by peritoneal dialysis fluid Yanli Liu a,1, Runsheng Guo b,1, Guojun Hao a, Jun Xiao a, Yi Bao a, Jing Zhou a, Qinkai Chen a,⁎, Xin Wei a,⁎ a b

Department of Nephrology, The First Affiliated Hospital of Nanchang University, Nanchang 330006, China Department of osteology, The First Affiliated Hospital of Nanchang University, Nanchang 330006, China

a r t i c l e

i n f o

Article history: Received 21 February 2015 Received in revised form 24 March 2015 Accepted 25 March 2015 Available online xxxx Keywords: Peritoneal fibrosis Long noncoding RNA MicroRNA Microarray analysis

a b s t r a c t Increasing amounts of evidence have indicated that noncoding RNAs (ncRNAs) have important regulatory potential in various biological processes. However, the contributions of ncRNAs, especially long noncoding RNAs (lncRNAs), to peritoneal fibrosis remain largely unknown. The aim of this study was to investigate miRNA, lncRNA and mRNA expression profiles and their potential roles in the process of peritoneal fibrosis. Microarray expression profiles of the miRNAs, lncRNAs and mRNAs were determined in normal control peritoneum and in a mouse model of peritoneal dialysis fluid (PDF)-induced fibrotic peritoneum. Differential expression, pathway and gene network analyses were developed to identify possible functional RNA molecules in peritoneal fibrosis. Compared to the normal control, 232 lncRNAs (127 up-regulated and 105 down-regulated), 154 mRNAs (87 upregulated and 67 down-regulated) and 15 miRNAs (14 miRNAs up-regulated and 1 down-regulated) were differentially expressed in the fibrotic peritoneum. Among the differentially expressed ncRNAs, 9 lncRNAs and 5 miRNAs were validated by real-time RT-PCR. Pathway analysis showed that the Jak-STAT, TGF-beta and MAPK signaling pathways had a close relationship with peritoneal fibrosis. Gene co-expression network analysis identified many genes, including JunB, HSP72, and Nedd9. It also identified lncRNAs AK089579, AK080622, and ENSMUST00000053838 and miRNAs miR-182 and miR-488. All of these species potentially play a key role in peritoneal fibrosis. Our results provide a foundation and an expansive view of the roles and mechanisms of ncRNAs in PDF-induced peritoneal fibrosis. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Peritoneal fibrosis is a major and severe complication in patients receiving continuous ambulatory peritoneal dialysis (CAPD). Long-term exposure to peritoneal dialysis fluid (PDF) is a cause of progressive peritoneal fibrosis, resulting in reduced ultrafiltration across the peritoneal membrane that ultimately leads to withdrawal from CAPD in many patients (Yanez-Mo et al., 2003). Histological changes in the peritoneum in response to long-term PD treatment include a loss of the mesothelial cell monolayer, excessive deposition of extracellular matrix, and angiogenesis (Mateijsen et al., 1999; Williams et al., 2002). However, the mechanisms and effective anti-fibrosis therapies of peritoneal fibrosis induced by PDF remain largely undefined.

Abbreviations: CAPD, continuous ambulatory peritoneal dialysis; PDF, peritoneal dialysis fluid; ncRNAs, non-coding RNAs; miRNAs, microRNAs; lncRNAs, long non-coding RNAs; ceRNA, competitive endogenous RNAs; MREs, miRNA response elements; EMT, epithelialto-mesenchymal transition. ⁎ Corresponding authors. E-mail addresses: [email protected] (Q. Chen), [email protected] (X. Wei). 1 These authors contributed equally to this study.

Non-coding RNAs (ncRNAs) are a type of RNA that do not code for proteins. They were previously regarded as “transcriptional noise” but are now known to have important regulatory potential in transcription and post-transcription. The most studied ncRNAs are the microRNAs (miRNAs), which are typically ~22 nt nucleotides long and negatively regulate target messenger RNAs (mRNAs) through the induction of mRNA degradation or translational inhibition. Accumulating evidence shows that miRNAs regulate diverse biological processes, including cell differentiation, proliferation, and apoptosis, and that aberrant expression of miRNAs may lead to severe diseases (Ambros, 2004; Krol et al., 2010). The miRNAs miR-21, miR-30a, miR-155 and miR-29 have been confirmed to be involved in fibrosis (Wang et al., 2012; Bhattacharyya et al., 2013; Yamada et al., 2013; Zhou et al., 2013). Furthermore, miR-29 and miR-30a were reported to have the ability to ameliorate peritoneal fibrosis in animal models of PD (Zhou et al., 2013; Yu et al., 2014), However, more research is needed to elucidate the complicated relationship between miRNAs and peritoneal fibrosis. Long non-coding RNAs (lncRNAs), which are defined as ncRNAs ranging in length from 200 nt to ~ 100 kb, have become an area of increased research focus (Lee, 2012). They have been shown to exert comprehensive effects on biological processes through a variety of mechanisms (Mercer et al., 2009; Hu et al., 2012) and are thought to

http://dx.doi.org/10.1016/j.gene.2015.03.050 0378-1119/© 2015 Elsevier B.V. All rights reserved.

Please cite this article as: Liu, Y., et al., The expression profiling and ontology analysis of noncoding RNAs in peritoneal fibrosis induced by peritoneal dialysis fluid, Gene (2015), http://dx.doi.org/10.1016/j.gene.2015.03.050

2

Y. Liu et al. / Gene xxx (2015) xxx–xxx Table 1 Primers for selected lncRNA. LncRNA

Forward primer

Reverse primer

uc007eib.1 ENSMUST00000053838 AK142426 uc008pwj.1 uc007dlv.1 AK080622 BC049991 AV310809 AK089579

TGGTGTTCTGCGTGATGAGG ACCAACCTGCTTGTCAGAG TGGTGTTCTGCGTGATGAGG GGAGTTGGGAAGAGTGTGTC GAAAGTCTTCTGGGTTGGG TTCACTTGGAACCTCAGGC TCTCCTTCATCCCTCTCCA CCACTAACCTTCCCTATTACAT GAGTGTCAGTTGGTGGAAAC

GAATAACACTTCCACCACAGG CCTTGCGATTCTATTAGTCAGC GCTGTGTCTCAGTTCAAGTTAGC GATGTAGTAACGGTAGGCAGC GTCAAGCAGCAAGCCTTAG CACTAACTGCTCCTTCGGTCTA TGAGCACTCTACCAACGGA AACTAATGGGTGCTGTGC CTTCTTACCTGTGTGTTCTGC

play important roles in the pathophysiology of several diseases, including cancer, pulmonary fibrosis and cardiovascular disease (Cao et al., 2013; Yang et al., 2013; Liu et al., 2014). Moreover, in recent studies, lncRNAs have been discovered that can act as competitive endogenous RNAs (ceRNA) of miRNAs by sharing common miRNA response elements (MREs), inhibiting miRNA activity and attenuating the repression of miRNA-target genes (Ebert and Sharp, 2010; Poliseno et al., 2010; Cesana et al., 2011). Although lncRNA studies predominate in other fields, such as cancer biology, studies exploring the signature of lncRNA expression, the possible roles of lncRNAs, and the relationships with miRNAs and mRNAs in fibrosis biology remain limited. In this study, the expression profiles of peritoneal lncRNAs, miRNAs and mRNAs were compared between normal control animals and a mouse model of peritoneal fibrosis induced by PDF. An integrative analysis combining the changes in the three groups of RNAs within different

genetic networks was used to identify genes and pathways related to peritoneal fibrosis.

2. Materials and methods 2.1. Animals BALB/c mice (male, 6–8 weeks of age) were provided by the Nanchang University Experimental Animal Center, Nanchang, China. A total of 36 BALB/c mice were randomly divided into 2 groups (n = 18 in each group): a normal control group and a peritoneal fibrosis model group. All animal experiments were carried out in accordance with the principles of the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Fig. 1. Pathological changes of the peritoneum in a mouse model of PD. Severe peritoneal fibrosis developed in the mouse model of PD. Compared to the normal control (A: HE staining and B: Masson's trichrome staining), the peritoneum in the model group (C: HE staining and D: Masson's trichrome staining) was markedly thickened and showed remarkable proliferation of the collagen fibers. E: The peritoneum was significantly thicker in the model group. F, G: Western blot analysis showed that E-cadherin was decreased. In contrast, the expression level of α-SMA, collagen I and fibronectin were increased in the peritoneum of the mouse model of PD. NPT, normal peritoneal tissue; FPT, fibrosis peritoneal tissue; FN, fibronectin. *P b 0.05 when compared with NPT. **P b 0.01 when compared with NPT. Original magnification, ×200 (A–D).


Y. Liu et al. / Gene xxx (2015) xxx–xxx

3

2.2. Peritoneal fibrosis model

2.4. Western blot analysis

The mice in peritoneal fibrosis model group were administered with 3–4 ml of 4.25% dextrose PDF (Baxter HealthCare, Deerfield, IL, USA) daily for 4 weeks through peritoneal injection (Nie et al., 2007). The normal control mice were untreated group. Peritoneal tissues were harvested on the 28th day following the treatment with PDF.

Protein from the peritoneal tissues was extracted as previously reported (Qin et al., 2011). The primary antibodies used in this study included α-SMA (abcam, USA) and collagen I, E-cadherin, fibronectin, and GAPDH (R&D, USA). Immunoreactive bands were visualized by the chemiluminescence dissolvent (Thermo, USA) and exposed to X-ray film (Kodak, USA). The relative protein levels of collagen I, E-cadherin, α-SMA, and fibronectin were normalized to GAPDH, and then, the ratio was compared with that of the normal control group.

2.3. Hematoxylin–eosin and Masson's Trichrome Staining After fixation in 4% paraformaldehyde overnight, dehydration in 70% ethanol and clearing in xylene, the peritoneal tissues were embedded in paraffin wax. Sections (4 μm) were prepared and stained with hematoxylin–eosin or Masson's trichrome stain. Nine random areas were examined at a magnification of ×400. The severity of the fibrosis was blindly determined by a semi-quantitative assay.

2.5. LncRNA microarray Briefly, samples (3 fibrotic peritoneal tissues and 3 normal peritoneal tissues) were used to synthesize double-stranded complementary DNA (cDNA). The double-stranded cDNA was labeled and hybridized

Fig. 2. Alterations in the RNA expression profiles between PD induced fibrotic peritoneal tissues and normal peritoneal tissues. Heat maps and hierarchical clustering of expression ratios (log2 scale) of lncRNAs (a), mRNAs (b) and miRNAs (c) in mouse peritoneal tissues. “Red” denotes a high relative expression and “blue” denotes a low relative expression. FPT, fibrotic peritoneal tissue; and NPT, normal peritoneal tissue.


4


to a 8 × 60 K Arraystar Mouse LncRNA Microarray v2.0 (Arraystar, Rockville, MD). The lncRNA expression microarray used in this study mainly classifies its probes according to the following subtypes. 1) Enhancer LncRNAs contain profiling data for all LncRNAs with enhancer-like function (Harrow et al., 2006). 2) HOX cluster probes contain profiling data of all probes in the four HOX loci, targeting 407 discrete transcribed regions, lncRNAs and coding transcripts (Rinn et al., 2007). 3) Rinn lincRNAs contain profiling data of all lincRNAs based on John Rinn's work (Guttman et al., 2009). 4) Enhancer LncRNAs near the coding gene contain the differentially expressed enhancer-like LncRNAs and their nearby coding genes (distance, 300 kb). 5) LincRNAs near the coding gene contain the differentially expressed lincRNAs and nearby coding gene pairs (distance, 300 kb). After hybridization and washing, the processed slides were scanned with the Axon GenePix 4000B microarray scanner (Molecular Devices, Sunnyvale, CA). Raw data were extracted as pair files using the NimbleScan software (version 2.5, Roche NimbleGen, Inc., Madison, WI). NimbleScan software's implementation of RMA offers quantile normalization and background correction. 2.6. MicroRNA microarray Microarray profiling for miRNA was performed using miRCURY™ LNA Arrays (v.18.0, Exiqon, Vedbaek, Denmark) according to the manufacturer's protocol. Briefly, 1 mg of total RNA from the tissues was labeled by polyA polymerase using the miRCURY™ Hy3™/Hy5™ Power labeling kit (Exiqon, Vedbaek, Denmark) according to the manufacturer's recommendations. RNA was hybridized to the miRCURY™ LNA Arrays as described in the array manual. Following hybridization, the slides were prepared, washed several times using a wash buffer kit (Exiqon), and dried by centrifugation for 5 min at 400 rpm. Then, the slides were scanned using the Axon GenePix 4000B microarray scanner (Axon Instruments, Foster City, CA).

Table 2 Selected up/down-regulated lncRNAs and miRNAs in the fibrotic peritoneal tissue. RNA length (bp) LncRNAs (seqname) Up-regulated AV310809 uc007eib.1 AK142426 ENSMUST00000053838 Down-regulated uc007dlv.1 AK080622 AK089579 BC049991 uc008pwj.1 miRNAs Up-regulated miR-182-5p miR-488-5p miR-296-3p miR-292-5p Down-regulated miR-200a-3p

Chromosome

Log2-fold change

325 981 409 363

Chr6 Chr10 Chr5 Chr10

5.19 4.69 3.66 3.55

259 1561 1710 2484 305

Chr1 Chr2 Chr6 Chr1 Chr3

4.79 4.77 4.02 3.97 3.77

25 21 22 21

Chr6 Chr1 Chr2 Chr7

11.64 8.54 5.37 4.47

21

Chr4

3.31

10 min at 95 °C, followed by 40 cycles of 95 °C for 15 s and 60 °C for 60 s. The relative amount of miRNA used was normalized against U6 snRNA, and the fold change for each miRNA was calculated by the 2−ΔΔCt method.

2.7. Realtime RT-PCR To validate the lncRNA microarray data, the expression level of 9 lncRNAs were analyzed using real-time RT-PCR. The 9 lncRNAs were selected according to the following criteria: 1) the top ten lncRNAs with the highest fold change (ten for each up- and down-regulated cluster, all clearly dysregulated) were included.; 2) those with nearby genes involved in biological processes, such as proliferation, apoptosis, differentiation and tumorigenesis, were included; 3) lncRNAs located in the sex chromosome were excluded. The total RNA was extracted from peritoneal tissues using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions. The expression patterns of the 9 selected lncRNAs were measured by real-time PCR using SYBRGreen assays (Transgene, Beijing, China). The primers used are listed in Table 1. The thermocycle conditions were as follows: 50 °C at 2 min for a hot-start, predenaturation at 95 °C for 30 s. and PCR amplification for 40 cycles at 95 °C for 15 s. and at 60 °C for 30 s. Relative quantification of the gene expression was performed using 2−ΔΔCt methods based on the CT values for both the target and reference genes (Livak and Schmittgen, 2001). Three independent experiments were performed, and the results were given as means ± SD. 2.8. QRT-PCR of miRNA To validate the miRNA microarray data, the expression levels of 5 miRNAs, which had the highest fold change, were detected by qRTPCR. The total miRNA were isolated from peritoneal tissues using the mirVana miRNA isolation kit (Ambion, Austin, TX, USA). Primers for the 5 selected miRNAs were purchased from Applied Biosystems. Expression of the selected miRNAs was detected by the TaqMan microRNA assay Kit (Applied Biosystems, Foster City, CA, USA), according to the manufacturer's protocol. The PCR conditions were 2 min at 50 °C and

Fig. 3. The differential expression of lncRNAs and miRNAs was validated by quantitative real-time PCR. The relative amount of each lncRNA (A) was normalized to 18 s rRNA, and each miRNA (C) was normalized to U6 snRNA. FPT, fibrotic peritoneal tissue; and NPT, normal peritoneal tissue. *P b 0.05, **P b 0.01 when compared with NPT (N = 15 in each group).



2.9. Data analysis 2.9.1. Significant differential gene analysis Agilent Feature Extraction software (version 10.7.3.1) was used to analyze the acquired array images. Quantile normalization and subsequent data processing were performed using the GeneSpring GX v11.5.1 software package (Agilent Technologies, Inc.). Differentially expressed genes were identified through the random variance model (Wright and Simon, 2003). A P value was calculated using the paired t-test. The threshold set for the up- and down-regulated genes was a fold change ≧ 2.0 and a P value b 0.05. Hierarchical clustering was performed based on the differentially expressed mRNAs, lncRNAs and miRNAs using Cluster_Treeview software from Stanford University (Palo, Alto, CA).

2.9.2. Pathway enrichment analysis Pathway analysis was used to identify significant pathways for the differentially expressed mRNAs according to the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. The Fisher's exact and EASEscore tests were used to select significant pathways (Hosack et al., 2003). The threshold for significant pathways was a P value b0.05.

5

2.9.3. MicroRNA target prediction Target mRNAs of the miRNAs were predicted based on the following software: TargetScan version 5.2 (http://www.targetscan.org/) PicTar 2005 (http://pictar.mdc-berlin.de/cgi-bin/PicTar_vertebrate. cgi). mRNAs were selected for co-expression network establishment when it was involved in both the prediction results and the differential expression gene results from the microarray data. 2.9.4. Co-expression analysis Gene co-expression networks were presented to identify interactions among genes. The method was performed as previously described (Yang et al., 2011, 2013). LncRNA–mRNAs were predicted by correlation analysis. In brief, LncRNA–mRNA co-expression pairs were based on the normalized signal intensity of microarray data. Pearson's correlation coefficients (PCCs) were used to calculate the LncRNA–mRNA pair correlation. To ensure a visual representation, only the strongest correlations (0.99 or greater) were drawn in these renderings. miRNA–mRNAs were predicted by TargetScan and PicTar software, the most widely

Fig. 4. Signaling pathways of differentially expressed RNAs. Signaling pathways of differentially expressed up-regulated mRNAs (A) and down-regulated mRNAs (B). Pathway analysis was predominantly based on the KEGG database. P-value b0.05 using the two-sided Fisher's exact test were classed as being significant. The vertical axis represents the pathway category and the horizontal axis represents the −log 10 (p value) of these significant pathways. Schematic diagram of the “MAPK signal pathway” shows all genes identified from differentially expressed mRNAs (C). Yellow marked nodes are associated with up-regulated genes; red marked nodes are associated with down-regulated genes; green nodes have no significance. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


6


Fig. 4 (continued).

used methods for miRNA target gene prediction. Each gene corresponds to a node in this network. Two genes connected by an edge indicate a strong correlation. The edges between two nodes were connected by different mechanisms. Within the network analysis, a degree is the simplest and most important measure of the centrality of a gene within a network, and it determines the relative importance. A degree is defined as the number of directly linked neighbors. Cytoscape software (v2.8.3) was used for drawing the gene co-expression network.

3. Results 3.1. Detection of PDF-induced peritoneal fibrosis in mouse model Mice were divided into the model group and normal control group according to a daily intraperitoneal injection of 4.25% dextrose PDF or saline. H&E and Masson's trichrome staining of peritoneum sections were used to evaluate the degree of fibrosis based on morphological changes. In the normal control group, a monolayer of mesothelium cells covering the surface of the peritoneum was observed. In contrast, the peritoneal tissue in the model group was markedly thickened and showed a remarkable proliferation of collagen fibers (Fig. 1A). In addition, western blot analysis showed that the peritoneal tissue in the model group had a loss of epithelial markers, such as E-cad, and a gain of the mesenchymal marker α-SMA. These changes were accompanied by a massive collagen matrix accumulation when compared with peritoneal tissue in the normal control group (Fig. 1B). These results

indicated that we successfully established a model of PDF-induced peritoneal fibrosis.

3.2. Expression of noncoding RNAs in the peritoneum To understand whether and which non-coding RNAs are involved in peritoneal fibrosis induced by PDF, microarray studies were performed in the model vs. normal control group (with three samples per group). An arraystar probe dataset, including 31,423 lncRNAs and 25,376 coding transcripts, was used to assess the lncRNAs and mRNA expression profiles. Up to 232 lncRNAs were differentially expressed in the PDFinduced peritoneal fibrosis group compared to the normal control group. Of these, 127 lncRNAs were up-regulated and 105 lncRNAs were down-regulated. In addition, a total of 154 mRNAs were differentially expressed: 87 mRNAs were up-regulated and 67 mRNAs were down-regulated. The fold change threshold was ≧2.0, and the P value was b 0.05 (Supplement data). The miRCURY™ LNA gene chip, which contained 3100 capture probes covering all human, mouse and rat mRNAs annotated in miRBase 18.0, was used to assess the miRNA expression profile in the peritoneal tissues. The results showed that in fibrotic peritoneal tissues, 14 miRNAs were up-regulated and 1 miRNA was down-regulated compared to normal peritoneal tissues (Supplement data). The fold change threshold was ≧ 2.0. Hierarchical clustering showed systematic variations in the expression of non-coding RNAs (lncRNAs and miRNAs) and proteincoding RNAs between the model and normal control group (Fig. 2).



3.3. Validation of dysregulated non-coding RNAs using qPCR After microarray analysis, we selected 9 lncRNAs and 5 miRNAs (Table 2) thought to play an important role in peritoneal fibrosis (according to the selection criteria presented in Section 2) for qRT-PCR validation of their differential expression levels in the model group vs. the normal control group (15 independent samples for each group). For the lncRNAs, AV310809, uc007eib.1, AK142426 and ENSMUST00000053838 were increased, while uc007dlv.1, AK080622, AK089579, BC049991 and uc008pwj.1 were decreased in the fibrotic peritoneum (all P b 0.05; Fig. 3A). For miRNAs, the expression levels of miR-182, miR-488, miR-292 and miR-296 were up-regulated, while miR-200a was down-regulated in the model group compared to the normal control group (all P b 0.05; Fig. 3B). All these results were consistent with the microarray data.

3.4. Microarray-based pathway analysis Pathway analysis was used to determine the biological pathways associated with the most differentially expressed mRNAs in peritoneal fibrosis. Up to 17 pathways were found to be involved in the differentially expressed mRNAs (Fig. 4). Among these, 10 pathways, including pathways linked to transcriptional misregulation in cancer, insulin secretion, Jak-STAT signaling, cytokine–cytokine receptor signaling, and MAPK signaling, corresponded to up-regulated mRNAs (Fig. 4A). In contrast, significant pathways corresponding to down-regulated mRNAs appeared to be responsible for prion diseases, protein processing in the endoplasmic reticulum, axon guidance, the TGF-β signaling pathway and the MAPK signaling pathway (Fig. 4B).

7

Among these pathways, the MAPK signaling pathway, which was reported to play an important role in renal and cardiac fibrosis, corresponded to both up- and down-regulated mRNAs (Fisher P = 0.011 and 0.041, respectively). Fifteen mRNAs that were differentially expressed in the fibrotic peritoneum were involved in the MAPK signaling pathway (Fig. 4C). Of these mRNAs, CACN, PKA, Rap1, IL1R, Mos, GCK, MUK, NFκB, cPLA2, Sapla, PTP and JIP1/2 were up-regulated, while HSP72, c-fos and GADD153 were down-regulated. 3.5. Establishment of gene co-expression network Gene co-expression network analysis was undertaken to investigate the potential interaction relationship between no-coding RNAs (lncRNAs and miRNAs) and coding genes. Selected non-coding RNAs (9 lncRNAs and 5 miRNAs) and 104 mRNAs that had the strongest correlation with the selected lncRNAs (PCC ≧ 0.99) or predicted as the target genes of selected miRNAs were presented in the network graph (Fig. 5). Blue marked nodes represent lncRNAs, red marked nodes represent miRNAs, and yellow marked nodes represent mRNAs. Solid edges represent the positive correlation between lncRNAs and mRNAs. Dotted edges represent the negative correlation between lncRNAs and mRNAs. T edges represent the negative regulation between miRNAs and their target mRNAs. The relative importance of a gene in this network was qualified by degrees. The higher the degree, the more centrally and the larger the gene is depicted within the network. This analysis indicated that AK089579, AK080622, and ENSMUST00000053838 potentially play a prominent role in the network, as 26, 26 and 27 mRNAs were targeted by these three lncRNAs, respectively. Ten mRNAs (Amt, Ascl3, Dok2, HSP72, Ier2, JunB, Mcam, Zbtb46, Nedd9 and Creld2) were co-regulated by the three lncRNAs, simultaneously.

Fig. 5. The LncRNA–miRNA–mRNA–network. The network represents 9 highly dysregulated lncRNAs and 5 highly dysregulated miRNAs that were co-expressed with 104 coding mRNAs. Red marked nodes are lncRNAs, yellow marked nodes are miRNAs, and green marked nodes are mRNAs. Solid edges represent the positive correlation between lncRNAs and mRNAs. Dotted edges represent the negative correlation between lncRNAs and mRNAs. T edges represent negative regulation between miRNAs and their target mRNAs.


8


In addition to the lncRNA–mRNA pairs, several differentially expressed mRNAs were predicted to be the target genes of the 5 highest expressed dysregulated miRNAs. Among these target genes, some were targeted by multiple miRNAs. For example, Ctdsp1 was targeted by both miR-182 and miR-488, and Zbtb46 was targeted by miR-488 and miR-200a. In this network, we also found that one mRNA can be co-regulated by lncRNAs and miRNAs. For instance, Nedd9, which was highly correlated with 5 lncRNAs (AK089579, AK080622, ENSMUST00000053838, uc008pwj.1 and AV310809), also predicted to be a target gene of miR-182. Ctdsp1 was targeted by 4 lncRNAs (AK089579, AV310809, uc007dlv.1 and AK142426) and 2 miRNAs (miR-182 and miR-488) simultaneously. Many targets mentioned here have been reported to be linked to cell proliferation or tissue fibrosis (Wang et al., 2006; Rajasekaran et al., 2012; Xiao et al., 2012; Deng et al., 2013; Ponticos et al., 2015). 4. Discussion As a renal replacement therapy for patients with end-stage renal failure, PD has been used worldwide, particularly in developing countries. However, the appearance of peritoneal fibrosis induced by PDF has limited the clinical use of PD. Emerging evidence shows that a set of miRNAs is involved in the mechanism for the development of peritoneal fibrosis (Zhou et al., 2013; Yu et al., 2014). However, the profile and the biological function of non-coding RNAs, especially lncRNAs, remain largely unclear. In the present study, we performed microRNA and lncRNA microarray analyses to evaluate the expression profiles of non-coding RNAs in the formation of fibrotic peritoneum. In the present study, we evaluated the expression profiles of lncRNAs, miRNAs and mRNAs in the peritoneum of a mouse model of peritoneal fibrosis for the first time. Compared to normal mice, we identified a set of differentially expressed RNA molecules, including 232 lncRNAs, 15 miRNAs and 154 mRNAs, expressed in the fibrotic peritoneum induced by PDF. An integrative method including pathway and coexpression network construction was developed to identify possible functional relationships between the different RNA molecules. Based on the differentially expressed mRNAs, pathway analysis was carried out to predict which biological functions and mechanisms were involved in the PDF-induced peritoneal fibrosis. We found that many important pathways, such as transcriptional misregulation in cancer, cytokine–cytokine receptor interaction, TGF-beta signaling, Jak-STAT signaling and MAPK signaling were abundant among the significantly enriched mRNAs. Most of these pathways have previously been reported to be involved in the process of tissue fibrosis. Pang reported that a selective inhibitor of STAT3 can attenuate renal fibrosis by inactivating interstitial fibroblasts in an obstructive nephropathy model (Pang et al., 2010). In addition, LIF, one of the cytokines that could induce the Jak-STAT pathway, has been shown to promote tubular cell proliferation and fibrosis in a model of experimental renal failure (Yoshino et al., 2003). These observations suggested that the Jak-STAT signaling pathway plays a role in renal fibrosis; however, the potential role of Jak-STAT signaling in peritoneal fibrosis has not been reported until now. TGF-beta is a central mediator of fibrogenesis. Studies in a wide range of experimental models have demonstrated the involvement of the TGF-beta signaling pathway in fibrosis (Fukasawa et al., 2004; Bujak et al., 2007; Ask et al., 2008). In peritoneal fibrosis, Dou found that the expression of TGF-beta1, Smad 3 and phosphorylated-Smad2/ 3 was up-regulated in the fibrotic peritoneum induced by high glucose peritoneal dialysate (Dou et al., 2005). They concluded that a high concentration glucose dialysate can activate the TGF-beta/Smad signaling pathway and induce peritoneal fibrosis. Among these associated pathways, the MAPK signaling pathway showed significant changes in both its up-regulated and downregulated mRNAs. Fifteen differentially expressed mRNAs were involved in the pathway, 12 of which were up-regulated and 3 of which were down-regulated. These results indicated that MAPK signaling

pathway may play an important role in the mechanisms of peritoneal fibrosis, which is consistent with previous reports. Aoki S found that fluid flow stress promotes hyperplasia and EMT in peritoneal mesothelial cells via the MAPK axis, suggesting that MAPK signaling may be involved in the pathogenesis of peritoneal fibrosis (Aoki et al., 2011). Recent studies have indicated that FR167653, a specific inhibitor of p38/MAPK, can reduced the number of cells positive for phosphorylated p38/MAPK, down-regulate the peritoneal expression of CCL2, a chemoattractant for fibrocytes, and attenuate the extent of peritoneal fibrosis (Kokubo et al., 2012). Further investigation of MAPK signaling may lead to a novel therapeutic approach for peritoneal fibrosis. In this study, a gene co-expression network was constructed to further investigate the relationship between lncRNAs, miRNAs and mRNAs in peritoneal fibrosis induced by PDF. In this gene co-expression network, 9 lncRNAs, 5 miRNAs and 104 mRNAs were involved. The results indicated that AK089579, AK080622, and ENSMUST00000053838 potentially play a key role in the network, as they were connected to the greatest number of adjacent genes and had the largest degrees. Ten mRNAs were targeted by the three lncRNAs. Among the targeted mRNAs, Dok2, Ier3, HSP72, Junb and Nedd9 were reported to be involved in the biological processes of the epithelial-to-mesenchymal transition (EMT), inflammation response, cell proliferation and tissue fibrosis (Wakisaka et al., 2007; Arlt and Schafer, 2011; Gunthner et al., 2013; Morimoto et al., 2014). HSP72, which was also targeted by another three lncRNAs (BC049991, AV310809, and AK080622) in the network and which plays a part in the MAPK signaling pathway, was closely related to tissue fibrosis. Takahashi reported that HSP72 can attenuate atrial fibrosis in three different experimental models of atrial fibrillation (Takahashi et al., 2012). Studies in obstructive nephropathy show that HSP72 can markedly reduce cell proliferation in the tubular epithelium, and can decrease interstitial fibroblast accumulation and collagen I deposition. Selective HSP72 expression inhibited EMT caused by TGFbeta1 in renal NRK52E cells (Mao et al., 2008). JunB, which was targeted by six lncRNAs (ENSMUST00000053838, AK089579, AK080622, uc007dlv.1, AV310809) in this network, was regarded as a member of the activator protein-1 (AP-1) family and was found to regulate many fundamental cell processes, including proliferation, differentiation, apoptosis and responses to stresses. Indeed, it is essential for many physiological functions at the whole organism level (Bakiri et al., 2000; Passegue and Wagner, 2000; Passegue et al., 2004). Recent studies have demonstrated that JunB accumulation due to altered mTOR/AKT signaling resulted in the over-expression of collagen and the progression of fibrosis in Scleroderma cases (Ponticos et al., 2015). Nedd9 is a Crk-associated substrate (Cas) family protein that plays a crucial role in the TGF-bate triggered epithelial–mesenchymal transition (EMT), an important mechanism in tissue repair and fibrosis. In our co-expression network, we found that Nedd9 was targeted by 5 lncRNAs (ENSMUST00000053838, AK089579, uc008pwj.1, AK080622, AV310809) and miR-182 simultaneously. Therefore, we speculated that one or several of these six ncRNAs may participate in peritoneal fibrosis by interacting with EMT through Nedd9. In addition, the network analysis results suggested that a single mRNA can be regulated by multiple lncRNAs and that lncRNAs can coregulate multiple target mRNAs. Furthermore, mRNAs can be coregulated by different lncRNAs in combination with different miRNAs. This phenomenon indicates that lncRNAs, miRNAs and mRNAs form a complicated regulation network in peritoneal fibrosis. However, the mechanism and function of these genes, and especially lncRNAs, in the network are still unclear. The study of this network could provide a more accurate direction for further investigation. In conclusion, the present study represents the first instance of microarray data being used to systematically and comprehensively analyze lncRNA and miRNA expressions in a mouse model of PDF-induced peritoneal fibrosis and a normal control group. Many of differentially expressed ncRNAs may play an important role in regulating peritoneal fibrosis through various mechanisms, including the MAPK, TGF-beta,



and Jak-STAT signaling pathways, by interacting with a series mRNAs. Based on these data, further studies of these differentially expressed genes' functions may help elucidate their specific roles in peritoneal fibrosis and should be performed. Further studies on ncRNAs will help expand our understanding of the genomic regulator networks in peritoneal fibrosis and may provide potential therapeutic targets for the treatment of peritoneal fibrosis. Disclosure The authors of this manuscript have no conflicts of interest to disclose. Acknowledgments Bioinformatics analysis was performed by Kangchen Biotech Co. Ltd., Shanghai, China. This study was supported by the National Natural Science Foundation of China (no. 81260120) and the Natural Science Foundation of the Jiangxi Province (no. 20122BAB215004). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.gene.2015.03.050. References Ambros, V., 2004. The functions of animal microRNAs. Nature 431, 350–355. Aoki, S., Makino, J., Nagashima, A., Takezawa, T., Nomoto, N., Uchihashi, K., Matsunobu, A., Sanai, T., Sugihara, H., Toda, S., 2011. Fluid flow stress affects peritoneal cell kinetics: possible pathogenesis of peritoneal fibrosis. Perit. Dial. Int. 31, 466–476. Arlt, A., Schafer, H., 2011. Role of the immediate early response 3 (IER3) gene in cellular stress response, inflammation and tumorigenesis. Eur. J. Cell Biol. 90, 545–552. Ask, K., Bonniaud, P., Maass, K., Eickelberg, O., Margetts, P.J., Warburton, D., Groffen, J., Gauldie, J., Kolb, M., 2008. Progressive pulmonary fibrosis is mediated by TGF-beta isoform 1 but not TGF-beta3. Int. J. Biochem. Cell Biol. 40, 484–495. Bakiri, L., Lallemand, D., Bossy-Wetzel, E., Yaniv, M., 2000. Cell cycle-dependent variations in c-Jun and JunB phosphorylation: a role in the control of cyclin D1 expression. EMBO J. 19, 2056–2068. Bhattacharyya, S., Kumar, P., Tsuchiya, M., Bhattacharyya, A., Biswas, R., 2013. Regulation of miR-155 biogenesis in cystic fibrosis lung epithelial cells: antagonistic role of two mRNA-destabilizing proteins, KSRP and TTP. Biochem. Biophys. Res. Commun. 433, 484–488. Bujak, M., Ren, G., Kweon, H.J., Dobaczewski, M., Reddy, A., Taffet, G., Wang, X.F., Frangogiannis, N.G., 2007. Essential role of Smad3 in infarct healing and in the pathogenesis of cardiac remodeling. Circulation 116, 2127–2138. Cao, G., Zhang, J., Wang, M., Song, X., Liu, W., Mao, C., Lv, C., 2013. Differential expression of long non-coding RNAs in bleomycin-induced lung fibrosis. Int. J. Mol. Med. 32, 355–364. Cesana, M., Cacchiarelli, D., Legnini, I., Santini, T., Sthandier, O., Chinappi, M., Tramontano, A., Bozzoni, I., 2011. A long noncoding RNA controls muscle differentiation by functioning as a competing endogenous RNA. Cell 147, 358–369. Deng, X., Xu, M., Yuan, C., Yin, L., Chen, X., Zhou, X., Li, G., Fu, Y., Feghali-Bostwick, C.A., Pang, L., 2013. Transcriptional regulation of increased CCL2 expression in pulmonary fibrosis involves nuclear factor-kappaB and activator protein-1. Int. J. Biochem. Cell Biol. 45, 1366–1376. Dou, X.R., Yu, X.Q., Li, X.Y., Chen, W.F., Hao, W.K., Jia, Z.J., Peng, W.X., Wang, X., Yin, P.D., Wang, W.J., Zheng, Z.H., 2005. The role of TGF-beta1/Smads in the development of peritoneal fibrosis induced by high glucose peritoneal dialysate and LPS. Zhonghua Yi Xue Za Zhi 85, 2613–2618. Ebert, M.S., Sharp, P.A., 2010. MicroRNA sponges: progress and possibilities. RNA 16, 2043–2050. Fukasawa, H., Yamamoto, T., Suzuki, H., Togawa, A., Ohashi, N., Fujigaki, Y., Uchida, C., Aoki, M., Hosono, M., Kitagawa, M., Hishida, A., 2004. Treatment with anti-TGF-beta antibody ameliorates chronic progressive nephritis by inhibiting Smad/TGF-beta signaling. Kidney Int. 65, 63–74. Gunthner, R., Kumar, V.R., Lorenz, G., Anders, H.J., Lech, M., 2013. Pattern-recognition receptor signaling regulator mRNA expression in humans and mice, and in transient inflammation or progressive fibrosis. Int. J. Mol. Sci. 14, 18124–18147. Guttman, M., Amit, I., Garber, M., French, C., Lin, M.F., Feldser, D., Huarte, M., Zuk, O., Carey, B.W., Cassady, J.P., Cabili, M.N., Jaenisch, R., Mikkelsen, T.S., Jacks, T., Hacohen, N., Bernstein, B.E., Kellis, M., Regev, A., Rinn, J.L., Lander, E.S., 2009. Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in mammals. Nature 458, 223–227. Harrow, J., Denoeud, F., Frankish, A., Reymond, A., Chen, C.K., Chrast, J., Lagarde, J., Gilbert, J.G., Storey, R., Swarbreck, D., Rossier, C., Ucla, C., Hubbard, T., Antonarakis, S.E., Guigo,

9

R., 2006. GENCODE: producing a reference annotation for ENCODE. Genome Biol. 7 (Suppl. 1), S4.1–S4.9. Hosack, D.A., Dennis, G.J., Sherman, B.T., Lane, H.C., Lempicki, R.A., 2003. Identifying biological themes within lists of genes with EASE. Genome Biol. 4, R70. Hu, W., Alvarez-Dominguez, J.R., Lodish, H.F., 2012. Regulation of mammalian cell differentiation by long non-coding RNAs. EMBO Rep. 13, 971–983. Kokubo, S., Sakai, N., Furuichi, K., Toyama, T., Kitajima, S., Okumura, T., Matsushima, K., Kaneko, S., Wada, T., 2012. Activation of p38 mitogen-activated protein kinase promotes peritoneal fibrosis by regulating fibrocytes. Perit. Dial. Int. 32, 10–19. Krol, J., Loedige, I., Filipowicz, W., 2010. The widespread regulation of microRNA biogenesis, function and decay. Nat. Rev. Genet. 11, 597–610. Lee, J.T., 2012. Epigenetic regulation by long noncoding RNAs. Science 338, 1435–1439. Liu, Y., Li, G., Lu, H., Li, W., Li, X., Liu, H., Li, X., Li, T., Yu, B., 2014. Expression profiling and ontology analysis of long noncoding RNAs in post-ischemic heart and their implied roles in ischemia/reperfusion injury. Gene 543, 15–21. Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative gene expression data using realtime quantitative PCR and the 2(−Delta Delta C(T)) method. Methods 25, 402–408. Mao, H., Li, Z., Zhou, Y., Li, Z., Zhuang, S., An, X., Zhang, B., Chen, W., Nie, J., Wang, Z., Borkan, S.C., Wang, Y., Yu, X., 2008. HSP72 attenuates renal tubular cell apoptosis and interstitial fibrosis in obstructive nephropathy. Am. J. Physiol. Renal. Physiol. 295, F202–F214. Mateijsen, M.A., van der Wal, A.C., Hendriks, P.M., Zweers, M.M., Mulder, J., Struijk, D.G., Krediet, R.T., 1999. Vascular and interstitial changes in the peritoneum of CAPD patients with peritoneal sclerosis. Perit. Dial. Int. 19, 517–525. Mercer, T.R., Dinger, M.E., Mattick, J.S., 2009. Long non-coding RNAs: insights into functions. Nat. Rev. Genet. 10, 155–159. Morimoto, K., Tanaka, T., Nitta, Y., Ohnishi, K., Kawashima, H., Nakatani, T., 2014. NEDD9 crucially regulates TGF-beta-triggered epithelial–mesenchymal transition and cell invasion in prostate cancer cells: involvement in cancer progressiveness. Prostate 74, 901–910. Nie, J., Dou, X., Hao, W., Wang, X., Peng, W., Jia, Z., Chen, W., Li, X., Luo, N., Lan, H.Y., Yu, X.Q., 2007. Smad7 gene transfer inhibits peritoneal fibrosis. Kidney Int. 72, 1336–1344. Pang, M., Ma, L., Gong, R., Tolbert, E., Mao, H., Ponnusamy, M., Chin, Y.E., Yan, H., Dworkin, L.D., Zhuang, S., 2010. A novel STAT3 inhibitor, S3I-201, attenuates renal interstitial fibroblast activation and interstitial fibrosis in obstructive nephropathy. Kidney Int. 78, 257–268. Passegue, E., Wagner, E.F., 2000. JunB suppresses cell proliferation by transcriptional activation of p16(INK4a) expression. EMBO J. 19, 2969–2979. Passegue, E., Wagner, E.F., Weissman, I.L., 2004. JunB deficiency leads to a myeloproliferative disorder arising from hematopoietic stem cells. Cell 119, 431–443. Poliseno, L., Salmena, L., Zhang, J., Carver, B., Haveman, W.J., Pandolfi, P.P., 2010. A codingindependent function of gene and pseudogene mRNAs regulates tumour biology. Nature 465, 1033–1038. Ponticos, M., Papaioannou, I., Xu, S., Holmes, A.M., Khan, K., Denton, C.P., Bou, G.G., Abraham, D.J., 2015. Failed degradation of JunB contributes to overproduction of type I collagen and development of dermal fibrosis in patients with systemic sclerosis. Arthritis Rheumatol. 67, 243–253. Qin, W., Chung, A.C., Huang, X.R., Meng, X.M., Hui, D.S., Yu, C.M., Sung, J.J., Lan, H.Y., 2011. TGF-beta/Smad3 signaling promotes renal fibrosis by inhibiting miR-29. J. Am. Soc. Nephrol. 22, 1462–1474. Rajasekaran, S., Vaz, M., Reddy, S.P., 2012. Fra-1/AP-1 transcription factor negatively regulates pulmonary fibrosis in vivo. PLoS One 7, e41611. Rinn, J.L., Kertesz, M., Wang, J.K., Squazzo, S.L., Xu, X., Brugmann, S.A., Goodnough, L.H., Helms, J.A., Farnham, P.J., Segal, E., Chang, H.Y., 2007. Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 129, 1311–1323. Takahashi, N., Kume, O., Wakisaka, O., Fukunaga, N., Teshima, Y., Hara, M., Saikawa, T., 2012. Novel strategy to prevent atrial fibrosis and fibrillation. Circ. J. 76, 2318–2326. Wakisaka, O., Takahashi, N., Shinohara, T., Ooie, T., Nakagawa, M., Yonemochi, H., Hara, M., Shimada, T., Saikawa, T., Yoshimatsu, H., 2007. Hyperthermia treatment prevents angiotensin II-mediated atrial fibrosis and fibrillation via induction of heat-shock protein 72. J. Mol. Cell. Cardiol. 43, 616–626. Wang, X., Venable, J., LaPointe, P., Hutt, D.M., Koulov, A.V., Coppinger, J., Gurkan, C., Kellner, W., Matteson, J., Plutner, H., Riordan, J.R., Kelly, J.W., Yates, J.R., Balch, W.E., 2006. Hsp90 cochaperone Aha1 downregulation rescues misfolding of CFTR in cystic fibrosis. Cell 127, 803–815. Wang, B., Komers, R., Carew, R., Winbanks, C.E., Xu, B., Herman-Edelstein, M., Koh, P., Thomas, M., Jandeleit-Dahm, K., Gregorevic, P., Cooper, M.E., Kantharidis, P., 2012. Suppression of microRNA-29 expression by TGF-beta1 promotes collagen expression and renal fibrosis. J. Am. Soc. Nephrol. 23, 252–265. Williams, J.D., Craig, K.J., Topley, N., Von Ruhland, C., Fallon, M., Newman, G.R., Mackenzie, R.K., Williams, G.T., 2002. Morphologic changes in the peritoneal membrane of patients with renal disease. J. Am. Soc. Nephrol. 13, 470–479. Wright, G.W., Simon, R.M., 2003. A random variance model for detection of differential gene expression in small microarray experiments. Bioinformatics 19, 2448–2455. Xiao, C., Wang, R.H., Lahusen, T.J., Park, O., Bertola, A., Maruyama, T., Reynolds, D., Chen, Q., Xu, X., Young, H.A., Chen, W.J., Gao, B., Deng, C.X., 2012. Progression of chronic liver inflammation and fibrosis driven by activation of c-JUN signaling in Sirt6 mutant mice. J. Biol. Chem. 287, 41903–41913. Yamada, M., Kubo, H., Ota, C., Takahashi, T., Tando, Y., Suzuki, T., Fujino, N., Makiguchi, T., Takagi, K., Suzuki, T., Ichinose, M., 2013. The increase of microRNA-21 during lung fibrosis and its contribution to epithelial–mesenchymal transition in pulmonary epithelial cells. Respir. Res. 14, 95. Yanez-Mo, M., Lara-Pezzi, E., Selgas, R., Ramirez-Huesca, M., Dominguez-Jimenez, C., Jimenez-Heffernan, J.A., Aguilera, A., Sanchez-Tomero, J.A., Bajo, M.A., Alvarez, V.,


10


Castro, M.A., Del, P.G., Cirujeda, A., Gamallo, C., Sanchez-Madrid, F., Lopez-Cabrera, M., 2003. Peritoneal dialysis and epithelial-to-mesenchymal transition of mesothelial cells. N. Engl. J. Med. 348, 403–413. Yang, F., Zhang, L., Huo, X.S., Yuan, J.H., Xu, D., Yuan, S.X., Zhu, N., Zhou, W.P., Yang, G.S., Wang, Y.Z., Shang, J.L., Gao, C.F., Zhang, F.R., Wang, F., Sun, S.H., 2011. Long noncoding RNA high expression in hepatocellular carcinoma facilitates tumor growth through enhancer of zeste homolog 2 in humans. Hepatology 54, 1679–1689. Yang, Y., Li, H., Hou, S., Hu, B., Liu, J., Wang, J., 2013. The noncoding RNA expression profile and the effect of lncRNA AK126698 on cisplatin resistance in non-small-cell lung cancer cell. PLoS One 8, e65309.

Yoshino, J., Monkawa, T., Tsuji, M., Hayashi, M., Saruta, T., 2003. Leukemia inhibitory factor is involved in tubular regeneration after experimental acute renal failure. J. Am. Soc. Nephrol. 14, 3090–3101. Yu, J.W., Duan, W.J., Huang, X.R., Meng, X.M., Yu, X.Q., Lan, H.Y., 2014. MicroRNA-29b inhibits peritoneal fibrosis in a mouse model of peritoneal dialysis. Lab. Invest. 94, 978–990. Zhou, Q., Yang, M., Lan, H., Yu, X., 2013. miR-30a negatively regulates TGF-beta1-induced epithelial–mesenchymal transition and peritoneal fibrosis by targeting Snai1. Am. J. Pathol. 183, 808–819.


Angiotensin II receptors and peritoneal dialysis-induced peritoneal fibrosis.

Periostin-Binding DNA Aptamer Treatment Ameliorates Peritoneal Dialysis-Induced Peritoneal Fibrosis.

Peritoneal morphology in children treated by continuous ambulatory peritoneal dialysis.

New developments in peritoneal fibroblast biology: implications for inflammation and fibrosis in peritoneal dialysis.

Effect of benfotiamine in podocyte damage induced by peritoneal dialysis fluid.

Phospholipids profiling and outcome of peritoneal dialysis patients.

Qualitative Profiling of Polyglucose Degradation Products in Peritoneal Dialysis Fluids.

A microrna screen to identify regulators of peritoneal fibrosis in a rat model of peritoneal dialysis.

[Bacteriological characteristics of the peritoneal dialysis fluid infections].

MicroRNA-29b inhibits peritoneal fibrosis in a mouse model of peritoneal dialysis.

Output of peritoneal cells during peritoneal dialysis.

Net ultrafiltration in peritoneal dialysis: role of direct fluid absorption into peritoneal tissue.

Peritoneal transfer of thiamphenicol during peritoneal dialysis.

Peritoneal dialysis.

Spironolactone to prevent peritoneal fibrosis in peritoneal dialysis patients: a randomized controlled trial.

The peritoneal dialysis system.

Peritoneal dialysis and peritonitis.

Peritoneal Dialysis in Asia.

Peritonitis in peritoneal dialysis.

Peritoneal dialysis in Peru.

Peritoneal dialysis in China.

Net fluid absorption under membrane transport models of peritoneal dialysis.

Complications of peritoneal dialysis.

Assessment of fluid and nutritional status using multifrequency bioelectrical impedance analysis in peritoneal dialysis patients.