Profiling of miRNAs and target genes related to cystogenesis in ADPKD mouse models

Autosomal polycystic kidney disease (ADPKD) is a common inherited renal disease characterized by the development of numerous fluid-filled cysts in both kidneys. We investigated miRNA-mediated regulatory systems and networks that play an important role during cystogenesis through integrative analysis of miRNA- and RNA-seq using two ADPKD mouse models (conditional Pkd1- or Pkd2-deficient mice), at three different time points (P1, P3, and P7). At each time point, we identified 13 differentially expressed miRNAs (DEmiRs) and their potential targets in agreement with cyst progression in both mouse models. These targets were involved in well-known signaling pathways linked to cystogenesis. More specifically, we found that the actin cytoskeleton pathway was highly enriched and connected with other well-known pathways of ADPKD. We verified that miR-182-5p regulates actin cytoskeleton rearrangement and promotes ADPKD cystogenesis by repressing its target genes—Wasf2, Dock1, and Itga4—in vitro and in vivo. Our data suggest that actin cytoskeleton may play an important role in renal cystogenesis, and miR-182-5p is a novel regulator of actin cytoskeleton and cyst progression. Furthermore, this study provides a systemic network of both key miRNAs and their targets associated with cyst growth in ADPKD.

phenotypic changes similar to those in Pkd1 f/f :HoxB7-cre mice ( Supplementary Fig. S2), except for survival rate, which indicated longer survival in Pkd2 f/f :HoxB7-cre mice than in Pkd1 f/f :HoxB7-cre mice (Fig. 1E). These data suggested that conditional inactivation of Pkd1-or Pkd2 may share common cyst formation processes in early stages, but different cyst expansion processes in later stages.
Integrated analysis of miRNA-seq and mRNA-seq. Given the phenotypic changes, both miRNA-seq and RNA-seq analyses were carried out using kidney tissues from the two mouse models at the indicated time points (P1, P3, and P7, n = 3). An integrative analysis of both types of NGS data was performed as shown in Supplementary Fig. S3. To profile transcripts and their expressions, we measured read counts instead of RPKM 27 or FPKM 28 values, because this was a more accurate measure for comparing differences between the mouse models at indicated time points 29 . miRNA expression was also measured based on known sequences. The read count values for each mouse model were normalized and used for further analysis. To enhance the candidate pools of mRNAs and miRNAs, mRNAs and miRNAs showing negative binomial distributions were identified independently and correlation analysis between mRNAs and miRNAs were performed. First, the most popular two methods (edgeR 30 and Deseq. 2 31 ) were performed in both wild-type (WT; Pkd1 f/f and Pkd2 f/f ) and knockout (KO; Pkd1 f/f :HoxB7-cre and Pkd2 f/f :HoxB7-cre) models to identify significantly differentially expressed mRNAs and miRNAs. For RNA-seq analysis, 3,040 transcripts in Pkd1 f/f :HoxB7-cre mice and 2,470 transcripts in Pkd2 f/f : HoxB7-cre mice were differentially expressed at the indicated time points (FDR < 0.05). In addition, we also identified 1,297 commonly regulated transcripts in both mouse models. For miRNA analysis, 243 miRNAs were differentially expressed, while 130 miRNAs were common in both mouse models (Supplementary Tables S1, 2, 3 and 4). Among the common miRNAs, 13 miRNAs, including 11 upregulated and two downregulated miRNAs ( Fig. 2A), were selected based on expression pattern comparison at each time point. Second, correlation analysis was carried out using 13 DEmiRs and differentially expressed genes (DEGs) in both mouse models. Then, pathway analysis was performed and well-known pathways involved in cystogenesis including ERK/MAPK, AKT/ mTOR, JAK/STAT, Wnt/β-catenin, PCP pathway, cell cycle, and TGF-β signaling pathways were identified along with novel pathways, such as the actin cytoskeleton pathway. miRNAs in the actin cytoskeleton pathway were highly enriched and delicately connected with other well-known signaling pathways (Supplementary Tables S5  and S6). Expression of candidate DEmiR targets associated with the actin cytoskeleton were gradually changed along with cyst progression in Pkd1 f/f :HoxB7-cre mice (Fig. 2B). Remarkably, miR-182-5p had multiple candidate targets involved in actin cytoskeleton signaling (Fig. 2C). The data indicated that 13 miRNAs were commonly associated with cystogenesis-related pathways, although they may target different mRNAs in two mouse models of ADPKD. Validation of key miRNAs and candidate target genes. To validate the correlation analysis, expression levels of 13 key miRNAs were confirmed in the kidney tissues of Pkd1 f/f :HoxB7-cre mice using qRT-PCR. Most miRNAs, including miR-182-5p, showed a similar expression pattern between qRT-PCR and sequencing results (Fig. 3A). Interestingly, we observed that all miR-182 family members, miR-182-3p, miR-182-5p, miR-183-5p, and miR-96-5p, were significantly upregulated in P7 Pkd1 f/f :HoxB7-cre mice kidney tissues. Upon considering both small RNA-seq raw data and qRT-PCR, we concluded that miR-182-5p was the most suitable subject for future research. To identify alterations of spatiotemporal miR-182-5p expression in the mouse postnatal kidney, section miRNA in situ hybridization (ISH) was performed. Expression of miR-182-5p was detected in the Pkd1 f/f medulla region and glomerulus. Interestingly, miR-182-5p-expressing cells were observed in the cuboidal epithelial lining cells of the cyst in Pkd1 f/f :HoxB7-Cre mice at P7 (Fig. 3B). miR-182-5p was localized in the overall kidney region at P1 and P3, including the medulla and cortex regions ( Supplementary Fig. S4). The glomerulus showed especially strong miR-182-5p expression in P1 Pkd1 f/f mice (Supplementary Fig. S4A-C). In the Pkd1 f/f :HoxB7-Cre mice, as expected, miR-182-5p was strongly expressed in the cyst-lining epithelial cells ( Supplementary Fig. S4D-F and J-L). In Pkd1 f/f at P3, interestingly, miR-182-5p expression was observed in the renal tubule domain, but the glomerulus and Bowman's capsule cells did not exhibit miR-182-5p-expressing cells ( Supplementary Fig. S4G-I). Candidate target mRNAs of miR-182-5p were downregulated in the kidney tissues of Pkd1 f/f :HoxB7-cre mice over time ( Fig. 3C and Supplementary Fig. S5A). Furthermore, the expression of genes involved in ERK/MAPK and AKT/mTOR signaling pathways was increased; these pathways regulate cell proliferation in ADPKD (Supplementary Fig. S6A and C). Indeed, when we examined cell proliferation at the three time-points by ki67 staining, proliferating cells were markedly observed in cyst-lining cells of both KO mouse models, but not in WT kidney tissues ( Supplementary Fig. S6B,D and E). In particular, in order to investigate the link between miR-182-5p and targets involved in actin cytoskeleton signaling as a novel cystogenesis pathway, a 3′UTR luciferase assay for miR-182-5p was performed. We found that miR-182-5p directly bound to the 3′UTR of these candidate target mRNAs related to actin cytoskeleton, negatively regulating their translation ( Fig. 3D and Supplementary Fig. S5B). Therefore, our results were consistent with those of previous studies and helped us identify miRNAs that bridge ADPKD actin cytoskeleton defects and cystogenesis.
Cyst formation upon actin cytoskeleton defects. The overexpression of PC1 induces cytoskeletal rearrangement, cell migration, and tubule formation 32 . To evaluate the relevance of this protein to the actin cytoskeleton in ADPKD cystogenesis, 3-dimensional (3D) culture and F-actin staining were performed after knockdown of either Pkd1 or Pkd2 in mouse inner medullary collecting duct (mIMCD) cells. Prior to staining, mRNA and protein levels of Pkd1, Pkd2, and miR-182-5p target genes were determined in recovered mIMCD spheroids ( Supplementary Fig. S7). Knockdown of either Pkd1 or Pkd2 in mIMCD cells resulted in remarkable defects of F-actin structure and accelerated cyst growth in 3D culture (Fig. 4A). In this study, we utilized two collecting duct-specific polycystin-1 knockout mouse models, Pkd1 f/f :HoxB7-cre and Pkd1 f/f :Aqp2-cre; two different promoters were used to drive Cre recombinase expression in the transgenic mice. Although Pkd1 f/f :HoxB7-cre mice and Pkd1 f/f :Aqp2-cre mice have similar PKD phenotypes, Pkd1 f/f :Aqp2-cre mice tend to survive longer than Pkd1 f/f :HoxB7-cre mice 33 . Upon staining these cells with both rhodamine-labeled phalloidin and marker genes related to the actin cytoskeleton in wild-type (Pkd1 f/f ) and Pkd1-deficient (Pkd1 f/f :Aqp2-cre and Pkd1 f/ f :HoxB7-cre) mouse kidneys, strong F-actin signals were observed in the apical membrane with weak signals outlining the tubules in WT kidneys. However, cyst-lining epithelial cells in Pkd1-deficient kidneys showed reduced and disorganized actin cytoskeletons with significantly decreased levels of marker genes such as Wasf2, Dock1, and Itga4 ( Fig. 4B and Supplementary Fig. S8). To investigate whether actin filament polymerization was involved in cyst development, actin structures were investigated after a wound healing assay. Five hours after scratching, cells deficient in Wasf2, Dock1, or Itga4 did not produce organized actin cytoskeletons like those observed in the control siRNA-transfected cells (Fig. 4C). In addition, knocking down Wasf2, Dock1, or Itga4 promoted cyst growth with actin cytoskeleton defects in 3D culture conditions (Fig. 4D). Moreover, western blotting analysis revealed that miR-182-5p negatively regulated the target genes Wasf2, Dock1, and Itga4 (Fig. 5A). Intriguingly, we found that overexpressing miR-182-5p inhibits actin cytoskeleton organization, particularly lamellipodium formation, by repressing marker genes related with the actin cytoskeleton, ultimately affecting cyst development in vitro ( Fig. 5B and C). Finally, overexpressing miR-182-5p increased cyst growth, while inhibiting miR-182-5p inhibited cyst formation (Fig. 5C). These data suggest that miR-182-5p may play a critical role as a novel inducer of cystogenesis by regulating the actin cytoskeleton.

Inhibiting miR-182-5p attenuates cyst growth in a Pkd1 conditional knockout mouse model.
To evaluate the therapeutic effect of miR-182-5p inhibition, a lentiviral particle-mediated miRNA inhibition system was introduced. The Pkd1 f/f :Aqp2-cre mouse model was selected for the injection study because Pkd1 f/ f :HoxB7-cre mice live for only 14-23 days. A control lentivirus or inhibitory miR-182-5p lentiviral particles (final conc.: 4 × 10 7 IU) were injected intraperitoneally five times on alternate days. Mice were then sacrificed the day after the final injection (Fig. 6A). Kidney weight-to-body weight ratios, BUN levels, and miR-182-5p expression in kidney tissues were reduced upon treatment with the inhibitory miR-182-5p lentivirus compared with those in con-lentivirus-treated Pkd1 f/f :Aqp2-cre mice ( Fig. 6B-E). Furthermore, we observed that the expression patterns of the target genes of miR-182-5p, Wasf2, Dock1, and Itga4, were restored in mice treated with inhibitory miR-182-5p lentivirus (Fig. 6F). Collectively, these results indicated that inhibiting miR-182-5p was effective for inhibiting cyst growth and restoring renal function.

Discussion
In the current study, we described parallel integrated analysis of miRNA-seq and RNA-seq data and validated the expression, direct interaction, and cystogenesis-related functions of key miRNAs and their targets. In particular, Relative 3′UTR luciferase activity of Wasf2, Dock1, and Itga4 genes upon transfecting mIMCD cells with Negative Control mimic (NC mimic) or miR-182-5p mimic. Mutating the seed sequence of miR-182-5p induced rescued luciferase activity of the psiCHECK-2 vector. WT (wild type), MT (mutant type). Data are presented as mean ± SD of three independent experiment in triplicate. ***P < 0.001; **P < 0.01.
we found commonly dysregulated pathways by comparing Pkd1-and Pkd2-deficient mouse models. These models had overlapping, differentially expressed miRNAs and signaling pathways related with cystogenesis, suggesting that these miRNAs play crucial roles in cyst formation and expansion in ADPKD patients with either PKD1 or PKD2 mutations. Although both mouse models have common key miRNAs, these miRNAs are considered to target different mRNAs in Pkd1 and Pkd2-deficient mice ( Supplementary Fig. S9). Indeed, our data showed that miR-182-5p was associated with cystogenesis by regulating the actin cytoskeleton signaling pathway in both mouse models. However, miR-182-5p directly targets and regulates Wasf2, Dock1, and Itga4 in Pkd1-deficient mice, but not in Pkd2-deficient mice ( Supplementary Fig. S10). This difference may be related to their phenotypic differences: even though these two models showed similar cyst progression and timing of renal failure, Pkd2-deficient mice remained alive longer than Pkd1-deficient mice, for maximum of about 20 days. This suggests that these mouse models share common molecular alterations in early stage cystogenesis but they have specific molecular changes in late stages, at least after P7. Therefore, our data showed that defects in Pkd1 or Pkd2 share some, but not all, molecular changes and signaling pathways. Moreover, the parallel analysis between time-dependent phenotypic characterization and miRNA-mRNA profiling allowed us to select and validate more precise targets involved in ADPKD cyst progression.
Recently, there have been several miRNA profiling studies in ADPKD rodent models, using microarrays.  miRNAs, including miR-182 and miR-96, by qPCR and suggested that several miRNAs may be associated with ADPKD-related signaling pathways. Dweep et al. identified eight significantly upregulated miRNAs in diseased kidneys of PKD/Mhm rats by a parallel microarray-based approach 35 . A major advantage of unbiased next-generation sequencing-based approaches such as small RNA-seq and RNA-seq is that they provide genome-wide and quantitative information, whereas microarray-based approaches do not. Our results are partially in agreement with two previous studies but have different candidate miRNAs and targeting signaling pathways. Although it is well-known that increased proliferation by pathways including ERK/MAPK is required for cyst formation, the precise mechanism of cyst formation and enlargement in ADPKD has not yet been described. This indicates that systemic analysis is necessary for uncovering novel relevant pathways. From integrated miRNA-seq and RNA-seq analysis, we observed that Pkd1 inactivation changes the expression of several key miRNAs involved in previously reported signaling pathways such as ERK/MAPK and Akt/mTOR signaling, as well a novel pathway, actin cytoskeleton signaling. In addition, our data showed that upregulation of miR-21a-5p in cyst-lining epithelial cells stimulates cyst growth in Pkd1-deficient mice, consistent with a previous study 23 .
Intriguingly, we provided evidence of a direct link between the actin cytoskeleton and cystogenesis in ADPKD. In a previous study, polycystins were found to be responsible for cell migration, polarity, and tissue morphogenesis 36 . PC1 regulates the actin cytoskeleton and cell migration mediated by the phosphatidylinositol 3-kinase (PI3-K) pathway in MDCK cells 32 . The PI3-K pathway was significantly enriched in diseased kidneys of Pkd1-deficient mice, suggesting that the downregulation of PI3K genes such as Pik3r1 and Pik3r3 may be associated with actin cytoskeleton defects and reduced migration capacity in cystic epithelial cells. In addition, PC1 is involved in the actin cytoskeleton by interacting with the Pacsin/WASP-2/Arp2/3 complex, which contributes to microfilament nucleation at the leading edge 18 . However, these two studies did not show direct evidence linking the actin cytoskeleton with cyst development. From our results, we identified that key proteins involved in the actin cytoskeleton were regulated by several miRNAs, including miR-182-5p, that control cyst progression. Wasf2, also known as Wave2, is a well-known key component of the actin regulatory machinery and mediates (C) miR-182-5p precursors or inhibitors (20 nM) were transfected into mIMCD cells, and then cells were observed in 3D culture conditions. The cysts in miR-182-5p precursor-treated cells were bigger than cysts in NC treated cells. miR-182-5p inhibitor-transfected mIMCD cells showed reduced of cyst lumen size. Immunostaining with Wasf2, Dock1, and Itga4 (green); F-actin (red); DAPI (blue). ***P < 0.001. actin cytoskeletal organization by cooperating with the Arp2/3 complex to promote the formation of actin filaments 37 . We identified, Wasf2 as a novel direct target of miR-182-5p. Inhibiting Wasf2 expression accelerated cyst growth, dysregulated actin structure, and decreased cell migration. We also found that Dock1 is a novel direct target of miR-182-5p. Dock1 is a guanine nucleotide exchange factor that activates the Rac1/WAVE-complex/ Arp2/3-complex, resulting in the formation of lamellipodial protrusions 38 . We observed that Dock1 repression also stimulated cyst development via alterations of the actin structure in mIMCD 3D cultures. Taken together, the results suggest that defects in the actin cytoskeleton and progression of cyst formation are tightly linked. Furthermore, actin dynamics seem to be linked to establishment of cell polarity and formation of apical junctions 39 . Defects in cell/matrix interaction is a suggested causative mechanism of cystogenesis in ADPKD, but these remain poorly understood. Integrins are the main cell surface receptors for proteins within the extracellular matrix (ECM) 40 . They contribute to cell movement and mediate adhesion junctions, which are essential for tissue integrity 41 . Integrin alpha 4 (Itga4), which we identified as a novel target of miR-182-5p, may play a crucial role in migratory events and we showed that it can affect cyst growth rate in vitro.
Considering our observations in this study, miR-182-5p should be a potential therapeutic target. We also observed that miR-182-5p inhibition by systemic injection of lentiviral particle-mediated inhibitory miRNA attenuated cyst progression in vivo. We confirmed that the expression of Wasf2, Dock1, and Itga4 were restored in the miR-182-5p lentivirus-injected Pkd1 f/f :Aqp2-cre mice. However, miR-182-5p inhibition could not slow cyst growth in Pkd2 f/f :Aqp2-cre mice because Wasf2, Dock1, and Itga4 play significant roles in Pkd1 but not in Pkd2 conditional KO mice. The miRNA and its target genes, other than miR-182-5p, may play more important regulatory functions in Pkd2 conditional KO mice. Overall, we present strong evidence that miR-182-5p upregulation is Pkd1-dependently associated with defects in the actin cytoskeletal pathway, which can cause symptoms of PKD. In conclusion, we applied a NGS-based methodology to identify several key miRNAs that are linked to Pkd1 or Pkd2 expression using two ADPKD mouse models, namely, conditional Pkd1 or Pkd2-deficient mice. The miRNA-mRNA integrated data for cyst progression at three time-points revealed genome-wide correlations between molecular changes and disease phenotype. Additionally, we provided evidence of a direct link between the actin cytoskeleton and cystogenesis in ADPKD. Therefore, we suggest that these key miRNAs and their possible potential targets play a central role in cystogenesis in ADPKD.  42 were obtained from Dr. Stefan Somlo at Yale University. HoxB7-Cre transgenic mice, which specifically drives cre expression in the kidney collecting duct, was obtained from Minho Shong at Chungnam National University. To characterize the phenotype resulting from Pkd1 or Pkd2 inactivation in the collecting duct, Pkd1 f/+ :HoxB7-Cre and Pkd1 f/f or Pkd2 f/+ :HoxB7-Cre and Pkd2 f/f mice were mated to generate a conditional knockout mouse model. Age-matched littermates with Pkd1 f/f or Pkd2 f/f genotypes were used as controls. Genotyping was performed using previously described protocols 26 . Tail DNA was PCR-amplified for Pkd1 f/f alleles using the forward primer 5′-CCGCTGTGTCTCAGTGCCTG-3′ and reverse primer 5′-CAAGAGGGCTTTTCTTGCTG-3′; for Pkd2 f/f alleles using the forward primer 5′-GGGTGCTGAAGAGATGGTTC-3′ and reverse primer 5′-TCCACAAAAGCTGCAATGAA-3′. HoxB7-cre transgene expression was confirmed using the forward primer 5′-GCGGTCTGGCAGTAAAAACTATC-3′ and reverse primer 5′-GTGAAACAGCATTGCTGTCACTT-3′. All experimental procedures were performed in accordance with relevant guidelines and regulations on laboratory animal care. All experimental protocols were reviewed and approved by the IACUC at Sookmyung Women's University.
Three-dimensional (3D) cell culture was performed according to previously reported protocol 43 . Briefly, suspended mIMCD cells (1 × 10 4 cells/well) were mixed with Matrigel at a 1:1 ratio and plated in 8-well chamber slides. Cells in Matrigel were grown in a humidified 5% CO 2 atmosphere at 37 °C for 5 days. After 5 days in 3D culture, phase contrast images of individual spheroids were obtained by a light microscope (IX70, Olympus) and cyst area was measured using ISP capture software (Olympus).
For phalloidin staining of the actin cytoskeleton in kidney tissues, tissues were fixed in 4% paraformaldehyde in PBS and embedded in OCT. Cryosections were blocked with 1% BSA in PBS for 1 h at room temperature and incubated with rhodamine-labeled phalloidin (Invitrogen, Catalog #R415) diluted 1:500 overnight at 4 °C. Slides were mounted using mounting medium with DAPI (Vector laboratory).
For immunocytochemistry, cultured cells were fixed in 4% paraformaldehyde for 10 min at room temperature, and then washed three times with PBS. Cells were permeabilized with 0.2% Triton X-100 containing 5% v/v bovine serum albumin, and incubated with the appropriate primary antibody at 4 °C overnight. Images were obtained with a confocal laser scanning microscope (LSM-700, Carl Zeiss).
For western blot analysis, equal amounts of protein were analyzed in duplicate by SDS-PAGE and electro-transferred to a polyvinylidene fluoride (PVDF) membrane (ATTO, Japan). Primary antibodies were diluted at 1:200~1:1000 in 1% skim milk in PBST at 4 °C overnight and secondary antibodies were incubated for 1 h at room temperature. Anti-β-actin (Bethyl Laboratories, Montgomery, TX, USA) was used as a loading control. Immunoreactive proteins were detected by horseradish peroxidase-conjugated secondary antibodies and enhanced using chemiluminescence (ECL) reagents (GE Healthcare Life Sciences, Piscataway, NJ).
Cystic index measurement and renal function test. The tubular cyst formation areas were quantified in whole kidney sections of Pkd1 f/f :HoxB7-Cre and Pkd2 f/f :HoxB7-Cre mice over time. Two sections from the mid-sagittal region of each kidney were analyzed on P1, P3, and P7 for each mouse model (n = 3). Whole kidney images stained with tubule markers were obtained using the scan slide module in a confocal laser scanning microscope (LSM-700, Carl Zeiss). Total kidney area and cystic area with DBA signal were measured using the tools provided by LSM-700. The total kidney area was divided into four parts according to their sizes: non-cystic area, dilated tubules (<1 × 10 5 μm 2 ), small cysts (1 × 10 5 -10 6 μm 2 ), and large cysts (>1 × 10 6 μm 2 ). Cyst index = (total cystic area/total kidney area) × 100 (%).
Small RNA and mRNA sequencing, alignment, and normalization. Total RNA was isolated using TRIzol according to the manufacturer's protocol (Invitrogen Life Technologies). To estimate expression levels and to determine alternatively spliced transcripts, the RNA-Seq reads were mapped to the mouse genome (Mus musculus, version mm10) downloaded from the UCSC website (http://genome.uscs.edu) using TopHat v2.1.0 44 , which is capable of reporting split-read alignments across splice junctions. Read counts of representative genes and their isoforms were measured by HTSeq 45 with default options using RefSeq annotation as a reference. Read counts under ten counts in at least three samples were filtered out before normalization, and normalization was performed using the TMM method 46 . Statistical significance of the transcript expression data was determined with criteria (two-fold change and false discovery rate (FDR < 0.05)) from the edgeR 30 Deseq2 31 implemented in the Bioconductor package. Differentially expressed genes (DEGs) were identified by pairwise comparisons between WT and KO groups at each time point. The DEG sets were collected for all three time points by combining the lists obtained by edgeR and Deseq2. Hierarchical clustering for DEGs was performed using complete linkage and Euclidean distance as a measure of similarity. All statistical analysis and visualization of differentially expressed genes was conducted using R 3.2.3 (www.r-project.org).
For miRNA-Seq analysis, raw reads were trimmed by removing the 3′ adaptor sequence and retaining more than 17 bp of the read sequence. For read mapping of miRNA, trimmed reads were aligned against the mouse genome or aligned to the known mature & hairpin miRNAs (miRBase version 21 47 , June, 2014) by the miRdeep2 program 48 . Raw read counts for miRNA were filtered and normalized by the same process of RNA-Seq data analysis. Differentially expressed miRNAs (DEmiR) were classified as genes having a 2 fold-change and P-value < 0.05. The DEmiR sets were determined for all three time points by combining the lists obtained by edgeR and Deseq2.
Integration of miRNA-mRNA data. We performed mRNA-miRNA data integration for matched samples in each data set. First, the DEG set and miRNA set were used to calculate Spearman's rank correlation between miRNA and mRNA expression in all data sets. This approach assumed that the expression of a given miRNA is negatively correlated with the mRNA expression of its targets. For miRNAs with the potential to regulate mRNA, a correlation test was performed in R to determine if the miRNA was statistically significant with a P-value < 0.05 and a negative correlation coefficient. Then we retained putative miRNA-target mRNA pairs with a relationship using miRwalk v2.0 49 (http://zmf.umm.uni-heidelberg.de/apps/zmf/mirwalk2/), a database that supplies the largest available collection of predicted and experimentally verified miRNA-target interactions.
Pathway and networking analysis. Putative target gene sets of differentially expressed miRNAs were evaluated by QIAGEN's Ingenuity Pathway Analysis (IPA, QIAGEN Redwood City, www.qiagen.com/ingenuity) for biological functions and canonical pathways or networks. Then, we identified canonical pathways associated with ADPKD phenotype by identifying genes correlated with miRNAs. The network analysis tool Cytoscape v3.3 (http://cytoscape.org/) was used to find the putative target genes for the predicted miRNAs in canonical pathways. Specially, gene-gene interactions were obtained from Pathway Studio 50 to identify highly connected modules within a canonical pathway.
To detect miRNA expression, cDNA was prepared from total RNA (500 ng) using a miScript II RT kit (Qiagen) according to the manufacturer's protocol. Real-time PCR quantification of mature miRNA was performed using target-specific miScript Primer Assays (Qiagen) and the miScript SYBR Green PCR Kit (Qiagen) according to the manufacturer's protocol.
3′UTR reporter analysis. The 3′UTR of mmu-miR-182-5p candidate target mRNAs were amplified by PCR using PrimeSTAR GXL DNA polymerase (Takara Bio Inc., Japan) and cloned into the XhoI and NotI site of a psiCHECK-2 vector (Promega) using an In-Fusion HD Cloning kit (Takara Bio Inc., Japan). The seed sequences of miR-182-5p were mutated using a PCR-based approach. For transfection, mIMCD cells (2 × 10 5 ) were plated into 6-well plates and 1.5 μg of luciferase constructs and 15 nM of negative control mimics or mmu-miR-182-5p mimics (Ambion) were co-transfected into the cells one day after seeding using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instruction. Cells were collected after 36 h for experiments and luciferase activity was measured by a Dual Luciferase Assay system (Promega). All transfection experiments were performed in triplicate and at least three times.
Section microRNA in situ hybridization. Kidneys were dissected at postnatal day 7 and subsequently fixed in 4% paraformaldehyde (PFA) overnight. After fixation, tissues were embedded in paraffin and sectioned at a thickness of 7 μm. Sections were incubated at 60 °C for deparaffinization using Histoclear II (National Diagnostics), re-hydrated through a graded series of alcohol washes and 1X DEPC-PBS. The specimens were treated with 10 μg/ml proteinase K for 5 min at 37 °C. Sections were incubated in pre-hybridization buffer (50% formamide, 5× SSC, 0.1% Tween-20, 50 ug/ml heparin, 500 ug/ml yeast tRNA) at 53 °C for 1 hr. Hybridization was conducted at 53 °C for overnight using miR-182 locked nucleic acid (LNA) modified probe (Exiqon). Sites of