Introduction

Interstitial cystitis (IC) is a non-specific chronic inflammatory bladder disease characterized by urinary frequency, urinary urgency, and pelvic pain induced by bladder filling [1, 2]. Patients with IC suffer significant declines in quality of life (QOL). The regional morbidity rate of IC varies, with 3–4 per 10 million in Japan, 60–70 per 10 million in the United States, and 18 per 10 million in Europe [3]. Based on cystoscopic findings, IC can be classified into 2 groups: Hunner type with ulcerative lesions (Hunner’s lesion), and non-Hunner type with bladder petechial oozing (glomerulation) after hydrodistension. Diagnosis and classification of IC is determined by cystoscopic findings and exclusion of other diseases based on subjective symptoms. Although several studies searching for diagnostic markers IC have been conducted, no definitive markers are known [4]. Moreover, no symptomatic treatments for IC such as bladder hydrodistension, drug therapy and food therapy, or curative therapy have been established [5]. Given the lack of consensus about IC etiology, a genomic approach to elucidate molecular mechanisms underlying IC pathogenesis is needed.

microRNA (miRNA) are small (19 to 22 nucleotide) non-coding RNAs that can finely control expression of protein-coding or non-coding RNAs [6]. A single miRNA can control sequence-dependent expression of an extremely large number of RNA transcripts [7]. Therefore, aberrantly expressed miRNAs can lead to dysregulation of intracellular RNA networks. Indeed, a large number of studies have shown aberrant miRNA expression in several diseases and cancers, thus highlighting the important role for miRNAs in human disease pathogenesis [8,9,10,11,12,13].

We have successfully identified several cancer networks that are regulated by antitumor miRNAs in various cancer types using genome-wide gene expression analyses and in silico database searches [14,15,16,17,18,19]. Here we adapted our miRNA analysis strategy to characterize the molecular pathogenesis of IC. First, we analyzed the miRNA expression signature of IC based on RNA sequencing of samples from clinical specimens of Hunner-type IC and non-Hunner type IC. We then examined dysregulated miRNA expression in IC tissues to identify novel molecular networks involved in IC pathogenesis.

Analyses of miRNA signature revealed that expression of 366 miRNAs (203 and 163 downregulated and upregulated, respectively) was dysregulated in IC tissues. Moreover, we identified several genes that are targeted by the dysregulated miRNAs in IC tissues. Novel approaches that identify aberrantly expressed miRNA regulatory networks in IC could provide new prognostic markers and therapeutic targets for this disease.

Materials and methods

Patients and IC clinical specimens

IC clinical specimens were obtained from patients who were diagnosed at Dokkyo Medical University Hospital and Chiba University Hospital. Written informed consent was obtained from each patient in advance of sample collection. All patients underwent bladder hydrodistension, and IC tissues were collected by bladder biopsy. Hunner type IC patients were diagnosed by the presence of lesions, whereas non-Hunner type IC was classified based on the presence of glomerulation during bladder hydrodistension. Patient characteristics and cystoscopic findings are summarized in Supplementary Table 1 and Fig. 1, respectively.

Fig. 1
figure 1

Cystoscopic findings of IC patients. Cystoscopic findings of IC patients used in deep sequencing analysis. No.1, 2, 6, 7 and 8 are Hunner type IC patients’ findings, indicating Hunner’s lesions. No. 3, 4 and 5 are non-Hunner type IC patients’ findings, showing bladder petechial oozing (glomerulation) after hydrodistension

Full size image

RNA extraction and cell culture

We extracted total RNA using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol, and RNA quality was assessed as previously described [15, 17, 20, 21]. The human bladder cancer (BC) cell line T24 was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA) and maintained as previously described [15, 17, 20, 21].

Construction of miRNA expression signature for IC

To create a miRNA expression signature for IC, we performed small RNA sequencing using a HiSeq 2000 (Illumina, San Diego, CA, USA) for 8 IC samples (Supplementary Table 1). Samples from 5 normal and 5 cancer tissues of bladder that we described in a previous study were used as comparative specimens [17]. We carried out small RNA sequencing and data mining procedures as described previously [16, 17, 19, 22]. A false discovery rate (FDR) of less than 0.05 was considered to be significant.

Transfection with mature miRNA

We used Ambion Pre-miR miRNA precursor for hsa-miR-320b (assay ID: PM13132; Applied Biosystems, Foster City, CA, USA) and hsa-miR-320c (assay ID: PM13133; Applied Biosystems)as mature miRNA species. miRNA transfection of T24 cells and analysis of transfection efficiencies were conducted as described previously [15, 17, 20, 21].

Identification of putative genes regulated by the miR-320 family in T24 cells

To identify genes regulated by the miR-320-family, we combined in silico and genome-wide gene expression analyses as previously described [15, 17, 20, 21]. We used the TargetScanHuman 7.1 (June, 2016 release, http://www.targetscan.org/vert_71) database and an oligo microarray (Agilent Technologies; Human Ge 60 K) for gene expression analyses. Microarray data were deposited into the GEO database (https://www.ncbi.nlm.nih.gov/geo/).

Immunohistochemistry

We incubated tissue specimens overnight at 4 °C with anti-E2F-1 antibodies (1:50 dilution; Cat. no. sc-251; Santa Cruz Biotechnology, Santa Cruz, CA, USA), anti-E2F-2 antibodies (1:50; Cat. no. sc-9967; Santa Cruz Biotechnology) and anti-TUB antibodies (1:50; product no. HPA049019; Sigma-Aldrich, St. Louis, MO, USA). The procedures were conducted as described previously [15, 17, 20, 21].

Results

Small RNA sequences of specimens from normal bladder, bladder cancer and IC specimens

We performed deep sequencing of 18 small RNA libraries from specimens from 5 normal bladders, 5 bladder cancer cases, and 8 IC cases. Patient backgrounds for the specimens from normal and bladder cancer cases are described in our previous report [17], and those for IC specimens are summarized in Supplementary Table 1. BC and normal tissues were reanalyzed for this study. We obtained between 13,554,054 and 22,747,525 total reads and 4,837,296–19,218,202 mapped reads (Table 1). We focused on previously annotated miRNAs and detected 745–1,417 miRNAs for each sample. From comparisons of normal bladder samples, BC samples, and IC samples, we constructed miRNA expression signatures that included miRNAs exhibiting significantly up- or downregulated expression (Tables 2–5; Supplementary Tables 2–7; (|log2FC| > 1 and FDR < 0.05).

Table 1 Annotation of reads aligned to small RNAs
Full size table
Table 2 Top 20 miRNAs with downregulated expression in IC compared to normal bladder
Full size table
Table 3 Top 20 miRNAs with upregulated expression in IC compared to normal bladder
Full size table
Table 4 Top 20 miRNAs with downregulated expression in IC compared to BC
Full size table
Table 5 Top 20 miRNAs with upregulated expression in IC compared to BC
Full size table

To extract miRNAs that had marked downregulation in IC tissues, we downselected candidate miRNAs according to the strategy shown in Supplementary Figure 1. Among the 203 miRNAs that were significantly downregulated in IC tissues compared to normal tissues, we excluded miRNAs that were also markedly upregulated or downregulated in comparisons of normal tissues with BC tissues. Of the 151 remaining miRNAs, we focused on the miR-320 family (miR-320a/b/c/d/e), which is composed of 5 miRNAs (miR-320a/b/c/d/e) having the same seed sequence, given the high possibility that these miRNAs will target similar genes (Supplementary Figure 2). In our profile, expression of all miR-320 family members was markedly downregulated in IC tissues.

Identification of molecular targets regulated by the miR-320 family

We hypothesized that genes regulated by miRNAs that have decreased expression in IC tissues could contribute to IC pathogenesis. Using the TargetScanHuman 7.1 database for initial analyses, we found 5158 genes that had candidate target sites for the miR-320 family. Next, genome-wide gene expression analyses were performed to narrow down candidate genes. This time, by transfection with miR-320b and miR-320c, which were mostly downregulated in IC, genes whose expression decreased in T24 cells were extracted. The microarray data were deposited into the GEO database (GEO accession number: GSE106791). These analyses identified 162 and 222 genes as candidates for regulation by miR-320b and miR-320c, respectively. Finally, we integrated these results, 85 genes remained as candidate genes for regulation by the miR-320 family (Table 6). The strategy for selection of target genes is shown in Supplementary Figure 3. Among putative targets of miR-320 family, several genes, e.g., E2F1, RAB23, ITGB3, and TERT, have been reported to be associated with bladder cancer [23,24,25,26]. However, as far as the PubMed database was searched, no reports were found about the relationship between these genes and IC. Additionally, we assigned genes to KEGG (Kyoto Encyclopedia of Genes and Genomes) pathways using the GENECODIS program and identified 6 significantly enriched pathways (Supplementary Table 8).

Table 6 Candidate target genes commonly regulated by the miR-320 family
Full size table

E2F-1, E2F-2 and TUB expression in IC clinical specimens

We focused on 3 transcription factors (E2F-1, E2F-2 and TUB) and further analyses were performed. Previous studies showed that dysregulated expression or mutation of several transcription factors were deeply involved in human diseases and syndromes [27]. Function of E2F family is the transcriptional activation or repression of its target genes and pivotal role of cell proliferation, differentiation and apoptosis [28, 29]. The involvement of these genes and IC is obscure. We performed immunohistochemistry to assess protein expression of E2F-1, E2F-2, and TUB in IC clinical specimens. All three proteins were strongly expressed in IC urothelial cells compared with normal bladder and BC tissues (Fig. 2).

Fig. 2
figure 2

E2F-1, E2F-2, and TUB expression in clinical bladder specimens. Immunochemical staining showed that E2F-1, E2F-2, and TUB proteins were strongly expressed in urothelial cells from IC specimens relative to normal tissues and BC tissues

Full size image

Discussion

Recently developed RNA sequencing technology is suitable for compiling miRNA expression signatures. We previously described several miRNA expression signatures of human cancers based on RNA sequencing and identified novel RNA networks regulated by antitumor miRNAs [14,15,16,17,18,19]. In this study, we presented the miRNA signature of IC determined by RNA sequencing. Identification of dysregulated miRNAs in IC tissues compared with normal bladder tissues or bladder cancer tissues will be useful for exploring molecular pathogenesis in IC cells.

Among the downregulated miRNAs in IC tissues, we focused on the miR-320 family because all members of this family showed downregulated expression of miR-320 in IC tissues compared with normal bladder tissues. In contrast to our study, a miRNA expression profile of IC generated by PCR-based microarray analysis in a previous study showed that several miRNAs (miR-449b, miR-500, miR-328 and miR-320) were upregulated in IC tissues [30]. As a factor that causes different results from our research, analysis methods and differences in patient background may be considered. This report also showed that IC patients had notably decreased expression of neurokinin-1 (tachykinin receptors) that was presumed to be due to upregulated miR-320 expression in IC tissues [30]. Downregulation of the miR-320 family and the antitumor activity of these miRNAs have been reported for several cancers [31,32,33,34,35]. In bladder cancer, previous studies showed that miR-320a and miR-320c acted as antitumor miRNAs by targeting integrin beta-3 (ITGB3) and cyclin-dependent kinase 6 (CDK6), respectively [25, 36].

To identify possible genes involved in IC pathogenesis, we searched molecular pathways containing genes targeted by the miR-320 family using genome-wide gene expression analyses and in silico database analyses. Using these approaches, we identified 85 genes as putative targets of miR-320 family regulation. Among these genes, we focused on transcription factors due to their potential to significantly affect downstream RNA networks. Several studies indicated that dysregulation of transcription factor activity is associated with various human diseases [37, 38]. More recently, the urothelial master transcription factors TP63, SHH and FOXA1 were reported to act as novel diagnostic markers and to have potential involvement in the molecular pathology of IC [39]. In this study, we focused on 3 transcription factors: E2F transcription factor 1 (E2F-1), E2F transcription factor 2 (E2F-2) and tubby bipartite transcription factor (TUB) and confirmed upregulated protein expression of all three in IC lesions.

The E2F family of transcription factors is involved in cell cycle regulation and synthesis of DNA in higher eukaryotes, and plays a pivotal role during the G1/S transition in mammalian cells [28, 29]. E2F-1, E2F-2, and E2F-3a function as transcription activators, whereas E2F-3b, E2F-4, E2F-5, E2F-6, E2F-7, and E2F-8 act as transcription suppressors [28]. E2Fs are associated with cancer and are major target genes for the Rb tumor suppressor protein (pRb). Inactivation of pRb causes overexpression of E2Fs that can act as a cancer promoter [40]. Another report described the involvement of E2F-1 and E2F-2 overexpression in inflammation associated with traumatic spinal cord injury (SCI) in a mouse model. Results from that report suggested that SCI-induced activation of E2F-1 and E2F-2 expression may result in movement disorder and hypersensitivity after trauma. Furthermore, E2F-1 and E2F-2 knockout significantly reduced the amount of neuron death, neuroinflammation and tissue damage in the SCI mouse model, together with an enhanced greater neuroprotective effect [41]. Based on these earlier findings, overexpression of E2F-1 and E2F-2 could promote inflammation associated with IC.

The tubby protein encoded by the TUB gene is a common upstream cell signalling protein in multicellular eukaryotes. The tubby protein acts as a signalling factor that potentially couples with G protein activity [42, 43]. Moreover, tubby proteins are involved in neuronal differentiation and development, whereas mutations in tubby genes in mammals are associated with delayed obesity, sensorineural hearing loss and retinal degeneration [43,44,45]. However, the functions of tubby proteins in humans are unclear and there is no report that they are involved in inflammatory diseases.

The expression changes in these three transcription factors and the networks induced by these changes may indicate that they play a significant role in IC onset and progression.

In conclusion, in this study we determined the miRNA expression signature in IC by RNA sequencing and successfully identified dysregulated expression of miRNAs. We found that all members of the miR-320 family were downregulated in IC tissues and that E2F-1, E2F-2, and TUB were putative targets of regulation by these miRNAs. Immunohistochemistry showing increased levels of E2F-1, E2F-2, and TUB proteins in IC lesions support the sequencing results. In order to accurately understand the RNA network within the cell, it is ideal to use cells derived from the disease. From the latest genome technology, it is important that IC derived cell lines can be established and analyzed for its RNA networks. The miRNA signature of IC generated by RNA sequencing and in silico analyses could provide a basis to develop novel therapeutic targets for IC.