Abstract

   Cholangiocarcinoma (CCA), a malignancy of the biliary epithelium, can
   arise at any point in the biliary system. We previously reported that
   CIAPIN1 is detectable in the sera and that its overexpression was
   associated with poor prognosis and metastasis of CCA patients. In this
   study, we investigated further its expression in CCA tissues,
   biological functions, and related signaling pathways in CCA cells.
   First, we examined CIAPIN1 expression in CCA tissues of 39 CCA patients
   using immunohistochemistry (IHC). Then, CIAPIN1-related proteins
   expressed in CCA cells were identified using RNA interference (siRNA)
   and liquid chromatography–mass spectrometry (LC–MS/MS). To predict the
   functions and signaling pathways of CIAPIN1 in CCA cells, the
   identified proteins were analyzed using bioinformatics tools. Then, to
   validate the biological functions of CIAPIN1 in the CCA cell line,
   transwell migration/invasion assays were used. CIAPIN1 was
   overexpressed in CCA tissues compared with adjacent noncancerous
   tissues. Its overexpression was correlated with lymph node metastasis.
   Bioinformatic analyses predicted that CIAPIN1 is connected to the
   TGF-β/SMADs signaling pathway via nitric oxide synthase 1 (NOS1) and is
   involved in the metastasis of CCA cells. In fact, cell migration and
   invasion activities of the KKU-100 CCA cell line were significantly
   suppressed by CIAPIN1 gene silencing. Our results unravel its novel
   function and potential signaling pathway in metastasis of CCA cells.
   CIAPIN1 can be a poor prognostic factor and can be a promising target
   molecule for CCA chemotherapy.

   Keywords: cholangiocarcinoma, CIAPIN1, prognosis, siRNA, proteomics,
   metastasis, signaling pathway

1. Introduction

   Cholangiocarcinoma (CCA) is highly endemic in Northeast Thailand in
   relation to the high incidence of infection with the carcinogenic liver
   fluke, Opisthorchis viverrini (OV), which is the significant risk
   factor for CCA [[38]1]. CCA is usually asymptomatic in the early stage
   and often diagnosed in the advanced stage, which highly compromises
   therapeutic options, resulting in a dismal prognosis [[39]2]. In
   addition, patient outcome is poor due to metastasis [[40]3]. Currently,
   combination chemotherapy with gemcitabine and cisplatin is the standard
   treatment for advanced or metastatic CCA patients. However,
   chemotherapy responses typically are limited and median overall
   survival is dismal (less than 12 months) [[41]4]. Therefore, the
   identification of novel target molecules for CCA therapy is still
   required. For that, a better understanding of the critical signaling
   pathways and genetic alterations in CCA is necessary. There are certain
   mitochondrial features that are common in many cancers including tumor
   metastasis [[42]5]. In the previous study of our group, Chua-On et al.
   identified a total of 281 mitochondrial proteins, 105 of which were
   overexpressed in the cancerous tissue of CCA compared with adjacent
   non-cancerous tissues [[43]6]. Among them, cytokine-induced apoptosis
   inhibitor 1 (CIAPIN1) was identified as a potential prognostic
   biomarker for CCA [[44]7]. In particular, the serum CIAPIN1 level was
   significantly higher in CCA patients with lymph node metastasis than
   those without metastasis [[45]7].

   The CIAPIN1 protein is encoded by the CIAPIN1 gene, is accumulated in
   the nucleolus, and localized in the cytoplasm, nucleus, and
   mitochondria [[46]6,[47]8]. CIAPIN1, also known as anamorsin, is an
   anti-apoptotic molecule that has no sequence similarities to a series
   of apoptosis-associated molecules such as Bcl-2 or caspase family
   members. It is a key mediator of RAS signaling pathways and mediates
   the maintenance of hematopoiesis in the fetal liver [[48]9]. CIAPIN1 is
   physiologically expressed in many tissues, especially
   self-renewing/proliferative tissues. It plays a biological role in many
   cancers as a candidate indicator for diagnosis, prognosis, and
   therapeutic target in various human cancers [[49]10]. Although the
   expression level of CIAPIN1 has been found to relate to an oncogene or
   tumor suppressor in many different solid tumors [[50]10], its
   expression, biological roles, and its mechanism in CCA have not been
   investigated. Remarkably, the involvement of CIAPIN1 in tumor
   metastasis mechanisms is still unknown. This study aimed to examine the
   expression of CIAPIN1 in CCA tumor tissue specimens and evaluate its
   prognostic value. Furthermore, to elucidate the biological role of
   CIAPIN1 in CCA cells, we attempted to prove the possible signal
   transduction pathways of CIAPIN1 and the associations of
   CIAPIN1-related proteins with chemotherapy drugs. The results
   demonstrated that CIAPIN1 might be a therapeutic target molecule for
   the prevention of CCA metastasis.

2. Materials and Methods

2.1. Patients and Specimens

   The formalin-fixed, paraffin-embedded (FFPE) pathologically confirmed
   CCA tumor tissue specimens from 39 patients were provided by the
   Cholangiocarcinoma Research Institute, Faculty of Medicine, Khon Kaen
   University (CARI, KKU), Thailand. The patients underwent surgical
   operations at the Srinagarind Hospital, KKU, between 2010 and 2012.
   According to the reports from pathologists, lymph node metastasis was
   positive in 26 of 39 patients, whereas it was negative in 13 patients.
   The tissues used in this study were all leftover specimens obtained
   during the surgical treatment. Written informed consent for the use of
   leftover specimens for research purposes was obtained from the
   attending physicians before surgery. The original documents were kept
   in CARI, KKU. This series of experiments has been approved
   ([51]HE631387) by the Ethics Committee of KKU, Thailand, and all
   specimens were anonymously coded according to the guidelines.

2.2. CCA Cell Lines

   The three human CCA cell lines, KKU-M055 (JCRB1551), KKU-100 (JCRB1568)
   [[52]11], and KKU-M213A, which was renamed from KKU-213 (JCRB1557)
   [[53]12], and an immortalized cholangiocyte cell line, MMNK1 (JCRB1554)
   [[54]13] were obtained from the Japanese Collection of Research
   Bioresources Cell Bank (Tokyo, Japan) and were kindly provided by CARI,
   KKU. The cell lines were cultured and subcultured as described
   previously [[55]14]. All cell lines in this study were confirmed
   mycoplasma-free by specific PCR.

2.3. Immunohistochemistry and Evaluation of CIAPIN1 Expression in CCA Tissues

   The IHC was conducted according to the previously reported procedures
   [[56]15]. The rabbit polyclonal antibody against human CIAPIN1
   (Cat#orb377996, Biorbyt, Cambridge, UK) was used for IHC staining at
   1:150 dilution.

2.4. Western Blot Analysis

   The Western blot analysis was performed for the detection of CIAPIN1
   protein in CCA cell lysate as described previously [[57]7]. Briefly,
   cells were lysed with lysis buffer (Cat#9806, RIPA, Cell Signaling
   Technologies, Danvers, MA, USA). Protein amounts were determined using
   the Bradford assay (Bio-Rad, Hercules, CA, USA). Twenty-five micrograms
   of cell lysate were dissolved in a sample buffer for Western blotting.
   The rabbit polyclonal antibody against human CIAPIN1 (Cat#orb377996,
   Biorbyt, Cambridge, UK) was used at 1:1000 dilution for Western blot
   analysis in this study. β-actin antibody (Cat#ab227387, Abcam,
   Cambridge, UK) was used for detecting a loading control.

2.5. Transient Silencing of CIAPIN1 Gene Using siRNA

   Since CIAPIN1 expression in the cell lysates of KKU-100 and KKU-213A
   cells was higher than that of the other CCA cell lines, we selected
   these cell lines for the CIAPIN1 gene silencing experiments. In brief,
   for the CIAPIN1 gene silencing using a siRNA technique, the cells (4.5
   × 10^5 cells/well) were seeded in a 6-well plate and cultured overnight
   before being transfected with 75 pM of siCIAPIN1 (Cat#sc-60168, Santa
   Cruz Biotechnology, Inc., Dallas, TX, USA), while scrambled siRNA
   (Cat#AM4611, Invitrogen, Carlsbad, CA, USA) was used as a negative
   control. Transfection was carried out using Lipofectamine 3000
   (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s
   instructions. After 6 h of transfection, the culture medium was
   replaced with a complete medium, and the plates were incubated at 37 °C
   for 24 h. To check the suppression of protein expression, the cells
   were harvested in lysis buffer and incubated at 4 °C for 10 min. The
   cell lysate was centrifuged at 20,000× g (4 °C) for 30 min. The level
   of CIAPIN1 protein was determined using Western blot analysis with
   β-actin as a loading control. In addition, CIAPIN1-silenced and control
   scrambled cells were tested for migration and invasion assay in vitro
   as given below.

2.6. Sample Preparation for Tryptic Digestion

   After transfection of siRNA or scramble RNA, cell lysates of CIAPIN1
   gene-silenced and scramble-treated control cells were prepared and
   their protein content was determined by Lowry assay using BSA as a
   protein standard. In-gel digestion was performed using an in-house
   method developed by the Functional Proteomics Technology Laboratory,
   National Center for Genetic Engineering and Biotechnology, Thailand.
   Briefly, 4 µg of total proteins were directly incorporated into a 12.5%
   polyacrylamide and congealed in a microtube, reduced disulfide bonds
   using 5 mM dithiothreitol (DTT) in 10 mM AMBIC at 60 °C for 1 h, and
   alkylated sulfhydryl groups using 15 mM iodoacetamide (IAA) in 10 mM
   AMBIC at RT for 45 min in the dark. For digestion, samples were mixed
   with 50 ng/µL of sequencing grade trypsin (1:20 ratio) (Promega,
   Walldorf, Germany) and incubated at 37 °C overnight. Prior to LC–MS/MS
   analysis, the digested samples were dried and protonated with 0.1%
   formic acid before injection into LC–MS/MS. The LC–MS/MS data analysis
   and protein identification were performed as described previously
   [[58]16]. The LC–MS analysis of each sample was done in triplicate.

2.7. Selection of CIAPIN1-Related Proteins Using Jvenn

   The effects of CIAPIN1 silencing on protein expression patterns of cell
   lysates were determined using the mass spectrometry data sets of siRNA
   and scramble-treated cells. Overlapping and unique proteins were
   identified using jvenn software which is an integrative web-tool for
   comparing lists with Venn Diagrams and accessed on 5 November 2020
   ([59]http://bioinfo.genotoul.fr/jvenn) [[60]17]. To speculate on the
   active role(s) of CIAPIN1 in CCA and its possible signaling pathway,
   proteins that were uniquely expressed in scramble-treated, but not in
   siRNA-treated KKU-213A and KKU-100 cells, were selected as the
   candidate protein group.

2.8. Protein–Protein Interaction Analysis for Prediction of CIAPIN1 Signaling
Pathway and Relationship with Chemotherapeutic Drugs

   The potential interaction of the identified proteins was analyzed using
   the search tool STITCH version 5.0 ([61]http://stitch.embl.de/,
   accessed on 15 November 2020). In brief, 150 gene names of related
   proteins were put into a box of multiple name items to predict the
   CIAPIN1 signaling pathway. Meanwhile, chemotherapeutic drugs
   (cisplatin, 5-fluorouracil, gemcitabine, doxorubicin) were added for
   exploring the interaction between CIAPIN1 and chemotherapeutic drugs.
   Then “Homo sapiens” was selected as the organism. In order to obtain an
   optimal network configuration, a default intermediate confidence score
   of >0.4 was used; and our set of protein and the first spheres of
   interaction were varied in the range of 10 nodes. To interpret the
   interaction view, strong associations are represented by the thicker
   lines, weak associations by thin lines, protein–protein interactions in
   grey, chemical–protein interactions in green, and interactions between
   chemicals in red lines.

2.9. KEGG Pathway Enrichment Analysis to Predict CIAPIN1 Signaling Pathway

   DAVID was used to identify significant groups of genes and pathways
   that are related to the data set used in this study. All 150
   CIAPIN1-related proteins were imported into the DAVID Bioinformatics
   resource version 6.8 ([62]https://david.ncifcrf.gov/, accessed on 15
   November 2020) for pathway annotation analysis. Generated files of KEGG
   pathway analysis were obtained. The EASE score threshold (maximum
   probability) = 0.1, as a default (p-value), was selected.

2.10. Correlation between CIAPIN1 and Related Proteins in the Signaling
Pathway

   To verify the correlation between CIAPIN1 and related proteins in the
   signaling pathways, we used mRNA expression data of CCA and a normal
   tumor from the open-access internet database of Gene Expression
   Profiling Interactive Analysis (GEPIA 2,
   [63]http://gepia2.cancer-pku.cn/, Peking University, Beijing, China,
   accessed on 15 November 2020) [[64]18]. The correlation of mRNA
   expression levels between CIAPIN1 and related proteins in the signaling
   pathways was determined using the pairwise analysis of the Spearman
   statistic tool. A p-value < 0.05 was considered for the statistical
   significance.

2.11. Molecular Docking Verification of CIAPIN1 Binding to Its Related
Proteins and Chemotherapeutic Drugs

   To identify potential interactions between a given protein and a
   ligand, CB-Dock was used ([65]http://cao.labshare.cn/cb-dock/ accessed
   on 16 May 2021). We obtained the three-dimensional (3D) X-ray crystal
   structure of human CIAPIN1 (PDB ID: 4M7R, solution 1.8 Å), NOS1 (PDB
   ID: 6CID, solution 1.75 Å), and SMAD2 (PDB ID: 6YIA, solution 1.30 Å)
   from the Protein Data Bank (PDB; accessed on 16 May 2021). The
   structure of doxorubicin (Compound CID: 31703) was obtained from
   PubChem ([66]https://pubchem.ncbi.nlm.nih.gov/). In brief, a protein
   file in the PDB format and a ligand file in the MOL2, MOL, or SDF, were
   derived from the protein database ([67]http://www.rcsb.org) to input
   into CB-Dock. To prepare the ligand files, Open Babel was used to
   convert NOS1 and SMAD2 to SDF format [[68]19]. Next, using CB-Dock, we
   predicted cavities of the protein and calculated the centers and sizes
   of the top N (N = 5 by default) cavities. The final results are
   displayed after the computation of N rounds, including binding scores,
   cavity sizes, and docking parameters of the predicted binding modes.
   The 3D structures of the highest binding modes on the web page were
   selected and presented. The lower the Vina scores are, the more stable
   the ligand binding to the receptor [[69]20], the reference of Vina
   scores as a cut-off was less than −5.6 kcal/mol [[70]21].

2.12. Transwell Cell Migration and Invasion Assay

   The effect of CIAPIN1-silencing on CCA cell motility was determined
   using a transwell 24-well plate. For the migration assay, the chamber
   was equipped with the membrane filter of 8 µm pore size (Corning,
   Kennebunk, ME, USA), and for the invasion assay, the chamber was
   equipped with the Matrigel-coated membrane (Corning, Kennebunk, ME,
   USA). CIAPIN1-silenced and scramble-treated KKU-100 cells were
   inoculated into the upper chamber with a cell density of 4 × 10^4
   cells/200 µL of serum-free medium. The lower chamber was filled with
   600 µL of complete medium. After incubation for 48 h, migrated cells on
   the lower surface of the membrane were fixed with 100% methanol and
   stained with 2% crystal violet in 2% ethanol for 15 min. They were then
   washed in 1xTBS 5 times. The migrated/invaded cells were counted as the
   number of nuclei under an inverted microscope fitted with an objective
   lens of 10×. Six randomized fields for each membrane filter were
   counted [[71]22].

2.13. Statistical Analysis

   The data are presented as the median ± quartile deviation or the mean ±
   standard deviation (with the minimum to the maximum range). The
   difference in the values between two independent sample groups was
   estimated using the Mann–Whitney U test. The association between serum
   CIAPIN1 levels and patients’ clinicopathological parameters was
   analyzed using Fisher’s exact test. X-tile software version 3.6.1
   [[72]23] was used to find out the optimal cut-off values of CIAPIN1
   expression in tissues in relation to the prognosis of CCA. This cut-off
   was used to dichotomize its relative intensity as low and high levels
   and their correlation with clinicopathological parameters including
   survival time was determined. The Cox proportional hazards regression
   model was used for univariate and multivariate analysis. p < 0.05 was
   considered statistically significant. IBM SPSS v.26 software (IBM
   Corp., Armonk, NY, USA) was used for statistical analyses.

3. Results

3.1. Immunohistochemical Detection of CIAPIN1 Protein in CCA Tissues

   The expression of CIAPIN1 in CCA was investigated immunohistochemically
   using 39 CCA tissues. The results showed that CIAPIN1 was stained
   predominantly in the cytoplasm of both cancerous and normal cells
   ([73]Figure 1b,c). [74]Figure 1a is a negative control staining of CCA
   tissue without primary antibodies. Using the H-score system, CIAPIN1
   was highly (p < 0.0001) expressed in CCA (mean ± SD of H-score = 216 ±
   25), compared with normal bile ducts in the CCA adjacent areas (mean ±
   SD of H-score = 51 ± 21) ([75]Figure 1d). In addition, comparing
   patients with and without lymph node metastasis, our data showed that
   CIAPIN1 expression in CCA of the patients in the lymph node metastasis
   group was significantly higher than that of the non-metastasis group (p
   < 0.0001, [76]Figure 1e).

Figure 1.

   [77]Figure 1

   [78]Figure 1
   [79]Open in a new tab

   Immunohistochemical detection and the representative images of CIAPIN1
   in CCA tissues (taken at a magnification of 400×). (a) The negative
   control of CCA tissue without primary antibody followed by secondary
   antibody. (b) Weak CIAPIN1 staining of the normal monolayer of
   cholangiocytes in normal tissues adjacent to CCA tissue. (c) Strong
   cytoplasmic staining in CCA cells. (d) The expression of CIAPIN1 in the
   cancerous region was higher than in the adjacent region with
   statistical significance. (e) CIAPIN1 expression in cancerous regions
   of the lymph node (LN) metastasis group was higher than the
   non-metastasis group with statistical significance. Data were presented
   as mean ± SD; Mann–Whitney U test was used; *** statistically
   significant with p < 0.0001.

3.2. Association between CIAPIN1 Expression in CCA Tissues and
Clinicopathological Features of Patients

   We analyzed the relationship between CIAPIN1 expression in cancerous
   tissue and the clinicopathological characteristics of CCA patients. At
   first, using X-tile software, the optimal cut-off value of CIAPIN1
   H-score to divide CCA patients into high and low CIAPIN1 expression
   groups was determined to be 207.5 ([80]Figure S1). Accordingly, 15 of
   39 CCA patients were in the low expression group and 24 in the high
   expression group. The patients’ characteristics such as age, gender,
   tumor size, the histopathological grade with non-papillary and
   papillary, vascular invasion, and lymph node metastasis were summarized
   in [81]Table 1. As can be seen, high CIAPIN1 expression was
   significantly correlated with lymph node metastasis (p = 0.010),
   intraductal invasion (p = 0.048), and overall survival time (p =
   0.013).

Table 1.

   Clinicopathological associations of CIAPIN1 expression in patients with
   CCA.
   Patients’ Characteristics CIAPIN1 Expression (H-Score) p
   Low ≤ 207.5(n = 15) High > 207.5(n = 24)
   Age (years) 56.67 ± 11.23 61.67 ± 6.53 0.091 ^a
   Gender 0.317 ^b
   Male 11 (73.3%) 13 (54.2%)
   Female 4 (26.7%) 11 (45.8%)
   Tumor size 0.176 ^b,c
   <5 cm 4 (26.7%) 12 (54.5%)
   ≥5 cm 11 (73.3%) 10 (45.5%)
   Histopathological grade 0.055 ^b
   Non-Papillary 10 (66.7%) 8 (33.3%)
   Papillary 5 (33.3%) 16 (66.7%)
   Lymph node metastasis 0.010 ^b,*
   No 11 (73.3%) 7 (29.1%)
   Yes 4 (26.7%) 17 (70.9%)
   Intraductal invasion 0.048 ^b,*
   No 11 (73.3%) 9 (37.5%)
   Yes 4 (26.7%) 15 (62.5%)
   Vascular invasion 0.711 ^b
   No 12 (80.0%) 17 (70.8%)
   Yes 3 (20.0%) 7 (29.2%)
   Survival time (days) 562 ± 428 291 ± 225 0.013 ^a,*
   [82]Open in a new tab

   The difference between high- and low-CIAPIN1 groups was estimated using
   ^a Mann–Whitney U test and ^b Fisher’s exact test; ^c n = 37; *
   Statistically significant correlation as p < 0.05.

3.3. High Expression of CIAPIN1 Is Associated with Poor Prognosis of CCA
Patients

   A log-rank test with Kaplan–Meier estimates was adopted to determine
   whether CIAPIN1 expression in CCA tissues can be a prognostic factor
   for the survival of CCA patients. The overall survival analysis using
   the Kaplan–Meier method revealed the mean (95%CI) survival time of CCA
   patients with low and high CIAPIN1 expression in CCA tissues was 799
   days (530–1067) and 462 days (258–667), respectively. In addition,
   higher CIAPIN1 expression was associated with shorter survival time
   ([83]Figure 2, p = 0.016). The Cox proportional hazards model analysis
   was conducted to examine the effect of the covariates, including age,
   sex, invasion, lymph node metastasis, histology grading, and the tissue
   CIAPIN1 expression level, showing that only CIAPIN1 is an independent
   factor for poor prognosis of CCA (HR = 4.01, 95% CI: 1.131–14.195; p =
   0.031, [84]Table 2). Then, we hypothesized that CIAPIN1 might affect
   cell behavior such as metastasis of CCA. To substantiate this
   hypothesis, we carried out CIAPIN1 gene silencing and bioinformatic
   analyses using CCA cell lines.

Figure 2.

   Figure 2
   [85]Open in a new tab

   Kaplan–Meier plots showing the comparison of overall survival of CCA
   patients according to CIAPIN1 expression. The curves represent the
   overall survival time of CCA patients having high (solid line) and low
   (dashed line) H-score levels. A significant difference of the survival
   time was observed between high and low H-score CIAPIN1 level groups by
   log-rank test and p < 0.05.

Table 2.

   Univariate and multivariate Cox regression analyses of
   clinicopathological parameters and CIAPIN1 protein in CCA tissues.
   Clinicopathological Factors Univariate Analysis Multivariate Analysis
   n HR 95% CI p n HR 95% CI p
   Age (years, median range) 0.551 0.740
   ≤60 21 1.00 - 20 1.00 -
   >60 18 1.31 0.541–3.160 17 0.84 0.307–2.316
   Gender 0.937 0.567
   Male 24 1.00 - 23 1.00 -
   Female 15 1.04 0.441–2.426 14 0.73 0.248–2.146
   Tumor size 0.466 0.453
   <5 cm 16 1.00 - 16 1.00 -
   ≥5 cm 21 1.40 0.565–3.481 21 1.57 0.481–5.150
   Histopathological grade 0.795 0.317
   Non-Papillary 18 1.00 - 17 1.00 -
   Papillary 21 1.12 0.469–2.686 20 2.68 0.389–18.422
   Lymph node metastasis 0.217 0.280
   No 18 1.00 - 16 1.00 -
   Yes 21 1.73 0.724–4.130 21 1.94 0.583–6.420
   Intraductal invasion 0.822 0.242
   No 20 1.00 - 19 1.00 -
   Yes 19 0.91 0.381–2.150 18 0.24 0.033–1.773
   Vascular invasion 0.882 0.847
   No 29 1.00 - 28 1.00 -
   Yes 10 1.08 0.392–2.977 9 0.89 0.281–2.834
   CIAPIN1 (H-score) 0.022 * 0.031 *
   Low (≤207.5) 15 1.00 - 15 1.00 -
   High (>207.5) 24 3.33 1.187–9.340 22 4.01 1.131–14.195
   [86]Open in a new tab

   HR = hazard ratio; CI = confidence interval. * Statistically
   significant as p < 0.05.

3.4. Expression of CIAPIN1 in CCA Cell Lines

   We first compared the expression of CIAPIN1 in three CCA cell lines and
   an immortal biliary epithelial cell line, MMNK1. As shown in [87]Figure
   3, the level of CIAPIN1 was remarkably variable among all four cell
   lines (KKU-100, KKU-213A, KKU-055, and MMNK1). From this result, we
   selected KKU-M213 and KKU-100 for CIAPIN1 gene silencing experiments.

Figure 3.

   [88]Figure 3
   [89]Open in a new tab

   Expressions of CIAPIN1 in CCA cell lines. (a) The protein levels of
   CIAPIN1 were determined by western blotting (WB) in CCA cell lines. (b)
   Relative expression of CIAPIN1 in CCA cell lines was examined by WB and
   the relative expression was normalized to β-actin, which is a loading
   control. The data were presented as the mean ± SD from triplicate
   independent experiments (** p < 0.01; *** p < 0.001).

3.5. Effects of CIAPIN1 Gene Silencing of CCA Cell Lines

   To investigate the biological roles of CIAPIN1 in CCA cells, the
   effects of CIAPIN1 gene suppression were examined using siRNA on both
   KKU-M213A and KKU-100 cells. As shown in [90]Figure 4, CIAPIN1
   expression of both CCA cell lines was transiently suppressed by siRNA
   at 24 h after transfection. Western blot analysis showed that the
   expression of CIAPIN1 was successfully suppressed in both KKU-M213A and
   KKU-100 cells with dominant suppression in KKU-100. Then, to predict
   biological function and possible signaling pathways of CIAPIN1,
   proteomic analysis of gene-silenced and scramble KKU-213A and KKU-100
   cells at 24 h was performed using mass spectrometry.

Figure 4.

   [91]Figure 4
   [92]Open in a new tab

   The effects of CIAPIN1 gene silencing on the CCA cell lines. (a)
   Western blot analysis showing suppressed CIAPIN1 protein after gene
   silencing. (b) Western blot examined the relative expression of CIAPIN1
   in CCA cell lines after gene silencing and normalization to β-actin
   expression. β-actin was used as a control for loading protein.
   scKKU-100 and scKKU213A are scramble-treated (without siRNA treatment)
   cells; siKKU-100 and siKKU213A are CIAPIN1-silenced-treated cells. The
   data were presented as the mean ± SD from triplicate independent
   experiments (* p < 0.05; ** p < 0.01).

3.6. Protein Expression Patterns of CIAPIN1 Gene-Silenced and
Scramble-Treated KKU-213A and KKU-100 Cell Lines

   Then, to predict possible signaling pathways of CIAPIN1, proteomic
   analysis of gene-silenced and scramble-treated KKU-213A and KKU-100
   cells was performed at 24 h. A total of 16,357 proteins were isolated
   from the cell lysates of CIAPIN1 gene-silenced and scramble-treated
   KKU-213A and KKU-100 cell lines. Then, using UniProt, 3752 proteins
   without gene names were excluded. Using jvenn software, we identified
   4410 proteins from scramble-treated KKU-213A, 1308 from siRNA-treated
   KKU-213A, 5015 from scramble-treated KKU-100, and 4636 from
   siRNA-treated KKU-100 ([93]Figure 5a). Among them, we identified 500
   proteins that were commonly expressed in scramble KKU-213A and KKU-100
   cells but not expressed in siRNA-treated cells ([94]Table S1).
   According to the MS intensity of each protein, the top 30% of 500
   proteins were selected for further analysis to predict functions and
   signaling pathways of CIAPIN1 in CCA cells ([95]Figure 5b).

Figure 5.

   [96]Figure 5
   [97]Open in a new tab

   Identification of CIAPIN1-related proteins. (a) Venn diagram presents
   the number of proteins in each sample and the degree of individual
   overlap among CIAPIN1 gene-silenced and scramble-treated KKU-213A and
   KKU-100 cells. The red circle illustrates the number of candidate
   proteins as CIAPIN1-related proteins. (b) Flowchart of selection
   CIAPIN1-related proteins and prediction of the signaling pathway in
   metastasis of CCA cells.

3.7. Construction of Protein–Protein Interaction Network of CIAPIN1 and its
Related Proteins

   To identify potential CIAPIN1 signaling pathways and functions, the
   protein–protein interaction (PPI) network was constructed by importing
   150 CIAPIN1-related proteins selected above to the STITCH version 5. As
   shown in [98]Figure 6, the PPI network indicated several possible
   signaling pathways. Among them, we speculated that the signaling
   pathway of CIAPIN1 might be mediated via nitric oxide synthase 1
   (NOS1), to TGFβR1/SMAD2 system. Then, the KEGG pathway enrichment
   analysis was performed by running DAVID version 6.8.

Figure 6.

   Figure 6
   [99]Open in a new tab

   Protein interaction networks of CIAPIN1 and related proteins of
   scramble-treated CCA cells. CIAPIN1 interacted directly with nitric
   oxide synthase 1 (NOS1), and indirectly with mothers against
   decapentaplegic homolog 2 (SMAD2) and transforming growth factor beta
   receptor 1 (TGFβR1). Thicker lines represent strong associations. Thin
   lines represent weak associations.

3.8. Prediction of CIAPIN1 Signaling Pathways in CCA Cells

   To elucidate possible functions and signaling pathways of CIAPIN1 in
   CCA cells, KEGG pathway enrichment analysis was performed by importing
   150 CIAPIN1-related proteins using DAVID version 6.8. The results show
   that nine CIAPIN1-related proteins including ERBB3 (Receptor
   tyrosine-protein kinase erbB-3), ITPR1 (Inositol 1,4,5-trisphosphate
   receptor type 1), CACNA1A (Voltage-dependent P/Q-type calcium channel
   subunit alpha-1A), ADRA1D (Alpha-1D adrenergic receptor), NOS1, ADRA1A
   (Alpha-1A adrenergic receptor), SMAD2, CSNK2A1 (Casein kinase II
   subunit alpha), and PARD3 (Partitioning defective 3 homolog) were
   involved in four KEGG signaling pathways with variable combinations
   ([100]Table 3). In terms of CCA biology, both the calcium signaling
   pathway and the adherens junction signaling pathway are assumed to be
   involved in tumor metastasis mechanisms. Moreover, these nine matching
   proteins were related to cell migration, invasion, and metastasis of
   tumor cells. Their names, abbreviations, and synonyms were used for
   further analysis.

Table 3.

   The KEGG pathway enrichment analysis of CIAPIN1-related proteins in the
   top 30% of scramble-treated CCA cell line samples by DAVID ver. 6.8.
   IDs Pathway Description p Count CIAPIN1-Related Matching Proteins
   hsa04020 Calcium signaling pathway 0.0071 6 ERBB3, ITPR1, CACNA1A,
   ADRA1D, NOS1, ADRA1A
   hsa04970 Salivary secretion 0.0206 4 ITPR1, ADRA1D, NOS1, ADRA1A
   hsa04730 Long-term depression 0.0630 3 ITPR1, CACNA1A, NOS1
   hsa04520 Adherens junction 0.0844 3 SMAD2, CSNK2A1, PARD3
   [101]Open in a new tab

   ERBB3 (Receptor Tyrosine-protein Kinase erbB-3), ITPR1 (Inositol
   1,4,5-trisphosphate Receptor Type 1), CACNA1A (Voltage-dependent
   P/Q-type Calcium Channel Subunit Alpha-1A), ADRA1D (Alpha-1D Adrenergic
   Receptor), NOS1 (Nitric Oxide Synthase, Brain), ADRA1A (Alpha-1A
   Adrenergic Receptor), SMAD2 (Mothers Against Decapentaplegic Homolog
   2), CSNK2A1 (Casein Kinase II Subunit Alpha), and PARD3 (Partitioning
   defective 3 homolog). EASE score (p-value) < 0.1 as default was
   obtained from DAVID ver. 6.8.

   To predict the potential signaling pathways of CIAPIN1 in CCA cells,
   these nine CIAPIN1-related proteins that appeared in calcium signaling
   and adherence junction signaling pathways were imported only to STITCH.
   PPI analysis indicated that CIAPIN1 interacts with this set of nine
   proteins (ERBB3, ITPR1, CACNA1A, ADRA1D, ADRA1A, SMAD2, CSNK2A1, and
   PARD3) through NOS1. Especially, it reacted to SMAD2, SMAD4, and TGFBR1
   via NOS1 ([102]Figure 7), which was accordant to the PPI network of 150
   CIAPIN1-related proteins as shown in [103]Figure 7. Thus, we speculate
   that CIAPIN1 was involved in the TGF-β/SMADs signaling pathway via NOS1
   in metastasis of CCA cells.

Figure 7.

   Figure 7
   [104]Open in a new tab

   Specific prediction of CIAPIN 1 signaling pathway in metastasis of CCA
   cells. CIAPIN1 was predicted to be involved in the transforming growth
   factor beta (TGF-β) and SMADs signaling pathway through NOS1 in the
   metastasis process of CCA cells (red rectangle). Thicker lines
   represent strong associations. Thin lines represent weak associations.

3.9. Correlation between CIAPIN1 and its Related Proteins in Signaling
Pathway

   Since, via NOS1, CIAPIN1 might be involved in TGF-β/SMADs signaling
   pathway in the metastasis process of CCA cells ([105]Figure 7). The
   Spearman correlation analysis tool on the open-access internet database
   of GEPIA2 was used to verify the relationship between CIAPIN1 and
   related proteins in CCA and normal tissue ([106]Figure 8). The scatter
   plots showed that the mRNA level of CIAPIN1 significantly correlated
   with that of SMAD2, SMAD4, and TGFβR1 which suggested strong and
   moderate positive association (R = 0.64, p < 0.0001; R = 0.54, p <
   0.0001; and R = 0.55, p < 0.0001; respectively).

Figure 8.

   [107]Figure 8
   [108]Open in a new tab

   Correlation of mRNA expression of pair genes. (a) CIAPIN1 and SMAD2;
   (b) CIAPIN1 and SMAD4; and (c) CIAPIN1 and TGFβR1 in CCA and normal
   tumor. Data were retrieved from GEPIA2 tool as log2 of transcript per
   million (TPM) using the Spearman correlation analysis, p < 0.05.

3.10. Molecular Docking of Protein–Ligand between CIAPIN1 and Each Related
Proteins in Signaling Pathway

   To verify the protein–protein interaction of CIAPIN1 and its related
   proteins in the predicted signaling pathway, CB-Dock analysis was
   performed to see the binding geometries of the key interactions.
   According to PPI networks in [109]Figure 6 and [110]Figure 7, we
   selected the docking between CIAPIN1 and NOS1, and NOS1 and SMAD2.
   After docking analysis, their Vina scores were less than −5.6 kcal/mol
   which is the cut-off. The highest Vina score between CIAPIN1 and NOS1
   was −7.6 kcal/mol and the largest cavity size was 7734 ([111]Figure
   9a), whereas those between NOS1 and SMAD2 were −6.6 kcal/mol Vina
   score, and 443 cavity size ([112]Figure 9b). In accordance with the
   results of PPI analysis using STITCH, CB-Dock indicated strong binding
   of CIAPIN1 to NOS1 which interacts with SMAD2.

Figure 9.

   [113]Figure 9
   [114]Open in a new tab

   The interactive 3D viewer illustrating the highest Vina score of
   binding mode and representative positions of protein–ligand docking at
   their pocket binding sites. Blue cartoon ribbons represent (a) CIAPIN1
   protein interface to the NOS1 (ball and stick) and (b) SMAD2 protein
   interface to the NOS1 (ball and stick). Interaction bond is presented
   by the dashed line.

3.11. Effect of CIAPIN1 Gene Silencing on Cell Motilities of CCA Cell Line

   Since CIAPIN1 was predicted to activate the TGF-β/SMADs signaling
   pathway via NOS1 in CCA metastasis, we investigated the role of CIAPIN1
   in the migration and invasion of CCA cells. For this purpose, we
   selected the KKU-100 cell line because the CIAPIN1 protein expression
   level in the cell lysate was much higher than the other two cell lines
   ([115]Figure 3). When the CIAPIN1 gene of KKU-100 cells was silenced
   using siRNA, both cell migration ([116]Figure 10a) and invasion
   ([117]Figure 10b) were significantly (p < 0.0001) suppressed.

Figure 10.

   [118]Figure 10

   [119]Figure 10
   [120]Open in a new tab

   The effects of CIAPIN1 gene silencing on cell migration (a) and
   invasion (b) of KKU-100. The cells treated without siRNA-CIAPIN1
   (scramble, SC) were used as controls. The data were presented as the
   mean ± SD from triplicate independent experiments (*** p < 0.0001).

3.12. The Correlation of CIAPIN1 and its Related Proteins to Chemotherapeutic
Drugs

   To elucidate whether CIAPIN1 can be a molecular target for cancer
   chemotherapy, we predicted the networks of CIAPIN1 and related proteins
   in the signaling pathway and chemotherapeutic drugs using STITCH. The
   results showed that CIAPIN1 interacts directly with doxorubicin and
   indirectly with cisplatin via NOS1 ([121]Figure 11a). In CB-Dock
   analysis, the highest Vina score between CIAPIN1 and doxorubicin was
   −8.4 kcal/mol with the largest cavity size of 7734 ([122]Figure 11b).
   Thus, CIAPIN1 can strongly bind to doxorubicin. Moreover, PPI analysis
   revealed that CIAPIN1, NOS1, doxorubicin, and cisplatin interacted with
   the TGFβ/SMADs system ([123]Figure 11a). Hence, CIAPIN1 might play a
   role in cancer drug resistance and be a molecular target for cancer
   chemotherapy.

Figure 11.

   [124]Figure 11
   [125]Open in a new tab

   Involvement of CIAPIN1 and related proteins in networks of
   protein–chemotherapeutic drug interaction. (a) The interactive 3D
   viewer illustrates the highest Vina score of binding mode and
   represents positions of protein–ligand docking at their pocket binding
   sites. Blue cartoon ribbons represent the CIAPIN1 protein interface to
   the doxorubicin ligand (ball and stick). The interaction bond is
   presented by the dashed line. (b) Protein–chemical interaction of
   CIAPIN1 (bold blue circle), related proteins (blue circles), and
   chemotherapeutic drugs cisplatin, 5-fluorouracil, gemcitabine, and
   doxorubicin (red circles). Thicker lines represent strong associations.
   Thin lines represent weak associations.

4. Discussion

   In oncogenesis research, the overexpression of CIAPIN1 has been
   reported in various solid tumors and is considered an oncogene or tumor
   suppressor in different tumor types. However, the expression and role
   of CIAPIN1 in CCA have remained unknown. Recently we have found that
   the serum CIAPIN1 level is associated with the prognosis of CCA
   patients [[126]7]. In the present study, we investigated the expression
   of CIAPIN1 in human CCA tissues and CCA cell lines to elucidate the
   role of CIAPIN1 and its molecular mechanisms in CCA tumorigenesis/tumor
   progression.

   Using immunohistochemical methods, we demonstrated that CIAPIN1 is
   overexpressed in the majority of CCA tissues ([127]Figure 1). Our data
   were in accordance with the previous data of bioinformatic analysis of
   CIAPIN1 expression in CCA tissue compared with normal samples using the
   GEPIA2 database [[128]7]. Furthermore, CIAPIN1 expression was higher in
   CCA tissues of patients with lymph node metastasis ([129]Figure 1e).
   Accordant with CIAPIN1 expression in CCA tissues, the serum CIAPIN1
   level in CCA patients with lymph node metastasis was higher than that
   without metastasis [[130]7]. Overexpression of CIAPIN1 was reported in
   epithelial ovarian cancer [[131]24] and metastatic ovarian serous
   carcinoma [[132]25]. These data indicate that CIAPIN1 may play an
   oncogenic role in diverse cancers. Our data suggest that CIAPIN1 might
   act as an oncogene to promote the malignant progression of CCA.

   In this study, CIAPIN1 is an independent prognostic factor for CCA. In
   addition, upregulation of CIAPIN1 is related to lymph node metastasis.
   In our previous study, the serum CIAPIN1 levels was an independent
   unfavorable prognostic factor for patients with CCA [[133]7]. Already,
   Lopes et al. reported that the expression analysis of CIAPIN1 may
   predict metastasis and poor prognosis of patients with gastric cancer
   [[134]26]. Furthermore, the decrease in CIAPIN1 expression is
   significantly associated with a longer survival time of diffuse large B
   cell lymphoma [[135]27], colorectal cancer [[136]28], pancreatic cancer
   [[137]29], and non-small-cell lung carcinoma [[138]30]. A further study
   with large sample size is needed to unravel the correlation between
   lymph node metastasis and CIAPIN1 expression, and also to identify
   whether CIAPIN1 can be a precise and reproducible prognostic marker and
   metastasis indicator for CCA patients. Moreover, this approach should
   expand to other types of cancers.

   In the present study, using a combination of gene silencing and mass
   spectrometry techniques followed by bioinformatic analysis revealed
   that CIAPIN1 and related proteins are involved in calcium signaling and
   adherens junction pathways, both of which are related to tumor
   metastasis [[139]31,[140]32]. Moreover, STITCH analysis combined with
   KEGG signaling pathway search revealed that, via NOS1, CIAPIN1
   interacted with nine proteins related to cell migration, invasion, and
   metastasis of tumor cells ([141]Figure 7). Thus, CIAPIN1 may play an
   important role in the metastatic process of CCA. To validate the role
   of CIAPIN1 in CCA metastasis, we examined the effects of CIAPIN1 gene
   silencing on cell migration and invasion of CCA cells. The results
   showed that cell motility was significantly suppressed after CIAPIN1
   silencing ([142]Figure 10). In contrast to our results, in the case of
   non-small-cell lung carcinoma (NSCLC), CIAPIN1 overexpression inhibited
   cell migration and invasion and suppressed the expression of MMPs and
   EMT-associated markers [[143]30]. Thus, CIAPIN1 might have a contrary
   function among different cancers or even within CCA via different
   signal pathways. Further experiments are needed to validate possible
   multifunctional roles of CIAPIN1 in cancer, more CCA cell lines, and
   also other tumor cell lines should be used to explore the biological
   function of CIAPIN1 in various cancers.

   In this study, PPI analysis illustrated that CIAPIN1 interacted with
   SMAD2, SMAD4, and TGFβR1 via NOS1 ([144]Figure 7). In addition, the
   Spearman correlation analysis of GEPIA2 showed a positive correlation
   of the mRNA level of CIAPIN1 with that of SMAD2, SMAD4, and TGFβR1
   ([145]Figure 8). Moreover, NOS1 and SMAD2 were expressed only in
   scramble-treated CCA cells. Thus, we speculated that NOS1 and SMAD2
   could be downstream of the CIAPIN1 signaling pathway. In fact, CB-Dock
   analysis revealed a strong binding between CIAPIN1 and NOS1, and also
   NOS1 and SMAD2 ([146]Figure 9). Recently, Medvedev et al. demonstrated
   the direct interaction between NOS1 and SMAD2 [[147]33], both of which
   are involved in TGF-β signaling [[148]34,[149]35], suggesting the
   involvement of TGF-β signaling downstream of CIAPIN1. In previous
   studies, NOS1 is overexpressed in various types of cancer, and its
   expression is associated with tumor progression. For instance, higher
   NOS1 expression promotes the proliferation and invasion of ovarian
   cancer [[150]36]. Accordingly, nitric oxide (NO) is synthesized from
   L-Arginine and oxygen by a family of enzymes termed nitric oxide
   synthases, including NOS1, which regulates the TGF-β signaling of
   endothelial cells [[151]34]. Moreover, TGF-β is one of the major
   signaling pathways that promote CCA progression. TGF-β induces EMT in
   CCA cell lines [[152]37]. Interestingly, tumor invasion and migration
   are also mediated via the TGF-β/SMAD4 signaling pathway [[153]38].
   Taken together, we speculate that, via NOS1, CIAPIN1 activates the
   TGF-β/SMADs signaling pathway to augment tumor metastasis ([154]Figure
   12). Since all those signaling pathways are speculated and based on
   results of in silico analysis, further direct experiments are required
   to confirm the interaction between CIAPIN1 and NOS1, and the modulation
   of CIAPIN1 on up and downregulation of these target expressions through
   TGF-β/SMADs signaling pathway.

Figure 12.

   Figure 12
   [155]Open in a new tab

   Schematic representation of the predicted signaling pathway of how
   CIAPIN1 promotes the metastasis and chemoresistance of CCA. CIAPIN1 was
   predicted to activate SMAD2 signal transduction and involve in the
   TGF-β signaling pathway via NOS1. The result indicated an increase in
   migration and invasion, which lead to metastasis in CCA. CIAPIN1
   interacted directly with doxorubicin, led to promote chemoresistance
   function in CCA cells. The molecules in red rectangles were
   demonstrated in the present study.

   In terms of multidrug resistance, Wang et al. reported that
   overexpression of CIAPIN1 contributes to multidrug resistance (MDR) in
   breast cancer. CIAPIN1 gene silencing enhanced the sensitivity of tumor
   cells to doxorubicin in drug-resistant breast cancer xenografts in this
   nude mouse model [[156]39]. Moreover, Lu et al. showed that, in the
   MCF7/ADM cell line of breast cancer, CIAPIN1 gene silencing by siRNA
   reduced the drug resistance against epirubicin, paclitaxel, and
   gemcitabine by regulating MDR1 and P53 expression [[157]40]. However,
   the mechanism of its effects on the chemoresistance of human CCA
   remained undefined. In the present study, the protein–chemotherapy drug
   interaction network analysis together with CB-Dock analysis revealed
   that CIAPIN1 interacts directly with doxorubicin ([158]Figure 11).
   Doxorubicin is routinely used for the treatment of several cancers
   [[159]41] and a combined treatment of doxorubicin with salinomycin has
   been used to enhance doxorubicin sensitivity in CCA patients [[160]42].
   In addition, our data indicated that CIAPIN1 interacted indirectly with
   cisplatin via NOS1 ([161]Figure 11a). Related to this finding, Zou et
   al. suggested that NOS1 promoted chemoresistance against cisplatin
   (DDP) in ovarian cancer cells, providing a potential target to reduce
   chemoresistance to ovarian cancer therapeutics [[162]36]. Thus, we
   speculate that CIAPIN1 may play an important role in chemoresistance
   against drugs such as enhancing doxorubicin resistance, and CIAPIN1 may
   be a novel molecular target for CCA chemotherapy. Further in vitro
   studies are required to validate the role of CIAPIN1 in the
   chemoresistance of CCA cells.

5. Conclusions

   In conclusion, this study revealed that the high expression of CIAPIN1
   in CCA tissues is associated with lymph node metastasis and poor
   survival of CCA patients. CIAPIN1 expression in CCA tissues can serve
   as a prognostic biomarker for overall survival time and can be an
   indicator for metastasis of CCA patients. In addition, CIAPIN1 gene
   silencing caused suppression of CCA cell motilities, suggesting an
   important role of CIAPIN1 in CCA metastasis. In addition, using
   bioinformatic analysis, we predicted that the signaling pathway of
   CIAPIN1 involves the TGF-β/SMAD signaling pathway via NOS1 in CCA
   cells. Moreover, CIAPIN1 may directly bind to doxorubicin to cause drug
   resistance. CIAPIN1 might serve as a novel promising molecular target
   for CCA chemotherapy.

Acknowledgments