Aberrantly expressed PLOD1 promotes cancer aggressiveness in bladder cancer: a potential prognostic marker and therapeutic target

Bladder cancer (BC) is the ninth most malignant tumor worldwide. Some BC patients will develop muscle‐invasive BC (MIBC), which has a 5‐year survival rate of approximately 60% due to metastasis. As such, there is an urgent need for novel therapeutic and diagnostic targets for MIBC. Analysis of novel antitumor microRNA (miRNA)‐mediated cancer networks is an effective strategy for exploring therapeutic targets and prognostic markers in cancers. Our previous miRNA analysis revealed that miR‐140‐5p acts as an antitumor miRNA in BC cells. Here, we investigated miR‐140‐5p regulation of BC molecular pathogenesis. Procollagen‐lysine, 2‐oxoglutarate 5‐dioxygenase 1 (PLOD1) was found to be directly regulated by miR‐140‐5p, and aberrant expression of PLOD1 was observed in BC clinical specimens. High PLOD1 expression was significantly associated with a poor prognosis (disease‐free survival: P = 0.0204; overall survival: P = 0.000174). Multivariate analysis showed PLOD1 expression to be an independent prognostic factor in BC patients (hazard ratio = 1.51, P = 0.0099). Furthermore, downregulation of PLOD1 by siRNAs and a specific inhibitor significantly decreased BC cell aggressiveness. Aberrant expression of PLOD1 was closely associated with BC pathogenesis. In summary, the present study showed that PLOD1 may be a potential prognostic marker and therapeutic target for BC.


Introduction
Bladder cancer (BC) is the ninth most malignant tumor worldwide, and approximately 430 000 cases were newly diagnosed in 2012 (Antoni et al., 2017). BC is clinically divided into two groups: muscle-invasive BC (MIBC) and non-muscle-invasive BC (NMIBC) (Lemke and Shah, 2018). Patients with the latter have a favorable prognosis (5-year survival rate: approximately 90%) after surgical resection. However, approximately 50% of cases develop intravesical recurrence after surgical resection, and approximately 15-40% of recurrent BC cases are invasive and exhibit distant metastasis (Lemke and Shah, 2018). Although radical cystectomy and cisplatin-based combination chemotherapy are the standard treatments for MIBC, the 5-year survival rate of patients with MIBC is approximately 60% (Chou et al., 2016;Lemke and Shah, 2018). In addition, the survival of patients with distant metastasis is only 15 months due to no effective treatment options (Abufaraj et al., 2018). Therefore, discovery of novel therapeutic and diagnostic targets is urgently needed.
A vast number of studies have shown that a large number of noncoding RNAs encoded by the human genome are functional and play critical roles in various cellular processes, for example, cell growth, migration, invasion, and apoptosis (Bartel, 2004). microRNAs (miRNAs), a class of noncoding RNAs, are endogenous single-stranded RNA molecules comprising 19-22 nucleotides that function as fine-tuners of RNA expression (Bartel, 2009;Goto et al., 2015b;Koshizuka et al., 2017a;Kurozumi et al., 2017). A single miRNA regulates a vast number of RNA transcripts, and a bioinformatics study showed that approximately 60% of protein-coding genes are controlled by miR-NAs (Bartel, 2009). Aberrantly expressed miRNAs are closely associated with cancer pathogenesis via disruption of RNA networks within cancer cells (Beermann et al., 2016).
Using the knowledge that a single miRNA controls numerous genes, we sequentially identified novel cancer pathways regulated by antitumor miRNAs in several cancers (Goto et al., 2015aMiyamoto et al., 2016). Identification of dysregulated miRNAs in cancer cells is the first step, and the latest RNA-sequencing technology is suitable for producing miRNA signatures. Interestingly, analyses of our RNA-sequencing-based signatures revealed that the passenger strand of some miRNAs is up-or downregulated in cancer tissues Koshizuka et al., 2017b). During miRNA biogenesis, the passenger strand of the miRNA duplex is degraded and does not play a role in gene regulation in cells (Mah et al., 2010). Our recent studies revealed that the passenger strand of certain miRNAs (e.g., miR-99a-3p, miR-144-5p, miR-145-3p, miR-455-5p, and miR-223-5p) acts as an antitumor miRNA by targeting several oncogenic genes closely involved in cancer pathogenesis Goto et al., 2017;Matsushita et al., 2015;Sugawara et al., 2018;Yamada et al., 2018aYamada et al., ,2018b.
Based on the miRNA signature of BC, we focused on miR-140-5p (the passenger strand of the miR-140duplex) to investigate the function of miR-140-5p and identify its target oncogenes as therapeutic and diagnostic targets for BC. Our data showed that procollagen-lysine, 2-oxoglutarate 5-dioxygenase 1 (PLOD1) is directly regulated by miR-140-5p in BC cells. Aberrant expression of PLOD1 was closely associated with BC pathogenesis. Notably, inhibition of PLOD1 by transfection of siRNA or a PLOD1 inhibitor significantly attenuated the malignant phenotype of BC cells.

Clinical specimen collection and cell culture
We obtained 15 BC tissues and normal adjacent tissues from patients undergoing total cystectomy at Chiba University Hospital between 2014 and 2015 (Table S1). All patients provided informed written consent forms, and the study protocol was approved by the Institutional Review Board of Chiba University (number: 484). The study methodologies conformed to the standards set by the Declaration of Helsinki. We used the human BC cell lines T24 and BOY. These cell lines were cultured in RPMI 1640 Medium supplemented with 10% fetal bovine serum (HyClone, Logan, UT, USA) as described previously (Yamada et al., 2018d).

Cell proliferation, migration, and invasion assays
Cell proliferation was evaluated by the XTT assay using the Cell Proliferation Kit II (Sigma-Aldrich). Cell migration was assessed by wound healing assays, and invasion was determined using modified Boyden chambers containing Matrigel-coated Transwell membrane filter inserts.

Cell-cycle assay
Bladder cancer cells were transiently transfected with either the transfection reagent only as the control or the 2,2 0 -dipyridyl, PLOD1 inhibitor, in six-well tissue culture plates. Cells were harvested by trypsinization 72 h after transfection. For cell-cycle analysis, cells were stained with propidium iodide using the Cycletest Plus DNA Reagent Kit (BD Biosciences, Bedford, MA, USA) according to the manufacturer's instructions and examined using the CyAn ADP Analyzer (Beckman Coulter, Brea, CA, USA). The percentages of cells in the G0/G1, S, and G2/M phases were calculated and compared. Experiments were performed in triplicate (Matsushita et al., 2015).

Apoptosis assays
Apoptotic cells were detected using the FITC Annexin V Apoptosis Detection Kit (BD Biosciences) according to the manufacturer's instructions and the BD FACSCelesta Flow Cytometer (BD Biosciences). Cells were identified as viable, dead, or early or late apoptotic cells, and the percentages of apoptotic cells under each experimental condition were compared. Anti-poly (ADP-ribose) polymerase (PARP) (#9542; Cell Signaling Technology, Danvers, MA, USA) was evaluated as a marker of apoptosis in this study .

Identification of candidate target genes regulated by miR-140
To identify candidate target genes regulated by miR-140-5p and miR-140-3p, we used a combination of in silico and genome-wide gene expression analyses. Genes potentially regulated by miRNAs in a sequencedependent manner are listed in the TargetScan database (release 7.2) (http://www.targetscan.org/vert_70/). Genes upregulated in BC were identified from a publicly available dataset in the Gene Expression Omnibus (GEO; accession number: GSE31684), and we narrowed down the list of candidate genes. Gene expression was also analyzed by our own oligonucleotide microarray analyses (Human GE 60K; Agilent Technologies), the data of which were deposited into the GEO (on June 14, 2018; http://www.ncbi.nlm.nih. gov/geo/) under accession number GSE115800.

Dual-luciferase reporter assay
The wild-type sequence of the PLOD1 3 0 -untranslated region (UTR) was inserted between the SgfI and PmeI restriction sites of the 3 0 -UTR of the hRluc gene within the psiCHECK-2 vector (C8021; Promega, Madison, WI, USA). We also generated PLOD1 3 0 -UTR sequences containing deletions in the miR-140-5p target sites (positions 43-49 and 725-731) for insertion into the psiCHECK-2 vector as described above. The psiCHECK-2 vector was used as a cloning vector for the synthesized DNA sequences.

Immunohistochemistry
Immunohistochemistry procedures were performed according to a previously described method. Clinical tissue sections were incubated overnight at 4°C with an anti-PLOD1 antibody diluted 1:10 (SAB1301577; Sigma-Aldrich).

Analysis of genes downstream of PLOD1
To investigate PLOD1-regulated pathways in BC cells, we assessed gene expression changes in T24 and BOY cells transfected with the PLOD1 inhibitor. Microarray analysis was performed to obtain expression profiles in these cells, and the microarray data were deposited into the GEO (on December 4, 2018; accession number: GSE123318).

Analysis of the clinical significance of PLOD1 expression
We investigated the clinical importance of miRNAs and genes in BC patients using RNA-sequencing data available in The Cancer Genome Atlas (TCGA; https://tcga-data.nci.nih.gov/tcga/). The gene expression and clinical data were obtained from cBioPortal (http://www.bioportal.org/), and provisional data were downloaded on October 5, 2018 (Anaya, 2016;Cerami et al., 2012;Gao et al., 2013).

Statistical analysis
Statistical comparisons involving two or three variables were performed using the Bonferroni-adjusted Mann-Whitney U-test. Spearman's rank tests were used to analyze the correlations among gene expression levels. These analyses were conducted using EX-PERT STATVIEW software (version 5.0, SAS Institute Inc., Cary, NC, USA). Multivariate analysis of prognostic factors for patient survival was conducted using JMP PRO 13 (SAS Institute Inc.).

Effect of miR-140-5p and miR-140-3p on apoptosis and cell-cycle assays in BOY cells
The percentage of apoptotic cells was significantly increased in miR-140-5p-and miR-140-3p-transfected cells compared with the control cells (Fig. S1A,B). Moreover, transfection of miR-140-5p and miR-140-3p upregulated the level of cleaved PARP (Fig. S1C). In a cell-cycle analysis, the proportion of cells in the G0/ G1 phase was significantly higher transfected with miR-140-5p compared with the control cells ( Fig. S1D).

Clinical significance and expression of PLOD1
Clinical data from BC patients were obtained from TCGA database, and information on survival revealed that patients with high PLOD1 expression had a significantly poorer prognosis compared with patients with low expression (disease-free survival: P = 0.0204; overall survival: P = 0.000174; Fig. 2B). High PLOD1 expression was also related to a highly malignant tumor morphology, advanced stage, and metastasis (Fig. S3A). According to multivariate Cox proportional hazards regression, high expression of PLOD1 was an independent predictive factor for overall survival in BC patients (hazard ratio: 1.51; 95% confidence interval: 1.1-2.07, P = 0.0099) (Fig. 2B). PLOD1 mRNA expression levels were significantly upregulated in BC tissues compared with normal adjacent tissues (P = 0.0464) (Fig. 2C). Immunostaining of PLOD1 in BC clinical specimens indicated high expression of PLOD1 in cancer lesions compared with adjacent noncancerous tissues at the same staining intensity (Fig. 2C).
In addition, expression levels of PLOD2 and PLOD3 were detected in BC clinical specimens (Fig. S4A,B). Also, immunohistochemical staining showed that overexpressed PLOD2 and PLOD3 were detected in cancer lesions (Fig. S4G,H). Interestingly, Fig. 1. miR-140 expression and antitumor functions in BC. (A-C) Expression levels of miR-140-5p and miR-140-3p in BC clinical specimens (P = 0.0013 and P = 0.0004, respectively). RNU48 was used as an internal control. P-values were calculated using Bonferroni-adjusted Mann-Whitney U-test. A positive correlation between miR-140-5p and miR-140-3p expression levels was detected by Spearman's rank test (R = 0.637, P = 0.0006). (D-F) Cell proliferation, migration, and invasion activities. Error bars are represented as mean AE SD (n = 5, n = 8, and n = 8, respectively). P-values were calculated using Bonferroni-adjusted Mann-Whitney U-test. *P < 0.001, **P < 0.0001. high expression of PLOD2 was significantly associated with poor prognosis of the patients with BC (Fig. S3B). Among PLOD family, expression of PLOD1 was the highest in BC tissues (Fig. S4C). Clinicopathological analysis was performed between PLODs expression and BC (NMIBC or MIBC) clinical specimens. However, no significant association was found in this study ( Fig. S4D-F).

PLOD1 was directly regulated by miR-140-5p
PLOD1 mRNA and protein levels were significantly decreased in T24 and BOY cells following transfection with miR-140-5p compared with mock-transfected cells or those transfected with miR-control (Fig. 2D,E). The TargetScan database indicated the presence of two miR-140-5p binding sites (positions 43-49 and 725-731) within the PLOD1 3 0 -UTR. We performed luciferase reporter assays using a vector containing these sequences to assess whether miR-140-5p directly regulates PLOD1 expression in a sequence-dependent manner. Cotransfection of miR-140-5p with vectors harboring the PLOD1 3 0 -UTR deletion constructs significantly decreased luciferase activity compared with the activity levels in mock-transfected and miR-control-transfected cells (Fig. 2F).

Knockdown and rescue studies of PLOD1
We confirmed that both PLOD1 mRNA and protein expression levels were suppressed by siRNA-mediated PLOD1 knockdown in BC cells (Fig. 3A and B). Transfection of si-PLOD1 suppressed cell proliferation, migration, and invasion activities (Fig 3C-E). The percentage of apoptotic cells was significantly increased in si-PLOD1-transfected cells compared with the control cells (Fig. S5A,B). Moreover, transfection of si-PLOD1 upregulated the level of cleaved PARP (Fig. S5C). In a cell-cycle analysis, the proportion of cells in the G0/G1 phase was significantly (F) Dual-luciferase reporter assays using vectors encoding the wild-type PLOD1 3 0 -UTR sequence containing two putative miR-140-5p target sites and 3 0 -UTR sequences with deletions of the target sites (Deletion). Normalized data were calculated as the ratio of Renilla/Firefly luciferase activities. Error bars are represented as mean AE SD (n = 3). P-values were calculated using Bonferroni-adjusted Mann-Whitney U-test. *P < 0.0001, **P < 0.005. higher, transfected with si-PLOD1_2 compared with the control cells, although G2/M phase was significantly elevated in si-PLOD1_1 transfection (Fig. S5D).
In addition, we performed a PLOD1 rescue study in T24 cells to validate whether oncogenic pathways regulated by PLOD1/miR-140-5p are crucial for BC development. PLOD1 and miR-140-5p transfection restored PLOD1 protein expression (Fig. 3F). Functional assays demonstrated that BC cell migration and invasion were significantly recovered by PLOD1 and miR-140-5p transfection compared with miR-140-5p alone (Fig. 3G-I).

Functional analysis of a PLOD1 inhibitor
After transfection of the PLOD1 inhibitor 2,2 0 -dipyridyl into BC cells, cell proliferation was suppressed in a dose-dependent manner (Fig. 4A). The IC 50 of 2,2 0 -dipyridyl was 82.8 lM in BOY cells and 37.1 lM in T24 cells. Cell migration and invasion were also decreased in a dose-dependent manner in cells transfected with the inhibitor (Fig. S6). In addition, the percentage of apoptotic cells was increased in PLOD1 inhibitor-transfected cells compared with the control cells (Fig. 4B). Moreover, transfection of PLOD1 inhibitor upregulated the level of cleaved PARP (Fig. 4B). In a cell-cycle analysis, the proportion of cells in the G0/G1 phase was significantly higher in BC cells transfected with the PLOD1 inhibitor compared with the control cells (Fig. 4C). In addition, we confirmed that the inhibitor suppressed the mRNA and protein levels of PLOD1 in a dosedependent manner (Fig. S7). Apoptosis and cell-cycle experiments gave similar results in BOY cells (Fig. S8).

Genes affected by the PLOD1 inhibitor
PLOD1 acts as lysyl hydroxylases that catalyze hydroxylation of collagen lysines, and it works under the following conditions, extracellular matrix maturation and remodeling. In order to explore the functional significance of PLOD1 on tumor progression, we examined the PLOD1-mediated downstream genes and pathways. As shown in the Venn diagram in Fig. S9, 1518 genes were considerably downregulated after transfection of the PLOD1 inhibitor in BOY and T24 cells. In a KEGG analysis of these genes, we identified 39 pathways enriched among the PLOD1-affected genes, including pathways related to cell cycle and apoptosis (Table 2).

Discussion
RNA sequencing is a suitable technology for creating miRNA expression signatures in cancer cells. Analyses of our miRNA signatures in cancers revealed that the passenger strand of some miRNA duplexes is functional in cancer cells by targeting cancer-related genes Goto et al., 2017;Matsushita et al., 2015;Sugawara et al., 2018;Yamada et al., 2018aYamada et al., ,2018b. This makes it possible to identify novel cancer pathways based on aberrantly expressed passenger strand miRNAs.
In this study, we focused on both strands of pre-miR-140 (miR-140-5p and miR-140-3p) and revealed their antitumor functions in BC cells. Previous reports showed that miR-140-3p is downregulated in squamous cell lung cancer and functions as a tumor suppressor by targeting bromodomain containing 9 in vitro and in vivo (Huang et al., 2019). As with miR-140-3p, a tumor-suppressive function of miR-140-5p has been reported in several cancers. miR-140-5p exerted a tumor-suppressive function and enhanced the effect of existing therapeutic drugs in non-small-cell lung cancer  (Flamini et al., 2017). Another report showed that miR-140-5p suppressed cell aggressiveness and suggested that miR-140-5p is a prognostic marker in gastric cancer (Fang et al., 2017). Downregulation of miRNAs was reported to be caused by epigenetic factors such as DNA methylation or histone deacetylation. Previous study showed that suppression of miR-140 expression was influenced by the hypermethylation of the promoter region in breast cancer (Wolfson et al., 2014). Elucidation of the detailed molecular mechanism of downregulation of miR-140-5p and miR-140-3p is also essential in BC cells. These studies indicate that both strands of pre-miR-140 act as critical miRNAs that prevent malignant transformation in cells. To our knowledge, this is the first study to identify a functional role of the miR-140 duplex and its oncogene targets in BC.
Our next focus was to investigate the molecular networks regulated by these miRNAs in BC cells. A total of 31 genes regulated by miR-140-5p and 33 genes regulated by miR-140-3p were identified as putative oncogenic targets in BC cells. Among these targets, the expression levels of eight genes (CERCAM, PLOD1, FADS1, PAFAH1B2, PAX6, ADAM17, CCDC103, and PLXNA4) were closely associated with BC pathogenesis. These genes are promising as therapeutic targets and prognostic markers, and further analysis is necessary to elucidate the molecular pathogenesis of BC. We focused on PLOD1 to investigate its oncogenic functions and clinical significance in BC. PLOD genes encode lysyl hydroxylases, which are crucial for collagen biosynthesis, cross-linking, and deposition (Qi and Xu, 2018). Collagen is a major component of the extracellular matrix (ECM), and collagen cross-linking is related to the stiffness of the ECM, which enhances cancer cell migration, invasion, and focal adhesion (Du et al., 2017;Peinado et al., 2008). The PLOD family consists of PLOD1, PLOD2, and PLOD3. A number of studies have demonstrated that overexpression of PLOD2 and PLOD3 promotes cancer progression and metastasis. Our previous studies showed that aberrant expression of PLOD2 was detected in BC and renal cell carcinoma tissues, and its overexpression enhanced cancer cell malignant transformation (Kurozumi et al., 2016;Miyamoto et al., 2016). We hypothesized that members of the PLOD family member are deeply involved in the molecular pathogenesis of BC. On the other hand, there are not many reports on the role of PLOD1 in cancer (Qi and Xu, 2018). Previous studies showed that aberrant expression of PLOD1 was significantly associated with shorter survival in patients with gastric or colorectal cancer (Wang et al., 2018). Overexpression of PLOD1 was also detected in esophageal squamous cell carcinoma and breast cancer (Gilkes et al., 2013;Li et al., 2017). Mutations in PLOD1 are the cause of PLOD1-related kyphoscoliotic Ehlers-Danlos syndrome, an autosomal recessive generalized connective tissue disorder (Giunta et al., 2005). The data from a large number of cohort analyses in TCGA database show that high expression of PLOD1 is significantly associated with a poor prognosis (overall survival: P = 0.000174), more strongly than are PLOD2 and PLOD3 (OS: P = 0.0097 and P = 0.315, respectively) (Fig. S3B,C). Furthermore, multivariate analysis showed that PLOD1 expression was an independent prognostic factor in patients with BC (hazard ratio = 1.51, P = 0.0099). Moreover, high expression of PLOD1 was significantly associated with tumor stage and presence of metastasis. Aberrant expression of PLOD1 has been shown to be closely related to the malignant phenotype of BC. Development of a new diagnostic strategy for BC using PLOD1 expression as a marker is desired.
Aberrant expression of PLOD1 was detected in BC clinical specimens, and inhibition of PLOD1 by siRNA-mediated knockdown or treatment with a PLOD1 inhibitor significantly reduced the malignant phenotype of BC cells (e.g., decreases in proliferation, migration, and invasion and an increase in apoptosis). We used 2,2 0 -dipyridyl, an iron chelator, as an inhibitor of PLOD1 in this study (Bernardes et al., 2018;Jover et al., 2018). Collagen lysyl hydroxylases reportedly depend on Fe2+ binding for stabilization, and 2,2 0 -dipyridyl prevents prolyl and lysyl hydroxylation (Barsh and Byers, 1981;Guo et al., 2018). A previous report showed that inhibition of PLOD1 and lysyl oxidase suppressed arterial smooth muscle cell calcification via ECM remodeling (Jover et al., 2018). Another study was conducted to investigate the effect of 2,2 0dipyridyl in combination with doxorubicin in breast cancer cells (Bernardes et al., 2018). In this study, we showed that PLOD1 expression and cell proliferation were suppressed after transfection of a PLOD1 inhibitor in a dose-dependent manner. Moreover, the PLOD1 inhibitor induced apoptosis and cell-cycle arrest at the G1-to-S phase transition.
The molecular mechanism of the antitumor effect of the PLOD1 inhibitor in BC cells was evaluated by global gene expression analysis. As a result, genes associated with cell cycle, ECM-receptor interactions, and apoptosis were differentially expressed in cells transfected with the PLOD1 inhibitor, supporting our current data. We focused on several genes (e.g., CCNB1, CCNB2, and SKP2) involved in 'cell-cycle pathway'. Expression of CCNB2 (cyclin B2) was upregulated in BC tissues, and suppression of its expression significantly inhibited invasive and metastatic abilities (Lei et al., 2016). Our recent study showed that CCNB1 (cyclin B1) was regulated by antitumor miR-223-5p in BC cells and its high expression was closely associated with poor prognosis of the patients with BC by TCGA database analysis (Sugawara et al., 2018). Moreover, overexpression of SKP2 (S-phase kinase-associated protein 2) was significantly related to advanced tumor stage and grade of the patients with BC (Kawakami et al., 2007).
Moreover, we performed rescue experiments by overexpressing PLOD1 and miR-140-5p. The results revealed that PLOD1 can counteract the antitumor effects, in terms of cell migration and invasion, of miR-140-5p in BC cells, indicating that the PLOD1/ miR-140-5p axis plays an important role in BC development.

Conclusion
Both strands of the miR-140 duplex (miR-140-5p and miR-140-3p) suppressed BC cell malignant transformation. Genes controlled by the miR-140-5p were found to be related to BC pathogenesis. PLOD1 expression was directly regulated by the miR-140-5p in BC cells. Aberrant expression of PLOD1 was closely contributed to BC development. Furthermore, inhibition of PLOD1 expression significantly attenuated to BC cell aggressive phenotypes. PLOD1 might be a novel biomarker and therapeutic target in BC. Further investigation is required for clinical application. Fig. S1. Effect of miR-140-5p and miR-140-3p on apoptosis and cell-cycle assays in BOY cells. Fig. S2. miR-140-5p and miR-140-3p localization within the RISC. Fig. S3. Clinical database analysis of PLOD1, PLOD2 and PLOD3 expression in BC patients. Fig. S4. Expression analysis of PLOD1, PLOD2 and PLOD3 in BC tissues. Fig. S5. Effect of si-PLOD1 on apoptosis and cell-cycle assays in BOY cells. Fig. S6. Effect of a PLOD1 inhibitor on the migration and invasion of BC cells. Fig. S7. Effect of a PLOD1 inhibitor on PLOD1 expression. Fig. S8. Effect of a PLOD1 inhibitor on apoptosis and cell-cycle assays in BC cells. Fig. S9. Downstream pathways affected by treatment with a PLOD1 inhibitor in BC cells. Table S1. Background characteristics of the BC patients.