Paternally expressed gene 10 (PEG10) has been associated with neuroendocrine muscle-invasive bladder cancer (MIBC), a subtype of the disease with the poorest survival. In this work, we further characterized the expression pattern of PEG10 in The Cancer Genome Atlas database of 412 patients with MIBC, and found that, compared with other subtypes, PEG10 mRNA level was enhanced in neuroendocrine-like MIBC and highly correlated with other neuroendocrine markers. PEG10 protein level also associated with neuroendocrine markers in a tissue microarray of 82 cases. In bladder cancer cell lines, PEG10 expression was induced in drug-resistant compared with parental cells, and knocking down of PEG10 resensitized cells to chemotherapy. Loss of PEG10 increased protein levels of cell-cycle regulators p21 and p27 and delayed G1–S-phase transition, while overexpression of PEG10 enhanced cancer cell proliferation. PEG10 silencing also lowered levels of SLUG and SNAIL, leading to reduced invasion and migration. In an orthotopic bladder cancer model, systemic treatment with PEG10 antisense oligonucleotide delayed progression of T24 xenografts. In summary, elevated expression of PEG10 in MIBC may contribute to the disease progression by promoting survival, proliferation, and metastasis. Targeting PEG10 is a novel potential therapeutic approach for a subset of bladder cancers.
Muscle-invasive bladder cancer (MIBC) is highly aggressive with poor survival rates (1, 2). Thorough understanding of disease progression is needed to guide treatments for this common, highly lethal malignancy. Recent molecular characterization from The Cancer Genome Atlas (TCGA) identified driver genes and pathways of MIBC (3). Among the many genomic, genetic, and epigenetic modifications, factors targeting RB1 and TP53 tumor suppressor genes are the most prevalent (4). Forty percent to 50% of MIBCs have inactivating mutations or reduced expression of RB1, which is strongly associated with poor clinical outcomes (5, 6). Loss-of-function mutations of TP53 are present in up to 60% of MIBCs (7) and are also associated with disease poor outcomes (8, 9). In addition, abnormalities of RB1 and TP53 genes coexist among 40%–50% of MIBCs (10).
Recently, the neuroendocrine-like subtype of MIBC has been recognized as a subgroup with the poorest survival in patients with MIBC (3, 11, 12). These tumors express relatively high levels of neuroendocrine markers, as well as neuronal differentiation and development genes (3, 11). Interestingly, in TCGA, unbiased nonnegative matrix factorization consensus clustering of RNA-sequencing (RNA-seq) data from 408 tumors revealed paternally expressed gene 10 (PEG10) as one of the genes associated with poor prognosis of neuroendocrine-like bladder cancers (3). This is in-line with previous studies from a patient-derived xenograft model that undergoes transdifferentiation from conventional prostatic adenocarcinoma to neuroendocrine prostate cancer (NEPC), where PEG10 was identified as a NEPC-specific therapeutic candidate (13). PEG10 promotes growth and invasion of prostate cancer cells, and its expression and function are tightly regulated by RB1 and TP53 whose genetic aberrations are hallmarks of NEPC.
PEG10 is a retrotransposon-derived placental gene, which structurally resembles human immunodeficiency virus (14). Although PEG10 no longer retains the reverse transcriptase activity, it is distinct among other mammalian genes by carrying an active −1 ribosomal frameshift element, allowing translation of two isoforms (RF1 and RF1/2) from overlapping reading frames from the same transcript (15). PEG10 also possesses two translation initiation sites, ‘a’ and ‘b,’ where ‘b’ is a non-ATG codon; besides this, PEG10 carries a domain for protease activity to generate a distinct self-cleavage product (termed cleaved n-terminal fragment; CNF). All these findings suggest a highly complex biology of PEG10 (16). Expression of PEG10 is low in adult tissues, but it is essential for placental development; heterozygous knockout PEG10+/− mice demonstrated lethality by embryonic day 10.5 (17). In addition to PEG10 proteins, a long noncoding RNA PEG10 has also been linked to several types of tumors (18, 19).
Because mutations of TP53 and inactivation of RB1 are common in MIBC, and because neuroendocrine gene expression appears to drive a particularly poor prognosis in a subset of patients, we hypothesized that PEG10 may contribute to disease progression and adverse prognosis. To test this hypothesis, we characterized PEG10 function in MIBC progression and investigated PEG10 as a novel therapeutic target in MIBC.
Materials and Methods
Bladder carcinoma cell lines were received from the Pathology Core of the Bladder Cancer SPORE at MD Anderson Cancer Center (Houston, TX). Cells were authenticated by DNA fingerprinting using AmpFISTR Amplification or AmpFISTR Profiler PCR Amplification Protocols (Life Technologies). Cells were maintained in minimum essential medium supplemented with 10% FBS, penicillin, streptomycin, vitamins, l-glutamine, nonessential amino acids, and pyruvate supplements, and routinely tested for Mycoplasma.
Analysis of TCGA dataset
DNA sequencing results for RB1 and TP53, including calls for deep deletions, truncation, and missense mutations, as well as normalized RNA-seq–derived gene expression data for PEG10 and neuroendocrine markers were downloaded from cbioportal.org (3). PEG10 expression as analyzed between different molecular subtypes (mRNA clusters; ref. 3) and as correlation with other neuroendocrine markers was calculated dependent on molecular subtype as well as RB1 and TP53 gene status.
Tissue microarray and IHC
The Vancouver Prostate Centre (VPC) bladder cancer tissue microarray (TMA) consists of 21 cases of MIBC and 61 cases of non-MIBC specimens obtained by transurethral resection. Written consents were obtained from the patients and the study was performed following the ethical guidelines of University of British Columbia (UBC, Vancouver, British Columbia, Canada) Clinical Research Ethics Board (CREB, #H09-01628) and reviewed by the chair of the UBC CREB. TMA preparation and IHC were performed as described previously (20, 21) and as outlined in the Supplementary Data. PEG10 IHC staining in the TMA was scored as: 0, no staining; 1, faint or focal staining; 2, convincing intensity in a minority of cells; and 3, convincing intensity in a majority of cells.
Western blotting and qPCR
Western blotting and qPCR were conducted as described previously (22, 23) and as outlined in the Supplementary Data. Primary antibodies and probes used in this study are listed in the Supplementary Tables S1 and S2, respectively.
Transfections of siRNAs (listed in the Supplementary Table S3) were carried out using RNAiMAX (Life Technologies) following the manufacturer's instruction. A PEG10 antisense oligonucleotide (ASO, 5′-GGCAGTGGTAGCGGCAGTAT-3′) and a scrambled oligonucleotide (SCRB, 5′-CCTTCCCTGAAGGTTCCTCC-3′) were purchased from IONIS Pharmaceuticals and transfected using oligofectamin following the protocols described previously (24). PEG10 plasmids were generated as described previously (13). All plasmids were transfected using Lipofectamine (Life Technologies) following the manufacturer's instruction.
Proliferation and cell-cycle analysis
Cell growth was evaluated with a Cell Counting Kit 8 (Dojindo) following the manufacturer's instruction. The cell-cycle distribution was examined by double stainings with bromodeoxyuridine (BrdU) and 7-aminoactinomycin D (7-AAD) using a BrdU-FITC Flow Kit (BD Biosciences) following the manufacturer's instruction. The double thymidine block was performed as described previously (13).
Cell migration and invasion assays
For migration assays, scratches were made using a sterilized aerosol pipet tip, and cells were maintained in serum-free medium containing TGFβ (0.1 ng/mL) for 24 hours. Mitomycin C (0.3 μg/mL) was added after scratching to suppress cell growth. Bright field images were taken at the same area right after the scratch and also at 24 hours after the scratching. Cell invasion was investigated using BioCoat Matrigel Invasion Chambers (BD Biosciences). Briefly, 5 × 104 cells were plated in the top chamber with serum-free condition, and medium containing 20% FBS was set in the bottom chamber. At 18 hours after seeding, the polycarbohydrate membranes from the bottom of the top chambers were dissected, fixed with methanol, and then stained with crystal violate to visualize the invaded cells. The number of invading cells was quantified in four microscopic fields and averaged.
Orthotopic bladder cancer xenograft model
The animal work was approved by the Institutional Review Board of UBC (Vancouver, British Columbia, Canada; A14-0291). Procedures were performed as described previously (25). Six-week-old female nude mice (Harlan Laboratories) were anesthetized using 2% isoflurane. Analgesia was achieved by a subcutaneous injection of meloxicam (Boehringer Ingelheim Vetmedica). Cells (4.0 × 105) in 50 μL Matrigel suspension were inoculated into the bladder wall of nude mice using a 30-G needle by percutaneous injection under ultrasound guidance. Beginning on 4th day after tumor inoculation, 15 mg/kg of PEG10- or Scr-ASO was administered systemically via intraperitoneal injection once per day for 5 days and then three times per week thereafter. For in vivo photoimaging [In Vivo Imaging Spectrum (IVIS) Lumina, PerkinElmer], cells transfected with a lentiviral construct for a firefly luciferase gene under Blasticidin Selection (Life Technologies) were employed. The direct association between the cell number, luciferase activity, tumor size, and bioluminescence was screened and controlled (R > 0.99) as described previously (25) using the Xenogen IVIS spectrum imager. Bioluminescence using the Xenogen IVIS Spectrum Imaging System (PerkinElmer) was used to assess tumor burden. Images were recorded at 10 and 15 minutes after luciferin injection. Averaged counts were used for statistical analysis. The terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay is outlined in the Supplementary Data.
Statistical analyses were carried out using the χ2 test, the unpaired t test, the ANOVA, and Wilcoxon test using JMP9 Software (SAS Institute). Statistical significance was determined as P < 0.05.
PEG10 associates with neuroendocrine markers in bladder cancer
The expression pattern of PEG10 was investigated in MIBC cases from TCGA database, which consists of 408 samples with RNA-seq data from chemotherapy-naïve, muscle-invasive, high-grade urothelial tumors (3). PEG10 mRNA was highly expressed in the neuronal subtype (neuroendocrine-like) MIBC compared with the basal, luminal, luminal-infiltrated, and luminal-papillary subtypes (Fig. 1A). Furthermore, PEG10 mRNA levels correlated with neuroendocrine markers in an all cases analyzed (Fig. 1B). This correlation was seen especially in cases with loss of RB1/TP53 (Supplementary Fig. S1A), but became less significant in cases with wild-type (WT) RB1/TP53 (Supplementary Fig. S1A). The observation that PEG10 associated with neuroendocrine markers in bladder cancer is consistent with other reports from both prostate (refs. 13, 26; Supplementary Fig. S1B) and small-cell lung cancer (27, 28), suggesting a broad positive correlation between PEG10 and neuroendocrine phenotypes. This finding was further confirmed in the bladder cancer TMA from the VPC, which demonstrated positive correlation between PEG10 protein and neuroendocrine markers (Fig. 1C).
Because TP53 and RB1 inactivation are common in MIBC (3, 4, 10), and PEG10 is a known target gene of RB/E2F and p53 (13, 29, 30), we next explored whether PEG10 mRNA levels correlated with the status of RB1/TP53 in TCGA database. As shown in Fig. 1D, PEG10 expression was enhanced in cases with RB1/TP53 loss (left), and this increase was further magnified in the 20 neuroendocrine-like cases (right). Interestingly, although PEG10 mRNA was not significantly correlated with overall survival in an analysis of all cases or when substratified by RB1/TP53 loss (Supplementary Fig. S1C), there was a statistically insignificant suggestion of worse survival with increased PEG10 mRNA levels in the 20 neuroendocrine-like cases (Fig. 1E).
PEG10 levels are elevated in chemotherapy-resistant cells
To define roles for PEG10 in bladder cancer progression, we first examined mRNA and protein levels of PEG10 in a panel of bladder cancer cell lines (31) by qPCR (Fig. 2A) and Western blotting (Fig. 2B). UM-UC14 and T24 cells showed high levels of PEG10 RF1/2 and RF1 isoforms (Fig. 2B). The levels of CNF1/2 isoforms were low in all bladder cancer cell lines (Fig. 2B). Because loss of RB1 and TP53 plays an important role during neuroendocrine transdifferentiation in prostate cancer (13), we therefore, compared RB and TP53 status with PEG10 levels in these bladder cancer cell lines. Five of seven cell lines carried TP53 mutations (mt) and two of the cell lines displayed undetectable phospho-Rb protein (pRb, Fig. 2B). Interestingly, UM-UC14 cells, which were featured with both mt TP53 and loss of Rb, showed highest mRNA and protein levels of PEG10 (Fig. 2A and B). Next, we explored whether manipulation of TP53 status would modulate the PEG10 level in bladder cancer cells. Introduction of WT TP53 into T24 cells (with mt TP53 and pRb protein) significantly attenuated both mRNA and protein levels of PEG10 (Fig. 2C and D). In contrast, introduction of WT TP53 in UM-UC14 cells (with mt TP53 and no pRb protein) did not significantly affect the PEG10 level (Supplementary Fig. S2A).
Because elevated PEG10 levels associate with neuroendocrine-like tumors, which display the poorest prognosis in MIBC, we hypothesized that PEG10 may regulate sensitivity of bladder cancer cells to chemotherapy. We, therefore, investigated whether PEG10 contributes to acquired chemotherapy resistance by establishing a stable cisplatin-resistant T24 cell line (T24R) that maintains constant growth rate in the presence of 10 μmol/L of cisplatin for at least 3 months (32). Both PEG10 protein and mRNA levels were highly elevated in resistant compared with parental cells (Fig. 2E). In addition, PEG10 was significantly induced by cisplatin over 5 days (Fig. 2F and G), indicating that induction of PEG10 might be an early event contributing to development of resistance. A similar phenomenon was observed in UM-UC14 cells (Supplementary Fig. S2B). PEG10 silencing triggered cell death in T24R cells to a greater extent than that in the parental cells, as measured by PARP cleavage (Fig. 2H); loss of PEG10 also retarded growth of the resistant T24R cells (Supplementary Fig. S2C), suggesting that PEG10 may promote bladder cancer cell growth and survival after chemotherapy.
Besides cisplatin, taxanes are also frequently used as second-line therapy of advanced MIBC. We developed a stable, paclitaxel-resistant T24 cell line (paclitaxel-re) by culturing cells with increasing concentrations of paclitaxel (starting from 0.125 μmol/L) with 2-fold increase at each cycle and eventually maintained cells at 2 μmol/L paclitaxel. The resistant cell line also exhibited increased PEG10 mRNA and protein levels compared with parental cells (Supplementary Fig. S2D), and PEG10 silencing triggered cell death in the resistant cells to a greater extent than in the parental cell line (Supplementary Fig. S2E), again indicating that chemotherapy-resistant cancer cell lines become more dependent on PEG10 for survival. Interestingly, analysis from a neoadjuvant chemotherapy (NAC) cohort (12, 33) revealed that, while mRNA levels of PEG10 were not predictive of response to NAC in the all case analysis (Supplementary Fig. S2F, left), there was an enrichment of PEG10 in the post-NAC luminal-like tumors, which had higher PEG10 levels compared with matched pre-NAC tumors (Supplementary Fig. S2F, middle and right).
PEG10 triggers Rb phosphorylation and promotes proliferation of bladder cancer cells
To further define roles for PEG10 in bladder cancer progression, we examined the effects of PEG10 on cell proliferation in T24 cells, which carry mt TP53 and pRb protein. PEG10 silencing attenuated T24 cell growth (Fig. 3A), enhancing levels of the cell-cycle regulators p21 and p27 and reducing cyclin D1 and pRb protein (Fig. 3B). Effects of PEG10 on cell-cycle progression were further evaluated in T24 cells using a BrdU incorporation assay in combination with 7-AAD staining in the cells released from thymidine block. As shown in the cell-cycle population bar graph (Fig. 3C) and the FACS profile (Supplementary Fig. S3A), siPEG10 delayed reentry of cells from G1- to S-phase. Analysis with Western blot indicated that Rb was hypophosphorylated after PEG10 silencing (Fig. 3D), likely because of increased p21 and p27 protein levels (Fig. 3D). Congruent with these observations, overexpression of PEG10 isoforms RF1 or RF1/2 in T24 cells promoted Rb phosphorylation and triggered cell growth (Fig. 3E and F).
PEG10 promotes proliferation of bladder cancer cells when total Rb protein is absent
The above findings indicate that PEG10 facilitates Rb phosphorylation to promote cell-cycle progression. We next investigated whether PEG10 modulates proliferation in bladder cancer cells lacking total Rb (tRb) protein (Fig. 2B). UM-UC14 cells bear mt TP53 and have no Rb protein, so that E2F transactivation is left unsupervised by Rb pathway. PEG10 silencing in UM-UC14 cells still repressed cell growth and induced p21 and p27 proteins (Fig. 4A and B). In the BrdU and 7-AAD combination assay, inhibition of PEG10 delayed entry of cells into S-phase after being released from the thymidine blocking (Fig. 4C; Supplementary Fig. S4A), elevating levels of p21 and p27 (Fig. 4D). In contrast, overexpression of PEG10 promoted UM-UC14 cell growth (Fig. 4E and F). In another cell line, UM-UC6, which carries WT TP53 and expresses tRb, but not pRb, protein so that E2F might be repressed by the Rb pathway, PEG10 overexpression still enhanced cell growth (Supplementary Fig. S4B and S4C). These data indicate that PEG10 promotes proliferation through molecular mechanisms involving Rb-dependent and Rb-independent pathways.
PEG10 promotes migration and invasion of bladder cancer cells
Considering the critical role of PEG10 in placental development where invasion of the maternal tissue presents a fundamental step (17), we next investigated roles of PEG10 on migration and invasion of bladder cancer cells. First, T24 and UM-UC14 cells transfected with siPEG10s or siCtrl were assessed in a migration assay using the wound-healing method. As shown in Fig. 5A, PEG10 silencing restrained cell migration, in contrast, overexpression of PEG10 accelerated cell migration (Supplementary Fig. S5A). Next, T24 and UM-UC14 cells transfected with PEG10 siRNAs or siCtrl were applied to an invasion assay recruiting the BioCoat Matrigel invasion chambers. Cells were stained with crystal violet (Fig. 5B) and numbers of invaded cells were quantified under microscope (Fig. 5C). Enumeration results indicate that PEG10 silencing significantly reduced invasion of bladder cancer cells.
To define molecular mechanisms of PEG10 modulation of migration and invasion in bladder cancer cells, protein levels of key regulators of cell mobility were investigated. Interestingly, TGFβ treatment, which is known to stimulate the epithelial-to-mesenchymal transition (EMT), induced PEG10 protein level and also the levels of SLUG and SNAIL (Fig. 5D), two key regulators of EMT (34). PEG10 silencing reduced TGFβ-stimulated SLUG and SNAIL protein levels (Fig. 5D), whereas PEG10 overexpression enhanced levels of these proteins (Supplementary Fig. S5B), indicating that PEG10 mediates TGFβ-induced EMT via SLUG- and SNAIL-dependent pathways.
PEG10 knockdown attenuates in vivo tumor growth
The above studies suggest that PEG10 may trigger bladder cancer progression by promoting survival (Fig. 2), proliferation (Figs. 3 and 4), and invasion (Fig. 5). We next evaluated whether targeting PEG10 with ASO can retard tumor progression. First, we assessed effects of PEG10-ASO in T24 cells in vitro. PEG10-ASO reduced both mRNA and protein levels of PEG10 (Fig. 6A; Supplementary Fig. S6A) in a dose- and sequence-dependent manner, and suppressed growth of T24 cells (Fig. 6B).
Next, an orthotopic bladder cancer xenograft model using T24 cells was established to evaluate effects of PEG10 inhibition in vivo. Tumor-bearing mice were treated with PEG10-ASO or Scrambled-ASO (Scr-ASO), and tumor burden was monitored with bioluminescence using the IVIS and tumor volume using ultrasonography. PEG10-ASO significantly delayed tumor growth compared with scrambled controls (Fig. 6C,–E). Furthermore, enhanced TUNEL staining signal was observed in the PEG10-ASO–treated tumor samples, suggesting increased apoptotic rates post-PEG10-ASO treatment (Fig. 6F). IHC staining and Western blot analysis on excised tumor tissues confirmed the reduction of PEG10 levels in the ASO-treated tumors (Supplementary Fig. S6B and S6C).
This study demonstrates that PEG10 is associated with invasive bladder cancer and that targeting PEG10 represses tumor progression in both in vitro and in vivo models. PEG10 is a placental gene essential for the development of mammalian placentation (17). Therefore, the origin of PEG10 is of interest considering its role in promoting migration and invasion of cancer cells (Fig. 5). PEG10 may also be oncogenic by modulating cell-cycle progression (35), reducing apoptosis mediated by SIAH1 (36), and/or by impeding TGFβ signaling via interaction with TGFβ receptor, ALK1 (37). PEG10 presents a wide diversity of functions, including regulating cell growth and differentiation, in addition to its key role in placental formation. As a potent growth promoter, PEG10 expression is tightly controlled. PEG10 gene is imprinted in the placenta/embryo and its expression is silenced in adult tissues. However, the expression of PEG10 is significantly enhanced in the neuroendocrine-like subtype of bladder cancer (Fig. 1A–C), a subtype which displays the poorest survival among the patients. Neuroendocrine differentiation is characterized by deregulated TP53 and RB1, and PEG10 is known to be reexpressed to drive proliferation and migration of cancer cells when left unchecked in the context of deregulated TP53 and RB1 pathways. While PEG10 does not directly contribute to neuroendocrine transdifferentiation (13), alterations in WT TP53 and tRb in neuroendocrine-like tumors de-repress PEG10 expression to enhance proliferation, survival, and migration of neuroendocrine-like tumors. In addition to altered transcriptional regulation, PEG10 may also undergo translational and posttranslational regulation that controls the protein level of PEG10 in cancer cells.
This study also links PEG10 to acquired chemotherapy resistance. PEG10 is induced in the bladder cancer cells after transient cisplatin and paclitaxel treatments and remains at high levels in stable drug-resistant cell lines. Silencing of PEG10 by siRNAs resensitizes treatment-resistant cancer cells to chemotherapy, indicating a role for PEG10 in stress responses and survival. Cancer cells stressed by treatment need to reprogram the transcriptome to adapt to varied microenvironments; reactivation of PEG10 is representative of lineage plasticity through activation of developmental pathways’ activation that support acquired treatment resistance in cancer. The unique genomic features (TP53 and RB1) of PEG10 regulation, along with its absence in most adult tissues, oncogenic characteristics, and intimate association with aberrant cancer cells, make it a distinct therapeutic target for a subset of advanced bladder cancers. While these data provide preclinical proof of principle for PEG10 inhibition in subtypes of advanced bladder cancer, assessment of PEG10 ASO combined with chemotherapy in preclinical models is required to further define clinical development path and roles of PEG10 in treatment response and resistance.
The ability of PEG10 to promote proliferation in both Rb functioning and Rb absent bladder cancer cells is of interest. Cellular division is well known to be controlled at the G1–S-phase transition by the Rb protein. Rb interacts with E2F and represses its transactivation of cell-cycle–regulating genes necessary for cellular division (38). During G1–S-phase transition, Rb protein is phosphorylated by cyclin-dependent kinases and their partners to release E2F protein from its suppression, allowing S-phase entry. The G1-phase arrest, however, does not always rely on Rb family members. Mouse embryos with triple knockout (TKO) of Rb protein and two family members, p107 and p130, live until days 9–11 of gestation, and the ability of TKO cells to arrest in G0–G1-phase is associated with repression of key E2F target genes (39), suggesting Rb-independent mechanisms regulating E2F transactivation. Mutations or deletion of RB1 are common in cancer, allowing escape from the antioncogenic senescence program. How Rb-negative tumor cells control proliferation rates in stressed environments remains undefined. The broad role of PEG10 in regulating proliferation in both Rb functional and Rb absent bladder cancer cells provide another rational for molecular-targeted therapy. Further study is required to define molecular mechanisms by which PEG10 regulates G1–S-phase transition in Rb-negative cells.
In conclusion, we demonstrate that PEG10 is associated with poor prognostic neuroendocrine subtype of bladder cancer, promoting cell survival, proliferation, and invasion. Inhibition of PEG10 may be a novel treatment strategy for certain subset of bladder cancers.
Disclosure of Potential Conflicts of Interest
S. Akamatsu reports grants and personal fees from AstraZeneca and Astellas Pharma, grants from Tosoh Corporation, and personal fees from Janssen outside the submitted work. F. Zhang reports grants from Vancouver Prostate Centre during the conduct of the study. H. Matsuyama reports grants from Janssen (research grant), and personal fees from Bayer (speaker honorarium), Pfizer (speaker honorarium), and MSD (speaker honorarium) outside the submitted work. A.W. Wyatt reports grants and personal fees from Janssen, and personal fees from AstraZeneca and Bayer outside the submitted work. P.C. Black reports a patent for cancer biomarkers and classifiers and uses thereof pending, a patent for molecular subtyping, prognosis ad treatment of bladder cancer pending, and a patent for systems methods and compositions for predicting metastasis in bladder cancer pending. No potential conflicts of interest were disclosed by the other authors.
Y. Kawai: Conceptualization, data curation, methodology, writing-original draft. K. Imada: Data curation, methodology. S. Akamatsu: Conceptualization, data curation, formal analysis, methodology, writing-original draft. F. Zhang: Conceptualization, data curation, methodology, writing-original draft, project administration, writing-review and editing. R. Seiler: Data curation, software, formal analysis, methodology, writing-original draft. T. Hayashi: Data curation. J. Leong: Data curation. E. Beraldi: Data curation. N. Saxena: Data curation. A. Kretschmer: Data curation. H.Z. Oo: Data curation. A. Contreras-Sanz: Data curation. H. Matsuyama: Data curation. D. Lin: Data curation, formal analysis. L. Fazli: Resources, data curation, formal analysis, visualization. C.C. Collins: Conceptualization. A.W. Wyatt: Conceptualization, writing-original draft. P. Black: Conceptualization, resources, writing-original draft, writing-review and editing. M.E. Gleave: Conceptualization, resources, supervision, funding acquisition, writing-original draft, writing-review and editing.
We thank Estelle Li, Igor Moskalev, Charan Tse, Joanna Pan, Brian Lee, and Teresa Huang for technical assistance. This study was supported by a Terry Fox Research Institute New Frontiers Program project grant (TFRI project #1062) for the whole author team.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.