MKAD-21 Suppresses the Oncogenic Activity of the miR-21/PPP2R2A/ERK Molecular Network in Bladder Cancer

Bladder cancer is one of the most common genitourinary malignancies, accounting for 3.1% and 1.8% of the overall cancer mortality in males and females, respectively (1). Urothelial carcinoma of the bladder, the most frequent histopathologic type of bladder cancer, has a variety of genetic and phenotypic characteristics. Many genetic factors, such as chromosomal abnormalities, genetic polymorphisms, mutations, and epigenetic alterations, contribute to tumorigenesis and the progression of this type of cancer (2). More than half of all bladder malignancies of all grades and stages contain chromosome 9 alterations, suggesting that chromosome 9 genes may be involved in early tumor development.

MiRNAs are single-stranded noncoding RNA molecules, 18–25 bases in length, that act as negative regulators of gene expression, through binding in the 3′ untranslated region (3′UTR) or coding sequence of genes (3). Growing evidence supports that dysregulation of miRNAs is associated with cancer initiation and progression in a variety of malignancies, suggesting that miRNAs may be used as molecular biomarkers for the diagnosis of cancer and prediction of prognosis or as therapeutic modalities. Pioneering data reported by several research groups have linked specific miRNA patterns associated with activation of different gene expression pathways in bladder cancer. For example, miRNA molecules aberrantly expressed in bladder cancers have been proposed to target components of the RAS kinase signaling pathway or the p53 pathway, such as FGFR3, p53-binding protein homolog (mouse), Mdm2-p53–binding protein homolog (mouse) and ataxia telangiectasia–mutated gene products (4). Thus, elucidation of the targets and molecular mechanisms triggered by aberrant miRNA expression is necessary for the potential clinical development of novel cancer therapeutics.

It has been reported that miR-21 is overexpressed in multiple human solid tumors, including bladder cancer (5, 6). A systematic review and meta-analysis showed that elevated miR-21 expression is associated with poor overall survival and proposed that miR-21 may serve as follow-up marker for early detection of bladder cancer progression or recurrence (7). Furthermore, miR-21:miR-205 ratio has been described as a tool for identifying the invasive capacity of bladder tumors with high sensitivity and specificity (6). Several studies aiming to determine the molecular mechanism by which miR-21 functions in bladder cancer showed that the oncogenic role of miR-21 might be attributed to suppression of PTEN (8, 9). MiR-21 has been shown to modulate cell proliferation and sensitivity to doxorubicin in transitional cell carcinomas of the urinary bladder. Particularly, BCL-2 upregulation upon miR-21 overexpression prevented T24 cells from apoptosis induced by doxorubicin (8). In addition, mechanistic studies demonstrate that downregulation of miR-21 results in cell-cycle arrest at the G₁ phase and p53 phosphorylation at Ser46, accompanied by significant attenuation of proliferation and migration of bladder cancer cells (9).

However, there is limited information regarding the molecular network that is regulated by miR-21 and its functional and clinical significance in bladder oncogenesis. In this study, we aimed to further define and characterize the miR-21 expression patterns associated with activation of certain gene expression pathways in bladder cancer.

RNA was extracted from 10 “normal” (adjacent nontumoral) and 74 bladder cancer tissues using the RNAeasy FPPE Kit (73504; Qiagen). All samples were classified according to the tumor–node–metastasis Classification of Malignant Tumors (UICC, 2014) staging and evaluated for histologic type. The formalin-fixed paraffin-embedded human bladder tissues were obtained from the School of Medicine at the University of Crete in Greece and the UCLA Department of Pathology and Laboratory Medicine. Patients provided written informed consent, and this study was performed in accordance with the Declaration of Helsinki and was approved by the Institutional Review Board committees of School of Medicine at University of Crete and School of Medicine at University of California, Los Angeles.

LNA miR-21 is an antisense oligonucleotide inhibitor against miR-21 seed sequence. LNA miR-21 sequence is 5′-T^mCAGTCTGATAAGCT-3′, where red color corresponds to the nucleotides with locked nucleic acid modification, ^mC corresponds to 5′-methy-cytosine, and the underlined sequence has phosphorothioate linkages between the nucleotides. All these chemical modifications in the LNA miR-21 inhibitor increase its resistance to nucleases, increasing its stability for in vivo studies.

Bladder cancer cell line 5637 was purchased from the ATCC (2013), and RT-112 was purchased from Sigma-Aldrich (2013) and were grown in RPMI 1640 (Gibco). Bladder cancer cell lines KU19-19 and HBCLS2 were purchased from DSMZ (2015 and 2013, respectively) and were grown in RPMI-1640 (ATCC). HBCLS1 cells were purchased from Cell Line Service (CLS; 2013) and were grown in RPMI-1640 (ATCC). Bladder cancer cell lines UMUC-1, UMUC-4, UMUC-5, UMUC-7, UMUC-9, UMUC-10, UMUC-11, UMUC-13, UMUC-14, UMUC-15, and UMUC-17 were purchased from MD Anderson (2013) and were grown in EMEM (ATCC). Bladder cancer cell lines HT-1197, TCCSUP, HT-1376, UMUC-3, SCABER, and J82 were purchased from the ATCC (2013) and were grown in EMEM (ATCC). SW1710 cells were purchased from DSMZ (2015) and were grown in DMEM (ATCC). Bladder cancer cell lines RT-4 and T24 were purchased from the ATCC and were grown in MCCOYS (ATCC). Bladder cancer cell lines BC-3C and CLS-439 were purchased from DSMZ (2015 and 2013, respectively) and were grown in MCCOYS (ATCC). SW780 cells were purchased from the ATCC (2013) and were grown in L-15 (ATCC). All the media used were supplemented with 10% FBS (Corning) and 1% antibiotic–antimycotic (Gibco). All cell lines were used in the described experiments before reaching the 7^th passage after thawing the cells out and were negative for mycoplasma prior performing these studies. Mycoplasma-free verification was checked by PCR using a mixed primer based on 16S rRNA of most common species of mycoplasma (10).

Treatment of RT-112 and 5637 with 50 or 100 nmol/L of LNA miR-21 or LNA scrambled negative control (LNA miR-scr; 199006-001, Exiqon) for 24 hours was followed by RNA extraction or functional assays. Bladder cancer cells were transfected with siRNAs for Protein Phosphatase 2 Regulatory Subunit Balpha (PPP2R2A), #1 (s5608) and #2 (s5609), and control (4390846) using Lipofectamine RNAiMax transfection reagent (13778150, Life Technologies) and incubated for 48 hours before RNA or protein extraction. Treatment of cells with 100 nmol/L miR-21 mimic or mimic miR-scramble was performed using Lipofectamine RNAiMax transfection reagent and incubated for 48 hours before RNA or protein extraction. For PPP2R2A transient overexpression experiments, RT-112 and 5637 cells were treated with miR-21 mimic or mimic miR-scramble and transfected with pPM-C-His vector (PV032532) or Blank-Control Protein Vector (PV001, ABM) using Fugene 6 (E2691, Promega). After 24 hours, cells were subjected to functional assays or serum starved for 16 hours, followed by 10-minute stimulation with FBS, and analyzed by Immnuoblot (IB) analysis.

Xenograft models of RT-112 and 5637 bladder cancer cell lines were established in 6-week-old CD-1 athymic nude mice (Charles River Laboratories). The following conditions were optimized for subcutaneous (s.c.) injection of both cell lines: 1.0 × 10⁷ cells with 50% matrigel/culture media (BD Biosciences). When tumors reached an average size of 150–300 mm³, mice were randomized into treatment groups.

For in vivo studies, we used MKAD-21 or LNA scramble negative control sequence (LNA miR-scr). Both were resuspended in RNAase-free water and diluted in sterile saline for intravenous (i.v.) or intraperitoneal (i.p.) injection at doses of 5 to 15 mg/kg. Tumor xenografts were measured with calipers 3 times/week, and tumor volume in mm³ was determined by using the formula: height × width × length. Mouse weights were recorded daily. Upon termination of the efficacy studies, xenograft tumor tissue was excised and snap-frozen in liquid nitrogen. Data were analyzed using StudyLog software from StudyDirector. All statistics were calculated using Microsoft Excel. All animal work was carried out under a protocol approved by the UCLA Institutional Animal Care and Use Committee and the UCLA Animal Research Committee.

RNA purified from cells and tissues with TRIZOL (15596026, Life Technologies) was reverse-transcribed to form cDNA using the Universal cDNA synthesis Kit (203300; Exiqon). Real-time PCR was carried out using the SYBR Green master mix (203450; Exiqon) and primers for miR-21 (204230, Exiqon) in a CFX384 Real Time PCR detection system (BioRad). MiR-21 expression levels were normalized to the levels of 5S rRNA (203906; Exiqon) and U6 snRNA (203907; Exiqon).

Real-time PCR was used to determine the expression levels of PPP2R2A. Reverse transcription was carried out using the iScript Reverse Transcription Supermix (1708841; Bio-Rad). Real-time PCR was carried out using IQ SYBR Green supermix (1708882; BioRad). The primer sequences used for real-time PCR were acquired from previous studies (11) or designed using the NCBI Nucleotide Database (http://www.ncbi.nlm.nih.gov/nuccore), Primer3 v.0.4.0 (http://bioinfo.ut.ee/primer3-0.4.0), and UCSC In-Silico PCR (http://genome.ucsc.edu/cgi-bin/hgPcr) and are included in Supplementary Table S3. Gene expression levels were normalized to the levels of Glyceraldehyde-3-phosphate dehydrogenase and β-actin. Normalized expression levels were quantified to the respective control. Bars represent mean ± SE; experiments were performed in quadruplicates for each condition.

Real-time cell proliferation analysis based on the application of electrical cell substrate impedance changes(https://lifescience.roche.com/wcsstore/RASCatalogAssetStore/Articles/BIOCHEMICA_4_08_p14-16.pdf) was performed using the xCELLigence RTCA instrument (ACEA Biosciences). The presence of cells affects the local ionic environment at the electrode solution interface. Cell status is represented by a dimensionless parameter termed Cell Index which is derived as the relative change in measured electrical impedance, after subtraction of the background measurements from media alone. Local ionic environment varies according to cell size, cell morphology, and strength of adhesion of the cells to the surface of the electrode, resulting in changes of the electrode impedance. RT-112 or 5637 cells (5 × 10³) were seeded in quadruplicates of an E-Plate 96 with interdigitated microelectrode arrays integrated in the bottom of each well. Subsequently, the E-Plate 96 was mounted on the SP Station of the xCELLigence RTCA system which is placed in a standard temperature-controlled CO₂ incubator under humidity saturation. The RTCA Software preinstalled on the RTCA control unit allows automatic selection of wells for measurement and real-time data acquisition within preprogrammed 15-minute time intervals. Bars represent mean ± SD; experiments were performed in quadruplicates for each condition.

Triplicate samples of 35 × 10⁴ RT-112 cells from each treatment were assayed in 48-well plates for colony formation using the CytoSelect Cell Transformation Kit (CBA-135; Cell Biolabs, Inc.). The number of colonies was quantified after 7 days by counting the entire area of each well divided in 5 fields, using a grid and an Evos microscope at a X20 magnification. Data are expressed as the mean number of colonies per field ± SE.

Invasion in matrigel has been conducted by using standardized conditions with BD BioCoat Matrigel invasion chambers (354480; BD Biosciences) according to the manufacturer's protocol. Assays for RT-112 and 5637 cells were conducted using 10% FBS-containing media as chemoattractant for 24 and 48 hours, respectively. Non-invading cells on the top side of the membrane were removed, whereas invading cells were fixed and stained with 0.1% crystal violet, 22 hours after seeding. The cells that migrated through the filter were quantified by counting the entire area of each filter divided in four fields, using a grid and an Evos microscope at a X20 magnification. Data are expressed as the mean number of invading cells per field ± SE.

5637 cancer cells were transfected with the PPP2R2A_pLightSwitch_ 3_UTR Reporter vector carrying human PPP2R2A 3′UTR (S812284; SwitchGear Genomics) and the pLenti CMV Puro LUC (w168-1; 17477; Addgene) containing luciferase reporter gene. At 24 hours, the cells were transfected with 100 nmol/L mimic miR-scr or miR-21 mimic, and 48 hours later, luciferase activity was measured using the Dual Luciferase Reporter Assay System (Promega). Data were expressed ±SE of the mean of three independent experiments.

Total cell extracts and protein lysates were separated by SDS-PAGE and transferred to PVDF membranes following standard procedures. Protein extraction from xenograft tissues was performed by grinding frozen specimens in liquid nitrogen into a powder with a mortar and a pestle, followed by the addition of RIPA buffer (R0278; Sigma). The samples were incubated with agitation for 1 hour at 4°C and clarified by centrifugation at 13,000 RPM for 20 minutes. The following antibodies were used for immunoblot analysis: PPP2R2A (5689), CREB (9104), phospho-p44/p42 MAPK (Thr202/Tyr204) (4370), total-p44/p42 MAPK (4695; Cell Signaling Technology).

Quantitative data are expressed as mean ± SD or SE of the mean, as indicated, or as boxes and whiskers representing the 25^th and 75^th percentiles (the lower and upper quartiles, respectively), using Origin 9.1 Software. Statistical analyses were performed using one-way ANOVA and Student t test (or both) or Pearson correlation. For animal studies, statistical differences between treatment arms at specific time points were performed using a two-tailed paired Student t test. P values of <0.05 were considered statistically significant.

The datasets generated during the current study are available in the GEO repository (http://www.ncbi.nlm.nih.gov/geo/) under the following accession numbers: GSE97782 and GSE97943.

To investigate the therapeutic potential of targeting miR-21 in bladder cancer in vitro, we efficiently suppressed its expression by treatment with LNA miR-21. LNA miR-21 compound was highly efficient in blocking miR-21 expression levels as determined by RT-qPCR analysis, in RT-112 and 5637 bladder cancer cell lines (Fig. 1A). The impact of miR-21 inhibition of expression on bladder cancer anchorage-independent growth was evaluated (Fig. 1B). Specifically, LNA miR-21 significantly reduced the number of colonies formed by RT-112 cells. Consistent with these findings, LNA miR-21 significantly decreased the invasive capacity of RT-112 and 5637 bladder cancer cells (Fig. 1C and D). LNA miR-21 (50 nmol/L) was determined as the dose to maximally block these miR-21–mediated phenotypic biological activities in bladder cancer cells, under the current experimental conditions.

To substantiate our findings in an in vivo experimental setting, we designed an experimental approach to study the effects of miR-21 blockade in mice bearing human bladder cancer xenografts. Nude mice were inoculated with either 5637 or RT-112 cells and developed palpable tumors in 5 days after cell injection. Mice were then treated with sterile saline or LNA miR-scr or increasing concentrations of MKAD-21 which is a chemically modified (phosphorothioate backbone and locked nuclei acid) antisense oligo against miR-21, by i.v. tail vein or i.p. injections every 5 days. In both 5637 and RT-112 xenografts, administration of MKAD-21 (15 mg/kg) resulted in decreased tumor size at the endpoint of the experiment relative to the respective controls, whereas minor alterations in the body weight of the mice were observed. Importantly, i.v. administration of MKAD-21 was more effective than i.p. administration. Specifically, 15 mg/kg of MKAD-21 given i.v. every 5 days for a total of 3 doses/cycles reduced the growth rate of the tumors, whereas mice body weight remained essentially stable (Fig. 2A and B). Furthermore, growth curves of the tumors originating from mice bearing 5637 xenografts, which were treated with increasing dosages of MKAD-21, show the threshold effect of the latter in bladder cancer progression (Fig. 2C). The efficiency of miR-21 inhibition of expression caused by the LNA miR-21 was validated by RT-qPCR in the excised tumors (Supplementary Fig. S1).

Using comparative qPCR analysis, we screened the levels of miR-21 transcript in a panel of 28 bladder cancer cell lines (Fig. 3A), which were further stratified into three groups (high, intermediate and low) according to their baseline miR-21 expression. Transcriptomic analysis revealed a gene signature consisting of 15 candidates that negatively correlate with miR-21 expression (Fig. 3B). As a next step, in silico complementarity analysis between miR-21 sequence and the 3′UTR sequences for all 15 genes found to be inversely correlated with miR-21 expression was used, using the RNA22 v2 microRNA target detection (12). Interestingly, the coding region of PPP2R2A was predicted to have a potent binding site for miR-21 (Fig. 3C).

Based on the findings above, we hypothesized that PPP2R2A is a downstream effector of miR-21 activity. Efficient miR-21–enforced expression by a miR-21 mimic in vitro, as assessed by RT-qPCR analysis (Fig. 4A), reduced PPP2R2A mRNA expression levels in both RT-112 and 5637 cells (Fig. 4B). A 3′UTR luciferase assay in 5637 bladder cancer cells transfected with a miR‐21 mimic revealed a significant decrease in the luciferase activity of PPP2R2A 3′UTR vector, relatively to mimic miR-scr, validating, at the molecular level, the direct interaction between miR-21 and the 3′UTR of PPP2R2A (Fig. 4C). To explore whether miR‐21 regulated PPP2R2A expression in vivo, RT-qPCR and IB were performed to detect the PPP2R2A levels in mice bearing 5637 bladder cancer xenografts treated with MKAD-21 or LNA miR-scr (Fig. 4D). The excised tumors demonstrated elevated PPP2R2A levels after treatment with LNA miR-21, suggesting their direct and inverse correlation.

The product of PPP2R2A gene belongs to the phosphatase 2 regulatory subunit B family, the major determinant of substrate specificity of the heterotrimeric serine/threonine Protein Phosphatase 2. PPP2R2A represents one of the four major Ser/Thr phosphatases implicated in the negative control of cell growth and division (13). Despite substantial evidence, pointing to its role as a tumor suppressor in bladder carcinoma (14, 15), PPP2R2A expression status and genomic alterations remain undefined. Given that PPP2R2A levels were found significantly decreased in bladder cancer patients (Fig. 5A), we sought to determine the functional role of PPP2R2A in bladder cancer cells. We first silenced its expression by using two different siRNAs (Supplementary Fig. S2) and conducted gene profiling analysis using the Agilent Human 44K expression array platform (Fig. 5B). In total, 498 and 914 genes were found to be differentially regulated by PPP2R2A expression in RT-112 cells by using siPPP2R2A#1 and siPPP2R2A#2, respectively. Notably, an overlap of 237 differentially expressed genes (DEGs) was consistently found with PPP2R2A suppression in RT-112 cells (Fig. 5C). Network analysis of the common deregulated genes, using the Ingenuity Pathway Analysis (IPA) Software, highlighted extracellular signal-regulated kinases (ERK) as the central hub (Fig. 5D) and predicted cancer/cellular growth and proliferation and as major functionally relevant networks downstream of PPP2R2A (Supplementary Table S1). In the same line, by mapping the differentially up- and downregulated genes to well-established ("canonical") pathways, cell-cycle–related events were distinguished as the top rated canonical pathways (Supplementary Table S2). The concept that PPP2R2A expression possibly regulates ERK pathway activation, as well as the proliferative capacity of bladder cancer cells, was further confirmed by IB analysis for the phosphorylated and thus activated p44/p42 MAPK (Fig. 5E) and real-time cell growth monitoring showing elevated proliferation of bladder cancer cells transiently depleted of PPP2R2A by an siRNA. Specifically, the characteristic kinetic trace of bladder cancer cells is reminiscent of an exponentially growing state (Fig. 5F). Differences in Cell Index measurements are significant after 24 or 3.3 hours of monitoring RT-112 and 5637 cells, respectively (Supplementary Table S3). Taken together, these data revealed that reduction of PPP2R2A levels results in activation of the ERK signaling pathway and elevation of bladder cancer cell–proliferative capacity.

To explore whether the oncogenic activity of miR-21 in bladder cancer is linked to the effects of its downstream target, PPP2R2A, we examined the association of miR-21 levels with ERK signaling pathway activation. Significantly higher levels of p44/p42 MAPK are observed upon PPP2R2A suppression both in bladder cancer cells and mouse xenografts treated with miR-21 inhibitor relatively to the respective controls (Fig. 6A and B). Importantly, simultaneous miR-21 and PPP2R2A overexpression in bladder cancer cells (Supplementary Fig. S3) showed reversal of the miR-21–induced ERK signaling activation (Fig. 6C), cell proliferation (Fig. 6D), and anchorage-independent growth (Fig. 6E) of RT-112 and 5637 cell lines. Statistically significant differences in Cell Index measurements of PPP2R2A-overexpressing versus control cells, upon miR-21 treatment, are shown in Supplementary Table S4. Taken together, these data reveal that miR-21 function and oncogenic activity in bladder cancer are mediated, at least in part, through the suppression of PPP2R2A expression and activation of the ERK signaling pathway.

Through integration of both molecular and clinical data from bladder tumors and interrogation of a large panel of human bladder cancer cell lines, we found a novel 15-gene signature that negatively correlates with miR-21 levels. Interestingly, one of these 15 genes, named Protein Phosphatase 2 Regulatory Subunit B Isoform A (PPP2R2A), was proven to be a direct miR-21 target and a negative regulator of the ERK signaling pathway and bladder cancer cell growth. Moreover, we determined the minimum dose and route of administration for MKAD-21 to effectively suppress bladder cancer tumor growth in vivo.

Current knowledge and concepts concerning the involvement of miRNAs in cancer have emerged from the study of cell culture and animal model systems featuring miRNA overexpression and/or ablation. These studies have demonstrated causal links between miRNAs and cancer development emphasizing their potential value in diagnostic, prognostic and possibly therapeutic strategies (16, 17). A growing body of evidence now suggests that miRNAs may contribute to bladder oncogenesis, progression, and metastasis. Examination of the differential expression of miRNAs between clinical bladder cancer and normal bladder tissue has led to the elucidation of an 11-miRNA expression signature. miRNAs downregulated in bladder cancer, such as miR-145, miR-143, and miR-125b, are known to be tumor suppressors, whereas miRNAs found to be upregulated, such as miR-21, miR-183, miR-96, miR-17-5p, and miR-20a, are oncogenic (17). Our analysis of the molecular and clinical data from bladder tumors as well as investigation of bladder cancer cell lines reveals a novel correlation of DEGs with elevated miR-21 levels and has a similar expression pattern in actual bladder cancer tumor specimens (18–20). By associating the levels of miR-21 with the expression status of its candidate downstream effectors, we provide molecular insights into the functional role of miR-21 as a mediator of key molecular drivers in bladder cancer.

MiRNA target prediction software identified a miR-21–binding site in the 3′UTR of PPP2R2A, and dual-luciferase reporter assays confirmed that the latter is a direct target of miR-21. Confirmation of this functional interaction was demonstrated with overexpression of miR-21, which reduced both PPP2R2A mRNA and protein expression in bladder cancer cell lines and human bladder cancer xenografts. Our mechanistic studies show that blockage of miR-21 expression results in the deactivation of the ERK signaling pathway, whereas overexpression of miR-21 triggers the same effects as PPP2R2A silencing in bladder cancer cells. Perhaps most importantly, concomitant enforced miR-21 and PPP2R2A expression rescued the miR-21–induced phenotype in bladder cancer cells lines, supporting the notion that PPP2R2A acts as a functional mediator of miR-21. Several findings lead to the conclusion that PP2A, as well as some of its regulatory subunits, such as PPP2R2A, has definitive roles as tumor suppressors in a large variety of cancers (21–23). Specifically, PPP2R2A is located on chromosome 8p21.2 and encodes B55α-regulatory subunit of PP2A. Loss of PPP2R2A is a common event in non–small cell lung cancer, and it results in the impairment of the homologous recombination repair pathway. In addition, shRNA-mediated silencing of PPP2R2A has been shown to be a potent oncogenic signal in experimental models of colon carcinogenesis (14). Furthermore, in patients with acute myeloid leukemia, PPP2R2A has been suggested to act as an adverse prognostic factor given that low levels of its expression in blast cells have been associated with shorter complete remission duration (24). The kinases that have been found to interact with PP2A include calcium-calmodulin–dependent protein kinase i.v. (25), p70 S6 kinase (26), p21-activated kinases (27), p38 kinase (28), casein kinase II (29), IκB kinases (30), MAPK (31, 32), and Akt (33, 34). In support of our current data, several reports using overexpression and RNA interference approaches, as well as in vitro dephosphorylation assays, have demonstrated the roles of the B55 regulatory subunits in the MAPK signaling pathway in multiple systems (35–37). Specifically, Van Kanegan and colleagues elucidated the role of PPP2R2A in targeting the PP2A heterotrimer to dephosphorylate and inactivate ERKs and indicated that PPP2R2A suppression leads to hyperactivation of ERK stimulated by constitutively active MAPK1. Along with the modulation of the Akt and ERK signaling pathways by PP2A-regulatory subunits, it was further demonstrated that PPP2R2Aα plays a role in regulating cell proliferation and survival, consistent with our results showing the tumor-suppressive actions of PPP2R2A in bladder cancer (38).

Previous studies have described the degradation of unmodified miRNAs (39). A locked 2′-O 4′-C methylene bridge in the structure of a modified oligonucleotide results in higher structural rigidity and increased selective affinity of the latter to the RNA counter strand and renders the oligo highly resistance to enzymatic degradation (40). The biological effects, as well as the safety issues derived from using LNA miR inhibitors for degradation of miRNAs, have been reported in several studies suggesting them as suitable candidates for therapeutic applications (39, 41). By derivatization with phosphorothioate linkages, enzymatic degradation is delayed and pharmacokinetics is enhanced. Phosphorothioate backbone modifications adequately stabilize the oligonucleotides against degradation and result in a high degree of binding to plasma proteins reducing rapid elimination (42). In a recent study, the pharmacokinetic and pharmacodynamic properties of LNA anti–miR-221 in NOD/SCID mice and Cynomolgus monkeys were reported. These studies show that LNA anti–miR-221 has a short half-life, optimal tissue bioavailability, minimal urine excretion in both mice and monkeys, and is detectable in mice vital organs and in xenografted tumors for up to 3 weeks after treatment (43). These studies highlight the suitability of LNA miRNAs for clinical use. MiR-21, which has been shown to be strongly overexpressed in glioblastomas, was silenced in vitro using LNA-modified antisense oligos and resulted in a significant reduction of tumor cell viability and elevated intracellular levels of caspases (44). Importantly, miR-21 inhibitors were found to be promising multiple myeloma therapeutic agents when delivered in vitro and in vivo (45). With this background, we tested LNA- and phosphorothioate-modified antisense oligo against miR-21 for its antitumor efficacy in human bladder cancer xenografts as a novel therapeutic approach. We show that i.v. administration of a low-dose of MKAD-21 drastically reduced human bladder cancer xenograft tumor growth when compared with i.p. administration. Further studies are warranted to investigate the molecular mechanisms underlying this novel inhibitor and to further evaluate its safety, in order to establish its potential value for treatment of bladder cancer patients with high miR-21 expression.

Collectively, the current study contributes to the understanding of the molecular mechanisms underlying the malignant features of the bladder cancer phenotype. We define a novel 15-gene signature that correlates with miR-21 levels in bladder cancer cells and demonstrate using several methodologic approaches that there is a direct targeting of PPP2R2A by miR-21. The tumorigenic potential of miR-21 was shown to depend on its downstream effector PPP2R2A, ultimately regulating ERK activation and bladder cancer cell growth. Finally, in view of our in vivo data showing the efficacy of a chemically modified antisense oligo against miR-21 (MKAD-21), the potential therapeutic applications of this compound should now be considered more closely for bladder cancer treatment.

D. Iliopoulos is cofounder at Algorithm Therapeutics Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: M. Koutsioumpa, D. Iliopoulos, A. Drakaki

Development of methodology: M. Koutsioumpa, N. O'Brien, D. Iliopoulos, A. Drakaki

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. O'Brien, F. Koinis, C. Vorvis, A. Soroosh, T. Luo, V. Georgoulias

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Koutsioumpa, H.-W. Chen, S. Mahurkar-Joshi, C. Vorvis, A.J. Pantuck, D. Iliopoulos

Writing, review, and/or revision of the manuscript: M. Koutsioumpa, H.-W. Chen, A.J. Pantuck, D. Iliopoulos, A. Drakaki

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Koutsioumpa, H.-W. Chen, D. Iliopoulos

Study supervision: M. Koutsioumpa, D. Iliopoulos, A. Drakaki

This work was supported by funding provided to Dr. Drakaki by the Division of Hematology/Oncology at David Geffen School of Medicine.

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.