Silencing of the miR-17∼92 Cluster Family Inhibits Medulloblastoma Progression

Medulloblastoma, a cerebellar tumor, is the most common malignant pediatric brain cancer (1). Molecular profiling classifies human medulloblastomas into four major subgroups with distinct mRNA and miRNA signatures: Sonic Hedgehog (SHH), WNT, group 3 (G3), and group 4 (G4; ref. 2). As current therapies induce adverse effects that compromise patients' quality of life (3, 4), novel therapeutic options have become a necessity. Small-molecule inhibitors of Smoothened (SMO) are effective at eliminating murine (5) and human (6) SHH medulloblastomas. However, treatment of young mice with the SHH antagonist induces long-bone defects (7). In the clinic, one patient with significant remission after treatment with GDC-0449 developed resistance due to SMO mutation (8). These recent results suggested that other approaches are necessary to treat SHH medulloblastomas.

MicroRNAs (miRNA) are endogenous noncoding RNAs of approximately 22 nucleotides, which play key roles in numerous biologic processes, including cancer (9). They bind to partially complementary sequences within the 3′-untranslated regions (UTR) of target mRNAs inducing translational repression or mRNA degradation (10). The miR-17∼92 cluster belongs to a family of 3 clusters encoded on different chromosomes (11). They include miR-17∼92, also called OncomiR-1, which encodes 6 miRNAs (miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92-1), miR-106b∼25, which encodes 3 miRNAs (miR-106b, miR-93m and miR-25), and miR-106a∼363, which encodes 6 miRNAs (miR-106a, miR-18b, miR-20b, miR-19b-2, miR-92-2, and miR-363). Individual miRNAs are classified into 4 families based on sequence similarities within their 6-nucleotide seed sequence (Fig. 1A; refs. 11, 12). The oncogenic potential of miR-17∼92 was discovered in B-cell lymphomas (13), with the miR-19a/b family responsible for tumorigenicity (14, 15).

Several mouse models of SHH medulloblastomas exist with mutation of Smo (16), Ptch1 alone (17), or with loss of Trp53 (18) or the cyclin-dependent kinase–inhibitory proteins, p18^Ink4c/Cdkn2c or p27^Kip1/Cdkn1b (19, 20). Mouse and human SHH medulloblastomas exhibit high levels of miRNAs encoded by miR-17∼92 and miR-106b∼25 (21, 22) with amplification of miR-17∼92 found in less than 10% of human medulloblastomas (22). Enforced expression of miR-17∼92 increases proliferation of postnatal day 7 (P7) cerebellar granule neuron progenitors (GNP; ref. 22) and collaborates with mutated Ptch1 to induce medulloblastoma in mice (21). We hypothesized that targeting members of the miRNA seed families from miR-17∼92 and miR-106b∼25 could serve as potential therapies for the treatment of SHH medulloblastomas. We recently reported that tiny locked nucleic acids (LNA) inhibit the function of miRNAs, including miRNA seed families (23–25). Here, we investigated the therapeutic potential of 8-mer LNA-anti-miRs in inhibiting miR-17, 20a, 106b, and 93 (anti-miR-17) and miR-19a and 19b-1 (anti-miR-19) in 2 murine SHH medulloblastoma models.

[Ptch1^+/−;Trp53^−/−] and [Ptch1^+/−;Cdkn2c^−/−] mice develop SHH medulloblastomas. Transplants were conducted in 1- to 2-month-old CD-1 nude (CD-1) mice (Charles River Laboratories). [NestinCre^+/−;miR-17∼92^{floxed/floxed};miR-106b-25^−/−] double knockout (DKO) mice were generated by conditionally deleting miR-17∼92 (11) in cells expressing the Cre recombinase driven by the Nestin promoter (26) in an miR-106b∼25--null background (11). Mice were housed in an accredited facility of the Association for Assessment of Laboratory Animal Care in accordance with the NIH guidelines. The Institutional Animal Care and Use Committee of SJCRH approved all procedures in this study.

Primary medulloblastoma cells from spontaneously arisen SHH medulloblastomas from [Ptch1^+/−;Cdkn2c^−/−] mice were purified and cultured as previously described (27). Medulloblastoma cells were treated with tiny LNAs targeting the seed region of miR-17, miR-19a, or the 8-mer scrambled LNA control (referred to as anti-miR-17, anti-miR-19, and scrambled LNA). After 48 hours, medulloblastoma cells were harvested with trypsin (Worthington Biochemicals) and ovomucoid (trypsin inhibitor, Worthington Biochemicals) solutions for further analysis. Fluorescence-activated cell sorting (FACS) was used to assess bromodeoxyuridine (BrdUrd) incorporation following a 2-hour pulse of 10 μmol/L BrdUrd (BD Pharmingen, 552598) and the percentage of Annexin V–positive cells. To determine the efficiency of uptake, medulloblastoma cells were treated with 8 μmol/L 6-carboxyfluorescein (FAM) tiny LNAs. Tiny LNA sequences were (5′–3′): anti-miR-17, AGCACTTT; anti-miR-19, ATTTGCAC; scrambled LNA control, TCATACTA, as previously published (23).

A total of 2 × 10⁶ medulloblastoma cells were treated with tiny LNAs during a 30-minute preplating step, harvested, resuspended in Matrigel (BD Biosciences), and injected subcutaneously into the flank of CD-1 mice. A subset of mice was sacrificed 14 days after transplant when flank tumors were palpable. Tumors were sectioned and stained for the FAM label using an anti-FAM–conjugated AlexaFluor 594. The presence of tiny LNAs was visualized by confocal microscopy. The remaining mice were imaged weekly using the VEVO-770 High-Resolution 3-Dimension Ultrasound Imaging System, and tumor volumes were measured by tracing tumors within virtual sections of the 3D image stack (VisualSonics). Mice were sacrificed when the size of the tumor was greater than 20% of the total animal body mass.

Three spontaneous medulloblastomas from [Ptch1^+/−;Trp53^−/−] mice were maintained by transplantation in the flank of CD-1 mice. Each tumor was purified and infected with lentiviruses encoding a spleen focus-forming virus-LTR driving 2A peptide–linked luciferase and yellow fluorescent protein (YFP; CL20-luc2aYFP). Forty-eight hours later, cells were subjected to FACS and 1 × 10⁶ YFP-positive cells were transplanted into the left flank of recipient animals (day 0). For cranial transplants, 0.5 × 10⁶ YFP-positive cells were injected into the cortices of CD-1 mice, as described previously (21). Bioluminescence imaging of luciferase activity monitored tumor growth twice weekly using a Xenogen IVIS-200 system prior and during LNA treatment. Three milligrams of d-luciferin (Caliper Life Sciences) was administered to each mouse intraperitoneally, 5 to 7 minutes before bioluminescence imaging. Radiance (photons/s/cm²/steridian) was determined within regions of interest (flank or head of mice) using Living Image Software version 4.3.1 (Caliper Life Sciences). Bioluminescence imaging detected luciferase in medulloblastoma cells the day after transplant (day 1) with an intensity ≥ 1.3 × 10⁶ radiance. Mice received intravenous (i.v.) injections via tail vein of saline-formulated tiny LNAs (scrambled, anti-miR-17, or anti-miR-19) or an equal volume of saline, with an initial loading dose of 25 mg/kg (day 2). For flank transplants, maintenance doses of tiny LNAs were administered twice weekly at 10 mg/kg (days 6, 9, 13, and 16). Mice were sacrificed 24 to 72 hours after the last dose of tiny LNAs. For cranial transplants, maintenance doses of 20 mg/kg were given on days 6 and 9. Mice were sacrificed when they became lethargic and displayed neurological signs (e.g., head tilting, head dome, failure to thrive), which occurred 5 to 8 days after transplant for mice treated with saline or scrambled LNA.

Three mice with intracranial transplantation of 0.5 × 10⁶ YFP-positive medulloblastoma cells from one [Ptch1^+/−;Trp53^−/−] tumor were injected i.v. with either 2 mL/kg of 2% Evans blue (Sigma-Aldrich, E2129) in 0.9% saline 6 days after transplant or 25 mg/kg FAM-labeled scrambled LNA 2 weeks after transplant. Evans blue is a dye with a high affinity for serum albumin not normally present in the brain (28, 29). Mice that received Evans blue were euthanized by cardiac perfusion with 4% paraformaldehyde 24 hours later. Mice that received tiny LNA injections were perfused after 30 and 60 minutes. Brains were cryoprotected in sucrose, frozen and sectioned for immunohistochemistry.

Total RNA from spontaneous medulloblastomas, purified tumor cells and flank allografts was extracted, reverse transcribed, and quantitative real-time PCR (qRT-PCR) for miRNAs was conducted, as described previously [ref. 21; ABI, catalogue number 4427975 and assay ID numbers = 001093 002308 (RNU6B), (hsa-miR-17), 002422 (hsa-miR-18a), 000580 (hsa-miR-20a), 000395 (hsa-miR-19a), 000396 (hsa-miR-19b), 000430 (hsa-miR-92), 000442 (hsa-miR-106b), 001090 (mmu-miR-93)]. Relative levels of microRNAs were quantified in triplicate. C_t values were normalized within the same RNA sample to the small nuclear RNA RNU6B and to P6 GNPs (relative value = 1).

The luciferase assay was conducted using the pmirGLO dual-luciferase miRNA target expression vector (pmirGLO; Promega). Oligonucleotides containing predicted or mutated miR-17 and miR-19 binding sites within Bmpr2, Smad5, and Smad4 3′-UTR sequences (see Supplementary Table S1) were cloned into pmirGLO according to manufacturer's instructions. SAOS-2 cells were maintained in Dulbecco's Modified Eagle's Media (DMEM) supplemented with 10% FBS, 4 mmol/L glutamine, and 100 units each of penicillin and streptomycin (GIBCO) at 37°C and 8% CO₂. Cells were obtained from Dr. Emma Lees (DNAX) in September 1994 and tested by immunoblotting with antibodies to p53 (SC#100 and Sc#99) and Rb (Pharmingen, G3-245) before being frozen down in September 1994. Cells were seeded in 24-well plates (2.0 × 10⁴ cells per well) and cotransfected the next day with 10 ng of pmirGLO reporter plasmids and 0 to 200 ng of MSCV-miR-17 ∼ 92-IRES-GFP, using Fugene HD (Promega). pCMV6 was used as control. Cells were lysed 48 hours later, and Dual-Luciferase Reporter Assays (Promega) were conducted on a Synergy 2 Biotek microplate reader.

P6 cerebella from wild-type and DKO mice were lysed using radioimmunoprecipitation assay (RIPA) buffer and proteins were extracted and 50 μg of proteins immunoblotted, as previously described (30), using antibodies to BMPR-II (BD Biosciences, 1:100) and actin (Santa Cruz Biotechnology, Inc; 1:2,000). The relative densities of Bmpr2 and actin bands were quantified using the gel analyzer plug-in on ImageJ64 (version 10.2; NIH) and normalized to relative Bmpr2 density in a NIH3T3 cell lysate (relative density = 1).

FISH for miR-17∼92 was conducted on 10 mouse SHH medulloblastomas [Ptch1^+/−;Cdkn2c^−/− (n = 6); Ptch1^+/−;Trp53^−/− (n = 4)] using the BAC clones RP23-132K20 for miR-17∼92 and RP23-35N6 for the chromosome 14 band A3 (30).

Medulloblastoma cells cultured on Matrigel-coated Lab-Tek-II CC2 4-Chamber Slides (Electron Microscopy Sciences) were treated with FAM-labeled tiny LNAs for 0.5, 1, 24, and 48 hours and fixed in 4% paraformaldehyde for 15 minutes at room temperature. Cells were counterstained with wheat germ agglutinin (WGA) conjugated to Alexa Fluor 594 and with 4′,6-diamidino-2-phenylindole, dihydrochloride (DAPI) according to Invitrogen's instructions. WGA is a lectin that binds to N-acetylglucosamine and N-acetylneuraminic acid (sialic acid) residues within the plasma membrane (31). Cells and tumor sections were stained with antibodies to Ki-67 (Leica, 1:1,000), FAM-AF594 (Invitrogen, 1:500), or GFP (Abcam, 1:500). Counterstained and immunostained cells were imaged using confocal microscopy (for details, see Supplementary Material). The number of FAM-, DAPI-, or Ki-67–positive cells was counted in each optical section using the cell counter plug-in on ImageJ64 (version 10.2; NIH). Four hundred or more cells, based on DAPI expression, were counted per treatment condition.

Macroscopic images of Evans blue dye within whole-mount mouse brain tumor and brain tumor sections were captured using the iSight camera on an iPhone 4S. Statistical significance was determined using the Student t test on GraphPad Prism software (version 5.0).

qRT-PCR analysis of miR-17, miR-18a, miR-19a, miR-20a, miR-19b-1, and miR-92-1 revealed that each member of the cluster was significantly overexpressed in purified medulloblastoma cells of independently derived medulloblastomas spontaneously arisen from 6 [Ptch1^+/−;Cdkn2c^−/−] mice (19) and 4 [Ptch1^+/−;Trp53^−/−] mice (18) compared with 1-month-old cerebellum from C57BL/6 mice. None of the 10 tumors showed amplification of miR-17∼92 (Supplementary Fig. S1A and S1B). miR-19a was the highest expressed in both SHH medulloblastoma mouse models followed by miR-20a in medulloblastomas from [Ptch1^+/−;Cdkn2c^−/−] and miR-17 in medulloblastomas from [Ptch1^+/−;Trp53^−/−] mice (Fig. 1B).

To assess tiny LNA uptake and subcellular localization, 8 μmol/L of FAM-labeled tiny LNAs were added directly to the culture medium. Medulloblastoma cells purified from 3 independently derived spontaneously occurring [Ptch1^+/−;Cdkn2c^−/−] tumors were incubated for different times before fixation and counterstained with WGA and DAPI. Confocal imaging revealed tiny LNAs in the nucleus and the cytosol of the cell soma and proximal neurites within 30 minutes after addition of tiny LNAs (Fig. 2A and B). FAM was observed in most cells treated with anti-miR-17, anti-mR-19, and scrambled LNA (Fig. 2C).

To assess the specificity of the LNA-anti-miRs in inhibiting miRNA function, RNA was extracted from medulloblastoma cells treated with anti-miR-17 and anti-miR-19a for 48 hours followed by qRT-PCR analysis of miR-17 and miR-19a levels within each sample from 3 tumors. The relative miR-17 and miR-19a levels in anti-miR-17- and anti-miR-19–treated cells, respectively, were decreased in each tumor compared with cells untreated or treated with scrambled LNA (Supplementary Fig. S2A and S2B).

To determine whether 8-mer anti-miRs led to suppression of medulloblastoma proliferation in vitro, medulloblastoma cells were treated for 48 hours with 8 μmol/L tiny LNAs based on a dose response (Supplementary Fig. S3A). Cells were harvested and counted. There was no significant change in the number of cells treated with the scrambled LNA compared with those left untreated. In contrast, we found significantly fewer cells after treatment with anti-miR-17, anti-miR-19, or anti-miR-17 with anti-miR-19 compared with untreated or scrambled LNA–treated controls; however, we did not observe an enhanced phenotype following treatment with both anti-miRs compared with either anti-miR alone (Supplementary Fig. S3B). Thus, each anti-miR was tested separately for the remainder of the experiments.

To assess whether the decrease in cell number was due to cell-cycle arrest or cell death, we measured proliferation by BrdUrd labeling and Ki-67 staining and apoptosis by Annexin V staining. Treatment of cells with anti-miR-17 or anti-miR-19 decreased the percentage of cells that incorporated BrdUrd and the number of Ki-67 immunoreactive cells without evidence of apoptosis (Supplementary Fig. S3C–S3E).

Tiny LNAs were added to 2 × 10⁶ tumor cells from 4 individually derived spontaneous medulloblastomas during the 30-minute preplating step before injection into the flank of naïve recipient mice. Secondary tumors monitored by ultrasound imaging arose between 2 weeks and 1 month after transplant. Tumors from cells treated with anti-miR-17 or anti-miR-19 were significantly smaller in volume (Fig. 3A) and weight (Fig. 3B) than those untreated or scrambled LNA–treated cells.

To determine whether the repression of flank tumor proliferation corresponded to a decrease in miRNA levels, RNAs from tumors were subjected to qRT-PCR. Tumors that arose from medulloblastoma cells treated with anti-miR-17 or anti-miR-19 exhibited a significant decrease in the relative levels of miR-17 and miR-19a seed sharing family members, respectively (Fig. 3C). The relative levels of miR-92, an miRNA that is not targeted by the 2 anti-miRs, were unchanged. Thus, 8-mer LNA-anti-miRs inhibit their cognate seed-sharing miRNA family members in vivo, as previously reported (23).

As the levels of miRNAs encoded by miR-17∼92 were similar between tumors from [Ptch1^+/−;Trp53^−/−] and [Ptch1^+/−;Cdkn2c^−/−] mice, and the loss of Trp53 in [Ptch1^+/−] mice facilitated tumor transplants compared with the [Ptch1^+/−;Cdkn2c^−/−] model, we used the [Ptch1^+/−;Trp53^−/−] medulloblastomas to test the effects of tiny LNAs in vivo. A total of 1 × 10⁶ YFP and luciferase-expressing tumor cells originating from 3 spontaneously occurring medulloblastomas from [Ptch1^+/−;Trp53^−/−] mice (tumor A, B, and C) were transplanted into the left flank of immunocompromised recipient CD-1 mice. In all 3 cases, i.v. administration of anti-miR-17 or anti-miR-19 resulted in inhibition of flank tumor progression, as measured by luciferase activity (Fig. 4A–D), and decreased tumor weight (Fig. 4E). The inhibition of tumor progression was correlated with a significant decrease in the relative levels of miR-17 and miR-19a seed-sharing families (Fig. 4F).

A major obstacle of using 8-mer LNA-anti-miRs for the treatment of brain tumors is their inability to cross the intact blood–brain barrier (BBB; ref. 23). Following intracranial transplant, Evans blue or FAM-labeled scrambled LNA were administered i.v. to 3 mice. We observed Evans blue dye accumulation around the implant site and tumor extravasation (Fig. 5A). FAM accumulation was similarly detected in tumors from three mice that received FAM-labeled scrambled LNA (Fig. 5B), suggesting that cranial implants compromise the BBB.

To determine whether treatment with 8-mer LNA-anti-miRs could be effective in suppressing the proliferation of medulloblastomas transplanted into the brain, we transplanted 0.5 × 10⁶ YFP and luciferase-positive cells from one of the medulloblastomas from [Ptch1^+/−;Trp53^−/−] mice into the cortices of CD-1 mice. Tumor-bearing mice were treated i.v. with saline-formulated tiny LNAs, and tumor growth was assessed using bioluminescence imaging. Tumors that developed in the cortices of mice treated with anti-miR-17 or anti-miR-19 exhibited less luciferase activity (Fig. 5C and D) and led to a 6-day and 3- to 4-day increase in mean survival, respectively, compared with saline and scrambled LNA–treated controls (Fig. 5E). Mice treated with tiny LNAs did not show any apparent signs of toxicity-related illness such as weight loss. qRT-PCR analysis revealed that tumors that arose from mice treated with anti-miR-17 had significantly reduced levels of miR-17, whereas its seed-sharing miR levels were not as significantly affected. Tumors treated with anti-miR-19 showed no significant reduction of the relative levels of miR-19a seed-sharing miRs (Fig. 5F).

The miR-17∼92 cluster targets SMAD2/3 (32) and SMAD4 (33), key effectors within the TGFβ signaling cascade, and BMPR2 within the BMP signaling cascade in human pulmonary arterial smooth muscle cells (34). Here, we tested whether miR-17 and miR-19 targeted members of the BMP signaling pathway, which dominantly and irreversibly induce the differentiation of SHH medulloblastomas (27). Using TargetScan, we found that Bmpr2, Smad5, and Smad4 contained predicted binding sites for miR-17 and miR-19a within their 3′-UTR.

We used the firefly–Renilla dual luciferase reporter assay to test whether miR-17∼92 induced posttranscriptional repression of Bmpr2, Smad5, or Smad4. Transfections were conducted in the SAOS-2 cell line due to its relatively low expression levels of the miR-17∼92 cluster (http://www.microrna.org). We observed a concentration-dependent repression of Bmpr2 luciferase reporter activity with increasing concentrations of miR-17∼92, which was not observed with Smad5 or Smad4 luciferase reporters (Supplementary Fig. S4A and S4B). Mutations in the miR-17 and miR-19a predicted binding sites of the Bmpr2 3′-UTR caused a derepression of luciferase activity (Fig. 6A), implicating Bmpr2 as a target of miR-17 and miR-19. These data were confirmed by Bmpr2 immunoblots of lysates from P6 cerebella from DKO and wild-type mice (Fig. 6B and C).

We investigated the role of miR-17 and miR-19a seed-sharing family members in medulloblastoma proliferation by inhibiting their function using 8-mer LNA-modified anti-miRs directed against their seed sequences, designated as anti-miR-17 and anti-miR-19. We found miR-19a, miR-17, and miR-20a to be the highest expressed miRNAs from miR-17∼92 within medulloblastomas from [Ptch1^+/−;Cdkn2c^−/−] and [Ptch1^+/−;Trp53^−/−] mice. Interestingly, overexpression of miR-19a and miR-19b are both capable of inducing B-cell lymphoma (14, 15). The differential expression of individual miRNAs from this cluster is explained, in part, by the tertiary structure of the miR-17∼92 pri-miRNA. Chaulk and colleagues revealed that this pri-miRNA cluster adopts a compact globular tertiary structure containing the 3′-miRNA hairpins (miR-19b and miR-92) within a core enveloped by the 5′-miRNA hairpins (miR-17, miR-19a, miR-20a), thereby affecting the efficiency of Drosha processing and the regulation of target mRNAs (35).

Tiny LNAs passively and efficiently entered medulloblastoma cells without the need of transfection or electroporation. In vitro treatment of purified medulloblastoma cells with anti-miR-17 or anti-miR-19 decreased their proliferation compared with untreated or scrambled LNA–treated controls. The diminished proliferative capacity of tumor cells after anti-miR-17 treatment may reflect the induction of a subset of antiproliferative genes, such as those responsible for differentiation. We found that Bmpr2 was a target of miR-17, as previously published (34), and extended these data by including miR-19. However, CLIP/SEQ experiments will be necessary to further identify bona fide targets for the miRNAs encoded by the miR-17∼92 cluster family within SHH medulloblastoma cells (36).

Our transplantation experiments provide several insights when considering the development of miRNA-based therapies for treatment of SHH medulloblastomas. First, we previously found variability in the expression of miRNAs from miR-17∼92 within human and mouse SHH medulloblastomas (21), which correlated with the fact that not all SHH medulloblastomas from [Ptch1^+/−;Cdkn2c^−/−] mice gave rise to secondary tumors in flank transplants. Interestingly, the relative expression levels of miR-17 and miR-19a were at least 2 to 3 times higher in transplantable tumors compared with those that were not (Supplementary Fig. S5A). Tumor sensitivity to 8-mer LNA-anti-miRs was correlated with the expression levels of miR-17∼92 (Supplementary Fig. S5B). Second, tiny LNAs were present in transplantable tumors for at least 14 days following medulloblastoma cell pretreatment (Supplementary Fig. S6). These data are consistent with previous findings showing the in vivo stability of tiny LNAs up to 21 days (23). Finally, we observed that tumors derived from cells pretreated with anti-miR-17 were smaller than those originating from cells pretreated with anti-miR-19. Similar results were obtained when tiny LNAs were administered intravenously into immunocompromised mice carrying flank and cranial allografts. One possible explanation is that anti-miR-17 targets the seed region of 4 miRNAs (miR-17, 20a, 106b, 93), possibly exhibiting a broader effect on medulloblastoma miRNA-mediated regulatory networks than anti-miR-19, which only targets 2 miRNAs (miR-19a and 19b-1).

Although systemic delivery of anti-miR-21 shows high levels of uptake and accumulation in the kidney cortex, liver, lymph nodes, bone marrow, and spleen of mice (23, 37), one of the major obstacles of using antisense oligonucleotides, including tiny LNAs, in the treatment of brain tumors is their inefficient delivery across the BBB (23). In this intracranial transplant model, the BBB was clearly compromised. However, it is unclear whether the integrity of the BBB is disrupted in children with medulloblastoma, which would facilitate brain penetration (38, 39). Other strategies to achieve direct delivery to tumors might be considered, including continuous infusion of tiny LNAs into the ventricles using minipumps (40) or the use of nanoparticles.

As the use of LNA-anti-miRs in non-human primates and humans is efficacious and safe (41–45), our data suggest that 8-mer LNA-anti-miRs targeting miR-17∼92 and miR-106b∼25 may have therapeutic application for the treatment of human SHH medulloblastomas expressing high levels of miR-17∼92 (22). The therapeutic value of LNA-anti-miRs in medulloblastoma could be extended to other oncomiRs such as, miR-1204 encoded by PVT1 and frequently fused to MYC in G3 medulloblastomas (46) and miR-182, which promotes G3 medulloblastoma cell proliferation and migration (47) and leptomeningeal spread of non-SHH medulloblastomas (48).

S. Obad and S. Kauppinen are and were, respectively, employees of Santaris Pharma, a clinical stage biopharmaceutical company that develops RNA-based therapeutics.

Conception and design: B.L. Murphy, S. Obad, S. Kauppinen, M.F. Roussel

Development of methodology: B.L. Murphy, O. Ayrault, S. Kauppinen, M.F. Roussel

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): B.L. Murphy, L. Bihannic, F. Zindy, M.F. Roussel

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): B.L. Murphy, S. Obad, L. Bihannic, O. Ayrault, M.F. Roussel

Writing, review, and/or revision of the manuscript: B.L. Murphy, S. Obad, F. Zindy, S. Kauppinen, M.F. Roussel

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Kauppinen, M.F. Roussel

Study supervision: M.F. Roussel

The authors thank Charles J. Sherr, Daisuke Kawauchi, and all members of the laboratory for their help during these experiments; Mark Hatley for the generous gift of pCMV6 and technical help with 3′-UTR assays; Tyler Jacks and Andrea Ventura for generously providing conditional miR-17∼92 and miR-106b∼25-null mice; and Fedor Karginov, Andrea Ventura, and Gregory Hannon for criticisms of the manuscript. They also thank Sarah Robinson and Shelly Wilkerson for managing the mouse colony; Jose Grenet and Dana Farmer for plasmid purification; Drs. Richard Ashmun and Ann-Marie Hamilton-Easton for flow cytometric analysis; Monique Payton, Cameron Ogg and Dr. Christopher Calabrese for ultrasound image acquisition and measurement of tumor volume; John Gray for lentiviruses; Sarah Robinson, Monique Payton, Shantel Brown, and Brittney Perdue for tail vein injections; Melissa Johnson and Shantel Brown for cranial implants; Dr. Yannan Ouyang for confocal imaging assistance; and Marc and Virginia Valentine for FISH and analysis.

This work was funded in part by NIH grant CA-096832 (M.F. Roussel), a core grant CA02165-29 (M.F. Roussel), Children's Brain Tumor Foundation Award (M.F. Roussel), the George J. Mitchell Endowed Fellowship (B.L. Murphy), and the American Lebanese-Syrian Associated Charities (ALSAC) of St. Jude Children's Research Hospital.

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.