Simultaneous downregulation of miR-21 and upregulation of miR-7 has anti-tumor efficacy

Dysregulation of miRNA expression has been implicated in cancer. Numerous strategies have been explored to modulate miR but sub-optimal delivery and inability to concurrently target multiple pathways involved in tumor progression have limited their efficacy. In this study, we explored the potential co-modulation of upregulated miR-21 and downregulated miR-7 to enhance therapeutic outcomes in heterogenic tumor types. We first engineered lentiviral (LV) and adeno-associated viral (AAV) vectors that preferentially express anti-sense miR against miR-21(miRzip-21) and show that modulating miR-21 via miRzip extensively targets tumor cell proliferation, migration and invasion in vitro in a broad spectrum of cancer types and has therapeutic efficacy in vivo. Next, we show a significantly increased expression of caspase-mediated apoptosis by simultaneously downregulating miR-21 and upregulating miR-7 in different tumor cells. In vivo co-treatment with AAV-miRzip-21 and AAV-miR-7 in mice bearing malignant brain tumors resulted in significantly decreased tumor burden with a corresponding increase in survival. To our knowledge, this is the first study that demonstrates the therapeutic efficacy of simultaneously upregulating miR-7 and downregulating miR-21 and establishes a roadmap towards clinical translation of modulating miRs for various cancer types.

MicroRNAs (miRs) are a class of small noncoding RNA that are involved in biological processes such as proliferation, differentiation, apoptosis and development 1,2 . Several studies have shown dysregulation of miRs in human cancers as a consequence of gene amplification, aberrant expression and gain of function mutations of oncogenic miRs (oncomiRs), as well as deletion or inactivation of tumor suppressor miRs (suppressor miRs) 3 .
miR-based therapeutic strategies are promising for cancer therapy 4 . However, targeting single miR is not sufficient to elicit significant therapeutic benefits in vivo. miR-21 is a key oncomiR, which is expressed highly in various cancer types [5][6][7][8][9][10][11][12][13][14][15][16] and several studies have shown over-expression miR-21 promotes proliferation, invasion, and migration by repressing expression of a range of tumor suppressor genes [5][6][7][8] . Down regulation of miR-21 induces apoptosis through targeting PI3K/AKT and JAK/STAT3 signaling pathways 7 and has a potential to collaborate or synergistically act with other miRs and fine-tune protein output 17 . microRNA-7 (miR-7) is an intronic microRNA that resides in the first intron of the heterogeneous ribonuclear protein K gene on chromosome 9 18 and is down regulated in different cancer types [19][20][21][22][23][24][25] . Recently, we have shown that forced expression of miR-7 in GBM cells results in down-regulation of EGFR and p-AKT, leading to activation of the NFkB signaling 26 . As both miR-21 and miR-7 activate different yet synergistic pathways and could be a contributing factor in tumor heterogeneity, it is essential to explore the potential of co-modulation of miR-7 and miR-21 that can target complementary pathways affecting tumor growth and development.
Two different strategies for miR-targeting therapy have been explored namely, activation or upregulation of tumor suppressor miRs and inhibiting the function of oncomiRs 4 . The strategies employed to block oncomiRs include: (1) anti-miR, (2) miR sponges, (3) miR masking, and (4) small molecule inhibitors 27 . Among these strategies anti-miRNA is widely used 27,28 . However, suboptimal delivery, poor stability, toxicity and off targets have limited their applications in vivo. Therefore, it is important to use effective and safe systems for successful delivery of anti-miR or miR mimics to cancer cells. Alternative delivery methods have been utilized in experimental and preclinical studies, such as liposomes, extracellular vesicles (EVs), polymer-mediated delivery systems, viral vectors, cell-based delivery systems and bacteriophage-based virus-like particles (VLPs) 28,29 . However, traditional methods have presented numerous caveats such as their transient nature of inhibition and several stoichiometric restrictions that have limited the success of the anti-miR approach. Novel expression systems such as miRzip have been developed to alleviate these issues 30 . miRzip anti-sense microRNAs are stably expressed RNAi hairpins that produce short, single-stranded anti-microRNAs that competitively bind their endogenous microRNA and inhibit its function 30 .The competitive nature of miRzip binding to its target results in the de-repression and subsequent changes in the transcripts targeted by the "zipped" microRNA. On the other hand, tumor suppressor miR can serve as therapeutic strategies classified as "miR-replacement" platforms and several studies have reported positive outcomes that are comparable to or greater than gene therapies for cancer 31 .
In this study, we first assessed the role of modulating oncogenic miR-21 via miRzip in a broad spectrum of cancer types and ultimately explored the therapeutic effects elicited by simultaneous targeting miR-21 and tumor suppressor, miR-7 in various cancer types in vivo.

miRzip-21 effectively downregulates miR-21, reduces invasion of cancer cells and induces apoptosis via suppression of AKT pathway.
The Cancer Genome Atlas data shows that miR-21 is up regulated in various human cancers and predominantly overexpressed in colon, prostate cancers and brain tumors (Fig. 1A). RT-PCR analysis of human prostate colon cancer and brain tumor (GBM) cell lines confirmed that miR-21 is widely overexpressed across different cancer types (Fig. 1B,C). To downregulate miR-21 in tumors cells, we packaged a lentiviral (LV) miRzip anti-sense miR-21 to stably express RNAi hairpins and investigated the efficacy of miRzip-21 mediated downregulation in different tumor cell types (Fig. 1D). RT-PCR analysis on LV-miRzip-21 transduced cells showed that miR-21 expression levels were significantly reduced as compared to control groups (Fig. 1E,F) resulting in a significant reduction in cell viability (Fig. 1G) in all cancer types including the primary patient derived GBM stem cells (GSCs). Furthermore, to understand the effect of LV-miRzip-21 following silencing of critical upstream targets, we transfected HCT116 and GBM8 tumor cells with siRNA for EGFR and AKT ( Supplementary Fig. 1A,B) prior to LV-miRzip-21 treatment. siEGFR or siAKT transfected HCT116 and GBM8 cancer cells when transduced with LV-miRzip-21 showed a reduced response in cell viability as compared to wild type cells ( Supplementary Fig. 1C). Morphological observations by light microscopy showed features of apoptotic cell death and reduced number of cells in miRzip-21 transduced cells as compared to controls ( Supplementary Fig. 2). . Data are presented as mean ± SD. Significant differences between miRzip-21 transduced cells and control groups are indicated by ***(P < 0.001), **(P < 0.01) and *(P < 0.05).
Next, we examined the role of miRzip-21 expression in survival, migration and invasion of tumor cells. Clonogenic assays revealed that miRzip-21 significantly inhibited the colony formation of tumor cells compared to control groups ( Fig. 2A and Supplementary Fig. 3). Wound-healing assay and transwell-based cell invasion assay indicated that downregulation of miR-21 robustly inhibits the migration and invasion of cancer cells compared to scramble and untreated control cells ( Fig. 2B and Supplementary Fig. 4). Western blot analyses showed that downregulation of miR-21 via miRzip-21 leads to reduced levels of activated AKT (p-AKT), a downstream target in the PTEN/AKT signaling pathway, in most cancer cells screened except Gli36d as compared to control treated cells. Moreover, the expression of PTEN, a known tumor suppressor protein, was significantly increased in tumor cells expressing miRzip-21. This resulted in tumor cells undergoing apoptosis as indicated by higher levels of cleaved-PARP (cl-PARP) and activation of caspases 3/7 and caspase 9 in miRzip-21 expressing cells compared to scramble and untreated control cells ( Fig. 2C-G). A diminished caspase 9 activation was observed when miRzip-21 was -expressed in GBM8 and HCT116 tumor cells treated with siEGFR or siAKT in, the as compared to wildtype cells ( Supplementary Fig. 1D). These data show that miRzip-21 effectively downregulates miR-21 and results in inducing caspase-mediated apoptosis via suppression of AKT pathway in a broad spectrum of tumor cell lines.

AAV-miRzip-21 but not MSC-miRzip-21 attenuates cancer growth in vitro and in vivo.
Mesenchymal stem cells (MSCs) have been demonstrated to be a safe and effective delivery vehicle for miRs, due to their ability to specifically target cancer cells and to transfer molecules via exosomal trafficking 32,33 . We hypothesized that MSC-shed exosomes will serve as carriers for anti-miR-21 and these MSC can be utilized to deliver miRzip-21 to the tumors in vivo (Fig. 3A). To determine the exosome enrichment of anti-miR-21 from transduced MSCs, exosomes were harvested from MSCs transduced with LV-miRzip-21 and control MSC. RT-PCR analysis revealed that MSCs shed exosomes were enriched in miRzip-21 (Fig. 3B). To investigate the therapeutic efficacy of MSC-miRzip-21, different cancer cells were cocultured with MSCs at 1:1 and 3:1 ratio. No change in tumor cell viability was observed in tumor cells post-co-culture with MSC-miRzip-21 in both tested ratios (Fig. 3C,D). To further investigate enrichment of anti-miRzip-21 western blot analysis of the MSC and isolated exosomes was performed. Western blot analysis using CD63 marker to identify exosomes revealed that exosomes are enriched from MSC engineered to express anti-miRzip-21 ( Supplementary Fig. 5). These data reveal that although MSC-shed exosomes are enriched in anti-miR-21, they are unable to influence tumor cell viability in vitro.
The delivery of transgenes via AAV provides long-term stable in vivo gene expression in both dividing and non-dividing cells with low risk of related genotoxicity, which makes it a useful and highly suitable option for cancer gene therapy [34][35][36][37][38] . AAV gene transfer technology has also shown promise in clinical trials 34,39 . To create an efficient delivery vehicle for targeting miR-21 in tumors, we created AAV bearing miRzip-21 ( Fig. 3E) and tested its efficacy in vitro. AAV-miRzip-21 transduced cells showed a reduction in cell viability (P < 0.01) as compared to control AAV treated or untreated cells (Fig. 3F). To evaluate the efficacy of AAV-miRzip-21 in vivo, we first created imageable tumors lines of prostate cancer (PC3-FmC), colon cancer (HT29-FmC) and GBM (LN229-FmC) by transducing them with LVs bearing a dual fluorescent and bioluminescent marker (Fluc mCherry; FmC). Next, mice bearing www.nature.com/scientificreports www.nature.com/scientificreports/ flank tumors of HT29-FmC and PC3-FmC (n = 10/tumor type) were injected intratumorally with three doses of 1 × 10 6 pfu of either AAV-scramble or AAV-miRzip-21 (n = 5/group) and mice were followed for changes in tumor volumes (Fig. 3G, H). A significant reduction in tumor volumes was seen in mice treated with AAV-miRzip-21 (P < 0.01) as compared to scramble treated groups (Fig. 3H). To evaluate the effect of AAV-miRzip-21 in intracranial GBM models, mice bearing LN229-FmC tumors were stereotactically injected intratumorally with two doses of 1 × 10 6 pfu of either AAV-scramble or AAV-miRzip-21 (n = 5/group) (Fig. 3I,J). A significant reduction in tumor volumes (P < 0.01) was seen in mice treated with AAV-miRzip-21 as compared to scramble treated groups (Fig. 3J). These data indicate that AAV-miRzip-21 effectively targets colon, prostate and brain cancer growth. However, the progression of colon and prostate tumor growth is halted more efficient than the GBM growth.
Co-modulation of miR-21 and miR-7 prolongs survival of mice bearing GBM xenografts derived from patient GSC. We have previously shown that forced expression of miR-7 in GBM cells results in down-regulation of EGFR and p-AKT and activation of the NFkB signaling 26 . In an effort to evaluate the effect of miR co-modulation on GBM tumor progression as compared to modulation of miR-21 alone (Fig. 3J), we explored the possibility of combing the downregulation of miR-21 and upregulation of miR-7 in GBMs. We tested a cohort of established and primary patient-derived GBM stem cells (GSC) for their response to a combination of miRzip-21 and miR-7. A significant reduction in GBM cell viability (P < 0.001) was seen in all the tested GBM cell lines post-combined treatment with AAV-miR-7 and AAV-miRzip-21 treatment as compared to monotherapy and control (Fig. 4A). Western blot analysis showed that combination treatment leads to substantial changes in the proteins involved in caspase-mediated apoptosis of tumor cells, in particular, caspase-3, caspase-8 and caspase-9 cleavage and subsequent PARP cleavage (Fig. 4B). Next, we evaluated the therapeutic efficacy in vivo using a primary patient derived GBM model. Specifically, mice bearing patient primary GSC (GBM18) expressing a bimodal imaging marker FmC, GBM18-FmC were challenged with 1 × 10 6 pfu of either AAV-GFP, AAV-miR-7, AAV-miRzip-21 or a combination of AAV-miR-7 and AAV-miRzip-21. A significant reduction (P < 0.001) in tumor burden was observed in mice treated with a combination of AAV-miR-7 and AAV-miRzip-21 as compared to the monotherapy and control (Fig. 4C). Kaplan Meier survival analysis showed a significant extension in survival of mice treated with the combination of AAV-miR-7 and AAV-miRzip-21 as compared to the other groups (Fig. 4D). Fluorescence imaging of brain sections revealed a robust infection of the GBM tumor following AAV injection (Fig. 4E) H&E analysis revealed a reduction in tumor burden in mice brains following dual modulation of miR-7 and miR-21 (Fig. 4F). These data reveal that modulation of miR-7 and miR-21 presents a therapeutic benefit in mice bearing GBM.

Discussion
In this study we show that delivery of miRzip-21 effectively induced apoptosis in different cancer types in vitro and mouse xenografts of prostate cancer, colon cancer and brain tumors. We also demonstrate a robust improvement in therapeutic benefit in mice bearing GBM brain tumors by co-modulating miR-21 and miR-7 thereby presenting a unique approach to target heterogenic tumors.
Several studies have shown miR-21 is among the most predominantly dysregulated miR in different cancer types including prostate cancer, colon cancer and highly malignant brain tumors and there is positive correlation between miR-21 expression levels and tumor stages [5][6][7][8] . miR-21 activates the EGFR/AKT signaling pathway via targeting tumor suppressor genes such as PTEN, PDCD4, APAF1, TPM1, TIMP3, RECK, FOXO1 and SPRY2 40 . Different systems have been used to inhibit oncomiR functions such as anti-miRs, miR sponges, miR-masking and small molecule inhibitors 27 . Despite great promise of available therapeutic anti-miR systems, the effectiveness of current anti-miR strategies is hindered by their transient nature of inhibition. In this study, we used the miRzip anti-sense microRNAs which are stably expressed RNAi hairpins that have anti-microRNA activity to downregulate miR-21. miRzip systems overcome the drawbacks by providing permanent expression of short hairpin anti-sense miRNAs and have been shown to efficiently prevent recurrence of cancers without adverse effects 41 .
Our findings indicate that intracellular constitutive expression of miRzip-21 has the potential to be more effective than other means of downregulating miRs in cancer cells with high level of miR-21 42 . microRNA-7 (miR-7) is an intronic microRNA that resides in the first intron of the heterogeneous ribonuclear protein K gene on chromosome 9 18 . We previously shown that forced expression of miR-7 via lentiviral vectors suppresses the EGFR/PI3K/ AKT signaling pathway in GBM 26 . miR-7 is also known to induce apoptosis through inhibition of BCL2 43 and inhibition of miR-21 activates caspase 9 and 3 7,44 .
Cancer cells activate multiple anti-apoptotic pathways to escape programmed cell death through overexpressing many of proteins. Therefore, targeting a single pathway maybe not be adequate for complete tumor control. miR-21 downregulation is known to induce apoptosis through inhibition of PI3K/AKT and JAK/STAT3 7 . Previous studies have also shown that miR-21 activates RAS/MAPK pathway through targeting antagonists of this pathway 45 and can simultaneously inhibit PI3K/AKT and RAF/MEK/ERK in GBM 46 . miR-7 is also known to induce apoptosis through inhibition of BCL2 43 and inhibition of miR-21 activates caspase 9 and 3 44,47 . In order to improve outcomes, simultaneously targeting both miR-21 and miR-7 pathways offers a promising strategy as they target parallel cell survival pathways and produce a synergistic effects 48,49 . Although there are few reports on forced co-expression or co-inhibition of miRNAs in cancer study [50][51][52][53]   www.nature.com/scientificreports www.nature.com/scientificreports/ co-inhibition of miR-21 and miR-10b can significantly induce apoptosis and reduce cell invasion in human GBM 50 . In an effort to improve therapeutic outcomes by miR modulation, we have explored co-modulation of miR-21 and miR-7 in this study as a strategy to target heterogenic tumors. Our findings indicate that simultaneous suppression of miR-21 and upregulation of miR-7 exert synergistic anti-cancer effects on human GBM through inhibition of EGFR and p-AKT and activation of caspase-mediated pathways, lead to prolonged survival rate of mice bearing xenografts model of patient-derived GBM 26 .
To gain insights into the molecular mechanisms underlying apoptotic effects of miRzip-21, our data showed the expression levels of cleaved-PARP, PTEN and caspase 3/7 and 9 activity were increased, while the level of p-AKT was decreased after miR-21 suppression. These results suggest that downregulation of miR-21 can increase PTEN which acts as negative regulator of PTEN/p-AKT pathway. Elevated level of PTEN results in activation of caspases 9 and 3/7. Previous studies have shown that inhibition of miR-21 in PTEN mutant cell lines suppresses activated AKT and EGFR 7 and leads to decreased levels of STAT3 and p-STAT3 7 . In our studies, we showed miRzip-21 mediated reduced levels of activated AKT and increased level of cleaved-PARP, PTEN and caspase 3/7 and 9 activity in PTEN mutant cell lines. We have also shown that silencing critical upstream receptors and targets diminished effects of LV-miRzip-21. This validates the involvement of the EGFR/PI3K/AKT pathway and proposed mechanism of action is depicted in Supplementary Fig. 6. This is in line with the previous studies showing that downregulation of miR-21 can induce apoptosis through inhibition of EGFR in JAK/STAT3 pathway, independent of the PTEN status 29 .
Our data indicates that modulation of specific miRs such as miR-21 could be an effective therapeutic strategy for personalized cancer treatment. However, sub-optimal delivery systems have hindered their progress into clinical settings. Exosomes offer a non-toxic and non-immunogenic delivery systems, however, efficiency of their drug-loading and retention is not sufficient to have substantial therapeutic effects in vivo 32,33 . Previously published studies have shown that intravenously injected exosomes are rapidly cleared from the blood and accumulate in lung, liver and spleen 54,55 . Industrial scale production of exosomes is a big hurdle to overcome before exosomes are tested in the clinic for treatment of a vast array of diseases 32 . In addition, there are no published results of trials, which used exosomes as miR carriers yet 56 . Previous studies have shown that MSCs can effectively deliver miRs to cancer cells via their exosomes [57][58][59] . We investigated the therapeutic efficacy of miRzip-21-carrrying MSC as source of exosomes containing miRzip-21. Our results showed no significant decrease in tumor cell numbers when co-cultured with MSC-miRzip-21 probably due to insufficient amount of anti-miR-21 in exosomes derived from MSCs.
The delivery of transgenes via AAV provides long-term stable in vivo gene expression in both dividing and non-dividing cells with low risk of related genotoxicity, which makes it a useful and highly suitable option for cancer gene therapy [34][35][36][37][38] . AAV gene transfer technology has also shown promise in clinical trials 34,39 . Currently, AAV-mediated gene delivery is performed in clinical trials of several human diseases such as Parkinson's, hemophilia A and B, cystic fibrosis, muscular dystrophy, and ocular disease [35][36][37][38] . Additionally, AAVs have been engineered to deliver anti-angiogenic, tumor suppressor, DNA repair, cytotoxic or suicide genes, cytokines, shR-NAs and antibodies. These studies showed promising results in preclinical stage 4 . So far, different serotypes of AAVs including AAV1, AAV-2, AAV1-AAV2 hybrids, AAV-6, AAV-7, AAV-8, AAV-9 and AAVrh10 have been applied in more than 200 gene delivery trial studies. AAV serotype-9 (AAV9) has the ability for effective cross the blood-brain barrier (BBB) and infects CNS cells 60 . Using AAV delivered miRzip-21 was shown to have efficacy both in vitro and in vivo in a broad spectrum of tumor lines, thus offering a potential delivery system for oncomiR silencing. Among all naturally occurring AAV serotypes, AAV2 is the best characterized and extensively studied and has served as the archetype for AAV biology 36 . AAV2 serotype additionally has a broad tissue tropism as compared to the other known serotypes thus making them a choice of selection for gene therapy 36 . In this study, we have used AAV2 serotype based on our previously published studies 26 and its efficacy in effectively delivering miR to brain tumors.
In conclusion, our data strongly reveal that AAV mediated stable silencing of miR-21 in combination with forced expression of miR-7 exerts superior anti-tumor outcomes and will serve as templates for a novel therapeutic approach in treating various cancer types.
24 hours later, cells were transduced with either LV-miRzip-21 or LV-Scr. Cells were incubated for ten days, colonies were fixed and stained with 0.2% crystal violet (Sigma-Aldrich) and counted.
Wound healing assay is a widely used technique to analyze the ability of cell migration in vitro. Adherent cancer cells were plated in 24 well plates and then transduced with LV-miRzip-21 or LV-Scr. After 24 hours, a straight scratch was introduced using a pipette tip. Cell migration was monitored at 4 different time points and quantified using Image J1.48 v.
Invasion assay. 1 × 10 5 LV-miRzip-21 or LV-Scr transduced cells in serum-free media were seeded on Matrigel-coated chambers of the 24 h well trans well plates (pore size 0.8 μm, Corning, USA). 700 µl of medium containing 15% FBS was used in the lower chamber. After 24 hours, cells were fixed with methanol and stained using 0.2% crystal violet. Cells were removed from upper side of the filters and stained cells on the lower side of the membranes were counted under a light microscope.
Silencing EGFR and AKT using siRNA. HCT116 and GBM8 tumor cells were transfected with either siRNA for AKT (Thermo Fisher Scientific, Waltham, MA), siRNA for EGFR (Thermo Fisher Scientific, Waltham, MA) or scramble RNA using Lipofectamine RNAiMax reagent (Life Technologies, Carlsbad, CA) per manufacturer's recommendations. 72 hours post transfection, cells were assayed for gene knockdown using western blotting as described below.
Western blotting. Total protein was extracted from tumor cells following transduction with either LV-miRzip-21 or LV-Scr and quantified using Bradford protein assay. For siRNA studies, the total protein from GBM8 or HCT116 tumor cells was extracted 48 h post transfection and quantified using Bradford Assay. For exosome analysis, total protein was extracted from MSC-miRzip-21 or MSC-Scr and also from exosomes purified from culture medium by ultracentrifugation. 20 µg of the protein lysates from these samples were loaded onto 10% Mini-PROTEAN TGX Precast Protein Gels (Bio-Rad, USA) and conducted at 110 V for 75 mins. Flowing the separation step, proteins were transferred to nitrocellulose membranes at 100 V for 1 h. Membranes were blocked with 5% dry milk or 5% bovine serum albumin (BSA) and 0.1% Tween-20 in TBS for 1 h at room temperature (RT), and incubated with primary antibodies in TBST overnight at 4 °C. After treated with HRP-conjugated secondary antibodies in TBST for 1 h at RT, membranes were developed with Super Signal West Pico system (Thermo Fisher Scientific, Waltham, MA). Then, signals were visualized using autoradiographic films and quantified with Image J. We used Anti-α-Tubulin or Vinculin as a loading and internal control. Adeno-associated virus packaging. Recombinant AAV was produced in HEK293T cells by standard triple transient transfection system with some modification. Briefly, three plasmids of pAAV-RC, pHelper and AAV transfer plasmid were co-transfected into HEK293T cells using polyethylenimine (PEI). Cells containing viral particles were collected 72 h after transfection, and the AAV particles were purified by iodixanol gradient (15%, 25%, 40% and 60%) ultracentrifugation, concentrated and formulated in phosphate buffered saline (PBS) containing 0.001% Pluronic F68 (Gibco). Virus titer was determined by measuring DNaseI-resistant genome copies using quantitative PCR.

Intracranial GBM cell implantation and in vivo bioluminescence imaging.
To understand the effect of AAV-miRzip-21 on GBM, LN229-FmC (2.5 × 10 5 cells per mouse, n = 10) were stereotactically implanted into the brains (right striatum, 2.5-mm lateral from bregma and 2.2-mm deep) of 8-week-old nude mice. Mice were then stereotactically injected with 1 × 10 6 pfu AAV-miRzip-21 (n = 5) or AAV-scramble (n = 5) on days 8 and 15-post tumor implantation. GBM burden was followed in in real time by bioluminescence imaging as described previously. To assess therapeutic benefit of the combination of anti-miR-21 and miR-7 modulation, GBM18-FmC cells (5 × 10 5 cells per mouse, n = 32) were stereo tactically implanted into the brains (right striatum, 2.5-mm lateral from bregma and 2.2-mm deep) of 8-week old nude mice. Mice were then stereotactically injected with 1 × 10 6 pfu AAV-miRzip-21 (n = 8), 1 × 10 6 pfu AAV-miR-7 (n = 8), 1 × 10 6 pfu AAV-miRzip-21 and 1 × 10 6 pfu AAV-miR-7 (n = 8) or AAV-scramble (n = 8) on days 15 and 22-post tumor implantation and followed for the GBM burden in real time by bioluminescence imaging (BLI) as described previously 26 and also followed for survival analysis. All in vivo procedures were approved by the Institutional Animal Care and Use Committee at BWH. Mice demonstrating any visual signs of distress as seen by impaired locomotion or difficulty in accessing food and water were excluded from the study. Treatment groups were randomly assigned from the total number of mice that were included in the study and the investigator was not blinded while administering the treatment.
Tissue processing, H&E staining and immunohistochemistry. GBM were prefused and brains were removed and sectioned for histological analyses. Brain sections were mounted on slides, washed in PBS and mounted and visualized for fluorescence on a confocal microscope (LSM Pascal, Zeiss, Oberkochen, Germany) 26 . For H&E staining, sections were incubated with Hematoxylin and EosinY (1% alcohol), dehydrated with 95% and 100% ethanol and mounted in xylene-based media 26 .
Statistical analysis. For in vitro experiments, one-way ANOVA was used to calculate the P values using SPSS 16.0 software. Pair-wise comparisons were conducted using two-sample T-Test with Bonferonni adjusted P-values for the cases of multiple pair-wise tests. Data were expressed as mean ± S.E.M. or S.D. and differences were considered significant at P < 0.05. For survival analysis, Kaplan-Meier estimates and the Log-Rank test for the comparison of survival curves were conducted for all time to event outcomes.