CRISPR-Cas knockout of miR21 reduces glioma growth

Non-coding RNAs, including microRNAs (miRNAs), support the progression of glioma. miR-21 is a small, non-coding transcript involved in regulating gene expression in multiple cellular pathways, including the regulation of proliferation. High expression of miR-21 has been shown to be a major driver of glioma growth. Manipulating the expression of miRNAs is a novel strategy in the development of therapeutics in cancer. In this study we aimed to target miR-21. Using CRISPR genome-editing technology, we disrupted the miR-21 coding sequences in glioma cells. Depletion of this miRNA resulted in the upregulation of many downstream miR-21 target mRNAs involved in proliferation. Phenotypically, CRISPR-edited glioma cells showed reduced migration, invasion, and proliferation in vitro. In immunocompetent mouse models, miR-21 knockout tumors showed reduced growth resulting in an increased overall survival. In summary, we show that by knocking out a key miRNA in glioma, these cells have decreased proliferation capacity both in vitro and in vivo. Overall, we identified miR-21 as a potential target for CRISPR-based therapeutics in glioma.


INTRODUCTION
Glioblastomas are the most common and lethal primary tumors of the central nervous system. They are known for their extensive cellular heterogeneity, among patients but also within tumors, which makes treatment challenging. 1 Despite efforts to improve existing treatments, the prognosis of this highly malignant glioma remains poor. In the last few decades there have been no significant improvements in mortality rates, generating an urgency to develop novel treatment strategies.
MicroRNAs (miRNAs) are short, evolutionarily conserved, non-coding RNA molecules; involved in translational repression, targeted cleavage, and deadenylation of messenger RNAs (mRNAs). 2 Dysregulation of miRNA expression has been observed in many different cancer types, including gliomas. 3 Interestingly, aberrant expression of miR-21 has been reported in many cancers and is highly upregulated in glioblastoma. [4][5][6] This miRNA is also involved in embryogen-esis, self-renewal, and development, and is among the most studied miRNAs in cancer research. 7 Murine miR-21 is located on chromosome 11 (human miR-21 is located on chromosome 17) within the intronic region of the gene for vacuole membrane protein 1 (Vmp1) (synonym Tmem49), a gene involved in autophagy and apoptosis. 8,9 Previously, miR-21 has been implicated in the pathogenesis of various cancers, making it a candidate biomarker and a potential target for therapeutics. To investigate the specific role of miR-21 in tumor growth, gain-and loss-of-function assays are frequently used to study which genes are regulated by this miRNA. 10 Many studies have been able to inhibit or reduce miRNA expression by using antisense oligonucleotide inhibitors or miRNA sponges (sponge RNAs contain complementary binding sites to an miRNA of interest and are produced from transgenes within cells). [11][12][13] However, the effectiveness of these inhibitors has not been robust, partly due to the short length-$22 nt-of miRNAs. 14 A more reliable technique to study the loss of function of miRNA with high efficiency and specificity is establishing a knockout (KO) of the genomic encoding sequences. Therefore, the aim of this study was to target and knock out the miR-21 sequences via clustered regulatory interspaced short palindromic repeats (CRISPR)-Cas technology. It has recently been described that expression levels of miR-21 in tumors are inversely correlated with the length of survival of glioblastoma patients. 15 In detail, SOX2 has been identified as a target of miR-21 in glioblastoma and has led to the characterization of a high-miR-21/low-SOX2 glioblastoma subtype associated with poor survival outcome. 16 In addition, elevated miR-21 levels in extracellular vesicles derived from cerebrospinal fluid (CSF) have been used as a biomarker linked to worse prognosis in glioma patients. 17 Together, these findings suggest that elevated miR-21 levels have an important role in tumor progression and poor survival among glioma patients.
Here, we present a strategy to KO miR-21 expression by targeting the alleles encoding miR-21 using an efficient CRISPR-Cas12a genomeediting system. Overall, we identify miR-21 as a potential target for CRISPR-based therapeutics in glioma.

Gene editing resulted in KO of miR21
To achieve the KO of miR-21 loci, we first tested five different CRISPR RNAs (crRNAs) for efficiency. Sanger sequencing and next-generation sequencing (NGS) were used to validate the CRISPR activity and to verify the miR-21 KO in mouse (GL261and CT2A) and human (U87) glioma lines ( Figure 1A). In detail, a CRISPR-Cas genome-editing system was used to KO miR-21 gene expression, as illustrated in Figure S1A. The CRISPR-Cas12a plasmids were designed to co-express a fluorescent marker to track transfection efficiency and identify cells that were potentially edited. Following this approach, we identified cells that expressed both the Cas12 (co-expressing GFP) and crRNA (co-expressing mCherry) by means of fluorescence microscopy ( Figure S1B). By sorting cells expressing both fluorescent proteins using fluorescence-activated cell sorting (FACS), we were able to select only cells where both gene-editing components were expressed. Subsequently, effective CRISPR activity was determined by disruption of the miR-21 sequence downstream of the specific crRNA sequence ( Figures S1C and S2). Subsequently, cells which had undergone CRISPR editing were single-cell sorted and expanded as clones, with all cells in the clone having the identical editing gene miR-21 KO events. These clones and all subsequent experiments were generated using the crRNA "JIR327," which showed high efficiency and specificity, determined on the basis of DNA Sanger sequencing ( Figure S3). First, the generation of miR-21 KO clones was confirmed by testing the expression levels of miR-21 using qRT-PCR and analyzed for the disrupted genomic sequences using CRISPResso2. 19 We compared the expression levels of miR-21 in GL261 wild-type (WT) cells with GL261 CRISPR-edited   cells, CT2A WT cells compared with CT2A CRISPR-edited cells, and  U87 WT compared with U87 CRISPR-edited cells labeled as miR-21 KO. As controls we used primary neonatal astrocytes isolated from C57BL/6 (miR-21 +/+ ) mice and miR-21 À/À mice. 20 As expected, miR-21 levels of the miR-21 KO lines (KO1-3) for GL261 and CT2A mouse and U87 human glioma cells were significantly lower than the WT cells from which they were derived and were similar to levels in miR-21 KO astrocytes, being at cycle threshold (CT) values that were considered background ( Figure 1B) Expression of miR-21 in GL261 cells was not affected by transfection of the Cas12 expression plasmid alone ( Figure 1B). To test whether the transcription of miR-21 was disrupted, we analyzed the expression of the pri-miR-21. We found no significant difference between pri-miR-21 WT and pri-miR-21 KO for GL261, CT2A, U87, and miR-21 KO mousederived astrocytes ( Figure 1C). We also evaluated expression levels of miRNA miR-29b and miR-15b as controls, whereby expression levels were similar in WT samples compared with KO1-3 for both GL261-and CT2A-derived lines ( Figures 1D and 1E). Next, the CRISPR-edited clones were aligned to the GL261 WT and the CT2A WT sequences to analyze the CRISPR-induced insertion-deletion (INDEL) ( Figure 1F). No changes were observed outside of the target of interest, as determined by Sanger sequencing of the top five off-targets (Table S1) for the miR-21 GL261 KO cell lines compared with the GL261 miR-21 WT cell line ( Figure S4). The CRISPR-edited colonies were analyzed using CRISPResso2 and compared with the unedited cells. We generated in total three miR-21 KO colonies each for GL261, CT2A, and U87 cell lines ( Figure S5).  Figure 1G). Importantly, NGS showed quantification of nucleotide percentages and deletions that occurred within the region of interest in CRISPR-generated miR-21 KO colonies 1-3 ( Figure 1H). INDEL quantification represents percentages of total reads of individual  Figure 1I). In summary, we demonstrated a successful approach to CRISPR editing of the miR-21 genomic locus and validated the generation of complete miR-21 KO mouse (GL261 and CT2A) and human (U87) glioma cells.

miR-21 KO did not affect Vmp1 expression
The genome-editing system caused an INDEL in the miR-21 sequence in an intron of the Vmp1 gene. To evaluate whether this affected the functionality of this gene, we first analyzed levels of mRNA expression using primer sets targeting different parts of the Vmp1 gene ( Figure S6A). The qRT-PCR results showed similar expression levels compared with GL261 WT cells, assuming that the sequence of the Vmp1 gene was not affected by the CRISPR editing ( Figure S6B). This was confirmed by both western blot analysis and immunohistochemistry of VMP1 protein expression, which did not show any difference in protein levels or intracellular location between WT and KO cells ( Figures S6C-S6E). Together, these results confirm that miR-21 KO has no off-target effects on the Vmp1 gene.

miR-21 KO resulted in upregulation of miR-21 regulated mRNA targets
We performed RNA sequencing (RNA-seq) to compare GL261 miR-21 KO1-3 with miR-21 WT and analyzed the 59 predicted miR-21 mRNA targets in the transcriptome of the GL261 cells. In total, 25 genes were upregulated and 27 genes were downregulated (Table 1) in miR-21 KO cells (significantly changed genes shown in red) compared with miR-21 WT cells (adjusted p < 0.1) ( Figure 2A). Next, we evaluated the expression of previously identified targets of murine miR-21. 21 Cdc25a and Cxcl10 were significantly upregulated (shown in boldface) based on differential expression analysis between miR-21 KO and miR-21 WT GL261 cells ( Figure 2B). Next, we restored the level of miR-21 in miR-21 KO cells by stable transduction using a lentiviral vector expressing miR-21 ( Figure S7), showing similar levels of miR-21 compared with GL261 WT ( Figure 2C). Using further qRT-PCR analysis we  validated a number of genes and found that those anti-proliferation genes-cell division cycle 25 homolog A (Cdc25a), C-X-C motif chemokine ligand 10 (Cxcl10), Krev interaction trapped protein 1 (Krit1), signal transducer and activator of transcription 3 (Smad7), signal transducer and activator of transcription 3 (Stat3), and cyclindependent kinase 6 (Cdk6)-were differentially expressed to a significant level when comparing miR-21 KO with the cells where we restored miR-21 ( Figure 2D). In brief, RNA-seq analysis screening downstream miR-21 targets shows that the lack of miR-21 results in enhanced levels of a number of miR-21 regulated mRNAs. This was validated by using qRT-PCR analysis and by overexpressing miR-21 in miR-21 KO cells, thereby rescuing the levels of miR-21 and restoring the target levels of these mRNAs to equality with the WT cells.

Reduced proliferation, migration, and invasion of miR-21 KO cells
In many cancers, upregulation of miR-21 is linked to increased cell proliferation. 22 Therefore, we compared the functional effect of WT and KO cells were measured using the WST reduction assay. Cell viability was measured every 24 h for 5 days. The results display the ratio of miR-21 KO compared with WT cells. miR-21 WT cells proliferated at a significantly higher rate compared with the miR-21 KO cells. Data represent triplicates and are presented as the mean with SEM (error bars). **p < 0.01, ***p < 0.001, ****p < 0.0001, multi-comparison two-way ANOVA. (C) Proliferation rates of U87 miR-21 WT and KO cells were measured using the WST reduction assay. Cell viability was measured every 24 h for 5 days. The results display the ratio of miR-21 KO compared with WT cells. miR-21 WT cells proliferated at a significantly higher rate compared with the miR-21 KO cells. Data represent triplicates and are presented as the mean with SEM (error bars). **p < 0.01, ***p < 0.001, ****p < 0.0001, multi-comparison two-way ANOVA. (D) Proliferation rates of GL261 miR-21 WT, KO-3, and the KO-3 restored cell were measured using the WST reduction assay. Cell viability was measured every 24 h for 5 days. The results display the ratio of GL261 miR-21 KO and the GL261 miR-21 restored cells compared with GL261 WT cells. WT and restored cells proliferated at a significantly higher rate compared with the miR-21 KO-3 clone. Data represent triplicates and are presented as the mean with SEM (error bars). **p < 0.01, ***p < 0.001, ****p < 0.0001, multi-comparison two-way ANOVA. (E) Representative images and quantification of colony-formation assay show a significantly lower number of colonies grown of the CT2A, U87, and GL261 miR-21 KO cells and larger colonies for miR-21 restored cells, as compared with the WT cells. miR-21 KO shows a significantly lower number of colonies compared with WT cells. ***p < 0.001, ****p < 0.0001, one-way ANOVA. Data represent three independent experiments and are presented as the mean with SEM (error bars) (F and G) Representative images and quantification of cell migration and invasion measured over 24 h across a transwell membrane with 8 mm pores with (invasion) or without (migration) Matrigel. Both migration (F) and invasion (G) ability of CT2A, U87, and GL261 miR-21 KO cell lines were significantly reduced compared with the miR-21 WT cells. GL261 KO cells were significantly reduced compared with GL261 miR-21 restored cells (****p < 0.0001, multicomparison two-way ANOVA). Histograms show the number of cells that grew out as clones and number of cells that crossed the cell membrane. Experiment was repeated three times, and data are presented as the mean with SEM (error bars). ***p < 0.001, ****p < 0.0001, one-way ANOVA. miR-21 KO using in vitro proliferation, migration, and invasion assays ( Figure 3A). Proliferation rates of GL261, CT2A, and U87 were measured comparing miR-21 WT, KO, and GL261 miR-21restored cells using the WST-1 reduction assay. Proliferation was significantly decreased in miR-21 KO in both GL261, CT2A, and U87 cells, as compared with miR-21 WT and restored cell lines measured over a 5-day period ( Figures 3B-3D). A colony-formation assay was performed to evaluate cell survival based on the ability of single cells to grow into colonies. GL261, CT2A, and U87 miR-21 KO cells produced significantly fewer and smaller colonies compared with the WT cells ( Figure 3E). Although the number of the GL261 restored colonies did not differ significantly compared with the GL261 WT colonies, the colony sizes showed enhanced proliferation in cells when the miR-21 was restored. The miR-21-restored cells had significantly more colonies compared with the miR-21 KO cells (Figure 3E, right panel). Interestingly, CT2A cells showed increased migration and invasion over GL261 cells, with a 2-fold higher degree of migration (p < 0.0020) and invasion (p < 0.0043) compared with the GL261. CT2A, U87, and GL621 miR-21 KO cells also showed reduced migration and invasion capacity compared with WT cells, while miR-21-restored GL261 cells (right panel) showed significantly more migrating cells compared with the GL261 miR-21 KO and WT cells ( Figures 3F and 3G). In sum, the depletion of miR-21 reduced glioma proliferation, migration, and invasion in vitro.

miR-21 KO showed reduced tumor growth in mice
To assess the impact of miR-21 KO on overall survival of tumorbearing mice, we injected GL261, CT2A, and U87 miR-21 KO and WT cells intracranially into the striatum of C57BL/6 mice or BALB/C nude mice. Representative In Vivo Imaging System (IVIS) images of the bioluminescence signal showed reduced tumor sizes in mice injected with the GL261 miR-21 KO cells ( Figure 4A). The mice injected with miR-21 KO GL261 cells showed a reduction in tumor growth compared with the mice injected with GL261 miR-21 WT cells, with a significant difference at day 20 ( Figure 4B). The miR-21 KO GL261 cell injected mice had a marked increase in survival compared with mice injected with GL261 miR-21 WT cells.
Mice harboring the GL261 miR-21 WT tumor showed a median survival of 21 days after tumor cell implantation, whereas GL261 miR-21 KO injected mice lived for at least 60 days in good health ( Figure 4C). When GL261 miR-21 KO injected mice were sacrificed, no tumors were observed based on GFP-positive tumor cells compared with GL261 miR-21 WT injected mice ( Figure 4D). The mice injected with miR-21 KO CT2A cells showed a reduction in tumor growth compared with the mice injected with CT2A miR-21 WT cells (Figure 4E). The miR-21 KO CT2A cell injected mice had an increase in survival compared with mice injected with CT2A miR-21 WT cells. CT2A miR-21 WT mice died with a median survival of 20 days after tumor cell implantation, whereas CT2A miR-21 KO injected mice lived for at least 30 days in good health ( Figure 4F). The mice injected with miR-21 KO U87 cells showed a reduction in tumor growth compared with the mice injected with U87 WT cells, with significant differences at days 18 and 21 ( Figure 4G). The miR-21 KO U87 cell injected mice had increased survival compared with mice injected with U87 WT cells. U87 WT mice died with a median survival of 21 days after tumor cell implantation, whereas U87 miR-21 KO injected mice lived for at least 25 days in good health ( Figure 4H). To summarize, our data show that miR-21 KO in both mouse and human glioma lines hampered tumor growth and increased overall survival in mice.

DISCUSSION
We successfully knocked out miR-21 in glioma using the CRISPR technique, as confirmed by non-detectable miR-21 expression levels and NGS. This resulted in decreased proliferation, migration, and invasion of these cells in vitro, as well as increased survival of tumorbearing mice in vivo. The cell lines used in this study differed in their growth rate, migration, and invasiveness. The CT2A glioma cell line presents a more aggressive form of glioma. Its higher levels of migration and invasion (as shown in Figure 3) result in faster tumor spread causing the overall survival to be reduced compared with GL261 cells (as shown in Figure 4). Interestingly, even without achieving 100% KO (GL261 KO1: 73%; KO2: 46%; CT2A KO2: 98%; KO3: 44%) based on the genomic integrity of the loci, miR-21 expression levels were markedly reduced, resulting in a functional effect on proliferation in all clones. The pri-miRNA expression did not show differences in levels on comparing miR-21 KO with WT, indicating that the transcription of primary miRNA was not impeded, but its processing and/or maturation to the mature miR-21 was disrupted due to gene editing. Furthermore, our results show that expression of the Vmp1 gene, in the intron of which miR-21 coding sequences are located, was not changed upon CRISPR editing of the miR-21 genomic sequences, and no off-target activity in sequence-predicted regions of the genome was detected.
Dysregulated expression of miRNAs in cancer cells can act as either a tumor suppressor or oncogene, resulting in inhibition or promotion of tumor development, respectively. 23 In many cancers, including glioma, miR-21 is highly abundant and functions as an oncogene regulating many different mRNA targets. 24 Using RNA-seq we screened the miR-21 targets, previously described in the literature, which have been found to be directly regulated by miR-21 in a variety of different cell types. 21 Using this initial screen, we identified specific miR-21 targets affected by the KO of miR-21 in glioma cells. Here we found several genes that are involved in cell proliferation and apoptosis. First, Cdc25a is overexpressed in a wide variety of cancers and has a crucial role in cell proliferation and apoptosis and, when overexpressed, promotes tumorigenesis. 25 For example, the tumorigenic effects of Cdc25a are displayed in the transformation of primary mouse embryonic fibroblasts. Here, together with expression of H-RasG12V or in the absence of RB1 these primary mouse embryonic fibroblasts will form tumors. 26 Counterintuitively, in colon cancer and glioblastoma, Cdc25a upregulation after inhibition of miR-21 resulted in decreased rather than increased tumor proliferation. 27,28 In addition, it was shown that Cdc25a has pro-apoptotic properties. 29 Here, caspase cleavage results in a catalytically active fragment that is localized to the nucleus where it induces apoptosis. 29,30 This indicates that the subcellular localization of Cdc25a is an important indicator for its effect on cell fate. 31 Possibly, the abundance of Cdc25a after miR-21 KO results in nuclear localization and increase in apoptosis in glioma cells. 27 A similar paradox can be found for Cdk6, which is also involved in cell-cycle regulation. mTOR inhibition and subsequent miR21 upregulation was shown to repress Cdk6 in T cell acute leukemia, resulting in decreased tumor proliferation. 32 Cdk6 protein levels are elevated in glioma, but this only leads to tumor progression after post-translational modification by SUMOylation, which stabilizes and enhances the kinase activity driving the cell cycle. 33 Like Cdc25a, the Cdk6 level dictates the miR-21 effect on cells. For example, high levels of Cdk6 have been shown to induce apoptosis rather than proliferation in Kaposi's sarcoma. 34 Thus, for cell-cycle regulators (Cdc25a and Cdk6) the cellular state or exposure to drugs in combination with how these proteins are post-translationally modified dictates their overall effect on proliferation and apoptosis. In this context it is important to note that prior to targeting miR-21, evaluation is needed as to whether this miRNA acts as a tumor suppressor or an oncogene depending on the mRNA targets that are important within certain cell types or under specific physiological conditions, such as after DNA damage due to irradiation. Also, Krit1 is a gene with a putative role in tumorigenesis, and miR-21 overexpression or silencing Krit1 mRNA levels increases tumor growth. 35 However, the function of this gene is not well documented, and it is not known whether the miR-21 KO leading to overexpression of Krit1 can tip the scale toward apoptosis, as has been shown for Cdc25a and Cdk6. 29,34 More clear-cut is the expression trend and consequences of Cxcl10 increase after miR21 KO. Cxcl10 is reported to be a negative regulator of cell growth in glioblastoma. 36,37 Elevated levels of Cxcl10 mRNAs are thus consistent with the observed reduction in miR-21 KO glioma cell proliferation. Moreover, Smad7 is a key regulator of the transforming growth factor b1, which suppresses the migration and invasion capacity of glioma cells. 38 Thus, increased levels of Smad7 mRNA can reduce the capacity of miR-21 KO cells to migrate. In addition, Stat3, a gene involved in cell proliferation and invasion, is activated in multiple cancers, 39 including glioma. It was shown that Stat3 and miR-21 closely interact, 40 and the complex interplay between non-coding RNAs (including miR-21) and Stat3 related to its function in glioma has recently been thoroughly reviewed. 41 Moreover, miR-21 downregulates human telomerase reverse transcriptase expression through its action on Stat3, inhibiting glioma cell growth. 42 Therefore, glioma cells lacking miR-21 might upregulate the Stat3 pathways involved in tumor progression. The KO of miR-21 resulting in elevation of mRNAs important in controlling the cell-cycle machinery can cause cell-cycle arrest and promotes apoptosis, together resulting in reduced cell proliferation observed in the miR-21 KO glioma cells.
The discrepancy between miR-21-KO-induced upregulation of genes known to be involved in tumorigenesis and tumor growth inhibition may reflect the ability of miRNAs to buffer the effect of variations in gene expression. 43 In glioma miR-21 levels are overexpressed, causing downregulation of mRNAs associated with controlling cell proliferation. However, in our miR-21 KO glioma models the feedback networks of miR-21 to regulate downstream mRNAs are lacking completely; this appears to collectively destabilize gene expression that, in the context of glioma, leads to decreased proliferation and invasion, as we observed in the subset of miR-21 targets we analyzed in this study. However, we should note that miR-21 dynamically regulates a network of multiple targets, many of them involved in tumor progression or apoptosis, and in our study we focused only on a subset of targets that might not represent the complete function of miR-21. Therefore, future studies will be needed to better understand the downstream pathways and mechanisms responsible for hampered tumor growth caused by miR-21 KO.
While the expression of Sox2 was previously shown to be affected by miR-21, in our gene-expression analysis using RNA-seq and qRT-PCR (data not shown) we did not find any significant changes in Sox2 mRNA. This was surprising, since this gene has been found to be targeted by miR-21 in other cell types, i.e., human mesenchymal cells, 44 neural crest stem cells, 45 and embryonic stem cells. 46 A possible explanation is that Sox2 is so highly expressed in glioma 47,48 and that miR-21 knockdown does not lower levels substantially. In our dataset, Sox2 is expressed in the top 500 (406 of 18,491) genes in GL261. Thus, the absence of miR-21 might not have a marked effect on this level of overexpression.
In prior studies, miR-21 was inhibited by transfection with either 2V-O-methyl-miR-21 or LNA/DNA-miR-21 to elucidate the functional target mRNAs of miR-21. 4,49 Inhibiting miR-21 has been shown to be efficient in reducing tumor size and eventually inhibiting tumor www.moleculartherapy.org progression in multiple cancer types, including breast cancer, 50 prostate cancer, 51 and pancreatic cancer. 52 However, the binding affinity between miRNA and those inhibitors was relatively weak and limited to short-term studies due to irreproducibility after transfection. The use of sponges seemed to overcome this pitfall by showing longterm stability; however, the miRNA sponge was found to be vulnerable to Ago2-mediated endonucleolytic cleavage, which decreased its efficiency. 14 The CRISPR system has been proved to be a successful genomic tool to KO specific miRNAs in vitro and in vivo to understand the function of specific miRNAs (i.e., miR-2188, miR-137, miR-182) in multiple different cancer types, including carcinoma, leukemia, and ovarian cancer. 14,[53][54][55][56] Others have used the CRISPR approach to target miR-21 in various human cell lines such as HEK293, colon cancer HCT-116, prostate cancer LNCaP, and breast cancer MCF-7 cell lines, 57 and in human colon cancer cell lines RKO and DLD1. 28 More recently, four different lentiviral vectors containing CRISPR components were used to target the pre-miR-21 hairpin structure at the DNA level, inducing several mutations in ovarian cancer cell lines. This resulted in upregulation of target mRNAs, including those for programmed cell death protein 4 and Sprouty RTK signaling antagonist 2, leading to reduced proliferation as well as increased migration and invasion. 58 The potential of miR-21 as a promising therapeutic candidate in various types of cancer has been vigorously studied and extensively reviewed. 59 Encouragingly, an interventional clinical trial is currently being conducted to evaluate six miRNAs (including miR-21) to define high-and low-risk colon cancer patients and determine which patients should receive adjuvant chemotherapy (ClinicalTrials.gov: NCT02 466113). Moreover, a clinical trial aimed at suppressing miR-21 in patients with Alport syndrome is under way, where RG-012, a chemically modified oligonucleotide that can bind to miR-21, is administered (ClinicalTrials.gov: NCT03373786). Although anti-miR-21 is an appealing cancer therapy, including for glioma, 15 further studies are needed to validate its success in large-scale studies. Therefore, a complete understanding of the complex mechanisms regulating the interaction between this miRNA and its targets is needed. Finally, other issues faced by miRNA therapeutics are the potential off-target toxicity and long-term effects. Optimistically the approach presented here, which is based on gene editing, avoids the side effects of current available strategies to inhibit miR-21 (e.g., antisense oligonucleotide inhibitors, locked nucleic acid-modified small interfering RNAs, and miRNA sponges) and should make it possible to safely and effectively target miR-21 in tumor cells in vivo for sustained effect.
It is collectively known that elevated miR-21 is involved in tumor progression. As extensively reviewed, miRNA inhibitors are increasingly being implemented in clinical trials. 60 However, current strategies to inhibit miR-21, such as antisense oligonucleotides, 61 in combination with S-TRAIL, 62 are limited due to their side effects and short halflife. 63 In our study, we used CRISPR technology so as not only to suppress or temporarily knock down the function of miR-21 but to completely remove miR-21 from glioma cells. These established miR-21 KO glioma cell lines allow further elucidation of the function of miR-21 in the context of glioma and also its effect on the tumor microenvironment. Now that we show the striking effect of loss of miR-21 in glioma cells on tumor progression, the next goal should be to focus on how to safely and efficiently deliver the CRISPR constructs, specifically targeting miR-21 selectively at the tumor site. We hypothesize that this could be done with adeno-associated virus (AAV) gene delivery, using AAV capsid variants which preferentially transduce glioma cells, 64,65 although other carriers (such as lentiviral vectors) can also be used to deliver the Cas protein and the single guide RNA (sgRNA) to the tumor. 66 In this study, we show that effective CRISPR KO targeting of miR-21 in mouse gliomas results in reduction of tumor growth. We identified miR-21 as an important target for miRNA-based therapeutics in glioma. Our study paves the way for the advancement of RNA-based therapy in glioma through targeting oncogenic miRNAs. Future studies could apply our approach to other cancers and other miRNAs that are correlated with tumor progression in the context of glioma or other tumor types, such as elevated miR-10b. 67 Although miR-21based biomarkers and therapeutics seem very plausible for future use, further large-scale validation studies will be necessary to reveal the clinical significance of miR-21 targeting, and further research is needed to focus on the delivery of these CRISPR therapeutics in combination with other treatment regimens.

Contact for reagent and resource sharing
Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Erik Abels (e.r.abels@lumc.nl).

Experimental model and subject details Animals
All animal experiments were conducted under the oversight of the Massachusetts General Hospital Institution Animal Care and Use Committee (IACUC). C57BL/6J, B6;129S6-Mir21atm1Yoli/J (miR-21 À/À ) mice 20 and CAnN.Cg-Foxn1nu/Crl (BALB/c) nude mice from Charles River Labs (IACUC protocol 2009N000054) were maintained with unlimited access to water and food under a 12-h/12-h light/dark cycle. To study tumor proliferation in vivo, we used a total of 20 C57BL/6J adult male mice randomly assigned to each group.

Cell culture
The National Cancer Institute (NCI) provided glioma cells (GL261, CT2A, and U87). All cells were cultured at 37 C in a 5% CO 2 humidified incubator. GL261 cells were cultured in medium containing Roswell Park Memorial Institute (RPMI) 1640 with L-glutamine (Corning, Glendale, AZ). CT2A and U87 cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Corning) supplemented with penicillin (100 units/mL)-streptomycin (100 mg/mL) (P/S) (Corning) and 10% fetal bovine serum (FBS) (Gemini Bioproducts, West Sacramento, CA). Cells were tested negative for mycoplasma contamination at periodic intervals throughout the study (Mycoplasma PCR Detection Kit G238; ABM, Richmond, BC, Canada). For the downstream mRNA expression analysis, both GL261 WT and GL261 miR-21 KO cells were stably transduced with a lentivirus vector to co-express the inserted miR-21 sequence and GFP under a cytomegalovirus (CMV) promoter (LentimiRa-GFP-mmu-mir-21 vector, mm10221; ABM). A schematic vector map is shown in Figure S7. Following transduction, cells were sorted based on the expression of GFP. miR-21 expression levels were measured using the TaqMan assay (000397; Applied Biosystems, Beverly, MA) to determine successful transduction.

Method details Intracranial tumor injection
Adult mice were anesthetized using 2.5% isoflurane in 100% oxygen via a nose cone. A total of 1 Â

Transfection
Glioma (GL261, CT2A, and U87) cells were transfected using Lipofectamine 2000 (Thermo Fisher Scientific, Cambridge, MA) in Opti-MEM using the manufacturer's protocol. In detail, 1 Â 10 6 glioma cells were plated in each well of 6-well plates (Corning) and transfected with plasmid DNA containing the CRISPR plasmids: 1 mL of Cas12a (RTW2896) (5 mg/mL) and 1 mL of pUC19-U6-AsCa-s12a_crRNA-miR-21-site3 (JIR327) (5 mg/mL). Transfection mixes were incubated at room temperature for 30 min before adding to the cells. Cells were subsequently incubated for 6 h with the transfection mix at 37 C. Transfection efficiency was monitored by level of GFP and mCherry fluorescence after 24 h, with corresponding microscopy images acquired on a Zeiss Axio Imager M2 (Carl Zeiss, Peabody, MA).

Fluorescence-activated cell sorting
First, to test the editing efficiency of the different sgRNAs, cells were sorted in bulk after transfection. Cells transfected with different plasmids were stained with 4 0 ,6-diamidino-2-phenylindole (DAPI; Invitrogen, Waltham, MA) at final concentration of 1 mg/mL. Live cells were sorted by selecting only DAPI-negative cells using a BD FACSAria II SORP Cell Sorter (Biosciences, Woburn, MA). Glioma cells co-transfected with GFP (CAS12a plasmid) and mCherry (sgRNA plasmid) were sorted by FACS. Gates were set to only sort cells positive for both GFP and mCherry. Live cells were plated as single cells in 96-well plates containing 100 mL of complete culture medium to select single clones. Collected cells were grown over a period of 2-3 weeks. After expanding the colonies, different aliquots were frozen in RPMI medium with 20% FBS and 10% DMSO and stored at À80 C.

Primary astrocyte cultures
Mixed glial cultures were isolated from cerebral cortices of P1 to P4 C57BL/6J or miR-21 KO mouse pups. 20 Meninges were removed, and cortical cells were dissociated using 0.05% trypsin/EDTA (Corning) followed by single-cell suspension preparation using 100 mm and 40 mm cell strainers (BD Falcon). Cells were cultured in DMEM with 20% FBS, 1% P/S, and 10 ng/mL M-CSF (Gibco, Waltham, MA) on poly-D-lysine (PDL; Sigma-Aldrich, Germany; 10 mg/mL) pre-coated T-75 culture flasks for 10-15 days. Primary   (Table S1). All primers were developed using primer-blast (http://www.ncbi.nlm.nih.gov/tools/primer-blast). After amplification, the PCR products were purified using the Qiagen PCR purification kit. Next, the PCR products were loaded on a 2% agarose gel (Bioresearch, Worcester, MA), and DNA fragments were resolved by gel electrophoresis. In parallel, the purified PCR products were sequenced using the forward primer 5 0 -TTG ACT GCA AAC CAT GAT GCT G-3 0 . To confirm CRISPR activity, we used both Sanger sequencing and NGS analysis, and determined CRISPR efficiency based on CRISPResso2 analysis by using the online tool http:// crispresso2.pinellolab.org. 19 qRT-PCR Total RNA was extracted using the Direct-Zol MicroRNA kit (Zymoresearch, Irvine, CA). For gene-expression analysis using qRT-PCR, cDNA was synthesized from 200 ng of total RNA and prepared using the SuperScript VILO cDNA Synthesis kit (Invitrogen). qRT-PCR analyses were performed using SYBR green (for Vmp1 mRNA expression analysis) ( Table 4) or TaqMan (for (pri)-miRNA and miR-21 downstream mRNA expression analysis). Gene expression was determined using the SYBR green protocol qPCR mix, as prepared following the manufacturing protocol of Power SYBR Green PCR Master Mix (Applied Biosystems, Beverly, MA). The cycling conditions using the fast protocol were 50 C for 2 min, 92 C for 10 min, and 40 cycles of 95 C for 1 s and 60 C for 20 s. The Vmp1 expression was normalized to the housekeeping gene b-Actin.
miRNA levels were analyzed using the TaqMan microRNA assay ( All qPCR reactions were performed using the QuantStudio 3 PCR system (Applied Biosystems) and normalized to the housekeeping gene Gapdh (Mm99999915_g1). The pri-miRNA expression was normalized to the housekeeping gene Hprt1 (Mm00446968). The Vmp1 expression was normalized to the housekeeping gene b-Actin.

Primers
Eleven sets of mmu-Vmp1 primers (Table 4) were developed using primer-blast (http://www.ncbi.nlm.nih.gov/tools/primer-blast) to specifically target the exons of the Vmp1 gene by qRT-PCR as listed below.

Western blot
Total protein was extracted from the cells using RIPA lysis buffer with presence of protease inhibitor cocktail (Sigma-Aldrich). Protein concentration was determined using the Pierce BCA protein assay (Thermo Fisher Scientific). Equal amounts of protein (20 mg) were loaded and resolved on 10% SDS-PAGE gels (Thermo Fisher and were processed using ImageJ 1.49v software. Brain slices on microscope slides (Fisherbrand) were fixed with 4% paraformaldehyde (PFA) for 10 min at room temperature. After fixation, the slices were rinsed with PBS for 5 min and blocked by using 5% BSA and 0.1% PBS-T for 1 h at room temperature. Brain slices were then incubated with the primary antibody GFP anti-mouse (1:400) (Invitrogen) at 4 C overnight. Cells were rinsed three times in PBS-T for 5 min each. Secondary antibody goat anti-mouse (1:400) (Invitrogen) was diluted in PBS-T and incubated for 1 h in the dark at room temperature. Slides were then mounted with DAPI (Vectashield; Vector Labs). Fluorescence microscopy images were acquired on a Keyence (Itasca, IL) microscope and processed using ImageJ 1.49v software.

RNA sequencing
The RNA concentration and integrity (RIN score) were determined using the Agilent 2100 Bioanalyzer Pico-chips (Agilent Technologies, Lexington, MA) according to the manufacturing protocol. RNA libraries were prepared from the extracted RNA using QuantSeq 3 0 mRNA-seq library prep kit for Illumina (Lexogen, Vienna, Austria). Library amplification and library barcoding were performed using the i7 Index Plate for QuantSeq for Illumina barcodes 7001-7096 (Lexogen), and 13 cycles of library amplification were completed. Equal molar individual libraries were pooled, and the pool concentration was determined using the SYBR FAST Universal qPCR Kit. Finally, libraries were diluted and denatured with the addition of 1% PhiX Sequencing Control V3 (Illumina, San Diego, CA). Single reads were generated using MiniSeq High Output kits (75 cycles) on a MiniSeq (Illumina).

Cell proliferation assay
Cell proliferation was assessed in vitro by the WST reduction assay to determine cell viability (cell counting kit-8 [CCK-8]; Dojindo, Rockville, MD). Cells were seeded at a low density (2 Â 10 3 cells/well) in a 96-well plate. After 24 h the medium was changed, and 10% WST solution was added to the cells. The cells were incubated at 37 C for 1 h, and fluorescence at wavelength 450 nm was measured using a microplate reader (SynergyH1; BioTek, Winooski, VT). Thereafter, the medium was changed and the cells were measured again 24 h later up until 80% confluency at day 3.

Cell migration and invasion
Cell migration and invasion assays 68 were performed in vitro using Boyden chamber transwell membranes with a permeable polycarbonate membrane containing 8 mm pores separating the two chambers (Corning). For the invasion assay, the membranes were first coated with 500 mL of Matrigel (Corning) diluted in serum-free medium (1:5) at 37 C for 1 h. One milliliter of migration buffer (medium containing 10% FBS) was added to the lower chamber and incubated for 30 min at 37 C. A total of 1.8 Â 10 4 cells was plated in the upper chamber in 0.5% FBS and incubated for 24 h at 37 C. After 24 h   the transwell was removed from the plate, 750 mL of 70% ethanol was added per well in a 24-well plate, and the transwell was placed into the well to fixate the cells which had migrated through the membrane and attached to the transwell. The transwell was removed and dried for 10 min. Next, 750 mL of 0.2% crystal violet (Sigma-Aldrich) was added to the well and incubated for 30 min. Migrated cells were imaged under an inverted microscope and counted using ImageJ.

Colony formation
Colony-formation assay was performed as described previously. 69 Cells were plated at low density ($100 cells per well) in 6-well plates and allowed to grow into colonies for 10-14 days. Cells were then fixed for 5 min in 4% PFA and stained with 0.05% crystal violet (Sigma-Aldrich) for 30 min. Cells were washed twice with PBS and imaged to count the colonies by ImageJ.

Data processing and analysis
Raw sequencing data were processed and aligned with the Bluebee platform using the Lexogen QuantSeq 2.3.6 FWD data analysis workflow. Data were aligned to mouse genome GRCm38. After aligning and read counting, differential expression analysis was performed with DESeq2 (version 1.30.1) 70 in R (version 4.0.5). Differential expression analysis, as performed in DESeq2, was subjected to statistical significance using Benjamini and Hochberg multiple testing adjusted p values, with an adjusted p value of <0.05 labeled as significant. Regularized logarithm (rlog) values were used for unsupervised clustering of the most differential expressed genes between samples, and heatmaps were plotted using the pheatmap package (version 1.0.12) in R using scripts previously described 71 . Raw and processed transcriptomic data described in this paper are deposited in NCBI Gene Expression Omnibus (GEO) and are accessible using GEO series accession number GEO: GSE182390.