Targeting non-coding genome for crispr-mediated oncolysis

Abstract

Provided herein are methods of killing cancer cells in an individual comprising delivering to the cancer cells guide RNAs (gRNAs) that specifically target cancer-selective induced sequence repeats present in the cancer cell, and a gRNA-guided nuclease that cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats. Also provided herein are compositions comprising a plurality of gRNAs that specifically target cancer-selective induced sequence repeats and a gRNA-guided nuclease that cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats. Also provided herein are methods of detecting cancer-selective induced sequence repeats in cancer cells in an individual comprising administering to the individual a genotoxic agent in a sub-lethal dose to the cancer cells, and sequencing the cancer cells.

Description

TARGETING NON-CODING GENOME FOR CRISPR-MEDIATED ONCOLYSISCROSS-REFERENCES TO RELATED APPLICATIONS

[0001] This application claims benefit of priority to U.S. Provisional Patent Application No. 63 / 581,757, filed September 11, 2023, which is incorporated by reference for all purposes.STATEMENT AS TO RIGHTS TO INVENTIONS MADE UNDER FEDERALLY SPONSORED RESEARCH AND DEVELOPMENT

[0002] This invention was made with government support under ESI R35 GM143124, R00 GM 118909, and K99 GM 118909 awarded by the National Institutes of Health. The government has certain rights in the invention.BACKGROUND

[0003] Primary glioblastoma (GBM) is a highly aggressive tumor of the central nervous system (CNS) that remains challenging to treat. Despite multimodal treatment regimens, median survival is 12-15 months. GBMs are highly diffuse and infiltrate normal brain parenchyma, rendering complete surgical resection improbable. Growth of residual tumor inevitably occurs, mostly near the resection cavity and with therapy resistance, ultimately resulting in death. Bulk and single-cell sequencing revealed that GBMs exhibit extensive intratumoral heterogeneity, with subpopulations of cells displaying distinct mutations, copy number aberrations, gene expression patterns, and epigenetic states. This heterogeneity renders therapies targeted at specific molecular processes ineffective as it enables tumor recurrence from minor clones with distinct genetic makeups.

[0004] Methylation status of the O-6-methylguanine-DNA methyltransferase (MGMT) promoter determines sensitivity to the alkylating agent temozolomide (TMZ), the current frontline chemotherapy for GBM. While TMZ -based therapies show relatively few side effects and extend survival, most patients experience disease progression. At recurrence, 23% of MGMT-silenced GBM patients present with hypermutated tumors that correlate with dismal survival (see, e.g., Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas, Nature 580, 517-523 (2020)). This percentage is higher for patients with recurrentastrocytomas and oligodendrogliomas. TMZ increases the rate of somatic mutation; coupled with tumor genomic instability and loss of DNA mismatch repair (MMR) pathways, this results in hypermutated tumors. Critically, no effective therapies currently exist against hypermutated gliomas. Hence, there is a need for treatment strategies that can effectively eliminate GBM, and other hypermutated cancer cells irrespective of their mutational and epigenetic profile.

[0005] Clustered regularly interspaced short palindromic repeats (CRISPR) - CRISPR- associated (Cas) methods, derived from bacterial and archaeal adaptive immune systems, enable RNA-guided introduction of DNA double-strand breaks (DSBs) at target loci throughout the genome. Paired with the rapid expansion of reference and personalized genome information, CRISPR technologies enable vast genetic manipulations, even in previously difficult contexts, including human cells. One potential caveat of CRISPR therapies are cytotoxic effects due to the generation of DNA DSBs. Hence, CRISPR may be used to induce DSBs to kill cancer cells.SUMMARY

[0006] Relevant to cancers, for example but not limited to CNS cancer, it has been shown that cytotoxic effects of DNA DSBs can be harnessed as a therapeutic strategy against cancers, for example, recurrent hypermutated gliomas, by targeting unique repeat sequences in tumor genomes, a concept deemed CRISPR-Cas “Cancer Shredding” herein. Some embodiments in which the repeat thresholds useful for abolishing cancer cells using Cas9 targeted against noncoding elements are provided. Additionally, TMZ-induced hypermutation in recurrent GBM is shown to lead to unique, cancer-specific repeat sequences. CRISPR targeting of these sequences leads to selective depletion of a patient-derived recurrent GBM line, while sparing normal cells. Thus, in some embodiments, the disclosure provides for combining local delivery of Cancer Shredding with immuno-oncology approaches to stimulate the tumor immune environment and synergistically enhance the effect of both approaches. The present disclosure presents Cancer Shredding, an innovative treatment paradigm independent of a tumor’s genetic and epigenetic origin that converts the tumor mutational burden (TMB) and TMZ or other genotoxic agent signature of hypermutated cancers into a possible avenue for treatment.

[0007] Provided herein are methods of killing cancer cells in an individual having cancer cells comprising a plurality of copies of cancer- selective induced sequence repeats (e.g., non-coding sequence repeats, coding sequence repeats or both) in a genome of the cancer cells, the method comprising delivering to the cancer cells: (i) guide RNAs (gRNAs) that specifically target thecancer-selective induced sequence repeats; and (ii) a gRNA-guided nuclease that cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats, thereby selectively killing the cancer cells.

[0008] In some embodiments, the cancer-selective induced sequence repeat has at least five copies in the genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeats have between 5 to 100 copies in the genome of the cancer cells. In some embodiments, the method further comprises sequencing the genome of the cancer cells prior to and / or after the delivering, thereby identifying the cancer-selective induced sequence repeats.

[0009] In some embodiments, the method further comprises, prior to the delivering: (i) sequencing non-cancerous cells of the individual; (ii) identifying cancer-selective induced sequence repeats in a genome of the cancer cells not substantially present in non-cancerous cells; and (iii) selecting gRNAs that specifically target the cancer-selective induced sequence repeats not substantially present in non-cancerous cells. In some embodiments, the cancer-selective induced sequence repeats are caused by a genotoxic agent. In some embodiments, the method further comprises administering the genotoxic agent to the individual having the cancer cells prior to the delivering. In some embodiments, the method further comprises sequencing the genome of the cancer cells after administering the genotoxic agent to the individual having the cancer cells and prior to the delivering, thereby identifying cancer-selective induced sequence repeats. In some embodiments, the genotoxic agent is administered to the individual having the cancer cells at a sub-lethal dose to the cancer cells. In some embodiments, genotoxic agent is a chemotherapeutic agent, for example temozolomide, cisplastin, paclitaxel, or 5 -fluorouracil. In some embodiments, the genotoxic agent is radiation.

[0010] In some embodiments, the delivering to the cancer cells comprises systemically administering the gRNA and gRNA-guided nuclease to the individual having the cancer cells. In some embodiments, the delivering to the cancer cells comprises locally administering the gRNA and gRNA-guided nuclease to the cancer cells. In some embodiments, the gRNA and gRNA- guided nuclease are delivered to the cancer cells in one or more vectors. In some embodiments, the one or more vectors are viral vectors, for example lentiviral vectors or retroviral vectors having or lacking independent replicative capacity. In some embodiments, the gRNA is delivered to the cancer cells in a replicating retroviral vector and the gRNA-guided nuclease is delivered to the cancer cells in a non-replicating retroviral vector.

[0011] In some embodiments, the cancer cells are primary cancer cells. In some embodiments, the cancer cells are metastatic cancer cells. In some embodiments, the cancer cells are glioblastoma cancer cells. In some embodiments, the gRNA guided nuclease is a Cas protein.

[0012] Also provided herein are compositions comprising a plurality of gRNAs that specifically target cancer-selective induced sequence repeats and a gRNA-guided nuclease that cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats (e.g., non-coding sequence repeats, coding sequence repeats or both). In some embodiments, the cancer-selective induced sequence repeat has at least 5 copies in the genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeat has between 5 to 100 copies in the genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeats are identified by sequencing the genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeats are caused by a genotoxic agent. In some embodiments, genotoxic agent is a chemotherapeutic agent, for example temozolomide, cisplastin, paclitaxel, or 5 -fluorouracil. In some embodiments, the genotoxic agent is radiation. In some embodiments, the cancer cells are primary cancer cells. In some embodiments, the cancer cells are metastatic cancer cells. In some embodiments, the cancer cells are glioblastoma cancer cells. In some embodiments, the gRNA guided nuclease is a Cas protein. Also provided are one or more vectors encoding the compositions described above or elsewhere herein.

[0013] Also provided herein are methods of detecting cancer-selective induced sequence repeats (e.g., non-coding sequence repeats, coding sequence repeats or both) in cancer cells in an individual comprising administering to the individual a genotoxic agent in a sub-lethal dose to the cancer cells, and sequencing the cancer cells. In some embodiments, the sequencing is performed prior to and / or after the administering. In some embodiments, the cancer-selective induced sequence repeats are caused by a genotoxic agent. In some embodiments, genotoxic agent is a chemotherapeutic agent, for example temozolomide, cisplastin, paclitaxel, or 5- fluorouracil. In some embodiments, the genotoxic agent is radiation. In some embodiments, the cancer-selective induced sequence repeat has at least 5 copies in the genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeat has between 5 to 100 copies in the genome in the cancer cells.

[0014] In some embodiments, the method further comprises (i) sequencing non-cancerous cells of the individual prior to and / or after the administering; (ii) identifying the cancer-selectiveinduced sequence repeats in a genome of the cancer cells not substantially present in non- cancerous cells; and (iii) selecting guide RNAs (gRNAs) that specifically target the cancer- selective induced sequence repeats not substantially present in non-cancerous cells. In some embodiments, the cancer cells are primary cancer cells. In some embodiments, the cancer cells are metastatic cancer cells. In some embodiments, the cancer cells are glioblastoma cancer cells.DESCRIPTION OF THE DRAWINGS

[0015] FIG. 1 A depicts a schematic of glioblastoma (GBM) disease progression with frontline therapy including radiation, temozolomide chemotherapy (TMZ), and surgical resection. Highlighted is the concept of CRISPR-Cas “Genome Shredding”, an alternative therapeutic strategy for recurrent tumors that induces DNA damage and cell elimination through targeting of repetitive elements in the genome.

[0016] FIG. IB is a graph showing quantification of protein coding and non-coding variants in a patient’s primary and recurrent GBM compared to the patient’s native genome. Variants were detected using Mutect2 based on whole-genome sequencing. Blood was used for the patient’s native control.

[0017] FIG. 1C is a Venn diagram showing the patient’s variants that are reported in ClinVar (NCBI) as risk factor, likely pathogenic, or pathogenic. Each section of the Venn diagram reports the number of coding and non-coding mutations. Select variants are highlighted. Notably, the recurrent tumor possesses numerous mutations not present in the primary GBM.

[0018] FIG. ID, left panel is a Venn diagram of the patient’s collection of all possible sgRNAs, termed the “sgRNA-ome”. FIG. ID, right panel is a graph showing the number of sgRNAs with the indicated amount of repeat targets (20-mer sequences, 5’-NGG-3’ PAM) in the patient’s native and cancer genomes.

[0019] FIG. IE is a violin plot of the repeat sgRNA-ome target counts in the human genome (hg38). Only sgRNAs with at least 10 target sites are represented.

[0020] FIG. IF depicts the target distribution of highly repetitive sgRNAs, termed “sgCIDE”, between coding and non-coding genome. The analysis was run for all sgCIDEs with at least 1,000 target repeats in the human genome. The inlay shows a zoom-in, highlighting that there are two discrete key classes of sgCIDEs.

[0021] FIG. 1G shows Circos plots of select sgCIDEs indicating the location of target sites across the genome. Each line of the inner circle represents one target locus. The outer circle highlights the chromosomes. Only fully complementary target sites (Hamming distance = 0) are shown. Target sequence and PAM (blue) are indicated. The target loci are equally spread throughout the genome with few exceptions of specific cytobands.

[0022] FIG. 2A is a diagram of a competitive proliferation assay to quantify cell depletion rates using flow cytometry. Lentiviral vector maps are shown.

[0023] FIG. 2B are graphs showing changes in ratios of sgRNA-transduced cells (mNeonGreen+) monitored by flow cytometry over seven days in a competitive proliferation assay in Cas9 expressing U251 (top panel) and LN229 (bottom panel) GBM cells. Wildtype cells not expressing Cas9 were used for normalization. The cell lines were stably transduced with the indicated guide RNAs inducing Genome Shredding (sgCIDE), targeting an essential gene (sgRPAl), or representing a non-targeting control (sgNT). Error bars indicate the SD (n=3).

[0024] FIG. 2C is a schematic of a Comet assay to assess genomic degradation at single-cell resolution.

[0025] FIG. 2D are images of a Comet assay demonstrating rapid and robust Genome Shredding at the single-cell level. Cas9 expressing U251 and LN229 cells were transduced with the indicated sgRNAs and genome fragmentation assessed at 24 hours post-transduction. Select representative images are shown.

[0026] FIG. 2E shows the quantification of genome fragmentation observed in the Comet assay in FIG. 2D.

[0027] FIG. 2F are growth curves showing the number of mNeonGreen+ cells (per 20X image) detected by real-time quantitative live-cell imaging (Incucyte) of Cas9 expressing U251 transduced with the indicated guide RNAs. The quantification indicates the average count per condition across n=4 images.

[0028] FIG. 3A shows quantification of cell viability after treatment of temozolomide (TMZ) sensitive and resistant glioblastoma cells with TMZ or Genome Shredding guide RNAs (sgCIDEs). Cas9 expressing U251 and T98G cells were subjected to TMZ (concentration gradient) or various sgRNAs (viral dilution gradient, 1 :X). Cell viability was quantified using a luminescent cell viability assay (CellTiter-Glo) at the indicated days after treatment initiation or transduction.

[0029] FIG. 3B is a schematic of propidium iodide (PI) based cell cycle analysis using flow cytometry.

[0030] FIG. 3C are scatter plots showing flow cytometry cell cycle analysis of TMZ and sgCIDE treated GBM cells. Cas9 expressing U251 cells were treated with TMZ (50 pM) or the indicated sgRNAs and cell cycle profiles analyzed after PI staining.

[0031] FIG. 3D are graphs showing the quantification of Sub-Gl fraction in cell cycle profiles shown partially in FIG. 3C. Cas9 expressing U251 and T98G cells were treated with DMSO, TMZ (50 pM) or the indicated sgRNAs, and stained with PI after one to five days. Error bars indicate the SD. Note, while T98G were resistant to TMZ, Genome Shredding was equally effective in U251 and T98G.

[0032] FIG. 3E are graphs quantification of cell viability 5 days after treating glioblastoma stem-like cells (GSCs) with TMZ (50 pM) or all-in-one lentiviral vectors (pCF827) to deliver Cas9 and sgRNAs (sgNT-1 / 2 / 3, sgCIDE- 1 / 2 / 3).

[0033] FIG. 4A is a schematic showing retreatment of GBM cells that have escaped a first cycle of Genome Shredding. During initial treatment (colony formation assay), GBM cells are consecutively transduced with a lentiviral vector expressing Cas9 and a puromycin selection marker (pCF226), followed by a vector expressing an sgRNA and mNeonGreen (pCF821). Rare “escapee” cells that survive this regimen (PuroR, mNeonGreen+) are then expanded and treated a second time. The retreatment consists either of a vector expressing only an sgRNA and mCherry (pCF820), or an all-in-one vector expressing an sgRNA and Cas9 linked to an mCherry marker (pCF826).

[0034] FIG. 4B shows colony formation in Cas9 expressing U251 and T98G. Cells were transduced with lentiviral vectors expressing the indicated sgRNAs and seeded at low density (100, 1,000, 10,000 cells) in 6-well plates. Controls were treated with DMSO or TMZ (50 pM). Cells were stained with crystal violet two weeks later.

[0035] FIG. 4C is graph showing quantification of cell depletion of the indicated U251-Cas9 escapee clones and control lines treated with an all-in-one vector (pCF826) expressing the indicated sgCIDE or non-targeting control guide RNAs and Cas9 linked to mCherry. sgCl, sgCIDE-1. sgC2, sgCIDE-2. Renewed expression of Cas9 and sgCIDEs led to effective retreatment of all escapee clones. For three-out-of-four clones, re-expression of Cas9 alone wassufficient for cell depletion, indicating that the previously introduced sgCIDEs were still active in those clones.

[0036] FIG. 5A is a schematic representation of experimental design for an in vivo assay.

[0037] FIG. 5B is a graph showing a sub-Gl assay confirming that concentrated sgCIDE-1 virus efficiently transduces and destroys Cas9-expressing LN229 cells in vitro. Error bars indicate the SD (n=3).

[0038] FIG. 5C is a graph showing in vivo competitive proliferation in mice harboring LN229 GBM intracranial xenografts treated with sgRNA virus. Mouse brains were collected when mice showed tumor-related symptoms, dissociated with the papain dissociation system, and analyzed by flow cytometry. The percentage of mCherry (tumor cells) and mNeonGreen (sgRNA virus) positive cells from mice treated with 5 pl concentrated sgNT-1 or sgCIDE-1 virus (4 mice each) and 15 pl concentrated sgNT-1 or sgCIDE-1 virus (6 and 5 mice, respectively) were quantified. The absence of mNeonGreen+ tumor cells in sgCIDE-1 treated mice showed that Genome Shredding efficiently eliminated transduced GBM cells in vivo. Significance was determined using the Mann-Whitney test. Data presented as mean ± SD. *, p<0.05. **, p<0.01

[0039] FIG. 5D are Kaplan-Meier curves of mice harboring LN229 GBM intracranial xenografts treated with 5 pl concentrated sgNT-1 or sgCIDE-1 virus (solid blue and orange lines, respectively) and 15 pl concentrated sgNT-1 or sgCIDE-1 virus (dashed blue and orange lines, respectively). Significance was determined using the log-rank test.

[0040] FIG. 5E is a graph showing flow cytometry analysis of wild-type (WT) Cas9 and TiCas9 edited HEK293T cells encoding GFP (HEK-RT1). HEK-RT1 cells were transduced with WT Cas9 (pCF226) or a tamoxifen-inducible Cas9 variant (TiCas9, pCF287), selected on puromycin (1.0 pg / ml), and further transduced with the indicated sgRNAs (pCF221; sgNT-1 / 2, sgGFP-8 / 9). Cells were then treated with DMSO (0.1%) or 4-OHT (1 pM in 0.1% DMSO, changed every day) for three days, followed by 24 h of doxycycline treatment (1 pg / ml) to induce GFP expression. Editing efficiency was quantified by flow cytometry of GFP expression. While no editing was observed in DMSO controls, addition of 4-OHT to TiCas9 expressing samples led to substantial GFP editing with sgGFP-8 / 9, though not fully matching the efficiency of WT Cas9.

[0041] FIG. 5F is a graph showing a competitive proliferation assay in TiCas9 expressing U251 GBM cells. U251-TiCas9 were stably transduced with the indicated sgRNAs (sgNT-1 / 2 / 3,sgCIDE- 1 / 2 / 3) and treated with 4-OHT (1 pM, renewed every day). Changes in ratios of sgRNA-transduced cells (mNeonGreen+) were monitored by flow cytometry over 5 days. Untreated cells (pre-treatment) were used for normalization. Error bars indicate the SD (n=3).

[0042] FIG. 5G are Kaplan-Meier curves of mice harboring TiCas9-sgRNA expressing U251 GBM intracranial xenografts. TiCas9 expressing U251 cells were transduced with a Genome Shredding guide RNA (sgCIDE-1) or a non-targeting control (sgNT-1) before injecting into the mouse brain. After one week, tumor-bearing mice were injected intraperitoneally (IP) with 100 mg / kg tamoxifen (dissolved in corn oil at 20 mg / ml) per day for 5 consecutive days (day 8 / 9 / 10 / 11 / 12) to induce TiCas9 expression. Survival benefits were quantified. Significance was determined using the log-rank test.

[0043] FIG. 6A is a schematic depicting an assay to evaluate CRISPR target requirements for efficient cell ablation by Genome Shredding. Cell depletion efficiencies of sgCIDEs were quantified with competitive proliferation assays in eukaryotic cells (human, mouse, chicken) having varying amounts of target sites within native chromatin.

[0044] FIG. 6B is a graph showing competitive proliferation assays in Cas9 expressing human (U251) and mouse (GL261) GBM cells, and chicken fibroblasts (DF1). Wildtype cells not expressing Cas9 were used for normalization. The cell lines were stably transduced with the indicated sgRNAs. Changes in ratios of sgRNA+ cells were monitored by flow cytometry (mNeonGreen+). Shown is the relative ratio at day 7 post-transduction.

[0045] FIG. 7A is a schematic showing identification of cancer-specific sgRNAs and assessment of Cancer Shredding in a patient-derived cell line (PDCL) of the recurrent GBM.

[0046] FIG. 7B is a graph showing the mutational signature of the primary and recurrent GBMs. Indicated is the frequency of variants in the primary and recurrent GBM after normalization to the patient’s native genome. The recurrent tumor shows a clear C>T mutational signature (TMZ-signature) characteristic of hypermutated GBM after TMZ treatment.

[0047] FIG. 7C is a Venn diagram showing the sgRNA-ome of the patient’s recurrent GBM. The mutational landscape and corresponding set of cancer-specific sgRNAs (sgRNA-ome) was determined in the patient’s recurrent GBM tissue as well as in two technical replicates of the recurrent GBM PDCL. Numbers indicate the amount of sgRNAs that are cancer-specific, i.e. that have at least one mismatch in the PAM-proximal eight nucleotides when compared to the patient’s native and reference genomes.

[0048] FIG. 7D shows identification of mismatch location of cancer repeat sgRNAs when compared to the patient’s native and reference genomes.

[0049] FIG. 7E shows enrichment and depletion of sgRNAs from a CRISPR screen in NHA control and SF11411 PDCL cells. The sgRNAs were ranked based on the log(2) fold-change (FC) of their representation from day 28 (D28) post-transduction compared to day 1 (TO). Specific subsets of sgRNAs, including sgCIDEs, cancer repeat sgRNAs, and non-target sgRNAs, are highlighted to facilitate visualization.

[0050] FIG. 7F shows quantification of cancer-specific sgRNA enrichment and depletion in NHA control versus SF11411 PDCL cells using MAGeCK RRA. Shown is the log(2)FC of normalized sgRNA representation between the two cell lines at the end of the screen (D28).

[0051] FIG. 7G shows validation of cancer-specific sgRNAs enabling TMZ-signature targeted Cancer Shredding. Statistical analysis of significantly depleted (FDR<0.05, MAGeCK RRA - gene level) sgRNAs. Cancer repeat sgRNAs showed the strongest cancer-specific depletion across the entire CRISPR library.

[0052] FIG. 7H is a schematic illustrating potential use cases of CRISPR Cancer Shredding. Cancer Shredding provides an innovative therapeutic strategy for hypermutated gliomas, and potentially other hypermutated cancers, that leverages cancer-specific mutations for targeted, CRISPR-mediated tumor cell ablation. Targetable mutations include therapy-induced TMZ signature mutations. sgHYP, hypermutated cancer-specific sgRNAs. sgTMZ, TMZ signature mutations-specific sgRNAs. sgREC, recurrent cancer-specific sgRNAs.

[0053] FIG. 8A shows distribution of tumor-specific mutations per genome-wide megabase (Mb). Violin plots showing the tumor mutational burden (TMB) of a patient-derived primary and recurrent GBM genome. The data was normalized to the patient’s native (blood) genome. Because of the logarithmic scale, mutation counts are shown as n+1. The recurrent GBM has significantly more mutations (p<0.01, Wilcoxon Rank Sum).

[0054] FIG. 8B shows TMB in coding genome. Mutations per exon for each exon of the primary and recurrent GBM genomes, showing the distribution of mutations affecting coding elements. Mutation counts are shown as n+1. The recurrent GBM has significantly more coding mutations (p<0.01, Wilcoxon Rank Sum).

[0055] FIG. 8C is a scatter plot Top repeat sgRNAs enable simultaneous targeting of multiple genes. The plot shows the number of exons targeted by each sgCIDE with at least 1,000 target sites in the human genome.

[0056] FIG. 8D is a Circos plots showing the genomic location (hg38) of select sgCIDE target sites. Giemsa-stained chromosome bands (UCSC Genome Browser, Chromosome Band Track) are indicated. Chromosomes are displayed in a clockwise fashion with the short arm (p) first. Note, CRISPR-Cas Genome Shredding target sites are spread throughout the entire genome with few exceptions corresponding to specific chromosome bands (cytobands). These cytobands include the variable heterochromatic regions on chromosomes 1, 9 and 16, the short arms of the acrocentric chromosomes 13, 14, 15, 21 (partially) and 22, the variable region at the q-terminus end of chromosome Y, and the pericentromeric regions on each chromosome. Many of these cytobands correspond to predominantly unsequenced regions represented by large gaps in the human reference genome.

[0057] FIG. 9A are graphs showing a competitive proliferation assay in U251 (left panel) and NHA (right panel) cells treated with all-in-one Cas9-sgRNA lentivirus (pCF827). The indicated cell lines were stably transduced with lentivirus to express Cas9, indicated sgRNAs, and mNeonGreen. Changes in ratios of transduced cells (mNeonGreen+) were analyzed by flow cytometry at day 2, 4, and 6 post transduction. Error bars indicate the SD (n=3).

[0058] FIG. 9B are representative real-time quantitative live-cell microscopy images of Cas9 expressing U251 cells transduced with the indicated guide RNAs. Cells transduced with the sgRNA expression vector (pCF821) are mNeonGreen+. Propidium iodide (red) was used to stain apoptotic cells.

[0059] FIG. 10A shows a luminescent cell viability assay (CellTiter-Glo) of temozolomide (TMZ) sensitive and resistant GBM cells treated with the indicated conditions. Cell lines stably expressing Cas9 were generated by transduction with the lentiviral vector pCF226 (U-251, LN- 229, T98G) or pCF823 (LN- 18).

[0060] FIG. 10B shows quantification of cell cycle analysis from temozolomide (TMZ) sensitive and resistant GBM cells treated with the indicated conditions. Cell lines stably expressing Cas9 were generated by transduction with the lentiviral vector pCF226 (U-251, LN- 229, T98G) or pCF823 (LN- 18).

[0061] FIG. 11 A is a schematic illustrating the generation and use of a CRISPR-resistant viral packaging cell line termed “CRISPR-Safe”. HEK293T cells were transduced with a lentiviral vector (pCF525-AcrIIA4) stably expressing the anti-CRISPR protein AcrIIA4 that inhibits Streptococcus pyogenes Cas9. Use of the resulting CRISPR-Safe packaging cell line enables high-titer production of all-in-one Cas9-sgCIDE viral particles.

[0062] FIG. 1 IB are graphs showing that use of the CRISPR-Safe viral packaging cell line rescues viral titers of all-in-one Cas9-sgCIDE vectors. Parental U251 cells (U251-pCF226- pCF821-sgNT-l #1) and U251 cells stably expressing AcrIIA4 (pCF525-AcrIIA4) were transduced with all-in-one lentiviral vectors (pCF826) expressing an mCherry-tagged Cas9 and the indicated sgRNAs. Viral particles were produced either using standard HEK293T packaging cells or the CRISPR-Safe packaging cell line. Viral titers were assessed by flow cytometry-based quantification of mCherry expression at day two post-transduction. Note, use of the CRISPR- Safe packaging cell line nearly completely rescued viral titers of all-in-one Cas9-sgCIDE vectors. Error bars indicate the SD (n=3).

[0063] FIG. 11C shows quantification of cell depletion of the indicated U251-Cas9 escapee clones and control lines secondarily treated with a lentiviral vector (pCF820) expressing various sgCIDE or non-targeting control guide RNAs and an mCherry fluorescence marker. sgCl, sgCIDE-1. sgC2, sgCIDE-2. Expression of new sgCIDEs did not affect cell proliferation of escapee clones.

[0064] FIG. 1 ID shows quantification of cell depletion of the indicated LN229-Cas9 escapee clones and control lines treated with a lentiviral vector (pCF820) expressing various sgCIDE or non-targeting control guide RNAs and an mCherry fluorescence marker. sgCl, sgCIDE- 1. sgC2, sgCIDE-2. Expression of new sgCIDEs did not affect cell proliferation of escapee clones.

[0065] FIG. 1 IE shows re-treatment of LN229-Cas9 escapee clones with an all-in-one vector (pCF826) expressing the indicated sgCIDE or non-targeting control guide RNAs and Cas9 linked to mCherry. sgCl, sgCIDE-1. sgC2, sgCIDE-2. Note, renewed expression of Cas9 and sgCIDEs led to effective retreatment of all escapee clones. For all four clones, re-expression of Cas9 alone was sufficient for cell depletion, indicating that the previously introduced sgCIDEs were still active.

[0066] FIG. 12A are violin plots of the repeat sgRNA-ome target counts (20-mer sequences, 5’-NGG-3’ PAM) in the mouse (mm 10) and chicken (galGal6) genomes. Only sgRNAs with at least 10 target sites are represented.

[0067] FIG. 12B are graphs showing competitive proliferation assays in Cas9 expressing human (U251) and mouse (GL261) GBM cells, and chicken fibroblasts (DF1). Wildtype cells not expressing Cas9 were used for normalization. The cell lines were stably transduced with the indicated sgRNAs. Changes in ratios of sgRNA+ cells were monitored by flow cytometry (mNeonGreen+) at the indicated days post-transduction. This is extended data from (Fig. 5b). Note, the most highly repetitive sequences are related to transposable elements that were either species-specific or present across mammals (human, mouse). Repetitive sequences with internal k-mer repeats (repeat expansion) can serve as efficient pan-vertebrate (human, mouse, chicken) targets for Genome Shredding.

[0068] FIG. 12C are graphs showing determination of DSB requirements for induction of cell death. Relative cell death was calculated as the ratio of cell death when normalized to the least and most efficient sgRNAs in the set. Shown is the relative cell death as a function of the total number of target sites that are fully complementary (HD 0) or have one mismatch (HD 0-1). For logarithmic representation, target site counts are given as n+1.

[0069] FIG. 13 A is a graph showing the mutational landscape of the patient’s genome and GBMs. Shown is the probability of variants seen in the patient’s native genome as well as primary and recurrent GBM. Mutation calling was done using Mutect2 against the human reference genome.

[0070] FIG. 13B is a graph showing Genome Shredding efficiently ablates recurrent hypermutated GBM cells. A PDCL (SF11411) was established from the patient’s recurrent hypermutated GBM and transduced with Cas9 (pCF823). Wildtype cells not expressing Cas9 were used for normalization. Cells were then transduced with the indicated sgRNAs. Changes in ratios of sgRNA-transduced cells (mNeonGreen+) were monitored by flow cytometry. Error bars indicate the SD (n=3).

[0071] FIG. 13C is a graph showing enrichment and depletion of sgRNAs from a CRISPR screen in NHA control and SF11411 PDCL cells. Shown is the log(2) fold-change (FC) of sgRNA representation from day 28 (D28) post-transduction compared to day 1 (TO). Different subsets of sgRNAs in the CRISPR library are highlighted, for NHA and SF11411 cells. Whilemost sgCIDEs depleted in both cell lines, as expected, cancer repeat sgRNAs depleted exclusively in SF11411 and were found enriched in NHAs.DEFINITIONS

[0072] As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural reference unless the context clearly dictates otherwise.

[0073] The use of any and all examples or exemplary language (e.g., “for example” or “such as”) provided herein, is intended merely to better illustrate the invention, and does not pose a limitation on the scope of the invention unless otherwise claimed.

[0074] The terms “may,” “may be,” “can,” and “can be,” and related terms are intended to convey that the subject matter involved is optional (that is, the subject matter is present in some examples and is not present in other examples), not a reference to a capability of the subject matter or to a probability, unless the context clearly indicates otherwise.

[0075] The use herein of the terms "including," "comprising," or "having," and variations thereof, is meant to encompass the elements listed thereafter and equivalents thereof as well as additional elements. Embodiments recited as "including," "comprising,” or "having" certain elements are also contemplated as "consisting essentially of and "consisting of’ those certain elements. As used herein, “and / or” refers to and encompasses any and all possible combinations of one or more of the associated listed items, as well as the lack of combinations where interpreted in the alternative (“or”).

[0076] The term “genotoxic agent” herein refers to substances that causes or increases the frequency of DNA or chromosomal damage. The term “genotoxic agent”, as used herein, encompasses mutagenic agents (i.e., substances which cause or increase the frequency of mutations). Generally, genotoxic agents herein may include chemotherapies and radiation. More specifically, genotoxic agents as used herein may include physical mutagens (e.g., ionizing radiation or ultraviolet radiation), alkylating agents or alkylating-like agents (e.g., cyclophosphamide or cisplatin), base analogs, intercalating agents (e.g., daunorubicin or doxorubicin), platinum-containing crosslinking agents, topoisomerase II inhibitors (e.g., etoposide or doxorubicin), antimetabolite drugs (e.g., 5 -fluorouracil or gemcitabine), or antimicrotubule drugs (e.g., paclitaxel).

[0077] The term “nucleic acid” or “nucleotide” refers to deoxyribonucleic acids (DNA) or ribonucleic acids (RNA) and polymers thereof, herein, “polynucleotides,” in either single- or double-stranded form. The nucleic acid molecule and polynucleotides may be derived from a variety of sources, including DNA, cDNA, synthetic DNA, RNA, or combinations thereof. Such nucleic acid sequences may comprise genomic DNA which may or may not include naturally occurring introns. Moreover, such genomic DNA may be obtained in association with promoter regions, introns, or poly A sequences.

[0078] The term “gene” or “transgene” can refer to the segment of DNA (e.g., a polynucleotide sequence) involved in producing or encoding a polypeptide chain. It may include regions preceding and following the coding region (leader and trailer) as well as intervening sequences (introns) between individual coding segments (exons). Alternatively, the term “gene” or “transgene” can refer to the segment of DNA involved in producing or encoding a nontranslated RNA, such as an rRNA, tRNA, guide RNA (e.g., a single guide RNA), or micro RNA.

[0079] The term “genome” is used to refer to the entire body of genetic material (i.e., nucleic acids or polynucleotides as described above) contained in each cell of an organism. As used herein, a “gene” refers to a defined region that is located within a genome and that comprises other, primarily regulatory, nucleic acid sequences responsible for the control of the expression, that is to say the transcription and translation, of the coding portion. Genes can include both coding and non-coding regions (e.g., introns, regulatory elements, promoters, enhancers, termination sequences and 5' and 3' untranslated regions). A gene typically expresses mRNA, functional RNA, or specific protein, including regulatory sequences. Genes may or may not be capable of being used to produce a functional protein. In some embodiments, a gene refers to only the coding region, (i.e., the region which ultimately results in protein expression).

[0080] As used herein, a “cell” can be in vivo, ex vivo or in vitro, and includes any eukaryotic cell. In some embodiments, the cell is a mammalian cell. In some embodiments, the cell is a human cell.

[0081] As used herein, the term “selective” or “selectively” in the context of killing cancer cells refers to an interaction between gRNAs, gRNA-guided nucleases, and cancer cells which results in killing cancer cells more often than non-cancerous cells. In some embodiments, 2X, 3X, 5X, 10X, 20X, 100X or more cancer cells are killed compared to non-cancerous cells.

[0082] The “CRISPR / Cas” system refers to a widespread class of bacterial systems for defense against foreign nucleic acid. CRISPR / Cas systems are found in a wide range of eubacterial and archaeal organisms. CRISPR / Cas systems include type I, II, and III sub-types. Wild-type type II CRISPR / Cas systems utilize an RNA-mediated nuclease, a Cas, in complex with guide RNAs (gRNAs) and activating RNAs to recognize and cleave foreign nucleic acid. gRNAs having the activity of both a gRNA and an activating RNA are also known in the art. In some cases, such dual activity guide RNAs are referred to as small guide RNAs (sgRNAs).

[0083] As used herein, the term “Cas” refers to CRISPR associated proteins that are RNA- mediated nucleases (e.g, of bacterial or archeal orgin, or derived therefrom). Exemplary Cas types include but are not limited to Cas9 proteins and homologs thereof or Cas 12a proteins and homologs thereof. A Cas may be wildtype or in other instances, the Cas is non-natural, artificial, engineered, synthetic, rationally designed, or man-made. A Cas used herein may be mutated (e.g., dCas9 or lbCasl2a RuvC mutant). A Cas may be a nuclease (an enzyme that cleaves both strands of a double-stranded nucleic acid), a nickase (an enzyme that cleaves one strand of a double-stranded nucleic acid), or a catalytically inactive (or dead) Cas fused to a nuclease (e.g., a Type IIS restriction endonuclease, such as FokI ). By way of example, a Cas having nuclease or nickase activity is referred to as a “catalytically active Cas.” A Cas lacking the ability to, on its own (i.e., without fusion to a nuclease), cleave or nick target nucleic acid is referred to as a “catalytically inactive Cas9 ” or a “dead Cas” (e.g., a dCas9 or lbCasl2a RuvC mutant).

[0084] The Cas may be derived from any species known in the art. For example, Cas 12a may be from Acidaminococcus sp. and Lachnospiraceae bacterium (“lbCasl2a”). Cas9 may be derived from bacteria of the following taxonomic groups: Actinobacteria, Aquificae, Bacteroidetes-Chlorobi , Chlamydiae-Verrucomicrobia, Chlroflexi, Cyanobacteria, Firmicutes, Proteobacteria, Spirochaetes, and Thermotogae. An exemplary Cas9 protein is the Streptococcus pyogenes Cas9 protein. Additional Cas9 proteins and homologs thereof are described in, e.g., Chylinksi, et al., RNA Biol., 10(5): 726-737 (2013); Nat. Rev. Microbiol., 9(6): 467-477 (2011); Hou, et al., Proc Natl Acad Sci USA., 110(39): 15644-9 (2013); Sampson et al., Nature, 497(7448):254-7 (2013); and Jinek, et al., Science, 337(6096):816-21 (2012). The Cas nuclease domain can be optimized for efficient activity or enhanced stability in the target cell.

[0085] As used herein, the phrase “delivering” in the context of delivering a nucleic acid or a viral vector refers a cell refers to the introducing into the cell the nucleic acid or viral vector. Asused herein “introducing” refers to contacting of the cell with the nucleic acid or viral vector such that translocation of the nucleic acid sequence or viral vector from outside a cell to inside the cell occurs. In some cases, introducing refers to transducing or infecting a cell or a population of cells with a viral vector or viral particle carrying one or more non-viral nucleic acids. In some cases, translocation of the nucleic acid from outside the cell to inside the nucleus of the cell occurs. Various methods of such translocation are contemplated, including but not limited to, viral infection, electroporation, transfection, transduction, contact with nanowires or nanotubes, receptor mediated internalization, translocation via cell penetrating peptides, liposome mediated translocation, and the like.

[0086] A “promoter” is defined as one or more a nucleic acid control sequences that direct transcription of a nucleic acid. As used herein, a promoter includes necessary nucleic acid sequences near the start site of transcription, such as, in the case of a polymerase II type promoter, a TATA element. A promoter also optionally includes distal enhancer or repressor elements, which can be located as much as several thousand base pairs from the start site of transcription.

[0087] A promoter is “operably linked” when it is placed into a functional relationship with another nucleic acid sequence, for example, a polynucleotide encoding a polypeptide provided herein. For example, a promoter is operably linked to a polynucleotide if it affects, either positively or negatively, the transcription of the polynucleotide.

[0088] The term "vector," as used herein, refers to a nucleic acid molecule capable of propagating another nucleic acid to which it is linked. The term includes the vector as a selfreplicating nucleic acid structure as well as the vector incorporated into the genome of a target cell into which it has been optionally introduced. A “vector” as used here refers to a recombinant construct in which a nucleic acid sequence of interest is inserted into the vector that is not normally present in the viral vector (i.e., a polynucleotide heterologous to the viral vector). In general, the heterologous nucleic acid, for example, a nucleic acid encoding a heterologous polypeptide, is flanked by at least one, and generally by two, long terminal repeat sequences (LTRs), for example, a 5’ LTR and a 3’LTR.

[0089] As used herein, "a viral vector" refers to a gene therapy vector used to deliver a polynucleotide construct to a cell. It is understood that the term viral vector encompasses recombinant vector particles or virions (i.e., viral particles comprising at least one capsid orenvelope protein and an encapsidated recombinant viral vector) and recombinant vector plasmids.

[0090] As used herein, “retroviruses” are RNA viruses wherein the viral genome is RNA. When a host cell is infected with a retrovirus, the genomic RNA is reverse transcribed into a DNA intermediate which is integrated very efficiently into the chromosomal DNA of infected cells. The integrated DNA intermediate is referred to as a provirus. Retroviruses are enveloped single-stranded RNA viruses that typically infect mammals, such as, for example, bovines, monkeys, sheep, and humans, as well as avian species. Retroviruses are unique among RNA viruses in that their multiplication involves the synthesis of a DNA copy of the RNA which is then integrated into the genome of the infected cell. As used herein, “a self-replicating retrovirus” or “replication retroviral vector (RRV)” is a retrovirus encoding all viral proteins necessary for viral replication (i.e., a retroviral gag protein, a retroviral pol protein and a retroviral env protein). The nucleic acid sequence encoding the gag, pol and env proteins are flanked by a 5’ LTR and a 3” LTR. The gag gene encodes the internal structural (matrix, capsid, and nucleocapsid) proteins; the pol gene encodes the RNA-directed DNA polymerase (reverse transcriptase), protease and integrase; and the env gene encodes viral envelope glycoproteins. The 5' and 3' LTRs of the retrovirus serve to promote transcription and polyadenylation of the virion RNAs. Lentiviruses have additional genes including vif, vpr, tat, rev, vpu, nef, and vpx (in HIV-1, HIV-2 and / or SIV). See, also, U.S. Patent Nos. 6,410,313, 6,899,871, and 8,741,279, incorporated herein in their entireties by this reference, for additional information regarding selfreplicating retroviruses.

[0091] The Retroviridae family consists of three groups: the spumaviruses (or foamy viruses) such as the human foamy virus (HFV); the lentiviruses, as well as visna virus of sheep; and the oncoviruses (although not all viruses within this group are oncogenic). The term "retrovirus" is used in its conventional sense to describe a genus of viruses containing reverse transcriptase. Retroviruses include lentiviruses. As used herein, a lentiviral vector is the lentivirus which include the "immunodeficiency viruses" which include human immunodeficiency virus (HIV) type 1 and type 2 (HIV-1 and HIV-2) and simian immunodeficiency virus (SIV). The oncoviruses are further subdivided into groups A, B, C and D on the basis of particle morphology, as seen under the electron microscope during viral maturation. Lentiviruses include non-replicating or replication deficient lentiviruses. In some embodiments, the lentivirus isrendered replication incompetent by modifying the lentivirus to comprise a modified (e.g., a U3 deletion (AU3)), self-inactivating (SCSI) 3’ LTR which renders the resulting lentiviral particles replication incompetent.

[0092] Retroviruses are defined by the way in which they replicate their genetic material. During replication the RNA is converted into DNA. Following infection of the cell a doublestranded molecule of DNA is generated from the two molecules of RNA which are carried in the viral particle by the molecular process known as reverse transcription. The DNA form becomes covalently integrated in the host cell genome as a provirus, from which viral RNAs are expressed with the aid of cellular and / or viral factors. The expressed viral RNAs are packaged into particles and released as infectious virion.

[0093] The retrovirus particle is composed of two identical RNA molecules. Each wild-type genome has a positive sense, single-stranded RNA molecule, which is capped at the 5' end and polyadenylated at the 3' tail. The diploid virus particle contains the two RNA strands complexed with gag proteins, viral enzymes (pol gene products) and host tRNA molecules within a core' structure of gag proteins. Surrounding and protecting this capsid is a lipid bilayer, derived from host cell membranes and containing viral envelope (env) proteins. The env proteins bind to a cellular receptor for the virus and the particle typically enters the host cell via receptor-mediated endocytosis and / or membrane fusion. After the outer envelope is shed, the viral RNA is copied into DNA by reverse transcription. This is catalyzed by the reverse transcriptase enzyme encoded by the pol region and uses the host cell tRNA packaged into the virion as a primer for DNA synthesis. In this way the RNA genome is converted into the more complex DNA genome. The double-stranded linear DNA produced by reverse transcription may, or may not, have to be circularized in the nucleus. The provirus now has two identical repeats at either end, known as the long terminal repeats (LTR). The termini of the two LTR sequences produces the site recognized by a pol product, the integrase protein, which catalyzes integration, such that the provirus is always joined to host DNA two base pairs (bp) from the ends of the LTRs. A duplication of cellular sequences is seen at the ends of both LTRs. Integration is thought to occur essentially at random within the target cell genome. However, by modifying the long terminal repeats it is possible to control the integration of a retroviral genome.

[0094] Transcription, RNA splicing, and translation of the integrated viral DNA is mediated by host cell proteins. Efficient infectious transmission of retroviruses requires the expression on thetarget cell of receptors which specifically recognize the viral envelope proteins, although viruses may use receptor-independent, nonspecific routes of entry at low efficiency. In addition, the target cell type must be able to support all stages of the replication cycle after virus has bound and penetrated.

[0095] As used herein, the retroviral vector can be a derivative of a murine, simian or human retrovirus. Examples of retroviral vectors in which a transgene (e.g., a heterologous polynucleotide sequence) can be inserted include, but are not limited to lentivirus, Moloney murine leukemia virus (MoMuLV), Harvey murine sarcoma virus (HaMuSV), murine mammary tumor virus (MuMTV), Rous Sarcoma Virus (RSV), murine leukemia virus (MLV), Gibbon ape leukemia virus (GALV), Feline Leukemia virus (FeLV), RD114, and xenotropic XMLV.

[0096] “Polypeptide,” “peptide,” and “protein” are used interchangeably herein to refer to a polymer of amino acid residues. As used herein, the terms encompass amino acid chains of any length, including full-length proteins, wherein the amino acid residues are linked by covalent peptide bonds.

[0097] The term “amino acid” refers to naturally occurring and synthetic amino acids, as well as amino acid analogs and amino acid mimetics that function in a manner similar to naturally occurring amino acids. Naturally occurring amino acids are those encoded by the genetic code, as well as those amino acids that are later modified, e.g., hydroxyproline, y-carboxy glutamate, and O-phosphoserine. Amino acid analogs refers to compounds that have the same basic chemical structure as a naturally occurring amino acid, i.e., an a carbon that is bound to a hydrogen, a carboxyl group, an amino group, and an R group, e.g., homoserine, norleucine, methionine sulfoxide, methionine methyl sulfonium. Such analogs have modified R groups (e.g., norleucine) or modified peptide backbones, but retain the same basic chemical structure as a naturally occurring amino acid. Naturally encoded amino acids are the 20 common amino acids (alanine, arginine, asparagine, aspartic acid, cysteine, glutamine, glutamic acid, glycine, histidine, isoleucine, leucine, lysine, methionine, phenylalanine, proline, serine, threonine, tryptophan, tyrosine, and valine) as well as pyrrolysine, pyrroline-carboxy-lysine, and selenocysteine.DETAILED DESCRIPTION

[0098] Provided herein are methods and compositions for killing cancer cells in an individual.Therapies for gliomas have made muted progress over the last decades, with alternative strategiesbeing critically needed. The inventors herein introduce the concept of CRISPR-mediated Genome and Cancer Shredding as an innovative therapeutic paradigm to treat recurrent gliomas. Cancer Shredding leverages CRISPR as anti-cancer modality through targeting of repetitive tumorspecific loci, leading to genome fragmentation and DNA damage-induced cell death. Specifically, the inventors have surprisingly discovered that ablation of cancer cells by Genome Shredding is rapid and efficient, and works independent of a cancer cell’s epigenetic status and sensitivity to frontline chemotherapy. Additionally, while most targeted cells succumb to initial DNA damage introduced by Genome Shredding, the inventors demonstrated that rare escapee clones can effectively be retreated, opening up the possibility for multi-cycle regimens.

[0099] The following description recites various aspects and embodiments of the present compositions and methods. No particular embodiment is intended to define the scope of the compositions and methods. Rather, the embodiments merely provide non-limiting examples of various compositions and methods that are at least included within the scope of the disclosed compositions and methods. The description is to be read from the perspective of one of ordinary skill in the art; therefore, information well known to the skilled artisan is not necessarily included.I. Methods of killing cancer cells

[0100] The methods provided herein comprise delivering to a cancer cell guide RNAs (gRNAs) that specifically target cancer-selective induced sequence repeats (e.g., non-coding sequence repeats, coding sequence repeats or a combination of both) present in a genome of the cancer cells; and (ii) a gRNA-guided nuclease that cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats, thereby selectively killing the cancer cells.

[0101] gRNAs generally have a nucleic acid sequence complementary to the nucleic acid sequence of a target genomic region (e g., the cancer-selective induced sequence repeats present in a genome of cancer cells). gRNAs may have sequences substantially complementary to the nucleic acid sequence of a target genomic region, i.e., having a single nucleic acid mismatch compared to the target nucleic acid sequence. gRNAs may thus confer sequence specificity to nuclease proteins, such as gRNA guided nucleases as used in the methods herein. In some embodiments, 2 or more (e.g., 2, 3, 4, 5, 6, 7, 8, 9 or 10) gRNAs are delivered to the cancer cell, thus targeting 2 or more different nucleic acid sequences (e.g., the cancer-selective inducedsequence repeats present in a genome of cancer cells). In some embodiments, the gRNAs are linked together consecutively and placed under the control of the same promoter, for example, as described in Port et al., Elife, 13(9):e53865 (2020). gRNA, as used herein, also encompasses multiple gRNAs when appropriate (e.g., delivery of a gRNA encompasses delivery of multiple gRNAs).

[0102] As used throughout, gRNA guided nuclease is a protein which has nicking (i.e., cleavage) activity and can induce single- or double-stranded breaks in a target genomic region (e.g., cancer-selective induced sequence repeats in the genome of a cancer cell) that is guided to the target genomic region by a gRNA.

[0103] In some embodiments, the gRNA guided nuclease is a Cas protein. Exemplary Cas proteins include but are not limited to Cas5, Cas6, Cas7, Cas8, Cas9, Casl2a, Casl2b, Casl2i, Casl2j, Casl2L, Casl2e, Casl2c, Casl2d, Casl2g, Casl2h, TnpB, or Casl4 and those described in, for example, Xi and Li, Comput. Struct. Biotechnol. J., 18:2401-2415 (2020). In any case, the Cas used herein has nicking activity, such that when bound to target a genomic region (e.g., cancer-selective induced sequence repeats in the genome of a cancer cell) as part of a complex with a guide RNA, a single strand break or nick is introduced into the target nucleic acid.

[0104] A Cas used herein can be an active variant, or fragment of a wild-type or modified Cas. A Cas can comprise an amino acid change such as a deletion, insertion, substitution, variant, mutation, fusion, chimera, or any combination thereof relative to a wild-type version of the Cas. A Cas can be a polypeptide with at least about 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 91%, 92%, 93%, 94%, 95%, 96%, 97%, 98%, 99%, or 100% sequence identity or sequence similarity to a wild-type exemplary Cas (i.e., the Cas listed above).

[0105] In some embodiments, the Cas used herein is a Casl2a or a Cas9. Casl2a (i.e., Cpfl) is a Cas appearing in many bacterial species, e.g., Acidaminococcus and Lachnospiraceae. Cas9 is derived from Streptococcus pyogenes. In some embodiments, the Cas used herein is Casl4 or Cas-phi. Cas-14 is a Cas appearing in Archaea. Cas-phi is a compact Cas derived from megaphages.

[0106] In some embodiments, a nuclease that is not gRNA-guided is used in the methods herein. The nuclease, for example, may be targeted to the target genomic region by a DNA- binding domain. Thus, in some embodiments, the method of killing cancer cells herein comprises delivering a nuclease targeted to the cancer-selective induced sequence repeats withoutdelivering gRNAs. Discussion herein of gRNAs and gRNA-guided nucleases applies to nucleases delivered without gRNAs when applicable (e.g., delivery methods).

[0107] In some embodiments, the nuclease is a zinc finger nuclease (ZFN). ZFNs typically comprise a zinc finger DNA binding domain and a nuclease domain. Generally, ZFNs include two zinc finger arrays (ZFAs), each of which is fused to a single subunit of a non-specific endonuclease, such as the nuclease domain from the FokI enzyme, which becomes active upon dimerization. Typically, a single ZFA consists of 3 or 4 zinc finger domains, each of which is designed to recognize a specific nucleotide triplet (GGC, GAT, etc.). A ZFN composed of two "3-finger" ZFAs is therefore capable of recognizing an 18 base pair target genomic region (i.e., recognition sequence); an 18 base pair recognition sequence is generally unique, even within large genomes such as those of humans and plants. By directing the co-localization and dimerization of the two FokI nuclease monomers, ZFNs generate a functional site-specific endonuclease that can target a particular locus (e.g., a cancer-selective induced sequence repeat in the genome of a cancer cell, gene, promoter, or enhancer) on the target genomic region.

[0108] Zinc-finger nucleases useful in the methods disclosed herein include those that are known and ZFN that are engineered to have specificity for one or more target genomic regions described herein (e.g., cancer-selective induced sequence repeats in the genome of a cancer cell). Publicly available methods for engineering zinc finger nucleases include: (1) Context-dependent Assembly (CoDA), (2) Oligomerized Pool Engineering (OPEN), (3) Modular Assembly, (4) ZiFiT (internet-accessible software for the design of engineered zinc finger arrays), (5) ZiFDB (internet-accessible database of zinc fingers and engineered zinc finger arrays), and (6) ZFNGenome. For example, OPEN is a publicly available protocol for engineering zinc finger arrays with high specificity and in vivo functionality, and has been successfully used to generate ZFNs that function efficiently in plants, zebrafish, and human somatic and pluripotent stem cells. OPEN is a selection-based method in which a pre-constructed randomized pool of candidate ZFAs is screened to identify those with high affinity and specificity for a desired target sequence. Additionally, ZFNGenome is a GBrowse-based tool for identifying and visualizing potential target genomic regions for OPEN-generated ZFNs. ZFNGenome provides a compendium of potential ZFN target genomic regions in sequenced and annotated genomes of model organisms. ZFNGenome includes more than 11 million potential ZFN target genomic regions, mapped within the fully sequenced genomes of seven model organisms; S. cerevisiae, C. reinhardtii, A.thaliana, D. melanogaster, D. rerio, C. elegans, and H. sapiens. ZFNGenome provides information about each potential ZFN target genomic regions, including its chromosomal location and position relative to transcription initiation site(s). Users can query ZFNGenome using several different criteria (e.g., gene ID, transcript ID, target genomic region sequence).

[0109] In some embodiments, the nuclease used in the methods herein is a is a transcription activator-like effector nuclease (TALEN). Transcription activator-like effectors (TALEs) are proteins secreted by Xanthomonas bacteria and play an important role in disease or triggering defense mechanisms, by binding target DNA and activating effector-specific target cell genes, see, e.g., Gu et al. Nature 435: 1122-5 (2005). A TALEN comprises a TALE DNA-binding domain fused to a DNA cleavage domain (i.e. a nuclease). The DNA binding domain interacts with target DNA in a sequence-specific manner through one or more tandem repeat domains. The repeated sequence typically comprises 33-34 highly conserved amino acids with divergent 12th and 13th amino acids. These two positions, referred to as the Repeat Variable Diresidue (RVD) are highly variable and show a strong correlation with specific nucleotide recognition (Boch et al., Science 326(5959): 1509-12 (2009); and Moscou and Bogdanove, 326(5959):1501 (2009)). This relationship between amino acid sequence and DNA recognition sequence has allowed for the engineering of specific DNA-binding domains by selecting a combination of repeat segments containing the appropriate RVDs.

[0110] The TALE DNA-binding domain can be engineered to bind to a target DNA sequence and fused to a nuclease domain, e.g., a Type IIS restriction endonuclease, such as FokI (see e.g., Kim et al. Proc. Natl. Acad. Sci. USA 93: 1156-1160 (1996)). The nuclease domain can comprise one or more mutations (e.g., FokI variants) that improve cleavage specificity (see, Doyon et al., Nature Methods, 8 (1): 74-9 (2011)) and cleavage activity (Guo et al., Journal of Molecular Biology, 400 (1): 96-107 (2010)). Other useful endonucleases that can be used as the nuclease domain include, but are not limited to, Hhal, Hindlll, Nod, BbvCI, EcoRI, Bgll, and AlwI. In some embodiments, the TALEN can comprise a TAL effector DNA binding domain comprising a plurality of TAL effector repeat sequences that bind to a specific nucleotide sequence (i.e., recognition sequence) in the target genomic region (e.g., cancer-selective induced sequence repeats in the genome of a cancer cell). TALENs function as dimers, and a pair of TALENs, referred to as the left and right TALENs, target sequences on opposite strands of DNA. TALENs are engineered as a fusion of the TALE DNA-binding domain and a monomeric FokI catalyticdomain. To facilitate FokI dimerization, the left and right TALEN target genomic regions are generally selected with a spacing of approximately 14-20 bases.[01111 In some embodiments, the nuclease used in the compositions herein is a meganuclease. Meganucleases generally refer to rare-cutting endonucleases or homing endonucleases that can be highly specific. Meganucleases can recognize DNA target genomic regions ranging from at least 12 base pairs in length. Meganucleases can be modular DNA-binding nucleases such as any fusion protein comprising at least one catalytic domain of an endonuclease and at least one DNA binding domain or protein specifying a nucleic acid target sequence. The DNA-binding domain can contain at least one motif that recognizes single- or double-stranded DNA. A meganuclease can optionally generate a double-stranded break. A double-strand break in DNA can result in DNA break repair which allows for the introduction of gene modification s) (e.g., nucleic acid editing as described in the methods herein). DNA break repair can occur via non -homologous end joining (NHEJ) or homology-directed repair (HDR). In HDR, a donor DNA repair template or template polynucleotide that contains homology arms flanking genomic regions of the target DNA can be provided. The meganuclease can be monomeric or dimeric. In some embodiments, the meganuclease is wildtype type, and in other instances, the meganuclease is non-natural, artificial, engineered, synthetic, rationally designed, or man-made. In some embodiments, the nuclease domain is catalytically inactive. In some embodiments, a meganuclease can bind DNA but cannot cleave the DNA. In some embodiments, the meganuclease of the present disclosure includes an I-Crel meganuclease, I-Ceul meganuclease, I-Msol meganuclease, I-Scel meganuclease, variants thereof, derivatives thereof, and fragments thereof. Detailed descriptions of useful meganucleases and their application in gene editing are found, e.g., in Silva et al., Curr Gene Ther, 11(1): 11-27 (2011); Zaslavoskiy et al., BMC Bioinformatics, 15: 191 (2014), and Takeuchi et al., Proc Natl Acad Sci USA, 111(11):4061-4066 (2014).

[0112] The gRNAs and gRNA guided nucleases used herein may be delivered to the cancer cells through methods known in the art and described elsewhere herein. In some embodiments, the gRNAs and gRNA guided nucleases are systemically administered to the individual having the cancer cells (e.g., by vector or nanoparticle through intravenous injection). In some embodiments, the gRNAs and gRNA guided nucleases are locally administered to the cancer cells (e.g., during surgery).

[0113] In some embodiments, the gRNA and gRNA-guided nuclease are delivered to the cancer cells in one or more vectors (e.g., viral vectors or plasmids). In some embodiments, the gRNA and gRNA-guided nuclease are delivered by the same vector. In some embodiments, the gRNA and gRNA-guided nuclease are delivered by two or more different vectors, e.g., the gRNA is delivered by a first vector and the gRNA guided nuclease (e.g., a Cas) is delivered by a second vector.

[0114] In some embodiments, the gRNA and gRNA-guided nuclease are delivered as one or more viral vectors. Exemplary viral vectors include but are not limited to adenovirus vectors (e.g., Ad2, Ad5, Ad7), adeno-associated viral vectors, herpes simplex viral vectors, retroviral vectors, pox viral vectors (such as vaccinia and avian poxvirus vectors, such as the fowlpox and canarypox vectors), lentiviral vectors, alphavirus vectors, poliovirus vectors, measles vectors, and other positive and negative stranded RNA viruses, viroids, and virusoids, or portions thereof. In some embodiments, the nucleic acid sequence encoding the gRNA and / or gRNA-guided nuclease may be codon optimized for delivery as a viral vector.

[0115] In some embodiments, the retroviral vector is a replicating retroviral vector (i.e., a retroviral vector that has independent replicative capacity). Replicating retroviral vectors have viral genes necessary for replication. Thus, replicating retroviral vectors can replicate and spread in dividing host cells, leading to persistent expression of genetic information contained in the replicating retroviral vector. In some embodiments, the retroviral vector lacks independent replicative capacity, i.e., is a non-replicating retroviral vector. Non-replicating retroviral vectors lack necessary genes for replication, and thus can delivered a larger transgene payload. However, non-replicating replication vectors, due to their lack of independent replicative capacity, lack the ability to spread through a cell (e.g., a cancer cell), thus causing low transgene delivery. In some instances, when two transgene payloads (e g., a gRNA and a gRNA-guided nuclease) are delivered by two vectors, for example, using both a replicating retroviral vector and a nonreplicating retroviral vector can be advantageous for delivery of both transgenes. In some embodiments, the gRNA is delivered to the cancer cells in a replicating retroviral vector and the gRNA-guided nuclease is delivered to the cancer cells in a non-replicating retroviral vector. In some embodiments, the gRNA is delivered to the cancer cells in a non-replicating retroviral vector and the gRNA-guided nuclease is delivered to the cancer cells in a replicating retroviral vector.

[0116] Certain replicating retroviral vectors, e.g., murine leukemia virus derived replicating retroviral vectors, mainly infect replicating cells due to a lack of nucleocapsid nuclear localization signals necessary for transport across nuclear membranes in quiescent cells (see, e.g., Logg et al, Replicating Retroviral Vectors in Cancer, Methods EnzyntoL, 507: 199-228 (2012)). As most cells are quiescent, replicating retroviral vector infection is largely restricted to rapidly proliferating cancer cells. Further, innate and adaptive immune responses against retrovirus infection are inactive in tumor cells. Thus, in some embodiments, the replicating retroviral vector is cancer- selective, i.e., the retroviral vector can substantially only replicate within cancer cells. For example, the gRNA and gRNA-guided nuclease may be delivered systemically to the individual having the cancer cells by one cancer-selective replicating retroviral vector such that the gRNA and nuclease are delivered substantially only to cancer cells.

[0117] In some embodiments, the viral vector is a lentiviral vector. In some embodiments, the lentiviral vector does not integrate viral genetic material into the target genome, i.e., it is a nonintegrating lentiviral vector. In other embodiments, the nucleic acid may be administered as a non-viral vector, including, but not limited to, as a plasmid, in a nanoparticle, (e.g., a lipid nanoparticle), or in a liposome.

[0118] In some embodiments, delivering comprises contacting the cell with nucleic acids encoding the gRNA and gRNA-guided nuclease not contained in a vector (e.g., as RNA), such that the gRNA and gRNA-guided nuclease are expressed in the cell. In some embodiments, delivering comprises contacting the cell with the gRNA and gRNA-guided nuclease in the form of a ribonucleoprotein (i.e., RNA associated with a protein). For example, the gRNAs used herein may form a ribonucleoprotein complex with the gRNA guided nuclease. The ribonucleoprotein complex may be delivered to the cancer cell through any of the methods described herein (e.g., as a polynucleotide or a in vector). In some embodiments, the gRNA and the gRNA guided nuclease forms a ribonucleoprotein, and the ribonucleoprotein is delivered to the cell as a polypeptide (e.g., through electroporation). In any case, Delivering may occur in vitro, ex vivo, or in vivo.

[0119] The methods provided herein comprise delivering to the cancer cell gRNAs that target cancer-selective induced sequence repeats present in the genome of the cancer cells. Genotoxic agents (e.g., chemotherapeutic agents or radiation) increase the rate of somatic mutations incancer cells. Compared to normal cells, cancer cells are typically genomically unstable and highly proliferative, focusing the mutational burden caused by systemic administration of genotoxic agents to cancer cells. The mutational burden from genotoxic agent administration is primarily borne by non-coding regions of the cancer cell genome, which represents the majority of the genome. Cancer-selective induced sequence repeats as used herein refers to such mutations in a genome of cancer cells associated with genotoxic agent administration. In some embodiments, cancer- selective induced sequence repeats are present in the non-coding regions of the genome of the cancer cell. In some embodiments, cancer-selective induced sequence repeats are present in the coding regions of the genome of the cancer cell.

[0120] In some embodiments, a cancer-selective induced sequence repeat has at least five copies in a genome. In some embodiments, the cancer-selective induced sequence repeats has between 5 to 100 copies, for example, 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100 copies. In instances where the gRNAs are fully complementary (i.e., have no mismatches) compared to the cancer-selective induced sequence repeat, fewer copies of the cancer-selective induced sequence repeat may be sufficient to enable gRNA-guided nucleases to induce double stranded breaks in the genome of the cancer cell to result in cancer cell death. For example, in some embodiments, the gRNAs are fully complementary to the cancer-selective induced sequence repeat and the cancer-selective induced sequence repeat has at least 30 (e.g., 31, 32, 33, 34, 35, 36, 37, 38, 39, or 40) copies in the genome of the cancer cell. In instances where the gRNAs are not fully complementary (i.e., have one or more mismatches) compared to the cancer-selective induced sequence repeat, more copies of the cancer- selective induced sequence repeat may be needed to enable gRNA-guided nucleases to induce double stranded breaks in the genome of the cancer cell to result in cancer cell death. For example, in some embodiments, the gRNAs are not fully complementary to the cancer-selective induced sequence repeat, e.g., the gRNAs have a single mismatch compared to the cancer-selective induced sequence repeat and the cancer-selective induced sequence repeat has at least 70 (e.g., 71, 72, 73, 74, 75, 76, 77, 78, 79, or 80) copies in the genome of the cancer cell.

[0121] In some embodiments, the cancer-selective induced sequence repeats comprise complementary nucleic acid sequences to any of the any of the nucleic acid sequences of sgCIDE sequences listed in Tables 2 or 4. In some embodiments, two or more cancer-selectiveinduced sequence repeats may be targeted, e.g., one cancer-selective induced sequence repeat having 10 copies and another cancer-selective induced sequence repeat having 50 copies.

[0122] The methods of killing cancer cells described herein may further comprise sequencing the cancer cells prior to and / or after delivering the gRNAs and gRNA-guided nucleases to the cancer cells, thereby identifying the cancer-selective induced sequence repeats. In some embodiments, the cancer cells are sequenced prior to delivering the gRNAs and gRNA-guided nucleases to the cancer cells. In some embodiments, the cancer cells are sequenced after delivering the gRNAs and gRNA-guided nucleases to the cancer cells. In some embodiments, the cancer cells are sequenced both prior to and after delivering the gRNAs and gRNA-guided nucleases to the cancer cells. In some embodiments, the whole genome of the cancer cells is sequenced. In some embodiments, the genome of the cancer cells is sequenced. Sequencing in some embodiments is from a biopsy of the target tumor or and otherwise representative tumor.

[0123] Sequencing methods useful in the present disclosure include, but are not limited to, shotgun sequencing, bridge PCR, Sanger sequencing (including microfluidic Sanger sequencing), pyrosequencing, massively parallel signature sequencing, nanopore DNA sequencing, single molecule real-time sequencing (SMRT) (Pacific Biosciences, Menlo Park, CA), ion semiconductor sequencing, ligation sequencing, sequencing by synthesis (Illumina, San Diego, Ca), Polony sequencing, 454 sequencing, solid phase sequencing, DNA nanoball sequencing, heliscope single molecule sequencing, mass spectroscopy sequencing, pyrosequencing, Supported Oligo Ligation Detection (SOLiD) sequencing, DNA microarray sequencing, RNAP sequencing, tunneling currents DNA sequencing, and any other DNA sequencing method identified in the future. One or more of the sequencing methods described herein can be used in high throughput sequencing methods. As used herein, the term “high throughput sequencing” refers to all methods related to sequencing nucleic acids where more than one nucleic acid sequence is sequenced at a given time.

[0124] In some embodiments, the methods of killing cancer cells described herein may further comprise, prior to delivering the gRNAs and gRNA-guided nucleases to the cancer cells, sequencing non-cancerous cells of the individual, identifying cancer-selective induced sequence repeats present in a genome of the cancer cells not substantially present in non-cancerous cells; and selecting gRNAs that specifically target the cancer-selective induced-sequence repeats not substantially present in the non-cancerous cells. As used in the methods herein, not substantiallypresent refers to cancer-selective induced sequence repeats present in the genome of cancer cells at least 2 times the amount of cancer-selective induced sequence repeats present in the genome of non-cancer cells. In some embodiments, not substantially present refers to cancer-selective induced sequence repeats present in the genome of cancer cells 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, or 10 times the amount of cancer-selective induced sequence repeats present in the genome of non-cancer cells. Thus, in some embodiments, the gRNAs do not target non-cancerous cells and the gRNA-guided nucleases do not cleave the genome of non- cancerous cells (i.e., the gRNA-guided nuclease selectively kills cancer cells).

[0125] In some embodiments, the cancer-selective induced sequence repeats are caused by a genotoxic agent. In some embodiments, the genotoxic agent is a chemotherapeutic agent. Exemplary genotoxic agents include but are not limited to altretamine, busulfan, carboplatin, crmustine, cisplatin, cyclophosphamide, dacarbazine, ifosfamide, lomustine, melphalan, temozolomide, trabectedin, 5-fluorouracil, 6-mercaptopurine, azacitidine, capecitabine, clofarabine, cytarabine, floxuridine, fludarabine, gemcitabine, methotrexate, pemetrexed, pentostatin, pralatrexate, a combination of trifluridine and tipiracil, vincristine, vinblastine, vinorelbine, paclitaxel (Taxol®), docetaxel, etoposide, teniposide, irinotecan, topotecan, daunorubicin, doxorubicin, doxorubicin liposomal, epirubicin, idarubicin, or valrubicin. In some embodiments, the chemotherapeutic agent is temozolomide, cisplastin, paclitaxel, or 5- fluorouracil. In some embodiments, the chemotherapeutic agent is radiation.

[0126] By way of example, temozolomide (TMZ) has a distinct mutational signature. TMZ is an oral alkylating prodrug used a frontline treatment for many solid tumors, such as gliomas, neuroendocrine tumors, leukemias, colorectal cancers, prostate cancers, breast cancers, and lung cancers. TMZ causes DNA adducts by methylating purine bases of DNA (O6-guanine; N7- guanine and N3-adenine). The cytotoxic O6-methylguanine then mispairs with thymine, inducing a DNA mismatch repair cycle and eventual cell death when O6-methylguanine DNA methyltransferase (MGMT) is absent. Thus a specific base change occurs upon treatment with TMZ (e g., cytosine to thymine conversions) and these changes can be readily identified and specifically targeted in a cancer cell in which they have been induced.

[0127] The method of killing cancer cells provided herein may further comprise administering the genotoxic agent to the individual prior to delivering the gRNA and gRNA-guided nuclease. The genotoxic agent may be administered to the individual by any suitable method, which willbe apparent to one skilled in the art depending on the agent used. In some embodiments, the genotoxic agent is administered to the individual at a dose that is sub-lethal to the cancer cells, i.e., less than the dose required to kill the cancer cells. Thus, the genotoxic agent may be administered to the individual at a dose such that the genotoxic agent causes cancer- selective induced sequence repeats in the genome of the of the cancer cell, but does not kill the cancer cell. For example, a dose of TMZ that is sub-lethal to cancer cells, yet causes cancer-selective induced sequence repeats may be between 1-400 pM.

[0128] In instances where the genotoxic agent is administered to the individual prior to delivering the gRNA and gRNA-guided nuclease, the methods of killing cancer cells provided herein may further comprise sequencing the genome of the cancer cells after administering the genotoxic agent to the individual and prior to delivering the gRNA and gRNA-guided nuclease. Thus, cancer-selective induced sequence repeats may be identified following the administration of the genotoxic agent and used in the engineering of gRNAs targeting cancer-selective induced sequence repeats. In some embodiments, the genotoxic agent is TMZ and the cancer-selective induced sequence repeats are cytosine to thymine substitutions.

[0129] In some embodiments, the cancer cells are brain cancer (e.g., glioma or meningioma), lung cancer (e.g., non-small cell lung cancer), ovarian cancer, prostate cancer, colorectal cancer, liver cancer (i.e., hepatocarcinoma), renal cancer (i.e., renal cell carcinoma), bladder cancer, breast cancer, lung cancer, thyroid cancer, pleural cancer, pancreatic cancer, uterine cancer, cervical cancer, testicular cancer, anal cancer, pancreatic cancer, bile duct cancer, gastrointestinal carcinoid tumors, esophageal cancer, gall bladder cancer, appendix cancer, small intestine cancer, stomach (gastric) cancer, skin cancer, choriocarcinoma, head and neck cancer, osteogenic sarcoma, fibrosarcoma, and melanoma. In some embodiments, the cancer cells are glioblastoma cells. In some embodiments, the cancer cells are primary cancer cells. In some embodiments, the cancer cells are cancer cells that have metastasized to other locations in the body compared to the region of primary cancer.II. Compositions

[0130] Provided herein are compositions comprising a plurality of gRNAs that specifically target cancer-selective induced sequence repeats in a genome of a cancer cell and a gRNA- guided nuclease that cleaves the genome of the cancer cells at the cancer-selective inducedsequence repeats, sparing normal cancer cells. The gRNAs and gRNA-guided nucleases used in the compositions provided herein may be those described above and elsewhere herein. As described above, and elsewhere herein, in some embodiments, the cancer-selective induced sequence repeat has at least 5 copies, or between 5 to 100 copies in the genome in the cancer cell, e.g., 5, 10, 15, 20, 25, 30, 35, 40, 45, 50, 55, 60, 65, 70, 75, 80, 85, 90, 95, or 100 copies, or as otherwise described herein. In some embodiments, the gRNAs are fully complementary to the cancer-selective induced sequence repeat and the cancer- selective induced sequence repeat has at least 30 copies in the genome of the cancer cell. In some embodiments, the gRNAs are not fully complementary, e.g., have a single mismatch, to the cancer-selective induced sequence repeat and the cancer-selective induced sequence repeat has at least 70 copies in the genome of the cancer cell.

[0131] In some embodiments, the cancer-selective induced sequence repeats are identified by sequencing the cancer cells. In some embodiments, the whole genome of the cancer cells is sequenced. In some embodiments, the genome of the cancer cells is sequenced. In some embodiments, the cancer-selective induced sequence repeats are caused by a genotoxic agent (e.g., a chemotherapeutic agent, such as temozolomide, cisplastin, paclitaxel, or 5-fluorouracil, or radiation). In some embodiments, the cancer-selective induced sequence repeats comprise complementary nucleic acid sequences to any of the any of the nucleic acid sequences of sgCIDE sequences listed in Tables 2 or 4. The cancer cells may be primary or metastatic cancer cells. In some embodiments, the cancer cells are glioblastoma cancer cells.

[0132] Also provided herein are one or more vectors encoding the composition of gRNAs and gRNA-guided nucleases described herein. The vectors aid in providing the composition to the cancer cells. In some embodiments, the vector is a viral vector, for example adenovirus vector, adeno-associated viral vector, retroviral vector, or lentiviral vector. In some embodiments, the vector is a retroviral vector and the retroviral vector is a replicating retroviral vector (i.e., it has independent replicative capacity. In some embodiments, the retroviral vector lacks independent replicative capacity. In some embodiments, gRNA is delivered to the cancer cells in a replicating retroviral vector and the gRNA-guided nuclease is delivered to the cancer cells in a nonreplicating retroviral vector.III, Methods of detecting cancer-selective induced sequence repeats

[0133] Provided herein are methods of detecting-cancer selective induced sequence repeats (e.g., non-coding sequence repeats, coding sequence repeats or both) in a genome of cancer cells in an individual comprising administering a genotoxic agent to the individual in a sub-lethal dose to the cancer cells and sequencing the genome of the cancer cells to detect cancer-selective induced sequence repeats. As discussed above, the genotoxic agent may be administered to the individual by any suitable method, which will be apparent to one skilled in the art depending on the agent used. The genotoxic agent is administered to the individual at a dose that is sub-lethal to the cancer cells, i.e., less than the dose required to kill the cancer cells.

[0134] In the present methods, sequencing cancer cells for detection of cancer-selective induced sequence repeats may be performed as described above or elsewhere herein. Suitable sequencing methods are described above. In some embodiments, the sequencing is performed prior to and / or after the delivering. Sequencing may be performed on the whole genome of the cancer cell or only on the non-coding genome of the cancer cells. In some embodiments, the cancer-selective induced sequence repeats comprise complementary nucleic acid sequences to any of the any of the nucleic acid sequences of sgCIDE sequences listed in Tables 2 or 4.

[0135] In some embodiments, the method of detecting cancer-selective induced sequence repeats further comprises sequencing non-cancerous cells of the individual, identifying cancer- selective induced sequence repeats present in a genome of the cancer cells not substantially present in non-cancerous cells; and selecting gRNAs that specifically target the cancer-selective induced-sequence repeats not substantially present in the non-cancerous cells. As described above, not substantially present refers to refers to cancer-selective induced sequence repeats present in the genome of cancer cells at least 2 times the amount of cancer-selective induced sequence repeats present in the genome of non-cancer cells. In some embodiments, not substantially present refers to cancer-selective induced sequence repeats present in the genome of cancer cells 2 times, 3 times, 4 times, 5 times, 6 times, 7 times, 8 times, 9 times, or 10 times the amount of cancer-selective induced sequence repeats present in the genome of non-cancer cells. In some embodiments, the selected gRNAs do not target non-cancerous cells.EXAMPLES

[0136] The examples below are intended to further illustrate certain aspects of the methods and compositions described herein and are not intended to limit the scope of the claims.ResultsCRISPR leveraging of the non-coding genome

[0137] To comprehensively assess mutational landscape progression in an individual GBM, a patient’s primary and recurrent GBMs were analyzed by whole-genome sequencing and compared to the patient’s native genome (FIG. 1 A). Upon initial diagnosis, the patient had undergone surgical resection as well as adjuvant radiotherapy and chemotherapy with TMZ. Eleven months later the patient relapsed. Quantification of the tumor mutational burden (TMB) showed a median of 123 mutations per megabase (Mb) for the primary tumor and 217 mutations per Mb for the recurrent GBM (p<0.01, Wilcoxon Rank Sum) (FIGS. 8A-B). This qualifies both tumors as hypermutated (see, e.g., Zehir, A. et al. Mutational landscape of metastatic cancer revealed from prospective clinical sequencing of 10,000 patients. Nat. Med. 23, 703-713 (2017) and Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas. Nature 580, 517-523 (2020)). By computational variants analysis using Mutect2, 4,434 and 11,601 protein coding mutations were identified in the primary and recurrent GBM, respectively (FIG. IB). Notably, a far larger number (451,484 and 698,557, respectively) of mutations were found in the non-coding genome. As coding regions represent -1-2% of the human genome, this implies a similar mutational burden across the entire genome (see, e.g., Venter, J. C. et al. The sequence of the human genome. Science 291, 1304-1351 (2001); Auton, A. et al. A global reference for human genetic variation. Nature 526, 68-74 (2015); and Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860-921 (2001)). Despite the considerable sequence space of the non-coding genome; however, current targeted therapies focus almost exclusively on protein-coding elements and their regulation. In GBM, this often means that recurrent tumors do not respond to initial treatment regimens as their genetic background differs from the primary tumor (see, e.g., Johnson, B. E. et al. Mutational analysis reveals the origin and therapy-driven evolution of recurrent glioma. Science 343, 189-193 (2014)). Hence, to track the patient’s genomic tumor progression, all variants were intersectedwith the ClinVar (NCBT) database to identify variants at loci annotated as pathogenic or likely pathogenic (shown in Table 1 below).Table 1. List of variants found in the patient’s native genome, primary GBM, and recurrent GBM that are reported in ClinVar (NCBI) as risk factor, likely pathogenic, or pathogenic. Genome = normal, primary, or reccurent.

[0138] Closer inspection revealed two germline missense mutations in the genes CHEK2 and ARID IB (FIG. 1C). Both genes are involved in DNA damage response (DDR). CHEK2 is a checkpoint kinase that acts as a tumor suppressor gene and has been implicated in a wide range of cancers, including gliomas (see, e.g., Bartek, J. & Lukas, J. Chkl and Chk2 kinases in checkpoint control and cancer. Cancer Cell 3, 421-9 (2003)). CHEK2 has also been associated with mismatch repair (MMR) through interaction with MSH2 (see, e.g., Adamson, A. W. et al. Methylator-induced, mismatch repair-dependent G2 arrest is activated through Chkl and Chk2. Mol. Biol. Cell 16, 1513-26 (2005)). ARID IB, another tumor suppressor, is part of the SWI / SNF chromatin remodeling complex and its alterations have been linked to cancer, including neuroblastoma and gliomas (see, e.g., Watanabe, R. et al. SWI / SNF factors required for cellular resistance to DNA damage include ARID1A and ARID1B and show interdependent protein stability. Cancer Res. 74, 2465-75 (2014); Ichimura, K. et al. Small regions of overlapping deletions on 6q26 in human astrocytic tumours identified using chromosome 6 tile path array- CGH, Oncogene 25, 1261-71 (2006); and Sausen, M. et al. Integrated genomic analyses identify ARID1A and ARID1B alterations in the childhood cancer neuroblastoma. Nat. Genet. 45, 12-7(2013)). Furthermore, both the primary and recurrent tumor had a somatic mutation in the TERT- promoter, referred to as “G228A”. 77A>7'-promotcr mutations are the most common non-coding mutation in cancer, occur in -83% of GBM, and enable binding of a GABPBIL-containing transcription factor complex that activates TERT expression (see, e.g., Mancini, A. et al. Disruption of the 01L Isoform of GABP Reverses Glioblastoma Replicative Immortality in a TERT Promoter Mutation-Dependent Manner, Cancer Cell 34, 513-528. e8 (2018)). Lastly, the recurrent tumor acquired a large set of additional pathogenic mutations, including a known coding variant in the oncogene PIK3CA, a missense mutation in the methyltransferase KMT2D that is mutated in -10% of cancer, and a missense mutation in the MMR nQMSH6 that has been identified as key driver of hypermutation in glioma (see, e.g., AACR Project GENIE Consortium. AACR Project GENIE: Powering Precision Medicine through an International Consortium, Cancer Discov. 7, 818-831 (2017) and Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas, Nature 580, 517-523 (2020)).

[0139] Overall, germline deficiencies were uncovered in DDR genes, which likely enabled the high mutational burden in the primary tumor, and a somatic mutation in the 77A>7-promoter critical for GBM cell immortalization. Moreover, the data revealed that the recurrent tumor is genetically divergent from the primary GBM, with several additional mutations associated with aggressive cancer progression and hypermutation. Consequently, no classical therapy is likely to yield lasting remission for this type of cancer. To explore alternative therapeutic strategies, CRISPR was leveraged and all possible Streptococcus pyogenes Cas9 (SpyCas9, 5’-NGG-3’ PAM) single-guide RNAs (sgRNAs) in the patient’s native genome as well as primary and recurrent GBM were computed (FIG. ID). Among the hundreds of millions of possible sgRNAs in this target space, termed “sgRNA-ome”, a large proportion that had multiple target loci in the genome, with the highest ones targeting over 100,000 sites each.

[0140] To better understand the origin of repetitive CRISPR target loci, all highly repetitive sequences (>1,000 repeats) were annotated. When aligned to annotation tracks of coding and non-coding elements (GENCODE v38, UCSC RepeatMasker), repeat sequences mostly mapped to the non-coding genome (FIGs. IF, 8C). Functionally, they were often derived from retrotransposons and other transposable elements (TEs), which account for up to -45% of the human genome, or microsatellites that account for -3% of the human genome (see, e.g., Lander, E. S. et al. Initial sequencing and analysis of the human genome, Nature 409, 860-921 (2001)and Ellegren, H. Microsatellites: simple sequences with complex evolution, Nat. Rev. Genet. 5, 435-445 (2004)). The TE class, representing the majority of repeat targets, was partitioned into two subgroups with differential representation among coding and non-coding genome (FIG. IF, inset). In contrast, the microsatellite class did not show this separation. TE-based targets had either a primary association with long interspersed nuclear elements (LINEs; -1-3% in coding genome) or short interspersed nuclear elements (SINEs; <1% in coding genome).

[0141] It was further assessed whether repetitive targets cluster in distinct chromosomes or genomic regions, but observed a nearly uniform distribution throughout the genome, with the exception of peri centromeric heterochromatin and stalks in acrocentric chromosomes (FIGs. 1G, 8D) which are often not fully represented in the human reference genome (see, e.g., Furey, T. S. & Haussler, D. Integration of the cytogenetic map with the draft human genome sequence, Hum. Mol. Genet. 12, 1037-1044 (2003)). Hence, while highly repetitive sgRNA targets are spread throughout the genome, they are mostly associated with discrete repeat elements of TE origin in the non-coding genome.Genome Shredding enables rapid cell elimination

[0142] To investigate the ability of CRISPR-Cas9 to eliminate GBM cells through targeting of repetitive sequences, a strategy referred to herein as “Genome Shredding” (FIGs. 1A. 2A), ten of the most repetitive sgRNAs in the human genome were selected (as shown in Table 2 below). Shown are the sgCIDE sequence and count of fully complementary (Hamming distance = 0) loci in the human genome. For each sequence, featuring -3,000-300,000 target sites per haploid genome, the number of off-target sites at Hamming distance (HD) of up to three, and allowing for NGG or NAG PAMs were determined, shown in below Table 2). Importantly, when tested in U-251 and LN-229 GBM cells, stable expression of Cas9 and all of the ten sgCIDEs led to rapid and robust cell depletion in competitive proliferation assays (FIG. 2C). While cells expressing non-targeting sgRNAs (sgNT) did not deplete, sgCIDE-expressing cells depleted more rapidly than cells expressing sgRNAs targeting a consensus essential gene, replication protein Al (RPA1). To assess Genome Shredding in a more therapeutically relevant model, without preengineered Cas9 expression, these experiments were repeated in native U-251 and normal human astrocytes (NHA) using an all-in-one Cas9-sgRNA vector (pCF827), with comparable results (FIG. 9A).Table 2. Genomic target count of select highly repetitive sgCIDEs.

[0143] To interrogate genomic consequences of Genome Shredding, conducted comet assays were conducted and DNA from lysed target cells on were run on agarose-coated slides (FIG. 2D). Single-cell analysis of Cas9 expressing U-251 and LN-229 showed very long tails for sgCIDE- 1 / 2 / 3 expressing cells at 24 hours post-transduction, while minimal degradation was observed for sgNT-1 / 2 / 3 controls (FIGs. 2E-F). This demonstrates extensive genomic fragmentation at early time points. Next, to assess the timing and dynamic effects of repetitive loci targeting, real-time quantitative live-cell imaging (Incucyte) was used to monitor Cas9 expressing U-251 stably transduced with lentivirus expressing mNeonGreen-coupled sgCIDEs (sgCIDE-1 / 2 / 3 / 6 / 8 / 10) or negative controls (sgNT-1 / 2 / 3) over five days (FIGs. Fig. 2F, 9B). Propidium iodide (PI) staining and quantification of live cells revealed that sgCIDEs induced growth inhibition starting at day one post-transduction, and cell death starting as early as day two. Together, these experiments show that targeting of highly repetitive sequences is a reliable strategy for rapid elimination of GBM cells.Genome Shredding is agnostic to genetic and epigenetic origin

[0144] Active MGMT removes TMZ-added methyl groups from the O6position of guanines, rendering chemotherapy ineffective (see, e.g., Pegg, A. E., Wiest, L., Foote, R. S., Mitra, S. & Perry, W. Purification and properties of 06-methylguanine-DNA transmethylase from rat liver. J. Biol. Chem. 258, 2327-33 (1983) and Hegi, M. E. et al. MGMT gene silencing and benefitfrom temozolomide in glioblastoma. N. Engl. J. Med. 352, 997-1003 (2005)). In GBM cells with promoter-methylated (silenced) MGMT, TMZ leads to DNA damage and a characteristic G2 / M cell cycle arrest, followed by p53-dependent cell death (see, e.g., Hirose, Y., Berger, M. S. & Pieper, R. O. p53 effects both the duration of G2 / M arrest and the fate of temozolomide-treated human glioblastoma cells. Cancer Res. 61, 1957-63 (2001)). Hence, it was investigated whether Genome Shredding is equally active in both TMZ sensitive and resistant GBMs. To evaluate cell elimination dynamics, Cas9 expressing TMZ-sensitive U-251 and LN-229, and TMZ-resistant T98G and LN-18 GBM cells were treated with TMZ or transduced them with lentiviral vectors expressing sgCIDEs (FIGs. 3 A, 10A). Luminescence-based quantification of cell viability over five days showed expected effects for TMZ dose-titrations, with lethality observed only in U-251 and LN-229. In contrast, sgCIDE- 1 / 2 / 3 / 6 / 8 / 10 expression induced viral titer-dependent lethality in all four tested GBM lines independent of MGMT promoter methylation status, while negative controls (sgNT-1 / 2) showed no effect.

[0145] To assess effects on cell cycle progression, cells were treated with TMZ or sgCIDEs for one to five days and stained them with PI post-fixation, followed by flow cytometry analysis (FIG. 3B). Control DMSO and sgNT-1 / 2 treatments, as well as guide RNAs targeting a non- essential gene (olfactory receptor, sgOR2B6-l / 2), showed comparable normal cell cycle profiles in Cas9 expressing U-251, LN-229, T98G, and LN-18 (FIGs. 3C-D, 10B). As expected, TMZ treatment (50 pM) showed the characteristic G2 / M cell cycle arrest with initial increase of the G2 peak, loss of Gl, and slow increase of the Sub-Gl population starting at day two, but only in the TMZ-sensitive lines. Increase of the Sub-Gl population was more prominent in U-251 (RRIDUVCL 0021) carrying the common TP53 R273H mutation compared to LN-229 (RRID:CVCL_0393, TP53 P98L) that retain a functional TP53 (FIG. 10B). This reflects previous observations that TP53 status affects resolution of the G2 / M arrest (see, e.g., Hirose, Y., Berger, M. S. & Pieper, R. O. p53 effects both the duration of G2 / M arrest and the fate of temozolomide-treated human glioblastoma cells. Cancer Res. 61, 1957-63 (2001)). Additionally, treatment with guide RNAs targeting the essential gene RPAl (sgRPAl-2 / 3) resulted in an accumulation in S-phase starting at day three, accompanied by increase of the Sub-Gl population, in all four GBM lines.

[0146] In contrast, Genome Shredding with sgCIDE-1 / 2 / 3 / 6 / 8 / 10 led to a rapid increase of the Sub-Gl population starting at day one post-transduction, combined with a drastic depletion ofthe G1 peak and slight increase of the S-phase population, in all four tested GBM lines (FIGs. FIGs. 3C-D, 10B), this change in cell cycle profde was consistent across all six sgCIDEs and for all four tested GBM cell lines, independent of MGMT promoter methylation status and TMZ- sensitivity. At day two post-transduction, the Sub-Gl population of sgCIDE transduced samples already represented -20-40%, and -30-60% at day 3, a signal of rapid cell death.

[0147] To further evaluate Genome Shredding in systems that model the therapy resistance seen in patients, glioblastoma stem-like cell (GSC) lines (GBM12, GBM85, GBM123) were transduced with all-in-one vectors (pCF827) expressing Cas9 and sgRNAs (sgNT-1 / 2 / 3, sgCIDE- 1 / 2 / 3), and assessed cell viability (FIG. 3E). While TMZ (50 pM) treatment efficiency varied across the three GSC lines, sgCIDE- 1 / 2 / 3 led to effective cell depletion in all cases. Thus, CRISPR-Cas Genome Shredding robustly induced cell death independent of the glioma cells’ genetic and epigenetic makeup, suggesting versatility in dealing with intratumoral heterogeneity.Genome Shredding escape is non-trivial and can be rescued with repeat treatment

[0148] To better understand what happened with these “escapee” cells, and whether there are biological barriers to cell ablation by Genome Shredding, a recurrent treatment assay was designed (FIG. 4A). Monoclonal cell lines were established by isolating escapee clones (sgCIDE-1 / 2), as well as negative controls (sgNT-1), and expanded them under puromycin selection to ensure continued presence of the genomically integrated Cas9 constructs (EFS-Cas9- P2A-Puro). It was further confirmed that all cell lines remained mNeonGreen positive, a marker for the sgRNA vector (U6-sgRNA EFla-mNeonGreen). Overall, four clones each from U-251 and LN-229 escapees, as well as respective controls, were prepared. To model recurrent treatment, a lentiviral vector expressing an sgRNA and mCherry (pCF820) as well as an all-in- one version encoding an sgRNA and mCherry-tagged Cas9 (pCF826) was developed. Additionally, for high-titer production of all-in-one vectors, a “CRISPR-Safe” packaging cell line was established through stable overexpression of the anti-CRISPR protein AcrIIA4 (FIGs. 11 A-B) using a lentiviral construct.

[0149] Competitive proliferation assays showed that retreatment of escapee clones with vectors providing only a new sgRNA (pCF820; sgNT-1, sgCIDE- 1 / 2) did not result in cell depletion (FIGs. 11C-D). However, retreatment with all-in-one vectors providing both a new sgRNA and Cas9 (pCF826) resulted in efficient cell ablation in all cases (FIGs. 4C, 1 IE). Interestingly, most (7 / 8) escapee clones could be retreated even with the Cas9-sgNT-l controlconstruct. Since all escapee clones were monitored for continued Cas9 expression (puromycin selection) and sgRNA vector presence (mNeonGreen+), this demonstrates that lack of a functional Cas9 was the key issue, with one escapee clone (U-251 sgCIDE-1 #1) also lacking a functional sgRNA, rather than a potential biological mechanism enabling cells to escape Genome Shredding. Cas9 may have been inactivated by in-frame mutations. Such mutations could have arisen already during plasmid propagation or viral preparation, rather than after cellular integration.

[0150] To further investigate the potential of Genome Shredding in an in-vivo tumor environment, sgNT-1 or sgCIDE-1 expressing lentivirus (5 pl, 15 pl) was locally injected into mice bearing Cas9-expressing LN-229 intracranial xenografts (FIG. 5A). Prior to injection, the efficacy of the concentrated virus in vitro, using Cas9-expressing LN-229, was evaluated, and showed efficient cell transduction and destruction for sgCIDE-1 (FIG. 5B). Post-mortem flow cytometry analysis of tumors treated in vivo with sgNT-1 constructs revealed -15-20% mNeonGreen+ cells (5 pl: 8.9-81.5%, 15 pl: 2.7-32.1%). In contrast, tumors injected with sgCIDE-1 showed only -0.2-0.5% mNeonGreen+ cells (5 pl: 0.0-0.4%, 15 pl: 0.0-0.9%), indicating that the vast majority of in-vivo sgCIDE-1 transduced cells were effectively eliminated (FIG. 5C). However, Kaplan-Meier analysis revealed no survival benefit of sgCIDE-1 treated groups (FIG. 5D), suggesting that the therapeutic delivery was insufficient and that elimination of -15-20% of tumor cells does not prolong overall survival.

[0151] To bypass delivery limitations and potentially achieve a more pronounced therapeutic effect, a tamoxifen-inducible Cas9 system, termed “TiCas9”, was developed based on arC9 variants (as described by, for example, Oakes et al., ” Profiling of engineering hotspots identifies an allosteric CRISPR-Cas9 switch, Nat Biotechnol. 2016 Jun;34(6):646-51). After evaluating TiCas9’s 4-hydroxytamoxifen-induced (4-OHT, 3 days) editing potential in GFP- containing HEK293T cells (FIG. 5E), Genome Shredding efficiency with sgCIDE-1 / 2 / 3 was assessed in TiCas9-encoding U-251 (FIG. 5F). While no leakiness in DMSO conditions was observed, 4-OHT treatment led to strong TiCas9 editing activity, albeit not fully matching wildtype Cas9 levels. Next, intracranial xenografts were established from TiCas9 and sgNT- 1 / sgCIDE-l encoding U-251 and treated them with intraperitoneal injections of tamoxifen (100 mg / kg per day for 5 days), starting one week after tumor implantation (FIG. 5G). Overall, Kaplan-Meier analysis, with cohorts of four and five mice, revealed a slight survival benefit ofthe sgCIDE-1 treated group, though not quite significant (p=0.0528). Together, these results demonstrate Genome Shredding-based tumor cell ablation in an in vivo microenvironment and support its potential as a therapeutic approach for brain cancer treatment.Minimal determinants of effective cell ablation

[0152] To delineate parameters of Genome Shredding, the minimum number of target repeats needed to effectively eliminate a cell was investigated. Generally, each target locus will lead to a CRISPR-induced DSB, followed by cellular repair through non-homologous end joining (NHEJ) or homology-dependent repair (HDR) or, if there is substantial DNA damage, cell death. However, target loci lead to DNA DSBs with varying efficiency depending on the sequence of the sgRNA, the complementarity of the sgRNA to the target sequence, and - to a lesser degree - the chromatin accessibility of the target site. Thus, to assess cell depletion efficiencies of sgCIDEs in eukaryotic cells with varying amounts of endogenous target sites, a set of sgRNAs was tested in human, mouse, and chicken genomic backgrounds (FIG. 6A]).

[0153] In addition to the human sgRNAs, all SpyCas9 (5’-NGG-3’ PAM) sgRNAs were computed for mouse (Mus musculus mm 10) and chicken (Gallus gallus, galGal6) genomes, and sgRNA occurrences were enumerated (FIG. 12A). Next, the number of target sites with full complementarity (HD 0) or 1-3 mismatches (HD 1-3) was quantified across all three genomes and a set of sgRNAs was selected for a DSB-induced cell death assay (see Tables 3 and 4 below; see also Table 2 above). Beyond non-targeting control sgRNAs (sgNT-1 / 2 / 3), three top human sgRNAs with fewer targets in the mouse genome and no targets in the chicken genome (sgCIDE- 1 / 2 / 3), four sgRNAs that target only the mouse genome (sgCIDE-21 / 22 / 23 / 24), six sgRNAs targeting the human and mouse genome (sgCIDE-31 / 32 / 33 / 34 / 35 / 36), and six microsatellitetargeting (k-mer repeat) sgRNAs (sgCIDE-41 / 42 / 43 / 44 / 45 / 46) with high target counts in all three genomes were chosen.Table 3.Table 4. Sequences for sgCIDEs in Table 3.

[0154] Competitive proliferation assays in Cas9 expressing human U-251 GBM, mouseGL261 GBM, and chicken DF-1 fibroblast cells demonstrated that pan-vertebrate sgCIDEs (#41- 46) rapidly deplete transduced cells independent of the species (FIG. 6B). Beyond the minimal number of 32 fully matching (HD 0) target sites in this set (sgCIDE-44, chicken genome), there was little correlation between cell death and the number of target sites, though all sgRNAs had at least several hundred target sites when counting loci with single mismatches (HD 1). Additionally, the top human sgRNAs (sgCIDE- 1 / 2 / 3) also efficiently eliminated mouse GBM cells despite having only -70-100 target sites with a single mismatch (HD 1) and only two fully complementary sites. Similarly, sgRNAs with hundreds of target sites in the human genome (HD 0) were equally efficient at depleting target cells as sgRNAs with hundreds of thousands of target sites. While it remains difficult to accurately determine the number of DSBs needed for efficient cell elimination, these results suggest that a minimum of -30 fully complementary target sites or -70 target sites with a single mismatch are sufficient.Leveraging mutational signatures in hypermutated GBM for targeted oncolysis

[0155] Cancer-specific mutations in hypermutated gliomas or other tumors, introduced during therapy, might lead to unique sequences that can be exploited for cancer-specific genome fragmentation, a process referred to herein as “Cancer Shredding” (FIG. 7A). Concretely, following rounds of TMZ treatment, GBMs often relapse as hypermutated tumors with characteristic cytosine-to-thymine (C>T) conversions mainly due to a combination of TMZ exposure and acquired MMR pathway deficiency (see, e.g., Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas. Nature 580, 517-523 (2020) and Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, 415-21 (2013)). These mutations render the genomic landscape of recurrent gliomas distinct from the patient’s native genome and should enable exclusive CRISPR targeting.

[0156] To assess whether such mutations exist in the patient data and generate cancer-specific sgRNA targets, the patient’s primary and recurrent GBM, as well as native genome were analyzed. Quantification of the mutational landscape across the coding and non-coding genome revealed stereotyped signatures not only in the primary and recurrent GBM, but also in the native genome (FIG. 6A). Both the native genome and primary GBM showed a high prevalence of OT and T>C mutations, a signature associated with germline MMR deficiency in absences of distinct exogenous stressors (see, e.g., Zhao, H. et al. Mismatch repair deficiency endows tumors with a unique mutation signature and sensitivity to DNA double-strand breaks. Elife 3, e02725 (2014)). Additionally, the recurrent GBM showed a specific increase in OT conversions, consistent with TMZ chemotherapeutic stress (see, e.g., Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas. Nature 580, 517-523 (2020)). When accounting for the mutational background present in the patient’s native genome, the recurrent GBM displayed the characteristic TMZ mutational signature (Signature 11) of hypermutated GBM (FIG. 7B) (see, e.g., Alexandrov, L. B. et al. Signatures of mutational processes in human cancer, Nature 500, 415-21 (2013) and Alexandrov, L. B. et al. The repertoire of mutational signatures in human cancer, Nature 578, 94-101 (2020)).

[0157] To determine whether TMZ signature mutations can be leveraged for GBM-specific targeting and cell ablation, a cell line was established from the patient’s recurrent GBM (FIG. 7A). Next, the sensitivity of this hypermutated patient-derived cell line (PDCL, SF11411) to general Genome Shredding was assessed, with observed effects similar to other GBM lines (FIG. 13B). To ensure that the SF11411 line represents the patient’s recurrent tumor, and not a distant sub-clone, whole genome sequencing was performed and a list of recurrent GBM specific sgRNAs (“sgRECs”) was computed. The mutational landscape of the SF11411 PDCL correlated well with that of the patient’s recurrent GBM tissue (FIG. 7C). For inclusion in the recurrent GBM sgRNA-ome, an sgRNA target sequence had to be found in both PDCL sequencing replicates and have a single mismatch at position 1-8 from the PAM compared to the patient’s native genome. Specifically, the candidate sgRNAs had zero targets at HD 0 and only one target at HD 1 in both the patient’s native and reference (hg38) genomes. Importantly, identified cancer-specific repeat sgRNAs were identified by computing targets that were both repetitive in the recurrent cancer and unique to the recurrent cancer (mismatch at position 1-8 from PAM); 10sgRNAs that meet these criteria were uncovered (FIG. 7D). Cancer-specific repeat sgRNAs are shown below in Table 5.Table 5. List of cancer-repeat sgRNAs and occurrence in occurrent GBM (HDO) and native genome (HDO)

[0158] To validate cancer-specific sgRNA predictions, a large-scale CRISPR screen was carried out in NHAs versus the hypermutated recurrent GBM SF1 141 1 PDCL (Fig. 7A). The library included guides targeting the coding and non-coding genome at a single site or multiple loci (including cancer repeat sgRNAs), all sgCIDEs, ~80 non-targeting controls, and -1000 safe-harbor references (with no off-targets up to HD 3), for a total of -5000 sgRNAs (see Table 3 above). NHAs and SF1141 Is transduced at single-copy (35% for NHA, 11% for SF11411) were compared with the CRISPR library by next-generation sequencing (NGS) at day 1 (TO) and 28 days (D28) post transduction (14 cell doublings). Most sgCIDEs were strongly depleted in both cell lines, with the exception of the mouse specific sgCIDE-21 / 22 / 23 / 24 and some sgCIDE-2 / 3 replicates that showed amplification biases in NHAs (FIGs. 7E, 13C). More importantly, while the non-targeting controls behaved neutrally in both lines, the cancer repeat sgRNAs strongly depleted only in the SF11411 line, while ending up enriched in NHAs.

[0159] To further quantify cancer-specific sgRNA depletion, the Mageck pipeline for statistical analysis (as described by, for example, “Li et al., MAGeCK enables robust identification of essential genes from genome-scale CRISPR / Cas9 knockout screens”, Genome Biol. 2014; 15(12): 554) was used. Among all sgRNAs in the CRISPR screen, the cancer repeat sgRNAs were most strongly depleted specifically only in the hypermutated SF11411 GBM cells (FIGs. 7F]) (see Table 3 above). Using gene-level statistics and an FDR<0.05, the cancer-repeat sgRNAs again stood out as the most highly depleted set (FIG. 7G). Together, this demonstrates that TMZ signature mutations in hypermutated recurrent glioma can result in unique sequences that enable selective CRISPR targeting of tumor cells and effective Cancer Shredding.

[0160] While this patient presented with a rare germline DDR deficiency and high mutational burden already in the native genome, for the many patients with acquired MMR deficiency (and a lower baseline burden of C>T mutations in the native genome (FIG. 13 A) it should be possible to identify even more hypermutated glioma-specific repeat sgRNAs (personalized cancerspecific repeat sgRNAs). This should hold particularly true for hypermutated astrocytoma and oligodendrogliomas that accumulate mutations over a much longer period of time (see, e.g., Jones, L. E. et al. Patient-derived cells from recurrent tumors that model the evolution of IDH- mutant glioma, Neuro-Oncology Adv. 2, vdaa088 (2020)). Taken together, Cancer Shredding provides a promising therapeutic approach for difficult-to-treat glioma and other hypermutated tumors.

[0161] CRISPR systems have revolutionized medicine by enabling genome engineering through RNA-guided introduction of DNA double-strand breaks. To identify an Achilles heel of hypermutation, the mutational landscape progression in a patient’s primary and recurrent GBM was quantified, and targetable repeat elements were identified. CRISPR-mediated genomefragmentation, by targeting highly repetitive loci, enables rapid elimination of GBM cells, an approach termed “Genome Shredding” herein. Minimal thresholds for efficient cell ablation were defined, and the non-coding genome was uncovered as a key source of targets for Genome Shredding. Importantly, unique repeat sequences in a patient’s recurrent GBM that carry a TMZ mutational signature were identified, demonstrating that their CRISPR targeting enables cancerspecific cell ablation. Together, CRISPR-mediated “Cancer Shredding” provides an innovative paradigm leveraging the non-coding genome and therapy-induced mutational signatures for selective glioma cell depletion and treatment of recurrent GBM, and possibly other tumors with hypermutated genomes.

[0162] Compared to conventional approaches, CRISPR strategies open up the vast space of the non-coding genome and non-driver mutations as targets for therapy. While protein coding sequences occupy -1-2% of the human genome, with driver genes being a small fraction of that, repeat elements make up over 50% (see, e.g., Martinez-Jimenez, F. et al. A compendium of mutational cancer driver genes. Nat. Rev. Cancer 20, 555-572 (2020) and Lander, E. S. et al. Initial sequencing and analysis of the human genome. Nature 409, 860-921 (2001)).Accordingly, the data show that the majority of the genomic sequence and variant space can serve as potential guide RNA targets for Genome Shredding. Moreover, the existence of tumorspecific repeat sequences would enable targeted Cancer Shredding without the need for cell typespecific delivery. Analyzing the patient’s cancer by whole genome sequencing (WGS), unique recurrent GBM-specific sgRNAs were identified. The targets of these sgRNAs, mainly in the non-coding genome, were generated by TMZ signature mutations characteristic of hypermutated gliomas (see, e.g., Touat, M. et al. Mechanisms and therapeutic implications of hypermutation in gliomas. Nature 580, 517-523 (2020) and Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, 415-21 (2013)). Importantly, using a large-scale CRISPR screen, it is demonstrated herein that Cancer Shredding enables selective elimination of hypermutated recurrent patient-derived GBM cells through targeting of therapy-induced TMZ signature mutations, thereby paving the way for innovative therapeutic interventions.

[0163] Prior work has shown that collateral nuclease activity of Casl3 can be exploited to inhibit the growth of certain cancer cells (as described by, for example, Wang et al., “The CRISPR-Casl3a Gene-Editing System Induces Collateral Cleavage of RNA in Glioma Cells”, Adv Set (Weinh). 2019 Aug 29;6(20): 1901299). However, since Casl3 targets RNA, thisapproach only allows targeting mutations in expressed regions of the genome, thereby restricting the sequence space by excluding large sections of the non-coding and coding genome. Cancer Shredding advantageously allows for targeting of unexpressed regions of the genome (e.g., noncoding regions).

[0164] While the guide RNAs employed here for Cancer Shredding are personalized to this patient’s recurrent GBM, identifying cancer-specific sgRNAs will be feasible for many patients with recurrent hypermutated gliomas and other tumors, particularly those with acquired MMR deficiency. As hypermutated astrocytoma and oligodendrogliomas accumulate mutations over a much longer period of time post MMR loss and TMZ treatment (years rather than months), they often have a higher TMB compared to hypermutated GBMs (see, e.g., Jones, L. E. et al. Patient- derived cells from recurrent tumors that model the evolution of IDH-mutant glioma. NeuroOncology Adv. 2, vdaa088 (2020)). Additionally, the baseline genome of most patients (without germline DDR / MMR deficiencies) will contain fewer pre-existing C>T mutations (FIG. 13A) making the TMZ signature mutations more prominent. Together, this should facilitate the identification of effective cancer-specific repeat sgRNAs.

[0165] Though the development of appropriate delivery modalities remains a challenge, Cancer Shredding establishes an innovative concept to tackle hypermutated gliomas. Recurrent cancers often lack effective treatment regimens because mutations driving growth are distinct from those in the primary tumor. Cancer Shredding provides an alternative to this paradigm by leveraging the non-coding genome and TMZ signature mutations. Of note, beyond salvage therapy for recurrent gliomas, Cancer Shredding could serve as non-systemic neo-adjuvant therapy to shrink hypermutated tumors prior to surgical resection. Additionally, stereotactic administration could enable therapy for tumors that are either too small to resect or are situated in regions of the brain where surgical resection is undesirable (FIG. 7H). Of note, the injury and destruction of tumor genomes by Cancer Shredding would likely boost the effect of immunotherapy in combating disease progression. Together, CRISPR-mediated Cancer Shredding provides an innovative therapeutic concept for patients with hypermutated gliomas, and possibly many other cancers with hypermutated genomes.Materials and MethodsMammalian cell culture

[0166] All mammalian cell cultures were maintained in a 37°C incubator at 5% CO2. HEK293T human kidney cells (293FT; Thermo Fisher Scientific, #R70007;RRID:CVCL_6911), HEK-RT1 (a monoclonal HEK293T cell line expressing a doxycycline- inducible GFP) (as described by, for example, Oakes et al., “CRISPR-Cas9 Circular Permutants as Programmable Scaffolds for Genome Modification”, Cell. 2019 Jan 10; 176(1 -2):254- 267. el6), DF-1 chicken fibroblast cells (ATCC, #CRL-12203; RRID:CVCL_0570), normal human astrocytes (NHAs; culture option 1, used for cell line establishment), and derivatives thereof were grown in Dulbecco’s Modified Eagle Medium (DMEM; Corning Cellgro, #10-013- CV) supplemented with 10% fetal bovine serum (FBS; Seradigm #1500-500), and 100 Units / ml penicillin and 100 pg / ml streptomycin (100-Pen-Strep; Gibco, #15140-122). U-251 human glioblastoma cells (Sigma-Aldrich, #09063001; RRID:CVCL_0021), LN-229 human glioblastoma cells (ATCC, #CRL-2611; RRID:CVCL_0393), T98G human glioblastoma cells (ATCC, #CRL-1690; RRID:CVCL_0556), LN-18 human glioblastoma cells (ATCC, #CRL- 2610; RRID:CVCL_0392), SF11411 hypermutated recurrent glioblastoma patient-derived cell line (PDCL) cells, GL261 mouse glioblastoma cells, normal human astrocytes (NHAs; culture option 2, used for experiments and CRISPR screen), and derivatives thereof were cultured in Dulbecco’s Modified Eagle Medium / Nutrient Mixture F-12 (DMEM / F-12; Gibco, #11320-033) or Dulbecco’s Modified Eagle Medium and Ham's F-12 (DMEM / F-12; Coming Cellgro, #10- 090-CV) supplemented with 10% FBS and 100-Pen-Strep. Glioblastoma stem-like cell lines (GSCs) GBM12, GBM85, and GBM123, were obtained from the Mayo Clinic Brain Tumor PDX national resource. GSCs were cultured in Neurocult NS-A Basal Medium (Stemcell Technologies; #05751), B27 Supplement minus vitamin A (ThermoFisher; #12587010), N2 Supplement (ThermoFisher; #17502001), GlutaMAX (ThermoFisher; #35050061), Sodium Pyruvate (ThermoFisher; #11360070), and Penicillin-streptomycin (ThermoFisher; #15140122) supplemented with 20 ng / ml EGF (ThermoFisher; #PHG0311), and 20 ng / ml FGF (PeproTech; #100-18B). GSCs were grown in monolayers in laminin-coated plates. U-251, LN-229, T98G, LN-18, and HEK293T cells were authenticated using short tandem repeat DNA profiling (STR profiling; UC Berkeley Cell Culture / DNA Sequencing facility). STR profiling was carried out by PCR amplification of nine STR loci plus amelogenin (GenePrint 10 System; Promega, #B9510),fragment analysis (3730XL DNA Analyzer; Applied Biosystems), comprehensive data analysis (GeneMapper software; Applied Biosystems), and final verification using supplier databases including American Type Culture Collection (ATCC) and Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ). U-251, LN-229, T98G, LN-18, and HEK293T cells were tested for absence of mycoplasma contamination (UC Berkeley Cell Culture facility) by fluorescence microscopy of methanol fixed and Hoechst 33258 (Poly sciences, #09460) stained samples.

[0167] The SF11411 PDCL was derived from the patient’s recurrent hypermutated glioblastoma tissue (David Raleigh, Tomoko Ozawa, and Kyounghee Seo; UCSF Brain Tumor Center Preclinical Therapeutics Core). The normal human astrocyte cell line NHA-N2, here simply referred to as “NHA”, is puromycin sensitive and GFP / ZsGreen negative. The monoclonal NHA-N2 cell line was derived from NHA-PC5 cells that are puromycin resistant and ZsGreen positive (as described by, for example, Ohba et al., “Mutant IDH1 Expression Drives TERT Promoter Reactivation as Part of the Cellular Transformation Process”, Cancer Res. 2016 Nov 15;76(22):6680-6689), through CRISPR deletion of the previously inserted puromycin and ZsGreen cassettes. The NHA-PC5 cells were edited by transduction with viruslike particles (VLPs) containing Cas9 and sgPuro-5 / 6 or sgZsGreen-2 / 3 (as described by, for example, Mangeot et al., “Genome editing in primary cells and in vivo using viral-derived Nanoblades loaded with Cas9-sgRNA ribonucleoproteins”, Nat Commun. 2019 Jan 3 ; 10(l):45).Expression vectors

[0168] The lentiviral vector pCF226, expressing Streptococcus pyogenes Cas9 and a puromycin selection marker, was described previously (see, e.g. Oakes, B. L. et al. CRISPR- Cas9 Circular Permutants as Programmable Scaffolds for Genome Modification. Cell 176, 254- 267. el6 (2019)). The lentiviral vector pCF823, expressing Streptococcus pyogenes Cas9 and a puromycin selection marker, was cloned based on the insert of pCF226 and the optimized lentiviral backbone of pCF821. Of note, the pCF823 (Cas9-Puro) vector yields higher viral titers than the otherwise comparable pCF226 (Cas9-Puro) vector. The lentiviral vector pCF821, encoding a U6-sgRNA cassette and an EFla driven mNeonGreen marker, was derived from the pCF525 backbone and the pCF221-based U6-sgRNA-EFla-mCherry insert (see, e.g., Watters, K. E., Fellmann, C., Bai, H. B., Ren, S. M. & Doudna, J. A. Systematic discovery of natural CRISPR-Casl2a inhibitors. Science 362, 236-239 (2018) and Oakes, B. L. et al. CRISPR-Cas9Circular Permutants as Programmable Scaffolds for Genome Modification. Cell 176, 254- 267. el6 (2019)). The mCherry fluorescence marker was replaced with a human codon optimized version of mNeonGreen (gBlock, Integrated DNA Technologies). Analogously, the lentiviral vector pCF820, encoding a U6-sgRNA-EFla-mCherry2 cassette, was derived from pCF821 by replacing the mNeonGreen marker with a human codon optimized version of mCherry2 (gBlock, Integrated DNA Technologies). Of note, both the pCF820 (mCherry 2) and pCF821 (mNeonGreen) sgRNA vectors yield higher viral titers than the otherwise comparable sgRNA vector pCF221 (mCherry). The all-in-one lentiviral vector pCF822, featuring a U6-sgRNA and EFS-Cas9-Puro cassette, was derived from pCF820 with an EFS-Cas9-Puro insert from pCF226. The all-in-one lentiviral vector pCF826, featuring a U6-sgRNA and EFS-Cas9-mCherry2 cassette, was derived from pCF822 by replacing the puromycin selection marker with mCherry2 from pCF820. The all-in-one lentiviral vector pCF827, featuring a U6-sgRNA and EFS-Cas9- mNeonGreen cassette, was derived from pCF822 by replacing the puromycin selection marker with mNeonGreen from pCF821. Vector sequences are provided in Table 2 above.Tamoxifen-inducible Cas9 (TiCas9)

[0169] A tamoxifen-inducible Cas9 system, referred to as “TiCas9”, was established for regulated in vitro and in vivo Cas9 expression. The lentiviral vector pCF287 encodes a CBX3- Delta ubiquitous chromatin opening element (UCOE) from pMHOOOl and an EFS-TiCas9-P2A- Puro cassette. TiCas9 is based on an allosterically regulated Cas9 (ArC9) (as described by, for example, Oakes et al., “Profiling of engineering hotspots identifies an allosteric CRISPR-Cas9 switch”, Nat Biotechnol. 2016 Jun;34(6):646-51) that contains an estrogen receptor ligandbinding domain (ER-LBD) inserted within the Cas9 CDS. To improve tamoxifen responsiveness and editing efficiency, ArC9 was modified by combining the ER-LBD mutations referred to as ER-T2 (G400V / M543A / L544A) (as described by, for example, Feil et al., “Regulation of Cre recombinase activity by mutated estrogen receptor ligand-binding domains”, Biochem Biophys Res Commun. 1997 Aug 28;237(3):752-7) with the V376A mutation (as described by, for example, Buskirk et al., “Directed evolution of ligand dependence: small-molecule-activated protein splicing”, Proc Natl Acad Sci U S A. 2004 Jul 20; 101(29): 10505-10) assumed to change the interaction with ER helix 12, and optimizing codon usage. For in vitro assays 4-OHT (1 pM in 0.1% DMSO; (Z)-4-Hydroxytamoxifen, Sigma-Aldrich #H7904) was used to induce TiCas9. Tamoxifen is a prodrug. It is metabolically converted to its active form, 4-OHT. Tamoxifen isusually used for in vivo studies and 4-OHT in tissue culture. Vector sequences are provided in Table 2 above.Design and cloning of sgRNAs

[0170] All sgRNAs were designed with a G preceding the 20-nucleotide spacer for better expression from U6 promoters. All sgRNA sequences were cloned into the pCF820, pCF821, pCF826, and pCF827 vectors using Esp3I restriction sites and enzymes (New England Biolabs).Viral transduction and generation of Cas9-expressing cell lines

[0171] Generally, lentiviral particles were produced in HEK293T cells or derivatives thereof using polyethylenimine (PEI; Poly sciences #23966) mediated transfection of plasmids, as previously described in Oakes, B. L. et al. CRISPR-Cas9 Circular Permutants as Programmable Scaffolds for Genome Modification, Cell 176, 254-267. el6 (2019). In brief, lentiviral transfer vectors were co-transfected with the lentiviral helper plasmid psPAX2 (Addgene #12260) and the VSV-G envelope plasmid pMD2.G (Addgene, #12259). Transfection reactions were assembled in reduced serum media (Opti-MEM; Gibco, #31985-070). For lentiviral particle production on 6-well plates, 1 pg lentiviral vector, 0.5 pg psPAX2 and 0.25 pg pMD2.G were mixed in 0.4 ml Opti-MEM, followed by addition of 5.25 pg PEI. After 20-30 min incubation at room temperature, the transfection reactions were dispersed over the HEK293T cells. Media was changed 12-14 h post-transfection, and virus harvested at 42-48 h post-transfection. Viral supernatants were filtered using 0.45 pm polyethersulfone (PES) membrane filters, diluted in cell culture media as appropriate, and added to target cells. Polybrene (5 pg / ml; Sigma-Aldrich) was supplemented to enhance transduction efficiency, if necessary. Importantly, unless described otherwise, all-in-one vectors encoding Cas9-sgRNA (e.g. sgCIDE) combinations were produced using the here established CRISPR-Safe packaging cell line to enable high viral titers. Cell lines stably expressing Streptococcus pyogenes Cas9 were generated by transduction with the lentiviral vector pCF226 or pCF823, followed by selection on puromycin (1-2 pg / ml; Invivogen, #ant-pr-l). To generate LN229-Luc-mCherry-Cas9 cells for in vivo experiments, a two-round transduction strategy was employed. First, LN229 cells were transduced with luciferase-IRES- mCherry lentivirus (GeneCopoeia, #LPP-HLUC-Lv214-100-C). The mCherry+ cells were then isolated by flow cytometry and subsequently transduced with pCF823 to generate LN229-Luc- mCherry-Cas9 cells.CRJSPR-Safe packaging cells

[0172] To prevent viral packaging cells from dying when transfecting all-in-one Cas9-sgRNA vectors expressing sgCIDEs, HEK293T human embryonic kidney cells (293FT; Thermo Fisher Scientific, #R70007; RRID:CVCL_6911) were transduced with the lentiviral vector pCF525- AcrIIA4 to stably express the anti-CRISPR protein AcrIIA4, a potent inhibitor of Streptococcus pyogenes Cas9. Transduced cells were selected on Hygromycin B (400 pg / ml; Thermo Fisher Scientific, #10687010) and the resulting cell line termed “CRISPR-Safe” packaging cells.Competitive proliferation assay

[0173] To assay the effect of Genome Shredding on cell viability and depletion, U-251, LN- 229, T98G, LN-18, SF11411, NHA, GL261, and DF-1 cells with or without Cas9 expression were transduced at -50-70% efficiency with lentiviral vectors to deliver mNeonGreen tagged sgRNAs (sgNTs, sgRPAls, sgCIDEs). The percentage of mNeonGreen+ cells was quantified by flow cytometry (Attune NxT flow cytometer, Thermo Fisher Scientific) at 1, 2, 3, 5, and 7 days post-transduction (or as indicated). To test the effect of Genome Shredding using an all-in-one vector, U-251 cells were transduced at -30% efficiency with pCF827 lentiviral vector to deliver mNeonGreen tagged Cas9 and sgRNAs (sgNT-1 / 2 / 3, sgCIDE- 1 / 2 / 3). The percentage of mNeonGreen+ cells was quantified by flow cytometry at 2, 4, and 6 days post transduction. To assay the cell elimination efficacy of Genome Shredding using TiCas9, TiCas9-expressing U- 251 cells were stably transduced at approximately 50-70% efficiency with lentiviral vectors carrying mNeonGreen-tagged sgRNAs (sgNT-1 / 2 / 3, sgCIDE- 1 / 2 / 3). Cells were then treated with 4-OHT (1 pM, renewed every day). The percentage of mNeonGreen+ cells was quantified by flow cytometry before treatment (day -1) as well as at day 1, 3, and 5 post 4-OHT treatment.Comet assay

[0174] To detect DNA double-strand breaks (DSBs), a neutral comet assay was conducted following the manufacturer’s protocol of the Comet Assay kit (Trevigen; Cat# 4250-050-K). In short, Cas9-expressing U-251 and LN-229 cells were transduced with lentiviral vectors expressing sgNTs or sgCIDEs (with 5pg / ml of polybrene). Cells were collected 24 hours posttransduction at a concentration of lxl0A5 / ml, embedded in molten LMAgarose at a ratio of 1 : 10 (v / v), and transferred onto a CometSlide. Slides were immersed in Lysis solution at 4°C overnight. The next day, slides were washed in IX TBE buffer, and gel electrophoresis usingTBE buffer was performed (22 volts, 20 minutes). Slides were then fixed in 70% ethanol for 5 minutes, dried at 37°C for 15 minutes, and stained with SYBR safe for 30 minutes. Images were taken using fluorescence microscopy. Comet tail length (in pixels) of individual cells was quantified using Imaged. Each comet tail length was measured from the bottom of the comet head.Live cell imaging with Incucyte SX5 Live-Cell Analysis Instrument

[0175] Cas9-expressing U-251 and LN-229 were plated at low density in a 96 well plate a day before live cell imaging. Right before transferring the 96-well plate into an Incucyte SX5 Live- Cell Analysis Instrument, cells were transduced with lentiviral vectors to express mNeonGreen- tagged sgRNAs (sgNT-1 / 2 / 3 and sgCIDE-1 / 2 / 3 / 6 / 8 / 10). Propidium iodide (PI) staining solution (Invitrogen) was also added to each well to visualize dying cells. Images (20X) were taken every 30 minutes until day 5 post-transduction. The number of mNeonGreen+ cells at each timepoint was calculated by the Incucyte SX5 Live-Cell Analysis Instrument.Cell viability assay

[0176] To determine the effect of different treatments on cell viability, 1,000-2,000 Cas9- expressing U-251, LN-229, T98G, and LN- 18 cells were plated in a 96-well plate. After an overnight incubation, cells were treated with different concentrations of TMZ (12.5-400 pM) or different dilutions of lentiviral vectors expressing sgNTs or sgCIDEs (with 5 pg / ml of polybrene). Cell viability was assessed at 1, 3, and 5 days post-treatment using the CellTiter-Glo Luminescent Cell Viability Assay (Promega, #G7572). To determine the effect of Genome Shredding on the viability of GSCs (GBM12, GBM85, and GBM123), 2000-4000 cells were plated in 96-well laminin-coated plates. After an overnight incubation, cells were treated with DMSO, 50 pM TMZ, or all-in-one lentiviral vectors to deliver Cas9 and sgRNAs (sgNT-1 / 2 / 3, sgCIDE- 1 / 2 / 3; with 5 pg / ml of polybrene). Cell viability was measured at 5 days post-treatment using the CellTiter-Glo assay. In brief, 20 pl of CellTiter-Glo reagent was added to the well containing cells in 100 pl of cell culture medium. The plates were incubated at room temperature for 25 minutes and the luminescent signal was measured using a multimode microplate reader. The assay was performed in triplicate for each experiment.Cell cycle analysis

[0177] To conduct cell cycle analysis and measure DNA fragmentation (sub-Gl), ~2xlOA4 Cas9-expressing U-251, LN-229, T98G, and LN- 18 cells were plated in 24 well plates. The next day, cells were treated with DMSO, TMZ (50 pM), or lentiviral vectors expressing mNeonGreen-tagged sgNTs, sgOR2B6s, sgRPAls, or sgCIDEs (with 5 pg / ml of polybrene). Cells were collected at 1-, 2-, 3-, 5-days post-treatment for cell cycle analysis using FxCycle PI / RNase Staining Solution (Invitrogen, #F 10797). In short, cells were fixed in cold 70% ethanol for at least 1 hour at -20°C, washed twice with PBS, and resuspended in 500 pL of FxCycle PI / RNase Staining Solution. Samples were incubated for 15 minutes at room temperature before being analyzed by a flow cytometer.Colony formation assay

[0178] To perform colony formation assays, 100, 1,000, and 10,000 Cas9-expressing U-251, LN-229, T98G, and LN- 18 cells were plated in 6 well plates. Cells were treated with DMSO, TMZ (50 pM), or lentiviral vectors expressing sgNTs or sgCIDEs (with 5 pg / ml of polybrene) the next day. The cell culture medium was changed every 3-5 days. Two weeks after the treatment, cells were fixed in 4% PFA for 15 minutes at room temperature, washed with PBS twice, and stained with 0.005% crystal violet for 20 minutes. Cells were then washed with dFLO three times before imaging.Virus preparation for in vivo experiments

[0179] pCF821-sgNT-l and pCF821-sgCIDE-l lentivirus using six 10 cm plates of HEK293T cells were generated. Per 10 cm plate, 8 pg of lentiviral vector, 4 pg of psPAX2 and 2 pg of pMD2.G were mixed in 2 ml Opti-MEM. Then, 42 pg polyethyleneimine (PEI; Polysciences #23966) was added to the mixture and incubated at room temperature for 25 minutes. The transfection mixture was then added to the 10 cm plate. Media was changed 16-18 hours after transfection, and virus was harvested 48 hours post-transfection. Viral supernatants were collected and filtered through 0.45 pm PES membrane filters to remove debris. The filtered supernatants were pooled and centrifuged at 25,000 rpm, 4°C for 2 hours using the SW28 rotor in an ultracentrifuge (Beckman). After centrifugation, the viral pellet was resuspended in 150 pl of D-PBS and transferred to a 1.5 ml tube. The viral suspension was then centrifuged at 10,000rpm, 4°C for 2 minutes to remove any remaining debris. Finally, the virus was aliquoted into 25 pl / tube and stored at -80°C until use.Orthotopic tumor xenografts

[0180] All animal experiments were conducted in accordance with the guidelines and regulations of the Institutional Animal Care and Use Committees at the University of California San Francisco. For the in vivo competitive experiments, intracranial tumor xenografts were established by injecting 2 x 10A5 LN229-Luc-mCherry-Cas9 cells (in 2 pl; 1 pl / minute) into the striatum of female athymic nude mice (Foxnlnu / nu; IMSR_JAX:007850; aged 5-7 weeks) at the following coordinate: AP +0.5 mm, ML +1.8 mm, DV -4mm relative to bregma, using a 27 gauge Hamilton syringe and automated stereotaxic equipment (NEUROSTAR Drill and Injection Robot). After one week, 5 pl or 15 pl of concentrated lentivirus (pCF821-sgNTl or pCF821- sgCIDEl) was injected into the tumor at the same coordinate via a CED needle at a rate of 1 pl / minute. For the TiCas9 experiment, intracranial tumor xenografts were established by injecting 3 x 10A5 U-251-TiCas9-sgNTl or U-251-TiCas9-sgCIDEl cells (in 3 pl; 1 pl / minute) as described above. After one week, tumor-bearing mice were injected intraperitoneally with 100 mg / kg Tamoxifen (dissolved in corn oil at 20 mg / ml) per day for 5 consecutive days to induce TiCas9 expression. Animals were euthanized according to university animal care guidelines upon the appearance of overt disease symptoms. For the in vivo competitive experiments, brain tissues were collected, dissociated with the Worthington Papain Dissociation System (LK003150) according to the manufacturer's protocol, and analyzed by flow cytometry.CRISPR sgRNA library composition

[0181] For the CRISPR screen, an sgRNA library consisting of a total of 4620 unique sgRNAs, was assembled. The sgRNA library included personalized recurrent GBM-specific sgRNAs, personalized cancer-specific repeat sgRNAs, positive controls (e.g., sgCIDEs), and negative controls (e.g., non-targeting sgRNAs). All cancer-specific guide RNA sets contain sgRNAs that are specific for the recurrent hypermutated GBM PDCL SF11411 (sequencing replicate 1 and 2). For the oligo pool (60-mers, Twist Bioscience), some sgRNAs were included multiple times for better representation (4965 oligos total, as shown in Table 3 above).

[0182] Concretely, the CRISPR library contained the following sets of sgRNAs: 1) Previously tested sgCIDEs. 2) Non-targeting sgRNAs. This set contains 80 sgRNAs that have no exacttarget or positional off-target up to Hamming Distance (HD) 3 in the human genome. 3) Safeharbor sgRNAs. A set of 1000 safe-harbor sgRNAs that are specific to their single target site in the human genome and have no off-targets up to HD 3 in the hg38 assembly. 4) Cancer-specific gene targeting sgRNAs. The “all-genes” set contains SF11411-specific, curated sgRNAs that target coding exons in the human genome. There are 1463 genes with 3165 sgRNAs. The “essential-genes” set is a subset of all-genes with sgRNAs that target known essential genes. There are 41 genes with 67 sgRNAs. 5) Cancer-specific MYC transcription factor binding site (TFBS) targeting sgRNAs. The “MYC-TFBS” set contains sgRNAs targeting MYC transcription factor binding sites (Encode 3). There are 95 MYC TFBS with 183 sgRNAs. 6) Cancer-specific CTCF targeting sgRNAs. The “CTCF” set contains sgRNAs targeting CCCTC-binding factor (CTCF) sites. There are 7 CTCF sites with 12 gRNAs. 7) Cancer-specific vista enhancer targeting sgRNAs. The “vista-enhancer” set contains sgRNAs targeting highly conserved vista enhancers. There are 83 vista enhancers with 130 sgRNAs. 8) Cancer-specific sgRNAs targeting repetitive sequences. The “cancer duplicate (a.k.a. cancer repeat)” set contains sgRNAs that are a) repetitive in the recurrent cancer and b) unique to the recurrent cancer. There are 10 cancer repeat sgRNAs that meet these criteria. The most repetitive sgRNA has 7 enumerated repeats at HD 0.Pooled sgRNA library generation and lentivirus production

[0183] A pooled oligo library was synthesized by Twist Biosciences and then PCR amplified using primers Amp_fw cttgtggaaaggacgaaacaccg, and Amp_rv gctatgctgtttccagcatagctc. The sgRNA inserts were then ligated into digested pCF822 backbone at a 5: 1 ratio via Gibson Assembly (NEB #E2611). The recombinant plasmids were precipitated with isopropanol and electroporated into Endura ElectroCompetent cells (Lucigen #60242). The cells were plated onto LB agar 245-mm square bioassay dishes (Corning #31111) supplemented with ampicillin (100 pg / mL) for 12-14 hours at 37°C. The colonies were harvested and plasmid DNA was extracted using the Qiagen Plasmid Maxi Kit (Qiagen #12163). To confirm the sgRNA distribution of the pooled library, an amplicon library containing the sgRNA cassettes was generated using NEBNext Ultra II Q5 Master Mix (NEB# M0544). The first step PCR primers contain partial Illumina adaptors: NGS fw ACACTCTTTCCCTACACGACGCTCTTCCGATCT ggactatcatatgcttaccg, and NGS rvTGACTGGAGTTCAGACGTGTGC TCTTCCGATCTggtgccactttttcaagttg. A second PCR was performed to attach Illumina adaptorsand to barcode samples. The amplicon library was then sequenced on Illumina platforms with at least lOOx coverage.

[0184] To produce lentivirus, HEK293FT cells were seeded at -75% confluency the day before transfection in a 10 cm plate containing 8 ml of DMEM with 10% FBS and 1% penicillin / streptomycin. Transfection was performed by mixing 14 pg of DNA (8 pg sgRNA library, 4 pg psPAX2, 2 pg pMD2.G) and 42 pg of PEI in 2 ml of Opti-MEM medium for each 10cm plate. The mixture was incubated for 25 minutes at room temperature before being added to cells. 24 hours post-transfection, the medium was changed to 5 ml of serum-free DMEM / F12 medium. 48 hours post-transfection, the virus was harvested and filtered through a 0.45 pm syringe filter. Aliquots were stored at -80°C.CRISPR screen

[0185] SF11411 and NHA cells were transduced with the pooled sgRNA library at a MOI <0.3. Sufficient cells were transduced to achieve a presentation of at least 2000 cells per sgRNA. One day post-transduction, half of the cells were collected for genomic DNA (gDNA) extraction (TO samples) using the Quick-DNA Miniprep Plus Kit (Zymo Research #D4068). Two days after transduction, puromycin (2 pg / ml) was added to the cells for selection. Throughout the screen, cells were split at a density to maintain a representation of at least 1000 cells per sgRNA, and cell count was performed to monitor cell growth.

[0186] Four weeks post-transduction (-14 cell doubling), cells were collected for gDNA extraction using the same kit mentioned above. sgRNA cassettes were PCR amplified following the protocol described earlier. The amplicon library was then sequenced on Illumina platforms with at least 500x coverage. The screen analysis was performed using MAGeCK (as described by, for example, Li et al., “MAGeCK enables robust identification of essential genes from genome-scale CRISPR / Cas9 knockout screens”, Genome Biol. 2014;15(12):554).Computational quantification of sgCIDE targets

[0187] Reference genomes for the hg38, mm 10, and galGal6 assemblies of the human, mouse, and chicken genomes were downloaded from the UCSC genome browser. FASTA files were extracted and the GuideScan software was utilized to determine the identity, coordinates, and target occurrences of sgRNAs in the SpyCas9 CRISPR system.Variant calling

[0188] Variant calling was conducted using the Mutect2 application from the GATK software. Variant calling was done on whole genome sequencing data of the patient’s wildtype, primary cancer, and recurrent cancer genome. As the whole genome sequencing data was computed in the hgl9 assembly of the human genome, variant calling with Mutect2 also used the hgl9 assembly as its reference. Variants called in the primary and recurrent tumors were filtered against the variants in the patient’s native genome, resulting in a list of cancer specific variants. Variant coordinates were assessed against coordinates of human exons in protein coding genes. If a variant coordinate overlapped an exon in a protein coding gene, the variant was labeled a coding variant. If the variant did not overlap an exon, then the variant was labeled non-coding.Suspected pathogenic variants

[0189] To identify variants that are likely pathogenic, the identified patient native genome, patient primary GBM, and patient recurrent GBM variants were intersected with the ClinVar (NCBI) database. Specifically, it was assessed whether patient-specific variants had a coordinate match with variants noted in ClinVar; however, it was not required that the identity of the mutation in the data matches the exact mutation reported in ClinVar. As possibly pathogenic, ClinVar entries from the following categories were considered: Pathogenic, Likely_pathogenic, risk factor, Pathogenic / Likely_pathogenic, Pathogenic|_risk_factor, and Conflicting interpretations of-pathogenicity for cases where the majority of the reported studies support the variant being pathogenic / likely_pathogenic / risk_factor.Determination of cancer specific sgRNAs

[0190] Variants were computed using Mutect2 from the GATK software. Variants passing a threshold of 6.3 TLOD score were deemed to be variants of high confidence. These variants were compared against coordinates of essential genes. Coordinates overlapping essential genes had sgRNAs computed that included the variant. The sgRNAs with the variant were assessed against known sgRNAs in the human genome to determine if the resulting sgRNA were specific to the cancer. An analysis of the variants to verify the TMZ chemotherapy signature was done with custom python scripts that processed the output files of the Mutect2 software.Circos plots

[0191] Circos plots were generated using the Circa software (OMGenomics). Input files for circos plots were generated from sgRNA genomic coordinates that were determined using the GuideScan software.Statistical analysis

[0192] Statistical analyses were performed using Prism (GraphPad), Excel (Microsoft), and SciPy (version 1.6.2) software. Quantified data represent mean ± standard deviation (SD), unless otherwise indicated. Values of “n” are listed in relevant methods and / or figure legends. Multiple hypothesis testing correction was computed using the Benjamini -Hochberg procedure. Nonparametric statistical tests (e.g. Wilcoxon Rank Sum) were used in instances where the underlying distribution of study was unknown or unlikely to follow Gaussian behavior.

[0193] It is understood that the examples and embodiments described herein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of this application and scope of the appended claims. All publications, patents, and patent applications cited herein are hereby incorporated by reference in their entirety for all purposes.

Claims

WHAT IS CLAIMED IS:

1. A method of killing cancer cells in an individual, wherein the individual has cancer cells comprising a plurality of copies of cancer- selective induced sequence repeats in a genome of the cancer cells, the method comprising delivering to the cancer cells:(i) guide RNAs (gRNAs) that specifically target the cancer-selective induced sequence repeats; and(ii) a gRNA-guided nuclease, wherein the gRNA-guided nuclease cleaves the genome of the cancer cells at the cancer-selective induced sequence repeats, thereby selectively killing the cancer cells.

2. The method of claim 1, wherein the cancer-selective induced sequence repeat has at least 5 copies in the genome of the cancer cell.

3. The method of claim 1, wherein the cancer-selective induced sequence repeat has between 5 to 100 copies in the genome of the cancer cell.

4. The method of any of claims 1-3, further comprising sequencing the genome of the cancer cells prior to and / or after the delivering, thereby identifying the cancer- selective induced sequence repeats.

5. The method of any of claims 1-4 further comprising, prior to the delivering:(i) sequencing non-cancerous cells of the individual;(ii) identifying cancer-selective induced sequence repeats in a genome of the cancer cells not substantially present in non-cancerous cells; and(iii) selecting gRNAs that specifically target the cancer- selective induced sequence repeats not substantially present in non-cancerous cells.

6. The method of any one of claims 1-5, wherein the cancer-selective induced sequence repeats are caused by a genotoxic agent.

7. The method of any one of claims claim 1-6, wherein the method further comprises administering the genotoxic agent to the individual having the cancer cells prior to the delivering.

8. The method of claim 7, further comprising sequencing the genome of the cancer cells after administering the genotoxic agent to the individual having the cancer cells and prior to the delivering, thereby identifying cancer-selective induced sequence repeats.

9. The method of claim 7 or 8, wherein the genotoxic agent is administered to the individual having the cancer cells at a sub-lethal dose to the cancer cells.

10. The method of any one of claims 1-9, wherein the genotoxic agent is a chemotherapeutic agent.

11. The method of claim 10, wherein the chemotherapeutic agent is temozolomide, cisplastin, paclitaxel, or 5 -fluorouracil.

12. The method of any one of claims 1-9, wherein the genotoxic agent is radiation.

13. The method of any one of claims 1-12, wherein the delivering to the cancer cells comprises systemically administering the gRNA and gRNA-guided nuclease to the individual having the cancer cells.

14. The method of any one of claims 1-12, wherein the delivering to the cancer cells comprises locally administering the gRNA and gRNA-guided nuclease to the cancer cells.

15. The method of any one of claims 1-14, wherein the gRNA and gRNA-guided nuclease are delivered to the cancer cells in one or more vectors.

16. The method of claim 15, wherein the one or more vectors are viral vectors.

17. The method of claim 16, wherein the viral vectors are lentiviral vectors18. The method of claim 16, wherein the viral vectors are retroviral vectors.

19. The method of claim 18, wherein the retroviral vectors have independent replicative capacity.

20. The method of claim 18, wherein the retroviral vectors lack independent replicative capacity.

21. The method of any one of claims 1-16, wherein the gRNA is delivered to the cancer cells in a replicating retroviral vector and the gRNA-guided nuclease is delivered to the cancer cells in a non-replicating retroviral vector.

22. The method of any one of claims 1-21, wherein the cancer cells are primary cancer cells.

23. The method of any one of claims 1-21, wherein the cancer cells are metastatic cancer cells.

24. The method of any one of claims 1-21, wherein the cancer cells are glioblastoma cancer cells.

25. The method of any one of claims 1-24, wherein the gRNA guided nuclease is a Cas protein.

26. A composition comprising:(i) a plurality of gRNAs that specifically target cancer-selective induced sequence repeats in a genome of a cancer cell; and(ii) a gRNA-guided nuclease, wherein the gRNA-guided nuclease cleaves the genome of the cancer cells at the cancer- selective induced sequence repeats, thereby selectively killing the cancer cells.

27. The composition of claim 26, wherein the cancer-selective induced sequence repeat has between 5 to 100 copies in the genome of the cancer cell.

28. The composition of claim 26 or 27, wherein the cancer-selective induced sequence repeats are identified by sequencing the genome of the cancer cells.

29. The composition of any one of claims 26-28, wherein the cancer-selective induced sequence repeats are caused by a genotoxic agent.

30. The composition of claim 29, wherein the genotoxic agent is a chemotherapeutic agent.

31. The composition of claim 30, wherein the chemotherapeutic agent is temozolomide, cisplastin, paclitaxel, or 5 -fluorouracil.

32. The composition of any one of claims 29, wherein the genotoxic agent is radiation.

33. The composition of any one of claims 26-32, wherein the cancer cells are primary cancer cells.

34. The composition of any one of claims 26-32, wherein the cancer cells are metastatic cancer cells.

35. The composition of any one of claims 26-32, wherein the cancer cells are glioblastoma cancer cells.

36. The composition of any one of claims 26-35, wherein the gRNA guided nuclease is a Cas protein.

37. One or more vectors encoding the composition of any one of claims 26-36.

38. A method of detecting cancer-selective induced sequence repeats in a genome of cancer cells in an individual, the method comprising:(i) administering to the individual a genotoxic agent in a sub-lethal dose to the cancer cells; and(ii) sequencing the genome of the cancer cells, thereby detecting a plurality of copies of cancer- selective induced sequence repeats.

39. The method of claim 38, wherein the sequencing is performed prior to and / or after the administering.

40. The method of claim 38 or 39, wherein the genotoxic agent is a chemotherapeutic agent.

41. The method of claim 40, wherein the chemotherapeutic agent is temozolomide, cisplastin, paclitaxel, or 5 -fluorouracil.

42. The method of claim 38 or 39, wherein the genotoxic agent is radiation.

43. The method of any one of claims 38-42, wherein the cancer-selective induced sequence repeat has at least 5 copies in the genome of the cancer cell.

44. The method of any one of claims 38-42, wherein the cancer-selective induced sequence repeat has between 5 to 100 copies in the genome of the cancer cell.

45. The method of any of claims 38-44 further comprising:(i) sequencing non-cancerous cells of the individual prior to and / or after the administering;(ii) identifying the cancer-selective induced sequence repeats in a genome of the cancer cells not substantially present in non-cancerous cells; and(iii) selecting guide RNAs (gRNAs) that specifically target the cancer-selective induced sequence repeats not substantially present in non-cancerous cells.

46. The method of any one of claims 38-45, wherein the cancer cells are primary cancer cells.

47. The method of any one of claims 38-45, wherein the cancer cells are metastatic cancer cells.

48. The method of any one of claims 38-45, wherein the cancer cells are glioblastoma cancer cells.

Images