T cell isolation and culture
Human peripheral blood Leukopaks enriched for peripheral blood mononuclear cells (PBMCs) from deidentified healthy donors were purchased from StemCell Technologies (200-0092). CD3+ T cells were isolated using the EasySep Human T cell isolation kit (100-069) per the manufacturer-provided instructions. After isolation, T cells were seeded at 1 × 106 cells per ml and activated with anti-CD3/CD28 Dynabeads (Life Technologies, 40203D). T cells were maintained in culture at a density of 1 × 106 cells per ml throughout and cultured in complete X-VIVO-15 (cX-VIVO), consisting of X-VIVO 15 (Lonza Bioscience, 04-418Q) supplemented with 5% fetal calf serum (R&D systems, lot M19187), 5 ng mL−1 IL-7 and 5 ng mL−1 IL-15, unless otherwise indicated. CD4+CD25low Tconvs were isolated from washed PBMCs using the EasySep Human CD4+CD127lowCD25+ Treg isolation kit (StemCell Technologies, 18063) according to manufacturer’s protocol. Tconvs were activated using Immunocult CD2/CD3/CD28 T cell activation reagent (StemCell Technologies, 10990) at 12.5 μl ml−1. Tconvs were maintained in culture in cX-VIVO supplemented with 300 IU per ml of IL-2 and passaged every 2–3 days.
mRNA production
Seven CRISPRoff mRNA products with varying cap structure (m7G, Cap1 and ARCA) and codon optimization sequences were purchased from Aldevron and stored at −80 °C. The CRISPRoff 143533B (CRISPRoff v7) mRNA design was used for in vitro transcription (IVT) to make mRNA in house using Cap1 (TriLink Biotechnologies #N-7113-5). For IVT reactions, plasmids containing the CRISPRoff-V2.3 codon-optimized sequence, CRISPRon-TETv3, CRISPRon-TETv4 or CRISPRon-TETv5 were cloned into a mutated T7 promoter plasmid as previously described67,68. IVT templates were produced by PCR amplification of CRISPRoff-V2.3 or CRISPRon variants with the forward primer correcting the T7 mutation and reverse primer appending a poly(A) tail, such that the final template contained the wild-type T7 promoter, 5′ untranslated region (UTR) including Kozak sequence, codon-optimized CRISPRoff-V2.3 coding sequence or CRISPRon variants, 3′ UTR and 145-bp poly(A) tail. The PCR product was purified using solid-phase reversible immobilization bead selection and stored at −20 °C until use. IVT reactions were performed with the HiScribe T7 high-yield RNA synthesis kit (New England Biolabs, E2040S) under full substitution of pseudo-UTP and in presence of 4 mM CleanCap AG (which encodes Cap1) (TriLink Biotechnologies, N-7113-5) with the addition of RNAse Inhibitor (New England Biolabs, M0314L) and yeast inorganic pyrophosphatase (New England Biolabs, M2403L). Transcribed mRNA was purified with lithium chloride and eluted in water. After quantification by NanoDrop spectrophotometer and normalization to 1 μg μl−1, mRNA product was assessed on an Agilent 4200 TapeStation system and subsequently stored at −80 °C. CleanCap Cas9 mRNA was purchased from TriLink (L-7606).
Epigenetic or genetic editing with mRNA electroporation
For experiments using Cas9, CRISPRi or CRISPRoff mRNA, fresh CD3+ T cells were activated with a 1:1 bead-to-cell ratio with anti-CD3/CD28 Dynabeads (Life Technologies, 40203D) in the presence of 5 ng μl−1 IL-7 and 5 ng μl−1 IL-15 at 1 × 106 cells per ml. After 2 days of stimulation, T cells were magnetically debeaded, washed with PBS and resuspended in TheraPEAK P3 buffer with supplement (Lonza, G4LP3-126000) at 0.75 × 106 cells in 20 μl. Cas9, CRISPRi and CRISPRoff mRNA were added to 20 μl of cells at an equimolar ratio (1, 1.07 or 1.6 μg, respectively) with 2 μg of chemically modified sgRNA (Synthego) and cells were electroporated on a Lonza 4D Nucleofector using pulse code DS137. Immediately after electroporation, 80 μl of prewarmed cX-VIVO was added to each electroporation well and cells were incubated for 30 min in a CO2 incubator at 37 °C followed by the distribution of each electroporation reaction into three wells of a 96-well round-bottom plate. Each well was brought to 200 μl with cX-VIVO. Cells were maintained and expanded by the addition of cX-VIVO every 2 or 3 days and restimulated with ImmunoCult Human CD2/CD3/CD28 T cell activation reagent (StemCell Technologies, 10990) every 9–10 days at 6.25 μl ml−1. All sgRNA sequences used are listed in Supplementary Table 1.
To evaluate CRISPRoff mRNA designs, we electroporated seven CRISPRoff mRNA designs across a range of doses along with an sgRNA targeting CD151. We then compared the CRISPRoff activity data across constructs using ordinary least square regression. We modeled CD151 expression as a function of dose and mRNA variant and then computed a P value for the difference between mRNA variants across all doses using the standard error. CRISPRoff 7 was the most potent CRISPRoff mRNA variant as assessed by the degree of CD151 silencing across CRISPRoff doses.
For experiments using CRISPRon mRNA, CD4+CD25low (Tconv) cells were isolated from PBMCs and activated using Immunocult CD2/CD3/CD28 T cell activation reagent (StemCell Technologies, 10990) at 12.5 μl ml−1. Then, 2 days after activation, Tconvs were electroporated with 1.6 μg of CRISPRon mRNA and 2 μg of chemically modified sgRNA (Synthego) with pulse code DS137 as described above. After electroporation, Tconv cells were maintained and expanded in cX-VIVO supplemented with 300 U per ml
Extracellular and intracellular flow cytometry
For all experiments with flow cytometry as a readout looking at cell surface markers, 0.5 × 105–1 × 105 cells per condition were transferred to a round-bottom 96-well plate, centrifuged, washed once with 200 μl of cell staining buffer and stained with antibodies (1:50 dilution) for 20 min at 4 °C in the dark (antibodies are listed in Supplementary Table 1). Samples were measured using an Attune NXT cytometer with a 96-well autosampler (Invitrogen) and analyzed using FlowJo version 10.9.0 unless otherwise stated. For experiments measuring PD1, LAG3 and CD39 surface expression over time, cells were stimulated with ImmunoCult Human CD2/CD3/CD28 T cell activation reagent (StemCell Technologies, 10990) at 6.25 μl ml−1 24 h before flow cytometry readout. To obtain comparable live-cell counts between conditions, events were recorded using a fixed volume for all samples. To determine the number of cell divisions in in vitro experiments over time, we plated 0.16 × 106 cells into 96-well round-bottom wells following electroporation. We then counted cells on an Attune NXT Cytometer every 48 h or at each passage time.
For intracellular flow cytometry staining, 0.5 × 105–1 × 105 cells per condition were transferred to a 96-well V-bottom plate, centrifuged and washed once with 200 μl of staining buffer. Cells were resuspended in 30 μl of staining buffer containing Ghost Dye red 780 (Tonbo, 13-0865-T500) and antibodies targeting surface proteins of interest and stained for 20 min at 4 °C in the dark. After staining, cells were washed once with 170 μl of staining buffer and then resuspended in 50 µl of 1× FOXP3 Fix/Perm buffer (BioLegend, 421403) and incubated at room temperature for 30 min in the dark. After fixation, cells were permeabilized in 200 µl of 1× FOXP3 Perm buffer for 15 min at room temperature in the dark. After permeabilization, cells were spun down and washed once with 1× FOXP3 Perm buffer and then resuspended in 30 µl of 1× FOXP3 Perm buffer containing antibodies targeting intracellular proteins and stained in the dark at room temperature for 30 min. Following intracellular staining, cells were washed once with the addition of 170 µl of staining buffer and centrifuged at 300g for 5 min; the supernatant was removed. Cells were resuspended in 200 μl of staining buffer and then measured using the Attune NXT Cytometer with a 96-well autosampler.
Bulk RNA-seq
Human primary T cells were harvested 27 days after electroporation (CD55 and CD81) or 7 days after electroporation (FAS, PTPN2, RC3H1, SUV39H1 and MED12). A total of 1 × 106 cells were harvested per condition and RNA was isolated using a Quick-RNA MicroPrep Kit (Zymo, R1050). Isolated RNA was treated with TURBO DNase (Invitrogen, AM2239) and concentrated using an RNA clean and concentrate kit (Zymo, NC0622892). Library preparation was carried out using the QuantSeq FWD V2 with UDI Set A1 kit and UMI module (Lexogen, 191.96). Final libraries were assessed using a 4200 TapeStation (Agilent), quantified using the Qubit ds HS assay kit (Invitrogen) and sequenced as single-end 50-bp reads on a HiSeq 4000 (Illumina) or NextSeq 500 (Illumina).
RNA-seq data were aligned and counts were generated using the RNA-seq pipeline of nf-core (version 3.18)69. Raw sequencing reads were quantified using Salmon and summarized to gene-level counts using tximport. Differential gene expression analysis was conducted using the limma-voom framework, with donor variation included as a covariate in the statistical model. Gene expression was normalized using the trimmed mean of M values method and lowly expressed genes were filtered before analysis. DEGs were identified by comparing treated samples to NTCs, with significance criteria set at an adjusted P value < 0.05 (empirical Bayes moderated statistics with Benjamini–Hochberg correction) and absolute log2 fold change > 1. Results were visualized using volcano plots displaying log2 fold change versus −log10(adjusted P value), with genes colored on the basis of significance thresholds or target gene.
The MED12-KO RNA-seq data shown in Supplementary Fig. 2n were from previously generated data in the A.M. lab40 and are representative of CD4+ cells collected 8 days after activation with ImmunoCult Human CD2/CD3/CD28 T cell activation reagent (StemCell Technologies, 10990). Genotyping measured by NGS showed ~80% editing at MED12. The RNA-seq reads were analyzed as previously described and genes with a false discovery rate (FDR)-adjusted P value < 0.05 were considered significant. We correlated our MED12 CRISPRoff KD (76.5% KD) with this dataset to better match the degree of gene disruption as our MED12 Cas9 KO data only had ~55% indel-editing efficiency (Supplementary Fig. 2b) and we observed fewer DEGs than expected.
Off-target predictions were generated through the Integrated DNA Technology (IDT) CRISPR–Cas9 guide RNA checker (https://www.idtdna.com/site/order/designtool/index/CRISPR_SEQUENCE) for both Cas9 KO and CRISPRoff KD sgRNAs. For CRISPRoff sgRNAs, predicted off-target loci were filtered for sites that fell within ±1 kb of a gene promoter. We also performed ‘on-target, off-gene’ analyses by assessing effects on proximal genes that fell within a 100-kb window around the intended target. Only predicted off-target or proximal genes that had an absolute KD log fold change > 1 and adjusted P value < 0.05 were considered potential true off-target genes.
WGBS
We generated WGBS libraries for 12 samples, corresponding to EE (empty electroporation), NTC and targeting for CD55 across two donors, each done in technical replicate. Genomic DNA was extracted using the QIAamp DNA Mini Kit (Qiagen) and 250 ng of DNA was diluted to 2.27 ng μl−1 in 110 μl with 2 μl of 0.5% lambda DNA spike-in and sheared using a Covaris E220 evolution with intensifier for 50 s to an average length of ~500 bp. Sonicated DNA was recovered using the MinElute reaction cleanup kit (Qiagen), bisulfite conversion was performed using the EZ DNA methylation-Gold kit (Zymogen) and the resulting ssDNA was quantified on the Qubit ssDNA assay kit (Invitrogen). Library preparation was performed using the xGen methylation-sequencing DNA library preparation kit (IDT, 10009860) and xGen Normalase UDI primers plate 1 (IDT, 10009796). The prepared libraries were quantified on a 4200 TapeStation system (Agilent) and Qubit double-stranded DNA HS assay kit (Invitrogen). Libraries were sequenced using paired-end 150-bp reads on a NovaSeqX with a 10% PhiX spike-in to diversify the sample pools.
Raw WGBS-seq FASTQ files were processed using the nf-methylseq:2.6.0 pipeline69 with the default parameters along with the ‘–three_prime_clip_R1 10’ and ‘–three_prime_clip_r2 10’ options. Differential CpG DNA methylation analysis was performed using the methylKit R package70. CpG methylation data from Bismark coverage files was imported. To search for differentially methylated tiles, the ‘tileMethylCounts’ function was used with options ‘win.size = 1000’ and ‘step.size = 100’. DMRs were scored by the percentage methylation difference and q values were calculated using the ‘calculateDiffMeth’ function with ‘overdispersion = MN’ and ‘adjust = BH’ options using the replicates as a covariate in fitting the model. Results were visualized as Manhattan plots to display −log10-transformed P values associated with individual methylation tiling windows. Statistically significant DMRs with FDR < 0.05 (Benjamini–Hochberg) were colored on the basis of their methylation status. To visualize the methylation status at individual loci in Integrative Genomics Viewer (IGV), the base-level methylation status was extracted from BedGraph files from the nf-methylseq pipeline. Then, results were converted into an IGV-friendly format and data were displayed as bar charts, in which methylated regions were considered as a methylation percentage of 50–100% shown in the range of 0.5 to 1 in red and unmethylated regions were considered as a methylation percentage of 0–50% shown in the range of −1 to −0.5 in blue.
PBAT-seq visualization
PBAT-seq files at the FOXP3 locus were provided50. PBAT-seq analysis was conducted as previously described50. PBAT-seq tracks were visualized using a sliding binning strategy with a bin size of 1,500 and step size of 300 in ggplot2 (version 3.5.1).
Targeted bisulfite sequencing
A total of 200,000 cells were collected for conditions coelectroporated with CRISPRon and Guide 3 targeting the TSDR or two AAVS1 control targets, spun down and frozen at −80 °C. Targeted bisulfite sequencing was conducted by EpigenDX at two sites across the FOXP3 locus (TSDR or TSS) and off-target sites including IL2RA and IKZF2.
Lysis, RNA extraction and reverse transcription (RT) for qPCR
Cells were lysed and reverse-transcribed as described below. Briefly, 0.1 × 105–0.2 × 105 T cells were spun down in 96-well U-bottom plates and washed once with DPBS (without Ca2+ and Mg2+) (StemCell Technologies, 37350). Cells pellets were either frozen on dry ice and then stored at −80 °C until further use or lysed in 50 μl of complete RNA lysis buffer (9.6 mM Tris-HCl (pH 7.8), 3 U per ml proteinase K, 300 U per ml DNAse 1, 0.5 mM MgCl2, 0.44 mM CaCl2, 10 μM DTT and 0.1% (w/v) Triton X-114). Cells were incubated in RNA lysis buffer for 8 min at room temperature and then 30 μl of lysed cells were added to 3 μl of RNA stop solution in a new 96-well PCR plate (1 mM proteinase K inhibitor, 90 mM EGTA and 113 μM DTT in UltraPure water) and incubated for 3 min at room temperature to stop the lysis reaction. Then, 32 μl of RevertAid RT kit (Thermo Fisher, K1691) was aliquoted in a separate 96-well PCR plate and 8 μl of the lysis samples were added and mixed. RT was performed in a thermocycler with samples incubated at 25 °C for 10 min, 37 °C for 60 min and 95 °C for 5 min. Samples were either immediately used for qPCR or frozen and stored at −80 °C.
A master mix was made using TaqMan Fast advanced master mix for qPCR (Thermo Fisher, 4444557) and primer probes (IDT) that targeted either the housekeeping gene, GAPDH, or a gene of interest (FAS, MED12, PTPN2, RASA2, RC3H1 or SUV39H1). The final concentration of primers was 0.5 μM and that of probes was 0.25 μM. Next, 15 μl of master mix was added to 9.6 μl of complementary DNA from the RT reaction above and qPCR was performed in 5-μl reactions with technical quadruplicates in a 384-well plate format using the QuantStudio real-time PCR system (Thermo Fisher). To analyze the data, the Ct values of the technical quadruplicates were first averaged and then the ∆Ct was calculated by subtracting the GAPDH housekeeping Ct value from the averaged experimental values. The ∆∆Ct was then calculated from subtracting the ∆Ct of the NTC from the ∆Ct of the experimental samples. The fold change in gene expression was then calculated (2−∆∆Ct).
Multiplex editing with CRISPRoff in T cells
T cells were electroporated as described above. In multiplexed conditions with either CRISPRoff or Cas9 mRNA, each gene targeted received 1.6 μg of sgRNA. Cells were prepared for flow cytometry to collect live-cell counts and cell surface protein expression as described above. FCS files were analyzed using FlowJo (version 10.9.0) to create a gating scheme; cells were gated on lymphocytes, then single cells and then live cells and a positive or negative gate was drawn for each target gene. To calculate the proportion of cells with a given number of knocked down genes, the FlowJo workspace was read into R using the ‘flowCore’, ‘CytoML’ and ‘openCyto’ packages. First, each individual cell was recorded as either positive or negative for each target gene, with negative indicating KD of a target, on the basis of thresholds set in FlowJo. These thresholds were verified through visualization with the ‘ggcyto’ R package. Each cell was then annotated with the total number of genes successfully knocked down, from zero to five target genes. Lastly, the proportion of all cells for each number of knocked down genes was calculated.
Epi-edited CAR-T cell production
For nonviral integration of a BCMA-specific CAR transgene at the TRAC locus, T cells were isolated and stimulated as described above. After 48 h of stimulation, cells were magnetically debeaded and prepared for electroporation. Briefly, to prepare the guide RNA targeting the TRAC locus, aliquots of crRNA and trans-activating crRNA (Edit-R, Dharmacon Horizon) were thawed and mixed 1:1 (v/v) and annealed by incubation at 37 °C for 30 min to form an 80 μM solution. ssDNAenh was mixed into the gRNA solution at a 0.8:1 volume ratio before adding 40 μM Cas9-NLS (Berkely QB3 MacroLab) at a 1:1 (v/v) to attain a molar ratio of sgRNA-Cas9 of 2:1. Final RNP mixtures were incubated at 37 °C for 15–30 min, after which 50 pmol of RNP was used for each electroporation. The TRAC-targeting Cas9-RNP was mixed with a 2,923-nt ssDNA HDRt encoding the BCMA-specific CAR transgene, incubated for 10–15 min and then mixed with cells that were resuspended in 20 μl of TheraPEAK P3 buffer with supplement (Lonza, G4LP3-126000). The CRISPRoff mRNA and sgRNA targeting RASA2 or NTC (Synthego) were added on top of the cells last and then cells were electroporated using the Lonza 4D Nucleofector with pulse code EH115. For any experiments incorporating an RNP in the electroporation and mRNA, we used pulse code EH115, as this code is the most effective for RNPs (demonstrated elsewhere) while still maintaining efficiency for mRNAs. For RASA2 silencing with CRISPRoff mRNA, we codelivered either a chemically modified full-length (20 nt) sgRNA or a chemically modified truncated sgRNA (17 nt with the last base pair mismatched to effectively make a 16-nt truncated guide) (Synthego).
For integration of a CD19-specific CAR transgene at the TRAC locus using AAV6, Alt-R A.s. Cas12a (Cpf1) Ultra (IDT, 10001272) was mixed with a crRNA targeting the TRAC locus (IDT) at room temperature for 10–15 min. Cells that were resuspended in 20 μl of TheraPEAK P3 buffer with supplement (Lonza, G4LP3-126000) were mixed with TRAC-Cas12a-RNP and CRISPRoff mRNA and sgRNA targeting either RASA2 or an NTC were added on top of the cells and electroporated using pulse code EH115. At 30 min after electroporation, cells were transduced with AAV encoding the CD19-CAR as previously described52. The AAV-ITR plasmids containing the 1928z CAR transgene and TRAC-targeting homology arms for HDR was packaged into AAV6 by transfection of HEK293T cells together with pHelper and pAAV Rep-Cap plasmids using polyethylenimine. AAVs were further purified using iodixanol gradient ultracentrifugation. AAVs were tittered using qPCR on DNase I (New England Biolabs)-treated, proteinase K (Qiagen)-digested samples. qPCR was performed with SsoFast EvaGreen Supermix (BioRad, 1725201) on a StepOnePlus real-time PCR System (Applied Biosystems). AAV was added to the cells at a multiplicity of infection of 1 × 105 and cells were incubated overnight in serum-free medium. Then, 1 day after electroporation, the AAV-containing medium was removed and the edited T cells were resuspended in fresh cX-VIVO and expanded using standard culturing conditions. The KI efficiency for both nonviral-mediated HDRT and AAV HDRT KI was evaluated by flow cytometry several days later.
Digital droplet PCR (ddPCR)
Genomic DNA from 1 × 106–2 × 106 cells was purified using the QIAamp DNA mini kit (Qiagen) following the manufacturer’s protocol. DNA was quantified using the NanoDrop One (Thermo Fisher Scientific). All DNA samples were digested with HindIII in 10× rCutSmart buffer (New England Biolabs) before the ddPCR. A ddPCR assay was designed to measure the occurrence of balanced translocations between TRAC and RASA2. The assays used a pair of primers targeting a balanced translocation with TRAC on the 5′ end and RASA2 on the 3′ end and a fluorescent FAM probe. A pair of primers targeting the housekeeping gene RPP30 were included as a reference using a fluorescent HEX probe. The percentage of the translocation occurrences was calculated on the basis of the number of FAM+ droplets normalized to the HEX+ droplets.
ddPCR was performed using a QX200 ddPCR system (BioRad) following the manufacturer’s protocols. The reaction mix consisted of ddPCR Supermix for probes (no dUTP; BioRad), 900 nM of each primer, 300 nM of the FAM probe, 450 nM of the HEX probe and 400 ng of purified, digested genomic DNA. A 20-µl PCR reaction was used to generate lipid droplets with an automated droplet generator (BioRad). PCR amplification was performed using the following conditions: 95 °C for 5 min and 42 cycles of 94 °C for 30 s (ramp: 2.5 °C s−1) and 62 °C for 1 min (ramp: 2.5 °C s−1), followed by an enzyme deactivation at 98 °C for 5 min. Readout was performed with QX200 droplet reader (BioRad) and ddPCR droplet reader oil (BioRad). Data analysis was conducted with the QX manager software (BioRad) and thresholds were set manually to obtain the number of positive droplets for each channel.
Western blotting
For immunoblotting experiments, 2 × 106–3 × 106 cells were harvested, resuspended in 70 μl of Pierce radioimmunoprecipitation assay buffer (Thermo Fisher, 89901) supplemented with protease and phosphatase inhibitor cocktail (Fisher Scientific, 78440) and incubated at 4 °C for 40 min. The protein concentrations were determined using the Qubit protein and protein broad-range assay kits (Invitrogen, Q33211). Then, 15 μg of protein per sample was loaded onto 4–15% Tris–glycine SDS gels (BioRad) followed by transfer to PVDF membrane (BioRad) using the Biorad Trans-Blot transfer system. After transfer, membranes were blocked with 5% (w/v) nonfat milk in PBS containing 0.1% Tween-20 for 30 min. Primary antibody incubations were performed for either 2 h at room temperature or overnight at 4 °C (antibodies provided in Supplementary Table 1).
In vitro repetitive stimulation assay
For in vitro cytotoxicity assays, we generated epi-silenced CAR-T cells with either TRAC BCMA-specific CAR KI using our nonviral approach or TRAC CD19-specific CAR-T cells using AAV as described above. For coculture assays, we generated CD19+ or BCMA+ nuclear-localized RFP+ A375 melanoma target cells. At 6 days after electroporation, 300 of these target cells were seeded in 50 μl of complete RPMI per well in a 384-well plate. Complete RMPI includes RPMI (Gibco, 21870076), 10% fetal calf serum (R&D systems, lot M19187), 1% L-glutamine, 1% penicillin–streptomycin, 10 mM HEPES solution (Sigma, H0887) and 1 mM sodium pyruvate (Gibco, 11-360-070). The next morning, epi-silenced TRAC CD19-specific CAR-T cells or BCMA-CAR-T cells were counted and CAR expression was assessed by flow cytometry. CAR-T cell numbers were normalized and added to the target cells according to the indicated E:T cell ratios. The final per-well volume was 100 μl. Target cell counts were measured using the Incucyte live-cell imaging system (Sartorius) with imaging at 6-h intervals based on RFP expression.
For repetitive stimulation assays, CD19-A375 or BCMA-A375 target cells were seeded in complete RPMI medium 1 day before coculture. The next day, half of the medium was replaced with cX-VIVO and CD19-CAR-T cells or BCMA-CAR-T cells were seeded on top of the target cells at a 1:1 E:T ratio. This was repeated every 48 h for up to 5–7 stimulations. For each coculture, CAR-T cells were collected, strained through a 70-μm filter and counted using an Attune NXT Cytometer (Invitrogen). CAR expression was assessed using flow cytometry before each repetitive stimulation to normalize CAR-T cell counts between conditions. Before using the CAR-T cells for any downstream assay, the T cells were collected, counted and purified using EasySep Release human CD45 positive selection kit (StemCell, 100-0105).
Heritability of CRISPRoff-induced silencing in CAR-T cells in vivo
All mice for animal experiments were housed and used in accordance with ethical guidelines approved by the University of California, San Francisco (UCSF) Institutional Animal Care and Use Committee (IACUC). All animal experiments were performed with 8–12-week-old female NOD-scid IL2rg−/− (NSG) mice were purchased from Jax. To assess whether CRISPRoff-mediated silencing persists in CAR-T cells in vivo, we generated epi-edited Cas12a-compatible TRAC CD19-CAR-T cells in combination with CRISPRoff mRNA and a pool of three sgRNAs targeting CD151 or an NTC. Mice were injected with 1 × 106 A375 melanoma cells (engineered to express CD19) through subcutaneous injection to the right flank. Then, 1 week later, mice were randomized on the basis of width and length of the tumors and 7.5 × 105 epi-edited or control-edited CAR-T cells were injected into the tail vein. Mouse health and tumor growth were monitored over time. At 14 days after CAR-T cell injection, mice were humanely killed and tumors and spleens were isolated and prepared for flow cytometry.
CD19-epi-silenced CAR-T cells and Nalm6 xenograft model
We generated Cas12a-compatible TRAC CD19-CAR-T cells treated with CRISPRoff mRNA and three guides targeting RASA2 or an NTC as described previously. Mice were intravenously injected with 0.5 × 106 FFluc–GFP NALM6 cells and then, 4 days later, injected with 0.1 × 106 RASA2-epi-silenced CD19-CAR-T cells or control-edited CD19-CAR-T cells. CRISPRoff silencing activity of RASA2 in CAR-T cells was validated using western blot or RT–qPCR before injection. If RASA2 silencing was not observed in CD19-CAR-T cells before injection (because of electroporation error), we excluded those conditions from analysis. Tumor burden was monitored using BLI over time and weight was assessed as were any signs of morbidity per our UCSF IACUC protocol guidelines. For all experiments, mice were randomized on the basis of the BLI signal from day 3 after Nalm6 injection to ensure equal tumor distribution in each group before T cells were transferred.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
