Targeted delivery of CRISPR-Cas9 and transgenes enables complex immune cell engineering

Abstract

As genome engineering advances cell-based therapies, a versatile approach to introducing both CRISPR-Cas9 ribonucleoproteins (RNPs) and therapeutic transgenes into specific cells would be transformative. Autologous T cells expressing a chimeric antigen receptor (CAR) manufactured by viral transduction are approved to treat multiple blood cancers, but additional genetic modifications to alter cell programs will likely be required to treat solid tumors and for allogeneic cellular therapies. We have developed a one-step strategy using engineered lentiviral particles to introduce Cas9 RNPs and a CAR transgene into primary human T cells without electroporation. Furthermore, programming particle tropism allows us to target a specific cell type within a mixed cell population. As a proof-of-concept, we show that HIV-1 envelope targeted particles to edit CD4⁺ cells while sparing co-cultured CD8⁺ cells. This adaptable approach to immune cell engineering ex vivo provides a strategy applicable to the genetic modification of targeted somatic cells in vivo.

Keywords: CAR-T cells; CRISPR delivery; CRISPR-Cas9; precision genome editing; viral engineering; virus-like particles.

Figure 1.. Production and characterization of Cas9-VLPs

(A) Schematic of plasmids for Cas9-VLP production. GP, glycoprotein; LTR, long terminal repeat; LV, lentiviral transfer plasmid. (B) Schematic of an immature Cas9-VLP produced through transient transfection. An HIV-1 protease cleavable linker (SQNYPIVQ) was inserted between Gag and Cas9. (C) Western blot of Cas9-VLP content with various ratios of Gag-pol and Gag-Cas9 plasmids used for production. Anti-FLAG (Cas9) and anti-p24 (capsid, CA) antibodies were used for detection. (D) Quantification of Cas9-VLPs produced per transfected p100 dish by CA ELISA; n = 2 technical replicates. (E) Jurkats were treated with B2M-targeting Cas9-VLPs and the transducing units (TUs) per milliliter titer was calculated. (F) Percentage of B2M indels plotted against the multiplicity of infection (MOI) using a sigmoidal 4-parameter logistic fit. Indels were quantified using Synthego’s ICE analysis tool. (G) The predicted MOI for each Cas9-VLP formulation to achieve 50% indels, interpolated from (F). EC₅₀, effective concentration at which a drug gives a half-maximal response. n = 3 technical replicates (E and F), except for formulation A (n = 2) (F). Error bars indicate standard error of the mean (D–F) and 95% confidence interval (G). ND, not detected.

Figure 2.. Cas9-VLPs efficiently mediate genome editing

B2M-targeting Cas9-VLP treatment results in genome editing of Jurkat and A549 cells. (A) Flow cytometry quantification of B2M expression at 3, 6, and 8 days post-treatment (dpt). Heatmaps represent the mean of technical replicates; n = 3, except for A549 at 8 dpt (n = 2). The highest treatment dose = 10% of Cas9-VLPs produced in a p100 dish. (B) Amplicon sequencing quantification of indels 3 dpt. Control = tdTom298 sgRNA. n = 3 technical replicates, except for A549 treated with 10 × 10⁴ pg CA (n = 2). (C) Flow cytometry quantification of B2M expression and transduction (mNeonGreen⁺) 6 dpt. Non-targeting control = guideless Cas9-VLP. n = 3 technical replicates. (D) Treatment with Cas9-VLPs that target B2M but do not co-package a lentiviral genome. Amplicon sequencing quantification of indels 3 dpt. Control = tdTom298 sgRNA. n = 3 technical replicates, except for Jurkats treated with 10 × 10⁴ pg CA traceless B2M Cas9-VLPs (n = 2). (E) Schematic of hypothetical lentiviral insertion at the Cas9 RNP-induced DNA break. (F) PCR assessment of targeted lentiviral integration. DNA was isolated from 293T cells 3 dpt with B2M-targeting or non-targeting Cas9-VLPs, and indicated primer pairs were used for analysis. Error bars indicate standard error of the mean.

Figure 3.. Generation of highly engineered CAR-expressing primary human T cells using Cas9-VLPs

(A) Cas9 RNP nucleofection and Cas9-VLP treatment of primary human T cells. Flow cytometry quantification of the mNeonGreen transduction marker and B2M expression 7 dpt. (B) Viability, transduction, and B2M expression in primary human T cells treated with Cas9-VLPs. B2M expression is plotted for CD4⁺ (red squares) and CD8⁺ (blue circles) subpopulations. (C) Simultaneous treatment with two Cas9-VLP preparations results in multiplexed genome editing. Cas9-VLPs targeting B2M and Cas9-VLPs targeting TRAC were used to co-treat primary human T cells. Surface expression of B2M and TCR was assessed by flow cytometry 13 dpt. n = 2 biological replicates from independent donors were used (A–C), and representative flow cytometry plots are shown for 1 donor (A and C). (D) Schematic of a single-step method to generate highly engineered CAR-T cells. (E) Primary human T cells were treated with CAR-Cas9-VLPs targeting B2M (top panels) or TRAC (bottom panels). Knockout was assessed for both CD4⁺ (red squares) and CD8⁺ (blue circles) subpopulations 12 dpt. (F) CAR-T cells generated by Cas9-VLP treatment, or untreated primary human T cells, were co-cultured with CD19⁺ Nalm-6 cells, and cytotoxic killing activity was measured at 24 h. Error bars indicate standard error of the mean.

Figure 4.. HIV-1 envelope pseudotyping targets Cas9-VLP genome editing to CD4⁺ T cells

(A) A viral glycoprotein is essential for Cas9-VLP-mediated genome editing. 293T and Jurkat cells were treated with B2M-targeting Cas9-VLPs pseudotyped with VSV-G (Cas9-VLP), without VSV-G (bald Cas9-VLP) or were left untreated (No tx). Indels were quantified by amplicon sequencing 3 dpt, n = 3. (B) B2M-targeting Cas9-VLPs pseudotyped with the HIV-1 envelope glycoprotein (Env-Cas9-VLPs) were used to treat primary human T cells (a mixture of CD4⁺ and CD8⁺ cells). (C) Viability, transduction (mNeonGreen), and B2M expression were assessed for CD4⁺ (red squares) and CD8⁺ (blue circles) subpopulations 7 dpt. n = 2 biological replicates from independent donors were used (B and C) and representative flow cytometry plots are shown for 1 donor (B). Error bars indicate standard error of the mean.