A platform to deliver single and bi-specific Cas9/guide RNA to perturb genes in vitro and in vivo

Abstract

Although CRISPR-Cas9 technology is poised to revolutionize the treatment of diseases with underlying genetic mutations, it faces some significant issues limiting clinical entry. They include low-efficiency in vivo systemic delivery and undesired off-target effects. Here, we demonstrate, by modifying Cas9 with phosphorothioate-DNA oligos (PSs), that one can efficiently deliver single and bi-specific CRISPR-Cas9/guide RNA (gRNA) dimers in vitro and in vivo with reduced off-target effects. We show that PS-Cas9/gRNA-mediated gene knockout preserves chimeric antigen receptor T cell viability and expansion in vitro and in vivo. PS-Cas9/gRNA mediates gene perturbation in patient-derived tumor organoids and mouse xenograft tumors, leading to potent tumor antitumor effects. Further, HER2 antibody-PS-Cas9/gRNA conjugate selectively perturbs targeted genes in HER2⁺ ovarian cancer xenografts in vivo. Moreover, we created bi-specific PS-Cas9 with two gRNAs to target two adjacent sequences of the same gene, leading to efficient targeted gene disruption ex vivo and in vivo with markedly reduced unintended gene perturbation. Thus, the cell-penetrating PS-Cas9/gRNA can achieve efficient systemic delivery and precision in gene disruption.

Keywords: CAR-T; Cas9; bi-specific; delivery; off-target effects; ovarian cancer.

Graphical abstract

Figure 1

Generation and characterization of cell-penetrating PS-Cas9/gRNA (A) Left panel, PS-Cas9-STAT3gRNA. A phosphorothioate-DNA oligo labeled with fluorescein isothiocyanate (FITC) was conjugated to a Cas9 protein. Right panel, representative flow cytometry plots of internalization of PS-Cas9-STAT3gRNA in Jurkat cells from three independent experiments (n = 3). (B) DNA sequencing showing the mutation frequency at the STAT3 loci in the PS-Cas9-STAT3gRNA-treated Jurkat cells. Cyan highlight, PAM; yellow highlight, insertion; red hyphen, deletion; lower case, mutation. (C) Western blotting shows the levels of STAT3 protein in the Jurkat cells treated with PS-Cas9 or PS-Cas9-STAT3gRNA. STAT3/Actin protein ratio was determined by the band intensities, quantified by Chemidoc MP imaging system and Image Lab software (Bio-Rad). (D) Representative flow cytometry plots showing penetration of PS-Cas9-Stat3gRNA in murine splenic CD3⁺ T cells with or without CD3/CD28 stimulation. The plots are representative of three independent experiments (n = 3). (E) Cell viability of murine splenic T cells on day 4 post PS-Cas9-CtrlgRNA treatment. (F) Representative western blotting showing the levels of STAT3 protein in murine splenic T cells receiving PS-Cas9 or PS-Cas9-Stat3gRNA treatment. The protein levels of STAT3 and Actin were quantified by the band intensity using ImageJ software.

Figure 2

PS-Cas9/gRNA-mediated PDCD1 gene perturbation preserves CD19 CAR-T viability and expansion in vitro and in vivo (A) CD19 CAR-T manufacturing and PS-Cas9-PDCD1gRNA-mediated PDCD1 gene perturbation. (B) Western blotting showing the level of PD-1 in PD-1-KO CD19 CAR-T cells. The protein levels were quantified by the band intensity using Chemidoc MP imaging system and Image Lab software (Bio-Rad), and subsequently PD-1 was normalized with Actin. (C) Representative flow cytometry plot indicating the level of surface PD-1 on PD-1-KO CD19 CAR-T cells. (D) Viability CD19 CAR-T cells with or without PS-Cas9-PDCD1gRNA treatments. (E) The intracellular interferon-γ levels of PD-1 KO and non-KO CD19 CAR-T cells as assessed by flow cytometry. (F) Degranulation of target tumor cells by PD-1 KO and non-KO CD19 CAR-T cells. (G) The cytolytic activity of PD-1 KO and non-KO CD19 CAR-T cells against PD-L1-expressing Raji cells at effector to target cells of 1:2. (H) Upper panel, the schematic of the experiment. The lower panel depicts bioluminescence measurements of tumor signals in Raji-PD-L1 tumor-bearing mice treated with non-KO and PD-1-KO CD19 CAR-T cells at various time points. (I) The level of truncated epidermal growth factor receptor (tEGFR) on the isolated PD-1 KO and non-KO CD19-CAR-T cells was assessed in CD45⁺CD3⁺ cell populations by flow cytometry. (E–G and I) Data shown are mean ± SD; unpaired student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005.

Figure 3

PS-Cas9/gRNA efficiently perturbs the PARG gene in patient-derived tumor organoids and mouse xenograft tumors (A) Representative fluorescent microscopy images of cell penetration of PS-Cas9-PARGgRNA (FITC labeled) in ovarian cancer PDOs D3 and D5 post treatment (n = 3). Scale bar, 100 μm. (B) Western blotting measuring the PARG protein levels in ovarian cancer PDOs 5 days post indicated treatments. (C) TIDE analysis showing the indel sizes and frequencies by the PARG gRNA in a pool of ovarian cancer PDOs treated with PS-Cas9-PARGgRNA. (D) Weight of OVCAR8 xenograft tumors from mice treated with PS-Cas9-CtrlgRNA or PS-Cas9-PARGgRNA systemically. Mean ± SEM is shown. Student’s t test. ∗p < 0.05. (E) Western blotting of the tumor tissue protein extracts showing PARG protein levels. All tumors in each group were pooled before protein extraction. (F) Western blotting of xenograft tumors from mice treated with either PS-Cas9-CtrlgRNA or PS-Cas9-PARGgRNA to determine the levels of poly-PARylation, ɣ-H2AX, and H2AX. (B, E, and F) The protein levels of PARG, PARylation, and ɣ-H2AX, as well as GAPDH and Actin, were quantified by the band intensity using the Chemidoc MP imaging system and Image Lab software (Bio-Rad), and PARG, PARylation, and ɣ-H2AX levels were subsequently normalized with GAPDH, Actin, or H2AX, respectively.

Figure 4

Antibody-PS-Cas9/gRNA conjugate selectively targets tumor cells in vitro and in vivo (A) Western blotting showing STAT3 protein levels in MCF7 (Her2⁻) and MCF7-Her2⁺ tumor cells after PS-Cas9-STAT3gRNA or trastuzumab-PS-Cas9-STAT3gRNA treatment. (B) TIDE analysis showing the indel sizes and frequencies by the STAT3 gRNA in MCF7-Her2⁺ cells treated with vehicle (control), PS-Cas9-STAT3gRNA, or trastuzumab-PS-Cas9-STAT3gRNA. (C) Western blotting of STAT3 protein levels in SKOV3 and OVCAR3 cells treated with vehicle (control), PS-Cas9-STAT3gRNA, or trastuzumab-PS-Cas9-STAT3gRNA. (D) Left panel, tumor growth kinetics of the SKOV3 xenograft tumors in mice systemically treated as indicated. Black arrows mark the days of treatments. Right panel, tumor weights at the endpoint. (E) Representative fluorescent IHC imaging showing the levels of Ki67 and STAT3 in the xenograft tumors. Red, Ki67; green, STAT3; blue, Hoechst 33342. Scale bar, 50 μm. The fluorescence intensities were quantified by Zen software and are presented in bar graphs on the right. (F) The bar graph showing the levels of STAT3 protein in the xenograft tumor extracts from the mice treated as indicated in (D). The STAT3 protein was quantified and normalized with GAPDH, as shown in Figure S5D. (A and C) STAT3/Actin protein ratio was determined by the band intensities, which were quantified by Chemidoc MP imaging system and Image Lab software (Bio-Rad). (D) Data shown are mean ± SEM. Tumor growth kinetics, two-way ANOVA. Tumor weight, unpaired student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005, ∗∗∗∗p < 0.001. (E and F) Data shown are mean ± SEM. Unpaired student’s t test, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.005.

Figure 5

Reducing off-target effects by dimer PS-Cas9/gRNAs ex vivo (A) Representative flow cytometry plot showing the internalization of the monomer or dimer of PS-Cas9-STAT3gRNA conjugates in human CD11b⁺ myeloid cells. n = 3 independent assays. (B) Representative T7E1 assays (n = 3) showing gene perturbation frequencies by monomer (PS-S-Cas9-STAT3gRNA2 or PS-AS-Cas9-STAT3gRNA3), two-monomer mixture (PS-S-Cas9-STAT3gRNAs mix), or bi-specific dimer (PS-Cas9-STAT3gRNA2+3) in CD11b⁺ cells. (C) Western blotting indicates the levels of STAT3 in human CD11b⁺ myeloid cells treated with monomer (PS-S-Cas9-STAT3gRNA2 or PS-AS-Cas9-STAT3gRNA3) or dimer (PS-Cas9-STAT3gRNA2+3). The STAT3 and Actin protein levels were quantified by the band intensity using the Chemidoc MP imaging system and Image Lab software (Bio-Rad). STAT3 levels were normalized with Actin. (D) T7E1 assays showing the off-target effects of the monomer, two-monomer mixture, or bi-specific dimer of PS-Cas9-STAT3gRNA conjugates in human PBMCs (from healthy donors, D). (B and D) Asterisk (∗) indicates the full length of PCR product amplified from the target region. Black arrows indicate the cleaved DNA fragments.

Figure 6

Targeting HPV E7 by cell-penetrating monomer or dimer PS-Cas9-E7gRNA in vitro and in vivo (A) Representative western blotting showing the effects of the indicated PS-Cas9-E7gRNA conjugates on E7 protein expression in CaSki cells. (B) The bar graph showing the effects of monomer and dimer PS-Cas9-E7gRNA conjugates on the sub-G1 phase in CaSki cells. n = 3 independent treatments. (C) The off-target effects by the monomer or dimer PS-Cas9-E7gRNA conjugate on chromosome 4 in CaSki cells were assessed by T7E1 assay. A representative image is shown. n = 3 independent assays. (D) The off-target effects of PS-Cas9-E7gRNA monomer (upper panel) or dimer (lower panel) on chromosome 4 in human PBMCs from four healthy donors (D1–4). (E) The effects of monomer or dimer PS-Cas9-E7gRNA conjugate systemic treatments on CaSki xenograft tumors. D0, tumor cell engraftment; D4, 9, and 11, PS-Cas9/gRNA treatments. Left panel, tumor growth kinetics; right panel, tumor weight at endpoint. (F) Representative immunohistochemical imaging of tumor tissues with anti-Ki67 (red) or anti-HPV E7 (green) antibodies. Scale bar, 50 μm. The fluorescence intensities of E7 and Ki67 were quantified by Zen software and are presented in bar graphs. (G) HPV E7 protein levels in the xenograft tumors in mice treated by the indicated PS-Cas9-E7gRNA conjugates. The E7 protein level was measured by western blotting and quantified by ImageJ software. The quantification of E7 and Tubulin is shown in the bar graph. (A) The protein levels of E7 and Actin were quantified by the band intensity using the Chemidoc MP imaging system and Image Lab software (Bio-Rad). E7 levels were normalized with Actin. (C and D) Asterisk (∗) indicates the full length of PCR product amplified from the target region. Black arrows indicate the cleaved DNA fragments. (E) Two-way ANOVA was used to analyze tumor growth kinetics. Student’s t test was used to analyze tumor weight. (F and G) Data shown are mean ± SEM. Student’s t test. ns, not significant; ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.005; ∗∗∗∗p < 0.001.