Lipopeptide-mediated Cas9 RNP delivery: A promising broad therapeutic strategy for safely removing deep-intronic variants in ABCA4

Abstract

Deep-intronic (DI) variants represent approximately 10%-12% of disease-causing genetic defects in ABCA4-associated Stargardt disease (STGD1). Although many of these DI variants are amenable to antisense oligonucleotide-based splicing-modulation therapy, no treatment is currently available. These molecules are mostly variant specific, limiting their applicability to a broader patient population. In this study, we investigated the therapeutic potential of the CRISPR-Cas9 system combined with the amphipathic lipopeptide C18:1-LAH5 for intracellular delivery to correct splicing defects caused by different DI variants within the same intron. The combination of these components facilitated efficient editing of two target introns (introns 30 and 36) of ABCA4 in which several recurrent DI variants are found. The partial removal of these introns did not affect ABCA4 splicing or its expression levels when assessed in two different human cellular models: fibroblasts and induced pluripotent stem cell-derived photoreceptor precursor cells (PPCs). Furthermore, the DNA editing in STGD1 patient-derived PPCs led to a ∼50% reduction of the pseudoexon-containing transcripts resulting from the c.4539+2001G>A variant in intron 30. Overall, we provide proof-of-concept evidence of the use of C18:1-LAH5 as a delivery system for therapeutic genome editing for ABCA4-associated DI variants, offering new opportunities for clinical translation.

Keywords: ABCA4 deep-intronic variants; CRISPR-Cas9 genome editing; MT: RNA/DNA Editing; Ribonucleoprotein; Stargardt disease; intron removal; lipopeptide; peptide-mediated delivery; retina.

Graphical abstract

Figure 1

Design of delivery approach and therapeutic strategy (A) Schematic representation of the proposed mechanism of internalization of the RPNC designed either for gene editing or partial intron removal (PIR). In the acidic endosomal compartments, the C18:1-LAH5 lipopeptide undergoes protonation, leading to the disassembly of the nanocomplexes. This protonation event triggers the release and subsequent escape of the payload from the endosomes, facilitating its delivery to the cytoplasm. Following this, the Cas9 ribonucleoprotein (RNP) enters the nucleus, aided by nuclear localization signals (NLSs) present on Cas9. (B) Schematic representation of intron 30 (upper panel) and intron 36 (lower panel) of ABCA4 gene. The red double-stranded DNA indicates the hot-spot area in which several intronic variants causing splicing defects have been reported. The PCR amplification primers are indicated with green arrows and the designed gRNAs are shown in their target position. The size of the full-length amplicon is indicated on the right. Below, the combination of the different gRNAs together with the scheme of the expected fragment and its size at genomic level are represented.

Figure 2

Characterization of RPNCs at different ratios of RNP and C18:1-LAH5 lipopeptide (A) The particle size and polydispersity index (PDI) of the RPNC prepared at increasing molar ratios of Cas9 RNP to C18:1-LAH5 lipopeptide (20–20 nM)/increasing molar ratios of lipopeptide (RPNC) were determined using DLS. Data shown as mean ± SD (n = 3). Z-Ave parameter represents the intensity-weighted average hydrodynamic diameter of the particles in the sample and therefore describes the particle size of a sample. The higher the Z-Ave value, the larger the particle size. (B) ζ-Potential of 20 nM Cas9/gRNA (RNP) complexed with increasing molar ratios of C18:1-LAH5 lipopeptide was measured; data shown as mean ± SD (n = 3 technical repeats). (C) Electrophoretic mobility shift assay (EMSA) showing the effect of increasing C18:1-LAH5 lipopeptide concentration on RPNC formation. Controls: 1:0:0 (only GFP Cas9), 0:1:0 (only ATTO550 gRNA), 1:1:0 (RNP [GFP Cas9 and ATTO 550 gRNA]), and 0:0:1 (only C18:1-LAH5 lipopeptide). Three experimental setups were employed: (1) a combination of GFP-Cas9, ATTO550-gRNA, and the lipopeptide (indicated as RNP); (2) GFP-Cas9 along with the lipopeptide (indicated as Cas9); (3) ATTO550-gRNA in conjunction with the lipopeptide (indicated as gRNA). Within the framework of each of these configurations, the resulting RNPs were complexed with the range of lipopeptide concentrations, ranging from 0 to 250 μM.

Figure 3

Cellular uptake of RPNC and functional delivery of Cas9 RNP as measured by heteroduplex cleavage with T7E1 at the targeted CCR5 locus HeLa cells were treated with two different RPNCs, consisting of RPNC including GFP-Cas9/gRNA/C18:1-LAH5, or non-labeled Cas9/ATTO550-gRNA/C18:1-LAH5, at a molar ratio of 1:150 for RNP formulation. Negative controls consisted of Cas9 RNP without the inclusion of C18:1-LAH5. The same contrast and brightness were used for control and C18:1-LAH5 conditions in each fluorescence channel. (A) HeLa cells were transfected with 20 nM eGFP-Cas9/gRNA and 20 nM non-labeled Cas9/ATTO550-gRNA either without (upper) or with (lower) C18:1-LAH5. Twenty-four hours after transfection, the lysosomes were stained with LysoTracker (red or green) for 30 min before fluorescence microscopy images were taken at 60× magnification. Scale bar, 50 μm. (B) Functional delivery was evaluated using a T7EI assay. Genomic DNA was extracted from both the control and treated samples, and an 800-nt fragment of the CCR5 locus was amplified by PCR. Subsequently, the PCR product was digested with T7EI to examine heteroduplex formation. The frequency of indels was calculated using the equation provided in the “materials and methods” section.

Figure 4

RPNC formulation optimization for gene editing in HEK293T stoplight and eGFP HEPA 1–6 reporter cells (A) Schematic illustration of HEK293T stoplight cells reporter construct. These cells constitutively express mCherry and will start to express eGFP upon introduction of a +1 +2 frameshift within the linker region downstream of the mCherry gene. (B) Gene-editing efficiency for both LAH5 peptide and C18:1-LAH5 lipopeptide is represented by the e-GFP induction. Formulations ranging from 0 to 5 μM peptide concentrations were tested on HEK293T stoplight cells and quantified by flow cytometry. Representative analysis of flow cytometry analysis can be found in Figure S18. (C) Schematic illustration of the mechanism for peptide-mediated RNP delivery and the mechanism of reporter eGFP HEPA 1–6 cells. The reduction in the eGFP signal serves as an indicator of gene editing. (D) The quantification of gene-editing efficiency was quantified using flow cytometry 5 days after transfection, wherein the percentage of eGFP knockout was measured. eGFP knockout efficiency of RPNC (RNP/peptide) compared with negative (NT) and positive (CRISPRMAX, in green) controls. To assess optimal RPNC dose for gene editing, eGFP HEPA 1–6 cells were exposed to RPNCs prepared with either LAH5 or C18:1-LAH5 lipopeptide at various molar ratios ranging from 1:50 to 1:250. Representative flow cytometry analysis can be found in Figure S19. Data are shown as the mean ± SD (n = 3), Tukey’s multiple comparisons test (ns, non-significant; ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Figure 5

Representation of the characterization, simultaneous delivery, and cellular uptake of RPNCs with two different gRNAs for the application of PIR (A) Particle size and ζ-potential characterization of RPNCs formulated with Cas9 and gRNA at a 1:2 M ratio and with increasing amounts of C18:1-LAH5 lipopeptide (molar ratios indicated as Cas9/gRNA1/gRNA2-lipopeptide), diameter values in nanometers (d.nm) and PDI shown as mean ± SD (n = 3), ζ-potential (mV) shown as mean ± SD (n = 3 technical replicates). (B) EMSA of RPNCs assembled with Cas9/RNP with two different gRNAs. GFP Cas9, ATTO550-gRNA1, and ATTO647-gRNA2 prepared using these labeled components were used as controls to visualize all separate components. Following complex formation in presence of non-labeled lipopeptide, the samples were analyzed using a 1.5% agarose gel. (C and D) RPNC characterization by NanoFCM, a flow cytometry method to determine size and fluorescence intensities of single nanoparticles >∼50 nm in size. (C) Free Cas9 RNP containing ATTO488-gRNA1 and ATTO647-gRNA2 show only a few events, most likely aggregated Cas9 RNP proteins with sizes >50 nm. (D) RPNC forming from Cas9 RNP and C18:1-LAH5 lipopeptide complexation showed more events, of which 84% were double positive for Cas9/ATTO488 gRNA1 and ATTO647. (E) Confocal imaging of time-dependent cellular uptake of RNP (GFP-Cas9, ATTO550-gRNA1 and ATTO647-gRNA2/C18:1-LAH5 [RPNC]) in HeLa cells. Scale bar, 30 μm.

Figure 6

Analysis of PIR efficacy of RPNC-mediated editing in fibroblasts at DNA level either in intron 30 or intron 36 of the ABCA4 gene (A) (Upper panel) A representative electrophoresis gel of the amplification of intron 30 by PCR in both control and patient fibroblast cell lines. Edited band 1 (E1) indicates the expected edited band after treating with gRNAs 30-1 and 30-2 while the edited band 2 (E2) represents the expected band after the editing with gRNAs 30-3 and 30-4. The 5′ UTR of RPE65 was amplified as loading control. (Lower panel) A graph chart representing the overall result of the conducted replicates (n = 4) indicating the percentage of the full-length (FL) amplicon, without editing and the percentage of editing (edited band, E) for each condition. Each bar is represented by the mean ± SD (n = 4). In addition, some unspecific amplifications were observed that did not correspond to the expected products following editing and amplification. (B) (Upper panel) A representative electrophoresis gel of the amplification of intron 36 of ABCA4 by PCR in control fibroblast. In this case, the edited band 1 (E4) represents the expected band after treating with RPNCs, including gRNAs 36-1 and 36-2, while edited band 2 (E3) indicates the expected edited band after gRNAs 36-1 and 36-3 mediated editing. The 5′ UTR of RPE65 was amplified as loading control. (Lower panel) A graph bar representing the percentage of FL, edited (E), and unspecific amplicons per condition. Each bar is represented by the mean ± SD (n = 2). MQ indicates the negative control of the PCR; NT, non-treated cells; Pept, cells treated with the lipopeptide but without Cas9 or gRNAs (negative control of the transfection). CHX indicates if cell were treated with (+) or without (−) cycloheximide 24 h before harvesting. Statistical significance with respect to the untreated condition (NT+) is indicated as ∗∗p < 0.01 or ∗∗∗∗p < 0.0001 using one-way ANOVA followed by Bonferroni correction.

Figure 7

Analysis of the effectiveness of the RPNC-mediated gene editing in photoreceptor precursor cells at genomic DNA level (A) Isogenic control PPCs and (B) patient-derived PPCs. (A) (Top) Representative electrophoresis gels of the amplification of the intron 30 (left) or intron 36 (right) of ABCA4 by PCR. In each, the FL amplicon and the expected edited bands (E1 or E4) after the treatment with the selected gRNAs (gRNAs 30 1 + 2 for intron 30 of ABCA4 and gRNAs 36 1 + 2 for intron 36 of ABCA4, respectively) were indicated on the right edge of the gel. The 5′ UTR of RPE65 was amplified as loading control. (Below) A graph chart representing the overall result indicating the percentage of the FL amplicon and of the edited band (E) for each condition. Each bar is represented by mean ± SD (n = 3). (B) On the left, a representative electrophoresis gel of the amplification of intron 30 of ABCA4 by PCR in patient-derived PPCs. FL amplicon and edited (E1) band were indicated at the right side of the gel. The 5′ UTR of RPE65 was amplified as loading control. On the right, a graph bar representing the average result in which each bar is represented by the mean ± SD (n = 3 for all conditions, except gRNAs 30 1 + 2 CHX+, n = 4). MQ indicates the negative control of the PCR; NT, non-treated cells; P + C, cells treated with lipopeptide and Cas9 RNP but not with gRNAs; Pept, cells treated only with lipopeptide; Cas9, cells treated only with Cas9 RNP. CHX indicates if cell were treated with (+) or without (−) cycloheximide 24 h before harvesting. In all graph bars, a statistical significance with respect to the untreated condition (NT+) is indicated as ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, and ∗∗∗∗p < 0.0001 using one-way ANOVA followed by Bonferroni correction.

Figure 8

Analysis of the effect of the genome editing of intron 30 of ABCA4 gene in PPCs at the RNA level (A) RT-PCR from exon 30 to exon 31 of ABCA4 (n = 3 for all conditions, except for gRNAs 1 + 2 CHX+, n = 4). (Top) A representative electrophoresis gel of the RT-PCR, below the bar graph showing the normalized PE-containing transcripts in comparison with the NT + condition of control or patient fibroblast. Each bar represents the mean +SD. (B) RT-PCR specifically amplifying the PE (n = 3 with the exception of gRNAs 30 1 + 2 CHX+, n = 4). (Top) A representative electrophoresis gel of a specific RT-PCR that specifically amplified the PE in patient PPCs (n = 3) is shown. (Bottom) The quantification of the transcripts was normalized against ACTB and the NT + condition (which is set to 1). (A and B) ACTB was amplified as loading control; MQ indicates the negative control of the PCR. CHX indicates if cell were treated with (+) or without (−) cycloheximide (CHX) 24 h prior to harvesting. Statistical significance with respect to the corresponding untreated condition (NT) is indicated. ∗p < 0.05 or ∗∗∗p < 0.001 (one-way ANOVA followed by Bonferroni correction). (C) qPCR analysis directly amplifying the PE in the control and patient line (n = 3 with exception of samples for gRNAs 30 1 + 2 whose n = 4). Each condition was normalized against GUSB and then compared with the NT+ condition of each cell line. (A–C) CHX indicates if cells were treated with (+) or without (−) CHX 24 h prior to harvesting. Statistical significance with respect to the untreated condition (NT+) is indicated as ∗p < 0.05 or ∗∗∗p < 0.001 by t test.