Development of Cellular Models to Study Efficiency and Safety of Gene Edition by Homologous Directed Recombination Using the CRISPR/Cas9 System

Abstract

In spite of the enormous potential of CRISPR/Cas in basic and applied science, the levels of undesired genomic modifications cells still remain mostly unknown and controversial. Nowadays, the efficiency and specificity of the cuts generated by CRISPR/Cas is the main concern. However, there are also other potential drawbacks when DNA donors are used for gene repair or gene knock-ins. These GE strategies should take into account not only the specificity of the nucleases, but also the fidelity of the DNA donor to carry out their function. The current methods to quantify the fidelity of DNA donor are costly and lack sensitivity to detect illegitimate DNA donor integrations. In this work, we have engineered two reporter cell lines (K562_SEWAS84 and K562GWP) that efficiently quantify both the on-target and the illegitimate DNA donor integrations in a WAS-locus targeting setting. K562_SEWAS84 cells allow the detection of both HDR-and HITI-based donor integration, while K562GWP cells only report HDR-based GE. To the best of our knowledge, these are the first reporter systems that allow the use of gRNAs targeting a relevant locus to measure efficacy and specificity of DNA donor-based GE strategies. By using these models, we have found that the specificity of HDR is independent of the delivery method and that the insertion of the target sequence into the DNA donor enhances efficiency but do not affect specificity. Finally, we have also shown that the higher the number of the target sites is, the higher the specificity and efficacy of GE will be.

Keywords: CRISPR/Cas9; DNA donor; dsRED; eGFP; efficacy; homologous directed recombination (HDR); models; on-target integration; safety; specificity.

Figure 1

Development of K562 SEWAS84 reporter cell line (eGFP-OFF/dsRED-ON). (A) Generation of the K562 SEWAS84 reporter cell line. Top: Representative scheme of the SEWAS84 lentiviral plasmid. Left: Representative plots of K562 cells transduced with SEWAS84 LVs at MOI=0.1. The eGFP+ population was sorted to generate the K562 SEWAS84 cell line expressing eGFP and containing 1 copy was obtained by single-cell cloning. (B) Scheme of the strategy for targeting the WAS sequences to silence eGFP in order to have an easy readout of the efficacy and specificity in the K562SEWAS84 cellular model. The sgRNA will cut in the WAS84 region of the integrated LV (top) promoting the on-target integration in the presence of appropriate DNA donors (bottom). The design of the DNA donor, harbouring an expression cassette for dsRED and a mutated eGFP flanked by the homology arms (5’HA and 3’HA), allow to measure efficiency of on-target donor integrations (% of GFP- dsRED+), as well as frequency of illegitimate insertion (% of eGFP+ dsRED+) as illustrated at the bottom of the figure. Nucleofection of plasmid in the absence of the CRISPR/Cas was used as control for background (se M&M for details).

Figure 2

Inclusion of the target sequence into the donor template increases the efficiency of on-target donor integration without increasing illegitimate integrations in K562 SEWAS84. (A) Scheme of CRISPR/Cas9 lentiviral plasmid expressing Cas9 and the gRNA (See M&M for details) (top) and different DNA donors (bottom): SdR*GFP donor (previously describe) and SdR*GFPWAS84 donor harbouring the CRISPR/Cas9 target site. (B) Drawing of the delivery methods used for the different donors (SdR*GFP donor or SdR*GFPWAS84) and the CRISPR/Cas9 system (Lg1SCas9) used alone as controls (1) and with the CRISPR/Cas9 system (Lg1SCas9) (2). Both systems were delivered to the target cells as plasmids by nucleofection. (C) Representative plots showing eGFP (FITC-A) and dsRED (PE-A) expression 48 h (top) and 30 days (bottom) of K562 SEWAS84 without any treatment (NT) and after nucleofection with donors only (SdR*GFP donor or SdR*GFPWAS84), and with L1SCas9 (all-in-one CRISPR/Cas9 lentiviral plasmid expressing Cas9 and the gRNA. See M&M for further detail) and SdR*GFP or SdR*GFPWAS84 donor as indicated on the top of each graph. with all-in-one CRISPR/Cas9 lentiviral plasmid expressing Cas9 and the xgRNA (See M&M for further detail) and SdR*GFP donor (left) or SdR*GFPWAS84 donor, (right). (D) Graphs showing the relative GE efficacy and specificity for both DNA donors. The relative efficacy of GE (left graph) was calculated by dividing the % of eGFP-dsRED+ cells (due to HDR or HITI) at day 30 by the % of double positive cells at 48 h (initial transfection efficacy). Right graph shows the frequency of illegitimate integration related to the efficacy of HDR. Values represent mean of three separate experiments and the error bars indicates the standard errors of the mean. The comparison between the different groups was carried out using the Mann–Whitney U test (* p < 0.05).

Figure 3

Development of K562GWP reporter cell lines (eGFP-ON/dsRED-OFF). (A) Scheme showing the procedure to generation of the K562GWP reporter cell lines. Top: Representative scheme of the GWP lentiviral plasmid used to produced LVs to transduce K562 cells (bottom). We used two different MOIs (1 and20) of the GWP LVs and obtained bulk populations harbouring 13VCN (13-K562GWP cell line) and 0.39 VCN. A representative plot of K562 cells transduced with GWP LV is shown to indicate absence of eGFP expression. The 0.39 VCN population was cloned and characterised in order to obtain 1 VCN (1-K562GWP reporter cell line). A scheme of the integrated vector (the target locus) is shown at the bottom. (B) Scheme of the strategy for targeting the WAS sequence to rescue eGFP expression through HDR using three different DNA donors; SEED, SECdR and WAS_SECdR. The sgRNA will cut in the WASI1 region (in purple) of the integrated LV (top) promoting the HDR process and giving rise to GFP+dsRED- cells (top-right) due to the presence of the SFFV-eGFP expression cassette flanked by a 5’HA and 3’HA. All donors have an additional dsRED expression cassette outside of the HAs which allow to measure specificity of HDR, asdonor insertions through a non-HDR mechanism will render GFP+ dsRED+ cells (expressed by the DNA donors), plot at the right.

Figure 4

CRISPR/Cas9 delivery systems achieved similar efficacies and specificities of HDR in K562GWP model. (A) Representative cytometry plots showing the differential fluorescence patterns of edited K562GWP cells, using different methods to deliver CRISPR/Cas9 system. The top of the figure illustrates the different delivery methods used to transfer CRISPR/Cas9 system to K562GWP cells: plasmid nucleofection, Cas9/gRNA ribonucleoprotein complexes (RNPs) nucleofection, and transduction with all-in-one Cas9/sgRNA IDLVs. The SEED plasmid donor was delivered by nucleofection, alone (as a negative control, left plots) or in combination with the CRISPR/Cas9. eGFP and dsRED levels were analysed at 48 h (upper plots) and at day 30 (lower plots). (B) Graphs showing the relative HDR efficacy (left) and specificity (right) of HDR for different delivery methods. The relative efficacy of HDR (left graph) was calculated by dividing the % of eGFP+dsRED- cells (on-target HDR) at day 30 by the % of double positive cells at 48 h (initial transfection efficacy). Right graph shows the frequency of illegitimate integration related to the efficacy of HDR. Values represent mean of three separate experiments and the error bars indicates the standard errors of the mean. The comparison between the different groups was carried out using the Mann–Whitney U test (* p < 0.05).

Figure 5

Efficacy and specificity of HDR in K562GWP cell line with different donor templates. (A) Representative cytometry plots showing the expression of eGFP (FITC-A) and dsRED (PE-A) in 13-K562GWP cells over time. Each of the donor templates (SEED, SECdR or WAS_SECdR), as well as the CRISPR/Cas9 system, was transferred to the target cells as plasmids by co-nucleofection. The expression of eGFP and dsRED was determined 48 h post-nucleofection, and 30 days later. (B) Graphs showing the relative HDR efficacy and specificity of HDR-GE for the different DNA donors. The relative efficacy of HDR (left graph) was calculated by dividing the % of eGFP+dsRED- cells (on-target HDR) at day 30 by the % of double positive cells at 48 h (initial transfection efficacy). Right graph shows the frequency of illegitimate integration related to the efficacy of HDR. Values represent mean of three separate experiments and the error bars indicates the standard errors of the mean. The comparison between the different groups was carried out using the Mann-Whitney U test (* p < 0.05).

Figure 6

Correlation of the efficacy of GE with the frequency of target sites. (A) Representative cytometry plots showing the expression of eGFP and dsREDon 13-K562GWP cells (13VCN) and 1-K562GWP (1VCN) over time. Reporter cells were nucleofected with WAS_SECdR donor alone or together with the Lg9SCas9 plasmid as indicated in the figure and the percentage of eGFP and dsRED expressing cells were determined 48 h (top plots) and 30 days (bottom plots) post-nucleofection. (B) Graphs showing the relative HDR efficacy (top) and specificity (bottom) of HDR for 13-K562GWP cells (13cc) and 1-K562GWP (1cc) reporter cells. The relative efficacy of HDR was calculated by dividing the % of eGFP+dsRED- cells (on-target HDR) at day 30 by the % of double positive cells at 48 h (initial transfection efficacy). The frequency of illegitimate integration is shown related to the efficacy of HDR. Values represent the mean of three separate experiments and the error bars indicate the standard errors of the mean. The comparison between the different groups was carried out using the Mann–Whitney U test (* p < 0.05).