Nanoparticle‐mediated genome editing in single‐cell embryos via peptide nucleic acids

Abstract Through preimplantation genetic diagnosis, genetic diseases can be detected during the early stages of embryogenesis, but effective treatments for many of these disorders are lacking. Gene editing could allow for correction of the underlying mutation during embryogenesis to prevent disease pathogenesis or even provide a cure. Here, we demonstrate that administration of peptide nucleic acids and single‐stranded donor DNA oligonucleotides encapsulated in poly(lactic‐co‐glycolic acid) (PLGA) nanoparticles to single‐cell embryos allows for editing of an eGFP‐beta globin fusion transgene. Blastocysts from treated embryos exhibit high levels of editing (~94%), normal physiological development, normal morphology, and no detected off‐target genomic effects. Treated embryos reimplanted to surrogate moms show normal growth without gross developmental abnormalities and with no identified off‐target effects. Mice from reimplanted embryos consistently show editing, characterized by mosaicism across multiple organs with some organ biopsies showing up to 100% editing. This proof‐of‐concept work demonstrates for the first time the use of peptide nucleic acid (PNA)/DNA nanoparticles as a means to achieve embryonic gene editing.


| INTRODUCTION
Genome editing has the potential to treat numerous genetic disorders.
Embryonic gene editing may have advantages over editing in adults due to the smaller number of cells, which are rapidly dividing to form the rest of the organism. Changes made in these stem cells may be durable and passed onto daughter cells, potentially allowing for high editing rates that may be maintained throughout adulthood. Embryonic gene editing could be used for many applications, such as the creation of genetic mouse models and improving the output of livestock products, and it may eventually have clinical utility during in vitro fertilization (IVF) for the treatment or prevention of human disease. Although preimplantation genetic diagnosis during IVF allows for selection of healthy embryos, there may be clinical scenarios where a higher number of potentially viable embryos for transfer may be desired. For example, due to higher numbers of affected embryos in X-linked and autosomal dominant genetic disorders, or due to advancing age of the oocytes of the biological mother and a decreased chance of a successful IVF cycle, families may wish for alternative treatments that could increase the number of healthy embryos available to improve the chance of the birth of a healthy child. In addition to couples that are known carriers of mutations, this approach may also be particularly applicable to couples undergoing IVF who are both homozygous for a monogenic disease.
Traditional embryonic gene editing approaches involve pronuclear microinjection or electroporation of engineered nucleases, such as CRISPR/Cas9. While capable of achieving high levels of gene editing, pronuclear microinjection is labor intensive, requires specially trained personnel, and both pronuclear microinjection and electroporation methods can result in significant embryo loss. [1][2][3][4][5] Recombinant AAV vectors have been shown to deliver CRISPR/Cas9 reagents through the zona pellucida (ZP) of mice without the need for electroporation or microinjection with high levels of gene editing without significant embryo loss. 6 However, integration of AAV into the genome can occur, 6,7 which may increase the possibility of off-target effects and limit potential applications for human embryonic gene editing.
Our approach uses nanoparticles (NPs) fabricated from a biodegradable polymer, poly(lactic-co-glycolic acid) (PLGA), which has been used in humans for over 40 years and is well known to be safe. To achieve gene correction, two active agents are loaded into the NPs: a triplex-forming peptide nucleic acid (PNA) that induces recombination at a specific, targeted site on a chromosome and a short donor DNA molecule that contains the desired gene sequence. Once administered, PLGA nanoparticles are taken up by cells where they degrade, releasing the loaded PNA and DNA. Once inside a cell, the engineered PNA binds to a specific genomic target site, forming a triplex, which induces DNA repair mechanisms that stimulate the recombination of the short, single-stranded donor DNA molecule containing the correct sequence, resulting in site-specific gene editing. This approach to gene editing has been previously shown to be capable of achieving sitespecific gene editing with extremely low levels of off-target sequence modifications and to be safe in both adult and fetal animals, [8][9][10][11][12][13] which may be an advantage for PLGA PNA/DNA-loaded NPs over nuclease-based gene editing techniques such as zinc-finger nucleases, CRISPR/ Cas9, and TALENs technologies, whose potential off-target, cytotoxic, and/or immunogenic effects may hinder in vivo gene editing applications. 14,15 In prior work, we demonstrated the feasibility and safety of PNA/DNA NP-mediated gene editing in adult mice and fetal mice.
Here, we sought to determine the feasibility, safety, and efficacy of embryonic gene editing mediated by treatment ex vivo with PNA/DNA-loaded NPs. We find that NP treatment at the single-cell zygote stage allows for normal physiological cell division with no significant cytotoxicity, and treated embryos are edited at substantial rates with no detected off-target effects above background by deep sequencing. Reimplanted murine embryos are capable of normal growth and differentiation with no observed developmental abnormalities. Reimplanted eGFP embryos also exhibit high levels of gene editing, albeit with significant mosaicism, and exhibit phenotypic expression of corrected eGFP, establishing the potential for embryonic gene editing with PNA/DNA NPs.  19 The PLGA NPs used for delivery, which were created with a double-emulsion solvent evaporation technique modified to encapsulate PNA and DNA oligomers as previously described, 8,20 had an average hydrodynamic diameter around 250 nm; additional characterization data are included in Table S1 and imaging of NPs is included in Figure S1.
Given the size of the PLGA NPs used, to investigate whether the NPs were capable of penetrating the ZP of fertilized mouse oocytes, NPs loaded with a fluorescent dye (coumarin-6, C6) were added to medium containing single-cell embryos. After incubation for 1 h, the embryos were washed and then imaged for the presence of fluorescence associated with the NPs. All tested doses (12.5 -500 μg ml À1 ) resulted in widespread NP accumulation throughout the embryo, indicating that the 250 nm PLGA NPs can traverse the ZP (Figure 1a).
Based on comparison to the brightfield images, C6-NP signal was seen within the cell but not seen in the perivitelline space, the zona pellucida (ZP), or extracellular medium. In addition, we observed nuclear clearing within the pronuclei, which would not occur if the signal was not intracellular and intranuclear. Based on these findings, we conclude that the NPs are intracellular ( Figure 1a) and that intracellular accumulation was dose dependent (Figure 1a).
Although NPs can penetrate the ZP, to achieve gene editing, PNA and donor DNA reagents must be released from the NP and then F I G U R E 1 Delivery of poly(lactic-co-glycolic acid) (PLGA) nanoparticles (NPs) to single-cell embryos. (a) Uptake of coumarin-6 dye loaded PLGA NPs 1 h after addition to medium containing single-cell embryos. Nuclei were stained with a live-cell nuclear stain (Hoescht dye) (blue), scale bars = 50 μm. (b) Representative image of embryo morphology and division for various treatment doses over 5 days, scale bars = 50 μm. (c, d) Development to the two-cell (c) and blastocyst stage (d) of treated and untreated groups. Nanoparticle dose in μg ml À1 is indicated on graph; vehicle indicates NPs at a concentration of 500 μg ml À1 that contained no cargo. The horizontal dotted line at 80% indicates the expected untreated embryo survival based on culture methods used. Survival percentages and number of embryos per group are included in Table S2 (two-cell stage) and Table S3 (blastocyst stage). reach the nucleus. One potential advantage of embryo editing with PNA/DNA NPs is the elimination of the technically challenging pronuclear microinjection step that is used for gene editing with CRISPR/ Cas9, but this requires that NPs provide nuclear delivery of the reagents. To determine whether NPs were reaching the nuclei of treated embryos, cells were stained with Hoechst, a nuclear stain that allows for live cell imaging, and imaged 1 h after fluorescent NP delivery. There was overlap between fluorescence from C6-loaded NPs and the Hoechst-stained embryo pronuclei (Figure 1a), indicating nuclear delivery. However, there was some nuclear clearing observed at the 1 h time-point, representing a region of the cell that contains fewer fluorescent NPs within the region of the nuclear staining ( Figure 1a), indicating less delivery to the nucleus relative to the intracellular space. However, at the beginning of mitosis, the nuclear envelope disassembles, allowing for mixing of the nuclear and cytoplasmic contents. During telophase, the nuclear envelope reforms around the condensed chromosomes at either pole of the cell. Nuclear envelope breakdown and reassembly could allow for the relocalization of NPs that were initially in the cytoplasm to the nucleus, and this could allow for high levels of editing even if NPs are not initially in high concentrations within the nucleus.
The safety of this approach in vitro was assessed. Two measures of embryo health in culture are development to the two-cell stage and blastocyst hatching from the ZP. 21 During the early stages of embryo development, the ZP prevents the embryo from implanting into the oviduct of a mouse, or the Fallopian tube of a human. Once the embryo traverses the oviduct and reaches the uterus, a healthily dividing embryo should hatch from the ZP to allow for implantation into the uterine wall. We observe that the NP-treated embryos hatch from the ZP in culture ( Figure 1b) and that NP treatment does not negatively impact the percentage of embryos reaching the two-cell stage or the blastocyst stage for any of the NP doses tested (Figure 1c,d).
Survival percentages and number of embryos per group are included in Table S2 (two-cell stage) and Table S3 (blastocyst stage). Untreated embryos are expected to have greater than 80% development to the two-cell and blastocyst stage based upon the medium and culture methods used; treatment with NPs does not appear to impair cellular division or to be otherwise cytotoxic.

| Delivery of PNAs into embryos
To examine the delivery of PNAs to embryos by NP treatment, PLGA NPs containing PNA fluorescently labeled with tetramethylrhodamine (TAMRA) were fabricated. Note that the labeled-PNA containing NPs are much dimmer on imaging than the C6 dye NPs, since the concentration of C6 in the NPs (0.3%) is much higher than the concentration of the labeled PNAs in the NPs ($0.02%). Embryos were treated with TAMRA-labeled PNA for 1 h then maintained in culture for 5 days, with daily imaging (Figure 2). Fluorophore signal correlates with PNA concentration. After 1 h, there is greater than background fluorescence present within the embryo; however, the signal is very faint. By the two-cell stage, PNA is widely distributed within the embryo, including the nuclei as shown by overlap of the Hoechst nuclear stain and the red TAMRA signal. The PNA signal persists throughout the cells for at least 5 days. The PNAs in the representative image at 120 h appear at the periphery since cells at this stage are located around the outside of the blastocyst. These cells include trophoblast cells that will go on to form the extraembryonic membranes and the placenta, and the inner cell mass, which will go on to form the fetus.
The center of the blastocyst is the cell-free fluid-filled blastocoel cavity, which assists in generating sufficient pressure for zona hatching.
In addition, the images indicate that treatment at the one-cell stage with PNA and DNA containing NPs did not impede further cell division or development into blastocysts.

| Analysis of editing in vitro at the blastocyst stage
To allow for analysis of gene editing both at the genotype and phenotype level, we used embryos derived from a transgenic reporter mouse (designated 654-eGFP) containing a GFP-beta globin fusion transgene with a IVS2-654 (C to T) mutation in the beta globin-  (Table S4) and even the lowest tested dose of 12.5 μg ml À1 resulted in an average level of editing of 67% (Figure 3a). At the highest dose tested of 500 μg ml À1 , a 94% average editing of each blastocyst was achieved; 19 of the 21 blastocysts investigated at the 500 μg ml À1 dose exhibited very high levels of editing approaching 100% and 2 embryos exhibited editing around 50% (Figure 3a). Deep sequencing was also performed on blastocysts to confirm on target editing (Table S5).
Additionally, eGFP fluorescence was observed in multiple treated embryos at the blastocyst stage (e4.5/e5.5) (Figure 3b). The eGFP signal appears localized around the outside of the blastocyst, which is consistent with the localization of cells at the periphery of the F I G U R E 2 Delivery of TAMRA-PNA by poly(lactic-co-glycolic acid) (PLGA) nanoparticles (NPs) to single-cell embryos in culture. Representative images for nanoparticle mediated delivery of TAMRA-labeled PNA NPs (red) to embryos in culture. Embryos were monitored for 5 days. Nuclei were stained with Hoescht dye (blue), scale bars = 50 μm.
F I G U R E 3 Gene editing quantified by droplet digital PCR (ddPCR) and visualized by fluorescence microscopy in 654-eGFP blastocysts following treatment of single-cell embryos with PNA nanoparticles. (a) Gene editing in blastocysts after PNA/DNA NP treatment by ddPCR analysis with mean ± s.e.m. Significance by one-way ANOVA analysis for various treatment groups compared with 500 μg ml À1 NP group is shown. Groups above 100 μg ml À1 NP dose were not significantly different, marked as ns. ANOVA of significance and number of samples for different treatment groups compared to untreated control is included in Table S4. Vehicle group is 500 μg ml À1 poly(lactic-co-glycolic acid) (PLGA) NPs containing no cargo. (b) Representative images for eGFP fluorescence in a treated and untreated e5.5 embryo, scale bars = 50 μm blastocyst, surrounding the blastocoel cavity. No eGFP fluorescence was observed in untreated embryos ( Figure 3b). Seventeen of the 21 analyzed blastocysts (81%) had eGFP fluorescence by a visual assessment of microscopy images. This is lower than what would be expected based on the ddPCR results of the 500 μg ml À1 dose, however, embryos may be edited but not expressing eGFP at high enough levels that it was able to be imaged on the microscope as above background fluorescence.

| Safety of embryonic nanoparticle treatment in reimplanted mice
Single-cell fertilized embryos containing two copies of the 654-eGFP transgene were harvested. Embryos were treated for 1 h with PLGA NPs containing the PNA and donor DNA designed to correct the IVS2-654 mutation, as previously described, 8 at a concentration of 500 μg ml À1 in the medium and then cultured for 1 day before transfer to surrogate mothers at the two-cell stage.
The 654-eGFP embryos treated with PNA/DNA NPs and reimplanted at the two-cell stage had a development rate after reimplantation similar to that expected for untreated embryos (Table 1); untreated reimplanted embryos yield pups at a rate of approximately 50%. 23,24 In our studies, 54 out of 123 NP-treated embryos  Note: Pups include both healthy fetuses harvested at e17.5 and pups born and survived to weaning.
Of note, mice reimplanted at the two-cell stage and analyzed at Untreated mouse 1 and 2 were plotted in (b), treated mice 1-4 (harvested at e17.5) were plotted in (c) and treated mice 4-9 (harvested at p21) were plotted in (d). Each mouse was assigned its own color in the plots. For (b), treated mouse 1 is assigned pink, treated mouse 2 is assigned blue, treated mouse 3 is assigned purple, and treated mouse 4 is assigned green. For (c), treated mouse 5 is assigned purple, treated mouse 6 is assigned pink, treated mouse 7 is assigned green, treated mouse 8 is assigned orange, and treated mouse 9 is assigned blue.  Table S6.  Table 2).

| Off-target analysis
Deep sequencing was also performed to analyze off-target effects in treated mice at seven sites in the genome with partial homology to the PNA binding site in both blastocysts and reimplanted pups. Four treated tissues from different mice were analyzed for each site of partial homology (Table 3). Two untreated mice were also analyzed for each site; no modifications were found in untreated controls nor the treated samples. Three treated blastocysts were also analyzed for each site with partial homology ( Analysis of gene editing revealed that the average level of editing achieved at the blastocyst stage at the dose of 500 μg ml À1 is higher than the level of editing seen in e17.5 or p21 pups derived from the one-cell and two-cell reimplantations, which may indicate that reimplantation at the earlier stages does not allow for the continued editing that could occur when the embryos are allowed to develop into blastocysts in culture. This is also consistent with results of reimplantation studies at different developmental timepoints, as we found higher levels of genotypic and phenotypic editing in mice reimplanted at the two-cell stage than the single-cell stage. In addition, during early embryonic development, the genome is transcriptionally silenced and the embryo contains all necessary signals for early cell divisions without any genomic transcription. 30 The embryo then undergoes the maternal-to-zygotic transition (MZT) when maternal mRNAs are degraded and the genome becomes transcriptionally active. 30 MZT occurs after the two-cell stage in mice and after the four-cell stage in humans. 30 Prior to the MZT, genomic DNA exists in a repressed state that is incompatible with transcription. 31 PNA-mediated editing acts through binding and formation of a PNA/DNA/PNA triplex that induces homologous recombination and genome editing. 32 Since DNA methylation and tight chromatin organization contributes to transcriptional repression in the early embryo, 31 T A B L E 3 Seven gene loci in the mouse genome with partial homology to the 18 bp γPNA target site in beta-globin intron 2 were previously identified, 2 with the sequences as indicated    35 Although perhaps unlikely, it is not impossible that embryo editing could be used in the future for the selection of superficial traits for nonmedical reasons, which would be of significant ethical concern. However, concerns over future nonmedical uses should not  The 654-eGFP mice used in the study were homozygous with two copies of a GFP-beta globin fusion transgene with a IVS2-654 (C to T) mutation causing an aberrantly expressed intron. Successful gene editing results in expression of a functional GFP mRNA transcript and eGFP fluorescence. 22

| Superovulation and embryo harvest
To obtain single-cell fertilized embryos, 5-to 8-week-old female mice

| Nanoparticle fabrication
NPs containing C6 were formulated using a single-emulsion solvent evaporation technique. 12

| Isolation of gDNA from blastocysts
Blastocyst lysis buffer was made as follows for 10 ml: 500 at 1000 Â g. Samples were then placed in À20 C freezer until further use. To check the fidelity of the WGA, a standard curve was constructed from known amounts of gDNA ( Figure S3) and the template appears to amplify with high fidelity (R 2 = 0.9996).

| Reimplantation studies
Embryos were harvested from female donor mice according to previously discussed methods. For embryos reimplanted at the single-cell stage, embryos were treated for 1 h in 50 μl KSOM-AA media containing 500 μg ml À1 PNA/DNA NPs before transfer to surrogate moms.
Embryos implanted at the two-cell stage were cultured according to above protocols. Only healthy appearing two-cell embryos were reimplanted. The two-cell embryos were transferred to a 50 μl droplet of KSOM-AA just prior to reimplantation. Embryos at either the one-cell or two-cell stage were transferred to pseudopregnant recipient females that had been mated with vasectomized males the previous night.

DATA AVAILABILITY STATEMENT
Sequencing data have been deposited in the Sequence Read Archive (https://www.ncbi.nlm.nih.gov/sra) under acquisition number SUB7430845.