Abstract Cas9 protein without sgRNAs can induce genomic damage at the cellular level in vitro. However, whether the detrimental effects occur in embryos after Cas9 treatment remains unknown. Here, using pig embryos as subjects, we observed that Cas9 protein transcribed from injected Cas9 mRNA can persist until at least the blastocyst stage. Cas9 protein alone can induce genome damage in preimplantation embryos, represented by the increased number of phosphorylated histone H2AX foci on the chromatin fiber, which led to apoptosis and decreased cell number of blastocysts. In addition, single-blastocyst RNA sequencing confirmed that Cas9 protein without sgRNAs can cause changes in the blastocyst transcriptome, depressing embryo development signal pathways, such as cell cycle, metabolism, and cellular communication-related signal pathways, while activating apoptosis and necroptosis signal pathways, which together resulted in impaired preimplantation embryonic development. These results indicated that attention should be given to the detrimental effects caused by the Cas9 protein when using CRISPR-Cas9 for germline genome editing, especially for the targeted correction of human pathological mutations using germline gene therapy. Keywords: MT: RNA/DNA Editing, guide-free Cas9, germline genome editing, genomic damage, transcriptome changes, pig embryo Graphical abstract graphic file with name fx1.jpg [57]Open in a new tab __________________________________________________________________ Wang and colleagues demonstrate that Cas9 protein, without sgRNA, could induce genome instability and alter transcriptome homeostasis in pig preimplantation embryos. This finding reveals a potential risk factor for using CRISPR-Cas9 in germline genome editing and in germline gene therapy. Introduction The clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) system has emerged as a powerful tool for genome editing in prokaryotes and eukaryotes. Direct microinjection of CRISPR-Cas9 components into fertilized zygotes provides an efficient way to generate genetically modified animals, which shows potential in disease modeling, trait improvement, and gene function investigation.[58]^1^,[59]^2 This procedure is particularly important for animal species, such as rabbits and monkeys, in which authentic germline-competent pluripotent stem cells are unavailable, and the efficiency of somatic cell nuclear transfer cloning is relatively low.[60]^3^,[61]^4 In humans, genome editing in zygotes by the CRISPR-Cas9 system provides potential for the correction of pathological sequences and understanding the basic regulation of early embryo development.[62]^5^,[63]^6 Previously, off-target mutations have been major safety concerns for the practice of germline genome editing with the CRISPR-Cas9 system, especially for human germline gene therapy, and numerous efforts have been exerted to minimize off-target effects.[64]^7^,[65]^8^,[66]^9 However, beyond off-target issues, recent studies reported that the expression of Cas9 protein alone can lead to negative effects on cells, which raised another serious safety concern regarding the application of the CRISPR-Cas9 system.[67]^10^,[68]^11^,[69]^12^,[70]^13^,[71]^14 In human cells, SpyCas9 can cause DNA damage, most likely because of its single-strand DNA cleavage activity in the presence of metal ions, without sequence preference.[72]^10 Another study reported that the Cas9 protein can activate the TP53 pathway and induce the emergence and expansion of TP53-inactivating mutations.[73]^10^,[74]^11 In addition, the Cas9 protein can disrupt the formation of the DNA-dependent protein kinase (DNA-PK) complex through interaction with KU86, which leads to defective DNA-PK-dependent repair of DNA double-strand break (DSB) damage via the non-homologous end-joining (NHEJ) pathway.[75]^12 Genome damage in embryos affects post-implantation embryo development and leads to miscarriage during pregnancy.[76]^15 Therefore, whether the expression of Cas9 protein alone can cause genomic instability in embryos must be verified. Human embryos are rare and precious clinical resources, thus not practical for the comprehensive analysis of Cas9 protein-meditated safety effects. Compared with rodents, pigs are more similar to human embryos in terms of preimplantation development stage.[77]^16 In addition, large numbers of porcine embryos are easily available from discarded ovaries in slaughterhouses, which decreases the need for laboratory animals.[78]^17^,[79]^18 Today, in vitro production embryos can be obtained by the somatic cell nuclear transfer (SCNT) or in vitro fertilization (IVF) and used to generate genetically modified pigs.[80]^19^,[81]^20 However, the complex operation and incomplete reprogramming of SCNT embryos and polyspermy of IVF embryos limited the availability of sufficient diploid embryos with normal development capacity to study the early embryonic development and genomic damage. Large numbers of parthenogenetically activated (PA) embryos can be easily available with simple operation. More importantly, PA embryos share similar development characteristics with biparental embryos in the preimplantation stage, though development capacity is different in the post-implantation stage between them, and the cell karyotypes of PA embryos will undergo self-duplication during the in vitro development, resulting in diploid embryos.[82]^21^,[83]^22^,[84]^23 Therefore, in this study, PA porcine embryos were selected as subject resources to evaluate the safety effects of the expression of the Cas9 protein alone. We observed that expression of the Cas9 protein alone can cause DNA damage in preimplantation embryos, which led to the upregulation of TP53 target genes and imbalance in embryonic transcriptome homeostasis, activation of the apoptosis and necroptosis signaling pathways, and a reduction in cell numbers in blastocysts. These results identify important safety concerns caused by the Cas9 protein for germline genome editing in different species, especially before considering the targeted correction of human pathological mutations using germline gene therapy. Results Persistence of Cas9 protein in germline genome editing in embryos The CRISPR-Cas9 system consists of two components, the Cas9 endonuclease and a single guide RNA (sgRNA). In theory, RNAs are easily degraded by endogenous RNA nucleases, whereas proteins are more stable, when we use CRISPR-Cas9 for germline genome editing.[85]^24^,[86]^25 Prolonged persistence of the proteins transcribed from Cas9 mRNA during embryo development may give rise to safety concerns. Therefore, we first tested the persistence of Cas9 in editing PA porcine embryos. We linked the tdTomato fluorescent protein downstream of Cas9 via a self-cleaving T2A peptide to ensure the integrity of the Cas9 protein structure and enable easy visualization of its expression ([87]Figure 1A). To verify the Cas9-T2A-tdTomato mRNA was functional, we chose IL2RG and RAG2 as the target genes, which had been previously validated to be editable in pig embryos by our group.[88]^26 Two sgRNAs targeting the regions of the exon 1 of porcine IL2RG and RAG2 genes were designed, respectively ([89]Figure S1A). In vitro-transcribed Cas9-T2A-tdTomato mRNAs with IL2RG- and RAG2-sgRNAs were simultaneously injected into 150 PA porcine embryos, and tdTomato mRNA without Cas9 was injected into 50 PA porcine embryos as the control, at 3 h after the activation of oocytes. These injected embryos were cultured in vitro for further reverse-transcriptase quantitative polymerase chain reaction (RT‒qPCR), immunofluorescence staining, and Sanger sequencing analysis ([90]Figure 1B). Under an inverted fluorescence microscope, red fluorescence can be observed at the two-cell, four-cell, morula, and even in blastocyst stages on day 6 after the injection of Cas9-T2A-tdTomato mRNA ([91]Figure 1C). Then, we randomly selected a total of 40 embryos that had been injected with Cas9-T2A-tdTomato and tdTomato mRNA to test the expression of Cas9 in single embryos at different stages by using RT-qPCR, and the other embryos were subjected to subsequent tests. Our results showed that all embryos at the four-cell and morula stages contained Cas9 mRNA, and Cas9 mRNA decreased significantly as the embryo developed to the blastocyst stage but was still detectable in some of the blastocysts ([92]Figure 1D). The Cas9 protein could also be detected in blastocysts with injection of Cas9-T2A-tdTomato mRNA (n = 20) by immunofluorescence staining with an anti-Cas9 antibody, but not in blastocysts with injection of tdTomato mRNA (n = 8) ([93]Figure 1E). We also tested whether the Cas9 protein would persist to the blastocyst stage in pig embryos injected with Cas9 mRNA or Cas9 protein only. The immunofluorescence staining results showed that Cas9 protein could persist to the blastocyst stage in all tested embryos with injection of Cas9 mRNA (n = 5) or Cas9 protein (n = 5) ([94]Figure 1E). We further collected nine blastocysts with co-injection of Cas9-T2A-tdTomato and IL2RG- and RAG2-sgRNA, and three tdTomato-injected blastocysts, to lyse individually for genotyping. DNA fragments surrounding the target site were amplified via PCR and subjected to Sanger sequencing. The Sanger sequencing data were analyzed by inference of CRISPR edits (ICE) analysis.[95]^27 The genotypes of Cas9 with IL2RG- and RAG2-sgRNA injected blastocysts showed that all the blastocysts harbored insertions or deletions (indels) at the IL2RG locus, while only three of them harbored indels at the RAG2 locus ([96]Figure S1B and [97]Tables 1 and [98]2). The mutation that occurred in these blastocysts suggested that the Cas9-T2A-tdTomato mRNA was functional. The different gene-editing efficiencies of CRISPR-Cas9 at different gene loci may be caused by many factors, such as sgRNA/target sequences and chromatin states.[99]^28 These results indicate that when Cas9 and sgRNA were microinjected into one-cell stage embryos for germline genome editing, Cas9 would persist at least until the blastocyst stage. Figure 1. [100]Figure 1 [101]Open in a new tab Detecting the Cas9 in pig germline genome-edited preimplantation embryos (A and B) Schematic of the vector of Cas9-T2A-tdTomato and microinjection of Cas9-T2A-tdTomato and sgRNA RNA into pig embryos. (C) Merged bright field and red fluorescence images at different developmental stages after Cas9-T2A-tdTomato mRNA injection; scale bar, 50 μm. (D) RT-qPCR analysis of the expression of Cas9 in Cas9-injected and tdTomato-injected embryos at the four-cell, morula, and blastocyst stages. Four-cell stage, Cas9-injected embryos, mean ± standard error of the mean (SEM) = 843.7 ± 171.3, n = 6; tdTomato-injected embryos, mean ± SEM = 12.77 ± 1.462, n = 6. Morula, Cas9-injected embryos, mean ± SEM = 580 ± 177.3, n = 5; tdTomato-injected embryos, mean ± SEM = 34.65 ± 12.78, n = 4. Blastocyst, Cas9-injected embryos, mean ± SEM = 99.05 ± 32.89, n = 11; tdTomato-injected embryos, mean ± SEM = 1.771 ± 1, n = 8. ∗∗p = 0.0017, ∗∗∗∗p < 0.0001, unpaired t test. (E) Fluorescence microscopic images of Cas9 (green), tdTomato (red), and DAPI (blue) in Cas9-T2A-tdTomato mRNA-injected (Cas9-T2A-tdTomato), Cas9 mRNA-injected (mCas9), Cas9 protein-injected (pCas9), tdTomato-injected (tdTomato), and control (PA) blastocysts; scale bar, 50 μm. Table 1. The information of IL2RG in each Cas9-injected gene editing embryo IL2RG Indels Contribution Sequence(5′–3′) Embryo-1 +2 92% ATTGAACACAAAACACTGAACCTTT|nnGGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCAT −21 1% TGAACACAA--------------------|-----------GTAGAGTGGA AACGTTGAGAGTCCCAGGGGGT −8 1% ATTGAACACAAAACACTGAACCTT-|----GTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT −7 1% ATTGAACACAAAACACTGAACCTTT|---GTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT Embryo-2 −4 1% ATTGAACACAAAACACTGAAC-----------|GGGAGGGGTAGAGTGGAAACG TTGAGAGTCCC −12 1% ATTGAACACAAAACACTGA-------|------GGTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT −10 1% ATTGAACACAAAACACTG-------|--AGGGGTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT −3 2% ATTGAACACAAAACACTGAACCTTT|-------AGGGGTAGAGTGGAAACGTTGAGAGTCCCA −10 2% ATTGAACACAAAACACTGA------|---GGGGTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT −11 2% ATTGAACACAAAACACTGA-------|----GGGTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT +3 2% ATTGAACACAAAACACTGAACCTTT|nnnGGGAGGGGTAGAGTGGAAACGTTGAGAGTCC +7 2% ATTGAACACAAAACACTGAACCTTT|nnnnnnnGGGAGGGGTAGAGTGGAAACGTTGAGAG −5 3% ATTGAACACAAAACACTGAAC------|-----GGAGGGGTAGAGTGGAAACGTTGAGAGTCCCA −8 5% ATTGAACACAAAACACTGA---|-GAGGGGTAGAGTGGAAACGTTGAGAGTCCCAGGGGGT −18 67% ATTGAACACAAAAC--------------|-----------GTAGAGTGGAAACGTTG AGAGTCCCAGGGGGT Embryo-3 −16 3% ATTGAACACAAAAC---------------|--------GGGTAGAGTGGAAACGTTG AGAGTCCCAGGGGT −15 8% ATTGAACACAA-------------------|--GGAGGGGTAGAGTGGAAACGTTG AGAGTCCCAGGGGT −12 17% ATTGAACACAAAAC-------------|--GGAGGGGTAGAGTGGAAACGTTGAGAG TCCCAGGGGT −16 24% ATTGAACACAA---------------------|--GAGGGGTAGAGTGGAAACGT TGAGAGTCCCAGGGGT −15 33% ATTGAACACAAAACACT-----------|-----------GTAGAGTGGAAACGTT GAGAGTCCCAGGGGT Embryo-4 −3 31% ATTGAACACAAAACACTGAACCTT--|---GAGGGGTAGAGTGGAAACGTTGAGAGTCCCAG −1 29% ATTGAACACAAAACACTGAACCTT--|GGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCA −9 15% ATTGAACACAAAACAC--------------|GGGAGGGGTAGAGTGGAAAC GTTGAGAGTCCCAGG −9 12% ATTGAACACAAAACACTGAAC-----|------GGGTAGAGTGGAAACGTTGAGAGTCCCAGGGG Embryo-5 −1 1% ATTGAACACAAAACACTGAACCTTT|-GGAGGGGTAGAGTGGAAACGTTGAGAGTCCCAG −1 1% ATTGAACACAAAACACTGAACCTT-|GGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCAG 0 98% ATTGAACACAAAACACTGAACCTTT|GGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCA Embryo-6 +2 87% ATTGAACACAAAACACTGAACCTTT|nnGGGAGGGGTAGAGTGGAAACGTTGAGAGTCCC −17 3% ATTGAACACA-----------------------|----GAGGGGTAGAGTGG AAACGTTGAGAGTCCCAGGGG −8 2% ATTGAACACAAAACACTGAAC-----|------GGGTAGAGTGGAAACGTTGAGAGTCCCAGGGG −7 2% ATTGAACACAAAACACTGAAC----|----GGGGTAGAGTGGAAACGTTGAGAGTCCCAGGGG −21 1% ATTGAACACA---------------------------|---------GGTAG AGTGGAAACGTTGAGAGTCCCAGGGG Embryo-7 +1 41% ATTGAACACAAAACACTGAACCTTT|nGGGAGGGGTAGAGTGGAAACGTTGAGAGTCCC 0 28% ATTGAACACAAAACACTGAACCTTT|GGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCA Embryo-8 −10 19% ATTGAACACAAAACACTGA--------|-----GGGGTAGAGTGGAAACGTTGAGAG TCCCAGGGG −7 13% ATTGAACACAAAACACTGA--------|---GGAGGGGTAGAGTGGAAACGTTGAGAGTCCCAG −10 10% ATTGAACACAAAACACTGAA------|------GGGTAGAGTGGAAACGTTGAGAGTCCCAGGGG −11 9% ATTGAACACAAAACAC--------------|---GAGGGGTAGAGTGGAAACGTTGAG AGTCCCAGGG −13 8% ATTGAACACAAAACA----------------|-----AGGGGTAGAGTGGAAACGTT GAGAGTCCCAGGG −11 7% ATTGAACACAAAACACTGA--------|--------GGGTAGAGTGGAAACGTTGAGA GTCCCAGGG −10 4% ATTGAACACAAAACACT------------|----GAGGGGTAGAGTGGAAACGTTGA GAGTCCCAGG −13 3% ATTGAACACAAA----------------------|GGGAGGGGTAGAGTGGA AACGTTGAGAGTCCCAGG −10 3% ATTGAACACAAAACAC-------------|---GGAGGGGTAGAGTGGAAACGTTG AGAGTCCCAGG −9 3% ATTGAACACAAAACACTGA-------|----AGGGGTAGAGTGGAAACGTTGAGAG TCCCAGGG −11 2% ATTGAACACAAAACACT------------|-----AGGGGTAGAGTGGAAACGT TGAGAGTCCCAGGG Embryo-9 −3 1% ATTGAACACAAAACACTGAACCTTT|-------AGGGGTAGAGTGGAAACGTT GAGAGTCCCA 0 96% ATTGAACACAAAACACTGAACCTTT|GGGAGGGGTAGAGTGGAAACGTTGAGAGTCCCA [102]Open in a new tab Table 2. The information of RAG2 in each Cas9-injected gene editing embryo RAG2 Indels Contribution Sequence(5′–3′) Embryo-1 0 79% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG −4 9% GTTACTTTTCGCTGCAGAGAGA---|-CTTGGTAGGAGATGTTCCTGAAGGCAGATATGGTCA −3 7% GTTACTTTTCGCTGCAGAGAGAAA-|--TTGGTAGGAGATGTTCCTGAAGGCAGATATGGTC −3 2% GTTACTTTTCGCTGCAGAGAGAAAG|---TGGTAGGAGATGTTCCTGAAGGCAGATATGGTC −16 1% GTTACTTTTCGCTGCAGA-------|---------GAGATGTTCCTGAAGGCAGATA TGGTCATTCCAT Embryo-2 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-3 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-4 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-5 −16 45% GTTACTTTTCGCTGCAGAGAGA---|-------------TGTTCCTGAAGGCAGATAT GGTCATTCCAT −1 30% GTTACTTTTCGCTGCAGAGAGAAA-|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGGT 0 8% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG −16 4% GTTACTTTTCGCTGCAGAGAGAA--|--------------GTTCCTGAAGGCAGATA TGGTCATTCCAT −16 4% GTTACTTTTCGCTGCAGAGAGAAA-|---------------TTCCTGAAGGCAGATA TGGTCATTCCAT −16 1% GTTACTTTTCGCTGCAGAGAG----|------------ATGTTCCTGAAGGCAGATAT GGTCATTCCAT −4 1% GTTACTTTTCGCTGCAGAGAGA---|-CTTGGTAGGAGATGTTCCTGAAGGCAGATATGGTCA −3 1% GTTACTTTTCGCTGCAGAGAGA---|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGGTC Embryo-6 +1 6% GTTACTTTTCGCTGCAGAGAGAAAG|nACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG 0 93% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-7 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-8 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG Embryo-9 0 100% GTTACTTTTCGCTGCAGAGAGAAAG|ACTTGGTAGGAGATGTTCCTGAAGGCAGATATGG [103]Open in a new tab Cas9 induces DNA damage without sgRNA in preimplantation embryos We next investigated whether the Cas9 protein can induce DNA damage in early embryos. The in vitro-transcribed Cas9-T2A-tdTomato mRNA with IL2RG-sgRNA or without sgRNA was injected into 600 PA porcine embryos in three independent experiments, and DNA damage was evaluated at the two-cell, four-cell, morula, and blastocyst stages (referred to as day 1, day 2, day 4, and day 6, respectively) ([104]Figure 2A). The in vitro-transcribed tdTomato mRNAs without Cas9 were injected into 300 PA porcine embryos and used as controls. The formation of phosphorylated histone H2AX (pH2AX) foci on the chromatin fiber, a sensitive marker for genomic damage, was determined by immunofluorescence staining analysis. We detected 28 and 69 nuclei in day-1 and day-2 embryos with injection of tdTomato (five day-1 embryos: n = 10 blastomeres; five day-2 embryos: n = 19 blastomeres), Cas9-T2A-tdTomato (four day-1 embryos: n = 8 blastomeres; five day-2 embryos: n = 20 blastomeres), and Cas9-T2A-tdTomato with sgRNA (five day-1 embryos: n = 10; nine day-2 embryos: n = 30 blastomeres), respectively ([105]Figure S5). The results showed that nearly no evident pH2AX foci formed in these three groups in day-1 and day-2 embryos ([106]Figures 2B, 2C, 2F, and 2G). However, obvious pH2AX foci were found in the Cas9-T2A-tdTomato or Cas9-T2A-tdTomato with sgRNA-injected embryos at the morula and blastocyst stages ([107]Figures 2D and 2E). We detected 65 and 278 nuclei in morula and blastocyst stage embryos with injection of tdTomato (15 morula stage embryos: n = 17 blastomeres; 30 blastocyst stage embryos: n = 54 blastomeres), Cas9-T2A-tdTomato (20 morula stage embryos: n = 26 blastomeres; 30 blastocyst stage embryos: n = 115 blastomeres), and Cas9-T2A-tdTomato with sgRNA (15 morula stage embryos: n = 22 blastomeres; 30 blastocyst stage embryos: n = 109 blastomeres), respectively ([108]Figure S5). We found that the numbers of pH2AX foci in the Cas9-injected embryos with and without sgRNA at both the morula and blastocyst stages were significantly higher than those in the tdTomato-injected embryos ([109]Figures 2H and 2I). Interestingly, the average number of pH2AX foci in the nuclei of morula embryos was higher than that in the blastocysts. The reduction in the average number of pH2AX foci at the blastocyst stage suggests that some cells/embryos with abnormalities were repaired or eliminated (e.g., via the arrest of affected embryos or apoptosis of abnormal cells), as suggested by Babariya et al.[110]^29 Figure 2. [111]Figure 2 [112]Open in a new tab Analysis of the guide-free Cas9-induced genome damage in PA porcine embryos (A) Schematic of the DNA damage test in Cas9-T2A-tdTomato mRNA-injected PA porcine embryos, with tdTomato mRNA-injected embryos as controls. We selected two-cell (day 1), four-cell (day 2), morula (day 4), and blastocyst (day 6) stage embryos to test DNA damage. (B–E) Fluorescence microscopic images at different embryo stages of pH2AX (green), tdTomato (red), and DAPI (blue) in tdTomato mRNA-, Cas9-T2A-tdTomato mRNA- and Cas9-T2A-tdTomato with sgRNA mRNA-injected embryos. Scale bar, 20 μm. (F–I) Statistical results of the number of pH2AX foci per nucleus in tdTomato mRNA-, Cas9-T2A-tdTomato mRNA-, and Cas9-T2A-tdTomato with sgRNA mRNA-injected embryos. For (H) and (I), we counted the number of pH2AX foci in the damaged nuclei. For (F), day 1 tdTomato-injected blastomeres, n = 10; Cas9-injected blastomeres, n = 8; Cas9 with sgRNA-injected blastomeres, n = 10. For (G), day 2 tdTomato-injected blastomeres, mean ± SEM = 0.4211 ± 0.2572, n = 19; Cas9-injected blastomeres, mean ± SEM = 0.35 ± 0.1094, n = 20; Cas9 with sgRNA-injected blastomeres, mean ± SEM = 0.8 ± 0.2218, n = 30. For (H), day 4 tdTomato-injected blastomeres, mean ± SEM = 7.588 ± 1.32, n = 17; Cas9-injected blastomeres, mean ± SEM = 34.96 ± 5.067, n = 26; Cas9 with sgRNA-injected blastomeres, mean ± SEM = 20.27 ± 2.22, n = 22. For (I), day 6 tdTomato-injected blastomeres, mean ± SEM = 6.13 ± 0.6177, n = 54; Cas9-injected blastomeres, mean ± SEM = 15.64 ± 0.9625, n = 115; Cas9 with sgRNA-injected blastomeres, mean ± SEM = 9.908 ± 0.6563, n = 109; ∗∗∗p = 0.0003, ∗∗∗∗p < 0.0001, unpaired t test. As reported, utilizing Cas9 with sgRNA to perform gene editing could induce genomic instability in preimplantation mouse and human embryos[113]^30^,[114]^31^,[115]^32 and off-target mutations in gene-edited mice, rats, and pigs.[116]^33^,[117]^34^,[118]^35 To eliminate the interference of sgRNA in porcine embryos and focus on the safety effects of Cas9 protein alone, we microinjected Cas9-T2A-tdTomato mRNA without sgRNA into PA porcine embryos in subsequent experiments. We microinjected one blastomere at the two-cell stage with Cas9-T2A-tdTomato or tdTomato mRNAs to further confirm that the embryonic genome damage was caused by the Cas9 protein ([119]Figure S2A). We also observed no evident pH2AX focus formation in these embryos at the day 2 stage, but we did in the morula and blastocyst stages. By counting 246 and 417 nuclei in one-blastomere-injection embryos with Cas9-T2A-tdTomato (12 morula stage embryos: n = 113 Cas9-injected +57 Cas9-un-injected blastomeres; 9 blastocyst stage embryos: n = 102 Cas9-injected +109 Cas9-un-injected blastomeres) and tdTomato mRNA (five morula stage embryos: n = 34 tdTomato-injected +42 tdTomato-un-injected blastomeres; eight blastocyst stage embryos: n = 120 tdTomato-injected +86 tdTomato-un-injected blastomeres) at morula and blastocyst stages, respectively ([120]Figure S5). We observed that the derivative of Cas9-injected blastomeres was more prone to generate pH2AX foci than tdTomato-injected and Cas9 uninjected blastomeres, whereas no significant difference was detected between tdTomato-injected and tdTomato uninjected blastomeres ([121]Figures S2B–S2E). These results confirmed that the Cas9 protein without sgRNA can also induce DNA damage in embryos, and the damage was reduced in the blastocyst stage. Cas9 protein negatively affects the development of preimplantation embryos Previous reports have shown that DNA damage can activate the TP53 signaling pathway, and increase the expression of TP53 target genes, such as PUMA, p21, PERP, and NOXA.[122]^12 To verify whether the TP53 signaling pathway was also activated in Cas9-injected embryos, we collected 24 blastocysts in three independent experiments and subjected them to RT-qPCR. Under an inverted fluorescence microscope, all collected embryos emitted red fluorescence, which indicated that Cas9-T2A-tdTomato or tdTomato mRNAs were successfully injected into the porcine embryos ([123]Figure S3A). RT-qPCR results revealed that the TP53 target genes were fully or partially activated in Cas9-injected blastocysts. Among them, four genes were significantly activated in embryo-2 and embryo-5; three genes were significantly activated in embryo-1, embryo-7, and embryo-8; two genes were significantly activated in embryo-4 and embryo-6; and one gene was significantly activated in embryo-9, embryo-10, embryo-11, and embryo-12 ([124]Figure 3A). We summarized the expression of these TP53 target genes in screened Cas9-injected blastocysts, and discovered that half of these blastocysts significantly elevated the summary score of TP53 target genes ([125]Figure 3B). These findings suggest that Cas9 can also activate the TP53 signaling pathway and increase the expression of TP53 target genes in preimplantation embryos. Figure 3. [126]Figure 3 [127]Open in a new tab Analysis of the guide-free Cas9 on the development of porcine preimplantation embryos (A) RT-qPCR analysis of the expression of the TP53 target genes PERP, PUMA, p21, and NOXA in 12 Cas9-injected blastocysts. ∗p < 0.033, ∗∗p < 0.002, and ∗∗∗p < 0.001; two-way analysis of variance. (B) Average activation of TP53 transcriptional targets in 12 Cas9-injected blastocysts. ∗p < 0.0332, unpaired t test. The dotted line represents the average value of the TP53 target gene in tdTomato-injected embryos. (C) Cell numbers of blastocysts in PA, tdTomato-injected and Cas9-injected embryos. PA, mean ± SEM = 51.68 ± 2.848, and n = 31; tdTomato, mean ± SEM = 45.62 ± 2.821, and n = 29; Cas9, mean ± SEM = 37 ± 2.987, and n = 23. ∗p = 0.0423, ∗∗∗p = 0.001; unpaired t test. (D) Annexin V-labeled apoptotic cells in PA, tdTomato-injected, and Cas9-injected blastocysts. Annexin V: green; tdTomato: red; and DAPI: blue, scale bar, 50 μm. Given that the activation of TP53 target genes initiates the program of cell-cycle arrest, cellular senescence, or apoptosis,[128]^36^,[129]^37 we next evaluated whether the developmental capacity of Cas9-injected embryos would be affected. Although the overall blastocyst rate in PA (39.66%), tdTomato-injected (30.94%), and Cas9-injected embryos (28.65%) did not significantly change in nine independent experiments ([130]Figure S3B), the average cell number of Cas9-injected blastocysts (n = 23, 37 ± 2.987) was significantly lower than those in the PA blastocysts (n = 31, 51.68 ± 2.848) and tdTomato-injected counterparts (n = 29, 45.62 ± 2.821) ([131]Figure 3C). In addition, cellular apoptosis was detected in Cas9-injected blastocysts by immunofluorescence staining for the expression of Annexin-V, a biological marker for apoptotic cells. We found that seven embryos had obvious apoptotic cells in 15 Cas9-injected blastocysts but not in tdTomato-injected (n = 15) and control PA (n = 15) blastocysts ([132]Figure 3D). These results indicate that the Cas9 protein can negatively affect the development of early preimplantation porcine embryos. Single-blastocyst RNA sequencing of Cas9-injected embryos To further elucidate the effects of the Cas9 protein on embryos, we injected the Cas9-T2A-tdTomato or tdTomato mRNA into 300 PA porcine embryos in three independent experiments, and used 37 Cas9-injected blastocysts and 20 tdTomato-injected blastocysts to perform single-blastocyst RNA sequencing (RNA-seq) analysis ([133]Figure S5). Principal-component analysis (PCA) showed 49% variance along the first component, which indicated the difference in the whole transcriptomes between Cas9-injected and tdTomato-injected embryos ([134]Figure 4A), and this finding was further confirmed by the hierarchical clustering heatmap ([135]Figure S4A). Then, we carried out weighted gene coexpression network analysis (WGCNA) on these RNA-seq data to identify the gene modules that were highly related to the presence of Cas9 protein. As shown in [136]Figures 4B and 4C, we detected four gene modules that were highly related to the Cas9 protein, and they were labeled by colors: turquoise (cor = 0.99, p < 1e^−200), blue (cor = 75, p < 1e^−200), yellow (cor = 0.82, p < 1e^−200), and red (cor = 0.61, p = 3.6e^−65). Among the four gene modules, only the turquoise expression pattern was positively related to the Cas9 protein, whereas the others showed a negative relationship ([137]Figures 4D, 4E, [138]S4B, and S4C). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis revealed that the turquoise module was related to apoptosis, the tumor necrosis factor signaling pathway, and necroptosis ([139]Figure 4F), which suggests that the Cas9 protein activated cell death-related signaling pathways. This finding could be the reason for the decreased cell number found in Cas9-injected blastocysts ([140]Figure 3C). The blue module was related to ubiquitin-mediated proteolysis, lysosome, peroxisome, protein export, carbon metabolism, citrate cycle (TCA cycle), cell cycle, and mitophagy, which indicates that the Cas9 protein disturbed protein degradation and energy metabolism, and all of these results may negatively affect embryonic development. Notably, the Fanconi anemia pathway,[141]^38 homologous recombination pathway, and nucleotide excision repair were also enriched in the blue module, which indicates that the function of DNA damage repair was repressed in Cas9-injected embryos ([142]Figure 4F). Several pathways similar to the blue module, including autophagy and carbon metabolism, focal adhesion, and adherens junction, which affect embryonic development and cellular communication, were also found in the red module ([143]Figure S4D). The total expression levels of all genes in each enriched pathway ([144]Figure 4G) and the expression levels of individual genes in each enriched pathway ([145]Figure 4H) were different between tdTomato-injected and Cas9-injected embryos. In Cas9-injected embryos, except for the apoptosis and necroptosis pathways, the total expression levels of all genes in other pathways were reduced compared with tdTomato-injected embryos. In detail, the transcription of genes in pathways such as the cell cycle, TCA cycle, focal adhesion, adherens junction, and DNA damage repair was downregulated, whereas those of cell death-related signal pathway genes were specifically activated. Both findings suggest the impaired development of embryos injected with Cas9. Figure 4. [146]Figure 4 [147]Open in a new tab Effects of guide-free Cas9 on the transcriptome homeostasis of porcine blastocysts (A) PCA of Cas9-injected and tdTomato-injected blastocyst RNA sequencing data. (B) Cluster dendrogram showing the identified coexpression gene modules labeled by different colors. (C) Scatterplot of gene significance and module membership showing the four most important modules related to Cas9 expression, the turquoise, blue, yellow, and red modules. (D and E) Heatmap and bar plot showing the correlation of module eigengene expression and Cas9 in the turquoise and blue modules. The rows correspond to genes and the columns to samples in the heatmap. The upper bar plot row shows the corresponding module eigengene expression values versus the samples. (F) Enriched KEGG pathways of the turquoise and blue modules; the x axis indicates significance. (G) Bar plot showing the average expression of the pathways of interest. (H) Heatmap revealing the expression details of genes in the pathways of interest. [148]Table S1 shows the gene names. Discussion CRISPR-Cas9 is currently the most popular tool for generating genetically modified animal models through microinjection of Cas9 and sgRNA into fertilized embryos. The CRISPR-Cas9 system can be easily used to edit genes in mammalian embryos, such as mice, rats, rabbits, dogs, pigs, monkeys, and even humans.[149]^39^,[150]^40 After implantation into surrogate mothers, these embryos can develop to full term to generate live-born animals. CRISPR babies also carry precise changes in their DNA.[151]^41^,[152]^42 For most studies, the off-target effects caused by sgRNA mismatches, repair events, or genomic rearrangement after sgRNA-induced DSBs and incomplete editing-mediated mosaicism are the major safety concerns for CRISPR-Cas9-based germline genome editing, especially for human germline gene therapy.[153]^43^,[154]^44^,[155]^45 In addition to the canonical safety concerns mentioned above, recent studies have reported that Cas9 protein alone is also a risk factor during gene editing of immortalized cell lines, primary cells, and pluripotent stem cells.[156]^11^,[157]^12 In this study, we verified whether the negative effects can also be caused by Cas9 protein, which may give rise to additional safety concerns for embryonic genome editing. After microinjection of Cas9 mRNA and sgRNA into one-cell stage embryos for embryonic gene editing, we observed that Cas9 would persist until at least the blastocyst stage. Thus, Cas9 is a “long-term problem” for embryos when CRISPR-Cas9 is used for germline genome editing, especially for future human germline gene therapy in the clinic. In this study, Cas9 mRNAs alone were injected into one-cell stage PA porcine oocytes, and DNA damage foci were significantly elevated in Cas9-injected embryos but not in tdTomato-injected embryos, as shown by immunofluorescence labeling of pH2AX. Thus, the Cas9 protein without sgRNA can also induce genome damage in embryos, as reported in mammalian cells.[158]^11^,[159]^12 These phenomena were further confirmed by the injection of Cas9-T2A-tdTomato or tdTomato mRNAs into one blastomere of the two-cell porcine embryos, in which we observed that the derivatives of Cas9-injected blastomeres were more prone to generate pH2AX foci than tdTomato-injected or uninjected blastomeres. Interestingly, no obvious pH2AX foci were observed in embryos injected with Cas9 either at the one-cell or two-cell stage, in which zygotic gene activation (ZGA) had not yet occurred, but significantly increased in the morula and blastocyst stage, in which ZGA had occurred. Therefore, ZGA may be associated with pH2AX focus formation and DNA damage response activation. At the molecular level, RT-qPCR showed that the DNA damage caused by Cas9 activated the TP53 signaling pathway, which increased the expression of TP53 target genes, such as PUMA, p21, PERP, and NOXA, in preimplantation embryos. Consistently, single-blastocyst RNA sequencing also showed that the expression of Cas9 protein altered the transcriptome profile of preimplantation embryos, which included the activation of cell death-related signal pathways, such as apoptosis and necroptosis signal pathways, and depression of cell cycle-, metabolism-, and cellular communication-related signal pathways, such as the TCA cycle, focal adhesion, and adherens junction signal pathways. In addition, the DNA damage repair signal pathways, nucleotide excision repair, homologous recombination, and Fanconi anemia pathway were repressed in the Cas9-injected group. All the above alterations at the molecular level may negatively influence preimplantation embryonic development, as pointed out by Palmerola et al.[160]^46 Therefore, apoptotic cells and decreased cell number, two indicators suggesting poor quality of blastocysts, were observed in Cas9-injected blastocysts. The poor quality of blastocysts resulting from genome damage may increase the odds of heritable mutations of functional genes, reduce the pregnancy rate, and cause fetal miscarriage or unhealthy offspring. Previously, the embryo lethality in constructing the Cas9-ubiquity expression chickens, and the reduced birth rates in constructing the base-editors expression mice have been reported by some researchers.[161]^47^,[162]^48 These possible detrimental effects were not pointed out in other animal germline gene editing experiments because their direct goal was to pursue generation of gene editing animals and the negative effects might be neglected. It appears reasonable, given that not all the Cas9-injected embryos had reduced quality, and some Cas9-injected blastocysts had comparable cell numbers to the control group. In practice, when germline gene editing is performed to create gene editing animals, usually, more embryos are required to be transplanted into recipient animals to make sure that gene edited animals could be achieved. In summary, we have proven that Cas9 protein without sgRNA in embryos can result in genomic damage, TP53 signaling pathway activation, disturbance of embryonic transcriptome homeostasis, and decreased quality of embryos. Therefore, attention should be given to Cas9 protein-induced detrimental effects when using CRISPR-Cas9 for germline genome editing, particularly the safety concerns for attempts at human embryo gene editing. Materials and methods Cumulus-oocyte complex collection and in vitro maturation Ovaries were obtained from the local abattoir (JiangGao, Guang Zhou, China) and transported in 0.9 mg/mL NaCl with 75 μg/mL kanamycin and 50 μg/mL streptomycin at 30°C. We collected the cumulus-oocyte complex (COCs) from follicles 3 mm–8 mm in diameter using a 10-mL syringe and injected them into a 50-mL centrifuge tube warmed in a 35°C water bath. The COCs were washed thrice in 1x PVA-TL HEPES, and COCs were selected with a sterile glass pipette. The number of the COCs was recorded, and they were transferred to the pre-equilibrated in vitro maturation medium for 42–44 h in a CO[2] incubator (38.5°C, 5% CO[2], and 100% humidity).[163]^49 Matured oocyte parthenogenetic activation and culture in vitro After the maturation period, the cumulus cells were removed by oocyte-denuding medium containing 0.1 mg/mL hyaluronidase. The denuded oocytes were washed twice by oocyte manipulation medium. The matured oocytes were collected and activated by electro cell manipulator under 1 DC pulses, 120 V, and 30 μs pulse length. After activation, the oocytes were washed twice and cultured in porcine zygote medium-3 (PZM-3) at 38.5°C in 5% CO[2] humidified air. Microinjection of RNA and protein into parthenogenetic embryos We transcribed Cas9-T2A-tdTomato mRNA, Cas9 mRNA, tdTomato mRNA, RAG2-sgRNA, and IL2RG-sgRNA using the HiScribe T7 ARCA mRNA Kit (NEB, E2060S) and HiScribe T7 High Yield RNA Synthesis Kit (NEB, E2040S) and harvested RNA using the RNeasy MinElute Cleanup Kit (Qiagen, 12243). The final concentrations of the Cas9-T2A-tdTomato, Cas9 and tdTomato mRNAs were 100 ng/μL, and those of RAG2-sgRNA and IL2RG-sgRNA were 50 ng/μL. The Cas9 protein was purchased from NEB (M0386M), and we diluted it to 200 ng/μL. Then, we injected the RNA or Cas9 protein into parthenogenetic embryos through an Olympus microscopic operating system. Embryo collection The injected embryos were randomly divided into four groups and cultured in drops with 500 μL porcine embryo-development medium, PZM-3. About 50–100 embryos were cultured in each drop. When the injected embryos of these four groups were cultured for 1 day, 2 days, 4 days, and 6 days, normal-developing two-cell, four-cell, morula, and blastocyst stage embryos were randomly and individually collected for immunofluorescence, RT-qPCR, or RNA-seq analysis. Genomic analysis of individual embryos The single blastocysts simultaneously injected with Cas9 and sgRNA were collected and lysed at 65°C for 6 min in 5 μL of QuickExtract DNA Extraction Solution (Lucigen, QE0905T). The lysates were then used as a template for amplifying the fragments covering the target sites of the IL2RG-sgRNA and RAG2-sgRNA with the Pig-IL2RG-F/R and Pig-RAG2-F/R primers, respectively ([164]Table S2). The PCR products were further subjected to Sanger sequencing. The genotypes of each blastocyst were further confirmed by inference of CRISPR edits (ICE) analysis ([165]ice.synthego.com) of Sanger sequencing results. Immunofluorescence The embryos were collected at different development stages, fixed with 4% para-formaldehyde for 15 min, and then washed thrice for 10 min with a washing solution buffer (phosphate-buffered saline [PBS] with 0.1% bovine serum albumin [BSA]). Fixed embryos were blocked and permeabilized in PBS buffer with 9% (v/v) Triton X- and 5% (v/v) goat serum for 3 h at room temperature. Then, the embryos were incubated overnight at 4°C in a 1:200 solution of the primary pH2AX antibody (CST, 9718) or Cas9 antibody (Huabio, ET1703-85) in an antibody diluent buffer. After washing three times in washing solution buffer, the embryos were incubated with the secondary antibody (anti-rabbit IgG-Alexa Fluor 488, CST, 4412) diluted 1:1,000 in antibody diluent buffer for 1 h at room temperature. The nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI) (Sigma‒Aldrich, F6057) for 10 min. Confocal image acquisition was performed using a Zeiss LSM 710 laser-scanning microscope. Single embryo RT-qPCR Pig qPCR primers (PUMA, p21, NOXA, PERP, spCas9, and glyceraldehyde-3-phosphate dehydrogenase) were mixed until the final concentration of 0.1 μM. A single blastocyst was placed in PBS buffer, and the embryo was pipetted with 0.5 μL PBS into reverse-transcribed reaction buffer, amplified by the Single-Cell Sequence Specific Amplification Kit (Vazyme, P621). Following the manufacturer’s instructions, fluorescence qPCR was performed with AceQ Universal SYBR qPCR Master Mix (Vazyme, Q511-02). [166]Table S2 shows the primer sequences. Blastocyst cell counting Blastocysts were selected and placed in PZM-3 droplets containing 10 mg/mL Hoechst 33342 (Thermo, 62249), and the droplets were cultured for 30 min. Then, the blastocysts were washed twice in PBS with 0.1% BSA using a sterile glass pipette, and the number of nuclei per blastocyst was counted under the fluorescence microscope. Embryo apoptosis detection Blastocyst apoptosis assays were performed by Annexin V-fluorescein isothiocyanate (FITC)/propidium iodide Apoptosis Detection Kit (Vazyme, A211). The blastocysts were placed in 100 μL 1x Binding Buffer, added with 5 μL Annexin V-FITC, and incubated at room temperature for 10 min in the dark. The blastocysts were collected and photographed by a fluorescence microscope. Single-blastocyst RNA-seq TruSeq RNA Sample Preparation kit (Illumina) was used to construct RNA-seq libraries. High-throughput sequencing was performed on a HiSeq 2500 system (Illumina). Quality control was performed to remove adapter sequences and low-complexity reads using Fastp (version 0.23.2). Subsequently, Salmon (version 0.8.2) was used to align clean reads to the Sus scrofa transcriptome (National Center for Biotechnology Information version GCF_000003025.6) to generate a gene expression matrix. Low-expression genes, which had TPM value <1 in all samples, were removed in downstream analysis. Next, DESeq, variance stabilizing transformation (vst), and functions of DESeq2 package (version 1.20.0) were used to normalize and transform data. The vst-transformed data were used as input to perform the WGCNA using WGCNA packages (version 1.69). To detect the gene modules, we selected the Soft threshold as 5 (power = 5) and set the network type as signed (networkType = “signed”). The KEGG pathway enrichment analysis of genes in the module was performed using clusterProfiler (version 3.10.1). Data and code availability The high-throughput sequencing data from this study have been submitted to the NCBI Gene Expression Omnibus and the accession number is [167]GSE224238 ([168]https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE224238). Acknowledgments