Human embryo gene editing: God's scalpel or Pandora's box?

Abstract

Gene editing refers to the site-specific modification of the genome, which mainly focuses on basic research, model organism construction and treatment and prevention of disease. Since the first application of CRISPR/Cas9 on the human embryo genome in 2015, the controversy over embryo gene editing (abbreviated as EGE in the following text) has never stopped. At present, the main contradictions focus on (1) ideal application prospects and immature technologies; (2) scientific progress and ethical supervision; and (3) definition of reasonable application scope. In fact, whether the EGE is ‘God's scalpel’ or ‘Pandora's box’ depends on the maturity of the technology and ethical supervision. This non-systematic review included English articles in NCBI, technical documents from the Human Fertilization and Embryology Authority as well as reports in the media, which performed from 1980 to 2018 with the following search terms: ‘gene editing, human embryo, sequence-specific nuclease (SSN) (CRISPR/Cas, TALENT, ZFN), ethical consideration, gene therapy.’ Based on the research status of EGE, this paper summarizes the technical defects and ethical controversies, enumerates the optimization measures and looks forward to the application prospect, aimed at providing some suggestions for the development trend. We should regard the research and development of EGE optimistically, improve and innovate the technology boldly and apply its clinical practice carefully.

Introduction

Gene editing refers to the precise insertion, knockout and alteration of the genome, in the whole set of genetic material of organisms [1]. Based on site-specific programmable nucleases, it induces target DNA double-strand breaks (DSBs) and triggers DNA repair mechanisms (include non-homologous end joining (NHEJ) and homologous recombination (HR)), so as to promote efficient and accurate genetic modifications, such as gene disruption, insertion and correction (Figure 1). Gene editing targets at somatic cells, germ cells, stem cells and embryos (Figure 2), which is widely used to develop improved crops, construct disease models, research gene function and gene therapy. To date, five embryo gene editing (EGE) experiments have been successfully completed (Table 1) [2–6]. Everything seems to be moving in the right direction, and the clinical application is beginning to dawn. On 25 November 2018, Jiankui He, an associate professor from Southern University of Science and Technology, announced that two babies with edited C-C chemokine receptor type 5 (CCR5) genes had been born in China. The research on EGE has been pushed to the waves of public opinion for a while. The discussions mainly focus on ‘safety, necessity, ethics supervision of EGE’.

Controversy about EGE

Whether the EGE is ‘God's scalpel’ or ‘Pandora's box’? Whether Jiankui He is ‘the first person to eat crabs’ or ‘over-edited devil’? At present, the contradiction of EGE mainly lies in (1) security—the contradiction between the prospect of promising application and the immature technology, (2) necessity—the lack of accurate definition in the scope of reasonable application and (3) ethical thinking—the balance between ethical torture and medical development.

Security

Off-target effects

The sgRNAin the CRISPR/Cas9 system matches the target DNA fragment and directs the Cas9 protein to recognize the genome. When the designed sgRNA is mismatched with the non-target DNA sequence, an unintended genetic mutation is introduced, which is named off-target effects. Although the odds of missing target are not high, the genetic material has been changed, which is inherited and passed on to the next generation. There are two problems: one is how to reduce the off-target effect from the technology itself; the other is how to improve the sensitivity of the detection method [7]. We can reduce off-target effects by the following feasibility strategies: (1) reduction of insertions and deletions introduced by NHEJ repair, such as the use of single-cutting enzyme-active mutants and BEs without DSB activation [8, 9]; (2) optimization of sgRNAs, such as shortening the length of sgRNA based on the efficiency of a certain sgRNA binding target [10]; (3) increasing the specificity of the nuclease protein, such as insertion of hydroxy tamoxifen stress intein at a specific site of the Cas9 protein (Hydroxytamoxifen(4-HT)-responsive intein) produces small molecules that activate Cas9 nuclease, which increases the specificity of Cas9 by 25 times [11]. At present, there are two methods for detecting off-target effects. One is based on genome-wide DSBs detection, and the other is to separate genomic DNA from Cas9 in vitro and then perform whole genome sequencing [12, 13]. Both methods need to refer to the whole genome and the sequencing depth is limited when detecting multiple sgRNAs, which needs to be improved.

Chimeric embryos

Due to the longer half-life of Cas9 protein and sgRNA, they can still maintain the nuclease activity after the first mitosis in the zygote, and cut the target gene in the genome after cleavage in different blastomeres to produce different modifications and form a mosaic of embryo, named mosaicism [4, 14]. Mosaicism is a major concern that cannot be addressed by Preimplantation Genetic Diagnosis (PGD), as we cannot sequence all cells in an embryo [15]. The hazards of chimeric embryos primarily depend on the degree of chimerism and the chromosomes/cell lineages where mosaicism occurred. Only precisely controlling the time of enzymatic activity, it is possible to avoid the development of mosaic embryos by gene modification in single cells. Considering that the embryo has a self-correcting mechanism, such as euploid cell dominant growth, abnormal cell self-correction and normal cell inward cell cluster aggregation, we should take a positive attitude toward the chimeric embryo problem.

Low repair efficiency

The third major challenge for clinical application of EGE is how to enhance the efficiency of gene editing, determined by the editing mode, cell type and locus sequence. Generally speaking, NHEJ is more efficient than Homology directed repair (HDR) [16, 17]. However, NHEJ usually results in insertions or deletions of target sites and frameshift mutations. HR can induce accurate point mutations or DNA insertions based on single/double-stranded DNA templates. Therefore, we need to improve the efficiency of HDR and decrease NHEJ-induced indel. A more desirable HDR-NHEJ ratio can be achieved through synchronization of the cell cycle or utilization of small molecules as well [18]. Gutschner et al. fused the Geminin gene into Cas 9 to induce its degradation in G1 stage and limited the target DNA cleavage into the S/G2 stage, where HDR can be increased up to five times [19, 20]. It’s reported that VE-822 and AZD-7762 can significantly increase the efficiency of gene editing in human pluripotent stem cell [21].

Bad gene and good gene

Fortune and misfortune depend on each other. For instance, mutations in hemoglobin will cause sickle cell disease, but also prevent malaria. Transforming FUT2 may reduce the probability of type I diabetes, but also increase the risk of infection with Novak virus [22]. Although CCR5-32 mutation can protect the offspring against HIV infection. It also lead diet-induced obese mice in crisis of glucose intolerance and increase the risk of West Nile virus infection in human [23, 24]. Recently, Wei and Nielsen [25]  estimate a 21% increase in the all-cause mortality rate in individuals who are homozygous for the ∆32 allele. Our limited knowledge of the role and characteristics of the human genome is not enough to interfere with nature. We must maintain humility and reverence for nature and take a cautious attitude before we make a permanent change to the human genetic stockpile.

Carcinogenicity and immune response

Matthew Porteus et al. found that Cas9 antibodies are ubiquitous in humans, whom treated with CIRSPR/Cas9 may cause a strong immune response and eventually lead to death [26]. Recently, the Swedish Karolinska Institute [27] and the Cambridge Novartis Institute of Biomedical Research [28] independently found that the DSB caused by CRISPR/Cas9 may activate the p53 pathway with a risk of carcinogenic. Fortunately, no cancers induced by p53 gene or gene editing have been found so far. At the same time, Korean scientists found that CRISPR gRNAs containing 5′-triphosphate (5′-ppp gRNA) trigger RNA-induced innate immune responses in human and mouse cells and result in cytotoxicity [29, 30]. It is a defect but also an opportunity. On the one hand 5′-ppp gRNAs in the cytoplasm are recognized by DDX58 and activate the Type I interferon response, causing up to about 80% cell death. On the other hand, given the sensitivity of T cells to the 5′ppp gRNAs, gene editing of T cells has great potential in the development of cell therapies for cancer and HIV.

Necessity

Should EGE be seen as an alternative to PGD?

At present, EGE is not the first choice for families with genetic diseases to obtain healthy offspring. In theory, PGD is sufficient to solve all hereditary diseases that meet Mendel's laws of inheritance [15]. The opponents suggest that they will not deny the enormous research value of EGE, but PGD will always be the last threshold for gene therapy. The supporters point out that PGD is not a panacea for blocking the transmission of genetic diseases to future generations. Strictly speaking, PGD is just a diagnostic procedure that can diagnose whether a particular embryo carries a single mutant gene, rather than a treatment that corrects the genetic variant gene in the embryo. For some special patients, genetic editing may be the only way to get healthy offspring: (1) When one of the parents is the autosomal dominant homozygote, the risk of transmission to the offspring is as high as 100%, and the embryo without mutation cannot be obtained by ART; (2) Both parents are autosomal recessive homozygotes, which means that they all carry two alleles; (3) Mitochondrial DNA mutations in oocytes and embryos.

Application prospects

Experts hope that EGE can make a contribution to four directions: completing the regulatory network containing human embryonic development cell lineage; drawing single-cell anatomical maps of human embryos from fertilized eggs to gastrula; ameliorating genetic defects in embryo development database; and establishing a shared tool and resource platform [31]. In view of the current research progress, EGE has made outstanding contributions in the following three aspects.

Preclinical studies

EGE produces gene knock-out/in animals by injecting engineered nucleases into the single-cell embryo stage, greatly facilitating the construction of disease models [32], research of disease pathophysiology and drug screening experiments. EGE can increase knowledge and understanding of human developmental processes and gene functions, which may benefit in developing or improving medical technology [33]. A humanized animal model for organ transplantation without graft-versus-host reaction was developed by constructing chimeric blastocysts with CRISPR/Cas9 [34].

Biotechnology

On the one hand, engineered nucleases are capable of developing transgenic crops and livestock that are resistant to disease or rich in certain nutrients [35, 36]. On the other hand, EGE is able to eliminate harmful species (e.g. dengue carrier Aedesaegypti) through ‘gene-driven’ mechanism [37].

Gene therapy

EGE aims to treat diseases by introducing normal genes or editing and repairing defective genes. For some maternal genetic diseases (hereditary mitochondrial diseases), monogenic diseases with a wide range of hereditary (muscle dystrophy) and genetic diseases difficult to treat (such as Huntington's disease), single somatic gene therapy is not effective. In this case, EGE may provide a better alternative for treatment. And some other common diseases, such as Alzheimer's disease, diabetes and some cancers, regulated by hundreds of genes. Although EGE cannot completely prevent these diseases, it can effectively reduce the risk.

Chromosomal diseases

Aneuploidy is an inherited disease caused by the addition or deletion of chromosomes, which can lead to significant morbidity and mortality in infancy or childhood [38]. The sgRNA can disrupt repeated sequences at multiple sites so that Cas9 repeatedly cleaves on this chromosome, where it is difficult to successfully be repaired. At present, the biggest problem of application is ‘specificity’. In this regard, Zuo Wei et al. proposed two solutions: on the one hand, considering that most chromosome-specific repeats are located in non-coding regions, we can minimize side effects by targeting small regions without obvious biological functions; on the other hand, due to the single nucleotide polymorphisms, we can specifically target one of the homologous chromosomes to avoid large deletions [39].

Mitochondrial gene

Mitochondrial disease is caused by the proportion (i.e. heterogeneity) of mtDNA mutations in the patient's cells [40]. Compared to other existing mitochondrial replacement therapies, EGE eliminates the need for additional donor oocytes and the procedure is relatively simple and less traumatic to the oocyte. However, when the ‘edit’ embryo has a lower mtDNA copy number below a certain threshold, it is also a risk that the embryo may not be implanted in the uterus [41]. Meanwhile, for mitochondrial diseases caused by homozygous mtDNA mutations, mitochondria-targeted nucleases cannot be used for therapeutic purposes.

Ethical consideration

Faced with the huge interest temptation of human EGE research, scientific researchers must adhere to the bottom line of scientific ethics and carry out responsible research and innovation. Based on the four basic principles of bioethics: respect, non-maleficence, beneficence and justice, we ask the following questions:

Design perfect baby

EGE opens up a technical channel for customizing ‘design babies’. On the one hand, by editing the defective genes in human embryos, we can let the offspring get rid of the nightmare of pathogenic genes. On the other hand, the feasibility of technical fields will inevitably lead some people to cross the ethical boundary for genetic enhancement of non-therapeutic properties. As a result, morality is out of orbit, such as ‘commercialization of life’ and ‘instrumentalization of the body’. In fact, even simple traits such as hair, skin and pupil color are subject to complex gene regulation. For example, variant in MC1R (rs1805008 T) may give rise to unique red hair in offspring, while it also makes his/her skin sensitive to sunlight [42]. In addition, due to the unique oocyte DNA repair mechanism, we cannot randomly add features not existing in parents to offspring. Creating a perfect baby is just a dream. At present, the most important thing is to optimize technology and standardize applications.

Selection of experimental materials

At present, the hEGE experiments mainly choose 3PN embryos as experimental materials. From the moral and ethical point of view, it’s an inevitable choice. In addition, 3PN embryos have extra one sperm nucleus and can develop a large number of cells in vitro [43]. However, when we attempt to decipher normal development, the embryos unable to survive or unsuitable may not be preferred. Since the ‘Warnock Report’ was published by the Committee on Human Fertilization and Embryology in 1985 [44], it is generally accepted that human embryos cannot be maintained in vitro for more than 14 days. However, to study the entire genome and the process of growth and development, it may require a longer window of time and safer technology.

Optimization strategy

In response to the above controversy, the improvement of hEGE in recent years includes (1) further optimization of the activity and specificity of the programmable nuclease and improvement of the DSB maintenance efficiency; (2) exploring more efficient EGE strategy; and (3) constructing an effective academic legal and ethics supervision system.

Optimization of nuclease

Repair efficiency

The ideal repair mode of HDR depends on the generation time of the DSB site, the concentration of the donor DNA and the length of the donor DNA homology-arms [45, 46]. The efficiency of EGE is related with nuclease concentration, activity and targeting. Hashimoto et al. [47] proposed that Cas protein instead of plasmid can improve editing efficiency without transcription and translation. However, the rapid degradation of Cas protein may reduce the target effect, thus, its application remains to be discussed. In addition, we can increase the mutagenesis efficiency by increasing nuclease expression levels, such as inhibiting intracellular nuclease degradation and selecting appropriate promoters to optimize nuclease expression [48, 49].

Nuclease variants

Nuclease optimization has two main directions: (1) Based on the existing nucleases, we improve the characteristics (increasing targeting efficiency, reducing off-target activity) of Cas9 through structural modification, residue replacement and so on [50–53]. (2) We can isolate Cas-like proteins from other prokaryotes, such as Cpf1 and C2c2 [54].

HypaCas9

The classical CRISPR system consists of SpCas9 and gRNA. Based on the structure and mechanism of SpCas9, substitution of basic amino acid (R/K/H) and polar charged amino acids (N/Q) residues with the neutral aliphatic amino acid alanine (A) residue may reduce non-specific binding of SpCas9 to potentially off-target DNA. Based on this hypothesis, scholars have developed SpCas9 variants: eSpCas9, SpCas9-HF1 and HeFSpCas9 [47, 51, 52, 54]. Thereafter, Dr Jennifer Doudna et al. find that REC3, a non-catalytic domain within Cas9, can recognize target mismatches and govern the HNH nuclease to regulate overall catalytic competence. By identifying residues involved in mismatch induction in REC3, they designed a novel high-precision Cas9 variant (HypaCas9) (Figure 3), which retains robust targeting activity in human cells [50].

Cas12a (Cpf1)

‘Cpf1’, the CRISPR binding protein from Francisellanovicida, own dual cleavage activity (DNA and RNA) [55]. Compared to the traditional CRISPR/Cas9 system, it has four advantages: (1) Cpf1 is a single RNA-mediated endonuclease independent of the tracrRNA. Li Wei et al. effectively improved the editing efficiency of Cas12a by optimizing the crRNA skeleton sequence; (2) Cpf1 form cohesive ends instead of blunt ends, in order to achieve precise insertion and integrate DNA more efficiently and accurately; (3) Li Wei et al. found that Cas12a recognizes a 5’T-rich protospacer adjacent motif (PAM), different from the 3’G-rich PAM utilized by Cas9, which can increase the recognition range of the genome. Moreover, Yang Hui et al. found that Cas12b/C2C1 maintains high enzymatic activity among a wide temperature range and pH range, which is suitable for multiple species with different physiological temperatures [56].

Programmable nickases

In order to reduce the off-target effect, Ran Fann et al. combined Cas9's D10A mutant nickase version with a pair of offset sgRNA complementary to the opposite strand of the target site [57], which converted the Cas9 nuclease into a single-stranded DNA nickase. Two sgRNAs recognize a target site of 40 (20 + 20) nucleotides for higher specificity than the classic (Figure 4). At the same time, the SSB can stimulate HDR without activating the error-prone NHEJ pathway, which can essentially prevent unwanted insertions [58]. The double nicking method can reduce off-target activity by 50–1500-fold without reducing the efficiency of knockout.

New CRISPR strategy

CRISPRi/CRISPRa

The key to the transition from ‘CRISPR Editor’ to ‘CRISPR Tuning’ is the catalytically inactive ‘dead’ Cas9 (dCas9). Although the mutated Cas9 (D10A, H840A) cannot cleave DNA double strands, it can still bind tightly to specific targeted DNA under the guidance of gRNA. (1) CRISPRi (CRISPR interference or inhibition): the dCas9 connected with a transcriptional repressor (such as KRAB (Kruppel associated box)) to form a complex (dCas9-KRAB fusion protein), which binds to the TSS site of target gene, inhibits transcription initiation and silences gene expression under the guidance of gRNA [59]. CRISPRi works similarly but slightly differently than RNAi. RNAi mainly targets mature RNA, while CRISPRi inhibits transcription initiation at the DNA level. (2) CRISPRa (CRISPR activation): dCas9 directly conjugates to a single transcriptional activator (such as dCas9-VP64 or dCas9-VP16) [60] or recruits transcriptional activators through molecular hooks (like scFvGCN4-sfGFP-VP64) [61] to activate transcriptional expression of the target gene. By designing multiple gRNAs, multiple target genes can be simultaneously regulated and permanent modification of the genome can be completed. Unfortunately, the limit that CRISPRi/CRISPRa complex can only work near the transcription start site will reduce the off-target effect.

CRISPR-repair/rescue

Zhang Feng et al. first described an RNA targeting nuclease C2C2 (Cas13) and developed a new RNA editing system based on this, named ‘REPAIR’ [54, 62]. Recently, they have made the ADAR2 enzyme a new function by enzymatic evolution, converting cytosine (C) into uracil (U), which named ‘Rescue’ [63]. Wei Wensheng et al. invented a novel RNA editing technique, named ‘LEAPER’, which recruited endogenous ADAR1 protein by arRNA to efficiently and accurately edit specific adenosine on target gene transcripts, avoiding the immune response and damage caused by foreign proteins. Compared to traditional DNA editing systems, RNA editing can be used for non-dividing cells without HDR and more flexible than Cas9/Cpf1 without the requirement of PAM. Furthermore, the off-target effects and insertions and deletions (indels) mediated by NHEJ can be rescued by reversible CRISPR-Repair system based on Cas13a. However, Cas13 cannot directly edit the genome due to lack of the RuvC and HNH domains.

Base editors

In April 2016, the David Liu laboratory of Harvard University first published the base editor (BE) editing technology, which doesn’t require DNA DSBs and a homologous template for single-base conversion [64]. The emergence of BE solves the problem of how to balance the accuracy and efficiency of DNA repair in traditional EGE (Table 2), and promotes the effectiveness and application scope of point mutation gene editing [65]. Yang Hui et al. demonstrated a great deal of RNA off-targets existed in multiple single-base editing techniques such as BE3, BE3-hA3A and ABE7.10 by whole transcriptome RNA sequencing. In addition, ABE7.10 also causes substantial oncogenes and tumor suppressor gene mutations with huge risk of cancer [66, 67]. Based on these, they optimized the three BEs by point mutation and completely eliminated the off-target effects. For the first time, three more high-precision single-base editing tools were obtained, which can provide an important basis for clinical treatment.

Brakes for gene editing

Previously, gene editing was like a car without brakes, dangerous and tempting. Fortunately, scientists found that the survival instinct also let the phage to countermeasures evolved out of the bacterial CRISPR system inhibitory protein, called ACr protein. At present, three kinds of Acrs, AcrIIC1, AcrIIC3 and AcrIIA4, have been found to inhibit Cas9 by different strategies (Figure 5). AcrIIC1, a broad-spectrum Cas9 inhibitor, prevents DNA cleavage of different Cas9 orthologs by directly binding to the conserved HNH catalytic domain. AcrIIC3 prevents the activity of individual Cas9 or thologs and induces Cas9 dimerization while blocking binding to target DNA [68]. AcrIIA4, a DNA mimic, prevents DNA binding by occupying a PAM recognition site of the relevant IIACas9 or thologs [68, 69]. These brakers can improve the safety of CRISPR and increase the accuracy of gene editing by reducing off-target effects.

Ethical regulation

Since 2015, genetic editing has been applied to human embryos for the first time, which lead to fierce debates among global biologists, bioethics scholars and the public. On the one hand, hEGE breaks through the insurmountable ethical ‘red line’. It should be ‘absolutely banned’, which rewrites the way humans evolve with serious technical risks, violates the right to self-determination, leads to new inequalities and damages the ‘genome integrity’ and ‘ethnic integrity’. On the other hand, hEGE is conducive to exploring the mysteries of life and promoting human health, which should be ‘fully open’ for its scientific and moral rationality. It is difficult to balance the seesaw, ‘limited openness’ is an inevitable trend. The construction of an effective supervision system for hEGE is an important guarantee for application.

Establish a sound legal system

At present, although many countries have issued the guides of hEGE, there has not yet formulated corresponding laws and regulations. To this end, all countries in the world should reach a consensus as soon as possible and formulate policy measures with international effectiveness to prevent the ‘empty field’ of the default penalty mechanism. The legal system of human EGE should be established following with the three aspects: (1) we should define the subject and relevant stakeholders of hEGE to bear corresponding responsibilities according to law; (2) we ought to establish a special ethics committee to review and supervise the process and results of hEGE, where the independence and impartiality of the ethics committee must be confirmed; (3) for the ‘deviant’ behavior in hEGE, it is necessary to formulate corresponding regulations to discipline and punish them.

Establish an effective public participation system

hEGE may trigger a range of social and ethical issues related to the future development of humanity. As the main receptor of EGE, the public has the right and responsibility to participate in the supervision. First, the government should respect the public's demands and ensure a benign interaction and organic balance between public participation and policy development in order to form a common will and enhance the legitimacy and rationality of public decision. We must respect the public's right of information and supervision. We should make the public understand the research and development process of hEGE, the benefits of clinical application, public value and overall interest to ensure the democratization and scientization.

Establish a sound risk response system

EGE is a dynamic process of innovation, so are the accompanying risks. Therefore, it’s urgent to estimate and analyze the possible risks and propose effective measures. The risk response system consists of three related links: risk prediction, risk monitoring and risk suppression. Risk prediction avoids the root causes of risks and hidden dangers for future R&D and clinical applications through assessing and predicting the defects of hEGE, based on existing expert experience and technical risk data. Risk monitoring means the effective organization and management of the whole process. Once the potential risk factor changes, it will take corresponding measures to avoid it. Risk inhibition means that once the risk occurs, it will take appropriate measures to minimize the loss, suppress the spread and expansion of risks and enrich the risk database as the future reference at the same time.

Conclusion

There is no right or wrong in technology. The root of social problems caused by science and technology is the flaw of social feedback mechanism. The dispute about hEGE never stops, which mainly focuses on immature technology and the lack of ethical regulation. With the development of science, the defects of the technology will be solved by continuously improving nuclease activity, targeting efficiency and improving the safety and accuracy of gene editing. Ethical issues need to be guarded against by establishing a sound legal, ethical and academic regulatory system. At present, EGE is still a bottom line that should not be crossed. He Jiankui was accidental and inevitable. In this world, there are not only people who ignite the entire forest and just take care of themselves, but also people who try to protect the fire in the dark and open the way for everyone. Therefore, we should be optimistic about the research and development of EGE, boldly try to improve technology and carefully apply its clinical practice.

Authors’ contributions

Qi Zhou is responsible for writing and editing. Yan Zhang and Yujie Zou are responsible for formulating topics and collecting information. Yang Jing and Tailang Yin are responsible for guiding the revision and providing professional advice.

Funding

This work was supported in part by National Key Research and Development Program of China (No. 2018YFC1002804, 2016YFC1000600), National Natural Science Foundation of China (No. 81571513, 81771662, 81771618), and the Major Technological Innovation Projects in Hubei Province (2017ACA101).

Yang Jing, second-level professor, chief physician, doctoral tutor, is the special professor called ‘Outstanding Scholars’ of Wuhan University and the famous IVF expert in China, which enjoys the special allowance of Hubei Provincial Government experts. She is mainly engaged in the diagnosis and treatment of reproductive endocrinology and infertility diseases.

References

Author notes