The Brave New World of Gene Editing and Molecular Medicine

Abstract

Gene therapy has long been a promise of molecular biology. So far, that promise has largely been unrealized. The advent of gene editing using technology adapted from bacteria may finally usher in the era of gene therapy.

Introduction

Humans have been modifying the genomes of diverse organisms for centuries. The extraordinary variety of dog breeds, for example, attests to the power of selective breeding. The familiar livestock and crops that comprise much of our diet are also the result of extensive genome modification resulting from a combination of random mutations and artificial selection.

With the advent of molecular cloning, genetic modification has taken the form of inserting foreign DNA into the genomes of plants and animals. The first published description of a transgenic animal was in 1974, a mouse into which had been introduced simian virus 40 DNA.¹ The first transgenic plant was reported in 1983.² Today, many transgenic crops have been engineered to carry foreign genes that confer insect resistance and herbicide tolerance and genetically engineered transgenic livestock carry foreign genes that enhance milk and meat production.

The application of genome modification in humans has been much more limited. The first gene therapy was performed in 1990 for two patients suffering from severe combined immunodeficiency due to lack of the enzyme adenosine deaminase (ADA). A functional ADA gene was introduced into cultured patient T cells using a recombinant virus, then the T cells were re-infused into the patient. Both patients showed somewhat improved immune responses after this treatment, although the gene therapy alone was not curative. Later advances in stem cell gene therapy improved on these initial results.³

ADA deficiency was an ideal candidate for gene therapy for several reasons:⁴

• the gene is relatively small, making a recombinant version for gene therapy technically easy to create;

• ADA expression levels vary widely across healthy individuals, implying that tight regulation of the therapeutic copy was not necessary;

• correct ADA expression in just some cells is sufficient for therapeutic effect, since the enzyme catalyzes a reaction that produces a product that can complement cells that still lack the enzyme;

• the effects of the disease are completely reversible;

• delivering the therapeutic gene to T lymphocytes is straightforward compared to other tissues or organs.

Clinical gene therapy was dealt a temporary setback with the death of Jesse Gelsinger, who died at the age of 18 four days after being injected with a recombinant adenovirus as part of a clinical trial to treat ornithine transcarbamylase deficiency. His death was the result of an overwhelming immune response triggered by the virus. After extensive investigations, the NIH and FDA devised new programs for patient protection that allowed clinical applications of gene therapy to resume. Currently, more than 2,600 gene therapy clinical trials have been approved, are underway or have been completed.⁵

So modifying genomes is not new. What is new is the unprecedented degree of precision afforded by the latest gene editing technology. The goal of genome editing is not simply to supplement the genome with additional genetic material, it is to edit the existing genetic information to correct or inactivate a gene.

How Does CRISPR/Cas9 Work?

The CRISPR/Cas9 system was first discovered in bacteria, which use it to attack infecting viruses. As the mechanism of the bacterial system has been worked out, the key elements have been identified and streamlined for use in any cell type.⁶ The basic mechanism of CRISPR/Cas9 DNA editing is cartooned in Figure 1. The two key components of the editing machinery are (1) a “guide” RNA that recognizes a specific site in the genome for editing and brings the Cas9 DNA-cleaving enzyme to the site, and (2) the Cas9 enzyme that cuts both strands of DNA at the target. There are two fates for the target DNA after cleavage by Cas9. One is that the two broken ends are reunited by an error-prone cellular repair mechanism called “non-homologous end joining.” This frequently results in the loss or gain of DNA nucleotide subunits at the site of cleavage, which can render the resulting DNA sequence non-functional if it encodes a protein or otherwise directs gene expression. The other outcome depends on the presence of an identical copy of the target DNA sequence carried by a separate DNA molecule. This mechanism, called “homology-directed DNA repair,” can be exploited in genome editing to replace the target DNA sequence with a modified sequence that either creates or corrects a mutation.

Figure 1.

Cartoon representation of the CRISPR/Cas9 DNA cleavage mechanism. The large orange oval highlights the CRISPR guide RNA (green strand) bound to one strand of target DNA (blue strands) across 20 consecutive subunits, or nucleotides. The small orange oval behind it represents the Cas9 enzyme that cleaves both strands of the targeted DNA. The cellular fates of DNA cleaved by the CRISPR/Cas9 are depicted: (left) non-homologous end joining and (right) homology-directed DNA repair. From reference ¹⁷.

Strategies for genome editing have existed for a couple of decades. What makes CRISPR/Cas9-based genome editing so exciting is the high specificity of targeting conferred by the pairing of the guide RNA with the target DNA. With six billion subunits, or nucleotides, of DNA in the human genome, the chance that there exists a close match to the desired editing target elsewhere in the genome is significant. Off-target edits must be avoided, as they could result in unknowable pathologies. Accordingly, much research is focused on maximizing the specificity of CRISPR/Cas9. Compared to the gene therapy for ADA deficiency, CRISPR/Cas9 editing doesn’t depend on the size of the target gene or how the gene is regulated normally.

When applied to human disease, there are two forms of clinical genome editing that are feasible, somatic cell genome editing and germline genome editing. Each strategy has its challenges.

Somatic Cell Gene Editing

Somatic cell genome editing involves destroying or correcting a mutant gene in order to restore healthy function to the patient. The effects of somatic cell editing are restricted to the patient and can’t be transmitted to their progeny. A partial list of some inherited diseases amenable to therapeutic somatic cell gene editing is given in Table I.

In November of 2018, Editas Medicine and Allergan received FDA approval for CRISPR/Cas9 somatic cell genomic editing to treat Leber’s congenital amaurosis type 10, the most common form of inherited childhood blindness. In January of 2019, the FDA announced fast-track approval for clinical trials of a CRISPR/Cas9 somatic cell genome editing strategy for sickle cell disease.

The biggest challenge for somatic cell genome editing is efficiently delivering editing CRISPR/Cas9 complexes to the appropriate target tissues in therapeutically meaningful amounts. There are a variety of potential strategies to deliver the editing molecules to the cells to be edited,⁷ the mechanistic details of which are beyond the scope of this review. For hematopoietic disorders, such as beta-thalassemia, sickle cell disease and severe combined immunodeficiency, culturing patient bone marrow stem cells, editing the stem cell genome ex vivo, and re-grafting the edited cells is technically straightforward. Furthermore, only a fraction of the hematopoietic stem cells need be successfully edited for the patient to experience substantial relief from disease symptoms. For genome editing delivered to other tissues or solid organs, viral vectors encoding the CRISPR/Cas9 components may prove to be the most efficient strategy.

Embryonic Genome Editing

Embryonic genomic editing targets the egg and sperm at the time of fertilization by co-injecting the sperm and the CRISPR/Cas9 editing complex into the egg. Because the editing complex is delivered by injection, editing efficiency is relatively high and the challenges posed by somatic cell editing are avoided. Some fraction of edited cells will certainly populate the presumptive germ line, making the edited chromosome heritable and thus part of the human gene pool. Importantly, this includes not only the desired edits, but also any unintended off-target modifications that might have occurred and gone undetected.

Complicating the risk/benefit of clinical gene editing, some variants may be protective for some conditions while increasing risk for others. An example of this is a common sequence variant found at the SLC39A8 locus.⁸ The SLC39A8 gene encodes a membrane protein that transports metal ions across cell membranes in various tissues, including the brain. An alanine-to-threonine variant at position 391 in the protein product of the gene is one of the few common variants implicated in schizophrenia susceptibility based on genome-wide association studies. However, other studies implicate the same variant in reduced risk for hypertension and Parkinson’s disease.^8,9 Thus, the cost-benefit to gene editing for this variant is not straightforward. Given how little we know about the pleiotropic effects of most genes in the human genome, any genome edit that could enter the human gene pool should be evaluated with extreme caution.

Ethics of Human Genome Editing

On November 28, 2017, Dr. He Jiankui, an associate professor at the Southern University of Science and Technology in Guandong China, shocked the audience at the Second International Summit on Human Genome Editing in Hong Kong by announcing the first babies — twin girls — born with CRISPR/Cas 9-edited genomes. The editing target in this case was the CCR5 gene, which encodes a cell surface receptor used by HIV to infect immune cells. The rationale was that the father was HIV-positive and that the girls might otherwise be born infected. By the next day, the organizing committee published a statement describing He’s claim as “deeply disturbing,” “irresponsible” and “failed to conform with international norms.” The Chinese government ordered He to stop doing science and claimed that what He did was illegal under Chinese law.

Apart from the issue of legality, it appears that the project was unnecessary, as it is effective (and much cheaper) to wash sperm free of virus before insemination. Moreover, there is evidence that people with a naturally occurring mutation that inactivates the CCR5 gene are more susceptible to infectious and chronic diseases.¹⁰ Finally, it isn’t clear that the specific edit He used corresponds to a known naturally occurring mutation, so there may be unknown side effects. The critical importance of establishing internationally recognized rules and policies concerning acceptable uses of human germline editing and to harmonize regulations, in order to discourage unacceptable activities while advancing human health and welfare was underscored at the first international summit on gene editing, sponsored by the U.S. National Academies of Sciences, Engineering, and Medicine.¹¹

Table 1.

Diseases that are candidates for somatic cell gene editing therapy

Disease	Affected gene	Clinical presentation	Prevalence
Alpha-1 antitripsin deficiency	SERPINA1	Lung and liver damage	ca. 1:1,500–3,500 in individuals of European ancestry
Amyloid transthyretin amyloidosis	TTR	problems with the nerves connecting the brain and spinal cord to muscles	3–4% of African-Americans
Beta-thalassemia	HBB	Severe anemia	ca. 1:10,000
Cystic fibrosis	CFTR	respiratory and digestive problems	ca. 1:2,500–3,500 Caucasians
Duchenne muscular dystrophy	DMD	Muscle weakness and damage	ca. 1:3,000 male births
Glycogen storage disease Ia	G6PC	problems with the liver, kidney and small intestine.	ca. 1:125,000
Hemophilia types A and B	F8 (type A) F9 (type B)	Failure of blood clotting	1:4000–5000 (type A) 1:20,000 males (type B)
Huntington’s disease	HTT	Severe progressive neurodegeneration, adult onset	ca. 1:14,000–33,000 individuals of European ancestry
Leber congenital amaurosis 10	CEP290	hereditary childhood blindness	ca. 1:33,000–50,000
Mucopolysaccharidosis types I and II	IDUA (type 1) IDS (type 2)	Multiple tissue and organ damage	ca. 1:100,000 (type 1) ca. 1:100,000 males (type 2)
Ornithine transcarbamylase deficiency	OTC	Development delay, intellectual disability, liver damage	Ca. 1:50,000–80,000
Primary hyperoxaluria type 1	AGXT	recurring kidney and bladder stones leading to kidney failure	ca. 1:72,000
Retinitis pigmentosa	NRL NR2E3	difficulty seeing at night and a loss of peripheral vision	ca. 1:3,500–4000
Severe combined immunodeficiency	IL2RG JAK3 ZAP70	ability to fight off bacterial, viral and fungal infections	ca. 1:50,000
Sickle cell disease	HBB	Anemia, pain, organ damage	1:500 African-Americans; ca. 1:1,000–1,400 Hispanic Americans
Sly syndrome (Mucopolysaccharidosis type VII)	GUSB	Affected growth and motor skills; mental retardation	Less than 1:250,000
Tay-Sachs disease	HEXA	Progressive neurodegeneration	1:3,600 people of Ashkenazi Jewish descent
Usher syndrome type 2a	USH2A	Progressive hearing and vision loss	ca. 1:10,000–100,000

Adapted from refs. ⁷, ¹⁹ and 20.

Ethics of Somatic Cell Genomic Editing

The ethics of somatic cell genomic editing are, in principle, no more problematic than any therapy, as long as the proper testing for safety and efficacy is conducted. The consequences of somatic cell genomic editing are borne entirely by the patient and the edited genome dies with the patient. Under those circumstances, informed consent from the patient or their guardian would be sufficient to implement the therapy.

Ethics of Human Germ Line Cell Genome Editing

Because human embryonic/germ line cell genome editing results in genome modifications that can enter the human gene pool, the ethical implications extend to our entire species. Thus, all of us are stakeholders in the future application of this technology. As illustrated by the swift and dramatic condemnation of Dr. He Jiankui discussed above, it is clear that human society is unwilling to extend blanket approval to the approach. Indeed, recent calls have been made for a moratorium on human germline genome editing (see below).

With the widespread use of in vitro fertilization and embryo selection, it is already possible for couples to choose a conceptus that is free from genetic disease. Accordingly, the cases in which germline editing is preferable to embryo selection are few.

Ethics of Human Genome Editing for Enhancement

The most ethically problematic application for genome editing is for genome enhancement, the editing of the human genome with the goal of increasing traits such as, e.g., intelligence, strength, endurance or physical attractiveness. This application falls into the category of eugenics and is fraught with the question of whose values are reflected in a decision to valorize a particular trait. Furthermore, as with the cases of the SLC39A8 and CCR5 genes, the benefit being sought may be offset by negative effects.

Other Applications of CRISPR/Cas9 Affecting Human Health

The potential for editing non-human genomes to advantage human health is huge. In addition to crop and livestock improvement, there is considerable interest in genome editing to eliminate the infectious diseases that have afflicted humanity for centuries.

For example, using CRISPR/Cas9 to modify the genomes of insect disease vectors to prevent disease transmission is a focus of research. In the case of malaria, promising results have been obtained for two strategies using CRISPR/Cas9 to (1) drive a mutation that causes recessive female sterility into the Anopheles mosquito population,¹² or (2) drive a mutation that makes the mosquito a poor vector for the malarial protozoan.^13,14 (Figure 2). In both reports, targeted transgene constructs carrying CRISPR/Cas9-based constructs result in >95% transmission of the mutant allele to progeny, where normal Mendelian inheritance would predict 50%. Unfortunately, neither of the reported strategies is yet ready for field-testing. Insects that carry a single copy of the CRISPR/Cas9 drive construct in each case are less genetically fit than wild-type mosquitoes, and thus the drive construct would be selected against in the wild. Further research is geared to making heterozygous transgenic mosquitoes at least as fit as wild-type mosquitoes. Also unknown is how likely it is that resistance to CRISPR/Cas9-based drive might appear, much like insecticide, herbicide, antibiotic and antiviral drug resistance arises over time. For example, variants in the CRISPR target sequences that make the targeting much less efficient will arise in large populations. Additionally, targets cleaved by Cas9 that are repaired by non-homologous end joining will result in a sequence immune to further CRISPR/Cas9 editing. These challenges may be addressed by using multiple CRISPR targeting RNAs in the drive construct.

Figure 2.

Mechanism and genetic transmission of CRISPR/Cas9-directed gene drives. (a) In an animal heterozygous for the gene drive CRISPR/Cas9 transgene, the Cas9 endonuclease (scissors) is targeted to the wild-type copy of the gene. When the cell repairs the resulting chromosome break using homologous recombination, it can use the gene drive chromosome as a repair template, thereby copying the drive onto the wild-type chromosome. (b) When a mosquito carrying the CRISPR/Cas9 endonuclease gene drive transgene (blue) mates with a wild-type mosquito (grey), the gene drive is preferentially inherited by most or all offspring. This can enable the drive-containing chromosome to spread over several generations until it is present in all members of the population. From ref. 18.

Other examples of insect- or tick-borne diseases that would be amenable to CRISPR-Cas9 vector control include Chagas and Lyme disease, chikungunya, Dengue and yellow fever, leishmaniasis, trypanosomiasis, West Nile, and Zika.

While the public health benefit to interrupting the life cycle of human pathogens is obvious, there may be unintended consequences to driving specific insect species to extinction. Therefore, strict control of field trials is essential.

Whither Genome Editing?

The genome editing genie is out of the bottle. Somatic cell genome editing is certain to become standard therapy for many inherited diseases, just as genetic modification of patient immune cells by CAR-T technology is moving into mainstream cancer therapy.

For germline genomic editing, the genie is also out of the bottle, despite the ethically fraught implication of eugenics. The Homo sapiens community will have to police the applications of germline genomic editing, since the human genome is ultimately the heritage of our species. The success or failure of germline editing regulation depends ultimately on the perception of “moral hazard,” the idea that bad behavior can be constrained by awareness that the consequences are borne by everyone. Encouragingly, the moral hazard of nuclear proliferation, both for power generation and for nuclear weapons has achieved a shaky global comity that has held for over 70 years. However, the recent surge in vaccine resistance in the West is a cautionary counter-example, where misplaced fear of vaccine-related harm has broken the implicit societal compact of universal vaccination.

Recently, the journal Nature published a call from an international group of researchers and ethicists for a moratorium on the clinical application of human germline genome editing for up to five years.⁸ The moratorium proposal has been endorsed by the U.S. National Academy of Sciences, the U.S. National Academy of Medicine, the U.K. Royal Society and the director of the National Institutes of Health.^15,16 During the moratorium period, the authors propose additional research on the safety of the technology and the development of an acceptable use policy. It is acknowledged that neither the moratorium nor any policies that emerge from it will be enforceable outside the borders of the nations that adopt them. What is certain is that the future will include human genome editing, with profound implications for our species.