September 15, 2017 (Vol. 37, No. 16)
Richard A. A. Stein M.D., Ph.D.
Developing Safety Standards for Germline Use
Technologies that would bias the inheritance of a gene or a group of genes in a population have been discussed for decades.
Such technologies, scientists have long proposed, could exploit translocation mechanisms to prevent, contain, and eradicate vector-borne infectious diseases, some of which are global public health emergencies. An especially interesting possibility was introduced back in 2003, when Austin Burt, Ph.D., an evolutionary geneticist at Imperial College London, described how site-specific “selfish genes,” such as homing endonuclease genes, could be engineered to target new host sequences and skew population sex ratios.
At the time, Dr. Burt’s suggestion could not be tested because a convenient means of retargeting selfish elements didn’t exist. Such a means, however, has come to the fore in recent years. It is, of course, the CRISPR/Cas9 genome-editing technology. It is already being used to construct gene drives that could be used to spread desirable mutations through populations in super-Mendelian fashion.
“In CRISPR gene-drive technologies, probably the biggest challenge is making sure that we understand the environmental consequences and the unintended consequences, if any,” George M. Church, Ph.D., professor of genetics at Harvard Medical School and MIT, tells GEN. Several years ago, Dr. Church’s group was the first to create a gene drive in the budding yeast Saccharomyces cerevisiae. In a more recent study on wild and laboratory strains of S. cerevisiae, Dr. Church and colleagues showed that CRISPR/Cas9 gene drive systems can bias inheritance over successive generations at efficiencies over 99%.
Historically, several model organisms have been used to study gene drives, and while each of them provided important lessons, there are key differences between them in terms of the types of information they provide and the challenges they help address. An area of particular concern in gene-drive studies revolves around the accidental escape from the laboratory of even a single organism, and the subsequent consequences over time on wild populations. “Because fruit flies are present outside of every laboratory, escape is easier in this model,” warns Dr. Church.
Dr. Church and colleagues recently developed and validated two molecular confinement methods. One method encodes Cas9 on an unlinked episomal plasmid and ensures that the gene drive element contains only the single guide RNA. (In this arrangement, the single-guide RNA-only gene drive is unable to spread in wild organisms, which lack Cas9.) The other method involves using exclusively synthetic target sequences, which are not encountered in wild-type organisms.
As part of these studies, Dr. Church’s laboratory illustrated the benefits of testing CRISPR/Cas9-based gene drives in the budding yeast before conducting work on multicellular organisms. Additional work with mathematical models led Dr. Church and colleagues to propose the use of alternative designs that could select against resistant alleles and improve the gene drive’s evolutionary stability.
One of the technical challenges in engineering mosquitoes is the need to perform the engineering within or near essential genes. “Engineering genes that are not important to the organism will quickly eliminate the gene drive because the organism does not need the target site,” explains Dr. Church.
Engineering Parasite-Resistant Mosquitoes
“As part of our efforts focusing on malaria, we are trying to create tools and generate mosquitoes that could be used for rigorous tests including large cage trials and subsequently, if the regulatory approvals are given, for field trials,” says Ethan Bier, Ph.D., professor of cell and developmental biology at the University of California, San Diego. Dr. Bier’s group was the first to show that a gene drive can be created in the fruit fly.
Overall, two competing strategies have been envisioned and developed for using gene-drive technologies. One strategy involves the use of mosquitoes to distribute or disseminate an immunizing gene cassette. If this strategy is implemented correctly, notes Dr. Bier, it would not have much or any impact on the health or fitness of the mosquitoes. The other strategy involves using gene drive to sterilize or reduce the population of mosquitoes. According to Dr. Bier, this is a version of genetic insecticide.
Dr. Bier’s laboratory is pursuing the first strategy in collaboration with a team of scientists based at the University of California, Irvine, and led by Anthony James, Ph.D., a professor of microbiology and molecular genetics, and of molecular biology and biochemistry. The collaboration is focusing on population-level mosquito modifications in which genes that confer a parasite-resistant phenotype are engineered into the mosquitoes that transmit the pathogen.
“The immunizing cassette, originally developed by Dr. James’ laboratory, would just stay in the population and not be subject to evolutionary pressures that try to rid those mosquitoes from the environment,” explains Dr. Bier. “They might therefore be present long enough to have a significant impact on the prevalence of the malaria parasite by blocking its transmission.”
In a recent study using Anopheles stephensi, a malaria vector on the Indian subcontinent, Dr. Bier and colleagues in the Dr. James’ group revealed that CRISPR/Cas9-directed homologous recombination drives gene conversion at a more than 99.5% efficiency in mosquito transgene heterozygotes. The technology to perform this work is based on the mutagenic chain reaction, which Dr. Bier and colleagues previously developed in Drosophila melanogaster, and in which a heterozygous mutation is converted to a homologous loss-of-function mutation in germline and somatic cells.
Gene-drive technologies have applications for other vector-borne diseases, such as leishmaniasis and Chagas disease, as well as for population reduction schemes to control crop pests. “Any scheme that goes after reducing the population of any insect or organism in the wild, even though it may be successful, will at the end, always be an uphill battle,” cautions Dr. Bier. A more desirable alternative involves modifying attributes that are undesirable, such as an organism’s ability to propagate disease or its preference for one type of crop versus another. “One can obtain more of an effect by just changing that characteristic and not trying to kill the organism,” advises Dr. Bier.
In introducing genetic modifications into mosquitoes, Dr. Bier and colleagues extensively rely on the ability to generate effector molecules that bind to parasites and render them unable to transfer to the body of the mosquito. One of the key requirements of these molecules is their ability to bind with high affinity to epitopes on the parasite.
“Technologies that we would value for this work include rapid protein evolution binders, which are molecules that are capable of binding random input peptides,” maintains Dr. Bier. After peptides are provided, genetically encodable binders that interact with them could be used in vivo to tether them to components in the mosquito to either kill the parasites or make them aggregate. “[Using] evolutionary synthetic biology approaches to make novel protein-interacting peptides,” adds Dr. Bier, “would be extremely valuable for our work.”
“We look at gene drive as something that can bias inheritance and is differentially included in the offspring, and that can be coupled to a trait that might be of use in terms of controlling the mosquito population,” says Tony Nolan, Ph.D., senior research fellow at Imperial College London. In a recent study, Dr. Nolan and colleagues designed a CRISPR/Cas9-based approach to individually target and disrupt three Anopheles gambiae genes that have high ovarian expression and tissue specificity.
“We disrupted key fertility genes,” informs Dr. Nolan. “That allowed us to introduce an element that can cause population reduction, which is viewed as the most successful strategy today to control malaria.” For two of these loci, the constructs were predicted to disappear from the population over time, but for the third one, the gene disruption met the minimum requirements for targeting female reproduction by gene drive in a mosquito population.
One of the challenges related to the implementation of gene drives is intimately related to the emergence of resistance. “Anything that tries to suppress a population would impose a selection pressure on the population,” states Dr. Nolan. An advantage, when using gene drives, is that some of the resistance mechanisms are foreseeable. “Therefore, one can plan in advance and make the emergence of resistance much less likely,” asserts Dr. Nolan.
Another challenge is the need to demonstrate that gene-drive technology, which is still new, can be trusted. Large amounts of data are needed to confirm that gene drive works and is safe. “There is a lot of testing that should happen between building something in the laboratory and making something the field,” advises Dr. Nolan. “This is going to be a very long process.”
Comparing Alternative Strategies
“We think of gene drives as having three categories of challenges or issues,” says Austin Burt, Ph.D., professor of evolutionary genetics at Imperial College London. The challenges, Dr. Burt suggests, are technological (the ability to “generate constructs that do what we want them to do”); regulatory (the ability to “obtain permission to use the technology that we develop”); and in a sense, communal (the ability to “broaden stakeholder acceptance in terms of people wanting to have this technology”).
Almost 15 years ago, Dr. Burt was the first investigator to propose the use of gene drives based on homing endonuclease genes. Homing endonuclease genes encode highly specific endonucleases with recognition sequences that occur only once in a genome and can activate recombination repair systems by inducing double-stranded chromosomal breaks in the homologous chromosome. As a result of the homology-directed repair process, the endonuclease gene is copied to the broken chromosome. This process can be used to spread the gene through a population.
In a recent modeling analysis, Dr. Burt and colleagues evaluated three different strategies—population suppression through dual-germline fertility disruption, population suppression with a driving-Y chromosome, and mosquito population replacement—to predict how each strategy would perform in a real-life setting from sub-Saharan Africa. Each strategy, despite presenting a unique set of challenges, was highly effective at reducing malaria transmission.
“The point of this work is to help define what it is that would constitute technical success,” declares Dr. Burt. A broader understanding of success, he adds, would encompass the gene-drive attributes that “need to be in place to predict the successful transition of the work from the laboratory to the field.”
Gene Drive in Practice
Many different techniques have been developed, and many more could be developed, to incorporate gene drives, says Zach N. Adelman, Ph.D., associate professor of entomology at Texas A&M University. What these techniques have in common, he suggests, is the need to address the issue of specificity.
The blessing of specificity is that off-target effects are minimized with a highly specific nuclease. The curse of specificity is that when sequence variation is pronounced in a natural population, or when new changes arise, a very specific nuclease will lack or come to lose the ability to recognize the genomic region that needs to be targeted. “This is what investigators have come up against recently,” insists Dr. Adelman.
Even though sequence variations may naturally occur at low levels in a population, such variations could quickly become more prevalent if a gene drive were to bring with it any kind of fitness cost. “We are trying to develop nucleases that are so specific that they do not cause undesirable changes, but are not so specific that it takes only a single change, one that might already occur in nature, to make them nonfunctional,” explains Dr. Adelman.
Dr. Adelman and colleagues recently proposed a two-step approach for gene editing in organisms that are difficult to manipulate genetically, such as mosquitoes. In this approach, the first step is to evaluate candidate site-specific nucleases. (Many synthetic guide RNA molecules are initially examined in vivo.) The second step is to carry out germline-based editing while constraining the choice of DNA repair response. (RNA interference is used to suppress components of the nonhomologous end-joining response.) Suppression of the Ku70 component substantially improved the rates of homology-directed repair and resulted in gene insertion frequencies of around 2–3%.
“The regulatory agencies are still coming to grips with what it means to have a technology that will be used in an environment that is beyond a containment barrier,” says Dr. Adelman. In the case of previous initiatives to generate genetically engineered products, such as salmon and crops, these were confined to a contained environment, did not move beyond where they were breeded (salmon) or planted (crops), and did not admix intentionally with wild population.
“But the goal in gene drive is to admix, and we are still working out the pathways,” declares Dr. Adelman. The pathways from the laboratory to the field will have to be constructed de novo at the same time that the regulatory frameworks are constructed, he suggests.