SELF-RECOMMENDING!

[ACC] Should Gene Editing Technologies Be Used In Humans?

[This is an entry to the 2019 Adversarial Collaboration Contest by Nita J and Patrick N.]

Introduction

In October 2018, the world’s first genetically edited babies were born, twin girls given the pseudonyms Lulu and Nana; Chinese scientist He Jiankui used CRISPR technology to edit the CCR5 gene in human embryos with the aim of conferring resistance to HIV. In response to the international furor, China began redrafting its civil code to include regulations that would hold scientists accountable for any adverse outcomes that occur as the result of genetic manipulation in human populations. Now, reproductive biologists at Weill Cornell Medicine in New York City are conducting their own experiment designed to target BRCA2, a gene associated with breast cancer, in sperm cells. While sometimes considered controversial, gene editing has been used as a last resort to cure some diseases. For example, a precursor of CRISPR was successfully used to cure leukemia in two young girls when all other treatment options had failed. Due to its convenience and efficiency, CRISPR offers the potential to fight cancer on an unprecedented level and tackle previously incurable genetic diseases. However, before we start reinventing ourselves and mapping out our genetic futures, maybe we should take a moment to reevaluate the risks and repercussions of gene editing and rethink our goals and motives.

How does CRISPR work?

CRISPR, which stands for clustered regularly interspaced short palindromic repeats, is an adaptive bacterial immune response that protects against repeat offenders. When exposed to a pathogenic bacteriophage, a bacterium can store some viral phage DNA in its own genome in “spacers,” which function as genetic mug shots, allowing the bacterium to quickly mount a defense in case of future invasions. When necessary, the CRISPR defense system will slice up any DNA matching these genetic fingerprints. In 2012, Jennifer Doudna and Emmanuelle Charpentier demonstrated how CRISPR could be used to slice any DNA sequence of choice. The CRISPR-Cas9 system allows researchers to not only recognize and remove DNA sequences but also modify them. The completion of the Human Genome Project in 2003 provided a copy of the genetic book of life; CRISPR offers a way to purportedly erase and “correct” certain words in that book.

Of course, this newfound power raises several ethical concerns. The major worry among scientists revolves around the long-term consequences of germline modification, meaning genetic changes made in a human egg, sperm, or embryo. Edits made in the germline will affect every cell in an organism and will also be passed on to any offspring. If a mistake is made in the process and a new disease inadvertently introduced, these changes will persist for generations to come. Human germline modification could also theoretically allow for the installation of genes to confer protection against infections, Alzheimer’s, and even aging. For many, the thought of controlling our own genetic destinies seems to be a very slippery slope, conjuring up dystopian images of Frankenstein or Brave New World. For these reasons and more, in 2015, Doudna and other scientists proposed a moratorium on the use of CRISPR-Cas9 for human genome editing until safety and efficacy issues could be more thoroughly addressed.

How safe and efficient is gene editing?

CRISPR is currently being used in clinical trials for cancers and blood disorders; since these interventions won’t lead to heritable DNA changes, these trials don’t face the same ethical dilemmas as Dr. He’s experiment but may nevertheless carry risks. Doubts persist about the safety and efficacy of the CRISPR gene editing system, as many other initially promising technologies have failed. Conventional gene therapies, which attempt to insert healthy copies of genes into cells using viruses, faced many early setbacks, including the tragic death of 18-year-old Jesse Gelsinger in 1999 during a gene therapy trial for ornithine transcarbamylase deficiency. However, the causes surrounding Gelsinger’s death may have included a systemic immune response triggered by the use of a viral vector.

While the death of Jesse Gelsinger marked a somber moment for the field, gene therapy also experienced successes when researchers from Paris treated two young infants who suffered from a fatal form of severe combined immunodeficiency disease (SCID), an inherited disorder characterized by low levels of T cells and natural killer cells, which leaves affected patients incredibly susceptible to infection. Fortunately, viral gene therapy was able to reverse the disease symptoms in this particular case. On the other hand, gene therapy trials using viral vectors were recently halted when 25-50 percent of gene therapy patients developed leukemia resulting from the insertion of a gene-carrying virus near an oncogene; a gene with the potential to cause cancer. Modern CRISPR technology is not affected by such hurdles, however, as it does not rely on the use of viral vectors. While more precise than traditional gene therapy, CRISPR nonetheless sometimes results in unintended edits, which may be especially problematic for certain gene targets. Some pairs of genes are “linked” due to physical proximity on the same chromosome and are therefore almost always passed on together. Any edits to a gene belonging to a linked pair may therefore inadvertently cause an edit in its neighboring partner.

Even intended cuts can have unexpected consequences. Two separate 2018 studies published in Nature Medicine, one conducted by the Karolinska Institute in Sweden and the other by Novartis Institutes for Biomedical Research, concluded that CRISPR edits might increase the risk of cancer via inhibition of a tumor suppressor gene called P53, which has been described as “the guardian of the genome” due to its crucial role in maintaining genomic stability. Double-stranded DNA breaks made by CRISPR activate repair mechanisms encoded by P53 that instruct the cell to either mend the damage or self-destruct. Making these types of edits successfully would therefore require inhibition of P53; however, cells could become more vulnerable to tumorigenic mutations and the development of cancer as a result. “We don’t always fully understand the changes we’re making,” says Alan Regenberg, a bioethicist at Johns Hopkins Berman Institute of Bioethics. “Even if we do make the changes we want to make, there’s still question about whether it will do what we want and not do things we don’t want.”

Nevertheless, a slight increase in cancer risk might be a worthwhile trade-off for many patients with genetic diseases, such as the aforementioned SCIDs, which affect 1 in 50,000 people globally. Usually, the only cure for SCIDs is a bone marrow transplantation, which requires a matched donor in order to avoid rejection by host immune cells or, alternatively, the depletion of T cells to avoid rejection in the case of an unmatched donor. CRISPR offers a safer, more efficient way to treat genetic diseases such as SCIDs. Bone marrow cells of a patient may be extracted and genetically modified using CRISPR, thereby avoiding rejection by the host immune system. Pre-clinical trials in mice are already underway to test the safety and efficacy of this approach. Stanford scientist Dr. Matthew Porteus demonstrated the efficiency of this technique and said in an interview, “We don’t see any abnormalities in the mice that received the treatment. More specifically, we also performed genetic analysis to see if the CRISPR-Cas9 system made DNA breaks at places that it’s not supposed to, and we see no evidence of that.”

CRISPR also offers the additional possibility of removing parts of a gene, providing extra value over standard viral gene therapy, which only allows for insertion of genes. This feature can be especially important for autosomal dominant genetic disorders, which are made manifest with only one copy of a deleterious mutation. In her book, A Crack in Creation, Jennifer Doudna speculates that as CRISPR becomes increasingly safe, the tool may be used to help people who aren’t fortunate enough to win the genetic lottery. Doudna intones, “Someday we may consider it unethical not to use germline editing to alleviate human suffering.” What was unthinkable just a few years ago will soon enter clinical praxis.

Are some genetic variants superior to others?

In biology, those organisms that are most suited to their environment exhibit the highest fitness, a measure that accounts for both survival and reproduction. The accumulation of mutations over time is thought to contribute to many disease processes, but genetic diversity can also be beneficial for an organism when faced with a changing environment or unanticipated stress, such as drought or illness. Discussions on rigid natural selection should give way to more nuanced conversations on “balancing selection, the evolutionary process that favors genetic diversification rather than the fixation of a single ‘best’ variant,” as described by Professor Maynard V. Olson at the University of Washington.

Evolution has allowed many potentially deleterious genes to remain in the gene pool due to their ability to impart a selective advantage to individuals with carrier status, a phenomenon referred to as heterozygote advantage. Sickle cell anemia is a disease inherited in an autosomal recessive pattern—two copies of the problematic gene variant are necessary for disease expression. However, having just one copy of that variant confers resistance to malaria, which may explain the increased prevalence of sickle cell anemia in areas where malaria is more common, namely India and many countries in Africa. In this manner, malaria acts as a selective evolutionary pressure maintaining the occurrence of the sickle cell variant in the gene pool.

Nevertheless, sickle cell disease has become prevalent in countries currently unaffected by malaria. In the United States, approximately 100,000 people suffer from sickle cell disease, but therapeutic options remain limited. Researchers have been investigating the possible insertion of wild-type, “anti-sickling” genes using viral vectors in affected patients as therapy. However, since the pathological mutation for sickle cell disease has already been clearly identified, correction of the mutated gene using CRISPR may offer a more straightforward approach. The biotech company CRISPR Therapeutics recently announced the results of a phase I clinical trial in which CRISPR technology was used to treat a patient with sickle cell disease, although the efficacy and safety of the intervention have not yet been evaluated.

Can gene editing eliminate disease?

To answer these questions, we need to first evaluate our understanding of genetics and weigh the importance of genetics against environmental factors such as diet and lifestyle.

How reliable is our understanding of gene-disease links?

A mutation is usually defined as a genetic sequence that differs from the agreed-upon consensus or “wild-type” sequence. After the completion of the Human Genome Project in 2003, the arduous process of genome annotation began. Genome-wide association studies, or GWAS, began examining population data over time to look for possible associations between genetic variants, or genotypes, and physical traits and diseases, or phenotypes. Unfortunately, these studies often fail to employ random sampling, and 96 percent of subjects included in GWAS have been people of European descent. In fact, scientific disciplines frequently disproportionately sample from WEIRD (western, education, industrialized, rich, democratic) populations, whether studying genetic diseases or human gut microbiota.

Given the sources of genetic information used to determine “wild-type” sequences, we may be using information that is relevant to one demographic but not another. According to Maynard Olson, one of the founders of the Human Genome Project, the wild-type human simply doesn’t exist, and “genetics is unlikely to revolutionize medicine until we develop a better understanding of normal phenotypic variation.” These words seem to have fallen on deaf ears, however, as evidenced by the burgeoning numbers of genome-wide association studies conducted over the last 12 years. Most of the associations discovered thus far are only correlative, and few studies have been conducted to determine whether observed associations are indeed causal.

Closer examination of the relationship between gene variants and certain diseases reveals weak associations in many cases. For example, the APOE gene, which encodes for the production of a protein known as apolipoprotein E, comes in three genetic forms- APOE2, APOE3, and APOE4 with the last being associated with an increased risk of developing Alzheimer’s disease (AD). However, the correlation is not determinative, as the Nigerian population exhibits high frequencies of the APOE4 allele but low frequencies of AD. Environment and nutrition also play significant roles in the disease pathophysiology, as illustrated by Dr. Dale Bredesen’s research demonstrating reversal of cognitive decline through a targeted dietary and lifestyle approach. In fact, the majority of afflictions commonly affecting the general population, such as type 2 diabetes, cardiovascular disease, cancer, Alzheimer’s, and Parkinson’s are not caused solely by mutations.

How often does disease arise as the result of genetic mutation alone?

Chronic diseases are the result of a complex interplay between host genetics and the environment. A study conducted by the Wellcome Trust Sanger Institute in Cambridge, England analyzed DNA sequencing data from 179 people of African, European, or East Asian origin as part of the 1000 Genomes Pilot Project and discovered that healthy individuals carried an average of 400 mutations in their genes, including around 100 loss-of-function variants that result in the complete inactivation of about 20 genes that encode for proteins. These findings indicate that deleterious mutations, even those that lead to protein damage, do not invariably give rise to disease. As Professor James Evans from the University of North Carolina, who was not involved in the study, summarized in an NPR health blog, “We’re all mutants. The good news is that most of those mutations do not overtly cause disease, and we appear to have all kinds of redundancy and backup mechanisms to take care of that.” The authors hypothesize that healthy individuals can carry disadvantageous mutations without showing ill effects for a number of possible reasons: an individual may carry just one copy of a gene mutation for a recessive disorder that requires two mutations in order to manifest the disease, the disease may exhibit delayed onset or require additional environmental factors for expression, or the reference catalogs used to identify gene-disease links may be inaccurate. One analysis found that 27 percent of database entries cited in the literature were incorrectly identified.

To account for the discrepancy between genetic predisposition and disease manifestation, in 2005, cancer epidemiologist Dr. Christopher Wild proposed the concept of the exposome, which encompasses “life-course environmental exposures (including lifestyle factors) from the prenatal period onwards” and accounts for factors such as socioeconomic status, chemical contaminants, and gut microflora. The risk of developing a chronic disease during one’s lifetime may be modeled by G×E: the interaction between a person’s genetics (G) and lifetime exposures (the exposome, E). Identical twin studies reveal that genotype alone cannot determine whether a given phenotype will be expressed, and the interaction between genes and the environment must be taken into account. 

In fact, the “genes load the gun, environment pulls the trigger” paradigm may be overly simplistic, as Dr. Alessio Fasano at Harvard Medical School has shown that loss of intestinal barrier function is likely also necessary for the development of chronic inflammation, autoimmunity, and cancer. Two particular gene markers, HLA-DQ2 and HLA-DQ8, are observed in the vast majority of celiac disease cases. While over 30 percent of the U.S. population carries one or both of the necessary genes, only around one percent of Americans are affected by celiac disease. This data suggests that exposure to gluten through ingestion of wheat, barley, or rye is not sufficient to trigger the development of celiac disease even in individuals with a genetic predisposition. Without the additional loss of intestinal tight junction function, celiac disease is not made manifest. Thus, factors besides genetics are necessary for the development of chronic disease.

How does gene expression contribute to disease risk?

The concept of genetic determinism purports that our genes are our destiny, but genes are not nearly as important as gene expression. When most people think of evolution, the first name that comes to mind is Charles Darwin, but a contemporary of Darwin’s named Jean Baptiste Lamarck had proposed a theory of “acquired characteristics” by which individuals evolved certain traits within their lifetimes. The most oft-cited example discrediting this theory is that of giraffes elongating their necks by stretching to reach the treetops and then passing on this trait of long necks to their progeny. In contrast, Darwin proposed that those giraffes that had the longest necks went on to find food, survive, and reproduce. Eventually, Darwin’s theory of natural selection prevailed, but his 18th century French naturalist contender may have simply foreseen the field of epigenetics, the study of those drivers of gene expression that occur without a change in DNA sequence. The prefix “epi-” means above in Greek, and epigenetic changes determine whether genes are switched on or off and also influence the production of proteins. If you imagine your genetic code as the hardware of a computer, epigenetics is the software that runs on top and controls the operation of the hardware. Epigenetic changes control the expression of genes through various mechanisms and are influenced by diet, exercise, lifestyle, sunlight exposure, circadian rhythms, stress, trauma, exposure to pollutants, and other environmental factors.

The epigenetic mechanism of DNA methylation involves tagging DNA bases with methyl groups, a process that tends to silence genes. DNA methylation is responsible for X-chromosome inactivation in females, a process necessary to ensure that females don’t produce twice the number of X-chromosome gene products as males. Methylation is also responsible for the normal suppression of many genes in somatic cells, allowing for cell differentiation. Every somatic cell in the human body contains nearly identical genetic material, but skin cells, muscle cells, bone cells, and nerve cells exhibit different properties due to different sets of genes being turned on or off. Dietary nutrients such as vitamin B12, folic acid, choline, and betaine double as methyl donors, so even small changes in nutritional status during gestation can result in markedly different effects on gene expression and varied physical characteristics in the offspring. If differential gene expression can produce such drastic changes, is genome rewriting really necessary? Perhaps the centrality of the gene in driving human health has been overstated. Indeed, why worry about a potentially pathogenic gene if it is never expressed?

Inappropriate DNA methylation has been referred to as a “hallmark of cancer,” along with uncontrolled cell growth and proliferation. Almost all types of human tumors are characterized by two distinct phenomena: global hypomethylation, which may result in the expression of normally suppressed oncogenes, genes that promote tumor formation, as well as regional hypermethylation near tumor suppressor genes. In other words, genes that promote tumor formation are turned on while genes that suppress tumor formation are turned off. Cigarette smoke has been shown to promote both demethylation of metastatic genes in lung cancer cells as well as regional methylation of other specific genes via modulation of enzymatic activities. To succinctly summarize, genes themselves are not driving tumor formation; rather inappropriate gene expression is increasing the risk of tumor development.

Can gene editing treat cancer?

Cancers are front and center among the conditions gene editing therapies are targeted to treat. To answer the question of whether CRISPR can be used to treat cancer, we need to first examine how cancer arises. Medical textbooks frequently attribute the development of cancer to the accumulation of mutations over time. However, the accumulation of genetic mutations is not sufficient to cause cancer; the tumor microenvironment must be taken into account. In other words, the same oncogenic mutation that is adaptive for cancer in altered tissue is not advantageous to cancer in healthy, homeostatic cells. 

James DeGregori at the University of Colorado School of Medicine offers the following analogy. When tackling drug dealing in the inner city, arresting all the drug dealers is unlikely to work; the ones left behind will be smarter and more conniving. Instead, one might focus on creating better jobs, schools, and infrastructure, so citizens won’t have to resort to crime as a means of survival. Addressing the environment that lead to the problem in the first place will provide a more stable long-term solution. Similarly, instead of simply targeting the cancer, altering the microenvironment to disfavor its proliferation may provide a more viable long-term strategy, as the former immediately selects for resistance, accounting for the difficulty in keeping a patient in remission. Highlighting the importance of the microenvironment in regulating development, homeostasis, and cancer, biologist Mina Bissell writes, “The sequence of our genes are like the keys on the piano; it is the context that makes the music.” Cancer depends on context, as should our approach to treatment.

Nevertheless, despite recent medical advances, cancer treatment has not seen significant improvement in decades. Standard therapies rely on toxic chemotherapy, which destroys both cancerous and healthy tissue. Furthermore, cancerous cells often evade detection and destruction by host immune defenses by expressing cell surface molecules that prevent killing by host T cells. A new and effective form of immunotherapy known as chimeric antigen receptor (CAR) T cell therapy attempts to harness the power of the human immune system to recognize and kill cancer cells. However, this method has several disadvantages. A patient must have a sufficient number of immune cells prior to beginning therapy, which may not be the case for patients who have already received chemotherapy. Additionally, the process is time-consuming, and the use of viral vectors may increase the risk of developing other cancers. 

To address the issues of T cell collection and manufacturing delays, researchers are now developing “off-the-shelf” CAR T cells, which utilize gene editing to prevent rejection by the host immune system and the development of graft-versus-host disease (GvHD), a condition in which foreign immune cells attack the recipient’s body. In 2017, two infants with relapsing leukemia were successfully treated with these “off-the-shelf” CAR T cells, which were modified using the genome editing tool TALEN. Short for transcription activator-like effector nucleases, TALEN can be considered the predecessor to CRISPR and uses enzymes that are specifically guided to a genomic sequence to induce a cut. However, designing these enzymes requires extensive work, making the process costly and time-consuming. Additionally, in vitro studies have demonstrated that CRISPR techniques exhibit better correction efficiencies and fewer off-target effects than TALEN. Moreover, the use of CRISPR can speed up the manufacturing of CAR T cells and drive down costs of such therapies from hundreds of thousands of dollars to a few hundred dollars.

Can gene editing prevent HIV?

Another prospective application for CRISPR technology is the treatment of HIV. Today, approximately 37 million people around the world live with HIV. The use of antiretroviral drugs has greatly reduced the death rate from 1.9 million in 2004 to less than one million in 2017. Challenges still exist, as human immunodeficiency virus (HIV) inserts itself into the host genome and mutates rapidly, making complete eradication of the disease very difficult. About one percent of the population is naturally immune to HIV due to a CCR5 gene mutation, which prevents the expression of a cell surface receptor that HIV binds to in order to gain entry into host cells. As previously mentioned, the first genetically edited babies were born in October 2018 after Chinese scientist Dr. He Jiankui used CRISPR technology to edit the CCR5 gene in human embryos.

According to Dr. He, a married couple with the pseudonyms Mark and Grace consented to in vitro fertilization with additional CRISPR treatment to provide immunity to HIV for their offspring. First, a process called sperm washing was used to separate sperm from semen, the fluid that carries HIV. Next, eggs were fertilized by sperm to create embryos, on which Dr. He performed CRISPR gene editing. After several implant attempts, successful pregnancy was achieved. Nine months later, twins with the pseudonyms Lulu and Nana were born healthy and purportedly suffered no off-target effects from the CRISPR therapy.
 
Testing indicated that gene editing did not successfully alter both copies of the CCR5 gene in one of the twins, however. Chinese researchers were apparently knowledgeable of the gene editing failure prior to the pregnancy attempt; the decision to proceed with implantation regardless has numerous ethical implications. “In that child, there really was almost nothing to be gained in terms of protection against HIV and yet you’re exposing that child to all the unknown safety risks,” said Dr. Kiran Musunuru, a professor of stem cell and regenerative biology at Harvard University. The choice to use the unedited embryo suggests that the researchers may have been more focused on testing the accuracy of the gene editing technology than providing immunity to disease.

According to the Chinese government and his employers, Dr. He acted without the knowledge or consent of his superiors. Chinese authorities suspended all of He’s research activities, saying his work was “extremely abominable in nature” and a violation of Chinese law. In fact, the procedure was not medically necessary. When only the father is HIV-positive, as in this case, sperm washing alone is usually sufficient to reduce transmission of the virus. A meta-analysis that investigated the efficacy of sperm washing did not find a single case where HIV was transmitted to offspring.
 
Dr. He claims that the CCR5 gene is already very well characterized, but a recently published study found that decreased function of the CCR5 gene enhances cognitive function in mice. At first glance, this new knowledge may appear to be a boon, but the potential benefit also invites a discussion on the possibility of designer babies. Another point to consider is the fact that the CCR5 mutation that confers HIV immunity more commonly appears in Caucasians and may make individuals more susceptible to infections that are common in Asia.

Can gene editing be used to create designer babies?

A discussion on human genome editing would not be complete without evaluating the potential to create “designer babies,” a term commonly used in the vernacular to refer to babies with genetic enhancements. Both the utility of gene editing for basic research and the use of somatic gene editing to heal individuals who are sick are generally widely accepted among the public. The waters become murkier when we consider germline editing and the possibility of preventing disease or altering traits unrelated to health needs. In the 1970s, scientists first began to establish distinctions between somatic and germline genome modifications; somatic edits only affect a single individual while germline edits can be passed down over generations. By the mid-1980s, bioethicists began to argue that the morally relevant line was between disease and enhancement rather than somatic and germline. Discussions of heritable enhancements in particular raise fears of a possible return to eugenics.

John Fletcher, former head of bioethics at the National Institutes of Health (NIH), once wrote, “The most relevant moral distinction is between uses that may relieve real suffering and those that alter characteristics that have little or nothing to do with disease.” Many scientists today share the sentiment that treatment and prevention of “disease” constitute acceptable uses of CRISPR technologies while “enhancement” applications should be discouraged, but the boundary between the two is riddled with semantic discord. Moreover, the line delineating disability and disease is often blurred, and many perceived shortcomings may in fact represent normal variation on the phenotypic spectrum.

The discussion of whether we can or should modify human characteristics may be a moot point since our knowledge of which genes affect complex traits such as height, intelligence, and eye color is still limited. Additionally, most traits are influenced not only by genetics but also environmental factors, and monozygotic twin studies demonstrate that genes alone cannot predict whether physical traits will be expressed. Furthermore, genes that encode for physical traits may also impart increased vulnerability to certain diseases. For example, variations in the MC1R gene responsible for red hair color may also increase the risk of developing skin cancer. As indicated earlier, Dr. He’s efforts to confer resistance to HIV may have also resulted in increased susceptibility to infection by West Nile virus or influenza. As always, trade-offs exist, and the idea of the “perfect specimen” is a fallacy. Any efforts to gain genetic advantages will always be subject to the limitations of biology.

How should society move forward with gene editing technology?

CRISPR technology holds invaluable potential as a research tool and possible treatment for diseases caused by single-point genetic mutations. As previously described, some genetic diseases can be treated by stem cell gene editing without the need for germline modification, thereby minimizing the risk for potential mistakes that could be passed on to subsequent generations. On the other hand, trying to correct an error after a certain point during development is sometimes problematic, as the error has already been incorporated into billions of cells. Jennifer Doudna offers the following visual: “Imagine trying to correct an error in a news article after the newspapers have been printed and delivered, as opposed to when the article is still just a text file on the editor’s computer.” Germline editing may therefore provide a more expedient option for the prevention of some genetic diseases such as sickle cell disease or cystic fibrosis.

One of the most compelling arguments against CRISPR gene editing, namely the potential for misuse, can also be considered the most compelling argument for CRISPR gene editing. Banning progress on gene editing technology may create a black market, but the continuation of research on gene editing will allow the scientific community to control its use and ensure patient safety. Research into CRISPR is continually finding ways to make the technology safer and more effective; a paper published in September 2019 reported on the potential for a novel CRISPR system to affect gene expression in human cells. The process is reversible in theory and doesn’t involve the cutting of DNA, thereby reducing the risk of human harm and leveraging the power of epigenetics.

Moreover, while gene expression and the tumor microenvironment are viable targets for cancer treatment, gene editing can be considered a last resort therapy for certain cases in which other interventions have failed. Common chronic diseases, such as Alzheimer’s, type 2 diabetes, and cardiovascular disease, likely require a more nuanced approach, as gene expression, governed by factors such as diet and lifestyle, plays a significant role in disease pathogenesis. The use of gene editing to mold favorable traits, such as eye or hair color, likely exposes individuals to unnecessary risks and does not constitute medical necessity. Nevertheless, many consider mainstream germline gene editing an inevitability. Joseph Fletcher, one of the founders of bioethics, wrote in 1971, “Man is a maker and a selector and a designer, and the more rationally contrived and deliberate anything is, the more human it is.” The establishment of gene editing guidelines should include input from scientists, policy makers, and the public and incorporate the most current knowledge available in order to prevent misuse and realize potential. As the custodians of such powerful technology, we must take care to use it in an ethical and responsible manner. Whether our efforts will alleviate human suffering or ensure the survival of our species, only time will tell.

This entry was posted in Uncategorized and tagged . Bookmark the permalink.

84 Responses to [ACC] Should Gene Editing Technologies Be Used In Humans?

  1. Sniffnoy says:

    This doesn’t really relate so much to the post as opposed to the subject in general (except insofar as I think it’s relevant that the post didn’t mention this), but, to repeat what I said on an earlier thread…

    The possibility that scares me isn’t people using gene editing to enhance their children — I mean, greater capabilities are a good thing — but to make their children more things-their-parents-want-but-that-are-not-in-fact-better. What specifically scares me is the possibility that a large portion of the population, given the ability to adjust the personalities of their babies — and I seem to recall that there are survey results to this effect, although I could not possibly cite them — would want children who are more agreeable or obedient. That’s… not really a good thing. The idea of a more agreeable, obedient humanity strikes me as not only horrifying but as maybe even an existential risk, not in the sense of something that could lead to extinction but in the sense of something that could permanently curtail human potential. That’s what scares me, not some idea that selecting for actual enhancements will somehow backfire…

    (I think I likely picked up this argument from someone else, but if so, I forget who, sorry…)

    • Aapje says:

      greater capabilities are a good thing

      Not always. What if people start competing on breast & butt size, duck lips, length, etc. We may get very non-optimal outcomes, especially if people overshoot the optimum.

      The idea of a more agreeable, obedient humanity strikes me as not only horrifying but as maybe even an existential risk, not in the sense of something that could lead to extinction but in the sense of something that could permanently curtail human potential.

      It would plausible reduce reproduction if everyone is passively waiting for their soul mate to make the first move.

    • Nancy Lebovitz says:

      Height is a trait where competition could lead to people going beyond optimum.

    • a reader says:

      What specifically scares me is the possibility that a large portion of the population, given the ability to adjust the personalities of their babies — and I seem to recall that there are survey results to this effect, although I could not possibly cite them — would want children who are more agreeable or obedient.

      What scares me more is the possibility that the state, not the parents, would do this, making it mandatory. Thinking that the world’s first genetically edited babies were born in China, that doesn’t seem impossible in the not so far future (although I don’t know about any genome-wide association studies for agreeableness or obedience).

    • enye-word says:

      Sorry if you were being purposefully coy and I’m outing you, but Kaczynski makes this argument fairly compellingly in Industrial Society and Its Future, so that might be where you got it from.

      • Sniffnoy says:

        No, never bothered to actually read that; but I might have picked it up indirectly from there. Still, thanks!

    • gwern says:

      What specifically scares me is the possibility that a large portion of the population, given the ability to adjust the personalities of their babies — and I seem to recall that there are survey results to this effect, although I could not possibly cite them — would want children who are more agreeable or obedient. That’s… not really a good thing.

      That was asking about toddlers. If you look at what people do, not say, in selecting mates, or sperm/egg donors, or what kind of people they discriminate in favor of, or life outcomes, you get very different results. I wouldn’t worry about that.

      • inhibition-stabilized says:

        Imagine you have an unedited kid, running around doing all the crazy things unedited kids do. You want another kid, but don’t think you can handle another one as crazy as your first. You go over to your neighbors’ house and see their kid sitting peacefully playing with Star Wars XXIV action figures. “My, how well-behaved he is!” “Well,” your neighbor confides, “we had him edited.” You start to imagine how much easier it would be to raise a second kid if he were like your neighbors’…

        I don’t think this would become universal — I think many if not most parents would realize this is a bad idea. But I think a significant fraction of parents would opt for this if it were an option. I should also note that if people actually prefer the opposite, as you suggest, results could be equally bad. Assertiveness (to take one example) might be very beneficial in an environment where other people’s personalities are fairly well distributed, but I doubt society would function very well if everyone were assertive. In general I think relying on our instincts in an environment (such as one in which we have access to powerful gene-editing tools) very different from our environment of evolutionary adaptation can lead to major issues. The distinguishing feature of gene editing is just how major and just how long-term those issues can be.

        All this being said, there are so many potential benefits that I honestly don’t know where I stand on the issue.

        Edit: Part of what I intended to point out with the parable at the beginning (but then forgot to mention) is that gene-editing will likely be subject to fads, which could lead to its own set of issues. We could end up with a world where each age cohort is fairly uniform, with a distinguishing set of features corresponding to whatever was popular those years. I’m not sure what kind of consequences this would have, or whether they would necessarily be bad — just another possibility to think about.

        • spork says:

          I think that if we find the genetic correlates to “desires to contribute to the care of retired parents,” that will give many people a strong incentive to select their kids’ genes.

        • gwern says:

          People already see differences. East Asian infants and kids are famously better behaved. Nevertheless, interracial marriage rates remain relatively modest. Clearly people can’t care that much about the issue.

          As far as this specific issue goes, I would further note that even if we extrapolated from a single survey about toddler hypotheticals and assume that all parents doing IVF henceforth would insist on upweighting Agreeableness, it wouldn’t matter because there are no usable Agreeableness PGSes and the near-zero SNP heritabilities indicate they will be none for the foreseeable future, and this may be because of deep evolutionary reasons ( http://www.unm.edu/~gfmiller/newpapers_sept6/penke%202007%20targetarticle.pdf http://www.larspenke.eu/pdfs/Penke_&_Jokela_in_press_-_Evolutionary_Genetics_of_Personality_Revisited.pdf ) which indicate why you need not worry overmuch about personality fads – because they will trigger their own backlash as parents realize that in a world of highly Agreeable people, being even more Agreeable is a terrible thing and being less Agreeable will help their kid get ahead, and this is part of why Agreeable people have not inherited the earth already. (And similarly with the other OCEAN personality traits: in a world with super-Conscientious people who are unable to stop doing things even when they are useless, someone less of a stick-in-the-mud and more able to cut their losses will outperform etc.)

      • newstorkcity says:

        I’m not so sure that this is comparable. They typical person deciding what they want in a mate is looking for traits that they want in a partner, giving little consideration for what the resulting child will be like (except sometimes in the case of disease). The sperm/egg donor example is much closer, though I don’t know what a typical receiver is selecting for and whether it supports your argument or not.

    • Milo Minderbinder says:

      A generation of accidental(?) Eloi is certainly something to think about. The diversity of schooling types desired by parents gives me hope that temperamental edits, if any, will hopefully be somewhat idiosyncratic.

      This precedent suggests to me that given the wide range of possible gene treatments, most parents will opt for safer, limited edits. A minority of those who opt for riskier or more marginal/aesthetic edits will unfortunately have children with some form of (potentially novel) disability.

      But isn’t the deliberate harm of gestating fetuses intended for delivery already permitted? I specifically mean that (in the U.S.) policy for newborns born to opioid-addicted (or any other drug that causes harm to developing fetuses) mothers is case-by-case, and mothers are frequently allowed to retain custody despite permanent damage done in vitro.

      Parents have extraordinarily expansive rights when it comes to child rearing. Minimizing the risk of gene treatments and ensuring broad access to the most efficacious seems to be the best way to contain the harms that will arise from the availability of such treatments, since it will likely be hard to stop parents from trying in the absence of legal gene editing.

      • notpeerreviewed says:

        There’s a stronger norm against medical providers causing harm than not preventing it. So, for example, a doctor can’t force a pregnant woman to stop drinking alcohol (at least not in most states), but that doesn’t mean a doctor can pursue treatments that risk having side effects as serious as fetal alcohol syndrome.

    • chaosmage says:

      We’ve been selecting more prosocial, more obedient, more agreeable children essentially forever, until we decided every child is a special little snowflake. Not before birth, but in social interactions later, in arranged marriages and definitely in inheritance.

      Was that a bad thing? I doubt it. I think the human niche is teamwork and breeding to be better at teamwork is obviously a winning strategy for us.

    • b_jonas says:

      Today’s Saturday Morning Breakfast Cereal (SMBC) strip raises that exact point: “http://www.smbc-comics.com/comic/selection”.

      However, I don’t see why you could make a moral distinction between making children more obedient and agreeable by transfering your culture and relegion to them, versus making them more obedient and agreeable by gene editing.

  2. JN340 says:

    Doctor here.

    Firstly, thank you for contributing to this collaboration. It’s an enormous amount of work and research.

    A few thoughts.

    1. Every disease is the product of genes and environment. Some diseases, like sickle cell anaemia, Huntington’s disease, and severe combined immunodeficiency, are predominantly genetic, while others, like cardiovascular disease, malaria, or obesity, are predominantly environmental. The first application of gene editing technology – the acid test, if you will – is to see if it works in predominantly genetic diseases for which the causes are well understood.

    2. The question “can gene editing eliminate disease?” is like the question “do quarks cause consciousness?” I mean, maybe? But I don’t know that the authors realised the enormity and complexity of the question they were taking on. To even attempt to answer something like this is hugely speculative and probably a waste of time.

    3. To this end, I think this debate would be better focussed around a few active biomedical or ethical issues surrounding gene editing technology. For example:
    – the ethics around “gene drive,” the possibility of eliminating an invasive species through CRISPR technology, usually focussed around Anopheles mosquitoes that spread malaria.
    – as discussed (briefly) by the authors, the possibility of human germline modification for enhancement
    – the argument for actively using this technology to treat genetic diseases – everyone agrees the potential is there, but scientists are treading lightly around this area – there was universal condemnation of Dr He’s group for prematurely using this technology. But think of the possibilities! Millions of people suffer and die from genetic diseases every year – if we could speed up the universal adaptation of CRISPR to prevent this, isn’t that worth a few casualties along the way??

    (playing devil’s advocate here)

    Anyway, it’s a super exciting field right now. One thing you mentioned briefly are CAR T cells. These could actually cure several varieties of leukaemia and lymphoma. The main issue is they cost $500 000 per application. Siddartha Mukherjee has a great article in the New Yorker written for a lay audience: https://www.newyorker.com/magazine/2019/07/22/the-promise-and-price-of-cellular-therapies

    • sclmlw says:

      There are other major issues with CAR-T besides cost alone. Another big problem is they take a long time to make, so patients can easily die while on ‘bridging therapies’ intended to tide them over until their new T-cells are ready.

      Another problem with CAR-T is that they often don’t work as well in solid tumors, where the tumor microenvironment often excludes newly infiltrating T-cells. It’s nice to get the cells inside the body, but getting them inside the tumor is another story.

      I think on the question of ‘eliminating disease’, people are always going to die of something. The question is whether they suffer and/or die of the limited subset of human disease that can be directly attributed to single mutations. That’s where the technology is headed in the near future, and I don’t think most people object to that kind of treatment. Whether it’s even possible to use it against complex multigenic diseases we’re still struggling to understand is not clear.

    • Dacyn says:

      The question “can gene editing eliminate disease?” is like the question “do quarks cause consciousness?” […] I don’t know that the authors realised the enormity and complexity of the question they were taking on.

      Funny, my reaction was “one of those questions is a lot easier than the other, and it’s not the gene editing one”. I think in many ways controversial philosophical questions are not very analogous to difficult empirical ones…

  3. Nancy Lebovitz says:

    It wouldn’t surprise me if a lot of the people who have genes for a disease but not the disease have other genes which compensate for the potential problem, so it wouldn’t just be environment or epigenetics. Those compensating genes might be problems themselves if the disease gene is absent.

    • paddl113 says:

      That is a very creative way of thinking and you are probably right in some cases. However, I would expect that some of those people would have been identified by now with all that frequent DNA testing nowadays.

  4. BlindKungFuMaster says:

    Maybe the original title of this collaboration was different, but I think discussing the applicablity of genome editing against cancer does little to answer the titular question. The current state of the technology is also not particularly relevant. If we can’t get the effect we want via gene editing (yet), obviously we shouldn’t edit genes (yet).

    My answer to the question is “of course we should”. At least as soon as the risk-benefit-tradeoff is reasonable. Why shouldn’t future generations be healthier, happier, longer-lived, nicer, smarter and more attractive?

    • albatross11 says:

      It doesn’t matter what ethical arguments are made against designer babies. When it is possible to make them, it will be possible for hundreds of thousands of different people/labs to do. At that point, there will be people offering designer babies, and the most ambitious-for-their-kids people will provide a ready market. Alice and Bob just happen to take a vacation in Brazil where they get pregnant for each of their kids, and their kids just happen to be extremely athletic and the smartest kids in their class and unusually tall and very healthy.

      In the modern US, parents spend great piles of money on private schools, test prep, resume-building activities, college admissions consultants, etc. The really competitive ones line up overnight and pull strings to get their kids into the most exclusive preschool. An illegal-but-available designer baby industry would find a huge pool of customers here.

      And once that happens, assuming the designer baby thing works, you get the children of the most ambitious people with money in the US having the designer kids who turn out to be, on average, smarter, taller, better-looking, harder-working, and more ambitious than everyone else. Twenty years after it becomes available, half the entering freshman class at MIT has the full package of IQ, attention span, and work ethic enhancements.

      • Randy M says:

        And once that happens, assuming the designer baby thing works, you get the children of the most ambitious people with money in the US having the designer kids who turn out to be, on average, smarter, taller, better-looking, harder-working, and more ambitious than everyone else. Twenty years after it becomes available, half the entering freshman class at MIT has the full package of IQ, attention span, and work ethic enhancements.

        I don’t think it’s possible to be egalitarian about this either. Either it will go to people with money, or people with connections. Top of the line stuff will always be more expense or rare than what is available to everyone, due to a bottleneck in the expertise for it.

        • notpeerreviewed says:

          I don’t think it’s possible to be egalitarian about this either. Either it will go to people with money, or people with connections. Top of the line stuff will always be more expense or rare than what is available to everyone, due to a bottleneck in the expertise for it.

          Counterweight: It’s probably easier in a technical sense to raise below-average traits to the average than it is to enhance average traits to superhuman levels, and CRISPR has made genetic engineering “cheap” in a certain sense (though I’m sure the total cost of a treatment intended for a human would still be high for a variety of reasons.)

      • Cliff says:

        I wonder at what point it becomes no different from using donor egg and sperm, or adopting. If you’re going to edit your kid’s athleticism, height, appearance, personality, intelligence… what’s left?

        • Milo Minderbinder says:

          Memes. You’re still raising the kid. It’s distinct from a donor/adopting in that you’re (theoretically) even more involved in shaping the child.

      • The Red Foliot says:

        Egalitarianism would be fueled by the fact that people want their fellow countrymen to be as productive and prosocial as possible. The government already spends unbelievable amounts of money on attempting to make people more productive through schooling. If a more efficacious means of doing this comes along why shouldn’t the government also want to spend hundreds of thousands of dollars per child effecting it? Also, rich people already use their status and wealth to obtain better genetics for their children — mate selection.

        The only anti-egalitarian force I can think of is that of the bioethicists who tend to make vain and apathetic statements about the technology in an attempt to arbitrarily stigmatize and ultimately outlaw it. If the technology becomes stigmatized and outlawed then it probably would be the case that only rich people with millions of dollars able to fly to Brazil would be able to afford it.

        • spork says:

          I think this is a good point. And unlike investments in schooling, which must be paid anew for every generation, investments in germline alterations carry over to the next generation for free. Thus a family that benefits from genetic enhancements benefits not just once, but in perpetuity. This means that in principle, it could really be a force for increasing egalitarianism, and it could make an attractive target for subsidies.

        • notpeerreviewed says:

          The only anti-egalitarian force I can think of is that of the bioethicists who tend to make vain and apathetic statements about the technology in an attempt to arbitrarily stigmatize and ultimately outlaw it. If the technology becomes stigmatized and outlawed then it probably would be the case that only rich people with millions of dollars able to fly to Brazil would be able to afford it.

          Bear in mind that in many developed countries, the government pays for most health care. So the point of contention might not be whether someone has to fly to Brazil to get gene therapy, but rather, whether gene therapy is considered to be part of the normal standard of care that governments will pay for.

      • Nancy Lebovitz says:

        Let’s go with a boring hypothesis. Instead of a nicely sculpted foot grinding down on a slightly misshapen face forever, let’s assume there’s a normal course of technological development.

        Designer babies become possible, but they’re very expensive. For a while. There’s huge pressure to make the tech cheaper.

        At a guess, the tech will become noticeably cheaper by the decade.

        Bioengineering of the people who already exist will also be improving.

        There will be charities and government programs to make the tech more available.

        I’m not expecting simple tyranny or egalitarianism.

      • Konstantin says:

        Is it really that easy from a technical standpoint? If you need millions of dollars in equipment and a team of skilled professionals, it’s hard for a black market to arise. You also have trust problems, even if it is easy to edit genes it is easier to do nothing and say you have edited the genes, and I can’t see a practical verification mechanism that doesn’t require trust.

        • phi says:

          Genome sequencing is getting pretty cheap, so I imagine that would be how it is verified. Of course, even if you can verify that the asked for changes were made, it would be difficult to know if those changes are actually going to lead to increased intelligence (, attractiveness, etc.) without trying it.

          • sclmlw says:

            You don’t need to sequence the whole genome if all you’re looking at is one or a defined set of genes. That makes it cheaper still.

        • notpeerreviewed says:

          Editing a single SNP is shockingly easy; it would be completely feasible for an ordinary fertility clinic to do so (maybe not literally right now, but it will be soon.) What’s unclear to me at this point is whether the same technology can ultimately be used to edit a large number of genes in a single embryo, because any sort of “enhancement” beyond curing a few genetic diseases is going to require changing more than just one or two genes.

          • sclmlw says:

            This is a great point. I’ve not personally worked with CRISPR before, but my sense is that – similar to other molecular biology techniques – there’s a huge difference between single-gene editing and wholesale changes to more than one gene. To do more than one, I’d think you’d need to make the change, implant the embryo, grow the human, and then introduce a second change in the next generation.

          • notpeerreviewed says:

            Me: We totally can’t do Brave New World yet because CRISPR only does one gene at a time.

            Science: Hold my beer.

            https://www.futurity.org/crispr-multiple-genes-2132952/

          • anonymousskimmer says:

            @sclmlw

            Somehow mIx embryonic stems cells with some placental stem cells and you could theoretically build an engineered human from serially-modiified ESCs.

  5. block_of_nihilism says:

    “Nevertheless, despite recent medical advances, cancer treatment has not seen significant improvement in decades.”

    This is simply incorrect. Decades? In breast cancer alone, we’ve seen the introduction of trastuzumab and other HER2-targeting therapies, letrozole, palbociclib, and (insert other new drugs I’m forgetting). Combine this with improved treatment protocols, discovery of prognostic gene panels like Oncotype Dx, and immunotherapy in TNBC and we’ve made dramatic progress in the last 2 decades alone… And the progress in other cancers, particularly hematologic, is even more impressive (imantinib, dasatinib, venetoclax, ibrutinib…)

    While this isn’t an essential part of your essay, I felt it was necessary to bring this up!

    • BlindKungFuMaster says:

      Yeah, and the link goes to a report that excitedly describes recent advances. I guess somewhere there is going to be a statistic that the same percentage of people is dying of cancer as 30 years ago, which ignores the fact that people have to die of something and simultaneous advances in different disease areas will tend to keep the death stats balanced.

    • albatross11 says:

      Yeah, I found that jarring, too. I think (I’m not an expert at all) there are some kinds of cancer for which they don’t have anything better than old-style chemo, or radiation treatment, or surgery. But there are other kinds of cancer for which they have extremely effective targeted treatments.

      My impression (mainly from reading _The Emperor of All Maladies_) is that the reason cancer progress is so slow is that most of the targeted treatments work only on a really narrow set of cancers, so instead of scientists producing a generic cure for cancer, they (with great effort and expense) produce a cure or near-cure (long term remission) for one tiny subset of cancer cases at a time.

    • sclmlw says:

      I agree. That statement was factually inaccurate. I also thought the discussion of microenvironment and epigenetics was misleading. They make it sound like cancer is driven by a bunch of hypomethylation, but much of those epigenetic changes are driven by the epithelial to mesenchymal transition characteristic of many cancers. Characterizing EMT as not inherently a gene-driven process is probably wrong. Yes, cell differentiation is driven by environmental factors during gastrulation, but post-natal changes in cellular identity like this are most likely driven by genetic insults. In other words, a cell pre-programmed to be a differentiated skin cell that reverses its programming is likely doing so because of genetic, not epigenetic, changes. It would be bizarre if somehow a bunch of random epigenetic ‘mutations’ all happened to create an EMT-like state in cancer cells.

      Unless I’m missing something here, the theory that cancer is driven by genetic changes (often introduced by environmental factors, yes) is still sound in a world where lots of hypomethylation is present in the cell. Because those epigenetic changes are driven by pre-existing genetic ‘programming’ that was already present and got activated inappropriately and lead to cancer.

      Also the microenvironmental discussion ignores the subtle interplay between the cancer cells themselves and the tumor infiltrating immune cells. Whether it’s driven by chemokine signalling, costimulatory molecules, or messing around with tumor vasculature, the cancer cells themselves are the ones playing the immune system to make it change to a pro-tumor response. This is a complex topic, so it’s understandable they didn’t go into it a lot, but the tone of the discussion was ‘genetics doesn’t matter because of microenvironment’, when it’s most likely genetic changes that are creating that microenvironment.

      • thompson says:

        New commentator.

        Minor quibble with your point (minor because you are very correct, it is a complex topic and no short comment is going to completely cover it). But I think you go too far towards the genetic direction and away from microenvironment.

        Characterizing EMT as not inherently a gene-driven process is probably wrong. Yes, cell differentiation is driven by environmental factors during gastrulation, but post-natal changes in cellular identity like this are most likely driven by genetic insults.

        This is what I think is a little misleading. EMT can be triggered by genetic changes, but also very much by the environment, even in a genetically normal cell. As a key example, EMT is an essential part of wound healing: epithelial cells will take on a more mesenchymal phenotype to close the wound. This is driven by cytokines and other local signals.

        This is relevant for tumorigenesis as damage, over-activation of cytokines in inflammatory diseases, etc, can likely help tilt the balance towards malignancy by pushing more towards a mesenchymal phenotype. There needs to be other genetic problems as well, I just wanted to push back a little against this line:

        when it’s most likely genetic changes that are creating that microenvironment.

        I think it is a little more bidirectional than that. Tumor drives the environment, but environment can drive the tumor as well. It is, as you say, a complex topic.

        • sclmlw says:

          Right, thanks for the correction. My point was to push back on what I think the ACC submission effectively claimed: ‘genetic mutations are only one way cancer arises, while epigenetic mutations are another possible route’. I know that’s putting words in their mouths, but I think it’s justified otherwise why include it in the discussion?

          For something like the development of cancer, repeated damage could tilt the balance, as you suggest, and there’s some great evidence that this is exactly what happens in certain types of cancers (for example alcohol+smoking enhances esophageal cancer). But I don’t know anyone who would suggest that repeat activation of a normal physiological process of wound healing would lead to tumorigenesis through epigenetic factors alone. At least, not that we’ve observed or have evidence to support. Thus, we would expect that there are genetic abnormalities that are required to maintain the tumor phenotype, and that fixing those genetic abnormalities would likely stop tumor growth – regardless of the epigenetic changes present in tumor cells.

          That’s what I was pushing back against. If we imagine the idealized scenario where we can ‘fix’ the mutations that caused any individual cancer, that would be tantamount to a ‘cure’ provided the tailored therapies are logistically feasible. There’s no reason to assume a large percent of cancers are not susceptible to this imagined therapy because their primary etiology is epigenetically-based, not genetically-based.

          • thompson says:

            But I don’t know anyone who would suggest that repeat activation of a normal physiological process of wound healing would lead to tumorigenesis through epigenetic factors alone.

            Yeah, and I should have tried to be more clear, I didn’t think you were incorrect, but that someone could get the wrong impression. I guess a minor crusade of mine is getting more people to recognize how much of cancer biology is just hijacking physiological processes for pathological ends.

            If we imagine the idealized scenario where we can ‘fix’ the mutations that caused any individual cancer, that would be tantamount to a ‘cure’ provided the tailored therapies are logistically feasible.

            Completely agree. Also agree with the implied logistical difficulties. Delivery is, IMO, the biggest challenge, and there is no Amazon Prime for CRISPR.

          • sclmlw says:

            Yeah, if anything the logistical challenges of CAR-T demonstrate how difficult individualized therapies can be. It’s conflicting logic people don’t recognize they’re doing when they imagine mass-produced personalized medicine. If it’s personalized, how is it mass-produced? I have some ideas for how that’ll happen in the future with oncology, but it’s not a simple matter of “we have CRISPR, now the world is our oyster!”

  6. Tulip says:

    I feel like this entry, moreso than the others thus far, has suffered from a degree of blatant “one author writes one paragraph, the next author writes the next paragraph, et cetera”.

    Like, maybe I’m wrong, but I have a very strong impression here that we’ve got The Author Who Likes Citing Individual Incidents And Quoting Individual Professors And Is Against Gene-Editing and we’ve got The Author Who Likes Citing A Few Studies Per Paragraph And Is In Favor Of Gene-Editing, and that the two never really bothered to come up with a unified message for their writeup beyond “here’s my pile of points, here’s your pile of points, let’s dump both piles onto the page and call it a day”.

    • inhibition-stabilized says:

      The citation patterns are a bit strange, now that you mention it…
      Regardless, I think you’re right that this collaboration suffers from a lack of a unified message. I don’t want to detract from the authors’ work — they clearly put a lot of research and thought into the essay. I think the main problem is that the question is too broad, and questions like “Is gene editing safe and efficacious?” are pretty distinct from questions like “Should we allow designer babies?” As JN340 said above, I think the article would have been better off focusing on a more narrow ethical question.

  7. DinoNerd says:

    Personally I expect:
    a) This will happen
    b) With plenty of unintended consequences

    If I were dictator, I’d insist on moving very slowly. Human generations are long; human lifespans are longer. Bad side effects could take a long time to become evident, and it would be unfortunate if the affected population was huge by the time they did. And for heavens sake, start with fixing serious (= lethal) problems that are well understood, and essentially entirely genetic, before trying to fix more complex issues – and all this before trying to “improve” perfectly functional people.

    Also in the “if I were dictator” category – incentivize people to provide a large control group. One novel I recall had “control naturals”, complete with an allowance large enough to live on. Enhanced children – and their descendants – didn’t get it – and the value rose over time, as the opportunity cost of non-enhancement increased. (Even though, in that novel, there weren’t any notable unintended consequences – the main value of “control naturals” was as a source of novel mutations, some potentially advantageous.)

    Even that’s not enough. People will generally want to make their children successful in their current environment, and will do so if they can. There’s always a tension between adaptability and over-specialization. Designer babies are likely to be significantly (over)specialized – bad if the environment changes; worse if a new design can’t be produced quickly.

    There’s also plenty of research on the value of diversity/disadvantages of group think. But if e.g. tall white extroverted males are most succesful in your culture, that type will be overproduced, and stereotypes against/exclusion of other types will intensify.

    And all this is before we start gene-editting our way to “peacock tails” – disadvantageous but highly visible features that are encouraged/mandated by fashion, and primarily show that our parents could afford to “optimize” their children.

    I’d bet that if this becoems technically feasible, we’ll have a peacock tail of some kind withinn the next 200 years, and 90% of the population will honestly believe those with the ‘tail’ are better people. (Not a literal peacock tail, of course; but e.g. ridiculous slimness, or height, or some such – yes, of course all babies are born by C-section, because of what this requires of female hips, but slimness is so much healthier …)

    • BlindKungFuMaster says:

      I think if the technology itself is safe, reducing mutational load would be pretty safe. If you change a variant that occurs in 0.1% of the population into a variant that occurs in 90% you are not going to break anything. So only allowing changes that reduce mutational load might be a pretty safe way to modify the genome. It would rule out fancy stuff like muscles without training or superstrong bones but I think there would still be a lot of possible optimisation.

      • notpeerreviewed says:

        I think if the technology itself is safe, reducing mutational load would be pretty safe. If you change a variant that occurs in 0.1% of the population into a variant that occurs in 90% you are not going to break anything. So only allowing changes that reduce mutational load might be a pretty safe way to modify the genome. It would rule out fancy stuff like muscles without training or superstrong bones but I think there would still be a lot of possible optimisation.

        I talked to a biologist about this over Thanksgiving, coincidentally enough. The problem is if you want to make meaningful improvements to someone’s genome by reducing their mutational load, you have to modify a *lot* of genes, and current CRISPR technology is up to that task. This seems like it should be one of the most interesting technological issues to keep an eye on, because yes, in theory it should be possible to make not-quite-superhumans-but-really-impressive-humans by simply going through and replacing every rare variant with a common one.

      • edmundgennings says:

        Indeed, this is the safest and most effective way to go.
        Incidentally, this probably means that in the medium run, genetic editing will be egalitarian in outcomes. At first, the rich will lessen their mutational load, but once prices come down the poor will as well and the genetic gap will not merely return to pre genetic editing time, but almost disappear.

      • DinoNerd says:

        Actually, what you break is the presence of that new mutation that’s in fact advantageous, but since we don’t know how, we get rid of it. Or that currently neutral mutation that’s going to save a fraction of the species when some future environmental challenge occurs. (Maybe those people will have enhanced resistance to the next major pandemic.)

        Most mutations are disadvantageous, but most != all. The ones that are disadvantageous tend to get rid of themselves, but with lots of suffering as it happens. I very little concern with fixing those – not none, because I can imagine somethign with unpleasant side effects that’s actually a net benefit. But going after all mutations is IMO excessive.

    • Nancy Lebovitz says:

      The book is Heinlein’s Beyond This Horizon. The cost to Control Natures is that they can’t win at much of anything, though there one person in the small group that’s running things who’s a Control Natural.

      If I were dictator, I’d put off all but the most cautious gene editing until bioengineering advanced to the point where errors could be undone. Then let the weirdness begin!

    • Nancy Lebovitz says:

      While we’re worrying, I suspect hypomania will be one of the desired traits, and that would be easy to get wrong.

    • spork says:

      But if e.g. tall white extroverted males are most succesful in your culture, that type will be overproduced, and stereotypes against/exclusion of other types will intensify.

      Generally, the more common something is, the less valuable it seems. In Japan, cute girls are now asking orthodontists to make their teeth more crooked. If you grew up with the standard genetic kit that makes you a dutiful, diligent, verbally and mathematically inclined workaholic, you might very well decide to scrap all that and make kids who are designed to turn out cool.

    • Aapje says:

      @DinoNerd

      Also in the “if I were dictator” category – incentivize people to provide a large control group.

      I see a lot of control groups being dissolved, after the treated group seemingly do a lot better, suggesting that trade-off between ‘withholding treatment to be surer that the treatment works’ vs ‘treating people with poor evidence’ is shifting to the latter.

      Note that there have been many cases where initial research showed significant impact of treatment (in effect size or for the scientific definition of significant), that more research showed to be false. So a low willingness to withhold treatment can easily cause bad/mediocre treatments to falsely be seen as (very) effective.

      There’s also plenty of research on the value of diversity/disadvantages of group think.

      Which, in my experience, progressives tend to present as being unequivocally in favor of diversity*, even though the actual results seem largely a wash. Sometimes performance improves due to diversity, sometimes it deteriorates. In most cases, it seems to make little difference, which makes sense, since most diversity that is studied logically has little impact on the outcome that is looked at.

      IQ is the one difference that does tend to have a major effect on performance, one that is rarely included in definitions of greater diversity, yet also one that is presumably one of the main things that people would want to gene edit (given the huge impact).

      *Of course, there are different definitions of diversity. Some people seem to desire diversity of skin color, gender, etc, but within a monoculture with very restrictive allowed opinions. Is that diversity or the lack of it?

  8. sclmlw says:

    On HIV and CCR5: I can understand the impetus for re-creating a CCR5 mutation that’s already seen in the normal human population for one-off cases where a mother has HIV, in an attempt to prevent transmission to the fetus. (Although we have other ways of producing offspring that are genetically similar, like with the use of surrogates.) I think it’s a bad idea to go around making this change to ALL humans. CCR5 is an immune system molecule that allows immune cells to follow signals sent out in response to potential threats. It’s one thing for some humans to have an error that creates genetic susceptibility to certain diseases. It’s quite another to make that susceptibility more broad. This is like introducing a secret back door in your security protocols and hoping bad actors don’t find it and exploit it. This was mentioned briefly in the essay, but I think I’m not sure the nuance is clear to an audience that doesn’t have a strong biology/immunology background.

  9. sclmlw says:

    There’s a very dark utilitarian perspective of medicine that sees treating disease as a way of perpetuating maladaptive genes within the population. The better we get at keeping alive people who suffer from genetically-driven diseases (which are only a subset of all disease, of course) to the point where they’re able to survive to pass on their genes to the next generation, the more we enrich for those genes and perpetuate that disease. In other words, for certain diseases medical treatment may be slowly altering the genome in negative ways. Does this mean germline removal of certain genetic abnormalities might reverse problems we accidentally created?

  10. metacelsus says:

    Biochemistry PhD student here, with experience using gene editing to generate cell lines.

    Some pairs of genes are “linked” due to physical proximity on the same chromosome and are therefore almost always passed on together. Any edits to a gene belonging to a linked pair may therefore inadvertently cause an edit in its neighboring partner.

    This is incorrect, it’s not what linkage means. CRISPR recognizes DNA sequences, not physical proximity.

    If differential gene expression can produce such drastic changes, is genome rewriting really necessary? Perhaps the centrality of the gene in driving human health has been overstated. Indeed, why worry about a potentially pathogenic gene if it is never expressed?

    Maybe, but what about loss-of-function mutations? And how are you going to specifically target a pathogenic gene for suppression, without using gene editing? RNAi could work but delivery is a challenge.

    And about epigenetics: it’s not the answer to everything. Often epigenetic disturbances are a result of genetic mutations (for example, chromosomal translocation causing activation of MMSET histone methyltransferase in multiple myeloma).

    Germline editing may therefore provide a more expedient option for the prevention of some genetic diseases such as sickle cell disease or cystic fibrosis.

    Pre-implantation diagnosis would be effective for these, and much safer.

    Research into CRISPR is continually finding ways to make the technology safer and more effective; a paper published in September 2019 reported on the potential for a novel CRISPR system to affect gene expression in human cells.

    I would also mention prime editing (published October 2019 here: https://www.nature.com/articles/s41586-019-1711-4)

  11. Freddie deBoer says:

    “Of course, this newfound power raises several ethical concerns. The major worry among scientists revolves around the long-term consequences of germline modification, meaning genetic changes made in a human egg, sperm, or embryo. Edits made in the germline will affect every cell in an organism and will also be passed on to any offspring. If a mistake is made in the process and a new disease inadvertently introduced, these changes will persist for generations to come. Human germline modification could also theoretically allow for the installation of genes to confer protection against infections, Alzheimer’s, and even aging. For many, the thought of controlling our own genetic destinies seems to be a very slippery slope, conjuring up dystopian images of Frankenstein or Brave New World.”

    Please forgive my insufferable urge to play copy editor. I’m confused by the presence of sentence number 5 in this sequence, as it speaks to the good things CRISPR could do in the context of a paragraph outlining the potential bad parts of CRISPR, and then reverts in the next sentence to the bad things. Just jarring to me.

  12. An Fírinne says:

    Opposition to gene editing is really just a general phobia of the unfamiliar. Eventually people will calm down and accept it like they do with things like rock and roll and divorce.

  13. Etoile says:

    One ethical dimension: at what point are we “playing God”? Is there a line when you’ve crossed too far, flown too close to the sun? I suppose this is a fundamentally question of your faith…. but viscerally, I at least look at something and feel like it’s too much. Wen you go to extremes – e.g. should we strive a 75-year-old woman pregnant trough IVF, even if we could do it physically? If you have a nonviable fetus that would have been rejected naturally, should you implant it in an artificial womb and develop it to maturity? Should we try to pursue immortality? Eternal youth? When does this sort of meddling become hubris?

    I think that’s why issues around fertility, childbirth, and abortion are so fraught: because they touch on something very primal, almost sacred. It’s hard to talk dispassionately about it — and that applies to “germline” human editing.

    • sclmlw says:

      Hypothetical question aimed at revealing preferences:

      Say you’ve identified a lentivirus present in 10% of the population, which inserts itself into germline cells only to manifest later in development. Further research reveals this virus makes infected people more susceptible to concomitant infections. Epidemiologically, they’re ten times more likely to die as a result of seasonal illnesses alone. You did some searching and have found this same virus has been endemic in human populations for over five hundred years, we just didn’t know until now that many people we thought just had weak immune systems were actually infected by this virus.

      Since this is a hypothetical, we can offer hope – you found a cure!

      1. Do you distribute your cure of this lentivirus to prevent future generations from suffering its ill effects?

      2. What if the lentivirus isn’t contagious? You discover that this shift in the lentivirus happened two hundred years ago, and now its only way to perpetuate through the population is through reproduction. Do you still distribute your cure?

      2a. How would this be different from gene editing, where you’re changing the representation of specific (presumably maladaptive) sequences in future generations?

      3. What’s the difference between a cure, a therapy with persistent impacts, and ‘playing god’? Is it just because we don’t want to mess with DNA? If so, why aren’t viruses off limits?

    • blacktrance says:

      As our capabilities expand, it’s proper for our estimation of them to expand accordingly as well. For example, performing heart surgery would be hubris if you’ve only read one Wikipedia article on the subject, but not if you’ve gone to med school and been properly trained. It’s bad to have an exaggerated evaluation of your abilities, but a goal can’t be hubristic in itself – and immortality and eternal youth are particularly awesome goals.

      • notpeerreviewed says:

        As our capabilities expand, it’s proper for our estimation of them to expand accordingly as well. For example, performing heart surgery would be hubris if you’ve only read one Wikipedia article on the subject, but not if you’ve gone to med school and been properly trained. It’s bad to have an exaggerated evaluation of your abilities, but a goal can’t be hubristic in itself – and immortality and eternal youth are particularly awesome goals.

        My main concern is that I hope our reach and our grasp advance at roughly the same rate. For example, it seems like we currently have all the knowledge we’d need to cure genetic diseases that are caused by single SNPs, without too much risk of anything going wrong. But if someone tried to take the knowledge we have today and use it to genetically engineer children with higher IQs, there would probably be a lot of unintended consequences – we simply don’t understand enough about those genes to do it safely at this point.

    • albatross11 says:

      I wonder how much this is just status-quo bias, though. I mean, half our kids used to die before they turned five, now a kid dying that young is a rare terrible tragedy. Compound fractures and stab wounds to the gut were almost uniformly fatal–now, they’re usually survivable. We have eliminated one of the most terrible plagues that used to kill us off, and are working on others, and we’ve found a way to prevent nearly all our kids from getting the childhood diseases that previously carried a certain number of them off every year. We’ve found ways to open closed-up arteries and replace failed joints and fix rotted-out teeth.

      All those might look like playing God to someone from a couple hundred years ago, but I sure don’t plan to turn them down if/when I need them. If in ten years it becomes possible to remove most known genetic diseases, and then after that basically no children in the US are born with Tay Sachs or Huntingtons or Sickle Cell, is that playing God, or just continuing the stuff we were previously doing?

      • sclmlw says:

        Yeah, the term ‘playing God’ has been used for a long time in relation to new medical advances. When the first effective anesthetics allowed surgery to become much more commonplace people talked about the the human body as sacrosanct, and complained surgeons were ‘playing God’.

        What would it really mean to ‘play God’ in the way most people imagine the term to mean? I’d think the de novo creation of life from non-living materials would fit. Tinkering around with something that used to be considered just a structural molecule a few decades ago seems like it should require more justification to qualify as ‘playing God’.

  14. Loris says:

    Some pairs of genes are “linked” due to physical proximity on the same chromosome and are therefore almost always passed on together. Any edits to a gene belonging to a linked pair may therefore inadvertently cause an edit in its neighboring partner.

    This is nonsense. Either the message has been garbled or someone got the wrong end of the stick.
    There are several concepts I think might have been intended, and this statement possibly conflates them.

    1) Some distinct genes (which are not necessarily neighbours) have similar sequences, and an edit to one might accidentally also change the other.
    2) If a gene linked to the targeted one is also ‘broken’, fixing the targetted defect may allow it to persist in the population.

    I’d tend to think anyone claiming (2) was protesting too much, but there’s a related concept they might mean:

    3) If a pair of genes are closely linked because they physically interact with each other, and matching sets have a fitness advantage, then changing a non-standard version of one of the genes to the more common version may stop it from interacting with its partner, revealing an unexpected deleterious phenotype.

    If you imagine your genetic code as the hardware of a computer, epigenetics is the software that runs on top and controls the operation of the hardware.

    Highlighting the importance of the microenvironment in regulating development, homeostasis, and cancer, biologist Mina Bissell writes, “The sequence of our genes are like the keys on the piano; it is the context that makes the music.”

    I think both these analogies are really bad, and will contribute more to confusion than understanding.

    • sclmlw says:

      I agree. I think this team struggled to explain a complex subject with analogies and short definitions (without which it would become unreadable to all but subject matter experts). At times I was impressed by their ability to succinctly explain the relevant aspects of an idea and move on.

      Other times I cringed at analogies that brought more confusion than illumination to the idea they were trying to get across. In these cases I was left wondering how much they really understand the subject, or whether they’re relying on others to frame the conceptual landscape for them. To be fair to them, a lot of the discussion crosses disciplines of virology, molecular biology, immunology, and oncology. Nobody can know everything, but some of these explanations were a little too tailored for easy public consumption which sacrificed accuracy.

    • Loris says:

      Correction to the above:
      3) If a pair of genes are closely linked because their products physically interact with each other, and matching sets have a fitness advantage, then changing a non-standard version of one of the genes to the more common version may stop a protein from interacting with its partner, revealing an unexpected deleterious phenotype.
      While I’m sure most people would accept the occasional error in a forum post, it still really annoys me when I notice one of my own.

      Yes, you’re right sclmlw, it’s perhaps unfair to dwell on the places the authors maybe trimmed away too much context without acknowledging the other parts where their approach succeeded in being clear, concise and readable in this rather arcane subject.

  15. Peter Gerdes says:

    I know it’s a hard topic and I may not have done any better but I don’t think this really got at the important enhancement issues which dominate costs and benefits.

    1. Complexity of effects of genes on people is already something current methods can handle. Steve Hsu’s company is already advising expecting parents with fertalized IVF eggs about probabilities for various outcomes. Currently the selection doesn’t move the probability estimates much but it will improve and even a 2% reduction in risk for depression would be a huge benefit.
    2. Fact that phenotype is heavily influenced by both genetics and enviornment does nothing to show that human enhancements wouldn’t have a massive effect. A farmer who finds out that cow milk production is affected both by enviornment and genes doesn’t throw up their hands and give up on selective breeding. They realize both matter.
    3. Almost certainly one of the largest factors in cost/benefit analysis will be the extent to which starting now will speed up the development of more powerful and accurate techniques.

    In short no you can’t just wave away enhancement because it’s kinda complicated and some people find it disturbing.

  16. Peter Gerdes says:

    I guess I feel this piece (while having lots of great info) used the objectionable technique of saying “it’s really complex” to brush off consideration of a morally fraught topic (enhancement) even though it almost surely ends up dominating cost benefit analysis.

    Even with dealing with just disease there is a sort of on the one hand and on the other hand thing going on which isn’t very useful if you don’t give any quantitative estimates.

    I mean if you are really sure that in the long run we can find some Gene edits that will ensure hypomania (high happiness little downside) or eliminate depression then it might make sense to just go nuts now and experiment on everyone you can convince to try. It would be hard to imagine that producing 100 experimental babies a year for 100 years could cost as much in utility as moving the time where you can implement these happiness upgrades on a wide scale forward by a single year.

  17. Jakub Łopuszański says:

    OK, so DNA has a lot of #ifdefs in it and we’ve learned how to apply patches yet we do not know what the master branch looks like, have no safety net of automated regression tests and haven’t figured how CMake maps environment to #defines.
    So… who’s working on the CMake part?

  18. lubelskie says:

    I think it’s important to keep in mind in the discussion of what is disease treatment and what is enhancement that it is often the state that decides who is “diseased” and who is a “public safety risk”

  19. Bram Cohen says:

    This post is very much about the exact current state of the the technology which is very limited in what it can do and has high probability of unintended side effects. Presumably this will improve massively in the future until we can precisely control the entire genome with high reliability, at which point the ethics of what can, should and will be done completely changes, which this essay mentions but disappointingly doesn’t explore details.

  20. TricksterPrinter says:

    -1 You have not addressed the evolutionary argument against gene editing.

    When the first truly gene-edited baby pops out, an evolutionary feedback loop has been created. The people that gene edit will have more fit babies and those in turn will be more likely to gene edit. This means that humanity will converge towards a single genome in evolutionary time. This argument gets laid out in the final chapter of the book “The Revolutionary Phenotype” by Dr. Gariepy in which he also describes the emergence of genetic layers and the evolution of sex.

    • Nicholas says:

      Sorry, what is the argument? If a single optimal genome exists, shouldn’t every person be able to have it? More than likely though, there will be several local maximum genomes that optimize for different traits. Also, when gene editing is this widely used, people might very well be able make their own fashion choices with their genes.:Changing hair and eye and skin colors as often as people dye their hair or paint their nails today.

    • sidereal says:

      Why would they converge? Seems to me they’d diverge (to the point of speciation and more) due to differing values and beliefs, and the fact that the modern economy has many more actual niches than the ancestral environment.

  21. hnau says:

    This collaboration tackles an important topic, which makes it disappointing to see how thoroughly it fails to reach any meaningful conclusion. Like the collaboration on circumcision, it adopts a bland, “committee report” style that’s nicely polished but gets in the way of actually clarifying the issue. The main difference is that the collaboration on circumcision at least reached some vague model of benefits and harms; this one doesn’t, as far as I can tell.

    It look like the authors did try to evaluate the potential upsides and downsides of gene editing, but came up short because of a shortage of definite scientific evidence on an immature technology. Here’s what I think they get wrong, though– in a scenario like this a shortage of evidence shouldn’t mean “eh, seems OK, go ahead” which is more or less what the authors say in their conclusion. That endorsement of the status quo will just result in adoption snowballing, possibly out of control. With a technology so potentially life-changing, the burden of proof should be on those who claim it will be used beneficially and responsibly if adopted.

    Overall, I got a few interesting pieces of information out of this write-up, but not what I’d have hoped to get out of an adversarial collaboration. I rate this collaboration as a 3/10 (no particular scale or judgment implied, this is just for my own reference). As always, many thanks to the authors for putting in the work to create this.

    • Aapje says:

      With a technology so potentially life-changing, the burden of proof should be on those who claim it will be used beneficially and responsibly if adopted.

      Is this a reasonable standard?

      Has democracy been used beneficially and responsibly in all cases? Could anyone have predicted how democracy would work out on the national scale without adopting it?

      • hnau says:

        Fair question. I recognize that there might be some cases where a standard like this prevents the move from a clearly worse equilibrium to a better one. But I don’t think democracy is a good example of such a case. I’d want an example of a technology and specifically one with as much obvious potential to change how society works as gene editing’s.

        I have to admit that what I had in mind was a more pragmatic, less principled objection. I worry that advocates of gene editing will keep saying “more research is required / we’re just doing science” until it’s on the cusp of wide adoption, at which point they’ll shift to “it’s cruel / regressive to take this away” and thus freeze out any actual chance for society to decide it doesn’t want it.

  22. SCPantera says:

    I wonder if future risk of designer babies could be offset by requiring storage of before-alteration clones or genetic data so in a really-bad-case scenario we could resume from where we started, or start from scratch with more experience.