TED Talk on CRISPR/Cas9 system of genetic engineering

Jennifer Doudna talks on the currently mainstream method of genetic engineering, using the site-specific CRISPR/Cas system.

In addition to the brief summary of how CRISPR works, she also talks about how genetic engineering is currently being used and predicts the first applications of gene therapy will be mostly for immune system diseases, as white blood cells can be removed from the body and modified ex vivo, or outside the body. I totally agree with this, with our current level of technology it’s far easier to engineer a cell outside the body rather than risk any of the adverse reactions to gene therapy in humans.

Most importantly for this blog, she talks about whether this technology could be used for genetic enhancement. She lists simple things that many of us might even consider no different to vaccines, like enhancing our resistance to cardiovascular disease, before quickly moving into the ‘designer humans’ idea of specifying or changing height or eye colour.

She backs up the moratorium on human germline genetic engineering that I have mentioned on this blog before. I have my objections to this idea (see my previous post for those details), but I have just thought of another problem. As mentioned, cells that can be removed from the body and modified in a dish are most likely the first ones we will be able to modify. In addition to blood cells, and perhaps therapies based on stem cells, our gametes (sperm and eggs) are cells that can be removed from the body (especially so with sperm) and modified outside the body, used to create embryos that can be re-implanted. Thus, it’s likely to be relatively easy to prevent certain genetic diseases before embryos with those disease genes are even created.

I suspect the pressure to cure diseases will be much greater than the pressures to create a clone, so a moratorium on human germline engineering is probably going to be more difficult to defend than the one on cloning.

Targeted gene therapy for HIV/AIDS

Some scientists (I still have no idea who) reported treating six patients with ex vivo gene therapy for the treatment of HIV/AIDS. Researchers used zinc finger nucleases to edit out the CCR5 gene from CD4-positive T-cells. Because HIV (or, most strains of it) uses the CCR5 protein to infect CD4+ T-cells, this essentially creates white blood cells that are immune to HIV. People with a deletion in their CCR5 gene (5-14% of Europeans have at least one such deletion, the CCR5-Δ32 allele) are essentially immune to AIDS (though they can sometimes get asymptomatic HIV infections).

All six patients showed some immune recovery, with five showing up to 6% of their cells modified, so that’s progress, I guess. Side effects were just a couple days of flu-like symptoms. In this study, the cells were removed from the patient and edited outside the body, probably because the ridiculous safety standards required for gene therapy that I recently blogged about pretty much rule out modifying the patient’s genes directly (i.e. in vivo). And I suppose the zinc finger nucleases probably aren’t as efficient as using a virus, so it wouldn’t really work in vivo just yet.

This is, however, the first time that human gene therapy has actually deleted a gene, rather than adding one. Pretty soon, addition, deletion, modification, replacement…it’ll all be possible.

Some scientists are, I think sensibly (with just 6% of cells modified), trying to avoid the work being overhyped as a ‘cure’. For example, Dr Michael Kolber, professor of medicine at the University of Miami:

“[This study] was a proof-of-principle that they could go in and do this. They demonstrated that the [genetically engineered] cells stayed in the patients, but the patients were not cured”

Of course this gets reported as:

Experts are reacting with cautious optimism…but they say the jury is out on whether the technique might ever spell an end to AIDS.

No, the jury isn’t out on whether the technique might ever work, just whether it has worked. It’s obvious beyond a reasonable doubt that gene therapy could provide both immunity and cures for HIV infections. And I’d say it’s obvious that gene therapy could cure all disease (and I’m not overhyping, just don’t expect these cures tomorrow and without side-effects during their early stages of development).

Gene therapy fixes (night) blindness

Those bionic eyes I blogged about have some competition – from genetic engineering. A form of blindness known as Leber’s congenital amaurosis is caused by a lack of a key gene, RPE65, in cells of the pigment epithelium cells. These are the cells just behind the photoreceptors, and they produce a key pigment molecule called 11-cis-retinal. This pigment is used primarily in rod photoreceptors, because the cones can make a similar molecule known 11-cis-retinaldehyde without relying on the pigment epithelium cells. However, the cones degenerate, meaning that Leber’s congenital amaurosis can cause blindness by middle-age.

Gene therapy was trialled on young adults (17-23), who still have reasonable colour vision thanks to functional cones. However, cone cells require more light to be stimulated, so the rod cells (which only show light and dark, or shades of gray) are used primarily in low-light conditions (and in peripheral vision).

The researchers used the common viral vector AAV (adeno-associated virus) plus an adeno 5 helper virus. This vector contained the DNA of the human RPE65 gene, including its promoter sequence (the terminator sequence was different though – bovine growth hormone polyadenylation sequence was used). They injected these into the subretinal space of one eye.

Unfortunately, though three patients underwent the surgery (at least, three were included in the study – I think it has been trialled on others), only one showed improvement. It was a dramatic improvement though – improved visual mobility, as measured by a navigation in a simulated street scene in low-light conditions – from 77seconds to 14 seconds, without bumping into the wall eight times like he did before the surgery. The researchers believe this lack of improvement in the other patients was due to their retinal degeneration being more severe, which if true would mean that this therapy should be targeted at children with the condition.

View full article in the New England Journal of Medicine.

Now the race is on. I wonder which will end up curing blindness first – bionics or genetics. This race is going to be in many areas aside from vision research, such as in giving mobility back to amputees (can we regrow limbs before we get fully functional prosthetics?). Personally, I’m willing to let both fields work, so that those who want it can be enhanced from both directions.

Those with germline-modified mitochondria walk among us

A news article on the mitochondrial transplants I blogged about previously has the headline: “First disease-free babies could be born in three years as doctors create embryo from THREE parents”

But, and I kick myself for not remembering (or, more accurately, rediscovering) a certain fact earlier – that babies have been born with genetic material from three parents over ten years ago. Let me tell you the story of the real first baby (technically, an embryo, although a healthy baby did result – see below) created from three parents. This story comes from the following source: J. Cohen et al “Birth of Infant after Transfer of Anucleate Donor Oocyte Cytoplasm into Recipient Eggs” The Lancet 350(9072): pp186-187, July 1997

In 1996-7, Jacques Cohen and his team from the Institute for Reproductive Medicine and Science at the Saint Barnabas Medical Center in Livingston, New Jersey (in the USA) were treating a 39-year old, whose efforts as assisted reproduction had all failed. She was offered, and consented to, a cytoplasm transfer. Ooplasm, containing mitochondria and all the other organelles and chemicals present in the cytoplasm of an oocyte, from the eggs of a 27-year old donor was extracted with a pipette and inserted into the eggs of the patient. The patient’s eggs were fertilised with her husband’s sperm, and four embryos were implanted. A single impregnation resulted, and a healthy baby girl (weighing 4356g) was born as a result.

The thing is, the researchers report the following:

We compared nuclear and mitochondrial DNA fingerprinting profiles from aspirated amniocytes with those from both parents and the egg donor, […] Donor mitochondria had been displaced by homologous mitochondria before 16 weeks’ gestation.

So, technically this wasn’t germline engineering, or at least wasn’t a ‘successful’ example of it. The now 11-year old girl currently contains none of the donor mitochondria, so can’t pass any but her own mitochondria to any children she may have. But the same is not the case in the other children (17 of them, I believe) born as a result of the same technique in that medical centre and others across America (as far as I know, it hasn’t happened outside of the US, but I could easily be wrong). According to the following article: Barritt et al “Mitochondria in human offspring derived from ooplasmic transplantation: A brief communication” Human Reproduction 16(3):pp513-516, March 2001

mtDNA fingerprinting analysis performed on blood samples from 1 year old children following ooplasmic transfer detected mitochondrial polymorphisms in which both alleles were present in the hypervariable region of the mitochondrial genome. […] These are the first reported cases of germline mtDNA genetic modification which have led to the inheritance of two mtDNA populations in the children resulting from ooplasmic transplantation. These mtDNA fingerprints demonstrate that the transferred mitochondria can be replicated and maintained in the offspring, therefore being a genetic modification without potentially altering mitochondrial function.

So, these children are actually the first germline-modified children to be born, as far as I know. But I shouldn’t use the words ‘germline genetically modified’, because since these words were used, the FDA tightened their restrictions on the procedure. In a meeting of the Biological Response Modifiers Advisory Commitee on May 9,2002, the FDA changed regulations to require each procedure to gain approval from the FDA before going ahead (read the minutes of the meeting here, or the full transcript here).

The only difference here between what I blogged about previously and the procedure discussed above is that the latter uses cytoplasmic transfer rather than nuclear transfer as in the former, but as I said in my previous blog on this, it is all relative to what is the defining characteristic of a cell – if a nucleus is the defining feature of a cell, then picking that up and putting in a different cell is actually a full cytoplasm transplant rather than nuclear transfer. (An analogy would be brain transplants vs body transplants – if personhood is localised in the brain, then a brain transplant is actually a body transplant).

Anyway, the point remains that the embryo created by British researchers is not the first to be created with three genetic parents. Let us hope that those members of the House of Commons in the UK are considering this fact when deciding whether to allow it in Britain.

Creating a knockout human

I said in my post about a possible genetic manipulation for viral immunity that a knockout human would “take too long and probably be unethical”. Well, I may have to eat my words (at least, the part about it taking too long). You see, I think there would be a way to make a knockout human in a shorter time than making a knockout mouse (of course, creating a knockout mouse with this method would be shorter still). I’ll explain how a knockout mouse is usually created, and then I’ll tell you the shortcut to be taken for long-lived creatures like humans.

Firstly, stem cells are isolated from a mouse embryo. These cells are genetically manipulated to insert a non-functional version of the gene of interest, hence “knocking out” the gene by overwriting it with a broken copy (a functional gene from another animal could be added to create a ‘knock in” mouse). The inserted gene is usually accompanied by a marker gene – a gene that can be detected (such as immunity to a toxic agent, which can be detected by growing the stem cells in a mixture containing that agent). Then, stem cells in which the gene has been successfully inserted are inserted into another (or the same) blastocyst, meaning that the mouse growing from that embryo will be a chimera – containing some cells that have been genetically modified and some that haven’t. If some of the germ cells (eggs or sperm) happen to have arisen from cells that have been genetically engineered, then those mice can be bred with another “knockout” mouse (or inbred) to create a mouse where both chromosomes contain genetically engineered genes.

This takes so long because to get your hands on a knockout mouse embryo requires you to wait for some mice to get old enough to breed. For mice, this is only about 5-8 weeks, but in humans it would be at minimum about 12 years and safely (and legally) at around 16/18/21 years. And not only is this a problem in terms of time, but forcing people to copulate, and specifying their partner, is a big ethical problem (even if we accept the ethical acceptability of genetic modification of humans). But this can be overcome, although it may bring up other ethical problems of its own.

The shortcut stems from the fact that we don’t actually need a man and a woman to create a child. Females would actually have all their primary oocytes (immature eggs) around 4-6 months before birth, but they mature and are released only after puberty. Likewise male embryos contain spermatogonia (although unlike in the female, they will be produced throughout life), which develop into spermatozoa after puberty. Turning primary oocytes into ova (mature eggs) is fairly simple to do in vitro, involving just a set of chemical signals (and has been done before to result in pregnancy). Turning spermatogonia into spermatozoa to achieve conception is harder, but it can be done in vivo (by putting the embryonic spermatogonia into a donor testicle, even one of a different species). So, one could possibly take a shortcut by performing some advanced IVF on germ cells harvested from the initial chimeric human embryos.

In solving this practical problem, numerous ethical issues could be raised. Firstly, one could object to creating a child whose genetic parents would have been embryos or foetuses at the time of conception (and need not even be implanted). Surely it is unethical to force two people to mate, but provided one accepts the discarding of embryos from IVF (and indeed normal reproduction), and also accepts abortion five months before birth, then why should this be an issue? If we can deny that the embryo or foetus is a person, or otherwise deny it a right to life, then we can surely deny the embryo or foetus the right to choose a sexual partner. One could also object to the xenotransplantation for maturation of human sperm (i.e. having human sperm grow in the testes of another animal), but this is likely a technical hurdle to be overcome, and so soon human sperm could be grown in a dish like we do for human eggs.

Perhaps an easier method, that gets around the above ethical hurdles, would be to transfer the nucleus of one of the genetically modified stem cells into an ovum and stimulate it to mature into a knockout human. This could negate the need to create the chimeric generation altogether. But, this is called cloning, and is frowned upon in most Western nations.

Of course, we are likely to come up with much better methods of genetic interventions in mice, and soon I expect the knockout procedure will be replaced with something more efficient. The easiest, I think, would be finding a way to dedifferentiate (turn back) the genetically modified stem cells from pluripotency to totipotency, allowing them to develop into an embryo without forming the chimera. In fact, I would not be at all surprised if such a procedure had already been done recently (but won’t that upset the pro-life crowd – “every embryonic-stem cell is sacred”?).

So, bottom line is that I was wrong. But in my defence, the press release did report that one of the researchers said something even more incorrect. So you can all look at his comment to distract yourself from the fact that I said something wrong:

Dr. Sonenberg explained that the process of knocking out genes is not possible in humans, but the researchers are optimistic new pharmaceutical therapies will evolve from their research. [emphasis mine to increase the power of the distraction]

Genetic Disease Immunity on the Horizon?

Researchers at McGill University have genetically engineered mice to be immune to viruses. Well, at least four viruses anyway (influenza virus, encephalomyocarditis virus, VSV and SNV).

Still, let us look at how they did it, and then whether I can get it done to myself so that I don’t need flu shots every year. Firstly, they made some knockout mice for the genes 4E-BP1 and 4E-BP2 (meaning, those genes are deactivated in those mice). Those genes are apparently repressor genes, and decrease the production of immune proteins called interferons. Knock out a repressor, no more repression = more interferons. Because interferons, especially type 1 interferons, have antiviral properties, the mice essentially became immune to the viruses tested.

And, the researchers report no noticeable side-effects. However, given that fact that genes have actually evolved to stop interferon production, I still think we can’t rule out side-effects in future trials.

So, where do we sign up for the genetic mods for viral immunity? You can’t. Because we can’t produce knockout humans (technically, we could, but it would take too long and probably be unethical), and somatic gene therapy (especially gene therapy to deactivate a specific gene*) is still a very young science. And germline modification, in case you were thinking about vaccinating your infants before they are even born, is oh so very illegal (although considering the comparison to infant vaccination I just gave, I challenge anyone to tell me why it should remain illegal).

The researchers, being the clever sorts that they are, suggest an alternative:

“If we are able to target 4E-BP1 and 4E-BP2 with drugs, we will have a molecule that can protect you from viral infection. That’s a very exciting idea.” Dr. Costa-Mattiolo said. “We don’t have that yet, but it’s the obvious next step.

I don’t agree that it is the ‘obvious next step’ (I’d say, ‘good near-term compromise’) but still the idea is ok. I’d rather take a pill to become immune to the flu than catch the flu, but I’d still rather be immune to the flu because my parents had be ‘genetically vaccinated’ as an embryo. What if a flu pandemic hits and the stockpile of the drugs runs out? What if I fall on hard times and can’t afford the drugs (assuming they aren’t fully covered by government health schemes in the country in which I happen to be living at the time)?

Oh well. I guess I’ll have to take this antiviral drug (or benefit from the herd immunity of everyone else taking theirs), and wait for my gene therapy.

*I think an alternative could be to insert additional genes that produce Irf7 – the regulatory protein that has its synthesis repressed by 4E-BP1 and 4E-BP2. Or even add a gene that represses (transcriptionally silences) 4E-BP1 and 4E-BP2.