Can Genome Editing Cure AIDS?

Hinco Gierman

Yes. At least, in theory. But, theory might soon become practice according to this week’s issue of the prestigious New England Journal of Medicine [1]. It published the first clinical trial of genome editing, in which the DNA of 12 human volunteers was “edited” to make them “genetically resistant” to the HIV virus.

What do I mean by genome editing and is it safe? How is this different from other gene therapies? Why would this cure AIDS? Is it unethical to genetically modify humans? And of course most importantly, should we be worried about “genetically modified super-humans” taking over the planet?


Genome editing means precisely changing a sequence of DNA in an exact location. In this study, the location was a gene called CCR5, which produces a protein that normally sticks to the outside of some of our immune cells (T cells). The HIV virus needs this protein to infect the T cells that it uses to multiply. The editing was deleting a piece of the CCR5 gene, which makes it inactive and leaves the HIV virus without a way to enter our cells. This deletion, called delta-32, naturally occurs in Northern Europeans [2,3]. Six years ago in Berlin, an HIV patient with leukemia received cells from a bone marrow donor that had the delta-32 mutation in both of the CCR5 gene copies (we inherit every gene from both our parents, so have two copies of each gene). The Berlin patient was cured from his leukemia, but also from his HIV infection, and to this day remains clear of HIV virus [4].

But bone marrow transplants are dangerous, costly and troublesome. So the authors of this study isolated T cells from the blood of 12 HIV patients, then edited the genomes of these cells, and put them back in the patients. Next, antiviral treatment was stopped in a few of the patients, which then lead to the HIV virus becoming active and trying to infect their T cells. The patients now had both normal and edited T cells. The edited T cells started taking over, as they were more resistant to the HIV virus. In other words, the mutated T cells are now fighting a Darwinian battle with the normal T cells. The HIV virus is literally digging it’s own grave, by providing the natural selection that is favoring the modified T cells (which it cannot use), ultimately leading to HIV’s own demise.

In one of the patients, levels of HIV virus became undetectable. But here it gets tricky: This patient already had one CCR5 delta-32 mutation of his own. The treatment mutated his other copy of the CCR5 gene in enough T cells (around 9%), so that they became completely resistant to HIV. This is a great success, yet at the same time points out a weakness of the treatment: in the other patients, although 11 to 28% of T cells were genetically modified in at least one CCR5 copy, only a 1-7% were modified in both CCR5 copies. And although one edited copy already makes T cells a bit more resistant, patients can probably only be cured if there are enough T cells with two both CCR5 genes edited.

Another important issue to point out is that these edited T cells were slowly but surely depleting. After about a year, half of them were gone. In other words, the treatment is temporary. The Berlin patient had an important advantage: he received bone marrow, which contains the (immortal) stem cells that actually produce our T cells. So, the Berlin patient will probably have permanently “modified” T cells al his life. In this trial it seems only one patient might have been cured (which needs to be seen). Maybe several transfusions are needed to cure some of the patients. This is especially important as HIV can hide in “reservoirs” of resting T cells.


Gene therapy means using DNA to cure disease and has been around for over 20 years. Gene therapy started by inserting an extra piece of DNA randomly into the genome of a patient. This approach has been successful in treating genetic diseases caused by a faulty gene, where inserting an extra healthy copy can treat the patient. One of the problems with that approach is that when you randomly insert a piece of DNA into a patient’s genome, this can have unforeseen consequences. For example, several patients that underwent such treatments got cancer as a result of the inserted gene activating a “cancer gene”.

More recent gene therapy approaches circumvent this danger by not inserting the healthy DNA into the genome. Instead, the healthy DNA forms a little circular mini-chromosome of it’s own (i.e. it floats around without integrating into your genome). The recently EU-approved gene therapy Glybera is an example of such a strategy [5]. Genome editing is different from the integrating and non-integrating approach in that it precisely fixes the mistake itself. This can be important when the mistake produces a “toxic” protein that needs to be fixed, like in the case of Huntington’s disease (HD).

HD is a deadly genetic disease for which no cure exists. And this is probably why the same company that is developing the CCR5 treatment is also working on a HD treatment. But curing it is not as easy as it sounds: for many diseases like HD, it is not entirely clear in which part of the body the genetic mistake needs to be fixed. Blood cells are easy to take out of the body and put back in, which is one of the reasons this study chose HIV patients. But for Huntington’s (which affects brain and muscle cells), it might be necessary to fix several types of cells. More importantly, it’s probably not a good idea to take your brain out of your body and put it back in. So you have to inject the patient with a virus that would deliver the genome editing proteins and DNA to the cells inside the body. Not so easy indeed.

Still, if I were to have Huntington’s and had only little time left, I would not have to think long before signing up for a clinical trial that might save my life. But I don’t know what I would do if I were an HIV patient. It is hard to say whether the potential benefits of genome editing outweigh the (unknown) risks, for patients of a disease that actually has a treatment.


The CCR5 delta-32 deletion already cured the Berlin patient of AIDS, by depriving the HIV virus of the CCR5 protein it needs to enter and infect T cells. If genome editing produces enough T cells with double CCR5 deletions, this technique could in theory cure any patient this way, if the treatment lasts long enough to deplete all HIV reservoirs.

Time to get into the fine print a bit more: The patients that all had normal CCR5 genes only had 1-7% of cells with double mutations after editing and still had detectable levels of HIV. So the technique might need to become more efficient before it can be fully successful in all patients. Also, it is not clear that this is safer than other forms of gene therapy. This study relied on so-called Zinc finger proteins, which are proteins that can be engineered to recognize specific DNA sequences. However, they are not always 100% accurate. Previous studies on this “CCR5” zinc finger, showed when checking 23 predicted “off-target” sites (i.e. DNA similar to the intended CCR5 target), 4 of them showed some editing, including the CCR2 gene [6]. It is unknown if other places of the genome can be targeted by accident, and also how many random mutations the procedure itself generates. Especially when the editing is a deletion or small change, the only way to check this, is by sequencing the entire genome of every single “edited” cell (which is nearly impossible). So it seems, that larger trials with longer follow-up will be needed to see if this technique is generally safe.

Other genome editing tools (TALENS, CRISPR) are in development, and it seems likely that the great unmet clinical need for lethal genetic diseases alone is sufficient to drive the development of gene therapies like genome editing.

But is genome editing a good or bad? One way to assess whether a (biotech) invention is good or bad is by weighing its risks and benefits for both individual and society. I’ve already discussed some of the potential risks and benefits for the individual, what about society?

Well, needless to say that curing chronic diseases like AIDS can lift a tremendous economic burden of society. Not just because of the cost of disease, but because AIDS affects young people in the workforce. But most genome editing therapies target disabling or lethal genetic diseases. Although potentially costly in treatment, these affect a relatively small group of people. So would genome editing have a large impact on society in the future?


Let’s be unscientific and speculate what could happen if genome editing becomes successful and applicable for any cell within the body without severe side effects. Of course, it would mean that many people carrying a deadly genetic disease like Huntington’s now have a potential cure. But once we do that, wouldn’t we want to start fixing other mistakes?

Take for example Angelina Jolie. She had her breasts removed because she had a mutation that increased her risk of breast and ovary cancer [7]. Maybe 10 years from now, she could have simply undergone a genome editing treatment to fix the mistake? And by now, you must have noticed we have started going down the famous slippery slope: Before you know it, we’re all genome editing ourselves and our kids into super-humans!

Let me start by saying that a slippery slope is not an argument in itself, as it assumes causality between an action and an (assumed) undesired consequence. Speculating about science is fun, but very unscientific, and can be more dangerous than the science itself if it leads to policies based on fear rather than fact. In other words, genome editing or gene therapy is not inherently dangerous just because of what it could be used for.

It might be good to point out here that we humans are being genetically modified all the time. For example, every HIV patient is a genetically modified human. The HIV virus inserts its own genetic material into our genome. Not to mention the mutations we accumulate every time we go out in the sun, or the countless number of DNA re-arrangements in our immune cells that form the basis of our adaptive immune system. There are other good arguments why genetically modified humans are not different from normal human per se. Hank Greely brings up a few good ones in his recent and excellent article discussing another novel Biotech treatment involving mitochondrial DNA donors [8].

But just like the CCR5 study, those are all examples of “somatic” mutations (meaning in the DNA of our body but not in the DNA of our sperm and egg cells, which get passed on to future generations). What if we start genome editing our babies? That sounds like science fiction, but the selection of healthy in vitro fertilized (IVF) embryos based on their DNA is already an existing practice.

So let me give one example of a potential danger if we start genome editing our babies and in effect, our species. We have identified thousands of DNA variations that are not deadly per se, but can cause disease or other undesirable effects. For example, a few well-characterized mutations in the hemoglobin genes can cause deadly anemic disease. Wouldn’t those be good candidates for genome editing? Well, just like the deletion in CCR5, these mutations turn out to protect against a disease, in this case malaria [9]. Although these mutations are rare in places without malaria, up to half of all people in a malaria-endemic region can carry these mutations. This is because natural selection in such places favors those mutations. Just like the CCR5 modified T cells, got favored by natural selection in the body of AIDS patients. My point is that in some cases, “bad mutations” can be life saving.

In conclusion, I think there is no inherent objection to genome editing or other forms of gene therapy. I do think we need to see what side effects can occur and weight them against the benefit, which might be different for every disease. It is important to realize that our genetic diversity is our main insurance policy against (future) disease. So we should be careful not to start editing all the “glitches” out of our genomes, as we can lose mutations that could protected us from future disease. So it may be those natural mutants who are the super-humans after all!

Hinco J. Gierman

Dr. Gierman, a frequent participant in CLB discussions, is finishing a post-doc at Stanford and will soon be moving to a position with Illumina. We’re glad we could get a blog post from him while he’s still at Stanford!

Footnotes & References

  1. Tebas, P. et al. (2014) N Engl J Med 370: 901–910.
  2. Note: The minor allele frequency of the CCR5 delta-32 allele is as high as 18% in some parts of Northern Europe (see footnote 3). The frequency drops close to 0% in Southern Europe. Importantly, only people carrying this deletion on both their chromosomes are resistant to infection by (most) HIV viruses. This would only be 3% of people in a population where the minor allele frequency is 18%, but is close to zero in most parts of the world. T cells with this mutation do have some drawbacks, but the reason it is high in some parts of Europe might be due to it conferring a similar advantage for other infectious diseases like smallpox.
  3. Novembre, J. et al. (2005) PLoS Biol 3: e339.
  4. Allers, K. et al. (2011) Blood 117: 2791–2799.
  5. Miller N (2012) Nat Rev Drug Discov 11: 419–419.
  6. Li L, et al. (2013) Mol Ther 21: 1259–1269.
  7. Jolie A (2013) New York Times. May 15.
  8. Greely H (2014) Law and Biosciences Blog. March 2.
  9. Note: In the case of hemoglobin disorders (e.g. sickle cell disease), it is actually beneficial to have one mutated and one normal copy of the gene. The mutated copy changes the shape of red blood cells, and prevents malaria parasites from multiplying. However, these mutated red blood cells are rigid and can cause deadly problems like . So, having one normal copy improves oxygenation of the body.
  10. Thanks to Jon Geisinger, Rutger Wielink, Ari Friedland and Hank Greely for critical reading.