What is genome modification? (Part II)
In part I of the “Running with scissors” post, I described for what purposes scientists would wish to use genome editing, especially in relation to humans. What is still lacking is a “look under the bonnet” of genome editing itself. This includes issues ranging from the mechanism with which CRISPR/Cas9 specifically changes DNA to side-effects and to the issue of embryonic DNA.
“Dumping” DNA into a cell: gene therapy
Prior to genome editing, it had been possible to insert a stretch of DNA into the genome of an organism. For example, in the 1990s, a Paris hospital treated children affected by an illness called “severe combined immunodeficiency” (SCID). Essentially, they did not have a functioning immune system, as a genetic disorder kept their white blood cells from doing their part. The approach taken by the physicians and researchers is called “gene therapy”: a virus was designed that would incorporate a functional equivalent of the defective gene into the children’s DNA.
However, the precise location of the insertion was beyond control, and for lack of a mechanism to cut DNA at a defined position, the irregular gene remained in place. This is important as the human genome involves well-tuned interactions between different sections of the genome. Any moderately complex organism self-regulates by means of refined genetic feedback loops: some genes, via the proteins they encode, have an effect on other genes. This way, genes are activated, suppressed, or regulated in other ways.
While gene therapy initially cured at least seventeen SCID patients, three of them went on to develop leukemia, a cancer of the blood cells. This led to a halt in treating SCID with gene therapy. It seems that in those affected by leukemia, “gene therapy” didn’t simply restore functionality, but caused new problems within the organism’s genetic feedback loops. A random integration of new genetic material may, for example, activate an oncogene (a gene causing cancerous growth) or compromise another gene that prevents such cancerous growth. “Dumping” genetic material into the genome at random is a dodgy business, as organisms are complex systems .
The first automated gene scissors
Things changed with the creation of so-called TALENs and ZFNs. Starting in the late 1990s, a group of researchers discovered that a certain protein occurring in certain bacteria first targets certain DNA sequences and then breaks, or “cuts”, the DNA at that target . Proteins are the complex building blocks of all organisms. They come in various kinds, depending on the sequence and the kinds of aminoacids from which they are built. With the protein in question here, it turned out that it consisted of two parts: one part that targets a particular DNA pattern and the scissors part (a so-called enzyme, a type of protein). Researchers then found that the ‘targeting part’ can be replaced with something similar in the lab. This allows for the definition of practically any genetic target. Automated scissors results that give you the option both to define the cutting target at will and to go ahead and make the cut, as it were. TALENs and ZFNs are such programmable scissors.
To be sure, any cell has a DNA-repair mechanism that fixes broken DNA strands. But even this repair mechanism often leaves the cut DNA dysfunctional. So the new gene scissors already give you the opportunity to “knock out” a gene. Both TALENs and ZFNs involve the cutting part of the original protein that was noted for its targeting and cutting capacity. TALENs and ZFNs differ, however, in the protein they use to home in on the site that is to be cut. ZFNs were developed earlier, but TALENs allow for greater flexibility in identifying that site.
Gene therapy + gene scissors = genome editing
But then you might also combine the new gene scissors with the ability to ferry a lab-made DNA stretch into the cell, the procedure on which gene therapy relied. Your new gene scissors can cut the faulty gene at the beginning and the end, while the new piece of DNA, a replacement for the cut-out, is delivered to the construction site. The cell’s self-repair mechanism would then forge a connection between the new DNA strand and the original DNA, at the site where a dysfunctional gene used to be. The new genetic material is thus incorporated, and the genome has been modified. This is the old gene therapy taken to an entirely new level. Of course just “knocking out”, or rendering dysfunctional, a genetic sequence without making any inserts is another option the new gene scissors give geneticists, and that is genome editing, too.
For all of the progress these tools mean, ZFNs and TALENs are still “a relatively cumbersome tool“. The target stretch of DNA typically differs between individual organisms, and the component that encodes this target is a protein. This means that for every new type of cut in every individual organism, a new protein needs to be tailor-made.
For example, in late 2015, an infant was rid of leukemia through the use of TALENs at a London hospital. For this purpose, white blood cells of a particular kind (T-cells) were donated by a healthy person. In a dish, researchers inserted a gene into these cells so they would eventually attack the cancerous cells of the ill child, and then TALENs were used to modify these cells genetically in two ways: they were kept from attacking anything in the recipient’s body besides the cancer cells, and they were further modified so they would evade the attack of a leukemia drug given to the patient. So modified, the new cells were injected into the patient’s body. Indeed they brought down the cancer cells to such a degree that a bone marrow transplant was possible, through which the patient would then produce regular white blood cells instead of cancerous ones. So for the two last genetic modifications of the donor cells, ‘bespoke proteins’ had to be created. They are the elements that indicate where the cell’s DNA would be cut in order to insert new material. This is rather laborious, given that proteins are fairly complex entities, consisting of many different amino acids that fold in intricate ways. For every individual T-cell donor, these target proteins need to be custom-made.
So what makes for a long-winded procedure in the use of a versatile new development is the difficulty in defining the target for the cut. This is where CRISPR/Cas9 enters the picture. CRISPR/Cas9 is at the forefront of research on genetic modification.
Similar to ZFNs and TALENs, CRISPR/Cas9 consists of two components, one that defines the target area for cleaving DNA – this is “CRISPR” – and one that cuts DNA at that location – that’s the “Cas9” enzyme. The crucial novelty, however, is that CRISPR is an RNA – often called a “guide RNA” – rather than a protein. RNA is a substance very similar to DNA. Indeed cells use RNA when copying their DNA or translating it into amino acids and proteins. This means that the definition of the target site is almost a 1:1 copy of the target site itself, which is significantly less complex than the targeting protein of ZFNs and TALENs. This reduces the complexity of cutting DNA significantly.
In addition, a set of CRISPR/Cas9 scissors can be equipped with several different guide RNAs, which allows for multiple successive cuts in one procedure. This option has been helpful, for instance, in the effort to rid the pig genome of those traces that pig-specific retroviruses wrote into pig-DNA. Here over 60 stretches of DNA had to be removed. The point of this effort would be to prepare a pig that could host potential donor organs, which should in turn not pose a risk of importing viral DNA from pigs into humans. I’ve covered the issue of donor organs grown in other organisms (“xenotransplantation”) in this post.
Interestingly, CRISPR is used by bacteria to defend themselves against the attacks of certain viruses. Once a bacterium has survived a viral attack, it incorporates the viral RNA into its own genetic information. This helps the bacterium in recognising this type of virus in a later encounter. The bacterium is now on its guard. CRISPR occurs, for example, in those bacteria involved in making yoghurt and cheese, so chances are that you eat CRISPR on a regular basis! Of course that does not mean that there is any genome editing in the technical sense going on in yoghurt or inside you.
Further technical issues
One problem with CRISPR/Cas9 is that it may make cuts in other locations of the DNA than encoded in its guide RNA after all. These would be so-called off-target effects. Certainly any medication has certain side-effects, but if CRISPR/Cas9 is ever used to combat disease in humans, this will involve making changes in millions of cells. An error ratio of 0.1 % may even be too high. So researchers are testing variations of Cas9 in the lab, and progress has been made to bring off-target effects down, perhaps even to undetectable levels.
Another problem is that even if the cutting is highly precise, merging the DNA strands with a newly provided piece of DNA may not work reliably enough. To replace the excised genetic material with new material, researchers are still relying on the cell’s mechanism of self-repair that was mentioned above, only that the repair mechanism is slipped new material. Is the traditional repair mechanism up to the task?
Unless these issues are solved properly, a problem called mosaicism can result: an organism with genetic information differing between cells. Sure our organism has lots of different types of cells, but crucially the genetic information in your skin and your brain cells is the same. With mosaicism, that would no longer be the case. If, by consequence, the number of dysfunctional cells remains above a critical threshold and, say, the body does not synthesize a large enough amount of a critical protein, or the wrong kind of substance accumulates in these cells, the whole procedure has failed.
Somatic and sex cells
A further critical issue is at what stage an organism would be treated with genome editing methods. Does a genetic modification affect the current organism only or are genetic changes heritable? The crucial point is the moment when, in the embryo, somatic cells start to differentiate from sex cells. The sex cells are the egg cells or the sperm cells. The cells that give rise to these might be changed as well, and those changes would be heritable as well. After a certain point early in development, these cells develop along a track that is separate from the rest of the body’s cells. These sex cells are often referred to as germ (line) cells. Eventually, the adult organism produces offspring in drawing on these sex cells, and on them only. So if a genetic modification is supposed to affect not only the current organism but, via genetic inheritance of modifications, the succeeding generations as well, it is no use modifying the somatic cells. The sex cells would be the target for heritable modifications. The technical method of modifying them would be to modify the genome before the differentiation of sex cells and somatic cells in the embryo. Technically speaking, this would also be the most efficient way to make sure that all the genetic changes will be manifest in a maximum number of cells in the current organism, regardless of inheritance.
So to make changes last throughout the generations, or to keep a disorder from spreading throughout the body in the first place, the genetic modification of an embryo is the procedure several scientists consider. It seems that this sort of work is already being carried out in four labs (in the UK, Sweden, and China). This procedure raises further questions that will be addressed in a later post.
 I adopt the admittedly casual expression of “dumping DNA” from John Parrington.
 See the article “Genome-editing tools storm ahead” from 2012 with further references.