This is a teaser chapter for my DNA book, and it might be hard to read outside the context of the book, unless you are already reasonably conversant with DNA.
Since life has its origin in DNA, one can claim that by extension, so does disease. And to a certain degree this is usefully true. Some diseases have an immediate genetic origin. For many more diseases, we know of genetic differences that significantly raise your chances of suffering from an affliction. And finally, as outlined in chapter X, we have genetic hints for many other maladies.
Sometimes when a gene malfunctions, things clearly fail. In such cases, it may be beneficial to either shut down a gene, to repair it, or to supply a fixed gene somehow.
In the lab, we have pretty good capabilities to engineer cells, where we can create individual cells to prescription. But for treating humans, that is mostly not what we need. If we treat thousands of cells and end up fixing just one, while messing up the rest, that is not something you can do to a living human being. A significant exception is when we can take the cells out of the patient for treatment, as with the CAR-T cells from chapter x. Such cells can be improved ‘offline’, multiplied, checked, and then be returned to the patient.
To generally fix genes in actual human beings with disease however, we need very targeted techniques that either deliver or do not harm a cell.
One of the hopes behind the Human Genome Project was that it would point us the way towards many diseases which ‘the genetic blueprint’ would then be able to address. This mostly did not happen, but we are getting tantalizingly close now.
The first diseases considered for gene therapy were of course the very lethal ones - treatment need not even be very good to have a net positive effect. In the 1980s, an obvious first target was @severe combined immunodeficiency@ (or SCID), which leads to newborns with no functioning immune system. Without treatment, any infection can then prove lethal. Sufferers without access to advanced medical care tend to die before they are two years old.
SCID has a number of different genetic causes, and a common one is a broken @ADA gene@.
Already in 1990, it proved possible to infect white blood cells from a then four year old SCID-ADA patient with a γ-retrovirus that incorporated a fixed ADA gene. Together with weekly ADA injections (which on their own are not sufficient), this restored her immune system. As of 2021, this initial patient is still alive. Similar experiments from that time however were not effective or even lethal - it turned out that the then chosen viruses could insert the fixed gene in very inconvenient places. In addition, some insertions appeared prone to very harmful behaviour like activating cancer causing @oncogenes@.
A virus aiming to integrate its genetic material with the host genome will tend to feature powerful ‘enhancers’, which enhance not only its own expression levels, but also those of surrounding genes which otherwise might have been dormant. Such genes may have been silenced for a good reason, for example because they cause cancer.
Gene inserts that turn out to cause cancer (often leukemia) are a common setback in gene editing experiments, but it is something we are getting a handle on. And since SCID is such a devastating disease with genetically quite clear causes, research continued apace.
After initial problems, it appears there is now great success inserting genes using what are called @lentivirus@es. This is the family of viruses to which HIV belongs, but press releases don’t dwell on this fact too much. And although researchers tend to not share too many details on the viral vector they designed, many of them appear to derive directly from HIV-1. Our huge stock of knowledge of HIV must surely have been helpful to craft the lentiviruses used for gene insertion.
As an example, in 2021, the results of a highly successful trial were reported, in which a lentivirus construct called EFS-ADA LV was used on SCID patients to insert a fixed ADA gene in @hematopoietic stem cell@s. These cells are the stem cells that give rise to other blood cells.
During three years of followup, around 95% of patients treated were living an ‘event free’ life, a massive success any way you measure it. This treatment is now entering phase 3 (‘real life’) trials under the name @OTL-101@. A similar lentivirus-borne treatment in 2018 for another common SCID variant was also remarkably successful.
In 2021 we also learned how another lentivirus was used highly successfully to treat sickle cell anemia in a small group of patients. Hematopoietic stem and progenitor cells were infected and provided with a replacement β-globin gene that encodes for antisickling hemoglobin. Out of 25 participants for which full data was available, all 25 saw complete resolution of severe vaso-occlusive events for the duration of observation (median 17.3 months).
Similarly, in 2022 it was reported that using an AAV5 vector targeted at the liver, 112 study subjects with hemophilia A were almost universally able to discontinue use of their previous treatment, and they saw a 83.8% reduction in bleeding events. Interestingly, 25% of the treatment candidates had to be excluded because they already carried AAV5 antibodies.
Although these treatments are very successful, it is found that the new genes do end up in loads of places in the genome. The lentivirus constructs are able to get their genes inserted near other expressed genes, which is pretty good work. But it is not very targeted. In this case since the alternative of not treating a scary disease is terrible, this can be tolerated. But there are clear risks, and in one trial it was noted that several gene copies landed in potentially dangerous places.
Crucially also, when modifying cells in the lab, it is not a problem if a lot of them don’t survive the treatment. As long as enough of them make it so a sufficient number can be cloned to be infused back into the patient. But this is not something you’d like to chance with the very cells you are trying to improve inside a human being. In a patient we can’t kill lots of the very cells we are trying to fix.
One way of improving on this situation is by using a virus that does not deposit itself in an existing chromosome, but instead inserts a separate ‘@episome@’ in the nucleus. This is like a little chromosome, except that it does not get replicated on cell division.
Such an episome can be inserted using an @Adeno-associated virus@. These are tiny viruses that can’t replicate on their own, but need a helper virus for that purpose. The original wild-type @AAV@ did attempt to integrate its DNA into a single specific place in our genome, but modified versions of the virus have had that ability removed.
AAV is a tiny virus, so it can’t carry a lot of genetic material. This means that for gene therapy, it can only deliver rather small genes. This limitation is so severe that attempts are being made to use several different AAV-constructs that deliver separate DNA strands that could then join up as a larger episome.
In addition, the episome does not replicate, so this can only be used to deposit a gene in non-dividing cells.
Even with these limitations, AAV has delivered useful results for several eye diseases. Specifically for Leber Congenital Amaurosis it has proven possible to infect retinal cells to get them a fixed copy of the @RPE65@ gene, which at least for a few years appears to deliver remarkable results in some patients. This is possible because RPE65, once stripped of introns, is small enough to fit in the AAV vector.
Work is continuing to find suitable viral delivery methods. Since a lot of viruses are around, and we don’t know them that well, chances are high we’ll eventually find better candidates.
Transiently modifying genes
All of the available viral vectors have issues. Lentiviruses may deposit a gene in the wrong place. AAV has very limited packaging capacity. Not all viruses can reach all cells. In addition, some of these viruses also lead to the formation of antibodies, so it may be a one time treatment. There is also pre-existing immunity which could hinder applicability.
Sometimes there are other ways of fixing genes. For example, if a patient has a gene that has damage within a single exon, it may be that if that exon is skipped, that a useful protein remains.
So how do you induce such skipping? DNA gets transcribed to RNA with all the exons (and introns) still in there. The RNA then gets processed by the spliceosome which cuts out the introns, and possibly also some exons. Crucial in this process are the splice acceptor & donor sites and the exonic splice enhancer (ESE) motifs. If such elements can be masked from the spliceosome, this will affect which exons will get spliced out or not.
One way of making this happen is by circulating molecules called phosphorodiamidate Morpholino oligomers. These are a modified form of DNA that can bind to RNA. Crucially, these ‘PMOs’ do not cause the degradation of the RNA. In addition, because they’ve been modified enough, they also won’t immediately be degraded by the cell and can hang around for a while.
There are a few diseases caused by defects in very large genes, where such exon skipping might be useful.
In Usher syndrome, which progressively degrades both vision and hearing, the large gene USH2A has a ton of exons. Mutations are often found in exon 13, and through intensive research, it was found that exon 13 could be skipped, and still leave a useful protein. Crucially, such a thing is only possible if the exon contains a whole number of codons. Otherwise, the skipping event would cause a frameshift.
If anyone is doing research on skipping USH2A exons 53 and 68, my family would love to hear from you.
In 2021, researchers (mostly) from The Netherlands announced that they’d been able to use PMOs to induce the skipping of exon 13 of USH2A and that this indeed seemed to lead to an improvement of the vision of the patients taking part in the trial compared to the sham-treated control group patients, at least for a while.
In an even more devastating disease, Duchenne muscular dystrophy, a similar situation exists, whereby the skipping of specific exons of the huge DMD gene might be helpful. This appears to work pretty well on paper, but not very convincingly as yet in reality. A problem is that PMOs need to get into the right cells to do their work, and this is not easy. Several PMO medicines (Eteplirsen, Casimersen, Casimersen, Golodirsen) are in various states of approval. When measured, all of these clearly improve the production of the DMD protein, sometimes by a factor of ten. The genetic therapy is therefore definitely working. But the effect on people’s lives has not been as convincing – perhaps because the DMD levels continue to be very low compared to non-affected people.
On a personal note, we might speculate on the possibility of using an AAV to deliver an episome that codes for antisense RNA that could durably influence exon skipping. This could sidestep the limited delivery capability of AAV, but still deliver a perhaps lasting result.
Despite tremendous hype, CRISPR suffers from similar challenges as the viral vectors: while it can effectively change our DNA, it may also change it where you don’t want any change. In addition, in order for CRISPR to do its work, it still needs to get into the right cells, which is a challenge in its own right.
In a recent first successful application, CRISPR was used in a clever way to edit red blood cells. In various diseases, the shape of red blood corpuscles is wrong. It turns out however that babies get born with a separate hemoglobin production system, one that eventually (mostly) shuts down, to be replaced by adult hemoglobin. It has been observed that people that theoretically should have sickle cell disease or β-thalassemia, this shutting down sometimes did not happen, and that this can be traced back to their fetal hemoglobin production staying active into adulthood. This spares these people from disease.
Further research found that the fetal hemoglobin production decreases as long as a gene called BCL11A is being expressed. Other experiments had also shown that suppressing (down-regulating) BCL11A amped up the concentration of fetal hemoglobin.
In 2021, first results were presented for two patients who’d had blood cells edited using CRISPR to reduce BCL11A expression. Both of these patients saw their sickle cell disease and β-thalassemia effectively resolved, which is great news. In addition, because this is well funded research, their cells were heavily scrutinized to see if CRISPR had made any off-target edits, but it hadn’t.
Although the primary diseases were well solved, both patients suffered a raft of quite serious side effects, which do not strictly appear to be CRISPR related.
Nevertheless, this appears to be proof that CRISPR can usefully applied to curing human disease, but for now still by doing the processing outside of the human body.
Gene therapy is already pretty successful in some places, and promising in others. The sickle cell results appear to be quite stunning, as do the SCID treatments. The results for eye disease are also sometimes very exciting, although initial enthusiasm may have been overblown.
A central challenge remains that the most useful viruses can’t go everywhere in our bodies, and on top of that deposit their gene payload in random places, places which might not be good. Other viruses are more surgical, but can only deliver smaller quantities of genetic information. In addition such delivery may not be suitable for dividing cells.
Meanwhile, temporary suppression/modification of genes by antisense PMO constructs does have effects on measured protein concentrations, but perhaps not so much yet on clinical outcomes.
And finally, despite all the massive hype, as of 2022 there is only a single CRISPR treatment that has delivered results, convincingly solving the primary disease, but possibly at an unacceptable cost in terms of adverse side effects.
Despite all this, the concepts have definitely been proven, and there are not lots of people living far better lives thanks to gene therapy, and from hereon things will only get better!
- Antisense oligonucleotide-based treatment of retinitis pigmentosa caused by USH2A exon 13 mutations, Mol Ther, 2021 Aug 4
- Lentiviral Gene Therapy Combined with Low-Dose Busulfan in Infants with SCID-X1, N Engl J Med, 2019 Apr 18
- Biologic and Clinical Efficacy of LentiGlobin for Sickle Cell Disease, N Engl J Med, 2021 Dec 12
- CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia, N Engl J Med, 2021 Jan 21
- Gene-Based Therapeutics for Inherited Retinal Diseases, Front Genet, 2021
- Viral vector platforms within the gene therapy landscape, Signal Transduct Target Ther, 2021 Feb 8
- Morpholino - Wikipedia
- Early Phase Clinical Immunogenicity of Valoctocogene Roxaparvovec, an AAV5-Mediated Gene Therapy for Hemophilia A, Mol Ther, 2021 Feb 3
- Valoctocogene Roxaparvovec Gene Therapy for Hemophilia A, NEJM, 2022 Mar 17