CRISPR Comes Back to Life
Baby KJ, LNPs, and Danaher
CRISPR has gotten some renewed attention recently, most notably due to the incredibly hopeful story about its use in treating a baby boy with a genetic liver disease (and partly due to Intellia Therapeutic’s reported adverse event for a candidate in phase 3 trials). Baby KJ was born with a severe CPS1 enzyme deficiency, which causes a buildup of ammonia in the body. This buildup is toxic and has a 50% survival rate for those who present symptoms in infancy. Ahead of a liver transplant, the initial treatment for newborns is medication and a very low protein diet (protein leads to elevated ammonia production). In a matter of months, a team of researchers created an individualized CRISPR treatment for KJ. He’s now received the treatment 3 times, eats a regular amount of protein, and is on a substantially lower dose of the initial medication.
It's worth understanding CRISPR and the current gene therapy landscape generally to appreciate how cool of an advancement this is. The CRISPR-Cas9 mechanism was initially discovered in bacteria, and serves to protect these bacteria against harmful foreign invaders. Like humans, bacteria have an immune system that responds to external threats. Part of the way an immune system functions (and why vaccines are helpful) is by learning from previous infections. The polio vaccine delivers an inactivated form of the polio virus, which causes my immune system to develop antibodies to fight the disease. Should I somehow then get exposed to polio, these antibodies recognize the virus, spring into action, and prevent serious illness.
In bacteria and archaea, CRISPR-Cas9 works somewhat analogously. When a bacteriophage (a virus that infects bacteria) injects its DNA into bacteria, an enzyme complex cuts out a short piece of the phage’s DNA and inserts it into a CRISPR array. This DNA segment is transcribed and forms a portion of guideRNA (gRNA), which works with the Cas9 enzyme to identify bacteriophages that have invaded in the past. The next time invasion occurs, the CRISPR-Cas9 complex recognizes the phage, and Cas9 makes a double-stranded break in its DNA, rendering the virus ineffective. In other words, the CRISPR-Cas9 complex is akin to an actually effective defense coordinator when the Eagles are on fourth and inches. Based on historical data, the DC recognizes that they’re going to use the tush push. One in existence just hasn’t yet figured out how to make the equivalent of a double-stranded break and prevent the first down.
In the case of the bacteriophage, once this double-stranded break occurs the phage is ideally further degraded and destroyed. If the effects of CRISPR-Cas9 stopped there, there wouldn’t be much to get excited about with its potential applications for treating disease. If there’s a method that definitely doesn’t work for fixing a faulty gene, it’s making a double-stranded cut and hoping the cellular repair machinery doesn’t rejoin it.
What makes CRISPR-Cas9 so intriguing is threefold:
1. The innate DNA repair mechanisms that spring into action after Cas9 makes its double-stranded break.
2. The modifications that can be made to guideRNA. Researchers can modify gRNA to direct the CRISPR-Cas complex towards a precise section of human DNA. This can either be a segment of DNA with a deleterious mutation, or a segment that’s responsible for up/downregulation of a gene. The ability to design gRNA sequences means one could theoretically use CRISPR to fix any disease caused by a DNA mutation.1
3. The modifications that can be made to the Cas9 protein. For treating some diseases the double-stranded break is helpful. In other cases we want the CRISPR-Cas9’s targeting ability, but don’t want it to make a cut.
Put simply, this CRISPR-Cas9 complex that evolved to protect bacteria from foreign invaders can be leveraged to treat and (hopefully) cure many human diseases.2
In the case of humans, the repair mechanism from point (1) that will most often spring into action is non-homologous end joining.3 This rejoins the double-strand, but imperfectly. An imperfect rejoin is incredibly helpful when the aim is to switch off, or knockout, a gene. This is the approach Vertex and Crispr Therapeutics took when developing CASGEVY, the only CRISPR treatment that has (thus far!) been FDA approved for a broad patient population. Casgevy is used to treat sickle-cell disease (SCD) and transfusion-dependent beta thalassemia (TDT), both of which are inherited blood disorders. Its design was based on the following observation:
“In most patients with SCD or TDT, HbF [fetal hemoglobin] disappears in infancy, at which point the symptoms of the disease begin to manifest. However, some patients have elevated levels of HbF that persist into adulthood, a condition known as hereditary persistence of fetal hemoglobin, or HPFH. These patients are often asymptomatic or experience much milder forms of disease because elevated HbF compensates for the defective adult hemoglobin.”4
Vertex/CRISPR’s strategy is based on upregulating (producing more of) the amount of fetal hemoglobin in a patient. To achieve this, CRISPR/Cas9 makes a double-stranded break in a region of the BCL11A gene, a gene responsible for repressing fetal hemoglobin transcription as one ages. The double-stranded break is repaired through non-homologous end joining, which then means the gene is unable to act as intended and fetal hemoglobin production is increased. This gene knockout approach is especially helpful when the aim is to treat more than one disease. Increased fetal hemoglobin helps those with both SCD and TDT, so you’re both helping more people and can amortize the R&D spend across a larger patient population. This isn’t a luxury Bluebird Bio enjoyed, which meant it was developing separate therapies for sickle cell and beta-thalassemia.
The CRISPR-Cas therapy to treat KJ went a step further, and is similar in philosophy to some ongoing clinical trials leveraging CRISPR (see Beam and Verve for further examples). As mentioned above, the gene knockdown approach depends upon the broken double-stranded DNA being rejoined imperfectly. There wouldn’t be a whole lot of value to the method if the break was then perfectly corrected! This isn’t a method, however, that comes in handy when the goal is to actually fix the mutation rather than just knock a gene out of commission. Fixing the mutation requires first altering the CRISPR-Cas9 complex so that it can no longer make double-stranded DNA breaks. Next, the CRISPR complex is modified to include deaminase, a type of enzyme that can alter a single nucleotide.
It's easy at this stage to get bogged down in overly scientific jargon and have your eyes glaze over. There are a few critical pieces to keep in mind:
DNA is a double-stranded helix composed of 4 different nucleotides: Adenine (A), Thymine (T), Guanine (G), and Cytosine (C). The double-strands of DNA are held together by hydrogen bonds between nucleotides on opposing strands. A pairs with T, and G pairs with C. In the case of KJ, he had an error in his genetic code where what should have been a cytosine was instead a thymine. Consequently, he had an A:T base pair rather than a G:C one.5 A point mutation like this is where a deaminase comes in handy.
In its natural state, CRISPR-Cas9 does not allow precise edits to one’s DNA. Achieving specific nucleotide edits requires modifying the complex and adding a deaminase enzyme. This approach is termed “base editing.”
This base-editing method is what was used to treat KJ. The CRISPR-complex was modified to include an adenine base editor, which as the name suggests edits an adenine base and turns it into a guanine.6 Thus, an A:T base pair at the variant location becomes a G:C one. The therapy was delivered intravenously and contained within a lipid nanoparticle (LNP), an equally vital part of the equation. I’ve discussed this before, but delivery vehicles are a vital component of getting gene therapies right. The right therapy with the wrong delivery mechanism means you don’t have a functioning therapy at all (see Verve’s recent announcement for an illustration of this). The individualized nature of the treatment for KJ made LNPs an especially good choice: the challenge with gene therapies delivered via AAV vector is you can only dose the patient once. Attempting to dose the patient again will drive the immune system into total overdrive. LNPs do cause an immune response, but it’s not severe enough that treatment is limited to one dose. Doctors didn’t know at the outset how much of the treatment KJ would need, so the LNP gave them much needed flexibility.
Thus far KJ has been dosed three times and shown meaningful improvement. At the time the New England Journal of Medicine paper was released, he was eating a regular amount of protein and on a much lower dose of medication. That the treatment method has (so far) worked is an important validation for a whole basket of CRISPR companies, many of which have experienced a tough few years in the aftermath of Covid (Beam, CRISPR Therapeutics, and Intellia are down 85%, 80%, and 95%(!) from their respective 2021 highs.) It’s also a good look for Danaher, whose companies Aldevron and IDT Technologies were both used in the development of KJ’s treatment. Aldevron was acquired in June 2021 as a bet on the growth of gene and cell therapy, which then fairly promptly entered a slump. Its involvement is at least evidence that Danaher bet on the right horse. Acuitas Therapeutics, which provided the LNP for the treatment, presumably also experienced a post-Covid drop in its business. Its partnership with Pfizer/BioNTech on the mRNA vaccine, and now with the treatment for KJ, indicates the organization is still poised to benefit from the theoretical growth of nucleic acid therapies in the future.
Disclaimer: The information in this post is not intended to be and does not constitute investment or financial advice. You should not make any decision based on the information presented without conducting independent due diligence
For those in the weeds on the science, you might object that this statement should be modified to any disease caused by a DNA mutation that is close to a PAM site. I am assuming scientists continue to make progress in designing a Crispr-Cas complex that doesn’t require a PAM site!
Cas9 actually isn’t the only enzyme that can work with CRISPR to make edits in DNA. Cas12a is another enzyme scientists are exploring using.
A second repair mechanism that less frequently kicks into gear is homology directed repair, which is much more accurate than NHEJ but requires a correct template strand of DNA to make the fix. There are companies trying to leverage this method to deliver a functioning copy of a gene to patients, but it’s much harder than a gene knockdown approach.
Page 5, CRISPR Therapeutics Annual Report.
There was also a genetic variant on the maternal allele (the copy of the gene that came from the mother) where what should have been a G became a T. However, the C to T mutation was focused on because it had been reported in a previous case of CPS1 deficiency.
There is technically an intermediate step here where adenine first becomes inosine.