“CRISPR” is the word in the headline that signals that the piece is going to be about gene editing. Doc Gumshoe did not want to put that word – actually, it’s not really a word, but an acronym – in the title of this piece because gene editing is a whole lot more than CRISPR, even though the discovery of CRISPR has hugely facilitated gene editing. I will attempt an accurate definition of CRISPR and the role of those entities a bit further on in this piece, but for now let’s look at the bigger picture.
Gene manipulation in some form has been around for a long, long time – centuries, at least, dating from those times when nobody had the foggiest notion of what a gene was, or how it was that some but not all of the characteristics of the parent were passed on to the offspring. The term “parent” as used here does not refer to a human Daddy or Mommy but to whichever being, animal or vegetable, that gave rise to offspring. When hunter-gatherers selected the largest fruit from trees, brought these prizes back to their huts for dinner, and threw the pits on the dung heap, they were unknowingly practicing genetic manipulation. From those seeds the next generation of fruit trees would grow, producing bigger fruit than their cousins in the woods. Tulip fanatics in Holland in the 17th century were doing genetic manipulation, as were breeders of prized hogs, as well as the agronomists who gave us such mixed-breed fruits as the pluot (a cross between a plum and an apricot) and perhaps the grapefruit (a cross between a sweet orange and a pomelo, which looks like a grapefruit but is quite a bit larger, and more sour). The grapefruit may have emerged accidentally, or it might have been a deliberate creation.
In any case, what the creators of those hybrids (and many, many more) were doing was taking genetic material from different “parents” and putting them together such that the offspring are genetically modified – essentially different from their parents. But they had no idea that all characteristic features of the animals and plants they were working with were encoded in genes, nor yet even of the existence of genes.
The concept of genes and genetics came into being in the early years of the 20th century, although as yet nobody knew what a gene actually was. Chromosomes had been identified as related to inherited characteristics, and had even been observed under the microscope at that time, but what the chromosomes actually were was still unknown. As the capacities of microscopes progressed, it became possible to get a closer look at chromosomes, which led to the discovery in 1953 by Francis Crick and James Watson of the configuration of the DNA molecule in the chromosome, i.e., the famous double helix. DNA in turn was found to be composed of a linking of four amino acids, termed “bases” (adenine, cytosine, guanine and thymine plus a phosphate group and a pentose sugar). The human genome consists of about 6.2 billion bases, paired and linked together in a very long twisting chain, which itself is wound around itself so as to take up minimum space. The genetic information is carried by these 3.1 billion base pairs, and the possibility of effecting changes to that genetic information, such as eliminating the genes that carried the information that would cause diseases or disabilities was immediately the subject of research.
The information carried in the genome might be compared with the information embodied in computer code. As by now everyone knows, computer code as read by the computer consists of strings of ones and zeros, which signify yesses and nos, impulse transmitted or impulse blocked. The human entering the code does not have to translate the numbers in our decimal system into the binary system of ones and zeros – the computer takes care of that. There doesn’t seem to be any limit to the amount of information that computer code can convey, but those strings of ones and zeros can get pretty long.
Genetic information is conveyed by that string of amino acids. As in computer code, the information is essentially digital, except that instead of two digits to work with, the genome can use four digits, for the four amino acids – A, C, G, and T. Thus, it can carry a good deal more information than computer code, in a limited space.
Tinkering with the genome in a way that would provide health benefits became a goal of scientific and medical research. One could call it the “holy grail.” But it was far from simple. A couple of early attempts to do actual gene editing were exceedingly complicated and difficult. One method relied on what were termed “zinc finger nucleases” (ZFNs). These would have to be designed specifically to recognize and bind to specific fairly short base-pair sequences, and sometimes editing a single short sequence would require the design and construction of three ZFN domains; longer base-pair sequences would require more elaborate ZFN domains.
The finding that opened the path to what we now call gene editing was the discovery that in those long strands of DNA, the gene sequences were separated by clusters of sequences that did not appear, at first glance, to convey any genetic information at all. We might call them nonsense sequences. They followed a certain pattern – they read the same way from either end: i.e., they were palindromic. They occurred at repeated intervals in the DNA strand. Thus, they came to be called Clustered Regularly Interspaced Palindromic Repeats – “CRISPR”. These sequences were discovered in 1980 and were originally dismissed as being of no interest – “junk DNA.”
It turned out that the CRISPR sequences, along with a CRISPR-associated (CAS) enzyme are of vital importance in conferring immunity to repeated infection by a virus. This was discovered in research on bacteria, which also (of course!) have CRISPR in their genome. When a virus invades a bacterium, the CAS enzyme snips off pieces of the viral genome, and stores it in the CRISPR part of the bacterial genome so that the bacterium essentially remembers the virus that attacked it so that it can mount a defense. A number of these enzymes have been discovered, but the one – CRISPR-Cas9 – that has up to this point been found to be most useful in gene editing was discovered in 2012 by Jennifer Doudna of the Broad Institute (Harvard / MIT) and Emmanuelle Charpentier of the Max Planck Institute for Infection Biology in Berlin.
CRISPR-Cas9 has been central in gene editing in the past seven years. The Cas9 enzyme makes it possible to “snip” entire gene sequences out of the genome without altering the gene itself. The Cas9 enzyme is guided by RNA, which has the capacity to locate specific DNA sequences. The enzyme binds to and breaks the CRISPR cluster on each end of the genetic material in a way that permits the genome to link together without a permanent break, resulting in the deletion, or “knock out”, of that gene. The word “snip” is of course used figuratively; there is no mechanical instrument that cuts the CRISPR cluster. The binding action of the enzyme separates the individual particles in the cluster. And the binding action is possible because of the configuration of the enzyme and the electric potential of its components.
What can gene editing actually accomplish?
The dream would be to cut out the genes that doom people to a number of diseases, such as amyotrophic lateral sclerosis (ALS), sometimes known as Lou Gehrig’s disease. A person one of whose parents has the ALS gene has about a 50% chance of developing the disease, which has a life expectancy of less than five years after onset of symptoms. In about 10% of ALS cases, the cause is a genetic factor. And a number of other diseases have a major genetic component. For example, cystic fibrosis, Down syndrome, Duchenne muscular dystrophy, hemophilia, Tay-Sachs disease, Tourette syndrome, and many others.
Genetic manipulation in the interest of protecting people from disease may not be as clearly a good thing as we might assume at first glance. For example, take sickle cell anemia, an inherited genetic disease, which is the fourth leading cause of death in children in many developed countries. Sickle cell presents a contradiction. One would expect natural selection to weed out a gene that has such dire consequences, but that isn’t happening. About 4.5 million people currently have the disease, and another 43 million have the sickle cell trait, but have not shown signs of the disease. A possible reason for this is that persons having the sickle cell trait – a single gene – have an approximately 30% lower probability of developing malaria, which is a major killer in the parts of the world where sickle cell is prevalent. So if scientists were able to eliminate the sickle cell gene, is it not likely that the number of malaria cases – and deaths – would zoom?
That tells us that tinkering with the human genome is not an area that should be entered blithely. There was a great deal of public controversy after the announcement, about a year ago, that a scientist in China had performed gene editing on a human embryo such that the babies, twin girls, were born with an edited gene that was supposedly associated with a much lower risk of HIV. Opinion on the question of actually altering the human genome was starkly divided. One faction asserted that it must never be done under any circumstances, pointing to the possibility of errors that would then be perpetuated in the offspring of those “designer babies.” Another possibility that was met with much suspicion and antagonism was that these “designer babies” would actually be superior beings – physically, intellectually, or in some other way – and grow up to lord it over the rest of us. But a contrary argument was that if it were indeed possible to edit out the genes that greatly increased the risk for a number of disabling or deadly diseases, enacting laws that deprived humans of that possibility might be considered equivalent to outlawing vaccinations. The possibility of error, however, hovers over the entire field of gene editing. How do we know, for example, that cutting out the gene that is associated with an increased likelihood of Down syndrome does not also have another, perhaps even more unfortunate result?
Other gene editing targets
There doesn’t appear to be any great ethical concern regarding genetic engineering when the subjects of the process are mice. For about the past 30 years, genetically-altered mice have been used in scientific research – it’s very handy to have a supply of mice whose genes have been altered in such a way as to make them especially likely to have the diseases that are being targeted in humans. In any number of studies of important drugs, the candidate drug is evaluated in transgenic mice prior to any kind of study in humans.
Initially, efforts to manipulate the genes of mice employed chemicals or viruses. These methods were clumsy, and resulted more often than not in random mutations that were of little clinical use. Genuine gene editing, first using the above-mentioned zinc finger nucleases (ZFNs) and transcription activator-like effector (TALE) nucleases, and later using the CRISPR-Cas9 system, have resulted in much more efficient and specific gene editing.
Engineering our own immune response
There is a constant war raging in our bodies between our immune system and harmful invaders of all kinds, whether these are microbes, viruses, foreign substances that don’t belong inside our bodies, and also our own cells when they mess up their own reproductive processes and become cancerous. The chief patrolling agents are T-cells, which are highly effective against a wide range of enemy agents, which they attack and kill. T-cells are equipped with receptors on their surface that permit them to recognize our own healthy cells and instruct the T-cells to leave those cells alone. These surface receptors, called immune checkpoints, act as brakes, so that the T-cells limit their attacks to the dangerous invaders. But cancer cells, with their own survival at stake, have developed high levels of proteins that also attach to those checkpoints, essentially fooling the T-cells into mistaking some cancer cells for normal cells, which then sneak through the protection of our immune response.
As has been reported in previous Doc Gumshoe sermons, those survival mechanisms in cancer cells had been an important target of drug development for 25 years before the approval of the first drug that addressed the checkpoints on T-cells. That was ipilimumab (Yervoy), from Bristol-Myers Squibb, which received FDA approval in March 2011 for the treatment of malignant melanoma. Several other drugs followed, including some of the most successful (and hugely profitable) cancer drugs, including nivolumab (Opdivo, also from BMS), and pembolizumab (Keytruda, from Merck).
A CRISPR CAR-T case history
Along about the time that Yervoy, Opdivo, and Keytruda were getting FDA approval, the first signs emerged that a treatment involving genetically modified T-cells could be effective in some forms of cancer. One of the first patients to be treated in this way was a seven-year-old girl who had relapsed acute lymphoblastic leukemia (ALL). The majority of children with ALL achieve complete remission after about two years of chemotherapy, but Emily Whitehead was in that 15% who were resistant to even the most intensive chemotherapy.
She was initially treated with chemotherapy, and when that first round of treatment failed, her parents took her to the Cancer Center of the Children’s Hospital of Philadelphia (CHOP), where the oncologist recommended a second round of chemotherapy for Emily. When she failed that second round, in February 2012, her parents agreed to enroll Emily in a clinical trial of advanced experimental T-cell therapy for advanced B-cell leukemias and lymphomas.
The therapy consisted of collecting Emily’s T-cells and modifying them genetically to attack the leukemic cells, termed B-cells. Emily’s T-cells were harvested and treated with an agent that incorporates a gene that encodes for a specific antigen receptor which recognizes and attaches to a protein called CD19 that is found only on the surface of those B-cells. The engineered T-cells were then reintroduced into Emily’s system, where they spread throughout her body and began to find and kill the cancerous B-cells. Genetically engineered T-cells are known as chimeric antigen-receptor modified T-cells, thus CAR-T and CRISPR CAR-T, since the CRISPR sequences are what permit the T-cells DNA to be edited.
Emily’s parents had been warned that the treatment was not without risk. They had expected that Emily would experience flu-like symptoms, but what took place was a good deal more serious. She was admitted to pediatric intensive care where it was learned that the growing presence of T-cells in her body gave rise to the proliferation of a cytokine that is involved in rheumatoid arthritis. The cytokine, interleukin-6 (IL-6), is customarily treated in patients with rheumatoid arthritis by tocilizumab (Actemra, from Genentech). Emily was treated with tocilizumab and her condition improved markedly. Almost overnight, her breathing improved, her fever dropped, and her blood pressure was back to normal. In the ensuing weeks, Emily complete recovered from her adverse reaction to the T-cell infusion, which is called cytokine release syndrome.
The doctors who were supervising her treatment were still unsure whether her underlying ALL was being successfully addressed by the T-cell treatment. Several weeks after the initial infusion, a bone marrow test was performed to assess the results. The medical team concluded that Emily was in complete remission. The bone marrow test was repeated after another three-month interval, and then after six months, and she had no signs of the disease. At the present time, five years after the T-cell infusion, she continues to be entirely disease-free, and the descendants of the original genetically-modified T-cells are still circulating in her body, armed and ready to attack any signs of cancer.
CRISPR CAR-T treatment targeting the CD19 protein has been used in about 150 patients with relapsed or refractory ALL, with mostly positive results; however, research continues on the best treatment modalities for this disease. A strategy that is under consideration for patients at high risk for refractory ALL is to harvest a supply of T-cells in advance, so as to have them ready for gene editing and reinfusion if the need arises.
Drugs using the CRISPR CAR-T mechanism
Two drugs based on CRISPR CAR-T have at this time received FDA approval. Both target the CD19 protein, which is found on many leukemic cells, called B cells. The first was Kymriah (tisagenlecleucel, from Novartis), which is approved for the treatment of relapsed/refractory B-cell precursor acute lymphoblastic leukemia, the form of cancer that Emily had. The second was Yescarta (axicabtagene ciloleucel, from Kite Pharmaceuticals), which is approved for the treatment of relapsed/refractory diffuse large B-cell lymphoma.
I should point out that in the CRISPR CAR-T world, the drugs are not directly given to the patients. What happens is that, as in the case of Emily, the patient’s T-cells are harvested and exposed to the chimeric antigen receptor (CAR) which is derived from mice and specifically targets the CD19 protein . That mechanism is what makes the T-cells able to find, attack, and destroy the cancer-causing B-cells. Then, the T-cells are reinfused into the patient’s body where they multiply and perfuse the patient’s circulatory system. The medical team then crosses its fingers and waits for the edited T-cells to do their work, which they mostly do.
The success rate for CAR-T therapy is a bit equivocal. The initial remission rates are excellent, 90% or higher. However, long term survival is another matter. What happens is that not all the leukemic cells express CD19 – a few express other cytokines, and after the CAR-T therapy has eliminated those B-cells with the CD19 protein, other B-cells start multiplying and spreading.
To attempt to combat this, research is ongoing on the possibility that T-cells could be marked with other antigens, such as the BCMA protein, which is commonly expressed in multiple myeloma, or other antigens such as CD20. There are also efforts underway to engineer CARs targeting many other blood cancer antigens, including CD30 in refractory Hodgkin’s lymphoma; CD33, CD123, and FLT3 in acute myeloid leukemia ; and BCMA in multiple myeloma.
A promising new CRISPR CAR-T drug from Celgene?
The drug in question, lisocabtagene maraleucel (JCAR017), being called lisocel for now, is being developed by Juno Therapeutics in collaboration with Celgene. A clinical trial in 262 patients with large B-cell lymphomas posted promising results, or, at least, promising in patients with what has been regarded as a deadly form of cancer. Among 255 subjects whose outcomes were evaluable for efficacy, the complete response rate was 53%. Overall, 73% of the subjects in the trial demonstrated a response, and the median duration of the response was 13.3 months.
These response rates with lisocel were 20 percentage points higher than those attained with Kymriah, and just slightly higher than those with Yescarta. (We need to note that the comparison in response rates between different clinical trials is at best inexact; this was not a head-to-head trial where the participants had comparable baseline status and received uniform treatment.)
Although lisocel and Yescarta were similar in terms of efficacy, lisocel’s safety data may give it a significant edge over Yescarta. Kite had announced that Yescarta had an 87% rate of neurotoxicity and a 13% or higher rate of grade 3 cytokine release syndrome (CRS), which, as you read above, is an adverse event threatening the CRISPR CAR-T treatment area. Patients treated with lisocel, in contrast, had a 30% rate of neurotoxicity and a 2% rate of CRS.
The Juno/Celgene collaboration also announced that the process of manufacturing lisocel is well under control. The average time from the separation of white blood cells from the patient and extracting the required T cells to the time when the T cells had been genetically engineered to become CAR-T was just 24 days. This is a highly important factor in patient treatment, since delay in initiation of treatment can be dangerous in the management of aggressive disease.
Celgene needs to win FDA approval of lisocel by the end of 2020 to trigger a $9 per share payout, to which they have committed under the terms of Celgene’s acquisition by Bristol-Myers Squibb.
Another strategy to increase the effectiveness of CRISPR CAR-T
As Doc Gumshoe has pointed out earlier in this piece, T-cells are equipped with surface receptors that recognize our own healthy cells and instruct the T-cells to leave those healthy cells alone. Those surface receptors are called immune checkpoints, and they act as brakes. Cancer cells, in their perpetual battle against our immune system, have developed high levels of proteins that attach to those checkpoints and deflect the attack from T-cells.
Currently, some of the most successful cancer drugs such as Yervoy and Opdivo (Bristol-Myers Squibb), Keytruda (Merck), Tecentriq (Genentech), and Imfizi (AstraZeneca) affect T-cell checkpoints and permit T-cells to go after the cancer cells. The checkpoint proteins targeted are labeled PD-1 and PD-L1.
At the Society for Immunotherapy meeting in November, it was announced that CAR-T engineered cells to which anti PD-1 molecule subunits were added demonstrated highly positive results. In a preliminary trial, 12 of 13 patients with relapsed/refractory non-Hodgkin’s lymphoma responded to treatment for an overall response rate of 92.3%. Seven of those 12 responders achieved a complete response.
The molecule added to the T-cells was a dominant-negative PD-1 molecule, which has lost some of its functionality, but still possesses effectiveness versus the endogenous PD-1 molecule. That molecule was added to the T-cell structure along with the CD19 targeting molecule.
The trial discussed above was conducted in China. The clinical development program now plans to focus on solid tumors, including thyroid, breast, pancreatic, prostate, and colorectal cancers.
As I am writing this, there are 640 clinical trials involving some aspect of CRISPR CAR-T under way or recruiting patients.
Improvements over CRISPR Cas9 in the works
I’m assuming that I am not the only one of us who was initially confused by those appellations; i.e., what’s the difference between CRISPR CAR-T and CRISPR Cas9?
CAR-T stands for chimeric antigen-receptor modified T-cells and refers only to the process through which T-cells are modified so that they overcome the defenses of cancer cells. Cas stands for CRISPR-associated enzymes, and there are a number of these, Cas9 being the first one pressed into service for gene editing. The Cas enzymes are what permit the CRISPR weapon to be used in deleting genes, by cutting through those palindromic repeats that separate the genes, and are also used in the process of inserting new genetic sequences in the DNA strand.
Catherine Freije and Cameron Myhrvols, researchers at the Broad Institute’s Infectious Disease and Microbiome Program are using a different protein that targets RNA rather than DNA. The protein, labeled Cas13, permits CRISPR to recognize and target infectious diseases that use RNA as their genetic material. The two viruses that they first targeted are Zika and dengue, both of which have presented great challenges not only to treatment, but also to diagnosis. Using Cas13, they designed a diagnostic procedure that was able to correctly identify both the Zika and the dengue viruses from a patient sample, with minimal sample processing. The next step that they are working on is using CRISPR Cas13 to attack those viruses, and to adapt the method to other pathogens.
The Cas13 protein is employed in a new system that may one day be used both to diagnose and treat a large number of infections, caused both by known and by emerging viruses. The system has been called CARVER for Cas13-Assisted Restriction of Viral Expression and Readout. The Broad Institute’s researchers identified thousands of sites in the genetic material of hundreds of viral species that would be effective targets for CARVER. Of the many possible viruses, the researcher’s experimentally tested the activity of the Cas13-based system on three: influenza A virus, vesicular stomatitis virus, and lymphocytic choiriomeningitis virus. They first introduced Cas13 and guide RNA into human cells, and then exposed those cells to the viruses. Within 24 hours, the CARVER system had reduced the level of viral RNA in the cell cultures by a factor of 40. The effect on viral infectivity was even more pronounced – within 8 hours, the system using Cas13 reduced the infectivity of the flu virus by a factor of 300.
Neither CARVER nor Cas13 have yet been used to treat a human being with those viral infections, and I have not seen predictions as to when they will enter mainstream medicine, whether for diagnosis or treatment. However, the science is moving fast, and a great deal has been learned in the two years since the Cas13 protein was identified.
… and yet another gene editing tool is created
This one has been called PRIME, and news about PRIME is what you might call “hot off the press.” It was first made public at a meeting at Cold Spring Harbor Laboratory, in October of this year. The principal investigator in the development of PRIME was David Liu, also of the Broad Institute. The chief difference between PRIME and CRISPR-based methods is that while CRISPR relies on the ability of cells to divide to help make the desired changes in DNA, PRIME does not require that trait to function. That means that PRIME could be used to correct genetic mutations in cells that do not commonly divide, such as those in the nervous system. Many diseases, such as Parkinson’s and Huntington’s, are caused by mutations in nervous system cells.
In addition, the new method doesn’t cut both strands of the DNA double helix, minimizing the chances of making unintended changes that could be dangerous. David Liu describes the method as follows: “Prime editors are more like word processors capable of searching for targeted DNA sequences and precisely replacing them with edited DNA strands.”
According to PRIME’s inventors, it has the potential to correct 89% of known disease-causing genetic variations in DNA, from the single-letter misspelling that causes sickle cell anemia to the superfluous four letters that cause Tay-Sach’s disease. They report making 175 edits in mouse cells and human cells.
PRIME editing improves on CRISPR-Cas9 as well as on the improvements that have been made to that system. For example, one of the CRISPR-Cas9 functionalities, base editing, can change only four of the possible amino acid bases – C (cytosine) to T (thymine) and the reverse, T to C; also A (adenine) to G (guanine) and the reverse. PRIME editing can make all 12 possible base changes.
PRIME editing is thought by genetic engineers to be capable of repairing point mutations that are the cause of about 7,000 inherited genetic diseases.
Where do we go from here?
Gene editing definitely has enormous potential to treat, or at least address, a huge range of diseases. Researchers at Western University in London, Ontario, have proclaimed that it is the next step in antimicrobial therapy, and pointed to the capacity of CRISPR-Cas9 to eliminate a species of antimicrobial-resistant Salmonella enterica. They employed a small packet of genetic material (a plasmid) that they tucked into E. coli bacteria, which would then transfer that packet to the Salmonella. The Salmonella would then be unable to reproduce and cease to exist. This would take place without harm to the beneficial bacteria in the patient’s biome. Maneuvers of that type could single out the specific pathogens while leaving the remaining bacteria alone.
The major reservation that applies to gene editing as therapy is that it will likely never be inexpensive. Therapy will need to be tailored to each individual patient, whether it’s by harvesting the patient’s T-cells, letting them reproduce in Petri dishes, teaching the T-cells to attack specific cancer cells through gene editing, and then reinfusing the T-cells (i.e., CRISPR CAR-T), or collecting E. coli from the patient as in the example above, making genetic modifications in the E. coli, and employing those tinkered E. coli to act as the antibiotic.
It doesn’t appear as though any of these versions of gene editing could ever be reduced to a single tablet or injection that is directly applicable to patient treatment in an “off the shelf” formulation. Gene editing provides a method of creating effective therapies for patients one at a time. The engineered T-cells that saved Emily’s life will do nothing for Peter or Paul. In each case the procedure will be expensive.
However, what we have to bear in mind is that 35,000 Americans died in 2018 from infections caused by pathogens against which no antimicrobials had any effect. Similarly, although cancer death rates have greatly declined, many people in all parts of the world die of cancer, and gene editing could have a large effect on cancer, both by editing the genes that make a person more susceptible to cancer and by attacking the cancers themselves.
I doubt that the U. S. Preventive Services Task Force will any time soon be recommending gene editing procedures as treatment for cancers, infections, or inherited susceptibility to serious ailments, but the horizon of possible benefits is very wide. I look forward to learning what happens in that area, and I’ll keep you in the loop.
* * * * * * *
Doc Gumshoe, as always, welcomes your questions and comments, so do please keep them coming. Best, Michael Jorrin (aka Doc Gumshoe)
[ed. note: Michael Jorrin is a longtime medical writer who shares thoughts and news on medicine and health with our readers a couple times a month. He is not a doctor, and does not generally write specifically about investments, but he has agreed to our trading restrictions. You can see his past articles here.]