The CRISPR revolution is about to hit the public stock markets with the upcoming Editas Medicine initial public offering. They are planning to price 5.9 million shares between $16-$18, which would give them a valuation of >$600M. While CRISPR is pervasive and widespread in its use and different applications in the biomedical research world right now, how to apply it to the clinical medicine is less certain. Toward that mission, though, Editas attempted to lay out its vision for capitalizing on Cas9 technology. Here are our thoughts on their roadshow presentation and their plan to make CRISPR therapeutics. Editas' S-1 can be found here.
How CRISPR works:
Cas9 nuclease is easily targeted to different sites in the genome through direction by a short guide RNA (gRNA) which basepairs to the target DNA sequence, making a double strand break. Note that there are Cas9 mutation variants that can only cut one DNA strand (nickase) or are completely defective in cutting, which are useful in different scenarios. Editas outlined the three basic paradigms of gene editing medicine.
1. Cut & Revise - ex: engineered T cells probably for knockout
2. Cut & Remove - ex: intron sequence from LCA10 protein
3. Cut & Replace - ex: introduce a DNA template to undergo homologous recombination into a target locus. Editing of hemoglobin beta as an example.
To identify which gRNAs induce cuts at the fewest off-target sites, they are using Keith Joung's GUIDE-Seq approach to identify where double strand breaks are occurring in the cells after adding Cas9 + the gRNAs. Editas believes this is the best way to identify potential off-target cuts, and to pick gRNAs that have the fewest or even no detectable off-target cutting, which was shown for a number of gRNAs they mentioned.
Pipeline Summary:
Eye diseases: LCA10 (CEP290) is Editas' lead indication, planning to begin phase I trial in 2017, induce small deletion in an intron, use AAV delivery to the retina (covered more below). Other planned genetic & infectious eye diseases in discovery phase, and the disease mechanism and possibly efficacy remains to be seen pending further disclosures.
T cells - NHEJ (knockout) ex vivo (covered more below)
Hematologic diseases (B-Thal, SCD) - NHEJ or HDR strategies, ex vivo
Muscle (DMD) - small & large deletions (presumably meaning exon skipping - delivery will be the big issue. See recent papers using AAV and CRISPR to induce exon skipping in DMD)
Lung (Cystic Fibrosis) - NHEJ & HDR (see the Vertex & CRISPR Therapeutics have a collaboration - $105M upfront for CF and hemoglobinopathies)
Genetic & infectious diseases of the liver (alpha-1 antitrypsin deficiency) - NHEJ and HDR, likely necessitating in vivo delivery directly to the liver.
In general, Editas believes they can leverage progress made by others in gene therapy, particularly in the aspects of delivery. Viral vectors using AAV seem to be the primary focus right now.
Editas showed a figure concerning the gross aspects of their platform. They don’t seem committed to any specific delivery modality (basically using what has been used previously - ie: AAV vs. nanoparticle vs. ex vivo cells), and said they are doing preclinical work in all of them. Given that delivery is the biggest issue facing CRISPR, one would have hoped for more details ahead of the IPO. They did not present any specific improvements in delivery they've made, except for using smaller Cas9 for AAV (potentially the S. aureus Cas9 as published previously by Feng Zhang.) to fit into the vector genome size.
Editas then showed the ability to edit genes in vivo, targeting the endogeneous Factor VII gene in mice. The exact delivery mechanism was unspecified however (most likely AAV) for the presumably liver targeting approach. The figure reported protein levels down of ~40-65% depending on guide RNA, as measured by ELISA in the serum. It's uncertain that this level of knockdown would be significant enough to modulate any disease to a great extent, although that will largely depend on the target.
 |
| From genengnews.com |
Finally, the competitive advantage over Zinc fingers, TALENs, and meganucleases is not very clear for single targets in genetic medicines, where you need to take time to develop the drug anyway (see the Genome biology figure for a comparison of zinc finger and CRISPR strategies). It is faster to use CRISPR for researchers & easy to design, which explains its widespread popularity, but this advantage is actually nullified in the clinical world. It is possible that Sangamo, Cellectis, and others championing these alternative genome editing platforms might see a resurgence in interest due to CRISPR-Cas9 patent exclusivity. The ability, in vivo, to make multiple double strand cuts using a single AAV vector may be an advantage for CRISPR-Cas9 over the other technologies, useful for the "Cut & Remove" types of edits - exemplified in their lead indication of LCA10 discussed further below.
T cell applications:
We like the potential of the Juno Therapeutics alliance that Editas has established, and that was emphasized many times throughout the presentation. The delivery and editing targets are fairly clear, with translational application direct to improving cancer immunotherapy. For the goals of the Juno collaboration, Editas outlined the need to improve persistence of T cells & overcome inhibitory microenvironment signals. Editas showed they could edit 80-90% of T cells for one target, while only achieving 40% editing of the second undisclosed target. There were also some effects on the viability of the edited cells, with a decrease to 50-60% normal viability, which Editas still considers well tolerated. The first CRISPR T cell application seems to be the knockout, and so it would not be a surprise if they are potentially targeting PD-1 and/or CTLA-4.
We have a couple of concerns though. First, there is risk in Juno being a big winner of the CAR-T cell battle, whose own story will continue to evolve over the next years. We also wonder if the IP will make it fall through on making this license special and exclusive. Other companies are doing genome editing of CAR T cells with different modalities, and some also having CRISPR partnerships. We want to know if they have specific IP on T cell applications, or if their claim is more on the general terms of owning the Cas9 technology. For example, the UCSF paper recently (which had Jennifer Doudna has an author) discussed CRISPR editing in T cells, and UCSF filed a patent on the technology. It is our presumption Editas does not own this due to the ongoing dispute, meaning this will flow to another company. Our fear is that CRISPR will be generalizable to all immunotherapy companies soon, meaning the Juno/Editas advantage is diminished, after which the revenue stream from the deal is solely based on the success of Juno's platform, which Editas has no control over.
We have a couple of concerns though. First, there is risk in Juno being a big winner of the CAR-T cell battle, whose own story will continue to evolve over the next years. We also wonder if the IP will make it fall through on making this license special and exclusive. Other companies are doing genome editing of CAR T cells with different modalities, and some also having CRISPR partnerships. We want to know if they have specific IP on T cell applications, or if their claim is more on the general terms of owning the Cas9 technology. For example, the UCSF paper recently (which had Jennifer Doudna has an author) discussed CRISPR editing in T cells, and UCSF filed a patent on the technology. It is our presumption Editas does not own this due to the ongoing dispute, meaning this will flow to another company. Our fear is that CRISPR will be generalizable to all immunotherapy companies soon, meaning the Juno/Editas advantage is diminished, after which the revenue stream from the deal is solely based on the success of Juno's platform, which Editas has no control over.
Leber's Congenital Amaurosis Type 10 as their First Treatment Target:
Editas will go after Leber Congenital Amaurosis 10 (LCA10) as their first indication. LCA10 is a disease that results in progressive blindness. LCA can be caused by many different mutations, but Editas will focus on the subset caused by CEP290 mutations (LCA type 10), of which there is no approved therapy. Spark therapeutics has used AAV vectors to treat LCA type 2 patients with mutations in a different gene (RPE65) by re-expressing the protein, with positive phase III results recently released for 1 year post-treatment. The mutation in the CEP290 gene is a single base pair mutation in an intron causing inappropriate splicing of the protein. This leads to truncation of protein by causing insertion of an abnormal exon that contains a premature stop codon. The theory is that if one just cuts out that cryptic splice site with 2 guide RNAs (as outlined in the figure below), removing that stretch of DNA, then normal splicing and protein expression will occur.
Concerning the data presented, Editas only has in vitro data from patient cells showing correction. By analyzing RNA in patient derived fibroblasts with LCA, they observed ~1.5 units normal CEP290 RNA, ~4 units mutant CEP290 RNA before treatment. After treatment, with either pair of gRNAs they tested, this caused a flip to ~3.5-4x normal RNA, ~1x mutant RNA (presumably no selection for cells that received Cas9 & guides). By protein, this RNA change approximately doubles the full length CEP290 protein produced in this bulk population of cells. No in vivo modeling of this approach or any phenotypic readout was shown. Indeed, Editas said they were finishing the in vivo work in 2016 in order to file the IND late this year for trials in 2017. While the genetics of the disease and logic of this therapeutic approach appear well understood, showing efficient delivery and editing in vivo will be important to further de-risk this approach.
There are a number of reasons this disease makes sense as a first target for Editas. The disease does not involve changing the coding sequence of the protein, meaning that the normal protein can be restored through deletion and an immune response against normal CEP290 will not happen (see the Editas figure above). The eye is immunopriveleged and should in theory have less or no reaction to the vector. Eye delivery efficiency using AAV is well documented in previous clinical trials (Spark has been successful (mostly) with an AAV to the eye, Avalanche and others have been less successful, but may be due to the AAV or the target), so Editas knows they can theoretically reach the target cells. The disease is in theory monogeneic and has a well defined mechanism. These points are all positives, which led to the selection of the LCA10 indiction.
As far as potential downsides for this target, one concern is how much protein will be needed for therapeutic efficacy. Is double the amount of correct protein enough for disease pathology, or will the presence of the pathologic protein in mixture in corrected cells still cause centrosome and cilia dysfunction and cell death over time?
We also have some questions about the age of treatment, since this disorder leads to cell death in the cones, and the time point of restoration might be crucial in order to rejuvenate them. Secondary injury to the retina could permanently put the cones on a path toward destruction even if the centrosome assembly is saved. We have concerns about their all-in-one vector approach for treating the disease. It's not specified when or how the Cas9 protein expression will ever be turned off, permanently marking those cells as immunogenic potentially. Furthermore, any AAV vector leaking out of the ocular cavity could prime an immune response to Cas9, which might eventually find its way to the eye, in spite of the anatomical barriers. Lastly, we wonder if a gene therapy competitor could simply deliver full length CEP290, although the very large size of the protein may make it less amenable to fitting the full coding sequence in an AAV. This also applies to many other applications of gene editing technology, that there could be competition from more traditional gene therapy approaches. In that regard, the stability of genome editing in host chromosomes is an advantage over episomal driven expression of a transgene from AAV, and while AAV is the current preferred in vivo gene therapy vector, a lentiviral approach could potentially solve this problem. As mentioned above, competition for this indication could come from other gene editing platforms that can cut the same sites that Editas is targeting, however, it may be easier to fit all the machinery to make the 2 different cuts here with Cas9 and 2 gRNAs versus multiple separate zinc fingers or TALENs.
Strengths of Editas Medicine:
The company appears to be the financial and intellectual leader in the CRISPR space so far. In particular, the financial aspect is crucial since they will have the most money to fund initial trials and future trials, as well as purchase more academic IP related to CRISPR technology.As far as potential downsides for this target, one concern is how much protein will be needed for therapeutic efficacy. Is double the amount of correct protein enough for disease pathology, or will the presence of the pathologic protein in mixture in corrected cells still cause centrosome and cilia dysfunction and cell death over time?
We also have some questions about the age of treatment, since this disorder leads to cell death in the cones, and the time point of restoration might be crucial in order to rejuvenate them. Secondary injury to the retina could permanently put the cones on a path toward destruction even if the centrosome assembly is saved. We have concerns about their all-in-one vector approach for treating the disease. It's not specified when or how the Cas9 protein expression will ever be turned off, permanently marking those cells as immunogenic potentially. Furthermore, any AAV vector leaking out of the ocular cavity could prime an immune response to Cas9, which might eventually find its way to the eye, in spite of the anatomical barriers. Lastly, we wonder if a gene therapy competitor could simply deliver full length CEP290, although the very large size of the protein may make it less amenable to fitting the full coding sequence in an AAV. This also applies to many other applications of gene editing technology, that there could be competition from more traditional gene therapy approaches. In that regard, the stability of genome editing in host chromosomes is an advantage over episomal driven expression of a transgene from AAV, and while AAV is the current preferred in vivo gene therapy vector, a lentiviral approach could potentially solve this problem. As mentioned above, competition for this indication could come from other gene editing platforms that can cut the same sites that Editas is targeting, however, it may be easier to fit all the machinery to make the 2 different cuts here with Cas9 and 2 gRNAs versus multiple separate zinc fingers or TALENs.
Strengths of Editas Medicine:
Financials:
$43M Series A
$25M upfront from Juno
$120M Series B
$150M Cash, at least 24 months runway before IPO
There is a plan to release more proof of concept (presumably additional preclinical) data each year at medical conferences for multiple programs. A potential challenge for Editas in the public markets might be investor patience, as they will not enter clinical trials until 2017. To this end, Editas emphasized these proof of concept data releases and potential further business development as important catalysts to keep investors interested in the story and maintaining more visible progress.
Lastly, there was only minor acknowledgement, in an appendix slide, of the upcoming patent interference hearings between University of California (Doudna), University of Vienna (Charpentier) versus MIT/Harvard/Broad patents. Editas has the licenses from the Broad/MIT/Harvard estate, so will be hoping for their victory. The legal case will not be covered here, but it could add significant volatility to the stock when the rulings are being made, although this will play out over a number of years.
Summary:
Editas led off the presentation referencing the 6,000 diseases with genetic mutations 95% and no approved therapy. While enticing, this statement overstates the potential of their platform in targeting all of these diseases, following the general hype in the media around CRISPR. However, there are other options like classical gene therapy and cell therapy, as well as other gene editing technologies that could go after the same indications. It is still uncertain which of these options will ultimately win. A major question for Editas' in vivo applications will be the efficiency of delivery to the target tissue, which appears to currently be piggybacking off of the routes of administration used for other gene and cell therapies. Thus, Editas will also share the same potential and difficulties as gene therapies in terms of delivering sufficient amounts of their therapy to target tissues, and target selection will be key. There is also the elephant in the room of who will really own Cas9 and genome editing? While carrying immense promise, the gene editing story will take some time to begin to unfold, and its full potential will not be known for many years. With those risks stated, the prospect of a dominant CRISPR therapeutic company on the public stock markets is very exciting, with a platform technology that could truly revolutionize medicine if properly executed.
Disclosure: DM owns shares of Sangamo.
