GSK acquisition of TSRO is very value-destructive and weakens faith in new GSK management. For $5bn, a drug that is a 50/50 to reach $500 million is quite the price tag.
GBT is probably a better long “up here” than before. The risk of an approval is gone. Commercial risks remain with Novartis’ drug seemingly superior in efficacy. A combination study would be wise. The market is big enough for both, though. I see GBT trading over $100 sooner or later.
DEEP DIVE – CELGENE – PRICE TARGET: $91.00
Revlimid (lenalidomide) is 2/3rds of Celgene’s revenue, with $10 billion in revenue expected for this year. The composition of matter patents on Revlimid expired (the IND for Revlimid was filed in 2000), as Revlimid is simply a derivative of Thalomid (thalidomide). We will spend some time on the indole-pyridine scaffold of thalidomide. While the structure of Revlimid is very similar to Thalomid, the drug does have some advantages over its predecessor. Nevertheless, nearly 30 patents which “protect” Revlimid will likely be found invalid or not infringed by would-be generic entrants. The first entrant, NatCo (partnered with Teva), has settled with Celgene for a partial distribution deal starting in 2022. The strongest patent is the ‘800 polymorph patent, expiring in 2027. NatCo is allowed a full launch in 2025, with their 2022 “low-single-digits” participation in Revlimid growing to one-third in that year. This would have been a fairly good deal for Celgene if no further ANDAs were filed. But when you have the world’s second best selling drug, you can count on competition. Dr. Reddy’s is at the plate as we speak, with expert discovery concluding in the near future and a trial likely for late 2019 or early 2020. More importantly, their 30-month stay will expire in late 2019. So, how strong is a polymorph patent? I initially felt that Celgene would have a very low likelihood of prevailing at trial, and their Natco settlement indicates weakness. Further research revealed polymorph patents do occasionally prevail, and I believe Celgene has a roughly 50-50 chance of having the patent upheld. The details of the ‘800 fight are beyond the scope of this review, but take a look at the docket and some case law in polymorph patents–many white papers are available. After Dr. Reddy’s, 5 more ANDAs have been filed: Zydus, Cipla, Lotus, Apotex and Sun are all waiting in the wings.
Ultimately, there isn’t enough Revlimid to go around for Celgene and the generics. The Natco settlement worked for one company, and there is enough room for Dr. Reddy’s, but ultimately Celgene will not be able to settle every generic as each subsequent filer finds each subsequent settlement offer less attractive than a trial. Imagine being the fourth ANDA here–do you really want the 3.5% of Revlimid starting in 2024 and up to 2025 but nothing if the patent is overturned in the future? Someone will break ranks and go to trial and overturn what is probably a flimsy patent. So, I have 2019 revenue of Revlimid at $11.5b, 2020 at $12.7b and 2021 at 50% of that: $6.3b, 2022 at $3.5b. Someone launches in 2021 is my best guess. One year later adds roughly $10 per share, so risk is weighted to the upside (probably) here. The three remaining points on Revlimid are the success of generics, replacement by the IMID portfolio and recent data.
Revlimid generics may not do so well, commercially, from the outset. Somewhat like the dynamics of a biosimilar, I predict that the first generic entrants for Revlimid may not find the marketplace too easy. Revlimid’s REMS program has a lot of mindshare with doctors and their assistants–switching to a Dr. Reddy’s program may not be facile and the near 100% generic switch rates we see with oral solids may not take place here. See Clozaril and Accutane for historical reference.
Revlimid is still alive, with the AUGMENT data showing remarkable efficacy. AUGMENT is actually a bad thing, I think, for Celgene, as it takes the wind out of potential NHL data for avodomide and iberdomide, the two named “IMID” follow-ons. If I wanted to transition revenue to those drugs, I’d delineate and differentiate them from Revlimid. Instead, we see Revlimid is potent in R/R NHL with Rituxan. So, where do we go with iberdomide and showing how it is different from generic Revlimid? It may not matter as clinical data you can see is better than hypothetical differences, but if I had my druthers I’d have saved non-MM/MDS for the follow-ons. I have no revenue at all in my model for the next-generation IMIDs. Part of the reason for that is they share the indole/pyridine scaffold. They are the same-old structure of thalidomide, which is really disappointing. A direct cereblon modulator should be doable at this point, and even Novartis has created such molecules. Again, I’m being conservative, but perhaps for good reason.
Pomalyst is also a thalidomide derivative and is relying on a polymorph patent. I have it going away in 2023. It is remarkable that practically the entire company is disappearing in a few years. The pharmaceutical world has never seen such a dramatic patent cliff combination in its history.
Otezla is also a thalidomide derivative, with a sulfonamide decoration which makes it a PDE4 inhibitor. It also relies on a weak polymorph patent, and there are many ANDAs on file, just like thalidomide, lenalidomide and pomalidomide. I suspect generics will enter in 2023, if not sooner.
Abraxane is going generic soon too. Like I said, the entire company disappears in a few years.
So, to counter the record large simultaneous patent cliffs, you have to build the world’s best pipeline, right? Celgene has tried their best, but nothing will replace the nearly $20 billion in peak sales that will be lost to generics. Ozanimod, luspatercept, lis-cel, BCMA CART, and fedratinib make my model and only reach about $6 billion in revenue for “the new Celgene”. This is still enough, as its growing and promising revenue that is conservatively forecast (could be $10 billion if everything goes right). But make no mistake, there is absolutely no way Celgene survives the patent cliff as we know it.
Ozanimod is the tortured S1P acquired from Receptos. I have it peaking at $1.6B and this is my most conservative forecast. It is possible ozanimod does far, far better than this. There are some other S1Ps, with Novartis actually having beaten Celgene’s refiling of ozanimod with their next-generation of Gilenya (fingolimod), siponimod. Still, with $4 billion of Gilenya sales despite a toxicity profile that shocks the conscience, it is blue skies for the fumbling ozanimod. Assuming approval in 2020, they won’t have much time to replace very much of Revlimid, but they might be able to soften the blow if they execute well, which I suspect they will.
I model $1.5B in net revenues from all BCMA CARTS (Juno and Bluebird). This might be conservative as well, but with the rapidly changing environment, it is hard to be confident of any CART revenue projections. Competition abounds from all fronts, CART and non-CART, so who knows if the numbers are accurate or not. Again, I tried to keep risk skewed to the upside. There are many that feel this paradigm shift will be a $5 billion+ opportunity for EVERY player. It is possible. The 80% response rates seen at ASH are remarkable for such late-stage patients. The 50-50 with Bluebird limits some revenue potential, however.
I model $1.0B for luspatercept, net of Acceleron’s share. This drug isn’t a miracle as the ASH data shows. It is still very good, and will change the lives of many MDS patients. Where can you price it though? You’re trying to wean people off of RBC transfusions and there are other options potentially coming. Some feel this will be far larger than I think, but the XLRN stock price is perhaps telling us something different. My numbers could be conservative here as well.
I only model $900m peak for lis-cel. I am a CART bear and I think drugs like MorphoSys’ will be seen as preferable. CART reminds me of Zevalin. You can squeeze out a tiny bit more performance relative to the mab, but is it worth it? Plus, you have Allogene and others making better CARTs. I’m just not ready to have $3B+ forecast for no reason. If Yescarta puts up a few more good quarters, perhaps a revision to $1.5B may be necessary. Again, conservative in most places, but I think I’m right on here.
Fedratinib: I don’t get this one. I see $400 million peak sales in the Jakafi-dominated myelofibrosis/PV indications. This isn’t going to be a blockbuster.
Celgene still has a few years before “impact” and that is really important. With $25B or so of high-margin Revlimid revenue prior to expiry to deploy, and $40B or so if you count pre-full cliff of Rev+Pom+Otezla, there is plenty of capital to do a few more deals. They passed on Tesaro, which is a good start. Celgene has generally been pretty good at BD. The Acceleron deal is a good example (signed for $25 million if I recall correctly). The Juno deal makes me nervous that they’re feeling desperate, but there is still plenty of firepower for acquisitions. A few smart deals will not save Revlimid, but they don’t need to. We’ve all digseted the impact of the cliff and what is important is to value the copious cash flows between now and then, and value the “stub” remainder that the pipeline represents. It’s worth a lot, more than the current fear-based stock price of $71. If you believe my nervous nature was too conservative on all the pipeline, the stock is probably worth $100 or $110 or perhaps more if they can execute Revlimid flawlessly. With Revlimid lasting a bit longer than I think and just one drug like ozanimod surprising to the upside, you could get $120 or $130 out of the stock. For such a big company, that is an attractive return. However, a dud of an acquisition (Tesaros abound) or further buybacks make the risk profile uncomfortable. Celgene has a gun against their head and most management teams are not known for patience during shareholder pain.
Papers I’ve Read
Timing the Landmark Events in the Evolution of Clear Cell Renal Cell Cancer: TRACERx Renal. Thomas J. Mitchell, et al. Cell 173, 1-13, 2018.
Workers here did serial biopsies of ccRCC progression and unveil some interesting findings. First, mutations in UTR region of TERT abound. Next, a tiny clonal population of the 3q deleted region found in this disease of just a few hundred cells can give rise to the visceral tumor decades later. Really cool work and likely to be therapeutically relevant as we screen for these chromothrpsis-bearing patients.
Activity-based E3 ligase profiling uncovers an E3 ligase with esterification activity. Kuan-Chuan Pao, et al. Nature 2018.
Fancy footwork in this paper to discover an E3 ligase that ubiquitinates atypical substrates.
A direct link between MITF, innate immunity, and hair graying. PLoS Biol 2018. Harris ML, et al.
Harris & colleagues do a nice job of showing the innate immune system impact on hair graying. We’ve seen similar work in alopecia and vitiligo, so here comes Rituxan for balding! Jokes aside, this was impressive work, using poly(I:C) to simulate infection. Actual infection would have been interesting, but probably not necessary. The MITF-inteferon connection is made plain here by the researchers.
My tooth abscess still bothers me from time to time. I am almost done with my course of amoxicillin. Fans, friends and family: Please give Trashy a good life if you don’t hear from me!