Skip to main content

Clinical Trials

Finger-Pointing at Celgene

Just what is going on at Celgene? They’ve had some odd recent setbacks (such as as the failure of Mongersen), and another such acquisition, Ozanimod, had its filing recently rejected by the FDA. Several things about that incident were eyebrow-raisers – for one, it’s rare for a large, established company to get an outright refusal-to-file notice. That’s one reason why you have a regulatory affairs group, to keep that sort of thing from happening – and if you’re really working with the agency during development, it doesn’t. (A refusal-to-file, for those outside the business, means that the FDA isn’t even going to bother reviewing your application for a new drug approval, because an initial inspection determined that the NDA package you submitted was obviously lacking. In other words, it’s a “Don’t waste our time” response).

Then it came out (during a recent American Academy of Neurology meeting) that most (maybe all?) of the drug’s activity is due to a metabolite. Now that certainly happens. The liver chain-saws most of the drugs that people take, and some of those chopped-down versions of the original structure are also active on the drug’s target. (Unfortunately, some of them might be active on other targets, too, which is another thing you’d like to discover as early as possible). And that’s just what was odd about this: how come the company didn’t know about this active metabolite earlier? And why wasn’t this issue fully addressed for the FDA?

It appears that the new compound is produced far more in humans than in other animal species, but that happens, too. You can try to get warning of this sort of thing by testing the compound against human liver microsomes or in human plasma (or whole blood), but those don’t recapitulate everything that can happen in vivo, of course. OK, so you find about it in Phase I, during first-in-man. Analyzing blood levels of the parent compound and metabolites is one of the biggest points of Phase I, actually, so it’s not like this could have been overlooked. If you find out that what you thought was your drug is apparently just a prodrug for what’s really working in vivo, well, you have more work to do. But it appears that lack of data about the metabolite could have been one of the main reasons the FDA found the NDA unworkable, which just makes no sense.

As a side note, I haven’t seen any discussion yet of what that metabolite is. You’d guess some sort of hydroxylation on the aryl rings, or removal of the O-isopropyl, perhaps, but those are the sorts of things that you might well have seen in liver microsomes, thus giving you more warning. I’ll guess that it’s trouble over at the aminoindane or the nitrile, via some oxidase that isn’t picked up well in the in vitro screens, but what do I know? Update: none of the above! Looks like it’s oxidation to the carboxylic acid.

This is all bad enough. But yesterday, the Financial Times reported a statement from Celgene’s Nadim Ahmed that surprised everyone some more. He basically blamed the fiasco on the company (Receptos) that Celgene acquired to get Ozanimod. Quoth he: I think that 99 per cent of folk at Celgene wouldn’t have submitted, but we had Receptos out on the West Coast and, for whatever reason, the decision was made to submit. . .We learned a lesson of humility and that when you do an acquisition it’s better to be more integrated rather than be completely away from the mothership.”

Nope. This is the the-dog-ate-my-homework school of regulatory affairs (OK, maybe the cat), and a company with twenty employees would get razzed for an excuse like that. Much less a Celgene. As one might expect, the ex-CEO of Receptos, Faheem Hasnain, isn’t having it. He says that Celgene had on-site control for over two years before that filing took place, and it’s hard to see how he isn’t right about that. Those crazy dudes out on the coast didn’t just whip up an NDA out in the garage and mail it off to the FDA while the managers in New Jersey weren’t watching. Right?

After Biogen’s compound for the same target (the sphingosine 1-phosphate receptor) dropped out in 2016, that left ozanimod as the main competitor to fingolimod, which has been on the market since 2011.  But that’s going to take longer than expected – Celgene now says that they’re going to re-file next year. It looks now like that didn’t have to happen, and maybe that phrase applies to many of Celgene’s lost billions during their acquisition era.

36 comments on “Finger-Pointing at Celgene”

  1. Anonymous says:

    Ozanimod had to be the main driver in acquiring Receptos. Doesn’t Celgene do scientific due diligence to make that decision? Celgene DMPK and clinical pharmacology scientists clearly should have recognized gaps in the metabolism data at the time of acquisition. A small company is going to do the bare minimum to complete clinical trials that will attract a buyer. Even if you let the Receptos group do the submission, Celgene scientists and regulatory staff should have provided a list of prioritized gaps that needed to be addressed. Making this excuse indicates Celgene management hasn’t really learned a thing.

  2. Philip says:

    I am in IT, not chemistry, not biology. If the following description of how S1P agonist works is off by a bit, I will not be offended by corrections or clarifications. If fact I will welcome them. S1P agonists sequester immune cells in the lymphatic system. This lowers the patient’s immune response, which can relieve symptoms of and reduce damage done by auto-immune diseases such as multiple sclerosis, ulcerative colitis and Chron’s. The down side is that with an attenuated immune system patients may not be able to fight off infections or even cancer. Of particular concern is JC virus which causes progressive multifocal leukoencephalopathy (PML). So the perfect S1P agonist would tamp down the immune response in a dose dependent way, not hit off target receptors and allow for quick return of the full immune system response when the drug is stopped.
    Ozanimod was going to replace fingolimod in the market because of its short half life. Half life being a surrogate measure for how fast the immune system recovers. Turns out that the metabolite’s half life is about the same as fingolimod’s. So Celgene has some hard decisions to make about the future of ozanimod. To me it looks like ozanimod can be approved. It also looks like it will never make a lot of money. People and corporations have a tendency to throw good money after bad. Celgene needs to be very careful here. They need to replace drugs going off patent, and I don’t think ozanimod is going to be one of those drugs. Over $7 billion down the tubes. Maybe some of the stock buyback money needs to be spent on their pipeline.

    1. cynical1 says:

      Philip – As far as I know, S1P agonists, although immumnosuppresives, are not associated with JC virus reactivation leading to PML. In fact, patients are often switched to fingolimod from natalizumab if they test positive for JC virus.

    2. Alchemist says:

      Fingolimod also has some cardiac side effects, which are supposedly due to its activity on another S1P isoform, S1P3. THere was a big rusha few years ago to make S1P1 or S1P1-S1P5 selective agents, to avoid this, and Ozanimod is one of those.
      And I guess (did not look at it honestly) that the side effects must have been lower in Ozanimod’s clinical trial.

      So a rationale could still be there, though the generics of Fingolimod will bbe hard to beat if there is not a big difference in tox.

      I do wonder what is the selectivity of the metabolite though…

      1. Philip says:

        I looked months ago when the metabolite hit the fan. So from memory, ozanimod looked a little better on the cardic side than fingolimod. Not enough IMHO to make it worth the price difference with generic fingolimod. Ozanimod’s big advantage was going to be fast immune system recovery. That disappeared with the appearance of the metabolite.

        The metabolite has opened up an opportunity for better S1P1-S1P5 agonists to be market leaders.

      2. Mike says:

        S1P3-mediated cardiac effects was the mantra for a long time based upon results from rodent models. However, the same sorts of cardiac events have shown up with S1P1-selective ligands (reference below), and it was suggested that this side effect may be mediated by different receptors depending upon the species. These side affects appear to be mediated by dose escalation…

    3. Tourettes of Chemistry says:

      Allosteric modulation of GPCR’s is another reality to bear in mind. Lessons from the phosphatases (albeit different biochemical function – enzymes vs. receptors) have been better characterized to my knowledge.

  3. Paco says:

    Metabolite is likely the phosphate. This is a well known phenom. In the S1P world.

    1. Barry says:

      If this phosphate is an active metabolite, does this mean that the primary alcohol is not necessary for the pharmacophore? Or is it “active” only off-target?

      1. Paco says:

        I think there is some evidence that non-phosphorylated alcohols can be potent against S1PR isoforms, suggesting that the alcohol does match the pharmacophore; however, the ionic component contributed by the phosphorylated alcohol can make compounds much more potent against the S1PR targets. Take a look at the endogenous ligand sphingosine-1-phosphate. Also check the story behind FTY720 which is a prodrug for the phosphate metabolite.

    2. Derek Lowe says:

      That certainly sounds possible, and you could expect both species differences in that secondary metabolite formation (and lack of clarity on it until you hit the human subjects). Update: and it also means that being surprised by this or not dealing with it in the NDA is an even more egregious miss. . .

      1. road says:

        That’s literally THE MOA for fingolimod. How could that possibly get overlooked?

  4. biotechtoreador says:

    Isn’t CELG’s SD facility a mile (at least as crow flies) from where Receptos used to be?

  5. Wow again says:

    Poking around it looks like fingolimod going off patent and generics will be out in 2019 or 2020. Does ozanimod have substantially a better efficacy or tox profile that will make it worth launching a year or two after a generic competitor is on the market?

  6. BernYeeBridges says:

    Celgene named the Phase III clinical trial in RMS using Ozanimod “Sunbeam”.

    Ow, my Pocketbook!

  7. gpcr says:

    Receptos is a great illustration of the academic industrial biotech pyramid scheme.

    The Scripps Research Institute has been crowing about the multibillion dollar acquisition as a demonstration of its “translational” prowess. Ray Stevens, who has since left TSRI for USC, I presume made out quite well as the founder. Look at the people who were part of that lab (for a long time) and were the hands that got all those GPCR crystal structures. Surely they got at least a bit from that sweet 7 billion dollar deal? Will we find them retired on a beach somewhere? Perhaps still employed as valuable, highly compensated employees at Receptos?

    1. Isidore says:

      Since when is it a crime or morally questionable to make money from one’s efforts? Are you suggesting that those who benefited financially from Celgene’s purchase of Receptos (and many of whom are no longer with the company) were responsible for Celgene’s “refusal-to-file” debacle? If so you may want to provide some additional information, because the way I read Derek’s post as well as the information in the various links he provided it would appear that Celgene’s scientists, regulatory people, and management were very cavalier about the whole filing, which predictably led to the problem.

  8. Roger Miller says:

    Hey gang. You don’t need to speculate about the ozanimod metabolites. One of the main metabolites, which is more active and has higher human exposure than parent, is the acid! Published earlier this year:
    Cardiac Safety of Ozanimod, a Novel Sphingosine‐1‐Phosphate Receptor Modulator: Results of a Thorough QT/QTc Study
    Jonathan Q. Tran et al.

    1. Derek Lowe says:

      Thanks – just updated the post.

      1. Hmmm…

        Presumably more than just (known) metabolism to the acid. That metabolite was identified by Receptos, in human samples, 4 years ago:

        Something else would seem to be going on, for example, a truly missed metabolite (surprising), or a fumbled characterization of the known (acid?) metabolite(s) (also surprising).

    2. Barry says:

      if the acid is more active than the primary alcohol, it sounds less likely that the alcohol is just a prodrug to the (active) phosphate

  9. John Wayne says:

    So … they were surprised when a major metabolite of a drug bearing an unhindered primary alcohol was the corresponding carboxylic acid? These guys are making below average medicinal chemists look like a bunch of geniuses.

  10. mallam says:

    Certainly Celgene has lost a bunch of money so far, but the original investors in Receptos accomplished their objective, one that is typical of small biotechs that never really intend to fully develop a drug candidate to a successful NDA filing and acceptance, which is to bring a reasonable looking drug candidate far enough along to sell the molecule or buy the entire company to a big payer. And this is why the offering(s) of biotechs need to be examined in exquisite detail.

  11. Passerby says:

    You would think that doing metabolite ID and testing the metabolites separately would be a standard part of the process.

  12. Justsayin’ says:

    There is no question that Celgene made a major blunder in not staying close to an asset that was being prepared for a BLA filing, as this capability is what Pharma should do best.
    Sadly their overinterpretation of this error to mean that “full integration” of an acquired assset is the solution only compounds the mistake, as illustrated by their subsequent decision to fully integrate Juno, a CAR-T company. Celgene has NEVER done discovery and early development well, and these capabilities are what they paid $9B to Juno for. Sadly, putting this responsibility into Celgene management’s hands instead of allowing the talent they acquired to drive the pipeline is a decision they (and patients) will regret.

  13. mfernflower says:

    So the ended up creating a prodrug for a drug they had no clue even existed?
    Didn’t they use a metabolic predictor software? Anybody or any program would envision the alcohol flopping about oxidizing in the liver

    It seems crazy that a multimillion dollar company would make a mistake a first year student would

  14. Emjeff says:

    To everyone who expresses surprise at this deal being done, you’re thinking about these deals all wrong. You are probably under the impression that highly-trained scientists are involved from start to finish, and that they have been able to vette the data thoroughly, make recommendations, and those recommendations are taken seriously by the company. Reality is much different. I have done quite a few due-diligence evaluations, and the simple truth is that , often, the scientists’ opinions are treated either indifference or disdain – if they are even asked for at all. I once did a DD where the scientists were told up-front:” We are doing this deal, no matter what.” Another project I inherited (which was a result of an in-licensing deal) was not even looked at by any scientists, and the business development people did not consider it important that we, as the acquiring company, receive the actual raw data from the partner. As a result, all we had were scanned PDF reports. So , needless to say, I am not shocked by any of this…

    1. Chrispy says:

      Your experience sounds very familiar to me: the business folks are driven to do their deals, and their ability to persevere and “Git ‘er done” even in the face of adversity commands respect among their peers. Scientists are confined to nonfiction and have a much narrower view of that is reasonable or possible. The MBAs often hold scientists and their views in contempt, as they see them as generating needless roadblocks.

    2. Been there, done that says:

      I’ve had a very similar experience with Celgene. Years ago, they came to purchase a program that we had been working for years on. We had substantial animal model SAR, efficacy, and tox data. They weren’t interested in any of it. Their only interest was to get the deal done and move on with their development plan–even though out data meant that the plan wouldn’t work. Needless to say, we weren’t surprised when that program took a massive detour and was delayed by several years. Clearly their culture hasn’t changed.

    3. milkshake says:

      While being at a small dysfunctional company with thieving and deranged management, I witnessed 1) A custom synthesis – contract kind of manufacturing deal that tied something like four man-years of our chemistry time and never earned any money (we are talking about a little company with three chemists, in times when the main clinical programs were struggling and falling behind) 2) Acquisition of another company for its two clinical “projects” that turned out to be total dogs, with non-existing PK, nonexisting biology and no meaningful analytical method to follow the drug in vivo – a real cherry of a project to develop in the clinic.

      Needless to say, the input of chemists to these two deal from which we never recovered was studiously avoided by the management. I got even temporarily blamed for the failure of the custom manufacturing deal, reprimanded and temporarily fired – and even though they later reconsidered and bought me a bottle as a sort of apology, i had this indelible impression from the affair that our CEO was perhaps a very good scientist, or a person.

  15. AML says:

    BLA…? Ozanimod is a small molecule, not a biologic drug.

  16. anon says:

    Receptos knew about the acid metabolite a long time ago, so this doesn’t make a lot of sense. Maybe there are gaps in knowledge of the fate of the acid metabolite. Perhaps it forms a glucuronide and nobody studied its tox? Weird.

    1. 10 Fingers says:

      Actually, that brings up a point about the acid that I hadn’t really considered until reading your comment. More often than many realize, carboxylates circulate (and recirculate) as fairly unstable acyl glucuronides. When isolated, if you don’t take care with your technique, it looks like there’s a lot of carboxylate – but 99% of it can be acyl glucuronide in circulation.

      I wonder if this issue could be part of the problem – two groups looking at samples differently?

  17. lennyjoels says:

    One only needs to look at Theranos to have another instance where the people in the trenches said “This isn’t going to work” and “Don’t do this” but allegedly brillant investors didn’t do their due diligence and board members appear to have been overwhelmed by greed and a field-specific Dunning–Kruger effect.

  18. Alan Goldhammer says:

    At least former CEO Bob Hugin has a chance to be the next Senator from New Jersey.

Comments are closed.