Skip to Content

Mongersen Fails

Readers may recall a post here last year about an odd trial of an antisense drug for Crohn’s disease. Celgene had acquired the drug (mongersen, GED-301) from Nogra Pharma of Ireland back in 2014 as a late-stage candidate, and for a while, things looked good. In fact, going back and reading the stories, you’d think that everything was pretty much on track:

Celgene ($CELG) bet big on the little-known Irish biotech Nogra Pharma when it partnered on a mid-stage drug for Crohn’s disease. And today Celgene spelled out the reasons why it gambled $710 million upfront on a Phase II drug, highlighting data that support a clear case that the therapy can help spur clinical remission in a broad group of patients.

An oral antisense agent is a pretty bold move, but then again, a Crohn’s drug of that sort just need to hit its target in the gut wall, not make it into circulation. And mongersen’s target is Smad7, a key player in the transcriptional signaling machinery for an important inflammation pathway (among other things). This is the sort of target that is very difficult indeed to go after with a small molecule, and that’s when you see the antibodies and oligonucleotide-based modes get tried.

Later results were not as encouraging, though. And the reason I called that trial almost exactly a year ago was that it had no control arm, making it very hard to tell the difference between mongersen’s effects and a placebo. On Friday, Celgene had an announcement on a full placebo-controlled Phase III trial, and guess what? It actually isn’t that different from a placebo. Fancy that. The announcement was that the trial was being discontinued due to futility; an interim analysis showed that nothing was happening.

Unfortunately, that candidate was actually a pretty important part of Celgene’s plans and revenue projections. When the company did their 2014 deal, it raised eyebrows because of the steep upfront price for a relatively unproven therapy from a relatively unknown (and very small) partner, but Celgene was (as they said at the time) into Planning Boldly For the Future, as well as Executing Transformative Deals on Late-Stage Clinical Assets and all that stuff. Unfortunately, the science crept up and sank its teeth into the ankle of this mighty deal, and one would assume that mongersen itself is no more. There’s a lot of finger-pointing about putting that much money into something so thin, but of course if the compound had worked, everyone would be taking visionary dealmaker victory laps. It’s evaluating that “if the compound works” part that is the tricky part, and a tiny company’s oral antisense agent for Crohn’s was always going to be a gamble. You just wonder if it had to be quite as expensive a gamble as it was.

And as always, whenever something like this happens, I will remind people that this is why you run big, well-controlled Phase III trials. Back in Phase II, mongersen looked as if it were going to work (as that quote above illustrates). It doesn’t seem to have any particularly bad safety issues, so under some regulatory proposals, that would have been the time to let suffering Crohn’s patients take it on a risk basis, speed up development, get the regulatory barriers out of the way, all that stuff. But that would have given everyone three years of useless placebo, at a no doubt stiff price. And since more drugs in clinical trials fail than work, I’m still baffled at how giving people a chance to pay for them at that point is supposed to improve health or save anyone money. It certainly wouldn’t have in this case. Celgene stuffed well over $700 million in real money into the shredder on this effort, and a million Crohn’s patients could have joined them.

 

31 comments on “Mongersen Fails”

  1. Emjeff says:

    “And as always, whenever something like this happens, I will remind people that this is why you run big, well-controlled Phase III trials.”

    More importantly, it’s why you use a control group in Phase 2. Doing a clinical trial without either a placebo arm or a standard-of-care arm is malpractice.

    1. tangent says:

      Can anyone explain why you’d do a trial design without a control arm? Or at least why that wouldn’t be a total red flag to an acquirer? I don’t get it.

      1. KeithD says:

        Because, just sometimes, you can get away with it!
        E.g. Yescarta just approved by FDA. Approval was based on a single-arm multicenter trial of 108 adult patients with aggressive B-cell non-Hodgkin lymphoma

      2. TC says:

        If your result is strong enough, you don’t need it. If your compound is supposed to regenerate limbs, and it works, you don’t need to have a control group of amputees to see it.

        1. tangent says:

          Okay, fair point, in that kind of case.

          Still, there you could also see your limb-growing in an experimental group that’s half the size, with the others on control. Almost universally, isn’t a half-and-half setup going to get you more power than all-experimental for the same cost to run?

          The kind of case I can construct for all-experimental is when your effect is “regrows lost heads in 1% of subjects.” There’s simply zero point in a control group, and you might as well put it all in experimental to minimize you chance of hitting on (0.99)^n bad luck.

  2. Anon says:

    I seem to recall that most or all of their phase 2 data was from a single study site. This can also be a risk factor for phase 3 futility.

    1. Calvin says:

      Yes. Many other companies looked at in-licensing the program and it was picked up by others that the clinical data was “odd” and “only worked in Italians”. UC/IBD is such an opportunity, that clearly Celgene thought it was worth the risk. Nevertheless, the valuation was way higher than anybody expected so you do have to wonder what Celgene thought they saw that nobody else saw. Clearly, it was a very expensive mistake to say the least!

  3. David says:

    Well, if patients pumped in a billion dollars into the drug over the three years between a “risk basis” approval and the Phase III data coming back, Celgene would have saved $700 million plus another $300 million for their trouble, but I doubt that’s the money-saving argument anyone wants to make.

  4. R.Thomas Winters says:

    Since Crohn’s is associated with a greater risk of CRC in younger people and there is evidence that SMAD7 is implicated in Metastasis of CRC to the liver, I wonder if the Phase III data might have some interesting things to say about the incidence of Stage IV CRC between the test and control groups? Others have suggested that SMAD7 might be a therapeutic target in CRC.

    1. Derek Lowe says:

      Yeah, that’s a good question. Is the compound a good enough probe to answer questions like this, I wonder, or is it too deficient in general?

      1. Jane says:

        If it is good enough to reduce the incidence of CRC in a phase III clinical trial, that in itself is a marketable indication. So my guess is probably not or Cellgene would be annoucing that although we didn’t meet our primary endpoint, we did find that….

        1. R. Thomas Winters says:

          CRC and cancer in general are not indications where a protective mechanism might show up in a study for like these unless someone makes a leap of faith to look for it in a longer term study. Look at all the studies that have been done concerning diet, exercise etc vs cancer. Very hard to draw conclusions unless you have a LOT of data and a lengthy amount of time to collect it. Someone susceptible to CRC via Crohn’s might take 7 to 8 years before a measurable visible tumor shows up that can be diagnosed, and usually only then via a colonoscopy and biopsy of the polyp! Sometimes the initial symptoms of CRC are mistaken for IBS/Crohns and or Diverticulitis. Still if the antisense drug used in this study could be used as a molecular probe in SMAD signaling in CRC that could be useful research provided as Derek indicated the compound is not a deficient molecular probe.

  5. Chrispy says:

    Well, it looks like a Phase 2 DID have a placebo control:
    http://www.nejm.org/doi/pdf/10.1056/NEJMoa1407250

    Looking at that data, you can see why the molecule looked so attractive. There was a clear dose-response, with the two higher doses seeing statistically significant remission vs. placebo. None of the authors appears to be affiliated with Celgene, though, and this study came out a year after the purchase of the molecule. So it is not clear what is going on.

    1. 10 Fingers says:

      That’s interesting. Given the differences between the lowest dose (not different from placebo in most criteria) and the two higher doses, I wonder if they were considering that “placebo”….

      If so, then the failure last year to show any differentiation between doses was an accurate signal….

      No excuse for failing to include a true placebo arm – more that they could well have known that they were heading towards pulling the plug on the trials sooner than they let on.

    2. Imaging guy says:

      You also should read the editorial accompanying that article, which expressed doubt over the results.
      http://www.nejm.org/doi/full/10.1056/NEJMe1415053#t=article

      1. 10 Fingers says:

        Thanks! Saw that – low CRP at entry for a large subset of patients seemed the most problematic call-out (to my naive eye).

        The most perplexing thing to me in this is the seeming surprise factor at these latest results. Either management was too blind to see what was coming, or they had a hint and failed to manage expectations appropriately. I try to not confuse ignorance or incompetence with malfeasance. However, this is quite a mess. Hard to know whether or not all three were in play.

      2. Chrispy says:

        It doesn’t seem like there is very much doubt expressed:

        “The remission rates of between 55 and 65% for the two highest doses are unprecedented when compared with those reported in the large pivotal induction studies of infliximab (32.5% glucocorticoid-free clinical remission at week 6 in the SONIC trial),4 adalimumab (36% clinical remission at week 4 in the CLASSIC-I trial),5 and more recently, vedolizumab (14.5% clinical remission at week 6 and 39% at week 54 in the GEMINI 2 trial),6 all of which also involved patients with moderate-to-severe active Crohn’s disease. ”

        There is a bit of hand-wringing about inclusion criteria and the lack of correlation with C-reactive protein, but they get pretty breathless about the long duration of action:

        “Another intriguing finding is that clinical remission was maintained for almost 3 months, even though the drug was administered for only 14 days…. If confirmed in future studies, this durable effect of mongersen would be unprecedented and would represent a first step toward curing the disease.”

        So I want to know what happened in this trial — how did it look so good only to blow out in Phase 3? I mean, the response rate really was phenomenal. Was this outright fraud? I have trouble understanding how such strong results could just be stumbled upon.

        1. Andy II says:

          So, why this phase 2 study result was published in New England J Medicine at the first place though the Editorial raised a set of very critical points that the reviewers should have addressed. Well, I have seen other similar cases of beautiful phase 2 studies in NEJM, whose phase 3 studies did not go as expected….

      3. Anonymous says:

        There was no endoscopic screening at entry. Without a confirmation that the patients enrolled had confirmed Crohn’s disease, their symptoms could have been from anything, including stress that might have been relieved by placebo or the active molecule.

  6. steve says:

    I think it’s clear what happened. Crohn’s is exacerbated by stress. It worked in Ph2 because that preceded the Nov 2016 election. It’s clear that another kind of antisense has predominated since then.

    1. You've been burned says:

      Bazinga!

  7. Barry says:

    The epidemiology of Crohn’s/ulcerative colitis shows a protective influence of parasitic worms. It’s rarely if ever seen in populations that carry worms in the gut. Some of these worms have been shown to provoke elevated levels of TGF-beta in the host. At least one seems to express/export an agonist of TGF-beta. But SMAD7 is downstream of the TGF-beta receptor. If SMAD7 is driving Crohn’s/ulcerative colitis, then these parasitic worms (as often as they’re protective) must be modulating something farther down the signaling cascade, too.

  8. Jon says:

    For a real tinfoil hat version, try this:

    Consider: It’s a zero-sum game. That seven hundred million didn’t disappear, it went to someone else. In Ireland. Who pays a lot less tax on corporate profits than us in the us.

    Less overhead, of course.
    Hmm. J.

  9. And D says:

    Just wondering about the delivery, is there good evidence such a large molecule can consistently permeate out of the lumen and reach target tissue, which presumably is somewhere in the colon wall?

  10. CuriousChemist says:

    One has to wonder how this oligo gets to enter the cell in the first place….mystery. Second, could this be case of clinical trial fraud? Does phase II data get audited for authenticity by an independent third party? Just asking folks here who might relevant experience.

  11. Curt F. says:

    I’m still baffled at how giving people a chance to pay for them at that point is supposed to improve health or save anyone money.

    I’m not sure I agree. I mean, sure you are right in this case, no question about that. But I’d be curious for your views on the topic more generally. Do you think drug discovery entrepreneurs (and the VCs that backed them) should never have options for liquidity prior to FDA approval? Surely there is some kind of balance between rewarding promising early-stage progress and over-stimulating dud projects. Is it just the price tag that was too high in this case? (The $700 million number reminds me of Sirtris, another case where you are right that health was not improved and no one saved any money on the path to new drugs.)

    1. tangent says:

      In an efficient market with perfect information, it would make sense. In our world, patients have so little power to distinguish between a “50% chance” drug and a “5% chance” drug before full trials, no rational company is going to invest in the small trials to push from 5% up to 50%, especially when those trials have a good chance of proving 0%. Just spam out drugs tested to the minimum legal standard (maybe that’s safety only), sit tight, and put the money into volume and marketing.

      1. Curt F. says:

        I wasn’t trying to say that patients should making decisions like this. But drug companies sometimes do. Celgene bought mongersen, which was apparently a big mistake. But Derek seemed to be saying that anyone (not just patients) ever buying a drug at an intermediate development point is a mistake, and I wonder if that is generally true. I think Derek would be much better positioned than I to have a useful opinion and I would love to hear him elaborate on this particular point .

  12. Norman Foster says:

    Chaos continues for Celgene. A cut in outlook, especially it’s long-term targets, is not an encouraging one. Also, a decrease in pharma company’s research and development expenses is not a good sign. http://alphastreet.com/b0662389

  13. Jan says:

    I am one of the patients who was about to join the phase 3 trials. I had just passed the screening tests and received a “smart phone”
    for the “daily digital health reports”, when my doctor phoned me and said that the study had been stopped…

    Until then, I had been sure that phase 2 trials were just as “scientific” as the phase 3 study would be, especially after all the medical tests I went through in the screening period that would tell if I was “go” or “no go” to join the phase 3 study.

    /Jan in Sweden

Leave a Reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.