Skip to main content


IDO Inhibitors Hit a Wall

Friday brought some very unwelcome news in oncology. I’ve written about IDO inhibitors before, partly in the context of an odd situation between Incyte and Flexus, and partly in response to a recent failed trial of a compound from NewLink. That last one shook people up a bit, but (as I mentioned in the post) there was still hope for combination of the Incyte drug with Merck’s Keytruda (pembrolizumab). The idea is that IDO inhibition would increase T-cell activity and add to Keytruda’s effects.

Unfortunately, that is not the case. The companies have announced that the Phase III trial of this combination in melanoma has utterly missed its endpoints, and that appears to be a massive setback for the whole idea. Add that to the earlier difficulties with other IDO compounds, and the picture is Not Good. Bristol-Myers Squibb has a big trial going with their own Opdivo (nivolumab) and their own IDO1 inhibitor (BMS-986205) in melanoma, and this bodes very ill for those results indeed. It has to be noted that they’ve announced what appear to be positive results for that combination in other trials, but hey: Merck and Incyte had what appeared to be decent-looking results early on, too. The feeling when this trial was announced was that since the company was not waiting for the Merck readout that they must have been pretty confident. But that could also reflect the company’s feeling – not unjustified at all – that they’re locked in a brutal struggle with Merck in this area and had no time to sit around.

This is not only bad news for melanoma patients, for Merck, and (especially) for Incyte. And it’s not only potentially bad news for BMS. It’s potentially bad news for everyone who thought that their understanding of these pathways was good enough to predict success for this trial. One of the biggest things going in immuno-oncology is the combination of the existing agents (PD pathway drugs) with other mechanisms to potentiate their effects and move them into new areas. This trial shows that we don’t know as much about how to do that as we thought we did. So although the IDO-based trials are now very uncertain, this news throws uncertainty around the entire field, whether it goes through that mechanism or not.

Now all eyes are on Bristol-Myers Squibb: will their current trial run to completion? I don’t know when they have any interim data reads coming up, but that’s going to be fraught with interest. The expectation at this point is that the trial will fail, for the same reasons – whatever those are – that the Merck/Incyte one just did. But there’s an outside chance that our ignorance of the field is even more profound and that it’ll actually work. (Update: I should emphasize that the binding mode of the BMS compounds is in fact different from the earlier ones – will this make a difference?) Wouldn’t that be a twist? But you’d have to be brave to bet that way. . .we shall see.

Update 2 (April 30): it appears that the company has halted their IDO trials. . .

38 comments on “IDO Inhibitors Hit a Wall”

  1. Imaging guy says:

    Recently Professor Vinay Prasad found that drugs lacking single-agent activity (monotherapy) in cancer offered only marginal benefits in combination therapy. His suggestion was that if a drug lacks single-agent activity, they should be abandoned since clinical trials are scarce resource.
    “Drugs that lack single-agent activity: are they worth pursuing in combination?”
    PMID: 28266519, DOI: 10.1038/nrclinonc.2017.27

    1. MD...B says:

      That line of thinking suggests that synthetic lethality isn’t a real thing, or at least is not a useful thing. I would suggest that once we better understand the translatability of such combinations, we will find that notion laughable. In the meantime, he may be correct.

    2. Barry says:

      professor Prasad is…er…umm…mistaken (to be uncharacteristically diplomatic). Inhibitors of beta-lactamase have proven (in combination with beta-lactams, and ONLY in combination with beta-lactams) to be good (effective, safe, profitable, sometimes life-saving) drugs.

      1. MP says:

        The quote specifically limits itself to anti-cancer activity; beta-lactams are anti-microbials.

        1. janex says:

          The exact example is incorrect, but the theory behind it is sound. If resistance developed to a useful cancer drug via degradation of that drug by the cancer cell. An inhibitor of the degrading enzyme would be efficacious in combination but not as a stand alone drug.

          Also at one point I worked on a project in which two pathways were required to be inhibited for efficacy. Inhibit pathway 1 and pathway 2 was upregulated, inhibit pathway 2 and pathway 1 was upregulated. The efficacy of either agent alone was minimal, but the efficacy of them together was synergistic (of course so where the side effects, which is why the project never left the early stages).

          Broadly, I think his statement is correct most of the time but that there will be rare occasions in which the underlying biochemistry of a system will require 2 agents simultaneously. The bigger problem industry wide is going to be self-deception i.e. my drug/system etc is one of the rare exceptions to the rule. People are emotionally invested and companies are financially invested, it’s hard to let go.

        2. Barry says:

          one great problem in Onco is that new potential drugs routinely get their first human test in patients who have already failed established therapies. These are advanced cancers with multiple driving mutations. A new drug–that might have been efficacious at an earlier, “pre-cancerous” state–may show no demonstrable efficacy in such a population. Clever selection of the Clinical population (as was done for Tarceva) can improve the odds. But still, treatment of these advanced cancers may require combinations of agents that fail in monotherapy.

          1. Mike says:

            Of course phase 1 trials (including oncology) are not primarily about efficacy. Obviously if you get efficacy, great but they’re more about tolerability and PK. PD should always be there but that won’t necessarily, and are in fact very unlikely to, show efficacy.

          2. Diver Dude says:

            The PD measure in any phase I study should be focused on demonstrating that your putative biology/MoA has survived the move into the species of interest not on anything efficacy related. Small steps.

          3. Barry says:

            re: Mike
            clinicals (phase I-III) for new oncology agents are often all in end-stage cancers that have failed all approved drugs first

            If you get PD in man and the progress of the disease is not affected, your project rationale was bad.

          4. Diver Dude says:

            Which probably the most important drug development question you can answer and Phase I is your first chance to answer it.

          5. Mike says:

            Phase 1 in patients who have failed multiple therapies. 2 and 3 will be in your intended population.

        3. Design Monkey says:

          Actually, Japanese did quite nice things with tegafur/gimeracil/oteracil combo, where the only active anticancer thing is tegafur, while others have about zilch of anticancer activity themselves, but they do inhibit and slow down tegafur’s metabolism, giving it more chance to act.

    3. paperclip says:

      I was prepared to disagree with him, but I see that he does qualify that “extraordinary preclinical evidence can warrant further exploration.” As for what is “extraordinary” — that should take into account the existing options for patients, but overall preclinical data indeed are preferable to wishful thinking or the sunk-cost fallacy.

  2. anon says:

    I think that BMS-986205 is more potent, more selective, and has much better PK than does the Newlink compound. I haven’t seen data on the other IDO inhibitors

    1. DCRogers says:

      Perhaps, but if the understanding of the underlying mechanism is flawed, all the potency, selectivity, and PK will get you nothing.

      For the sake of the sick, I’m rooting for BMS to succeed; but agree with Derek that the odds just got harder.

      1. anon says:

        I’m rooting for the same, but I am just saying, we can’t know one way or another if the mechanism has value, if it was tested using a species without sufficient exposure to even possibly have an effect.

    2. It’s also an irreversible inhibitor (acquired in the purchase of Flexus Biosciences). in contrast to epacadostat and the (failed) iTeos compound.

  3. Peter Kenny says:

    It can be useful to think how patients might sense the benefits of a different binding mode. That said, I think we need to be more open minded about assessing the aesthetic aspects of molecular interactions and this theme is explored in the blog post that I’ve linked as the URL for this comment.

  4. Curious Wavefunction says:

    Are there good examples where a different binding mode (not allosteric) for the same target led to enhanced efficacy? Epacadostat which is the Incyte/Flexus drug binds with the heme still in there. The BMS compound displaces the heme and occupies the rest of the binding site previously occupied by the heme. Not sure how this difference will translate to a qualitatively different outcome in efficacy.

    1. Peter Kenny says:

      Hi Ash, the difficulty is that one can’t do the controlled experiment of just changing the binding mode while keeping everything else (e.g. affinity; exposure) constant. If different binding modes corresponded to different types of inhibition (e.g. competitive versus non-competitive) one might anticipate different behavior in vivo. If we don’t have information about unbound drug concentration at the binding site, we can’t distinguish a bad target from inadequate exposure. I’ve linked a blog post, in which I discuss this problem, as the URL for this comment

    2. Barry says:

      There are (at least) two binding modes of B-RAF inhibitors. One locks B-RAF into a conformation in which it potentiates C-RAF activity (and give paradoxical enhanced signalling in the cell-based assay). The other locks B-RAF into a conformation in which it does not form active heterodimers with C-RAF (nor active homo-dimers). So yes, in at least one case, mode of binding matters.

    3. anon says:

      HIV protease?

  5. Curious says:

    Is there an escape mechanism? Could human cancer cells start expressing TDO2 under selective pressure of an IDO1 inhibitor?

  6. ScientistSailor says:

    Remember that Avastin failed in PII in breast cancer before looking great in CRC. We’ve got a long way to go on this mechanism, and IO in general

    1. imaging guy says:

      I don’t know what stage of colorectal cancer (CRC) you are talking about. The article I mentioned in my first post said (in supplementary information) that median improvement in overall survival (OS) for Avastin plus chemotherapy over chemotherapy alone was only 1.4 months (for metastatic CRC). Progression free survival improvement was 1.6 months.

      1. ScientistSailor says:

        oops, Lung, not CRC…

  7. Srini says:

    How about a dual IDO1-TDO2 inhibitor vis-a-vis a selective IDO inhibitor. Will it have better handle to contribute synergy with PD1 antibodies.

  8. Pennpenn says:

    “This trial shows that we don’t know as much about how to do that as we thought we did.” If pharmaceutical development was a song, this would probably be the chorus.

    1. AR says:

      It would be the ‘stand-up and key change’ 3rd chorus with the backing singers involved.

  9. johnnyboy says:

    Interesting that the patient tumours were not evaluated for IDO1 expression (no mention of it in According to literature, only somewhere around 40% of melanomas are IDO1+. You can’t inhibit something that isn’t there. Curious to know why they didn’t include IDO1 expression, at least as a stratification biomarker, since they did require pre-treatment biopsies.

    1. DJ says:

      New around here,

      but this seems like a valid statement and logical choice to include. Not too familiar with what can and cannot be done in trials.

  10. Derek, interim data from bladder and cervical cancer patients was presented at the November 2017 SITC meeting. Small numbers, and as usual in Phase II IO combination studies, hard to interpret.

    1. Derek Lowe says:

      So we’ll all just have to take our chances in Phase III, I guess. Which is also “as usual”. . .

  11. steve says:

    I think johnnyboy hit the nail on the head – you need better patient selection. If 60% of your patients don’t express the target then you’re starting from a difficult position. From a scientific point of view, what you really want to do (before running a clinical trial!) is see if there’s a correlation between patients who fail to respond Keytruda and IDO expression. This should be a pretty straightforward study with archived material. The problem is a rush to the clinic before doing the background science.

  12. Justanothermedicinalchemistwithnoclue says:

    How sure are we that only IDO in tumor cells plays a role? What about different immune cell populations? Yes, PD read-out from PhI seems to be one of the interesting questions around IDO trials. But would you look primarily for a proof that the compound hits the target (I think Incyte showed data about kynurenine/tryptophan ratios in patient plasma) or for signs of an immune response? It gets more complex the longer you think about it…

  13. friend says:

    My friend has started the enrollment process but not dosed for epacadostat/Pembro and is still hung up on the original data and not ready to let go. The trial so far is still open… She is relapse from Oncosec pIL12/ Pembro.
    Any advice?

  14. loupgarous says:

    The huge amounts of money Big Pharma firms are investing in “promising” biologics, while gratifying to a cancer patient like me, are also unnervingly similar to Dutch Tulip Mania, or the “irrational exuberance” surrounding tech stocks in the late 1990s.

    I don’t want research and development on biologics (for cancer, AD and other dread diseases) to stop. But I wish there was more discernible reason to the business side of things. One way or another, my life literally depends on Big Pharma staying in business and developing new cures for rare cancers.

  15. bassam says:

    Tryptophan pathways are controlled by IDO1, IDO2 and TDO. Immune suppression is mainly due to Kynurenine and its metabolites. If IDO is blocked and TDO is activated then downstream enzymes of the Kynurenine pathway and kynurenine itself (AHR inhibitor) will still suppress the immune system.

Comments are closed.