Skip to main content


MTH1: From Hot Topic to Devalidation?

Here’s an argument that’s been going on inside the oncology drug development world over the last few years: is the enzyme MTH1 a great cancer target or a complete waste of time? MTH1, also known as NUDT1, degrades phosphates of 8-oxoguanine, an oxidized form of the regular guanine that cells know and love. That enzymatic activity keeps the oxoguanine from being incorporated into DNA, where it is definitely not a completely functional replacement for regular G. Such incorporation has been noted as a side effect of aging in animal models, and the theory has been that some cancer lines also depend on MTH1 to keep the nucleotide supply cleaned up while their do their rapid-division phenotype. Given the importance of DNA repair enzymes and related systems in tumor biology, this sort of prep work seems to be a reasonable idea.

In 2014, papers came out from researchers in Sweden and a Sweden/Austria/UK collaboration with strong evidence for that hypothesis. Both of them showed small molecule inhibitors of MTH1 (with crystal structures) and demonstrated that they had strong effects on various cancer cell types. These reports were followed, though, by others saying that MTH1 inhibitors didn’t seem to show such activity at all. Some of the original researchers responded in this 2016 paper, saying that the problem was that the cellularly-inactive compounds turned out not to actually lead to oxidized nucleotides being incorporated into DNA, but why that should be the case was still a mystery:

Here, we cannot give an explicit biochemical explanation to why the potent MTH1 inhibitors previously described by Kettle and Petrocchi do not result in incorporation of 8-oxodG into DNA nor why they do not kill cancer cells. There are four different isoforms of MTH1, which can be differentially inhibited by various MTH1 inhibitors (data not shown) and there are further emerging post-translational modifications on MTH1 that may affect the efficiency of MTH1 inhibition. Clearly, more in depth understanding of the complex MTH1 biology is required to answer these questions. Another possibility is that the TH588 and TH1579 compounds have relevant off-target effects, which work together with MTH1 inhibition to provide the cell killing effects.

The AstraZeneca paper that failed to validate MTH1 had suggested just that latter possibility: that the activity of such compounds as TH588, from the Swedish group, might well just be due to off-target cytotoxic effects. Now a group from Bayer has jumped into the controversy with a new paper that also fails to validate the whole concept:

Attractive target rationale combined with previous success in identifying potent and cellularly active inhibitors prompted us to develop novel MTH1 matter. We developed structurally distinct, potent and selective MTH1 inhibitors with high solubility, metabolic stability, cell permeability and cellular target engagement. However, these properties did not translate into in vitro or in vivo anti-cancer efficacy either in mono- or in combination-therapies. Based on these observations we concluded MTH1 is not essential for cancer cell survival or for intracellular sanitization of damaged nucleotides and thus not a viable target to be exploited for drug development.

What’s more, when they compared their own compound (BAY-707) to the previously reported TH588, they could not reproduce the blood levels reported for the latter compound, found it to be quite insoluble, and could not demonstrate any activity in three different animal models with it. BAY-707 itself had far better pharmacokinetics, but treatment in two different xenograft models, alone or in combination with radiation or three other standard chemotherapies, showed no effect at all for MTH1 inhibition. The Bayer group also notes that MTH1 knockout mice seem to display no phenotypic differences to wild-type. That’s not always a solid indicator, due to developmental compensation, but taken together with the pharmacologic results it has be considered another mark against the target. They’re ready to throw dirt on the grave of the whole idea:

Identifying the right targets is of utmost importance for drug discovery in the era of targeted therapies. It relies on the quality of chemical probes, knockdown and knockout reagents with minimized off-target activities. In this study we describe a novel class of highly potent MTH1 inhibitors, which despite superior biochemical potency, cellular target engagement and pharmacokinetic profile to other MTH1 tool compounds, exert no in vitro or in vivo anti-cancer efficacy either in mono- or in combination-therapies. With BAY-707, a representative member of this compound class, we de-validate MTH1 as a broad-spectrum non- oncogenic cancer dependency and provide the scientific community with a chemical probe having no off-target-related cytotoxicity, to elucidate the biology of MTH1 in cell cultures and living organisms.

As for the idea that some MTH1 inhibitors somehow don’t allow oxo-guanine incorporation into DNA, the Bayer group found that BAY-707 did indeed demonstrate this effect, while TH588 did show the production of the altered DNA. But they went on to show that this doesn’t seem to have anything to do with the MTH1 pathway – you can get the same effect by treatment of cells with several other distinct cytotoxic drugs, or even with siRNAs directed toward other targets than MTH1. So rather than this being a complication of MTH1 biology, as the earlier response suggested, it seems to be just an artifact that has nothing to do with the target at all.

After all this, I have to agree with the conclusion that MTH1 is not a valid cancer target. The whole story is a good illustration of how what looks like a perfectly reasonable idea, backed up with what seems to be higher-than-usual amount of solid data, can still turn out to be completely mistaken. It also show just how complicated it can be to unravel what’s really happening inside cells, and the critical importance of good small-molecule probes, thoroughly characterized, in answering those questions. There are pitfalls and complications waiting everywhere, and new ideas like this one have to be checked from every angle you can think of. Kick the tires, slam the doors – it’s the only way.

35 comments on “MTH1: From Hot Topic to Devalidation?”

  1. Barry says:

    Actual validation of a drug target is a big deal (is e.g. CETP there yet?) and is never complete until you have clinical results. We invest hundreds of $millions and many man-years (often in vain) to get there.

  2. DLIB says:

    The pitfalls of cancer target selection post from a few days ago seems to hit the mark….maybe people will take these lessons to heart

  3. You get what you reward says:

    Original paper was published in Nature. Why wasn’t the paper that disproved the original published in Nature?

    1. Johnny says:

      Totally agree.

    2. Wavefunction says:

      Probably at least two reasons; firstly, it would have made them look bad for doing a shoddy job with the original paper, and secondly, journals are far more interested in publishing incorrect “original” research than correct criticism.

      1. T says:

        Results that turn out not to be replicable/conclusions that later turn out to be flawed don’t necessarily mean the journal did a “shoddy” job handling the paper. It seems the content was reasonable based on what everyone (authors/referees) knew at the time. Reasonable enough that a pharma company invested considerable time and money in following up. That the conclusions were later shown to be flawed isn’t evidence of a failure of peer review but rather science working as it should, with knowledge refined and corrected as we learn more. That said, I agree that it would have been better to have the correction in the same journal (although we can’t be sure that it was actually sent there).

        1. mjb says:

          yes, the AstraZeneca paper was sent to Nature back in 2015, but the editors were not interested in publishing negative data.

    3. tangent says:

      So are there any citation metrics that account reasonably for the validity of a paper? A naive metric will just see “new paper in ACS Chem Bio cites existing paper in Nature” and think that’s praise of the Nature paper, and of Nature as a journal. When in fact to any observer it lowers the assessment of the paper and of Nature.

  4. yestoDNA says:

    DNA repair is a very redundant system for obvious reasons. Maybe MTH1 is inhibited and some other repair protein picks up the slack (looking at you mismatch repair and transcription couple repair).

  5. Cell says:

    Just saw US jobs report. 25$/hr or 49k/year is the average OVERALL salary for US workers. Postdoc salary just offered to me: 40K. Salary of my PhD adviser: 200k ( not including consulting fees). Willingness of my PHD or Postdoc advisers to write letter or rec for me after multiple Nature papers: 0. These guys are garbage.

    1. Cell says:

      Ill also add…amount of work ive witnessed my PhD adviser do… He sits there all day long, writting ” grants” that are never funded.

      1. halbax says:

        If you have multiple advisors who will not write you a recommendation than I suspect the problem might be with you and not them. Plus, if at this stage you think that writing grant applications does not require actual work then you must be pretty clueless.

        1. c says:

          No, its not me. My publication record is strong across multiple labs. They dont like to be challenged or engage in discussion of science, and they retaliated. And im aware grant applications are work. Just not hard or difficult work. Most professors ive worked with are in meetings or traveling a very large fraction of each week (ie. playtime ) and if they had significant work to do, that would not be possible.

          1. What? says:

            Cell- Go ahead and write your own grants then and let everybody else in on the secret of how its not “hard or difficult” to get funded in <10% funding rates. As the guy halbax said, you sound clueless. And again, if multiple labs and PIs had an issue with you, look in the mirror. Most PIs aren't great, but they aren't out to wreck your career either.

          2. Pajas says:

            Dude, take you personal problems somewhere else!

          3. C says:

            You guys have no clue what direction im comming from so you need to rethink you dismissal of my comments. Of all the 6 students that were in grad schoool when i did, all but 1 got fired and 1 committed suicide. I spoke to my “mentor ” 1 time every 6 mo – 1 year, and he ultimately literally stole my PhD project from me. There is widespread and systemic abuse of “students” and until you have the humility to accept that this is an issue, you just look like callous idiots.

  6. a says:

    Cell; stop spamming this forum with your off topic rants: Gyofb

    1. C says:

      Sure, man grad student emplyment has no bearing at all on chem/bio ( although that very topic has been the feature of mutiple posts here ). No, your right, we scholars have better things ro attend to.

      1. Kent G. Budge says:

        Your student employment has no bearing at all on the validity of MTH1 as a target for cancer therapies, which is the actual topic of this post.

        What are you hoping to accomplish by posting about your employment difficulties here? I’m genuinely puzzled.

        1. C says:

          Sorry, but trying to get me banned by baiting me with provocations is unlikely to be sucessful. I am speaking about objective issues regarding systemic grad student abuse that can, should, and will be heard by the community.

          1. Derek Lowe says:

            My advice is to start your own blog, actually. You’ll have more room to maneuver there.

          2. Passerby says:

            To be fair to Cell, in spite of the tangential relevance of his comment to the topic of the post, he/she has a point. Most grad school advisors are indeed far more interested in travel, grant writing and self-promotion than actual mentoring, writing recommendation letters etc. Also true that they are usually averse to being challenged and have Titanic-sized egos. Graduate school has turned into a morass in so many ways. (and before anyone accuses me of being another disgruntled grad student, my years-old grad school experience was phenomenal and my advisors were super-supportive)

          3. Kent G. Budge says:

            “Most grad school advisors are indeed far more interested in travel, grant writing and self-promotion than actual mentoring, writing recommendation letters etc. Also true that they are usually averse to being challenged and have Titanic-sized egos.”

            This was not my experience, but then I finished graduate school a quarter century ago.

            If this is an accurate description of the modern graduate school, it certainly bears some discussion. In the appropriate forum. I suggest Chronicles of Higher Education may be a good starting place.

      2. C says:


        Ok, i may “start my own blog”. I actually have a lot of respect for you, as do some of the relavant parties who i reference, which is why i chose this forum to engage regarding the problems with their management. However, its unlikely that any start up blog would reach the level of readership needed to make an impact ( i understand that your blog is read internationally as an authority on this field, which i obviously cannot replicate given my status). I reserve the right to comment on directly related posts you make, provided that you maintain your excellent level of free discourse appropriate to academic discussion ( sadly this is uncommon now ).

        Bryan Lanning

  7. Mister B. says:

    If we turn this target validation issue the other way round, how much money and effort would it cost to validate (100% more or less) a possible drug target ?

    To be more specific, what are the need to validate a target ?

    From time to time, as Derek is reporting many success and failure in drug discovery, I’ve seen many times the word “luck” when it comes to discovery of a new drug. Is the quality of the choosen target a part of this “luck” ?

    1. Janex says:

      Actually yes. If you wait to start on a target until you are 100% sure it is a good target, your competitor will have a launched drug and 5 other competitors will have me toos following close behind. So when you are picking a target you do the best you can with the information available which is generally a few papers on the target (at best), some information on the pathways the target is involved in etc and a testable hypothesis. Almost all of these targets which look like they have potential will wind up being non-viable for one reason or another, that is why early discovery groups work on so many of them. Every so often you will get lucky and one of these targets will be good. Where the not luck part can help out is in designing experiments to efficiently test your hypothesis as accurately and as quickly as possible (keeping in mind that in some cases it will take phase II clinical trials to test the hypothesis).

      1. Mister B. says:

        Thank you Janex for this answer !
        One more question about it.
        Would it be interesting to have in-house some “fundamental-like” biological research ?
        I mean, some experimented biologists looking for better understanding of some diseases. I know it’s more the academic field, but if I think as a businessman (I am not 😉 ) would it be a good strategy ?

        1. Thoryke says:

          That was the original goal of the cancer research projects at the Memorial Sloan-Kettering Cancer Research Center in New York City. Basic research, applied research, opportunities for collaboration. And while some projects led in unproductive directions [Laetrile, anybody?], other researchers identified a little thing called ‘tumor necrosis factor’…not so great for cancer treatment, but revolutionary for understanding auto-immune diseases:

        2. Janex says:

          Big Pharma used to do a lot of in house research of that type back in the old days. These days they are moving away from that business model in favor of collaborations with academic scientists. Both approaches have their pluses and minuses.

          The do the original research yourself approach can give you a big lead on the competition if you find something plus there is nothing like being buried in basic biology to give you a deep understanding of the targets/pathways you are investigating, which can in itself be very valuable when designing accurate experiments to test hypothesis.. But the returns (i.e. a marketable drug) can be 20, 30 or even 50 plus years down the road. That tends to be too long term for today’s business models. Plus there is a huge failure rate. Most of these basic research projects wind up producing nothing actionable.

          The collaboration approach can be useful if it’s managed correctly. The academics and the industry guys have different goals though and that can produce conflict if not addressed in a constructive manner. It’s also harder to keep what you are discovering a secret long enough to get a jump on the competition (one of the major goals of academics is publications). And you lose the advantage of having your own employees gaining the in depth knowledge of the system you are studying. However, on the plus side, academics are set up to do basic research and usually are already involved in doing basic research in the area you are interested in i.e. no need to re-invent the wheel.
          Also, the collaboration doesn’t just allow you to tap into the knowledge of your official collaborator but allows you to tap into the knowledgebase of his/her network/colleagues as well. That informal knowledgebase can be incredibly valuable too.

    2. Mediocrates says:

      Unfortunately, you only get close to 100% validation of a drug target when you’ve successfully treated human patients. And even that’s not 100%, as evidenced by the successful drugs that get their true mechanism of action worked out years or decades after making their way into the clinic.

  8. Barry says:

    If we look at the statins, validation of HMGCoA reductase as a drug target came with a natural product (compactin) and the race was on. But if it had happened that cerivastatin were the first in class to go to clinic, we might have judged that inhibition of HMGCoA reductase was not a tolerable intervention; the tox is on-target. Only years later would we figure out that the difference between a blockbuster like atorvastatin and a drug that has to be withdrawn from the market like cerivastain is tissue distribution.

  9. r says:

    There are so many targets that start out hot but the data cannot be replicated in the lab. Usually pharma doesn’t have time to refute each one and moves on to the next “hot” target. I’m a bit curious given that background at the publicity MTH1 devalidation is getting! Perhaps do-ability and confidence in biology?

    1. VC says:

      1. Strong claim about “eradication” of cancer
      2. Seemed plausible (if you do a bit of googling it looks like Bayer had actually “validated” the target in 2015 according to published AACR abstract?
      3. As a result many co’s jumped on it

  10. Boris says:

    Why there is no action from Nature after all these controversies against the original claims?

Comments are closed.