Skip to main content


Threshold Goes Under the Threshold

And now we turn to “Why you can’t say a thing about whether a drug is any good until it’s been through all the clinical trials”, part nine hundred and seventy-five thousand, two hundred and ten. Today’s example is evofosfamide, TH-302 from Threshold Pharmaceuticals. The idea behind it seems perfectly reasonable on paper: it’s a prodrug of a nitrogen mustard compound, a classic (and classically nasty) variety of chemotherapy. In this case, though, the reactive species doesn’t get released until the drug gets into a low-oxygen environment. It has a nitroimidazole group on it, a class of compounds known to be activated under those conditions, and in this case, when there’s not enough oxygen around to intercept the pathway, the compound will fall apart to reveal the nitrogen mustard end.

A prodrug of this sort that’s activated by hypoxia would seem to be useful in a number of solid tumors, whose cellular environments are notoriously oxygen-poor. There are a number of programs underway to exploit this exact idea, with Threshold being one of the most advanced. They signed a deal with Merck KGaA (Merck-Darmstadt) to develop the compound, and earlier this year (after the drug won fast-track status from the FDA for advanced pancreatic cancer) you could read articles that would tell you about how Threshold could turn into a big deal once their trial results came in.

Or not. To be fair, one of the analysts quoted in that last link did say that the company was in for an “eventful next twelve months”, which is a bit like J. P. Morgan’s prediction that the market would fluctuate. Threshold now has the Phase III data, and it’s a wreck. The compound completely missed, which crosses out Merck KGgA’s top oncology candidate and has sent Threshold’s stock down into delisting territory. It traded at $6 to $8 dollars back when they signed the development deal, and at $4 to $6 since then, but as I write, it’s at sixty-three cents, and the survival of the company is very much in doubt.

Which is a pity. We certainly need something that will affect advanced pancreatic cancer, a diagnosis which is pretty much a death sentence under current medical knowledge. And Threshold’s idea was not a bad one at all. It just didn’t work, for reasons that are not clear. A lot of clinical results are not clear – even the good results aren’t always good for the reasons that people were expecting. Next time you feel like putting some money down on a small company that’s in Phase II with an oncology deal, consider Threshold, who were once as handsome and tall as any of them. . .

10 comments on “Threshold Goes Under the Threshold”

  1. William Gerber says:

    I recall reading something a few years ago by a chemist who thought that the chemistry behind TH-302 didn’t stack up. I’m not in science but I recall his argument was that hypoxic conditions would be unlikely to provide the energy required for the reaction you describe above (prodrug –> drug). Does that make any sense? Also, I believe I read somewhere (probably an anonymous blog for what that’s worth) that this compound doesn’t penetrate tumor tissue enough to have a big impact on sarcoma. Maybe a lack of scientific basis for this drug was why the market cap was so low going into data release? Thanks.

  2. biotechtoreador says:

    I’m amazed THLD was able to keep fooling people that this was somehow a viable asset.

    The hypoxia bit is a good story with a veneer of plausibility, but there’s a reason none of these things have actually worked since Denny first published on them in the 80s (and why Promacta’s and Vion’s similar drug failed).

    What people seem to miss in waving their hands about how hypoxic tumors are is that that doesn’t necessarily translate into redox potential. Best I could tell, the redox potential of a tumor cell is about -240 mEV, well short of the 1.1 V required to reduce an aromatic nitro group to an amine.

    THLD, for a second time (its glufosamine drug was an equally implausible mechanism and failed a few years ago), provides a great example of how understanding chemistry can help one make money in the stock market.

    1. William Gerber says:

      Dear BiotechReader, what you wrote above sounds a lot like what I read a few years ago and may be the kernel of doubt that caused me to look into this further and sell my position. For that, (assuming it was you) I thank you. I always assumed, maybe incorrectly, that if Threshold really wanted to know if the compound worked as advertised (hypoxia activated), they should have figured out some way (in Phase II preferably) to test tumor hypoxia levels pre and post treatment in TH-302 arm and control arm. Instead, they showed those mice slides for about 4 years. My guess is that they didn’t want to know if it really worked in humans as advertised. Given the size and power of the trial, my guess is that they saw the positive trend in the EORTC and other trials w/ Dox + Ifosfamide that did not reach statistical significance and thought that they might eek into significance with a larger trial and get approval for what is essentially Ifosfamide.

  3. Morten G says:

    How are oxygen levels in bone marrow?

    1. Scrub Ninja says:

      Morten: Bone marrow is not a low oxygen environment. It has extensive vasculature and a strong blood supply.

      In fact, bone marrow is so highly vascular that we can use it as an entryway into the bloodstream. In a medical emergency, if we can’t rapidly obtain IV access the normal way, we can get access by sticking a big needle into a bone. Infusing fluid and drugs into the marrow space gets them into circulation just as fast and as completely as injecting them into a vein.

      1. fajensen says:

        A…AAaaaaaahgh! Next up is the eyeballs!!

  4. Bon Vivant says:

    As an executive at Threshold I was able to steer quite a bit of money given to the company, for research, my way, so who are you to say anything about the value of such an enterprise.

    1. BeenLongInTheIndustry says:

      So, did any promising drug candidate ever come out of the other projects that you spent this money on?

  5. JBstein says:

    the tumor mouse model tells you as much about the clinical chances of a treatment as the number plate of your car says about it’s ability to go from a to b, however, no mouse model pictures available just mean you have no engine under the hood in the first place …

    1. Falanx says:

      It tells you very much less than a numberplate. Numberplates tend to be assigned in batches to batches of identical cars off the production line, so you’ve got an idea what is under the hood, in the ECU and plastered all over the dash and seats from the plate…

Comments are closed.