Skip to main content

Animal Testing

Enough With the Mouse Behavioral Models?

This piece in STAT is well worth a read. The author, Adam Rosenberg of Rodin Therapeutics, is ready to ditch rodent-centric models for human CNS disease, and I can see where he’s coming from. I’ve often said that when I think back on my Alzheimer’s and schizophrenia drug discovery days (back when I was first starting out), and I remember all those compounds I made whose crucial assays were things like whether a mouse ran into the dark half of a cage or not, it makes me want to bang my head on something. The Alzheimer’s work, for example, was literally that: mice in general have an instinct to run into a dark area when they get illuminated. But if you electrify the floor of the dark area so that it tingles their feet, they can learn not to do that. Young mice, though, forget that a lot more easily than adult mice, so our assay for our Alzheimer’s candidates (which were selective muscarnic antagonists) was to give them to young mice to see if they could remember better not to run across the metal strips when the light came on in their cage.

Now, you tell me what relevance that has to Grandma forgetting where the house key is. These were not transgenic mice; they were not suffering from anything remotely like Alzheimer’s. All we knew was that for some reason young mice had less robust memory formation, and that the human disease we were targeting most definitely had effects on memory, so, well. . .Even at the time, I had my doubts, of course, but (1) I was just starting out in the business and wasn’t in any position to tell anyone how to do things any better, and (2) I had no ideas about how to do things better, anyway. I’m not sure if I still do, other than to stop pretending that some of these mouse assays are real just in order to reassure ourselves. And that’s where Rosenberg is as well, it looks like:

To begin, it helps to have this discussion frankly — and to have it with investors, board members, and executive teams. We must all be hesitant to overweight behavioral phenotypes when making key decisions in late-preclinical neuroscience drug discovery. . .

We’ll never know how many compounds were moved into the clinic based on questionable behavioral data. We’ll also never know how many otherwise promising compounds were shelved for failure to show “efficacy” in improving cognition in a flawed mouse model.

He’s suggesting that looks at neuronal circuitry and function (while still quite black-boxy) are still better than trying to infer efficacy from behavioral models, which are a whole level (or two or three) removed. Basically, he’s calling for people to get real and admit that we don’t understand these things very well, but that one thing that we probably do understand is that mouse behavior doesn’t really translate well to human behavior. The sorts of assays that Rosenberg is proposing are a bit fuzzy and hard to interpret, but what do you think we have now? They at least would seem to have a somewhat better chance of translating to something real.

That would be quite a step for a lot of people in the field, though. I don’t know of too many who would stand up and defend the rodent passive avoidance response assay or the Morris swim maze or what have you as great front ends for Alzheimer’s research, but at the same time, they would be rather nervous about abandoning them completely. But by now, I think this stuff really does probably do more harm than good: as Rosenberg says, these models reassure us when we have no real basis to be reassured. In the worst cases, we’re spending time, effort, money, and mice in order to fool ourselves.

This is not a blast against all mouse models or animal testing in general. There are many areas where rodent models are really useful – in fact, crucial – and others where they have their false-positive problems but are still good gatekeeper assays. But human CNS work dealing with higher brain functions (memory, cognition, and behavior) is just not one of those areas. A crap assay is *not* better than no assay at all.

30 comments on “Enough With the Mouse Behavioral Models?”

  1. John Wayne says:

    “A crap assay is *not* better than no assay at all.”

    The choir applauds!

    1. johnnyboy says:

      This is actually one of the most frustrating part of my work – convincing the in vivo guys that their disease model is actually crap and does not show what they think it does. “But it’s been used in all these publications !” is the usual refrain – that and “a crap model is better than no model”. Not their fault, they’ve got Research management who want to see the pretty graphs before they’ll let anything move forward.

      1. Jesse Brodkin says:

        Sounds like it’s the higher ups who want to be confirmed with “pretty graphs” who are the issue here (agree)

  2. cynical1 says:

    Historically, I found that there was an urban legend in the industry that suggested that the FDA would not allow you to take forward a molecule that had not proven itself “efficacious” in these types of behavioral models. In fact, that is simply not true. I remember trying to argue this point to management and it fell on deaf ears.

  3. Jamil says:

    A view from the other side…..

  4. Jb says:

    Rodent studies still give information with regards to ‘Is it safe?’. Safety is a whole different question vs. ‘Will this work?’. In general, many publications over the years have indicated that nearly 70-90% of human toxicities can be caught in animal models when you use one or more species (one usually being a rodent), which is pretty good. This says nothing, however, about whether said investigational article will work or not though. Phase I studies primarily focus on safety. As long as your test article isn’t killing mice, damaging organs or causing other tox, you may be cleared to proceed even without the most robust proof of concept.

    So what’s the solution, go back and open up research to great apes like chimps? Chimps naturally get Alzheimer’s like signs and are the only animal known to IIRC.

  5. Eric says:

    This argument applies to every biochemical assay and every cell-based model that has been used in Alzheimer’s drug development. With a 100% failure rate in the clinic, clearly none of the models used to date have been predictive. It’s not clear to me what alternative Rosenberg is proposing. Until something works in the clinic I don’t know how we can assess a ‘crap’ assay vs a predictive assay. Should we just skip all that efficacy stuff and move everything into the clinic that passes preclinical safety hurdles? That may actually be the best approach right now, but that’s an expensive proposition.

    1. Adam says:

      Eric, thanks for reading and for your comment. I certainly don’t claim to have all the answers, but I did hope to catalyze a conversation. So I’m really thrilled that Derek commented on the piece, and that it has generated considerable discussion both here and across the various social media channels. Most of it has been in agreement, some a bit defensive of behavioral neuropharmacology, but at least people are talking about it rather than ignoring what to many of us seems like an obvious conclusion…perhaps as obvious (if not more so) than over-investment in amyloid intervention.

      Anyway I fully agree with you that until we see true clinical success, none of these approaches are truly validated or predictive. That said, as Derek points out I do believe that some of the newer tools (and indeed some of the older ones) are far better than using rodent behavior as a key go/no-go to assess whether your compound is getting into the brain and having the desired mechanistic effect.

      Just a few examples of these tools: high content imaging; implanted EEG; LTP and other systems biology-based methods, etc. Afraxis ( and some other CROs are doing good work here, and there are many talented groups within academia and industry looking at similar approaches. And while other platforms such as ex vivo / patient derived cell lines and ips neurons are promising from a screening perspective, again I’m not advocating to abandon all in vivo research to assess drug effect; far from it. Rather, I’m proposing that we rely more heavily on quantitative measures to assess molecular, structural, functional and circuitry effects, rather than the behavior that so few seem to believe in, but so many still want to see.

      And while the tools depend on the mechanism, a meaningful translational benefit is that many of these quantitative effects are also then more easily measurable in humans (eg w/ the SV2A PET tracer, or the considerable recent work around fluid biomarkers such as SNAP25, NFL, neurogranin etc). This allows for a deeper view on both target engagement and PD effect.

      So, are the newer tools still black boxy? Yes, for sure. There is obviously much we do not yet understand about the human brain (or the rodent brain, for that matter). None of these more quantitative assays are perfect, and none should be viewed in isolation. But are they also more measurable than rodent behavior, from both a preclinical and translational perspective? I certainly think so. Hopefully, this then leads to an earlier go/no-go based on proof of mechanism – and a more iterative, learning-based clinical trial design – rather than a swing for the fences on clinical endpoints based on non-predictive rodent behavior.

      1. Eric says:

        Thanks for the response Adam. It’s not clear to me that these newer imaging assays are predictive of a desired clinical outcome, but they can be evidence of target engagement. Will that be good enough? I don’t know, but to your point it’s probably better to try something new rather than another mouse water maze study. At least a new approach has a chance of yielding new information. Eventually clinical data will validate or invalidate these models as well. Neuroscience research is a tough field. Drug development in neuroscience is even harder and seems a little bit like playing darts while blindfolded.

  6. In Vivo Veritas says:

    And behavioral neuroscience work remains some of the most challenging work to to well and reliably. It’s way too easy to put you own bias, conscious or not, into the Von Frey hair, the “in or out” status in place preference, on or off the platform in Morris. Did that mouse on the hotplate just withdraw it’s hind limb!?!? With that said, being trained in this stuff and being good at it – means you now officially have reached the pinnacle of in vivo work. Because after CNS work in mice, everything seems automated, regular and a little too simplistic. Most companies have a person they give the tough new in vivo model development work to. Chances are, that person is a neuroscientist.

    1. George says:

      Hi, Paramus,
      Once was an in vivo guy in drug discovery, i couldn’t agree more with you as below sentence…


      “Because after CNS work in mice, everything seems automated, regular and a little too simplistic.”

  7. oldnuke says:

    The replacement for the mouse model is now the investment banker model. Put a couple of slides of a faux biologist up on the wall (wearing black sweaters is cheating) and time how long it takes each one to home in on the slides and toss all of their money at it.

    1. Nick K says:

      Comment of the day!

    2. a. nonymaus says:

      But we’re trying to move from a mouse model towards humans, not towards lemmings!

  8. luysii says:

    Knocking out a gene calls Pals1 produces mice with no cerebral cortex whatsoever. Despite this they can (1) survive (2) ate (3) breed.

    For details
    either read [ Neuron vol. 66 pp. 69 – 84 ’10 ]
    or see

    Along these lines, what made Einstein so good was not his ability to run a maze and find food.

    1. Einstein's fridge says:

      The ability to find food is vitally important to all scientific work, although there are anecdotal reports of survival on coffee alone.

  9. Anonymous says:

    Moe: What are you!? A man or a mouse?!
    Curly: Well, I like cheese.

  10. MrXYZ says:

    What are peoples’ thoughts on animal behavioral assays as a form of preclinical safety assessment? For example, I can think of examples of therapeutic targets that are present both peripherally and in the brain but where hitting the target peripherally leads to a potential therapeutic effect but hitting the target in the brain could lead to neurological effects (behavioral). Obviously one would look for brain penetrance during a preclinical PK study but would a behavioral safety study add value?

    1. Mike says:

      The Irwin (neurobehavioral) study is required before FTiH

  11. Wavefunction says:

    I have also seen pharmacologists and management suffering from the classic sunk cost fallacy in this regard. They have spent so much time, effort and money on creating the transgenic mouse model that they need to now justify it and succumb to wishful thinking by believing it’s real.

  12. Wash says:

    Derek, please! Most of biology is just simply hogwash that not reproducible. Many therapeutics cause more harm than good.

  13. dearieme says:

    “A crap assay is *not* better than no assay at all.”

    Other fields of scholarship espouse the “crap assay” doctrine. “It’s the only evidence we’ve got” they say. “Then you have no evidence at all” says I.

  14. Anon says:

    Each time I read about a “breakthrough cure for Alzheimer’s Disease” in BBC News, it is invariably based on some dodgy mouse model. Now I just shrug and say to myself, “whatever, am I bovvered?”.

  15. anon says:

    The CNS director for a large pharma company once explained to me that the animal models for CNS exist to provide some sort of rationale, likely bogus, for taking a drug into humans. You then watched carefully in the clinical trial to discover what, it anything, the drug did for depression, anxiety, schizophrenia, or anything else that might show up in your clinical population to tell what the drug might be useful for in humans.

  16. franko says:

    Aren’t you neglecting the serendipity aspect? You only have to recall Viagra but in fact many useful treatments were found using an assay for something else.

    In fact that might make for a great post: Treatments that were found while looking for something else. It might be embarrassing for the industry though!

    If you don’t have a better assay, and there is at least some argument for a connection between an effect you are testing for in the lab and an effect you want to see for Alzheimer’s in the clinic, then go for it. You will at least learn something. When you get a better assay then, yeah, go with that.

    1. Russ says:

      I don’t think that should be embarrassing. James Black observed that one way to find new drugs was to look at side effects of existing ones.

    2. Anonymous says:

      Drug Repurposing was a topic In The Pipeline a while ago, 28 August 2014. I’ll just repurpose that blog entry by putting the link in my handle.

  17. Falanx says:

    Seeing as it’s been over a year since people had to start admitting the mouse pain model was garbage…

  18. Paramus says:

    As a behavioural pharmacologist who has worked in Industry for many years, I totally agree with many of the comments. The major problem is that many scientists that carry out behavioural pharmacology studies lack an understanding of what the assays are really telling you. These studies are not ‘models’ of disease, and the sooner that term is dropped, the better. These assays very often only have predictive validity, i.e. It is a test a benzodiazepine or SSRI modifies. That does not mean that non-benzodiazepine compounds will work in the model and it is not a model of anxiety! These assays are an important part of the ‘overall jigsaw’, they allow you to understand part of the PD/PK relationships in a complex organism, which may help predict how the compound will behave in man. It is not just restricted to neuroscience, I once worked on an IBS project, everybody agreed that IBS models were unpredictable at best and ‘crap’ most of the time. When presenting to the investors, they all agreed with that assessment of the models. At the end of the presentation, they wanted to know why we didn’t have any preclinical efficacy data!

    1. George says:

      Yeah, I have same points with you, Paramus.

      For a pharmacologist, behavior in mice is just one of endpoints to understand the target engagement of a chemical in vivo, like PK/PD.

Comments are closed.