Skip to main content

Alzheimer's Disease

A Prospective Alzheimer’s Trial Reports

For the past several years, a clinical trial from Washington University (St. Louis) has been underway in people with genetic mutations that lead to early-onset Alzheimer’s. The Dominantly Inherited Alzheimer’s Network (Trials Unit), DIAN-TU, has been dosing 194 such patients with one of two anti-amyloid antibodies, either Lilly’s solanezumab or Roche/Genentech’s gantenerumab (or placebo), and looking for signs of slowing cognitive decline. Here’s a blog post back when that started.

Results were announced today, and sadly you already know what they are. No effect for either drug, the doses of which were increased over time. I’ve written several times about solanezumab here over the years (most recently here), and it has an unbroken record of clinical failure. Lengthy, expensive, complete clinical failure, and this latest trial is more of the same. Gantenerumab is the same: never a sign of benefit. The unusual thing about this trial was that it was a preventative effort in people whose genetic background makes their development of Alzheimer’s virtually certain, and was thus started in patients very early. Another such effort has been going for some years now in a patient population in Colombia, with another Roche/Genentech antibody (crenezumab). Baseline data have been released from that one, but (to my knowledge) no treatment results. Odds of a beneficial effect of treatment must at this point be rated as extremely small.

I really don’t know what else to say at this point. I’ve been writing this blog since 2002, several years after my own experiences with Alzheimer’s drug discovery, and ever since then I have chronicled the failure of drug after drug aimed at the amyloid hypothesis. Different mechanisms, different dosing, different patient populations, different trial designs: none of it has worked. Not once. I myself include the most recent Biogen/Eisai antibody efforts in that list – yes, the one that they’ve submitted to the FDA and are trying to get approved, a decision about which I have made my position abundantly clear.

Today’s announcement doesn’t bear directly on the Biogen/Eisai results – different antibody, different patient population – but it does dump another shovelful of cold dirt on the entire idea of treating Alzheimer’s disease with such amyloid-targeting antibodies. Nothing has ever worked. Never once. So when another effort comes along with another antibody, whose initial results were a clear failure (as stated by the investigators themselves), is this a surprise? Not at all: the surprise was when Biogen came back around and said that gosh, a re-examination of the results makes them think that they stopped the trial too early, that if you stand at a particular angle at a particular time of day it almost looks like something was starting to work. Squint, they say. It’s there.

As that last-linked blog post makes clear, I don’t find that evidence compelling, and most certainly not compelling enough to approve the drug as it stands. But with the way the FDA is behaving these days, God only knows what’s going to happen. People are investing on just that basis, although rarely stated in quite those words. But if you can look at the ruins of all the attempts to show efficacy with anti-amyloid antibody therapies and decide that the Biogen/Eisai data represent a real ray of hope, you have a sunnier disposition than I do.

47 comments on “A Prospective Alzheimer’s Trial Reports”

  1. anon says:

    Biogen has not yet submitted their BLA for aducanumab

  2. Dr. Viktor Frankenstein says:

    That is seriously bad news. Breaking up already existing amyloid plaques in older patients is obviously a major challenge for antibodies, but in these cases it should have been easier to stop the amyloid at a really early stage …
    Presumably at this stage all we have is a deteriorating mental ability of patients who are still alive. It will take decades to see if their amyloid deposits look any different from untreated patients. Who knows, maybe they are plaque-free (i.e. the antibodies did their job), but this is not the actual cause of the disease? Something interesting may still emerge from this trial in the long run!

  3. James Millar says:

    Given that targeting amyloid is looking like a dead end at this point, is there anything that looks like a reasonable next focal point?

    1. bernie says:

      I thought there is some consensus emerging that amyloid is inert or protective and that the oligomers are the toxic species causing cell death. Valerie Daggett has proposed an alpha-sheet structure for the oligomers that I find very interesting, but I don’t know how widely accepted her theory is.

      1. Harrison says:

        Unfortunately the soluble oligomer hypothesis is an example of the goal-post moving that has occurred in the amyloid field over the past 30 years. Amyloid does not correlate with cognitive impairment. In contrast, the first studies using tau imaging suggest better correlation with cognition. There are several on-going trials targeting tau, so the verdict is still out.

    2. James Millar says:

      Thanks for the answers. This is totally foreign to me, I don’t have any scientific background, but I find it intensely interesting.

      What’s a little disheartening is that it sounds like both LLY and RO have every intent to keep on pushing at this (from what endpoints quotes at least)

    3. Crocodile Chuck says:

      ALZ is like most other pathologies: a infection [chronic]

      Keep your Blood: Brain Barrier healthy!

    4. JasonP says:

      There are plenty of targets for treatment. If one is interested in learning go read the research that gets reported on at Seems like the solution is going to be a multi-faceted one and not just a “one pill” silver bullet. Here is a summary of the attempts on ‘the line to gain’ in 2019.

      Some of the most recent results suggest that chaperones are a new target. Others are focusing on microglial dysfunction/regulation. Appears as though increased funding by Congress as initiated more basic cell science research.

    5. Dionysius Rex says:

      Right now I’d be backing startups in the virology (HSV) field.

    6. Binky says:


  4. luysii says:

    This is very sad news, particularly so for a good friend and classmate whose wife of 60 years thinks he’s her father.

    Which brings up SERF1A (in man) and MOAG-4 in yeast. SERF1A is clearly doing something important as it has changed little from the humble single yeast cell to man. There are only 68 amino acids in SERF1A of which 14 are Lysine, 9 are Arginine, 5 are Glutamic acid and 5 are Aspartic acid. That’s interesting in itself, as we have 20 different amino acids, and if they occurred randomly you’d expect about 3 – 4 of each. The mathematicians among you should enjoy figuring out just how improbable this compared to random assortment. So just four amino acids account for 33 of the 68 in SERF1A Even more interesting is the fact that all 4 are charged at body pH. This means that positive and negative can bind to each other (something energetically quite favorable). How many ways are there for the 10 acids to bind to the 23 bases? Just 23 x 22 x 21 X 20 X 19 X 18 x 17 x 16 x 15 x 14 or roughly 20^10 ways. This means that SERF1A doesn’t have a single structure, but many of them. It is basically a disordered protein.

    Several conformations of SERF1 are seen in solution, and there it binds to Abeta monomers and oligomers forming a ‘fuzzy complex’, in which the number of Abetas and SERF1s are not fixed — e.g. there is no fixed stoichiometry — something chemists are going to have to learn to deal with.

    So what does SERF1A do? It increases the rate at which Abeta monomers and oligomers form the nucleus of the amyloid fiber. It does not help the amyloid fiber grow. This means that the amyloid fiber is good and the monomers and oligomers are bad, otherwise why would natural selection keep SERF1A around so long and so unchanged. Not only that, but SERF1A has exactly the same effect with alpha-synuclein, the main protein of the Lewy body of Parkinsonism.

    This is likely why the billions and billions spent by big Pharma trying to get rid of the amyloid fiber haven’t helped Alzheimer’s disease (and in some cases have made it worse). For more detail please see —

    1. ScientistSailor says:

      “That’s interesting in itself, as we have 20 different amino acids, and if they occurred randomly you’d expect about 3 – 4 of each”

      I would never expect amino acids to occur randomly, or even their frequency of occurrence to be evenly distributed. Seems like an unnecessary point that distracts from your argument.

  5. Ed says:

    This one is especially disheartening since it’s aimed at the genetic forms of the disease, which is where the amyloid hypothesis was strongest. Or to be more precise, there’s no question that amyloid is the root cause of the disease in these patients. After all, most of them have mutations in APP that increase the production of more toxic forms of ABeta peptides. And anyone who inherits these mutations gets the same disease. So the causal amyloid hypothesis is very well confirmed for this genetic form of Alzheimer’s disease at least.

    However, it seems to be the final nail in the coffin for the therapeutic amyloid hypothesis. Even when patients have a disease because of mutant amyloid, clearing the plaques with an antibody does not have any clinical benefit.

    1. Andre Brandli says:

      “It seems to be the final nail in the coffin for the therapeutic amyloid hypothesis.”

      You cannot make this conclusion based on the data presented. The data indicate that the two antibodies, gantenerumab and solanezumab, were ineffective in slowing disease progression in patients with a genetic predisposition to develop early-onset familial Alzheimer disease. Importantly, the patients were treated for at least 5 years, in some a cases for up to 7 years with one of the two antibodies. In addition, the treatment doses were increased during the course of the clinical trial.

      In my opinion, the data are unequivocally the final nails in the coffins of gantenerumab and solanezumab as therapeutic agents to treat AD. They do however in no means disqualify the amyloid hypothesis of AD. The hypothesis is strongly supported by clear-cut genetic evidence, where mutations in APP, PSEN1 and PSEB2 or trisomy 21 (with an elevated APP gene doses) predispose to early-onset AD. Failure to acknowledge these facts will lead to more failures in AD drug discovery.

      1. James Millar says:

        The drugs’ parents seem to disagree with even that opinion!

      2. Red Agent says:

        The nails have long been in the coffin, and the coffin buried for years. This is another failed attempt at resurrection.

      3. Harrison says:

        The problem with the amyloid hypothesis is that the patients with the familial APP-related mutations represent somewhere between 1-5% of all Alzheimer’s patients, and may not be representative of the common, late-onset form of the disease. APOE4, the most common genetic risk factor may account for 60%+ of late-onset patients and there is a clear link between APOE4 and amyloid accumulation, although APOE4 could be acting through multiple mechanisms unrelated to amyloid. The real question is what is leading to AD in the other 35% of cases?

      4. AOM says:

        Also remember the ever important “correlation does not always mean causation.” Especially true with anything biological.

      5. Ed says:

        I’m 100% with you on the weight of the genetic evidence for amlyoid as a cause of familial AD. I was trying to make the distinction between a causal hypothesis (amyloid build up causes the disease) which is incontrovertible for the familiar forms of the disease, and the therapeutic hypothesis (getting rid of amyloid via a mAb, protease inhibitor, etc. will treat the disease).

        Still, you’re right to point out that this trial only indicates the clinical failure of two particular amyloid mAbs. I had conflated this new trial with the previous trial showing aducanumab cleared plaques but did not alter disease progression, in non-familial AD. If this trial shows that gantenerumab and solanezumab don’t clear plaques, there’s still hope that aducanumab might be effective. But prospects will be much more bleak if the current trial shows plaque clearance but no clinical improvement for patients with familial AD.

        1. Andre Brandli says:

          @Ed: We are on the same page regarding the interpretation of the failed DIAN-TU clinical trial. As mentioned in my comments below, we currently do not know whether the antibodies clear the plaques building up in the patients. My prediction is as good as yours…..

      6. stistica says:

        If you test 20 different mAbs antibodies against amyloid at alpha=0.05, what do you think is the chance to commit at least 1 type I error?

        1. DrSAR says:

          I’d suspect it’s not 100%, or do you?

    2. luysii says:

      In one way it IS the last nail in the coffin for ‘the therapeutic amyloid hypothesis’, by which is meant ‘removing amyloid is therapeutic’, but not if ‘the therapeutic amyloid hypothesis’ is read literally e.g. amyloid is therapeutic — which is the point of [ Proc. Natl. Acad. Sci. vol. 116 23040 – 23049 ’19 ] to which I referred earlier. If the implications of the paper are correct this is a 180 degree therapeutic paradigm shift which is totally unexplored.

  6. loupgarous says:

    “As that last-linked blog post makes clear, I don’t find that evidence compelling, and most certainly not compelling enough to approve the drug as it stands. But with the way the FDA is behaving these days, God only knows what’s going to happen. People are investing on just that basis, although rarely stated in quite those words.”

    Derek, I’m trying to find the exact quote from CDER head Janet Woodcock about Sarepta’s stock price being a factor in the eteplirsen approval, only finding references to it by you and other commenters.

    Tthere’s already plenty of investment buzz <a href=" around Sarepta regarding their third drug candidate for treatment of Duchenne’s muscular dystrophy, casimersen, which skips exon 45. Several pharma blogs like this one mentioned casimersen, and Sarepta’s claim that they’ll be covering 30% of DMD patients with eteplirsen, golodirsen and casimersen. Not all the buzz about Sarepta and casimersen is good: Barron’s ran with this headline on January 23rd: “Sarepta Got a Scathing Rejection from the FDA. That’s Why Its Stock Is Falling.”

    Yet, if you look back, similar articles in investor news blogs ran during the FDA application/rejection/rescue by Woodcock and Stein process surrounding eteplirsen and golodirsen. Would it be a terrible surprise if, in a few weeks, CDER overruled scientific opinion once more and approved casimersen?

    I wouldn’t necessarily call this a “pump and dump” operation, but the stink of regulatory capture at CDER is strong.

    This repeating pattern of application/rejection/appeal/approval of Sarepta’s DMD drugs benefits people who bought Sarepta just after “scathing rejections”, held on to it, and find they made money by betting on a kinder, gentler CDER that doesn’t base its decsions purely on scientific data. The investment press now recognizes the FDA has gone from being purely a regulatory agency with a mandate to protect the public from drugs that promise and don’t deliver to an agency with a role in advocacy for approval of marginal drugs like Sarepta’s exon-skipping drugs.

    An investigation into who at CDER, Congress and the White House has invested in Sarepta, and when they did it might be in order.

  7. ScientistSailor says:

    I hope these patients we followed with imaging to show if plaque growth was inhibited…

    1. John Wayne says:

      Yes, I agree – it would be great to know the plaque levels in these patients.

      While it is clear that amyloid is somehow related to Alzheimer’s disease, therapeutic intervention hasn’t worked well. If I was one of those people who had 250-500 million dollars and a desire to impact dementia, it would be hard to convince myself to put it towards anything amyloid-related.

  8. DanielT says:

    Is there any indication that these antibodies slowed the build up of beta amyloid? From the abstract Derek linked to it seems only the results of the primary objective (cognitive decline) were released. It would only be the death of the beta amyloid hypothesis if there was significantly less beta amyloid in the treated patient population, otherwise it is just evidence these antibodies don’t work.

    1. Andre Brandli says:

      Daniel, you are raising the KEY question regarding the DIAN-TU clinical trial. At present, we only know from the press releases that the primary objective, halting or slowing down cognitive decline, was not met in this 5-7 year long trial. The analysis of secondary outcome measures, including the extensive cerebrospinal fluid testing, magnetic resonance imaging sequences, and amyloid PET and tau PET scans should be reported later this year. They may hopefully answer your question.

      If there were no evidence for significant clearing of amyloid plaques, then we can conclude that these two antibodies are ineffective at removing plaques from the brains of treated patients. This could be attributed either to poor penetration of the blood-brain-barrier or the inability of the antibodies to engage with the target in vivo. For example, the amyloid plaques might be too extensive or compact for antibody-mediated removal. In either case, it would indicate that these anti-Abeta antibodies are ineffective at target engagement.

      To to conclude that the amyloid hypothesis is dead on the basis of the failure of gantenerumab and solanezumab in clinical trials is foolish.

      By contrast, Biogen’s aducanumab has been shown to clear plaques from patients brains. Unfortunately, Biogen’s aducanumab drug trials (ENGAGE, EMERGE) were not performed with patient populations prone to early-onset familial AD due to pathogenic mutations in APP, PSEN1 or PSEN2. This would be, in my opinion, the most stringent test for proving (or disproving) aducanumab’s efficacy in AD treatment. The FDA should consider this option in responding to Biogen’s anticipated filing for aducanumab approval later this year.

  9. Gretchen says:

    If you get a computer virus, one thing antivirus programs do is to lock the virus up in a virus vault. Maybe plaque is like a virus vault. Its concentration would track with the disease, but it wouldn’t be causative.

    1. luysii says:

      Gretchen — that’s exactly the point of [ Proc. Natl. Acad. Sci. vol. 116 23040 – 23049 ’19 ] referred to earlier

  10. anon says:

    Given all of this, why are companies still spending money on degrading amyloid using protacs?

    1. John Wayne says:

      If you stumble upon something that works, you will also make >$10 billion a year. High risk, high reward.

  11. Barry says:

    Until we understand the etiology of Alzheimer’s we can’t even know that these familial forms are the same disease. We may eventually cure the familial form and learn that it merely shares symptoms with AD itself

  12. loupgarous says:

    Inflammation has been claimed to correlate to disease progression in Alzheimer’s disease. Amyloid plaques are correlated to infective processes or other inflammation in the brain. Addressing infections in and near the brain has been proposed as a treatment (in other words, “stop what triggers formation of amyloid plaques”). It’s a theory that ought to get some of the attention and resources which the amyloid hypothesis has grabbed. It would have the advantage of being able to be studied with drugs which are handy and relatively cheap.

  13. MarkNelson says:

    Loupgarous is right. Neuroinflammation is enacted by microbes, from bacteria to viruses, and the abeta forms mesh networks around the pathogen, embedding metals which generate ROS that resorb the pathogen as well as the abeta network. These mechanisms were studied by Dr. Robert Moir at Harvard MGH along with Dr. Rudy Tanzi, until recently when Rob died of GBM at an early age.
    I met him last summer and was struck by his science, his passion and his detailed studies (not published) that showed how the microbiome changed over time in autopsied humans from early age to geriatrics- and the fact that a leaky BBB was suspect in the cause of microbial attack and establishment of AD. He told me this was an evolutionary artifact from the days we were coelacanths, and that he was at odds with the antibody crowd on this. He also had a hard time getting funding for his theories, shunned by NIH reviewers although he had support and private benefactors who funded his work.
    I’ve met thousands of scientists over the years, but his theories on AD were the most intriguing, the most detailed, and made the most sense of any I’ve heard to date.

    So go ahead a drive a nail in the antibody coffin, and after reading Dr. Moirs’ work and his descriptions of abeta as an antimicrobial peptide, you’ll wish and hope others will continue these studies and find small molecule therapies that can affect AD.

    1. Isidore says:

      Not an expert, nor a biologist, so please forgive the naivete of the question: As I understand it Robert Moir’s hypothesis was that plaques form in the process of beta-amyloid sequestering pathogens that enter the brain. If this is the case wouldn’t one expect to find evidence of such brain “intruders” in the plaques? Beta amyloid plaques have been studied extensively post-mortem for decades now?

  14. Nick K says:

    What, no post from Lane Simonian?!

    1. Thoryke says:

      Let us not poke at people who are not beating old drums. Reward the behaviour you want to see!

    2. sgcox says:

      Derek and Lane have a gentlemen agreement that Lane avoids provocative AD discussion and Derek let him comment on anything else. It worked very well so let it be, please.

  15. anon says:

    Is it possible that breaking down the larger plaques, or lessening their formation, is wholly ineffective simply because smaller aggregates are much more toxic? Thus you need to stop these smaller aggregates from forming in the first place to see efficacy?

    1. flem says:

      sounds like the cholesterol hypotheses. “its the small dense LDL particles that cause the problem”

  16. Michel Vounatsos says:

    Good thing brogen doesn’t need to show a clinical effect to get its drug approved 👏

  17. Danish Viking says:

    There was talk about GLP-1 analouges as alzheimer drug recently. Does anybody have an opinion about this ?

  18. metaphysician says:

    So, my question:

    What experiment *would* be considered a definitive disproof of the Amyloid Hypothesis? And if no one can come up with such an experiment, wouldn’t that mean it wasn’t a hypothesis in the first place?

    1. Matthew TKK says:

      It’s not a true hypothesis in the Popperian sense because there is no definitive falsification criterion. It’s in a grey area of correlation with some medium quality causative evidence, but because the process of amyloid deposition is itself poorly understood, it’s not possible to test what factor in the deposition process might be causative, for its causality to be tested. So we have a disease in which the critical pathology leading to cognitive decline is not known, and the deposition process of amyloid is poorly understood, and the hypothesis amounts to “something in this fuzz is causal to something in that fuzz”. The best evidence remains that families with genes causing amyloid deposition and/or aggregation-prone forms of Abeta, get early-onset dementia very reliably, and people with trisomy 21 in which an extra allele of the amyloid precursor is also expressed, likewise get early onset dementia very reliably. That’s why this study is concerning – the previous antibody trials tried to clear up after the amyloid deposition, whereas this one was attempting to shift the trajectory of people with this reliable early onset, and didn’t succeed. There are any number of excuses but as an amyloid believer, I really expected this one would show something.

Comments are closed.