Skip to main content

Alzheimer's Disease

A New Alzheimer’s Approval in China

Edit: tomorrow’s post will be on this subject too, with some information that I’ve been learning today. . .

Today brought the rather surprising news that the Chinese government’s National Medical Product Administration (NMPA) has approved a new Alzheimer’s drug. What’s more, it’s an unusual (and unexpected) mechanism, so this is worth a close look. What’s going on?

A key paper to read is this one from September published in September in Cell Research. That’s a Nature group journal published in collaboration with the Chinese Society for Cell Biology, and the paper describes work from the Shanghai Institute of Materia Medica and Green Valley Pharmaceutical Co. on a seaweed-derived oligomannate preparation (GV-971). The clinical trials for this have been going for a while now, although the whole effort, it’s safe to say, has not gotten very much attention (here’s a 2018 article on the Phase III results, which to the best of my knowledge have not yet been published).

Why a polysaccharide and why for Alzheimer’s? The story is that there was epidemiological data that suggested a lower incidence of Alzheimer’s among populations that consumed this variety of seaweed. GV-971 appears to be “a mixture of acidic linear oligosaccharides ranging from dimers to decamers”. So this would be polymers of mannose, beta 1,4-linked, that have some of the hydroxy groups sulfated. These are pretty numerous in some seaweed species – apparently the genus Codium (to pick one) uses such polymers as major component of its cell walls, largely replacing cellulose. These sorts of sulfated polysaccharides have been investigated as antiviral agents, but there’s a huge amount of structural variety in this area, and generalizations could be tricky.

I note that the proposed mechanism of action of GV-971 seems to have gone through some stages. This paper from August talked about its ability to interact with and sequester various beta-amyloid species, while noting that its MOA was still up for debate. The September paper linked to in the second paragraph, though, concentrates instead on a gut-bacteria immunology/inflammation mechanism, which frankly is more plausible in its way than any effect on beta-amyloid. Not least because it’s quite difficult to imagine a mannan polysulfate making its way past the blood-brain barrier. But the gut-bacteria explanation, which quite interesting, does make some leaps along the way. People who have been following this field (gut microbiota influencing human disease) will appreciate that it can be very difficult to prove real connections, and that the number of confounding variables is huge. Still, there can be little doubt that this point that such connections are/can be real – the question is, are they real for Alzheimer’s? Here’s the current hypothesis:

We have revealed that gut microbiota patterns and amino acid-derived metabolites are important for the infiltration of specific types of immune cells, which drives neuroinflammation during AD progression. GV-971 therapeutically harnesses the abnormal production of amino acids, infiltration of immune cells to the brain, and in turn neuroinflammation via remodelling the gut microbiota.

The team uses the 5XFAD mouse model, one in which amyloidogenic mutations have been knocked in (similar to what happens in human cases of familial early-onset Alzheimer’s). Like all AD mouse models, there is room to argue about this one. There is no doubt that these animals develop plenty of amyloid plaques and other brain pathology, but I note that there are ongoing arguments about their memory deficits and how those are measured (see that link above), depressingly similar arguments to the ones I used to hear first-hand when I worked in the AD field myself in the early 1990s. Here’s more on that model, specifically on attempts to get it out of the inbred C57 black mouse strain and into a more genetically diverse population that might recapitulate the heterogeneous human etiology of the disease.

The paper notes significant changes in the gut microbiota of the transgenic mice as they age, as compared to the relative stability of wild-type controls, and they relate that to inflammatory changes seen in the brain tissue (such as the infiltration of immune cells from the blood supply). Treatment with GV-971 did seem to alter this progression, and the treated mice also showed improvement on cognitive tests.

Thus the human trials. As mentioned, the data from these has not been published, so it’s very hard to comment on the NMPA decision. The Phase II trials are apparently discussed in this reference (Chin. J. Pharmacol Toxicol. 31, 459-460, 2017), but so far I cannot find the paper itself, only references to it, so I don’t know how detailed it is, either. Here’s all that’s available on the Phase III, from Green Valley’s press release:

The Phase 3 clinical trial is a multicenter, randomized, double-blind, placebo-controlled, parallel-group 36 week study led by Peking Union Hospital and Shanghai Jiaotong University Medical School Mental Health Center. The study was conducted in 34 Tier-1 hospitals across China. A total of 818 patients with the diagnosis of mild to moderate Alzhemer’s disease  completed the study. The trial was conducted in collaboration with IQVIA (formally Quintiles) and Signant Health (formerly Bracket) among other partners.

Trial results demonstrated that Oligomannate statistically improve cognitive function in mild-to-moderate AD patients as early as week 4 and the benefit was sustained at each follow-up assessment visit. The mean difference between Oligomannate and placebo groups in ADAS-Cog12 Score (a standard cognitive measure commonly used in AD studies) was 2.54 (p< 0.0001), with sustained efficacy from first month of treatment to the end of 9 months of treatment. Oligomannate was safe and well tolerated with side effects comparable to the placebo.

This is a conditional approval, with monitoring for safety. The company says that it plans a multinational Phase III trial next year to support regulatory filings in other countries, so we’re going to have to wait to see how that goes to be able to say anything more.

My take: new mechanistic approaches to Alzheimer’s are very much needed, and this certainly is one of those. But the improvements noted in this trial seem to come on very quickly, which makes me wonder, frankly, if they’re real and reproducible. Alzheimer’s is not a disease where one expects such results, and we have to remember that modulating neuroinflammation has been tried before as an AD therapy, without seeing anything like this. In addition, it’s very hard to comment on the ADAS-Cog scores mentioned, because there are significant variations in the way such tests are administered and I am quite leery of generalizing from a single number in a press release. I am very glad, then, that there’s going to be another large Phase III. If this is real, it could be quite significant, and I certainly hope that turns out to be the case. But I can’t say it’s real – none of us can – until we see more data.


39 comments on “A New Alzheimer’s Approval in China”

  1. Me says:

    The flipside being: If it is false, these people will – literally – be lined up and shot.

  2. John Wayne says:

    Don’t count this out. Decades ago the Chinese published artemisinin and everybody in the west thought it was hilarious (the structure, use in malaria, etc.); now it is a critical medicine used wordwide.

    I do agree with Derek that the phase 3 study is critical.

  3. Lane Simonian says:

    I promised not to comment on any Alzheimer’s entries on this blog until there was a breakthrough against this disease. I consider this to be a qualified breakthrough.

    First of all this is a fair and good review of the drug and its purported mechanism. Green Valley (the company producing this drug) has been under pressure to find a mechanism by which its compound works and the above linked research article is a good first attempt at doing so. The connections between gut inflammation and Alzheimer’s disease are still quite loose, but a few studies are making a relatively strong case for this hypothesis at least.

    But there are other routes to Alzheimer’s disease besides through the gut (through the nose for example). So it is likely that drugs/natural products that inhibit inflammation and amyloid formation in the gut are also doing the same thing in the brain (assuming they cross the blood-brain barrier).

    So if you are searching for a mechanims for how a compound derived from brain algae has a positive effect on Alzheimer’s disease, this is likely the most important one:

    Several other studies using plant compounds affecting the same enzymes and pathways have produced nearly the same results:

    The problem is that many of these clinical trials were open label with few participants. This may be the trial that makes people look at Alzheimer’s disease in a different way since it was placebo-controlled, randomized, and double-blinded with large numbers of participants.

    I have long insisted Alzheimer’s disease can be effectively treated with small molecule plant compounds. Maybe I was one kingdom and a larger molecule away.

    1. tangent says:

      I have long insisted Alzheimer’s disease can be effectively treated with small molecule plant compounds. Maybe I was one kingdom and a larger molecule away.

      🙂 aside from that, Mrs. Lincoln…

  4. Not me too says:

    What will this drug sell for in China?
    How does this impact the plans that Biogen has for their antibody?

  5. luysii says:

    The 5xFAD mouse contains 5 different known familial Alzheimer disease mutations distributed over two proteins (the amyloid precursor protein and presenilin1).

  6. NorthernBioSkeptic says:

    It would be nice to see that 2.54 number in more context. How much decline on ADAS-Cog12 would we typically expect from mild-to-moderate AD patients in 9 months?

    1. Lane Simonian says:

      The charts in this link provide an approximate answer.

    2. electrochemist says:

      The following article (which quotes Derek’s blog!) gives comparative data for Aricept, based on a similar instrument:

      “For example, data supporting the approval of the commonly used Aricept (donepezil) showed patients taking the drug experienced score declines of a relatively similar magnitude on a similar, but slightly narrower cognitive test. Average differences in score changes between two Aricept doses and placebo were 2.8 and 3.1 points, respectively.”

  7. LeeH says:

    Is a 2.54 units of change in the ADAS-Cog12 score a useful effect (as opposed to just being non-random)?

    1. NorthernBioSkeptic says:

      It’s hard to evaluate without more context but it’s not nothing. Even just adding a few months of time before institutionalization would be very much worth it if the treatment is basically seaweed in a capsule with a good side-effect profile.

    2. old neuroscientist says:

      Yes, about what one expect nowadays, although in the beginning (before the available clinical trial population had no treatment options) a drug like donepezil produced ~6 points.

      1. old neuroscientist says:

        forgot to mention that the trial was relatively short at 36 weeks, typically an AD trial would be 104 weeks to allow for adequate decline in the placebo population, so that speaks to the magnitude of effect being clinically meaningful. conversely, the placebo group showed a precipitous drop from week 24 to week 36, which is odd if not concerning.

        1. MrXYZ says:

          The placebo drop really concerns me. Any thoughts on this?

          1. old neuroscientist says:

            a miracle? strange indeed.

          2. Tim Bergel says:

            I wasn’t able to find this out anywhere – was the drop in the placebo group large relative to that 2.54? I would be very … interested if it was

          3. old neuroscientist says:

            Reply to Tim Bergel’s question: YES

  8. Anon says:

    Looks similar to the kind of compounds that Canadian company Neurochem tested – and got nowhere with.

  9. Harrison says:

    I find this phrase from the Endpoints summary a bit worrying: “None of the secondary endpoints — including CIBIC-plus, ADCS-ADL or NPI — were met.”  My first thought on hearing about a drug with an ill-defined mechanism and a trial conducted in a single non-Western country was “Is this another dimebon?”

  10. Skeptic says:

    Somebody has been digging through papers from the lead scientist behind this drug and finding a lot of image manipulation.

    1. Derek Lowe says:

      Followup post tomorrow on this and some other worrying stuff. . .

  11. old neuroscientist says:

    More bizarre to my mind is the clinical data where the difference between placebo and drug on the ADAS-Cog at 36 weeks was 2.54 points (p<0.0001), without concomitant effect on ADAS-ADL or any other functional endpoint. The FDA criteria is a statistically-significant effect on co-primary endpoints comprised of 1 cognitive and 1 functional measure in two independent clinical trials (thus, the data upon which China granted conditional approval would not cut it in the US (or EU for that matter)). It is difficult to conceive of a drug producing that magnitude of effect on cognition without having any effect at all on function. How the hell does that work??!

  12. Simon Auclair says:

    Here is more evidence against amyloid hypothesis and an interesting lead for research:

  13. Andre Brandli says:

    “5xFAD mice express human APP and PSEN1 transgenes with a total of five AD-linked mutations: the Swedish (K670N/M671L), Florida (I716V), and London (V717I) mutations in APP, and the M146L and L286V mutations in PSEN1.”

    The 5xFAD mouse model represents in my humble opinion not a valid mouse of human FAD. 1) Human FAD patients are heterozygous for a pathogenic mutation in either APP, PSEN1 or PSEN2. To my knowledge, no patient carrying APP AND PSEN1 mutations has ever been observed. 2) In the 5xFAD model, the mutated gene versions are not expressed under control of endogenous APP and PSEN1 gene promotors, respectively. Hence, the results obtained from experiments with the 5xFAD mice are by know means predictive for the pathophysiological processes seen in human FAD patients.

    1. Luysii says:

      Andre Brandli — Excellent points. But the mice get sick and all sorts of neuropathology in a hurry, so they at least provide a model on which to try various treatments to get rid of the plaques and tangles. All this assumes that the plaques and tangles are in fact causative and not just an epiphenomenon of an (as yet unknown) pathologic process.

      1. Doctor Jay says:

        I agree. It’s an imperfect model (all AD mouse models are) but at least provides some aggressive pathology to challenge. I think their study was weakened (flawed) by not including drug-treated WT control groups to rule out confounding off-pathway drug effects (cognitive enhancers, locomotion stimulants etc).

  14. Scott says:

    36 months in a human trial for Alzheimer’s seems rather short to me, but I’m not a medical chemist.

    Anyone actually working in the field care to comment about that length of trial?

    1. old neuroscientis says:

      The standard trial is 104 weeks.

    2. Scott says:

      Yeesh, I need more coffee before I post on ITP… that was a 36 *week* trial in humans.

      Smells like snake oil to me, boss.

  15. Another Guy says:

    Just to be the devil’s advocate, if the treatment is basically seaweed in a capsule, wouldn’t we have previously noticed a clinical benefit in patients who eat sushi or other foods that contain seaweed? Even a small difference in the rate of onset or severity of Alzheimer’s disease? I’m not saying this treatment won’t work, and maybe there is something special about how the seaweed is processed, but is there any epidemiological data that might link seaweed-like foods and a lowered rate of neuropsychological disease prevalence?

    1. SomeGuyThatReadThePost says:

      Did you read the post? To quote:

      >Why a polysaccharide and why for Alzheimer’s? The story is that there was epidemiological data that suggested a lower incidence of Alzheimer’s among populations that consumed this variety of seaweed.

      1. Anonymous says:

        SGTRTP: I read that, too, but I can’t access the literature. If those diet studies were done in remote areas with a somewhat stable gene pool, that could surely confound the significance of the dietary claims. It could be a genetically protected population, not a diet-protected population.

        If you could move a bunch of genetically diverse families to those villages (or towns or cities) and blend them in, that could reveal more about the significance of the seaweed diet and other local factors over a single adult lifetime. Of course, if it’s like some other dietary studies, there could be some really confusing epigenetic 2nd, 3rd, and nth generation effects (e.g., starvation diets or high-fat diets that impact children and their descendants).

        From the lay / popular press descriptions, it sounds like a relatively safe trial to run, so I say to go for it. Per usual, make sure it’s done correctly and to high standards.

        (Note to self: more maki and less nigiri on my future sushi adventures, just in case they’re right!)

        1. DrOcto says:

          Probably best that you also take into account the heavy metal content of apex predator fish like tuna then.

          1. tangent says:

            Or it could be the effect of seaweed is from arsenosugars [], to bring this back to some good old arsenic life.

  16. Simon Auclair says:

    Sir! General Heimlich is on manuevers Sir!
    General Alzheimer’s…is uh….um…

  17. Radioactive Man says:


    You mention several times that the compound is sulphated, in particular commenting that it’s hard to imagine a mannan polysulfate crossing the blood-brain barrier. But the structure in the paper doesn’t seem to show any sulfates. Am I missing something? Would the compound have a reasonable chance of getting into the brain without the sulfates?

    1. anon says:

      Good catch! The claimed structure actually has multiple carboxylates:

      Still highly negatively charged at physiological pH, though.

  18. Lane Simonian says:

    I can read the chart better on my cell phone, although I still might not be reading it right. The drop in the placebo group appears to be 1.34 points (from 1.5 to .16). This is not a precipitous drop, in fact it is in line with the historic placebo drop in Alzheimer’s disease during this time period (from 1 to 2 points).

    That the placebo group is still above baseline at 9 weeks is a bit unusual–one would expect about a one point drop in Adas-cog scores in non-APOE4 carriers and about a 2 point drop in APOE4 carriers at this point. The percentage of APOE4 carriers in southern part of China is relatively low so perhaps being slightly above baseline at 36 week is not particularly unusual either.

    Inferences can lead to wrong or premature conclusions. The important question is whether these results hold up over a longer time frame such as two years.

  19. Skeptical says:

    “Why a polysaccharide and why for Alzheimer’s? The story is that there was epidemiological data that suggested a lower incidence of Alzheimer’s among populations that consumed this variety of seaweed.”

    Related to that, is this article (doi: 10.1038/s41598-019-41399-4): Dietary Sargassum fusiforme improves memory and reduces amyloid plaque load in an Alzheimer’s disease mouse model

    The authors postulate an effect by a phytosterol LXR agonist.

Comments are closed.