Shubhabrata Mukherjee, research associate professor internal medicine, University of Washington School of Medicine.
“The vast majority of researchers studying Alzheimer’s now acknowledge that the disease has multiple contributing factors that involve cascades of cellular and molecular processes that we don’t yet understand. “Alzheimer’s disease, as it is currently understood, is really not a single disease. It’s a number of pathologies that all result in cognitive impairment.”
Costantino Ladecola, neurologist at Weill Cornell Medical College in New York.
If the diagnosis ‘Alzheimer’s’ is an end-stage …
In 2018, Nature Molecular Psychiatry published a study identifying subgroups within patients diagnosed with Alzheimer’s disease, based on substantial patient heterogeneity, both genome-wide including the APOE ε4 allele, as well as with regard to cognition and function, and the location in the brain where disease manifestation occurs.
In November 2021, New Scientist made an overview of - almost - all different theories that have been put forward as causes for Alzheimer’s, demyelination and mitochondrial/cellular dysfunction being absent. Its title is: ‘Why Alzheimer’s is not a single disease – and why that matters’.
Still in November 2021, a Swiss/French study published in Nature Reviews Neuroscience which came to the conclusion that there are not one, but three forms of Alzheimer’s disease based on risk factors, characteristics of their disease and clinical fate. The study concluded that the amyloid cascade theory (amyloid->tau->neurodegeneration->cognition loss) was only confirmed in one of these three groups, where patients carry an inherited genetic mutation known as “autosomal dominant” and develop Alzheimer’s early in life. The second group is carrier of the APOE ε4 allele. The third group, which represents about half of the entire AD population, consists of people with no associated genetic mutation for which the presence of neurotoxic proteins seems to be an important but not unique risk factor.
…with several disease-provoking pathways…
Alzheimer’s is currently considered to develop through multiple pathways. The disease develops differently in individuals, complicating efforts to diagnose early. Using CellDeveloper, researchers have created an overview of pathways involved in Alzheimer’s disease (AlzPath), which comes close to looking like the entire tube network of both London and New York combined.
One can say at least, if there is even one disease (and I believe there isn’t), it’s extremely complex, and the 240 drug trials already from 2002 to 2021 that have gone to waste to prove it.
…then why do biotech/big pharma companies not distinguish between patient populations?
The big question here is: why is precision-medicine so blatantly absent in AD with the science clearly pointing in that direction. I have no answer to that question.
Over the past year, I have investigated many companies in the neurodegenerative and mostly AD space, each of them claiming they have a unique approach to Alzheimer’s. Yet almost each time, these companies seem to lack including the logical consequence of their unique approach, which is the proper identification of subcohorts they would be willing to target with their drug candidate.
Each of these companies cannot have invented the light, which means they are probably omitting something, missing something, and in my humble opinion, either way on their way to probable failure at the end of the road, or at least a substantial part of non-responders.
Cortexyme has changed its stance in 2022… yet that change was the result of a trial failure. It is only after its drug failed that Cortexyme started considering retargeting an AD subpopulation, namely that with gum disease.
INmune Bio has been overtly focusing on the inflamed subcohort of AD patients since at least 2019, when they announced their Phase 1 trial in Australia: “A key element of the small trial is to identify people with Alzheimer’s that are most likely to benefit from the treatment. Enrollment is limited to patients with evidence of peripheral inflammation.” And lately, in its quarterly call, specifically denominates a patient group within AD: “The study showed that XPro 1 milligram per kilogram once a week as a subcu injection, decreases neuroinflammation in patients with ADi, that's capital A, capital D, small i. ADi is the term we have coined for patients with AD -- with Alzheimer's disease who have biomarkers of inflammation and neuroinflammation. The Phase 1 trial demonstrated downstream benefits of decreasing neuroinflammation, including decreased neurodegeneration, or nerve cell death, improved synaptic function, arguably the most important target in Alzheimer's disease and remyelination. We’ve provided anecdotes of improved cognition in the Phase 1 trial but definitive evidence of the effects of XPro on cognition in patients with ADi awaits the results of the blinded randomized Phase 2 trials. The company understands it must deliver clinically relevant data in these trials. We have presented detailed descriptions of the ADi Phase 2 trials previously. Here I will highlight the two unique aspects of those trials. Both use enrichment strategies to enroll patients, and both use EMACC that's capitalized E-M-A-C-C, to test cognition. EMACC stands for early Alzheimer's disease cognition composite. EMACC is ideally suited to measure cognitive changes in patients with MCI and mild AD.”
I see none of the other AD companies make such distinction. Even though other companies target other alleged pathways of Alzheimer’s, they don’t follow suit, and do not limit their trials to patients who may preferentially benefit from their treatment. Annovis, Anavex, Cognition Therapeutics, Longeveron, Synaptogenix, Cassava Sciences, Athira Pharma, Alzamend Neuro, all target the entire AD population, even though their therapy stories are so diverse that you would sometimes think that they are treating a different disease. And what if they were (as all the more scientists are now saying – cfr. supra)?
Even Biogen did not do so with Aduhelm, nor did Roche with gantenerumab, or Eli Lilly with donanemab, even if they must have read the above articles on the limitations of the amyloid approach as well.
Would it be that the call for identification of subcohorts in this massive multibillion dollar market would be too much to ask for biotech/big pharma management? Or is it perhaps that they do not know better, and think their solution is all-inclusive, in blatant disrespect of the current scientific views?
Alzheimer’s disease drug development largely remains in baby shoes….
In a way, identifying subcohorts correctly and targeting them would make it possible for the AD biotech and big pharma investing space to step out of its baby shoes and move on to adulthood.
If one sees the oncology space as having reached adulthood, Curis - a company I believe currently comes at a bargain price after a recent much oversold clinical trial hold - could serve as an example. Curis targets primarily the KRAS- and FLT-mutational subcohorts in AML, having launched different trials whether or not in combination with standard of care. That’s specific, and if it weren’t for a recent toxicity issue and ensuing clinical hold, they were well on their way to achieve something (and may still be).
In major depressive disorder, after seeing the blockbuster SSRI’s limitations, with one third of patients not being helped by these drugs, the field is also looking for alternatives, such as the many companies involved in the psychedelic space. There, also, science clearly points in the direction of targeting inflammation in a subgroup of patients who have elevated markers of neuroinflammation (read the book ‘The Inflamed Brain’ by Prof. Ed Bullmore if you want to have that science in a more popularized version).
Even Parkinson’s disease drug development seems to be somewhat more evolved. The need for precision medicine here is clearly called for, with some subtypes having been identified. Mission Therapeutics, for example, targets the Parkin pathway, which is mutated in certain Parkinson’s disease patients. Both Gain Therapeutics and Eli Lilly through the acquisition of Prevail Therapeutics target Parkinson’s patients with confirmed mutations in the GBA1 gene. Still in PD, Denali Therapeutics, Cerevel Therapeutics and Neuron23 target the LRRK2 mutation, with Neuron23 using a companion diagnostic to select the patients most likely to respond to this treatment.
In Alzheimer’s disease, I only see INmune Bio leading the way, using the following companion diagnostics: two or more biomarkers of inflammation, either CRP>1.5 mg/L , ESR>10 sec, HgbA1c>6%, ApoE4 positive, thereby targeting about 50% of the total AD population. I don’t understand why other AD companies aren’t doing so. Do they seriously believe they are targeting the entire AD population? Do they bring sufficient scientific results to underpin that? In most cases, I don’t see much of that underlying science, if already it is not the object of heavy debates.
… would any trial not distinguishing between AD subgroups not be at high risk of failure?
Cortexyme must have known, at least halfway through their trial, that they wouldn’t succeed in all patients, and hence any statistical significance they would want to obtain would be at risk due to a patient group not responding to treatment and essentially doing no better or worse than placebo. Unfortunately the company couldn’t tell its investors at that point that this could happen.
What if Cassava Sciences’ simufilam does not solely target filamin A, and could affect an additional pathway. After all the fuzz around Cassava Sciences’ science, what if simufilam’s actual mechanism of action was in part undiscovered, in that it would e.g. be restoring cellular housekeeping, miR-132 expression, phosphorylated tau, or addressing remyelination, or neuroinflammation. Its trial denomination suggests that Cassava Sciences is rethinking and refocusing Alzheimer’s, but what if it needs further, or different, rethinking?
If I then see that XPro puts down more favorable credentials - in my eyes – targeting the inflamed subgroup and not targeting filamin A at all, then that gets me very suspicious.
I have similar ideas on other companies.
My view on things
Alzheimer’s drug development remains in baby shoes, and I believe even the Parkinson’s space is more advanced.
Not distinguishing or identifying sub-cohorts in trials means in practice the drug should work in all AD patients, which for me goes against many peer-reviewed publications in Alzheimer’s, makes no sense looking at the science from a broader scientific perspective and does not feel prudent. Precision-medicine exists in every branch of medicine, why wouldn't it in the disease we don't seem to be able to address?
Certainly in placebo-controlled randomized trials, the risk is that a non-responding sub-cohort diminishes the overall treatment benefit of the cohort that is being helped by the drug, if any.
In conclusion, if any trial not distinguishing between subgroups in AD is more logical at first sight, with insight and research progressing I feel such trial may include non-responding patients. I would feel comfortable to say that, in the current state of the science, such trials are high-risk. They may not follow the latest state of the science, their trial results may be influenced by non-responders, and they are contrary to the normal evolution of science and drug development.
This is, of course, my personal opinion, which ought not to be taken for investment advice.