Cassava Sciences: Trying To Manage Expectations (NASDAQ:SAVA)

An explanation is needed as to why the first group of fifty trial participants on Cassava Sciences’ (NASDAQ:SAVA) simufilam experienced a 3.2 improvement in ADAS-Cog scores while the second group experienced a .2 point decline. Critics have asserted that at least some of those in the first group did not have Alzheimer’s disease. Cassava Sciences advocates apparently taking their cue from the company now have tried to flip the script by arguing that some in the second group did not have Alzheimer’s disease but instead had some other form of dementia or mixed dementia, such as Alzheimer’s disease and vascular dementia. That is why they did not perform as well.

It is impossible to resolve this controversy with the data at hand, but there are clues to a possible answer. According to Cassava Sciences, the first group had a mean ADAS-cog score of 15.5 and a mean MMSE score of 22.1, but a mean ADAS-cog score of 15.5 corresponds to a mean MMSE score of 25.5 (table one). An MMSE score of between 21 and 24 is usually indicative of mild Alzheimer’s disease. Some of those with higher MMSE scores in the first group may have been allowed into the trial based on being amyloid positive. But being amyloid positive is not the same thing as having Alzheimer’s disease. Individuals with strong antioxidant levels in their brain can have significant amounts of amyloid and not have Alzheimer’s disease (study). It is likely, then, that some people in the first group had mild cognitive impairment, very early Alzheimer’s disease, or were cognitively normal.

The results from the second group of 50 patients are even more difficult to interpret because there is no public record of their baseline scores and characteristics: MMSE scores, ADAS-cog scores, age, gender, ethnicity, APOE4 status, etc. Nothing. The contention made by some of the company’s backers is that some individuals in this group did not go under intensive testing to determine if they had Alzheimer’s disease. If that is the case, then some participants were admitted into the trial based solely on their MMSE scores. Just like amyloid, MMSE scores alone cannot be used to provide a definitive diagnosis of Alzheimer’s disease. So it is possible that a few of these individuals had either some other form of dementia or mixed dementia. But even if this is the case, treatments that lessen oxidative stress in Alzheimer’s disease are likely to have at least some benefits in terms of cognition for those with other forms of dementia (possibility).

Knowing the actual MMSE scores of the trial participants in both groups would not completely resolve this debate, but it would go a long way in that direction. If there is a significant gap in terms of baseline MMSE scores between the two groups, this would strongly indicate (but not prove) that more people in the second group had mild Alzheimer’s disease than in the first group.

If the .2 point decline is reflective of the effects of simufilam for mild Alzheimer’s disease, this would put it right in line with acetylcholinesterase inhibitors for mild Alzheimer’s disease (chart). And this may be no accident. Acetylcholinesterase inhibitors such as Aricept inhibit oxidative stress as sigma-1 agonists and are peroxynitrite decomposition catalysts (a nitro-oxidant that probably plays a key role in Alzheimer’s disease). Simufilam potentially at least can do the same (previous article). About 20 percent of the population does not have a functioning sigma-1 receptor, so they do not benefit from this form of treatment. Sigma-1 receptor activity declines with the progression of Alzheimer’s disease, so this type of treatment does not work nearly as well for moderate Alzheimer’s disease as it does for mild Alzheimer’s disease. This may explain why some participants responded better to simufilam than others.

Not only the results but probably the mechanism of action of simufilam is in need of adjustment. Cassava Sciences has always presented itself outside the realm of various other anti-amyloid treatments, but it too is an amyloid-centric approach. The company’s stated mechanism of action is that by restoring filamin A to its normal configuration, it reduces amyloid toxicity by “disabling its signalling” via nicotinic acetylcholine and toll-like receptors (publication). Yet, in trials where nearly all amyloid was removed, Alzheimer’s disease continues to progress. It does not take a very small amount of amyloid to contribute to Alzheimer’s disease, it requires a lot (such as the case with APOE4 carriers). In non-APOE4 carriers who have less amyloid, anti-amyloid drugs don’t work at all (latest example from Eli Lilly).

It is possible that simufilam produces beneficial results even in the absence of amyloid. For instance, simufilam may be a zinc and copper chelator, which may be of some significance in that zinc and copper contribute to the misfolding of proteins as well as to their toxicity. If this is the case, then it is the presence of copper and zinc and not any particular misfolded protein that can at least early on contribute to the progression of Alzheimer’s disease. Eventually, though, copper and zinc are entombed in amyloid plaques, so their role in Alzheimer’s disease declines. It is possible, though, that a small part of simufilam’s early benefit is as a copper and zinc chelator.

The role of misfolded proteins by themselves in Alzheimer’s disease is somewhat uncertain. Misfolded amyloid by itself probably does no damage to the brain. Misfolded tau proteins can inhibit neurotransmissions. It is unclear at this point if misfolded filamin A proteins do any damage to the brain or not.

Simufilam’s mechanism of action appears to only vary slightly from acetylcholinesterase inhibitors. If that is the case, then, its impact should start to decline at 18 months. The results from Cassava Sciences’ 18-month cognitive maintenance study are therefore critical.

Cassava Sciences’ Remi Barbier has been very bullish on simufilam, which is what one would expect from an aggressive CEO (at six months, at nine months, at one year). Recently, though, he has been more vocal in stating that efficacy can only be established via the company’s phase 3 clinical trials. Perhaps, this change in tone is in response to reports of a Department of Justice investigation as to whether the company in the past has potentially misled investors, government officials, and the general public, or perhaps it is in response to middling results in the second group of 50 patients (Cassava Sciences partially buried these results by only providing the average between the two groups rather than the numbers for the second group). In any case, one should take Remi at his word now. Simufilam appears to be a safe drug, but it is not close to establishing efficiency. I would not recommend buying it, and I would only recommend holding it until further trial results are close. The stock has gone up recently as a result of insider buying and the finding by the Journal of Prevention of Alzheimer’s Disease that there was no conclusive evidence that Cassava Sciences had altered critical early data. It may go up further based on additional “exoneration” or fear of missing out, but at some point, the company has to establish efficacy. If I understand the mechanism of action of the drug correctly, that is not likely to happen.