Brain Shrinkage As A Side Effect

 BY DEREK LOWE

I wrote here recently about lecanemab, the latest anti-amyloid antibody being tested as an Alzheimer’s therapy. But there’s another aspect to this drug (and apparently to the other attempts at such antibodies) that hasn’t had as much attention. Madhav Thambisetty, a neurologist at the National Institute on Aging, has an opinion piece at Stat where he lays it out: the Phase II trials of lecanemab showed a notable amount of brain shrinkage in the treatment group as compared to the controls.

This can be seen in that paper’s supplementary information, specifically Figure S4B (changes in whole brain volume) and S4C (changes in ventricular volume). Whole brain volume goes down, while ventricular volume increases, both in a dose-responsive manner. The 10mg/kG bi-weekly dosing schedule (which is what was used in the Phase III trial, and will be used after accelerated approval) was significantly different from the placebo controls starting from the earliest time point and this difference continued to increase out to the last measurement at 18 months.

It’s not just an effect of lecanemab, but the situation is complicated. Donanemab (another antibody, as you can tell by the name) didn't really hit anything clinically useful in trials, but the treatment group showed accelerated brain volume loss (figure 3C in this paper). This was also observed with bapineuzumab in 2016, and indeed with another antibody (AN1792) as far back as 2005. Patients taking crenezumab did show brain atrophy as well, but unfortunately this is not compared to placebo controls in the published data, so it’s impossible to say if it was greater with antibody treatment. Lilly’s solenezumab did not show an effect versus controls, however, and neither was it seen with gantenerumab. To make things more interesting, brain volume loss was also noted with the beta-secretase inhibitors atabecestat and verubecestat, although the latter seemed to be an effect of the first thirteen weeks of dosing and did not progress afterwards (it was the subject of a paper all its own).

Now, there’s a substantial literature on brain volume changes in Alzheimer’s and other neurodegenerative diseases, and you can find a lot of details about various regions. Much of this has focused on hippocampal volume, but it’s interesting to note that (in the examples above where this is broken out) that anti-amyloid therapy doesn’t seem to particularly have a differential effect on hippocampal volume. Several trials have had this as a particular measurement to collect. Total brain volume tends to decrease with aging as well, although the relationship of this to preclinical or undiagnosed neurodegenerative disease is not well worked out. In general, though, it’s safe to say that brain volume reduction is considered to be a pathological sign associated with reduced cognition.

It really does merit attention, and seeing it accelerate with some anti-amyoid therapies is something to think about. An explanation that has been advanced for this finding is the clearance of the amyloid itself, a volume decrease due to the sheer loss of amyloid plaques in the brain tissue. But as one of those references above shows, this makes very little sense. The total amyloid load in the brain of a clinical trial subject has been calculated to be between five and 10 milligrams, and the brain volume changes with drugs like donanemab is several milliliters. Given the density of brain tissue, those figures mismatch by about three orders of magnitude. And the same paper notes that the volume changes don’t follow the same time course as amyloid removal, either. I would add that the fact that no such volume changes have been measured with other antibodies that also seem to remove amyloid is a strong argument against this hypothesis, too.

In his Stat article, Thambisetty advances several questions that really should be answered. First off is of course whether this brain volume reduction tracks with effects on cognition, and whether they have any relationship to known biomarkers of neurodegeneration. Second is whether there’s a relationship between it and the ARIA and ARIA-E effects that are seen on brain imaging during most of the trials. (Some of those lead to localized brain swelling, to be sure, but perhaps things are smaller after that goes down?) And finally, what are the long-term consequences of such therapies as opposed to the effects seen during the clinical trials?

As he notes, we really haven’t seen any of these addressed specifically by either the companies involved or the FDA. If lecanemab is approved and moves into more general use, we would be doing the field a real disservice by not collecting the relevant data.

https://www.science.org/content/blog-post/brain-shrinkage-side-effect