December 30th, 2024
Here for your holiday cheer is what I expect will be the penultimate lifespan-reporting update on RMR1! Just four mice have made it to 2025 (yeah, I'm 55 hours optimistic there...), all of them male, and once they die I will launch into the next phase of updates, which will consist of lots of statistical analyses of the survival curves, followed by various reports on our healthspan data.
For now, though, I will continue the habit of highlighting some of the qualitative impressions that the survival curves are telling us. First let's recall that there are really big sex differences in terms of the impact of the various interventions. Thus, as in a few of the previous updates, I'll discuss the two sexes separately.
First let's look at the females, which are somewhat easier to interpret. The most conspicuous feature is a rather disappointing one that I've mentioned before: most of the signal seems to be from the one non-damage-repair intervention, rapamycin, in that the mice receiving only rapa have a survival curve almost identical to that of the mice receiving everything.
But one beauty of our study design is that that's not the only relevant comparison when we're asking about how much a given intervention contributes: we can (and must) also look at how the group getting all three OTHER interventions did relative to the group getting nothing. And here we have the highly intriguing result that the all-but-rapa group reached the 50% survival mark just one week younger than the all-four group, but then fell off a cliff and went extinct at pretty much the exact same age as the no-treatments group.
This screams that the damage repair interventions WERE working, but that their benefits waned after a year, hence that in future studies we should look closely at repeating such interventions, maybe every 6-9 months. (Recall that whereas the damage-repair trio were only administered at the start of the study, rapa was in the mice's chow throughout.) Also, it has been reported that even a rather brief period of rapa supplementation extends mouse lifespan - but that's a separate issue.
Now for the males. First, let me highlight something that I am pretty sure is a pure case of statistical scatter, with no actual meaning: that the last six mice in the all-four group all died within just a couple of weeks of each other. That's the group that one would initially have bet on as giving the oldest mouse in the study (since males of this strain live longer than females). But that's why we use 50 mice per group, and honestly we'd be happier if we could use 500... with this many groups, it's actually more likely to see one or two anomalies of this sort of magnitude than not to.
So, setting that aside, let's look again at rapa versus the damage repair interventions. Even though the last three rapa-only males outlived the entire all-four group, the main message is that in contrast to females, rapa mice were doing considerably worse than the all-four group for most of the study. Additionally, the all-but-rapa group behaved in the exact opposite way for males than for females when compared to the no-treatment group: it actually did a lot worse than the no-treatment group at first, but then the five mice that were alive nine months ago did amazingly and one of them is still kicking. Now, again that may be a statistical anomaly, of the opposite kind than the one I just mentioned for the all-four males - I can't yet rule that out, and in future posts I will be providing the rigorous statistical analysis to answer such questions.
We also need to ask why Gal-Nav and mTERT, each one on its own, did so very badly. Their underperformance relative to no-treatment controls was enough that I can't put it down to statistical noise, but also we must remember that nothing of the kind was seen in females. One thing I'll be looking closely at in that regard is mock versus naive controls, to see whether the delivery vehicle can give an answer - again this is not just a two-group comparison, it's a slice through the data from several groups. Watch this space, for that and many other analyses!
