ATF5 mitoUPR pre-print

Our paper on the role of ATF5 in the mammalian mitochondrial unfolded protein response (UPRmt) is now on BioRxiv as a pre-print. It’s also been submitted to a regular journal, so hopefully will be in press some time this millenium.

For the uninitiated, UPRmt is a mitochondrion specific unfolded protein response, much of which was figured out in worms (C. elegans). The key mediator is the transcription factor ATFS1, which has both mito’ and nuclear targeting domains. It is normally made and imported into mito’s and destroyed by proteases. When mito’ proteostasis is upset (e.g., a mis-match in stoichiometry between the mtDNA and nuclear DNA encoded subunits of the electron transport chain, which causes the leftover bits to mis-fold or aggregate), the import is blocked and ATFS1 goes to the nucleus. There, it upregulates a bunch of things to restore mito’ homestasis including chaperones, antioxidants, and glycolysis to take care of energy needs while the mito’s are undergoing repair.

All those things (chaperones, antioxidants, glycolysis) sound like they might be useful to have around in a situation such as cardiac ischemia (i.e., what we study), so we reasoned that activating mito-UPR might be cardioprotective. To induce it in mice, we used oligomycin (yes, that ATP synthase inhibitor) or doxycycline (a tetracycline antibiotic that disrupts mito’ ribosomes – which is a good reason not to use it in experiments). Sure enough, after 6 hours (as expected for a gene program) the hearts were protected against ischemia-reperfusion injury.

Within the mito’ UPR field there’s a debate about what the mammalian ortholog of the worm ATFS1 protein is. Some folks have said the mito UPR is perhaps just a branch of the integrated stress response, mediated mainly by ATF4. This has been accompanied by a general perception in the field that maybe mito UPR is just a worm thing and might not really exist in mammals at all?  That changed a couple of years ago when our collaborator Cole Haynes showed that the mammalian trasncription factor ATF5 can rescue the ATFS1 knockout worm.

So, we tested the UPRmt inducers for cardioprotection in an Atf5-/- mouse, and the protection was lost. To the best of our knowledge, this is the first in-vivo demonstration that ATF5 is a component of the mito’ UPR in mammals (previously it had been shown in cell culture).  We also did some RNA-Seq, trying to find out the pathways induced, and while we did see the classical UPRmt target genes such as hsp60 were up by qPCR, there wasn’t really much going on in the seq’ data – certainly no gene signatures or big pathways. In some ways this is a good thing – it would be kinda boring if we just saw an ISR or Nrf2/Keap or NFKB or any of those other “classical” stress signatures. Instead, it’s likely we’ll have to drill down at the individual target level to really identify the downstream molecular mechanism by which ATF5 is inducing protection.

The other interesting aside to this paper, is there’s quite a sizeable literature on the use of tetracycline antibiotics for cardioprotection, including some human clinical trials. To date, the party line has been that doxycycline confers protection by inhibiting matrix metalloproteases (they’re induced during and after heart attack). But, our data showing a requirement of ATF5 for dox-induced protection suggests an alternative mechanism.

Finally, a note on the practicalities of this operation.  The paper’s lead author, Yves Wang, undertook a Herculean effort just to get the study done, because the Atf5-/- mice are a total nightmare to deal with!  Due to an olfactory neural defect, the homozygotes exhibit ~80% neonatal mortality – the pups can’t find the teat and so can’t get milk. As such, the vivarium bill for maintaining the Atf5-/- colony is currently running about $2500 a month, even with our exceptionally low per-diem rates. Hopefully once this is published we can down-size a bit!