iiinova — where your digital self keeps evolving

Payam Gammage @mito-oncogene.bsky.social · 14/08/2026

A series of short perspectives on the emerging complexities and exciting possibilities for mitochondria in health and disease (to which I gratefully contributed two cents on the developments in mitochondrial genetics and cancer). www.cell.com/molecular-ce...

cell.com

Mitochondria with (distinct) personality

Cells owe a lot to their mitochondria—to their many mitochondria. Recent discoveries and emerging technologies point to functional distinctions within that population. We asked a group of researchers ...

030

Payam Gammage @mito-oncogene.bsky.social · 03/06/2026

That’s really kind Iain, thank you. Yeah, almost certainly the connection - I think before eukaryotes the ‘cytoplasm’ appeared to have capacity to make it’s own, but this ability was lost right at the beginning of the eukaryotic pact

010

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

A huge thank you to all of our collaborators and colleagues within @cruk-si.bsky.social @glasgow.ac.uk and @johninnescentre.bsky.social + massive congrats to Flora (currently writing her thesis). Work in our lab was supported by CRUK, @erc.europa.eu NIH/NCI @embo.org

240

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

We hope this gives the community a new framework for understanding mitochondrial biology, and perhaps a way to manipulate mitochondrial content in the future - whether it’s treating mito disease patients, depleting mitos in tumours or helping out exhausted T cells - the opportunities are there.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So, there you have it. ABCB7 & SLC25A39 dictate a regulatory logic underpinning mitochondrial homeostasis that exists across major eukaryotic superfamilies spanning ~1.6-1.8 billion years of evolution

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

PGC1a/b, TFAM, POLG – these guys are not found in plants. TFAM/POLG in opisthokonts, the PGCs only emerge late in metazoa…mTOR/TOR and GCN2/ISR emerge early, but even iron is regulated differently in plants. Somehow, despite all of this, ABCB7/SLC25A39 still regulate mtDNA content.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

That single result holds significant implications. This mechanism, controlling mtDNA homeostasis in both plants and humans, actually pre-dates key parts of the mtDNA replication machinery… it may even represent a mechanism for mito homeostasis dating back to the last eukaryotic common ancestor

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

This Arabidopsis result was the knockout punch for me – plant mtDNA is totally different from metazoan or fungal. The mtDNA replisome is not conserved, there is no POLG or TFAM in plants. Plant mtDNA is megabases in size, with subgenomic circles & linear fragments. There are even transposons!

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

With Alberto and Janneke’s help, we assessed the impact of silencing/partial loss of function mutants across yeast, Drosophila larvae and Arabidopsis seedlings. Same result everywhere. Decreased ABCB7 function = more mtDNA.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

We went hunting through ENSEMBL & OrthoDB to discover that SLC25A39 is almost as deeply conserved as ABCB7 – so we set off to find some collaborators, who we found in the awesome labs of @albertosanzmon1.bsky.social and Janneke Balk, who brought models of flies, yeast and plants to work on

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

At this point we found ourselves wondering how fundamental this might be… a transporter, ABCB7, conserved from bacteria, and its partner transporter, SLC25A39, forming a regulatory circuit integrating multiple inputs? Sounded pretty fundamental to us

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So there’s a single transporter in mitochondria which can coordinate cellular/mito iron, cytosolic amino acid sensing and mitochondrial cysteine abundance, all in the interest of controlling how many mitochondria there are... pretty neat.

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Ok, ok. Time to take a breath and try to tie this all together.

static.klipy.com

ALT:

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

This cystine in mitochondria is probably coming from the imported GSH, and it appeared to be feeding mitochondrial Fe-S biogenesis, as has been described previously by @ginadenicola.bsky.social and Pearce labs in cysteine starvation www.cell.com/cell/fulltex... www.nature.com/articles/s41...

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Now the whole story took a turn, right in the final innings, when we realised ABCB7 depleted mitochondria demonstrate an ~18-fold increase in cystine abundance, happening right at the same moment cytosolic cystine depletion was being sensed by GCN2.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

GCN2 is one of the four canonical ISR kinases, which senses uncharged tRNAs (low cysteine anyone) to shut down cap-dependent translation and engage cap-independent translation of ATF4. So, is the ISR responsible for our change in mtCN? We blocked ISR effector function with ISRIB and confirmed it.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Taking a step back to get the wider view we realised a few other things… the ISR, mTORC1 signalling and Myc signalling were through the roof in ABCB7-silenced cells. DEPTOR depleted ~4 fold, TFEB depleted, Myc elevated ~2 fold. And a GCN2-driven signature emerging.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Turns out it was the cysteine. If you take it out of the medium mtDNA doubles in 5 days, and SLC25A39 KO cells still respond to this stimulus. Friends at NYU recently reported effects on mitochondria when modulating cysteine in the diet too – a long running field www.nature.com/articles/s41...

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So… mitochondrial glutathione accumulation/cytosolic cysteine depletion then, who’s the culprit? Well, we overexpressed SLC25A39 and used mitoGshF (again, hat tip Birsoy lab www.nature.com/articles/s41...) to boost mitochondrial GSH too. No dice.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Was this GSH uptake via SLC25A39 essential to the mtCN modifying function of depleting ABCB7? Turns out yes, if you knock it out, ABCB7 silencing no longer changes mtDNA copy number.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Something we noticed is that SLC25A39, an Fe-S sensitive mitochondrial GSH transporter (chapeau, Birsoy lab www.science.org/doi/10.1126/...) was upregulated in ABCB7-silenced cells, leading to substantial accumulation of glutathione in mitochondria, coupled to cellular cystine depletion.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

We were even able to see some Fe-S cluster ligand cysteines being reduced (no Fe-S) or oxidised (Fe-S bound) following exactly the same pattern, thanks to cool redox proteomics methods from Sergio Lilla and @zanivanlab.bsky.social www.cell.com/cell-reports...

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

In keeping with this all the hallmarks of iron starvation response can be seen at the protein level, and a really nice separation of Fe-S containing proteins began to emerge: for the most part, if it increases in abundance it’s inside the mito matrix, if it decreases it’s outside.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Cells get larger, with more mitochondria, consuming more oxygen and keep on proliferating.

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Turns out that is exactly what happens. Silence ABCB7, iron starvation response engages and iron floods the cell

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So perhaps, if FBXL5 doesn’t get Fe-S clusters the cell thinks its run out of iron? Iron shoots up, Fe-S clusters are made in mitos, but don't leave to stop the starvation response? Classic feed-forward loop, and mitos being constrained by iron seemed appropriately prokaryotic

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Fe-S clusters in the cytosol, via the F-box protein FBXL5, are thought to regulate iron homeostasis in cells by ubiquitination of the mRNA iron regulatory element (IRE) binding protein IREB2/IRP2. When stable, IRP2 compels release of intracellular iron stores and iron uptake into cells.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Sidebar 2 - ABCB7 mutations cause rare human disease (X-linked sideroblastic anemia with ataxia) & subtypes of myelodyplastic syndrome – our first clue came from histopathology of these patients; non-heme iron accumulation in ringed sideroblasts. 10.1182/blood-2006-04-015768

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

In humans, Fe-S clusters are made in mitochondria and used in a diverse array of enzymes (DNA polymerases, nucleotide biosynthesis, you name it) in the nucleus, cytosol etc. So, there’s something important about mito iron/Fe-S clusters… or lack of cytosolic Fe-S clusters? both??

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

A little background on our new favourite gene – ABCB7 is a highly conserved ABC-type, homodimeric transporter, known to transport Fe-S clusters/intermediates conjugated to glutathione from the mitochondrial matrix to the cytosol, a function confirmed from plants to humans.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

ABCB7 is an essential gene, so no joy to be had knocking it out (DepMap had the answer, but we still had to learn the hard way). We validated using siRNA. The results were clear – 2-3 fold increase in mtDNA when we efficiently (~96%) silenced ABCB7 in U20S & HEK cells.

120

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

With hindsight, it was about as clear a result as we could have asked for. The mitochondrial iron importer, SLC25A28, a prominent hit at the bottom end (positive regulator), & ABCB7, the mitochondrial Fe-S intermediate exporter, as our #1 hit at the top end (negative regulator).

130

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

~250 million sorted cells later, the results came in and much head-scratching ensued. Technical controls worked, but nothing slapped us in the face. However, after a week or two, in darkest February 2025, Flora first noticed the signal that led us to a truly incredible answer…

static.klipy.com

ALT:

220

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Now, TFAM is well known to positively correlate with mtDNA levels in cells, so we decided to perform a pooled whole genome screen, sorting the top and bottom 5% of fluorescent cells, looking for genes that regulate mtDNA copy number.

130

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Sidebar 1 – this was rough, most fluorophores we tried aggregated horribly and interfered with TFAM function. Luckily, mCherry did the job, and we isolated a homozygous knock-in clone with beautifully labelled mtDNA nucleoids, validating nicely across the board.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

We set out to ask this question w/o holding too many preconceptions – by performing a whole genome CRISPR screen. To do this we engineered a model where the endogenous mitochondrial histone-equivalent, TFAM, has a long linker, epitope tags and mCherry knocked into its c-terminus.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So, time to get into what we actually did, and what biology taught us about mtDNA regulation and homeostasis.

static.klipy.com

ALT:

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

The most established players are AMPK, PGC-1a/PGC-1b, PPARg and components of the mtDNA replisome. Much elegant work on these, and while these factors can certainly influence the amount of mitos/mtDNAs in a cell, I don’t think many would claim these are the sole homeostatic regulators

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Huge amounts of work in the area over several decades, which I’m going to summarise uncharacteristically fast.

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

So the mechanism isn't about maintaining a certain number of circles, or about the number of origins of replication present, nor even the unit length, but is a cell autonomously-restrained parameter.... but set by what?

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

And from the annals of history - a lesson on how mtDNA copy number is controlled from Eric Schon and Mike King - mtDNA can be big (duplicated), small (partially deleted) or normal size - and this changes the number of copies of those mtDNAs per cell - however total mtDNA (pg/cell) stays the same.

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

We also found out a couple of years back through elegant work from @edreznik.bsky.social @mskcancercenter.bsky.social that mtDNA and nuclear DNA scale in polyploid states readily found in cancer www.nature.com/articles/s41...

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

How tissues maintain mtDNA copy numbers across this huge range is not clear. Even more curious, cells have ‘memory’ for how many mtDNAs they ‘should’ have – if you deplete mtDNA in cells with EtBr or ddC, once the drug is gone it recovers to the starting level..?!?

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

You wouldn’t want to use up all that delightful oxygen you just inhaled by having tonnes of mitochondria in alveoli or blood, would you?)

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

One of the many curious features of mtDNA is that it is multi-copy: 100’s – 1000’s of copies per cell, and in tissues the amount of mtDNA broadly scales with the amount of mitochondria needed to meet metabolic requirements. Heart is up the top end, blood and lung down the bottom.

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Let’s begin at the beginning. Human mitochondrial DNA – mtDNA - is a compact, circular genome contained in mitochondria that encodes a handful of proteins required for oxidative phosphorylation – the rest of the proteins that make up mitochondria are encoded in the nucleus.

110

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

Fair warning before you go any further – this is more of a tapestry than a thread, so go find somewhere comfy to sit down and strap in.

static.klipy.com

ALT:

100

Payam Gammage @mito-oncogene.bsky.social · 02/06/2026

For those who might think ‘mtDNA copy number *YAWN*’ – stick with me – mitochondria have been around since day dot, and it turns out mtDNA regulation has friends in high places: integrated stress response folks, mTOR enthusiasts, redox ppl, there’s something in here for everyone.

static.klipy.com

ALT:

110