Children with a rare genetic disease lose the cellular structures their brains, muscles and other organs need to produce energy, and many die within their first years of life. Experiments in patients’ cells and mice suggest that activating the energy sensor AMP-activated protein kinase (AMPK) can protect mitochondria that still function and restore energy production – pointing towards the first targeted treatment while overturning the idea that mitochondrial recycling is always beneficial.
Cells routinely break down old or damaged mitochondria, which convert nutrients into ATP – the energy cells use to function. This recycling process, termed mitophagy, usually protects cells. But in FBXL4-related mitochondrial DNA depletion syndrome, it becomes destructive: cells remove mitochondria that can still produce energy, leaving the brain, muscles, liver and other energy-demanding organs with too little power.
More than 100 people with disease-causing FBXL4 variants have been reported worldwide, and previous clinical research has estimated the median age of death at about 3 years. Symptoms often begin in infancy and may include severely delayed development, low muscle tone, feeding difficulties, brain and liver damage and metabolic acidosis, in which acid accumulates in the body. No treatment currently alters the course of the disease.
“We found a fundamental difference between the mitophagy pathways. One pathway targets damaged mitochondria, but in the NIX–BNIP3 pathway, the mitochondria still retained functional capacity – they could consume oxygen and potentially make ATP,” says Ian Ganley, Professor of Cellular Homeostasis at the University of Dundee in Scotland and Principal Investigator at the MRC Protein Phosphorylation and Ubiquitylation Unit.
Ganley and colleagues had also discovered that AMPK acts as a cellular fuel gauge, regulating the two pathways in opposite directions.
“When energy levels are low, you probably want to enhance the turnover of dysfunctional mitochondria because they are not working. But you probably want to block the turnover of functional mitochondria because they can still make ATP. What we found is that AMPK does exactly that,” says Ian Ganley.
The experiments published in EMBO Molecular Medicine remain preclinical and do not show that AMPK activators can help children with the disease. But they identify a specific fault that drug-like compounds may be able to correct – and therefore a possible path towards the first targeted treatment.
When mitochondrial recycling becomes excessive
Ian Ganley’s laboratory studies mitophagy, the specialised form of cellular recycling through which mitochondria are delivered to lysosomes and broken down. Much of the field has focused on the PINK1-Parkin pathway, which marks damaged mitochondria for removal and is disrupted in some inherited forms of early-onset Parkinson’s disease.
“Mitophagy has been linked to Parkinson’s disease because it makes sense that if damaged mitochondria cannot be removed, they can accumulate and cause problems. ATP production goes down, reactive oxygen species can go up, and that can contribute to disease,” says Ian Ganley.
But mitophagy is not a single pathway with a single purpose. Ganley’s group became interested in another route driven by the mitochondrial proteins NIX and BNIP3.
“That was a fundamental difference. In the PINK1-Parkin pathway, the mitochondria were damaged and dysfunctional. In the NIX-BNIP3 pathway, they still retained functional capacity. Whether they were completely normal we did not know, but they were capable of consuming oxygen and potentially making ATP,” says Ian Ganley.
In addition, other studies were revealing how variants in FBXL4 cause mitochondrial DNA depletion syndrome 13, also known as MTDPS13.
“What FBXL4 normally does is keep that mitophagy pathway very low. But among patients with mutations in this gene, NIX and BNIP3 slowly accumulate on mitochondria. At some point, they trigger mitophagy and you get turnover of the mitochondria,” says Ian Ganley.
In healthy cells, FBXL4 acts as a brake by marking NIX and BNIP3 for destruction before they accumulate. Without functional FBXL4, the proteins build up on mitochondria until they signal that the structures should be dismantled.
A chance discovery reveals a possible way to intervene
The result is severe mitochondrial depletion. Ganley’s group, however, had recently found that AMPK regulates the two pathways in opposite directions: under energy stress, it enhances PINK1-Parkin while inhibiting NIX-BNIP3.
“We found that AMPK turns on the mitophagy of dysfunctional mitochondria yet inhibits the mitophagy of functional mitochondria. So the PINK1-Parkin pathway is enhanced, but the NIX-BNIP3 pathway is blocked. It was a way the cell balances these two pathways,” says Ian Ganley.
The two findings fitted together. The pathway stripping mitochondria from cells in MTDPS13 was also the pathway that AMPK could suppress.
“We had always been thinking about how to enhance mitophagy to treat disease. Then we found a way to inhibit a very specific mitophagy pathway, and at the same time it turned out that this pathway is overactivated in this rare disease. It was a bit of serendipity: we were working on this at the same time these other discoveries were being made, and then it linked in with what we were doing,” says Ian Ganley.
Putting AMPK to the test
The researchers tested AMPK activation in engineered human cells, skin-derived fibroblasts from people with MTDPS13 and a mouse model that reproduces parts of the cellular damage caused by loss of FBXL4.
“Our first strategy was to take patient cells, where we know this mitophagy pathway is excessive, and treat them with AMPK activators to see whether we could bring mitophagy back down to normal levels,” says Ian Ganley.
They began by disrupting FBXL4 in human cells and treating them with the direct AMPK activator MK-8722. They repeated the experiments without active AMPK to determine whether any rescue depended on the enzyme itself.
The team then used a fluorescent marker called mito-QC to follow mitophagy directly. It labels mitochondria with two colours, but one signal disappears in the acidic environment of a lysosome. The remaining colour therefore reveals which mitochondria have entered the cell’s degradation system.
“We did not want to rely on one compound or one readout. We used patients’ cells, engineered cells, different AMPK activators and direct measurements of mitophagy to ask whether the pathway was really being controlled,” says Ian Ganley.
From engineered cells to patients’ cells and mice
The researchers repeated the experiments with the direct activators PXL-770 and BI-9774 and measured mitochondrial proteins, oxygen consumption and ATP production. They then tested fibroblasts from people with MTDPS13 to determine whether the mechanism also appeared in cells carrying disease-causing variants.
Finally, they used a chemically induced mouse model because mice completely lacking FBXL4 die shortly after birth. The team examined mitophagy and mitochondrial proteins in the liver and brain after 5 days of treatment and also tested metformin, an indirect and more tissue-specific AMPK activator.
Disrupting FBXL4 caused NIX and BNIP3 to accumulate, increased mitophagy and reduced mitochondrial proteins. MK-8722 reversed much of this loss, but only when AMPK was active. Mito-QC showed that fewer mitochondria entered lysosomes, while increased oxygen consumption confirmed that the preserved ones could still produce energy.
“That comes back to our original idea that this pathway is turning over functional mitochondria. When we block it, we restore mitochondrial function,” says Ian Ganley.
Patients’ cells regain part of their energy supply
Fibroblasts from people with MTDPS13 brought the experiments closest to the children affected. Compared with control cells, they showed more mitophagy and contained fewer mitochondria.
“Importantly, it restored mitochondrial activity and ATP production. In the patients’ cells, where there is more mitophagy, there are fewer mitochondria and mitochondrial activity is very low. Activating AMPK blocked this pathway and restored that level of function,” says Ian Ganley.
In the chemically induced mouse model, AMPK activation reduced excessive mitophagy in the liver and in some hippocampal neurons. Metformin produced a similar but weaker liver effect, supporting the mechanism rather than pointing to metformin as an obvious treatment.
“Metformin gave a similar effect in the animal model, but it is much less potent and very tissue-specific. The other AMPK activators tend to be less tissue-specific,” says Ian Ganley.
The mice did not carry disease-causing FBXL4 variants, and the experiments were short-term. Even so, engineered cells, patients’ cells and living tissue pointed in the same direction: suppressing NIX-BNIP3 mitophagy preserved mitochondria that could still produce energy.
“The animal model is not the same as a child with FBXL4 mutations, but it allowed us to ask whether AMPK activation could suppress excessive mitophagy in organs that matter for the disease,” says Ian Ganley.
From rescued cells towards a possible treatment
The study does not show that AMPK activators improve symptoms or prolong children’s lives. But drug-like compounds corrected part of the underlying defect in patients’ cells and affected the same pathway in mouse liver and brain. PXL-770 has already entered clinical testing for another disease, although it has not been approved.
“These AMPK activators have not been clinically approved yet, but some are in clinical trials. One possibility could be small studies, maybe even on compassionate grounds, because these children normally die. We do not know yet, but this could be a therapeutic approach for a disease that is currently incurable,” says Ian Ganley.
Researchers must still determine whether AMPK activation improves symptoms, organ function or survival and whether suitable compounds can reach the brain and other affected tissues. A genetic animal model would also provide stronger evidence.
The principle may extend beyond this rare disease. Parkinson’s research has largely focused on what happens when cells fail to remove defective mitochondria.
“AMPK activation stops the turnover of good mitochondria and enhances the turnover of bad mitochondria. This is probably also relevant for Parkinson’s disease, in which we think bad mitochondria accumulate, and it opens up new points of regulation that we may be able to use to swing the mitophagy balance between dysfunctional and functional mitochondria,” says Ian Ganley.
Simply increasing or decreasing mitophagy may therefore be the wrong goal. The key may be to block the pathway that removes working mitochondria without disrupting the pathway that clears damaged ones.
Ganley began by asking how cells decide which mitochondria to remove. That basic question may now point towards a treatment for children who currently have none.
“It really hits home how the work you do in the lab could actually have translational impact. A lot of the time, we are not thinking about that – we are thinking about mechanisms and cells in a dish. But if this could be used therapeutically, that would be absolutely amazing,” says Ian Ganley.
