The biology of fasting

Mitophagy

Also called: Mitochondrial autophagy, PINK1/Parkin pathway

Written by Aaron CuhaReviewed Sep 2026

Autophagy aimed at one target: mitochondria that have stopped working properly are tagged, wrapped and digested, so a cell keeps the ones that still make energy cleanly.

What it is

Mitochondria are where most of your ATP is made, and a damaged one is worse than useless. It leaks electrons, generates reactive oxygen species, and can trigger cell death. Mitophagy is the quality control that finds those and removes them selectively, leaving the healthy population behind.

This matters more with age. Mitochondrial function declines measurably in ageing muscle, and one of the leading explanations is that the removal system slows down faster than the damage does, so broken mitochondria accumulate.

The machinery

The main pathway is PINK1 and Parkin, and it is an elegant piece of engineering. PINK1 is a kinase constantly imported into healthy mitochondria and immediately chewed up once inside, so it never accumulates. When a mitochondrion loses its membrane potential, which is what damage looks like electrically, import fails. PINK1 gets stuck on the outer surface and builds up. That accumulation is the damage signal.

Stranded PINK1 then phosphorylates ubiquitin at serine 65 and recruits Parkin, an E3 ubiquitin ligase, which coats the mitochondrion in ubiquitin chains. Receptors read those chains, bind LC3, and the autophagosome forms around it.

There is a second, receptor-mediated route that skips ubiquitin entirely. BNIP3, NIX and FUNDC1 sit in the outer mitochondrial membrane and bind LC3 directly. This route matters in low oxygen and during normal developmental turnover.

Mitochondria also have to be physically separated from the network before they can be removed, which is why the fission protein DRP1 is part of the story.

What switches it on

Loss of mitochondrial membrane potential is the direct trigger. Upstream, the things that raise mitophagy are the things that stress energy supply: exercise, hypoxia, nutrient restriction, and AMPK activation.

On the timing
No human study establishes an hour at which fasting begins clearing mitochondria, and the studies that looked for a fasting effect in human muscle did not find one. What has been measured in people is that the PINK1 and Parkin pathway can be observed in muscle biopsies and differs by training status, and separately that a mitophagy-activating compound produces measurable changes in human muscle across randomised trials. Neither of those is a clock, and neither is evidence that a fast does the same thing.

What has been measured

In people

The PINK1 and Parkin pathway has been measured directly in human muscle biopsies, and the pathway has been pharmacologically engaged in three randomised controlled trials of urolithin A. Those trials are the positive evidence. The studies that went looking for mitophagy raised by fasting in human muscle did not find it, and one found the fasted state lower than the fed state. Both halves belong here.

In 10 endurance-trained male runners versus 9 sedentary controls, muscle biopsies in the fasted state showed significantly higher mitophagy and mitochondrial dynamics protein markers in the trained group. Phospho-PINK1 at threonine 257 correlated with phospho-Parkin at serine 65 (r = 0.64, p < 0.005), and phospho-Parkin S65 with phospho-DRP1 at serine 616 (r = 0.70, p < 0.05), in the trained group only. General autophagy markers did not differ between groups, so this was a mitochondria-specific difference (Molecular Metabolism, 2017).
First-in-human trial of urolithin A, a mitophagy activator, in healthy sedentary elderly adults. Four weeks at 500 mg and 1,000 mg modulated plasma acylcarnitines and skeletal muscle mitochondrial gene expression. The primary outcome was safety, which it met (Nature Metabolism, 2019).
Randomised placebo-controlled trial in middle-aged adults, two doses for four months (NCT03464500). Muscle strength improved about 12%, plasma acylcarnitines and C-reactive protein fell, and expression of proteins linked to mitophagy and mitochondrial metabolism rose significantly in skeletal muscle. The primary endpoint, peak power output, was not significantly improved (Cell Reports Medicine, 2022).
Randomised double-blind placebo-controlled trial in 66 adults aged 65 to 90, 1,000 mg daily for four months (NCT03283462). Muscle endurance improved significantly at two months in both muscles tested, and acylcarnitines, ceramides and C-reactive protein fell. The two primary endpoints, six-minute walk distance and maximal ATP production measured by magnetic resonance spectroscopy, were not significantly better than placebo (JAMA Network Open, 2022).
Mitophagy dysfunction has been characterised in the muscle of patients with mitochondrial muscle disease, showing the pathway can be assessed in human tissue in a clinical population (2022).
Looking for the opposite result: in 7 trained athletes cycling 2 hours at 70% of peak oxygen uptake, once fed and once fasted, with vastus lateralis biopsies before, immediately after and 1 hour after, the authors found no activation of mitophagy in either state. LC3bII and p62 in the mitochondrial fraction were unchanged and the LC3bII to LC3bI ratio fell in the fasted state (P = 0.019), while fission and the mitophagy genes Drp1, Bnip3 and Bnip3L rose only in the fed state (Medicine and Science in Sports and Exercise, 2017). Treat the falling ratio with care: work in human muscle since has shown that a falling LC3B-II can sit on top of rising flux, so this is best read as an absence of positive evidence rather than as proof of suppression.
In 10 healthy men fasted with and without added exercise, with biopsies at 4 and 8 hours, PGC-1alpha protein fell rather than rose. That paper states directly that the rapid activation of mitochondrial biogenic pathways seen in rodent muscle within 6 hours is absent in human muscle across 10 to 72 hours of fasting (Pflugers Archiv, 2021).

In other species and in cell culture

The pathway's mechanics, and most of what is known about how fasting specifically affects it, come from mice and from cultured cells, including the fluorescent reporter systems that allow mitophagy to be watched in a living animal.

Reporter mice carrying mt-Keima or mito-QC allow mitophagy to be visualised and quantified in intact tissue over time. No equivalent exists for a living human, which is the main reason human mitophagy work relies on protein markers in biopsies.
Urolithin A was shown to stimulate mitophagy and improve muscle function in old animals before the human trials were run, and in preclinical models of ageing and of muscular dystrophy.

Why it matters

Of everything fasting is claimed to do at the cellular level, mitophagy is the one where a specific intervention has been taken all the way into randomised human trials with muscle biopsies. That makes it the best-evidenced part of this whole area, and it is worth being precise about what those trials showed.

They showed pathway engagement: mitophagy-linked protein expression in muscle went up, and the plasma metabolite signature moved in the direction of better mitochondrial efficiency. On function, results were mixed. Muscle strength and endurance improved in two trials. Both trials missed their primary endpoint. That combination is worth more than a marketing summary of either one.

A claim you will see repeated

Two limits belong with the urolithin A evidence. No urolithin A trial has measured mitophagic FLUX; they measure pathway protein and gene expression, plasma metabolites and function. And an independent trial not run by the manufacturer, in 10 people with heart failure with reduced ejection fraction, was null. Separately, these trials are frequently cited as though they prove fasting produces the same effect. They do not. Those are trials of an orally dosed compound derived from ellagitannins by gut bacteria, and the research programme is run by Amazentis, which sells it. What the trials establish is that the mitophagy pathway can be engaged and measured in living people. They say nothing about what a 24 or 48 hour fast does to the same pathway, which has not been tested the same way.

Citations

Human2017
Skeletal muscle autophagy and mitophagy in endurance-trained runners before and after a high-fat meal
Molecular Metabolism
Muscle biopsies from 10 endurance-trained male runners and 9 sedentary non-obese controls. Mitophagy and mitochondrial dynamics protein markers were significantly higher in trained muscle in the fasted state, while general autophagy markers were similar. Phospho-PINK1 Thr257 correlated with phospho-Parkin Ser65 (r = 0.64, p < 0.005) in the trained group only. A high-fat meal did not change autophagy or mitophagy in either group.
Human2019
The mitophagy activator urolithin A is safe and induces a molecular signature of improved mitochondrial and cellular health in humans
Nature Metabolism
First-in-human trial in healthy sedentary elderly adults, single dose and 4 weeks of repeated dosing. Favourable safety profile as the primary outcome. At 500 mg and 1,000 mg, four weeks modulated plasma acylcarnitines and skeletal muscle mitochondrial gene expression.
Human2022
Urolithin A improves muscle strength, exercise performance, and biomarkers of mitochondrial health in a randomized trial in middle-aged adults
Cell Reports Medicine
Randomised placebo-controlled trial, two doses over 4 months (NCT03464500). Muscle strength improved about 12%. Plasma acylcarnitines and C-reactive protein fell significantly. Expression of mitophagy and mitochondrial metabolism proteins in skeletal muscle rose significantly. The primary endpoint, peak power output, was not significantly improved.
Human2022
Effect of Urolithin A Supplementation on Muscle Endurance and Mitochondrial Health in Older Adults: A Randomized Clinical Trial
JAMA Network Open
66 adults aged 65 to 90 randomised to 1,000 mg urolithin A or placebo for 4 months (NCT03283462). Muscle endurance improved significantly at 2 months in both the first dorsal interosseus and tibialis anterior. Acylcarnitines, ceramides and C-reactive protein fell. The primary endpoints, six-minute walk distance and maximal ATP production by magnetic resonance spectroscopy, were not significantly different from placebo. No difference in adverse events.
Human2022
Mosaic dysfunction of mitophagy in mitochondrial muscle disease
Cell Metabolism
Mitophagy assessed in the skeletal muscle of patients with mitochondrial muscle disease, demonstrating that the pathway can be characterised in human tissue in a clinical population.
Human2017
Lack of Activation of Mitophagy during Endurance Exercise in Human
Medicine and Science in Sports and Exercise
7 trained athletes cycled 2 hours at 70% of peak oxygen uptake in a fed and a fasted state, with vastus lateralis biopsies at baseline, before, immediately after and 1 hour after. Mitophagy was not activated in either state. LC3bII and p62 in the mitochondrial fraction were unchanged, and the LC3bII to LC3bI ratio fell after exercise in the fasted state only (P = 0.019), indicating reduced mitophagy. Fission by phospho-DRP1, and the genes Drp1, Bnip3 and Bnip3L, rose only in the fed state.
Human2021
Increasing whole-body energetic stress does not augment fasting-induced changes in human skeletal muscle
Pflugers Archiv, European Journal of Physiology
10 healthy males, two supervised fasts with and without 2 hours of arm ergometer exercise, vastus lateralis biopsies before, at 4 hours and at 8 hours. PGC-1alpha protein decreased over both fasts. The paper states that the rapid activation of mitochondrial biogenic pathways seen in rodent muscle within 6 hours of fasting is absent in human muscle across 10 to 72 hours.