Autophagy: When Does Cellular Self-Cleaning Really Begin?
The famous 16-hour number is too sweeping. Autophagy is a continuum that depends on your prior nutrition, your insulin status and your movement.
You have closed your eating window. Now the clock is running. Hour 12, hour 14, and somewhere in your head a thought is ticking: at 16 the switch flips, and cellular cleaning begins.
I understand the appeal of this number. It gives control. It makes something invisible feel plannable. It is only, unfortunately, too simple to be true.
It has become normal to treat autophagy like a timer. Cellular cleaning, however, is not a light switch that turns on at 16:00 sharp. It is a dial that runs constantly and gets louder or quieter depending on insulin, your last meal and movement. Anyone who watches only the number of hours misses the levers that really count.
The honest answer first
Many people are looking for the one number. From when? From 12? From 16? From 18?
The most honest answer science can give today: in humans there is no cleanly documented from-hour-X threshold. Not because no one has looked, but because autophagy does not work that way.
Autophagy comes from the Greek and means self-eating. Imagine a night shift in a huge factory. A cleaning crew collects broken machines, takes them apart and reuses the parts. This crew never fully clocks off. It works all the time, in every cell, around the clock. Fasting does not switch it on. Fasting turns it up.
That is the first important shift in perspective. We are not talking about a start, but about an increase. And an increase has no fixed start time, it has a pace.
Autophagy is a continuum, not a switch. The question is not off or on, but quiet or loud. And how loud it becomes is decided not by the clock on your wall, but by the chemistry in your blood. And now you know why the search for the one number leads nowhere.
Where the 16-hour number even comes from
So where does the number get its power? It sounds so precise that it has to be true. That is exactly the trap.
A large part of the hype can be traced back to animal studies. In a much-cited paper in mice, short fasting massively raised autophagy in nerve cells. That was an important discovery, because for a long time the brain was thought to be exempt. But it was a mouse. And the effect showed up after 24 to 48 hours, not after 16.
Alirezaei and colleagues showed in 2010 that short fasting strongly ramps up autophagy in nerve cells in mice, accompanied by falling mTOR activity. They observed this after 24 to 48 hours of food withdrawal. For you this means: this is mouse neurobiology, a pointer to a mechanism, but not a human 16-hour threshold.
Alirezaei M et al. Short-term fasting induces profound neuronal autophagy. Autophagy. 2010;6(6):702-710. DOI: 10.4161/auto.6.6.12376Worlds lie between mouse and human. A mouse has a much faster metabolism. It empties its stores in hours, where a human needs days. Transferring a number of hours from a rodent to a human is about as sensible as transferring the life expectancy of a mouse onto you.
There is no single, clean human study showing that autophagy in a healthy person switches on exactly at hour 16. If someone sells you this number as a fact, they are selling you a simplification. Fasting in this time frame can be sensible. The precision of the number, however, is not covered by the data.
The actual lever: insulin, mTOR and AMPK
If not the clock, then what? Here it is worth looking through the metabolic lens of Clinical Psychoneuroimmunology. Because your cells do not count hours. They measure energy and building materials.
Two sensors take the lead. One is called mTOR, you can think of it as the build switch. When food is plentiful, above all amino acids and anything that drives insulin, mTOR shifts toward growth and storage. In this state the cleaning crew powers down. Why tidy up when building material is being delivered right now?
The second sensor is called AMPK, the energy guardian. When energy runs low, insulin falls, the stores empty, then AMPK rises. And AMPK gives the autophagy start kinase called ULK1 the signal to turn the crew up louder. Simplified: mTOR brakes, AMPK hits the gas.
Kim and colleagues decoded in 2011, in Nature Cell Biology, how this switch works at the molecular level. They showed that during glucose shortage AMPK directly activates the kinase ULK1, while high mTOR activity blocks it. For you this means: it is not the time since eating that counts, but whether your body is currently signalling a nutrient surplus or an energy shortage.
Kim J, Kundu M, Viollet B, Guan KL. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nat Cell Biol. 2011;13(2):132-141. DOI: 10.1038/ncb2152This explains something practical that pure hour-counting cannot. Two people have both been fasting for 16 hours. One had a steak with plenty of protein beforehand, little movement, a lot of stress. The other trained until empty and had eaten early and light the day before. Their autophagy stands at completely different points, even though the clock says the same thing.
Your cells do not read a clock. They read your insulin level. A splash of milk in your coffee, a small protein bar, a spoon of honey can bring back the insulin stimulus and wake mTOR again. Not because you have eaten calories, but because you have signalled building material to the body. And now you know why the question about prior nutrition matters more than the one about the hour.
What has actually been measured in humans
Do you know the feeling when a claim is asserted everywhere but rarely backed up? With human autophagy that is exactly the case. The reason is uncomfortable: it can barely be seen directly in a living human.
Autophagy is highly dynamic. A marker such as LC3 can rise because more is being cleared, or because breakdown is stalling and rubbish is piling up. The same signal, two opposite meanings. That is why the measurement is so tricky.
Bonam and colleagues summarised in 2020, in the journal Cells, how difficult it is to measure autophagy via the marker LC3. They stress that the process is tissue dependent and dynamic and that single markers easily mislead. For you this means: a simple hour figure for humans cannot be seriously derived from this at all.
Bonam SR, Bayry J, Tschan MP, Muller S. Progress and Challenges in The Use of MAP1LC3 as a Legitimate Marker for Measuring Dynamic Autophagy In Vivo. Cells. 2020;9(5):1321. DOI: 10.3390/cells9051321Nevertheless there is first serious human data. And it paints a sober, intriguing picture.
Jamshed and colleagues had eleven people eat early in 2019, between 8am and 2pm. In the morning the gene expression of an autophagy gene was raised, together with raised ketones; in the evening, by contrast, the growth sensor mTOR rose. For you this means: eating window and time of day shift the signals noticeably, even though raised gene expression is not yet the same as more actual breakdown.
Jamshed H et al. Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans. Nutrients. 2019;11(6):1234. DOI: 10.3390/nu11061234Chaudhary and colleagues tested intermittent fasting in mice and humans in 2022. In mouse liver the autophagy markers rose, but in human muscle they did not; after 24 hours of fasting little more than a single marker changed in humans. For you this means: what happens in one organ or one species cannot be transferred one to one onto your muscle.
Chaudhary R et al. Intermittent fasting activates markers of autophagy in mouse liver, but not muscle from mouse or humans. Nutrition. 2022;101:111662. DOI: 10.1016/j.nut.2022.111662Bensalem and colleagues measured in 2025, in the Journal of Physiology, for the first time the real autophagic flux in blood cells of 121 people over six months. Interval-style eating showed a higher flux than the standard group after six months, but the difference was small and not significant within the fasting group itself. For you this means: fasting might raise autophagy, but measurably rather over months than from a certain hour.
Bensalem J et al. Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysis. J Physiol. 2025;603(10):3019-3032. DOI: 10.1113/JP287938Dang and colleagues showed in 2025, in the FASEB Journal, that autophagic flux varies strongly between different immune cells. Above all certain monocytes responded to amino acid withdrawal, and the flux differed by sex and age. For you this means: autophagy is cell and tissue dependent, a single value for the whole body is an illusion.
Dang LV et al. Cell-Type-Specific Autophagy in Human Leukocytes. FASEB J. 2025;39(12). DOI: 10.1096/fj.202402377RPut these puzzle pieces together and a clear pattern emerges. Not the tidy number of hours, but an honest landscape: the effects are real, but tissue dependent, individual and visible rather over time. Even current research admits this openly.
Masedunskas and colleagues described in 2024 the design of a study on three-day water fasting with and without exhausting exercise. Already in the introduction they name the striking lack of human studies on autophagic flux through fasting. For you this means: even the field openly says that we do not yet know the human answer well.
Masedunskas A et al. Investigating the Impact of Glycogen-Depleting Exercise Combined with Prolonged Fasting on Autophagy and Cellular Health in Humans. Nutrients. 2024;16(24):4297. DOI: 10.3390/nu16244297Movement: the underrated autophagy lever
Now comes the part that the clock fixation almost always overlooks. Autophagy does not hang only on the plate. It also hangs on your muscles.
Movement creates exactly the energy stress that your cells read as a signal. Muscle work empties glycogen, drives AMPK up and thereby pushes the same start buttons as fasting. Only often faster.
Moller and colleagues studied healthy people in 2015 after one hour of cycling, once during a 36-hour fast and once with glucose intake. The autophagy signals in muscle rose in both cases, independent of nutrient status. For you this means: exercise starts cellular cleaning in muscle, whether or not you have fasted.
Møller AB et al. Physical exercise increases autophagic signaling through ULK1 in human skeletal muscle. J Appl Physiol. 2015;118(8):971-979. DOI: 10.1152/japplphysiol.01116.2014Schwalm and colleagues observed in 2015 that autophagic flux in muscle rose above all after high-intensity training, measured by falling p62 and rising AMPK activity. Intensity was more decisive than the diet. For you this means: how demanding your movement is often counts more for autophagy than the question of whether you have fasted.
Schwalm C et al. Activation of autophagy in human skeletal muscle is dependent on exercise intensity and AMPK activation. FASEB J. 2015;29(8):3515-3526. DOI: 10.1096/fj.14-267187That is a liberating message. You are not at the mercy of the number of hours. A brisk walk, a training session, strength work, all of these are stimuli in their own right that do not wait for the fasting clock.
The individual factors instead of a clock
Let us sum up what shifts your personal autophagy curve. Not as a recipe with clock times, but as a map of the levers.
Prior nutrition
A protein and carbohydrate rich meal keeps insulin and amino acids up longer and brakes on the mTOR side for longer. Anyone who eats lighter and earlier reaches energy shortage sooner.
Insulin status
Insulin resistance and constant snacking keep insulin chronically high. Then the build switch mTOR stays on longer, even with the same fasting time.
Glycogen stores
Full stores from liver and muscle supply energy and dampen the shortage stimulus. Empty stores, for example after exercise, pull AMPK up faster.
Movement & training
Movement starts cellular cleaning independent of eating. Trained people also mobilise energy more efficiently, which shifts the curve.
Do you see how much lies under the surface of the number of hours? That is exactly why the question is framed wrongly. More useful than from when? is the question: how do I create the conditions under which autophagy can become louder?
- Calm insulin. Lengthen the pauses between snacks, structure the day with a stable eating rhythm and avoid sugary drinks in the fasting window. Black coffee, water and unsweetened tea barely disturb the stimulus.
- Place movement in the fasting window. A brisk walk or a short more intense session toward the end of the fasting time uses the energy stress as a stimulus in its own right, independent of the clock.
- Look at the whole picture, not at the minute. According to the human data, regularity over weeks and months matters more than hitting a magical hour exactly on a single day.
Autophagy is not a reward that is handed out on time. It is a state you can invite. Do not ask yourself whether you cracked the 16. Ask yourself whether your body experienced energy shortage and movement today. That is the language your cells really understand. And now you know why the clock never tells the whole story.
If you want not only to read but to look closely at your own metabolism, for example insulin status, ferritin or inflammation values, you will find below this article the option to book an appointment.
Frequently asked questions
From when does autophagy begin in humans?
There is no cleanly documented from-hour-X threshold. Autophagy runs constantly in every cell as a baseline process and is turned up during energy shortage. How quickly depends on your insulin status, your last meal and your movement, not on a fixed number of hours.
Is it true that autophagy starts after 16 hours of fasting?
This number is a simplification without a clean human basis. Part of the hype goes back to a mouse study in which 24 to 48 hours of fasting raised autophagy in nerve cells. That is mouse neurobiology, not a human 16-hour threshold.
What controls autophagy in the body?
Two nutrient sensors. mTOR is activated by insulin, glucose and amino acids and slows autophagy down. AMPK rises during energy shortage and starts it via the kinase ULK1. Nutrients are the switch, not the clock.
Do protein and coffee with milk slow autophagy?
Amino acids, especially leucine, and anything that drives insulin can activate mTOR and dampen autophagy. A splash of milk, a protein shake or a small meal shift the curve. Black coffee, water and tea without sugar barely do.
Can autophagy even be measured in humans?
Only with difficulty and indirectly. Markers such as LC3B-II or elaborate flux measurements in blood cells are used. These methods are young, and single markers are easy to misread. That is why any exact hour figure is not scientifically covered.
Do human studies show that fasting raises autophagy?
There are early signals. In a study with 121 people, autophagic flux after six months of interval-style eating was higher than in the control group. The effect was small and showed up over months, not from a certain hour.
Does movement start autophagy even without fasting?
Yes. In human muscle, autophagy signals rise after roughly 60 minutes of exercise, regardless of whether you have fasted. Intensity appears to matter more than the diet.
How long do I have to fast to encourage autophagy?
An honest answer is: there is no guaranteed number. Longer without an insulin stimulus, with empty glycogen stores and with movement pushes the curve upward. A 16 to 18 hour window is a sensible frame, but not a switch that clicks on time.
Is autophagy the same in every organ?
No. Autophagy is cell and tissue dependent. In one study, only certain immune cells responded to amino acid withdrawal. What happens in the liver does not automatically happen in muscle or in the brain.
Read on in the fasting and metabolism cluster
Intermittent fasting for women over 40
Why the female body responds differently to fasting windows.
Insulin resistance and weight loss
The insulin level that also co-decides your autophagy.
Avoiding blood sugar spikes
How meal order and choices dampen the insulin stimulus.
Understanding zone 2 training
Movement as a lever in its own right for energy and cell health.
Sources
- Kim J, Kundu M, Viollet B, Guan KL. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nat Cell Biol. 2011;13(2):132-141. DOI: 10.1038/ncb2152 [In vitro, human cells]
- Alirezaei M et al. Short-term fasting induces profound neuronal autophagy. Autophagy. 2010;6(6):702-710. DOI: 10.4161/auto.6.6.12376 [In vivo, mouse]
- Bagherniya M, Butler AE, Barreto GE, Sahebkar A. The effect of fasting or calorie restriction on autophagy induction: A review of the literature. Ageing Res Rev. 2018;47:183-197. DOI: 10.1016/j.arr.2018.08.004 [Review]
- Møller AB et al. Physical exercise increases autophagic signaling through ULK1 in human skeletal muscle. J Appl Physiol. 2015;118(8):971-979. DOI: 10.1152/japplphysiol.01116.2014 [RCT, crossover, human]
- Schwalm C et al. Activation of autophagy in human skeletal muscle is dependent on exercise intensity and AMPK activation. FASEB J. 2015;29(8):3515-3526. DOI: 10.1096/fj.14-267187 [RCT, human]
- Jamshed H et al. Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans. Nutrients. 2019;11(6):1234. DOI: 10.3390/nu11061234 [RCT, crossover, n=11]
- Chaudhary R et al. Intermittent fasting activates markers of autophagy in mouse liver, but not muscle from mouse or humans. Nutrition. 2022;101:111662. DOI: 10.1016/j.nut.2022.111662 [RCT, human n=50; in vivo, mouse]
- Bensalem J et al. Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysis. J Physiol. 2025;603(10):3019-3032. DOI: 10.1113/JP287938 [RCT, real-world, n=121]
- Dang LV et al. Cell-Type-Specific Autophagy in Human Leukocytes. FASEB J. 2025;39(12). DOI: 10.1096/fj.202402377R [Real-world, human, ex vivo]
- Bonam SR, Bayry J, Tschan MP, Muller S. Progress and Challenges in The Use of MAP1LC3 as a Legitimate Marker for Measuring Dynamic Autophagy In Vivo. Cells. 2020;9(5):1321. DOI: 10.3390/cells9051321 [Review, methods]
- Masedunskas A et al. Investigating the Impact of Glycogen-Depleting Exercise Combined with Prolonged Fasting on Autophagy and Cellular Health in Humans. Nutrients. 2024;16(24):4297. DOI: 10.3390/nu16244297 [RCT, study protocol]