Mitochondria, Part 2: What Damages Mitochondria, and What Actually Helps
Written By Kevin Wallace.
Human studies, graded. Not the internet list.
1. The short answer first
Part 1 met the mitochondrion: a former bacterium that still carries a scrap of its own DNA, makes ATP, and does several other jobs. This article asks the practical question. What damages mitochondria, and what actually helps?
From human studies, not test tubes or mice, the answer is less exotic than the internet suggests. Insulin resistance has the strongest everyday link. Physical inactivity and short sleep are real, on a longer clock than a bad weekend. The best-proven repair is not a pill. It is regular exercise, which builds mitochondrial respiratory capacity at any age.
Toxins and a few medications matter too. The evidence there is narrower, and I will say where it is solid. One rule runs through the medication section. Do not stop a prescribed drug because of an article. Take the question to your physician.
2. Can you tell if your mitochondria are struggling?
Not directly. Anyone selling a simple home test deserves skepticism. In the studies below, researchers took small muscle biopsies and measured them in the lab. That is not an order your doctor writes for ordinary fatigue.
Waking unrefreshed, an afternoon crash, workouts that feel harder, slow recovery, a falling heart-rate variability or a rising resting heart rate are worth noticing. None of them measures mitochondria. Fatigue has many causes. Persistent symptoms deserve a workup, not a self-diagnosis.
The best everyday proxy is aerobic fitness. In 146 adults aged 18 to 89, lower mitochondrial ATP production in muscle tracked with lower fitness and poorer glucose handling. (1) Training that raises fitness has repeatedly improved muscle mitochondria, as the studies below show.
3. How damage actually happens
Most of the harm that has been measured runs through three routes. They are numbered because they are three different mechanisms, not a list of tips.
Oxidants. As electrons move down the energy chain, a small fraction leak and form reactive oxygen species. Small, controlled amounts, hydrogen peroxide in particular, are signals. They help drive the response to exercise, immunity, and low oxygen. Trouble starts when production outruns the cell’s defenses and damages proteins, membranes, and mitochondrial DNA. The dose is the difference. (2)
Fuel the machinery cannot clear. Mitochondria can process fat and sugar only so fast. When more arrives than they can burn, partly processed fuel backs up. In muscle, that backlog travels with insulin resistance.
Failed cleanup. Healthy cells remove damaged mitochondria and build new ones. When cleanup slows, as it tends to with age and inactivity, damaged units linger and the network works less well.
4. Insulin resistance and nutrient overload
This is the common threat, and the wording has to be careful. A 2022 analysis of 55,081 U.S. adults found that in 2017 to 2018 only 6.8 percent met a five-part definition of optimal cardiometabolic health: adiposity, blood glucose, blood lipids, blood pressure, and no clinical heart disease. (3) That is not the same thing as insulin resistance. It is a stricter score, and most adults miss it. Separately, CDC estimates put prediabetes at about 38 percent of U.S. adults, most of them undiagnosed. (4) Both numbers say the same practical thing. Fuel handling is off in a large share of people.
The mitochondrial link is documented, in a selected group. In a 2004 study, lean young adults who were insulin resistant, and whose parents had type 2 diabetes, had about 30 percent lower mitochondrial phosphorylation in muscle than matched controls. They also had about 80 percent more fat stored inside muscle cells, and about 60 percent lower insulin-driven glucose uptake. The phosphorylation comparison was 13 people against 10. (5) The authors read this as a possible inherited mitochondrial weakness leading to insulin resistance. That is a hypothesis the study could not prove.
Later experimental work cut the other way, and it was not in people. At Washington University, rats with an iron-deficiency cut in the muscle energy chain did not become insulin resistant. On a high-fat diet, the iron-deficient animals were protected. (6) The honest picture is a two-way street, not a settled arrow. Overload strains mitochondria. Strained mitochondria handle fuel worse. Inherited weakness may matter in some families. For practical purposes the lever is the same either way. The changes that improve insulin resistance are the changes that support muscle mitochondria.
High blood sugar has a second proposed route. In a 2001 Nature paper, Michael Brownlee argued that excess glucose pushes the mitochondrial chain to overproduce superoxide, and that this one oxidant switches on several pathways behind diabetic damage in eyes, kidneys, nerves, and arteries. (7) He offered it as a framework for research. It has been influential. It is not a settled fact, and it should not be quoted as one.
5. Inactivity: use it or lose it, but not overnight
A few days of rest do not wreck muscle mitochondria. Bed-rest studies, where healthy volunteers lie down around the clock, are the cleanest test, and they are still small.
After 10 days of bed rest, 10 recreationally active men lost peak aerobic capacity, and the early bottleneck was upstream of the mitochondria. Microvascular blood flow was impaired. Mitochondrial mass and maximal respiration were unchanged, and sensitivity to ADP was slightly better, not worse. (8) I am not hanging a percent on the fitness drop. The paper says peak oxygen uptake fell. It does not need a round number to make the point.
Three weeks reaches the mitochondria. In a crossover study of 9 young men, 21 days of bed rest in energy balance lowered insulin sensitivity and lowered mitochondrial respiration per gram of muscle. Resistive vibration exercise in bed prevented that respiration loss and blunted the insulin-sensitivity drop. Once respiration was scaled to a marker of mitochondrial content, most of the difference disappeared. Content fell. The remaining mitochondria were not uniformly broken. (9) This is one small study. It is a full paper, not a conference abstract, and it is still a small paper.
Short breaks do not undo the network. Weeks of inactivity do. Modest muscle work during illness or recovery is worth keeping if a physician says it is safe.
6. Sleep: five short nights are enough to show
The clearest human test is a 2021 controlled study of 24 healthy young men, eight in each group. One group had 8 hours in bed for five nights. One had 4 hours. One had 4 hours and three sessions of high-intensity intervals.
In the short-sleep group, mitochondrial respiratory function fell by 15.9 pmol of oxygen per second per milligram of muscle, and the glucose area under the curve rose by 149 units. In the group that exercised while sleep-restricted, neither change was significant. (10)
Two cautions. The study was small, short, and limited to young men. It does not tell us about women, older adults, or years of poor sleep. And exercise protected in that controlled setting. It is not a license to skimp on sleep. The effects of losing it showed up within days.
7. After 65: a slope, not a cliff
This belongs here, not in the supplement article. The question after retirement is the same question as the rest of this piece: what still damages the network, and what still repairs it. The pill trials do not have a better senior record than the training trials do.
The decline is real, and it is gradual. In 146 healthy adults aged 18 to 89, muscle mitochondrial DNA, mitochondrial transcripts, and ATP production all fell with age. ATP production fell on the order of 8 percent per decade. Oxidative damage to DNA was higher in the older muscle. The older group in that study ran into the late eighties, 34 people in the oldest band. (1) There is no separate break in the data at 65, 70, or 80. The same slope is still operating.
Age and inactivity are tangled in that survey. Most people move less as they get older, and the study could not fully pull those apart. A 2024 review of the lifespan data makes the same point from the other side: the mitochondrial gap is smaller in older adults who stay active, and in some comparisons it shrinks further when younger and older groups are matched for how much they move. (23)
The repair still works in the years people actually retire into. In the Mayo trial, adults 65 to 80 gained more from 12 weeks of interval cycling than adults 18 to 30: maximal mitochondrial respiration rose 69 percent, against 49 percent in the younger group. (18) The groups were about a dozen people. It was 12 weeks, not a lifetime. It did not erase deleted mitochondrial DNA. It did show that the machinery still answers the signal in the seventh and eighth decade.
What I will not do is invent a statistic for “the average 75-year-old mitochondrion.” The human samples at 80 and beyond are small. The honest senior numbers are the per-decade slope through age 89, and the training response at 65 to 80.
8. Membranes, cardiolipin, and the lists that outrun the data
A mitochondrion has two membranes. The inner one is folded into cristae, holds the energy chain, and is where electrons leak. A phospholipid called cardiolipin sits almost only in that inner membrane. It helps hold the chain complexes together as supercomplexes. When cardiolipin content, fatty-acid makeup, or oxidation goes wrong, function suffers. That makes it a plausible weak point. (11) A peptide called elamipretide, or SS-31, was designed to bind cardiolipin. It is an investigational drug, not a supplement. The trial record belongs in the supplement article, not here.
Popular lists then add processed fats, pollution, heavy metals, infections, antibiotics, and chronic stress. I am not carrying those as facts. The antibiotic claim has a real paper behind it, and a real limit. Clinical-range doses of three bactericidal antibiotics, ciprofloxacin, ampicillin, and kanamycin, raised mitochondrial oxidant stress in human cells in a dish and in mice. Bacteriostatic antibiotics did not do the same in that study. (12) It was not shown in patients. It is not a reason to refuse an antibiotic you need.
The other items on those lists are open questions until someone measures them in people. Flagging them is not the same as establishing them.
9. Inflammation and infection
When the immune system goes to work, its cells change how they make energy. A resting T cell runs mainly on mitochondrial respiration. Once activated, it leans on glycolysis, outside the mitochondria, and turns glucose into lactate quickly. Glycolysis nets 2 ATP per glucose. Full oxidation through mitochondria nets on the order of 30, depending on the shuttle. The activated cell wants speed and building blocks more than efficiency. Mitochondria also signal during infection, including the alarm when their own DNA escapes into the cell, as Part 1 described.
That switch is healthy when it turns off after the threat passes. The idea that chronic infection, chronic stress, or poor recovery leaves cells stuck in defense mode is a reasonable reading of the biology. I have not found human studies that measure it directly. Useful idea. Not an established fact.
10. Toxins: where the evidence is real
Pesticides. Some attack mitochondria directly. Rotenone blocks complex I. Paraquat generates oxidants. In a case-control study nested in the Agricultural Health Study, 110 people with Parkinson disease and 358 controls, use of rotenone was associated with 2.5 times the odds of Parkinson disease (95 percent confidence interval 1.3 to 4.7). Paraquat use had the same point estimate (1.4 to 4.7). (13) That is an occupational association, not proof that a single exposure causes the disease, and it is not a finding about every pesticide. For anyone who handles these chemicals, it is one of the better arguments for not using them, and for real protection if you must.
Alcohol. Chronic heavy drinking damages liver mitochondria. It increases oxidant production, harms mitochondrial DNA, and reduces respiratory capacity. Processing alcohol also pushes the liver toward storing fat. Structural and functional changes have been seen in patients and in animal models. Most of the step-by-step mechanism is from the animal work. (14)
11. Medications: proven, possible, and overstated
Proven, and mostly historical: certain older HIV drugs. Some nucleoside reverse transcriptase inhibitors, notably stavudine and zidovudine, can inhibit DNA polymerase gamma, the enzyme that copies mitochondrial DNA. The result can be a falling mitochondrial DNA copy number, rising lactate, and, at worst, lactic acidosis. (15) These drugs are largely retired from current regimens. The lesson remains the textbook one. A drug can be a mitochondrial toxin. That does not make every drug one.
Possible: metformin during training. In a randomized trial, 53 older adults at risk of diabetes did 12 weeks of supervised aerobic training on metformin, up to 2,000 mg a day, or placebo. In the placebo group, a key measure of mitochondrial respiration rose about 25 percent. Metformin erased that gain, and it blunted the improvement in insulin sensitivity. The blunting of the fitness gain fell just short of statistical significance. (16) One trial, in people without diabetes. A conversation with your doctor. Not a reason to stop.
Overstated: coenzyme Q10 for statin muscle aches. Statins can lower circulating CoQ10, because the pathway that makes cholesterol also makes the side chain of CoQ10. Whether that causes the aches is unproven, and muscle CoQ10 does not always follow the blood level. Meta-analyses of supplementation disagree. A 2021 pooled analysis of randomized trials found no significant improvement in muscle pain or creatine kinase. (17) Routine CoQ10 with a statin is not established. Do not stop a statin over it.
12. What actually helps
Exercise comes first, by a wide margin. It is the intervention with the most consistent human evidence that it builds mitochondrial respiratory capacity, at any age.
Twelve weeks of high-intensity interval cycling raised maximal mitochondrial respiration 49 percent in adults 18 to 30, and 69 percent in adults 65 to 80. Groups were small. Resistance training in that study did not raise respiration. (18)
Steady aerobic training works too. In the metformin trial above, 12 weeks of supervised aerobic exercise raised mitochondrial respiration about 25 percent in the placebo group. (16)
Exercise protects, in the settings tested. Interval sessions prevented the mitochondrial and glucose changes of five short nights. Resistive exercise in bed preserved respiration across 21 days of bed rest. (9, 10)
Volume and intensity do different jobs. A 2018 review of training studies found that total work best predicted gains in mitochondrial content, while hard efforts near maximal power best predicted gains in respiratory function. The two often move independently. (19) In practice that argues for both: regular steady work, such as brisk walking or cycling, plus some short hard intervals if your health allows. If you have heart disease or another condition, ask your physician before vigorous work.
Every workout sends a signal. Repetition builds the result. In seven men, a three-hour bout of knee-extensor exercise raised PGC-1 alpha transcription 10-fold to more than 40-fold within two hours, and its messenger RNA 7-fold to 10-fold. It was back toward baseline by 24 hours. (20) That bout was far longer than a normal workout, and a spike in a signal is not new mitochondria. The lasting gains come from repeating the stimulus over weeks.
Strength training counts, with a narrower record. Eleven young men did 12 weeks of resistance training. Coupled respiration supported by complex I doubled. Respiration supported by complex I plus II rose 1.4-fold. Lean mass rose 4 percent and quadriceps strength 15 percent. Transcripts usually tied to building new mitochondria did not change. Complex I protein did. (21) Read that as better function, and some added machinery, not as a second copy of the aerobic result. Pairing strength work with aerobic work is the sensible way to get both.
Food: work on insulin resistance, and call the inference what it is. I have not found strong human trials that measure muscle mitochondria after diet alone. The link runs through the markers that do move: waist, fasting glucose, triglycerides, HDL, and blood pressure. Whole, minimally processed food is the foundation for those markers. Track them with your physician. Do not expect a biopsy study to exist for every plate.
Sleep: protect it. Five nights of four hours in bed measurably reduced mitochondrial respiration in young men. Regular hours and enough of them are among the cheapest protections you have.
Skip megadose antioxidants around training. In a 2009 study, vitamin C at 1,000 mg a day and vitamin E at 400 IU a day during four weeks of hard training blocked the improvement in insulin sensitivity seen in the unsupplemented men. (22) The protocol was heavy for untrained people, about 85 minutes a day, five days a week, and the finding is debated. It fits the oxidant paradox: some oxidant signal is part of how exercise works. Antioxidants from food were not the thing tested.
Limit the exposures with a mechanism and a human association. Keep alcohol moderate. Handle rotenone and paraquat as the occupational data say to handle them, or do not handle them.
13. The evidence at a glance
| Factor | What the human data show | Grade |
|---|---|---|
| Interval and steady aerobic exercise | Builds respiration and, with volume, content. Seen in young and older adults. | Strong |
| Insulin resistance | About 30 percent lower muscle phosphorylation in a small study of lean offspring. Cause unsettled. | Strong link, causality open |
| Older NRTIs (stavudine, zidovudine) | Inhibit polymerase gamma. Can deplete mitochondrial DNA and raise lactate. | Strong, mostly historical |
| Bed rest | Little mitochondrial change at 10 days. Respiration per gram down at 21 days in 9 men. | Moderate, small studies |
| Four hours of sleep, five nights | Respiration down 15.9 pmol O2/s/mg. Young men only. Exercise prevented it. | Moderate, one small trial |
| Rotenone and paraquat | Odds of Parkinson disease about 2.5 in occupational users. Association. | Moderate |
| Chronic heavy alcohol | Liver mitochondrial damage and fat storage. Mechanism mostly animal. | Moderate |
| Metformin during training | Erased an about 25 percent respiration gain. One trial, no diabetes. | Moderate |
| Resistance training | Respiration up 1.4- to 2-fold in 11 young men. | Moderate, one small study |
| Megadose vitamins C and E | Blocked a training gain in insulin sensitivity. Debated. | Moderate, debated |
| CoQ10 for statin aches | Meta-analyses conflict. A 2021 pool found no significant benefit. | Weak |
| Brownlee superoxide hypothesis | Framework for diabetic complications. Not a settled measurement. | Influential hypothesis |
| Bactericidal antibiotics | Oxidant stress in human cells and mice. Not shown in patients. | Weak |
| Cells stuck in defense mode | Fits immune biology. Not measured in people. | Unverified |
14. Where we land
The internet has a long list. The human record has a short one. Insulin resistance is the common link, and the arrow of cause is not settled. Ten days in bed do not erase muscle mitochondria. Three weeks start to. Five bad nights were enough to measure, in young men. Rotenone, paraquat, heavy alcohol, and two older HIV drugs have a real mitochondrial story. Most of the rest of the popular list does not, yet.
What helps is plainer. Repeat the training stimulus. Protect sleep. Do not megadose antioxidants around hard workouts. Do not stop a medicine because a mechanism sounds neat.
In zone 5 the evenings are already closing in. That is a useful month for the two levers with the least argument: a walk most days, a short hard effort if your physician has cleared it, and a night that is actually a night. The supplement trials are the next article. Nothing in a bottle gets ahead of this one.
15. Sources
Short et al., PNAS, 2005. https://www.pnas.org/doi/10.1073/pnas.0501559102
Holmström and Finkel, Cellular mechanisms and physiological consequences of redox-dependent signalling, Nature Reviews Molecular Cell Biology, 2014. https://pubmed.ncbi.nlm.nih.gov/24854789/
O’Hearn et al., JACC, 2022. https://pubmed.ncbi.nlm.nih.gov/35798448/
CDC, National Diabetes Statistics Report. Prediabetes, about 38 percent of U.S. adults. https://www.cdc.gov/diabetes/php/data-research/index.html
Petersen, Dufour, Befroy, Garcia, and Shulman, NEJM, 2004. https://www.nejm.org/doi/full/10.1056/NEJMoa031314
Han et al., PLoS ONE, 2011. Rat iron-deficiency model. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0019739
Brownlee, Nature, 2001. https://www.nature.com/articles/414813a
Zuccarelli et al., Journal of Physiology, 2021. https://pubmed.ncbi.nlm.nih.gov/34505290/
Kenny et al., Diabetologia, 2017. https://pubmed.ncbi.nlm.nih.gov/28500394/
Saner et al., Molecular Metabolism, 2021. https://pubmed.ncbi.nlm.nih.gov/33137489/
Paradies et al., Biochimica et Biophysica Acta, 2014. https://pubmed.ncbi.nlm.nih.gov/24183601/
Kalghatgi et al., Science Translational Medicine, 2013. https://pubmed.ncbi.nlm.nih.gov/23825301/
Tanner et al., Environmental Health Perspectives, 2011. https://ehp.niehs.nih.gov/doi/10.1289/ehp.1002839
Nassir and Ibdah, World Journal of Gastroenterology, 2014. https://www.wjgnet.com/1007-9327/full/v20/i9/2136.htm
Gardner, Côté, and colleagues, reviews of NRTI polymerase-gamma toxicity. Lewis and Dalakas, Nature Medicine, 1995, is the classic clinical description. https://pubmed.ncbi.nlm.nih.gov/7585189/
Konopka et al., Aging Cell, 2019. https://pmc.ncbi.nlm.nih.gov/articles/PMC6351883/
Wei et al., Irish Journal of Medical Science, 2022. https://pubmed.ncbi.nlm.nih.gov/33993403/
Robinson et al., Cell Metabolism, 2017. https://www.cell.com/cell-metabolism/fulltext/S1550-4131(17)30099-2
Granata, Jamnick, and Bishop, Sports Medicine, 2018. https://pubmed.ncbi.nlm.nih.gov/29934848/
Pilegaard, Saltin, and Neufer, Journal of Physiology, 2003. https://pubmed.ncbi.nlm.nih.gov/12563009/
Porter et al., Medicine & Science in Sports & Exercise, 2015. https://pubmed.ncbi.nlm.nih.gov/25539479/
Ristow et al., PNAS, 2009. https://www.pnas.org/doi/10.1073/pnas.0903485106
Lee et al., Delineating the contribution of ageing and physical activity to changes in mitochondrial characteristics across the lifespan, Molecular Aspects of Medicine, 2024. https://pubmed.ncbi.nlm.nih.gov/38626488/
16. On the Sanctuary site
The Cell Inside the Cell, Mitochondria, Part 1, once posted.
