This is Chris MasterJohn’s post on Urolithin A.
͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏
The claims are it will renew your mitochondria, boost your strength, bolster your VO2max, and act like exercise in a pill. Is this fact or fiction? Here’s what you need to know.
Chris Masterjohn, PhD
Jul 14 ∙ Paid
READ IN APP
Urolithin A is claimed to be exercise in a pill, a potent weapon against aging that will rejuvenate your mitochondria and make you stronger and younger even without exercising, eating well, or fasting.
Dr. Mark Hyman lists it as the best way to clean up old mitochondria and create new ones, central to fixing the feeling of having no energy. Hyman has endorsed Mitopure, the major brand behind the product, as “a real scientific breakthrough. It’s the first product to unlock a precise dose of purified urolithin A.”
Dr. Rhonda Patrick lists it as a central way to address age-related mitochondrial decline.
The Swiss company Amazentis makes Mitopure. They developed it in conjunction with the Swiss Federal Institute of Technology de Lausanne, protected it with 56 patents, and launched a long-term research program on it. “At our core,” they write, describing the purpose of their research program, “we stand against pseudoscience.”
A 2016 paper in Nature Medicine by Amazentis authors launched the mechanistic claims that urolithin A lengthens lifespan in worms and leads to increased endurance and strength in rodents by stimulating mitophagy. Mitophagy is the clearance of bad mitochondria to make way for new, freshly made mitochondria.
Ever since this paper was published, nearly every other paper and review that studies urolithin A cites it as foundational evidence that stimulating mitophagy is its central mechanism of action.
Amazentis-affiliated papers love to call urolithin A a “first.” It is a “first in class” natural plant-based stimulator of mitophagy. In 2019, they published the “first-in-human” randomized, double-blind, placebo-controlled trial with urolithin A.
According to Amazentis-affiliated randomized controlled trials, four weeks of 500-1000 milligrams per day of urolithin A has a similar impact on mitochondrial health as ten weeks of aerobic exercise or twelve weeks of high-intensity interval training.
Better yet, these trials show that 1000 milligrams per day of urolithin A for four months will make you 12% stronger and increase your VO2 max by 14% without even exercising!
Most urolithin A papers take the mitophagy claims at face value without probing deeper into what urolithin A is doing at a mechanistic level. Close examination of the foundational 2016 paper suggests it is stimulating more fundamental pathways shared by other polyphenols, and these include inhibition of complex I of the mitochondrial respiratory chain.
That puts urolithin A into the class of “Good-For-You Toxins” that I described in my Hormesis Lesson, where a little bit of something bad for our cells can provide the stimulus we need to improve our cellular defenses.
It also raises important questions about how much is too much over what period of time, whether it needs to be cycled, and whether it is equally good for everyone or could be hurting some people who are bad candidates for it.
There are some indications from the trials that headaches and gastrointestinal problems are fairly common side effects, indicating not everyone’s metabolism is being optimized.
Further, Mitopure costs $100-125 per month. Is this doing something special over and above what a good diet does for you that is worth that kind of money?
Finally, while Amazentis stands against pseudoscience at its core, the claim that you can increase strength and VO2 max without exercising deserves some closer scrutiny before we throw $125 per month at it.
In this article, we examine the findings of the randomized controlled trials, put urolithin A into its appropriate context as a plant polyphenol, and form a working model of its mechanism of action that can help us properly interpret the trials. From this, we form conclusions about who should be taking urolithin A and why, and how to know if it might be right for you.
This Article Is Reserved for Masterpass Members
You can sign up for the Masterpass here. You can learn more about the Masterpass here.
This is educational in nature and not medical or dietetic advice. See terms for additional and more complete disclaimers.
The Short Answer
- 500-1000 milligrams of urolithin A per day could be used in place of resveratrol in the fasting-feeding reset.
- It does not matter, according to limited evidence, whether it is taken with water or food. Its effects should kick in within 6 hours and be gone within four days.
- It could be one option among many to serve as a substitute for polyphenol component of a diet rich in fruits and vegetables for those who should be eating such a diet but won’t. (Most people should by default, unless lab work suggests metabolic and antioxidant optimization without fruits and vegetables.)
- Long-term use of 500-1000 milligrams is probably mostly safe but good for endurance and bad for power and velocity.
- Testing your fasting and postprandial lactate, aiming for 0.5-0.9 mmol/L fasting and 0.5-1.2 mmol/L postprandially, would be useful to test whether urolithin A is benefitting your mitochondrial health. A rise in either metric would indicate complex I inhibition is doing more harm than good.
- Outside of targeted use in the fasting-feeding reset, urolithin A is an overpriced and partial substitute for a diet rich in fruits and vegetables, a healthy fasting-feeding cycle, and a well-constructed physical practice.
Findings From the Randomized Trials
There are four randomized, controlled trials of urolithin A.
One was an Iranian paper that found four weeks of 500 milligrams twice a day had no effect on low ejection fraction in individuals with heart failure.
The other three are Amazentis-affiliated trials.
In older adults, 500-2000 milligrams of urolithin A upregulated the mitochondrial genes that are downregulated in the pre-frail subjects, non-significantly appeared to increased mitochondrial biogenesis, and decreased acylcarnitines. A decrease in acylcarnitines suggests improved efficiency of fatty acid oxidation. In conjunction with this, several genes involved in fatty acid oxidation were increased. The authors likened these effects to those demonstrated for aerobic and high-intensity interval exercise programs.
In untrained middle-aged adults aged 40-64 who were overweight and had poor physical endurance, four months of 500 and 1000 milligrams of urolithin A increased muscle strength, almost significantly boosted estimated VO2max, and non-significantly tended to increased distance on a 6-minute walk test.
In older adults aged 65-90, 1000 milligrams of urolithin A improved endurance in the index finger and shin muscles at the two-month time point, but was no different from placebo by the end of the study at four months. The six-minute walk test seemed to have greater improvement in the urolithin A group, but it did not reach statistical significance. There was no difference in post-exercise rate of recovery in creatine phosphate stores.
The problem with these papers is they reported almost all of their results as change-from-baseline. As I described in How a Study Can Show Something to Be True When It’s Completely False — Regression to the Mean, this is never appropriate to report without reporting the underlying data and the primary analysis should always be focused on the differences in ending values between the treatment and placebo groups.
Extraordinary claims require extraordinary evidence, a standard popularized by Carl Sagan. While it may not be “extraordinary” to claim a supplement can boost strength and VO2max with no exercise, it is at least a strong claim we should be skeptical of, and it should at least have a rigorous standard of evidence behind it.
The paper in older adults did show absolute data for the muscle contractions:
You can see that the placebo contractions increase in speed , since the number of contractions per unit time is going up from baseline through the trial, while the urolithin A group adds endurance in the form of length of time without ever increasing speed at all. In fact, the increase endurance appears as a horizontal trickle with fewer contractions per unit time being stretched out to the right.
That was for the index finger.
The data for the shin muscle look similar:
They only performed their statistical analysis on the change from baseline in total number of contractions. As stated above, this favored the urolithin A group at two months, but the difference disappeared by four months.
They also reported in the text the absolute values for the six-minute walk. At baseline, the placebo group had 6% more distance than the urolithin A group, which favors better performance in the urolithin A group by random chance. The placebo group improved from 464.6 to 507.2, while the urolithin A group improved from 436.6 to 497.5. In their statistical analysis, this was a 14.9% improvement in the urolithin A group and a 10.1% improvement in the placebo group, which appears as a 48% relative outperformance of the urolithin A group that does not achieve statistical significance. What it actually is, though, is slightly better distance in the placebo group — 507.2 versus 497.5 — that does not reach statistical significance.
Thus, baseline stats were in their favor to get a spurious finding, and they still couldn’t get one that was statistically significant.
In the untrained middle-age adults, the baseline VO2max was reported, and it was very similar across groups. Thus, the improvement in the 1000-milligram group is more likely to be real. However, baseline strength measurements were not reported, so the claims about improved strength cannot be properly evaluated.
Overall, these trials are poorly reported, but there is some indication that urolithin A provides a shift toward improved fatty acid oxidation and better endurance, while doing little for or perhaps antagonizing improvements in power and velocity.
In older adults, urolithin A did not seem to cause greater side effects than the placebo. In middle-aged adults, there was no statistical evidence that side effects differed between groups, and total side effects were about the same, but there appeared to be a dose-dependent increase in headache (4.8% placebo, 16.7% 500 mg, 22.2% 1000 mg), bloating (2.4% placebo, 4.2% 500 mg, 11.1% 1000 mg), and loose stools or diarrhea (0% placebo, 4.2% 500 mg, 8.3% placebo).
When the trials are interpreted in the light of the better mechanistic insights from animal studies discussed below, it is likely that urolithin A is acting similarly to many other polyphenols by stimulating antioxidant defense, inhibiting complex I in the respiratory chain, shifting metabolism away from carbohydrate and toward fat, and promoting mitochondrial turnover.
To genuinely understand urolithin A, we first situate it in its context as a “post-biotic” of plant polyphenols and then examine the mechanistic data in animals.
Urolithin A is a Post-Biotic
Urolithin A is a “postbiotic,” that is, a gut microbiome metabolite of dietary precursors. Specifically, it is derived from microbial metabolism of ellagic acid, which is generated in the human digestive tract from ellagitannins found in a variety of plant foods, especially berries, tea, and nuts.
Urolithin A in Its Polyphenolic Context
Ellagitannins are a class of tannins. Tannins were named from the Latin tannum for “oak bark” as a result of the use of oak bark in converting animal hides to leather. This then became to tan the hide, and someone who tans the hides became a tanner . Upon chemical study of this process, a tannin was designated a biological molecule capable of forming insoluble precipitates with proteins. Later, the active components of oak bark were realized to be polyphenols.
Phenol is the alcohol form of benzene. Phenols as a class are derivatives of phenol, the simplest of the phenols. Polyphenols are molecules made of multiple phenols bound together into one molecule. There are over 8,000 different kinds of polyphenols, and tannins are generally those that are able to precipitate proteins. However, now that tannins are defined by their chemical taxonomy, we know that protein precipitation is a general feature of the class of molecules but does not always apply to each individual molecule within the class.
tannic acid, a representative tannin, belonging to the larger class of polyphenols.
The tannins can be broadly classed into proanthocyanidins, also known as condensed tannins, and the hydrolyzable tannins. Proanthocyanidins play an important role in the flavor and colors of berries, grapes, and flowers. They are related to the catechins found in green tea and chocolate, resembling a stack of epicatechin as their general structure.
The four catechins found in green tea, part of the much larger group of polyphenols. Epicatechin, epigallocatechin, epicatechin gallate, and epigallocatechin gallate.
The general structure of proanthocyanidins, resembling a stack of four epicatechins.
The hydrolyzable tannins can be digested into smaller products. Those that generate gallic acid are gallotannins, and those that generate ellagic acid are ellagitannins. However, some generate both acids, and some generate more complicated variations of them. Proanthocyanidins and hydrolyzable tannins are generally inverse to each other, with the former found in young tissues and dominant in the spring, and the latter found in mature tissues and dominant in the autumn.
Ellagitannins Are Natural Insecticides
There are over 500 different ellagitannins identified. The variation within plant species can be enormous. For example, two birch trees standing side by side have been shown to have 50-fold differences in their ellagitannin content. Insects have a major influence on the composition of specific ellagitannins, but not on the total ellagitannin content.
Ellagitannins generally are either good at precipitating proteins, or have pro-oxidant effects, both of which are thought to serve as defenses against insects. Protein precipitation in the insect digestive system will hurt the nutritional value of the plant, and pro-oxidant effects will damage the health of the insect. Both effects are facilitated by the high pH of the insect gut, and the pro-oxidant effects can be mitigated by the presence of vitamin C or proanthocyanidins, which are antioxidants. Digestion of the ellagitannins down to the simple ellagic acid protects against these effects.
From among foods commonly eaten in modern society, ellagitannins are rich in berries, tea, and nuts, with pomegranate being most often mentioned but with walnuts and jabuticaba, the fruit of the Brazillian grapetree, being highest.
Punicalagin is the ellagitannin that dominates in pomegranate, and its high concentration in yellow-wood has been identified as the cause of liver and kidney toxicity in cattle, though it doesn’t seem to have such toxicity in rats.
Punicalagin, the dominant ellagitannin in pomegranate.
Urolithin A as a “Post-Biotic”
In the human digestive tract, ellagitannins are digested down to ellagic acid. Some microbes of the gut then convert ellagic acid to some combination of four urolithins: A, B, C, and D. Of these, urolithin A is best studied, partly because it is the predominant urolithin produced, and partly because it appears to have a superior longevity benefit in worms.
In healthy adults ages 20-80, 40% of people produce considerable urolithin A from pomegranate, 27% produce small amounts, and 33% are non-converters. Upon stool analysis, urolithin A production is associated with a high ratio of Firmicutes to Bacterioidetes, and with the abundance of the Clostridiales and Ruminococcaceae families. The high-converters have a microbiome especially rich in Akkermansia muciniphilia .
500 milligrams of Mitopure urolithin A produced the equivalent of high-converters drinking six cups of pomegranate juice. However, it is noteworthy that walnuts contain 6-15 times the amount of ellagitannin as pomegranate, so walnuts could easily compete with Mitopure in a high-converter.
Pomegranate juice produces peak urolithin A 24 hours after consumption, whereas Mitopure produces the peak 6 hours after consumption. This is because ellagitannins have to make their way down to the colon to get converted, whereas urolithin A supplements do not require any conversion.
Urolithin A Is Fairly Similar in Structure to Epicatechin and Resveratrol
Note that, to take two other example polyphenols, urolithin A is fairly similar in shape and size to resveratrol and epicatechin. Epicatechin and the other catechins are flavonoids, and flavonoids as a class are similar to it in structure. Urolithin A and epicatechin both have two phenolic rings on either side of an oxygenated ring, whereas resveratrol has two phenolic rings on either side of a carbon-carbon double-bond. Urolithin A has a simpler and more condensed arrangement compared to epichatechin.
Small differences in structure can lead to large differences in some effects, but they are close enough that we should not be surprised if they share some effects in common.
Urolithin A Is Treated As a Toxin
No studies have identified the percent of urolithin A that is absorbed from supplements or from conversion of ellagitannins, but it is likely very small.
Indeed, a study in mice found that encapsulating urolithin A in nanoparticles increased its bioavailability 7-fold, indicating that it cannot possibly have higher than 14% absorption, though even that is likely to be an extreme overestimate.
Urolithin A is unambiguously treated as a toxin by the body and subject to extensive transformation by the liver’s detoxification system. As such, most of it appears in plasma conjugated to glucuronide or sulfate — two major detoxification substances — and it is almost fully eliminated from the body within 72-96 hours.
A very small amount of unconjugated urolithin A sneaks its way into tissues and can be detected in a muscle biopsy.
In general, polyphenols are subject to small single-digit percent absorption as a result of the intestinal detoxification system, and that is likely true for urolithin A.
For example, for epicatechin, 3.4% of the dose is absorbed, but half of that is almost immediately effluxed into the bile, bringing the net absorption to 1.2%. 20% of what is absorbed reaches the circulation, mostly as sulfated, glucuronidated, or methylated metabolites of the liver’s detoxification system, and it is all gone within 3-4 days.
Polyphenols and Hormesis: The “Good-For-You Toxins”
As covered in my lesson, Beneficial Toxins: Phytochemicals, Hormesis, and Nrf2, the reason our bodies treat polyphenols as toxic is because they are. If the bioavailability were high and they were retained in the body, they would kill our cells. However, since the bioavailability is low and they are rapidly eliminated, a diet rich in polyphenols provides low concentrations that stimulate our ability to defend ourselves against worse toxins and against reactive oxygen species that we make in our own metabolism. The principle that a little bit of something bad can be good by promoting an adaptive response is known as hormesis.
This also implies that too much can be bad, and that some people may have nutritional or genetic limits to their ability to ramp up their defenses, making relatively low doses become harmful. This is supported by many case reports of green tea supplements causing liver failure.
One of the central mechanisms by which polyphenols engage in hormesis is by forming adducts with sulfur-based thiol groups. This means they can form adducts with glutathione and deplete it. But it also means they can stimulate the cellular thiol-based sensor of oxidative stress, Keap1, which makes the transcription factor Nrf2 enter the nucleus and upregulate genes involved in the defense against cellular stress and xenobiotics.
Since this principle broadly applies to polyphenols, we should use it as a default framework for interpreting the urolithin A research.
Urolithin A Lengthens Lifespan in Worms
In the worm C elegans, urolithin A lengthens maximal lifespan from just under 30 days to just under 40 days. The other urolithins also lengthened lifespan, but urolithin A had the greatest effect.
In this 2016 paper, the authors commented that “until now, there has not been an in-depth investigation into the mechanism of action of urolithins and their benefits following chronic administration,” and that while there had been some research in this area, “a clear biological pathway has not yet been described.” They then showed that it simulates mitophagy, the breaking down of damaged mitochondria, and called it a “first-in-class natural compound that induces mitophagy.”
However, they did not map out the exact pathway of mitophagy induction, and the hints they developed point back to a Nrf2-centric explanation with hints of respiratory chain inhibition also shown for other polyphenols like resveratrol and catechins.
So, it’s unlikely urolithin A is first in anything, except to have its lifespan-lengthening research take on a mitophagy-centered approach.
What Is the Mechanism of Urolithin A?
The lifespan effects were partially antagonized in worms missing AMPK, a sensor of nutrient depletion, and they were completely abolished in worms missing succinate dehydrogenase, an enzyme that feeds food energy directly from the citric acid cycle into the complex II of the mitochondrial respiratory chain.
On day 1, urolithin A decreased mitochondrial density and nearly cut ATP levels in half. By day 8, however, the “old” worms had their level of ATP and mitochondrial density recover. Moreover, they were protected against age-related declines in motility and in muscle fiber organization quality. The fact that their motility stayed high suggests that the lifespan-lengthening was not simply due to spending more time in a dormant “dauer” state, which can often be the case in these worms, especially under conditions of stress or nutrient deprivation.
These data were consistent with the idea that urolithin A first degraded less healthy mitochondria and then led to the synthesis of new healthier mitochondria to replace the old ones. That is, it stimulated mitophagy followed by mitochondrial biogenesis.
The authors dismissed a relation to “oxidative stress” by showing there was no interaction with a specific glutathione S-transferase enzyme, but that is not sufficient to dismiss a relation to “oxidative stress” as a whole.
They claim that urolithin A did not increase reactive oxygen species by showing it did not increase activity on the Mitosox assay, but the Mitosox assay is specific to superoxide, not “oxidative stress.”
Mitochondria are widely claimed to generate superoxide that is then converted to hydrogen peroxide by superoxide dismutase. However, it has been shown that they can generate hydrogen peroxide directly. For example, in complex I, there are at least two sites that generate reactive oxygen species, one that generates superoxide and one that generates hydrogen peroxide. Under high concentrations of NADH, superoxide production declines and hydrogen peroxide production picks up. Thus, a negative result on the Mitosox assay cannot rule out or even comment on whether hydrogen peroxide levels were increased.
They showed robustly that mitophagy was stimulated in the worms and in mammalian cells in multiple different ways. Further, silencing multiple different genes involved in autophagy or mitophagy all blocked the lifespan extension, showing that mitophagy was necessary for the benefit to lifespan.
But why was mitophagy stimulated?
Silencing skn-1 blocked the lifespan extension as well. Skn-1 is the worm version of Nrf2. As the authors noted, it had been shown the year earlier that skn-1 is a central regulator of mitophagy and mitochondrial biogenesis.
Indeed, urolithin A protects mice from acetaminophen toxicity by stimulating Nrf2, with computer modeling suggesting it directly binds to Keap1-Nrf2 complex, and Nrf2 regulates mitophagy in mammals, just like skn-1 does in worms.
So, rather than “first in class” we are now back to “polyphenols stimulate the Nrf2 pathway.”
Still, there is more to the story than this.
The authors claim that urolithin A did not inhibit the respiratory chain because it did not inhibit the basal consumption of oxygen.
But consumption of oxygen is not a reliable metric of respiratory chain function. Inhibiting complex IV in cancer cells allows the persistence of other oxygen-consuming pathways of generating ATP. As covered in my article, Methylene Blue: Biohacker’s Delight, or Playing With Fire?, methylene blue rewires the respiratory chain in multiple different ways that allow ATP to be produced while oxygen is converted to hydrogen peroxide instead of being converted to water in complex IV.
These are inefficient methods of ATP production that produce considerably less ATP than the standard canonical pathway of the respiratory chain. Thus, oxygen consumption can persist despite a loss of respiratory chain efficiency and the generation of less ATP.
The authors tested the functioning of several respiratory chain complexes during urolithin A treatment in mouse muscle stem cells:
This is looking at oxygen consumption in response to different fuels rather than the more rigorous metric of monitoring electrons flows through specific complexes to their targets.
Point 1 in the figure represents fueling the cells with pyruvate, malate, glutamate, and NAD+, all of which allow formation of NADH that feeds into complex I.
The figure makes it look like point 2 is measuring complex I and II together, but according to the description that point represents fueling the cells with succinate, which goes directly to complex II. However, the authors seem very convinced that this figure shows specific complex I inhibition, so it may be that the description of the methods is misleading, and point 2 reflects the addition of succinate to the other fuels. I have reached out to the authors and will update this if I hear from them, but tentatively I will accept their interpretation that this shows urolithin A decreasing complex I activity but not complex II activity.
The authors seemed confused about basic biochemistry and stated that the impaired oxidation of pyruvate, malate, and glutamate was “indicating that [urolithin A] downregulates aerobic glycolysis and thereby indirectly inhibits complex I (CI) respiration,” but there is no glycolysis involved at all in feeding cells with pyruvate, malate, or glutamate, and there is no evidence in the entire paper that the complex I inhibition is indirect.
They showed that the protein expression of succinate dehydrogenase, part of complex II, was upregulated 10-fold, and proteins of complexes III, IV, and V were upregulated 4-5-fold. This led them to conclude complex I was specifically but indirectly impaired and the rest of the respiratory chain was upregulated in compensation, especially complex II.
When they fed the mouse cells with a little bit of glucose and a large amount of fatty acids, urolithin A increased oxygen consumption:
On the left is without an uncoupler, and on the right is with an uncoupler. We can get the main point by focusing on the left.
Whereas urolithin A hurt the oxygen consumption response to pyruvate, glutamate, and malate as fuel (point 1 in the earlier figure), it helped the oxygen consumption response to fatty acids as fuel (left side of the figure above).
Misunderstanding another point of basic biochemistry, the authors argued that fatty acid oxidation was feeding FADH2 into the upregulated complex II. However, FADH2 from fatty acid oxidation does not feed into complex II. It feeds into the enzyme ETF dehydrogenase, which directly feeds energy into CoQ10.
Even though the authors misunderstood how this works, it still makes sense that complex I activity was impaired and complex II activity upregulated. Fatty acid oxidation generates much less NADH than carbohydrate oxidation, so the use of complex I to oxidize NADH will be dramatically lower. For fatty acids, every two carbons require one turn of the citric acid cycle, whereas for glucose every three carbons do. So the citric acid cycle enzymes have to increase, including succinate dehydrogenase, to metabolize even the same number of carbons. Less ATP is generated per carbon when skipping over complex I, so the citric acid cycle would have to increase even further to accommodate a larger flux of carbons to maintain ATP concentrations constant.
These results are consistent with the lifespan extension in worms being dependent on succinate dehydrogenase.
In the mouse muscle stem cells, the mitochondrial membrane potential, which reflects the use of energy to pump hydrogen ions across the membrane to use their backflow to synthesize ATP, was decreased. Alongside this, ATP levels dropped. The decrease in the membrane potential then stabilized PINK-1, a mediator of mitophagy.
Nrf2 acts in the nucleus to increase the expression of the PINK-1 protein, so these processes are likely synergizing. Nrf2 activation (or skn-1 activation in the worms) allows PINK-1 production, and impairing mitochondrial ATP generation activates PINK-1 to initiate mitophagy.
While the PINK-1 mechanism is almost certainly only one part of the picture, as broad principles it is likely that Nrf2 represents the nuclear arm of mitophagy induction and the decrease in the mitochondrial membrane potential represents the mitochondrial arm of mitophagy induction.
The decline in ATP also activates the AMPK pathway, a major signaling arm of nutrient deprivation, fasting, and caloric restriction.
The authors seem to throw up their hands at explaining the decrease in membrane potential, because, supposedly, no reactive oxygen species were generated and there was no inhibition of the respiratory chain. Yet they showed neither of these things. They showed with Mitosox that there was no increase in superoxide generation, and they showed that oxygen consumption was not different, yet the oxidation of NADH-generating fuels was impaired. It is entirely possible that urolithin A directly inhibited complex I in a manner that enhanced its production of hydrogen peroxide but not superoxide.
To complete their story, they should have looked at structural interactions between urolithin A and the respiratory chain complexes rather than prematurely dismissing a direct effect.
A more recent paper showed green tea catechins lengthen lifespan in C elegans by inhibiting complex I. They showed a decline in mitochondrial function on day 1 followed by a later restoration, just like with urolithin A. The authors suggested this increased reactive oxygen species and stimulated hormesis through skn-1 . They did not investigate mitophagy, but their results produced hints of striking similarity to the urolithin A paper. The extension of lifespan was small compared to urolithin A, but elsewhere it had been shown that much higher concentrations of EGCG produce results comparable to urolithin A.
Resveratrol has been shown to inhibit complex I in some contexts, and separately to lengthen lifespan in C. elegans with the gains comparable to urolithin A if high enough concentrations are used.
While nearly all urolithin A papers published since this 2016 paper refer to it as a stimulator of mitophagy, this seems to be a result of its combined stimulation of the Nrf2 pathway and complex I inhibition. These are better characterized as broad effects of polyphenols rather than specific effects of urolithin A. Indeed it is highly likely that many other polyphenols are doing the exact same thing.
Other Effects in Animals
In the 2016 Nature paper, several animal experiments documented in much less detail than the worm experiments were described in the last section. These showed 25 milligrams per kilogram bodyweight urolithin A increased grip strength and running endurance. Analysis of muscle tissue showed less assembled complex I, a non-significant trend toward greater assembled complex II, and a stimulation of autophagy.
In 2019, it was shown that in mice fed a high-fat diet, urolithin A slightly improved glucose tolerance, substantially lowered fasting glucose and increased insulin sensitivity, stimulated mitochondrial biogenesis in liver and adipose tissue, and increased the expression of antioxidant enzymes like superoxide dismutase and genes involved in fatty acid oxidation like CPT-1. Secondary to all of this it predictably lowered inflammation. The upregulation of both superoxide dismutase and CPT-1 can be attributed to Nrf2 activation, although Nrf2 was not investigated in this study.
Another 2019 paper paper that year showed that in mice made diabetic with a high-fat diet and a low dose of streptozotocin, usually used in higher doses to induce type 1 diabetes but here said to induce type 2, urolithin A decreased fasting glucose, excessive water consumption and loss, and glycated hemoglobin, and helped normalize antioxidant and inflammatory markers. They showed this was associated with mTOR activation and blocked by “the autophagy inhibitor chloroquine,” and they interpreted this as part of stimulating autophagy. However, chloroquine is not a specific inhibitor of autophagy and this paper is mechanistically weak.
In yet another 2019 paper, “induction of mitophagy” with NAD+, urolithin A, and the antimicrobial compound actinonin improved Alzheimer’s-related pathology in mouse and worm models of the disease. They demonstrated the dependence of this effect on PINK-1, parkin, and DCT-1, all involved in mitophagy, but they didn’t look at Nrf2/skn-1, which regulate all these genes.
In a fourth 2019 paper, urolithin A encapsulated in nanoparticles for higher bioavailability protected mice against acute kidney injury caused by the anticancer drug cisplatin. Nrf2-regulated genes were massively upregulated by cisplatin and mitigated by urolithin A. This seems to argue against Nrf2 stimulation as a primary mechanism of action, yet this paper did not look at anything happening in the liver. Glutathione protects against cisplatin injury, and the liver is the major site of glutathione production. A major portion of cisplatin’s effect on kidney toxicity is from increasing the liver’s production of indoxyl sulfate from tryptophan metabolites. So it is possible urolithin A promoted Nrf2 activation in the liver and this protected against the kidney toxicity, leading to a lower requirement for Nrf2 activation in the kidney.
In 2022, it was shown in mice that urolithin A protected against acetaminophen toxicity. This is very well established to occur through glutathione depletion. Inhibiting mitophagy by suppressing the expression of the Atg5 gene, which codes for an important gene involved in autophagy and mitophagy, did not antagonize the benefit, but silencing Nrf2 abolished it. Computer modeling suggested urolithin A binds directly to the Keap1/Nrf2 complex.
Conclusions on the Mechanism
These papers offer little basis to revise any mechanistic implications of the 2016 paper, although the last paper in mice strengthens the centrality of Nrf2.
Thus, the best working model for understanding how urolithin A works is that it activates Nrf2 and inhibits complex I. Part of this effect results in Nrf2 increasing the transcription of the nuclear PINK-1 gene, and part of it results from impairing the mitochondrial membrane potential and thereby stabilizing and activating PINK-1, collectively leading to the stimulation of miotphagy. More broadly, Nrf2 stimulation represents the nuclear arm of mitophagy induction, and impaired complex I activity represents the mitochondrial arm of mitophagy induction. Both Nrf2 activation and complex I inhibition will have many other effects, such as AMPK activation and upregulation of antioxidant and xenobiotic defense. Overall, a major net effect is to degrade existing mitochondria and replenish them with fresher mitochondria that have a higher proportion of complex II to complex I and are primed to handle more fatty acid oxidation and less glucose oxidation. This is probably good for endurance and bad for power and velocity, as hinted at in the randomized trials.
What Can We Learn From the Mechanism?
It is likely that pharmacological inhibition of complex I with high-dose metformin or polyphenols has some value as a partial fasting mimetic.
On the one hand, ATP levels drop, and AMPK is activated. This is a general feature of fasting.
Inefficient extraction of ATP from food will lower insulin release, since insulin responds directly to ATP (or rather to ATP-induced influx of potassium), not to glucose. Lower insulin is also a general feature of fasting.
Flux through complex II and ETF dehydrogenase will rise, accommodating increased fatty acid oxidation. Flux through complex I due to glucose utilization will decrease. This is a general feature of the fasting state.
On the other hand, the NADH/NAD+ ratio will rise, a general feature of the fed state.
This will cause the levels of citrate to rise, a general feature of the fed state.
The ratio of ATP to ADP or AMP, the ratio of NADH to NAD+, and the concentration of citrate are all major signals of the fasting and fed states.
During exercise, muscle-specific changes mimic fasting, but power and velocity require more glucose utilization and flux through complex I, while endurance requires more fatty acid oxidation and flux through complex II and ETF dehydrogenase.
Neither under fasting or exercising conditions will you ever get a higher NADH/NAD+ ratio or a rise in citrate.
Thus, urolithin A is a partial mimetic of the fasting state and the endurance exercise state, but is not a complete mimetic of either.
For people or animals who are generally overfed and under-exercised, having some stimulus of the fasting state instead of none is likely a major benefit.
But is it something we want if we also go through our own natural fasting-feeding cycle and get abundant movement of the right types? That is not so clear.
It is likely that a diet rich in polyphenols provides beneficial activation of Nrf2. Does this have the same value in people who exercise abundantly, generating hydrogen peroxide that will also oxidize the thiol groups of Keap2 and stimulate Nrf2 transcription? Probably not.
One of the major problems with all of the urolithin A studies is that they never take into account the background polyphenol intake, let alone exercise, of the study populations.
As such, there is no basis to believe that urolithin A is preferable to, or even adds anything to, a proper fasting-feeding cycle with a well-constructed exercise program.
The Bottom Line
Here are the uses of urolithin A as I see them:
- 500-1000 milligrams of urolithin A per day could be used in place of resveratrol in the fasting-feeding reset.
- It does not matter, according to limited evidence, whether it is taken with water or food. Its effects should kick in within 6 hours and be gone within four days.
- It could be one option among many to serve as a substitute for polyphenol component of a diet rich in fruits and vegetables for those who should be eating such a diet but won’t. (Most people should by default, unless lab work suggests metabolic and antioxidant optimization without fruits and vegetables.)
- Long-term use of 500-1000 milligrams is probably mostly safe but good for endurance and bad for power and velocity.
- Testing your fasting and postprandial lactate, aiming for 0.5-0.9 mmol/L fasting and 0.5-1.2 mmol/L postprandially, would be useful to test whether urolithin A is benefitting your mitochondrial health. A rise in either metric would indicate complex I inhibition is doing more harm than good.
- Outside of targeted use in the fasting-feeding reset, urolithin A is an overpriced and partial substitute for a diet rich in fruits and vegetables, a healthy fasting-feeding cycle, and a well-constructed physical practice.
