Sunday, May 25, 2008

Is the secret to good health in a bowl of cherries?

In a word, no. As we all know the secret to good health is NO.

I received a question regarding a product that mentions specific antioxidant activity against peroxynitrite. Rather than leave the answer as a comment I am posting it as a blog. I looked at the CherryActive website and while their product may have abundant antioxidants of certain (and uncharacterized) types there is no compelling evidence on their website or in the literature that consuming supplemental antioxidants (of any type) has any positive health value and some modest evidence that modest levels supplemental antioxidants are actually harmful.

Cherries as a unique neutraceutical? Unlikely at best.

Even supposing that supplemental antioxidants had good effects, the only place those good effects could occur would be inside cells at the sites of free radical formation. Ingested antioxidants must survive digestion in the stomach and gut, be absorbed through the gut into the blood stream, be transported through the portal system, be metabolized by the liver (most likely) and then distributed by the blood stream throughout the body in exactly the correct amounts of each type of antioxidant in each different tissue compartment simultaneously. Is there any evidence that such things might occur with intact cherry antioxidants from this specific type of cherry? (I say intact because that is what the in vitro tests of cherry juice measured, antioxidant capacity before digestion and absorption, not afterward.) Not really. Perhaps it could happen with antioxidants that human physiology evolved to work best with, but such evolution would take a long time, many generations. For specific antioxidants to be a continuous and integral part of the human diet for many generations, those antioxidants could not be unique to a single food type which would only be available in the "wild" for a short period while in season (perhaps 1 month out of the year for cherries).

The only way humans could evolve physiology adapted to cherries as the ideal source of specific and unique antioxidants would be if the lack of the ability to utilize cherry antioxidants produces such a detriment to survival and reproduction that humans who could utilize those antioxidants had a substantial advantage. There certainly has been no report of the kinds of adverse effects from not consuming cherries that would be required to provide sufficient evolutionary pressure for humans to evolve a physiology that matched the phytochemical profile of cherries, particularly not a particular type of cherries. In any case, cherries were domesticated pretty recently, in Europe, Asia and North America (not Africa where most of human evolution occurred).

Free radicals, peroxynitrite, and superoxide: They say it like it is a bad thing.

They cite an article on NO and peroxynitrite (with the author's name misspelled). The article is a pretty good review; however the authors don't seem to appreciate the (extreme) importance of the location of the formation of superoxide to the location of the formation of peroxynitrite. Superoxide is a charged anion (as is peroxynitrite). Lipid membranes block anions very well. Superoxide can't pass through a lipid membrane except through an anion channel. Virtually all superoxide is generated inside of vesicles and is isolated from the cytoplasm by a lipid membrane. Mitochondria do generate a lot of superoxide, but it is vectorally produced to the inside of the inner matrix. There are 2 lipid membranes confining the superoxide, and the mitochondrial potential (~140 mV) keeps anions inside too. Superoxide (as superoxide) can't get out of mitochondria while the potential is intact and/or while the membranes are intact. Superoxide is dismutated to H2O2 and that being uncharged can diffuse out. Acute effects of superoxide or peroxynitrite in mitochondria would be expected to be confined to the inner matrix. Peroxynitrite does decompose producing NO2 which is uncharged and so can diffuse through the lipid membranes. NO2 could diffuse out, but NO2 is quite reactive and so can't diffuse very far. Nitration of mitochondrial proteins not in the inner matrix might be mediated through NO2. There might be a number of important control mechanisms utilizing nitration of mitochondrial proteins with NO2. Nitration of mitochondrial proteins is known to be important. The detailed mechanisms of that are still mostly unknown.

The other main sites of superoxide production are microsomes where a lot of metabolism of xenobiotic (and non-xenobiotic) chemicals takes place utilizing the cytochrome P450 enzymes. In microsomes too, the superoxide is vectorally produced to the inside of the vesicle. Cytochrome P450 enzymes are involved in a great many synthetic operations in the body. That is how cholesterol is synthesized, also steroids, and also how many toxins are rendered less toxic (and sometimes more toxic). Virtually all of the P450 enzymes are highly uncoupled, that is they make a lot of superoxide. Mitochondria only make a few percent superoxide from the O2 they consume, the P450 enzymes make a lot more, in some cases as much as 50% of the O2 consumed ends up as superoxide. This superoxide is important in the regulation of the activity of the P450 enzymes.

Normally the P450 enzymes are inhibited by NO. NO binds to the active site where O2 binds and prevents O2 from binding and beginning the reaction. Normally there is a little activity, so that electrons trickle through and generate enough superoxide to keep the NO level in balance. When the P450 enzyme becomes activated, the superoxide level goes up, the NO level goes down, and the activity of the P450 enzyme goes up, metabolizing what ever normal or xenobiotic chemical needs metabolizing. The pulse of superoxide and other free radicals produced by that the pulse of metabolic activity has important regulatory effects.

The oxidation of xenobiotic chemicals is called Phase I of the drug metabolizing process. This is followed by Phase II of the drug metabolizing process where the oxidation products are tagged with molecules that make them excretable. This includes glutathione, sulfate and even glucose. Glutathione and glucose tag the chemical for excretion in the bile, sulfate tags it for excretion in the urine.

The free radicals and electrophiles produced during the Phase I process cause the expression of the Phase II genes which then detoxify those oxidation products by conjugating them so they can be excreted. There is considerable thought that a significant part of the observed health effects of different phytochemicals in food is due to their metabolism by the P450 system, the normal generation of superoxide and other free radicals, and the stimulation of the Phase II system by those free radicals. It may be the production of free radicals rather than the destruction of free radicals that is mechanism by which fruits and vegetables are protective (if they actually are protective which is not clear). There have actually been no blinded studies of food intake on health.

In vitro, peroxynitrite is only observed when there is near stoichiometric production of superoxide and NO. When either is in several fold excess, peroxynitrite and effects attributed to peroxynitrite are not observed. In other words the problem of peroxynitrite (if is actually is a problem) is likely due to either not enough superoxide or not enough NO, rather than from too much of either. The association of nitrated proteins with adverse health states may be an effect not a cause. Peroxynitrite only occurs when the quantities are essentially equal and equal production occurs for extended periods of time. To me that sounds more like a signaling problem, that the "signal" of peroxynitrite wasn't high enough to switch physiology hard enough to get it back into the "good" state of either excess NO or excess superoxide. If that is the case, blocking peroxynitrite signaling would be a bad thing and would exacerbate peroxynitrite problems.

There is extremely good evidence that free radicals are extremely important signaling molecules that are absolutely essential in the right amounts at the right time in the right place and for the right duration. Free radicals tend to be small, are extremely reactive toward many different molecules, and have a very low background (because they are so reactive). Those properties make free radicals excellent signaling molecules capable of achieving a high signal to noise ratio even at very low levels and with very fast kinetics on a large variety of biological molecules simultaneously over multiple different time scales. The pathways controlled by free radical reactions include just about everything, energy production, gene transcription, the cell cycle, DNA replication, ischemic preconditioning. There are just about no physiological pathways that are not closely coupled to pathways utilizing free radicals as signaling molecules. Those ideal properties of free radicals as signaling molecules also make experiments to understand the signaling pathways those free radicals are involved in very challenging. It is not at all surprising that over evolutionary time organisms would evolve physiology to use such ideal signaling molecules many times, at many places, in many tissue compartments, for many reasons to do many things by many different pathways over many different time scales by many different mechanisms.

Signaling by free radicals as signaling molecules are only going to work well in the media they evolved to work in. That media is quite complex (multiple different phases, i.e. lipid and aqueous of different sizes and compositions), and is different in each tissue compartment (blood, plasma, CFS, cytoplasm, mitochondrial matrix, depot fat, liver, kidney, etc, etc.). Adding different compounds that bind strongly to free radicals (that would be supplemental antioxidants) is virtually certain to perturb that media and interfere with that signaling in complex ways.

All the large, long term double-blind placebo controlled trials of supplemental antioxidants have shown either no effect or a slight negative effect. A good recent meta-analysis summarizing many of those antioxidant trials is in the NEJM. A more recent Cochrane review on the subject (which I have only seen the abstract of) found essentially no benefit from supplemental antioxidants.

To me, that is not at all a surprise. My own feeling is that free radical signaling is too important to organisms for them to allow the background environment that signaling occurs in to be set "at random" by arbitrary (and in the wild highly variable) dietary levels of antioxidants. Humans (and every other organism) didn't evolve for their physiology to be optimized by consuming a certain number of a certain type of cherries (or other specific foods) per day. There simply has to be feedback control(s) that regulate the internal environment(s) the free radical signal(s) propagate in. Oxidative stress is such an important parameter to regulate that organisms "learned" how to regulate their own state of oxidative stress billions of year ago. Bacteria regulate their level of oxidative stress (as does every other organism). Stress response genes, including genes for responding to oxidative stress are among the most highly conserved genes in organisms. Why would we think that humans somehow lost that ability? Especially when there is no evidence that humans have lost it. If a particular human did lose the ability to regulate oxidative stress, death would occur in a few minutes (or less).

There are many compounds in vivo that have antioxidant properties in vitro. The natural modulation of the concentrations of a few dozen such compounds in vivo would be difficult to measure, particularly since the antioxidant properties of these molecules can be different by many orders of magnitude. For example NO is an excellent antioxidant. It has a free electron and reacts with free radicals at near diffusion limited kinetics. It is a better antioxidant than is Vitamin E by about a factor of 5,000.

NO rapidly reacts with peroxidized lipids, quenching the free radical mediated chain reaction. NO reacts with peroxidized lipid radicals considerably faster than does vitamin E (~2E9 M-1 s-1 vs. 5E5), and considerably faster than the propagation reaction (1.3E3). In vitro, 7 nM/min NO was as effective as 50 µM/L α- tocopherol in free radical catalyzed linoleic acid oxidation. Oxidative consumption of α- tocopherol did not occur until NO concentrations fell below ~10 nM/L. The combination of NO and α- tocopherol prevented lipid oxidation better than either alone. A slight increase in the NO level has a much larger effect on antioxidant activity than a much larger increase in Vitamin E levels.

10 nM/L NO is somewhat higher than the normal background (which is around 1 nM/L). 1 nM/L is about 30 parts per trillion by weight. There are no techniques to measure that in vivo on the length and time scales that are important. The only reason we know that the basal level is about that is because at higher levels NO activates sGC and causes local vasodilation which brings in oxyhemoglobin which removes the NO.

Summary antioxidant effects

There is no compelling evidence that supplemental antioxidants of any type have any positive health value. Everything important in physiology is regulated by feedback regulation. Oxidative stress is no exception. Attempting to perturb physiology artificially will simply force it to work harder to maintain the setpoint it is trying to achieve. I think this is the reason why all the large double blind placebo controlled studies show slight negative effects of supplemental antioxidants. Diet choice is a major compensatory pathway by which an organism regulates its state of oxidative stress. When people have high levels of oxidative stress they choose to eat foods with lower levels of antioxidants so the body doesn't need to destroy them metabolically (by generating even more superoxide). Diet choice is an effect of ill health, not the cause.


They mention that cherries seem to have pharmacological effects at reducing urate levels, and imply (or actually state) that lowering urate levels would be a good thing. They cite a paper that does show an acute lowering of plasma urate levels (funded in part by the California Cherry Advisory Board). The reductions reported in plasma urate are quite small and (in my opinion) not physiologically important. The "normal range" for plasma urate (in females) is 139-393 micromoles/L (serum). The levels measured in this study were all within the normal range, the reduction was small and it was an extremely short study (5 hours). Urinary levels of urate did increase in the cherry group. It isn't clear what is the mechanism for the very modest plasma urate reduction. If it is due to an acute inhibition due to cherry flavonoids, in the longer term (days, weeks, months), there may be compensatory increase in the metabolic systems that produce urate such that there would be no chronic effect, or the chronic effect could be to increase urate levels. If it is increased urinary excretion, levels in the urine are what determine the formation of sodium urate kidney stones. Physiology is inherently non-linear with feedback and hysteresis. The only way to determine if a treatment has a long term effect is to measure the effects of that treatment long term.

Gout is caused by the formation and deposition of crystals of sodium urate. Urate is the terminal metabolite of purines which are heterocyclic nitrogen containing bases used in RNA, DNA, ATP and some other fundamental biochemical pathways. Normally they are ultimately degraded to xanthine and then oxidized by xanthine oxidoreductase into urate. For gout to occur, the solubility of sodium urate crystals has to be exceeded in the relevant tissue compartment and nucleation of sodium urate has to occur. If the sodium urate level is below the solubility limit, further reductions in urate levels will have no preventative effect.

The usual treatment for gout is administration of allopurinol which inhibits xanthine oxidoreductase. A great many natural flavonoids also inhibit xanthine oxidase in vitro. Many natural flavonoids have urate lowering activity in vivo. Hypersensitivity to allopurinol is a significant triggering mechanism for Stevens Johnson Syndrome. The mechanism by which that happens is not understood. I would be concerned that any pharmacologically active compound that inhibits xanthine oxidase might have side effect profiles similar to allopurinol. That includes the unidentified compounds in cherries.

It is not clear that a lower plasma urate level is actually a benefit. There is pretty good evidence of an association of higher plasma urate with a reduced incidence of Parkinson's disease. There may also be a protective effect of urate in a number of diseases (I have only seen the abstract of this), but which is cause and which is effect is not clear. Urate is a pretty good antioxidant itself. Trading an insignificant reduction of the possibility of getting gout for a significant increase in the possibility of getting Parkinson's or Alzheimer's disease isn't a choice that I would make.

Without knowing one's urate status, taking pharmacologically active agents to lower it does not seem prudent to me. If one did need to lower one's urate level, using a compound of known composition, purity, efficacy and side-effect profile produced under controlled conditions seems more prudent to me than using some fruit juice. Allopurinol is generic so it is cheap, most insurance would pay for it with the proper medical indication (which is the only reason you want to take anything pharmacologically active), and you need to have your urate status actually measured by your MD in case that joint pain is actually something more serious.

Muscle damage

The CherryActive website claims

"Research study from University of Vermont and published in the British Journal of Sports Medicine shows that drinking cherry juice can help repair damaged muscle leading to an increased recovery rate from strenuous exercise"

The actual report (funded by Cherrypharm Inc and done by researchers who each own 2.5% equity in Cherrypharm Inc) only talks about symptoms and says:

"Although the results of this study indicate a protective effect of cherry juice, it is not possible to conclude that cherry juice supplementation prevented muscle damage, because only two of four indirect markers of damage showed an effect."

They used a proprietary blend of cherry juice and apple juice.

"The cherry juice blend was prepared by mixing freshly prepared tart cherry juice with commercially available apple juice in a proprietary ratio (Cherrypharm Inc, West Hartford, Connecticut, USA). Frozen tart cultivar Montmorency cherries were used to prepare the cherry juice following standard procedures that simulate industrial processing."

They only used subjective measures of muscle function, how the subjects felt about how their muscles were functioning. They specifically chose to not use objective measures of muscle damage, myoglobin and/or creatine kinase. They rationalize their decision in terms of potential confounding due to unauthorized muscle activity not within the protocol they were testing. Why they considered that objective instrumental measures of muscle injury would be more subject to confounding than subjective measures is not explained. I suspect it has to do with each of them owning 2.5% of Cherrypharm Inc. They suggest that future tests might be run using actual measurements but no future papers have been forthcoming. Either they never did the tests, or they did the tests and did not publish them.

They suggest no plausible mechanism by which consuming a cherry juice apple juice blend would have any effects on muscle damage or healing.

The statements on the CherryActive website are clearly not supported by the research they cite. They are being dishonest to pretend that it does.


Cherries are a fine and tasty food. I have nothing against cherries, I like cherries and I eat cherries when in season. Moderate consumption of cherries as part of a varied diet is likely no more and no less healthful than moderate consumption of any other fruits and vegetables. Cherries do tend to be somewhat pricy for me even when in season. The cherry products sold by CherryActive are extremely pricy. I see no reason to consume any of CherryActive's products other than as simple foods, and virtually any other source of cherries would be more economical as are many other fruits and vegetables.

I didn't know very much about cherries when I started looking into CherryActive. I suspect that I will keep my cherry intake more moderate primarily because of the xanthine oxidoreductase inhibition effects. An extremely important property of xanthine oxidoreductase is the reduction of nitrate to nitrite and the reduction of nitrite to NO. NO from nitrite is extremely important in many emergency and extreme stress situations. Nitrite is substantially protective against infarcts due to acute ischemia in heart muscle, liver, kidney, brain. Messing with the xanthine oxidoreductase system by inhibiting it with cherries is not something one should do without good reasons. The reasons that CherryActive provide are not good enough for me.