Placebo and nocebo effects

Involovement of nitric oxide

Some of this derives from a couple posters I presented at a scientific conference. Send me your email address (via posting a message with email and then deleting it) and I will send you a copy. I get quite a bit into ATP physiology, but that is necessary to understand why the placebo effect actually works to promote healing.

One of the strongest effects in medicine is the placebo (I will please) effect, yet some of the known physiological mechanisms behind this effect are not well known or appreciated, and the details are still not fully understood. The nocebo (I will harm) effect is similar, and is usually considered to be the opposite of the placebo effect, but actually is somewhat different.

Contrary to popular belief, both of these effects are misnamed, in that harmful effects can occur via placebos, and beneficial effects via nocebos. Contrary to popular belief, these placebo and nocebo effects are quite real and can be completely indistinguishable from effects due to efficacious treatments, including improved healing, decreased pain, and these effects can often be detected instrumentally.

Virtually any treatment can have some effect, including those that have no conceivable physical mechanism for working, including homeopathy, Chi manipulation, prayer, sacrificing animals, treatment of surrogates, sham devices and pharmacologically inert pills. An interesting rapid response to this article is from a Dr. Jane Woo (who I presume is perhaps "the expert" on placebos), saying "One of my residents once said that he advocated morphine injections, as opposed to tablets, because, "There's something about steel hitting skin and having a doctor say, 'This is going to make you feel better.' Injections simply work better than pills.""

One of the earliest "treatments" that children receive from their mothers is known as "kiss it and make it better". Any parent anecdotally knows that this is an "effective" treatment. While saliva does have nitrite from the reduction of salivary nitrate by commensal bacteria on the tongue, usually a motherly healing kiss is insufficiently slobbery to transfer sufficient nitrite, and the therapeutic effect is faster than nitrite or NO transfer from the treated boo-boo would allow.

Most placebo research has been associated with pain relief, and increased analgesia from placebo effects is well documented. I will focus this discussion more on non-pain effects of placebos and nocebos, and specifically on how placebos actually do improve healing.

It is well known that physiology is extremely complicated, and the regulation achieved by normal physiology is exquisite. So exquisite, that it has been endowed with the imaginary and mythical property of "homeostasis". In reality, nothing in physiology is static, rather our inability to measure the changes that we know must be present simply leaves us ignorant of those changes. While the default assumption of stasis is simple, it is clearly an assumption based on ignorance, and is clearly wrong. But this blog is about placebos, not homeostasis.

One of the best regulated physiological parameters is the ATP concentration. Not surprising because ATP is used by just about every physiological process, hundreds of thousands, if not millions of different pathways, in each cell, regulated simultaneously.

So how is ATP regulated? The answer is, extremely well!

ATP is considered to be one of the mythic "homeostatic" parameters, that is regulated to be constant. But that cannot be correct. The only way a parameter can be regulated is via feedback which necessitates a deviation from a setpoint followed by a compensatory response. Our inability to measure that deviation does not mean it does not exist. It only shows that our instruments are insufficiently sensitive and precise.

ATP cannot be measured in individual cells on the length and time scales where it matters, at least not non-destructively. The usual way is to freeze the tissue by clamping it with liquid nitrogen cooled copper tongs, then taking a small piece and assaying it. Depending on technique, the assay might be the average of only 100 cells or so. More likely a few thousand.

We know that ATP is regulated within individual cells because ATP doesn't diffuse through lipid membranes. It is difficult to get any information on the dynamics of ATP production and consumption by destructively measuring the average of a few hundred cells. Most any parameter would look pretty "constant" if the only measurements were averages of a few thousand values. If the only measurements of heart rate were averages over an hour (3600 seconds) wouldn't it look pretty constant too? You could measure the increase due to a marathon (2-3 data points), but not from a 4 minute mile.

ATP is not stored; the instantaneous production always equals the instantaneous consumption. If this were not the case, then there would be rapid accumulation or depletion of ATP. Each mole of glucose that is oxidized produces about 38 moles of ATP. A 2000 calorie diet then produces some 55 kg of ATP per day. Obviously, the ADP and P are recycled, and ATP production and consumption is very precisely matched.

In all control systems, the sooner you can start pulling levers to change things, and the more levers you have to pull, the better control you can exert. Physiology is no exception. Invoking physiological pathways in anticipation of a need for that pathway improves performance of the system. Neurological control of some aspects of physiology is well known. It would not be surprising if other aspects were controlled as well.

The key to understanding physiology is to understand that everything is a compromise, just like engineering. Every task that physiology needs to do has a cost in terms of ATP production to do the task, manufacture of molecules to carry out the task, DNA to code for the information to make the molecules to do the task, machinery to turn the DNA into what ever molecules are needed when they are needed, and the control system to do all of these things at the right time and in the right place, and time for all of this to happen. All of these things take up space, and have a "cost" in terms of maintenance and a "cost in forgone reproduction. Cells, and ultimately organisms that did these things most "efficiently" had more resources to use on reproduction, and so had more descendents, and so are the extant organisms we observe.

I will consider two extreme metabolic states, the "fight or flight" state (FoF), and the resting and relaxed state (RnR). It is well known that organisms can invoke such states and FoF results in the characteristic physiological effects of "stress", and if prolonged can cause physical disorders associated with "stress". RnR reverses the effects of stress, but to do so, has to occur within a certain time period.

I suggest that placebos invoke the RnR response, and that nocebos invoke the FoF response. Usually then, a placebo will promote healing and other effects associated with rest, and a nocebo will promote effects associated with stress, which usually are detrimental, but can be beneficial, as in the reduction in nausea, also in this example. The expectation of low nausea via a placebo produced greater nausea than the expectation of high nausea via the identical substance as a nocebo. In the last example, gastric tachyarrhythmia was instrumentally measured to be greater with placebo than nocebo, thus the increased gastric symptoms were not merely subjective. An explanation for this is discussed later.

We know that ATP production and consumption can vary by large amounts. Muscle for example can increase its oxygen consumption by a factor of 10. An interesting property of muscle is that it can be worked to exhaustion that is until ATP levels fall enough that the cells necrose and die. A useful "feature", when you are running from a bear. Death from exhaustion is balanced by death from being eaten by a bear. Physiological systems evolved to minimize death due to the sum of both events. Presumably, organisms do this "efficiently". That is they allocate ATP in an optimum manner to maximize organism survival. That is, ATP consumption is prioritized. Since ATP not used by a low priority pathway is as good as ATP produced, the "optimum" control system for ATP consumption will allocate ATP to the most vital pathways first, and when demand exceeds supply, turn off low priority pathways. What are "low priority" pathways? Well, anything that takes a long time (longer than the ATP crisis) must be low priority.

In the running from a bear example, anything that takes longer than the escape time can be put off. If what ever that pathway is going to produce won't happen until after you have either escaped or been eaten, it can't contribute to your escape, and so can be shut off to supply a few more molecules of ATP. In the limit of a perfectly regulated ATP system, the longest term pathways would be shut off first, then shorter term and later, still shorter term pathways as ATP demand exceeds supply. When the time horizon of the pathways being shut down reaches the present is when you drop dead from exhaustion.

One thing that can be put off is cell maintenance. Maintaining cells consumes ATP. Damaged proteins are ligated to ubiquitin, carried to the proteasome for disassembly, first unfolded by ATP powered unfoldases, and then broken into little bits by ATP powered proteases, and then replacement proteins are manufactured. All of these steps require ATP. No doubt some damaged proteins don't need to be removed immediately, but can be stored until later, until after the ATP crisis is over.

Accumulation of damaged proteins, as in amyloidosis, is common in virtually all of the disorders that are exacerbated by "stress" including obesity, diabetes, end stage kidney failure, dilative cardiomyopathy, neurodegenerative diseases, and so on. Accumulation of damaged proteins is harmful, but cells can tolerate quite large quantities and still remain viable, and there is some suggestion that such aggregation is actually protective. From my own experience (as a bachelor living alone), greater quantities of garbage can be tolerated if they are aggregated in a few places, rather than when distributed uniformly. I suggest that the "problem" is not so much increased production of these damaged proteins, but rather reduced removal. In "steady state", removal must equal production. If there is accumulation, then production exceeds removal. Presumably aggregate removal is regulated, and the "problem" of accumulation is either dysregulation, or precise regulation about a bad setpoint. The dysregulation hypothesis requires multiple cells and cell types to simultaneously develop the same dysregulation, an implausible coincidence. Since removal requires ATP, and some accumulation can be tolerated, I suggest that chronically low ATP will cause accumulation due to a bad setpoint. The "setpoint" is a function of energy status, and if there isn't enough ATP, clearing bad proteins gets put off until later. How much later? Until ATP is back up where it "should" be. What if that never happens? Then you are SOL (shit out of luck). Remember, physiological pathways can't "compensate" because it is precisely those pathways that are affected.

Reduction in metabolic rate is well known in degenerative diseases, for example in Alzheimer's there is a very well documented reduction in brain metabolism that precedes pathology. For a very striking image of this look here. As well as a reduction in metabolic activity, there is a reduction in blood flow. Blood flow is regulated by the vasodilatation produced by NO, and the vascular changes observed in Alzheimer's are consistent with low NO.

What form does the reduction in metabolism observed take? Is it a failure of 1, 10, or 100 or more pathways (of the millions the cell regulates) each pathologically consuming a little less ATP? I would presume that if only a few pathways were involved, they would necessarily represent a large fraction of normal brain metabolism, and disruption of such presumably important pathways would likely have effects more prompt than slow degradation over years. If many pathways are involved, how can many pathways simultaneously "go bad" in diverse areas of the brain? On the other hand, if it is a "bad setpoint", that is if ATP is low because of low NO (discussed later), then the brain would "gracefully" consume less ATP via the extremely robust ATP consumption hierarchies by turning off the least important pathways first, the long term maintenance pathways. This could go on for years, and may even reverse itself at times as the cells go down the low NO death spiral.

The proteasome disassembles proteins one at a time. Larger damaged assemblies including damaged mitochondria can only be disposed of via autophagy. Mitochondria have a finite life. In the rat, CNS mitochondria turnover in about a month. In other organs the turnover is faster. Mitochondria are tricky to recycle, particularly damaged mitochondria because they can be sources of superoxide and hydrogen peroxide, and because they contain abundant Fenton reactive metals which turn hydrogen peroxide into hydroxyl radicals which will damage anything they touch. I will discuss the mechanisms for putting off of autophagy during times of low ATP in a future blog. This is also directly mediated by low NO and low ATP.

Mitochondria are unique, in that they have their own DNA and ribosomes, and manufacture some of their own proteins. The vast majority of mitochondrial proteins (perhaps a couple thousand) are coded in the nucleus, synthesized in the cell's ribosome, and ported into mitochondria during mitochondria biogenesis. Only 13 proteins are coded for by mitochondrial DNA, the active sites of the respiration chain complexes. The vast majority of the complexes are coded for in the nucleus, but these are regulatory subunits, not the active sites.

The number of pathways that consume ATP is not small. For the purposes of this analysis, we need to look at each pathway separately. Rather than look at generic "protein synthesis", we need to consider synthesis of proteins (protein aaaa, protein aaab, protein, aaac… protein zzzz) all separately because that is how they are regulated. Under conditions of ATP depletion, expression of some proteins is upregulated, heat shock proteins and others. I have denoted each protein by 4 letters because that is about how many different proteins are expressed, 26^4 ~ 10^5. Each protein has on the order of a few hundred amino acids, so the number of individual steps that are involved is many millions. We know that the expression of each protein is controlled "just right" because if it wasn't, either there would be not enough, or the cell would explode from too much.

So, how does a cell control a few million pathways and prioritize them based on ATP level? What can it use as a "signal"? I suggest that it must use ATP itself. There are not enough other molecules for it to use a different molecule for each one; some must be controlled by the same molecule, but by different concentrations. For this discussion I am not particularly concerned with the mechanism(s) involved. No doubt there are many.

In a cell, there are 3 ATP parameters, ATP concentration, ATP production rate, and ATP consumption rate. These 3 parameters are independent, and can (and are) controlled independently. Since muscle can consume ATP to the point of death, low ATP will necessarily stimulate maximum ATP production. Just short of death, the cell will "want" to turn off all non-essential systems to stave off ATP depletion for as long as possible. So low ATP turns off the "housekeeping" pathways. So what sets the ATP concentration? In part, that is set by NO via soluble guanylyl cyclase and cGMP.

When physiology calls for maximum ATP production, one of the first things it does is lower NO levels, to disinhibit cytochrome c oxidase. Under basal conditions, cytochrome oxidase is mostly inhibited by NO, which blocks O2 from binding and being reduced to water, the ultimate sink for electrons. O2 consumption can go up an order of magnitude. That means an order of magnitude more O2 must diffuse to the mitochondria and be reduced to water. O2 is only transported by passive diffusion. In the lungs, O2 diffuses into the blood and is absorbed by hemoglobin forming oxyhemoglobin. The blood carries the O2Hb to tissues where the O2 comes off and diffuses to the mitochondria down a concentration gradient. The lowest O2 concentration in the body is at the mitochondria where the O2 is consumed. For the flux of O2 to increase by an order of magnitude, the O2 concentration gradient must increase an order of magnitude. How does this happen? The concentration in the blood doesn't change, the spacing between vessels and mitochondria doesn't change much, so to increase the flux, the concentration at the mitochondria must drop by an order of magnitude. Then with the higher gradient, more O2 can diffuse to the more active mitochondria and more ATP can be produced. The O2 consumption by cytochrome c oxidase increases an order of magnitude while the O2 concentration drops an order of magnitude. The specific O2 consumption (moles O2/Torr O2/mg protein) must go up 2 orders of magnitude. This is accomplished by lowering the NO level local to the mitochondria.

So the low NO necessary for disinhibition of cytochrome c oxidase also serves to lower the ATP setpoint. This lowers the ATP concentration, which turns off non-essential systems. The lower ATP concentration upregulates ATP production by the mitochondria. When mitochondria don't have enough O2, the respiration chain becomes reduced, what little O2 is present becomes reduced by single electrons, not on cytochrome c oxidase, and superoxide is formed. This superoxide destroys NO at diffusion limited kinetics, pulls down the NO level, disinhibits cytochrome c oxidase which then pulls down the O2 level allowing more O2 to diffuse to the mitochondria.

So, under conditions of FoF, the NO level is lowered. The more severe the FoF, the lower the NO level is taken. NO is a small uncharged molecule that diffuses readily through lipid membranes. The only barrier to NO in the body is crystalline bone. A state of low NO, is then propagated to all cells, so that the metabolic status of all cells can be regulated in sync. This is important because to maximize the ability to run from a bear, O2 and glucose consumption by non-essential systems must be curtailed as well as ATP consumption by muscle repair systems.

Is the hypothesis of ATP hierarchies plausible? Well, we know that physiology does behave this way. There is an effect called ischemic preconditioning, where a brief ischemic event induces a transient state where a prolonged ischemic event will produce less damage. This is well observed in a number of different organs. Transient ischemia reduces ATP demand and so cells can survive ischemia that would otherwise kill them. This behavior is what the ATP hierarchies hypothesis would predict. The mechanisms behind ischemic preconditioning are mostly unknown. No doubt as a stress response from deep evolutionary time there are many pathways involved in very complex and redundant ways, which may (is likely to) be different for different organs. Oxidative stress is known to be involved in some aspects of ischemic preconditioning.

Presumably ischemic preconditioning has some detrimental long term effects, otherwise cells would evolve to be in that state continuously. They don't, therefore there must be long term negative consequences. Those negative consequences might not show up for some time, but they must be present. This is a danger of short term endpoints in clinical trials. A treatment may prevent short term damage but if continued may cause increased long term damage.

This is one of the dangers of pain relief. If it merely masks the pain symptoms, and people then behave as if they are in the RnR state when they are actually still in the FoF state, then running themselves to death is much easier. Similarly, what do "stimulants" actually do? Do they increase the ability of cells to make ATP? Doubtful that a drug could improve on a few billion years of evolution. They do increase ATP availability (otherwise they wouldn't be stimulants), most likely by invoking the FoF state and turning off non-voluntary pathways like long term maintenance, but without the pain that normally warns of degraded repair systems.

So what happens under conditions of RnR? Well, to activate all the repair pathways, ATP needs to be high, so via sGC, NO levels have to be high too. What triggers mitochondria biogenesis is NO, so to make more mitochondria NO levels need to be high too. So RnR is a state of high NO.

How is this state of high NO produced? One mechanism is by a reduction in mitochondrial potential. To generate high ATP flux, mitochondria increase their potential to increase the driving force for ATP production. This does increase the rate, but it also increases superoxide production, a valuable feature, which pulls down the NO level to increase O2 diffusion. When the demand for ATP drops, the potential drops, the superoxide formation rate drops, the NO destruction rate drops, and the NO concentration rises provided there is sufficient basal NO production to begin with. If the basal NO production rate is too low, then the reduction in the NO destruction rate doesn't raise the NO level.

This presents a problem, if the state of FoF is prolonged sufficiently that mitochondria biogenesis suffers. The only reason that organisms have the ability to increase their metabolic activity over basal levels is because there are "excess" mitochondria. That is mitochondria in excess of the minimum necessary to supply basal ATP requirements.

Fewer mitochondria can supply the same ATP by increasing mitochondrial potential. This results in greater superoxide production, and also greater "slip", that is a reduction in the number of ATP molecules produced per mole of O2 reduced. A hallmark of many of degenerative diseases is weight loss, often inappropriately termed malnutrition, where the actual problem is increased basal metabolism. Elevated basal metabolism is observed in dilative cardiomyopathy, chronic renal failure, HIV infection, liver cirrhosis, chronic obstructive pulmonary disease, Does an increased basal metabolism mean the body is doing "more stuff"? Likely not, rather it is doing the same "basal metabolism stuff" but using ATP generated less efficiently with fewer mitochondria as observed in heart failure. It might even be doing less, because low ATP has turned off the repair pathways which is why the liver, kidneys, heart are failing in the first place. In HIV, a standard treatment is via highly active anti-retroviral therapy (HAART). A side effect of this treatment is reductions in mitochondria biogenesis. This can result in hyperlactatemia because of increased glycolysis to supply ATP. But if the liver and kidneys don't have sufficient mitochondria to recycle the lactate via the Cori cycle, where does it go? Perhaps into ectopic fat. I suspect that this is one of the problems of obesity. NO selectively partitions into lipid, and adipose tissue is a source of inflammation and oxidative stress. If NO drops sufficiently to impact mitochondria biogenesis, there may be no internal mechanism to raise it sufficiently for a long enough time to reverse the mitochondria depletion.

So how does all of this relate to the placebo effect? Well, if healing and cellular repair is accomplished most effectively during periods of RnR, then invoking that state will promote healing, well being, and long life. One of the things that does invoke feelings of rest and relaxation is love. It is well known that married people live longer lives (and it isn't just that it seems longer). The well known maternal "kiss it and make it better" treatment does relieve pain and presumably resets the RnR state. Presumably regular episodes of love and affection from a romantic partner can reset the RnR state too.

At the heart of energy metabolism is nitric oxide. A major determinant of whether an organism is in the FoF state, or the RnR state is the level of NO. Because NO is freely diffusible, and is created and destroyed at many sites in the body, the basal level has an impact on the signaling effects of NO. Low basal NO will affect every NO mediated signal with no threshold. This is an extremely important point. Anything that increases basal NO will shift physiology to the RnR state and away from the FoF state. There are many things that will do this, placebos are one of them. The relaxation response causes the production of NO. My own favorite method is via commensal ammonia oxidizing bacteria on the skin. No matter what the basal NO level is, physiology can always destroy that NO very rapidly with superoxide. Mitochondria have an essentially unlimited capacity to make superoxide, limited only by the supply of O2 and reducing equivalents. What ever the NO level is, mitochondria can pull it down to zero. This has important implications in acute respiratory distress syndrome, and is what is responsible for the multiple organ failure which sometimes occurs.

Long term meditation does result in reduced age-associated loss of cortical white matter. I presume by increased repair, improved energy status, reduced apoptosis, better clearing of damaged proteins, and perhaps increased axonogenesis. Many neurotrophic factors have effects mediated through NO.

Meditation modulates the immune system and increases antibody titers due to vaccination. Meditation reduces the symptoms of the metabolic syndrome and improves a number of heart health parameters.

If placebos increase NO levels and invoke the RnR state, then nocebos likely reduce NO levels and induce the FoF state. What conditions might be improved by the FoF state? In the earlier example, nausea was reduced by a nocebo. Much of the enteric nervous system is nitrergic that is the nerves generate NO. If the basal level of NO is reduced by a nocebo, then the response of the enteric nervous system to CNS generated nausea signals mediated by NO will be reduced by a nocebo and enhanced by a placebo. When running from a bear, it is a "feature" to delay vomiting.

When coaches try to motivate athletes, usually it is via negative and violent symbolism, not by restful and peaceful symbolism. Invoking FoF is good when going into combat, even the ritualized combat of athletic events. However, the FoF state has costs associated with forgone cell repair and maintenance. It is a state used when necessary, but not a state that can be sustained long term. It would therefore be desirable to have a mechanism to terminate the FoF state, and to invoke the RnR state. This is the "relaxation response". Young children haven't yet learned to invoke this state, so it can be invoked for them by a parent by the "kiss and make it better" treatment.

So how does this all relate to pain? In this context, pain is a signal from your body telling you that your ATP consumption is exceeding what physiology can provide without shutting important stuff down. Your body will let you run yourself to death, because escaping from a bear is more important than any other damage short of death.

Implications of the placebo effect being mediated by NO. Every disease and disorder that is characterized by low NO will be helped by increasing NO, and so will be helped by placebos. This is not an imagined improvement, but an actual improvement. ASDs are caused by low NO, so they are helped by placebos and made worse by nocebos. This is why bullying is particularly bad for people with ASDs. They already have low NO, so bullying which invokes the FoF state makes that worse. What ASDs need is love and affection. As do children, and as do adults. As does everyone.