Can Mice Show Placebo Effects? And Does It Even Matter?

Chances are, if you follow science news, that you already have - or will soon - come across a headline along the lines of "Scientists prove acupuncture works" or the likes. Or if you - like I do - tend to prefer quality blogs as sources for your science news, you might have instead come across posts lamenting the questionable quality of a recent Nature Neuroscience publication purporting to have found the biological mechanism of how acupuncture works.

I responded yesterday to some of these posts, but emails and comments  that I have received in response have moved me to edit some of this content. In particular I want to  point out two things, before going into the details of the study:

1. For one, the study is a study on mice. It is - like any single study - incomplete and it certainly does NOT prove that acupuncture can do anything for your health that goes beyond a mere Placebo effect (in particular if you are not a laboratory mouse yourself!). In particular, the study does not include a Placebo control trial, which many of my favorite science writers have criticized harshly. In consequence, any interpretation of results as proving the efficacy of acupuncture should be viewed as inherently overblown; immensely so if we take acupuncture to include the esoteric notions of cosmic energy flows, Qi and Meridians etc.
2. On the other hand, the study DOES show a clear biochemical link between sticking needles into inflamed mouse tissue and the release of a certain biochemical called adenosine. The study builds on this finding by highlighting the pain-reducing effect of this chemical. In my opinion, saying that sticking needles into tissue causes release of a pain killer is not the same as saying that the entire esoteric belief system around acupuncture is validated, but I do think that this does suggest some possible mechanism for why people who receive acupuncture report pain relief; even if this relief can be achieved equally well through Placebo treatment.

For the rest of this edited post I remain most of the original text, in which I raise the question of whether the lack of a Placebo control, which seems to be at the center of much of the criticism surrounding this paper, is really that relevant to the research that was conducted. You decide.

The main question that triggered me to write about this study - in contrast to my usual random content - was the question of whether you specifically need a Placebo control for animal trials? In my head at least, the Placebo effect has to do with expectations (that's what makes Placebo controls different from other simple co-variates for your regression analysis) and it’s hard for me to phantom that laboratory mice (except for these two guys) would have any mental concept of acupuncture that could bias results.

Googleing the issue of Placebo effects in animals didn’t really help a lot, so I read the original Nature Neuroscience piece instead to make up my own mind about how crude a mistake the absence of a Placebo trial for this study really was.

Up to this point, I was totally expecting a botched study, but as it turns out I think the study is very well done, and I do not think it absurd that it was accepted by Nature Neuroscience editors. Here’s my summary on the study:

In brief, I think it’s fair to say that the study is essentially about the role of Adenosine A1 Receptors in mediating the pain-relieving effects of sticking needles into inflamed tissue (whether you want to call this acupuncture or not, is up to you).

In the crudest terms what the 16-person strong research team around Nanna Goldman shows is that when you stick a needle into a hurting mouse close to where the pain is, adonesine levels increase in the mouse’s interstitial fluids. Additionally, they show that increasing adenosine into hurting mice reduces pain. But only for those mice who actually have adeonsine A1 receptors.

In a little more detail the study goes like this:

The researchers start their exposition on the (obviously difficult) assumption that there is indeed something real going on when people receive acupuncture, and therefore the question they try to address is 'what is the biological basis for this?'.

For their first experiment the research team stuck needles into (I assume, healthy) mice and measured the amount of adenosine that was released into the mice’ extracellular cavities as a consequence. The graph below makes the effect very clear, as adenosine concentration appears to increase 24-fold during the 30 minutes of needle puncturing, and then returns to baseline in the next 30 minutes thereafter (around the 60 minute mark).

It is well worth noting that the authors are very explicit that there are other ways to increase extracelleluar adenosine besides acupuncture:

“Notably, previous studies have shown that deep brain stimulation is also associated with a severalfold increase in extracellular ATP and adenosine. Similar to electroacupuncture and transcutaneous electrical nerve stimulation, deep brain stimulation delivers electrical stimulation that triggers an increase in extracellular adenosine concentration”. 

So it is very well possible that the Sham Treatments that everybody seems to be missing in this study, might do the same thing. Who knows? I don't find it that relevant for what follows:

Having shown that sticking needles into mice causes increases in adenosine, the researchers investigated whether increased adenosine levels are actually related to pain reduction in any relevant way.

The first step in doing so involved evoking inflammatory pain in a group of mice by injecting them with some nasty agent.

Mice thus treated showed typical signs of inflammation which included reduced thresholds for mechanical and thermal pain (mice with inflammations in their right paw will retract this paw at lower intensities of touch and heat than mice without inflammation, or than they do for their non-inflamed paw). Next, each mouse’s inflamed paw was injected with what is basically adenosine (2-chloro-N(6)-cyclopentyladenosine (CCPA) to be exact), and it was then determined whether pain thresholds were increased (i.e. whether there was any measurable pain reduction in the mice).

The control here was provided by including into the trial a particular subset of mice that had been genetically manipulated to lack adenosine receptors. If the adenosine injections are directly related to pain reductions, then mice without the receptor should not show decreased pain, and this is exactly the result that was obtained:

Regular mice injected with adenosine became more tolerant to touch and heat, while mice lacking the adenosine A1 receptor maintained their reduced low thresholds to even innocuous mechanical and thermal stimulation.

Somewhat importantly here, adenosine was always injected into the hurting paw, and these injections did not influence behavior regarding healthy paws.

I consider this important as it comes to play in the next experiment: Here, mice were again caused pain (this time through Sciatic nerve ligation), and again CCPA was injected. However, this time CCPA was injected into the hurting leg for one group of mice and into the contralateral leg for another group. As you can expect, pain reduction only occurred for those who received adenosine to the right spot (i.e. into the ipsilateral leg).

Even more importantly

“substituting CCPA injection in the ipsilateral leg with an equal volume of saline failed to change the threshold to either thermal or mechanically induced pain” 

To further understand whether CCPA reduced pain via directly acting on local nerve tracks, the researchers went on to measure responses in the mice' anterior cingular cortex (ACC), while they received painful stimulation. Since generally the ACC responds to painful stimuli, the task was to find out whether CCPA reduced the ACC response to pain.

Again, one group of poor mice was provided CCPA into their hurting, right leg, while another group of mice received CCPA into their healthy left leg. Pain response measurements in the ACC clearly showed that only those mice who received CCPA into the hurting, right leg showed marked pain decrease over the 60 minute time span following the injection. This seems to rule out the idea that adenosine acts generally, but suggests that the effect is of local nature.

Hence the authors hypothesize that

“CCPA acts locally, probably on unmyelinated C fibers in the superficial peroneal nerve, which travels in close proximity to the Zusanli points” 

 and cite previous research that corroborates this idea.

Also importantly, mice lacking the adenosine A1 receptor again failed to show any pain reduction in response to CCPA even for this part of the study, which allows for the more specific hypothesis that

“CCPA reduced painful stimulation by activating adenosine A1 receptors on unmyelinated C fibers, and possibly A∂ fibers, in the superficial peroneal nerve”

As you see, at this point it’s mostly biochemistry and not too much about acupuncture. However, the authors returned to the questionable medical practice in another experiment in which they tried to replicate the results they had just achieved by injecting adenosine , but this time via the use of needles alone [as Vaughan pointed out in a comment to my original post, the notion of "acupuncture" is difficult, or even "bizarre" in the context of mice, but for what it's worth: the researchers put away the chemicals and tried to replicate the result by sticking needles where it hurts].

To make it short from here on: Exactly as in the CCPA injection experiments, use of acupuncture needles appeared to reduce pain only in those mice which a) received  treatment to their hurting foot, and b) only when treatment was performed “correctly” (i.e. it wasn’t enough to just stick needles to the right spot, the needles had to be turned after insertion, as per usual acupuncture practice), and c) treatment only worked for those mice who were not genetically robbed of their adenosine A1 receptors.

All of this makes it very reasonable in my opinion to assume – as the study’s authors seem to do – that it is indeed increased adenosine release following needle puncturing that lead to the observed pain reduction. Since acupuncture also involves puncturing tissue with needles, this seems a plausible mechanism for why some people report pain reduction after such treatment...

The final part of the paper shows how additional chemical agents which prolong the time that adenosine remains in extracellular space cavities also appear to prolong the pain relieving effect that adenosine exhibits.

In their final discussion the authors again mention that

“Notably, needle penetration has been reported to not confer an analgesic advantage over nonpenetrating (placebo) needle application, as opposed to our observations and those of others. However, it is possible that ATP release from keratinocytes in response to mechanical stimulation of the skin results in an accumulation of adenosine that transiently reduces pain, as A1 receptors are probably expressed by nociceptive axon terminal in epidermis. In fact, vibratory stimulation applied to the skin depressed the activity of nociceptive neurons in the lower lumbar segments of cats by release of adenosine.” [citations omitted].

In retrospect I am still in search of a good answer to my original question of whether animals can exhibit true placebo effects, but additionally I am also confused about why reactions to this study have been so hostile. It is certainly not terrible science as some have suggested, and the fact that mainstream media is exaggerating the shit out of  this study doesn't mean it is of itself bad either.
Of course there is the issue of a potential conflict of interest that Ed was correct to point out, but even this does not immediately mean that the methods and results used are rubbish. Are they somewhat overstated? Possibly. But not more than 99% of other studies out there (where p-values of 0.049 are grand truths while the p=0.051's are barely reported).
As far as previous research is concerned, I have no problem in drawing a loose connection between the results of this study and the practice of acupuncture. I also have no qualms about accepting this study side-by-side with others that show no significant difference between acupuncture and  other alternatives (such as physiotherapy) or even Placebo. Other mechanical treatments may tap  into the same chemical mechanism, and finding that a treatment performs like Placebo does not necessarily imply a null effect either (If the effects are non-additive, it simply means that the effect is indistinguishable in size to the Placebo effect, which is after all a true effect).

Feel free to leave your comments below, or tweet me your thoughts.

[Update: Here is the most comprehensive discussion of the paper I've seen so far on the web].
[Note: Reader comments to this article were to the original (unedited) post and may be somewhat out of context]

Main Reference:

Nanna Goldman et al. (2010). Adenosine A1 receptors mediate local anti-nociceptive effects of acupuncture Nature Neuroscience : 10.1038/nn.2562