This is a follow-on from my previous post, where I laid out the beginnings of my ideas for influencing food choice. To summarise, I’ll highlight a few sentences from that post: “…it appears that AgRP neurones are a fundamental link between sensory detection of food, hunger, and the learned seeking of high caloric foods. More specifically, the drop in AgRP neurones activity upon sensory detection of a food seems to be the determinant of how much of that food the animal wants to eat. Now, what if we could (briefly) activate AgRP neurones during consumption of an unhealthy meal? I say activate, but it could equally mean limit the inhibition […] Over time, with repeated exposures to the same high calorie meal and activation of the AgRP neurones, you would drive a preference away from that unhealthy meal in the future.”
This is still conjecture on my part, so if you have any reason to suggest my conclusions may be faulty, please do leave a message to explain where my reasoning falls short. Anyway, in order to further my hypothesis, I had suggested an experiment where we provide a mouse with continuous access to two foods of differing reward value (eg. normal chow and high fat/high sugar diet, what we usually term high energy diet or HED). The mouse would have its AgRP neurones optogenetically activated every time it eats some of the high reward HED, thus limiting the drop in AgRP neurone activity that is the teaching signal for how rewarding that food is.
However, upon further musing, I don’t think we should do this with our usual ChR2 stimulation paradigm, as this would drive action potentials upon stimulation – we are interested in limiting the drop rather than driving further hunger. What we want instead is mild depolarisation of the membrane. In that case, it might be best to use a stabilised step-function opsin (SSFO), because they have a longer milder activation; however, their typical on-time is about 30 mins, which might be too long for our purposes.
It is important to consider the time-course of any neuronal control. Ideally, we should make sure that our manipulations match the normal time-course for a response. We could either do this crudely, where we allow hungry mice access to just HED for 30 mins, and trigger a single SSFO activation when they start eating, or we could do a more advanced (and challenging, and hopefully more informative) paradigm, where we allow access to chow and HED and only trigger depolarisation for the expected time of response for each bite of HED. I will keep SSFO activation during a purely HED meal as a plan B, but I would prefer do allow a fair choice between foods, as I think this allows us to capitalise on the instantaneous nature of optogenetics, and should allow the mice to redirect their appetite in real time.
To have any hope of influencing food choice, we need to drive an appropriate AgRP neurone response every time they consume HED. And to do this, there are two parameters we need to determine experimentally:
- The expected AgRP neurone response to each consumption of HED, which we can check using GCaMP photometry in vivo. The important aspect to quantify is the timecourse of the response to a predetermined meal. Probably the easiest way is to use “bitesize” pellets (such as these sucrose pellets that I’ve used in the past).
- An appropriate optogenetic stimulation paradigm to elicit a mild depolarisation without triggering action potentials. I could do this quite easily using patch clamping of the opsin-transfected AgRP neurones and test a range of stimulation paradigms. I think likely a long (4-8 seconds) stim at a lower than usual brightness would be likely to produce the outcome I want.
Once I have determined these two parameters, it will be time to perform the experiment, which should be relatively straightforward as opto studies go:
- Implant optic fibre into mice expressing ChR2 in AgRP neurones
- Allow continuous access to chow and sucrose pellets throughout a long (6 hours?) stimulation session, with no food available at other times
- Track consumption of each food, and trigger continuous “light-on” for the experimentally derived response time after consumption of each sucrose pellet
- Perform repeated stimulation sessions and see if preference shifts away from sucrose
Assuming this shows the outcome I predict, the next step would be to investigate this effect using pharmacology. I would give mice set “meals” with normal chow or HED, and try to shift their perception of the reward using a low dose of AgRP neurone modulators eg. PYY antagonist. Here’s a possible plan:
- Mice given calorie restricted meals for chow and HED separated by a few hours, then a calorically unlimited meal of both for an hour in the evening.
- Mice given injection of low dose PYY at onset of chow meal, and PYY antagonist at onset of HED meal
- Track food intake and body weights
The question here would be, does the treatment shift preference away from the unhealthy HED? I’ll be honest and say that I really have no idea whether these experiments would pan out the way I envisage. But if they do succeed in influencing food choice, I think it could pave the way for some very interesting (and completely novel) therapeutics to combat obesity.