So hungry, so sad: Calorie restriction and depression

Mood is intimately connected to feeding behavior, which in turn impacts mood. Given that a fairly radical departure from normal feeding — calorie restriction (CR) — is one of the most widely discussed and promising means of extending lifespan in a wide range of organisms, it seems reasonable to ask whether CR significantly impacts psychology. (It has already been reported that CR in humans can result in behavioral changes including hoarding.)

Chandler-Laney et al. investigated the neurochemistry and behavior of rats subjected to a regimen of periodic calorie restriction, and conclude that the treatment has a deleterious effect on the animals’ mood:

A history of caloric restriction induces neurochemical and behavioral changes in rats consistent with models of depression

A history of dieting is common in individuals suffering from eating disorders for which depression and mood disturbances are also comorbid. We investigated the effect of a history of caloric restriction (HCR) in rats that involved cyclic food restriction and refeeding with varying levels of access to palatable food (PF) on: 1) responses to the SSRI, fluoxetine; 2) monoamine levels in brain regions central to the control of feeding, reward, and mood regulation; and 3) behavioral tests of anxiety and depression. HCR coupled with intermittent but not daily access to PF exaggerated rats’ anorectic response to fluoxetine (p < 0.05); was associated with a significant 71% and 58% reduction of 5-HT and dopamine, respectively, in the medial prefrontal cortex; and induced behaviors consistent with models of depression. HCR, irrespective of access to PF, abolished the strong association between 5-HT and dopamine turnover in the nucleus accumbens in control rats (r = 0.71 vs. – 0.06, p < 0.01). Access to PF, irrespective of HCR, reduced hypothalamic dopamine. Together, these findings suggest that a history of frequent food restriction-induced weight fluctuation imposes neurochemical changes that negatively impact feeding and mood regulation.

Taken at face value, this would seem to have dire ramifications for humans considering undergoing CR for the purposes of health or lifespan extension, inasmuch as we would like to not only increase the extent of life but also its quality. In its most severe forms, depression is a crippling (even fatal) illness. Even if the average effect is small on the hypothetical median human, this same slight change could serve as a triggering event in people already prone to depression. At the very least, a history of depression would be a significant counter-indication for someone considering CR.

But they don’t pay us the big bucks to take the results of studies at face value. A question that must be answered, before we begin extrapolating from rats to humans, is: Does this study employ a fair model of CR? The authors were informed by an interest in pathological food restriction (i.e., eating disorders), suggesting that achieving a good fit to life-extension CR wasn’t high on their list. Furthermore, the feeding pattern in this study is cyclical, rather than constant. It seems possible that the frequent changes in food availability, rather than the restriction of supply per se, are what’s bothering the subject animals.

So what we’d like to see is a study in which food availability is chronically restricted, without temporal variation. Fortunately (for us, not the rats), another group has done just that. In Jahng et al., the authors subjected young rats to CR, and found 5-HT reductions and behavioral alterations that are consistent with the paper discussed above:

Chronic food restriction in young rats results in depression- and anxiety-like behaviors with decreased expression of serotonin reuptake transporter

Evidence of semi-starvation is commonly found in patients with eating disorders. This study was conducted to examine the adverse effects of chronic caloric restriction in young rats, since there have been increasing incidence of eating disorders especially among young populations. Food restriction group was supplied daily with 50% of chow consumed by its ad libitum fed control group from postnatal day 28. After 5 weeks of food restriction, brain contents of serotonin (5-hydroxy-tryptamine; 5-HT) and its metabolite 5-hydroxyindol acetic acid were analyzed by high-performance liquid chromatography and mRNA expression of 5-HT reuptake transporter (5-HTT) by in situ hybridization. Plasma corticosterone levels were determined by radioimmunoassay. Behavioral assessments were performed with Porsolt swim test for depressive behavior and with elevated plus maze test for anxiety. Five weeks of food restriction markedly increased plasma level of corticosterone, and significantly decreased 5-HT turnover rates in the hippocampus and the hypothalamus. 5-HTT mRNA expression decreased in the raphe nucleus of food restricted rats compared with free fed controls. Immobility time during the swim test increased in the food restricted group, compared to the control group. Food restricted rats spent more time in the closed arms, less time in the open arms, of elevated plus maze compared with control rats. These results suggest that chronic caloric restriction in young rats may lead to the development of depressive and/or anxiety disorders, likely, in relation with dysfunction of brain 5-HT system.

Once again, these authors are approaching the question from the perspective of eating disorders rather than life extension, but I’m not sure that matters: The body doesn’t know why it’s not getting a certain amount of food, just that it isn’t. On the other hand, the CR regimen in this study is very stringent (50% ad libitum, as opposed to the ~60% used in most life extension studies), so it’s possible that these rats are outside the critical CR zone and into the realm of frank starvation and malnutrition. Still: there’s no cycling, and the results are the same as for Chandler-Laney et al., so I’m comfortable with the tentative conclusion that the cycling wasn’t the problem.

Taking both papers together, along with the cautions about their applicability to CR as biogerontologists understand it, my overall impression is that I really wish someone would couple a long-term CR experiment (in which life extension and other CR benefits, such as delayed cancer and diabetes, are observed) with behavioral and neurochemical studies of this kind. The CR field has been good about measuring not just quantity of life but also quality — but thus far, “quality of life” has been limited to readily quantified metrics of health, rather than evidence of psychological well-being (or its lack).

One of the reasons that I’m prone to believe the results, as well as their extrapolated interpretation, is that a connection between CR and depression makes a certain amount of sense from an evolutionary standpoint: If an organism believes (because of feeding history and other cues such as smells) that the environment is currently stressful and hostile — empty of food, but full of predators trying to eat it — one adaptive strategy would be to just stay in the burrow, doing nothing much and waiting it out. If the mechanism by which this occurs is psychological depression, too bad. No one said that Darwinian fitness had to be fun. Depression, clearly maladaptive when chronic and unrelated to stimuli, may well have adaptive benefits under certain circumstances.

On a final note: Assuming for the moment that some degree of depression is a consequence of even healthy well-managed CR, it’s possible that efficacious calorie restriction mimetic drugs might also trigger similar mood problems. As such pharmaceuticals become available (and that will be within a few years), determining where in the pathway a drug target acts, relative to the signaling events that result in psychological phenotypes, will be critical.