The neurotoxicology of hard foraging and fat-melts

In this issue of PNAS, Maswood et al. (1) extend the already remarkable effects of caloric restriction (CR) to 1-methyl-4-phenyl-1,2,3,6, tetrahydropyridine (MPTP)-induced parkinsonism in rhesus monkeys. MPTP was identified in synthetic street heroin as a cause of rapid-onset parkinsonism (2). This lipophilic protoxin rapidly crosses the blood–brain barrier and is converted in astrocytes by monoamine oxidase B to 1-methyl-4-phenylpyridinium (MPP⁺) (3), which selectively damages dopaminergic neurons.

CR in the range of 10–40% with balanced micronutrients commensurately increases the lifespan of many short-lived species (4, 5, 6). The basis for the robust effects of CR on longevity, while far from clear, have been associated with the attenuation of many aging processes (e.g., accumulation of oxidation products in nucleic acids, proteins, and lipids) and chronic diseases (e.g., tumors and renal neurodegeneration). Blood glucose and insulin are lowered, but glucocorticoids are elevated, which is paradoxic for neuroprotective effects (7). Ongoing studies with rhesus monkeys show benefits to health particularly in the middle-aged, which tend to suffer from obesity; however, effects on longevity remain to be proven (4, 8). Recently, shorter-term CR has been applied to models of neurological disease and, during a span of months, has positive effects on Alzheimer's disease transgenics (9) and in the present Parkinson's disease model.

The protocol used in these studies of adult rhesus monkeys reduced the food intake by ≈23% for 6 months (10) before injection with MPTP, with 2-fold benefits to dopamine loss and movement disorders. Notably, these improvements did not require severe weight loss: Total body weight was modestly decreased (–12%); effects on visceral fat were probably greater (4). These findings agree with rodent models in which a more severe CR attenuated the neurotoxicity of MPTP (11) and of excitotoxins (kainate and ibotenate), which cause broader neurodegeneration (12). However, another group (13) did not find neuroprotection by CR against MPTP in the same genotype at a slightly later age. It also was found that CR did not benefit a model of amyotrophic lateral sclerosis (14).

Many mechanisms may be at work in protection against MPTP by CR, as discussed by Maswood et al. (1). Oxidative stress, which can synergize with dopaminergic neurotoxicity, is decreased in the brain, as in other tissues, possibly through induction of antioxidant enzymes. Neurotoxicity also may be decreased by the induction of cytoprotective “heat shock” chaperones. The authors favor the role of BDNF and glial cell line-derived neurotrophic factor (GDNF), which were induced by CR in these brains. Both growth factors show neuroprotection in Parkinson models; GDNF is now in clinical trials for Parkinson's disease. I discuss several aspects of these interesting findings: the role of locomotor activity (exercise) in these effects of CR; an evolutionary perspective on the adaptive role of CR in cytotoxicity to xenobiotics, for which the effects of CR on MPTP neurotoxicity may be a model; and possible hazards of CR in liberating lipophilic organochlorine toxins.

Are neuroprotective effects of caloric restriction due to induction of growth factors by increased motor activity?

In rodent models, “voluntary exercise” through access to running wheels increases brain BDNF and neuroprotection against MPTP during ad libitum feeding (15, 16). However, CR itself also rapidly induces voluntary running in rodents (17, 18). Human anorectics also show increased activity, although other mechanisms may be at work (19). Could the neuroprotective effects of CR be due to the induction of BDNF and other growth factors by increased motor activity? Because the present 23% CR induced much smaller increases of activity in rhesus monkeys than did 10% CR in rats (17), the threshold for activity induction of BDNF may be lower in monkeys than in rats. Clearly, large animals are more resistant to starvation than mice are (8). The elevation of glucocorticoids by CR raises another question: Because BDNF expression in another brain region (the hippocampus) is repressed by elevated corticosterone or behavioral stress (20), would the neuroprotection by CR be greater if the glucocorticoid elevations of CR (9) were suppressed?

From an evolutionary perspective, hunger may increase locomotor activity as an adaptive foraging behavior (18, 21). A specific case of optimum foraging theory is that hunger stimulates locomotion to search more widely for food sources. In herbivores, diet selection is optimized to maximize nutrients while minimizing the ingestion of plant toxins (22). But hunger-driven searching for nutrients and consumption of atypical foods implies increased hazards of toxicity from xenobiotics (5). Perhaps that is why CR and starvation induce detoxifying enzymes such as the P450 cytochromes (23, 24, 25), which should accelerate toxin clearance as well as induce the cytoprotective heat shock proteins noted above. Returning to the Parkinson model, it is cogent that MPTP is metabolized by the P450 enzyme CYP2D6, which is found in nigral neurons (26) and, if overexpressed, can decrease MPP⁺ cytotoxicity (27). Future studies could address the effects of CR on toxin clearance in the brain by induction of CYP2D6 and other enzymes. These findings on reduced MPTP neurotoxicity bring further depth to the hypothesis that CR is an evolved adaptation to food scarcity that enables reproduction to be postponed by slowing aging, pending better times (5, 8, 28).

Last, I note the potential consequences of neurotoxins stored in fat and caused by dieting in human life-styles. Fat depots universally accumulate manmade organochlorines, such as the polychlorinated biphenyls (PCBs). Brains of patients with Parkinson's disease may also accumulate PCBs (29). This is important because sublethal PCB levels have dopaminergic neurotoxicity in animal models (e.g., ref. 30). In view of the increasing popularity of fat-decreasing diets, it is sobering to consider that blood PCB levels increase during chronic weight loss (31, 32). It seems timely to compare the incidence of parkinsonism in those who strive to reduce what they consider to be excessive body fat by means of CR or yo-yo diets and in those who have long maintained a low fat intake and body mass index, which appear to lower the risk of parkinsonism (33).

Genetics may also figure here, because carriers of the CYP2D6 allele associated with “poor catabolism” of organochlorines show a higher prevalence of Parkinson's disease (34, 35). More generally, the development of strategies to optimize diet and body composition for long-term health of blood vessels and the brain may be guided by many loci in an individual's genotype.