The present study examined the intersection of both aging and diet as risk factors for AD with both male and female wildtype and ApoE ɛ4 knock-in rats after four months on HFD/HSD or “Western” diet. At four months of age, all rats were put on the HFD/HSD diet ad libitum. Four months later a period equivalent to ca 8.7 human years 30 they were tested for cognitive behaviors and imaged for differences in resting state BOLD functional connectivity (rsFC). To our surprise male and female ApoE ɛ4 carriers were unaffected by the HFD/HSD diet as compared to wild-type rats; instead, it was the male wild-type, but not the female that showed behavioral and neuroradiological changes. These finding are discussed below with respect to preclinical and clinical studies on the behavior and neurobiology of ApoE ɛ4 carriers and the influence of diet.
Behavior
Eight-month-old male wildtype rats on HFD/HSD showed a modest deficit in cognitive function when evaluated in the Barnes maze while the male and female ApoE ɛ4 carriers and female wildtype showed no deficits. In a previous study using the same genetic rat model we reported 6-month-old human ApoE ɛ4 knock-in rats feed a normal diet were more likely to have cognitive and behavioral problems as compared to female carriers early in life 24. In the same rat model, we reported 8-month-old female carriers on normal diet show enhanced microvascular density in the hippocampus hypothesized to be a neuroadaptive response helping to spare cognitive function 23. These and other studies on rodent models of ApoE ɛ4 show genotype-specific changes in behavior and brain structure characteristics of early neurodegenerative pathology 22, 31–34. Given the risk associated with HFD/HSD for cardiovascular disease, Alzheimer’s disease, and early dementia, we hypothesized ApoE ɛ4 male and female rats would show exacerbated signs of cognitive dysfunction. Instead, ApoE ɛ4 male and female rats were spared the anticipated negative effects of the HFD/HSD, while wildtype males, but not females showed the most changes. Was the HFD/HSD protective in these 8-month-old ApoE ɛ4 rats, evidence of antagonistic pleiotropy, i.e., genes that provide resilience for developmental maturation and reproduction but are detrimental in old age35? Indeed, Janssen et al., reported female ApoE ɛ4 mice feed HFD/HSD showed no deficits in spatial learning together with a reduction in neuroinflammation in the hippocampus as compared to wildtype controls on HFD/HSD 36. In a recent review on feeding behavior and metabolism, Alcantara and colleagues argue that evolution has driven a neurobiology that favors the overconsumption of calories during times of plenty to offset the dearth of calories during times of famine29. ApoE ɛ4’s role in lipid metabolism, transport of cholesterol and complex lipids needed in neural development, membrane maintenance, and repair would be essential in handling the lipid load in HFD/HSD 11. Jones et al., studied the behavioral and metabolic changes in male and female human ApoE ɛ3 and ApoE ɛ4 knock-in mice fed HFD/HSD as compared to low fat diet and reported male ApoE ɛ4 mice were more susceptible to metabolic disturbances but not female ApoE ɛ3 and ApoE ɛ4 mice 21. This is in contrast to our data that showed male ApoE ɛ4 were unaffected by the HFD/HSD. The metabolic sensitivity of the male ApoE ɛ4 mice to HFD/HSD but not female ApoE ɛ4 mice was corroborated by Matter et al., with the added distinction that females showed impairment in cognitive function 20. In a subsequent study, Jones, and coworkers feed ApoE ɛ3 and ApoE ɛ4 mice HFD/HSD for 12 weeks and found the ApoE ɛ3 genotype but not ApoE ɛ4 presented with significant neuroinflammation in the hippocampus as measured by activated microglia and astrocytes 37. Interestingly, they argued the neuroinflammation was neuroprotective in the ApoE ɛ3 mice creating a neuroadaptive response to the HFD/HSD that would be beneficial in the future. Unfortunately, in our studies on ApoE ɛ4 rats we did not do postmortem histology looking for evidence of gliosis and neuroinflammation.
Functional Connectivity - Neural circuitry of feeding and metabolism
Male wildtype showed whole brain hyperconnectivity as compared to the male carrier while the female ApoE ɛ4 showed greater connectivity as compared to the female wildtype. The hyperconnectivity for the male wildtype on HFD/HSD was unexpected. However, in a previous study we reported female ApoE ɛ4 rats at six months of age show hyperconnectivity in amygdala, paraventricular nucleus of the hypothalamus and ascending reticular activating system 24. In this study the male wildtype and female carrier both showed greater global hippocampal connectivity as compared to the male ApoE ɛ4 carrier. As noted above this connectivity was not reflected in any change in cognitive function for females but may be associated with the deficit in learning for the male wildtype. In the latter case, the hyperconnectivity may reflect neuroplastic compensation in response to the early stages of brain injury and loss of function 38, 39
The sensitivity of the male wildtype to the HFD/HSD as measured by alterations in cognitive function and functional connectivity suggested the risk factor was not ApoE ɛ4, but instead, a metabolic disorder. While body weights were only recorded through the first 8–9 weeks of HFD/HSD there were no significant changes between wildtype and carriers for either sex (Supplementary Fig S1). So obesity was not a contributing factor. Metabolism involves a balance in appetite and energy utilization. In a recent review conducted by Alcantara et al., a model was proposed that describes the interconnected neural circuitry responsible for regulating the different phases of feeding, including appetitive, consummatory, and terminating phases. 29. The circuits that control feeding are complex and involve multiple regions of the brain, most notably the hypothalamus. Within the hypothalamus, various nuclei host distinct populations of neurons that have significant roles in controlling appetite and energy balance40–42. These include neurons that respond to or produce orexigenic peptides such as ghrelin 41, and neuropeptide Y 43, and those that respond to or produce anorexigenic signaling molecules e.g. leptin 44 and melanocortin 45. More specifically, the arcuate nucleus is home to a specific group of neurons known as agouti-related peptide (agRP) neurons in the arcuate nucleus that play a central role 46 in regulating all aspects of feeding in coordination with other hypothalamic areas 29, 47.
When the many brain areas controlling feeding (Table 1) are combined as a single hub it was the wild-type males on HFD/HSD that showed a significant global increase in connectivity exceeding all other experimental conditions. The global connectivity of the hypothalamus as a single node is also extensive in wildtype males and significantly greater than the other experimental conditions. However, the global connectivity of the three key hypothalamic areas i.e., acuate, lateral, and paraventricular nuclei, noted for their role in controlling the distinct phases of feeding present with only a modest number of connections and are not significantly different between experimental groups. A seminal paper by Low and colleagues identified neurons in the cerebellar nuclei, which are characterized by their unique molecular and topographical properties, were found to be activated in response to feeding and nutrient infusion resulting a decrease in food intake 48. Activation of the deep cerebellar nuclei terminates feeding by altering dopaminergic signaling to the striatum. Our study looking at the interaction between genes and diet spotlighted these cerebellar nuclei e.g. lateral (dentate) intermediate (interposed) and medial (fastigial) and analyzed their connectivity. Indeed these cerebellar nuclei had an extensive network of connections far exceeding the hypothalamic nuclei but only in the male wildtype rats. How this hyperconnectivity is affecting feeding and metabolism in male wildtype is unknown.
Limitations
The study was not designed to follow feeding and metabolism, points of interest only raised after analyzing the connectivity data. Circadian measures of feeding, metabolism, blood and urine chemistry for each experimental group would have enhanced our understanding of the effect of HFD/HSD on wild type and ApoE ɛ4 carriers and why this genetic rat model is resilient - an example of antagonistic pleiotropy? Equally important would have been the same four experimental groups but maintained on normal rat chow. Unfortunately the study had only a limited number of ApoE ɛ4 rats preventing the study of controls on a normal diet. Hence all comparisons are with respect to HFD/HSD. What would be typical behavior or connectivity is uncertain. As noted above, there was no post-mortem histology to confirm the presence or absence of neuroinflammation as reported by Jones et al., 21. Still another limitation was the collection of functional connectivity data under light isoflurane anesthesia to minimize motion and physiological stress 49. What constitutes a “resting state” condition in a head restrained, awake rodent could be debated. The administration of anesthesia can decrease the strength of the BOLD signal; however, it does not disrupt the connectivity between brain regions. This has been observed in various species and under different physiological conditions. 50–54.
Summary
The present study examined the effect of high fat, high sugar diet on male and female, wildtype and ApoE ɛ4 knock-in rats with the expectation that carriers would present with deficits in cognition and functional connectivity. The results were unexpected. The genetic risk was overshadowed by the diet. Male wildtype rats were most sensitive to the HFD/HSD presenting with a deficit in cognitive performance and enhanced functional connectivity in neural circuitry associated with food consumption and metabolism. The deep cerebellar nuclei key in the regulation of feeding behavior showed hyperconnectivity in male wildtype but not in female wildtype or female and male ApoE ɛ4 rats.