Both the increased availability of energy-dense, highly palatable foods and dieting are believed to contribute to the development of obesity and certain forms of eating disorders [1
]. Highly palatable food drives overeating often followed by dieting, which is culturally driven by norms for thinness or health [3
]. Similar to what is observed in drug addiction, abstaining from highly palatable foods is hypothesized to contribute to the emergence of a negative emotional state, which in turn sustains cravings and promotes relapse, resulting in a vicious circle of palatable food withdrawal and compulsive eating [4
Individuals affected by forms of eating disorders and obesity have been shown to display impairments in multiple domains of cognition, including hippocampal-dependent memory function [5
]. The hippocampus is an important brain area that plays a key role not only in contextual learning and memory processes [8
], but also in feeding behavior [9
], as well as affective and addiction disorders [10
]. One prominent feature of hippocampal plasticity is adult neurogenesis, the ability of the dentate gyrus subgranular zone (SGZ) to give rise to new neurons throughout life [12
]. Accumulating evidence suggests that neurogenesis plays a pivotal role in affective as well as addictive disorders [10
A series of experiments was performed to evaluate the effects of alternation of palatable food withdrawal and refeeding on memory function and hippocampal neurogenesis. The first aim of this study was to assess whether withdrawal from chronic, intermittent access to a highly palatable diet impairs memory function. For this purpose, we used a battery of behavioral tasks not involving appetitive reinforcers or food restriction/deprivation to evaluate different aspects of memory function following an established rodent model of compulsive eating induced by palatable food cycling [15
]. The second aim of the study was to assess whether withdrawal from chronic, intermittent access to a highly palatable diet impairs key neurogenic processes, such as the number of proliferating cells and the number of newborn neurons in the dentate gyrus SGZ of the hippocampus. Furthermore, the third aim of this study was to evaluate the effects of the uncompetitive N
-aspartate receptor (NMDAR) uncompetitive antagonist memantine, a pro-cognitive drug used for the treatment of Alzheimer’s disease, on palatable food withdrawal-induced memory dysfunction. Lastly, the fourth aim of this study was to evaluate the effects of memantine on food intake during palatable food withdrawal and refeeding.
The major findings of the present study were the following: (1) withdrawal from intermittent access to palatable food is accompanied by hippocampus-dependent memory function impairment; (2) deficit in memory function is associated with disruption of dentate gyrus neurogenesis in rats exposed to intermittent access to palatable food; (3) the uncompetitive NMDAR antagonist, memantine, fully rescues the memory function impairment in rats withdrawn from intermittent access to palatable food; (4) memantine fully blocks excessive intake of palatable food and ameliorates hypophagia of the standard diet in cycled rats.
In this study, we found that withdrawal from a highly palatable diet was responsible for memory function impairment. Notably, the deficit in memory function observed in palatable food-withdrawn rats was selective for hippocampus-dependent memory function, as revealed by both the inability of cycled rats to recognize the novel location of a familiar object, and the preference for a stimulus-response at the expenses of spatial learning. Indeed, an intact hippocampus is required in the mnemonic functions that allow the recognition of a novel position of a specific object [36
], as well as the recognition of a novel contextual environment over a novel cue [27
]. Conversely, palatable food-withdrawn rats, when tested in an NOR task, showed an intact recognition memory, as shown by an increased spontaneous tendency to explore a novel object over a familiar one, a memory function that is dependent on extra-hippocampal cortices (e.g., perirhinal cortex) [37
]. Similarly, when diet cycled rats were withdrawn from the palatable diet and tested in the SAB task, no deficits in working memory were observed, a memory function that is mainly dependent on the prefrontal cortex [38
Memory impairment was observed when rats were withdrawn from the highly palatable diet (i.e., during the C phase), but not when access to the palatable diet was renewed (i.e., during the P Phase). We can confidently exclude that the observed memory and neurogenesis deficits are due to the negative energy homeostatic balance resulting from the hypophagia of the standard diet during palatable food withdrawal. Indeed, caloric restriction has been extensively shown to produce a beneficial effect to both memory function and neurogenesis (see review [39
]). Instead, according to a negatively reinforced mechanism, a more plausible explanation for the observed results is that the stressful emotional state induced by palatable food withdrawal is responsible for the observed memory deficit. Indeed, we have previously shown that, in rats undergoing similar diet cycling procedure, withdrawal from palatable food is accompanied by the emergence of a negative affect characterized by anxiety-like and depressive-like behaviors, similar to what observed in drug addiction [15
]. Accordingly, stress has been extensively shown to negatively impact mnemonic processes [41
], facilitating dorsal striatum-dependent “habit” memory at the expense of hippocampus-dependent “cognitive” memory [42
The present results are in line with human studies showing that dietary factors are associated with the emergence of hippocampal dysfunction and worse hippocampal-dependent memory performance [43
]. Interestingly, previous studies have shown that exposure to energy-dense diets impairs place, but not object, recognition memory in rats [47
]. Our results provide evidence of a differential role for palatable food withdrawal and refeeding in spatial memory performance in rats undergoing a diet alternation procedure.
Hippocampal neurogenesis has been shown to critically support spatial learning and memory processing [49
]. Indeed, specific knockdown of adult neurogenesis results in an impairment of both spatial and object recognition memory in adult rats [50
]. Consistent with a hippocampal-dependent memory deficit, dentate gyrus neurogenesis in cycling rats was disrupted as compared to rats, which were instead monotonously fed the standard chow diet. Interestingly, while the reduction in proliferating cells was diet phase-dependent, as observed by a selective reduction in Ki-67 in palatable food withdrawal, newborn neuron disruption was diet phase-independent, as measured by an overall decrease of DCX in both C and P phases. A possible explanation for these apparently incongruent effects observed in proliferating cells and newborn neurons in the two diet phases may be related to the temporal dynamics of the different stages of the neurogenic process. Cell proliferation in SGZ is a relatively short phenomenon (~1 day, [51
]) that precedes the differentiation into young immature neurons [12
], a stage that instead can last up to four weeks [52
]. Not surprisingly, the number of proliferating cells correlates with the number of newborn immature neurons [32
]. Therefore, in the conditions of these experiments, Ki-67 measurement represents a snapshot of how many new cells were proliferating in a specific diet-phase, while DCX measurement reflects the number of newborn neurons within the previous few weeks of diet cycling. Therefore, since stress is known to be an extrinsic negative modulator of cell proliferation [12
], it is plausible to assume that while every stressful palatable food withdrawal event negatively affected the number of proliferating cells in that specific phase, it also decreases the overall number of newborn neurons, irrespective of the diet phase.
Eating disorders generally occur more frequently in women than in men [53
]. In addition, women appear to be at greater risk to develop affective and stress-related disorders following traumatic experiences, from adolescence throughout adulthood [54
]. Moreover, it has been shown that a gender difference in spatial navigation exists, with women typically performing worse than males [55
]. Although the animal model of diet alternation used here has been developed in both female and male rats, no direct sex comparisons have ever been performed [15
]. Future studies will be important to make direct comparisons between sexes in the consummatory, stress-related, learning and memory outcomes following this diet alternation procedure.
Interestingly, there is evidence suggesting that hippocampal pathology is associated with the onset of increased food intake and body weight gain. Indeed, damage in the hippocampus has been related with a reduced ability to inhibit caloric intake and ultimately, to increased body weight [9
]. Moreover, selective hippocampal lesions as well as neurodegenerative diseases affecting this brain structure are often associated with increased energy intake [9
]. Therefore, while the present data do not allow for any causal relationship between feeding and altered hippocampal function induced by diet alternation, we speculate that hippocampal damage may also contribute to the excessive intake of palatable diet observed here. Once again, we can confidently exclude that the observed disruption of hippocampal neurogenesis process is due to the energy deprivation as caloric restriction has been consistently shown to increase hippocampal neurogenesis [39
Our results show that treatment with the uncompetitive NMDAR uncompetitive antagonist memantine was able to fully restore the mnemonic deficit in rats withdrawn from the palatable diet. In addition, memantine both blocked the excessive intake of palatable food and ameliorated the withdrawal-induced hypophagia of the otherwise acceptable standard diet in rats undergoing the palatable diet alternation procedure. Therefore, the same drug was able to improve both the mnemonic and consummatory maladaptive behavioral responses to palatable diet alternation. Memantine is an FDA approved medication that ameliorates mnemonic symptoms in patients with moderate and severe disease, and has also proved to be effective in individuals affected by binge eating disorders [57
]. Indeed, memantine has been shown to reduce the frequency of binging days and episodes, the severity of the eating disorder in humans [58
], and forms of excessive eating in animals [19
]. Here, we provide evidence that memantine can be a potential treatment option for both memory deficits and aberrant feeding patterns induced by the same diet alternation procedure. Future studies will be helpful to determine whether memantine treatment in this animal model can also reverse impairment in neurogenesis.
In summary, the present study shows that a pattern of palatable food withdrawal dieting/overeating leads to a significant impairment in hippocampal-dependent tasks paralleled by a decrease in hippocampal neurogenesis expressed as a reduced number of immature neurons in the dentate gyrus and a reduction in progenitor cells proliferation. Remarkably, treatment with the uncompetitive NMDAR antagonist memantine proved effective in preventing the phase-specific learning and memory deficits as well as in both reducing the exacerbated binge-like eating when palatable food was presented, and attenuating withdrawal-induced hypophagia of regular food. The present data supports the hypothesis that a history of repeated discrete alternations in food palatability across time profoundly impacts hippocampal function and plasticity, which may in turn facilitate aberrant feeding behavior. Moreover, this study provides evidence that supports the use of memantine as a potential pharmacological treatment for mnemonic and consummatory symptomatology in disorders of eating behavior.