The effect of ketogenic diet on behaviors and synaptic functions of naive mice

Abstract Introduction Beyond its application as an epilepsy therapy, the ketogenic diet (KD) has been considered a potential treatment for a variety of other neurological and metabolic disorders. However, whether KD promotes functional restoration by reducing the pathological processes underlying individual diseases or through some independent mechanisms is not clear. Methods In this study, we evaluated the effect of KD on a series of behaviors and synaptic functions of young adult naive mice. Wild‐type C57BL/6J mice at age of 2–3 months were fed with control diet or KD for three months. Body weight and caloric intake were monitored throughout the experiments. We assessed behavioral performance with seizure induction, motor coordination and activity, anxiety level, spatial learning and memory, sociability, and depression. Synaptic transmission and long‐term potentiation were also recorded. Results KD‐fed mice performed equivalent to control‐diet‐fed mice in the behavioral tests and electrophysiological assays except exhibiting slower weight gain and increased seizure threshold. Conclusions Our results contribute to the better understanding of effects of the KD on physiological behaviors and synaptic functions.


| INTRODUC TI ON
A ketogenic diet (KD) is diet that is rich in fat, have adequate protein and low carbohydrates content, and known to induce and sustain a ketotic state in the body by producing high levels of ketone bodies (Hallbook, Ji, Maudsley, & Martin, 2012;Boison, 2017). Since the 1920s, the KD has been extensively utilized as a useful strategy to treat refractive epilepsy (Youngson, Morris, & Ballard, 2017). During the last decades, application of KD has also been extended to other neurological or non-neurological diseases, and demonstrated its beneficial effects (Koppel & Swerdlow, 2018). For example, KD was reported to reduce the anxiety and improve motor behavior in mice with Rett Syndrome (Mantis, Fritz, Marsh, Heinrichs, & Seyfried, 2009); rescue hippocampal memory defects in mice with Kabuki syndrome (Benjamin et al., 2017); exhibit anxiolytic and cognitionsparing properties or improves motor performance in mouse models of Alzheimer's disease (Brownlow, Benner, D'Agostino, Gordon, & Morgan, 2013;Kashiwaya et al., 2013); improve motor performance in SOD1-G93A transgenic mice of amyotrophic lateral sclerosis (ALS) (Zhao et al., 2006); delay weight loss in the R6/2 1J mouse model of Huntington's disease ; reverse behavioral abnormalities in an acute NMDA receptor hypofunction model of schizophrenia (Kraeuter, Loxton, Lima, Rudd, & Sarnyai, 2015); and reverse diabetic nephropathy, which is a profound diabetic complication (Poplawski et al., 2011).

Despite a variety of beneficial effects of KD on various diseases
in animal studies, it is not clear whether KD offers beneficial effects in normal animals. In other words, whether the beneficial effects of KD are restricted to behavioral impairments occurring in diseases or it affects physiological behavioral performance in general is not well understood. Answer to this question is important because people should be cautious that there may be risks of using KD by healthy users and overuse should be avoided (Schugar & Crawford, 2012). In the last decade, several studies investigated effects of KD on naive animals, however, findings are quite contradictory. For example, Zhao, Stafstrom, Fu, Hu, and Holmes (2004) reported KD have detrimental effects on spatial learning memory and brain growth, while the other study claimed its beneficial effects in novel object recognition test (Brownlow, Jung, Moore, Bechmann, & Jankord, 2017).
Many studies even failed to find its influence on cognitive function (Thio et al., 2010;Fukushima et al., 2015). Electrophysiologically, several studies demonstrated impaired long-term potentiation (LTP) by KD in freely behaving rats (Koranda, Ruskin, Masino, & Blaise, 2011;Blaise, Ruskin, Koranda, & Masino, 2015), despite of that the other find negative results (Thio et al., 2010). In addition, KD is shown to increase sociability in young male rats (Kasprowska-Liskiewicz et al., 2017). However, it exhibited minimal effect in other studies (Ruskin et al., 2013;Verpeut, DiCicco-Bloom, & Bello, 2016;Castro, Baronio, Perry, Riesgo, & Gottfried, 2017). The reasons for these inconsistencies are not very clear, possibly due to variation in animal husbandry and/or experimental designs. Moreover, most of earlier studies employed short-term feeding strategy, ranging from 2 weeks to 2 months, to examine the effect of KD. It is not known whether long-term feeding of KD would show more profound effect on physiological behaviors.
In this study, we conducted a series of behavioral tests and electrophysiological recordings on male adult naive mice fed with KD for 3 months to investigate the effect of KD on physiological behaviors and synaptic functions.

| Animals
Adult male C57BL/6J mice at age of 2-3 months were purchased from Guangdong Medical Laboratory Animal Center. Animals were group-housed and allowed to acclimate to the facility for 1 week prior to experiments. Throughout experimental procedure, mice were single-housed in a room with ad libitum access to food and water. Housing conditions were maintained at a temperature of 22 ± 1°C, at >30% humidity and a standard 12 hr light/dark cycle (08:00-20:00). All experiments were performed in accordance with National Institutes of Health guide for the care and use of Laboratory animals, and approved by the Animal Ethics Committee of Guangzhou Medical University.

| Diets and feeding
The content of customized diets (Research Diets) is as follows (percalorie macronutrient): control (D10070802), 10% protein, 80% carbohydrates, and 10% fat; ketogenic diet (KD, D10070801), 10% protein and 90% fat (also see Table 1). The fat sources are soybean oil and cocoa butter. Micronutrient content, fiber, and preservatives are matched according to a per-calorie basis. Detailed description of the diet composition is shown in the

| Blood ketones
Blood ketone levels were measured using the blood glucose and ke-

| Behavioral analysis
All behavioral tests were conducted consecutively with an interval of 2 days, to minimize the interference of each test on the others.
All tests were performed between 1 and 5 p.m. on the test day.
Mice were handled by test performers for 3 days before behavioral tests. On the test day, mice were acclimated to the testing room for 1 hr before testing. Locomotor activity was measured as described previously (Sun et al., 2016). Briefly, mice were placed in a chamber (40 × 40 × 20 cm) and movement was monitored for 30 min using an overhead camera and tracking software (SMART 3.0, Panlab).
The center 20 × 20 cm region was artificially defined as the center region. Number of entries and duration spent in the center region were recorded. Their movement was recorded for 5 min using an overhead camera and tracking software (SMART 3.0, Panlab). The time that mice spent in the open arms, closed arm, center area, and the number of entries were quantified autonomously. The percentage (%) of time that mice spent in a given area was defined as time spent in the given area (sec)/the total time (300 s) × 100.
During Y-maze test, mice were placed at the center of a Y-shaped maze with three arms (35 cm) at an angle of 120°, and allowed to move freely through the maze for 8 min. The total number and series of arm entries were recorded. Nonoverlapping entrance sequences (e.g., ABC, BCA) were defined as spontaneous alternations. The percentage of spontaneous alternation was defined as the ratio of total number of spontaneous alternations to (total arm entries-2) × 100.
Morris water maze test was performed as described previously (Sun et al., 2016;Ou et al., 2018). Briefly, a day before training, mice were placed in a pool (120 cm in diameter) and scored for ability to find the visible platform (10 cm in diameter) within 60 s. Mice that failed to locate and climb onto the platform twice were excluded from further test. The platform was then moved to a new location and submerged 1 cm beneath surface of white water. Mice were trained for 5 days with 4 trials (60 s in duration) per day to climb on the hidden platform. Seven positions were used to ensure that visual spatial memory was used by mice to find the hidden platform. On the 6th day, the platform was removed and mice were placed into the pool at a new start position and scored for time spent in platform area (N30: 30 cm as diameter) and number of platform crossings within 60 s.
A clear glass cylinder at a height of 40 cm and diameter of 12 cm was used for the forced swimming test (FST). During the test, the cylinder was filled with water (22 ± 1°C) to 23 cm. Mice were gently put into the cylinder recorded for 6 min. The duration of their immobility was quantified autonomously by software (SMART 3.0,

Panlab).
For tail suspension test, a mouse was hanged by wrapping the tail with adhesive tape onto a hook, which connects to the ceiling of the rectangular compartment (70 cm in height, 30 cm in width, 30 cm in depth), and video-recorded for 6 min. The duration of immobility was quantified autonomously by software (SMART 3.0, Panlab).
The social interaction test was performed as described previously with minor modification Wang et al., 2018). Briefly, a rectangular apparatus consisted of two chambers (20 × 25 × 25 cm) with a neutral middle zone (12 × 25 × 25 cm) that allowed for unbiased entry into either chamber. During phase I test, mice were put into the middle zone and habituated for 5 min. An unfamiliar mouse (Stranger 1, S1) was then introduced into a wire cage in left chamber and an empty wire cage put on the right chamber. The test mouse was allowed to freely explore all three chambers for 5 min. The animal remains in the chamber for an extra 5 min to better acquire cues from S1 mouse. During phase II test, a novel stranger mouse (Stranger 2, S2) was put into the previously empty wire cage and the test animal was allowed to explore for another 5 min. Time spent in each chamber was calculated by SMART 3.0

software (Panlab).
For the rotarod test, mice were placed on the rotarod, facing away from the direction of rotation. The rotarod was set with a start speed of 4 rpm, accelerating gradually to 40 rpm in 5 min. Mice were trained twice to adapt to the rod. Briefly, mice were repeatedly placed on the rod if they fall down, until the 5-min session ends.
After rest of 1.5 hr, mice were placed on the rod again, rotating with a speed of 4 rpm to 40 rpm to record the latency when mice fall and speed at which mice fall. Tests were repeated three times on the same day, with interval time at ~1.5 hr. Mean speed at fall was defined as (4 + S)/2, where S is the speed at which mice fall.
To measure the seizure susceptibility, mice were injected with PTZ at dose of 50 mg/kg (i.p.), as previously described (Sun et al., 2016). Right after PTZ injection, mice were put back into their home cages and the time was recorded. The latency till the onset of generalized convulsive seizures (GS) was recorded. Behavioral seizures were scored based on the criteria by Racine (Racine, 1972): stage 0, no seizure; stage 1, head nodding; stage 2, sporadic full-body shaking, spasms; stage 3, chronic full-body spasms; stage 4, jumping, shrieking, falling over; and stage 5, violent convulsions, falling over, death. Stages 4 and 5 were considered as GS. If GS was not observed in 20 min, 20 min were scored.

| Electrophysiological recording
Hippocampal slices were prepared as described previously (Lu et al., 2014). Briefly, mice were anesthetized with isoflurane and brains were extracted and chilled in ice-cold modified artificial cerebro-  used to compare data from two groups. All tests were two-sided.

| Statistical analysis
Data are represented as mean ± SEM. Value of p < 0.05 was considered to be statistically significant.

| KD increases seizure threshold without alterations of locomotor activity and anxiety level
It is well known that KD is an effective treatment for refractory epi- Because of slower weight gain in KD-fed mice, we suspected that KD may affect their motor capacity. In fact, it was reported that KD is able to improve the motor coordination and cognition recovery in young rats suffering from traumatic brain injury (Appelberg, Hovda, & Prins, 2009). To verify this hypothesis, we first examined the motor coordination by the rotarod test, a performance test based on a rotating rod with forced motor activity being applied. The length of time that a given animal stays on this rotating rod at accelerating speed reflects its capacity to balance and coordinate (Deacon, 2013). As shown in Figure 2b, the mean speed at mice falling and the latency to fall were similar between the two groups, suggesting that KD does not

| No effect of KD on spatial memory
To examine whether KD has an influence on cognitive function, we first measured their spatial recognition memory, a kind of short-term memory, by the Y-maze test which is based on the innate tendency of mice to explore novel environments (Dudchenko, 2004;Jiao et al., 2017).

| Unaltered social interaction and depressivelike behaviors by KD
It has been reported that KD has a beneficial effect on sociability of patients and animal models of autism spectrum disorder, as well as naive young rats (Ahn, Narous, Tobias, Rho, & Mychasiuk, 2014;Kasprowska-Liskiewicz et al., 2017). However, whether KD, especially in the case of long-term intervention, exhibits a similar effect on adult male mice is not clear. We utilized a modified three-chamber assay to test voluntary social interaction . After habituation to the three-chamber box, mice were given a choice of either interacting with an empty wire cage or a stranger mouse.
Time that mice spent in either compartment was recorded. As shown in Figure 4a, mice in the KD group exhibited similar preference for the stranger mouse (S1) over the empty wire cage to that in the control group (Two-way ANOVA, F(1,28) = 0.0106, p = 0.9186).
Furthermore, while a second stranger mouse (S2) was placed into the previously empty wire cage, mice in the two groups still showed comparable preference for S2 over S1 (Figure 4b, two-way ANOVA, F(1,28) = 0.1307, p = 0.7204). These results suggest that long-term feeding of KD displayed no effect on social behaviors of adult naive mice.
We further investigated the effect of KD on depression-like behaviors by using forced swimming test (FST) and tail suspension test (TST), two classical methods to detect acute behavioral despair, which is a major symptom of depression (Arauchi & Hashioka, 2018).
As shown in Figure 4c, there was a trend that the duration of immobility for the KD group was a little shorter than controls, although  These data are in agreement with a previous report (Stafstrom, Wang, & Jensen, 1999). These results suggest unaltered basal synaptic transmission. We also examined their synaptic plasticity by recording long-term potentiation (LTP) with tetanic stimulation. As shown in Figure 5b, the induction and maintenance of LTP was not In short, these findings suggest that KD feeding has no effect on basal synaptic transmission and plasticity in adult naive mice.

| D ISCUSS I ON
In this study, we found that, with 3-month feeding of KD, adult naive mice displayed elevated blood ketone level and seizure threshold with slower weight gain, compared with control. However, there was no effect of KD on motor coordination and locomotor activity, non-neurological disorders, its effect on naive animals is less investigated and existing data is not conclusive. Distinct reasons may account for the discrepancy, such as strain, animal age, composition of KD, feeding strategy. Interestingly, most studies adopted shortterm feeding strategy, ranging from 2 weeks to 2 months. In addition, naive rats were frequently used to examine the effect of KD on physiological function. In contrast, very few studies used normal mice as research objects. To our knowledge, this is the first study to specially investigate the influence of long-term feeding of KD on the young adult naive mice.
In our experiments, we successfully induced therapeutic ke-  (Newman et al., 2017;Roberts et al., 2017). These results suggest that the KD may function in an age-dependent manner.
The reason for decreased weight gain under circumstance of equal caloric intake and locomotor activity is not very clear. One possible mechanism may be due to the alteration of resting energy expenditure (Gershuni et al., 2018). This could be caused by decreased insulin level and higher levels of plasma ketone bodies which induce a higher use of lipids by cells, or by changed level of hormones including glucocorticoid, which has been associated with growth suppression (Jequier, 2002;Peres, Nogueira, Paula, Costa, & Ribeiro, 2013 (Fukushima et al., 2015;Kasprowska-Liskiewicz et al., 2017). It is widely known that side effect of KD in clinics includes constipation, vomit, and inappetence, which frequently deteriorates one's emotion. Our observations suggest that with long-term feeding of KD, emotion state is hardly affected in mice.
The Y-maze and water maze tasks indicate that the learning ability, short-and long-term memory are not impaired by KD. In contrast, it seems that KD-fed mice actually learn faster than controls, although it shows only a tendency without statistical significance.
However, this is not consistent with a previous study showing that the KD impaired performance on the Morris water maze test (Zhao et al., 2004). One possible reason is difference in feeding period. In their study, KD was fed on immature rats starting from P21 to P60, a developmental stage when neuronal circuit is still in the process of modification and maturation. However, young adult mice, in which the brains were considerably developed, were utilized in our study to dissect the effect of KD. Consistent with results from Y-maze and water maze tasks, synaptic plasticity in the hippocampus was not altered either.
We also tested the social behaviors of KD-fed mice with threechamber apparatus and found no change in the social exploration preference, suggesting the social behavior is little affected by longterm feeding of KD. These results are interesting in light that numerous studies have reported beneficial effect of KD in social behaviors in autism patients, animal models of autism and wildtype young adult rats (Ruskin et al., 2013;Ahn et al., 2014;Verpeut et al., 2016;Castro et al., 2017;Kasprowska-Liskiewicz et al., 2017). The real reason for the discrepancy is not clear. A proposed mechanism is that the beneficial effect of KD is age-dependent. In the literature, animals at age of 3 weeks or 1 month were fed with KD, feeding process lasted 10 days or 1 month. However, in our study, mice for experiments were considered adult, a developed stage when neuronal plasticity is much different from that of young mice. Another reason may be due to the context of social test. It has been reported that social interaction test carried out in the home cage of the tested animal is more sensitive than three-chamber test (Kasprowska-Liskiewicz et al., 2017). It will be interesting to verify whether the effect of KD on sociability is age-and context-dependent in future studies.
Interestingly, mice with KD exhibited dramatically elevated seizure induction threshold, in accordance with numerous studies showing that application of the KD to multiple animal epilepsy models has demonstrated therapeutic effects, including increased induced-seizure threshold, delayed seizure development, attenuated seizure risk and decreased the seizure severity (Todorova et al., 2000;Mantis et al., 2004;Maalouf et al., 2009;Kawamura, Ruskin, & Masino, 2016). A number of mechanical studies suggest potential therapeutic mechanisms of KD, including metabolic effects of ketonemia, decreased blood glucose and insulin levels; neuronal effects involving ATP-sensitive potassium (KATP) channel modulation, enhanced purinergic signaling, glutamatergic and/ or GABAergic neurotransmission, increased brain-derived neurotrophic factor (BDNF) expression, attenuation of neuroinflammation, as well as improved mitochondrial function, indirect effect through gut microbiota (Boison, 2017;Puchalska & Crawford, 2017;Youngson et al., 2017;Augustin et al., 2018). Further investigations of the mechanisms underlying the specific effect of KD on epilepsy susceptibility rather than other behaviors in naive mice are warranted.
In summary, we examined the effect of long-term feeding of KD on adult naive mice with a bunch of behavioral tests and electrophysiological recordings. Results indicate that KD exhibits little effect in naïve mice, except alteration of weight gain and seizure threshold.
These observations suggest that KD may function most probably in the condition of behavioral impairments occurring in diseases. Our study contributes to the better understanding of effects of the KD on physiological behaviors and synaptic functions.

ACK N OWLED G M ENTS
We thank all the members in the laboratories for helpful discussions and suggestions. We gratefully thank Dr. Bhupesh Singla for revising our manuscript. This study is supported in part by the Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.