Accepted Manuscript

Attenuation of stress-induced weight loss with a ketogenic diet

Elizabeth Sahagun, Lauren M. Ward, Kimberly P. Kinzig

PII: S0031-9384(19)30326-9
DOI: https://doi.org/10.1016/j.physbeh.2019.112654
Article Number: 112654
Reference: PHB 112654
To appear in: Physiology & Behavior
Received date: 28 March 2019
Revised date: 29 July 2019
Accepted date: 15 August 2019

Please cite this article as: E. Sahagun, L.M. Ward and K.P. Kinzig, Attenuation of stress-
induced weight loss with a ketogenic diet, Physiology & Behavior, https://doi.org/10.1016/
j.physbeh.2019.112654

This is a PDF file of an unedited manuscript that has been accepted for publication. As
a service to our customers we are providing this early version of the manuscript. The
manuscript will undergo copyediting, typesetting, and review of the resulting proof before
it is published in its final form. Please note that during the production process errors may
be discovered which could affect the content, and all legal disclaimers that apply to the
journal pertain.
 ACCEPTED MANUSCRIPT

Attenuation of stress-induced weight loss with a ketogenic diet

Keywords: 1) ketogenic diet, 2) chronic stress, 3) hypothalamic-pituitary-adrenal axis, 4) mood

disorders, 5) animal model

Elizabeth Sahagun, Lauren M. Ward, & Kimberly P. Kinzig

Department of Psychological Sciences, Purdue University, West Lafayette, Indiana, USA

T
Corresponding author: Elizabeth Sahagun

IP
Email: esahagun@purdue.edu

CR
Abstract

US
Ketogenic diets (KDs) are high-fat, low-carbohydrate diets that have been used
AN
therapeutically for decades, most notably for the treatment of epilepsy and diabetes. Recent data,

however, suggest that KD may impart protective effects on mood disorders. The current
M

experiments test the hypothesis that KDs can protect from stress-induced symptoms of mood
ED

disorders. To test this, we assessed behavioral and neuroendocrine effects of KD in male and

female Long Evans rats. Animals experienced three weeks of chronic mild stress (CMS) while
PT

consuming KD or control chow (CH). Body weight and food intake data were recorded daily and
CE

behaviors were assayed after three weeks. Plasma beta-hydroxybutyrate (HB), corticosterone

(CORT) and interleukin-1 beta (IL-1) were measured after behavioral testing, along with
AC

hypothalamic corticotropin-releasing hormone (CRH) and neuropeptide Y (NPY) mRNA

expression. CMS induced weight loss in the CH groups, however the KD-fed rats were resistant

to CMS-induced weight loss. Female rats fed KD were protected from CMS-induced reductions

in plasma CORT and hypothalamic NPY expression. Collectively, these data suggest protective

potential of KDs against chronic stress, particularly in females.

 ACCEPTED MANUSCRIPT

1. Introduction
Lifestyle changes known to decrease stress and improve mental health are commonly

suggested for individuals reporting symptoms of affective disorders (Walsh, 2011). Improved

nutrition through dietary modifications, or implementing relatively undefined “healthy eating

habits”, are suggested as beneficial to overall well-being. Even so, data supporting the health

T
IP
benefits of various dietary approaches have produced equivocal findings. Incomplete

CR
understanding of how dietary changes can impact mood and mental health can result in

individuals turning to non-empirical sources of information, such as anecdotes from media

US
sources making indirect untested claims. For example, the sustained popularity of ketogenic diets

(KD) for improved health is largely fueled by lay reports in which there are widespread claims
AN
that severe restriction of carbohydrates can increase energy and mood.
M

Empirical data demonstrate that KDs are effective in the treatment of epilepsy and
ED

diabetes, and are hypothesized to have other positive neuronal benefits. (Neal et al., 2008; Paoli,

Rubini, Volek, & Grimaldi, 2013). KD has been shown to protect from harmful diet-induced
PT

deficits in the brain such as impaired performance on cognitive tasks, decreased brain derived
CE

neurotropic factor (BDNF), as well as increased hippocampal inflammation and blood brain

barrier permeability (Davidson et al., 2013; Hargrave, Davidson, Lee, & Kinzig, 2015; Hargrave,
AC

Davidson, Zheng, & Kinzig, 2016). KDs can also protect from metabolic insult and peripheral

immune challenges (Dupuis, Curatolo, Benoist, & Auvin, 2015; Kimberly P. Kinzig, Honors, &

Hargrave, 2010; Kimberly P Kinzig, Scott, Hyun, Bi, & Moran, 2005; Youm et al., 2015). More

recently, data suggest that KD may be neuroprotective in response to the pathophysiology of

mood disorders and adverse effects of chronic stress (Brietzke et al., 2018; Brownlow, Jung,

Moore, Bechmann, & Jankord, 2017).

 ACCEPTED MANUSCRIPT

The neuroprotective effects of KD are suggested to rely on the induction of ketosis and

elevation of beta-hydroxybutyrate (βHB). The elevation of βHB using a 80% lard-based diet has

shown to protect from diet-induced cognitive deficits and blood brain barrier leakage (Davidson

et al., 2013; Hargrave et al., 2015, 2016). βHB is produced by the liver and elevated in the blood

stream in response to consuming a high-fat/low-carbohydrate diet, prolonged fasting, or

T
exercising (Mattson, 2014). βHB enters the brain by active transport via monocarboxylate

IP
transporters (Achanta & Rae, 2016). The presence of βHB in the CNS has been demonstrated to

CR
decrease anxiety-related behavior and influence activity levels (Ari et al., 2016; Murphy,

Likhodii, Hatamian, & Burnham, 2005).

US
Mood disorders constitute neurological changes that result in emotional dysregulation
AN
and cognitive deficits. The core of some symptoms can also be the result of energy dysregulation

driven by poor diet. Existing evidence shows that diets high in fat and sugar can cause
M

depression-like symptoms, and epidemiological research shows that depression is often

ED

comorbid with metabolic disorders (Luppino, Floriana S.; de Wit, Leonore M.; Bouvy, 2010).

There is evidence suggesting KD can reverse the adverse metabolic effects of high- fat/ high-
PT

sugar diet, however the effects on mental health are unclear. Collectively, data suggest that KD
CE

may be beneficial to both mental and metabolic health.

Chronic stress increases one’s susceptibility to develop a mood disorder. Stressful stimuli
AC

activate the sympathetic nervous system, which leads to release of hormones that motivate a

“fight or flight” response. Under normal conditions, this is an acute adaptive response of

survival that is mediated by the hypothalamic-pituitary- adrenal (HPA) axis. The HPA axis has

several feedback mechanisms to return the brain and body back to homeostasis once a stressful

event is over. Chronic activation can lead to dysregulatio n of the HPA axis and becomes

detrimental to homeostatic processes such as energy regulation (Torres & Nowson, 2007),
 ACCEPTED MANUSCRIPT

desensitization of neurotransmitter function in higher up brain regions (Lucassen et al., 2014),

and immune function (Wohleb, Franklin, Iwata, & Duman, 2016).

The goal of the following experiments was to elucidate the effects of KD on the

development of anxiety- and depressive-like behaviors, analogous to symptoms of mood

disorders, that result from exposure to chronic mild stress (CMS). Rats were given a KD with

T
macronutrient ratios similar to a strict KD adopted by humans (80% fat, 15% protein, and 5%

IP
carbohydrate) (Kimberly P. Kinzig et al., 2010), and exposed to a series of mild stressors over

CR
the course of three weeks. We hypothesized that rats consuming KD during CMS will be

resilient to deleterious effects typically induced by the model, as measured by behavioral and

US
neuroendocrine outcomes.
AN
2. Materials Methods
M
ED

2.1 Animals and Diets

PT

Adult (16-17 week old) male (N=33) and female (N=37) Long-Evans rats were

individually housed in suspended wire cages in humidity (55-65%) and temperature (20.5 ± 1
CE

°C) controlled rooms that were maintained with a 12:12 hour light/dark cycle unless stated in the
AC

stress schedule (Table 1). Males and females were housed in separate rooms with light/dark

cycles staggered by 1 hour to maintain consistency of stress and testing times between groups.

Animals were weight matched and then assigned to either KD or chow (CH) diet. The KD

(Research Diets D06040601, New Brunswick, NJ) consisted of 80% fat, 5% carbohydrates, and

15% protein by calories. This lard-based diet has been used to effectively impart therapeutic

effects of elevated βHB (Hargrave et al., 2016; Kimberly P. Kinzig et al., 2010; Kimberly P
 ACCEPTED MANUSCRIPT

Kinzig et al., 2005). The control diet, CH (Teklad 2018, Envigo, Indianapolis, IN), consisted of

18% fat, 58% carbohydrates, and 24% protein. Diet compositions are further described in Table

2. After four days of consuming the assigned diet, animals were randomly assigned to a stress (+)

or non-stress (-) subgroup for the remainder of the experiment. There were four experimental

groups per sex: CH-, CH+, KD-, KD+. Group n for males were 7, 9, 8, 9, and for females 9, 9, 9,

T
10, respectively. All animals had ad libitum access to food and water in the home cage. Animals

IP
were weighed daily and food intake was measured by weighting food container daily and

CR
subtracting food spillage. All procedures were approved by the Purdue Animal Care and Use

Committee (PACUC).

US
AN
2.2 Chronic Mild Stress (CMS)
M

Stress groups underwent three weeks of CMS by exposure to daily stressors (Willner,
ED

2017), which are described in Table 1. In short, animals were switched to their respective diets

and given four days to habituate to their new diet. Then on days 5-26, stressors were presented
PT

once a day for various amounts of time in a pseudorandom manner to prevent habituation and

predictability. After CMS, behavioral responses to the elevated plus maze (EPM), open field
CE

(OFT), and forced swim (FST) were evaluated on consecutive days. Rats were sacrificed
AC

approximately two hours after FST for plasma and hypothalamic tissue analyses.

2.3 Behavioral Assessments

Food intake was measured daily and final behavioral tests were done for three

consecutive days following the end of CMS. All final testing was done in a room adjacent to the
 ACCEPTED MANUSCRIPT

vivarium during the first half of the light cycle. Animals were given 10-15 minutes to habituate

to the room before the start of each test. Test order was counterbalanced within each sex to

minimize test time bias.

2.3.1 Elevated Plus Maze

T
Anxiety-like behavior was assessed using the elevated plus maze (EPM). The EPM

IP
apparatus is 3 feet high, and has with four arms at 90-degree angles, each 6 inches wide and 3

CR
feet long. Two of the opposite side arms are enclosed and the other two are open. Animals were

placed in the center of the EPM and behavior was video recorded from a ceiling mounted camera

US
for 5 minutes. Rats were returned to home cages immediately after testing. The percent of time
AN
spent in open arms and in the center was scored manually by an observer blinded to the

experimental conditions because AnyMaze tracking was not sensitive enough to track the head
M

of Long Evans on the apparatus.

ED

2.3.2 Open Field Test

PT

The open field test (OFT) quantifies locomotor activity and anxiety-like behavior. The
CE

apparatus is 4 feet by 4 feet, with a grid printed on the floor. The rat is placed in the center of the

apparatus and video recorded for five minutes on a ceiling view mounted camera. Videos were
AC

later scored automatically using Any-maze behavior tracking software (Stoelting Co.) Total

distance traveled, percent time spent in the center and time spent in the corners of the apparatus

were quantified.

2.3.3 Forced Swim Test

 ACCEPTED MANUSCRIPT

Depressive-like behaviors were assessed in rats by forced swim test (FST). The FST

apparatus was a 22x17x25 inch glass container with enough water to prevent rats from stepping

on the bottom of the apparatus. Water was maintained at 25-30° C. Rats were monitored for the

duration of the test and the test was stopped if rats did not maintain a swim or float behavior. The

test was recorded for five minutes using a ceiling mounted camera. After five minutes, fur was

T
immediately dried with a towel and rats were returned to the home cage. Any-maze behavior

IP
tracking software (Stoelting Co.) was used to score behaviors. An observer blinded to treatment

CR
conditions evaluated time spent immobile, latency to immobility, and time spent displaying

escape behaviors.

2.4 Terminal Assessment US

AN
Rats were sacrificed by pentobarbital (Beauthanasia, D, Merck Animal Health, Madison,
M

NJ) overdose approximately two hours after FST test for collection of blood and removal of
ED

brains. Following rapid decapitation, 3 mL of trunk blood was collected in VACUETTE®

K3EDTA blood collection tubes (Greiner Bio-One, Monroe, NC) and centrifuged for 15 minutes
PT

at 4°C, and 2500 rpm. Plasma was isolated and aliquoted into three tubes and stored in a -80°C

freezer for later analyses. Plasma analyses included circulating levels of corticosterone, βHB,
CE

and IL-1β. Brains were rapidly removed and rinsed with 1x phosphate buffered saline (PBS).
AC

The hypothalamus of each rat was micro-dissected on ice, flash frozen on dry ice, then stored at -

80°C for later analysis of CRH and NPY.

2.4.1 Corticosterone Assay

Plasma corticosterone (CORT) concentration was quantified in duplicates using an EIA

kit according to manufacturer instructions (Enzo Life sciences, protocol ADI-900-097). Samples
 ACCEPTED MANUSCRIPT

were diluted 1:40 using a steroid displacement reagent and standards were serially diluted 1:50.

100 μl of sample or standard were loaded on donkey anti-sheep IgG coated plates, followed by

50 μl of conjugate, then 50 μl of sheep antibody to rat CORT. Plates were then incubated at room

temperature on a plate shaker for 2 hours at 500 rpm, and then washed 3 times. After washing,

200 μl of pNpp substrate solution was added to each well, and then incubated at room

T
temperature for one hour. 50 μl of stop solution was then added to every well, and optical density

IP
was immediately read at 405 nm absorbance using a microplate reader (Multiskan Ascent,

CR
Thermo Scientific).

US
2.4.2 Beta-Hydroxybutyrate Assay

Plasma βHB was measured in duplicate by β-Hydroxybutyrate LiquiColor® assay kit

AN
(Stanbio, Laboratory, Boerne, TX) using a microplate reader. First, 1 mM standard solution was
M

diluted with deionized water into five concentrations: 1, 0.75, 0.5, 0.25, and 0.125 mM. 2.5 μl of
ED

sample or standard were loaded on a 96 well plate. Reagent 1 was incubated at 25°C for 10 min

then 90 μl were dispensed into each well. The absorbance of 505 nm was immediately taken as
PT

the baseline reading. After adding 15 μl of reagent 2 per well, the plate was incubated in 25°C

for 10 min and a second reading was taken. The difference between the second and the first
CE

measurements was used to quantify expression and the concentration of plasma samples were
AC

interpolated from the standard curve.

2.4.3 IL-1β Assay

Plasma IL-1β was quantified in duplicate with R & D Quantikine ELISA Rat IL-1β /IL-

1F2 immunoassay according to the manufacturer’s instructions. In summary, Plasma samples

were diluted 1:3 using the provided calibrator diluent and the assay standard was serially diluted
 ACCEPTED MANUSCRIPT

with a dilution factor of 2 in 8 concentrations. 50 μl of assay diluent was added to rat IL-1β

coated well plates, then 50 μl of sample or standard were loaded into each well plate then

incubated at room temperature for 2 hours. After the incubation period, wells were washed 5

times, loaded with 100 μl of conjugate, incubated again at room temperature for 2 hours, then

washed again five times. After second wash, 100 μl of substrate solution was added to each well,

T
incubated for 30 min at room temperature, protected from light, then 100 μl of stop solution was

IP
added. Immediately following the addition of stop solution, 450 nm absorbance was measured in

CR
a microplate reader (Multiskan Ascent with Ascent software v2.6, ThermoFisher Scientific,

Waltham, MA).

2.4.4 RNA Extraction and cDNA Synthesis US

AN
Tissues were homogenized in 1 ml TriReagent (RT111, Molecular Research Center,
M

Cincinnati, OH) then 50 ul bromanisole (BCP, Molecular Research Center, Cincinnati, OH) was
ED

used to extract RNA. RNA was quantified by spectrophotometer at 260 nm absorbance

(Eppendorf BioPhotometer 22331, Hauppauge, NY). First Strand cDNA Synthesis Kit
PT

(ThermoFisher Scientific, Waltham, MA) was used to synthesize cDNA from 2 μg total RNA.
CE

2.4.5 CRH and NPY RT-qPCR

AC

Real-time quantitative PCR (qPCR) was performed with PowerUp SYBR Green Master

Mix (Applied BioSystems, ThermoFisher Scientific, Waltham, MA) on an iCycler iQ + MyiQ

Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules, CA). The primer sequences

for β-actin (NM_031144), NPY(NM_012614), and CRH (NM_031019) were designed and/or

validated using Primer-BLAST (National Center for Biotechnology Information, NIH), and the

thermal profile for each primer was optimized to 90-110% efficiency. Primer sequences are
 ACCEPTED MANUSCRIPT

listed in Table 3, as are sequence and thermal conditions. Target gene expression levels were

normalized to β-actin by ΔΔCt method.

2.5 Statistical Analyses

Data were analyzed using R Studio (version 1.1.464, R Studio, Inc. Boston, MA2018).

T
Kruskal-Wallis test was first run to evaluate whether combined male and female data were

IP
normally distributed. χ2 values for sex comparisons are reported for each outcome measure. If

CR
data were parametric, then an analysis of variance (ANOVA) was run with three factors: stress,

diet, and sex. If Kruskal-Wallis test indicates male and female groups are non-parametric,

US
ANOVA was run with two factors, stress and diet, per sex. Time was also included as a repeated
AN
factor, when necessary. Follow-up Tukey’s post hoc multiple comparison test are reported when

significant. Combined male and female data are illustrated using GraphPad Prism software
M

(v7.00, GraphPad software Inc., La Jolla, CA). Significance was set at 95% percent confidence
ED

intervals (p < 0.05) for all analyses, and effects of stress (#), diet (*), and sex (†) are reported.

Mean ± standard error of the mean (SEM) are illustrated in figures.

PT
CE

3. Results
AC

3.1 Beta-Hydroxybutyrate levels

Final βHB was measured to confirm animals on KD had elevated circulating βHB.

Parametric analyses found male and female plasma βHB levels were normally distributed

(Kruskal-Wallis test, χ2 = 2.143, p = 0.143), therefore, three way analyses were appropriate.
 ACCEPTED MANUSCRIPT

Animals consuming KD were found to have significantly higher βHB than animals on CH

(F(1,57)= 45.689, p= <0.001), as shown in Figure 2. There were no significant effects of stress

(F(1,57)=0.035, p= 0.853), or sex (F(1,57)= 2.565, p= 0.115), and no significant interactions.

3.2 Daily Caloric Intake

Combined caloric intake data were not normally distributed between sexes (Kruskal-

T
Wallis χ2 = 50.918, p = <0.001). Because this failed to meet assumptions to analyze sex as an

IP
ANOVA factor, male and female data were analyzed separately with stress and diet as factors.

CR
Analysis of average daily food intake (Figure 3) over the three weeks suggests there were no

significant effects of stress (F(1,29)= 2.220, p= 0.147) or diet (F=(1,29)3.997, p=0.055) within

US
male groups however a significant effect of stress within females (F(1,33)=6.035, p= 0.019).
AN
Tukey’s multiple comparison tests suggest this was driven by CH+ females, which had

significantly lower average food intake than KD- females (p=0.042). No effects of diet (F(1,33)=
M

2.201, p=0.147) were observed in females.

ED

3.3 Body Weight

PT

Baseline body weights were subtracted from final weight to calculate body weight gain

from the start to the end of the experiment. Data were non-parametric (Kruskal-Wallis χ2 =
CE

7.133, p = 0.007) so data were analyzed by stress and diet (Figure 4). For males, there were
AC

significant effects of stress (F(1,29)= 10.676, p= 0.003) and diet (F(1,29)= 19.712, p= <0.001).

Tukey’s post hoc test analysis found that these effects were driven by CH+ group. CH+ gained

significantly less weight than CH- (p=0.022), KD- (p=<0.001), and KD+ groups (p=0.001). For

females, there was only a significant effect of stress (F(1,33)=9.281, p= 0.005) and not diet

(F(1,33)=3.329, p= 0.077). Tukey’s post hoc analysis found that body weight gain differences

were also driven by the CH+ group. CH+ females gained significantly less weight than CH- (p=
 ACCEPTED MANUSCRIPT

0.003), KD- (p = 0.001), and KD+ (p = 0.049). Collectively, CMS males and females consuming

CH gained less weight than their respective non-stress CH consuming group. No differences

were observed in KD consuming animals.

3.4 Behavioral Assessments

T
3.4.1 Elevated Plus Maze (EPM)

IP
Behaviors on the EPM were measured three weeks after CMS and analyzed for percent of

CR
time in closed, center, and open parts of the apparatus. Compiled sex data were normally

distributed for all three measures (closed arms: Kruskal-Wallis χ2 = 0.799, p = 0.371, open arms:

US
Kruskal-Wallis χ2 = 1.614, p = 0.204, center quadrant: Kruskal-Wallis χ2 = 0.779, p = 0.377).
AN
There was a significant effect of stress on time spent in the closed arms of the apparatus

(F(1,62) = 7.127, p= 0.01) (Figure 5A). Tukey’s multiple comparison test suggests this was
M

largely driven female CH+ group, as they significantly less time spent in the closed arms than
ED

was spent by CH- females (p=0.034). There were no significant effects of diet (F(1,62)= 1.203,

p= 0.277) or sex (F(1,62)= 1.697, p= 0.197).

PT

There were no general treatment effects of stress, diet or sex on time spent in center of
CE

apparatus, however there was an interaction of stress and diet (F(1,62)= 4.043, p=0.049) (Figure

5B). Tukey’s multiple comparison test suggests that stressed animals consuming KD spent less
AC

time in the center of the apparatus (p= 0.054).

There was also a significant effect of stress on time spent in the open arms of the EPM

(F(1,62)=7.346, p= 0.009) (Figure 5C). Tukey’s multiple comparison test suggests this was

largely driven by female stress groups (p=0.02). There were no significant effects of diet

(F(1,62)= 0.042, p=0.839) or sex (F(1,62)= 3.597, p=0.063).

 ACCEPTED MANUSCRIPT

3.4.2 Open Field Test (OFT)

OFT behaviors were measured on second day of behavior testing. Locomotor activity

was measured as distance traveled in apparatus and percent of time spent in center or corner

quadrants of apparatus as anxiety- like behavior. Compiled sex data were normally distributed for

T
IP
all three behavior measures (center: Kruskal-Wallis χ2 = 3.198, p = 0.074, corner: Kruskal-

CR
Wallis χ2 = 1.059, p = 0.303, distance: Kruskal-Wallis χ2 = 0.287, p = 0.592). There were no

significant effects of sex, stress, or diet on locomotor activity or anxiety- like behavior in OFT

US
(Figure 6A-C). There were also no interactions between the three factors.
AN
3.4.3 Forced Swim Test (FST)

FST behaviors were measured on third day of behavior testing. Sex data were not
M

normally distrubuted across all FST measures (immobility: Kruskal-Wallis χ2 = 4.369, p =

ED

0.037, latency to immobility: Kruskal-Wallis χ2 = 0.717, p = 0.397, escape: Kruskal-Wallis χ2 =

PT

6.699, p = 0.009). Therefore, FST data are reported by two-way ANOVA per sex when they are

not normally distributed and three way ANOVA when they are.
CE

There were no significant effects of stress (F(1,29)= 0.003, p=0.956) or diet (F(1,29)=

0.064, p= 0.801) on time spent immobile in males or females (stress: (F(1,32)= 0.960, p= 0.335;
AC

diet (F(1,32)= 0.858, p= 0.361) (Figure 7A).

There were no signficant effects of stress (F(1,61)= 2.222, p= 0.141) or diet (F(1,61)=

0.000, p= 0.992) or sex (F(1,62)= 0.450, 0.505) on latency to immobility (Figure 7B).

There were no signficant effects of stress (F(1,29)= 0.327, p=0.572) or diet (F(1,29)=

0.006, p= 0.937) on time spent escaping in males or females (stress: (F(1,32)= 0.030, p= 0.864;
 ACCEPTED MANUSCRIPT

diet (F(1,32)= 0.462, p=0.501) (Figure 7C).

3.5 Neuroendocrine Assessment

3.5.1 Corticosterone (CORT)

Terminal corticosterone (CORT) was analyzed using plasma samples collected

T
IP
approximately 2 hours after forced swim test (FST). Data were not normally distributed across

CR
sexes (Kruskal-Wallis χ2 = 4.686, p = 0.03) so each sex was analyzed by two-way ANOVA by

stress and diet (Figure 8). There were no effects of stress (F(1,25)= 0.000, p= 0.985), or diet

US
(F(1,25)=0.072, p= 0.791) observed in male groups with no significant interaction.

Significant effects of terminal CORT were observed in females. Although there were no
AN
main effects of stress (F(1,31)=0.708, p=0.406) or diet (F(1,31)=0.297, 0.589), there was a
M

significant interaction (F(1,31)=7.806, p=0.009). Tukey’s multiple comparisons found that

ED

females that underwent CMS while consuming CH had lower CORT than CH female controls

(p=0.038).
PT

3.5.2 Interleukin-1β (IL-1β)

CE

Terminal plasma samples were used to analyze circulating levels of cytokine, IL-1β.

Combined sex data were normally distributed (Kruskal-Wallis χ2 = 3.554, p = 0.059) so three-
AC

way ANOVA was run with stress, diet, and sex as factors (Figure 9). There were no effects of

stress (F(1,62)=0.003, p= 0.956), diet (F(1,62)= 0.095, p= 0.758), or sex (F(1,62)= 0.0489,

p=0.826). Notably, all values were toward the lower reliably detectable limit (9 ng) according to

the kit manufacturer, suggesting minimal to no inflammation in any of the groups.

 ACCEPTED MANUSCRIPT

3.5.3 Neuropeptide Y (NPY) and Corticotropin-Releasing Hormone (CRH) Expression

Hypothalamic tissue was collected and micro dissected approximately two hours after final

behavior test. NPY expression was not normally distributed across sexes (Kruskal-Wallis χ2 =

3.926, p- = 0.048) so each sex was analyzed by two-way ANOVA (Figure 10A). Within male

groups, there were no significant effects of stress (F(1,29)= 0.148, p= 0.703), or diet (F(1,29)=

T
0.727, p= 0.401). Within female groups, there was a significant effect of stress (F(1,30)= 4.291,

IP
p= 0.047), but not diet (F(1,30)=0.857, p= 0.362). Tukey’s post hoc analysis revealed the diet

CR
effect in females was driven by a significant difference between CH- and CH+ groups (p=

0.045). Changes in hypothalamic NPY expression were only observed in CH+ females.

US
CRH expression was normally distributed (Kruskal- Wallis χ2 = 0.106, p = 0.744), therefore
AN
a three-way ANOVA was used, with stress, diet, and sex as factors (Figure 10B). There were no

significant effects of stress (F(1,59)= 0.665, p= 0.418), diet (F(1,59)=0.344, p= 0.560), or sex
M

(F(1,59)= 0.017, p= 0.897), on hypothalamic CRH expression and no significant interactions

ED

between the three factors.

PT
CE

4. Discussion
AC

Male and female rats that experienced chronic mild stress (CMS) while consuming a

chow diet (CH+) showed some phenotypic characteristics of affective disorders, such as

decreased body weight. Differences in time spent in the open arms of the EPM, plasma CORT,

and hypothalamic NPY expression were also observed, but only in female groups. Additionally,

our data did not demonstrate any behavioral or neuroendocrine differences in either KD+ or KD-

condition. This suggests that KD animals were protected from stress-induced changes in body
 ACCEPTED MANUSCRIPT

weight, CORT, and NPY. Collectively, these data suggest KDs have some therapeutic potential

in normalizing physiological changes related to energy expenditure, particularly in females.

Increased time spent in the open arms of EPM and decreased CORT in females is

consistent with chronic stress experiments from other groups, and indicative of anxiety-like

behavior (McCormick, Smith, & Mathews, 2008). The FST served as an acute stressor, and

T
females that undergo acute stress have been shown to have higher CORT than those that have

IP
undergone chronic stress (Kudielka & Kirschbaum, 2005). The direction of behavioral changes is

CR
opposite of what is typically observed in males. The dampened response in females may be

related to differences in adaptive response to chronic stress (Cook & Wellman, 2004). Females

US
exhibit hypoactivity in the HPA axis in response to chronic stress, whereas males typically
AN
exhibit hyperactivity. This sexual dimorphism has been attributed to the multifaceted roles of

estrogens and reproductive cycles. This phenomenon is still not fully understood, albeit its
M

critical importance in understanding how and why there are sex differences in not only the
ED

development of mood disorders, but also metabolism (Baker, Kentner, Konkle, Santa-Maria

Barbagallo, & Bielajew, 2006; Mauvais-Jarvis, 2015). Estrous cycle was not measured in this
PT

experiment as typical methods, such as vaginal swabs, are stressful and an extra stress variable to
CE

females could have influenced the data interpretation of general sex differences.

Some of our data were inconsistent with what has been reported by others using CMS
AC

models. Inconsistencies in stress responses include a lack of change in time spent floating in the

FST, no changes in time in the center of OFT, no changes in plasma CORT and IL-1β, or

changes in hypothalamic CRH (Hill, Hellemans, Verma, Gorzalka, & Weinberg, 2012). This

could be attributed to a number of issues such as external environmental factors, stressors being

too mild, habituation to stress, and even the sex of experimenters (Sorge et al., 2014; Willner,

2016b, 2016a). Because of these additional factors that can influence the paradigm, criticism for
 ACCEPTED MANUSCRIPT

this particular model for its issues of reproducibility across labs has been growing (Antoniuk,

Bijata, Ponimaskin, & Wlodarczyk, 2019; Cryan, Markou, & Lucki, 2002; McGonigle, 2014;

Nestler, Gould, & Manji, 2002)

Lack of changes in CRH and IL-1β could also be due to how plasma and tissue were

processed. Animals were sacrificed approximately 2 hours (+/- 15min) after FST, when animals

T
were expected to return to a basal level. The high variability suggests there were some

IP
differences in some animals, and it is possible that repeated blood collections at different time

CR
points post-FST would have revealed differences in stress responses in our experiments. It is

known, for example, that males that experience CMS have higher basal levels, but respond

US
differently at 15, 30, and 60 minutes post- acute stress (Bielajew, Konkle, & Merali, 2002). Pre-
AN
behavior testing samples would clarify basal levels, however, the stress of blood collection could

introduce an acute stress episode before testing and influence behavior testing. Additionally,
M

NPY expression levels were more variable in males under experimental conditions in this
ED

experiment than is typically reported. This greater variability may also be attributed to the timing

of tissue collection, given that CORT, CRH, and NPY mechanisms are interrelated.
PT

Other high fat diet studies, such as those using a ‘Western diet’ (WD) that is high in both
CE

fat and carbohydrate, demonstrate that calorie-dense diets high in fat that are not ketogenic

disrupt HPA activity. This has been related to disordered feeding behavior and induction of a
AC

depressive-like phenotype (Dutheil, Ota, Wohleb, Rasmussen, & Duman, 2015; Rho, 2015;

Tannenbaum et al., 1997; Zemdegs et al., 2016). Our current data demonstrate that minimizing

the carbohydrate level of a very high fat diet does not produce the negative effects of WD, and

do not increase susceptibility to the development of symptoms consistent with affective

disorders. Our data show that KD can play a role in preventing stress-induced decreases in body

weight which is indicative of symptoms of anxiety and depression identified in the DSM-V. This
 ACCEPTED MANUSCRIPT

could be attributed to the elevation of BHB observed in KD-consuming animals. Our

experiments demonstrate no depressive- or anxiety- like behavior, body weight, food intake, or

hypothalamic dysregulation when animals consumed a diet with 80% fat. This suggests that the

effects of high fat diets (containing 42-60% fat) are non-linear and that it is likely that other

elements of these diets, such as simple carbohydrates, alone or in combination with high fat

T
induce deleterious effects on mental health. This is consistent with previous experiments from

IP
our lab suggesting that KDs may impart protective effects on metabolic and cognitive health

CR
(Hargrave et al., 2015; K P Kinzig & Taylor, 2009; Kimberly P. Kinzig et al., 2010)

The goal of this experiment was to elucidate the protective effects of KDs to the

US
development of stress-induced mood disorders in a pre-clinical model. The model chosen is
AN
reportedly influenced by numerous factors that are difficult to replicate between labs, thus

limiting the findings in this study. Although the range of depressive and anxiety- like behaviors
M

that are analogous to those observed in mood disorders were not reached with this model, we did
ED

find that KD somewhat increased resiliency to stress-induced changes observed in groups

consuming a normal chow diet. Perhaps a different pre-clinical model, such as chronic
PT

psychosocial stress would further test our hypothesis, however, sex differences in these models
CE

are challenging to interpret as males and females respond to different psychosocial stressors

(Fuchs, 1999). Further research is warranted not only on the effects of KD on mental health, but
AC

also on what other aspects of high fat/high carbohydrate diets play roles in mood disorders.

6. Acknowledgements
 ACCEPTED MANUSCRIPT

We would like to express our appreciation to Dr. Julia Chester for assisting with

corticosterone assays, and Melissa McCurley for technical assistance with animal husbandry,

behavioral testing, and tissue collection.

7. References

T
IP
Achanta, L. B., & Rae, C. D. (2016). Beta-Hydroxybutyrate in the Brain: One Molecule,

CR
Multiple Mechanisms. Neurochemical Research, 42(1), 1–15.

https://doi.org/10.1007/s11064-016-2099-2

US
Antoniuk, S., Bijata, M., Ponimaskin, E., & Wlodarczyk, J. (2019). Chronic unpredictable mild
AN
stress for modeling depression in rodents: Meta-analysis of model reliability. Neuroscience

and Biobehavioral Reviews, 99(December 2018), 101–116.

M

https://doi.org/10.1016/j.neubiorev.2018.12.002
ED

Ari, C., Kovács, Z., Juhasz, G., Murdun, C., Goldhagen, C. R., Koutnik, A. M., … D’Agostino,

D. P. (2016). Exogenous Ketone Supplements Reduce Anxiety-Related Behavior in

PT

Sprague-Dawley and Wistar Albino Glaxo/Rijswijk Rats. Frontiers in Molecular

CE

Neuroscience, 9(December), 1–10. https://doi.org/10.3389/fnmol.2016.00137

Baker, S. L., Kentner, A. C., Konkle, A. T. M., Santa-Maria Barbagallo, L., & Bielajew, C.
AC

(2006). Behavioral and physiological effects of chronic mild stress in female rats.

Physiology & Behavior, 87(2), 314–322. https://doi.org/10.1016/j.physbeh.2005.10.019

Bielajew, C., Konkle, A. T. ., & Merali, Z. (2002). The effects of chronic mild stress on male

Sprague–Dawley and Long Evans rats: I. Biochemical and physiological analyses.

Behavioural Brain Research, 136(2), 583–592. https://doi.org/10.1016/S0166-

4328(02)00222-X
 ACCEPTED MANUSCRIPT

Brietzke, E., Mansur, R. B., Subramaniapillai, M., Balanzá-martínez, V., Vinberg, M., González-

pinto, A., … Mcintyre, R. S. (2018). Neuroscience and Biobehavioral Reviews Ketogenic

diet as a metabolic therapy for mood disorders : Evidence and developments. Neuroscience

and Biobehavioral Reviews, 94(July), 11–16.

https://doi.org/10.1016/j.neubiorev.2018.07.020

T
Brownlow, M. L., Jung, S. H., Moore, R. J., Bechmann, N., & Jankord, R. (2017). Nutritional

IP
Ketosis Affects Metabolism and Behavior in Sprague-Dawley Rats in Both Control and

CR
Chronic Stress Environments. Frontiers in Molecular Neuroscience, 10(May), 1–17.

https://doi.org/10.3389/fnmol.2017.00129

US
Cook, S. C., & Wellman, C. L. (2004). Chronic stress alters dendritic morphology in rat medial
AN
prefrontal cortex. Journal of Neurobiology, 60(2), 236–248.

https://doi.org/10.1002/neu.20025
M

Cryan, J. F., Markou, A., & Lucki, I. (2002). Assessing antidepressant activity in rodents: recent
ED

developments and future needs, 23(5), 238–245. https://doi.org/10.1016/S0165-

6147(02)02017-5
PT

Davidson, T. L., Hargrave, S. L., Swithers, S. E., Sample, C. H., Fu, X., Kinzig, K. P., & Zheng,
CE

W. (2013). Inter-relationships among diet, obesity and hippocampal-dependent cognitive

function. Neuroscience, 253, 110–122. https://doi.org/10.1016/j.neuroscience.2013.08.044

AC

Dupuis, N., Curatolo, N., Benoist, J. F., & Auvin, S. (2015). Ketogenic diet exhibits anti-

inflammatory properties. Epilepsia, 56(7), e95–e98. https://doi.org/10.1111/epi.13038

Dutheil, S., Ota, K. T., Wohleb, E. S., Rasmussen, K., & Duman, R. S. (2015). High Fat Diet

Induced Anxiety and Anhedonia: Impact on Brain Homeostasis and Inflammation.

Neuropsychopharmacology, 41(October), 1–30. https://doi.org/10.1038/npp.2015.357

Fuchs, E. (1999). Defeat is a major stressor in males while social instability is stressful mainly in
 ACCEPTED MANUSCRIPT

females : Towards the development of a social stress model in female rats, 50(1), 33–39.

Hargrave, S. L., Davidson, T. L., Lee, T. J., & Kinzig, K. P. (2015). Brain and behavioral

perturbations in rats following Western diet access. Appetite, 93, 35–43.

https://doi.org/10.1016/j.appet.2015.03.037

Hargrave, S. L., Davidson, T. L., Zheng, W., & Kinzig, K. P. (2016). Western diets induce

T
blood-brain barrier leakage and alter spatial strategies in rats. Behavioral Neuroscience,

IP
130(1), 123–135. https://doi.org/10.1037/bne0000110

CR
Hill, M. N., Hellemans, K. G. C., Verma, P., Gorzalka, B. B., & Weinberg, J. (2012).

Neurobiology of chronic mild stress: Parallels to major depression. Neuroscience and

US
Biobehavioral Reviews, 36(9), 2085–2117. https://doi.org/10.1016/j.neubiorev.2012.07.001
AN
Kinzig, K P, & Taylor, R. J. (2009). Maintenance on a ketogenic diet: voluntary exercise,

adiposity and neuroendocrine effects. International Journal of Obesity (2005), 33(8), 824–
M

830. https://doi.org/10.1038/ijo.2009.109
ED

Kinzig, Kimberly P., Honors, M. A., & Hargrave, S. L. (2010). Insulin sensitivity and glucose

tolerance are altered by maintenance on a ketogenic diet. Endocrinology, 151(7), 3105–

PT

3114. https://doi.org/10.1210/en.2010-0175
CE

Kinzig, Kimberly P, Scott, K. a, Hyun, J., Bi, S., & Moran, T. H. (2005). Altered hypothalamic

signaling and responses to food deprivation in rats fed a low-carbohydrate diet. Obes Res,
AC

13(10), 1672–1682. https://doi.org/10.1038/oby.2005.205

Kudielka, B. M., & Kirschbaum, C. (2005). Sex differences in HPA axis responses to stress: A

review. Biological Psychology, 69(1 SPEC. ISS.), 113–132.

https://doi.org/10.1016/j.biopsycho.2004.11.009

Lucassen, P. J., Pruessner, J., Sousa, N., Almeida, O. F. X., Marie, A., Rajkowska, G., … Czéh,

B. (2014). Neuropathology of stress, 109–135. https://doi.org/10.1007/s00401-013-1223-5

 ACCEPTED MANUSCRIPT

Luppino, Floriana S.; de Wit, Leonore M.; Bouvy, P. F. (2010). Overweight, Obesity, and

Depression, 67(3), 220–229. https://doi.org/10.1001/archgenpsychiatry.2010.2

Mattson, M. P. (2014). Challenging oneself intermittently to improve health. Dose-Response,

12(4), 600–618. https://doi.org/10.2203/dose-response.14-028.Mattson

Mauvais-Jarvis, F. (2015). Sex differences in metabolic homeostasis, diabetes, and obesity.

T
Biology of Sex Differences, 6(1), 1–9. https://doi.org/10.1186/s13293-015-0033-y

IP
McCormick, C. M., Smith, C., & Mathews, I. Z. (2008). Effects of chronic social stress in

CR
adolescence on anxiety and neuroendocrine response to mild stress in male and female rats.

Behavioural Brain Research, 187(2), 228–238. https://doi.org/10.1016/j.bbr.2007.09.005

US
McGonigle, P. (2014). Animal models of CNS disorders. Biochemical Pharmacology, 87(1),
AN
140–149. https://doi.org/10.1016/j.bcp.2013.06.016

Murphy, P., Likhodii, S. S., Hatamian, M., & Burnham, W. M. (2005). Effect of the ketogenic
M

diet on the activity level of wistar rats. Pediatric Research, 57(3), 353–357.
ED

https://doi.org/10.1203/01.PDR.0000150804.18038.79

Neal, E. G., Chaffe, H., Schwartz, R. H., Lawson, M. S., Edwards, N., Fitzsimmons, G., …
PT

Cross, J. H. (2008). The ketogenic diet for the treatment of childhood epilepsy: a
CE

randomised controlled trial. The Lancet Neurology, 7(6), 500–506.

https://doi.org/10.1016/S1474-4422(08)70092-9
AC

Nestler, E. J., Gould, E., & Manji, H. (2002). Preclinical models: Status of basic research in

depression. Biological Psychiatry. https://doi.org/10.1016/S0006-3223(02)01405-1

Paoli, A., Rubini, A., Volek, J. S., & Grimaldi, K. A. (2013). Beyond weight loss : a review of

the therapeutic uses of very-low-carbohydrate ( ketogenic ) diets. European Journal of

Clinical Nutrition, 67(8), 789–796. https://doi.org/10.1038/ejcn.2013.116

Rho, J. M. (2015). How does the ketogenic diet induce anti-seizure effects? Neuroscience
 ACCEPTED MANUSCRIPT

Letters, 637, 4–10. https://doi.org/10.1016/j.neulet.2015.07.034

Sorge, R. E., Martin, L. J., Isbester, K. A., Sotocinal, S. G., Rosen, S., Tuttle, A. H., … Mogil, J.

S. (2014). Olfactory exposure to males, including men, causes stress and related analgesia

in rodents. Nature Methods, 11(6), 629–632. https://doi.org/10.1038/nmeth.2935

Tannenbaum, B. M., Brindley, D. N., Tannenbaum, G. S., Dallman, M. F., McArthur, M. D., &

T
Meaney, M. J. (1997). High-fat feeding alters both basal and stress-induced hypothalamic-

IP
pituitary-adrenal activity in the rat. The American Journal of Physiology, 273(6 Pt 1),

CR
E1168-77. Retrieved from

http://ajpendo.physiology.org/content/273/6/E1168.full%5Cnhttp://www.ncbi.nlm.nih.gov/

US
pubmed/9435533
AN
Torres, S. J., & Nowson, C. A. (2007). Relationship between stress, eating behavior, and obesity.

Nutrition, 23(11–12), 887–894. https://doi.org/10.1016/j.nut.2007.08.008

M

Walsh, R. (2011). Lifestyle and mental health. American Psychologist, 66(7), 579–592.
ED

https://doi.org/10.1037/a0021769

Willner, P. (2016a). Reliability of the chronic mild stress model of depression: A user survey.
PT

Neurobiology of Stress, 6, 68–77. https://doi.org/10.1016/j.ynstr.2016.08.001

CE

Willner, P. (2016b). The chronic mild stress (CMS) model of depression: History, evaluation and

usage. Neurobiology of Stress, 6, 78–93. https://doi.org/10.1016/j.ynstr.2016.08.002

AC

Wohleb, E. S., Franklin, T., Iwata, M., & Duman, R. S. (2016). Integrating neuroimmune

systems in the neurobiology of depression. Nature Reviews Neuroscience, 17(8), 497–511.

https://doi.org/10.1038/nrn.2016.69

Youm, Y.-H., Nguyen, K. Y., Grant, R. W., Goldberg, E. L., Bodogai, M., Kim, D., … Dixit, V.

D. (2015). The ketone metabolite β-hydroxybutyrate blocks NLRP3 inflammasome-

mediated inflammatory disease. Nature Medicine, 21(3), 263–269.

 ACCEPTED MANUSCRIPT

https://doi.org/10.1038/nm.3804

Zemdegs, J., Quesseveur, G., Jarriault, D., Pénicaud, L., Fioramonti, X., & Guiard, B. P. (2016).

High-fat diet-induced metabolic disorders impairs 5-HT function and anxiety- like behavior

in mice. British Journal of Pharmacology, 2095–2110. https://doi.org/10.1111/bph.13343

T
IP
CR
US
AN
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Table 1: CMS Details

Stressor Description Duration

Restraint Restrained in a clear PVC device with a perforated front at an 30 minutes

unpredictable time during the light cycle.

T
IP
Lights Moved to a neighboring housing room in novel cages, where the 12 hours*

CR
lights remained on during the dark cycle.

US
Flooding 200 mL of room temp tap water added to bedding in novel cage 16 hours*

during the light cycle.

AN
M

Cold Moved to a novel cage and placed in a 4 degrees C chamber at an 30 min

unpredictable time during the light cycle.

ED
PT

Crowding Temporarily housed with 3-4 unfamiliar conspecifics also 6 hours*

undergoing CMS during the light cycle.

CE
AC

Static Static noise (80 dB) played in a neighboring housing room in the 6 hours*

middle of the light cycle.

Cage Tilt Moved to a novel cage during different times of the day tilted 30-45 6 hours*

degrees in home room.

*Note: Animals had access to food and water in conditions lasting 6 or more hours.
 ACCEPTED MANUSCRIPT

Table 2: Diet Compositions (% kcal)

T
Diet type Carbs % kcal Protein % kcal Fat % kcal Calorie Density Product ID

IP
Chow 58 24 18 3.1 kcal/g Teklad

CR
(CH) (2018)

Ketogenic 5 15 80
US 6.1 kcal/g Research Diets
AN
(KD) (D06040601)
M
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

Table 3: Primer Sequences and Thermal Profiles

Target Gene Primer Sequences NCBI Sequence

P T
Thermal Profile Efficiency

R I
CRH Forward: 5’ – CTC TCT GGA TCT CAC CTT CCA C- 3’

Reverse: 5’ - CTA AAT GCA GAA TCG TTT TGG C - 3’

S C
NM_031019.1 95°C/15s

60°C/15s
100.4%

N U 72°C/60s

A
Forward: 5’- TAT CCC TGC TCG TGT GTT TG- 3’M
NPY

E D
Reverse: 5’- TGT CGC AGA GCG GAG TAG TA - 3’
NM_012614 95°C/15s

58°C/15s
97.6%

P T 72°C/60s

C E
β -actin

A C
Forward: 5’ - ATT GGT GGC TCT ATC CTG GC - 3’

Reverse: 5’- AAA CGC AGC TCA GTA ACA GTC - 3’

NM_031144.3 95°C/15s

60°C/15s
90.4%

72°C/60s
 ACCEPTED MANUSCRIPT

T
IP
CR
US
Figure 1. Experimental Timeline. Rats were exposed to either ketogenic diet (KD) or chow (CH)
AN
diet for 4 days, then underwent three weeks of chronic mild stress (CMS). After CMS, animals

were tested for three consecutive days on elevated plus maze (EPM), open field (OFT), and
M

forced swim (FST). They were then sacrificed approximately two hours after FST for plasma and
ED

tissue analysis.
PT
CE
AC
 ACCEPTED MANUSCRIPT

T
IP
* CH-
P la s m a B H B ( m M )

CR
* * *
1 .0
CH+

US
KD-
0 .5

KD+
AN
0 .0
M

M a le F e m a le

Figure 2. Terminal plasma beta-hydroxybutyrate (βHB). βHB levels were normally distributed
ED

(Kruskal-Wallis test, χ2= 2.143, p= 0.143). Animals consuming KD had significantly higher
PT

plasma βHB after three weeks of diet exposure than those consuming CH (*: p < 0.0001). Values
CE

are means +/- SEM; n= 6-10 per group.

AC
 ACCEPTED MANUSCRIPT

A v e r a g e D a ily F o o d In t a k e (k c a l)

† CH-

T
100 CH+

IP
KD-
#
KD+

CR
50

US
0
M a le F e m a le
AN
Figure 3. Average daily food intake. Food intake was not normally distributed between sexes
M

(Kruskal-Wallis χ2= 50.918, †: p < 0.001). There were no significant changes in male groups.
ED

Females undergoing CMS has lower food intake, particularly between CH+ and KD- groups (#:

p= 0.042). Values are means +/- SEM; n= 7-10 per group.

PT
CE
AC
 ACCEPTED MANUSCRIPT

† CH-
60

B W C h a n g e (g )
CH+

KD-
40
KD+

20

T
# *

IP
0
M a le F e m a le

CR
Figure 4. Body weight change from start to end CMS. Body weight change was not normally

US
distributed by sex (Kruskal-Wallis χ2 = 7.1334, †: p= 0.008). CH+ males gained less weight than

CH- (#: p = 0.022), KD- (#: p < 0.001), and KD+ (#: p= 0.001). The same was observed within
AN
females. CH+ females gained less weight than CH- (*: p= 0.003), KD- (*: p= 0.001), and KD+
M

(*: p= 0.049). Values are means +/- SEM; n= 7-10 per group.
ED
PT
CE
AC
 ACCEPTED MANUSCRIPT

A. 80 # *

T im e in C lo s e d A r m s (% )
CH- B. 40

CH+

T im e i n C e n t e r ( % )
60
KD- 30

40 KD+
20

20
10

0 0

T
M a le F e m a le M a le F e m a le

IP
C.

CR
#
40 #
T im e in O p e n A r m s ( % )

30

US
20

10
AN
0
M a le F e m a le
M

Figure 5. Elevated Plus Maze (EPM) behavior. (A) Time spent in closed arms of apparatus was
ED

normally distributed between sexes (Kruskal-Wallis χ2= 0.799, p= 0.371). There was a
PT

significant effect of stress, driven by CH+ females (#: p= 0.034). (B) Time spent in center of

apparatus was normally distributed between sexes (Kruskal-Wallis χ2= 0.779, p= 0.377). There
CE

was a significant interaction between stress and diet. KD+ groups spent less time in the center of
AC

the apparatus than CH+ groups (*: p= 0.049). (C) Time spent in open arms of apparatus was also

normally distributed (Kruskal-Wallis χ2 = 1.615, p= 0.204). There was a significant effect of

stress. Tukey’s multiple comparison test suggests this was largely driven female stress groups (#:

p= 0.02). Values are means +/- SEM; n= 7-10 per group.

 ACCEPTED MANUSCRIPT

A. 20
B .. 8
CH-
D is t a n c e T r a v e le d ( m )

T im e in C e n t e r ( % )
CH+
15 6
KD-

10 KD+ 4

5 2

T
IP
0 0
M a le F e m a le M a le F e m a le

CR
US
C. 100
T im e in C o r n e r s ( % )

80
AN
60

40
M

20

0
ED

M a le F e m a le

Figure 6. Open field test (OFT) behavior. (A) Distance traveled was normally distributed
PT

between sexes (Kruskal-Wallis χ2= 0.287, p= 0.592). There were no significant effects of stress,
CE

diet, or sex. (B). Time spent in center was normally distributed (Kruskal-Wallis χ2= 3.198, p=

0.074. There were no effects of stress, diet, or sex. (C) Time spent in corners was also normally
AC

distributed (Kruskal-Wallis χ2= 1.059, p= 0.303). There were no significant effects of stress,

diet, or sex. Values are means +/- SEM; n= 7-10 per group.
 ACCEPTED MANUSCRIPT

B.
A. † 200

L a t e n c y to Im m o b ilit y ( s )
150
CH-
T im e Im m o b ile (s )

CH+ 150

100 KD-

KD+ 100

50
50

T
IP
0 0
M a le F e m a le M a le F e m a le

CR
C. 150 †

US
T im e E s c a p in g ( s )

100
AN
50
M

0
M a le F e m a le
ED

Figure 7: Forced swim test (FST) behavior. (A) Time spent immobile. Sex distributions were
PT

significantly different (Kruskal-Wallis χ2= 4.46, †: p= 0.034). There were no significant effects
CE

within male or female groups of stress or diet. (B) Latency to immobility. Sex distributions were

not significantly different (Kruskal-Wallis χ2= 0.717, p= 0.397). There were no significant
AC

effects of stress, diet, or sex observed. (C) Time spent escaping. Sex distributions were

significantly different (Kruskal-Wallis χ2= 6.699, †: p= 0.01). There were no significant effects

within male or female groups of stress or diet. Values are means +/- SEM; n= 7-10 per group.
 ACCEPTED MANUSCRIPT

200
†
C o r t ic o s t e r o n e ( n g /m l) CH-

150
# CH+

T
IP
100 KD-

CR
50 KD+

US
0
M a le AN F e m a le

Figure 8. Terminal plasma corticosterone (CORT). Data were not normally distributed across

sexes (Kruskal-Wallis χ2= 4.686, †: p= 0.03). Females undergoing stress while consuming chow
M

(CH+) had lower CORT than control females (CH-) (#: p= 0.038) and there were no changes
ED

observed between male groups. Values are means +/- SEM; n= 6-10 per group.
PT
CE
AC
 ACCEPTED MANUSCRIPT

15
CH-
IL - 1  ( n g /m L )

10
CH+

T
KD-

IP
5

CR
KD+

US
M a le F e m a le
AN
Figure 9. Final plasma IL-1β. Combine sex data were normally distributed (Kruskal-Wallis χ2=
M

3.553, p= 0.059). There were no significant effects of CMS exposure or exposure to KD in either
ED

males or females. Values are means +/- SEM; n= 7-10.

PT
CE
AC
 ACCEPTED MANUSCRIPT

A. 2 .5
†
CH-

N P Y ( 2 ^ -   C t)
2 .0 CH+

KD-
1 .5 #
KD+

1 .0

0 .5

T
IP
0 .0
M a le F e m a le

CR
B.
2 .5

US
C R H (2 ^ -   C t)

2 .0

1 .5
AN
1 .0
M

0 .5

0 .0
ED

M a le F e m a le

Figure 10. Hypothalamic neuropeptide Y (NPY) and corticotropin-releasing hormone (CRH)

PT

mRNA expression. Cycle thresholds (Ct) normalized to ß-actin, and ∆∆ Ct was normalized to
CE

chow control condition (CH-). (A) Male and female groups had different NPY expression

distributions (Kruskal-Wallis χ2= 3.927, †: p= 0.048), therefore sexes were analyzed separately.
AC

There were no significant differences within males. Within female analyses found CH+ group

had lower NPY expression than CH- (#: p= 0.045). (B) Male and female groups had normal CRH

distributions (Kruskal- Wallis χ2 = 0.106, p = 0.744). There were no significant effects of stress,

diet, or stress on CRH expression in the hypothalamus. Values are means +/- SEM; n = 7-10 per

group.
 ACCEPTED MANUSCRIPT

Highlights:
 Ketogenic diet does not induce adverse effects, or increase vulnerability to behavioral or
neuroendocrine effects of chronic mild stress
 Body weight differences in stressed animals consuming chow were not observed in
ketogenic diet consuming animals
 Sex-specific stress effects of chronic stress on chow consuming animals were not observed
in ketogenic diet consuming animals

T
IP
CR
US
AN
M
ED
PT
CE
AC