Supplementation with Combined Lactobacillus helveticus R0052 and Bifidobacterium longum R0175 Across Development Reveals Sex Differences in Physiological and Behavioural Effects of Western Diet in Long–Evans Rats

The gut microbiome affects various physiological and psychological processes in animals and humans, and environmental influences profoundly impact its composition. Disorders such as anxiety, obesity, and inflammation have been associated with certain microbiome compositions, which may be modulated in early life. In 62 Long–Evans rats, we characterised the effects of lifelong Bifidobacterium longum R0175 and Lactobacillus helveticus R0052 administration—along with Western diet exposure—on later anxiety, metabolic consequences, and inflammation. We found that the probiotic formulation altered specific anxiety-like behaviours in adulthood. We further show distinct sex differences in metabolic measures. In females, probiotic treatment increased calorie intake and leptin levels without affecting body weight. In males, the probiotic seemed to mitigate the effects of Western diet on adult weight gain and calorie intake, without altering leptin levels. The greatest inflammatory response was seen in male, Western-diet-exposed, and probiotic-treated rats, which may be related to levels of specific steroid hormones in these groups. These results suggest that early-life probiotic supplementation and diet exposure can have particular implications on adult health in a sex-dependent manner, and highlight the need for further studies to examine the health outcomes of probiotic treatment in both sexes.


Introduction
The microbes that inhabit the animal gut (i.e., gut microbiota) have been consistently reported to affect physiological and psychological functioning [1]. These microbes-especially bacteria-have important physiological functions in the host, including helping to regulate digestion, the immune system, central nervous system, and hormone levels (see [2] for a review). Indeed, the composition of the microbiota has been linked to various diseases (see [3] for a review) and is highly vulnerable to environmental exposures of the host (e.g., antibiotic use, infection) [4]. These environmental experiences can alter the microbiome in a way that promotes disease development (i.e., dysbiosis) [4]. Importantly, For breeding, two groups, each containing four naïve breeding pairs, were placed in separate colony rooms that would eventually become the probiotic or placebo administration rooms. The colony rooms were virtually identical in appearance and setup. Breeding in both rooms was conducted at the same time, and involved placing each of the eight pairs of rats together in a standard housing cage for seven days. After this time, the female was assumed to be pregnant, and the male was removed from the cage and housed alone until sacrifice. Females were pair-housed after breeding to minimise isolation stress until estimated gestational day 17 when they were given a fresh cage to await parturition. Females were weighed bi-weekly during pregnancy, and seven of eight bred females gave birth after 21 gestational days, with one female in the placebo-designated room not producing a litter. Within 24 h of parturition, females and their litters were transferred into large cages with a burrow section (semi-naturalistic housing, SNH, Figure 2A) [36]. Female breeders were kept with their pups until weaning at postnatal day (PD) 21. For breeding, two groups, each containing four naïve breeding pairs, were placed in separate colony rooms that would eventually become the probiotic or placebo administration rooms. The colony rooms were virtually identical in appearance and setup. Breeding in both rooms was conducted at the same time, and involved placing each of the eight pairs of rats together in a standard housing cage for seven days. After this time, the female was assumed to be pregnant, and the male was removed from the cage and housed alone until sacrifice. Females were pair-housed after breeding to minimise isolation stress until estimated gestational day 17 when they were given a fresh cage to await parturition. Females were weighed bi-weekly during pregnancy, and seven of eight bred females gave birth after 21 gestational days, with one female in the placebo-designated room not producing a litter. Within 24 h of parturition, females and their litters were transferred into large cages with a burrow section (semi-naturalistic housing, SNH, Figure 2A) [36]. Female breeders were kept with their pups until weaning at postnatal day (PD) 21. A total of 46 pups were live-born (1 still-born) to four probiotic-treated dams with an additional 4 pups not surviving to weaning. In the placebo group, 33 pups were live-born (2 still-born) to three dams with an additional 3 pups not surviving to weaning. Due to the discrepancy in sample sizes (i.e., 42 probiotic offspring, 30 placebo offspring), pups in the probiotic group were culled to 32 offspring (i.e., 10 culled and used for molecular assay optimisation), which left 8 pups in each diet condition with an equal sex ratio. Overall, for this experiment, breeding produced 62 offspring for eight experimental offspring groups separated by treatment (32 probiotic, 30 placebo), by diet condition (31 Western diet, 31 control diet), and by sex (30 males, 32 females). Offspring were housed in same-sex pairs in groups of two or three in standard housing cages as described. Ethically, to reduce the number of offspring produced for this experiment, eight dams were bred, and offspring were sorted into the eight experimental groups randomly but in a way that maximised diversity. Due to litter sex ratios and ethical responsibility to pair house or house in threes, all litters could not always be represented in each group, but complete litter statistics are summarised in Table 1. Sample size was chosen from previous similar experiments by our laboratory group [36] and kept as small as possible to reduce the number of animals bred for experimental purposes. A total of 46 pups were live-born (1 still-born) to four probiotic-treated dams with an additional 4 pups not surviving to weaning. In the placebo group, 33 pups were live-born (2 still-born) to three dams with an additional 3 pups not surviving to weaning. Due to the discrepancy in sample sizes (i.e., 42 probiotic offspring, 30 placebo offspring), pups in the probiotic group were culled to 32 offspring (i.e., 10 culled and used for molecular assay optimisation), which left 8 pups in each diet condition with an equal sex ratio. Overall, for this experiment, breeding produced 62 offspring for eight experimental offspring groups separated by treatment (32 probiotic, 30 placebo), by diet condition (31 Western diet, 31 control diet), and by sex (30 males, 32 females). Offspring were housed in same-sex pairs in groups of two or three in standard housing cages as described. Ethically, to reduce the number of offspring produced for this experiment, eight dams were bred, and offspring were sorted into the eight experimental groups randomly but in a way that maximised diversity. Due to litter sex ratios and ethical responsibility to pair house or house in threes, all litters could not always be represented in each group, but complete litter statistics are summarised in Table 1. Sample size was chosen from previous similar experiments by our laboratory group [36] and kept as small as possible to reduce the number of animals bred for experimental purposes. Table 1. Litter diversity details for each of the eight experimental groups including total sample size, maximum subjects per litter in each group, and represented litters of total by treatment (four probiotic, three placebo). PL: placebo; PR: probiotic; WD: Western diet; CD: control diet; M: male; F: female.

Probiotic Administration
In this experiment, each dam and her offspring were placed into the same treatment group and received probiotic (CEREBIOME ® ) or its placebo. For all animals, the daily probiotic dosage was 1 billion colony forming units (CFUs) per mL, per day (i.e., 0.05 g of probiotic powder per 0.5 mL of reverse osmosis water, based on the CFU/g value for this specific batch of probiotic). An equivalent amount of placebo (i.e., the probiotic vehicle that contains malic acid, xylitol, and maltodextrin as excipients) solution was prepared (i.e., 0.05 g of placebo powder per 0.5 mL of reverse osmosis water). Solutions were made fresh each morning and maintained at 4 • C ± 4 • C until administration, according to the protocol developed in [37]. For both dams and offspring, probiotic or placebo were delivered via syringe feeding during the dark phase of the light-dark cycle (i.e., 11:00 ± 1 h) as per the protocol developed by Tillmann and Wegener [38] for adult male rats. Briefly, the syringe feeding protocol involves a training period that varies by age and sex of the rat, whereby the rat is lap-fed by an experimenter. After successful training (three to four days for adult male rats), rats will voluntarily approach the syringe and feed from the cage.
There is evidence suggesting that probiotics are transferred through the act of breastfeeding [39,40], and there are further reports that infant microbial gut colonisation may begin in utero [41]. Thus, although this experiment focused on offspring outcomes, solutions were delivered daily to the dam during pregnancy and lactation to maximise any potential benefits for the offspring. To expedite syringe training with the female breeders and to not interfere with breeding, females were trained by adding a 0.25 M sucrose solution (i.e., 0.04 g of sucrose in 0.5 mL of reverse osmosis water) to the solution for four days prior to breeding. Four of the eight females voluntarily took the syringe with sucrose-sweetened probiotic or placebo solution after the four days. After breeding, probiotic and placebo solution were delivered without added sucrose, and all females voluntarily fed from the cage at that time. Solutions were delivered to the offspring daily from PD 22 until day of sacrifice (PD 71 to 74; i.e., for equal to or greater than seven weeks). As the probiotic vehicle contained a polysaccharide food additive that was slightly sweet, sucrose was not added to the offspring probiotic and placebo solutions during training. All offspring learned to take the solution voluntarily from the cage after 27 days.

Diet Manipulation
From weaning, all rats were fed a specific chow and provided water ad libitum until sacrifice. Diets were purchased from Research Diets (New Brunswick, NJ, USA) and comprised equivalent protein matched by the supplier and a similar number of kilocalories. Offspring rats in both the probiotic and placebo groups received either control diet (product #D14042701; carbohydrate 73% kcal, fat 10% kcal, protein 17% kcal, kcal/g = 3.9) or Western diet (product #D12079B; carbohydrate 43% kcal, fat 40% kcal, protein 17% kcal, kcal/g = 4.7). In the control diet, the main sources of carbohydrates were corn starch followed by maltodextrin, whereas they were sucrose, maltodextrin, and corn starch in the Western diet. The main source of fat for both diets was anhydrous milk fat, but the Western diet contained more anhydrous milk fat relative to the control diet; they contained equal amounts of corn oil. Protein content, vitamins, minerals, and added fibre were identical between the two diets, but the Western diet contained cholesterol, whereas the control diet did not.

Weight and Food Intake
Offspring rats were weighed bi-weekly from weaning until sacrifice to monitor for any changes in health status and also at an average age of PD 67 to obtain a final adult weight. All birth, wean, and adult weights are provided in grams (g). Additionally, when offspring reached adulthood, food intake was measured for five consecutive days between PD 63 and PD 70. Briefly, a 24 h ± 1 h change in food weight for each cage was determined and divided by the number of rats in the cage. A 4-day average amount of food consumed (in kcal, by multiplying each average food weight in grams by the kcal/g values above) was calculated from the change in food across each of the five measurement days, and differences by treatment, diet, or sex were determined.

Anxiety-Like Behaviours
To assess anxiety-like behaviour, all offspring were tested for 5 min in each of the open field test (OFT; between PD 60 and PD 61) and light-dark box (LDB; between PD 62 and PD 63). For both tests, rearing and line crosses were scored in real-time, while remaining behaviours were scored later from video. Behavioural apparatuses were cleaned with 70% ethanol between rats and before the first rat was tested each day. The testing room was lit by a 60 W red light during both tests.
Behaviours measured included line crosses (between equal squares marked by white Fisherbrand TM labelling tape (Thermo Fisher Scientific, Whitby, ON, Canada), total rearing, supported rearing, unsupported rearing, transitions between light and dark areas (in the LDB), time in centre (in the OFT; not including initial time taken to exit centre), time in light (in the LDB), latency to move from centre (in the OFT), and latency to move from the dark area (in the LDB). As described by Kalueff and Tuohimaa [42], increased time spent by a subject in (or transitions into) more anxiety-inducing areas (i.e., the centre area in the OFT; the light section in the LDB) was interpreted as showing less anxiety. Additionally, more time taken to leave a more-anxiety-inducing area when initially placed there (e.g., centre of OFT) is indicative of higher anxiety, whereas less time taken to voluntarily enter a more-anxiety-inducing area (e.g., light section of LDB when initially placed in the dark section) is indicative of lower anxiety [42]. Increased locomotion (e.g., line crosses) is usually regarded as a measure of lower anxiety in behavioural tests such as the OFT and LDB [42], but should be interpreted with caution when weight varies between tested groups [43]. Rearing frequency is another commonly examined behaviour, which may indicate more exploration (i.e., low anxiety) but can also vary due to overall locomotor ability, which is influenced by more than just anxiety (e.g., by weight) [44]. It has recently been suggested that rearing always be separated out by supported (locomotion measure to be interpreted cautiously) and unsupported (exploratory and low-anxiety behaviour), as these are distinct behavioural variables [45].

Open Field Test
The OFT was constructed of black plexiglass (80 cm long × 80 cm wide × 35 cm high) with no lid ( Figure 2B). Animals were initially placed in the centre of the arena [46,47], facing a consistent side of the testing apparatus.

Light-Dark Box
The LDB ( Figure 2C; overall 60 cm long × 30 cm wide × 45 cm high, 6 cm × 6 cm middle passage door; adapted from designs in [48,49]) had a 50/50 light/dark split where the dark box was constructed to fit into half of an overall apparatus constructed of clear plexiglass. The dark box was constructed of clear plexiglass on one side (to allow for video and live scoring) and opaque black plexiglass on the remaining three sides and hinged lid. At the start of testing, all rats were placed in the centre of the dark side, facing the opening to the light side. A 60 W white light was mounted 45 cm above the light section of the LDB, which was illuminated during testing.

Predator Odour Exposure, Sacrifice, and Tissue Collection
Between PD 71 and PD 74, animals were sacrificed 90 min following a 5-min acute predator odour stressor comprised of exposure to a cotton swab containing 0.5 mL of pure cat urine in a metal tea strainer hung from the side of the apparatus in the light section of the LDB ( Figure 2C). This modified LDB was cleaned with 70% ethanol between each exposure and before the first exposure of the day. All surgical equipment was cleaned with 70% ethanol before the first sacrifice of the day and between animals. Testing, sacrifices, and dissections from the placebo group were always performed prior to the probiotic group.
All 62 rats were anaesthetised with Euthanyl (sodium pentobarbital) via intraperitoneal injection and, once fully unconscious, were quickly decapitated. Brains were removed and immediately placed on a plexiglass plate over dry ice, and the hippocampus and hypothalamus were gross dissected from a coronal section. Adrenal glands, colon, caecum, small intestine, stomach, fat, liver, and spleen were collected and immediately placed on dry ice until they were stored at −80 • C. The presence or absence of both L. helveticus R0052 and B. longum R0175 in all animals was confirmed in caecum contents collected at sacrifice, as previously described [37]. Briefly, we confirmed by qPCR that, on the day caecum was collected, CEREBIOME ® (previously known as Probio'Stick ® ) was detected in all probiotic-treated animals and not detected in any placebo-treated animals [37].

Plasma Leptin Enzyme-Linked Immunosorbent Assay (ELISA)
For the detection of leptin concentration (pg/mL) in plasma, samples were previously prepared, as described in Section 2.7. Two leptin rat ELISA kits (ThermoFisher Scientific, Waltham, MA, USA, cat. #KRC2281, lot #218111-003) were counterbalanced by group type and for location on the plate. Samples were assayed in duplicate and diluted to 1:11 based on previous dilution optimisation. It was decided that samples with greater than 20% coefficient of variation (CV) would be redone, but no samples were assayed with greater than 20% CV. Interassay and intraassay variability are reported by the manufacturer as 4.6% and 3.6%, respectively.

Statistical Analyses
All statistical analyses were conducted using IBM SPSS Statistics (Version 25; IBM Corporation, Armonk, NY, USA) software. For behavioural analyses, outliers were removed from analyses if they were greater or below three standard deviations (SD) from the mean (M). There were no outliers found for adult weight or food intake variables. For plasma leptin levels, one extreme value just greater than 3 SDs from the mean (a placebo male exposed to Western diet) was noted, but the value remained in analyses because it represented a possible biological value on the standard curve with a CV of 0.87% between duplicates. For inflammation, outliers were removed from analyses if values were greater or less than 3 SDs from the mean for each analyte individually. Values below the standard curve were reported as the minimum for the standard curve for each analyte. Of the 23 analytes tested, only those deemed measurable were reported, with measurable being defined as 80% or greater of the subjects tested falling within the standard curve for that analyte.
All behavioural, adult weight, food intake, leptin, and inflammation dependent variables were analysed using 2 × 2 × 2 (treatment by diet by sex) factorial analyses of variance (ANOVAs). Results from all ANOVA analyses are presented with descriptive statistics (i.e., M and SD) provided for all measures, with the goal of conveying the amount of variability of data from the mean. Wean weight was analysed using a 2 × 2 (treatment by sex) factorial ANOVA, as offspring were not exposed to diet until after weaning. Further 2 × 2 × 2 ANOVAs requiring a co-variate (ANCOVAs) were conducted on all behavioural measures (controlling for adult weight). All 2-way and 3-way interactions were analysed post hoc by simple effects analyses. While the alpha level was chosen to be 0.05 and main effects and interactions are presented when results are below this threshold, due to the large number of ANOVAs being conducted in this study, it should be noted that p-values between 0.01 and 0.05 should be interpreted with caution as the possibility of type I error is increased with multiple comparisons. Correlations were analysed by Spearman's rho (r s ) also at an alpha level of 0.05 with p-values between 0.01 and 0.05 also interpreted with caution. Effect sizes were reported as partial eta squared (η p 2 ), see [50,51] for considerations when using η p 2 for power analyses and meta-analyses.

Anxiety-Like Behaviours
A series of treatment by diet by sex factorial ANOVAs were conducted to analyse the six behavioural variables recorded from the OFT. Overall, there were no main effects of diet on any of these behavioural measures, but there were specific main effects of treatment and sex. First, for total rearing, a main effect of treatment was found (p = 0.002), in that placebo rats reared more overall compared to probiotic rats (Table 2). Specifically, there was a main effect of treatment for supported rearing (p < 0.001), whereby placebo rats performed more supported rears than probiotic rats (Table 2). No main effects of treatment were found for line crosses, unsupported rearing, time in centre, or latency from centre. Next, a main effect of sex was found for line crosses (p < 0.001), in that females crossed more lines than males (Table 2). Two main effects of sex were found for total rearing, in that females showed more overall rearing (p = 0.046) and supported rearing (p = 0.041) than males ( Table 2). No sex differences were found for unsupported rearing, time in centre, or latency from centre. Another series of treatment by diet by sex factorial ANOVAs were conducted to analyse the seven behavioural variables recorded from the LDB. Unsupported rearing in the LDB was not reported because the behaviour was too infrequently performed, but it is included in total rearing. Similar to OFT results, there were no main effects of diet on any of these behavioural measures, but there were main effects for treatment and sex, along with an interaction effect for the supported rearing behaviour. First, a main effect of treatment was found for supported rearing (p = 0.013), in that placebo rats engaged in more supported rears than probiotic rats and a main effect of treatment for transitions (p = 0.002) was found in that probiotic rats transitioned more between the light and dark compartments compared to placebo rats (Table 3). There were no main effects of treatment for line crosses, total rearing, time in light, or latency to light. Next, main effects of sex were found for line crosses (p < 0.001), supported rearing (p = 0.003), transitions (p < 0.001), and latency to light (p < 0.001), whereby females demonstrated more line crosses, more supported rearing, more transitions, and took less time to enter the light compartment of the LDB (Table 3). There were no main effects of sex for total rearing or time in light. Table 3. Overview of treatment and sex main effects in the light-dark box from factorial ANOVA analyses for line crosses, supported rearing, transitions, and latency to enter light, including means (M) and standard deviations (SD).

Behaviour, df
Probiotic ANOVA analyses from LDB data revealed a treatment by sex interaction for supported rearing (F 1, 54 = 4.10, p = 0.048, η p 2 = 0.070). Simple effects analyses showed that placebo males performed more supported rears compared to probiotic males (p = 0.002; Figure 3A). Furthermore, the probiotic females performed more supported rears than the probiotic males (p = 0.001; Figure 3A). A further treatment by diet by sex interaction was found for supported rearing (F 1, 54 = 5.62, p = 0.021, η p 2 = 0.094), and simple effects analyses showed that placebo Western males performed more supported rears compared to probiotic Western males (p = 0.008; Figure 3B). As well, probiotic Western females performed more supported rears compared to probiotic Western males (p = 0.004; Figure 3B). With control diet females, those given the placebo reared more with support than those given the probiotic (p = 0.036), and with placebo females, those given the control diet reared more with support compared to those given the Western diet (p = 0.022). Additionally, females reared more than males in both placebo control animals (p = 0.019) and probiotic control animals (p = 0.036).

Wean and Adult Weights
A treatment by sex factorial ANOVA showed that wean weight was lower in offspring from mothers consuming probiotic during lactation (treatment main effect;

Wean and Adult Weights
A treatment by sex factorial ANOVA showed that wean weight was lower in offspring from mothers consuming probiotic during lactation (treatment main effect; F1, 58

Wean and Adult Weights
A treatment by sex factorial ANOVA showed that wean weight was lower in offspring from mothers consuming probiotic during lactation (treatment main effect; F1, 58    From the treatment by diet by sex factorial ANOVA on adult weight, a sex by treatment interaction was found (F 1, 54 = 5.83, p = 0.019, η p 2 = 0.097). Simple effects analyses showed that males in the placebo group weighed more in adulthood compared to males in the probiotic group (p < 0.001; Figure 4B), whereas no difference by treatment or diet was seen with female rats. A sex by diet by treatment interaction was also found (F 1, 54 = 8.85, p = 0.004, η p 2 = 0.141), where male rats in the Western diet condition weighed more in the placebo group than in the probiotic group (p < 0.001; Figure 4B). Furthermore, analyses of this three-way interaction showed that of males in the placebo group, those in the Western diet condition weighed more than males in the control diet condition (p = 0.004; Figure 4B).

Average Daily Calorie Intake
Food intake was measured as average amount of calories consumed, calculated from grams of food eaten (see Section 2.4 above). It is noteworthy that adult weight was positively correlated with daily calorie intake (r s = 0.87, p < 0.001, N = 62). of the treatment by sex interaction, it was first found that males in the placebo group consumed more daily calories compared to males in the probiotic group (p = 0.002; Figure 5A). On the contrary, probiotic females consumed more calories compared to placebo females (p < 0.001; Figure 5B). Analysis of the two-way diet by sex interaction also showed that in males only, animals given the Western diet consumed more calories compared to control diet animals (p < 0.001; Figure 5A).
group weighed more in adulthood compared to males in the probiotic group (p < 0.001; Figure 4B), whereas no difference by treatment or diet was seen with female rats. A sex by diet by treatment interaction was also found (F1, 54 = 8.85, p = 0.004, ηp 2 = 0.141), where male rats in the Western diet condition weighed more in the placebo group than in the probiotic group (p < 0.001; Figure 4B). Furthermore, analyses of this three-way interaction showed that of males in the placebo group, those in the Western diet condition weighed more than males in the control diet condition (p = 0.004; Figure  4B).

Average Daily Calorie Intake
Food intake was measured as average amount of calories consumed, calculated from grams of food eaten (see Section 2.4 above). It is noteworthy that adult weight was positively correlated with daily calorie intake (rs = 0.87, p < 0.001, N = 62). A treatment by diet by sex factorial ANOVA revealed a main effect of diet (F1, 54 Figure 5A). On the contrary, probiotic females consumed more calories compared to placebo females (p < 0.001; Figure 5B). Analysis of the two-way diet by sex interaction also showed that in males only, animals given the Western diet consumed more calories compared to control diet animals (p < 0.001; Figure 5A). A treatment by diet by sex interaction was found for daily calories consumed (F1, 54 = 11.44, p = 0.001, ηp 2 = 0.175) and simple effects analyses showed that in Western-diet-exposed males, placebo- A treatment by diet by sex interaction was found for daily calories consumed (F 1, 54 = 11.44, p = 0.001, η p 2 = 0.175) and simple effects analyses showed that in Western-diet-exposed males, placebo-treated animals consumed more calories compared to probiotic-treated animals (p < 0.001, Figure 5A). Additionally, within placebo-exposed males, Western diet rats consumed more calories than control diet rats (p < 0.001, Figure 5A). In females given the Western diet, probiotic animals consumed more calories compared to placebo animals (p < 0.001; Figure 5B).

Plasma Leptin
For leptin concentration in plasma, while the treatment by diet by sex factorial ANOVA did not reveal main effects of treatment or diet, there was a main effect of sex (F 1, 54 = 30.11, p < 0.001, analyses revealed that probiotic females had higher leptin levels than placebo females (p = 0.036; Figure 6). Simple effects analyses further revealed that males in the Western diet group had higher leptin levels than males in the control diet group (p = 0.010; Figure 6).

Plasma Leptin
For leptin concentration in plasma, while the treatment by diet by sex factorial ANOVA did not eveal main effects of treatment or diet, there was a main effect of sex (F1, 54  Simple effects analyses evealed that probiotic females had higher leptin levels than placebo females (p = 0.036; Figure 6). imple effects analyses further revealed that males in the Western diet group had higher leptin levels han males in the control diet group (p = 0.010; Figure 6). Plasma leptin levels were positively correlated with adult weight (r s = 0.77, p < 0.001, N = 62; Figure 7A) and daily calorie intake (r s = 0.75, p < 0.001, N = 62; Figure 7B). Figure 6. Plasma leptin levels (pg/mL) in males and females by treatment and diet condition. Data expressed as M(SD). * p < 0.05, ** p < 0.01, *** p < 0.001. PL: placebo; PR: probiotic; WD: Western diet; CD: control diet.
Plasma leptin levels were positively correlated with adult weight (rs = 0.77, p < 0.001, N = 62; Figure 7A) and daily calorie intake (rs = 0.75, p < 0.001, N = 62; Figure 7B).   Due to main effects and interaction effects of treatment, diet, and sex on adult weight, along with correlations between weight and certain behaviours, a series of ANCOVAs on all dependent behavioural variables in the OFT were conducted. The main effects of treatment for total rearing (F1, 53 = 8.53, p = 0.005, ηp 2 = 0.139) and supported rearing (F1, 53 = 12.73, p < 0.001, ηp 2 = 0.197) remained, with placebo-treated rats rearing more overall and performing more supported rears than probiotictreated rats. Females still performed more line crosses than males (sex main effect; F1, 53 = 7.32, p = 0.009, ηp 2 = 0.121), but females no longer performed more total rears (F1, 53  In the LDB, similar to the OFT, there was a negative correlation between line crosses and adult weight (r s = −0.29, p = 0.020, N = 62; Figure 8B) that was no longer present when examining the correlations by treatment group or by sex. Transitions in the LDB were higher in animals that weighed less (r s = −0.38, p = 0.002, N = 62). The negative correlation between transitions and adult weight was not present when only examining placebo animals (r s = −0.22, p = 0.243, n = 30; Figure 9A) or by examining results by sex, but was present with probiotic animals (r s = −0.49, p = 0.005, n = 32; Figure 9B). As well, overall, heavier animals in adulthood took longer to enter the light area (r s = 0.43, p < 0.001, n = 61). There was no correlation between latency to light and weight in placebo rats only (r s = 0.37, p = 0.052, n = 29) or by examining results by sex, but there was a positive relationship in probiotic rats (r s = 0.48, p = 0.006, n = 32). There were no other correlations for the remaining LDB behaviours and weight found, but upon examining all variables broken down by sex, a positive association was found in males between supported rearing and adult weight (r s = 0.52, p = 0.003, n = 30), showing that more supported rears were related to higher weight in males only. ANCOVAs were also conducted for all LDB behavioural measures to examine differences in these behaviours after controlling for adult weight. Results showed that there was no longer a main effect of treatment on supported rearing (F1, 53

Discussion
The goal of this work was to increase knowledge on the health effects of L. helveticus R0052 and B. longum R0175 (i.e., CEREBIOME ® ) using a developmental model. To summarise, two tests of anxiety, the OFT and LDB, were used in the present study with consistent results. After controlling for adult weight, placebo-treated rats showed evidence of higher anxiety [42,45,49,61]. Again, after factoring in weight, females showed evidence of lower anxiety [42]. This is the first animal study to our knowledge that has examined the metabolic consequences of CEREBIOME ® , and our metabolic results suggest drastically different weight and food intake patterns in male and female rats with respect to probiotic treatment and diet. In females, treatment with the probiotic resulted in higher leptin levels compared to placebo and increased calorie intake, especially with Western diet exposure, without the expected change in weight. In males, the probiotic counteracted the increase in daily calories consumed and adult weight that was seen in the placebo-treated animals given the Western diet, with no corresponding change in leptin levels by treatment (i.e., leptin levels were higher only with Western diet exposure). It is also important to note that wean weight was lower in the offspring of dams given probiotic compared to placebo, even with no difference in birth weight, suggesting that the probiotic could be modulating weight during lactation either directly (i.e., the probiotic is being transferred to the pups through the breast milk) or indirectly (i.e., the pups are receiving probiotic from being in the same environment as their mother). Finally, approximately 90 min after rats were exposed to acute predator odour stress, specific inflammatory analytes, seemingly independent of analyte category, were higher in those given CEREBIOME ® , in those treated with a Western diet, and in males.
The fact that placebo-treated rats reared more with support than probiotic-treated rats is noteworthy, as probiotic treatment is reported to improve anxiety in animal studies [18,24] and decrease the stress response [15,16,62,63]. Rearing is a complex behaviour (especially supported rearing) [45], and should be interpreted with caution as the act of performing a supported rear can vary based on the state of the animal performing the behaviour (i.e., escape behaviour or exploratory behaviour) [61]. From our results, it can only be theorised that the placebo animals may be performing more supported rears due to a desire to escape the testing area. This idea is strengthened by the fact that probiotic-treated animals transitioned more in the LDB (a behaviour consistent with lower anxiety) and reveals that multiple means for estimating anxiety levels is ideal. Future studies would benefit from additional measured variables such as number of defecations to help distinguish between exploratory-versus escape-related rearing [64].
From previous literature describing the effects of diet on the stress response system [27,65], it is of interest that we did not see any effect of diet on our anxiety measures. This is consistent with Abildgaard and colleagues [62], who also found no differences in the OFT by probiotic treatment or by diet, but they solely measured locomotion. It may be that results are affected by the type of diet, length of stress exposure, and choice of behavioural tests. One important note is that a high-carbohydrate control diet such as that used here leads to insulin resistance, excess fat storage, and weight gain [66], and low-carbohydrate diets are used when treating insulin resistance and type 2 diabetes [67]. High-carbohydrate diets are defined as containing greater than 45% of their energy from carbohydrates [67]-markedly lower than commonly used control diets, which are frequently comprised of 70% carbohydrates. In fact, Prasad and Prasad [68] report that rats given a 90% high-fat diet for one week had decreased observed anxiety on the elevated plus maze (EPM) test compared to baseline when compared to a 90% carbohydrate and 90% protein diet. Additionally, while Bridgewater et al. [46] found that male mice (but not females) given a 60% high-fat diet showed more anxiety-like behaviours in the OFT and EPM when also exposed to a chronic unpredictable stress paradigm. Importantly, the study by Bridgewater and colleagues [46] obtained similar results to the present study, in that male and female mice responded uniquely with respect to their observed anxiety.
Lactobacillus and Bifidobacterium strains have been documented as agents that can aid in the prevention and reversal of obesity via their impact on modulating the gut microbiome (i.e., affecting energy balance/storage, food intake hormones, and the lining of the gut) [69]. Previous research has shown that weight gain and other markers of obesity are improved with probiotic treatment (e.g., Bifidobacterium strains [20]; L. helveticus R0052 [22]). At least for males, our results suggest that probiotic treatment may lessen the negative health impacts of the Western diet. Our results are consistent with literature characterising the effects of probiotics on weight gain in response to unhealthy diet (e.g., in mice [22]; in zebrafish [70]). Although sex-specific effects of high-fat diet have been documented [46], there is a stark lack of studies examining how the combination of diet and probiotic treatment might affect males compared to females. For instance, Karlsson et al. [19] reported lower weight gain, fat around organs, and plasma leptin with Lactiplantibacillus plantarum treatment, but results were not reported by sex. In a recent study, individuals with major depressive disorder were supplemented with the CEREBIOME ® formulation, a galactooligosaccharide prebiotic, or placebo, and it was found that only the probiotic was associated with decreased reported depressive symptoms on the Beck Depression Inventory [71]. Of interest was that this probiotic was related to an increase in the ratio of tryptophan to isoleucine (a branched-chain amino acid, BCAA), with no change in overall tryptophan levels, suggesting that the probiotic may be related to decreased levels of isoleucine [71]. Lee et al. [72] recently found that overall BCAA levels in plasma (i.e., sum of valine, leucine, and isoleucine) was associated with metabolic dysfunction and diabetes, so further studies that aim to characterise the metabolic effects of CEREBIOME ® would benefit from including a measure of overall or specific BCAAs.
Because leptin is reported to inhibit food intake [73], it is of interest that the female rats given probiotic had both increased leptin and increased calorie intake compared to those given placebo. There is evidence of sex differences in gut microbiota compositions. For instance, Tennoune et al. [74] report Escherichia coli K12 DNA presence at baseline in female rat feces only. Interestingly, the group further reports increased weight in females after administration of this same strain but both decreased weight and decreased food intake in males [74]. As Taraschenko and colleagues [75] suggested, evidence shows that diet-induced obesity occurs differently in each sex, highlighting the need for studies to report on results in both sexes when examining the anti-obesity effects of probiotic treatment. One study that aimed to correct high-fat diet-induced obesity was only successful with male rats [75]. Furthermore, Alonso-Caraballo and colleagues [76] found that increased anxiety-like behaviours in the EPM were associated with increased weight gain and plasma leptin in male rats only. Interestingly, when Harris et al. [77] induced leptin resistance (by administering high-fat diet and leptin injections) females still lost weight in response to the leptin administration with no change in food intake (i.e., they did not become leptin resistant like males) [77]. The results from Harris et al. [77] combined with the present results may indicate a time-delay in the development of leptin resistance in female rats compared to male rats. As Gruzdeva et al. [78] describe in their review, increased leptin can reduce appetite and body weight to a certain extent, but if the body becomes resistant to leptin as seen with obesity, increased leptin signals are no longer effective in decreasing food intake. Although leptin resistance is difficult to characterise and the mechanisms underlying leptin resistance have not been fully elucidated [78], the male rats in the present study had greater plasma leptin levels compared to females, which may be indicative of leptin resistance. Thus, if females were not in a leptin resistance state, higher leptin levels in the probiotic-treated compared to the placebo-treated females may indicate a more metabolically healthy outcome at the time leptin was measured.
Our results from inflammatory marker analysis are intriguing as we expected that probiotic treatment would both decrease pro-inflammatory markers and increase anti-inflammatory markers. In fact, probiotic animals displayed higher levels of pro-inflammatory analyte, IL-1α, but also of the anti-inflammatory analyte, IL-10, and the regulatory analytes, M-CSF and MIP-3α. Prior studies have shown that probiotic treatment decreases pro-inflammatory analytes and factors in response to chronic stress (e.g., IL-6 in rats [23]; IL-1β and NF-κB in Syrian golden hamsters [24]). Furthermore, Dai and colleagues [79] administered a probiotic including Bifidobacteria, Lactobacilli, and one Streptococcus species in a rat model of colitis and found that treatment similarly induced production of IL-10 (anti-inflammatory), but concurrently, reduced production of pro-inflammatory analytes (e.g., TNF-α and IL-6 in colon and serum). In addition, Bisson et al. [80] showed that a probiotic given to rats-containing L. helveticus R0052, B. longum R0175, Lacticaseibacillus rhamnosus R0011 and Saccharomyces boulardii yeasts-reduced pro-inflammatory analytes in serum (e.g., IL-1α, IL-6) and induced production of IL-4 and IL-10 after Escherichia coli infection. However, research is conflicted with respect to the effects of probiotic treatment on the inflammatory response, especially after stress exposure. In rats, Ait-Belgnaoui et al. [15] found no changes in plasma pro-inflammatory cytokines (i.e., IL-1β, IL-6) after probiotic (Companilactobacillus farciminis) treatment or acute stress exposure (partial restraint stress), but did find lower levels of hippocampal inflammatory analytes (IL-1β, IL-6, TNF-α mRNA) and stress hormones after stress exposure (i.e., CRF mRNA, plasma corticosterone and adrenocorticotropic hormone, ACTH) with the probiotic treatment. Conversely, in response to a multi-species probiotic treatment with prior stress exposure, Abilgaard et al. [62] found that mononuclear cells isolated from rat blood produced a greater amount of the cytokines IFN-γ, IL-2 and IL-4 when stimulated. As well, mice that were immune challenged and then exposed to an acute footshock stressor had increased glucocorticoid release and reduced IL-1α and IL-1β in the brain [81].
A complex relationship exists between the interleukin-1 family and the HPA axis where low levels of IL-1 (e.g., IL-1α and IL-1β) might be beneficial for efficient coping with stressors, but under severe or chronic stress conditions, increased IL-1 might be an adaptive method in reducing the stress response and preventing widespread damage [82]. In fact, one study that used the same combination of L. helveticus R0052 and B. longum R0175 as the present study found that, in rats, the probiotic prevented increased stress hormone release (i.e., corticosterone, noradrenaline, adrenaline) in plasma and prevented downregulation of glucocorticoid receptors, seen in response to chronic stress [16]. Wagner and Johnson [83] reported that cells from a vaginal epithelial cell line infected with Candida albicans and exposed to Lacticaseibacillus rhamnosus GR-1®and Limosilactobacillus reuteri RC-14®responded by increasing expression of IL-1α and IL-1β mRNA. Taken together, if the placebo animals had a greater physiological stress response to the potent predator odour stress (i.e., severe acute stress), this may explain the relatively lower amounts of plasma IL-1α measured in placebo compared to probiotic animals. Although research is severely lacking on the effects of stress and probiotic treatment on less-studied immunological analytes such as M-CSF and MIP-3α, it would not be surprising if these analytes are also affected by the stress response.
Our findings that the Western diet-exposed animals had increased inflammatory marker release is interesting when considering the relationship between glucocorticoids (steroid hormones) and the immune response. Glucocorticoids, when administered therapeutically, can modulate the immune system in response to autoimmune diseases and organ transplantation but can also promote insulin resistance and diabetes development (see [84] for a commentary). Furthermore, Wang et al. [85] showed that animals on a high-sugar diet had increased plasma corticosterone levels. Taking into consideration the research reported above, increased carbohydrate intake from the control diet may have led to increased presence of stress hormones, which in turn, could help to explain a lower inflammatory response in our control diet group.
We noted a number of sex differences in plasma analyte levels that may be explained, at least in part, by levels of circulating sex steroid hormones, and, is of importance to further studies characterising the effects of probiotic treatment in both sexes. Pyter et al. [86] reported that chronic stress exposure led to increased hippocampal inflammatory marker expression in male, but not female rats, after an immune challenge with lipopolysaccharide. Although there were no sex differences in corticosterone levels after the immune challenge, female rats had increased plasma estradiol, which may have prevented hippocampal expression of genes for the inflammatory analytes measured (e.g., IL-1β, TNF-α) [86]. Both testosterone (indirectly) and estradiol (directly) affect the HPA axis related to their actions on androgen and estrogen receptors, respectively [87] and males naturally have lower circulating levels of estradiol compared to females [88]. Furthermore, in response to stress, a recent review reports that female rats have increased corticosterone and ACTH compared to males [87], which may also suppress the inflammatory response as described above and may be further affected by circulating levels of estradiol. Theoretically, increased estrogens and/or increased corticosterone release may affect the inflammatory response in females and warrants investigation in studies aiming to elucidate sex differences in the mechanistic link between stress and inflammation.

Conclusions
We present evidence that the probiotic-treated animals showed less anxiety in the behavioural testing arenas used in this study (i.e., the OFT and LDB). Future studies would benefit from combining further anxiety behavioural tests (e.g., the EPM) with the aim of increasing the duration and complexity of testing in order to characterise the more subtle effects of probiotic treatment (e.g., to increase the incidence of lower-frequency behaviours). Importantly, by examining behavioural results by sex while factoring in weight, we show that weight is not the only factor affecting the differences in the observed anxiety-like behaviours seen between males and females. With the drastically different pattern of results between males and females seen for our metabolic measures (i.e., adult weight, food intake, and plasma leptin), it becomes important for future studies to include sex as a factor in the study design. This is the first study to our knowledge that examined inflammatory levels after acute exposure to predator odour stress with CEREBIOME ® treatment. These results provide evidence that male, probiotic-treated, and Western-diet-administered rats have a greater inflammatory response after acute stress, which may be related to increased levels of specific steroid hormones and warrants further study (e.g., estradiol in males, corticosterone in placebo-treated and control diet-fed animals).
Overall, this study adds to what we know about a combination L. helveticus R0052 and B. longum R0175 probiotic with respect to adult anxiety behaviours, weight, food intake, and plasma levels of leptin and inflammatory analytes in a developmental Long-Evans rat model also exposed to a control or Western diet. The examination of these factors in the same group of animals is important, as all of the abovementioned health outcomes have been shown to be affected by probiotic treatment and each other. Importantly, our results demonstrate specific benefits of this probiotic that vary by sex, which is a critical consideration for developmental animal studies, which often examine effects in males exclusively. Over and above the specific findings in this study, it is our hope that results demonstrate the importance of including both sexes in animal studies that examine health outcomes and call attention to the idea of appropriate caution when choosing a control diet for a high-fat or Western diet. In future studies examining the effects of Western diet exposure, it would be enlightening to compare to the rat chow traditionally provided by animal care facilities rather than a high-carbohydrate control diet.