- Email: [email protected]
Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline
Journal Pre-proofs Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline Kyoko Hasebe, Leni R Rivera, Craig M Smith, Theo Allnutt, Tamsyn Crowley, Tiffanie M Nelson, Olivia M Dean, Sean L McGee, Ken Walder, Laura Gray PII: DOI: Reference:
S0889-1591(18)31241-8 https://doi.org/10.1016/j.bbi.2019.09.001 YBRBI 3848
To appear in:
Brain, Behavior, and Immunity
Received Date: Revised Date: Accepted Date:
17 January 2019 28 August 2019 3 September 2019
Please cite this article as: Hasebe, K., Rivera, L.R., Smith, C.M., Allnutt, T., Crowley, T., Nelson, T.M., Dean, O.M., McGee, S.L., Walder, K., Gray, L., Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline, Brain, Behavior, and Immunity (2019), doi: https://doi.org/10.1016/j.bbi.2019.09.001
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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.
© 2019 Published by Elsevier Inc.
Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline.
Kyoko Hasebea, Leni R Riveraa, Craig M Smithab, Theo Allnuttd, Tamsyn Crowleya, Tiffanie M Nelsonef, Olivia M Deanbc, Sean L McGeea, Ken Waldera, Laura Grayab
aSchool
of Medicine, Centre for Molecular and Medical Research, Deakin University,
Geelong, Australia; bThe Florey Institute of Neuroscience and Mental Health, Parkville, Australia; cSchool of Medicine, IMPACT Strategic Research Centre, Deakin University, Geelong, Australia; dTheo Allnutt Bioinformatics, Geelong, Australia; eGeelong Centre for Emerging Infectious Diseases, Geelong, Victoria, 3220, Australia; fDeakin University, Geelong, Victoria, 3220, Australia.
Corresponding author: Dr. Laura Gray, School of Medicine, Centre for Molecular and Medical Research, Deakin University, Geelong, Australia; [email protected]
Word count: 6208 words
Abstract An emerging novel therapeutic agent for major depressive disorder, minocycline, has the potential to influence both gut microbiome and inflammatory status. The present study showed that chronic high fat diet feeding led to changes in both behaviour and the gut microbiome in male mice, without an overt inflammatory response. The diet-induced behavioural changes were characterised as increased immobility in the forced swim test and 1
changes in locomotor activities in the open field test. Minocycline significantly altered the gut microbiome, rendering a community distinctly different to both untreated healthy and diet-affected states. In contrast, minocycline did not reverse high fat diet-induced changes in behaviour.
Key words High fat diet feeding, minocycline, gut microbiome
2
1. Introduction Emerging therapies for major depressive disorder (MDD), such as the promising agent minocycline (Rosenblat and McIntyre, 2018), have the potential to influence both the gut microbiome composition and inflammatory status. Given our developing understanding of how changes in the gut microbiome, often driven by diet, can influence mood and behaviour, it is critical to investigate how this nexus of intersecting factors can be exploited by novel therapies with strong potential for impact on patient outcomes. There is now significant evidence to support an association between diet quality and the risk of developing MDD (Jacka et al., 2014; Rahe et al., 2015; Yu et al., 2014). Although the development of MDD can alter dietary patterns, longitudinal studies confirm that poor diet quality often precedes MDD onset, suggesting that diet can be a causal factor (Akbaraly et al., 2009; Rienks et al., 2013; Sanchez-Villegas et al., 2009; Sanchez-Villegas et al., 2012). Furthermore, dietary patterns modulate inflammatory status. Several studies reported that healthy dietary patterns lower levels of inflammatory mediators, such as C-reactive protein (CRP) (Neale et al., 2016; Schwingshackl and Hoffmann, 2014) and Interleukin-6 (IL-6) (Schwingshackl and Hoffmann, 2014). Due to the established link between inflammation and mood (Cepeda et al., 2016; Kubera et al., 2013; Remus and Dantzer, 2016; Vichaya et al., 2018), poor diet quality could therefore induce elevations of pro-inflammatory mediators, which subsequently may predispose to a depressed mood. Changes in gut microbiota are now considered to be mechanistically important to the interaction between poor diet and mood. Alterations in gut microbial communities induced by poor diets have been reported in multiple studies (Brown et al., 2012; de La Serre et al., 2010; Turnbaugh et al., 2008; Turnbaugh et al., 2009). A significant proportion of the gut microbial population are capable of modulating the host inflammatory response, via key 3
molecules including bacterial lipopolysaccharides (LPS), which are major outer surface membrane components present in almost all gram-negative bacteria (Alexander and Rietschel, 2001). An increase in permeability of the intestinal barrier (tight junctions) and resultant bacterial translocation to the systemic circulation (often referred to as a “leaky gut”), is also implicated in the pathophysiology of MDD (Garate et al., 2011; Kelly et al., 2015; Maes et al., 2012; Slyepchenko et al., 2017). Experimental administration of LPS in rodents induces a suite of behavioural symptoms sufficiently akin to depressive states, which is considered an animal model of MDD (Kubera et al., 2013; Remus and Dantzer, 2016). Several studies have suggested that changes in the gut microbiome may be an initiating or causal factor in the development of mood changes. For instance, transplantation of the faecal microbiome from patients with MDD into germ-free mice resulted in an increase in depressive-like behaviours relative to mice transplanted with microbiota from healthy control individuals (Zheng et al., 2016). Furthermore, transplantation of the cecum and colonic microbiome from mice fed a high fat diet induced changes in anxiety-like behaviours in recipient mice (Bruce-Keller et al., 2015). These findings suggest the gut microbial community may play a causal role in the pathogenesis of depressive-like behaviour, and that diet may be a key factor in driving changes in the microbiome, and thus mood. Minocycline (MINO) is a broad spectrum antibiotic which has been investigated for its potential antidepressant effects. Although MINO exerts direct effects on the gut microbial community when administered orally, several reports suggest the potential antidepressant effects are primarily caused by pharmacological effects within the brain, via modulation of neuroinflammation and suppression of microglial activity (Dean et al., 2012; Garrido-Mesa et al., 2013; Henry et al., 2008; Zheng et al., 2015). MINO ameliorates the depressive-like 4
behaviour observed in various animal models of depression (Henry et al., 2008; Liu et al., 2015; Majidi et al., 2016). For example, Wong et al. (Wong et al., 2016) reported that chronic restraint stress-induced depressive-like behaviour in mice was reversed by oral MINO, and this was associated with changes in gut microbiota. Adjunctive MINO treatment also alleviated depressive symptoms in human MDD patients (Dean et al., 2017; Husain et al., 2017; Miyaoka et al., 2012). This study therefore investigated the effects of high fat diet-induced changes in gut microbiota and behaviour in mice, and on inflammatory and microbial signalling pathways in the periphery and brain. In addition, the present study first showed minocycline-induced changes in the gut microbiome were not associated with reversals of behaviour induced by high fat diet.
2. Materials and methods 2.1 Animals and diet All experiments were performed in accordance with the guidelines of Deakin University Animal Ethics Committee and the National Health and Medical Research Council. Four-week old male C57BL/6J mice were purchased from Animal Resources Centre (Perth, Western Australia, Australia). Animals were group housed (four per cage) and maintained on a 12h light/dark cycle (7am – 7pm) at 20-22 °C with ad libitum access to food and water. After two weeks of acclimatisation, animals were assigned into two groups: high fat diet (HFD), and standard chow (STD), and were maintained on each respective diet until the end of the study. Groups were matched at this baseline time point across body weight, fat mass and lean mass measured by an EchoMRI Whole Body Composition Analyzer. The STD group was fed with standard chow (BARASTOC Rat and Mouse Cubes, Ridley AgriProducts, Victoria, 5
Australia) that contains 5% kcal of fat and 20% kcal of protein. The HFD group was fed with high fat diet (SF04-001, Speciality feeds, Western Australia, Australia) that contains 45% kcal of fat (45% of available energy in the form of fat) and 22.6% kcal of protein. Detailed dietary compositions of standard chow and high fat diet are shown in Table 1. Body weight was measured once per week throughout of the study. Mice were 24 weeks old by the time of behavioural assessments. Table 1. Nutritional content of standard chow and high fat diets Standard chow diet
High fat diet
Barastoc, Mice and Rodent cube
Specialty Feeds SF04-001
Protein
20.0%
22.6%
Total Fat
5.0%
23.5%
Crude Fibre
5.0%
5.4%
12.6MJ/Kg
19MJ/Kg
Digestible Energy
2.2 Experimental design and timeline The Experimental design and timeline used for this study is shown in Figure 1.
Figure 1. Schematic diagram of experimental design and timeline.
2.3 Sacrifice and tissue collection 6
The mice of each group were humanely killed by cervical dislocation, the brain removed, and the hypothalamus and bilateral whole hippocampal formations dissected. Hypothalamic and hippocampal tissues were snap frozen in liquid nitrogen and stored at -80 °C. Plasma was collected from trunk blood and stored at -80 °C for later use
2.4 Drugs Minocycline hydrochloride (MINO) (PCCA, Houston, USA) was prepared in 0.9% saline at concentrations of 10 mg/ml for dosing at 50 mg/kg of MINO per day. The dose of MINO was in line with other studies (Henry et al., 2008; McKim et al., 2016; Zheng et al., 2015) and a pilot study conducted in our laboratory (unpublished data). Although body weights were relatively stable after 13 weeks of HFD or STD, dosages were re-adjusted weekly to ensure they were aligned with body weights, and solutions were pH-matched to that of the vehicle. MINO and saline were orally gavaged daily for a 4 week period, which began 13 weeks after assigning HFD and STD groups.
2.5 Behavioural testing All behavioural testing was conducted between 10:00 am and 14:00 pm and was recorded and analysed using Noldus EthoVision XT8 (Noldus Information Technology, The Netherlands).
2.6 Open field test The open field test (OFT) was used to measure locomotion and exploratory behaviour in a novel environment. The apparatus comprised a clear plexiglass box (50 cm x 50 cm, with 36 cm high walls). The light intensity was low in the examination room and there was no 7
difference in the light intensity between the centre and the outer zones. Mice were placed in a corner of the area, and allowed to move freely for 6 min and filmed from above. Time spent in the centre, travel distance and mobility were analysed using a video tracking system (Ethovision XT 8.5, Noldus Information Technology, The Netherlands).
2.7 Porsolt forced swim test The Porsolt forced swim test (FST) displays predictive validity for testing the efficacy of antidepressants (Krishnan and Nestler, 2011). Mice were placed individually in a glass cylinder (diameter: 18.5 cm; height: 26.5 cm) filled with 25± 1 °C water to a height of 13.5 cm for 6 min. Mice were filmed from the side, and the duration of immobility and latency to first immobility were scored manually by a blinded experimenter, where immobility (i.e. the Porsolt posture) was defined as immobility in all four limbs.
2.8 Sucrose preference test The sucrose preference test (SPT) was used to measure anhedonic behaviour (Krishnan and Nestler, 2011). To avoid stress associated with social isolation, mice were housed 2 per cage and were given free access to two water bottles, one with tap water and another with 1 % sucrose solution. Mice were split from 4/box to 2/box at 5 pm, and returned to 4/box at 9 am. Both pairs were placed in new boxes, and given time (approximate 15 minutes) to acclimatise before the addition of sucrose. At the conclusion of the test, bottles were removed and consumptions of sucrose solution and tap water were measured by weighing each bottle, which was conducted by a blinded experimenter.
8
2.9 16S metagenomic sequencing Faecal samples were collected during the final week of behavioural testing. DNA was isolated using QIAamp DNA stool kits (Qiagen Pty Ltd, Victoria, Australia), modified to include a bead-beating step. The extracted DNA was quantified and purity was assessed using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA). Sequencing was performed at Macrogen Inc (Seoul, South Korea). The extracted DNA was prepared for sequencing on the Illumina MiSeq platform using the Illumina 16S Metagenomic Sequencing Library Preparation protocol. Primers targeting the 16S rRNA gene’s V3-V4 region were amplified from DNA. Primer sequences included Illumina’s adaptor sequence (shown in lowercase): 341F, 5'-tcgtcggcagcgtcagatgtgtataagagacagCCTACGGGNGGCWGCAG-3’ and 805R, 5'-gtctcgtgggctcggagatgtgtataagagacaggactachvGGGTATCTAATCC-3’ (Klindworth et al., 2013). The PCR product was purified using AMPure XP beads. Amplicons were indexed using the Illumina Nextera XT Index kit, and purified again using AMPure XP beads prior to a fluorometric quantification method.
2.10 Bioinformatics Sequenced amplicons were filtered to a minimum read length of 300 base pairs (bp), resulting in an average read length of 100 bp. The UPARSE pipeline for illumine paired reads was conducted (Edgar, 2013) with a 3% cluster radius. The RDP 16S database v16 was used to taxonomically classify operational taxonomic units (OTUs) with a 90% confidence threshold. OTUs with abundance less than 0.5% of the total number of reads or occurrence of less than two across all samples were removed from the analysis. QIIME 1.8 (Caporaso et al., 2010) was used to calculate diversity metrics and statistically compare alpha diversity (on rarefied data) and OTU abundance between species. OTU abundance was cumulative 9
sum scaling (CSS) normalised prior to statistical tests (Paulson et al., 2013). All commands and associated python scripts in the analysis are provided in the GitHub repository: https://github.com/theo-allnutt-bioinformatics/Hasebe_2019. Principal coordinates analysis (PCoA) was performed on weighted UniFrac distances (Lozupone and Knight, 2005) calculated from a maximum likelihood tree (FastTree v2.3.1) (Price et al., 2010) of the filtered OTUs on raw counts and CSS normalised counts. Beta diversity was analysed by Analysis of Molecular Variance (AMOVA) (Excoffier et al., 1992).
2.11 Analyses of inflammatory markers Plasma levels of inflammatory markers were assessed using commercially available ELISA kits: lipopolysaccharide binding protein, LBP (Boster Bio, mouse LBP ELISA kit), soluble CD14, IL6 and TNFα (R&D systems, Quantikine ELISA kits).
2.12 Real time quantitative PCR analysis Hypothalamic and hippocampal tissues were then homogenised and processed for total RNA extraction using RNeasy mini Kits (Qiagen Pty Ltd, Victoria, USA). RNA was quantified and purity assessed using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA). Complimentary strand DNA was synthesised using Dynamo cDNA synthesis kit (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction (qPCR) was performed using Taqman Gene Expression Assays (Applied Biosystems, Foster City, CA, USA) for ionized calcium-binding adapter molecule 1 (Iba1) (Mm00479862_g1), Glial fibrillary acidic protein (gfap) (Mm01253033_m1), Major histocompatibility complex) class II molecules (MHC II) (Mm658576_m1), cluster of differentiation molecule 11b (CD11b) (Mm00434455_m1),
Glutamate
aspartate 10
transporter
(Mm00600697_m1)(GLAST),
Excitatory amino acid transporter (GLT1) (Mm00441457_m1), Cysteine-glutamate antiporter (xCT) (Mm00442530_m1), inducible Nitric oxide synthase (iNOS) (Mm00440502_m1), Occludin (Mm00500912_m1) and 18S rRNA (Mm03928990_g1). Each sample was amplified in triplicate. Relative quantification was calculated by the ΔΔCt method and compared with endogenous levels of 18S rRNA.
2.13 Statistical analyses All behavioural and biochemical graphs were generated and data analysed using Prism (GraphPad Software, Inc. La Jolla, CA, USA). Mean ± standard errors of the mean are displayed. Normality of each dataset was assessed by Kolmogorov-Smirnov test. Two-way Analysis of Variance (ANOVA) was conducted to examine main effects of diet and treatment and interaction effects of diet and treatment, followed by Bonferroni post hoc comparisons to
determine
between
and
within
group
effects.
For
alpha
diversity
and
Firmicutes/Bacteroidetes ratio, Kruskal-Wallis test followed by Dunn’s multiple test was used to determine statistical significance. A weighted distance-based analysis of molecular variance (AMOVA) was used to assess spatial distance of each treatment group observed based on the weighted UniFrac distances in each treatment group in principal coordinate analysis (PCoA). For gene expression analyses, differences in levels of target genes were assessed by unpaired t tests. Statistical significance was set for all analyses at p < 0.05.
11
3. Results 3.1 High fat diet increased adiposity while MINO did not affect body weight or composition HFD fed mice were significantly heavier than STD fed mice at the conclusion of the study, and this effect was similar within both the vehicle and MINO treated groups (Two-way ANOVA; main effect of diet, F (1,68) = 178.2, p < 0.0001; Figure 2A, Supplemental Figure 1). HFD induced an increase in fat mass, in both vehicle and MINO treated mice (main effect of diet, F (1,68) = 325.0, p < 0.0001; Figure 2B). In contrast, lean mass was not affected by HFD (main effect of diet, F
(1,68)
= 3.08, p > 0.05). Furthermore, MINO treatment did not alter
body weight (Two-way ANOVA; main effect of treatment, F (1,68) = 0.67, p > 0.05), fat mass (main effect of treatment, F (1,68) = 0.44, p > 0.05) or lean mass (main effect of treatment, F (1,68)
= 0.006, p > 0.05). In line with these observations, MINO and HFD did not interact to
affect body weight (diet x drug treatment interaction, F (1,68) = 0.1, p > 0.05), fat mass (diet x drug treatment interaction, F
(1,68)
= 2.04, p > 0.05) or lean mass (diet x drug treatment
interaction, F (1,68) = 2.31, p < 0.05).
12
Figure 2. Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle, on: (A) body weights; (B) fat mass, and; (C) lean mass. Data are mean ± SEM. n = 18-22/mice per group. 2-way ANOVA, main effect of diet #### (p<0.0001). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.2 High fat diet mediated behavioural changes and the influence of MINO 3.2.1 High fat diet increased forced swim test immobility, which was not altered by MINO HFD fed mice spent significantly more time immobile in the FST (main effect of diet, F (1, 76) = 12.11, p < 0.001), compared to STD mice (Figure 3A). There was no main effect of MINO treatment (F (1, 76) = 2.54, p > 0.05; diet x drug treatment interaction, F (1, 75) = 0.18, p > 0.05).
13
Because increased immobility might conceivably result from increased fat mass (rather than diet-induced neurological/behavioural changes), a correlation analysis between fat mass and the duration of immobility was performed which included mice from the HFD groups only. However, a linear regression model indicated that fat mass was not significantly associated with the time spent immobile in HFD fed mice (F (1, 38) = 0.11, p > 0.05; Figure 3B). The increased immobility also did not correlate with body weight (Supplemental figure 3(A)).
3.2.2 High fat diet or MINO treatment did not affect sucrose preference Sucrose preference of HFD mice did not differ from STD mice, and there was no effect of MINO treatment (main effect of diet, F
(1, 36)
= 3.96, p > 0.05; main effect of drug, F
2.06, p > 0.05; diet x drug interaction, F (1, 36) = 0.48, p > 0.05; Figure 3C).
14
(1, 36)
=
Figure 3. Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle, on: (A) Duration of immobility in the Porsolt forced swim test; (B) Correlation between duration of immobility and fat mass; and (C) Sucrose preference (%); Sucrose consumption divided by total fluid intake (sucrose and water consumption overnight); Data are mean ± SEM. n = 1822/mice per group (for Sucrose preference test, n = 9-11/mice per group). 2-way ANOVA, main effect of diet ### (p<0.001) and # (p <0.05). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.2.3 MINO altered time spent in the centre of the open field test
15
MINO treatment significantly increased the time spent in the centre zone in the OFT overall (Two-way ANOVA; main effect of drug treatment, F
(1,76)
= 10.37, p < 0.01; Figure 4A), and
this effect appears to be most noticeable within HFD fed mice. In contrast, diet did not affect the time spent in the centre region (main effect of diet, F (1,76) = 0.08, p > 0.05; diet x drug treatment interaction, F
(1,76)
= 1.81, p > 0.05). Two-way ANOVA indicated that MINO
treatment did not affect the distance travelled (main effect of drug treatment, F (1,76) = 0.84, p > 0.05; Figure 4B), suggesting that the increased time that MINO treated mice spent in the centre region (see above) was not simply due to hyperactivity.
3.2.4 High fat diet altered locomotor activity patterns, which were not influenced by MINO HFD mice travelled significantly less distance compared to STD controls (main effect of diet, F
(1,76)
= 7.83, p < 0.01). This effect was similar within both vehicle and MINO treatment
groups (diet x drug treatment interaction, F (1,76) = 0.63, p > 0.05). Although fat mass can potentially alter patterns of locomotor activity, correlation analysis of mice from both HFD groups using a linear regression model indicated that distance travelled and fat mass were not associated (F (1,38) = 0.50, p > 0.05; Figure 4C). The reduced distance travelled in OFT was also not correlated with body weight (Supplemental figure 3 (B)). Duration of mobility was defined as the total amount of time that the animal was moving in the field during the OFT. MINO treatment did not alter duration of mobility relative to vehicle controls (Two-way ANOVA; main effect of treatment, F (1,76) = 2.02, p > 0.05; diet x drug interaction, F (1,76) = 0.19, p > 0.05; Figure 4D). HFD mice were significantly less mobile compared to STD controls overall (main effect of diet, F (1,76) = 19.52, p < 0.0001). Interestingly, HFD mice moved significantly faster than STD mice on average during each movement episode compared to STD mice control overall (Two-way ANOVA; main effect of 16
diet, F
(1,76)
= 19.41, p < 0.0001; Figure 4E). Combined with the finding that HFD mice
displayed reduced duration of mobility and distance travelled, this suggests that HFD mice engaged in a different pattern of locomotor behaviour, which was associated with fast running bouts that were shorter and/or less frequent than the slower yet more consistent locomotor activity of STD controls. MINO treatment did not affect locomotor activity speed (main effect of treatment, F (1,76) = 2.89, p > 0.05; diet x drug interaction, F (1,76) = 0.33, p > 0.05). A linear regression model showed no association between locomotor activity speed and fat mass within HFD fed mice (F (1,38) = 1.59, p > 0.05; Figure 4F). In addition, there was no association between locomotor activity speed and body weight (Supplemental figure 3 (C)).
17
Figure 4 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on the open field test: (A) Time spent in the centre; (B) Travel distance; (C) Correlation between distance travelled and fat mass; and (D) Duration spent mobile; (E) Speed; and (F) Correlation between speed and fat mass. Data are mean ± SEM. n = 18-22/mice per group. 2-way 18
ANOVA, main effect of diet #### (p < 0.0001) and ## (p <0.01); main effect of treatment $$ (p <0.01). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.3 High fat diet and MINO mediated alteration in gut microbiota 3.3.1 Principal coordinate analysis (PCoA) Principal coordinate analysis (PCoA) was used to identify whether treatment groups cluster together based on gut microbial community members and their abundances. The PCoA plot was constructed with the weighted UniFrac distance matrix, which incorporates the relative abundance of gut microbiota and their phylogenetic history based on gene sequence similarity (Figure 5A). The gut microbiota in STD/Vehicle, STD/MINO, HFD/Vehicle and HFD/MINO clustered distinctly from each other. HFD (PC1) explained 45.3% of the variation and MINO treatment (PC2) explained 15.6% of the variation between groups. A weighted distance-based analysis of molecular variance (AMOVA) was used to assess the statistical significance of the spatial separation observed between treatment groups. Statistically significant dissimilarities were observed between STD/Veh, HFD/Veh, STD/MINO and HFD/MINO groups with respect to bacterial diversity (p = 0.001).
3.3.2 Relative abundance at phylum level Figure 5B shows the relative abundance of bacterial phyla in each treatment group. At the phylum level, it appeared that Proteobacteria and Baceteroidetes were substantially reduced with a HFD and replaced by Verrucomicrobia. Following from MINO treatment, STD/MINO group was subsequently replaced with Candidatus-Saccharibacteria (also known as TM7) and Firmicutes, while HFD/MINO group was replaced with Actinobacteria. 19
3.3.3 Alpha diversity (Chao1 richness index) Alpha diversity is a measurement to assess structure and function of microbial diversity within a community (a sample) which accounts for the number of species, commonness and rareness of abundant bacteria (Lozupone and Knight, 2008). A Kruskal-Wallis test followed by Dunn’s multiple comparison test indicated that the Chao 1 richness index suggested a significant reduction of richness in the gut microbial diversity in HFD/Veh mice compared with STD/Veh mice (p < 0.05; Figure 5C). MINO treatment also contributed to a significant reduction of diversity of the gut microbiota that was observed within both STD (p < 0.001) and HFD mice (p < 0.01).
3.3.4 Firmicutes to Bacteroidetes (F/B) ratio Bacteroidetes and Firmicutes are two dominant phyla of the gut microbiota. Firmicutes to Bacteroidetes (F/B) ratio (Figure 5D) is used to assess differences in the main composition of the gut microbiota. A Kruskal-Wallis test followed by Dunn’s multiple comparison test indicated that HFD significantly increased the F/B ratio compared to the gut microbiota in STD/Veh mice (p < 0.001).
20
Figure 5 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on gut microbial community: (A) Principal coordinate analysis; (B) Relative abundance at phylum level; (C) Chao1 richness index; and (D) Firmicutes/Bacteroidetes ratio. n = 9-17/mice per group. Kruskal-Wallis test followed by Dunn’s multiple test was used for Chao1 richness index and F/B ratio. Upper and lower quartiles defined the box with median midline, and the whiskers were assessed using Tukey’s method. ***p < 0.001; **p < 0.01; *p < 0.05.
3.4 Effects of High fat diet and MINO on inflammatory mediators in the periphery Two-way ANOVA indicated that HFD significantly increased plasma levels of LBP (main effect of diet, F (1,67) = 4.70, p < 0.05), compared to STD fed mice (Figure 6A). In addition, MINO 21
treatment significantly decreased plasma LBP levels (main effect of treatment, F (1,67) = 5.73, p < 0.05; diet x drug interaction, F (1,67) = 0.18, p > 0.05). On the other hand, neither HFD nor MINO treatment affected the levels of soluble CD14 in plasma (Two-way ANOVA; main effect of diet, F (1,46) = 0.003, p > 0.05; main effect of treatment, F (1, 46) = 0.038, p > 0.05; diet x drug interaction, F (1,46) = 0.28, p > 0.05; Figure 6B). The levels of IL6 and TNFα in the periphery were under detection level (data not shown).
3.5 Effects of MINO on inflammatory mediators in the CNS As markers of glial activation, levels of Iba1 and GFAP in the hippocampus and the hypothalamus were assessed using real time QPCR. In the hypothalamus, two-way ANOVA indicated that MINO treatment significantly reduced levels of GFAP expression (main effect of treatment, F (1,31) = 8.76, p < 0.01; Figure 6C). The effect was independent from HFD (main effect of diet, F (1,31) = 2.89, p > 0.05; diet x drug interaction, F (1,31) = 0.62, p > 0.05). Levels of Iba 1 expression did not alter with either HFD or MINO in the hypothalamus (Two-way ANOVA; main effect of diet, F
(1,32)
0.05; diet x drug interaction, F
= 0.07, p > 0.05; main effect of drug F
(1,32)
(1,32)
= 0.70, p >
= 0.02, p > 0.05). In the hippocampus, levels of Iba1
expression also did not alter with either HFD nor MINO (Two-way ANOVA; main effect of diet, F
(1, 34)
= 0.27, p > 0.05; main effect of treatment, F
(1,34)
= 0.83, p > 0.05; diet x drug
interaction, F (1,34) = 0.91, p > 0.05), and nor did GFAP levels (Two-way ANOVA; main effect of diet, F (1, 34) = 0.28; main effect of drug, F (1,34) = 0.04, p > 0.05; diet x drug interaction, F (1,34)
= 0.24, p > 0.05). Levels of MHC II, CD11b, GLAST, GLT1, xCT, iNOS and Occludin did not
differ between groups (data not shown).
22
Figure 6 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on: (A) Plasma LBP levels. ; (B) Plasma soluble CD14 levels; C) Gene expression levels of glial activation markers 23
in the hippocampus and hypothalamus. Data are mean ± SEM. n = 7-11/mice per group. 2way ANOVA, main effect of diet # (p <0.05) and main effect of treatment $$ (p <0.01) and $ (p <0.05).
24
4. Discussion This study examined the effect of MINO on the behavioural changes, inflammatory state and gut microbiome alterations elicited by a HFD. HFD induced behavioural changes was paralleled by significant shifts in the population of the gut microbiome. This was accompanied by elevations of LBP, which might point to the way the immune system responds to microbiome changes and the means by which such signals are linked to changes in the brain. HFD mice displayed several behavioural changes relative to STD controls, such as significantly longer times spent immobile in the FST, however these changes were not influenced by MINO treatment. MINO did however result in changes in the gut microbial community.
4.1 HFD induced behavioural changes The HFD mice exhibited an increase in immobility time in the FST, which is consistent with other studies (Sharma et al., 2013; Yamada et al., 2011). Other studies, however, demonstrated less conclusive effects of a high fat diet (Andre et al., 2014; Krishna et al., 2016; Pyndt Jorgensen et al., 2014; Takase et al., 2016), which is possibly due to the large variability in the proportion and type of fat, timing and duration of feeding, animal age and sample size. For instance, the age of animals and the timing of HFD feeding in our study aligns with the Sharma et al. (2013) and Yamada et al. (2011) studies. This consistency of timing and age among the three studies may indicate that there may be a vulnerable age window by which HFD can mediate specific depressive-like behaviours. For the SPT, although a trend existed for our HFD mice to display less sucrose preference compared with STD mice, this did not reach statistical significance, which does not support the presence of anhedonia.
Potential interpretations of sucrose preference data as 25
correlating with anhedonia need to be considered cautiously however, as chronic HFD feeding may alter taste perception and reward. Anhedonia, or reduced capacity to experience hedonic pleasure in response to previously rewarding stimuli, is one of the core symptoms of MDD (Krishnan and Nestler, 2011). Anhedonic responses to a broad range of stimuli (ranging from social to gustatory and others) are often present in MDD patients, which suggests persistent dysregulation of common or final downstream reward systems such as the mesolimbic dopaminergic pathway, irrespective of the upstream stimulus. However, previous taste exposures are able to specifically influence the perception and rewarding palatability of subsequent taste stimuli, without altering other non-gustatory rewards more generally. For example, prior exposure to sweet tastes can habituate animals such that subsequent sweet stimuli are less rewarding (De Luca, 2014). Furthermore, internal homeostatic conditions are powerful modulators of appetitive behaviour, taste perception and reward, and this variable was also likely to be different between STD and HFD mice. Interestingly, several other studies have observed HFD-induced changes in sucrose preference, regardless of mouse strain, animal age, dietary fat content or duration of HFD feeding (Sharma et al., 2013; Takase et al., 2016; Yamada et al., 2011). However, it remains relatively unknown whether these behavioural changes indicate specific differences in taste and consummatory behaviour specifically, or anhedonia more broadly (Lockie et al., 2015).
We observed that HFD induced an increase in moving speed but a decrease in mobility during the OFT, although HFD significantly decreased travel distance. Importantly, correlation analysis found no conclusive evidence that these behavioural changes were attributable to fat mass itself or motor impairment due to obesity, and hence may instead 26
represent the effects of neurological changes. This is also in agreement with the similar conclusion that the altered behaviour in the FST was not due to physical obesity-related changes in swimming capacity. HFD influenced changes in locomotor activity have also been reported in previous experiments, with some studies reporting that HFD increased locomotor activity, including a longer travel distance than controls (Krishna et al., 2015; Sharma and Fulton, 2013).
4.2 Effects of HFD on the gut microbiota HFD feeding of mice resulted in changes in the gut microbiota that were characterised by significant reductions in the richness of the microbial community (alpha diversity), alteration of beta diversity and changes in F/B ratio. These changes were aligned with numerous studies (Boulangé et al., 2016; Cani and Delzenne, 2010; Geurts et al., 2011; Turnbaugh et al., 2008). On the other hand, Pyndt Jorgensen et al reported that HFD-induced changes in gut microbiome was not necessarily associated with depressive-like behaviours, such as anhedonia and behavioural despair (Pyndt Jorgensen et al., 2014). One possible reason contributing to the difference between our study and the Pyndt Jorgensen et al study might be a difference in behavioural reactivity in the FST by different mouse strains (Lucki et al., 2001).
4.3 Effects of MINO on diet-induced behavioural changes MINO significantly increased the time spent in the centre zone in HFD mice of the OFT. Given that MINO did not affect the distance travelled, duration of mobility or speed in the OFT, the increased time that MINO treated HFD mice spent in the centre region of the OFT 27
indicates that this change is not simply due to increased locomotor activity, and may instead reflect a possible reduction in anxiety-like behaviour. However, the interpretation of increased time spent in the centre zone of an OFT indicating reduced anxiety is often criticised as an oversimplification, which has limited validity and often lacks robust reproducibility (Spruijt et al., 2014; Walsh and Cummins, 1976). Thus, other behavioural tests which are more suitable to assess anxiety-like behaviour, such as the light/dark box or the elevated plus maze, may be required in future investigations to determine whether reduced anxiety-like behaviour is truly present. Lastly, MINO did not affect the duration of FST immobility nor sucrose consumption. Therefore, we were unable to confirm the therapeutic effects of MINO on our diet-induced model of behavioural change.
4.4 Effects of MINO on the gut microbiota Our study showed significant shifts in the gut microbial population in response to MINO treatment. MINO treatment resulted in a significant reduction in the Chao 1 richness index and membership of abundant bacteria within the composition of the gut microbial community in the STD/MINO and HFD/MINO groups of mice. Weighted UniFrac PCoA analysis showed clear separation between the four treatment groups, indicating that both diet and MINO exerted significant effects on the microbial communities. Intriguingly, MINO mediated significant changes in membership of the abundant bacteria and the gut microbiome community profiles in both HFD and control animals, rather than restoring the profile of the HFD fed mice to the original state. These results were confirmed using AMOVA on weighted UniFrac distances, which indicated significant group differences in the microbial communities (p = 0.001). Our results thus suggest that MINO might have differential effects on STD and HFD fed gut microbiota. The findings here suggest that 28
different key populations in the STD and HFD fed gut microbiota might be affected by MINO treatment and in turn, may result in different alterations in the balance of the memberships of abundant bacteria between the STD and HFD fed gut microbiome communities. This intriguing finding needs to be further investigated by exploring what drives the differential effects of MINO, as well as evaluating deeper levels of bacterial compositions, such as class or species levels, in future studies. Surprisingly, despite the clinical use of MINO as an antibiotic, the specific effects of MINO on the gut microbiota have been documented in very few studies. A previous study showed that MINO treatment (via intraperitoneal injection, 5mg/kg; 21 days) in parallel with chronic restraint stress prevented stress-induced depressive-like behaviour in mice in the FST (Wong et al., 2016). Our study had different experimental designs and hypotheses from Wong et al (2016), and therefore the findings are not readily comparable. However, both studies demonstrated that MINO did not restore the gut microbiome to the original state in the time frame of this study. The impacts of antibiotic use of the gut microbiome have been shown to have lasting and varying impacts over time; with some studies showing changes lasting up to 12 months (Dethlefsen et al., 2008; Jernberg et al., 2010). Different antimicrobial agents can influence the normal gut microbiome in divergent ways. MINO exerts a significant impact on a wide range of bacteria, however, it appears that the bacteriostatic effect of MINO mediates a shift in the balance of the gut microbiota; as the growth of existing bacteria is slowed, opportunities are provided for other bacteria to grow (Kim et al., 2017). MINO did not restore the original ecosystem in mice which had been exposed to a high fat diet. Hence, the gut microbiota population elicited by MINO treatment is different from both healthy (control) and diet-affected states. This raises the question of
29
which underlying pathways link the changes in the gut microbiome elicited by MINO to health, disease, and behaviour in the host.
5. Conclusion Our study demonstrated that high fat diet consumption resulted in significant changes in both behaviour and the gut microbial community. Our study also highlighted that MINO treatment significantly altered gut microbial profiles. MINO altered the balance of the gut microbiome in both mice fed standard chow and a high fat diet, as indicated by reductions of bacterial abundance profiles and increases in the F/B ratio, but did not push the microbiome profile of the HFD group back towards that of the control group. However, MINO-induced alterations of the gut microbial community were not associated with significant modulation of the behavioural changes induced by the HFD. Our findings indicate that MINO treatment can exert striking changes in the gut microbiome profile, but also that these changes do not necessarily correlate with changes in behaviour. Together, this data highlights the complex relationships between gut microbial profiles and behaviour, and indicates that further exploration of both pharmacological and dietary factors is warranted.
6. Acknowledgement We thank Drs Martin O’Hely, Fiona M Collier, Bruna Panizzutti and Chiara Bortolasci for technical assistance.
7. Funding This research was funded in part by the Rebecca L. Cooper Medical Research Foundation. OMD is supported by an R.D. Wright Biomedical NHMRC Research Fellowship (1145634).
30
8. Disclosure OMD has received grant support from the Brain and Behavior Foundation, Simons Autism Foundation, Stanley Medical Research Institute, Deakin University, Lilly, NHMRC and ASBDD/Servier. She has also received in kind support from BioMedica Nutracuticals, NutritionCare and Bioceuticals. Other authors declare no competing interests.
31
9. References Akbaraly, T.N., et al., 2009. Dietary pattern and depressive symptoms in middle age. Br J Psychiatry. 195, 408-13. Alexander, C., Rietschel, E.T., 2001. Bacterial lipopolysaccharides and innate immunity. J Endotoxin Res. 7, 167-202. Andre, C., et al., 2014. Diet-induced obesity progressively alters cognition, anxiety-like behavior and lipopolysaccharide-induced depressive-like behavior: focus on brain indoleamine 2,3-dioxygenase activation. Brain Behav Immun. 41, 10-21. Boulangé, C.L., et al., 2016. Impact of the gut microbiota on inflammation, obesity, and metabolic disease. Genome Medicine. 8, 42. Brown, K., et al., 2012. Diet-Induced Dysbiosis of the Intestinal Microbiota and the Effects on Immunity and Disease. Nutrients. 4, 1095-1119. Bruce-Keller, A.J., et al., 2015. Obese-type gut microbiota induce neurobehavioral changes in the absence of obesity. Biol Psychiatry. 77, 607-15. Cani, P.D., Delzenne, N.M., 2010. Involvement of the gut microbiota in the development of low grade inflammation associated with obesity: focus on this neglected partner. Acta Gastroenterol Belg. 73, 267-9. Caporaso, J.G., et al., 2010. QIIME allows analysis of high-throughput community sequencing data. Nat Meth. 7, 335-336. Cepeda, M.S., Stang, P., Makadia, R., 2016. Depression Is Associated With High Levels of CReactive Protein and Low Levels of Fractional Exhaled Nitric Oxide: Results From the 2007-2012 National Health and Nutrition Examination Surveys. J Clin Psychiatry. 77, 1666-1671.
32
de La Serre, C.B., et al., 2010. Propensity to high-fat diet-induced obesity in rats is associated with changes in the gut microbiota and gut inflammation. Am J Physiol Gastrointest Liver Physiol. 299, G440-8. De Luca, M.A., 2014. Habituation of the responsiveness of mesolimbic and mesocortical dopamine transmission to taste stimuli. Front Integr Neurosci. 8, 21. Dean, O.M., et al., 2012. Minocycline: therapeutic potential in psychiatry. CNS Drugs. 26, 391-401. Dean, O.M., et al., 2017. Adjunctive minocycline treatment for major depressive disorder: A proof of concept trial. Aust N Z J Psychiatry. 51, 829-840. Dethlefsen, L., et al., 2008. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLoS Biol. 6, e280. Edgar, R.C., 2013. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Meth. 10, 996-998. Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of Molecular Variance Inferred from Metric Distances among DNA Haplotypes: Application to Human Mitochondrial DNA Restriction Data. Genetics. 131, 479-491. Garate, I., et al., 2011. Origin and consequences of brain Toll-like receptor 4 pathway stimulation in an experimental model of depression. J Neuroinflammation. 8, 151. Garrido-Mesa, N., Zarzuelo, A., Galvez, J., 2013. Minocycline: far beyond an antibiotic. Br J Pharmacol. 169, 337-52. Geurts, L., et al., 2011. Altered Gut Microbiota and Endocannabinoid System Tone in Obese and Diabetic Leptin-Resistant Mice: Impact on Apelin Regulation in Adipose Tissue. Frontiers in Microbiology. 2, 149.
33
Henry, C.J., et al., 2008. Minocycline attenuates lipopolysaccharide (LPS)-induced neuroinflammation, sickness behavior, and anhedonia. J Neuroinflammation. 5, 15. Husain, M.I., et al., 2017. Minocycline as an adjunct for treatment-resistant depressive symptoms: A pilot randomised placebo-controlled trial. J Psychopharmacol. 31, 11661175. Jacka, F.N., et al., 2014. Dietary patterns and depressive symptoms over time: examining the relationships with socioeconomic position, health behaviours and cardiovascular risk. PLoS One. 9, e87657. Jernberg, C., et al., 2010. Long-term impacts of antibiotic exposure on the human intestinal microbiota. Microbiology. 156, 3216-3223. Kelly, J.R., et al., 2015. Breaking down the barriers: the gut microbiome, intestinal permeability and stress-related psychiatric disorders. Front Cell Neurosci. 9, 392. Kim, S., Covington, A., Pamer, E.G., 2017. The intestinal microbiota: Antibiotics, colonization resistance, and enteric pathogens. Immunological reviews. 279, 90-105. Klindworth, A., et al., 2013. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Research. 41, e1-e1. Krishna, S., et al., 2015. Neurochemical and electrophysiological deficits in the ventral hippocampus and selective behavioral alterations caused by high-fat diet in female C57BL/6 mice. Neuroscience. 297, 170-81. Krishna, S., et al., 2016. Time-dependent behavioral, neurochemical, and metabolic dysregulation in female C57BL/6 mice caused by chronic high-fat diet intake. Physiol Behav. 157, 196-208.
34
Krishnan, V., Nestler, E.J., 2011. Animal Models of Depression: Molecular Perspectives. Current Topics in Behavioral Neurosciences. 7, 121-147. Kubera, M., et al., 2013. A new animal model of (chronic) depression induced by repeated and intermittent lipopolysaccharide administration for 4months. Brain, Behavior, and Immunity. 31, 96-104. Liu, Y.N., et al., 2015. TNFalpha mediates stress-induced depression by upregulating indoleamine 2,3-dioxygenase in a mouse model of unpredictable chronic mild stress. Eur Cytokine Netw. 26, 15-25. Lockie, S.H., et al., 2015. Diet-induced obesity causes ghrelin resistance in reward processing tasks. Psychoneuroendocrinology. 62, 114-120. Lozupone, C., Knight, R., 2005. UniFrac: a New Phylogenetic Method for Comparing Microbial Communities. Applied and Environmental Microbiology. 71, 8228-8235. Lozupone, C.A., Knight, R., 2008. Species Divergence and the Measurement of Microbial Diversity. FEMS microbiology reviews. 32, 557-578. Lucki, I., Dalvi, A., Mayorga, A.J., 2001. Sensitivity to the effects of pharmacologically selective antidepressants in different strains of mice. Psychopharmacology (Berl). 155, 315-22. Maes, M., et al., 2012. Increased IgA and IgM responses against gut commensals in chronic depression: further evidence for increased bacterial translocation or leaky gut. J Affect Disord. 141, 55-62. Majidi, J., Kosari-Nasab, M., Salari, A.A., 2016. Developmental minocycline treatment reverses the effects of neonatal immune activation on anxiety- and depression-like behaviors, hippocampal inflammation, and HPA axis activity in adult mice. Brain Res Bull. 120, 1-13. 35
McKim, D.B., et al., 2016. Neuroinflammatory Dynamics Underlie Memory Impairments after Repeated Social Defeat. J Neurosci. 36, 2590-604. Miyaoka, T., et al., 2012. Minocycline as adjunctive therapy for patients with unipolar psychotic depression: an open-label study. Prog Neuropsychopharmacol Biol Psychiatry. 37, 222-6. Neale, E.P., Batterham, M.J., Tapsell, L.C., 2016. Consumption of a healthy dietary pattern results in significant reductions in C-reactive protein levels in adults: a meta-analysis. Nutr Res. 36, 391-401. Paulson, J.N., et al., 2013. Differential abundance analysis for microbial marker-gene surveys. Nat Meth. 10, 1200-1202. Price, M.N., Dehal, P.S., Arkin, A.P., 2010. FastTree 2 – Approximately Maximum-Likelihood Trees for Large Alignments. PLOS ONE. 5, e9490. Pyndt Jorgensen, B., et al., 2014. A possible link between food and mood: dietary impact on gut microbiota and behavior in BALB/c mice. PLoS One. 9, e103398. Rahe, C., et al., 2015. Associations between depression subtypes, depression severity and diet quality: cross-sectional findings from the BiDirect Study. BMC Psychiatry. 15, 38. Remus, J.L., Dantzer, R., 2016. Inflammation Models of Depression in Rodents: Relevance to Psychotropic Drug Discovery. The international journal of neuropsychopharmacology. 19, pyw028. Rienks, J., Dobson, A.J., Mishra, G.D., 2013. Mediterranean dietary pattern and prevalence and incidence of depressive symptoms in mid-aged women: results from a large community-based prospective study. Eur J Clin Nutr. 67, 75-82. Rosenblat, J.D., McIntyre, R.S., 2018. Efficacy and tolerability of minocycline for depression: A systematic review and meta-analysis of clinical trials. J Affect Disord. 227, 219-225. 36
Sanchez-Villegas, A., et al., 2009. Association of the Mediterranean dietary pattern with the incidence of depression: the Seguimiento Universidad de Navarra/University of Navarra follow-up (SUN) cohort. Arch Gen Psychiatry. 66, 1090-8. Sanchez-Villegas, A., et al., 2012. Fast-food and commercial baked goods consumption and the risk of depression. Public Health Nutr. 15, 424-32. Schwingshackl, L., Hoffmann, G., 2014. Mediterranean dietary pattern, inflammation and endothelial function: a systematic review and meta-analysis of intervention trials. Nutr Metab Cardiovasc Dis. 24, 929-39. Sharma, S., Fernandes, M.F., Fulton, S., 2013. Adaptations in brain reward circuitry underlie palatable food cravings and anxiety induced by high-fat diet withdrawal. Int J Obes (Lond). 37, 1183-91. Sharma, S., Fulton, S., 2013. Diet-induced obesity promotes depressive-like behaviour that is associated with neural adaptations in brain reward circuitry. Int J Obes (Lond). 37, 382-9. Slyepchenko, A., et al., 2017. Gut Microbiota, Bacterial Translocation, and Interactions with Diet: Pathophysiological Links between Major Depressive Disorder and NonCommunicable Medical Comorbidities. Psychother Psychosom. 86, 31-46. Spruijt, B.M., et al., 2014. Reproducibility and relevance of future behavioral sciences should benefit from a cross fertilization of past recommendations and today's technology: "Back to the future". J Neurosci Methods. 234, 2-12. Takase, K., et al., 2016. High-fat diet feeding alters olfactory-, social-, and reward-related behaviors of mice independent of obesity. Obesity (Silver Spring). 24, 886-94. Turnbaugh, P.J., et al., 2008. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe. 3, 213-23.
37
Turnbaugh, P.J., et al., 2009. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 1, 6ra14. Vichaya, E.G., et al., 2018. Motivational changes that develop in a mouse model of inflammation-induced depression are independent of indoleamine 2,3 dioxygenase. Neuropsychopharmacology. Walsh, R.N., Cummins, R.A., 1976. The Open-Field Test: a critical review. Psychol Bull. 83, 482-504. Wong, M.L., et al., 2016. Inflammasome signaling affects anxiety- and depressive-like behavior and gut microbiome composition. Mol Psychiatry. 21, 797-805. Yamada, N., et al., 2011. Impaired CNS leptin action is implicated in depression associated with obesity. Endocrinology. 152, 2634-43. Yu, Z.M., Parker, L., Dummer, T.J., 2014. Depressive symptoms, diet quality, physical activity, and body composition among populations in Nova Scotia, Canada: report from the Atlantic Partnership for Tomorrow's Health. Prev Med. 61, 106-13. Zheng, L.S., Kaneko, N., Sawamoto, K., 2015. Minocycline treatment ameliorates interferonalpha- induced neurogenic defects and depression-like behaviors in mice. Front Cell Neurosci. 9, 5. Zheng, P., et al., 2016. Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism. Mol Psychiatry. 21, 786-96.
38
Highlights
Long-term high fat diet consumption was associated with behavioural changes but not depressive-like behaviour in male C57BL/6J mice.
Chronic high fat diet feeding mediated significant alteration of the gut microbiome and behaviour in the host.
Minocycline treatment altered gut microbiome profiles (Chao1 richness index, Principal coordinate analysis and relative abundance) were different from both healthy and diet-affected mice.
Our study first showed minocycline-induced changes in the gut microbiome do not necessarily correlate with changes in behaviour.
39
S0889-1591(18)31241-8 https://doi.org/10.1016/j.bbi.2019.09.001 YBRBI 3848
To appear in:
Brain, Behavior, and Immunity
Received Date: Revised Date: Accepted Date:
17 January 2019 28 August 2019 3 September 2019
Please cite this article as: Hasebe, K., Rivera, L.R., Smith, C.M., Allnutt, T., Crowley, T., Nelson, T.M., Dean, O.M., McGee, S.L., Walder, K., Gray, L., Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline, Brain, Behavior, and Immunity (2019), doi: https://doi.org/10.1016/j.bbi.2019.09.001
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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.
© 2019 Published by Elsevier Inc.
Modulation of high fat diet-induced microbiome changes, but not behaviour, by minocycline.
Kyoko Hasebea, Leni R Riveraa, Craig M Smithab, Theo Allnuttd, Tamsyn Crowleya, Tiffanie M Nelsonef, Olivia M Deanbc, Sean L McGeea, Ken Waldera, Laura Grayab
aSchool
of Medicine, Centre for Molecular and Medical Research, Deakin University,
Geelong, Australia; bThe Florey Institute of Neuroscience and Mental Health, Parkville, Australia; cSchool of Medicine, IMPACT Strategic Research Centre, Deakin University, Geelong, Australia; dTheo Allnutt Bioinformatics, Geelong, Australia; eGeelong Centre for Emerging Infectious Diseases, Geelong, Victoria, 3220, Australia; fDeakin University, Geelong, Victoria, 3220, Australia.
Corresponding author: Dr. Laura Gray, School of Medicine, Centre for Molecular and Medical Research, Deakin University, Geelong, Australia; [email protected]
Word count: 6208 words
Abstract An emerging novel therapeutic agent for major depressive disorder, minocycline, has the potential to influence both gut microbiome and inflammatory status. The present study showed that chronic high fat diet feeding led to changes in both behaviour and the gut microbiome in male mice, without an overt inflammatory response. The diet-induced behavioural changes were characterised as increased immobility in the forced swim test and 1
changes in locomotor activities in the open field test. Minocycline significantly altered the gut microbiome, rendering a community distinctly different to both untreated healthy and diet-affected states. In contrast, minocycline did not reverse high fat diet-induced changes in behaviour.
Key words High fat diet feeding, minocycline, gut microbiome
2
1. Introduction Emerging therapies for major depressive disorder (MDD), such as the promising agent minocycline (Rosenblat and McIntyre, 2018), have the potential to influence both the gut microbiome composition and inflammatory status. Given our developing understanding of how changes in the gut microbiome, often driven by diet, can influence mood and behaviour, it is critical to investigate how this nexus of intersecting factors can be exploited by novel therapies with strong potential for impact on patient outcomes. There is now significant evidence to support an association between diet quality and the risk of developing MDD (Jacka et al., 2014; Rahe et al., 2015; Yu et al., 2014). Although the development of MDD can alter dietary patterns, longitudinal studies confirm that poor diet quality often precedes MDD onset, suggesting that diet can be a causal factor (Akbaraly et al., 2009; Rienks et al., 2013; Sanchez-Villegas et al., 2009; Sanchez-Villegas et al., 2012). Furthermore, dietary patterns modulate inflammatory status. Several studies reported that healthy dietary patterns lower levels of inflammatory mediators, such as C-reactive protein (CRP) (Neale et al., 2016; Schwingshackl and Hoffmann, 2014) and Interleukin-6 (IL-6) (Schwingshackl and Hoffmann, 2014). Due to the established link between inflammation and mood (Cepeda et al., 2016; Kubera et al., 2013; Remus and Dantzer, 2016; Vichaya et al., 2018), poor diet quality could therefore induce elevations of pro-inflammatory mediators, which subsequently may predispose to a depressed mood. Changes in gut microbiota are now considered to be mechanistically important to the interaction between poor diet and mood. Alterations in gut microbial communities induced by poor diets have been reported in multiple studies (Brown et al., 2012; de La Serre et al., 2010; Turnbaugh et al., 2008; Turnbaugh et al., 2009). A significant proportion of the gut microbial population are capable of modulating the host inflammatory response, via key 3
molecules including bacterial lipopolysaccharides (LPS), which are major outer surface membrane components present in almost all gram-negative bacteria (Alexander and Rietschel, 2001). An increase in permeability of the intestinal barrier (tight junctions) and resultant bacterial translocation to the systemic circulation (often referred to as a “leaky gut”), is also implicated in the pathophysiology of MDD (Garate et al., 2011; Kelly et al., 2015; Maes et al., 2012; Slyepchenko et al., 2017). Experimental administration of LPS in rodents induces a suite of behavioural symptoms sufficiently akin to depressive states, which is considered an animal model of MDD (Kubera et al., 2013; Remus and Dantzer, 2016). Several studies have suggested that changes in the gut microbiome may be an initiating or causal factor in the development of mood changes. For instance, transplantation of the faecal microbiome from patients with MDD into germ-free mice resulted in an increase in depressive-like behaviours relative to mice transplanted with microbiota from healthy control individuals (Zheng et al., 2016). Furthermore, transplantation of the cecum and colonic microbiome from mice fed a high fat diet induced changes in anxiety-like behaviours in recipient mice (Bruce-Keller et al., 2015). These findings suggest the gut microbial community may play a causal role in the pathogenesis of depressive-like behaviour, and that diet may be a key factor in driving changes in the microbiome, and thus mood. Minocycline (MINO) is a broad spectrum antibiotic which has been investigated for its potential antidepressant effects. Although MINO exerts direct effects on the gut microbial community when administered orally, several reports suggest the potential antidepressant effects are primarily caused by pharmacological effects within the brain, via modulation of neuroinflammation and suppression of microglial activity (Dean et al., 2012; Garrido-Mesa et al., 2013; Henry et al., 2008; Zheng et al., 2015). MINO ameliorates the depressive-like 4
behaviour observed in various animal models of depression (Henry et al., 2008; Liu et al., 2015; Majidi et al., 2016). For example, Wong et al. (Wong et al., 2016) reported that chronic restraint stress-induced depressive-like behaviour in mice was reversed by oral MINO, and this was associated with changes in gut microbiota. Adjunctive MINO treatment also alleviated depressive symptoms in human MDD patients (Dean et al., 2017; Husain et al., 2017; Miyaoka et al., 2012). This study therefore investigated the effects of high fat diet-induced changes in gut microbiota and behaviour in mice, and on inflammatory and microbial signalling pathways in the periphery and brain. In addition, the present study first showed minocycline-induced changes in the gut microbiome were not associated with reversals of behaviour induced by high fat diet.
2. Materials and methods 2.1 Animals and diet All experiments were performed in accordance with the guidelines of Deakin University Animal Ethics Committee and the National Health and Medical Research Council. Four-week old male C57BL/6J mice were purchased from Animal Resources Centre (Perth, Western Australia, Australia). Animals were group housed (four per cage) and maintained on a 12h light/dark cycle (7am – 7pm) at 20-22 °C with ad libitum access to food and water. After two weeks of acclimatisation, animals were assigned into two groups: high fat diet (HFD), and standard chow (STD), and were maintained on each respective diet until the end of the study. Groups were matched at this baseline time point across body weight, fat mass and lean mass measured by an EchoMRI Whole Body Composition Analyzer. The STD group was fed with standard chow (BARASTOC Rat and Mouse Cubes, Ridley AgriProducts, Victoria, 5
Australia) that contains 5% kcal of fat and 20% kcal of protein. The HFD group was fed with high fat diet (SF04-001, Speciality feeds, Western Australia, Australia) that contains 45% kcal of fat (45% of available energy in the form of fat) and 22.6% kcal of protein. Detailed dietary compositions of standard chow and high fat diet are shown in Table 1. Body weight was measured once per week throughout of the study. Mice were 24 weeks old by the time of behavioural assessments. Table 1. Nutritional content of standard chow and high fat diets Standard chow diet
High fat diet
Barastoc, Mice and Rodent cube
Specialty Feeds SF04-001
Protein
20.0%
22.6%
Total Fat
5.0%
23.5%
Crude Fibre
5.0%
5.4%
12.6MJ/Kg
19MJ/Kg
Digestible Energy
2.2 Experimental design and timeline The Experimental design and timeline used for this study is shown in Figure 1.
Figure 1. Schematic diagram of experimental design and timeline.
2.3 Sacrifice and tissue collection 6
The mice of each group were humanely killed by cervical dislocation, the brain removed, and the hypothalamus and bilateral whole hippocampal formations dissected. Hypothalamic and hippocampal tissues were snap frozen in liquid nitrogen and stored at -80 °C. Plasma was collected from trunk blood and stored at -80 °C for later use
2.4 Drugs Minocycline hydrochloride (MINO) (PCCA, Houston, USA) was prepared in 0.9% saline at concentrations of 10 mg/ml for dosing at 50 mg/kg of MINO per day. The dose of MINO was in line with other studies (Henry et al., 2008; McKim et al., 2016; Zheng et al., 2015) and a pilot study conducted in our laboratory (unpublished data). Although body weights were relatively stable after 13 weeks of HFD or STD, dosages were re-adjusted weekly to ensure they were aligned with body weights, and solutions were pH-matched to that of the vehicle. MINO and saline were orally gavaged daily for a 4 week period, which began 13 weeks after assigning HFD and STD groups.
2.5 Behavioural testing All behavioural testing was conducted between 10:00 am and 14:00 pm and was recorded and analysed using Noldus EthoVision XT8 (Noldus Information Technology, The Netherlands).
2.6 Open field test The open field test (OFT) was used to measure locomotion and exploratory behaviour in a novel environment. The apparatus comprised a clear plexiglass box (50 cm x 50 cm, with 36 cm high walls). The light intensity was low in the examination room and there was no 7
difference in the light intensity between the centre and the outer zones. Mice were placed in a corner of the area, and allowed to move freely for 6 min and filmed from above. Time spent in the centre, travel distance and mobility were analysed using a video tracking system (Ethovision XT 8.5, Noldus Information Technology, The Netherlands).
2.7 Porsolt forced swim test The Porsolt forced swim test (FST) displays predictive validity for testing the efficacy of antidepressants (Krishnan and Nestler, 2011). Mice were placed individually in a glass cylinder (diameter: 18.5 cm; height: 26.5 cm) filled with 25± 1 °C water to a height of 13.5 cm for 6 min. Mice were filmed from the side, and the duration of immobility and latency to first immobility were scored manually by a blinded experimenter, where immobility (i.e. the Porsolt posture) was defined as immobility in all four limbs.
2.8 Sucrose preference test The sucrose preference test (SPT) was used to measure anhedonic behaviour (Krishnan and Nestler, 2011). To avoid stress associated with social isolation, mice were housed 2 per cage and were given free access to two water bottles, one with tap water and another with 1 % sucrose solution. Mice were split from 4/box to 2/box at 5 pm, and returned to 4/box at 9 am. Both pairs were placed in new boxes, and given time (approximate 15 minutes) to acclimatise before the addition of sucrose. At the conclusion of the test, bottles were removed and consumptions of sucrose solution and tap water were measured by weighing each bottle, which was conducted by a blinded experimenter.
8
2.9 16S metagenomic sequencing Faecal samples were collected during the final week of behavioural testing. DNA was isolated using QIAamp DNA stool kits (Qiagen Pty Ltd, Victoria, Australia), modified to include a bead-beating step. The extracted DNA was quantified and purity was assessed using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA). Sequencing was performed at Macrogen Inc (Seoul, South Korea). The extracted DNA was prepared for sequencing on the Illumina MiSeq platform using the Illumina 16S Metagenomic Sequencing Library Preparation protocol. Primers targeting the 16S rRNA gene’s V3-V4 region were amplified from DNA. Primer sequences included Illumina’s adaptor sequence (shown in lowercase): 341F, 5'-tcgtcggcagcgtcagatgtgtataagagacagCCTACGGGNGGCWGCAG-3’ and 805R, 5'-gtctcgtgggctcggagatgtgtataagagacaggactachvGGGTATCTAATCC-3’ (Klindworth et al., 2013). The PCR product was purified using AMPure XP beads. Amplicons were indexed using the Illumina Nextera XT Index kit, and purified again using AMPure XP beads prior to a fluorometric quantification method.
2.10 Bioinformatics Sequenced amplicons were filtered to a minimum read length of 300 base pairs (bp), resulting in an average read length of 100 bp. The UPARSE pipeline for illumine paired reads was conducted (Edgar, 2013) with a 3% cluster radius. The RDP 16S database v16 was used to taxonomically classify operational taxonomic units (OTUs) with a 90% confidence threshold. OTUs with abundance less than 0.5% of the total number of reads or occurrence of less than two across all samples were removed from the analysis. QIIME 1.8 (Caporaso et al., 2010) was used to calculate diversity metrics and statistically compare alpha diversity (on rarefied data) and OTU abundance between species. OTU abundance was cumulative 9
sum scaling (CSS) normalised prior to statistical tests (Paulson et al., 2013). All commands and associated python scripts in the analysis are provided in the GitHub repository: https://github.com/theo-allnutt-bioinformatics/Hasebe_2019. Principal coordinates analysis (PCoA) was performed on weighted UniFrac distances (Lozupone and Knight, 2005) calculated from a maximum likelihood tree (FastTree v2.3.1) (Price et al., 2010) of the filtered OTUs on raw counts and CSS normalised counts. Beta diversity was analysed by Analysis of Molecular Variance (AMOVA) (Excoffier et al., 1992).
2.11 Analyses of inflammatory markers Plasma levels of inflammatory markers were assessed using commercially available ELISA kits: lipopolysaccharide binding protein, LBP (Boster Bio, mouse LBP ELISA kit), soluble CD14, IL6 and TNFα (R&D systems, Quantikine ELISA kits).
2.12 Real time quantitative PCR analysis Hypothalamic and hippocampal tissues were then homogenised and processed for total RNA extraction using RNeasy mini Kits (Qiagen Pty Ltd, Victoria, USA). RNA was quantified and purity assessed using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Massachusetts, USA). Complimentary strand DNA was synthesised using Dynamo cDNA synthesis kit (Thermo Fisher Scientific). Quantitative real-time polymerase chain reaction (qPCR) was performed using Taqman Gene Expression Assays (Applied Biosystems, Foster City, CA, USA) for ionized calcium-binding adapter molecule 1 (Iba1) (Mm00479862_g1), Glial fibrillary acidic protein (gfap) (Mm01253033_m1), Major histocompatibility complex) class II molecules (MHC II) (Mm658576_m1), cluster of differentiation molecule 11b (CD11b) (Mm00434455_m1),
Glutamate
aspartate 10
transporter
(Mm00600697_m1)(GLAST),
Excitatory amino acid transporter (GLT1) (Mm00441457_m1), Cysteine-glutamate antiporter (xCT) (Mm00442530_m1), inducible Nitric oxide synthase (iNOS) (Mm00440502_m1), Occludin (Mm00500912_m1) and 18S rRNA (Mm03928990_g1). Each sample was amplified in triplicate. Relative quantification was calculated by the ΔΔCt method and compared with endogenous levels of 18S rRNA.
2.13 Statistical analyses All behavioural and biochemical graphs were generated and data analysed using Prism (GraphPad Software, Inc. La Jolla, CA, USA). Mean ± standard errors of the mean are displayed. Normality of each dataset was assessed by Kolmogorov-Smirnov test. Two-way Analysis of Variance (ANOVA) was conducted to examine main effects of diet and treatment and interaction effects of diet and treatment, followed by Bonferroni post hoc comparisons to
determine
between
and
within
group
effects.
For
alpha
diversity
and
Firmicutes/Bacteroidetes ratio, Kruskal-Wallis test followed by Dunn’s multiple test was used to determine statistical significance. A weighted distance-based analysis of molecular variance (AMOVA) was used to assess spatial distance of each treatment group observed based on the weighted UniFrac distances in each treatment group in principal coordinate analysis (PCoA). For gene expression analyses, differences in levels of target genes were assessed by unpaired t tests. Statistical significance was set for all analyses at p < 0.05.
11
3. Results 3.1 High fat diet increased adiposity while MINO did not affect body weight or composition HFD fed mice were significantly heavier than STD fed mice at the conclusion of the study, and this effect was similar within both the vehicle and MINO treated groups (Two-way ANOVA; main effect of diet, F (1,68) = 178.2, p < 0.0001; Figure 2A, Supplemental Figure 1). HFD induced an increase in fat mass, in both vehicle and MINO treated mice (main effect of diet, F (1,68) = 325.0, p < 0.0001; Figure 2B). In contrast, lean mass was not affected by HFD (main effect of diet, F
(1,68)
= 3.08, p > 0.05). Furthermore, MINO treatment did not alter
body weight (Two-way ANOVA; main effect of treatment, F (1,68) = 0.67, p > 0.05), fat mass (main effect of treatment, F (1,68) = 0.44, p > 0.05) or lean mass (main effect of treatment, F (1,68)
= 0.006, p > 0.05). In line with these observations, MINO and HFD did not interact to
affect body weight (diet x drug treatment interaction, F (1,68) = 0.1, p > 0.05), fat mass (diet x drug treatment interaction, F
(1,68)
= 2.04, p > 0.05) or lean mass (diet x drug treatment
interaction, F (1,68) = 2.31, p < 0.05).
12
Figure 2. Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle, on: (A) body weights; (B) fat mass, and; (C) lean mass. Data are mean ± SEM. n = 18-22/mice per group. 2-way ANOVA, main effect of diet #### (p<0.0001). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.2 High fat diet mediated behavioural changes and the influence of MINO 3.2.1 High fat diet increased forced swim test immobility, which was not altered by MINO HFD fed mice spent significantly more time immobile in the FST (main effect of diet, F (1, 76) = 12.11, p < 0.001), compared to STD mice (Figure 3A). There was no main effect of MINO treatment (F (1, 76) = 2.54, p > 0.05; diet x drug treatment interaction, F (1, 75) = 0.18, p > 0.05).
13
Because increased immobility might conceivably result from increased fat mass (rather than diet-induced neurological/behavioural changes), a correlation analysis between fat mass and the duration of immobility was performed which included mice from the HFD groups only. However, a linear regression model indicated that fat mass was not significantly associated with the time spent immobile in HFD fed mice (F (1, 38) = 0.11, p > 0.05; Figure 3B). The increased immobility also did not correlate with body weight (Supplemental figure 3(A)).
3.2.2 High fat diet or MINO treatment did not affect sucrose preference Sucrose preference of HFD mice did not differ from STD mice, and there was no effect of MINO treatment (main effect of diet, F
(1, 36)
= 3.96, p > 0.05; main effect of drug, F
2.06, p > 0.05; diet x drug interaction, F (1, 36) = 0.48, p > 0.05; Figure 3C).
14
(1, 36)
=
Figure 3. Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle, on: (A) Duration of immobility in the Porsolt forced swim test; (B) Correlation between duration of immobility and fat mass; and (C) Sucrose preference (%); Sucrose consumption divided by total fluid intake (sucrose and water consumption overnight); Data are mean ± SEM. n = 1822/mice per group (for Sucrose preference test, n = 9-11/mice per group). 2-way ANOVA, main effect of diet ### (p<0.001) and # (p <0.05). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.2.3 MINO altered time spent in the centre of the open field test
15
MINO treatment significantly increased the time spent in the centre zone in the OFT overall (Two-way ANOVA; main effect of drug treatment, F
(1,76)
= 10.37, p < 0.01; Figure 4A), and
this effect appears to be most noticeable within HFD fed mice. In contrast, diet did not affect the time spent in the centre region (main effect of diet, F (1,76) = 0.08, p > 0.05; diet x drug treatment interaction, F
(1,76)
= 1.81, p > 0.05). Two-way ANOVA indicated that MINO
treatment did not affect the distance travelled (main effect of drug treatment, F (1,76) = 0.84, p > 0.05; Figure 4B), suggesting that the increased time that MINO treated mice spent in the centre region (see above) was not simply due to hyperactivity.
3.2.4 High fat diet altered locomotor activity patterns, which were not influenced by MINO HFD mice travelled significantly less distance compared to STD controls (main effect of diet, F
(1,76)
= 7.83, p < 0.01). This effect was similar within both vehicle and MINO treatment
groups (diet x drug treatment interaction, F (1,76) = 0.63, p > 0.05). Although fat mass can potentially alter patterns of locomotor activity, correlation analysis of mice from both HFD groups using a linear regression model indicated that distance travelled and fat mass were not associated (F (1,38) = 0.50, p > 0.05; Figure 4C). The reduced distance travelled in OFT was also not correlated with body weight (Supplemental figure 3 (B)). Duration of mobility was defined as the total amount of time that the animal was moving in the field during the OFT. MINO treatment did not alter duration of mobility relative to vehicle controls (Two-way ANOVA; main effect of treatment, F (1,76) = 2.02, p > 0.05; diet x drug interaction, F (1,76) = 0.19, p > 0.05; Figure 4D). HFD mice were significantly less mobile compared to STD controls overall (main effect of diet, F (1,76) = 19.52, p < 0.0001). Interestingly, HFD mice moved significantly faster than STD mice on average during each movement episode compared to STD mice control overall (Two-way ANOVA; main effect of 16
diet, F
(1,76)
= 19.41, p < 0.0001; Figure 4E). Combined with the finding that HFD mice
displayed reduced duration of mobility and distance travelled, this suggests that HFD mice engaged in a different pattern of locomotor behaviour, which was associated with fast running bouts that were shorter and/or less frequent than the slower yet more consistent locomotor activity of STD controls. MINO treatment did not affect locomotor activity speed (main effect of treatment, F (1,76) = 2.89, p > 0.05; diet x drug interaction, F (1,76) = 0.33, p > 0.05). A linear regression model showed no association between locomotor activity speed and fat mass within HFD fed mice (F (1,38) = 1.59, p > 0.05; Figure 4F). In addition, there was no association between locomotor activity speed and body weight (Supplemental figure 3 (C)).
17
Figure 4 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on the open field test: (A) Time spent in the centre; (B) Travel distance; (C) Correlation between distance travelled and fat mass; and (D) Duration spent mobile; (E) Speed; and (F) Correlation between speed and fat mass. Data are mean ± SEM. n = 18-22/mice per group. 2-way 18
ANOVA, main effect of diet #### (p < 0.0001) and ## (p <0.01); main effect of treatment $$ (p <0.01). STD = standard diet; HFD = high fat diet; Vehicle = vehicle group; MINO = minocycline treatment group.
3.3 High fat diet and MINO mediated alteration in gut microbiota 3.3.1 Principal coordinate analysis (PCoA) Principal coordinate analysis (PCoA) was used to identify whether treatment groups cluster together based on gut microbial community members and their abundances. The PCoA plot was constructed with the weighted UniFrac distance matrix, which incorporates the relative abundance of gut microbiota and their phylogenetic history based on gene sequence similarity (Figure 5A). The gut microbiota in STD/Vehicle, STD/MINO, HFD/Vehicle and HFD/MINO clustered distinctly from each other. HFD (PC1) explained 45.3% of the variation and MINO treatment (PC2) explained 15.6% of the variation between groups. A weighted distance-based analysis of molecular variance (AMOVA) was used to assess the statistical significance of the spatial separation observed between treatment groups. Statistically significant dissimilarities were observed between STD/Veh, HFD/Veh, STD/MINO and HFD/MINO groups with respect to bacterial diversity (p = 0.001).
3.3.2 Relative abundance at phylum level Figure 5B shows the relative abundance of bacterial phyla in each treatment group. At the phylum level, it appeared that Proteobacteria and Baceteroidetes were substantially reduced with a HFD and replaced by Verrucomicrobia. Following from MINO treatment, STD/MINO group was subsequently replaced with Candidatus-Saccharibacteria (also known as TM7) and Firmicutes, while HFD/MINO group was replaced with Actinobacteria. 19
3.3.3 Alpha diversity (Chao1 richness index) Alpha diversity is a measurement to assess structure and function of microbial diversity within a community (a sample) which accounts for the number of species, commonness and rareness of abundant bacteria (Lozupone and Knight, 2008). A Kruskal-Wallis test followed by Dunn’s multiple comparison test indicated that the Chao 1 richness index suggested a significant reduction of richness in the gut microbial diversity in HFD/Veh mice compared with STD/Veh mice (p < 0.05; Figure 5C). MINO treatment also contributed to a significant reduction of diversity of the gut microbiota that was observed within both STD (p < 0.001) and HFD mice (p < 0.01).
3.3.4 Firmicutes to Bacteroidetes (F/B) ratio Bacteroidetes and Firmicutes are two dominant phyla of the gut microbiota. Firmicutes to Bacteroidetes (F/B) ratio (Figure 5D) is used to assess differences in the main composition of the gut microbiota. A Kruskal-Wallis test followed by Dunn’s multiple comparison test indicated that HFD significantly increased the F/B ratio compared to the gut microbiota in STD/Veh mice (p < 0.001).
20
Figure 5 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on gut microbial community: (A) Principal coordinate analysis; (B) Relative abundance at phylum level; (C) Chao1 richness index; and (D) Firmicutes/Bacteroidetes ratio. n = 9-17/mice per group. Kruskal-Wallis test followed by Dunn’s multiple test was used for Chao1 richness index and F/B ratio. Upper and lower quartiles defined the box with median midline, and the whiskers were assessed using Tukey’s method. ***p < 0.001; **p < 0.01; *p < 0.05.
3.4 Effects of High fat diet and MINO on inflammatory mediators in the periphery Two-way ANOVA indicated that HFD significantly increased plasma levels of LBP (main effect of diet, F (1,67) = 4.70, p < 0.05), compared to STD fed mice (Figure 6A). In addition, MINO 21
treatment significantly decreased plasma LBP levels (main effect of treatment, F (1,67) = 5.73, p < 0.05; diet x drug interaction, F (1,67) = 0.18, p > 0.05). On the other hand, neither HFD nor MINO treatment affected the levels of soluble CD14 in plasma (Two-way ANOVA; main effect of diet, F (1,46) = 0.003, p > 0.05; main effect of treatment, F (1, 46) = 0.038, p > 0.05; diet x drug interaction, F (1,46) = 0.28, p > 0.05; Figure 6B). The levels of IL6 and TNFα in the periphery were under detection level (data not shown).
3.5 Effects of MINO on inflammatory mediators in the CNS As markers of glial activation, levels of Iba1 and GFAP in the hippocampus and the hypothalamus were assessed using real time QPCR. In the hypothalamus, two-way ANOVA indicated that MINO treatment significantly reduced levels of GFAP expression (main effect of treatment, F (1,31) = 8.76, p < 0.01; Figure 6C). The effect was independent from HFD (main effect of diet, F (1,31) = 2.89, p > 0.05; diet x drug interaction, F (1,31) = 0.62, p > 0.05). Levels of Iba 1 expression did not alter with either HFD or MINO in the hypothalamus (Two-way ANOVA; main effect of diet, F
(1,32)
0.05; diet x drug interaction, F
= 0.07, p > 0.05; main effect of drug F
(1,32)
(1,32)
= 0.70, p >
= 0.02, p > 0.05). In the hippocampus, levels of Iba1
expression also did not alter with either HFD nor MINO (Two-way ANOVA; main effect of diet, F
(1, 34)
= 0.27, p > 0.05; main effect of treatment, F
(1,34)
= 0.83, p > 0.05; diet x drug
interaction, F (1,34) = 0.91, p > 0.05), and nor did GFAP levels (Two-way ANOVA; main effect of diet, F (1, 34) = 0.28; main effect of drug, F (1,34) = 0.04, p > 0.05; diet x drug interaction, F (1,34)
= 0.24, p > 0.05). Levels of MHC II, CD11b, GLAST, GLT1, xCT, iNOS and Occludin did not
differ between groups (data not shown).
22
Figure 6 Effects of 17 weeks HFD or STD, and 4 weeks of MINO or vehicle on: (A) Plasma LBP levels. ; (B) Plasma soluble CD14 levels; C) Gene expression levels of glial activation markers 23
in the hippocampus and hypothalamus. Data are mean ± SEM. n = 7-11/mice per group. 2way ANOVA, main effect of diet # (p <0.05) and main effect of treatment $$ (p <0.01) and $ (p <0.05).
24
4. Discussion This study examined the effect of MINO on the behavioural changes, inflammatory state and gut microbiome alterations elicited by a HFD. HFD induced behavioural changes was paralleled by significant shifts in the population of the gut microbiome. This was accompanied by elevations of LBP, which might point to the way the immune system responds to microbiome changes and the means by which such signals are linked to changes in the brain. HFD mice displayed several behavioural changes relative to STD controls, such as significantly longer times spent immobile in the FST, however these changes were not influenced by MINO treatment. MINO did however result in changes in the gut microbial community.
4.1 HFD induced behavioural changes The HFD mice exhibited an increase in immobility time in the FST, which is consistent with other studies (Sharma et al., 2013; Yamada et al., 2011). Other studies, however, demonstrated less conclusive effects of a high fat diet (Andre et al., 2014; Krishna et al., 2016; Pyndt Jorgensen et al., 2014; Takase et al., 2016), which is possibly due to the large variability in the proportion and type of fat, timing and duration of feeding, animal age and sample size. For instance, the age of animals and the timing of HFD feeding in our study aligns with the Sharma et al. (2013) and Yamada et al. (2011) studies. This consistency of timing and age among the three studies may indicate that there may be a vulnerable age window by which HFD can mediate specific depressive-like behaviours. For the SPT, although a trend existed for our HFD mice to display less sucrose preference compared with STD mice, this did not reach statistical significance, which does not support the presence of anhedonia.
Potential interpretations of sucrose preference data as 25
correlating with anhedonia need to be considered cautiously however, as chronic HFD feeding may alter taste perception and reward. Anhedonia, or reduced capacity to experience hedonic pleasure in response to previously rewarding stimuli, is one of the core symptoms of MDD (Krishnan and Nestler, 2011). Anhedonic responses to a broad range of stimuli (ranging from social to gustatory and others) are often present in MDD patients, which suggests persistent dysregulation of common or final downstream reward systems such as the mesolimbic dopaminergic pathway, irrespective of the upstream stimulus. However, previous taste exposures are able to specifically influence the perception and rewarding palatability of subsequent taste stimuli, without altering other non-gustatory rewards more generally. For example, prior exposure to sweet tastes can habituate animals such that subsequent sweet stimuli are less rewarding (De Luca, 2014). Furthermore, internal homeostatic conditions are powerful modulators of appetitive behaviour, taste perception and reward, and this variable was also likely to be different between STD and HFD mice. Interestingly, several other studies have observed HFD-induced changes in sucrose preference, regardless of mouse strain, animal age, dietary fat content or duration of HFD feeding (Sharma et al., 2013; Takase et al., 2016; Yamada et al., 2011). However, it remains relatively unknown whether these behavioural changes indicate specific differences in taste and consummatory behaviour specifically, or anhedonia more broadly (Lockie et al., 2015).
We observed that HFD induced an increase in moving speed but a decrease in mobility during the OFT, although HFD significantly decreased travel distance. Importantly, correlation analysis found no conclusive evidence that these behavioural changes were attributable to fat mass itself or motor impairment due to obesity, and hence may instead 26
represent the effects of neurological changes. This is also in agreement with the similar conclusion that the altered behaviour in the FST was not due to physical obesity-related changes in swimming capacity. HFD influenced changes in locomotor activity have also been reported in previous experiments, with some studies reporting that HFD increased locomotor activity, including a longer travel distance than controls (Krishna et al., 2015; Sharma and Fulton, 2013).
4.2 Effects of HFD on the gut microbiota HFD feeding of mice resulted in changes in the gut microbiota that were characterised by significant reductions in the richness of the microbial community (alpha diversity), alteration of beta diversity and changes in F/B ratio. These changes were aligned with numerous studies (Boulangé et al., 2016; Cani and Delzenne, 2010; Geurts et al., 2011; Turnbaugh et al., 2008). On the other hand, Pyndt Jorgensen et al reported that HFD-induced changes in gut microbiome was not necessarily associated with depressive-like behaviours, such as anhedonia and behavioural despair (Pyndt Jorgensen et al., 2014). One possible reason contributing to the difference between our study and the Pyndt Jorgensen et al study might be a difference in behavioural reactivity in the FST by different mouse strains (Lucki et al., 2001).
4.3 Effects of MINO on diet-induced behavioural changes MINO significantly increased the time spent in the centre zone in HFD mice of the OFT. Given that MINO did not affect the distance travelled, duration of mobility or speed in the OFT, the increased time that MINO treated HFD mice spent in the centre region of the OFT 27
indicates that this change is not simply due to increased locomotor activity, and may instead reflect a possible reduction in anxiety-like behaviour. However, the interpretation of increased time spent in the centre zone of an OFT indicating reduced anxiety is often criticised as an oversimplification, which has limited validity and often lacks robust reproducibility (Spruijt et al., 2014; Walsh and Cummins, 1976). Thus, other behavioural tests which are more suitable to assess anxiety-like behaviour, such as the light/dark box or the elevated plus maze, may be required in future investigations to determine whether reduced anxiety-like behaviour is truly present. Lastly, MINO did not affect the duration of FST immobility nor sucrose consumption. Therefore, we were unable to confirm the therapeutic effects of MINO on our diet-induced model of behavioural change.
4.4 Effects of MINO on the gut microbiota Our study showed significant shifts in the gut microbial population in response to MINO treatment. MINO treatment resulted in a significant reduction in the Chao 1 richness index and membership of abundant bacteria within the composition of the gut microbial community in the STD/MINO and HFD/MINO groups of mice. Weighted UniFrac PCoA analysis showed clear separation between the four treatment groups, indicating that both diet and MINO exerted significant effects on the microbial communities. Intriguingly, MINO mediated significant changes in membership of the abundant bacteria and the gut microbiome community profiles in both HFD and control animals, rather than restoring the profile of the HFD fed mice to the original state. These results were confirmed using AMOVA on weighted UniFrac distances, which indicated significant group differences in the microbial communities (p = 0.001). Our results thus suggest that MINO might have differential effects on STD and HFD fed gut microbiota. The findings here suggest that 28
different key populations in the STD and HFD fed gut microbiota might be affected by MINO treatment and in turn, may result in different alterations in the balance of the memberships of abundant bacteria between the STD and HFD fed gut microbiome communities. This intriguing finding needs to be further investigated by exploring what drives the differential effects of MINO, as well as evaluating deeper levels of bacterial compositions, such as class or species levels, in future studies. Surprisingly, despite the clinical use of MINO as an antibiotic, the specific effects of MINO on the gut microbiota have been documented in very few studies. A previous study showed that MINO treatment (via intraperitoneal injection, 5mg/kg; 21 days) in parallel with chronic restraint stress prevented stress-induced depressive-like behaviour in mice in the FST (Wong et al., 2016). Our study had different experimental designs and hypotheses from Wong et al (2016), and therefore the findings are not readily comparable. However, both studies demonstrated that MINO did not restore the gut microbiome to the original state in the time frame of this study. The impacts of antibiotic use of the gut microbiome have been shown to have lasting and varying impacts over time; with some studies showing changes lasting up to 12 months (Dethlefsen et al., 2008; Jernberg et al., 2010). Different antimicrobial agents can influence the normal gut microbiome in divergent ways. MINO exerts a significant impact on a wide range of bacteria, however, it appears that the bacteriostatic effect of MINO mediates a shift in the balance of the gut microbiota; as the growth of existing bacteria is slowed, opportunities are provided for other bacteria to grow (Kim et al., 2017). MINO did not restore the original ecosystem in mice which had been exposed to a high fat diet. Hence, the gut microbiota population elicited by MINO treatment is different from both healthy (control) and diet-affected states. This raises the question of
29
which underlying pathways link the changes in the gut microbiome elicited by MINO to health, disease, and behaviour in the host.
5. Conclusion Our study demonstrated that high fat diet consumption resulted in significant changes in both behaviour and the gut microbial community. Our study also highlighted that MINO treatment significantly altered gut microbial profiles. MINO altered the balance of the gut microbiome in both mice fed standard chow and a high fat diet, as indicated by reductions of bacterial abundance profiles and increases in the F/B ratio, but did not push the microbiome profile of the HFD group back towards that of the control group. However, MINO-induced alterations of the gut microbial community were not associated with significant modulation of the behavioural changes induced by the HFD. Our findings indicate that MINO treatment can exert striking changes in the gut microbiome profile, but also that these changes do not necessarily correlate with changes in behaviour. Together, this data highlights the complex relationships between gut microbial profiles and behaviour, and indicates that further exploration of both pharmacological and dietary factors is warranted.
6. Acknowledgement We thank Drs Martin O’Hely, Fiona M Collier, Bruna Panizzutti and Chiara Bortolasci for technical assistance.
7. Funding This research was funded in part by the Rebecca L. Cooper Medical Research Foundation. OMD is supported by an R.D. Wright Biomedical NHMRC Research Fellowship (1145634).
30
8. Disclosure OMD has received grant support from the Brain and Behavior Foundation, Simons Autism Foundation, Stanley Medical Research Institute, Deakin University, Lilly, NHMRC and ASBDD/Servier. She has also received in kind support from BioMedica Nutracuticals, NutritionCare and Bioceuticals. Other authors declare no competing interests.
31
9. References Akbaraly, T.N., et al., 2009. Dietary pattern and depressive symptoms in middle age. Br J Psychiatry. 195, 408-13. Alexander, C., Rietschel, E.T., 2001. Bacterial lipopolysaccharides and innate immunity. J Endotoxin Res. 7, 167-202. Andre, C., et al., 2014. Diet-induced obesity progressively alters cognition, anxiety-like behavior and lipopolysaccharide-induced depressive-like behavior: focus on brain indoleamine 2,3-dioxygenase activation. Brain Behav Immun. 41, 10-21. Boulangé, C.L., et al., 2016. Impact of the gut microbiota on inflammation, obesity, and metabolic disease. Genome Medicine. 8, 42. Brown, K., et al., 2012. Diet-Induced Dysbiosis of the Intestinal Microbiota and the Effects on Immunity and Disease. Nutrients. 4, 1095-1119. Bruce-Keller, A.J., et al., 2015. Obese-type gut microbiota induce neurobehavioral changes in the absence of obesity. Biol Psychiatry. 77, 607-15. Cani, P.D., Delzenne, N.M., 2010. Involvement of the gut microbiota in the development of low grade inflammation associated with obesity: focus on this neglected partner. Acta Gastroenterol Belg. 73, 267-9. Caporaso, J.G., et al., 2010. QIIME allows analysis of high-throughput community sequencing data. Nat Meth. 7, 335-336. Cepeda, M.S., Stang, P., Makadia, R., 2016. Depression Is Associated With High Levels of CReactive Protein and Low Levels of Fractional Exhaled Nitric Oxide: Results From the 2007-2012 National Health and Nutrition Examination Surveys. J Clin Psychiatry. 77, 1666-1671.
32
de La Serre, C.B., et al., 2010. Propensity to high-fat diet-induced obesity in rats is associated with changes in the gut microbiota and gut inflammation. Am J Physiol Gastrointest Liver Physiol. 299, G440-8. De Luca, M.A., 2014. Habituation of the responsiveness of mesolimbic and mesocortical dopamine transmission to taste stimuli. Front Integr Neurosci. 8, 21. Dean, O.M., et al., 2012. Minocycline: therapeutic potential in psychiatry. CNS Drugs. 26, 391-401. Dean, O.M., et al., 2017. Adjunctive minocycline treatment for major depressive disorder: A proof of concept trial. Aust N Z J Psychiatry. 51, 829-840. Dethlefsen, L., et al., 2008. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLoS Biol. 6, e280. Edgar, R.C., 2013. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat Meth. 10, 996-998. Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of Molecular Variance Inferred from Metric Distances among DNA Haplotypes: Application to Human Mitochondrial DNA Restriction Data. Genetics. 131, 479-491. Garate, I., et al., 2011. Origin and consequences of brain Toll-like receptor 4 pathway stimulation in an experimental model of depression. J Neuroinflammation. 8, 151. Garrido-Mesa, N., Zarzuelo, A., Galvez, J., 2013. Minocycline: far beyond an antibiotic. Br J Pharmacol. 169, 337-52. Geurts, L., et al., 2011. Altered Gut Microbiota and Endocannabinoid System Tone in Obese and Diabetic Leptin-Resistant Mice: Impact on Apelin Regulation in Adipose Tissue. Frontiers in Microbiology. 2, 149.
33
Henry, C.J., et al., 2008. Minocycline attenuates lipopolysaccharide (LPS)-induced neuroinflammation, sickness behavior, and anhedonia. J Neuroinflammation. 5, 15. Husain, M.I., et al., 2017. Minocycline as an adjunct for treatment-resistant depressive symptoms: A pilot randomised placebo-controlled trial. J Psychopharmacol. 31, 11661175. Jacka, F.N., et al., 2014. Dietary patterns and depressive symptoms over time: examining the relationships with socioeconomic position, health behaviours and cardiovascular risk. PLoS One. 9, e87657. Jernberg, C., et al., 2010. Long-term impacts of antibiotic exposure on the human intestinal microbiota. Microbiology. 156, 3216-3223. Kelly, J.R., et al., 2015. Breaking down the barriers: the gut microbiome, intestinal permeability and stress-related psychiatric disorders. Front Cell Neurosci. 9, 392. Kim, S., Covington, A., Pamer, E.G., 2017. The intestinal microbiota: Antibiotics, colonization resistance, and enteric pathogens. Immunological reviews. 279, 90-105. Klindworth, A., et al., 2013. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Research. 41, e1-e1. Krishna, S., et al., 2015. Neurochemical and electrophysiological deficits in the ventral hippocampus and selective behavioral alterations caused by high-fat diet in female C57BL/6 mice. Neuroscience. 297, 170-81. Krishna, S., et al., 2016. Time-dependent behavioral, neurochemical, and metabolic dysregulation in female C57BL/6 mice caused by chronic high-fat diet intake. Physiol Behav. 157, 196-208.
34
Krishnan, V., Nestler, E.J., 2011. Animal Models of Depression: Molecular Perspectives. Current Topics in Behavioral Neurosciences. 7, 121-147. Kubera, M., et al., 2013. A new animal model of (chronic) depression induced by repeated and intermittent lipopolysaccharide administration for 4months. Brain, Behavior, and Immunity. 31, 96-104. Liu, Y.N., et al., 2015. TNFalpha mediates stress-induced depression by upregulating indoleamine 2,3-dioxygenase in a mouse model of unpredictable chronic mild stress. Eur Cytokine Netw. 26, 15-25. Lockie, S.H., et al., 2015. Diet-induced obesity causes ghrelin resistance in reward processing tasks. Psychoneuroendocrinology. 62, 114-120. Lozupone, C., Knight, R., 2005. UniFrac: a New Phylogenetic Method for Comparing Microbial Communities. Applied and Environmental Microbiology. 71, 8228-8235. Lozupone, C.A., Knight, R., 2008. Species Divergence and the Measurement of Microbial Diversity. FEMS microbiology reviews. 32, 557-578. Lucki, I., Dalvi, A., Mayorga, A.J., 2001. Sensitivity to the effects of pharmacologically selective antidepressants in different strains of mice. Psychopharmacology (Berl). 155, 315-22. Maes, M., et al., 2012. Increased IgA and IgM responses against gut commensals in chronic depression: further evidence for increased bacterial translocation or leaky gut. J Affect Disord. 141, 55-62. Majidi, J., Kosari-Nasab, M., Salari, A.A., 2016. Developmental minocycline treatment reverses the effects of neonatal immune activation on anxiety- and depression-like behaviors, hippocampal inflammation, and HPA axis activity in adult mice. Brain Res Bull. 120, 1-13. 35
McKim, D.B., et al., 2016. Neuroinflammatory Dynamics Underlie Memory Impairments after Repeated Social Defeat. J Neurosci. 36, 2590-604. Miyaoka, T., et al., 2012. Minocycline as adjunctive therapy for patients with unipolar psychotic depression: an open-label study. Prog Neuropsychopharmacol Biol Psychiatry. 37, 222-6. Neale, E.P., Batterham, M.J., Tapsell, L.C., 2016. Consumption of a healthy dietary pattern results in significant reductions in C-reactive protein levels in adults: a meta-analysis. Nutr Res. 36, 391-401. Paulson, J.N., et al., 2013. Differential abundance analysis for microbial marker-gene surveys. Nat Meth. 10, 1200-1202. Price, M.N., Dehal, P.S., Arkin, A.P., 2010. FastTree 2 – Approximately Maximum-Likelihood Trees for Large Alignments. PLOS ONE. 5, e9490. Pyndt Jorgensen, B., et al., 2014. A possible link between food and mood: dietary impact on gut microbiota and behavior in BALB/c mice. PLoS One. 9, e103398. Rahe, C., et al., 2015. Associations between depression subtypes, depression severity and diet quality: cross-sectional findings from the BiDirect Study. BMC Psychiatry. 15, 38. Remus, J.L., Dantzer, R., 2016. Inflammation Models of Depression in Rodents: Relevance to Psychotropic Drug Discovery. The international journal of neuropsychopharmacology. 19, pyw028. Rienks, J., Dobson, A.J., Mishra, G.D., 2013. Mediterranean dietary pattern and prevalence and incidence of depressive symptoms in mid-aged women: results from a large community-based prospective study. Eur J Clin Nutr. 67, 75-82. Rosenblat, J.D., McIntyre, R.S., 2018. Efficacy and tolerability of minocycline for depression: A systematic review and meta-analysis of clinical trials. J Affect Disord. 227, 219-225. 36
Sanchez-Villegas, A., et al., 2009. Association of the Mediterranean dietary pattern with the incidence of depression: the Seguimiento Universidad de Navarra/University of Navarra follow-up (SUN) cohort. Arch Gen Psychiatry. 66, 1090-8. Sanchez-Villegas, A., et al., 2012. Fast-food and commercial baked goods consumption and the risk of depression. Public Health Nutr. 15, 424-32. Schwingshackl, L., Hoffmann, G., 2014. Mediterranean dietary pattern, inflammation and endothelial function: a systematic review and meta-analysis of intervention trials. Nutr Metab Cardiovasc Dis. 24, 929-39. Sharma, S., Fernandes, M.F., Fulton, S., 2013. Adaptations in brain reward circuitry underlie palatable food cravings and anxiety induced by high-fat diet withdrawal. Int J Obes (Lond). 37, 1183-91. Sharma, S., Fulton, S., 2013. Diet-induced obesity promotes depressive-like behaviour that is associated with neural adaptations in brain reward circuitry. Int J Obes (Lond). 37, 382-9. Slyepchenko, A., et al., 2017. Gut Microbiota, Bacterial Translocation, and Interactions with Diet: Pathophysiological Links between Major Depressive Disorder and NonCommunicable Medical Comorbidities. Psychother Psychosom. 86, 31-46. Spruijt, B.M., et al., 2014. Reproducibility and relevance of future behavioral sciences should benefit from a cross fertilization of past recommendations and today's technology: "Back to the future". J Neurosci Methods. 234, 2-12. Takase, K., et al., 2016. High-fat diet feeding alters olfactory-, social-, and reward-related behaviors of mice independent of obesity. Obesity (Silver Spring). 24, 886-94. Turnbaugh, P.J., et al., 2008. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe. 3, 213-23.
37
Turnbaugh, P.J., et al., 2009. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med. 1, 6ra14. Vichaya, E.G., et al., 2018. Motivational changes that develop in a mouse model of inflammation-induced depression are independent of indoleamine 2,3 dioxygenase. Neuropsychopharmacology. Walsh, R.N., Cummins, R.A., 1976. The Open-Field Test: a critical review. Psychol Bull. 83, 482-504. Wong, M.L., et al., 2016. Inflammasome signaling affects anxiety- and depressive-like behavior and gut microbiome composition. Mol Psychiatry. 21, 797-805. Yamada, N., et al., 2011. Impaired CNS leptin action is implicated in depression associated with obesity. Endocrinology. 152, 2634-43. Yu, Z.M., Parker, L., Dummer, T.J., 2014. Depressive symptoms, diet quality, physical activity, and body composition among populations in Nova Scotia, Canada: report from the Atlantic Partnership for Tomorrow's Health. Prev Med. 61, 106-13. Zheng, L.S., Kaneko, N., Sawamoto, K., 2015. Minocycline treatment ameliorates interferonalpha- induced neurogenic defects and depression-like behaviors in mice. Front Cell Neurosci. 9, 5. Zheng, P., et al., 2016. Gut microbiome remodeling induces depressive-like behaviors through a pathway mediated by the host's metabolism. Mol Psychiatry. 21, 786-96.
38
Highlights
Long-term high fat diet consumption was associated with behavioural changes but not depressive-like behaviour in male C57BL/6J mice.
Chronic high fat diet feeding mediated significant alteration of the gut microbiome and behaviour in the host.
Minocycline treatment altered gut microbiome profiles (Chao1 richness index, Principal coordinate analysis and relative abundance) were different from both healthy and diet-affected mice.
Our study first showed minocycline-induced changes in the gut microbiome do not necessarily correlate with changes in behaviour.
39