If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, GermanyGerman Center for Diabetes Research (DZD), Neuherberg, GermanyUniversity of Potsdam, Institute of Nutritional Science, Molecular and Experimental Nutritional Medicine, Nuthetal, Germany
German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, GermanyUniversity of Potsdam, Institute of Nutritional Science, Molecular and Experimental Nutritional Medicine, Nuthetal, Germany
German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, GermanyGerman Center for Diabetes Research (DZD), Neuherberg, GermanyUniversity of Potsdam, Institute of Nutritional Science, Molecular and Experimental Nutritional Medicine, Nuthetal, Germany
German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, GermanyGerman Center for Diabetes Research (DZD), Neuherberg, GermanyUniversity of Potsdam, Institute of Nutritional Science, Molecular and Experimental Nutritional Medicine, Nuthetal, Germany
Address correspondence to: André Kleinridders, Department of Molecular and Experimental Nutritional Medicine, Institute of Nutritional Science, University of Potsdam, Arthur-Scheunert-Allee 114-116, 14558 Nuthetal, Germany, Phone: 49.33200.885230
German Institute of Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, GermanyGerman Center for Diabetes Research (DZD), Neuherberg, GermanyUniversity of Potsdam, Institute of Nutritional Science, Molecular and Experimental Nutritional Medicine, Nuthetal, Germany
Diabetic patients exhibit an increased prevalence for emotional disorders compared to healthy humans, partially due to a shared pathogenesis including hormone resistance and inflammation, which is also linked to intestinal dysbiosis. The preventive intake of probiotic lactobacilli has been shown to improve dysbiosis along with mood and metabolism. Yet, a potential role of lactobacillus rhamnosus (Lacticaseibacillus rhamnosus 0030, LR) in improving emotional behavior in established obesity and the underlying mechanisms are unknown.
Methods
Female and male C57BL/6N mice were fed a low-fat diet (LFD, 10%kcal from fat) or high-fat diet (HFD, 45%kcal from fat) for 6 weeks, followed by daily oral gavage of vehicle or 1x108 CFU of LR, and assessment of anxiety- and depressive-like behavior. Cecal microbiota composition was analyzed using 16S rRNA Sequencing, plasma and CSF were collected for metabolomic analysis, and gene expression of different brain areas was assessed using RT-qPCR.
Results
We observed that 12 weeks of HFD feeding induced hyperinsulinemia which was attenuated by LR application only in females. On the contrary, HFD-fed male mice exhibited increased anxiety and depressive-like behavior, where the latter was specifically attenuated by LR application which was independent of metabolic changes. Furthermore, LR application restored the HFD-induced decrease of tyrosine hydroxylase, along with normalizing cholecystokinin gene expression in dopaminergic brain regions, both signaling pathways impacting emotional disorders.
Conclusion
Our data show that LR attenuates depressive-like behavior after established obesity with changes in the dopaminergic system in male mice and mitigates hyperinsulinemia in obese females.
Across the globe, diabetes, obesity, and depression are still on the rise. While obesity and diabetes represent metabolic disorders, depression is a mental illness and a leading cause for disability, which affects about 300 million people worldwide. Being obese already at a young age increases the risk to lower self-esteem (
). While there are multiple causes for the association between obesity, diabetes and emotional disorders, research progress over the last decade has established a potential common metabolic basis for these disorders. It has been proposed that depression and metabolic disorders share inflammation, oxidative stress and hormone resistance due to mutually dysregulated signaling pathways. Obesity and diabetes are promoted in the presence of insulin resistance, which causes deteriorated glucose metabolism, hyperglycemia and uncontrolled lipolysis. Disrupted brain insulin signaling, induced by genetic deletion of the insulin receptor (IR) using the Nestin-Cre or the GFAP-Cre mouse, causes behavioral abnormalities with depressive-like behavior (
). The intake of high caloric, high fat diet (HFD) with elevated levels of saturated long chain fatty acids, is sufficient to induce brain insulin resistance (
), disorders characterized by abnormal dopamine signaling. Interestingly, HFD feeding in rodents reduces the expression of TH, the rate-limiting enzyme in dopamine synthesis (
). HFD-induced alterations of the microbiome are linked to emotional disorders and affect dopaminergic neurotransmission. Accordingly, treating HFD-fed mice with an antibiotic cocktail, to erase an altered gut microbiota composition induced by HFD, attenuates anxiety and depressive-like behavior (
Additionally, the use of pre- and probiotics to support a healthy microbiota composition have been tested to prevent the establishment of metabolic deteriorations and cognitive function (
). Lactobacilli represent a prominent member of probiotics, which are part of many dietary products that claim to be “good for metabolism”. The preventive supplementation of different lactobacillus strains during obesity development can protect against metabolic dysfunction and improve mood (
). However, whether sex-specific differences exist or such a supplementation alters metabolism or behavior in established obesity has not been tested.
In this study we investigated the effect of lactobacillus rhamnosus (Lacticaseibacillus rhamnosus 0030, LR) on emotional behavior in diet-induced obese mice. We show that a daily oral gavage for a minimum of six weeks reduces depressive-like behavior in male mice with established obesity. Moreover, this is linked to elevated gene expression of TH and a reversal of HFD-altered CCK mRNA levels in the nucleus accumbens. In females, HFD feeding causes hyperactivity, which can be mitigated by LR application. Furthermore, LR utilization attenuates HFD-induced hyperinsulinemia in female mice.
Overall our data show that the daily application of LR is able to reverse HFD-induced behavioral alterations in males, alters behavior-related gene expression, and further counteracts HFD-induced hyperinsulinemia in females.
Methods and Materials
See the Supplement for full animal study description and methods relating to metabolic and behavioral phenotyping, final procedures, along with molecular analyses and ‘omics’ analyses.
Animals
C57BL/6N female and male mice were obtained from Charles River Laboratories (Sulzfeld, Germany) and group-housed in a temperature-controlled room (22±1°C) on a 12-h light-dark cycle with free access to food and water. Mice were fed with either semi-synthetic low-fat diet (LFD, 10% of kcal from fat) or high-fat diet (HFD, 45% of kcal from fat) (both from Research Diets, Inc.). From week 12 of age until the end of experiments, all mice received a daily peroral gavage of either 100μl vehicle (phosphate-buffered saline, PBS, Gibco) or 100μl Lactobacillus rhamnosus (Lacticaseibacillus rhamnosus 0030, LR) (1x108 CFU of LR in PBS) to ensure accurate dosage.
Behavior tests
Behavioral assessment was performed weekly starting at week 14 of age. All tests were conducted during the light cycle. The tests were performed in the following order: Open Field Test (OFT), Dark/Light Box Test (DLB), Elevated Plus Maze Test (EPM), and Mousetail Suspension Test (MST). In a naïve cohort of male mice, tests were performed in reverse order: MST, DLB, and the Splash Test. An animal study overview can be found in Suppl. Fig. 1A.
Figure 1Preventive intervention with Lactobacillus rhamnosus attenuates HFD-induced dysregulation of tyrosine hydroxylase in the caudate putamen. (A) Relative mRNA levels of tyrosine hydroxylase and (B,C) monoamine oxidase A/B in the caudate putamen of C57BL/6J male mice after 12 weeks of HFD. * P<0.05 after 1-way ANOVA with Dunnett’s Post-hoc test (n= 8-11). All data are presented relative to Tbp (2ΔCT) and as mean±SEM. LFD: low-fat diet, HFD: high-fat diet, LR: lactobacillus rhamnosus.
Total RNA was isolated from brain tissue using ReliaPrep RNA Tissue Miniprep System (Promega) according to manufacturer instructions. Following isolation, RNA was quantified by NanoDropTM Spectrophotometer (Thermo Fisher Scientific) and reverse-transcribed to cDNA using Thermo Fisher Scientific dNTP-Set, oligo(dT)15 primers, random hexameric primers, and M-MLV reverse transcriptase (all from Promega). RT-qPCR was performed using GoTaq 1-Step RT-qPCR System mix (Promega) and 200nM of each forward and reverse primer (obtained from Sigma-Aldrich/Merck, see Supplementary Table S1). Fluorescence was monitored using the ViiA7 Real-Time PCR System (Applied BiosystemsTM, Thermo Fisher Scientific). Each run was followed by a melt curve (90°C to 60°C) for quality control. Samples were analyzed in duplicate and relative quantification of gene expression levels was performed according to the ΔΔCT method using TATA-box-binding protein (Tbp) as reference gene. Data were expressed as 2ΔΔCT and relative to the respective control group, if not stated otherwise.
Statistical analysis
General statistical analysis was performed using GraphPad Prism 9. Data is represented as mean±SEM. For comparison of two groups, unpaired two-tailed Student’s t-test was employed for parametric samples, Mann-Whitney U test was used for non-parametric samples. Comparison of multiple groups was performed by 1-way ANOVA (parametric samples) followed by Dunnett’s multiple comparison test to compare groups with a defined control group. To compare all groups with each other, Tukey’s multiple comparison test was utilized. For non-parametric samples, Kruskal-Wallis test and Dunn’s multiple comparison test were employed. Statistical significance is considered as P value<0.05 (*), P value<0.01 (**), P value<0.001 (***), P value<0.0001 (****).
Results
Preventive application of LR counteracts high fat diet feeding-induced reduction of TH gene expression
To determine whether diet-induced obesity and a preventive application of probiotics can alter dopamine homeostasis, C57BL/6J male mice were fed a semi-synthetic low-fat diet (LFD) or HFD for 12 weeks and in parallel, received daily applications of either PBS (vehicle) as control or lactobacillus rhamnosus (Lacticaseibacillus rhamnosus 0030, LR) (now referred to as HFD LR). Interestingly, LR attenuated the HFD-induced reduction of the rate-limiting enzyme of dopamine synthesis, in a dopaminergic brain region. While HFD feeding resulted in a ∼42% decrease in TH mRNA levels in the striatum compared to LFD vehicle, HFD LR treatment attenuated this decrease, revealing indistinguishable gene expression levels of TH compared to LFD (Fig. 1A). Yet dopamine degrading enzymes monoaminoxidases A and B were unaffected (Fig. 1B, C).
HFD-induced metabolic deteriorations in female and male mice prior to LR application
Leading on from this, we examined whether the LR application can counteract emotional alterations in established obese conditions. Moreover, it is unclear whether behavioral alterations in obesity are caused by an overall excess of calorie intake or can be specifically linked to increased intake of lipids and fatty acids. To answer both questions, we fed male and female C57BL/6N mice a semi-synthetic low-fat diet (LFD) as control and HFD to establish diet-induced obesity.
After the initial six weeks of HFD feeding, female HFD-fed mice weighed more than LFD control, exhibited a minor, yet significant increase in blood glucose and showed unaltered plasma insulin levels, while HFD-fed males exhibited increased body weight and hyperinsulinemia, but no changes in blood glucose (Suppl. Fig. 2A-C, Suppl. Fig. 4A-C). After randomization of HFD-fed mice, half of all HFD-fed mice received a daily oral gavage of LR for an additional six weeks, while the second group received PBS as control (HFD vehicle).
Figure 2LR application regulates metabolism of obese female mice independent of changes in body weight. (A) Body weight and (B) adipose tissue weights of female mice after 13 weeks of HFD. (C) Plasma leptin levels of female mice after 10 weeks of HFD. (D) Blood glucose levels and (E) adrenal glands weight of female mice after 13 weeks of HFD. (F) Plasma corticosterone, (G) plasma epinephrine, and (H) plasma insulin levels of female mice after 10 weeks of HFD. * P<0.05, ** P<0.01, *** P<0.001 and **** P<0.0001 after 1-way ANOVA with Dunnett’s Post-hoc test (A,B,E,G) or after Kruskal-Wallis test with Dunn’s Post-hoc test (C,H) (n= 18-20). All data are presented as mean±SEM.
Figure 4LR intervention does not modulate metabolism of obese male mice. (A) Body weight and (B) white adipose tissue (WAT) weights of male mice after 13 weeks of HFD. (C) Plasma leptin levels of male mice after 10 weeks of HFD. (D) Blood glucose levels male mice after 13 weeks of HFD. (E) Plasma insulin levels of male mice after 10 weeks of HFD. *** P<0.001 and **** P<0.0001 after 1-way ANOVA with Dunnett’s Post-hoc test (A-D) or after Kruskal-Wallis test with Dunn’s Post-hoc test (B: gonadal WAT) (n= 15-19). * P<0.05 after Mann-Whitney U test (E). All data are presented as mean±SEM.
LR application after established obesity reduces plasma insulin levels in female mice but does not alter depressive-like behavior
HFD feeding continued to increase body weight in female mice, and caused elevated fat mass along with increased leptin levels compared to LFD, but did not alter other organ weights or blood glucose levels between all tested groups (Fig. 2A-D, Suppl. Fig. 2D-F). In females HFD feeding caused an increase in weight of adrenal glands compared to LFD, while LR application attenuated this increase, leading to similar adrenal gland weights as in LFD mice (Fig. 2E). Yet, basal corticosterone levels were indistinguishable between all groups despite increased plasma epinephrine levels in LR treated females (Fig. 2F, G). In line with increased adrenal gland weight, HFD feeding resulted in elevated plasma insulin levels compared to LFD control. Again, the daily application of LR mitigated HFD-induce hyperinsulinemia, showing beneficial effects on insulin sensitivity (Fig. 2H).
Unaltered depressive-like behavior in LR-treated, obese females
Next, we assessed emotional behavior in lean and obese mice as well as in obese mice treated with LR. Firstly, we assessed explorative behavior using an open field test (OFT) in females which were fed a HFD for 8 weeks and received LR for two weeks. HFD feeding caused hyperactivity in females compared to LFD determined by increased distance traveled, mean speed during the OFT and overall mobility time (Fig. 3A-C). LR application was able to mitigate both HFD-induced increase in distance and mean speed, but did not affect overall mobility time. Interestingly, while locomotor activity was increased by HFD, exploration was unaffected by dietary interventions as shown by an indistinguishable amount of time spent in the center of the field between tested groups (Fig. 3D, E). Moreover, HFD feeding neither affected anxious behavior as determined by the Elevated Plus Maze Test (EPM) and the Dark/Light Box Test (DLB), nor did it change depressive-like behavior, which was examined by the Mousetail Suspension Test (MST) in female mice (Suppl. Fig. 3). Nevertheless, LR application increased the time spent in the open compartment during the EPM, indicating increased risk behavior (Suppl. Fig. 3B). Thus, HFD feeding induces hyperactivity in females, while LR treatment does not grossly alter emotional behavior in established HFD-induced obesity.
Figure 3LR application regulates hyperactivity in diet-induced obese female mice. (A) Distance, (B) mean speed, and (C) active time of female mice after 8 weeks of HFD during the Open Field Test. (D) Exemplary track plots and (E) time spent in the center of the field of female mice after 8 weeks of HFD during the Open Field Test. Grey area represents the center of the field. * P<0.05 and ** P<0.01 after 1-way ANOVA with Dunnett’s Post-hoc test (n= 19-20). All data are presented as mean±SEM.
LR application does not impact metabolism in obese male mice
Unexpectedly and different to females, LR treatment did not attenuate or reverse any metabolic effects of HFD feeding in males, displaying indistinguishable elevated body weight, obesity, hyperleptinemia and hyperinsulinemia, while also not changing blood glucose levels or organ weights compared to HFD males (Fig. 4A-E, Suppl. Fig. 4D-H). However, similar to female mice, LR increased plasma epinephrine levels in males without affecting plasma corticosterone concentrations (Suppl. Fig. 4G, I). Additionally, LR treatment did not alter markers of inflammation in diet-induced obesity (Suppl. Fig. 4J).
Depressive-like behavior in obese males is attenuated by LR application
In contrast to the females, two weeks of LR application did not change activity during an OFT in HFD-fed male mice, and explorative behavior also remained unaltered (Suppl. Fig. 5A-E). Furthermore, LFD vehicle, HFD vehicle and HFD LR fed mice did not differ significantly in the time they spent in the light compartment in a DLB, showing that LR did not alter anxious behavior in this experimental setting (Fig. 5A, Suppl. Fig. 5F). Similarly, stress-induced anxiety was unaffected, as all groups traveled a similar distance and spent similar amount of time in open arms during an elevated X maze along with unaltered corticosterone levels after the test (Suppl. Fig. 5G-J).
Figure 5LR intervention alleviates aspects of HFD-induced depressive-like behavior but does not regulate anxiety. (A) Time spent in the light compartment of male mice after 9 weeks of HFD during the Dark Light Box (DLB) Test. (B) Immobility time of male mice after 12 weeks of HFD during the Mousetail Suspension Test (MST). (C) Immobility time of male mice after 9 weeks of HFD during the MST in a naïve cohort. (D) Time spent in the light compartment of male mice after 12 weeks of HFD during the DLB Test in a naïve cohort. (E) Latency to initiate grooming of male mice after 13 weeks of HFD during the Splash Test in a naïve cohort. Grey area highlights behavior tests which were conducted in week 12 of HFD-feeding. * P<0.05 after 1-way ANOVA with Dunnett’s Post-hoc test (B,E) or after Mann-Whitney U test (D) (n= 13-20). All data are presented as mean±SEM.
Moreover, after 12 weeks of HFD feeding, LFD vehicle, HFD vehicle and HFD LR fed male mice differed significantly in their immobility time during a MST, as an indicator of altered emotional behavior. HFD-fed mice receiving daily oral gavage of LR for six weeks, displayed reduced depressive-like behavior compared to HFD vehicle, as assessed by a ∼22% decrease in immobility time during the test (Fig. 5B).
LR application affected depressive-like behavior but, unexpectedly, not anxiety (Fig. 5A, B). As the LR treatment was three weeks shorter during the assessment of anxiety using a DLB compared to the MST, it raised the possibility that only a prolonged LR application was able to modulate anxious behavior. Thus, anxiety and depressive-like behavior were analyzed in a second naïve cohort in a reversed chronological order for the DLB and the MST. Feeding male mice a HFD for only 9 weeks with or without daily application of LR for three weeks revealed no differences in immobility time between all three tested groups (Fig. 5C). After six weeks of daily LR application and 12-weeks of HFD feeding, only HFD feeding (combination of HFD and HFD LR groups) decreased the time, which mice spent in the light compartment of the DLB, by ∼33% compared to LFD vehicle group, indicative of HFD-induced anxiety. Yet, there was no difference in time spent in the light compartment between HFD vehicle and the HFD LR male mice (Fig. 5D).
To confirm that prolonged LR application was still able to alter depressive-like behavior in males, we further assessed the motivation of self-care during a splash test one week later. HFD feeding increased the time of grooming latency in mice compared to LFD control, as LR treatment of HFD-fed mice attenuated this phenotype resulting in a similar grooming latency as the LFD control mice, showing that LR treatment specifically improves motivation and reduces HFD-induced depressive-like behavior in males (Fig. 5E).
LR treatment exhibits minor effects on microbiota composition in both sexes
To gain insights on how HFD feeding as well as the therapeutic application of LR were able to specifically modulate emotional behavior, gut microbiota composition was determined in cecal samples using 16S rRNA gene amplicon sequencing, revealing only minor alterations in microbiota diversity.
In females, diet (PvalShannon= 0.016 on phylum level) and LR (PvalShannon= 0.041 on phylum level) application caused minor, yet significant changes in alpha diversity, while in males, alpha diversity remained unchanged (data not shown). Beta diversity revealed again a sexual dimorphism highlighting the importance to investigate both sexes. While HFD feeding significantly changed cecal microbiota composition in both sexes, LR only modulated beta diversity in female mice (Fig. 6A, B). To get more insights into microbiota composition, we determined the differential abundances of operative taxonomic units (OTUs) in both sexes and between diets and treatment. In females, the abundance of 41 OTUs was significantly different due to HFD feeding. HFD predominantly increased the abundance of genera like Lactococcus, Oscillospira, Ruminococcus and Akkermansia, while the abundance of genera like Allobaculum und Coprococcus was decreased. In contrast, LR only changed the abundance of two OTUs in total, with increased Allobaculum otu84 abundance compared to the HFD group (Fig. 6C). In males, only HFD feeding caused significantly different abundances of 18 OTUs, including the genera Allobaculum (which was decreased) and Oscillospira (which was increased) (Fig. 6D).
Figure 6HFD intervention but not LR application influences cecal microbiota composition. (A) PCA plot of Unifrac distances on species level visualizing the effect of diet on the cecal microbiota beta diversity (R2=0.246, P= 0.001) and PCA plot of Unifrac distances on species level visualizing the effect of LR on the cecal microbiota beta diversity (R2=0.093, P= 0.006) of female mice after 13 weeks of HFD. (B) PCA plot of Unifrac distances on species level visualizing the effect of diet (R2=0.18525, P= 0.028) and LR (R2=0.04523, P= 0.641) on the cecal microbiota beta diversity of male mice after 14 weeks of HFD. (C) Differential abundance analysis in female mice between LFD and HFD (’diet effect’) as well as HFD vs LR (’lactobacillus effect’). (D) Differential abundance analysis in male mice between LFD and HFD (’diet effect’; ratio= LFD/HFD) as well as HFD vs LR (’lactobacillus effect’; ratio= HFD/LR). (C,D) adjPvalue<0.05. Numbers within the graphs indicate individual animal IDs and correspond to individual samples. PCA: principal component analysis, OTU: Operative Taxonomic Unit.
), we only investigated the metabolome of plasma and CSF samples from male mice. In the CSF, 44 metabolites (including lipids species) were significantly altered by a log2-fold change between HFD and LFD (’diet effect’) with 10 metabolites being upregulated and 34 downregulated in HFD compared to LFD (Suppl. Fig. 6A). Additionally, 144 metabolites were significantly different between LR and HFD (’lactobacillus effect’), with 46 upregulated and 98 downregulated metabolites in the LR compared to HFD groups (Suppl. Fig. 6A). A similar result was observed in the plasma. Here, the abundance of 78 metabolites was found to be different between HFD and LFD, with 29 up- and 49 downregulated in HFD. Between LR and HFD, the abundance of only 30 metabolites was significantly different with 24 up- and six downregulated and only one metabolite appeared in both comparisons which was not annotated (Fig. 7A). To gain more insights about overall regulation, a KEGG pathway enrichment analysis (PAE) of the CSF and plasma metabolome was only performed on annotated metabolites for the individual comparisons (HFD vs LFD, LR vs HFD) (p<0.05). This analysis revealed that taste transduction (map04742) and glycerophospholipid metabolism (map00564) pathways were differentially regulated by diet or LR application in the CSF (Suppl. Fig. 6B). In contrast, HFD exposure predominantly altered metabolites in plasma which belong to fatty acid metabolism, whereas treatment with LR changed multiple pathways, including synthesis and degradation of branched-chain amino acids (BCAAs) (Fig. 7B). Based on these results, we specifically determined the relative abundance of single BCAAs in CSF and plasma. While diet compared to LR treatment had a stronger effect on BCAAs in CSF (Suppl. Fig. 6C), LR was the driving factor for reduced peripheral BCAA levels without altering Slc7a5 gene expression for BCAA uptake (Fig. 7C, Suppl. Fig. 6D). Although plasma serotonin was not detected in our unbiased metabolomic approach, the abundance of its precursor L-Tryptophan was decreased in HFD feeding, suggesting a modulation of serotonin production (Fig. 7D, Suppl. Fig. 7A). Overall, HFD and LR application caused mild alterations in serotonergic-related metabolites in male mice (Suppl. Fig. 7A, B).
Figure 7LR application but not HFD alters abundance of branched chain amino acids in the plasma of male mice. (A) Visualization of significant annotated and non-annotated plasma metabolites within and between both comparisons (‘diet effect’ and ‘lactobacillus effect’) of male mice after 13 weeks of HFD. (B) KEGG pathway analysis of annotated plasma metabolites and fatty acids of male mice after 13 weeks of HFD (adjPvalue<0.05 after Fisher’s Exact test). (C) Relative abundance of L-Valin, L-Leucin, L-Isoleucin, and (D) L-Tryptophan in plasma of male mice after 13 weeks of HFD. * P<0.05, ** P<0.01 and *** P<0.001 after 1-way ANOVA with Dunnett’s Post-hoc test (C) or unpaired two-tailed Student’s t-test (D) (n= 7-10). All data are presented as mean±SEM. ↑: significantly more abundant, ↓: significantly less abundant.
), we further analyzed gene expression of key enzymes in the caudate putamen (striatum), ventral tegmental area/ substantia nigra (VTA/SN), and nucleus accumbens (NAcc). HFD exposure did not affect gene expression of tyrosine hydroxylase (TH) or monoamine oxidase A and B (MaoA and B) in samples of CPu as well as VTA/SN from males (Suppl. Fig. 8A, B). Importantly, HFD decreased TH gene expression by ∼38% compared to LFD in the NAcc (Fig. 8A). Again, mRNA levels of MaoA and MaoB along with gene expression of dopamine transporter were not regulated (Fig. 8B, Suppl. Fig. 8C). The HFD-induced decrease of TH mRNA levels was reverted by LR treatment in NAcc (Fig. 8A), but this result could not be confirmed on protein level using western blot (Suppl. Fig. 8D) and immunohistochemistry analysis (Suppl. Fig. 8E). There was no direct correlation between Th mRNA levels and immobility time or body weight (Suppl. Fig. 9A-C). As Tryptophan hydroxylase mRNA levels, a marker for serotonergic metabolism, was also unaltered (Suppl. Fig. 9D-F), our data suggest that additional factors might contribute to this behavioral phenotype.
Figure 8LR attenuates HFD-reduction in tyrosine hydroxylase gene expression in the nucleus accumbens and reveals a signature of altered cholecystokinin expression. (A) mRNA expression of tyrosine hydroxylase (TH) and (B) monoamine oxidase A/B in the nucleus accumbens of male mice after 13 weeks of HFD using RT-qPCR. (C) Visualization of significant (Pvalue<0.05) differentially expressed genes (DEGs) within and between both comparisons (‘diet effect’ and ‘lactobacillus effect’) in the nucleus accumbens of male mice after 13 weeks of HFD (n= 6). (D) Significantly enriched biological processes (gene ontology) for all annotated genes of DEGs within and between both comparisons (‘diet effect’ and ‘lactobacillus effect’) of male mice after 13 weeks of HFD using Fisher’s Exact test and no correction for multiple testing (raw Pvalue). (E) Cholecystokinin (CCK) mRNA levels using RNA sequencing and (F) RT-qPCR for validation. (G) Correlation analysis of relative gene expression of Th and Cck. All data are presented relative to Tbp (2ΔCT). Continuous line represents the mean and the dotted line represents the error after linear regression analysis. * P<0.05 after 1-way ANOVA with Dunnett’s Post-hoc test (A,E,F) or Pearson correlation (G). All data are presented as mean±SEM.
To gain a better understanding of HFD and LR-induced alterations on gene expression patterns of the dopaminergic system, we further performed RNA sequencing analysis on NAcc samples of the different groups. The samples were matched and selected according to their variability in gene expression levels of TH, resulting in the same HFD-induced reduction by ∼36% compared to LFD control and displaying a similar variation in the LR group (n= 6 samples per group). HFD compared to LFD lead to 176 differentially expressed genes (DEGs; log2-fold change [log2FC]>|1.0|; raw p<0.05) (Fig. 8C). Unexpectedly, TH was not found among those DEGs because of a [log2FC]<|1.0| (HFDvsLFD [log2FC]=-0.6021 and LRvsHFD [log2FC]=0.5939). Nevertheless, TH mRNA levels were significantly decreased by HFD feeding (p= 0.0034 after unpaired two-tailed Student’s t-test) and there was a strong trend towards increased TH gene expression by LR application (p= 0.0942 after unpaired two-tailed Student’s t-test) (Suppl. Fig. 10A). Of the 176 differentially regulated genes between HFD and LFD, 153 had an annotated gene name, with 65 genes being upregulated in the HFD-fed group, and 88 genes were downregulated in HFD (Fig. 8C). To evaluate the effect of LR application in this context, DEGs between LR and HFD were determined. In total, 266 DEGs were identified of which 233 had an annotated gene name. Within those 233 DEGs, 189 were upregulated in the LR treated group while 44 were downregulated. Interestingly, 61 DEGs overlapped in both comparisons and were subsequently evaluated using a heatmap by plotting the respective log2FoldChange (Fig. 8C). Interestingly, all genes which were significantly downregulated by HFD feeding (compared to the LFD group) were significantly upregulated due to LR application (compared to HFD vehicle) and vice versa, suggesting a ’rescue effect’ by LR on gene expression level in HFD-fed male mice (Suppl. Fig. 10B). 58 out of 61 DEGs were successfully mapped to proteins (STRING protein network) and subsequent network analysis revealed altered regulation of neuropeptide hormone activity (GO:0005184), neuropeptide signaling pathway and peptide hormone binding (CL:19959), and neuroactive ligand-receptor interaction (mmu04080) (Suppl. Fig. 10C). To confirm this finding, gene names of DEGs between LFD and HFD as well as between HFD and LR were reanalyzed using GO Enrichment Analysis of the PANTHER Classification System, which revealed significant modulation of pathways involved in the regulation of behavior (Fig. 8D). Cholecystokinin (CCK) was one of five genes which were involved in the regulation of behavior and showed a significant upregulation of CCK gene expression in the HFD vehicle group compared to LFD (Fig. 8E). Interestingly, while Cck was upregulated in diet-induced obesity, LR application in HFD-fed mice normalized Cck levels (Fig. 8E). Further, Cckbr but not Cckar expression was significantly increased in HFD-fed mice which was not observed in mice receiving LR application (Suppl. Fig. 10D). Strikingly, we confirmed the dysregulation of Cck using a targeted qPCR approach between tested groups with significantly reduced CCK gene expression due to LR application in HFD-induced obesity (Fig. 8F). Lastly, correlation analysis revealed a significant negative correlation between relative expression of Th and Cck (Fig. 8G) (
). In summary, LR prevented HFD-induced increase of plasma insulin and adrenal gland weight in female C57BL/6N mice with established obesity, and modulated HFD-induced hyperactivity. In contrast, LR did not exert metabolic effects in diet-induced obese male mice, but specifically regulated depressive-like and goal-directed behavior with altered TH and CCK gene expression levels in the nucleus accumbens which are potential mediators of LR-induced attenuation of HFD-induced emotional alterations.
Discussion
The intake of an unhealthy high caloric diet causes a dysbiosis and is linked to an altered emotional behavior in humans (
). As this dietary intervention also impacts the microbiome, the preventive supplementation of probiotics has been shown to improve gut health, metabolism and emotional behavior in rodents. Yet, it is not well understood how efficient such an intervention in established obesity might be and how this influences the fat content and the quality of fat have in this context. We show that A) the increased content of unhealthy fat (lard) in the diet is sufficient to cause hyperactivity in females and depressive-like behavior in males while B) the supplementation of lactobacillus LR ameliorates HFD-induced depressive-like behavior in males and C) attenuates elevated insulin levels in diet-induced obesity in females.
In order to make arguments about the quality of fat, we used a semi-synthetic LFD as a control in our study, containing 10% calories from fat. Both diets (LFD and HFD) contained the same amount of sucrose and soybean oil, while the HFD (45% calories from fat) contained a higher amount of lard, indicating that the observed effects were not only dependent on a difference in fat (Δ fat= 35%) but also in lard content, which is rich in e.g. LCSFAs and omega-6 PUFAs. Palmitate has been shown to be especially detrimental for brain insulin sensitivity (
) in males. Thus, the observed impact on metabolism and emotional behavior can be allocated to a difference in fat content. This is important as the percentage of calories from fat along with different fat sources can exert different behavioral and metabolic effects. It has been shown that different dietary fat types cause specific alterations in anxiety- and depressive-like behavior (
). Thus, the moderate effect of feeding a 45% HFD on basic metabolic parameters and insulin sensitivity, compared to a semi-synthetic LFD accounts for the mild effect on emotional behavior. Furthermore, the biological sex influences emotional behavior in mice (
). While we show that in established obesity, lactobacillus rhamnosus supplementation is able to improve depressive-like behavior in males, HFD LR fed females exhibited unaltered insulinemia and adrenal glands weight compared to control, thus alleviating the detrimental effect of HFD-induced hyperinsulinemia. Reasons for this sex difference are so far unknown.
Although HFD feeding exhibited an increase in adrenal gland mass in females, both sexes showed elevated plasma epinephrine levels exclusively in the HFD LR group without affecting plasma corticosterone or blood glucose levels. Besides its classical role in stress signaling, epinephrine has been shown to improve cognitive function, memory for emotionally arousing experiences, depression, but also motivation (
), this might offer an explanation for the observed decrease in depressive-like behavior in males as both performed tests partially rely on differences in motivation and dopaminergic signaling (
In males, LR treatment did not affect insulin sensitivity in obesity but ameliorated depressive-like behavior. This is in contrast to the preventive application of lactobacillus in mice, which attenuates the development of diet-induced obesity with insulin resistance (
). In established obesity, LR treatment did not alter markers of inflammation in diet-induced obesity (Suppl. Fig. 4J), suggesting that the reversal of inflammation is needed to improve metabolism in male mice.
Another reason for the observed differences in modulating either behavior or metabolism relates to the duration of administration (diet and/or LR) in our experimental setting. Prolonged probiotic administration exerts more profound effects on health (
Long-term use of probiotics Lactobacillus and Bifidobacterium has a prophylactic effect on the occurrence and severity of pouchitis: a randomized prospective study.
). A time dependent effect is also known from drugs to treat psychiatric diseases. These drugs need to reach a certain threshold to induce their therapeutic effect (
). Similar data about differences in behavioral outcomes can also be observed for dietary interventions. A transient effect of HFD-feeding on exploration in male mice has been only observed after 3 weeks (
). In our study, after 8 weeks of HFD, activity and exploration remained unchanged, which was also observed after 10 weeks of HFD during the EPM. Interestingly, HFD-feeding only affected anxiety and depressive-like behavior after 12 (DLB and MST) and 13 weeks (Splash Test), indicating that a long-term exposure is necessary to disturb behavior. Similarly, male mice receiving a 45% HFD for a minimum of 3 months exhibited changes in locomotion, while it took at least 5 months of HFD-feeding to decrease exploration and increase anxiety (
An unexpected observation of our study represented the specific attenuation of depressive-like behavior by LR in males while anxiety was unaffected (Fig. 5). Anxiety and depressive-like behavior share common dysregulated pathways and are often observed in the same models (
). Elevated CCK in the brain as well as the administration of CCK receptor agonists has been shown to deteriorate emotional behavior by altering dopaminergic signaling (
Effects of citalopram treatment on behavioural, cardiovascular and neuroendocrine response to cholecystokinin tetrapeptide challenge in patients with panic disorder.
). We have identified a specific genetic signature of alterations in dopaminergic and CCK signaling indicating that the presence of alterations in both pathways modulate depressive-like but not anxious behavior. An interaction between these signaling pathways is affected in depression and that their involvements can contribute to depressive-like behaviors (
). Lastly, on a molecular level, palmitate has been shown to induce CCK signaling in hypothalamic neurons, connecting the exposure to a lard-based HFD to elevated CCK in the brain (
). Moreover, a decrease in tyrosine hydroxylase as well as low levels of dopamine due to increased turnover have already been associated with alterations in emotional behavior (
), and we confirmed this observation, showing that TH mRNA levels were decreased in HFD conditions (RNAseq and qPCR) as sign of altered dopaminergic signaling. Why TH protein levels were unaffected in this scenario remains unknown, but can be explained by spatial differences in transcription and translation as well as regulation of the catalytic activity by phosphorylation ((
Neurochemical and electrophysiological deficits in the ventral hippocampus and selective behavioral alterations caused by high-fat diet in female C57BL/6 mice.
). Interestingly, the increase in physical activity, as observed in female mice fed a HFD, can be mediated by the mesolimbic pathway, where estrogen stimulates dopamine release in the NAcc, supporting the potential presence of differentially affected dopamine signaling in both sexes. Conversely, inhibition of dopaminergic signaling decreases physical activity (
Only minor changes in cecal microbiome composition were detected in female and male mice, however, both HFD-feeding and LR application had an effect on beta diversity in female mice (Fig. 6C), while male mice only exhibited alterations in response to the HFD. Analysis of differential abundances show that Allobaculum is linked to metabolic and behavioral alterations in female mice (Fig. 6E). Here, HFD-feeding decreased abundance of Allobaculum, which was increased after LR treatment. Interestingly, reduced abundance of Allobaculum is linked to chronic stress and depressive-like behavior (
), confirming the association of an altered abundance of the genera Allobaculum to emotional disorders. On the other hand, elevated Allobaculum abundance is linked to improved insulin signaling (
) (see also Fig. 2H). The differences in abundance of Allobaculum can be explained by the decreased amount of carbohydrates in HFDs, as Allobaculum grows in the presence of exopolysaccharides (
Oral administration of live exopolysaccharide-producing Pediococcus parvulus, but not purified exopolysaccharide, suppressed Enterobacteriaceae without affecting bacterial diversity in ceca of mice.
Pyridostigmine Protects Against Diabetic Cardiomyopathy by Regulating Vagal Activity, Gut Microbiota, and Branched-Chain Amino Acid Catabolism in Diabetic Mice.
) and thus connects our observed alteration in microbiota to altered BCAA metabolism, as seen in our KEGG analysis (Fig. 7B,C). Interestingly, a reduction of plasma BCAA levels in HFD-fed mice has been shown to specifically improve depressive-like behavior via serotonergic signaling (
), which supports the presence of altered emotional behavior.
Taken together, our data show that LR application exhibit sex-specific effects in established obesity. While metabolism is positively affected in obese females, males exhibit improved emotional behavior.
Author contributions
M.S. contributed to the design of the study, researched data, and wrote the manuscript. K.W., R.H., M.R. and S.C. contributed to researching data. A.K. designed the study, supervised all work and wrote the manuscript. A.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Funding
This work was supported by the Deutsche Diabetes Gesellschaft (DDG) and by a grant from the German Ministry of Education and Research (BMBF) (Project 031 B0569) and the State of Brandenburg (DZD grant 82DZD00302).
Acknowledgements
We thank Organobalance GmbH, Berlin for providing the lactobacilli and performing the analysis of the cecal gut microbiota. We thank Benjamin Anderschou Holbech Jensen and Ida Søgaard Larsen (Laval University, Institute Universitaire de Cardiologie et de Pneumologie de Québec) for providing brain samples of the preventive probiotic study. We thank Markus Jähnert for critical input regarding the RNA sequencing analysis and Christine Gumz for measuring RNA quality using the Bioanalyzer (Experimental Diabetology, German Institute of Human Nutrition Potsdam-Rehbruecke).
Conflict of interest statement
The authors report no biomedical financial interests or potential conflicts of interest.
Long-term use of probiotics Lactobacillus and Bifidobacterium has a prophylactic effect on the occurrence and severity of pouchitis: a randomized prospective study.
Effects of citalopram treatment on behavioural, cardiovascular and neuroendocrine response to cholecystokinin tetrapeptide challenge in patients with panic disorder.
Neurochemical and electrophysiological deficits in the ventral hippocampus and selective behavioral alterations caused by high-fat diet in female C57BL/6 mice.
Oral administration of live exopolysaccharide-producing Pediococcus parvulus, but not purified exopolysaccharide, suppressed Enterobacteriaceae without affecting bacterial diversity in ceca of mice.
Pyridostigmine Protects Against Diabetic Cardiomyopathy by Regulating Vagal Activity, Gut Microbiota, and Branched-Chain Amino Acid Catabolism in Diabetic Mice.
00020-4&id=gr1.jpg){kind=link}
00020-4&id=gr2.jpg){kind=link}
00020-4&id=gr4.jpg){kind=link}
00020-4&id=gr3.jpg){kind=link}
00020-4&id=gr5.jpg){kind=link}
00020-4&id=gr6.jpg){kind=link}
00020-4&id=gr7.jpg){kind=link}
00020-4&id=gr8.jpg){kind=link}