Sex-dependent effects of carbohydrate source and quantity on caspase-1 activity in the mouse central nervous system

Abstract

Background: Mounting evidence links glucose intolerance and diabetes as aspects of metabolic dysregulation that are associated with an increased risk of developing dementia. Inflammation and inflammasome activation have emerged as a potential link between these disparate pathologies. As diet is a key factor in both the development of metabolic disorders and inflammation, we hypothesize that long term changes in dietary factors can influence nervous system function by regulating inflammasome activity and that this phenotype would be sex-dependent, as sex hormones are known to regulate metabolism and immune processes.

Methods: 5-week-old male and female transgenic mice expressing a caspase-1 bioluminescent reporter underwent cranial window surgeries and were fed control (65% complex carbohydrates, 15% fat), high glycemic index (65% carbohydrates from sucrose, 15% fat), or ketogenic (1% complex carbohydrates, 79% fat) diet from 6 to 26 weeks of age. Glucose regulation was assessed with a glucose tolerance test following a 4-h morning fast. Bioluminescence in the brain was quantified using IVIS in vivo imaging. Blood cytokine levels were measured using cytokine bead array. 16S ribosomal RNA gene amplicon sequencing of mouse feces was performed to assess alterations in the gut microbiome. Behavior associated with these dietary changes was also evaluated.

Results: The ketogenic diet caused weight gain and glucose intolerance in both male and female mice. In male mice, the high glycemic diet led to increased caspase-1 biosensor activation over the course of the study, while in females the ketogenic diet drove an increase in biosensor activation compared to their respective controls. These changes correlated with an increase in inflammatory cytokines present in the serum of test mice and the emergence of anxiety-like behavior. The microbiome composition differed significantly between diets; however no significant link between diet, glucose tolerance, or caspase-1 signal was established.

Conclusions: Our findings suggest that diet composition, specifically the source and quantity of carbohydrates, has sex-specific effects on inflammasome activation in the central nervous system and behavior. This phenotype manifested as increased anxiety in male mice, and future studies are needed to determine if this phenotype is linked to alterations in microbiome composition.

Keywords: Behavior; Caspase-1; Glucose intolerance; High glycemic index diet; Inflammasome; Inflammation; Ketogenic diet; Microbiome; Neuroinflammation; Sex.

Fig. 1

Ketogenic diet leads to weight gain and glucose intolerance in male and female mice. A Body weight was measured prior to diet intervention and weekly thereafter until the end of the experiment, at 26 weeks. Difference in body weight was assessed using a mixed effects model for high glycemic index (HGD) and ketogenic (KD) diets relative to control diet (CD). B Fasting glucose was measured following a 4-h morning fast at 26 weeks. C Mice were injected with a glucose solution and blood glucose was measured at 15-, 30-, 60-, and 120-min post injection. Area under the curve (AUC) for the glucose tolerance test (GTT) was calculated as an index of whole glucose excursion. D Male mice on the KD first displayed impaired glucose tolerance at 10 weeks, which did not correspond to weight gain, while E female mice on the KD first showed impaired glucose at 14 weeks, coinciding with significant weight gain. All data are shown as mean ± SEM and n = 5–11 per group. Statistical differences for weight over time were calculated with a mixed effects model. One-way ANOVA followed a Dunnett’s test was used to compare multiple groups (A–C) and a Student’s t test was used to compare two groups (D, E). For all statistical tests, *, **, ***, ****p < 0.05, 0.01, 0.001, and 0.0001, respectively

Fig. 2

High glycemic index diet (HGD) in males and ketogenic diet (KD) in females causes increased caspase-1 activation in the brain. A Representative images of caspase-1 biosensor mice with cranial windows at 26-weeks. B Cleaved caspase-1 signal was measured in mice at 10- and 26-weeks of age and the difference between the two time points was calculated for each animal. C Caspase-1 signal at 26 weeks in males and females. All data are shown as mean ± SEM and n = 4–11 per group. Statistical differences were determined using one-way ANOVA followed by a Dunnett’s test, comparing the HGD or KD to their sex-matched CD, where *, **p < 0.05 and 0.01 respectively

Fig. 3

IL-1β and TNF in the serum correlate with caspase-1 activation in the brain. Cytokine levels in serum collected from 26-week-old mice were quantified using a CBA. A IL1-β and B TNF levels were measured in the serum of male mice. Correlation between C IL-1β and D TNF in the serum and caspase-1 activation in the CNS was assessed using simple linear regression in males. E IL-1β and F TNF levels were measured in the serum of female mice. Correlation between G IL-1β and H TNF in the serum and caspase-1 activation in the CNS was assessed using simple linear regression in females. All data are shown as mean ± SEM and n = 4–10 per group. Statistical differences were determined using one-way ANOVA followed by a Dunnett’s test, comparing the HGD or KD to their sex-matched CD. Correlation analysis of serum cytokines and caspase-1 activation were performed using a simple linear regression

Fig. 4

Diet alters the male mouse microbiome. DNA was extracted from 26-week-old male mice and 16S ribosomal RNA gene amplicon sequencing performed. A Bray–Curtis Dissimilarity Index identified clusters of samples that were similar in composition. B Alpha diversity analysis was performed by calculating richness (ace, chao1, and observed richness), evenness (Pielou) and overall diversity (Shannon and Simpson). C Beta diversity was assessed by PCoA using the Bray–Curtis Dissimilarity Index at the ASV level and resultant values tested for significance with the experimental covariates using the PERMANOVA test. N = 8–10 per group. Statistical comparisons of bacterial abundance were conducted using the Phyloseq package and involved pairwise Chi-Squared Tests, and results were filtered based on a significance threshold of Benjamini-Hochberg adjusted p-value < 0.01

Fig. 5

Specific genera of bacteria are enriched in the high glycemic index and ketogenic diets. DNA was extracted from 26-week-old male mice and 16S ribosomal RNA gene amplicon sequencing performed. A The top 30 genera displayed as a relative abundance heatmap. B DESeq analysis of bacterial abundance between diets, where individual dots represent different amplicon sequence variants within the genus. N = 8–10 per group. Statistical comparisons of bacterial abundance were conducted using the DESeq package and involved pairwise Chi-Squared Tests, and results were filtered based on a significance threshold of Benjamini-Hochberg adjusted p-value < 0.01

Fig. 6

Carbohydrate source and quantity alters mouse anxiety behavior without impacting motor function or learning and memory. C57BL/6 mice were randomly assigned to control (CD), high glycemic index (HGD), or ketogenic (KD) diet at 6 weeks of age and put through a battery of behavior tests at 26 weeks. A Accelerated rotarod was used to evaluate motor coordination and learning by recording latency to fall from the rod. B Percent freezing in cued fear conditioning was used as a measure of associative memory. C Novel object recognition assessed spatial learning. Preference index was defined Time_novel/(Time_novel + Time_familiar). Elevated zero maze evaluated anxiety-like behavior via D percent time spent in open zone, E number of open zone entries, and F total distance traveled. All data are shown as mean ± SEM and n = 6–11 per group. One-way ANOVA followed by a Dunnett’s test was used to assess statistical significance, comparing the HGD or KD to their sex-matched CD, wherein *, **, ***p < 0.05, 0.01, and 0.001, respectively