Sex-Specific Regulation of Inflammation and Metabolic Syndrome in Obesity

Figures

• 

  Figure 1.. Characteristics of the 300-OB cohort. A, Overview of some of the parameters measured in the 300-OB cohort. Each outlined square indicates a different data type and schematically describes the different measurements. circ. markers, circulating markers of inflammation. B, Baseline characteristics, also split by sex. C, Percentage of individuals with the metabolic Syndrome as defined by the National Cholesterol Education Program (NCEP)-ATP III criteria. Left: overall percentage, right: subdivided into women and men. Yellow: people with the condition, green: people without the condition. D, Percentage of individuals with other conditions. E–I, Percentage of individuals with each of the 5 criteria used in defining metabolic syndrome, displayed in a similar style to (C): purple indicates people on medication for the condition (E) high blood pressure, (F) high glucose, (G) abdominal obesity, (H) low HDL cholesterol (HDL-c), (I) high triglycerides *P value ≤0.05 calculated using χ2 test. ***P value ≤0.005.

• 

  Figure 2.. Associations of various immune and host parameters in the 300-OB cohort. A, Mean and SDs of several circulating inflammatory markers for women and men. B, P values of associations between age and sex with markers of inflammation* (BH-false discovery rate [FDR] multiple testing corrected). C, P values of associations between markers of inflammation, corrected for age, sex, and season (BH-FDR multiple testing corrected). D, Example association between IL (interleukin)-6 and leptin. E, P values of associations between age and sex with immune cell counts* (BH-FDR multiple testing corrected). F, P values of associations between immune cell numbers and circulating markers of inflammation, corrected for age, sex, and season (BH-FDR multiple testing corrected). Note: all P values were calculated using linear regression and testing the null hypothesis that β=0 for relationship between x and y, where x and y are the parameters of interest. The data in the example plots (D) were corrected for age, sex, and season and transformed using IRT (see section Methods). *Associations with age corrected for sex and season, associations with sex corrected for age and season.

• 

  Figure 3.. The effect of metabolic syndrome (MetS) on immune parameters. A, Significance of the difference in levels of several inflammatory markers between individuals with and without MetS. Red indicates that the marker is significantly higher in individuals with metabolic syndrome and blue means it is higher in those without. The first column is for all individuals, and the second and third are for women and men, respectively (false discovery rate [FDR] corrected per column). B, Significance of the interaction effect between sex and metabolic syndrome. Red indicates that the increase of women with metabolic syndrome vs those without was significantly lower than men with metabolic syndrome vs those without, that is, red means men show a stronger increase than women or women show a stronger decrease than men. C, Interaction effects for each individual medical condition used to score metabolic syndrome with sex. Each individual is scored as either having the condition or not, and interaction effects are calculated in a similar way as in (B). Similarly, red means men show a stronger increase than women, or women show a stronger decrease than men; blue means the opposite. The left part shows…

• 

  Figure 4.. A, Left, same as (A) but for IL (interleukin)-6, IL-1β, and TNFα production in peripheral blood mononuclear cells (PBMCs) after different stimulations. Middle, similar to the heatmap on the left but for high triglycerides instead of metabolic syndrome (right) similar to (D) but solely plotted for high triglycerides (false discovery rate [FDR] corrected). The arrows mark those cytokines after stimulation that show a significantly stronger increase in men than in women. Note: if men show a nonsignificant difference between men with and without MetS and women also show a nonsignificant difference, the increase in men show with getting MetS can still be significantly different from the change that women show. B, Example plot showing the effect of metabolic syndrome and sex on IL-6 production in response to Candida albicans stimulation. C, Example plot showing the effect of having high triglycerides and sex on IL-6 production in response to C albicans stimulation. Note: P values were calculated in a similar fashion as for Figure 3.

• 

  Figure 5.. A, Principle component analysis (PCA) plot of PC1 vs PC2 for the lipidomics data. B, Lipid groups that are significantly different between women and men. Blue indicates higher levels in women. C, Left, significance (false discovery rate [FDR]-corrected P values) of the difference of lipid concentrations between individuals with or without metabolic syndrome. Calculated for the whole 300-OB cohort and calculated for women a women separately. Right, FDR-corrected P values for the interaction effect between metabolic syndrome and sex in predicting the lipids, that is, showing if the increase/decrease of these lipids seen in metabolic syndrome is the same in women and men. D, PCA plot of the metabolomics data after inverse rank transformation and standardization. The first 2 PCs are plotted, and the samples are split into 4 groups: women without metabolic syndrome (“Female_NoMetS”), women with metabolic syndrome (“Female_YesMetS”), men without metabolic syndrome (“Male_NoMetS”), men with metabolic syndrome (“Female_YesMetS”). PC2 shows some separation between women and men. E, Significance of sex differences for the 7 metabolites from the valine, leucine, and isoleucine biosynthesis…

• 

  Figure 6.. Metabolic comparison between normal weight and overweight individuals. A, Visualization of the top 5% of metabolites (lowest P values) associated with each of the circulating markers of inflammation to give an impression of the number of significant metabolites. For each marker, the false discovery rate (FDR)-corrected P values were sorted independently to show the relative number of significant metabolites for each marker. B, Top 10 metabolic pathways associated with circulating markers of inflammation in the 300-OB cohort. See Methods for details. C, Example pathway plot of Central Carbon Metabolism in Cancer for VEGF in the 300-OB cohort. Significance is indicated as −log10 (P values), any P value below P=1×10−4 is rounded up to 1×10−4 for this visualization. Red indicates a positive association, blue indicates a negative association, and white indicates that the metabolite was not present in our data. Visualization was performed using Pathview.23 The portion of the figure showing the TCA cycle is shown as a zoomed-up inset.