Table 1.
Energy density and composition of experimental diets.
Fig 1.
Venn diagrams showing the number of differentially expressed genes (q<0.001) in each diet-comparison for each tissue.
SKM and BMC showed little transcriptional response to diet, whereas hundreds of genes were affected in the other four tissues. In BAT and WAT there was mainly an effect of HFD. In SPL there was mainly an effect of HFD-P. In LIV there was an influence of both HFDs per se, but also additionally, separate effects of HFD-P and HFD-S. SKM, skeletal muscle; BMC, bone marrow cells; BAT, brown adipose tissue; WAT, white adipose tissue; HFD, high fat diet; SPL, spleen; HFD-P high fat diet rich in polyunsaturated fatty acids; LIV, liver; HFD-S, high fat diet rich in saturated fatty acids.
Fig 2.
(A) PCA plot based on normalized gene expression data for differential expressed genes that were highly significant (q<1e-7) in any diet comparison and any tissue. As expected samples clustered tissue-wise, but interestingly, the WAT/HFD samples separated from the LFD samples and appeared closer to the two immune tissues SPL and BMC. It is also apparent in SPL, the HFD-P samples separated from the LFD and HFD-S. These observations suggest that diet effects on tissue transcriptomes were mainly observed in WAT and SPL, influenced by HFD per se in the former and by HFD-P specifically in the latter. (B) A hierarchical tree of the samples was constructed based on the same genes used in the PCA plot. Here, the separation of the WAT samples was clear as they divided into the two main clusters, the LFD samples together with LIV, SKM and BAT, and the HFD samples with BMC and SPL.
Fig 3.
A heatmap showing significant (q<0.05) immune system GO-terms for the different diet comparisons across all tissues.
Changes in diet affected immune system related gene expression in primarily WAT and SPL. HFDs resulted in major changes in WAT (indicated by black arrows) whereas HFD-P, specifically, resulted in the changes observed in SPL (indicated by green arrows). The figure is based on the adjusted non-directional p-values (S1–S6 Tables). Color indicates GO-terms affected by transcriptional up-regulation (red) or down-regulation (blue). Green indicates a significant yet unspecific regulation and white indicates no significant regulation. WAT, white adipose tissue; SPL, spleen; HFDs, high fat diets; HFD-P high fat diet rich in polyunsaturated fatty acids.
Fig 4.
(A) A heatmap of expression values from all samples for the subset of genes that were significantly changed (q<0.001) in WAT by HFD (HFD-S vs. LFD or HFD-P vs. LFD). Again, the samples divided into the same two main clusters as in Fig 2B (hierarchical tree not shown). The heatmap enables the identification of groups of genes that show similar expression values in the WAT/HFD samples as in the BMC and SPL samples and differ from the expression values in WAT/LFD and the remaining tissues. (B) In order to extract the genes in an unbiasedly manner that drive the separation of the WAT samples into the two clusters, we performed a sample classification using supervised learning, based on the same genes as in (A). In this process each gene is scored based on its influence on classification. The plot shows the ordered scores and it can be seen that roughly 200 genes are important for the separation into the two clusters, whereas the remaining genes have little or no importance. The top box shows the 12 most important genes and the bottom box shows the enrichment of GO-terms (Biological Processes) for the top 200 genes, mainly illustrating their involvement in immune system related processes. PCA, principle component analysis; WAT, white adipose tissue; HFD, high fat diet; LFD, low fat diet; SPL, spleen; BMC, bone marrow cells; HFD-P high fat diet rich in polyunsaturated fatty acids; HFD-S, high fat diet rich in saturated fatty acids; LIV, liver; SKM, skeletal muscle; BAT, brown adipose tissue; GO, gene ontology.
Fig 5.
(A) Organ weight of retroperitoneal WAT and (B) gonadal and (C) subcutaneous WAT from mice fed LFD, HFD-S and HFD-P. One-way ANOVA with Tukey’s post-hoc test, n = 8 mice per group. Data are shown as mean + SEM. mRNA levels of macrophage marker (D) Adgre1 (F4/80), M1 markers (E) Cd40 (CD40) and (F) Ccl2 (MCP-1) in retroperitoneal WAT from mice fed LFD HFD-S and HFD-P. One-way ANOVA with Tukey’s post-hoc test, n = 8–10 mice per group. Data are shown as mean + SEM. (G) Number of CLS/μm2 in WAT from mice fed LFD, HFD-S and HFD-P. CLS was analyzed by Kruskal-Wallis test (global p = 0.001) followed by Mann-Whitney test, with n = 6 mice per group. Data are shown as scatter plots with the line indicating the median. (H) Representative micrographs showing F4/80 staining in WAT from mice fed LFD, HFD-S and HFD-P. Scale bar represents 200 μm. *, **, ***: p<0.05, <0.01 and <0.001 respectively. WAT, white adipose tissue; LFD, low fat diet; HFD-S high fat diet rich in saturated fatty acids; HFD-P high fat diet rich in polyunsaturated fatty acids; CLS, crown like structures.
Fig 6.
(A) Fasting B-Glucose levels, (B) HOMA-IR index and (C) OGTT AUC from mice fed LFD, HFD-S and HFD-P. For (A) and (C) data are shown as mean + SEM, for (B) data are shown as back transformed geometric mean + geometric SEM. One-way ANOVA with Tukey post-hoc test, n = 5–6 mice per group. HOMAR-IR, homeostasis model assessment of insulin resistance; OGTT, oral glucose tolerance test; AUC, area under the curve; LFD, low fat diet; HFD-S, high fat diet rich in saturated fatty acids; HFD-P, high fat diet rich in polyunsaturated fatty acids (HFD-P).
Fig 7.
(A) Organ wight of LIV, (B) percentage ORO positive areas in LIV and (C) representative micrographs from mice fed LFD, HFD-S or HFD-P. Scale bar represents 200 μm. For (A) two-way ANOVA with experiment as nuisance factor, n = 10 mice per group. Data are shown as estimated marginal mean + SEM. For (B) two-way ANOVA followed by Tukey’s multiple comparisons test. **, ***: p<0.01, <0.001 respectively. LIV, liver; ORO, oil red O; LFD, low fat diet; HFD-S, high fat diet rich in saturated fatty acids; HFD-P, high fat diet rich in polyunsaturated fatty acids (HFD-P).