Adipose tissue in health and disease through the lens of its building blocks

Abstract

Understanding adipose tissue cellular heterogeneity and homeostasis is essential to comprehend the cell type dynamics in metabolic diseases. Cellular subpopulations in the adipose tissue have been related to disease development, but efforts towards characterizing the adipose tissue cell type composition are limited. Here, we identify the cell type composition of the adipose tissue by using gene expression deconvolution of large amounts of publicly available transcriptomics level data. The proposed approach allows to present a comprehensive study of adipose tissue cell type composition, determining the relative amounts of 21 different cell types in 1282 adipose tissue samples detailing differences across four adipose tissue depots, between genders, across ranges of BMI and in different stages of type-2 diabetes. We compare our results to previous marker-based studies by conducting a literature review of adipose tissue cell type composition and propose candidate cellular markers to distinguish different cell types within the adipose tissue. This analysis reveals gender-specific differences in CD4⁺ and CD8⁺ T cell subsets; identifies adipose tissue as rich source of multipotent stem/stromal cells; and highlights a strongly increased immune cell content in epicardial and pericardial adipose tissue compared to subcutaneous and omental depots. Overall, this systematic analysis provides comprehensive insights into adipose tissue cell-type heterogeneity in health and disease.

Figure 1

TissueDecoder the analysis of phenotypic traits. and the deconvolution with CIBERSORT.

Figure 2

AT21 Signature Matrix: (A) Heatmap showing the correlations of signatures for the reference dataset compiled from various studies reporting isolated cell type transcriptome profiles. The dendrogram (on the left) shows the clustering within the reference dataset. The dendrogram is constructed using hierarchical clustering with 1-correlation as the distance and average linkage as the linkage criteria. The number of samples contained for each cell type (N), and the GEO accession numbers per study is shown on the right. Three groups revealed higher correlation and clustered together in the reference dataset in accordance with our expectations: Highest correlation is observed between subcutaneous and pericardial adipocytes followed by the cell types of the mesenchymal stem/stromal cell origin (osteoblasts, mesenchymal stem/stromal cells, chondrocytes and Adipose Stem/Stromal Cells - ASCs) and finally the T cell group (CD4⁺ T and CD8⁺ T cell subsets) follow. (B) Bar plots showing top ten cell type markers identified by the primary criterion and the secondary criterion of CellMaDe. Below each column conventional markers are demonstrated with their associated score. Bold font indicates the presence of the gene in the AT21 signature matrix and red colored bars indicate the value of the (primary/secondary) criterion was negative. Letters in brackets (M: membrane, E: extracellular, ME: both membrane and extracellular, or O: other) specify the gene ontology cellular location of the corresponding protein.

Figure 3

Cell Type Specificity of TissueDecoder: Heatmaps of AT21 signature matrix showing (A) cell type composition prediction from CIBERSORT with AT21 signature matrix, (B) the normalized expression of selected conventional markers from literature and (C) the normalized expression of CellMaDe predicted primary markers. The common xaxis above the figure denotes the samples used for annotating cell types from the AT21 signature matrix. CellMaDe and CIBERSORT are both trained with AT21 signature matrix, therefore (A,C) are optimal results for both techniques whereas (B) represents the separation power of conventional markers on the AT21 signature matrix. Green boxes indicate for each cell type (columns) the row with its corresponding marker.

Figure 4

Review of Adipose Tissue Cellular Composition: Literature review of reported adipose tissue cellular composition in comparison to our results using CIBERSORT (in green). Shown is the mean (dots), minimum, and maximum (arrows) percentage of total cells (calculated according to assumptions and formulas stated in the supplementary methods) for five different cell types - macrophages (A), ASCs (B), CD4⁺ T cells (C), CD8⁺ T cells (D), and endothelial cells (E). The color specifies the method used for cell counting as indicated in the legend. The gray dots and arrows in (A) are our results where the macrophage score and monocyte score have been added together. It is shown as a comparison, since the macrophage markers CD68, HAM56, and CD14 also stain monocytes. The gray dot in (B) represents the sum of ‘supra adventitial-adipose stromal cells’ (black dot in the same row) and ‘endothelial progenitor cells’, which were distinguished in the respective study, but are likely both covered in the ‘adipose stem/stromal cell’ score from our AT21 signature matrix. On the left hand side of each plot the references to the studies from which the results were taken (see Supplementary Data S4) and the utilized markers are indicated. For ASCs and endothelial cells a combination of markers was used (see Supplementary Data S4). A star attached to the study reference letter indicates that the study participant had an average body mass index above 35. It is included in the figure since it has been reported that the macrophage frequency is increased in people with severe obesity.

Figure 5

Estimated Percentage of Cell Types per Adipose Depot: (A) Pie chart showing the overall cell type distribution of various adipose tissue depots (Subcutaneous – n = 616, Omental – n = 51, Pericardial – n = 66, and Epicardial – n = 46) in terms of four main archetypes of cells in adipose tissue (Immune cells, Stem/stromal cells, Adipocytes, and other). (B) Bar plots showing the detailed distribution of the immune cell compartment from each adipose tissue depot. All values are averages of the analyzed samples from the respective depot.

Figure 6

Adipose Tissue Composition Across Phenotypic Traits: (A) Heatmap showing the significant over (red)/under (blue) represented cell type across different categories based on the z-score (denotes the number of standard deviations each group is away from the other). Only significant results (uncorrected p < 0.05) are colored. Stars label results that stay significant after correction for multiple testing. The bean plots showing (B) all significant cell types and (C) ASCs and subcutaneous adipocytes (not detected as significant), for the heavier vs leaner twin, discordant study. The y axis describes differences in estimated cell fractions between the heavier and leaner twin within a twin pair.