Differences in the Inflammatory Response of White Adipose Tissue and Adipose-Derived Stem Cells

Abstract

The application of liposuctioned white adipose tissue (L-WAT) and adipose-derived stem cells (ADSCs) as a novel immunomodulatory treatment option is the currently subject of various clinical trials. Because it is crucial to understand the underlying therapeutic mechanisms, the latest studies focused on the immunomodulatory functions of L-WAT or ADSCs. However, studies that examine the specific transcriptional adaptation of these treatment options to an extrinsic inflammatory stimulus in an unbiased manner are scarce. The aim of this study was to compare the gene expression profile of L-WAT and ADSCs, when subjected to tumor necrosis factor alpha (TNFα), and to identify key factors that might be therapeutically relevant when using L-WAT or ADSCs as an immuno-modulator. Fat tissue was harvested by liposuction from five human donors. ADSCs were isolated from the same donors and shortly subjected to expansion culture. L-WAT and ADSCs were treated with human recombinant TNFα, to trigger a strong inflammatory response. Subsequently, an mRNA deep nextgeneration sequencing was performed to evaluate the different inflammatory responses of L-WAT and ADSCs. We found significant gene expression changes in both experimental groups after TNFα incubation. However, ADSCs showed a more homogenous gene expression profile by predominantly expressing genes involved in immunomodulatory processes such as CCL19, CCL5, TNFSF15 and IL1b when compared to L-WAT, which reacted rather heterogeneously. As RNA sequencing between L-WAT and ADSCS treated with TNFα revealed that L-WAT responded very heterogeneously to TNFα treatment, we therefore conclude that ADSCs are more reliable and predictable when used therapeutically. Our study furthermore yields insight into potential biological processes regarding immune system response, inflammatory response, and cell activation. Our results can help to better understand the different immunomodulatory effects of L-WAT and ADSCs.

Keywords: TNFalpha; adiposederived stem cells; immunomodulation; inflammation; white fat tissue.

Figure 1

Experimental timeline and setup. Human L-WAT was harvested from five different donors through water-jet-assisted liposuction. Fresh fat samples were incubated with TNFα for 48 h, to mimic an inflammatory milieu that triggers a strong immune response. Simultaneously, ADSCs were isolated from the same donors’ fat samples harvested during earlier liposuction. After an expansion period of 72 h, ADSCs were also treated with TNFα. Subsequently the incubation with TNFα, RNA from fat and cell samples was isolated and sequenced. A portion of the same sample without TNFα was implemented in parallel and used as reference control.

Figure 2

Significantly differentially expressed genes in L-WAT and ADSCs. Significantly upregulated genes in ADSCs compared to L-WAT and in ADSCs and TNFα compared to L-WAT and TNFα (A). Significantly, downregulated genes in ADSCs compared to L-WAT and in ADSCs and TNFα compared to L-WAT and TNFα (B).

Figure 3

Principal Component Analysis (PCA) of L-WAT and ADSCs. The PCA revealed a clear separation of L-WAT (blue dots) and ADSCs (red dots) along the main component PC1. In addition, TNFα treatment showed a separation along PC4, with a more homogeneous response of ADSCs (dark-red dots), when compared to L-WAT (dark-blue dots). There was no clear clustering of samples from the abdomen (black border) or the thighs (no border).

Figure 4

Hierarchical cluster analysis of the 30 most contributing genes of PC1 and PC4. The 30 most to PC1 contributing genes are fat-related marker genes including LEP, FABP4 and ADIPOQ (A). The analysis of the PC4 reveals the separation among genes associated with immunomodulatory processes (B).

Figure 5

Significantly changed biological pathways in ADSCs and L-WAT after incubation with TNFα. Differential gene expression analysis revealed hundreds of significant upregulated (A) and downregulated (B) biological pathways in ADSCs treated with TNFα, when compared to L-WAT treated with TNFα.

Figure 6

Visualization of the five most positively (NES > 0) or negatively (NES < 0) TNFα-dependent hallmark gene sets in ADSCs, when compared to L-WAT. Blue arrowheads indicate the difference of interferon alpha and gamma response of ADSCs and L-WAT upon TNFα treatment. NES: normalized enrichment score. Adjusted p-value < 0.05.