Adipogenic Differentiation of Mesenchymal Stem Cells Alters Their Immunomodulatory Properties in a Tissue-Specific Manner

Abstract

Chronic inflammation is associated with formation of ectopic fat deposits that might represent damage-induced aberrant mesenchymal stem cell (MSC) differentiation. Such deposits are associated with increased levels of inflammatory infiltrate and poor prognosis. Here we tested the hypothesis that differentiation from MSC to adipocytes in inflamed tissue might contribute to chronicity through loss of immunomodulatory function. We assessed the effects of adipogenic differentiation of MSC isolated from bone marrow or adipose tissue on their capacity to regulate neutrophil recruitment by endothelial cells and compared the differentiated cells to primary adipocytes from adipose tissue. Bone marrow derived MSC were immunosuppressive, inhibiting neutrophil recruitment to TNFα-treated endothelial cells (EC), but MSC-derived adipocytes were no longer able to suppress neutrophil adhesion. Changes in IL-6 and TGFβ1 signalling appeared critical for the loss of the immunosuppressive phenotype. In contrast, native stromal cells, adipocytes derived from them, and mature adipocytes from adipose tissue were all immunoprotective. Thus disruption of normal tissue stroma homeostasis, as occurs in chronic inflammatory diseases, might drive "abnormal" adipogenesis which adversely influences the behavior of MSC and contributes to pathogenic recruitment of leukocytes. Interestingly, stromal cells programmed in native fat tissue retain an immunoprotective phenotype. Stem Cells 2017;35:1636-1646.

Keywords: Adipocytes; Endothelial cells; Inflammation; Mesenchymal stem cells; Neutrophils; Tissue-specific.

Figure 1

Effects of differentiation of MSC into adipocytes on their immunomodulation of neutrophil or lymphocyte recruitment. Cocultures formed by seeding EC with bone marrow MSC (BMMSC) or BMMSC‐derived adipocytes (A, C) on opposite sides of 0.4 µm porous filters or (B) seeding EC onto a collagen gel containing BMMSC or BMMSC‐derived adipocytes for 24 hours. Cultures were stimulated with (A, B) 100 U/ml tumor necrosis factor α (TNFα) for 4 hours or (C) 100 U/ml TNFα and 10 ng/ml interferon γ for 24 hours. In (A) and (C), a 4 minute bolus of purified leukocytes was perfused over the endothelium. (A) Neutrophil or (C) PBL adhesion was assessed and expressed as the number of cells adherent/mm² per 10⁶ cells perfused. In (B), neutrophils were allowed to adhere to the EC cultured on collagen gel containing MSC or adipocytes for 20 minutes before adhesion was assessed at 2 hours and expressed as a percentage of cells added. Data are mean ± SEM from n = 3 experiments incorporating a different EC and leukocyte donor in each experiment and two different BMMSC donors. ANOVA showed a significant effect of culture conditions on neutrophil (p < 0.01 in A, p < .05 in B) and PBL adhesion (p < 0.01 in C). *, p < .05; **, p < 0.01 by Tukey post‐test. Abbreviations: EC, endothelial cells; MSC, mesenchymal stem cells; PBL, peripheral blood lymphocytes; PMN, neutrophil.

Figure 2

Secretion and role in immunosuppression of IL‐6 in cocultures of MSC or MSC‐derived adipocytes. (A) IL‐6 and (B) sIL‐6R release into supernatants from EC, bone marrow MSC (BMMSC), and BMMSC‐derived adipocyte monoculture and cocultures was assessed after 24 hours. (C) BMMSC or (D) BMMSC‐derived adipocyte cocultures were treated with neutralizing antibodies against IL‐6 for the duration of the coculture and cytokine treatment. Neutrophil adhesion was assessed and expressed as the number of cells adherent/mm² per 10⁶ cells perfused. (E) SOCS3 gene expression analyzed by quantitative polymerase chain reaction in EC cultured alone, with MSC or with MSC‐derived adipocytes. Data are expressed as 2^–ΔCT relative to SOCS3 expression by EC cultured alone. Data are mean ± SEM (A) n = 16–19 and (B, C) n = 4, (D) n = 3, or (E) n = 3 independent experiments using a different EC and (C, D) neutrophil donor in each experiment. Three different BMMSC donors were used in all experiments, except (A) where 10 different donors were used. ANOVA showed significant effects of treatment in (A) (p < 0.01), (C and D) (each p < 0.05) *, p < 0.05 and **, p < 0.01 by Tukey post‐test. ¥¥ = p < 0.01 compared to the sum of the EC and respective MSC monoculture supernatant by paired t test. Abbreviations: AD, adipocytes; EC, endothelial cells; IL‐6, interleukin‐6; MSC, mesenchymal stem cells; sIL‐6R, soluble IL‐6 receptor; SOCS3, suppressor of cytokine signaling 3.

Figure 3

Expression and function of TGFβ receptors in endothelial cells cocultured with MSC or MSC‐derived adipocytes. Cocultures formed by seeding EC with MSC or MSC‐derived adipocytes on opposite sides of 0.4 µm porous filters for 24 hours. Gene expression for (A) TGFβ R1, (B) TGFβ R3, and (C) decorin was analyzed in endothelial cells cultured alone or with either MSC or MSC‐derived adipocytes by qPCR. Data are expressed as 2^–ΔCT relative to 18S. Data are mean ± SEM from (A) n = 5–6, (B) n = 3 or (C) n = 4 independent experiments using a different donor for each cell type in each experiment. ANOVA showed significant effects of treatment in (A) and (B) (both p < 0.01). *, p < 0.05 and **, p < 0.01 by Tukey post‐test. Abbreviations: EC, endothelial cells; MSC, mesenchymal stem cells; TGFβ, transforming growth factor beta.

Figure 4

Content and immunosuppressive effects of conditioned media from MSC or adipocyte cocultures. (A) Conditioned media from EC alone or EC in coculture were analyzed by a cytokine expression array. Representative immunoblots from a single experiment. EC monocultures were treated with conditioned media from (B) BMMSC or (C) BMMSC‐derived adipocytes either cultured alone (monocultures) or cocultured with EC for 24 hours. Neutrophil adhesion was assessed at 2 minutes post perfusion and expressed as the number of cells adherent/mm² per 10⁶ cells perfused. In (B), ANOVA showed a significant effect of treatment on neutrophil adhesion, p < 0.05. Data are mean ± SEM, (B) n = 3–5 and (C) n = 5 independent experiments using a different EC and neutrophil donor in each experiment. (B) 2 or (C) 3 different BMMSC donors were used. *, p < 0.05 by Dunnett post‐test. Abbreviations: AD, adipocytes; EC, endothelial cells; MSC, mesenchymal stem cells.

Figure 5

Immunosuppressive effects by ADSC and adipocytes on neutrophil recruitment. EC were cocultured with (A) ADSC, (B) ADSC‐derived adipocytes (ADSC adipo) or (C) mature adipocytes (Mature adipo) for 24 hours prior to stimulation with tumor necrosis factor α (TNFα) for 4 hours. (D) ADSC cocultures were treated with neutralizing antibodies against IL‐6 for the duration of the coculture and cytokine treatment. Neutrophil adhesion was assessed at 2 minutes postperfusion and expressed as the number of cells adherent/mm² per 10⁶ cells perfused. In (D), ANOVA showed a significant effect of treatment on neutrophil adhesion, p < 0.05. Data are mean ± SEM, (A) n = 6, (B) n = 3, and (C, D) n = 4 independent experiments using a different EC and neutrophil donor in each experiment. Three different stromal cell donors were used in all experiments, except (D) where one donor was used. *, p < 0.05 compared to EC monoculture by paired t test. ¥, p < 0.05 and ¥¥, p < 0.01 by Tukey post‐test. Abbreviations: ADSC, adipose‐derived stromal cells; EC, endothelial cells; IL, interleukin.

Figure 6

Immunosuppressive effects of ADSC and adipocytes on lymphocyte recruitment. EC were cocultured with (A) ADSC, (B) ADSC‐derived adipocytes, or (C) mature adipocytes (mature adipo) for 24 hours prior to stimulation with TNFα and IFNγ for 24 hours. PBL adhesion was assessed at 2 minutes postperfusion and expressed as the number of cells adherent/mm² per 10⁶ cells perfused. Data are mean ± SEM from n = 3–4 independent experiments using a different donor for each cell type in each experiment. *, p < .05; **, p < .01 compared to EC monoculture by paired t test. Abbreviations: ADSC, adipose‐derived stromal cells; EC, endothelial cells; PBL, peripheral blood lymphocytes.