IL-1β mediated nanoscale surface clustering of integrin α5β1 regulates the adhesion of mesenchymal stem cells

Abstract

Clinical use of human mesenchymal stem cells (hMSCs) is limited due to their rapid clearance, reducing their therapeutic efficacy. The inflammatory cytokine IL-1β activates hMSCs and is known to enhance their engraftment. Consequently, understanding the molecular mechanism of this inflammation-triggered adhesion is of great clinical interest to improving hMSC retention at sites of tissue damage. Integrins are cell-matrix adhesion receptors, and clustering of integrins at the nanoscale underlies cell adhesion. Here, we found that IL-1β enhances adhesion of hMSCs via increased focal adhesion contacts in an α5β1 integrin-specific manner. Further, through quantitative super-resolution imaging we elucidated that IL-1β specifically increases nanoscale integrin α5β1 availability and clustering at the plasma membrane, whilst conserving cluster area. Taken together, these results demonstrate that hMSC adhesion via IL-1β stimulation is partly regulated through integrin α5β1 spatial organization at the cell surface. These results provide new insight into integrin clustering in inflammation and provide a rational basis for design of therapies directed at improving hMSC engraftment.

Figure 1

hMSCs retain their phenotype in the presence of IL-1β. (A) Quantification of the concentration of IL-1β in the cell medium upon initial addition of either control or IL-1β medium to cells and after 1 and 7 days in culture as measured by ELISA. N = 10 replicates for each condition. Parametric one-way ANOVA, Tukey multiple comparison test. (B) Quantification of DNA concentration of hMSCs after 1 and 7 days in culture. N = 27 replicates for each condition. Non-parametric Kruskal–Wallis test with Dunn’s multiple comparison test. ***p < 0.001. Box and whisker plots represent minimum to maximum. (C,D) Representative flow cytometry dot plots of forward scatter (FSC-H) versus APC fluorescence intensity for the cell surface markers CD44, CD73, CD90 and CD105 for control and IL-1β treated (10 ng/mL) hMSCs following (C) 1 day or (D) 7 days in culture. Gating strategy and analysis are in Supplementary Figure S1.

Figure 2

IL-1β increases hMSC adhesion via α5β1. (A) Representative cell area plots of control or IL-1β treated (10 ng/mL) hMSCs. Scale bar = 50 μm. (B) Quantification of hMSC area and length following IL-1β treatment. N = 73–101 (5 images per condition). Non-parametric unpaired two-tailed t-test, Mann–Whitney post hoc. (C) Representative confocal images of hMSCs in control media, treated with IL-1β, or IL-1β treatment after IL-1 receptor (IL-1R) blocking for 1 h, labeled for vinculin (magenta), actin (green), and nuclei (cyan). Areas highlighted with the magenta box are shown in the bottom row. Scale bar = 50 μm top row, 10 μm bottom row. Blue arrows indicate actin ruffles and filopodia. (D) Fold change in relative number of vinculin focal adhesions normalized to controls at day 1. N = 15–18 images total for each condition. Parametric one-way ANOVA, Tukey multiple comparison test. (E) Immunolabeling of actin (green) and active integrin α5β1 (magenta). Scale bar = 20 μm. (F) Representative confocal images of hMSCs treated with the α5β1 antagonist ATN-161 for 1 h followed by control media or IL-1β treatment, labeled for vinculin (magenta), actin (green), nuclei (cyan). Areas highlighted with the magenta box are shown in the bottom row. Scale bar = 50 μm top row, 10 μm bottom row. (G) Fold change in relative number of vinculin focal adhesions normalized to ATN-161 controls at day 1. N = 15 images total for each condition. Parametric unpaired two-tailed t-test. *p < 0.05, **p < 0.01, ns = not significant. Box and whisker plots represent 5th–95th percentile. Bar charts represent mean ± SEM.

Figure 3

IL-1β induces α5β1 surface activation and clustering. (A) Schematic representing immunolabeling of active α5β1. (B) Representative dSTORM reconstructed image of integrin α5β1 clusters on the surface of an hMSC. Scale bar = 500 nm, zoom of green box = 100 nm. (C) Representative dSTORM reconstructed images of integrin α5β1 clusters on the surface of control and IL-1β treated (10 ng/mL) hMSCs following 1 day in culture. Scale bar = 200 nm. (D) Fold change in relative number of clusters, number of surface detections and cell surface cluster density, normalized to controls at day 1. (E) Representative dSTORM reconstructed images of integrin α5β1 clusters on the surface of control and IL-1β treated (10 ng/mL) hMSCs following 7 days in culture. Scale bar = 200 nm. (F) Fold change in relative number of clusters, number of surface detections and cell surface cluster density, normalized to controls at day 7. (G) Area of integrin α5β1 clusters at day 1 and 7. N = 3 independent experiments, 29–75 ROIs total for each condition. (H) Fold change in relative number of clusters and surface detections on hMSCs seeded on fibronectin-coated glass normalized to controls at day 1. N = 5–55 ROIs total. Non-parametric unpaired two-tailed t-test, Mann–Whitney post hoc. *p < 0.05, **p < 0.01, ***p < 0.001, ns = not significant. Bar charts represent mean ± SEM. Box and whisker plots represent 5th–95th percentile.

Figure 4

α5β1 total protein levels do not change. (A) Representative western blots of total protein membrane and cytosolic α5 and β1 integrin subunits in control and IL-1β treated hMSCs following 1 day in culture. Control and IL-1β samples were run on the same blot, with dividing lines delineating the samples from the housekeeping proteins used for normalization, probed on the same blot as the subunit protein of interest. (B) Fold change in α5 protein expression in the membrane and cytosol normalized to control. N = 6. (C) Fold change in β1 protein expression in the membrane and cytosol normalized to control. N = 6. Non-parametric unpaired two-tailed t-test, Mann–Whitney post hoc. ns = not significant. Bar charts represent mean ± SEM.

Figure 5

Blocking of IL-1R or TNF-α treatment prevents activation of surface α5β1 clustering. Representative dSTORM reconstructed images of integrin α5β1 clusters on the surface of hMSCs at day 1 following (A) IL-1β treatment (10 ng/mL) after IL-1 receptor (IL-1R) blocking for 1 h or (C) TNF-α treated (10 ng/mL). Scale bar = 200 nm. Fold change in relative number of clusters, number of surface detections and cell surface cluster density, normalized to respective controls at day 1 of (B) IL-1β treated (10 ng/mL) following IL-1R blocking for 1 h or (D) TNF-α treated (10 ng/mL) hMSCs. IL-1R + control data is plotted in Suppl. Fig. S4. N = 3 independent experiments, 50–100 ROIs total for each condition for IL-1R + IL-1β, 55–65 ROIs total for each condition for TNF-α. Non-parametric unpaired two-tailed t-test, Mann–Whitney post hoc. ns = not significant. Bar charts represent mean ± SEM. (E) Schematic of integrin α5β1 activation and clustering at the cell surface in the presence of IL-1β. This activation is prevented by prior antibody blocking of the IL-1R and does not occur in the presence of TNF-α.