Designing dense connective tissues: identifying micro-environmental factors that direct human MSC differentiation in 3D

摘要

Introduction: Microenvironmental factors such as biomaterial physical and biochemical properties, external mechanical stimuli, and soluble chemical cues integrate in vivo to direct the differentiation of mesenchymal stromal cells (MSCs). Identifying what factor combinations promote the differentiation of MSCs into myofibroblasts (the tissue producing cells of developing valves, ligaments, skin, tendons and periodontal ligaments) is a significant challenge when designing for dense, type Ⅰcollagen-rich connective tissues. To date, select combinations of microenvironmental factors have been studied in 2D, but systematic consideration of the integration of factors in 3D culture has not been explored fully. Materials and Methods: We developed a 3D screening platform to systematically study MSC responses to microenvironmental stimuli. Permutations of the cell adhesion sequences RGD, DGEAand YIGSR at different concentrations (0-2 mM) were incorporated Into 5-11 wt% (5-22 kPa compressive moduli) polyethylene glycol norbornene (PEG-NB) hydrogel arrays. Since PEG-NB is inert and easily crosslinked with UV light, adhesion ligands can be tuned independently from biomaterial properties, enabling us to identify the contribution of individual factors in a combinatorial fashion. Human bone marrow-derived MSCs were encapsulated and cultured in the 3D arrays for 7 days with 0 or 5 ng/mL TGF-β1. The extent of α-smooth muscle actin (αSMA) expression and collagen type Ⅰdeposition were modeled using least squares estimation and regression on 3D confocal data (Figure 1A) that was analysed using Imaris (Figure 1B). Each experiment was replicated four times and statistical analyses were performed in JMP on the pooled data. Results and Discussion: Regression analyses of the screening data revealed that in the presence of TGF-β1, PEG-NB hydrogel wt% was the most significant factor influencing myofibrablast differentiation (p ＜ 0.0001), with greater aSMA staining intensity manifesting at lower gel wt%. These data are in contrast to MSC myofibroblastic differentiation in 2D, which is promoted by stiffer substrates. With TGF-β1, RGD concentration was correlated with aSMA staining intensity in a positive, biphasic manner and demonstrated a significant, non-linear dependence with PEG-NB wt% (p = 0.02) (Figure 2A). Interestingly, aSMA staining intensity could only be predictably modeled by regression in cultures that were supplemented with TGF-β1. These observations suggest that while MSCs can express aSMA in the absence of exogenous TGF-β1, aSMA expression is only sensitive to material properties in the presence of TGF-(51. Collagen deposition was predominantly negatively correlated with PEG-NB hydrogel wt% (p＜0.0001) with or without TGF-β1, with slight non-linear dependence between RGD and DGEA (p = 0.01) (Figure 2B) or linear dependence with RGD (p = 0.04) in the absence or presence of TGF-β1, respectively. Conclusion: The newly developed 3D screening platform enables the systematic identification of optimal conditions for connective tissue engineering when considering multiple interacting microenvironmental factors. The data presented here demonstrate that conditions which promote a myofibroblast phenotype in 2D do not translate completely to the more complex and integrative 3D environment.