Student Award for Outstanding Research Winner in the Ph.D. Category for the 9th World Biomaterials Congress, Chengdu, China, June 1-5, 2012: The interplay of bone-like extracellular matrix and TNF-α signaling on in vitro osteogenic differentiation of mesenchymal stem cells

Abstract

As an initial step in the development of a bone tissue engineering strategy to rationally control inflammation, we investigated the interplay of bone-like extracellular matrix (ECM) and varying doses of the inflammatory cytokine tumor necrosis factor alpha (TNF-α) on osteogenically differentiating mesenchymal stem cells (MSCs) cultured in vitro on 3D poly(ε-caprolactone) (PCL) microfiber scaffolds containing pregenerated bone-like ECM. To generate the ECM, PCL scaffolds were seeded with MSCs and cultured in medium containing the typically required osteogenic supplement dexamethasone. However, since dexamethasone antagonizes TNF-α, the interplay of ECM and TNF-α was investigated by culturing naïve MSCs on the decellularized scaffolds in the absence of dexamethasone. MSCs cultured on ECM-coated scaffolds continued to deposit mineralized matrix, a late stage marker of osteogenic differentiation. Mineralized matrix deposition was not adversely affected by exposure to TNF-α for 4-8 days, but was significantly reduced after continuous exposure to TNF-α over 16 days, which simulates the in vivo response, where brief TNF-α signaling stimulates bone regeneration, while prolonged exposure has damaging effects. This underscores the exciting potential of PCL/ECM constructs as a more clinically realistic in vitro culture model to facilitate the design of new bone tissue engineering strategies that rationally control inflammation to promote regeneration.

Figure 1

Overall study design using pregenerated bone-like ECM. Mesenchymal stem cells (MSCs) were isolated from rat bone marrow and cultured on electrospun PCL scaffolds in the presence of dexamethasone (+dex) and complete osteogenic media, which has been shown to stimulate production of bone-like ECM. These MSCs were removed and then replaced with a second batch of freshly isolated MSCs, which were subjected to the experimental conditions shown in Figure 2.

Figure 2

Schematic of study design for ECM and TNF-α interplay. The four control groups consist of two negative controls seeded with MSCs precultured in “+dex” media (as shown) or in “−dex” media; a positive control seeded with MSCs precultured in “+dex” media that continued to receive dexamethasone during days 0–16, and a group of acellular PCL/ECM scaffolds. OM = Osteogenic Media (α-MEM, 10% v/v fetal bovine serum, 50 μg gentamicin/mL, 0.5 μg amphotericin-B/mL, 10 mM β-glycerophosphate, 50 μg/mL ascorbic acid) either with or without 10⁻⁸ M dexamethasone (“+dex” vs. “−dex”).

Figure 3

Representative scaffold morphology via SEM. Representative scaffolds from the electrospun PCL meshes generated for this study, imaged via SEM at varying magnifications, (a) 300×, (b) 600×, and (c) 2000×. Scale bar for (a) is 200 μm, for (b) is 100 μm, and for (c) is 25 μm.

Figure 4

MSC cell count on preformed ECM with varying TNF-α dose. Scaffolds with pregenerated bone-like ECM (PCL/ECM) were seeded with MSCs precultured in +dex media (except for the group designated “(pre: −dex),” which was seeded with MSCs precultured in −dex media). Experimental groups were exposed to varying TNF-α doses. The control groups included acellular PCL/ECM, a positive control cultured in +dex media (0 ng/mL +dex), and two negative controls cultured in −dex media (0 ng/mL TNF) containing MSCs precultured in either +dex or −dex media (the latter is denoted with “pre: −dex”). Bars represent average ± standard deviation of (n = 4) scaffolds per group at each timepoint. Within a timepoint, groups with * differ from all other groups (p < 0.05), while letters A–D indicate groups that differ from all other groups but not groups with the same letter (p < 0.05). On day 16, + indicates that d4 and d16 values of that group differ (p < 0.05), while # indicates that values at all timepoints differ (p < 0.05).

Figure 5

ALP activity of MSCs on preformed ECM with varying TNF-α dose. Scaffolds with pregenerated bone-like ECM (PCL/ECM) were seeded with MSCs precultured in +dex media (except for the group designated “(pre: −dex),” which was seeded with MSCs precultured in −dex media). ALP activity on a per scaffold basis is shown in (a), while the same data on a per cell basis is shown in (b) with the exception of the acellular group. Experimental groups were exposed to varying TNF-α doses. The control groups included acellular PCL/ECM, a positive control cultured in +dex media (0 ng/mL +dex), and two negative controls cultured in −dex media (0 ng/mL TNF) containing MSCs precultured in either +dex or −dex media (the latter is denoted with “pre: −dex”). Bars represent average ± standard deviation of (n = 4) scaffolds per group at each timepoint. Within a timepoint, groups with * differ from all other groups (p < 0.05), while letters A–F indicate groups that differ from all other groups but not groups with the same letter (p < 0.05). On day 16, + indicates that d4 and d16 values of that group differ (p < 0.05), while # indicates that values at all timepoints differ (p < 0.05).

Figure 6

MSC calcium deposition on PCL/ECM with varying TNF-α dose. The baseline calcium content of the PCL scaffolds with pregenerated bone-like ECM (PCL/ECM) was 140 ± 13 μg Ca²⁺/scaffold, which is indicated by the dashed line. PCL/ECM scaffolds were seeded with MSCs precultured in +dex media (except for the group designated “(pre: −dex),” which was seeded with MSCs precultured in −dex media). Experimental groups were exposed to varying TNF-α doses. The control groups included acellular PCL/ECM, a positive control cultured in +dex media (0 ng/mL +dex), and two negative controls cultured in −dex media (0 ng/mL TNF) containing MSCs precultured in either +dex or −dex media (the latter is denoted with “pre: −dex”). Bars represent average ± standard deviation of (n = 4) scaffolds per group at each timepoint. Within a timepoint, groups with * differ from all other groups (p < 0.05), while letters A–C indicate groups that differ from all other groups but not groups with the same letter (p < 0.05). On day 16, + indicates that d4 and d16 values of that group differ (p < 0.05), while # indicates that values at all timepoints differ (p < 0.05).

Figure 7

Surface morphology of PCL/ECM after 16 days of culture. On the final day of the study, the surface of one half of a representative scaffold from each group was imaged via SEM. Images in each row are from the same group, and images in each column have the same magnification. Specifically, the images depict: (a–c) the positive control (0 ng/mL +dex in other Figures); (d–f) the negative control cultured in −dex media (0 ng/mL TNF); (g–i) the negative control seeded with dexamethasone-naïve MSCs and cultured in −dex media (0 ng/mL TNF (pre: −dex)); and (j–l) acellular PCL/ECM (Acellular). The final three rows of images depict the experimental groups cultured with varying doses of TNF-α: (m–o) 0.1 ng/mL, (p–r) 5 ng/mL, and (s–u) 50 ng/mL. The original magnification of the images in the columns (from left to right) is: 200× (scale bar is 200 μm), 400× (scale bar is 100 μm), 1000× (scale bar is 25 μm). For all columns, the scale bars at the top and bottom of each column apply to all images within that column.

Figure 8

PCL/ECM construct cell and mineral distribution visualized by histology after 16 days of culture (cross-sectional view). Group names are the same as previous figures. 5 μm thick sections were stained with von Kossa, which stains mineralized matrix black (blue arrows), and eosin, which stains cells and non-mineralized matrix reddish-pink (yellow arrows). Images were captured at 10× original magnification. Scale bar in the lower right corner represents 200 μm and applies to all images.