Mechanical characterization of differentiated human embryonic stem cells

Abstract

Human embryonic stem cells (hESCs) possess an immense potential in a variety of regenerative applications. A firm understanding of hESC mechanics, on the single cell level, may provide great insight into the role of biophysical forces in the maintenance of cellular phenotype and elucidate mechanical cues promoting differentiation along various mesenchymal lineages. Moreover, cellular biomechanics can provide an additional tool for characterizing stem cells as they follow certain differentiation lineages, and thus may aid in identifying differentiated hESCs, which are most suitable for tissue engineering. This study examined the viscoelastic properties of single undifferentiated hESCs, chondrogenically differentiated hESC subpopulations, mesenchymal stem cells (MSCs), and articular chondrocytes (ACs). hESC chondrogenesis was induced using either transforming growth factor-beta1 (TGF-beta1) or knock out serum replacer as differentiation agents, and the resulting cell populations were separated based on density. All cell groups were mechanically tested using unconfined creep cytocompression. Analyses of subpopulations from all differentiation regimens resulted in a spectrum of mechanical and morphological properties spanning the range of hESCs to MSCs to ACs. Density separation was further successful in isolating cellular subpopulations with distinct mechanical properties. The instantaneous and relaxed moduli of subpopulations from TGF-beta1 differentiation regimen were statistically greater than those of undifferentiated hESCs. In addition, two subpopulations from the TGF-beta1 group were identified, which were not statistically different from native articular chondrocytes in their instantaneous and relaxed moduli, as well as their apparent viscosity. Identification of a differentiated hESC subpopulation with similar mechanical properties as native chondrocytes may provide an excellent cell source for tissue engineering applications. These cells will need to withstand any mechanical stimulation regimen employed to augment the mechanical and biochemical characteristics of the neotissue. Density separation was effective at purifying distinct populations of cells. A differentiated hESC subpopulation was identified with both similar mechanical and morphological characteristics as ACs. Future research may utilize this cell source in cartilage regeneration efforts.

Figure 1

Illustration of the creep cytocompression apparatus. A piezoelectric actuator drives a 50.8 μm tungsten probe axially toward cells seeded onto a culture dish and the free end of the probe is simultaneously tracked by a laser micrometer. The difference in recorded displacement by the laser micrometer and piezoelectric motor results in a probe deflection (δ), which is correlated to a reaction force using cantilever beam theory. Through a negative feedback algorithm, the position of the probe is continuously altered to hold a step load of 100 nN onto the single cells. An inverted objective located below the stage is used to position the probe and the cell, as well as measure cell diameters.

Figure 2

Histological sections of EBs differentiated with TGF-β1 (column 1) and KOSR (column 2). Original magnification, 40X. Collagen type I, collagen type II, and s-GAGs well all present, indicating a cartilaginous differentiation of the hESCs.

Figure 3

Density separation of differentiated hESC subpopulations with the Percoll gradient technique. For this figure, chondro-induction was achieved with KOSR. Differentiated hESCs were centrifuged through Percoll solutions ranging from 10 to 60% and the cell interface between each density layer was counted and seeded for cytocompression testing. The majority of the KOSR cells (52.7%) fell within the 20–30% density interfaces.

Figure 4

Representative creep curves of single cells. The experimental data points were fitted to a viscoelastic model to yield an instantaneous modulus, relaxed modulus, apparent viscosity, and a creep time constant. Undifferentiated hESCs typically exhibited a greater deformation, suggestive of a lower stiffness, and a faster time to equilibrium, suggestive of a lower apparent viscosity and time constant, than differentiated hESCs (example shown from TGF 40–50 group) in response to the same applied load. In addition, the equilibrium deformation of single cells from the TGF 40–50 subpopulation was akin to that of native chondrocytes, indicative of their similar stiffness values. For clarity, only one out of every 1000 experimental data points is shown for each curve. Representative cells were of 12 μm height.

Figure 5

Viscoelastic material properties of undifferentiated and differentiated single hESCs, as well as mesenchymal stem cell and articular chondrocyte controls. Differences in instantaneous and relaxed moduli were observed between density interfaces (TGF 20–30 vs TGF 30–40 or TGF 40–50), differentiation regimens (TGF-β1 vs. KOSR), and differentiation state (hESC vs. MSC vs. TGF 40–50) (A). Moreover, differentiated cell subpopulations (TGF 30–40 and TGF 40–50) were identified which were not different than native ACs. In addition, the apparent viscosities of the TGF 30–40 and TGF 40–50 groups were greater than that of undifferentiated hESCs (B). Data presented as mean ± standard deviations.

Figure 6

Morphological properties of undifferentiated and differentiated single hESCs, as well as mesenchymal stem cell and articular chondrocyte controls. The cell height-to-width ratios of all differentiated cell subpopulations were greater than that of undifferentiated hESCs and approaching 1.0, which is indicative of a more rounded cell morphology. The values for differentiated hESCs all fell between the range of MSCs to ACs. Moreover, there was no difference between the cell height-to-width ratio of the TGF 40–50 and KOSR 30–40 cells and native ACs. Data presented as mean ± standard deviations.