Fibrocartilage tissue engineering: the role of the stress environment on cell morphology and matrix expression

Abstract

Although much is known about the effects of uniaxial mechanical loading on fibrocartilage development, the stress fields to which fibrocartilaginous regions are subjected to during development are mutiaxial. That fibrocartilage develops at tendon-to-bone attachments and in compressive regions of tendons is well established. However, the three-dimensional (3D) nature of the stresses needed for the development of fibrocartilage is not known. Here, we developed and applied an in vitro system to determine whether fibrocartilage can develop under a state of periodic hydrostatic tension in which only a single principal component of stress is compressive. This question is vital to efforts to mechanically guide morphogenesis and matrix expression in engineered tissue replacements. Mesenchymal stromal cells in a 3D culture were exposed to compressive and tensile stresses as a result of an external tensile hydrostatic stress field. The stress field was characterized through mechanical modeling. Tensile cyclic stresses promoted spindle-shaped cells, upregulation of scleraxis and type one collagen, and cell alignment with the direction of tension. Cells experiencing a single compressive stress component exhibited rounded cell morphology and random cell orientation. No difference in mRNA expression of the genes Sox9 and aggrecan was observed when comparing tensile and compressive regions unless the medium was supplemented with the chondrogenic factor transforming growth factor beta3. In that case, Sox9 was upregulated under static loading conditions and aggrecan was upregulated under cyclic loading conditions. In conclusion, the fibrous component of fibrocartilage could be generated using only mechanical cues, but generation of the cartilaginous component of fibrocartilage required biologic factors in addition to mechanical cues. These studies support the hypothesis that the 3D stress environment influences cell activity and gene expression in fibrocartilage development.

FIG. 1.

Schematic of preparation, conditioning, and mechanical testing of cell-populated collagen matrices. Solubilized collagen and cells are mixed in Dulbecco's modified Eagle's medium (a) and poured into Teflon casting wells (b). After a 48 h incubation (c), the mandrel is removed from the well (d) and the remodeled matrix ring is removed from the mandrel (e). The matrix ring is then mounted between spacers for long-term culture (f). Finally, the matrix is examined for cell morphology and gene expression (g). Color images available online at www.liebertonline.com/tea.

FIG. 2.

Ring-shaped CSCMs were cyclically loaded in vitro (top left). The in vitro loading setup resulted in regions of the matrix that loaded in tension and regions of the matrix that were loaded in compression. This is similar to the loading environment of tendons of the hand that wrap around bony pulleys (bottom left). Cells in the tensile region (top right) were spindle shaped and aligned with the direction of loading, whereas cells in the compressive region were rounded without a preferred orientation (bottom right, 20× objective). CSCM, mesenchymal stromal cell seeded collagen matrix.

FIG. 3.

(a) Mechanical modeling was used to examine stress field in a CSCM wrapped around a cylindrical loading bar or, equivalently, a pulley-like representation of a wrap-around tendon. (b), (c) Symmetric boundary conditions were applied on the left and lower surfaces. The free end of the tendon was displaced downward uniformly and allowed to contract freely in the horizontal direction. This motion required a net force F. (d) The domain Ω, a quarter of the entire domain, was modeled using ∼10,000 quadratic interpolation plane stress elements.

FIG. 4.

The radial stress distribution along the contact surface as a function of angular (counterclockwise) position. The radial stress is negative (i.e., compressive) throughout the contact region. The compressive radial stress remains fairly constant along the contact surface. The compressive stress increases at the point where the collagen matrix begins to lose contact with the post. Finally, the radial stress goes to zero as the collagen matrix loses contact with the post. (a) Compressive stresses increased with specimen thickness (note: every 10th data point is plotted in (a)). The transition zone before the peak compressive stress was reached increased with specimen thickness. Results pictured are for an isotropic CSCM and loading bar displacement of 0.02a. (b) Anisotropy increased the angular width of the transition zone, but had little effect on the magnitude of the peak compressive stress. Results pictured are for b/a = 1.05, and G_rθ/E_θ = 0.005.

FIG. 5.

The analytical stress field within the compressive zone plotted as a function of the degree of anisotropy. The range of anisotropy shown on the abscissa corresponds to the CSCMs in vitro on the lower extreme and to tendons in vivo on the upper extreme. Data are plotted for (b/a) = 1.2. Stress in the radial direction, shown by σ_rr, is always negative (i.e., compressive), whereas stress in the θ direction is always positive (i.e., tensile). The hydrostatic pressure predicted in the contact region of a wrap-around tendon is tensile, supporting the hypothesis that only a single compressive stress component is needed for a development of the fibrocartilage. Note that the hydrostatic stress is (σ_r + σ_θ + σ_z)/3, where σ_z ≈ 0. Labels B and C delineate stresses at the locations indicated on the schematic of the loading system.

FIG. 6.

Kernel curve histograms for cell shape factor are shown for one representative cyclically loaded collagen matrix. Cells in the compressive region of the matrix were rounder than cells in the tensile region.

FIG. 7.

Kernel curve histograms for cell orientation are shown for one representative cyclically loaded collagen matrix. Cells in the tensile region of the cyclically loaded matrix were more aligned compared to the cells in the compressive region. Cell alignment was centered around zero degrees, defined as the direction of cyclic loading.

FIG. 8.

Scleraxis and collagen I mRNA expression levels were significantly higher in the tensile region (T) of the cyclically loaded CSCMs compared to the compressive region (C). Cyclic loading led to an upregulation of both Scleraxis and Collagen I relative to static loading. Both tendon-specific genes were upregulated relative to time zero control CSCMs (CTL). Scleraxis was significantly downregulated due to TGF-β3. Aggrecan and Sox9 expression was significantly upregulated due to TGF-β3 relative to time zero and control media CSCMs (^#p < 0.05 for comparisons under the bars, *p < 0.05 compared to CTL, ^&p < 0.05 Static vs. Cyclic). TGF-β3, transforming growth factor beta3.

FIG. 9.

Immunohistochemistry results for type II collagen (cell nuclei shown in blue; type II collagen shown in green; 40× objective; scale bar = 100 μm). Color images available online at www.liebertonline.com/tea.