Cooperative effects of fibronectin matrix assembly and initial cell-substrate adhesion strength in cellular self-assembly

Abstract

The cell-dependent polymerization of intercellular fibronectin fibrils can stimulate cells to self-assemble into multicellular structures. The local physical cues that support fibronectin-mediated cellular self-assembly are largely unknown. Here, fibronectin matrix analogs were used as synthetic adhesive substrates to model cell-matrix fibronectin fibrils having different integrin-binding specificity, affinity, and/or density. We utilized this model to quantitatively assess the relationship between adhesive forces derived from cell-substrate interactions and the ability of fibronectin fibril assembly to induce cellular self-assembly. Results indicate that the strength of initial, rather than mature, cell-substrate attachments correlates with the ability of substrates to support fibronectin-mediated cellular self-assembly. The cellular response to soluble fibronectin was bimodal and independent of the integrin-binding specificity of the substrate; increasing soluble fibronectin levels above a critical threshold increased aggregate cohesion on permissive substrates. Once aggregates formed, continuous fibronectin polymerization was necessary to maintain cohesion. During self-assembly, soluble fibronectin decreased cell-substrate adhesion strength and induced aggregate cohesion via a Rho-dependent mechanism, suggesting that the balance of contractile forces derived from fibronectin fibrils within cell-cell versus cell-substrate adhesions controls self-assembly and aggregate cohesion. Thus, initial cell-substrate attachment strength may provide a quantitative basis with which to build predictive models of fibronectin-mediated microtissue fabrication on a variety of substrates.

Statement of significance: Cellular self-assembly is a process by which cells and extracellular matrix (ECM) proteins spontaneously organize into three-dimensional (3D) tissues in the absence of external forces. Cellular self-assembly can be initiated in vitro, and represents a potential tool for tissue engineers to organize cells into modular building blocks for artificial tissue fabrication. Fibronectin is an ECM protein that plays a key role in tissue formation during embryonic development. Additionally, the cell-mediated process of converting soluble fibronectin into insoluble, ECM-associated fibrils has been shown to initiate cellular self-assembly in vitro. In this study, we examine the relationship between the strength of cell-substrate adhesions and the ability of fibronectin fibril assembly to induce cellular self-assembly. Our results indicate that substrate composition and density play cooperative roles with cell-mediated fibronectin matrix assembly to control the transition of cells from 2D monolayers into 3D multicellular aggregates. Results of this study provide a quantitative approach to build predictive models of cellular self-assembly, as well as a simple cell-culture platform to produce biomimetic units for modular tissue engineering.

Keywords: Biomimetic material; Cell adhesion; Extracellular matrix; Fibronectin; Self-assembly.

Fig. 1

Fibronectin matrix mimetics. (A) Schematic representation of a fibronectin subunit and fibronectin fusion proteins. (B) FN-null MEFs were seeded (1.6 × 10⁵ cells/cm²) onto wells pre-coated with FNIII1H^RGD (100 nM) in the presence of integrin function-blocking antibodies, isotype-matched control antibodies, or EDTA. Cell adhesion was determined as described in “Methods”. Data are presented as mean fold difference compared to IgG (α₅, α_v, β₃) or IgM (β₁) controls ± SEM of 3 experiments performed in triplicate. *Significantly different from +IgG, p < 0.05 (ANOVA). (C) Tissue culture plates were coated with increasing concentrations of proteins. The relative amount of protein bound to wells was determined by ELISA. Data are presented as mean absorbance of triplicate wells ± SEM and represent of 1 of 4 experiments performed. (D) FN-null MEFs were seeded (1.6 × 10⁵ cells/cm²) onto protein-coated wells and allowed to attach for 4 h. The number of adherent cells was determined as described in “Methods”. Data are presented as mean absorbance of triplicate wells ± SEM and represent of 1 of 3 experiments performed. *Significantly different from FNIII1H,8^RGD and FNIII1H,8-10, p < 0.05 (ANOVA).

Fig. 2

Fibronectin matrix mimetics support cellular self-assembly. FN-null MEFs were seeded (2 × 10⁴ cell/cm²) on tissue culture plates pre-coated with FNIII1H^RGD, FNIII1H,8^RGD, or FNIII1H,8-10 at coating concentrations ranging from 12.5 – 800 nM. Soluble plasma fibronectin (25 nM) was added 4 h post-seeding. Phase contrast microscopy images were taken at 72 h post-seeding. Images represent 1 of 3 experiments performed. Scale bar, 50 μm.

Fig. 3

Effect of substrate density on aggregate diameter. FN-null MEFs were seeded (2 × 10⁴ cell/cm²) on wells pre-coated with indicated coating concentrations of FNIII1H^RGD (A, B; black bars), FNIII1H,8^RGD (B; grey bars) or FNIII1H,8-10 (B; white bars). Soluble plasma fibronectin (25 nM) was added 4 h post-seeding. Phase contrast microscopy images were obtained 72 h post-seeding, capturing at least 30 aggregates per sample. Spheroid diameter was quantified from images using ImageJ. Data are presented as mean aggregate diameter ± SEM of 3 experiments. (A) *Significantly different from 12.5 nM, p < 0.05; (B) *Significantly different, p < 0.05 (ANOVA).

Fig. 4

Initial cell-substrate binding strength of fibronectin matrix mimetics. Tissue culture plates (96-well) were pre-coated with increasing concentrations of FNIII1H^RGD, FNIII1H,8^RGD, or FNIII1H,8-10. FN-null MEFs (1.6 × 10⁵ cell/cm²) were added to wells on ice and centrifugal adhesion assays were performed as described in “Methods”. GraphPad Prism was used to fit a sigmoidal curve to the data. Data are presented as mean absorbance of triplicate wells ± SEM and represent 1 of at least 3 independent experiments performed.

Fig. 5

Bimodal response to soluble fibronectin. FN-null MEFs were seeded (2 × 10⁴ cells/cm²) on tissue culture plates pre-coated with fibronectin matrix mimetics at concentrations above and below the initial C^c50 value (400 nM and 800 nM for FNIII1H^RGD; 100 nM and 400 nM for FNIII1H,8^RGD; 25 nM and 50 nM for FNIII1H,8-10). Four h after seeding, cells were treated with various concentrations of soluble fibronectin (6.25 – 100 nM). Microscopy images were taken at 72 h post-seeding. Images represent 1 of 3 experiments performed. Scale bar, 50 μm.

Fig. 6

Fibronectin-induced cellular self-assembly at early and late stages of cell spreading. FN-null MEFs were seeded (2 × 10⁴ cells/cm²) on tissue culture plates pre-coated with FNIII1H,8^RGD at various concentrations (25 - 800 nM). At either 4 h or 24 h after seeding, cells were treated with soluble fibronectin (25 nM). Phase contrast microscopy images were taken at 72 h post-seeding. Images represent 1 of 3 experiments performed. Scale bar, 50 μm.

Fig. 7

Fibronectin fibril formation within 3D aggregates. FN-null MEFs were seeded (2 × 10⁴ cells/cm²) onto plates pre-coated with 400 nM FNIII1H^RGD, allowed to adhere for 4 h, and then treated with 25 nM fibronectin. Cells were processed for immunofluorescence microscopy at 24 (A), 48 (B), or 72 (C,D) h post-seeding and immunostained for fibronectin. Images were collected along the z-axis at 1μm intervals using two-photon microscopy. Image in (C) is a representative slice taken 40 μm above the surface of the culture plate. (D) Z-slice images were reconstructed in 3D and projected along the x-y plane. Images represent 1 of 3 experiments performed. Scale bars, 50 μm.

Fig. 8

Fibronectin matrix assembly maintains aggregate cohesion. FN-null MEFs were seeded (2 × 10⁴ cells/cm²) onto plates pre-coated with 400 nM FNIII1H^RGD. At 4 h post-seeding, 25 nM fibronectin was added to initiate aggregate formation. At 72 h post-seeding, phase contrast microscopy images were obtained, and FUD peptides (250 nM) were added directly to wells to inhibit fibronectin matrix assembly. Control wells received the inactive control, del29 (250 nM). Samples were incubated for an additional 2 d and images were obtained on each day, at 96 h and 120 h post-seeding. Images represent 1 of 3 experiments performed. Scale bars, 50 μm.

Fig. 9

Fibronectin decreases cell-substrate attachment strength during cellular self-assembly. FN-null MEFs were seeded (4 × 10⁴ cells/cm²) onto tissue culture plates pre-coated with increasing concentrations of FNIII1H^RGD, FNIII1H,8^RGD, or FNIII1H,8-10 and allowed to adhere for 4 h. Cells were then incubated in the absence (−sFN) or presence (+sFN) of soluble fibronectin (50 nM) for an additional 20 h. Centrifugal cell adhesion assays were performed and the number of cells that remained adherent was determined, as described in “Methods”. Data are presented as mean absorbance ± SEM of at least 3 separate experiments, each performed in triplicate. *Significantly different from corresponding ‘−sFN’, p < 0.05 (t-test). The dotted lines denote the initial Cc⁵⁰ values.

Fig. 10

Effect of Rho inhibition on fibronectin-mediated cellular self-assembly. FN-null MEFs were seeded (2 × 10⁴ cells/cm²) onto tissue culture plates pre-coated with 200 nM (A, B) or 400 nM (C-F) FNIII1H^RGD, in the presence of either 10 μM Y-27632 (B, D, F) or an equal volume of the vehicle, PBS (A, C, E). At 4 h post-seeding, cells were treated with 25 nM soluble fibronectin (+sFN) and incubated for up to 3 d. At 72 h post-seeding, phase contrast microscopy images were obtained (A-D). At 24 h post-seeding, some wells were processed for immunofluorescence microscopy and fibronectin fibrils were visualized using polyclonal anti-fibronectin antibodies (E, F). Images represent 1 of 3 experiments performed. Scale bars, 50 μm.