Engineering a 3D hydrogel system to study optic nerve head astrocyte morphology and behavior

Abstract

In glaucoma, astrocytes within the optic nerve head (ONH) rearrange their actin cytoskeleton, while becoming reactive and upregulating intermediate filament glial fibrillary acidic protein (GFAP). Increased transforming growth factor beta 2 (TGF β2) levels have been implicated in glaucomatous ONH dysfunction. A key limitation of using conventional 2D culture to study ONH astrocyte behavior is the inability to faithfully replicate the in vivo ONH microenvironment. Here, we engineer a 3D ONH astrocyte hydrogel to better mimic in vivo mouse ONH astrocyte (MONHA) morphology, and test induction of MONHA reactivity using TGF β2. Primary MONHAs were isolated from C57BL/6J mice and cell purity confirmed. To engineer 3D cell-laden hydrogels, MONHAs were mixed with photoactive extracellular matrix components (collagen type I, hyaluronic acid) and crosslinked for 5 minutes using a photoinitiator (0.025% riboflavin) and UV light (405-500 nm, 10.3 mW/cm²). MONHA-encapsulated hydrogels were cultured for 3 weeks, and then treated with TGF β2 (2.5, 5.0 or 10 ng/ml) for 7 days to assess for reactivity. Following encapsulation, MONHAs retained high cell viability in hydrogels and continued to proliferate over 4 weeks as determined by live/dead staining and MTS assays. Sholl analysis demonstrated that MONHAs within hydrogels developed increasing process complexity with increasing process length over time. Cell processes connected with neighboring cells, coinciding with Connexin43 expression within astrocytic processes. Treatment with TGF β2 induced reactivity in MONHA-encapsulated hydrogels as determined by altered F-actin cytoskeletal morphology, increased GFAP expression, and elevated fibronectin and collagen IV deposition. Our data sets the stage for future use of this 3D biomimetic ONH astrocyte-encapsulated hydrogel to investigate astrocyte behavior in response to injury.

Keywords: Biomechanical Strain; Collagen IV; Extracellular matrix; Fibronectin; GFAP; Glaucoma; Reactive gliosis; Transforming growth factor beta 2.

Fig. 1.

Schematic of MONHA-encapsulated hydrogel formulation.

Fig. 2.

MONHA-encapsulated hydrogel stiffness, viability, and proliferation. (A) Elastic modulus of acellular and MONHA-encapsulated hydrogels (N = 4/group, *p = 0.0374). (B) Representative Live (green)/Dead (red) fluorescence images from 3 independent MONHA isolations immediately after crosslinking (N = 3/group). Pink percentage values depict average % live cells per hydrogel. Scale bars: 500 μm (top), 250 μm (bottom). (C) Longitudinal Live (green)/Dead (red) fluorescence images of representative hydrogels. Scale bar 250 μm. (D) Normalized cell proliferation over time (7 d, 14 d, 21 d, and 28 d; shared significance indicator letters represent nonsignificant difference (p > 0.05), distinct letters represent significant difference (p < 0.05)). Significance was determined by unpaired t-test (A) and two-way ANOVA using multiple comparisons tests (D) (*p < 0.05).

Fig. 3.

MONHA stellate morphology and astrocytic marker expression in hydrogels. (A) Representative fluorescence images of astrocyte morphology within hydrogels at 7 d, 14 d, 21 d, and 28 d. Scale bar: 100 μm. (B) Representative fluorescence images of astrocytes expressing GFAP (green) and (C) CX43 (green) in hydrogels at 28d. Scale bar: 100 μm. (B′-C′) Magnified images from boxed regions (orange) showing astrocytes expressing either GFAP (upper right, white arrows) or CX43 puncta (lower right, white arrows). Scale bar: 100 μm.

Fig. 4.

MONHA process length and branching in hydrogels. (A–C) F-actin staining and tracing of astrocytic morphology over time. Scale bar: 100 μm (A–B), 50 μm (C). (D–E) Process length and degree of branching of astrocytes. (F) Sholl analysis indicating the number of process intersections at each increasing radius from cell body for 7d, 14d, and 21d (N = 10 cells/group). Statistical significance was determined using one-way ANOVA (****p < 0.0001) for process length (D) and degree of branching (E), and two-way ANOVA main effects only model for process complexity over time (F) (main effect of time F (149, 2012) = 8.778, ****p < 0.0001, and main effect of process complexity F (2, 2012) = 3.992, *p < 0.05).

Fig. 5.

TGFβ2 effect on F-actin network in MONHA-encapsulated hydrogels. (A) Representative confocal fluorescence images of F-actin expression levels in control versus TGFβ2-treated MONHA-encapsulated hydrogels (2.5 ng/ml, 5 ng/ml, 10 ng/ml). Scale bar: 100 μm. (A′) Magnified images from boxed regions (yellow) of F-actin cytoskeletal changes (white arrows) in control versus TGFβ2-treated MONHA-encapsulated hydrogels (2.5 ng/ml, 5 ng/ml, 10 ng/ml). Scale bar: 50 μm. (B) Quantification of fold change in F-actin intensity (N = 10 fields of view/group). Statistical significance was determined using one-way ANOVA (****p < 0.0001) for F-actin expression levels (B).

Fig. 6.

Effect of TGFβ2 on ECM protein and GFAP levels. (A–C) Representative fluorescence images of collagen IV, fibronectin, GFAP expression in control versus 5 ng/ml TGFβ2-treated MONHA-encapsulated hydrogels. Scale bar: 100 μm. (D–F) Quantification of fold change in signal intensity for collagen IV, fibronectin deposition, and GFAP expression shows significant difference between groups. (N = 5–7 fields of view/group for 2 strains). Statistical significance was determined using unpaired t-test (**p < 0.001, ****p < 0.0001).