Strain rate-dependent induction of reactive astrogliosis and cell death in three-dimensional neuronal-astrocytic co-cultures

Abstract

A mechanical insult to the brain drastically alters the microenvironment as specific cell types become reactive in an effort to sequester severely damaged tissue. Although injury-induced astrogliosis has been investigated, the relationship between well-defined biomechanical inputs and acute astrogliotic alterations is not well understood. We evaluated the effects of strain rate on cell death and astrogliosis using a three-dimensional (3-D) in vitro model of neurons and astrocytes within a bioactive matrix. At 21 days post-plating, co-cultures were deformed to 0.50 shear strain at strain rates of 1, 10, or 30 s(-1). We found that cell death and astrogliotic profiles varied differentially based on strain rate at 2 days post-insult. Significant cell death was observed after moderate (10 s(-1)) and high (30 s(-1)) rate deformation, but not after quasi-static (1 s(-1)) loading. The vast majority of cell death occurred in neurons, suggesting that these cells are more susceptible to high rate shear strains than astrocytes for the insult parameters used here. Injury-induced astrogliosis was compared to co-cultures treated with transforming growth factor beta, which induced robust astrocyte hypertrophy and increased glial fibrillary acidic protein (GFAP) and chondroitin sulfate proteoglycans (CSPGs). Quasi-static loading resulted in increased cell density and CSPG secretion. Moderate rate deformation increased cell density, GFAP reactivity, and hypertrophic process density. High rate deformation resulted in increased GFAP reactivity; however, other astrogliotic alterations were not observed at this time-point. These results demonstrate that the mode and degree of astrogliosis depend on rate of deformation, demonstrating astrogliotic augmentation at sub-lethal injury levels as well as levels inducing cell death.

Fig. 1. Schematic of the 3-D neuronal-astrocytic co-culture and mechanical deformation models (not to scale)

Mechanical deformation was imparted to cell-containing matrices through the action of the 3-D Cell Shearing Device (3-D CSD), a custom-built electromechanical device utilizing a linear actuator under closed-loop proportional-integral-derivative control with positional feedback from an optical position sensor (A). Neuronal-astrocytic co-cultures in 3-D were plated throughout the thickness of a bioactive matrix and were laterally constrained by an elastomer mold. Shear deformation of the elastomer mold and cell-embedded matrices was induced through horizontal displacement of the cell chamber top-plate, which was coupled to the linear actuator (B). Input to the system was a symmetrical trapezoidal input to a constant displacement (corresponding to 0.50 bulk shear strain) at strain rates of 1, 10, or 30 s⁻¹ (corresponding to rise times of 500, 50, or 16.7 ms, respectively; hold time was held constant at 5 ms) (C).

Fig. 2. Culture viability and cell density following mechanical loading or TGF-β1 treatment

3-D neuronal-astrocytic co-cultures were subjected to mechanical deformation, TGF-β1 treatment, or control conditions at 21 DIV and culture viability was assessed two days later. Fluorescent confocal reconstructions of representative co-cultures after control conditions (A), TGF-β1 treatment (B), or mechanical loading at 0.50 strain at low (1 s⁻¹, C), moderate (10 s⁻¹, D), or high (30 s⁻¹, E) strain rates (live cells shown green and nuclei of dead cells shown red; scale bar = 50 μm). High rate deformation resulted in a significant reduction in culture viability* (p < 0.05) while quasi-static deformation or TGF-β1 treatment had no effect on culture viability (F). Furthermore, there was a significant increase in cell density following mechanical loading at low and moderate strain rates compared to controls^† (p < 0.05), suggesting a hyperplasic response. There was also a significant increase in the density of dead cells following moderate and high rate deformation compared to control cultures ^‡ (p < 0.05).

Fig. 3. Assessment of the phenotype of living/dead cells at two days following high strain rate mechanical injury

A fluorescent TUNEL assay was used in conjunction with immunocytochemistry to identify the phenotype of cells undergoing death. GFAP (A) was used to label astrocytes, while MAP-2 (B) was used as a marker for neurons. TUNEL staining labeled the nucleus of dead cells (C), and Hoechst was used as a counterstain for identification of all nuclei within the co-culture (D). TUNEL staining was most prominently found in neurons, with few TUNEL⁺ astrocytes at this time-point. Scale bar = 20 μm.

Fig. 4. Derivation of markers of astrogliosis

In order to develop the reactive astrogliotic outcome measures utilized in this study, control co-cultures (A, C) were compared to co-cultures treated with TGF-β1 for 14 days (B, D). Alterations in astrocyte morphology (GFAP) were observed following TGF-β1 (B) as compared to untreated control cultures (A) as many hypertrophic processes were observed (arrows). Also, there were robust increases in CSPG expression in the matrix following 14 day TGF-β1 treatment (D) compared to untreated controls (C). Scale bar = 20 μm.

Fig. 5. Astrocyte GFAP reactivity and hypertrophy following mechanical deformation or TGF-β1 treatment

Co-cultures immunolabeled for GFAP (red) with nuclear counterstain (blue) at two days following control (A), TGF-β1 (B), quasi-static (C), moderate rate (D) and high rate (E) deformation (scale bar = 20 μm), revealed increased GFAP reactivity and process density following TGF-β1 treatment and deformation. TGF-β1 treatment and moderate rate deformation induced signicant increases in the density of hypertrophic processes at this time-point compared to control cultures* (p < 0.05) (F).

Fig. 6. Expression and localization of CSPGs following deformation or TBF-β1 treatment

Fluorescent micrographs of representative neuronal-astrocytic co-cultures at 23 DIV, two days following control conditions (A), TGF-β1 treatment (B) or high rate deformation (C) (scale bar = 20μm); cultures were immunolabeled for CSPGs (red) and GFAP (green) with nuclear counterstain (blue). GFAP⁺ hypertrophic astrocytes (green) (D) were observed adjacent to CSPGs (red) (E); overlay of previous two photomicrographs (F) following mechanical loading (scale bar = 10μm). There was a significant increase in CSPG deposition in the matrix following TGF-β1* (p < 0.001) (G). There were also significant increases in the CSPG content in the medium following quasi-static (1 s⁻¹) deformation and TGF-β1 treatment compared to controls* (p < 0.05) PG = proteoglycan; GAG = glycosaminoglycan (H).

Fig. 7. Postulated astrocytic responses based on mechanical insult severity

Conceptual framework describing a proposed continuum in astrocytic response based on insult severity, ranging from reactive astrogliosis (i.e. programmed physiological response) to necrotic/apoptotic cell death (e.g., surpassing biophysical thresholds).