Glucocorticoid-Induced Reductions of Myelination and Connexin 43 in Mixed Central Nervous System Cell Cultures Are Prevented by Mifepristone

Abstract

Repeated stress induces systemic elevations in glucocorticoid levels. Stress is also associated with alterations in central nervous system astrocytes and oligodendrocytes that involve connexins and myelin proteins. Corticosteroid elevation seems a major factor in stress-induced neuropathology. Changes in astrocyte connexins and myelin components may be important mediators for the neurological effects of corticosteroid elevations. Two primary cell culture models, myelination culture from rat embryonic spinal cord (SC) or cerebral cortex (CC) consisting of neurons and glial cells (oligodendrocytes, microglia and astrocytes), and mixed astrocyte-and-oligodendrocyte culture prepared from postnatal rat CC, were used in this study. Cell cultures were treated with either vehicle, corticosterone (CORT) with or without glucocorticoid receptor antagonist mifepristone, or dexamethasone (DEX) during the period of in vitro myelination. Immunoreactivity of astrocyte connexin 43 (Cx43) and oligodendrocyte myelin basic protein (MBP), or the myelination index (co-localization of MBP and phosphorylated neurofilament) was determined by double immunofluorescent labeling. Oligodendrocyte morphology was evaluated by Sholl analysis. Prolonged exposure to CORT or DEX induced dose-dependent reduction of the myelination index, and of immunostaining for MBP and Cx43 in SC and CC myelination cultures, which was prevented by mifepristone. In glial cultures single CORT or DEX exposure caused shrinkage and simplification of/' MBP- or CNPase-positive oligodendrocyte processes. The results support that concurrent effects of glucocorticoids on myelination and astrocyte Cx43 immunoreactivity are mediated by glucocorticoid receptors and may partially account for the involvement of CNS glia in the pathological effects of prolonged stress.

Keywords: astrocytes; corticosteroids; development; dexamethasone; glucocorticoid receptors; myelin.

Fig. 1:

Myelination in spinal cord (SC)-derived cultures based on MBP immunoreactivity. Micrographs taken from a microscopic field containing structures labeled by fluorescence immunolabeling of MBP (A) and pNF (B), in addition to nuclear stain DAPI (C) in a mixed culture from the SC at DIV28. The three stains are merged in D to illustrate MBP+ labeling along pNF+ axons (arrows) and the presence of MBP+ oligodendrocytes (arrowheads). The lightning bolt-shaped arrows point to an MBP-immunoreactive membrane-like expansions that were often found in control cultures at this stage of the myelination process. Calibration bar: 100 μm.

Fig. 2:

CORT and DEX treatment significantly reduced myelination index and MBP area fraction in the SC cultures. Top panels: Micrographs from microscope fields of CONTROL (left panel), 50 μM CORT (CORT50) (middle) and 50 μM DEX (DEX50) (right) cultures at DIV28 with nuclear DAPI staining (blue), and immunofluorescence labeling of MBP (green or yellow) and neurofilament (red). Note that linear stretches of MBP immunostaining are readily identifiable (white arrows) along pNF immunolabeled axons in the CONTROL culture, but are mostly absent from axons in CORT50 or DEX50 cultures. However, both CONTROL and DEX50 cultures contained membrane like MBP-immunoreactive expansions (lightning bolt-shaped arrows), which were very rare in CORT50 cultures. Calibration bar: 100 μm.

Fig. 3:

Graphs A-E: Quantification of the length of MBP-immunoreactive stretches aligned with pNF+ axons relative to the number of total cells (DAPI stained cell nuclei) (A) or relative to the total area fraction of pNF+ axons (B) (this latter quotient is referred as myelination index in the main text) in control, 5 μM CORT, 50 μM CORT, and 50 μM DEX treated cultures. The total area fraction of MBP immunoreactivity (that includes all positive cell bodies, processes, membrane-like expansions and myelinated portions of axons) relative to total cell numbers is presented in C, and the ratio of myelinated length of axons to the total area fraction of MBP in D. Panel E illustrates the area fraction (%) of pNF immunoreactivity for the same cultures as above. Since all ANOVAS (in panel A, F(3,68)= 9.315, p≤ 0.0001; B, F(3,68)= 7.894, p≤ 0.0001; C, F(3,68)= 5.469, p≤ 0.002; D, F(3,68)= 7.693, p≤ 0.0002), except that of pNF+ area fraction (panel E; F(3,68)=2.643, p=0.0562) showed a significant effect of group, stars (*) denote where the pairwise Dunnett’s tests for multiple comparisons to the control group showed a statistical difference at p≤0.01.

Fig. 4:

Exposure to glucocorticoids suppressed Cx43 expression in SC myelination cultures. Micrographs from microscopic fields of Control (top panels), 5 μM CORT (CORT5) (second row), 50 μM CORT (CORT50) (third row) and 50 μM DEX (DEX50) (bottom panels) at DIV28 with immunofluorescence labeling of MBP (green), Cx43 (red), and the merged image (MBP+Cx43+DAPI) illustrating the reduced amount of Cx43 and MBP immunolabeling by corticosteroid treatment. Calibration bar: 150 μm. Graphs A and B: Quantification of the area fraction of Cx43 immunoreactivity relative to the total number of cells (A) and average numbers of cells in the same samples (B). Since the ANOVA showed a significant effect of group for Cx43 area fraction (F(3,35)= 8.336 p≤ 0.0003), stars denote where the pairwise Dunnett’s tests (for multiple pairwise comparisons to the control group) revealed a statistically significant difference at p≤0.01.

Fig. 5:

Mifepristone significantly prevented CORT-induced myelination deficits in CC-derived cultures. Micrographs from microscope fields of glucocorticoid-untreated Control (top panels), 50 μM CORT-treated (middle panels) or 50 μM CORT plus 10 μM mifepristone-treated (bottom panels) in primary CC-derived cultures at DIV28 with nuclear DAPI staining (blue), and immunofluorescence labeling of MBP (green or yellow in left panel, green in the right panel) and pNF(red). Note that linear stretches of MBP immunostaining are readily identifiable (white arrows) along pNF+ immunolabeled axons in the Control and CORT50+mifepristone cultures but are mostly absent from axons in cultures treated only with CORT50 (middle panels). Scale bar: 150 μm.

Fig. 6:

Quantification of the myelination index in CC-derived cultures treated with glucocorticoids. The length of MBP-immunoreactive segments aligned with pNF-positive axons (myelin length) relative to the total area fraction of pNF+ axons (myelination index, in A) or relative to the total number of cells (DAPI stained cell nuclei, in B), and the total area fraction of MBP immunoreactivity relative to cell numbers (C), were quantified in Control, 5 μM CORT, 50 μM CORT, 50 μM DEX, 50 μM CORT+10 μM mifepristone, and 10 μM mifepristone alone treated CC cultures. No significant changes in total cell number (D) or area fraction of pNF+ immunoreactivity (E) were caused by any treatments as compared to controls. Since most ANOVAS showed a significant effect of group for all variables studied (panel A, F(4,43)= 8.4, p≤ 0.0001; B, F(5,52)= 8.963 p≤ 0.0001; C, F(5,52)= 30.15, p≤ 0.0001) pairwise Sidak’s tests for pairwise multiple planned comparisons to control or the CORT50-treated group were performed and revealed a statistically significant difference as compared to controls (stars, *), or significant differences between the CORT50 only treatment and the CORT50+mifp or the mifp only treatments (pound symbol, #) at p≤0.01. mifp = mifepristone

Fig. 7:

Glucocorticoid exposure resulted in a dramatic reduction of Cx43 immunoreactive puncta but not GFAP immunoreactivity in CC-derived myelination cultures. Micrographs from microscope fields of controls (left 3 panels), 50 μM CORT-treated (middle column of panels) or 50 μM CORT plus 10 μM mifepristone-treated (right 3 panels) in CC cultures at DIV28 immunofluorescence with labeling for GFAP (red, top row of panels), Cx43 (green, middle row of panels. The bottom 3 panels are the merged images of the overlying two panels plus the nuclear DAP I staining. Note that CORT50 treatment dramatically reduced Cx43 immunoreactive puncta, but co-incubation of CORT50 with mifepristone largely prevented that reduction (middle and bottom rows of panels). Calibration bar: 200 μm.

Fig. 8:

Bar graph summarizing the quantification of the area fraction of Cx43 immunoreactivity relative to the number of cells in CC-derived cultures untreated with glucocorticoids (Control) or treated with 5 μM CORT, 50 μM CORT, 50 μM DEX, 50 μM CORT+10 μM mifepristone (mifp), or 10 μM mifp only. ANOVA showed a significant effect of treatment group for Cx43 area fraction (F(5,56)= 8.01, p≤ 0.0001). Thus, stars denote where the pairwise Sidak’s tests showed a statistical difference as compared to the control cultures at p≤0.01, and the pound symbol (#) denotes significant differences between the CORT50-only treatment and the CORT50+mifepristone treatment in the same Sidak’s test, with significance at p≤0.01. mifp = mifepristone.

Fig. 9:

Effect of glucocorticoid exposure on MBP-immunoreactive oligodendrocyte morphology in mixed astrocytes-oligodendrocyte cultures from the cerebral cortex. Top and middle panels: Micrographs from microscope fields of mixed glial cultures exposed to vehicle (Control), 5 μM CORT (CORT5), 50 μM CORT (CORT50) and 50 μM DEX (DEX50) for 4 days. Immunofluorescence labeling of MBP (green) (top panels) and the merged image (MBP+DAPI) (middle panels) illustrate the shrunk appearance of MBP immunoreactive cells in corticosteroids-treated cultures as compared to controls. Calibration bar: 200 μm. Bottom graphs: Quantification of Sholl analysis parameters for MBP-immunoreactive cells in mixed glial cultures from the control (CONT), 5 μM CORT (C5), 50 μM CORT (C50) or 50 μM DEX (DEX50) treatments. The parameters determined were: Sum of intersections of MBP immunoreactive cell processes with the concentric circles of the templates (A), Ending radius of the longest processes (B), Ramification index (C) and Critical value (D). Since all ANOVAS but one demonstrated a significant effect of group for all four variables (panel A, F(3, 136)= 8.057, p≤ 0.0001; B, F(3, 136)= 2.493, p= 0.0628; C, F(3, 136)= 3.495, p≤0.02; D, F(3, 131 )=12.32 p≤ 0.0001), stars (*) denote where the pairwise Dunnett’s tests showed a statistical difference between a particular treatment and the untreated control cultures at p≤0.01.

Fig. 10:

Effect of glucocorticoid exposure on CNPase-immunoreactive oligodendrocyte morphology in mixed astrocytes-oligodendrocyte cultures from the cerebral cortex. Top and middle panels: Micrographs from microscope fields of control, 5 μM CORT (CORT5), 50 μM CORT (CORT50) and 50 μM DEX (DEX50)-treated mixed glial cultures immunostained with myelin proteins CNPase (green in the top panels) and the merged image (CNPase+DAPI) (middle panels) illustrate the shrunk appearance of CNPase immunoreactive oligodendrocytes. Calibration bar: 200 μm. Bottom panels: Quantification of Sholl analysis parameters for CNPase immunoreactive cells in CC-derived cultures control (CONT) or cultures treated with 5 μM CORT (C5), 50 μM CORT (C50) or 50 μM DEX (DEX50). The parameters determined were: Sum of intersections of CNPase immunoreactive cell processes with the concentric circles of the templates (A), Ending radius of the longest processes (B), Ramification index (C) and Critical value (D). Since all ANOVAS demonstrated a significant effect of group for all four variables (panel A, F(3, 293)= 12.84, p≤ 0.0001; B, F(3, 293)= 8.2, p< 0.0001; C, F(3, 293)= 3.137, p≤ 0.03; D, F(3, 293)= 17.39, p≤ 0.0001) stars (*) denote where the pairwise Dunnett’s tests showed a statistical difference between a particular treatment and the untreated control cultures at p≤0.01.