Hydrogen peroxide alters membrane and cytoskeleton properties and increases intercellular connections in astrocytes

Owing to their high consumption of oxygen and enrichment of polyunsaturated fatty acids in their membranes, cells in the central nervous system (CNS) are highly susceptible to oxidative damage. Oxidative stress induced by excess production of reactive oxygen species (ROS) has been implicated in pathologic processes associated with brain injuries and other neurodegenerative diseases (Coyle and Puttfarcken, 1993). H₂O₂, a major form of ROS, has been shown to mediate cell damage either through direct oxidation of lipids, proteins and DNA or acting as a signaling molecule to trigger cellular apoptotic pathways (Butterfield and Lauderback, 2002; Howe et al., 2004; Huang et al., 2004; Lin et al., 2004; Shin et al., 2004; van Rossum et al., 2004). Although neurons are generally more susceptible to oxidative injury than astrocytes, it is clear that oxidative stress also alters astrocyte functions (Robb and Connor, 1998). Astrocytes are known to provide vital functions necessary for proper neural performances, including maintenance of ion gradients, metabolism of neurotransmitters, and protecting neurons from oxidative injury (Desagher et al., 1996). Therefore, it is important to understand how oxidative stress alters astrocyte functions.

Lipid peroxidation is associated with modifications of membrane proteins (e.g. receptors, ion transporters and channels) leading to altered Ca²⁺ homeostasis, and Ca²⁺-dependent kinases (proteases, phospholipases and nucleases), leading to mitochondrial dysfunction and activation of apoptotic pathways (Butterfield and Lauderback, 2002; Gyulkhandanyan et al., 2003; Jacobson and Duchen, 2002; Robb et al., 1999). Despite the metabolic changes, little is known about the direct effects of H₂O₂ on membrane properties of astrocytes. In this study, the effects of H₂O₂ on changes in phase properties of astrocyte membranes were examined by fluorescence spectroscopy of 6-dodecanoyl-2-dimethylaminonaphthalene (Laurdan).

In addition to lipid peroxidation, the cytoskeleton network is also one of the earliest targets of oxidative stress (Dalle-Donne et al., 2001; Zhao and Davis, 1998). There is a growing body of evidence supporting the hypothesis that perturbation of cytoskeletal proteins is the initial step of oxidant-induced cell damage. Although oxidative injury is known to selectively alter cytoskeletal proteins (Aksenov et al., 2001), the mechanism by which oxidants change the structure and the spatial organization of actin filaments is still unclear (Dalle-Donne et al., 2001). It has been reported that H₂O₂ induces rapid formation of focal adhesion complexes and reorganization of actin network in endothelial cells (Huot et al., 1998). In endothelial cells, H₂O₂ also causes activation of the p38 MAPK and this is followed by both activation of the MAPK-activated protein kinase-2/3 and the phosphorylation of the small heat shock protein (HSP) (Huot et al., 1998; Pearl-Yafe et al., 2004). Oxidative stress has been shown to increase gap junction communication among astrocytes through reorganization of the actin network (Rouach et al., 2004).

In this paper, we report the effects of oxidative stress induced by H₂O₂ on plasma membrane properties, cytoskeleton arrangements and cell-cell interaction in rat astrocytes. Specifically, we quantitatively examined how H₂O₂ alters the phase properties of plasma membranes using the technique of fluorescence spectroscopy of Laurdan. We also studied the effects of H₂O₂ on actin polymerization and on the colocalization of myosin Va and F-actin by confocal immunofluorescence microscopy. Finally, we demonstrated that these oxidative changes involved stimulation of the p38 MAPK pathway, and the p38 inhibitor SB203580 was able to block many effects of H₂O₂.

Stock solutions of 1-stearoyl-2-oleoyl-sn-glycero-3-phosphocholine (SOPC), 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and cholesterol (Avanti Polar Lipids, Alabaster, AL, USA) were made in chloroform. Stock solution of Laurdan (Molecular Probes, Eugene, OR, USA) was made in methanol. FeSO₄ and H₂O₂ were from Sigma (St Louis, MO, USA). Dulbecco's modified Eagle's medium (DMEM), F12 medium, phosphate-buffered saline (PBS), Tris-buffered saline (TBS), trypsin-EDTA, streptomycin-penicillin and amphotericin B solutions were from Gibco (Gaithersburg, MD, USA). Fetal bovine serum (FBS) was from US Bio-Technologies (Parkerford, PA, USA). SB203580 was purchased from Promega (Madison, WI, USA).

Lipids in rat brains were extracted according to the procedure described by Zhang and Sun (Zhang and Sun, 1995). Briefly, brain tissue was homogenized in 10 ml of PBS and 40 ml of chloroform-methanol (2:1 v/v) was added. The mixture was centrifuged at 2000 g for 10 minutes. The lower organic phase containing the lipids was filtered through a Pasteur pipette column packed with glass wool and anhydrous Na₂SO₄. The lipid extract was collected and stored at –20°C until use.

For preparation of lipid vesicles, brain lipids (or other purchased phospholipids) were redissolved in chloroform and about 0.001 mol% of Laurdan was added. The preparation of vesicles was accomplished by film rehydration and followed by evaporation of the chloroform under vacuum for 3 hours (Lee et al., 2001). Final addition of degassed PBS with minimal agitation led to spontaneous budding of vesicles off the glass wall and into suspension.

Based on Fenton's reaction, H₂O₂ is converted to OH^– in the presence of Fe²⁺. A FeSO₄ stock solution (8 mM) and a 30% H₂O₂ stock solution were prepared in ice-cold water and kept at 0°C to make Fenton's reagent. Lipid peroxidation was initiated by addition of Fenton's reagent (200 μM Fe²⁺ in H₂O₂) to the solution containing lipid vesicles. The vesicles were then incubated in the dark at 37°C for 2 hours prior to fluorescence spectroscopy of Laurdan. To study the effects of p38 MAPK inhibitor, Laurdan-labeled vesicles were preincubated with the inhibitor for 30 minutes before the addition of Fenton's reagent.

Astrocyte cultures were obtained from newborn rats using a standard stratification/cell shaking procedure (McCarthy and de Vellis, 1980). This procedure yielded confluent mixed glial cultures within 7-9 days, after which the flasks were shaken at 200 r.p.m. at room temperature for 4 hours to remove microglial cells. These astrocytes (>95% as quantified by anti-glial fibrillary acidic protein labeling) were subsequently subcultured onto coverslips coated with poly-D-lysine (0.4 mg/ml), and fed every 48 hours with fresh DMEM culture medium and 10% FBS. Cells were maintained at 37°C in a 5% CO₂ humidified incubator.

After subculturing, astrocytes were grown to approximately 60% confluency. Before the addition of H₂O₂, astrocytes were incubated in a serum-free medium containing DMEM:F12 (1:1 v/v) for 2-3 hours. In the study of the role of the MAPK pathway, cells were pre-treated with different concentrations of SB203580 for 30 minutes prior to the addition of H₂O₂. After incubation of astrocytes with H₂O₂ for 2 hours at 37°C, cells were washed twice with PBS and incubated with DMEM containing 1% Laurdan for 15 minutes. Excess Laurdan was removed by washing cells three times with PBS.

Spectroscopic measurements were performed at 37°C on a temperature controlled FP-750 spectrofluorometer (Jasco, Japan). Lipid vesicles were diluted in PBS and incubated for 2 hours in the presence or absence of H₂O₂ before measurement. A coverslip with attached astrocytes was put into the cuvette in Krebs-Ringer Hepes buffer (pH 7.4), and was positioned at 45° with respect to both excitation and emission ports of the fluorometer. Calculations for the generalized polarization (GP) of Laurdan followed the definition (Parasassi et al., 1990): GP=(I_B–I_R)/(I_B+I_R), where I_B and I_R are the intensities at 440 nm and 490 nm, respectively, with fixed excitation wavelength of 350 nm.

Astrocytes were cultured in 60 mm dishes until 90% confluent for western blot analysis. After treatment with test agents (i.e. H₂O₂ and/or SB203580) for 30 minutes, protein lysates were prepared by adding 200 μl 1× sodium dodecyl sulfate (SDS) sample buffer (Cell Signaling Technology, Beverly, MA, USA), sonicated for 10-15 seconds to shear the DNA and reduce sample viscosity, and then boiled for 5 minutes. Protein concentration of each sample was assayed by the Bradford method. After adjusting for proteins (Bradford, 1976), aliquots were applied to a 10% SDS-polyacrylamide gel (PAGE; BioRad, Hercules, CA, USA) and blotted to nitrocellulose membrane. The sample was blocked with 5% nonfat milk, 0.1% Tween 20-PBS, pH 7.4, for 1 hour. The membrane was incubated at 4°C overnight with primary polyclonal antibodies (1:1000 dilution) against p38 MAPK and phospho-p38 MAPK (Cell Signaling Technology, Beverly, MA, USA) containing 1× TBS, 5% BSA and 0.1% Tween 20 with gentle shaking. Then the membrane was washed three times with 0.1% Tween 20-PBS for 5 minutes and incubated for 1 hour with diluted (1:1000) secondary horseradish-conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at room temperature. The membrane was then washed as described above and visualized on Iso-Max imaging film (SciMart, St Louis, MO, USA) using the Lumi GLO chemiluminescence reagent.

Cells were fixed using 4% paraformaldehyde solution and permeabilized by 0.1% Triton X-100 in PBS before staining. 5% normal goat serum (NGS) in PBS was applied to the cells for 30 minutes to block non-specific binding. F-actin was fluorescently labeled with Oregon Green-phalloidin (250 nM) (Sigma, St Louis, MO, USA). To label myosin in the cells, rabbit polyclonal anti-myosin Va was added at the final concentration of 0.8 μg/ml (Chemicon, Temecula, CA, USA). This was followed by labeling with secondary antibody, Alexa Fluor^®-594 donkey anti-rabbit IgG (Molecular Probes, Eugene, OR, USA) at a final concentration of 2 μg/ml.

Bright-field illumination and fluorescence microscopy were performed with a Nikon TE-2000 U fluorescence microscope and a 40×, NA 0.95 objective. Image were acquired using a cooled CCD camera controlled with a computer that ran a MetaVue imaging software (Universal Imaging, PA, USA). Fluorescence excitation source was controlled with a Uni-Blitz mechanical shutter.

To analyze the colocalization between myosin and F-actin, high-resolution immunofluorescence microscopy was performed with a BioRad Radiance 2000 (Carl Zeiss Microimaging, NY, USA) confocal system coupled to an Olympus IX70 (Tokyo, Japan) inverted microscope. Confocal images were acquired with a 60× NA 1.2 water immersion objective. Background was subtracted for all images prior to analysis.

Actin polymerization was quantified by integrating the intensity of Oregon Green-phalloidin-labeled F-actin over the cell body. The integrated intensity was then normalized by the integrated intensity of the labeled F-actin in control cells (i.e. without H₂O₂ treatment). In this manner, the normalized intensity of cells greater than one is used to indicate that actin polymerization is enhanced compared to the control.

Relative percentages of astrocytic connections in cultures were measured by the ratio of the number of cells with TNT-like connections to the total number of cells in the image field.

The colocalization of F-actin and myosin was quantified by normalizing the area of coincident intensity to the area of noncoincident intensities.

Data are presented as mean ± s.d. from at least three independent experiments. Comparisons between groups were made with one-way ANOVA followed by Bonferroni posttests. Values of P<0.05 are considered statistically significant.

We applied the fluorescence spectroscopy of Laurdan integrated into plasma membranes of primary astrocytes and membranes of artificial lipid vesicles to study the possible changes in membrane phase properties resulting from oxidative stress produced by H₂O₂ in the presence of Fe²⁺ (Fenton's reaction). Consistent with Laurdan being an extremely hydrophobic molecule, Fig. 1A shows that this molecule preferentially partitioned into the bilayer membranes of vesicles made from rat brain lipid extract. Data in Fig. 1B show that Laurdan integrates into the plasma membranes of astrocytes. To further prove that Laurdan only partitions into plasma membranes of cells, spherical erythroblasts and neuroblastoma SH-SY5Y cells were stained with Laurdan. In fluorescence images edge-brightness of these Laurdan-labeled cells shows that Laurdan preferentially integrates into plasma membranes (Fig. S1 in supplementary material). Owing to the hydrophobic properties of Laurdan, it is reasonable that once it is integrated into the plasma membranes of astrocytes, it would be unlikely to diffuse further into the intracellular organelles. Therefore, the spectral changes of Laurdan integrated in the plasma membranes can be used as an indicator of the changes in membrane phase properties.

Oxidative stress has been reported to stimulate signaling pathways including the mitogen-activated protein kinases (MAPK) (Choi et al., 2004; Huot et al., 1998; Kurata, 2000; Rosenberger et al., 2001; Song et al., 2002; Usatyuk et al., 2003). In this study, western blot analysis was used to demonstrate that H₂O₂ triggers the p38 MAPK pathway in astrocytes and phosphorylation of p38 MAPK increased with increasing dose of H₂O₂ (Fig. 2A). SB203580 (at 20-30 μM), a specific inhibitor for the p38 MAPK, was capable of suppressing the phosphorylation of p38 MAPK induced by H₂O₂ (Fig. 2B). Therefore, 20 μM SB203580 was found to be the optimal concentration, and was used to study whether H₂O₂ induced MAPK pathway mediates the changes in morphology, membrane phase properties and cytoskeletal organization in astrocytes.

We applied the fluorescence spectroscopy of Laurdan integrated into plasma membranes of primary astrocytes and membranes of artificial lipid vesicles to study the possible changes in membrane phase properties resulting from oxidative stress produced by H₂O₂ in the presence of Fe²⁺ (Fenton's reaction). Laurdan possesses both an electron donor and an electron receptor, so that fluorescent excitation induces a large excited-state dipole. This strong dipole tends to locally align the surrounding molecules (e.g. water), which dissipates a small fraction of the excited state energy and produces a red shift of the emission spectrum. A membrane with a lower interfacial tension (i.e. liquid crystalline phase) allows more water molecules to partition into the membrane core, thus shifting Laurdan's emission maximum further into the red (Lee et al., 2001; Parasassi et al., 1990). To quantify this shift, Gratton and co-workers (Parasassi et al., 1990) have defined the generalized polarization (GP) and applied it to phase transitions of different lipid systems (Parasassi et al., 1991; Parasassi et al., 1993; Parasassi et al., 1994) as well as natural membranes (Parasassi et al., 1992; Waschuk et al., 2001). A higher GP value indicates that the membrane becomes more gel-like, and a lower GP indicates that the membrane becomes more liquid crystalline-like.

In our data analysis, we plotted GP-GP_o, where GP_o is the GP of control experiments (i.e. concentration of H₂O₂ is zero). Therefore, GP-GP_o of the control is always zero, serving as a common reference datum. We found that the GP-GP_o of Laurdan in astrocyte membranes was increased by increasing dose of H₂O₂ (Fig. 3), suggesting that H₂O₂ causes astrocyte membranes to become more gel-like. These effects of H₂O₂ on astrocyte membranes could be suppressed by the MAPK inhibitor, SB203580 (Fig. 3).

In order to rule out the cellular complexity in astrocytes, and to reveal the effects of ROS induced by H₂O₂ on bilayer membranes through direct oxidation, we also measured the GP-GP_o values for membrane vesicles made from (a) rat brain lipid extract, (b) DMPC, a symmetric saturated phospholipid, and (c) SOPC, an asymmetric unsaturated phospholipid, which were incubated with Fenton's reagent. Contrary to the results from astrocyte membranes, the GP-GP_o values of Laurdan in membranes of brain lipid vesicles and SOPC decreased (i.e. becoming more liquid crystalline-like) with increasing doses of H₂O₂, whereas that in membranes of DMPC remained almost zero or unchanged (Fig. 4A). Such decreases in GP-GP_o values of Laurdan in membranes of brain lipid vesicles and SOPC were observed only in the presence of Fe²⁺ (data not shown). These results suggest that ROS generated by H₂O₂ reacts with unsaturated acyl chains, and produces more asymmetric phospholipids, thereby loosening the molecular packing in the hydrophobic core of the lipid bilayer. Subsequently, membranes made of brain lipids and SOPC became more liquid crystalline-like with increasing H₂O₂.

In addition to unsaturated phospholipids, cholesterol is another major lipid component in astrocyte membranes. However, we did not observe significant effects of H₂O₂ on the phase properties of membranes made of cholesterol/DMPC (1:1) (Fig. 4B). Therefore, GP-GP_o values of cholesterol/SOPC membranes were smaller than those of SOPC membranes, which was due to different compositions of the two membrane systems and not to the oxidation of cholesterol (Fig. 4B). SB203580 did not reduce membrane phase change induced by H₂O₂ in vesicle bilayer membranes (Fig. 4C). Since the reactions of ROS with unsaturated phospholipids cause the lipid bilayer membranes to become more liquid crystalline-like, whereas the plasma membranes of astrocytes become more gel-like, this would suggest that changes in the phase properties of plasma membranes of astrocytes during oxidative insults are not simply due to direct oxidation of phospholipids and cholesterol.

Besides the effects on plasma membranes, oxidative stress is known to cause cytoskeletal reorganizations in cells (Qian et al., 2003; Rosado et al., 2002; Rouach et al., 2004; Zhao and Davis, 1998). Fig. 5E,G shows the effects of oxidative stress on the formation of actin-enriched protrusions (white arrowheads), similar to cytonemes described by Ramirez-Weber and Kornberg (Ramirez-Weber and Kornberg, 1999), and the formation of tunneling nanotube (TNT)-like connections (white arrows) similar to the TNT described by Rustom et al. (Rustom et al., 2004). These structures are F-actin-enriched (Fig. 5E), and contain myosin (Fig. 7J). TNT-like connections and cytonemes were observed only on solid surfaces supporting cell cultures of astrocytes. After treatment with H₂O₂, astrocytes tend to establish connections with each other, resulting in a significant increase (P<0.01) in the number of actin filament-enriched TNT-like structures (Fig. 5H). Such reorganization of F-actin in astrocytes after H₂O₂ treatment is dramatic when compared with non-treated astrocytes, in which the actin stress fibers spread out evenly, forming a few cell motility structures (e.g. rosette-like dots) at the cell leading edges and within the cell bodies.

Enhanced actin polymerization may be a part of the process in the formations of cytonemes and TNT-like connections in astrocytes under oxidative stress. Therefore, we quantified the fluorescence intensity of Oregon Green-phalloidin-labeled F-actin in astrocytes. The intensity of fluorescence in astrocytes treated with H₂O₂ was significantly increased (P<0.01) as compared to that without H₂O₂ treatment (Fig. 6).

In addition, SB203580 was found to suppress the enhanced formation of actin-enriched structures, cytonemes and nanotubes (Fig. 5C,F,H), and the polymerization of actin induced by H₂O₂ (Fig. 6).

Myosin is an actin-associated motor protein, and is known to play an important role in cellular transportation and communication. Since H₂O₂ enhances actin polymerization and induces F-actin reorganization, it may also alter the colocalization of myosin to F-actin. We quantified the effect of H₂O₂ on the colocalization of myosin with F-actin by analyzing confocal images of fluorescently labeled actin and myosin in astrocytes. Fig. 7 shows images of fluorescently labeled actin, myosin and the colocalization images used for quantitative analysis. Our results indicated a significant increase in colocalization of myosin Va with F-actin after H₂O₂ treatment in astrocytes, while pre-treatment with SB203580 could partially reduce this enhanced colocalization induced by H₂O₂ (Fig. 7K). Myosin Va was also observed inside the TNT-like connections (Fig. 7J).

In this study, we examined the effects of oxidative stress on astrocyte membranes and the cytoskeleton by exposing astrocytes to H₂O₂, an oxidant commonly used to examine oxidative mechanisms mediating cell death (Misonou et al., 2000). A study by Hyslop et al. (Hyslop et al., 1995) reported that under ischemic conditions, the concentration of H₂O₂ in the rat striatum can possibly reach 200 μM. In our study, we observed astrocyte membrane phase properties and the phosphorylation of p38 MAPK at concentrations of H₂O₂ ranging from 0 to 400 μM and we chose to use 200 μM to study morphological changes and for observation and quantification of changes in actin and myosin using immunofluorescence.

Since Fe²⁺ is present in the cell culture medium and inside astrocytes, addition of H₂O₂ alone to the cell culture is sufficient to initiate Fenton's reaction. Results of this study demonstrate that oxidative stress induced by H₂O₂ decreases astrocyte membrane fluidity, induces cytoskeletal reorganization, and increase the formation of cytonemes and TNT-like connections. In addition, the p38 MAPK inhibitor, SB203580, could suppress these effects produced by H₂O₂. As oxidative stress is involved in a number of neurodegenerative diseases, this study, aimed at understanding the effects of H₂O₂ on astrocytes, should prove relevant and significant.

In order to investigate the origins of the effects of H₂O₂ on astrocyte membranes, we systematically characterized phase properties of astrocyte membranes and compared them to vesicles made from (a) brain lipid extract, (b) unsaturated phospholipids (SOPC), (c) symmetric saturated phospholipids (DMPC), (d) cholesterol/SOPC (1:1) and (e) cholesterol/DMPC (1:1). Total brain lipid extract is composed of acidic and neutral phospholipids, gangliosides, cholesterol, sphingolipids and isoprenoids (McLaurin et al., 2002). Thus, vesicles derived from brain lipids serve as a simple model system to mimic astrocyte membranes without any complications produced by cellular processes. Our results show that increasing doses of H₂O₂ cause membranes of brain lipid vesicles to become more liquid crystalline-like (Fig. 4) but membranes of astrocytes to become more gel-like (Fig. 3). Since the results from brain lipid vesicles represent the direct effects of lipid peroxidation, the contrasting results between membranes of brain lipid vesicles and plasma membranes of astrocytes suggest that the effects of H₂O₂ on astrocyte membranes are not simply due to lipid peroxidation, but also to alterations of cell membrane proteins and organization of intracellular structures. Clearly, more studies are needed to explore how oxidative stress alters cell membrane structures through intracellular signaling pathways.

We also examined the effects of oxidative stress on the F-actin fibers in astrocytes. Actin filaments are involved in a wide variety of cellular processes, including cell motility, cell cycle control, cellular structure and cell signaling. They function in cellular processes by undergoing dynamic structural reorganization or remodeling, leading to the formation of discrete structures at the periphery for attachment to the substratum in response to different signals. In this study, we found that H₂O₂ increases polymerization of actin and the formation of cellular protrusions (cytonemes), and TNT-like connections in primary rat astrocytes. These results are consistent with studies in other cell types indicating a rapid remodeling of the structure of actin filaments upon oxidative stress (Qian et al., 2003; Zhao and Davis, 1998). However, there is also evidence that oxidation of actin by strong oxidant compounds may disrupt the actin network in cells (Dalle-Donne et al., 2002; Valen et al., 1999), leading to inhibition of actin polymerization and actin filament depolymerization. Our data indicate that oxidative stress significantly enhances the formation of intercellular TNT-like connections and the association of myosin to actin filaments. Rustom and his coworkers (Rustom et al., 2004) regarded TNT as an independent form of mammalian cell-cell communication. These TNTs are F-actin-enriched structures, and can transport organelles from one cell to another. It is a highly coordinated process regulated by cell signaling cascades. In astrocytes, oxidative stress is known to cause the formation of gap junction, which is another form of intercellular connection (Rouach et al., 2004). With the notion of TNTs functioning as transportation highways (Rustom et al., 2004), our results lead to the hypothesis that oxidative stress enhances communication between astrocytes. Therefore, our results lay the groundwork to study the possible communication among astrocytes altered by oxidative stress.

Our findings suggest that changes in actin polymerization are not caused by direct oxidation of actin by H₂O₂, but may involve signaling pathways. Indeed, oxidative stress is known to induce signaling cascades and subsequent cytoskeletal remodeling in a number of neural and non-neural cell systems (Bundy et al., 2005; Carvalho et al., 2004; Choi et al., 2004; Huot et al., 1998; Kevil et al., 2001; Kurata, 2000; Usatyuk and Natarajan, 2004; Usatyuk et al., 2003). In dopaminergic neurons, phosphorylation of p38 MAPK by oxidative stress is linked to activation of caspases and apoptotic pathways (Choi et al., 2004) and p38 MAPK mediates the formation of actin stress fibers induced by β-amyloid peptide (Song et al., 2002). Our study, with primary rat astrocytes, demonstrated a dose-dependent increase in phosphorylation of p38 MAPK upon treatment with H₂O₂, and inhibition of p38 MAPK reduced H₂O₂-mediated changes in membrane phase properties and cytoskeletal reorganization. These results suggest that oxidative stress-induced activation of MAPK is important in modulation of cytoskeletal organization as well as downstream pathways leading to the cell death mechanism.

In conclusion, results of this study demonstrate that H₂O₂ causes astrocyte membranes to become more gel-like and induces actin polymerization and, subsequently, enhances formation of cytonemes and cell-to-cell TNT-like connections. Based on these results, we suggest that oxidative stress may have an important impact on astrocyte membranes, signaling pathways and cytoskeletal arrangements. All these effects may lead to altered astrocyte functions.

This work was supported by grants DHHS 5P01 ES10535 and 1P01 AG18357 from NIH and an MU Research Board grant, URB-04-038.