Characterization of an improved procedure for the removal of microglia from confluent monolayers of primary astrocytes

Journal of Neuroscience Methods 150 (2006) 128–137

Characterization of an improved procedure for the removal of microglia from confluent monolayers of primary astrocytes Mary E. Hamby, Tracy F. Uliasz, Sandra J. Hewett ∗∗ , James A. Hewett ∗ Department of Neuroscience MC 3401, University of Connecticut Health Center, 263 Farmington Avenue, CT 06030-3401, USA Received 21 October 2004; received in revised form 15 June 2005; accepted 15 June 2005

Abstract Cultures of astrocytes can be readily established and are widely used to study the biological functions of these glial cells in isolation. Unfortunately, contamination by microglia can confound results from such studies. Herein, a simple and highly effective modification of a common procedure to remove microglia from astrocyte cultures is described. After becoming confluent, astrocytes were exposed to a mitotic inhibitor for 5–6 days then treated with 50–75 mM l-leucine methyl ester (LME) for 60–90 min. Unlike previous protocols that employed lower LME concentrations on subconfluent cultures or during passage of astrocytes, this protocol effectively depleted microglia from highdensity astrocyte monolayers. This was evidenced by the selective depletion of microglial-specific markers. Purified monolayers appeared morphologically normal 24 h after LME treatment and expressed nitric oxide synthase-2 (NOS-2) and cyclooxygenase-2 (COX-2) proteins upon stimulation with LPS plus IFN␥, albeit to a lower level than unpurified monolayers. This difference could be attributed to removal of contaminating microglia from monolayers and not to astrocyte dysfunction, since LME treatment did not alter global protein synthesis and a reactive phenotype could be induced in the purified monolayers. Thus, this modified protocol selectively depletes microglia from high-density primary astrocyte monolayers without compromising their functional integrity. © 2005 Elsevier B.V. All rights reserved. Keywords: Cell culture; Astrocytes; Microglia; l-Leucine methyl ester; Inducible nitric oxide synthase (iNOS; NOS-2); Nitric oxide (NO); Cyclooxygenase-2 (COX-2); Reactive astrocytosis; Lipopolysaccharide (LPS)

1. Introduction Astrocytes are the predominant cell type in the central nervous system (CNS). They function to maintain normal brain physiology, including survival and guidance of migrating neurons during development, formation and preservation of the blood–brain barrier, and maintenance of neuronal homeostasis and plasticity (reviewed in Araque et al., 2001; Doetsch, 2003; Fields and Stevens-Graham, 2002; Janzer, 1993; Montgomery, 1994; Parri and Crunelli, 2003). In addition, astrocytes become reactive following injury to the CNS and serve important modulatory roles during CNS inflammation (Dong and Benveniste, 2001; Malhotra et al., 1990; Smith and Lassmann, 2002). In this regard, the expression of a ∗

Corresponding author. Tel.: +1 860 679 4131; fax: +1 860 679 8766. Corresponding author. Tel.: +1 860 679 2871; fax: +1 860 679 8766. E-mail addresses: [email protected] (S.J. Hewett), [email protected] (J.A. Hewett). ∗∗

0165-0270/$ – see front matter © 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2005.06.016

number of genes can be upregulated in astrocytes in response to proinflammatory stimuli (John et al., 2005). Among these are nitric oxide synthase-2 (NOS-2) and cyclooxygenase-2 (COX-2), which have been implicated in the pathogenesis of CNS inflammation in several instances (Murphy, 2000; O’Banion, 1999). The elucidation of astrocyte function has benefited from the ability to study these glial cells under defined conditions in vitro. Although astrocyte cultures are relatively easy to establish, these cultures can be contaminated by microglia, which have been shown to modify astrocyte responses under certain circumstances (Brown et al., 1996; Ciccarelli et al., 2000; Xiong et al., 1999). Furthermore, because proinflammatory stimuli can affect the expression of similar genes in microglia and astrocytes, interpretation of results using astrocyte cultures can be confounded by the presence of contaminating microglia. This could be especially problematic when sensitive biochemical measures are employed. Control of microglial growth is particularly difficult in astrocyte

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

cultures because astrocytes are a primary source of colony stimulating factor-1 (Hao et al., 1990; Thery et al., 1992), which is a potent and selective growth factor for microglia (Giulian and Ingeman, 1988). Thus, microglia will proliferate extensively in cultures of astrocytes if not controlled. Anti-mitotic agents have been employed to inhibit microglial growth in primary astrocyte cultures (Hewett, 1999; Solenov et al., 2004; Swanson et al., 1997; Tedeschi et al., 1986; Uliasz and Hewett, 2000). However, such an approach can be applied only on high-density monolayers after the astrocytes have entered a non-proliferative state induced by cell–cell contact. This allows a period of microglial cell growth prior to mitotic inhibition; thus, such cultures can still have a considerable number of microglia present. l-Leucine methyl ester (LME), a lysosomotropic agent originally used to selectively destroy macrophages (Thiele et al., 1983), has also been employed to deplete microglia from neural cultures including astrocytes (Giulian et al., 1993; Guillemin et al., 1997) and oligodendrocytes (Hewett et al., 1999). Typically, such protocols have employed 1–10 mM LME for this purpose (Bowman et al., 2003; Simmons and Murphy, 1992). However, this approach is most effective when performed on cells at low density or during cell passage. Even then, astrocyte cultures cannot be considered to be free from microglia (Giulian and Baker, 1986). In this report, we describe a new method for eradicating microglia from high-density primary astrocytes. In this protocol, treatment of confluent astrocyte monolayers with a mitotic inhibitor followed by a brief exposure to high concentrations of LME (50–75 mM) generates highly purified astrocytes without the need for cell passage.

2. Materials and methods 2.1. Primary cultures Primary astrocyte cultures were derived from 1 to 3 day postnatal CD1 mice (Charles River Laboratories, Wilmington, MA) as described previously (Trackey et al., 2001). Following an aseptic dissection, cerebral cortices were dissociated and cells plated at a density of 1–1.5 hemispheres/24-well plate (Falcon Primaria, BD Biosciences, Lincoln Park, NJ) or 1.2–1.6 hemispheres/6-well plate (Falcon Primaria, BD Biosciences, Lincoln Park, NJ). Plating medium consisted of media stock (MS) containing 10% bovine growth serum (Hyclone, Logan, UT), 10% iron-supplemented calf serum (CS; Hyclone, Logan, UT), 10 ng/ml epidermal growth factor (Invitrogen, Carlsbad, CA), 2 mM l-glutamine (Mediatech; Herndon, VA), 50 IU/ml penicillin and 50 ␮g/ml streptomycin (Mediatech; Herndon, VA). MS was comprised of modified Eagle’s medium (Earle’s salt; Mediatech, Herndon, VA) supplemented with glucose (J.T. Baker, Phillipsburg, NJ) and sodium bicarbonate (FisherChemicals, Fair Lawn, NJ) to a final concentration

129

of 25.7 and 28.2 mM, respectively. Unless stated otherwise, upon reaching confluence, astrocyte monolayers were treated with 8 ␮M cytosine ␤-d-arabinofuranoside (Ara-C; Sigma–Aldrich, St. Louis, MO) for 5–6 days to eliminate dividing cells (i.e., microglia). This treatment is not cytotoxic to quiescent contact-inhibited astrocytes. Cultures were subsequently maintained in growth medium consisting of MS containing 10% CS, 2 mM l-glutamine, 50 IU/ml penicillin and 50 ␮g/ml streptomycin. Such cultures are termed primary astrocytes because they are never subjected to cell passage. Cells were grown, maintained and stimulated at 37 ◦ C in a humidified atmosphere of 6% CO2 . All studies were performed on monolayers between 14 and 31 days in vitro (DIV). 2.2. LME treatment LME (Sigma–Aldrich, St. Louis, MO) was dissolved in MS and added to cultures at concentrations of either 5 mM for 8 h or 25–75 mM for 60–90 min. All LME solutions were adjusted to pH ∼7.4 and filter sterilized prior to addition to cells. Between 45 and 90 min after addition of high concentration LME, cultures were visually inspected to ensure maximal microglial lysis with minimal astrocytic toxicity. Thereafter, monolayers were washed thoroughly with growth medium and allowed to recover for 1 day in growth medium prior to experimentation. 2.3. Monolayer stimulation To induce NOS-2 protein expression, monolayers were exposed to 2 ␮g/ml lipopolysaccharide (LPS; DIFCO, Kansas City, MO) plus 3 ng/ml recombinant mouse interferon-␥ (IFN␥; R&D Systems, Minneapolis, MN) for 24–35 h. Both LPS and IFN␥ are required to induce NOS2 in mouse primary astrocyte cultures (Hewett et al., 1993; Vidwans et al., 2001). Stimulation medium contained DMEM (Gibco BRL, Gaithersburg MD) supplemented with 5% CS, 2 mM l-glutamine, 50 IU/ml penicillin and 50 ␮g/ml streptomycin. NOS-2 protein was assessed by immunoblot or immunofluorescence analysis. Nitric oxide (NO) was assessed by nitrite accumulation in cell supernatants. To induce COX-2 protein expression, astrocyte monolayers were treated with 0.5 mM dibutyryladenosine 3 ,5 -cyclic monophosphate (dbcAMP, Sigma–Aldrich, St. Louis, MO) in MS supplemented with 2 mM l-glutamine. COX-2 protein expression was assessed 6 h later by immunoblot analysis. Monolayers were exposed to 5 mM dbcAMP in stimulation medium to induce reactive astrocytosis. The reactive phenotype was assessed 48 h later by expression of glial fibrillary acidic protein (GFAP) depicting a transition from a cobblestone to elongated morphology. 2.4. Measurement of nitrite accumulation Production of nitric oxide (NO) was assessed indirectly by measurement of nitrite, an oxidative breakdown prod-

130

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

uct of NO (Green et al., 1982). Cell culture supernatants were collected (100 ␮l) and mixed with an equal volume of Griess Reagent consisting of a 1:1 mixture of 1% (w/v) sulfanilamide in 60% (v/v) acetic acid:0.1% (w/v) napthylenediamine dihydrochloride in H2 O. Nitrite accumulation was measured spectrophotometrically at 550 nm in a microtiter plate reader (Thermolabs; Chantilly, VA). 2.5. Immunofluorescence analysis Cell-type specific protein expression was assessed by indirect immunofluorescence. Astrocytes were detected by expression of GFAP. Monolayers were fixed for 5 min with 1:1 solution of freshly prepared acetone and methanol

then incubated for 7 min with PBS containing 0.25% Triton X-100. After blocking non-specific binding sites with PBS containing 10% normal goat serum (30 min), 1 ␮g/ml rat anti-bovine GFAP monoclonal antibody (Zymed, South San Francisco, CA) was added in PBS containing 2% normal goat serum (Jackson ImmunoResearch, West Grove, PA) and monolayers incubated overnight at 4 ◦ C. Unbound antibody was washed out with PBS and monolayers incubated in the dark (1 h) with PBS containing 15 ␮g/ml Cy3-conjugated goat anti-rat IgG antibody (Jackson ImmunoResearch, West Grove, PA). Microglia were detected by expression of CD11b as above except that monolayers were incubated with 10 ␮g/ml rat anti-mouse CD11b monoclonal antibody (Pharmingen, San Jose, CA). Stained monolayers were

Fig. 1. Treatment with 5 mM LME (8 h) does not eradicate microglia from confluent astrocyte monolayers. Untreated monolayers (a and b), monolayers treated with either Ara-C alone (c and d) or Ara-C followed by an 8 h exposure to 5 mM LME (e and f) were processed for CD11b immunolabeling 48–60 h after LME. Representative phase contrast images (a, c, e) and corresponding CD11b immunolabeled images (b, d, f) are shown for each treatment group (20× magnification).

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

washed with PBS and fluorescent images acquired using a CRX digital camera (Digital Video Camera Co., Austin, TX) mounted on an Olympus IX50 inverted microscope outfitted with epifluorescence. Data was processed using Adobe Photoshop software. To examine NOS-2 protein expression specifically in astrocytes, monolayers were double labeled with rabbit polyclonal anti-mouse NOS-2 antibody (2 ␮g/ml, Upstate Cell Signaling Solutions, Lake Placid, New York) and rat antibovine GFAP monoclonal antibody (1 ␮g/ml, Zymed, So San Francisco, CA). Primary antibodies were visualized with Rhodamine-Red-conjugated goat anti-rabbit IgG antibody (14 ␮g/ml; Jackson ImmunoResearch, West Grove, PA) and FITC-conjugated goat anti-rat IgG antibody (15 ␮g/ml; Jackson ImmunoResearch). 2.6. Reverse transcriptase-polymerase chain reaction (RT-PCR) Microglia but not astrocytes have been reported to express the mRNA for the CSF-1 receptor, c-fms (Hao et al., 1990; Krady et al., 2002). Thus, expression of the c-fms transcript was assessed as a sensitive indicator for microglial contamination. Total RNA was isolated from monolayers using TRIzol Reagent (Invitrogen) and resuspended in RNAasefree H2 O (Cellgro, Mediatech, Hernandon, VA). RNA was quantified spectrophotometrically at 260 nm (GeneQuant) and 0.5 ␮g reverse transcribed using reverse transcriptase and oligo (dT)(12–18) as previously described (Hewett et al., 1999). cDNAs for c-fms and ␤-actin were amplified in a thermal cycler (Bio-Rad, Hercules, CA) using amplimers for mouse c-fms (5 -CTGAGTCAGAAGCCCTTCGACAAAG3 and 5 -CTTTGCCCAGACCAAAGGCTGTAGC-3 ) and ␤-actin (5 -GTGGGCCGCTCTAGGCACCAA-3 and 5 CTCTTTGATGTCACGCACGATTTC-3 ). c-fms was amplified for 27 cycles (94 ◦ C/30 s, 63 ◦ C/45 s, 72 ◦ C/60 s) and ␤-actin was amplified for 23 cycles (94 ◦ C/30 s, 65 ◦ C/45 s, 72 ◦ C/60 s). PCR products (423 and 540 bp, respectively) were separated in a 2% agarose gel containing ethidium bromide (0.5 ␮g/ml) and visualized with a UV transilluminator (UVP, Kodak, Rochester, NY). Ethidium bromide fluorescence was imaged using the Kodak Electrophoresis Documentation and Analysis System 120 and images processed using Adobe Photoshop.

131

followed by 95% ethanol (400 ␮l; 2 × 5 min), and then solubilized using 0.1 M NaOH (250–300 ␮l/well). Total protein was determined via the BCA assay (Pierce Chemical, Rockville, IL). Radiolabel incorporation was determined in 150–200 ␮l of sample using a Packard Tricard 4000 scintillation counter and normalized to total TCA-precipitable protein. 2.8. Immunoblot analysis Cultures in 6 well plates were washed twice with 2 ml of ice-cold PBS and incubated in PBS containing EDTA (5 mM) and EGTA (5 mM) for ∼20 min on ice. Cells were harvested by gentle scraping, pelleted (1000 × g; 5 min, 4 ◦ C) and lysed in 50 ␮l of a modified RIPA buffer [50 mM Tris, pH 7.5, 0.5% deoxycholic acid, 0.1% SDS, 1% NP40, 150 mM NaCl, 5 mM iodoacetamide and protease inhibitor cocktail (Roche Diagnostics, Indianapolis, IN)]. Lysates were incubated on ice for 30 min, cellular debris pelleted (10,000 × g; 5 min, 4 ◦ C) and supernatants stored at −20 ◦ C. After determining the protein concentration of each lysate (BCA, Pierce Chemical Co., Rockville, IL), 15 (NOS-2) or 25 (COX-2) ␮g protein was separated by 8% SDS-PAGE and electrophoretically transferred to nitrocellulose (0.2 ␮m; Bio-Rad Laboratories, Hercules, CA). Membranes were washed twice

2.7. Total protein synthesis Total protein synthesis was assessed by incorportation of 3 H-l-leucine into trichloroacetic acid (TCA, Sigma, St. Louis, MO)-precipitated protein. Growth medium was exchanged with medium devoid of l-leucine and 3 H-lleucine (5 ␮Ci/ml, Perkin-Elmer, Boston, MA) was added 45 min later. Monolayers were thoroughly washed with PBS (3 × 400 ␮l) at various times thereafter (0–16 h) and proteins precipitated for 30 min in 400 ␮l 10% cold TCA. The precipitate was washed with 10% cold TCA (400 ␮l; 3 × 5 min)

Fig. 2. Treatment with 5 mM LME does not reduce NOS-2 protein expression or NO production. All monolayers were treated with Ara-C without additional LME treatment (−LME) or with 5 mM LME for 8 h (+LME). NOS-2 protein expression (A) and NO production (B) were assessed 24 h after LPS (2 ␮g/ml) plus IFN␥ (3 ng/ml) treatment. Basal = medium alone (unstimulated). (A) NOS-2 protein expression was assessed by immunoblot analysis. ␤-Actin protein expression was assessed on the same blot. (B) NO production was quantified by accumulation of nitrite and expressed as mean ± S.E.M.; n = 24 culture wells taken from six separate dissections. There were no statistically significant between group differences as determined by two-way ANOVA.

132

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

with H2 O (20 ml; 5 min) and proteins detected using speciesspecific WesternBreeze Immunodetection Kits (Invitrogen) per the manufacturer’s instructions. Antibodies used to detect NOS-2, COX-2 and ␤-actin were rabbit anti-NOS-2 polyclonal (1 ␮g/ml, Upstate Cell Signaling Solutions), rabbit anti-COX-2 polyclonal (125 ng/ml; Cayman Chemical, Ann Arbor, MI) and mouse anti-␤-actin monoclonal (290 ng/ml, Sigma, St. Louis, MO), respectively. Results were recorded on X-ray film (Amersham, Arlington Heights, IL).

3. Results Monolayers of primary astrocytes grown in the absence of both Ara-C and LME treatment contain a substantial number of contaminating microglia, as evidenced by the high level of CD11b staining (Fig. 1a and b). CD11b-positive cells were considerably diminished, but not eradicated, after incubation for 5–6 days with 8 ␮M Ara-C (Fig. 1c and d). The remain-

Fig. 4. Effect of LME treatment on c-fms expression. Ara-C treated astrocyte cultures were exposed to medium alone (Lane 1), 5 mM LME for 8 h (Lane 2) or 75 mM LME for 1 h (Lane 3). One day later, total RNA was isolated, first strand cDNA was synthesized and c-fms mRNA expression was assessed by PCR as described in methods. ␤-actin mRNA was assessed in all RNA samples to control for the amount of RNA in each sample. Results are representative of two separate experiments.

ing microglia were not affected by subsequent exposure of monolayers to 5 mM LME for 8 h (Fig. 1e and f). Because microglia can express NOS-2 in response to proinflammatory stimuli (Boje and Arora, 1992), we reasoned that NO production as measured by nitrite accumulation would be reduced in parallel with microglia depletion. Indeed, nitrite accumulation induced by LPS plus IFN␥ was significantly reduced by Ara-C treatment (56.93 ± 5.46 and 31.48 ± 4.38 ␮M, respectively). Additionally, in agreement with CD11b staining (Fig. 1), treatment with 5 mM LME

Fig. 3. NO production and microglial cell number are reduced in parallel following brief, high concentration LME exposure. Ara-C treated cultures were exposed to increasing concentrations of LME (0–75 mM) for 60–90 min. The following day, LPS (2 ␮g/ml) plus IFN␥ (3 ng/ml) was added to induce NOS2. (A) Thirty hours later, cell culture supernatants were collected and nitrite accumulation was determined as a measurement of NOS-2 catalytic activity. Results are expressed as mean ± S.E.M. (n = 9 cultures from three separate dissections). (*) Indicates values significantly different from non-LME treated cultures (0 mM) as determined by one-way ANOVA followed by a Dunnett’s t-test (P < 0.01). (B) Representative phase photomicrograph and corresponding CD11b immunolabeled image from a 75 mM LME-treated culture (60–90 min) assessed in A (20× magnification) NB: the small spherical structures evident in the phase contrast image represent the residual cellular debris that sometimes follows the LME treatment protocol.

Fig. 5. NOS-2 and COX-2 protein expression following treatment with 75 mM LME. Ara-C-treated cultures were either left alone (−) or treated with 75 mM LME for 60–90 min (+). The next day, cultures were stimulated with (A) LPS plus IFN␥ for ∼30 h or (B) 0.5 mM dbcAMP for 6 h. Total protein was isolated and 15 ␮g (A) or 25 ␮g (B) was separated via SDS-PAGE. Western blot analysis was performed using antibodies specific for NOS-2 (A), COX-2 (B) or ␤-actin (A and B). Results are representative of at least two blots.

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

for 8 h did not reduce NOS-2 protein expression or NO production compared to monolayers treated with Ara-C alone (Fig. 2). We next determined if depletion of microglia in Ara-Ctreated monolayers could be improved by exposure to higher concentrations of LME. Nitrite accumulation (Fig. 3A) and CD11b staining (not shown) were reduced in parallel in a concentration-dependent manner following treatment for 1–1.5 h with 25–75 mM LME. Treatment with 75 mM LME, a condition that markedly depleted CD11b staining (Fig. 3B), reduced nitrite accumulation to ∼20% of monolayers treated with Ara-C alone (Fig. 3A). In agreement with CD11b stain-

133

ing, c-fms mRNA was also depleted by treatment with 75 mM LME for 1 h, whereas an 8 h exposure to 5 mM LME minimally affected the expression level of this microglial marker compared to monolayers treated with Ara-C alone (Fig. 4). Thus, sequential treatment of confluent astrocyte monolayers with Ara-C and 75 mM LME successfully depleted microglia. Consistent with nitrite accumulation (Fig. 3A), treatment with 75 mM LME reduced but did not abolish NOS-2 protein expression induced by LPS plus IFN␥ (Fig. 5A). Moreover, the remaining NOS-2 protein colocalized with GFAP (Fig. 6d–f) indicating that the NO produced by the

Fig. 6. NOS-2 co-localizes with GFAP in monolayers treated with 75 mM LME. Ara-C treated cultures only (a–c) or those that additionally received LME (75 mM; 60–90 min) (d–f) were exposed to LPS (2 ␮g/ml) plus IFN␥ (3 ng/ml) for 24 h. Protein levels of NOS-2 and GFAP were examined by indirect immunofluorescence. (a and d) anti-GFAP labeled cells detected with FITC (green)-conjugated secondary antibody. (b and e) anti-NOS-2 labeled cells detected with CY3 (red)-conjugated secondary antibody. (c and f) merged photos depicting co-localization of the two antigens. (a–c) and (d–f) are from the same microscopic field taken at 80× magnification.

134

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

Fig. 7. Total protein synthesis is not altered by LME treatment. Parallel sister Ara-C-treated cultures were either left alone (−LME) or exposed to 75 mM LME (+LME). The next day, the rate of global protein synthesis was assessed by measuring the incorporation of 3 H-l-leucine into TCA-precipitated proteins over increasing periods of time. Results are expressed as the mean cpm × 103 /␮g protein ± S.E.M. (n = 6–8 wells from 2 to 3 separate dissections). There were no statistically significant between group differences as determined by two-way ANOVA.

enriched monolayers (Fig. 3A) was derived from NOS-2 protein expressed by astrocytes. COX-2 expression, another gene induced in both astrocytes and microglia (Hewett, 1999; Minghetti et al., 1999), was similarly reduced, but not eliminated, by pretreatment with 75 mM LME, providing further support for the notion that protein expression induced in enriched monolayers was astrocyte-derived (Fig. 5B). These results indicate that the purified astrocyte monolayers remained responsive to proinflammatory stimuli. Importantly, 75 mM LME did not alter total protein synthesis compared to monolayers treated with Ara-C alone (Fig. 7). Finally, the ability of dbcAMP to elicit a reactive phenotype was retained after the treatment paradigm with Ara-C and 75 mM LME (Fig. 8).

4. Discussion Microglia, the resident CNS phagocytic cells (Kreutzberg, 1996), are frequent contaminants of astrocyte cultures and can potentially confound interpretation of results derived from analyses with such cultures. Microglial growth can be controlled by treatment of astrocyte cultures with mitotic inhibitors (Hewett, 1999; Solenov et al., 2004; Swanson et al., 1997; Tedeschi et al., 1986; Uliasz and Hewett, 2000). This treatment, which is initiated after astrocytes become confluent, is probably effective because the microglia remain in a proliferative state, whereas contact inhibition induces astrocytes to become quiescent. The ability of Ara-C to significantly antagonize microglial growth in primary astrocyte monolayers was confirmed herein. However, some microglia still remained upon completion of the Ara-C treatment period. These surviving cells might reflect a population of differentiated, non-proliferating microglia. Surprisingly, subjecting

the Ara-C-treated monolayers to 5 mM LME did not further reduce microglia numbers, even after an 8 h exposure to LME. Since microglia frequently reside below the astrocyte monolayer, it is possible that this population was less accessible to cytotoxic levels of 5 mM LME. Indeed, this LME concentration is reported to be most effective when administered to low-density astrocyte cultures or to cells during passage (Giulian and Baker, 1986). Alternatively, it is possible the residual microglia represent a subpopulation that is less sensitive to the cytotoxic actions of LME. Regardless of the reason, we set out to test the hypothesis that higher LME concentrations would efficiently remove the remaining microglia from confluent astrocyte monolayers. Our results indicate that AraC treatment must be followed by a brief exposure (1–1.5 h) to high concentrations of LME (50–75 mM) to maximally deplete microglia from high-density astrocyte monolayers. Successful removal of microglia is attained without the need for cell passage and without causing damage to astrocytes. Microglia depletion was assessed by CD11b immunolabeling and by c-fms mRNA expression. CD11b is a component of the CD11b/CD18 heterodimer complex (Mac-1), which is a member of the ␤2 integrin family (Ehlers, 2000). CD11b is selectively expressed on cells of the myeloid lineage and as such has been employed as marker for microglia (Hewett et al., 1999). The absence of CD11b immunolabeling was consistent with the removal of microglia from monolayers treated with 75 mM LME. However, because microglia have been shown to exhibit heterogeneity for CD11b expression (Santambrogio et al., 2001), these results were confirmed by expression of c-fms mRNA. This transcript encodes the receptor for macrophage colony stimulating factor, which is also expressed by microglia but not astrocytes (Hao et al., 1990; Krady et al., 2002). The absence of both CD11b immunolabeling and c-fms expression after treatment with 75 mM LME provided independent confirmation of the effective removal of the surviving rogue microglia in Ara-Ctreated monolayers. Expression of NOS-2 and COX-2, which can be induced in both microglia and astrocytes by proinflammatory stimuli, was assessed to determine the influence of contaminating microglia on studies with astrocyte cultures. The significant reduction in protein levels of these enzymes after treatment with 75 mM LME not only provides further support of the effectiveness of the protocol, but importantly, also demonstrates that even the low numbers of microglia that remain in Ara-C-treated monolayers can significantly impact results. This further emphasizes the need to completely remove these cells when the goal is to determine astrocyte-specific responses in vitro. The expression of NOS-2 and COX-2 in the astrocyte monolayers following treatment with the high concentrations of LME suggests that the treatment paradigm did not damage the astrocytes, as they retained the ability to respond to proinflammatory stimuli. It should be noted that, while not all GFAP-positive cells in the purified monolayer expressed NOS-2, the proportion of positive cells was similar to that

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

135

Fig. 8. Induction of astrocytic reactive phenotype is unaltered by LME treatment. Ara-C-treated cultures only (a, b, e, f) or those that additionally received treatment with 75 mM LME (60–90 min) (c, d, g, h) were exposed to medium alone (a–d) or 5 mM dbcAMP for 48 h (e–h). Protein levels of GFAP were examined by indirect immunofluorescence. Phase contrast micrographs (a, c, e, g) and their corresponding GFAP immunolabeled images (b, d, f, h) are shown at 40× magnification.

136

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137

observed in monolayers treated with Ara-C in the absence of LME. This suggests that the expression of NOS-2 in only a subpopulation of cells after treatment with 75 mM LME was due to an innate heterogeneity and not to a subtle functional alteration induced by the LME treatment paradigm. Heterogeneous expression of NOS-2 in primary astrocytes has been reported previously lending credence to this possibility (Galea et al., 1992). Further, additional results indicate the astrocyte monolayer was unharmed. First, morphologically, the purified monolayers appeared normal 24 h following exposure to 75 mM LME. Second, total protein synthesis was identical in 75 mM LME-treated cultures as compared to cultures that received Ara-C alone. Finally, the ability of dbcAMP to elicit a reactive phenotype (Eddleston and Mucke, 1993; Fedoroff et al., 1984; Freeman et al., 1989) was retained after our treatment paradigm. Care should be taken to closely adhere to this time frame (1–1.5 h) as longer exposure times can result in significant damage to the astrocyte monolayer (unpublished observations). In sum, the paradigm described in this report depletes microglia, generating highly enriched astrocyte monolayers that remain viable and functional. This goal is attained in confluent monolayers without the need for cell passage. Such purified primary astrocyte monolayers will facilitate elucidation of the molecular mechanisms underlying astrocytespecific responses pertaining to, although not limited to, CNS inflammation.

Acknowledgements This work was supported by Grant NS36812 from NINDS, National Institutes of Health (S.J.H., J.A.H.) and by a grant from the American Heart Association. S.J.H. is an Established Investigator of the American Heart Association.

References Araque A, Carmignoto G, Haydon PG. Dynamic signaling between astrocytes and neurons. Annu Rev Physiol 2001;63:795–813. Boje KM, Arora PK. Microglial-produced nitric oxide and reactive nitrogen oxides mediate neuronal cell death. Brain Res 1992;587:250–6. Bowman CC, Rasley A, Tranguch SL, Marriott I. Cultured astrocytes express toll-like receptors for bacterial products. Glia 2003;43:281–91. Brown DR, Schmidt B, Kretzschmar HA. A neurotoxic prion protein fragment enhances proliferation of microglia but not astrocytes in culture. Glia 1996;18:59–67. Ciccarelli R, Di Iorio P, D’Alimonte I, Giuliani P, Florio T, Caciagli F, et al. Cultured astrocyte proliferation induced by extracellular guanosine involves endogenous adenosine and is raised by the co-presence of microglia. Glia 2000;29:202–11. Doetsch F. The glial identity of neural stem cells. Nat Neurosci 2003;6:1127–34. Dong Y, Benveniste EN. Immune function of astrocytes. Glia 2001;36:180–90. Eddleston M, Mucke L. Molecular profile of reactive astrocytes— implications for their role in neurologic disease. Neuroscience 1993;54:15–36.

Ehlers MR. CR3: a general purpose adhesion-recognition receptor essential for innate immunity. Microbes Infect 2000;2:289–94. Fedoroff S, McAuley WA, Houle JD, Devon RM. Astrocyte cell lineage. V. Similarity of astrocytes that form in the presence of dBcAMP in cultures to reactive astrocytes in vivo. J Neurosci Res 1984;12: 14–27. Fields RD, Stevens-Graham B. New insights into neuron-glia communication. Science 2002;298:556–62. Freeman MR, Beckmann SL, Sueoka N. Regulation of the S100 protein and GFAP genes is mediated by two common mechanisms in RT4 neuro-glial cell lines. Exp Cell Res 1989;182:370–83. Galea E, Feinstein DL, Reis DJ. Induction of calcium-independent nitric oxide synthase activity in primary rat glial cultures. Proc Natl Acad Sci USA 1992;89:10945–9. Giulian D, Baker TJ. Characterization of ameboid microglia isolated from developing mammalian brain. J Neurosci 1986;6:2163–78. Giulian D, Ingeman JE. Colony-stimulating factors as promoters of ameboid microglia. J Neurosci 1988;8:4707–17. Giulian D, Vaca K, Corpuz M. Brain glia release factors with opposing actions upon neuronal survival. J Neurosci 1993;13:29–37. Green LC, Wagner DA, Glogowski J, Skipper PL, Wishnok JS, Tannenbaum SR. Analysis of nitrate, nitrite, and [15N]nitrate in biological fluids. Anal Biochem 1982;126:131–8. Guillemin G, Boussin FD, Croitoru J, Franck-Duchenne M, Le Grand R, Lazarini F, et al. Obtention and characterization of primary astrocyte and microglial cultures from adult monkey brains. J Neurosci Res 1997;49:576–91. Hao C, Guilbert LJ, Fedoroff S. Production of colony-stimulating factor-1 (CSF-1) by mouse astroglia in vitro. J Neurosci Res 1990;27:314–23. Hewett JA, Hewett SJ, Winkler S, Pfeiffer SE. Inducible nitric oxide synthase expression in cultures enriched for mature oligodendrocytes is due to microglia. J Neurosci Res 1999;56:189–98. Hewett SJ. Interferon-gamma reduces cyclooxygenase-2-mediated prostaglandin E2 production from primary mouse astrocytes independent of nitric oxide formation. J Neuroimmunol 1999;94:134–43. Hewett SJ, Corbett JA, McDaniel ML, Choi DW. Interferon-gamma and interleukin-1 beta induce nitric oxide formation from primary mouse astrocytes. Neurosci Lett 1993;164:229–32. Janzer RC. The blood–brain barrier: cellular basis. J Inherit Metab Dis 1993;16:639–47. John GR, Lee SC, Song X, Rivieccio M, Brosnan CF. IL-1-regulated responses in astrocytes: relevance to injury and recovery. Glia 2005;49:161–76. Krady JK, Basu A, Levison SW, Milner RJ. Differential expression of protein tyrosine kinase genes during microglial activation. Glia 2002;40:11–24. Kreutzberg GW. Microglia: a sensor for pathological events in the CNS. Trends Neurosci 1996;19:312–8. Malhotra SK, Shnitka TK, Elbrink J. Reactive astrocytes—a review. Cytobios 1990;61:133–60. Minghetti L, Nicolini A, Polazzi E, Greco A, Perretti M, Parente L, et al. Down-regulation of microglial cyclo-oxygenase-2 and inducible nitric oxide synthase expression by lipocortin 1. Br J Pharmacol 1999;126:1307–14. Montgomery DL. Astrocytes: form, functions, and roles in disease. Vet Pathol 1994;31:145–67. Murphy S. Production of nitric oxide by glial cells: regulation and potential roles in the CNS. Glia 2000;29:1–13. O’Banion MK. Cyclooxygenase-2: molecular biology, pharmacology, and neurobiology. Crit Rev Neurobiol 1999;13:45–82. Parri R, Crunelli V. An astrocyte bridge from synapse to blood flow. Nat Neurosci 2003;6:5–6. Santambrogio L, Belyanskaya SL, Fischer FR, Cipriani B, Brosnan CF, Ricciardi-Castagnoli P, et al. Developmental plasticity of CNS microglia. Proc Natl Acad Sci USA 2001;98:6295–300. Simmons ML, Murphy S. Induction of nitric oxide synthase in glial cells. J Neurochem 1992;59:897–905.

M.E. Hamby et al. / Journal of Neuroscience Methods 150 (2006) 128–137 Smith KJ, Lassmann H. The role of nitric oxide in multiple sclerosis. Lancet Neurol 2002;1:232–41. Solenov E, Watanabe H, Manley GT, Verkman AS. Sevenfold-reduced osmotic water permeability in primary astrocyte cultures from AQP4-deficient mice, measured by a fluorescence quenching method. Am J Physiol Cell Physiol 2004;286:426–32. Swanson RA, Liu J, Miller JW, Rothstein JD, Farrell K, Stein BA, et al. Neuronal regulation of glutamate transporter subtype expression in astrocytes. J Neurosci 1997;17:932–40. Tedeschi B, Barrett JN, Keane RW. Astrocytes produce interferon that enhances the expression of H-2 antigens on a subpopulation of brain cells. J Cell Biol 1986;102:2244–53. Thery C, Stanley ER, Mallat M. Interleukin 1 and tumor necrosis factoralpha stimulate the production of colony-stimulating factor 1 by murine astrocytes. J Neurochem 1992;59:1183–6. Thiele DL, Kurosaka M, Lipsky PE. Phenotype of the accessory cell necessary for mitogen-stimulated T and B cell responses in

137

human peripheral blood: delineation by its sensitivity to the lysosomotropic agent, l-leucine methyl ester. J Immunol 1983;131: 2282–90. Trackey JL, Uliasz TF, Hewett SJ. SIN-1-induced cytotoxicity in mixed cortical cell culture: peroxynitrite-dependent and -independent induction of excitotoxic cell death. J Neurochem 2001;79:445– 55. Uliasz TF, Hewett SJ. A microtiter trypan blue absorbance assay for the quantitative determination of excitotoxic neuronal injury in cell culture. J Neurosci Methods 2000;100:157–63. Vidwans AS, Uliasz TF, Hewett JA, Hewett SJ. Differential modulation of prostaglandin H synthase-2 by nitric oxide-related species in intact cells. Biochemistry 2001;40:11533–42. Xiong H, Yamada K, Jourdi H, Kawamura M, Takei N, Han D, et al. Regulation of nerve growth factor release by nitric oxide through cyclic GMP pathway in cortical glial cells. Mol Pharmacol 1999;56: 339–47.