Journal of Neuroimmunology, 30 (1990) 239-243 Elsevier
239
JNI 01028
Short Communication
Human astrocytes proliferate in response to tumor necrosis factor alpha Barbara P. Barna 1, Melinda L. Estes 2, Barbara S. Jacobs 1, Susan H u d s o n 3 and Richard M. Ransohoff 3,4,5 Departments of j Irnmunopathology, 2 Pathology, 3 Molecular Biology, ~ Neurology and 5 Mellen Center for Multiple Sclerosis Treatment and Research, The Cleveland Clinic Foundation, Cleveland, OH 44195, U.S.A.
(Received 11 July 1990) (Revised, received 9 August 1990) (Accepted 10 August 1990)
Key words: Astrocyte; Tumor necrosis factor-a; Central nervous system
Summary Two different human astrocytic cell lines derived from adult epilepsy surgical specimens were exposed in vitro to concentrations of 1-100 ng/ml recombinant tumor necrosis factor alpha (TNFa). Results indicated dose-dependent stimulation of DNA synthesis and proliferation. Both of these effects were abrogated by treatment with monoclonal antibody specific for TNFa but not by irrelevant murine IgG. Immunocytochemical characterization of TNFa-treated and control cultures indicated that greater than 98% of proliferating cells contained cytoplasmic glial fibrillary acidic protein (GFAP), and were therefore astrocytic in nature. These studies demonstrate that growth of adult human non-neoplastic astrocytes is stimulated by TNFa, an inflammatory cytokine produced primarily by macrophages but also by astrocytes.
Introduction
Tumor necrosis factor alpha (TNFa) was originally described by Carswell et al. (1975) as a
Address for correspondence: Dr. Barbara P. Barna, Ph.D., Head, Cellular Immunology Section, Department of Immunopathology, One Clinic Center, 9500 Euclid Avenue, Cleveland, OH 4,1195-5131, U.S.A. This work was supported by a Clinical Investigator Development Award (K08-01265, NINDS) (R.M.R.), a grant from the Reinberger Foundation (R.M.R.), and in part by CA-49950 from the National Institutes of Health (B.P.B.). Dr. Ransohoff is a Harry M. Weaver Neuroscience Scholar of the National Multiple Sclerosis Society (JF, 2032 A-l).
factor producing hemorrhagic necrosis in vivo. TNFa has since been shown to be a potent immunoregulatory molecule and the mediator of a broad range of host inflammatory responses (Beutler and Cerami, 1989). The primary sources of TNFa are monocytes and macrophages which synthesize high levels when activated (Beutler and Cerami, 1989). Among central nervous system (CNS) cells, both microglia (Frei et al., 1987) and astrocytes (Lieberman et al., 1989; Chung and Benveniste, 1990) have been shown to secrete TNFa in vitro. TNFa has also been detected in the cerebrospinal fluid of patients with bacterial meningitis (Leist et al., 1988), and more recently in ghal fibrillary acidic protein (GFAP)-positive
0165-5728/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
240
astrocytes within multiple sclerosis (MS) brain tissue (Hofman et al., 1989). These findings strongly suggest that T N F a is produced within the CNS during inflammatory injury. Vilcek and co-workers (1986) have shown that T N F a promotes growth of human diploid fibroblasts. Human T N F a has also been found to stimulate proliferation in bovine astrocyte cultures (Selmaj et al., 1990). The present study demonstrates that T N F a augments DNA synthesis and proliferation in cultures of adult human astrocytes as well (Barna, 1989a). Our results* support the concept of T N F a regulation of astrocyte inflammatory responses in vivo.
Materials and methods
Cell cultures. Two human glial cell lines (PIN and W3N) were established from temporal iobectomy tissue specimens obtained during surgery for intractable epilepsy. Tissues were finely minced and cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY, U.S.A.), L-glutamine, and antibiotics, as previously described (Barna et al., 1985). Monolayers of established cultures were dispersed with 0.05% trypsin in versene (Gibco), and subcuitured once to provide a stock of primary passage cells for cryogenic storage and characterization. Studies were carried out with second and third passage cells from cryogenic storage. Detailed characterization of cell line P1N, which demonstrates greater than 95% GFAP-positive cells, has previously been described (Barna et al., 1989b). In cell line W3N, greater than 80-85% of cells contained cytoplasmic G F A P (Estes et al., 1990). Cell lines were morphologically and phenotypically distinct from fibroblasts, leptomeningeal cells, oligodendrocytes, and endothelial cells (Estes et al., 1990). DNA synthesis. Evaluation of DNA synthesis by tritiated ([3H]) thymidine incorporation was carried out as previously described (Barna et al., 1985). Briefly, microtiter plate cultures of 5000 * These results were presented in part at the annual meeting of the American Neurological Association, New Orleans, LA, September 25, 1989.
cells per well were incubated for 90-96 h in control medium and in medium containing 1-100 n g / m l rTNFa. Tritiated thymidine (5 ~ C i / m l ) was included during the last 2 6 h of culture. After enzymatic (0.05% trypsin in versene) dispersion, cells were aspirated onto filter paper, washed with distilled water to remove unincorporated radiolabel, and radioactive contents were determined in a liquid scintillation counter. Results were expressed as mean counts per minute (cpm) + SD. Proliferation, Proliferation was assessed in microtiter plate cultures as described previously (Viicek et al., 1986). Seven days after culture of 5000 cells per well in medium with and without 1-100 n g / m l rTNFcc, cells were washed in saline and stained with 0.1% crystal violet for 30 rain. After rinsing with tap water and air-drying, crystal violet was eluted in 95% ethanol, and optical density (OD) at 550 nm was determined in an enzyme-linked immunosorbent assay (ELISA) plate reader (Flow Laboratories, McLean, VA, U.S.A.). Results were expressed as mean OD__+ SD. OD significantly ( p < 0.01) correlated with cell number (r = 1.0) in microtiter plate cultures. Characterization of proliferating cells. Bromodeoxyuridine (BrdU) incorporation was used to detect proliferating cells as described by Yong and Kim (1987) with modification. Cultures established on glass coverslips in Petri dishes were incubated for 18 h with 50 /.tM BrdU in medium with and without 50 n g / m l rTNFa. Coverslips were then fixed in 70% ethanol and permeabilized in 0.1% Triton X-100. After washing in phosphate-buffered saline (PBS), cells were treated for 10 min with 1 M HCI to denature native DNA, then neutralized in 0.1 M sodium borate, pH 9.2 for an additional 10 min. Coverslips were then immunostained sequentially for 1 h per stain with: rabbit anti-bovine GFAP, 1:100 (Dako): fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit serum; monoclonal anti-BrdU (Beeton-Dickinson, Mountain View, CA, U.S.A.); and rhodamine-conjugated rabbit anti-mouse immunoglobulin. Two 5-rain washes in PBS followed each staining step. Coverslips were then mounted in 90% glycerol for microscopic determination of the percentage of GFAP-positive cells containing nuclear BrdU in 10 high power fields.
241
Reagents. Recombinant T N F a was obtained from ICN (Cleveland, OH, U.S.A.) and contained a biological activity of 20 units/ng. Monoclonal antibody to T N F a was obtained from Boehringer-Mannheim Biochemicals (Indianapolis, IN, U.S.A.), and for treatment of r T N F a preparations was coupled to Protein A-Sepharose beads (Pharmacia, Piscataway, N J, U.S.A.). Irrelevant murine IgG obtained from Becton-Dickinson was similarly coupled to Protein A-Sepharose beads (Pharmacia). Statistical analysis. Comparisons between rTNFa-treated and control responses were evaluated by Student's t-test. Differences were considered significant if p < 0.05.
";
..
) I
(-_? k
:%.~.
R ~
Effect of rTNFa on DNA synthesis and proliferation. Exposure of culture cell lines PIN and W3N to 1.0-100 n g / m l r T N F a for 96 h resulted in significant ( p
239
JNI 01028
Short Communication
Human astrocytes proliferate in response to tumor necrosis factor alpha Barbara P. Barna 1, Melinda L. Estes 2, Barbara S. Jacobs 1, Susan H u d s o n 3 and Richard M. Ransohoff 3,4,5 Departments of j Irnmunopathology, 2 Pathology, 3 Molecular Biology, ~ Neurology and 5 Mellen Center for Multiple Sclerosis Treatment and Research, The Cleveland Clinic Foundation, Cleveland, OH 44195, U.S.A.
(Received 11 July 1990) (Revised, received 9 August 1990) (Accepted 10 August 1990)
Key words: Astrocyte; Tumor necrosis factor-a; Central nervous system
Summary Two different human astrocytic cell lines derived from adult epilepsy surgical specimens were exposed in vitro to concentrations of 1-100 ng/ml recombinant tumor necrosis factor alpha (TNFa). Results indicated dose-dependent stimulation of DNA synthesis and proliferation. Both of these effects were abrogated by treatment with monoclonal antibody specific for TNFa but not by irrelevant murine IgG. Immunocytochemical characterization of TNFa-treated and control cultures indicated that greater than 98% of proliferating cells contained cytoplasmic glial fibrillary acidic protein (GFAP), and were therefore astrocytic in nature. These studies demonstrate that growth of adult human non-neoplastic astrocytes is stimulated by TNFa, an inflammatory cytokine produced primarily by macrophages but also by astrocytes.
Introduction
Tumor necrosis factor alpha (TNFa) was originally described by Carswell et al. (1975) as a
Address for correspondence: Dr. Barbara P. Barna, Ph.D., Head, Cellular Immunology Section, Department of Immunopathology, One Clinic Center, 9500 Euclid Avenue, Cleveland, OH 4,1195-5131, U.S.A. This work was supported by a Clinical Investigator Development Award (K08-01265, NINDS) (R.M.R.), a grant from the Reinberger Foundation (R.M.R.), and in part by CA-49950 from the National Institutes of Health (B.P.B.). Dr. Ransohoff is a Harry M. Weaver Neuroscience Scholar of the National Multiple Sclerosis Society (JF, 2032 A-l).
factor producing hemorrhagic necrosis in vivo. TNFa has since been shown to be a potent immunoregulatory molecule and the mediator of a broad range of host inflammatory responses (Beutler and Cerami, 1989). The primary sources of TNFa are monocytes and macrophages which synthesize high levels when activated (Beutler and Cerami, 1989). Among central nervous system (CNS) cells, both microglia (Frei et al., 1987) and astrocytes (Lieberman et al., 1989; Chung and Benveniste, 1990) have been shown to secrete TNFa in vitro. TNFa has also been detected in the cerebrospinal fluid of patients with bacterial meningitis (Leist et al., 1988), and more recently in ghal fibrillary acidic protein (GFAP)-positive
0165-5728/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
240
astrocytes within multiple sclerosis (MS) brain tissue (Hofman et al., 1989). These findings strongly suggest that T N F a is produced within the CNS during inflammatory injury. Vilcek and co-workers (1986) have shown that T N F a promotes growth of human diploid fibroblasts. Human T N F a has also been found to stimulate proliferation in bovine astrocyte cultures (Selmaj et al., 1990). The present study demonstrates that T N F a augments DNA synthesis and proliferation in cultures of adult human astrocytes as well (Barna, 1989a). Our results* support the concept of T N F a regulation of astrocyte inflammatory responses in vivo.
Materials and methods
Cell cultures. Two human glial cell lines (PIN and W3N) were established from temporal iobectomy tissue specimens obtained during surgery for intractable epilepsy. Tissues were finely minced and cultured in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) (Gibco, Grand Island, NY, U.S.A.), L-glutamine, and antibiotics, as previously described (Barna et al., 1985). Monolayers of established cultures were dispersed with 0.05% trypsin in versene (Gibco), and subcuitured once to provide a stock of primary passage cells for cryogenic storage and characterization. Studies were carried out with second and third passage cells from cryogenic storage. Detailed characterization of cell line P1N, which demonstrates greater than 95% GFAP-positive cells, has previously been described (Barna et al., 1989b). In cell line W3N, greater than 80-85% of cells contained cytoplasmic G F A P (Estes et al., 1990). Cell lines were morphologically and phenotypically distinct from fibroblasts, leptomeningeal cells, oligodendrocytes, and endothelial cells (Estes et al., 1990). DNA synthesis. Evaluation of DNA synthesis by tritiated ([3H]) thymidine incorporation was carried out as previously described (Barna et al., 1985). Briefly, microtiter plate cultures of 5000 * These results were presented in part at the annual meeting of the American Neurological Association, New Orleans, LA, September 25, 1989.
cells per well were incubated for 90-96 h in control medium and in medium containing 1-100 n g / m l rTNFa. Tritiated thymidine (5 ~ C i / m l ) was included during the last 2 6 h of culture. After enzymatic (0.05% trypsin in versene) dispersion, cells were aspirated onto filter paper, washed with distilled water to remove unincorporated radiolabel, and radioactive contents were determined in a liquid scintillation counter. Results were expressed as mean counts per minute (cpm) + SD. Proliferation, Proliferation was assessed in microtiter plate cultures as described previously (Viicek et al., 1986). Seven days after culture of 5000 cells per well in medium with and without 1-100 n g / m l rTNFcc, cells were washed in saline and stained with 0.1% crystal violet for 30 rain. After rinsing with tap water and air-drying, crystal violet was eluted in 95% ethanol, and optical density (OD) at 550 nm was determined in an enzyme-linked immunosorbent assay (ELISA) plate reader (Flow Laboratories, McLean, VA, U.S.A.). Results were expressed as mean OD__+ SD. OD significantly ( p < 0.01) correlated with cell number (r = 1.0) in microtiter plate cultures. Characterization of proliferating cells. Bromodeoxyuridine (BrdU) incorporation was used to detect proliferating cells as described by Yong and Kim (1987) with modification. Cultures established on glass coverslips in Petri dishes were incubated for 18 h with 50 /.tM BrdU in medium with and without 50 n g / m l rTNFa. Coverslips were then fixed in 70% ethanol and permeabilized in 0.1% Triton X-100. After washing in phosphate-buffered saline (PBS), cells were treated for 10 min with 1 M HCI to denature native DNA, then neutralized in 0.1 M sodium borate, pH 9.2 for an additional 10 min. Coverslips were then immunostained sequentially for 1 h per stain with: rabbit anti-bovine GFAP, 1:100 (Dako): fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit serum; monoclonal anti-BrdU (Beeton-Dickinson, Mountain View, CA, U.S.A.); and rhodamine-conjugated rabbit anti-mouse immunoglobulin. Two 5-rain washes in PBS followed each staining step. Coverslips were then mounted in 90% glycerol for microscopic determination of the percentage of GFAP-positive cells containing nuclear BrdU in 10 high power fields.
241
Reagents. Recombinant T N F a was obtained from ICN (Cleveland, OH, U.S.A.) and contained a biological activity of 20 units/ng. Monoclonal antibody to T N F a was obtained from Boehringer-Mannheim Biochemicals (Indianapolis, IN, U.S.A.), and for treatment of r T N F a preparations was coupled to Protein A-Sepharose beads (Pharmacia, Piscataway, N J, U.S.A.). Irrelevant murine IgG obtained from Becton-Dickinson was similarly coupled to Protein A-Sepharose beads (Pharmacia). Statistical analysis. Comparisons between rTNFa-treated and control responses were evaluated by Student's t-test. Differences were considered significant if p < 0.05.
";
..
) I
(-_? k
:%.~.
R ~
Effect of rTNFa on DNA synthesis and proliferation. Exposure of culture cell lines PIN and W3N to 1.0-100 n g / m l r T N F a for 96 h resulted in significant ( p
