0306-4522/91$3.00+ 0.00 Pergamon Press plc 0 1991IBRO
Neuroscience
Vol. 41, No. I, pp. 1499158,1991 Printed in Great Britain
ISOLATION AND CHARACTERIZATION OF HUMAN FETAL BRAIN-DERIVED MICROGLIA IN IN FrlTRO CULTURE N. F. HASSAN, D. E. CAMPBELL, S. RIFAT and S. D. DOUGLAS* Division of Infectious Diseases and Immunology, Clinical Immunology Laboratory and The Joseph Stokes Jr. Research Institute, Children’s Hospital of Philadelphia, Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, U.S.A. Abstract-Human brain microglia may play a central role in immunopathogenesis of CNS diseases including HIV infection, multiple sclerosis and Alzheimer’s disease. In order to investigate the possible relationship between microglia and the mononuclear phagocyte system, human brain microglia were isolated from 14-18-week-old fetal brains, and maintained in in vitro culture. Enriched fetal brain microglia were stained for different monocyte/macrophage and glial cell markers. Fresh dissociated brain cells lacked macrophage surface markers. Isolated microglial cells stained positive for complement receptor C3bi, Class II buman leukocyte antigen-DR (HLA-DR)] antigen and with the lectin Ricinus communis. Microglia also share several functional properties with monocyte/macrophages, which include generation of superoxide anion and histochemically demonstrable intracellular acid phosphatase and non-specific esterase. Primary human dissociated brain cultures were maintained in culture for at least 28 weeks. Although microglia were not observed above the astrocyte cell layer after 5 weeks in culture, micoglia-like cells appear below the astrocyte layer after 12 weeks in culture. These cells stained positive for non-specific esterase and displayed oxidative burst activity upon activation with phorbol myristate acetate. Thus, we have successfully isolated an enriched population of microglia from human fetal brain and have demonstrated that these cells possess markers and properties which are characteristics of mononuclear phagocytes.
Brain macrophages or micro&a have been the subject of considerable controversy since their initial description by Del Rio-Hortega using a weak silver carbonate stain, 50 years ago.’ Del Rio-Hortega proposed that
microglia were intrinsic cells of the brain, of mesodermal origin, distributed throughout the CNS, and found in two predominant forms: the ameboid, or macrophage-like, cells which occurred in developing brain and at sites of injury, and the ramified, or highly branched cells, considered to be like “quiescent” cells in mature CNS.” The cell of origin of the brain microglia is very controversial. Some investigators believe that microglia are derived from migrating blood monocytes during the neonatal period,8,” whereas others maintain that the microglia are of neuroectodermal origin.7~‘3Several studies have been performed to identify microglia in situ in undamaged human brain tissue. These cells were non-reactive using anti-monocyte or anti-B,-microglobulin antibodies raised against peripheral blood monocytes,i4 and therefore it was concluded that the resting human microglia are likely to be of neural origin since they *To whom correspondence should be addressed. Abbreviations: DMEM, Dulbecco’s Eagle medium; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; GALC, galactocerebroside; GFAP, glial fibrillary acidic protein; HBSS, Hank’s balanced salt solution; NBT, nitroblue tetrazolium; PBS, phosphate-buffered saline; PMA, phorbol myristate acetate; RCA-l, Ricks communis agglutinin-1 .
do not express known mononuclear phagocyte cell markersi It was further postulated that macrophages which occur in inflammatory diseases of the CNS are blood monocytes migrating in response to chemotactic stimuli from damaged neuronal cells, rather than microglia cell~.~ Recently, lectin histochemistry has provided probes for identification of microglia. Ricinus communis agglutinin-1 (RCA-l), which binds specifically to galactose residues on the brain microglia cell surface, has been used to demonstrate the presence of microglia in human nervous tissues.‘6 RCA-l-positive microglia were observed throughout the CNS, more frequently in white than gray matter.’ Although several studies have attempted to characterize brain microglia in brain slices, the cellular heterogeneity of the CNS and the presence of contaminating blood monocytes in choroid plexus and blood vessels in brain sections have led to conflicting results concerning the origin and characteristics of brain microglia cell~.‘~J~ In an effort to circumvent this problem, several in vitro techniques were developed to isolate and purify brain microglia in in oitro culture. The McCarthy and de Vellis isolation technique has been widely applied to obtain highly purified populations of astrocytes and oligodendrocytes,‘” and has also been used to obtain enriched microglia preparations in in vitro culture derived from rodent brains.6 In this study, we describe the isolation of human fetal brain microglia using a modification of the McCarthy and de Vellis isolation technique. Partially 149
150
N. F. HASSAN et al.
purified cultured human brain microglia obtained from fetal brain cell dissociation cultures were characterized using different human monocyte/macrophage markers. EXPERIMENTAL
PROCEDURES
Materials Human fetal brain materials from elective abortions (1418 weeks) were provided by the International Institute for the Advancement of Medicine (Havertown, PA). Penicillin, streptomycin, 2.5% trypsin, L-glutamine and non-essential amino acids were obtained from Gibco Laboratories, Grand Island, NY. Deoxyribonuclease, insulin, non-specific esterase and acid phosphatase histochemical staining kits were obtained from Sigma Chemicals, St Louis, MO. Petri dishes (60 mm) were purchased from NUNC, Naperville, IL. Nylon mesh and Leukostat stain were obtained from Fisher Scientific, Fair Lawn, NJ. Antibodies used in this study include: mouse monoclonal antibodies to CD3, CD14, CD1 lb, A2B5 and anti-human leucocyte antigen-DR (HLA-DR), and rabbit polyclonal antibodies anti-galactocerebroside and anti-glial fibrillary acidic protein. Monoclonal antibodies anti-human Leu-M3 (CD14) (Becton and Dickinson, Mountain View, CA), OK T3 (CD3) (Ortho Diagnostic Systems, Inc, Raritan, NJ), A2B5, hybridoma supernatant (American Tissue Culture Collection, Bethesda, MD), mouse anti-human C3bi (CDllb) (provided by Dr Donald Anderson, Baylor College of Medicine, Houston, TX), mouse anti-human HLA-DR (12) (Coulter, Hialeah, FL), rabbit anti-galactocerebroside (GALC) (Chemicon, El Segundo, CA) and rabbit anti-glial fibrillary acidic protein (GFAP) (Advanced Immunochemical Services Inc., Los Angeles, CA) were used in these studies. Fluorescein isothiocyanate (FITC)-conjugated goat F(ab’), anti-mouse immunoglobulin was obtained from Tago, Burlingame, CA. Flow cytometric analyses were performed using an EPICS C (Coulter, Hialeah, FL). Positive antibody staining was detected using a Zeiss epifluorescence microscope equipped with 35 mm camera and an HBO 100 W high-pressure mercury lamp. Immunoperoxidase histochemical staining FETAL HUMAN (14-15 week
was performed using a Vectastain ABC kit (Vector Laboratories, Burlingame, CA). Human recombinant gamma interferon was kindly provided by Dr Peter Sorter, Hoffman La Roche, Nutley, NJ. In vitro cells
culrure
of primary
dissociated
human
fetal
Human fetal brain materials were processed following elective abortions. Figure I summarizes the technique used for the isolation of human fetal brain microglia. Brain tissues were rinsed several times with cold Hank’s balanced salt solution (HBSS) supplemented with 50 U/ml penicillin and 50 pg/ml streptomycin, and meninges with attached blood vessels were macroscopically dissected from brain tissues. Following dissection, the brain tissues were sliced into small pieces and transferred to 60 mm sterile Petri dishes on ice and mechanical dissociation was carried out using sterile surgical scalpels. Prior to enzymatic treatment, 500 ~1 aliquots of dissociated cells were removed for cell characterization including flow cytometric analysis and enzyme histochemistry. Cold HBSS was then added to bring the cell volume to 16 ml in each Petri dish and 4ml of 2.5% trypsin and deoxyribonuclease (10 pg/ml) were then added to each dish. The cell mixture was shaken for 15 min at 200 r.p.m., 37°C and then centrifuged three times for 10 min at 1200 r.p.m., 4°C. The cell suspension was then filtered through a sterile nylon mesh with a pore size of 200 pm. The filtered cells were suspended in culture medium, counted and seeded into 75 cm* flasks at a cell concentration of 108-lo9 cells/ml for continuous culture in a humidified atmosphere of 95% air and 5% CO, at 37°C. The culture medium consists of Dulbecco’s Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS), 50 U/ml penicillin, 50 pg/ml streptomycin, 2.5 pg/ml fungizone, 4 mM L-glutamine, 100 PM non-essential amino acids and 5 pg/ml insulin. Five days after initial inoculation the flasks were rinsed several times with warm DMEM to remove non-adherent cells and fresh culture medium supplemented with 10% conditioned medium from human astrocyte culture was added. The medium from primary dissociated human brain cultures was changed every 5 days and brain cells were maintained in long-term culture for 12-28 weeks.
BRAIN old)
Long term (up to 2:
Mechanical dissociation
-
Microglia analysis
--)
Microglia analysis
culture weeks)
Orbital ‘shaking (Detachment of microglial I
/ Centrifugation (1200rpm. 10 min
x 31
a
_
brain
Culture (4-5 weeks1 Filtration
Fig. 1. Isolation and culture of human fetal brain microglia in in vifro culture. performed in freshly dissociated fetal brain cells, purified microglia after short-term and microglia in long-term culture (12-28 weeks).
Cell analyses were culture (4-5 weeks)
Human Preparation of human (conditioned medium)
astrocyte
fetal brain-microglia:
culture
supernatant
Human primary brain dissociated cultures maintained for 3-4 weeks were shaken for 16 h at 200 r.p.m., 37°C. Following overnight shaking of the cultures, loosely adherent and detached cells were removed and cultures were trypsinized and subcultures were prepared. The cells were maintained in culture medium supplemented with 10% FCS, and the culture medium was changed every 5 days. Supernatants of confluent astrocyte cultures were harvested by centrifugation (2OOOg, 10 min), aliquoted and stored at -20°C until used. Characterization offreshly dissociated human fetal brain cells Cell surface antigen analysis. Mechanically dissociated human fetal brain cells werk stained with the monoclonal antibodies anti-human Leu-M3. OK T3 and A2B5, hybridoma culture supernatant. The cells were incubated with the monoclonal antibodies for 30 min at 4”C, washed in phosphate-buffered saline (PBS) containing 0.01% sodium azide and incubated with 1: 20 dilution of FITC-conjugated goat F(ab’)? anti-mouse immunoglobulin for 30 min. Flow cytometric analyses were performed using a fluorescence activated cell sorter, EPICS C (Coulter). Forward and right angle light scatter were used to establish appropriate gates on the cells by excluding non-viable cells. The fluorescence intensity distribution of 5000 cells was accumulated for analysis. A threshold of fluorescence was set for each experiment and the background from control cells labeled with isotype-matched mouse immunoglobulin was subtracted from the test antibody-labeled preparations. Histochemical staining. Cytospins of freshly dissociated human fetal brain cells were stained with Leukostat stain, and tested for non-specific esterase and acid phosphatase using histochemical staining kits. For each preparation 600 cells were counted under oil immersion and cells containing dark cytoplasmic granules were scored as positive. isolation of human fetal brain-derived microglia. After 7-10 days, a confluent layer of fibroblast-like cells was observed in the primary human brain cultures. Four to five weeks later, rounded and spindle-shaped phase-bright cells appear in small colonies above the confluent monolayer. The cultures were then shaken in an orbital shaker for 16 h at 150 r.p.m., 37°C. The supematant media containing the detached cells were centrifuged for 10 min at 1200 r.p.m., 4°C resuspended, counted and plated on poly+-lysinecoated coverslips at a density of 5-10 x lo4 cells/coverslip in culture medium. After 2 h, coverslips were washed with DMEM and stained for different intracellular and surface markers. Characterization ofpuriJied humanfetal brain derived microglia in in vitro culture Ceil surface antigen analysis. Isolated human brain microglia were stained for cell surface and intracytoplasmic antigens using different human monocyte/macrophage markers. Indirect immunofluorescence microscopy and the immunopcroxidase technique were applied to study the expression of different markers. The following antibodies were used: mouse anti-human C3bi (1: 100 dilution), mouse anti-human HLA-DR (12) (I:2 dilution), rabbit anti-GALC (I:50 dilution) and rabbit anti-GFAP (1:40 dilution). Rabbit immunoglobulin or isotype-matched mouse immunoglobulin were used as a control antibody for each of the reagents mentioned above. Positive antibody staining was detected using fluorescence microscopy or immunoperoxidase histochemical staining. Fluorescence microscopy was performed with a Zeiss epifluorescence microscope equipped with a 35 mm camera. For each prepartion 200 cells were counted and positive cells were scored in each case.
Immunoperoxidase histochemical staining was performed using a Vectastain ABC kit. Microglia were also stained with biotinylated lectin RCA-l and visualized with immunoperoxidase histochemical staining. In order to study the
isolation
151
and culture
of HLA-DR on microglia, cells were treated with 1000 U/ml human recombinant gamma interferon for 18 h and stained with mouSe anti-human HLA-DR monoclonal antibody (12). Histochemical staining. Cytospins of isolated human fetal brain microglia were tested for non-specific esterase and acid phosphatase using histochemical staining kits. Cells were regulation
also assessed for superoxide anion generation using the nitroblue tetrazolium (NBT) slide test.” Cells were incubated
with NBT dye and phorbol myristate acetate (PMA) (1 pg/mI) as stimulant for 90 min at 5% CO,, 37°C. Coverslips were washed several times and fixed in methanol for 20min and counterstained with safranin. For each preparation 200 cells were counted under oil immersion and cells with dark blue formazan granules were scored as positive.
Characterization of primary dissociated human fetal brain cells in long-term culture Human brain cells in long-term culture were assessed for superoxide generation by the NBT slide test and tested for intracellular non-specific esterase using hi&chemical staining kits as described above.
RESULTS
Histochemistry and cell surface antigen analysis primary dissociated human fetal brain cells
of
Histocytochemical staining of dissociated human fetal brain cell suspensions using Leukostat (Modified Wright’s stain) reveal the presence of neuroblast-like cells 20-40 pm in diameter with large nuclei and thin cytoplasm (Fig. 2A). A small percentage of the cells stained positive for non-specific esterase and acid phosphatase (Fig. 2B,C). The percentage of cells stained positive for acid phosphatase (seven experiments) was higher than cells stained positive for non-specific esterase (1.4-2.8 % and 0.7-2.5 %, respectively). In only one experiment the cells positive for non-specific esterase scored higher than cells with acid phosphatase (3.0 and 1.4%, respectively; Table 1). Flow cytometric analysis was carried out in order to determine whether the non-specific esterase and acid phosphatase-positive cells were peripheral blood monocytes contaminating the brain cell suspensions. Human fetal brain cell suspensions derived from 14-18-week-old fetuses did not express either CD14 or CD3 using the monocyte/macrophage marker Leu-M3 and T-lymphocyte surface marker 0KT3 monoclonal antibodies, respectively (Fig. 3). However, at least 50% of mechanically dissociated fetal brain cells stained positive for A2B5, a surface antigen expressed on neuroglial precursor cells (Fig. 3). Moreover, surface staining for other cell antigens including CDllb, CD4 and Class II (HLA-DR) was negative (data not shown). Isolation and characterization brain -derived microglia
of isolated human fetal
Primary dissociated fetal brain cells were plated and maintained in in vitro culture in the presence of 10% conditioned medium from human fetal astrocyte culture. Increased number of cells and cell spreading were observed in cultures incubated with conditioned medium (Fig. 4A,B). After 3-4 weeks, cells 15-25 pm
152
N.
F.
HASAN
rt ul
Fig. 2. Cytospin preparations of primary dissociated human fetal brain cells stained with Leukostat stain (modified Wright’s stain) (A), non-specific esterase (arrow) (B) and acid phosphatase (arrowj (C). Light microscopy. x 900.
in di,ameter with fine plasma membrane processes and vact rolated cytoplasm were observed growing on the top of a confluent layer (bed layer) of fibroblast-like cells (Fig. 4C) that consisted primarily of astrocytes
(data not shown). With increased time in culture these cells become more rounded and loosely attach1 ed to the bed layer (Fig. 4D) and shaking of the cultu tes is usually Performed at this time (4-5 weeks).
Human Table human
1. Histochemical fetal brain cells
fetal
Non-specific
esterase
isolation
staining of primary dissociated for non-specific esterase and acid phosphatase* Experiment
Acid phosphatase
brain-microglia:
1
2
3
4
0.7
1.7
1.1
1.4
1.6 2.8
3
1.4 2
*Results are expressed as percentage seven different experiments.
6
7
2.5
1
1.4 4.1 2.5
positive
cells
from
Characterization of primary dissociated brain cells in long-term culture
Histochemistry Primary dissociated fetal brain cells were shaken after 4-5 weeks in in vitro culture. Brain microglia recovered from six different experiments gave yields of 0.993.6 x lo6 cells with 54-90% viability (Table 2). The enriched microglia stained positive for nonspecific esterase and acid phosphatase (Fig. 5A,B and Table 2) and generated superoxide anion upon activation with PMA using the NBT slide test (15-49%).
Enriched brain microglia were stained for surface and intracellular monocyte/macrophage and glial cell markers. Using the indirect immunofluorescence technique, cells stained variably for C3bi receptor (1674%; Fig. 5C) and HLA-DR (12) antigen (lo-53%)
Exp.
Cell no. ( x 106)
no. 1 2 3 4 5 6
Viability (%I
0.9 1.1 3.6 1.3 1.2 1.2
*Non-specific
2. Cell characterization
67 82 62 54 90 83
of human
Anti-C3bi (% + ve)
fetal
brain
Ia + INFy
71 16 74 ND ND 60
10 12 ND ND ND 53
human fetal
Primary dissociated human fetal brain cultures were shaken 3-4 weeks after culture and were then maintained in vitro for more than 28 weeks. Cultures were provided with fresh media every 5 days and maintained at 37°C and 5% CO,. Following culture shaking, microglia were not seen growing above the astrocyte cell layer. However, after 12 weeks, groups of cells were observed under the astrocyte cell layer in all brain cultures. These cells were present in small clusters of 7-10 cells heterogeneous in size, highly branched with large nuclei and thin cytoplasm (Fig. 6A). They stained positive for non-specific esterase (Fig. 6B) and generated superoxide anion upon activation with PMA as assessed by NBT
Cell surface antigen expression
Table
153
and did not stain for GALC, an oligodendrocyte cell marker (Table 2). Moreover, the addition of 1000 U/ ml human recombinant gamma interferon to brain microglia cultures for 18 h increased the expression of HLA-DR (12) cell surface antigen (19-69%; Fig. 5D and Table 2). Using immunoperoxidase histochemical staining, human fetal brain microglia stained positive for RCA-l (86-lOO%), and a small percentage of cells (3-15%) were positive for GFAP, an astrocyte cell marker (Table 2).
no. 5
and culture
microelia
in
in vitro culture Acid phosphatase (% + ve)
GALC
34 19 ND ND ND 69
0 0 0 ND ND 0
10 15 ND 0 3 9
100 100 ND 100 86 ND
82 78 ND 87 82 83
87 80 ND 92 85 85
esterase.
i”““:h:i”:I
A: Control (MslgGBb)
F: A2B5 (*ve cells= 58%)
64
36
128
160 192 22425
I
32
64
Fluorescence
96
I28
160
192 2242561
i.
32
64
96
128
160
192 224 25
Intensity
Fig. 3. Flow cytometric fluorescence histograms of mechanically dissociated human fetal brain cells. Cells were analysed for expression of blood monocyte surface marker LeuM3 (B), lymphocyte surface marker OKT3 (D) and glial cell precursor surface marker A2B5 (F). Cells were also stained with appropriate control antibodies (A,C,E). Fluorescence intensity is plotted on the abscissa and cell number (10,000 cells analysed) on the ordinate.
__
-.,--
--~-
_
-
Fig.’ 4. in U&O &iture of primary &ssoc~a&d human fetal brain cetls. Calls were mkintained in in tlifro ctdture for 3 days in DMEM FCS in the absence (A) and the Presence (Ij) of 10% conditioned medium from human fetal astrocyte culture supernatant. Spindle-shaped micro&a-like cells appear above the astrwyte 041 layer after 17-Z days in culture (C) After 5 7 day% ~41~ lwome more rounded. refractile and loosely adherent to astrocyte cell layer (D). Phase contrast micro=oPy. X 5@0.
Fig. 5. Cell analysis of purified human fetal brain microglia. Cytospin preparations showing microglia stained positive for non-specific esterase (A) and acid phosphatase (B). Light microscopy. x 1200. Cells were plated on polylysine-coated glass coverslips and cell surface antigen expression were characterized. Microglia were stained for macrophage cell surface antigen CD1 lb and Class II HLA-DR using anti-human C3bi (C) and 12 (D) monoclonal antibodies, respectively. Indirect fluorescence microscopy. x 1200.
1%
N. F. HASAN et ul.
Fig. h.
Human fetal brain-microglia: isolation and culture reduction
(Fig. 6C). However,
acid phosphatase
they did not stain for
or GFAP (data not shown). DISCUSSION
Brain microglia are considered to be the components of the reticuloendothelial system present within the CNS. Although increases in the number of microglia have been observed in several CNS diseases, including HIV encephalopathy, multiple sclerosis and Alzheimer’s disease and may be important in the immunopathogenesis of these conditions, their cell of origin is controversial. Microglia share several functional similarities with mononuclear phagocytes including the blood peripheral monocyte/macrophage. In this study, we have successfully isolated brain microglia from 14-18-week-old human fetal brains using the McCarthy and de Vellis technique and have investigated the characteristics of these cells (Fig. 1). Mechanically dissociated human fetal brain cells are mostly neuroglial cell precursors (Fig. 2A), which stain positive for A2B5 monoclonal antibody (Fig. 3F). A2B5 is expressed on neuronal cells and neuroglial precursor cell~.~ Moreover, cell surface staining of these cells using CD14 or CD3 monoclonal antibodies was not detectable (Fig. 3B,D). The absence of monocyte/macrophages or CDlCpositive cells in the primary dissociated brain cells virtually exclude the possibility that the microglia, which are observed in culture (3-4 weeks), are derived from proliferation of contaminating peripheral blood monocyte/macrophages. The possibility, however, that the microglia are derived from differentiating stem cells is not excluded by these experiments. Moreover, the small percentage of cells seen in primary dissociated brain cells that stained positive for non-specific esterase and acid phosphatase (Table 1) may represent the microglia in fetal brain tissue. The modification of the McCarthy and de Vellis technique with the addition of astrocyte culture conditioned medium to the dissociated brain cultures led to significant enhancement of cell spreading and number (Fig. 4A,B). Astrocytes are known to secrete several glial peptides including an interleukin 3-like factor (multi-colony stimulating factor) which enhance the growth and proliferation of rodent brain microglia and also peritoneal macrophages.4 The detailed characterization of the cell population by indirect immunofluorescence microscopy and
157
immunoperoxidase histochemical staining indicates that an enriched population of microglia is obtained. Oligodendrocytes, the myelin secreting cells, were not detected in our preparations; however, a small population of astrocytes (O-15%) was present (Table 2). Although the complement receptor C3bi (Mac 1) is expressed variably on the cells (16-74%; Table 2), the majority of the microglia stain positive for RCA-l (86-lOO%), a specific brain microglia marker which binds to galactose residues present on microglia cell membrane.16 The expression of the surface antigen C3bi may be down-regulated after separation of cells from astrocyte feeder layer or there may be two populations of microglia, one of which lacks the expression of C3bi antigen. Enriched microglia express Class II antigen (HLA-DR; lo-53%), a cell surface antigen that may play an important role in antigen presentation in the CNS and which can be up-regulated by interferon gamma (Table 2), as has been previously shown for mouse brain microglia.” Class II antigen (HLA-DR) is also expressed on brain astrocytes3 Human brain microglia also share several histochemical characteristics with the peripheral blood monocyte/macrophage. Microglia stained positive for non-specific esterase and acid phosphatase and generated superoxide anion upon activation with PMA using the NBT reduction slide test (Table 2). Primary dissociated human fetal brain cultures are maintained in in vitro culture for more than 28 weeks. Cells with microglia-like morphology are seen growing under an astrocyte cell layer, and stained positive for non-specific esterase and generated superoxide anion (Fig. 6B,C). These cells may represent migrating microglia from the upper surface of an astrocyte cell layer or newly developing cells. However, since these cells lack intracellular acid phosphatase (data not shown) it is possible that these cells may represent newly developing microglia rather than migrating cells. Further studies of the relationship between human microglia and mononuclear phagocytes require an understanding of the role of growth factors on proliferation and differentiation of glial cells, which will allow the development of long-term cultures with adequate cell numbers for investigations. Acknowledgements-This work was supported in part by grants from the National Multiple Sclerosis Society, Grant RG 1919-A- 1, Biomedical Research Support Institutional Grant from the National Institutes of Health RR-05506 and W. W. Smith Charitable Trust A-4089.
Fig. 6. Long-term in vitro culture of primary dissociated human fetal brain cells. Microglia-like cells were observed in primary dissociated human fetal brain cultures below the astrocyte cell layer after 12 weeks. (A) Cells are highly branched with large nuclei and thin cytoplasm. Phase contrast microscopy. x 500. (B) Cells stained positive for non-specific esterase. Light microscopy. x 500. (C) These cells generated superoxide anion upon activation with PMA, as assessed with NBT reduction to dark blue intracellular insoluble formazan precipitate. Light microscopy. x 1200.
158
N. F. HASSAN YI al. REFERENCES
1. Del Rio-Hortega P. (1932) Microglia. In Penj?eld’s Cytology and Cellular Pathology oj’the Nervous Sysfem (ed. Penfield W.), pp. 483-534. Harbert and Row, New York. and glial surface antigens on cells in culture. In Cell Culture in the Neurmciences 2. Fields K. (1985) Neuronal (eds Bottenstein J. E. and Sato G.), pp. 45-93. Plenum Press, New York. 3. Fontana A., Fierz W. and Wekerle H. (1984) Astrocytes present myelin basic protein to encephalitogenic T-cell line. Nature 301, 273-276. 4. Frei K., Bodmer S., Schwerdel C. and Fontana A. (1986) Astrocyte-derived interleukin 3 as a growth factor for microglia cells and peritoneal macrophages. J. Immun. 137, 3521-3527. 5. Giulian D. (1987) Ameboid microglia as effecters of inflammation in the central nervous system. J. Neurosci Res. 18, 155-171. 6. Guilian D. and Baker T. J. (1986) Characterization of ameboid microglia isolated from the developing mammahan brain. J. Neurosci. 6, 2163-2178. I. Kitamura T., Miyake T. and Fujita S. (1984) Genesis of resting microglia in gray matter of mouse hippocampus. J. camp. Neurol. 226, 421-433. 8. Ling E. A. (1981) The origin and nature of microglia. In Advances in Cellular Neurobiology (eds Federoff S. and Hertz L.), Vol. 2, pp. 33-82. Academic Press, London. 9. Mannoji H., Yeger H. and Becker L. E. (1986) A specific histochemical marker (lectin Ricinus communis agglutinin-I) for normal human microglia and application to routine histopathology. Acta nkropath. 71, 341-343. -10. McCarthv K. D. and de Vellis J. (1980) Preparation of seoarate astronlial and olieodendroelial cell cultures from rat cerebral iissue. J. Cell Biol. 85, 840-9Ifl2. L 11. Murabe Y. and Sano Y. (1983) Morphological studies on neuroglia. VII. Distribution of brain macrophages in brains of neonatal and adult rats, as determined by means of immunohistochemistry. Cell Tirs. Res. 229, 85 -95. 12. Oehmichen M. (1983) Inflammatory cells in the central nervous system. Prog. Neuropath. 5, 277 335. 13. Oehmichen M. (1982) Functional properties of microglia. In Recent Advances in Neuropathology (eds Smith W. T. and Cavanagh J. B.), pp. 833107. Churchill Livingstone, Edinburgh. 14. Oehmichen M., Wietholter H. and Greaves H. F. (1979) Immunological analysis of human microglia: lack of moncytic and lymphoid membrane differentiation antigens. J. Neuropath. exp. Neurol. 38, 99-103. 15. Perry V. H., Hume D. A. and Gordon S. (1985) Immunohistochemical localization of macrophages and microglia m the adult and developing mouse brain. Neuroscience 15, 313-326. 16. Suzuki H., Franz H., Yamamoto T., Iwasaki Y. and Konno H. (1988) Identification of the normal microgha population in human and rodent nervous tissue using lectin-histochemistry. Neuropath. appl. Neurobiol. 14, 221 227. 17. Suzumura A., Mezitis S. G. E., Gonatas N. K. and Silberberg D. H. (1987) MHC antigen expression on bulk isolated macrophage-microglia from newborn mouse brain: induction of Ia antigen expression by gamma interferon. J. Neuroimmun. 15, 263-278. 18. Volkman D. J., Buescher E. S., Gallin J. I. and Fauci A. S. (1984) B cell lines as a model for inherited phagocytic diseases: abnormal supcroxide generation in chronic granulomatous disease and giant granules in Chediak-Higashi syndrome. J. Immun. 133, 3006-3009. (Accepted 5 September
1990)
Neuroscience
Vol. 41, No. I, pp. 1499158,1991 Printed in Great Britain
ISOLATION AND CHARACTERIZATION OF HUMAN FETAL BRAIN-DERIVED MICROGLIA IN IN FrlTRO CULTURE N. F. HASSAN, D. E. CAMPBELL, S. RIFAT and S. D. DOUGLAS* Division of Infectious Diseases and Immunology, Clinical Immunology Laboratory and The Joseph Stokes Jr. Research Institute, Children’s Hospital of Philadelphia, Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, U.S.A. Abstract-Human brain microglia may play a central role in immunopathogenesis of CNS diseases including HIV infection, multiple sclerosis and Alzheimer’s disease. In order to investigate the possible relationship between microglia and the mononuclear phagocyte system, human brain microglia were isolated from 14-18-week-old fetal brains, and maintained in in vitro culture. Enriched fetal brain microglia were stained for different monocyte/macrophage and glial cell markers. Fresh dissociated brain cells lacked macrophage surface markers. Isolated microglial cells stained positive for complement receptor C3bi, Class II buman leukocyte antigen-DR (HLA-DR)] antigen and with the lectin Ricinus communis. Microglia also share several functional properties with monocyte/macrophages, which include generation of superoxide anion and histochemically demonstrable intracellular acid phosphatase and non-specific esterase. Primary human dissociated brain cultures were maintained in culture for at least 28 weeks. Although microglia were not observed above the astrocyte cell layer after 5 weeks in culture, micoglia-like cells appear below the astrocyte layer after 12 weeks in culture. These cells stained positive for non-specific esterase and displayed oxidative burst activity upon activation with phorbol myristate acetate. Thus, we have successfully isolated an enriched population of microglia from human fetal brain and have demonstrated that these cells possess markers and properties which are characteristics of mononuclear phagocytes.
Brain macrophages or micro&a have been the subject of considerable controversy since their initial description by Del Rio-Hortega using a weak silver carbonate stain, 50 years ago.’ Del Rio-Hortega proposed that
microglia were intrinsic cells of the brain, of mesodermal origin, distributed throughout the CNS, and found in two predominant forms: the ameboid, or macrophage-like, cells which occurred in developing brain and at sites of injury, and the ramified, or highly branched cells, considered to be like “quiescent” cells in mature CNS.” The cell of origin of the brain microglia is very controversial. Some investigators believe that microglia are derived from migrating blood monocytes during the neonatal period,8,” whereas others maintain that the microglia are of neuroectodermal origin.7~‘3Several studies have been performed to identify microglia in situ in undamaged human brain tissue. These cells were non-reactive using anti-monocyte or anti-B,-microglobulin antibodies raised against peripheral blood monocytes,i4 and therefore it was concluded that the resting human microglia are likely to be of neural origin since they *To whom correspondence should be addressed. Abbreviations: DMEM, Dulbecco’s Eagle medium; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; GALC, galactocerebroside; GFAP, glial fibrillary acidic protein; HBSS, Hank’s balanced salt solution; NBT, nitroblue tetrazolium; PBS, phosphate-buffered saline; PMA, phorbol myristate acetate; RCA-l, Ricks communis agglutinin-1 .
do not express known mononuclear phagocyte cell markersi It was further postulated that macrophages which occur in inflammatory diseases of the CNS are blood monocytes migrating in response to chemotactic stimuli from damaged neuronal cells, rather than microglia cell~.~ Recently, lectin histochemistry has provided probes for identification of microglia. Ricinus communis agglutinin-1 (RCA-l), which binds specifically to galactose residues on the brain microglia cell surface, has been used to demonstrate the presence of microglia in human nervous tissues.‘6 RCA-l-positive microglia were observed throughout the CNS, more frequently in white than gray matter.’ Although several studies have attempted to characterize brain microglia in brain slices, the cellular heterogeneity of the CNS and the presence of contaminating blood monocytes in choroid plexus and blood vessels in brain sections have led to conflicting results concerning the origin and characteristics of brain microglia cell~.‘~J~ In an effort to circumvent this problem, several in vitro techniques were developed to isolate and purify brain microglia in in oitro culture. The McCarthy and de Vellis isolation technique has been widely applied to obtain highly purified populations of astrocytes and oligodendrocytes,‘” and has also been used to obtain enriched microglia preparations in in vitro culture derived from rodent brains.6 In this study, we describe the isolation of human fetal brain microglia using a modification of the McCarthy and de Vellis isolation technique. Partially 149
150
N. F. HASSAN et al.
purified cultured human brain microglia obtained from fetal brain cell dissociation cultures were characterized using different human monocyte/macrophage markers. EXPERIMENTAL
PROCEDURES
Materials Human fetal brain materials from elective abortions (1418 weeks) were provided by the International Institute for the Advancement of Medicine (Havertown, PA). Penicillin, streptomycin, 2.5% trypsin, L-glutamine and non-essential amino acids were obtained from Gibco Laboratories, Grand Island, NY. Deoxyribonuclease, insulin, non-specific esterase and acid phosphatase histochemical staining kits were obtained from Sigma Chemicals, St Louis, MO. Petri dishes (60 mm) were purchased from NUNC, Naperville, IL. Nylon mesh and Leukostat stain were obtained from Fisher Scientific, Fair Lawn, NJ. Antibodies used in this study include: mouse monoclonal antibodies to CD3, CD14, CD1 lb, A2B5 and anti-human leucocyte antigen-DR (HLA-DR), and rabbit polyclonal antibodies anti-galactocerebroside and anti-glial fibrillary acidic protein. Monoclonal antibodies anti-human Leu-M3 (CD14) (Becton and Dickinson, Mountain View, CA), OK T3 (CD3) (Ortho Diagnostic Systems, Inc, Raritan, NJ), A2B5, hybridoma supernatant (American Tissue Culture Collection, Bethesda, MD), mouse anti-human C3bi (CDllb) (provided by Dr Donald Anderson, Baylor College of Medicine, Houston, TX), mouse anti-human HLA-DR (12) (Coulter, Hialeah, FL), rabbit anti-galactocerebroside (GALC) (Chemicon, El Segundo, CA) and rabbit anti-glial fibrillary acidic protein (GFAP) (Advanced Immunochemical Services Inc., Los Angeles, CA) were used in these studies. Fluorescein isothiocyanate (FITC)-conjugated goat F(ab’), anti-mouse immunoglobulin was obtained from Tago, Burlingame, CA. Flow cytometric analyses were performed using an EPICS C (Coulter, Hialeah, FL). Positive antibody staining was detected using a Zeiss epifluorescence microscope equipped with 35 mm camera and an HBO 100 W high-pressure mercury lamp. Immunoperoxidase histochemical staining FETAL HUMAN (14-15 week
was performed using a Vectastain ABC kit (Vector Laboratories, Burlingame, CA). Human recombinant gamma interferon was kindly provided by Dr Peter Sorter, Hoffman La Roche, Nutley, NJ. In vitro cells
culrure
of primary
dissociated
human
fetal
Human fetal brain materials were processed following elective abortions. Figure I summarizes the technique used for the isolation of human fetal brain microglia. Brain tissues were rinsed several times with cold Hank’s balanced salt solution (HBSS) supplemented with 50 U/ml penicillin and 50 pg/ml streptomycin, and meninges with attached blood vessels were macroscopically dissected from brain tissues. Following dissection, the brain tissues were sliced into small pieces and transferred to 60 mm sterile Petri dishes on ice and mechanical dissociation was carried out using sterile surgical scalpels. Prior to enzymatic treatment, 500 ~1 aliquots of dissociated cells were removed for cell characterization including flow cytometric analysis and enzyme histochemistry. Cold HBSS was then added to bring the cell volume to 16 ml in each Petri dish and 4ml of 2.5% trypsin and deoxyribonuclease (10 pg/ml) were then added to each dish. The cell mixture was shaken for 15 min at 200 r.p.m., 37°C and then centrifuged three times for 10 min at 1200 r.p.m., 4°C. The cell suspension was then filtered through a sterile nylon mesh with a pore size of 200 pm. The filtered cells were suspended in culture medium, counted and seeded into 75 cm* flasks at a cell concentration of 108-lo9 cells/ml for continuous culture in a humidified atmosphere of 95% air and 5% CO, at 37°C. The culture medium consists of Dulbecco’s Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS), 50 U/ml penicillin, 50 pg/ml streptomycin, 2.5 pg/ml fungizone, 4 mM L-glutamine, 100 PM non-essential amino acids and 5 pg/ml insulin. Five days after initial inoculation the flasks were rinsed several times with warm DMEM to remove non-adherent cells and fresh culture medium supplemented with 10% conditioned medium from human astrocyte culture was added. The medium from primary dissociated human brain cultures was changed every 5 days and brain cells were maintained in long-term culture for 12-28 weeks.
BRAIN old)
Long term (up to 2:
Mechanical dissociation
-
Microglia analysis
--)
Microglia analysis
culture weeks)
Orbital ‘shaking (Detachment of microglial I
/ Centrifugation (1200rpm. 10 min
x 31
a
_
brain
Culture (4-5 weeks1 Filtration
Fig. 1. Isolation and culture of human fetal brain microglia in in vifro culture. performed in freshly dissociated fetal brain cells, purified microglia after short-term and microglia in long-term culture (12-28 weeks).
Cell analyses were culture (4-5 weeks)
Human Preparation of human (conditioned medium)
astrocyte
fetal brain-microglia:
culture
supernatant
Human primary brain dissociated cultures maintained for 3-4 weeks were shaken for 16 h at 200 r.p.m., 37°C. Following overnight shaking of the cultures, loosely adherent and detached cells were removed and cultures were trypsinized and subcultures were prepared. The cells were maintained in culture medium supplemented with 10% FCS, and the culture medium was changed every 5 days. Supernatants of confluent astrocyte cultures were harvested by centrifugation (2OOOg, 10 min), aliquoted and stored at -20°C until used. Characterization offreshly dissociated human fetal brain cells Cell surface antigen analysis. Mechanically dissociated human fetal brain cells werk stained with the monoclonal antibodies anti-human Leu-M3. OK T3 and A2B5, hybridoma culture supernatant. The cells were incubated with the monoclonal antibodies for 30 min at 4”C, washed in phosphate-buffered saline (PBS) containing 0.01% sodium azide and incubated with 1: 20 dilution of FITC-conjugated goat F(ab’)? anti-mouse immunoglobulin for 30 min. Flow cytometric analyses were performed using a fluorescence activated cell sorter, EPICS C (Coulter). Forward and right angle light scatter were used to establish appropriate gates on the cells by excluding non-viable cells. The fluorescence intensity distribution of 5000 cells was accumulated for analysis. A threshold of fluorescence was set for each experiment and the background from control cells labeled with isotype-matched mouse immunoglobulin was subtracted from the test antibody-labeled preparations. Histochemical staining. Cytospins of freshly dissociated human fetal brain cells were stained with Leukostat stain, and tested for non-specific esterase and acid phosphatase using histochemical staining kits. For each preparation 600 cells were counted under oil immersion and cells containing dark cytoplasmic granules were scored as positive. isolation of human fetal brain-derived microglia. After 7-10 days, a confluent layer of fibroblast-like cells was observed in the primary human brain cultures. Four to five weeks later, rounded and spindle-shaped phase-bright cells appear in small colonies above the confluent monolayer. The cultures were then shaken in an orbital shaker for 16 h at 150 r.p.m., 37°C. The supematant media containing the detached cells were centrifuged for 10 min at 1200 r.p.m., 4°C resuspended, counted and plated on poly+-lysinecoated coverslips at a density of 5-10 x lo4 cells/coverslip in culture medium. After 2 h, coverslips were washed with DMEM and stained for different intracellular and surface markers. Characterization ofpuriJied humanfetal brain derived microglia in in vitro culture Ceil surface antigen analysis. Isolated human brain microglia were stained for cell surface and intracytoplasmic antigens using different human monocyte/macrophage markers. Indirect immunofluorescence microscopy and the immunopcroxidase technique were applied to study the expression of different markers. The following antibodies were used: mouse anti-human C3bi (1: 100 dilution), mouse anti-human HLA-DR (12) (I:2 dilution), rabbit anti-GALC (I:50 dilution) and rabbit anti-GFAP (1:40 dilution). Rabbit immunoglobulin or isotype-matched mouse immunoglobulin were used as a control antibody for each of the reagents mentioned above. Positive antibody staining was detected using fluorescence microscopy or immunoperoxidase histochemical staining. Fluorescence microscopy was performed with a Zeiss epifluorescence microscope equipped with a 35 mm camera. For each prepartion 200 cells were counted and positive cells were scored in each case.
Immunoperoxidase histochemical staining was performed using a Vectastain ABC kit. Microglia were also stained with biotinylated lectin RCA-l and visualized with immunoperoxidase histochemical staining. In order to study the
isolation
151
and culture
of HLA-DR on microglia, cells were treated with 1000 U/ml human recombinant gamma interferon for 18 h and stained with mouSe anti-human HLA-DR monoclonal antibody (12). Histochemical staining. Cytospins of isolated human fetal brain microglia were tested for non-specific esterase and acid phosphatase using histochemical staining kits. Cells were regulation
also assessed for superoxide anion generation using the nitroblue tetrazolium (NBT) slide test.” Cells were incubated
with NBT dye and phorbol myristate acetate (PMA) (1 pg/mI) as stimulant for 90 min at 5% CO,, 37°C. Coverslips were washed several times and fixed in methanol for 20min and counterstained with safranin. For each preparation 200 cells were counted under oil immersion and cells with dark blue formazan granules were scored as positive.
Characterization of primary dissociated human fetal brain cells in long-term culture Human brain cells in long-term culture were assessed for superoxide generation by the NBT slide test and tested for intracellular non-specific esterase using hi&chemical staining kits as described above.
RESULTS
Histochemistry and cell surface antigen analysis primary dissociated human fetal brain cells
of
Histocytochemical staining of dissociated human fetal brain cell suspensions using Leukostat (Modified Wright’s stain) reveal the presence of neuroblast-like cells 20-40 pm in diameter with large nuclei and thin cytoplasm (Fig. 2A). A small percentage of the cells stained positive for non-specific esterase and acid phosphatase (Fig. 2B,C). The percentage of cells stained positive for acid phosphatase (seven experiments) was higher than cells stained positive for non-specific esterase (1.4-2.8 % and 0.7-2.5 %, respectively). In only one experiment the cells positive for non-specific esterase scored higher than cells with acid phosphatase (3.0 and 1.4%, respectively; Table 1). Flow cytometric analysis was carried out in order to determine whether the non-specific esterase and acid phosphatase-positive cells were peripheral blood monocytes contaminating the brain cell suspensions. Human fetal brain cell suspensions derived from 14-18-week-old fetuses did not express either CD14 or CD3 using the monocyte/macrophage marker Leu-M3 and T-lymphocyte surface marker 0KT3 monoclonal antibodies, respectively (Fig. 3). However, at least 50% of mechanically dissociated fetal brain cells stained positive for A2B5, a surface antigen expressed on neuroglial precursor cells (Fig. 3). Moreover, surface staining for other cell antigens including CDllb, CD4 and Class II (HLA-DR) was negative (data not shown). Isolation and characterization brain -derived microglia
of isolated human fetal
Primary dissociated fetal brain cells were plated and maintained in in vitro culture in the presence of 10% conditioned medium from human fetal astrocyte culture. Increased number of cells and cell spreading were observed in cultures incubated with conditioned medium (Fig. 4A,B). After 3-4 weeks, cells 15-25 pm
152
N.
F.
HASAN
rt ul
Fig. 2. Cytospin preparations of primary dissociated human fetal brain cells stained with Leukostat stain (modified Wright’s stain) (A), non-specific esterase (arrow) (B) and acid phosphatase (arrowj (C). Light microscopy. x 900.
in di,ameter with fine plasma membrane processes and vact rolated cytoplasm were observed growing on the top of a confluent layer (bed layer) of fibroblast-like cells (Fig. 4C) that consisted primarily of astrocytes
(data not shown). With increased time in culture these cells become more rounded and loosely attach1 ed to the bed layer (Fig. 4D) and shaking of the cultu tes is usually Performed at this time (4-5 weeks).
Human Table human
1. Histochemical fetal brain cells
fetal
Non-specific
esterase
isolation
staining of primary dissociated for non-specific esterase and acid phosphatase* Experiment
Acid phosphatase
brain-microglia:
1
2
3
4
0.7
1.7
1.1
1.4
1.6 2.8
3
1.4 2
*Results are expressed as percentage seven different experiments.
6
7
2.5
1
1.4 4.1 2.5
positive
cells
from
Characterization of primary dissociated brain cells in long-term culture
Histochemistry Primary dissociated fetal brain cells were shaken after 4-5 weeks in in vitro culture. Brain microglia recovered from six different experiments gave yields of 0.993.6 x lo6 cells with 54-90% viability (Table 2). The enriched microglia stained positive for nonspecific esterase and acid phosphatase (Fig. 5A,B and Table 2) and generated superoxide anion upon activation with PMA using the NBT slide test (15-49%).
Enriched brain microglia were stained for surface and intracellular monocyte/macrophage and glial cell markers. Using the indirect immunofluorescence technique, cells stained variably for C3bi receptor (1674%; Fig. 5C) and HLA-DR (12) antigen (lo-53%)
Exp.
Cell no. ( x 106)
no. 1 2 3 4 5 6
Viability (%I
0.9 1.1 3.6 1.3 1.2 1.2
*Non-specific
2. Cell characterization
67 82 62 54 90 83
of human
Anti-C3bi (% + ve)
fetal
brain
Ia + INFy
71 16 74 ND ND 60
10 12 ND ND ND 53
human fetal
Primary dissociated human fetal brain cultures were shaken 3-4 weeks after culture and were then maintained in vitro for more than 28 weeks. Cultures were provided with fresh media every 5 days and maintained at 37°C and 5% CO,. Following culture shaking, microglia were not seen growing above the astrocyte cell layer. However, after 12 weeks, groups of cells were observed under the astrocyte cell layer in all brain cultures. These cells were present in small clusters of 7-10 cells heterogeneous in size, highly branched with large nuclei and thin cytoplasm (Fig. 6A). They stained positive for non-specific esterase (Fig. 6B) and generated superoxide anion upon activation with PMA as assessed by NBT
Cell surface antigen expression
Table
153
and did not stain for GALC, an oligodendrocyte cell marker (Table 2). Moreover, the addition of 1000 U/ ml human recombinant gamma interferon to brain microglia cultures for 18 h increased the expression of HLA-DR (12) cell surface antigen (19-69%; Fig. 5D and Table 2). Using immunoperoxidase histochemical staining, human fetal brain microglia stained positive for RCA-l (86-lOO%), and a small percentage of cells (3-15%) were positive for GFAP, an astrocyte cell marker (Table 2).
no. 5
and culture
microelia
in
in vitro culture Acid phosphatase (% + ve)
GALC
34 19 ND ND ND 69
0 0 0 ND ND 0
10 15 ND 0 3 9
100 100 ND 100 86 ND
82 78 ND 87 82 83
87 80 ND 92 85 85
esterase.
i”““:h:i”:I
A: Control (MslgGBb)
F: A2B5 (*ve cells= 58%)
64
36
128
160 192 22425
I
32
64
Fluorescence
96
I28
160
192 2242561
i.
32
64
96
128
160
192 224 25
Intensity
Fig. 3. Flow cytometric fluorescence histograms of mechanically dissociated human fetal brain cells. Cells were analysed for expression of blood monocyte surface marker LeuM3 (B), lymphocyte surface marker OKT3 (D) and glial cell precursor surface marker A2B5 (F). Cells were also stained with appropriate control antibodies (A,C,E). Fluorescence intensity is plotted on the abscissa and cell number (10,000 cells analysed) on the ordinate.
__
-.,--
--~-
_
-
Fig.’ 4. in U&O &iture of primary &ssoc~a&d human fetal brain cetls. Calls were mkintained in in tlifro ctdture for 3 days in DMEM FCS in the absence (A) and the Presence (Ij) of 10% conditioned medium from human fetal astrocyte culture supernatant. Spindle-shaped micro&a-like cells appear above the astrwyte 041 layer after 17-Z days in culture (C) After 5 7 day% ~41~ lwome more rounded. refractile and loosely adherent to astrocyte cell layer (D). Phase contrast micro=oPy. X 5@0.
Fig. 5. Cell analysis of purified human fetal brain microglia. Cytospin preparations showing microglia stained positive for non-specific esterase (A) and acid phosphatase (B). Light microscopy. x 1200. Cells were plated on polylysine-coated glass coverslips and cell surface antigen expression were characterized. Microglia were stained for macrophage cell surface antigen CD1 lb and Class II HLA-DR using anti-human C3bi (C) and 12 (D) monoclonal antibodies, respectively. Indirect fluorescence microscopy. x 1200.
1%
N. F. HASAN et ul.
Fig. h.
Human fetal brain-microglia: isolation and culture reduction
(Fig. 6C). However,
acid phosphatase
they did not stain for
or GFAP (data not shown). DISCUSSION
Brain microglia are considered to be the components of the reticuloendothelial system present within the CNS. Although increases in the number of microglia have been observed in several CNS diseases, including HIV encephalopathy, multiple sclerosis and Alzheimer’s disease and may be important in the immunopathogenesis of these conditions, their cell of origin is controversial. Microglia share several functional similarities with mononuclear phagocytes including the blood peripheral monocyte/macrophage. In this study, we have successfully isolated brain microglia from 14-18-week-old human fetal brains using the McCarthy and de Vellis technique and have investigated the characteristics of these cells (Fig. 1). Mechanically dissociated human fetal brain cells are mostly neuroglial cell precursors (Fig. 2A), which stain positive for A2B5 monoclonal antibody (Fig. 3F). A2B5 is expressed on neuronal cells and neuroglial precursor cell~.~ Moreover, cell surface staining of these cells using CD14 or CD3 monoclonal antibodies was not detectable (Fig. 3B,D). The absence of monocyte/macrophages or CDlCpositive cells in the primary dissociated brain cells virtually exclude the possibility that the microglia, which are observed in culture (3-4 weeks), are derived from proliferation of contaminating peripheral blood monocyte/macrophages. The possibility, however, that the microglia are derived from differentiating stem cells is not excluded by these experiments. Moreover, the small percentage of cells seen in primary dissociated brain cells that stained positive for non-specific esterase and acid phosphatase (Table 1) may represent the microglia in fetal brain tissue. The modification of the McCarthy and de Vellis technique with the addition of astrocyte culture conditioned medium to the dissociated brain cultures led to significant enhancement of cell spreading and number (Fig. 4A,B). Astrocytes are known to secrete several glial peptides including an interleukin 3-like factor (multi-colony stimulating factor) which enhance the growth and proliferation of rodent brain microglia and also peritoneal macrophages.4 The detailed characterization of the cell population by indirect immunofluorescence microscopy and
157
immunoperoxidase histochemical staining indicates that an enriched population of microglia is obtained. Oligodendrocytes, the myelin secreting cells, were not detected in our preparations; however, a small population of astrocytes (O-15%) was present (Table 2). Although the complement receptor C3bi (Mac 1) is expressed variably on the cells (16-74%; Table 2), the majority of the microglia stain positive for RCA-l (86-lOO%), a specific brain microglia marker which binds to galactose residues present on microglia cell membrane.16 The expression of the surface antigen C3bi may be down-regulated after separation of cells from astrocyte feeder layer or there may be two populations of microglia, one of which lacks the expression of C3bi antigen. Enriched microglia express Class II antigen (HLA-DR; lo-53%), a cell surface antigen that may play an important role in antigen presentation in the CNS and which can be up-regulated by interferon gamma (Table 2), as has been previously shown for mouse brain microglia.” Class II antigen (HLA-DR) is also expressed on brain astrocytes3 Human brain microglia also share several histochemical characteristics with the peripheral blood monocyte/macrophage. Microglia stained positive for non-specific esterase and acid phosphatase and generated superoxide anion upon activation with PMA using the NBT reduction slide test (Table 2). Primary dissociated human fetal brain cultures are maintained in in vitro culture for more than 28 weeks. Cells with microglia-like morphology are seen growing under an astrocyte cell layer, and stained positive for non-specific esterase and generated superoxide anion (Fig. 6B,C). These cells may represent migrating microglia from the upper surface of an astrocyte cell layer or newly developing cells. However, since these cells lack intracellular acid phosphatase (data not shown) it is possible that these cells may represent newly developing microglia rather than migrating cells. Further studies of the relationship between human microglia and mononuclear phagocytes require an understanding of the role of growth factors on proliferation and differentiation of glial cells, which will allow the development of long-term cultures with adequate cell numbers for investigations. Acknowledgements-This work was supported in part by grants from the National Multiple Sclerosis Society, Grant RG 1919-A- 1, Biomedical Research Support Institutional Grant from the National Institutes of Health RR-05506 and W. W. Smith Charitable Trust A-4089.
Fig. 6. Long-term in vitro culture of primary dissociated human fetal brain cells. Microglia-like cells were observed in primary dissociated human fetal brain cultures below the astrocyte cell layer after 12 weeks. (A) Cells are highly branched with large nuclei and thin cytoplasm. Phase contrast microscopy. x 500. (B) Cells stained positive for non-specific esterase. Light microscopy. x 500. (C) These cells generated superoxide anion upon activation with PMA, as assessed with NBT reduction to dark blue intracellular insoluble formazan precipitate. Light microscopy. x 1200.
158
N. F. HASSAN YI al. REFERENCES
1. Del Rio-Hortega P. (1932) Microglia. In Penj?eld’s Cytology and Cellular Pathology oj’the Nervous Sysfem (ed. Penfield W.), pp. 483-534. Harbert and Row, New York. and glial surface antigens on cells in culture. In Cell Culture in the Neurmciences 2. Fields K. (1985) Neuronal (eds Bottenstein J. E. and Sato G.), pp. 45-93. Plenum Press, New York. 3. Fontana A., Fierz W. and Wekerle H. (1984) Astrocytes present myelin basic protein to encephalitogenic T-cell line. Nature 301, 273-276. 4. Frei K., Bodmer S., Schwerdel C. and Fontana A. (1986) Astrocyte-derived interleukin 3 as a growth factor for microglia cells and peritoneal macrophages. J. Immun. 137, 3521-3527. 5. Giulian D. (1987) Ameboid microglia as effecters of inflammation in the central nervous system. J. Neurosci Res. 18, 155-171. 6. Guilian D. and Baker T. J. (1986) Characterization of ameboid microglia isolated from the developing mammahan brain. J. Neurosci. 6, 2163-2178. I. Kitamura T., Miyake T. and Fujita S. (1984) Genesis of resting microglia in gray matter of mouse hippocampus. J. camp. Neurol. 226, 421-433. 8. Ling E. A. (1981) The origin and nature of microglia. In Advances in Cellular Neurobiology (eds Federoff S. and Hertz L.), Vol. 2, pp. 33-82. Academic Press, London. 9. Mannoji H., Yeger H. and Becker L. E. (1986) A specific histochemical marker (lectin Ricinus communis agglutinin-I) for normal human microglia and application to routine histopathology. Acta nkropath. 71, 341-343. -10. McCarthv K. D. and de Vellis J. (1980) Preparation of seoarate astronlial and olieodendroelial cell cultures from rat cerebral iissue. J. Cell Biol. 85, 840-9Ifl2. L 11. Murabe Y. and Sano Y. (1983) Morphological studies on neuroglia. VII. Distribution of brain macrophages in brains of neonatal and adult rats, as determined by means of immunohistochemistry. Cell Tirs. Res. 229, 85 -95. 12. Oehmichen M. (1983) Inflammatory cells in the central nervous system. Prog. Neuropath. 5, 277 335. 13. Oehmichen M. (1982) Functional properties of microglia. In Recent Advances in Neuropathology (eds Smith W. T. and Cavanagh J. B.), pp. 833107. Churchill Livingstone, Edinburgh. 14. Oehmichen M., Wietholter H. and Greaves H. F. (1979) Immunological analysis of human microglia: lack of moncytic and lymphoid membrane differentiation antigens. J. Neuropath. exp. Neurol. 38, 99-103. 15. Perry V. H., Hume D. A. and Gordon S. (1985) Immunohistochemical localization of macrophages and microglia m the adult and developing mouse brain. Neuroscience 15, 313-326. 16. Suzuki H., Franz H., Yamamoto T., Iwasaki Y. and Konno H. (1988) Identification of the normal microgha population in human and rodent nervous tissue using lectin-histochemistry. Neuropath. appl. Neurobiol. 14, 221 227. 17. Suzumura A., Mezitis S. G. E., Gonatas N. K. and Silberberg D. H. (1987) MHC antigen expression on bulk isolated macrophage-microglia from newborn mouse brain: induction of Ia antigen expression by gamma interferon. J. Neuroimmun. 15, 263-278. 18. Volkman D. J., Buescher E. S., Gallin J. I. and Fauci A. S. (1984) B cell lines as a model for inherited phagocytic diseases: abnormal supcroxide generation in chronic granulomatous disease and giant granules in Chediak-Higashi syndrome. J. Immun. 133, 3006-3009. (Accepted 5 September
1990)
