Journal of Neuroscience Kesearch 27:36-42 (1990)
Chemotaxis by a CNS Macrophage, the Microglia J. Yao, L. Harvath, D.L. Gilbert, and C.A. Colton Department of Physiology and Biophysics, Georgetown University School of Medicine, Washington, D.C. (J.Y.. C.A.C.); Division of Blood and Blood Products, Center lor Biologics Evaluation and Rescarch. Food and Drug Administration (L.H.) and Laboratory of Biophysics, NINCDS, National Institutes of Health (D.L.G.), Bethesda, Maryland
Microglia demonstrate many characteristics similar to those seen in rnonocytes and tissue-specific rnacrophages, including phagocytosis, production of oxygen radicals, and growth factors and expression of MHC antigens. We have examined the ability of microglia, cultured from the cerebral cortices of neonatal rats, to demonstrate another important functional characteristic of monocytic-derived cells, that is, chemotaxis. Our results show that cultured rat microglia demonstrate chemotaxis to complement dependent chemoattractants such as recombinant C5a, zymosan activated serum, and to rat serum as well as to transforming growth factor-beta, a chemoattractant produced by platelets. Microglia fail to migrate to bacterial dependent chemoattractants such as the Nformyl peptides. The failure to respond is not dependent on maturational state of the microglia. Treatment with DMSO or casein, agents known to induce morphological and functional changes in cultured microglia reminescent of a “resting” and an “activated” macrophage, respectively, do not alter the response to f’Met-Leu-Phe. In addition, the chemotactic response to serum in DMSO or casein-treated cells is the same as the response seen in untreated day 10 cultured microglia or untreated age-matched controls. The ability of microglia to migrate in response to inflammatory mediators suggests that these cells can move to sites of injury, thereby enabling them to participate in an inflammatory response. Key words: glia, migration, macrophage, transforming growth factor-beta, Met-Leu-Phe INTRODUCTION Jn 1932, del Rio Hortega proposed that microglia represented the reticuloendothelial system in the CNS and, as such. were capable of performing macrophagic functions. This view has been supported by many investigators who have shown that both microglia in the CNS and cultured microglia possess morphological, cyto0 1990 Wiley-Liss, Inc.
chemical. and membrane antigenic similarities to monocytcs and macrophages (Giulian, 1987; Giulian and Baker, 1986; Ling, 1981; Ling et al., 1982; Oehtnichen, 1983; Perry et al., 1985; Perry and Gordon, 1988; Streit et al.. 1988). For example, microglia are reactive for non-specific esterase and express antigens such as the C3b and Fc receptors (Giulian and Baker, 1986; Ling et al., 1982; Perry et al., 1985). In addition, microglia exhibit functional characteristics of macrophages including phagocytosis, sccretion of regulatory factors such as interleukin- 1 (IL- I ) and superoxide anion production (Colton and Gilbert, 1987; Giulian and Baker, 1986; Giulian et al., 1986). Another important function of macrophages is thc ability to migrate to sites of tissue infection and injury. Typically, monocytes and tissue macrophages respond to complement components such as C5a (Hugli. 198 1; Snyderman and Pike, 1984; Yancey et al.. 1985). Complement components, in general, play a key role in the recruitment of effector cells in the microenvironment of the inflammatory rcsponse (Hugli, 1981; Yancey et al., 1985). Other inflammatory mediators also play a role in the chemotaxis of monocytic derived cells and include transforming growth factor-beta (Wahl et al., 1987) and N-formyl peptides (Falk et al., 1982; Harvath and Aksamit, 1989; Marasco et al., 1984; Rot et al., 1987; Schiffman, 1982; Snyderman and Pike, 1984). N-formyl peptides are released by various types of bacteria (Marasco et al., 1984; Rot et al.. 1987). It is not clear, however, whether microglia migrate to these same agents or, if, in fact, they exhibit chernotaxis at all. Studies by Oehmichen (1983) and others (del Rio Hortega, 1932; Perry and Gordon. 1988) imply that microglia can migrate. The purpose of this study was, thus,
Received December 20. 1989; revised March 26, IYYU; accepted April 1.2, 1990. Address reprint requests to C. Colton, Depaittnent of Physiology and Biophysics. Georgetown University School of Medicine, Washington D.C. 20007.
Migration in Cultured Rat Microglia
to determine whether cultured microglia exhibit chemotaxis in rcsponsc to known chcmoattractants.
MATERIALS AND METHODS Microglia Preparation Primary glial cell cultures were obtained from 2day-old rat (Sprague-Dawlcy) ccrebral hemispheres as described previously (Colton and Gilbert, 1987; McCarthy arid de Vellis, 1980; Kaff et al., 1979). Briefly, the cortices were dissected, placed in Leibovitz’s (L- 15) medium, and the meninges removed carefully using a binocular microscope. The cortices were then placed in fresh L-1.5 media, mechanically dissociated by tituration, and plated into 75 cm2 Falcon flasks. Cells were cultured at 37°C using a modified Dulbecco’s Medium (DMEM) with 10% fetal calf serum, 4.5 gil glucose, 3 mM glutainine, and 25 kgiml gentamycin. Thcsc conditions have been shown to be unfavorable for neuronal growth and produce a mixed glial culture (McCarthy and de Vellis, 1980). After growing to conflucncy (about 10 to 12 days), the flasks were placed on a rotary shaker for 15 to 20 hr (250 rpm) at 37°C. This allowed separation of the microglia from the underlying astrocytic layer by using their differential adherdnt properties. The supernatant, enriched in microglia, was centrifuged gently (400 x g) and the cells rcsuspended in Earle’s solution for the migration assay.
37
Alteration of Maturational State In some cases, cultures were grown in the presence of alpha casein (0.5mgiml). dimethyl sulfoxide (DMSO, 0.5% viv) or normal DMEM for an additional 14 days. Cells werc then isolated as described above and used in the cheniotaxis studies.
Microglia Chemotaxis Migration of microglia was studied using a multiwell niicrochcmotaxis assay described by Harvath et al. (1 980) and by Falk et al. ( 1980). This technique has been successfully employed to study chemotaxis in a variety of cell types including monocytes, neutrophils, macrophage cell lines and tumor cell lines (Aksamit et al., 1981; Harvath and Aksainit, 1989: Harvath et al.. 1980; Liotta et al., 1986). Cell suspensions containing 50,000 microglia were added to the upper wells of the chemotaxis chamber (Neuroprobe, Inc.). These cells were separatcd from the lower wells containing chernoattractant solution by a polycarbonate (Polyvinylpyrrolidone free) filter with 5 - p n diameter pores. To promote adhesion of the suspended microglia to the upper surface of the polycarbonate filter, the filters were treated with polyD-lysine (Syg/ml). allowed to dry, and then used immediately in the chemotaxis assay. After incubation at 37°C for 2 hr, nonmigrated cells on thc upper surface of the filter werc wiped off. Cells on the lower surface of the filter were fixed in methanol, Cell Identification stained with Leukostatfr‘a (Fisher) and counted using an Standard histochemical and imrnunocytochemical Optomax System 1V Image Analyzer (Optomax, Inc., techniques were used to identify cells in the microglial Hollis, NH). Data are presented as the average number enriched fraction. Cells from the supernatant fraction of cellsirnm’ of filter surface (k SE) and were obtained were plated onto poly-D-lyaine coated coverslips and by averaging 12 fields ( . 3 to .6 mm’ifield) from each of stained using known niicroglial and astrocytic markers. triplicate wells for a minimum of 3 separate microglial Microglia were identified by their positive reaction for cultures. Total number of cells migrating per well was non-specific esterase as described by Ling et al. ( 1982) found by multiplying the number of cellsimm* by a facand by the presence of membrane receptors for the lec- tor of 8. In this manner, the rlitio of cells migrated to tins, Ricinius cornmiinis agglutinin (Mannoji et al., percenlagc of input cells could be calculated. For poor 1986) or wheat germ agglutinin (Colton, unpublished). migratory responses, cellular debris occassionally inter,4strocytes were identified by their positive reaction for fered with the accuracy of the image analysis. In these GFAP (McCarthy and de Vellis, 1980: Raffet al., 1979). cases, cells were counted using a binocular microscope When prepared as described above, the supernatant con- and an average obtained from 10 random fields (0.021 tained 85-95% microglia, with GFAP positive cells rcp- mm2/field) of each assay group. Significance was deterrcscnting most of the remaining cells. Cultures were neg- mined using an unpaired Students t-test. ative for the presence of neurofilament. Chemotaxis was distinguished from chemokinesis, Cells that migrated to the lower surface of the poly- where chemokinesis was defined as thc increased rancarbonate filter were identified by directly fixing and dom movement of cells due to a general enhancement of staining thc cells on the filter. Due to the properties of the cell activity by the putative chemoattractive agent. In poly-carbonate filters, immunofluorescent staining could these experiments, the same concentration of Lhe putative not be used. Thus, cells were idcntified using nonspe- cheinoattractant was placed in both upper and lower cific esterase only. Migrated cells were 9.5 2 1.2% (n = wells of the chamber, thereby eliminating a chemotactic 3 filters, 100 cells countedifilter) nonspecific esterase gradient. The number of cells migrating under these conpositive. Remaining cells were indistinquishablc. ditions was taken a s a measure of chemokincsis (Falk et
38
Yao et al.
al., 1980; Richards and McCullough, 1984). Random migration was evaluated using Earle's solution as well as Earle's solution with ethanol concentrations equivalent to the ethanol concentration in fMet-Leu-Phe assays.
Peritoneal Macrophages Macrophages from the peritoneal cavity of Sprague-Dawley rats were harvested by injecting ice cold Hank's ringer into the peritoneal cavity. After heveral minutcs the fluid was withdrawn and centrifuged at 400 x g. The pellet containing macrophages was resuspended into Earles, the cells counted and then used immediately in the migration assay. Chemotactic Factors fMet-Leu-Phe and casein were purchased from Sigma Chemical Co. Met-Leu-Phe was prepared in a stock concentration of 10-' M in absolute ethanol and diluted in Earle's solution immediately before use in the chcmotaxis assays. TGF-beta (human) was a gift of Dr. Monique Dubois-Dalq, Lab of Molecular Genetics, NIH. Zymosan activated serum (ZAS) was prepared as described by Metcalf et a]. (1 986) using rat serum. LPS activated serum was prepared as described by Meade et al. (1984). Recombinant human C5a was gcncrously provided by Dr. Henry Showell, Pfizer Drugs, Groton, CT. RESULTS Migration to Complement Components; Serum, ZAS and C5a The migration of day 10 cultured microglia in rcsponse to a chemotactic gradient for normal rat serum is shown in Figure 1 . Microglia responded in a dose-dependent fashion such that thc number of cells migrating increased with increasing concentration of serum. Maximal migration was seen at a serum dilution of 1:10 to 1: 100 (Fig. 1) and represented 2 1.8% of the total number of cells added to the uppcr well. Movement in response to Earles solution only, i.e., random migration, accounted for an average of 108 2 20 cellsimn? or only 1.7% of the total number of cells in the upper well. Microglia also exhibited a chemokinetic response to rat serum. In other words, thc presence of serum, itself, induced an enhanced degree of random, nondirected movement in microglia and accounted for 10% or less of the total input cells. The number of microglia migrating in response to a chemotactic gradient, however, was significantly higher than that sccn moving chemokinetically (p < 0.001). Movement of cells in response to the chernotactic gradient for serum was time dependent and increased over 4 hr, reaching an indistinct plateau at about 2 hr.
I
I I? )',I
11
~;bi,
Fig. 1 . Migration of cultured microglia to rat serum. Data points represent the avcragc number of microgliainim' (2 SEM) that have inigrated in response to a chemotactic gradient for various concentrations of serum (closed circles), in response to Earles solution only (open triangles) and in rcsponse to equal concentrations of serum placed on both sides of the filter (chemokinesis)(open circles); n = at least 35 wells from a minimum of 3 different culture groups. Random movement remained at about 2% of total input cell number over the same timc period; 2 hr was chosen to be the routine length of incubation for the remaining studies. Because serum contains many nonspecific and specific factors that may affect cell movement, further experiments were done to clarify the cheniotactic agent. Migration in response to a nonspecific protein was tested by analyzing the chemotaxis of microglia to bovine serum albumin (BSA). Dilutions of 1:10, 1:100, and I: 1000 of a 1.O% (wtivol) BSAiEarles solution were tested for chcmotactic ability. The number of cells migrating under these conditions were not significantly different from random movement, indicating that BSA does not servc as a chemotactic agent to cultured microglia. Migration to xymosan activated serum was also studied and is shown in Figure 2. Zymosan activation promotes formation of complement cascade components including C5a, a potent chemoattractant (Evarerts ct al., 1985). As indicated, cultured microglia migrated in a dose-dependent manner to a chemotactic gradient for ZAS with maximal migration seen at a dilution of 1 :100. This value represented 18.9% of the total number of cells placed initially in the upper well. To further clarify the role of complement coniponents in microglial chemotaxis, studies were done using recombinant human C5a. As shown in Figure 3 . C5a serves as a chemoattractant for cultured microglia, with maximal migration at 0.9 kgiml representing 17.8% of the input cells responding. C5a also acted as a chemokinetic agent and at 9 pgiml, it was not possible to distinquish directed migration from enhanced random migration.
Migration in Cultured Rat Microglia
39
/
Ni'?MAL
Fig. 2. The effect of zymosan activated serum on microglial chemotaxis. Closed circles-average (2 SEM) numbcr of microglia migratedimin* in response to chemotactic gradients of ZAS. Random migration indicated by the straight line; n = at least 9 wells assayed for a minimum of 3 different culture groups.
CA'j-1
,,
rGEn
nM i(-1
Fig. 4. Changes in microglial chemotaxis with maturational state. Bar height represents the average number of microgliai mm' ? SEM. Open bars-chemotactic response to a 1:10 dilution of rat serum; diagonal striped bars-chemokinetic response to a 1 : l O dilution of rat serum; crosshatched barsmigration to Earle's only. Data obtained from each of 4 treatment protocols, i.e., normal day 10 of culture microglia, casein-treated (0.5 mgiml) microglia, DMSO-treated (0.5% mg/ml) microglia, and age matched controls. Minimum number of wells assayed = 27, for a minimum of 3 different
culture groups.
Fig. 3. The effect of recombinant human C5a on microglial chemotaxis. Data points rcprcscnt the average nuniber of cells/ mm2 ? SEM exhibiting chemotaxis (open circles) and chemokinesis (closed circles) in response to varying conccntrations of recombinant human C5a; n = a minimum of 15 wells assayed for 3 difl'erent culture groups.
Migration in Different Maturational States Microglia, like other tissue macrophagcs, exhibit different maturational states (Adams and Hamilton, 1984; Streit et al., 1989). Exposure of cultured microglia to DMSO has been shown to induce characteristics similar to a "resting" state (Giulian, 1987; Giulian and Baker, 1986) whereas exposure to alpha-casein produces cellular activation (Gruys, 1980). In this study microglia were grown an additional 14 days in the continual presence of 0.5% DMSO or of 0.5 mgiml casein. Untreated cells were cultured for the same period o f time and used as aged matched controls. Chemotaxis, chemokinesis and random migration to varying concentrations of rat serum was thcn rc-examined in each of these groups. The
results for a serum dilution of 1 : 10 (maximal chemotactic response) are shown in Figure 4. Generally. cells treated with casein dcmonstratcd higher levels of chemotaxis, chemokinesis, and random migration that either the aged or normal cells but not higher than the DMSO-treated. 'This eflect is an apparent effect only, however, because when the number of cells migrating randomly in response to chemokinetic factors are subtracted from the nurnbcr of cells moving in response to a chemotactic gradient, there was no significant difference between any of the groups. This indicated, then, that microglia treated with casein or DMSO exhibited a similar level of directed migration when compared to normal, day 10 or normal, day 24 cultured microglia. Two other well-known chemotactic factors for monocytes and neutrophils were examined for their efrcct on cultured microglia. that is, transforming growth factor-beta (TGF-beta) and n-formyl peptide (met-LeuPhe). As shown in Figure 5 , microglia were found to migrate in a dose-dependent fashion to TGF-beta, with an optimal chemotactic concentration of 10 ng/ml. In the lower concentration ranges (0.01 ngiml to 1 ngiml), there was no significant difference between chemotaxis and chcmokincsis or random movement. The response of microglia to a chemotactic gradient of fMet-Leu-Phe is shown in Figure 6. Unlike human peripheral blood neutrophils and monocytes, cultured microglia did not exhibit chemotaxis to met-Leu-Phe over the concentration range of M to lo-" M Met-Leu-Phe. Microglial migration was slightly elevated at some Met-Leu-Phe concentrations ( 10p9M to
40
Yao et al. TABLE 1. Microglial Response to Met-Leu-Phe After Treatment With DMSO or Casein
$
I
233'34
Normal Day 10
i ' :
Chemotaxis
I
,/'
Random
3 lO(l1
0 0 I ii
0 1682
'
:JOG
1.r
:'Or;
143 -+ 14 (3) 118 i 21 (3)
DMSO Treated
Casein Treated
Untreated Day 24
-t 6 (3) 54 12
14?3
65f6 (3) 112 t 21
44
*
13)
(3) 32 2 4 (3)
(3
Data points represent the average number of microglia/rnm' (1.SEM) migrating in response to lo-' M tMet-Leu-Phe for normal. day 10 of culture microglia, for DMSO-treated microglia, for casein-treated mi10: 2'24.) croglia, and for untreated, day 24 of culture (age-matched) controls.
CONCENTRtTl:q~N:)I I:,, h c t j ( , - , , ~ , ' i - ~ I ' i
Fig. 5. The effect of TGF-beta on microglial migration. Open circles-average number of microgliaimm' exhibiting chemotactic migration in response to various concentrations of transforming growth factor-beta (TGF-beta). Closed circles-avcrage number of microgliaimm' migrating in response to Earle's solution only. Open triangles-average number of microgliai mm2 exhibiting chemokinetic movement in response to TGFbeta; n = minimum of 9 wells assayed for 3 different culture groups.
seen over the concentration range of lo-" M to lo-' M fMet-Leu-Phe. Migration was seen, however. to LPS activated serum, indicating that the macrophages were capable of directed movement.
DISCUSSION
Since microglia are of monocytic origin, it is not surprising that they demonstrate chemotaxis in response to known leukocyte chemoattractants and to about the 400 same degree as that seen in other types of macrophages. For example, the percentage of migrating cells obtained from macrophagic cell lines, using the multiwell chamber and a polycarbonate filter with 5-pin pores, varies from 2 to 4870 of the input cells (depending on the cell line), whereas approximately 20 to 35% of freshly isolated human monocytes migrate in response to m e t Leu-Phe under similar conditions (Aksamit et al., 198 I ; Harvath et al., 1980; Richards and McCullough, 1984). This is compared l o about 20% for the maximal chemotactic response of microglia to serum in our experiments using a similar chamber and 5-pm pore filters. The total number of cells exhibiting chemotaxis is Fig. 6. Effect of Met-Leu-Phe on rnicroglial chemotaxis. variable and, in part, depends on the type of chemoatData points represent the average number of microgliairnm2 tractant. Among the agents tested, complement conipomigrating in response to varying chemotactic concentration nents such as C5a are potent attractants for cultured migradients of met-Leu-Phe (closed circles). Random migration croglia and probably account for at least part of the indicated by the straight line; n = a minimum of 9 wells ability of normal rat serum and zymosan activated serum assayed for 3 separate culture groups. to serve as positive chemoattractants. In the CNS, microglia as well as newly recruited monocytes from the lo-"); however, this increase was not significantly dif- circulation are the prominent cellular components of the ferent from random migration. To eliminate the possibil- response to injuries such as stab wounds (Ling, 1981; ity that day 10 inicroglia may not express functional Schelper and Adrian, 1986; Streit ct al., 1988). The abilfMet-Leu-Phc rcceptors, the response to lo-' M m e t - ity of microglia to migrate in response to complement Leu-Phe was re-examined in casein-treated, DMSO- factors could explain, in part, their increased number at treated, and day 24 of culture microglia. Chemotactic these injury sites. Othcr types of chernoattractive agents may also be migration was not observed under any of these condiinvolved in the response of microglia to injury in the tions (Table I). In some experiments, peritoneal macrophages from CNS. particularly those in which the blood-brain barrier the same species of rat were also examined for thcir has bccn disrupted. We have demonstrated that cultured ability to migrate to Met-Leu-Phe. No chemotaxis was microglia can migrate in response to transforming 7
Migration in Cultured Rat Microglia
41
growth factor-beta (TGF-beta). TCF-beta is produced by functional parameters in the liver macrophage, the hcmopoietic cells such as platelets (Wahl ct al., 1987) Kupffer cell (Gruys. 1980) as well as to increase superand is involved in nionocyte recruitment during wound oxide production i n cultured microglia (Colton, unpublished). Treatment with casein at the same concentrations healing. Not all known leukocyte chernoattractants are pos- that caused an increased production of reactive oxygen itive chenioattractants for microglia. For example, cul- species did not change the overall percentage of cells tured microglia do not respond to thlet-Leu-Phe, a syn- exhibiting chemotaxis to serum. There was, however, a thetic N-formyl peptide known for its efrective general increase in activity such that nondirected movechemoattractant action on circulating monocytes and ment (chcmokincsis and random migration) was inneutrophils (Falk et al., 1982: Harvath and Aksaniit, creased. The microglia still failed to respond to M e t 1989; Marasco et al., 1984; Rot et al., 1987; Schiffman Leu-Phe under these conditions. l‘hese results suggcst. then. that rat microglia can 1982; Snyderman and Pike, 1984). This peptide is identical to N-formyl peptides released by E. coli (Marasco not respond to n-formyl peptides even when their degree et al., 1984) and is structurally similar to other N-formyl of maturation has been altered. In addition, treatment with DMSO or casein did not change the overall dcgrcc peptides released by S. sureus (Rot et al., 1987). Thc unresponsiveness of microglia to met-Leu- of microglial chernotaxis to other factors such as serum, Phe appears to be a characteristic shared by other tissue despite the fact that morphological and functional macrophages (Aksamit et al., 1981; Harvath and Ak- changes have bccn shown with these agents. The possisamit, 1989; Wharheit et id., 1988). This characteristic is bility remains, however, that different concentrations of species specific since Warheit ct al. ( 1988) have reported DMSO or casein, differcnt treatment conditions or difthat pulmonary macrophages from hamster migrated to ferent types of “activating” agents could produce a fMet-Leu-Phe, whereas mouse and guinea pig pulmo- change in microglial chemotaxis or in the response to nary macrophages responded only marginally. In the met-Leu-Phe. The ability of microglia to migrate in response to same study, rat pulmonary macrophages did not respond at all to N-formyl peptides. Rat peritoneal macrophages inflammatory mediators strengthens the idea that microalso do not respond to fik-Leu-Phe. These observations glia serve as CNS-specific macrophages. As such, their indicate that macrophages from some species either fail ability to move to sites of injury may serve as an importo express receptors or have a very low level of receptor tant step in the overall response of the CNS to that injury. expression for fMet-Leu-Phe. Cultured rat microglia, thus, appear to resemble rat pulmonary and peritoneal macrophages in their failure to respond to fMet-Leu-Phe. ACKNOWLEDGMENTS Since tissue macrophages possess numerous physWe thank Dr. Monique Dubois-Dalq for the gift of iological and biochemical properties that can be up- or TGF-beta, Dr. Elliott Schiffmann for hclpful suggesdown-regulated during macrophagc maturation (Adams tions, and Dr Larry Lilienfield for constant support. and Hamilton, 1984), it is possible that the chemoattrac- This work wa5 supported, in part, by a grant to C . Colton tan1 receptors Tor TMet-Leu-Phe are not expressed at all from the Alzheirner’s Disease and Related Disorders stages of maturation. At least two different states of mat- Foundation. uration for CNS microglia, i.c., resting (ramified) and ameboid. have been documented (del Rio Hortega, 1932; Oehmichen, 1983). In cultured microglia, Giulian and REFERENCES Baker ( 1986) have reported that treatment with DMSO or Adams DO. Hamilton TA (1984): The cell biology of macrophage retinoic acid induces morphological and functional charactivation. Ann Rev Immunol 2:1-83-3 18. acteristics reminiscent of ramified CNS microglia. We. Akxamit R . Falk W. 1,eonard E. (1981): Chemotawis by mouse mactherefore, used DMSO in our study to induce a rophagc cell lines. J Immunol l26:2 194-2 199. “resting” state. Although niorphological changes simi- Colton C. Gilbert D (1987): Production o f superoxide anions by a CNS macrophage. the microglia. FEBS Lttrs 223:284-288. lar to that described by Giulian and Bakcr are seen in del Kio Hortega P (1932): Microglia. In Penfield W (edi: “Cytology microglia treated with 0.5% DMSO, the cells retained and Cellular Pathology o f the Nervous System,” Vol. 2. New the ability to migrate to serum and to the same degree as York: Harper and Row, pp 483-543. that seen in untreated, aged matched or normal, day 10 Evsraerts MC, Berghe GVD, Saint-Reruy JM, Corbcsl L (1985): Effect of ape-dependent enzymatic degradation of‘ zymosan into of culture cells. Microglia treated in this manner reuligosaccharidcs during incubation with srrum on its upsonizamained unresponsive to met-Leu-Phe. tion by ccimplerrieiit. Pedia Res 19: 1293-1297. We also examined chemotaxis to Met-Leu-Phe Falk W. Coodwin RfI, Leonard EJ (1980): A 48-we11 niicrochemoand to complement factors in microglia treated with an taxis assembly for rapid arid accurate measurement of leuko“activating” factor. Casein has been shown to enhance cyte migration. J Immunol Methods 33:239-247.
42
Yao et al.
Ferrer I , Sarmiento J (1980): Reactive niicroglia in the developing brain. Acta Neuropathol (Berl) 50:6Y -76. Giulian U (1987): Ameboid microglia as cffcctors of inflamination in thc ccntral nervous system. J Neurosci Kes 18:155-171. Giulian D, Baker TJ (1986): Characterization of ameboid microglia isolated from developing mammalian brain. J Neurosci 6: 2 163-2 178. Giulian D, Baker T, Shih L. Lachrnan L (lY86): Interleukin I of the central nervous system is produced by amoeboid microglia. J Exp Med 164394-604. Gruys F. ( 1980): Pathogenical studies in casein-induced amyloidosis of the liver in hamsters. In Glenner G, Pinho e Costa P. de Freivas A (eds): “Amyloid and Amyloidosis.” Princeton. NJ: Excerpta Medica, pp 482-485. Harvath L, Aksainit K (1989): Human granulocytes and granulocytes froin othcr species demonstrate differences in chemotactic responsiveness to oxidized N-formyl-methionyl-lcucyl-phenylalanine. Comp Biochem Physiol 92A:97-100. Harvath L, Falk W, Leonard EJ (1980): Rapid quantitation of neutrophi1 chernotaxis: use of a poiyvinylpyrrolidone-free polycarbonate mcmbrane in a multiwell assembly. J Imrnunol Methods 37:3Y -45. Hugli TE (1981): The structural basis for anaphylatoxin and chemotactic function of C3a. C4a and CSa. CRC Crit Rev lmmunnl 2321-366. Leonard EJ, Nocr K, Skeel A (1985): Analysis of human rnonocyte chemoattractant binding by flow cytometry. J 1,eucocyte Biol 38:403-413. Ling EA (1981): The origin and nature of microglia. In Fcdcroff S , Hertz L (eds): “Adcances in Cellular Neurobiology.” Vol. 2. New York: Academic Press. pp 33-82. Ling EA. Kaur C: Wong WC (1982): Light and electron microscopic demonstration of non-specific esterase in amoeboid microglial cells in the corpus callosurn in postnatal rats: A cytochemical link to monocytes. J Anat 135385-394. Liotta L, Mandler R, Murano G, Katz D, Gordon R, Chiang P, Schiffinan E (1986): Tumor cell autocrine motility factor. Proc Nat Acad Sci 83:3302-3306. McCarthy K. de Vellis J ( 1980): Prcparation of scparatc astroglial and oligodendroglial cell cultures from rat cercbral tissuc. J Ccll Riol 85:890-902. Mannoji H. Ycgcr H. Beckcr L (1986): A specific histochemical marker (Icctin Kicinus cornrnunis agglutinin-1) for normal human microglia, and applicatin to routine histopathology. Acta Neuropathol (Berlj 71:341-343. Marasco WA, Phan SH, Krutzsch H, Showell HJ. Feltner DE, Nairn R, Recker El,, Ward PA (1984): Purification and idcntification o i iormyl-methionyl-leucyl-phenylalanine as the major peptidc
neutrophil chemotactic Factor produced by Escherishia coli. J Biol Cheni 259:5430-5439. Meade B, Kind P, Ewell J , McGrath P, Manclark C (1984): In vitro inhibition of Inurine macrophage migration by Bordetella pertussis lymphocytosis-promoting factor. Infect Imniun 45:718725. Metcalf J , Gallin J , Nausccf W, Root K 11986): In Metcalf J , Gallin J , Nauseef W, Root R (eds): “I.aboratory Manual o i Neutrophi1 Function.” New York: Raven Press, p 23. Oehrnichen M (1983): lnflaniniatory cells in the central nervous systcni: an intcgrating concept based on recent research in pathology, immunology, and forensic medicine. In Ziminerman M (cd): “Progrcss in Neuropathology,” Vol. 5 . New York: Raven Prccs, pp 277-206. Pcrry. VH, Gordon S (1988): Macrophages and microglia in the nervous system. TINS l1:273-277. Perry VH, Hume DA, Gordon S (1985): Immunohistochemical localization of macrophages and microglia in the adult and developing mouse brain. Neurosci 15:313-326. Raff MC. Field KL, Hakoinori S , Mirsky R , Pruss R, Winter J (1979): Cell-type-specific markers for distinguiching and studying neurons and the major classes o i glial cells in cultures. Brain Res 374:283-308. Richards K , McCullough J (1984): A modified microchamber method for chemotaxis and chemokinesis. lrnmunol Comm 13:49-62. Rot A, Hcndcrson LE, Copeland TD, Leonard EJ (1987): A series of six ligantls for the human formyl peptidc receptor: tetrapeptides with high chemotactic potency and efficacy. Proc Nat Acad Sci 847967-797 1. Schelpcr K , Adrian E (1986): Monocytes become macrophages; they do not become microglia: A light and electron microscopic autoradiographic study using 125-Iododeoxyuridine. J Keuropathol Exp Neurol 45:l-19. Schiffinann E (1982): Leukocyte chemotaxis. Ann Rev Physiol 44: 553-568. Snyderman R, and Pike MC (1984): Chemoattractant receptors o n phagocytic cells. Ann Rev lnimunol 2257-28 I . Streit W, Graeber M. Kreutzberg C (1988): Functional plasticity of microglia: A review. Glia 1 :301-307. Wahl SM, Hunt UA. Wakefield LM, McCartney-Francis N, Wahl LM, Roberts AB, Sporn MB (1987): Transforming growth factor type beta induces monocyte chemotaxis and growth factor production. Ccll Biology 84:5788- 5792. Warheit DB, Hartsky MA, Stefaniak MS (1988): Comparative physiology of rodent pulmonary macrophages: in vitro functional responscs. J Appl t’hysiol 64: 1953-1959. Yancey KB. Hammer CH, Harvath L. Renfer L, Frank M M . Lawley TJ ( I 985): Studies of human CSa as a mediator of inrlammation in normal human skin. J Clin Invest 75:486-495.
Chemotaxis by a CNS Macrophage, the Microglia J. Yao, L. Harvath, D.L. Gilbert, and C.A. Colton Department of Physiology and Biophysics, Georgetown University School of Medicine, Washington, D.C. (J.Y.. C.A.C.); Division of Blood and Blood Products, Center lor Biologics Evaluation and Rescarch. Food and Drug Administration (L.H.) and Laboratory of Biophysics, NINCDS, National Institutes of Health (D.L.G.), Bethesda, Maryland
Microglia demonstrate many characteristics similar to those seen in rnonocytes and tissue-specific rnacrophages, including phagocytosis, production of oxygen radicals, and growth factors and expression of MHC antigens. We have examined the ability of microglia, cultured from the cerebral cortices of neonatal rats, to demonstrate another important functional characteristic of monocytic-derived cells, that is, chemotaxis. Our results show that cultured rat microglia demonstrate chemotaxis to complement dependent chemoattractants such as recombinant C5a, zymosan activated serum, and to rat serum as well as to transforming growth factor-beta, a chemoattractant produced by platelets. Microglia fail to migrate to bacterial dependent chemoattractants such as the Nformyl peptides. The failure to respond is not dependent on maturational state of the microglia. Treatment with DMSO or casein, agents known to induce morphological and functional changes in cultured microglia reminescent of a “resting” and an “activated” macrophage, respectively, do not alter the response to f’Met-Leu-Phe. In addition, the chemotactic response to serum in DMSO or casein-treated cells is the same as the response seen in untreated day 10 cultured microglia or untreated age-matched controls. The ability of microglia to migrate in response to inflammatory mediators suggests that these cells can move to sites of injury, thereby enabling them to participate in an inflammatory response. Key words: glia, migration, macrophage, transforming growth factor-beta, Met-Leu-Phe INTRODUCTION Jn 1932, del Rio Hortega proposed that microglia represented the reticuloendothelial system in the CNS and, as such. were capable of performing macrophagic functions. This view has been supported by many investigators who have shown that both microglia in the CNS and cultured microglia possess morphological, cyto0 1990 Wiley-Liss, Inc.
chemical. and membrane antigenic similarities to monocytcs and macrophages (Giulian, 1987; Giulian and Baker, 1986; Ling, 1981; Ling et al., 1982; Oehtnichen, 1983; Perry et al., 1985; Perry and Gordon, 1988; Streit et al.. 1988). For example, microglia are reactive for non-specific esterase and express antigens such as the C3b and Fc receptors (Giulian and Baker, 1986; Ling et al., 1982; Perry et al., 1985). In addition, microglia exhibit functional characteristics of macrophages including phagocytosis, sccretion of regulatory factors such as interleukin- 1 (IL- I ) and superoxide anion production (Colton and Gilbert, 1987; Giulian and Baker, 1986; Giulian et al., 1986). Another important function of macrophages is thc ability to migrate to sites of tissue infection and injury. Typically, monocytes and tissue macrophages respond to complement components such as C5a (Hugli. 198 1; Snyderman and Pike, 1984; Yancey et al.. 1985). Complement components, in general, play a key role in the recruitment of effector cells in the microenvironment of the inflammatory rcsponse (Hugli, 1981; Yancey et al., 1985). Other inflammatory mediators also play a role in the chemotaxis of monocytic derived cells and include transforming growth factor-beta (Wahl et al., 1987) and N-formyl peptides (Falk et al., 1982; Harvath and Aksamit, 1989; Marasco et al., 1984; Rot et al., 1987; Schiffman, 1982; Snyderman and Pike, 1984). N-formyl peptides are released by various types of bacteria (Marasco et al., 1984; Rot et al.. 1987). It is not clear, however, whether microglia migrate to these same agents or, if, in fact, they exhibit chernotaxis at all. Studies by Oehmichen (1983) and others (del Rio Hortega, 1932; Perry and Gordon. 1988) imply that microglia can migrate. The purpose of this study was, thus,
Received December 20. 1989; revised March 26, IYYU; accepted April 1.2, 1990. Address reprint requests to C. Colton, Depaittnent of Physiology and Biophysics. Georgetown University School of Medicine, Washington D.C. 20007.
Migration in Cultured Rat Microglia
to determine whether cultured microglia exhibit chemotaxis in rcsponsc to known chcmoattractants.
MATERIALS AND METHODS Microglia Preparation Primary glial cell cultures were obtained from 2day-old rat (Sprague-Dawlcy) ccrebral hemispheres as described previously (Colton and Gilbert, 1987; McCarthy arid de Vellis, 1980; Kaff et al., 1979). Briefly, the cortices were dissected, placed in Leibovitz’s (L- 15) medium, and the meninges removed carefully using a binocular microscope. The cortices were then placed in fresh L-1.5 media, mechanically dissociated by tituration, and plated into 75 cm2 Falcon flasks. Cells were cultured at 37°C using a modified Dulbecco’s Medium (DMEM) with 10% fetal calf serum, 4.5 gil glucose, 3 mM glutainine, and 25 kgiml gentamycin. Thcsc conditions have been shown to be unfavorable for neuronal growth and produce a mixed glial culture (McCarthy and de Vellis, 1980). After growing to conflucncy (about 10 to 12 days), the flasks were placed on a rotary shaker for 15 to 20 hr (250 rpm) at 37°C. This allowed separation of the microglia from the underlying astrocytic layer by using their differential adherdnt properties. The supernatant, enriched in microglia, was centrifuged gently (400 x g) and the cells rcsuspended in Earle’s solution for the migration assay.
37
Alteration of Maturational State In some cases, cultures were grown in the presence of alpha casein (0.5mgiml). dimethyl sulfoxide (DMSO, 0.5% viv) or normal DMEM for an additional 14 days. Cells werc then isolated as described above and used in the cheniotaxis studies.
Microglia Chemotaxis Migration of microglia was studied using a multiwell niicrochcmotaxis assay described by Harvath et al. (1 980) and by Falk et al. ( 1980). This technique has been successfully employed to study chemotaxis in a variety of cell types including monocytes, neutrophils, macrophage cell lines and tumor cell lines (Aksamit et al., 1981; Harvath and Aksainit, 1989: Harvath et al.. 1980; Liotta et al., 1986). Cell suspensions containing 50,000 microglia were added to the upper wells of the chemotaxis chamber (Neuroprobe, Inc.). These cells were separatcd from the lower wells containing chernoattractant solution by a polycarbonate (Polyvinylpyrrolidone free) filter with 5 - p n diameter pores. To promote adhesion of the suspended microglia to the upper surface of the polycarbonate filter, the filters were treated with polyD-lysine (Syg/ml). allowed to dry, and then used immediately in the chemotaxis assay. After incubation at 37°C for 2 hr, nonmigrated cells on thc upper surface of the filter werc wiped off. Cells on the lower surface of the filter were fixed in methanol, Cell Identification stained with Leukostatfr‘a (Fisher) and counted using an Standard histochemical and imrnunocytochemical Optomax System 1V Image Analyzer (Optomax, Inc., techniques were used to identify cells in the microglial Hollis, NH). Data are presented as the average number enriched fraction. Cells from the supernatant fraction of cellsirnm’ of filter surface (k SE) and were obtained were plated onto poly-D-lyaine coated coverslips and by averaging 12 fields ( . 3 to .6 mm’ifield) from each of stained using known niicroglial and astrocytic markers. triplicate wells for a minimum of 3 separate microglial Microglia were identified by their positive reaction for cultures. Total number of cells migrating per well was non-specific esterase as described by Ling et al. ( 1982) found by multiplying the number of cellsimm* by a facand by the presence of membrane receptors for the lec- tor of 8. In this manner, the rlitio of cells migrated to tins, Ricinius cornmiinis agglutinin (Mannoji et al., percenlagc of input cells could be calculated. For poor 1986) or wheat germ agglutinin (Colton, unpublished). migratory responses, cellular debris occassionally inter,4strocytes were identified by their positive reaction for fered with the accuracy of the image analysis. In these GFAP (McCarthy and de Vellis, 1980: Raffet al., 1979). cases, cells were counted using a binocular microscope When prepared as described above, the supernatant con- and an average obtained from 10 random fields (0.021 tained 85-95% microglia, with GFAP positive cells rcp- mm2/field) of each assay group. Significance was deterrcscnting most of the remaining cells. Cultures were neg- mined using an unpaired Students t-test. ative for the presence of neurofilament. Chemotaxis was distinguished from chemokinesis, Cells that migrated to the lower surface of the poly- where chemokinesis was defined as thc increased rancarbonate filter were identified by directly fixing and dom movement of cells due to a general enhancement of staining thc cells on the filter. Due to the properties of the cell activity by the putative chemoattractive agent. In poly-carbonate filters, immunofluorescent staining could these experiments, the same concentration of Lhe putative not be used. Thus, cells were idcntified using nonspe- cheinoattractant was placed in both upper and lower cific esterase only. Migrated cells were 9.5 2 1.2% (n = wells of the chamber, thereby eliminating a chemotactic 3 filters, 100 cells countedifilter) nonspecific esterase gradient. The number of cells migrating under these conpositive. Remaining cells were indistinquishablc. ditions was taken a s a measure of chemokincsis (Falk et
38
Yao et al.
al., 1980; Richards and McCullough, 1984). Random migration was evaluated using Earle's solution as well as Earle's solution with ethanol concentrations equivalent to the ethanol concentration in fMet-Leu-Phe assays.
Peritoneal Macrophages Macrophages from the peritoneal cavity of Sprague-Dawley rats were harvested by injecting ice cold Hank's ringer into the peritoneal cavity. After heveral minutcs the fluid was withdrawn and centrifuged at 400 x g. The pellet containing macrophages was resuspended into Earles, the cells counted and then used immediately in the migration assay. Chemotactic Factors fMet-Leu-Phe and casein were purchased from Sigma Chemical Co. Met-Leu-Phe was prepared in a stock concentration of 10-' M in absolute ethanol and diluted in Earle's solution immediately before use in the chcmotaxis assays. TGF-beta (human) was a gift of Dr. Monique Dubois-Dalq, Lab of Molecular Genetics, NIH. Zymosan activated serum (ZAS) was prepared as described by Metcalf et a]. (1 986) using rat serum. LPS activated serum was prepared as described by Meade et al. (1984). Recombinant human C5a was gcncrously provided by Dr. Henry Showell, Pfizer Drugs, Groton, CT. RESULTS Migration to Complement Components; Serum, ZAS and C5a The migration of day 10 cultured microglia in rcsponse to a chemotactic gradient for normal rat serum is shown in Figure 1 . Microglia responded in a dose-dependent fashion such that thc number of cells migrating increased with increasing concentration of serum. Maximal migration was seen at a serum dilution of 1:10 to 1: 100 (Fig. 1) and represented 2 1.8% of the total number of cells added to the uppcr well. Movement in response to Earles solution only, i.e., random migration, accounted for an average of 108 2 20 cellsimn? or only 1.7% of the total number of cells in the upper well. Microglia also exhibited a chemokinetic response to rat serum. In other words, thc presence of serum, itself, induced an enhanced degree of random, nondirected movement in microglia and accounted for 10% or less of the total input cells. The number of microglia migrating in response to a chemotactic gradient, however, was significantly higher than that sccn moving chemokinetically (p < 0.001). Movement of cells in response to the chernotactic gradient for serum was time dependent and increased over 4 hr, reaching an indistinct plateau at about 2 hr.
I
I I? )',I
11
~;bi,
Fig. 1 . Migration of cultured microglia to rat serum. Data points represent the avcragc number of microgliainim' (2 SEM) that have inigrated in response to a chemotactic gradient for various concentrations of serum (closed circles), in response to Earles solution only (open triangles) and in rcsponse to equal concentrations of serum placed on both sides of the filter (chemokinesis)(open circles); n = at least 35 wells from a minimum of 3 different culture groups. Random movement remained at about 2% of total input cell number over the same timc period; 2 hr was chosen to be the routine length of incubation for the remaining studies. Because serum contains many nonspecific and specific factors that may affect cell movement, further experiments were done to clarify the cheniotactic agent. Migration in response to a nonspecific protein was tested by analyzing the chemotaxis of microglia to bovine serum albumin (BSA). Dilutions of 1:10, 1:100, and I: 1000 of a 1.O% (wtivol) BSAiEarles solution were tested for chcmotactic ability. The number of cells migrating under these conditions were not significantly different from random movement, indicating that BSA does not servc as a chemotactic agent to cultured microglia. Migration to xymosan activated serum was also studied and is shown in Figure 2. Zymosan activation promotes formation of complement cascade components including C5a, a potent chemoattractant (Evarerts ct al., 1985). As indicated, cultured microglia migrated in a dose-dependent manner to a chemotactic gradient for ZAS with maximal migration seen at a dilution of 1 :100. This value represented 18.9% of the total number of cells placed initially in the upper well. To further clarify the role of complement coniponents in microglial chemotaxis, studies were done using recombinant human C5a. As shown in Figure 3 . C5a serves as a chemoattractant for cultured microglia, with maximal migration at 0.9 kgiml representing 17.8% of the input cells responding. C5a also acted as a chemokinetic agent and at 9 pgiml, it was not possible to distinquish directed migration from enhanced random migration.
Migration in Cultured Rat Microglia
39
/
Ni'?MAL
Fig. 2. The effect of zymosan activated serum on microglial chemotaxis. Closed circles-average (2 SEM) numbcr of microglia migratedimin* in response to chemotactic gradients of ZAS. Random migration indicated by the straight line; n = at least 9 wells assayed for a minimum of 3 different culture groups.
CA'j-1
,,
rGEn
nM i(-1
Fig. 4. Changes in microglial chemotaxis with maturational state. Bar height represents the average number of microgliai mm' ? SEM. Open bars-chemotactic response to a 1:10 dilution of rat serum; diagonal striped bars-chemokinetic response to a 1 : l O dilution of rat serum; crosshatched barsmigration to Earle's only. Data obtained from each of 4 treatment protocols, i.e., normal day 10 of culture microglia, casein-treated (0.5 mgiml) microglia, DMSO-treated (0.5% mg/ml) microglia, and age matched controls. Minimum number of wells assayed = 27, for a minimum of 3 different
culture groups.
Fig. 3. The effect of recombinant human C5a on microglial chemotaxis. Data points rcprcscnt the average nuniber of cells/ mm2 ? SEM exhibiting chemotaxis (open circles) and chemokinesis (closed circles) in response to varying conccntrations of recombinant human C5a; n = a minimum of 15 wells assayed for 3 difl'erent culture groups.
Migration in Different Maturational States Microglia, like other tissue macrophagcs, exhibit different maturational states (Adams and Hamilton, 1984; Streit et al., 1989). Exposure of cultured microglia to DMSO has been shown to induce characteristics similar to a "resting" state (Giulian, 1987; Giulian and Baker, 1986) whereas exposure to alpha-casein produces cellular activation (Gruys, 1980). In this study microglia were grown an additional 14 days in the continual presence of 0.5% DMSO or of 0.5 mgiml casein. Untreated cells were cultured for the same period o f time and used as aged matched controls. Chemotaxis, chemokinesis and random migration to varying concentrations of rat serum was thcn rc-examined in each of these groups. The
results for a serum dilution of 1 : 10 (maximal chemotactic response) are shown in Figure 4. Generally. cells treated with casein dcmonstratcd higher levels of chemotaxis, chemokinesis, and random migration that either the aged or normal cells but not higher than the DMSO-treated. 'This eflect is an apparent effect only, however, because when the number of cells migrating randomly in response to chemokinetic factors are subtracted from the nurnbcr of cells moving in response to a chemotactic gradient, there was no significant difference between any of the groups. This indicated, then, that microglia treated with casein or DMSO exhibited a similar level of directed migration when compared to normal, day 10 or normal, day 24 cultured microglia. Two other well-known chemotactic factors for monocytes and neutrophils were examined for their efrcct on cultured microglia. that is, transforming growth factor-beta (TGF-beta) and n-formyl peptide (met-LeuPhe). As shown in Figure 5 , microglia were found to migrate in a dose-dependent fashion to TGF-beta, with an optimal chemotactic concentration of 10 ng/ml. In the lower concentration ranges (0.01 ngiml to 1 ngiml), there was no significant difference between chemotaxis and chcmokincsis or random movement. The response of microglia to a chemotactic gradient of fMet-Leu-Phe is shown in Figure 6. Unlike human peripheral blood neutrophils and monocytes, cultured microglia did not exhibit chemotaxis to met-Leu-Phe over the concentration range of M to lo-" M Met-Leu-Phe. Microglial migration was slightly elevated at some Met-Leu-Phe concentrations ( 10p9M to
40
Yao et al. TABLE 1. Microglial Response to Met-Leu-Phe After Treatment With DMSO or Casein
$
I
233'34
Normal Day 10
i ' :
Chemotaxis
I
,/'
Random
3 lO(l1
0 0 I ii
0 1682
'
:JOG
1.r
:'Or;
143 -+ 14 (3) 118 i 21 (3)
DMSO Treated
Casein Treated
Untreated Day 24
-t 6 (3) 54 12
14?3
65f6 (3) 112 t 21
44
*
13)
(3) 32 2 4 (3)
(3
Data points represent the average number of microglia/rnm' (1.SEM) migrating in response to lo-' M tMet-Leu-Phe for normal. day 10 of culture microglia, for DMSO-treated microglia, for casein-treated mi10: 2'24.) croglia, and for untreated, day 24 of culture (age-matched) controls.
CONCENTRtTl:q~N:)I I:,, h c t j ( , - , , ~ , ' i - ~ I ' i
Fig. 5. The effect of TGF-beta on microglial migration. Open circles-average number of microgliaimm' exhibiting chemotactic migration in response to various concentrations of transforming growth factor-beta (TGF-beta). Closed circles-avcrage number of microgliaimm' migrating in response to Earle's solution only. Open triangles-average number of microgliai mm2 exhibiting chemokinetic movement in response to TGFbeta; n = minimum of 9 wells assayed for 3 different culture groups.
seen over the concentration range of lo-" M to lo-' M fMet-Leu-Phe. Migration was seen, however. to LPS activated serum, indicating that the macrophages were capable of directed movement.
DISCUSSION
Since microglia are of monocytic origin, it is not surprising that they demonstrate chemotaxis in response to known leukocyte chemoattractants and to about the 400 same degree as that seen in other types of macrophages. For example, the percentage of migrating cells obtained from macrophagic cell lines, using the multiwell chamber and a polycarbonate filter with 5-pin pores, varies from 2 to 4870 of the input cells (depending on the cell line), whereas approximately 20 to 35% of freshly isolated human monocytes migrate in response to m e t Leu-Phe under similar conditions (Aksamit et al., 198 I ; Harvath et al., 1980; Richards and McCullough, 1984). This is compared l o about 20% for the maximal chemotactic response of microglia to serum in our experiments using a similar chamber and 5-pm pore filters. The total number of cells exhibiting chemotaxis is Fig. 6. Effect of Met-Leu-Phe on rnicroglial chemotaxis. variable and, in part, depends on the type of chemoatData points represent the average number of microgliairnm2 tractant. Among the agents tested, complement conipomigrating in response to varying chemotactic concentration nents such as C5a are potent attractants for cultured migradients of met-Leu-Phe (closed circles). Random migration croglia and probably account for at least part of the indicated by the straight line; n = a minimum of 9 wells ability of normal rat serum and zymosan activated serum assayed for 3 separate culture groups. to serve as positive chemoattractants. In the CNS, microglia as well as newly recruited monocytes from the lo-"); however, this increase was not significantly dif- circulation are the prominent cellular components of the ferent from random migration. To eliminate the possibil- response to injuries such as stab wounds (Ling, 1981; ity that day 10 inicroglia may not express functional Schelper and Adrian, 1986; Streit ct al., 1988). The abilfMet-Leu-Phc rcceptors, the response to lo-' M m e t - ity of microglia to migrate in response to complement Leu-Phe was re-examined in casein-treated, DMSO- factors could explain, in part, their increased number at treated, and day 24 of culture microglia. Chemotactic these injury sites. Othcr types of chernoattractive agents may also be migration was not observed under any of these condiinvolved in the response of microglia to injury in the tions (Table I). In some experiments, peritoneal macrophages from CNS. particularly those in which the blood-brain barrier the same species of rat were also examined for thcir has bccn disrupted. We have demonstrated that cultured ability to migrate to Met-Leu-Phe. No chemotaxis was microglia can migrate in response to transforming 7
Migration in Cultured Rat Microglia
41
growth factor-beta (TGF-beta). TCF-beta is produced by functional parameters in the liver macrophage, the hcmopoietic cells such as platelets (Wahl ct al., 1987) Kupffer cell (Gruys. 1980) as well as to increase superand is involved in nionocyte recruitment during wound oxide production i n cultured microglia (Colton, unpublished). Treatment with casein at the same concentrations healing. Not all known leukocyte chernoattractants are pos- that caused an increased production of reactive oxygen itive chenioattractants for microglia. For example, cul- species did not change the overall percentage of cells tured microglia do not respond to thlet-Leu-Phe, a syn- exhibiting chemotaxis to serum. There was, however, a thetic N-formyl peptide known for its efrective general increase in activity such that nondirected movechemoattractant action on circulating monocytes and ment (chcmokincsis and random migration) was inneutrophils (Falk et al., 1982: Harvath and Aksaniit, creased. The microglia still failed to respond to M e t 1989; Marasco et al., 1984; Rot et al., 1987; Schiffman Leu-Phe under these conditions. l‘hese results suggcst. then. that rat microglia can 1982; Snyderman and Pike, 1984). This peptide is identical to N-formyl peptides released by E. coli (Marasco not respond to n-formyl peptides even when their degree et al., 1984) and is structurally similar to other N-formyl of maturation has been altered. In addition, treatment with DMSO or casein did not change the overall dcgrcc peptides released by S. sureus (Rot et al., 1987). Thc unresponsiveness of microglia to met-Leu- of microglial chernotaxis to other factors such as serum, Phe appears to be a characteristic shared by other tissue despite the fact that morphological and functional macrophages (Aksamit et al., 1981; Harvath and Ak- changes have bccn shown with these agents. The possisamit, 1989; Wharheit et id., 1988). This characteristic is bility remains, however, that different concentrations of species specific since Warheit ct al. ( 1988) have reported DMSO or casein, differcnt treatment conditions or difthat pulmonary macrophages from hamster migrated to ferent types of “activating” agents could produce a fMet-Leu-Phe, whereas mouse and guinea pig pulmo- change in microglial chemotaxis or in the response to nary macrophages responded only marginally. In the met-Leu-Phe. The ability of microglia to migrate in response to same study, rat pulmonary macrophages did not respond at all to N-formyl peptides. Rat peritoneal macrophages inflammatory mediators strengthens the idea that microalso do not respond to fik-Leu-Phe. These observations glia serve as CNS-specific macrophages. As such, their indicate that macrophages from some species either fail ability to move to sites of injury may serve as an importo express receptors or have a very low level of receptor tant step in the overall response of the CNS to that injury. expression for fMet-Leu-Phe. Cultured rat microglia, thus, appear to resemble rat pulmonary and peritoneal macrophages in their failure to respond to fMet-Leu-Phe. ACKNOWLEDGMENTS Since tissue macrophages possess numerous physWe thank Dr. Monique Dubois-Dalq for the gift of iological and biochemical properties that can be up- or TGF-beta, Dr. Elliott Schiffmann for hclpful suggesdown-regulated during macrophagc maturation (Adams tions, and Dr Larry Lilienfield for constant support. and Hamilton, 1984), it is possible that the chemoattrac- This work wa5 supported, in part, by a grant to C . Colton tan1 receptors Tor TMet-Leu-Phe are not expressed at all from the Alzheirner’s Disease and Related Disorders stages of maturation. At least two different states of mat- Foundation. uration for CNS microglia, i.c., resting (ramified) and ameboid. have been documented (del Rio Hortega, 1932; Oehmichen, 1983). In cultured microglia, Giulian and REFERENCES Baker ( 1986) have reported that treatment with DMSO or Adams DO. Hamilton TA (1984): The cell biology of macrophage retinoic acid induces morphological and functional charactivation. Ann Rev Immunol 2:1-83-3 18. acteristics reminiscent of ramified CNS microglia. We. Akxamit R . Falk W. 1,eonard E. (1981): Chemotawis by mouse mactherefore, used DMSO in our study to induce a rophagc cell lines. J Immunol l26:2 194-2 199. “resting” state. Although niorphological changes simi- Colton C. Gilbert D (1987): Production o f superoxide anions by a CNS macrophage. the microglia. FEBS Lttrs 223:284-288. lar to that described by Giulian and Bakcr are seen in del Kio Hortega P (1932): Microglia. In Penfield W (edi: “Cytology microglia treated with 0.5% DMSO, the cells retained and Cellular Pathology o f the Nervous System,” Vol. 2. New the ability to migrate to serum and to the same degree as York: Harper and Row, pp 483-543. that seen in untreated, aged matched or normal, day 10 Evsraerts MC, Berghe GVD, Saint-Reruy JM, Corbcsl L (1985): Effect of ape-dependent enzymatic degradation of‘ zymosan into of culture cells. Microglia treated in this manner reuligosaccharidcs during incubation with srrum on its upsonizamained unresponsive to met-Leu-Phe. tion by ccimplerrieiit. Pedia Res 19: 1293-1297. We also examined chemotaxis to Met-Leu-Phe Falk W. Coodwin RfI, Leonard EJ (1980): A 48-we11 niicrochemoand to complement factors in microglia treated with an taxis assembly for rapid arid accurate measurement of leuko“activating” factor. Casein has been shown to enhance cyte migration. J Immunol Methods 33:239-247.
42
Yao et al.
Ferrer I , Sarmiento J (1980): Reactive niicroglia in the developing brain. Acta Neuropathol (Berl) 50:6Y -76. Giulian U (1987): Ameboid microglia as cffcctors of inflamination in thc ccntral nervous system. J Neurosci Kes 18:155-171. Giulian D, Baker TJ (1986): Characterization of ameboid microglia isolated from developing mammalian brain. J Neurosci 6: 2 163-2 178. Giulian D, Baker T, Shih L. Lachrnan L (lY86): Interleukin I of the central nervous system is produced by amoeboid microglia. J Exp Med 164394-604. Gruys F. ( 1980): Pathogenical studies in casein-induced amyloidosis of the liver in hamsters. In Glenner G, Pinho e Costa P. de Freivas A (eds): “Amyloid and Amyloidosis.” Princeton. NJ: Excerpta Medica, pp 482-485. Harvath L, Aksainit K (1989): Human granulocytes and granulocytes froin othcr species demonstrate differences in chemotactic responsiveness to oxidized N-formyl-methionyl-lcucyl-phenylalanine. Comp Biochem Physiol 92A:97-100. Harvath L, Falk W, Leonard EJ (1980): Rapid quantitation of neutrophi1 chernotaxis: use of a poiyvinylpyrrolidone-free polycarbonate mcmbrane in a multiwell assembly. J Imrnunol Methods 37:3Y -45. Hugli TE (1981): The structural basis for anaphylatoxin and chemotactic function of C3a. C4a and CSa. CRC Crit Rev lmmunnl 2321-366. Leonard EJ, Nocr K, Skeel A (1985): Analysis of human rnonocyte chemoattractant binding by flow cytometry. J 1,eucocyte Biol 38:403-413. Ling EA (1981): The origin and nature of microglia. In Fcdcroff S , Hertz L (eds): “Adcances in Cellular Neurobiology.” Vol. 2. New York: Academic Press. pp 33-82. Ling EA. Kaur C: Wong WC (1982): Light and electron microscopic demonstration of non-specific esterase in amoeboid microglial cells in the corpus callosurn in postnatal rats: A cytochemical link to monocytes. J Anat 135385-394. Liotta L, Mandler R, Murano G, Katz D, Gordon R, Chiang P, Schiffinan E (1986): Tumor cell autocrine motility factor. Proc Nat Acad Sci 83:3302-3306. McCarthy K. de Vellis J ( 1980): Prcparation of scparatc astroglial and oligodendroglial cell cultures from rat cercbral tissuc. J Ccll Riol 85:890-902. Mannoji H. Ycgcr H. Beckcr L (1986): A specific histochemical marker (Icctin Kicinus cornrnunis agglutinin-1) for normal human microglia, and applicatin to routine histopathology. Acta Neuropathol (Berlj 71:341-343. Marasco WA, Phan SH, Krutzsch H, Showell HJ. Feltner DE, Nairn R, Recker El,, Ward PA (1984): Purification and idcntification o i iormyl-methionyl-leucyl-phenylalanine as the major peptidc
neutrophil chemotactic Factor produced by Escherishia coli. J Biol Cheni 259:5430-5439. Meade B, Kind P, Ewell J , McGrath P, Manclark C (1984): In vitro inhibition of Inurine macrophage migration by Bordetella pertussis lymphocytosis-promoting factor. Infect Imniun 45:718725. Metcalf J , Gallin J , Nausccf W, Root K 11986): In Metcalf J , Gallin J , Nauseef W, Root R (eds): “I.aboratory Manual o i Neutrophi1 Function.” New York: Raven Press, p 23. Oehrnichen M (1983): lnflaniniatory cells in the central nervous systcni: an intcgrating concept based on recent research in pathology, immunology, and forensic medicine. In Ziminerman M (cd): “Progrcss in Neuropathology,” Vol. 5 . New York: Raven Prccs, pp 277-206. Pcrry. VH, Gordon S (1988): Macrophages and microglia in the nervous system. TINS l1:273-277. Perry VH, Hume DA, Gordon S (1985): Immunohistochemical localization of macrophages and microglia in the adult and developing mouse brain. Neurosci 15:313-326. Raff MC. Field KL, Hakoinori S , Mirsky R , Pruss R, Winter J (1979): Cell-type-specific markers for distinguiching and studying neurons and the major classes o i glial cells in cultures. Brain Res 374:283-308. Richards K , McCullough J (1984): A modified microchamber method for chemotaxis and chemokinesis. lrnmunol Comm 13:49-62. Rot A, Hcndcrson LE, Copeland TD, Leonard EJ (1987): A series of six ligantls for the human formyl peptidc receptor: tetrapeptides with high chemotactic potency and efficacy. Proc Nat Acad Sci 847967-797 1. Schelpcr K , Adrian E (1986): Monocytes become macrophages; they do not become microglia: A light and electron microscopic autoradiographic study using 125-Iododeoxyuridine. J Keuropathol Exp Neurol 45:l-19. Schiffinann E (1982): Leukocyte chemotaxis. Ann Rev Physiol 44: 553-568. Snyderman R, and Pike MC (1984): Chemoattractant receptors o n phagocytic cells. Ann Rev lnimunol 2257-28 I . Streit W, Graeber M. Kreutzberg C (1988): Functional plasticity of microglia: A review. Glia 1 :301-307. Wahl SM, Hunt UA. Wakefield LM, McCartney-Francis N, Wahl LM, Roberts AB, Sporn MB (1987): Transforming growth factor type beta induces monocyte chemotaxis and growth factor production. Ccll Biology 84:5788- 5792. Warheit DB, Hartsky MA, Stefaniak MS (1988): Comparative physiology of rodent pulmonary macrophages: in vitro functional responscs. J Appl t’hysiol 64: 1953-1959. Yancey KB. Hammer CH, Harvath L. Renfer L, Frank M M . Lawley TJ ( I 985): Studies of human CSa as a mediator of inrlammation in normal human skin. J Clin Invest 75:486-495.
