Cellular and Molecular Neurobiology, Vol. 10, No. 3, 1990
/3-Adrenergic Receptor Regulation, Through Cyclic AMP, of Nerve Growth Factor Expression in Rat Cortical and Cerebellar Astrocytes J o a n P . S c h w a r t z t'2 a n d K a t h l e e n M i s h l e r a Received February 14, 1990; accepted March 29, 1990
KEY WORDS: nerve growth factor (NGF) mRNA; catecholamines,forskolin;cerebeUar astrocytes, cortical astrocytes. SUMMARY
1. Type 1 astrocytes prepared from 3-day rat cortex and cerebellum express the 1.3-kb nerve growth factor (NGF) mRNA and synthesize and release ~-NGF. 2. Isoproterenol (IP), a fl-adrenergic agonist, stimulates NGF mRNA content in cortical astrocytes; this increase is blocked by the fl-adrenergic antagonist propranolol but not the a~-antagonist phenoxybenzamine. The ECs0 for the effect of IP is 5 nM. 3. IP increases astrocyte cyclic AMP as does forskolin, which directly activates adenylate cyclase and also increases NGF mRNA content. Cerebellar astrocytes contain about one-third as much NGF mRNA, which can also be increased by forskolin and cyclic AMP. 4. These results suggest that CNS astrocytes can serve as a source of NGF and that the NGF gene is one of the class of cyclic AMP regulated genes. INTRODUCTION fl-Nerve growth factor (NGF) has long been recognized as a neurotrophic factor for certain populations of peripheral neurons but only recently has there been evidence suggesting synthesis (Shelton and Reichardt, 1984; Korsching et al., 1Clinical Neuroscience Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland20892. 2To whom correspondenceshould be addressed at CNB, NINDS, NIH, Building9--Room lWl15, Bethesda, Maryland20892. 447 0272-4340/90/0900-0447506.00/0© 1990 Plenum Publishing Corporation
448
Schwartzand Mishler
1985) as well as biological actions (Gnahn et al., 1983; Mobley et al., 1985) in the brain. The presence of NGF in the brain raised the question of which cells were responsible for its synthesis. Several recent publications demonstrate that • cultured astrocytes can secrete NGF (Tarris et al., 1986; Furukawa et al., 1986, 1987b; Yamakuni et al., 1987). However, all of these cultures were prepared from whole brain or a combination of brain regions and recent data demonstrate that certain properties of astrocytes are region specific (Kimelberg and Katz, 1986; Chamak et al., 1987; Lehmann and Hansson, 1987; Shinoda et al., 1989). In addition, morphological and surface marker characteristics have led to the separation of astrocytes into two types, polygonal (Type 1) and process-bearing or stellate (Type 2) (Raft et al., 1983; Wilkin et al., 1983). Since polygonal astrocytes express a high level of the fl-adrenergic receptor, whereas stellate astrocytes express almost none (Trimmer and McCarthy, 1986), and since we had previously demonstrated that the C6 glioma cell line (a CNS-derived glial tumor) expressed fl-adrenergic receptors which regulated expression of NGF (Schwartz et al., 1977; Schwartz and Costa, 1977; Schwartz, 1988), we have chosen to study regulation of NGF expression via the fl-adrenergic receptor in pure cultures of type 1 astrocytes from rat cerebral cortex. EXPERIMENTAL PROCEDURES
Cell Culture. Both cortical and cerebellar astrocytes were prepared according to the method of McCarthy and DeVellis (1980) from 3-day-old rats with only minor modifications (Noble and Murray, 1984; Dubois-Dalcq, 1987). After the third overnight shake, the cells were trypsinized and cultured for 24 hr in 10 IxM cytosine arabinoside, to kill off fibroblasts. Cultures used for experiments were 99% astrocytes (GFAP positive) (Fig. 1). Indirect immunofluorescence showed fewer than 1 per 100 cells to be microglia [MAC-1 positive (Springer and Ho, 1982)], type 2 astrocytes or neurons [A2B5-positive (Dubois-Dalcq, 1987)], oligodendrocytes [galactocerebroside positive (Raft et al., 1978)], or fibroblasts (fibronectin positive). Culture conditions were Dulbecco's modified Eagle's medium-10% fetal bovine serum (Whittaker-MA Bioproducts hybridoma serum), in a humidified atmosphere of 90% air-10% CO2. R N A Preparation and Analysis. Cells (0.5-1 x 106/100-mm petri dish) were isolated and total RNA was prepared and analyzed as described (Schwartz, 1988). RNA was extracted with 6 M guanidine HC1 and precipitated with ethanol twice, followed by proteinase K digestion. Three serial dilutions were spotted onto nitrocellulose using a slot-blot apparatus. Each blot was hybridized at 55°C in 40% formamide and washed at 60°C to 0.1 x SSC-0.1% SDS (1 x SSC is 0.15 M sodium chloride-0.017 M sodium citrate). Blots were hybridized with random hexamer-labeled probes for mouse NGF (Ullrich et al., 1984), followed by 1B15 (cyclophilin) (Milner and Sutcliffe, 1983), an internal standard which corrects for mRNA recovery and/or degradation. The area of the peak for the NGFhybridized band, obtained by densitometric scan of the autoradiogram, was divided by the area of the peak for 1B15--the value obtained is expressed as units of NGF mRNA.
Cyclic AMP Regulation of Nerve Growth Factor Synthesis
449
A
Fig. 1. GFAP immunohistochemistry of astrocyte cultures. Cortical (A) and cerebellar (B) cultures were stained using a rabbit polyclonal antibody for GFAP (1/500 for 30min), followed by a goat anti-rabbit antibody labeled with fluorescein. (1/250 for 30 min; Jackson Laboratories). Control cultures, to which normal rabbit serum was added instead of the first antibody, showed no cell staining, x200; reduced 20% for reproduction.
NGF Assay. N G F was analyzed using a two-site immunoassay based on a protocol provided by Boehringer-Mannheim. The first antibody was a rabbit polyclonal to mouse fl-NGF. The I g G fraction, purified by H P L C on a TSK-2000 S W G column, was used at a final dilution of 1130,000. The second antibody was a fl-galactosidose-linked monoclonal (Boehringer-Mannheim). The assay will detect 0.3 p g N G F . Cell samples were p r e p a r e d by homogenization in 5 0 m M Tris, p H 7 , 2 0 0 m M NaC1, 2 m M E D T A , 2 0 U / m l aprotonin (BoehringerMannheim), and 20/~M phenylmethylsulfonyl fluoride. A f t e r removal of an
450
Schwartz and Mishler
aliquot for protein assay (Lowry et al., 1951), BSA and Triton X-100 were added to final concentrations of 1.5-0.15%. Cell medium was concentrated approximately 10-fold using a Centricon-10 unit (Amicon) prior to assay. Recovery of added NGF was 90-95%. Cyclic A M P was assayed as previously described (Schwartz and Passonneau, 1974).
RESULTS
Type 1 rat cortical and cerebellar astrocytes express authentic fl-NGF mRNA, -1.3 kb in size, by Northern blot analysis (Fig. 2). Exposure of cortical astrocytes to the fl-adrenergic agonist isoproterenol (IP; 100 nM) resulted in a twofold stimulation of NGF mRNA content (Fig. 3B). The maximum increase occurred at 2 hr, at a time when the cyclic AMP content was returning toward control but still -150-fold elevated (Fig. 3A). That the effect of IP is mediated via a fl-adrenergic receptor is illustrated in Fig. 4: the fl-blocker propranolol (500 nM) completely blocked the stimulation of NGF mRNA produced by IP, whereas the a~-blocker phenoxybenzamine (500 nM) was ineffective. Neither blocker had any effect when added alone. The specific tr-agonist, phenylephrine, had no effect at the same dose (100 nM) as IP. The dose-response relationship for the effect of IP on NGF mRNA (Fig. 5) was comparable to that observed for fl-adrenergic receptors in other systems, with an ECs0 of 5 nM.
Cx Cb
Fig. 2. Northern blot analysis of NGF mRNA in cortical and cerebellar astrocytes. Total RNA, prepared as described under Experimental Procedures, was applied to a 1% agaroseformaldehyde gel (10 #g for cortical cells and 25 #g for cerebellar cells). The nitrocellulose blot was probed for both NGF mRNA, as shown, and for 1B15 (not shown): the corresponding units of NGF mRNA calculated are 2.9 (cortical astrocytes; Cx) and 1.0 (cerebellar; Cb). The arrows mark the position of 28 S (upper) and 18 S (lower) ribosomal RNA.
Cyclic AMP Regulation of Nerve Growth Factor Synthesis
451
A 8000
E 6000
Z E v
4000
2000 I
d U
0
. . . .
i
. . . .
i
30
. . . .
i
60
. . . .
90
i
. . . .
120
150
TIME WITH ISOPROTERENOL ( m i n u t e s )
B 2O
. 8-
.6v
< Z cr
E Z
4-
2 -
10
08
• 0
i 2
•
, 4
,
i
•
6
i 8
•
i
,
I0
i 12
•
J 14
,
i 16
18
TIME WITH ISOPROTERENOL (hours)
Fig. 3. Time course of isoproterenol stimulation of NGF mRNA (B) and cyclic AMP (A) in rat cortical astrocytes. Confluent cultures were exposed to 100 nM IP for the times indicated, after which they were extracted and analyzed for either NGF mRNA, expressed as units relative to 1B15 mRNA (B), or cyclic AMP (pmol/mg protein) (A). Values are means + SE: n = 3-4 for NGF mRNA; n = 4 for cyclic AMP.
< Z
2 ~e
Fig. 4. Effect of isoproterenol + 0:- and fladrenergic blockers on NGF mRNA content of cortical astrocytes. Confluent cultures were exposed to 100 nM IP + 500 nM propranolol (prop) or phenoxybenzamine (phen) for 3 hr prior to extraction and analysis of mRNA. Values are means + SE (n = 4). (*) P < 0.01 vs control; ( # ) P < 0 . 0 1 vs IP.
~ z
i
o
CON
IP
IP ÷ PROP
IP + PHEN
452
Schwartz and Mishler
z~ It.
L~ z
1.
I
I
I
1
0
-10
-9
-8
-7
-6
ISOPROTERENOL (M)
Fig. 5. Dose-response curve for IP stimulation of NGF mRNA. Confluent cultures were incubated with the indicated doses of IP for 3 hr prior to extraction and analysis of NGF mRNA. Values are means + SE (n = 3-6).
600 •
A
500"
E
400 •
E
o_ 300"
< 200" I u LD I00"
U
0 I
.
.
.
.
I
.
30
.
.
.
I
.
.
.
.
I
60
.
.
90
.
.
I
.
120
.
.
.
150
TIME WITH FORSKOLIN (minutes)
~ ,2.4
2.0
g
/3-Adrenergic Receptor Regulation, Through Cyclic AMP, of Nerve Growth Factor Expression in Rat Cortical and Cerebellar Astrocytes J o a n P . S c h w a r t z t'2 a n d K a t h l e e n M i s h l e r a Received February 14, 1990; accepted March 29, 1990
KEY WORDS: nerve growth factor (NGF) mRNA; catecholamines,forskolin;cerebeUar astrocytes, cortical astrocytes. SUMMARY
1. Type 1 astrocytes prepared from 3-day rat cortex and cerebellum express the 1.3-kb nerve growth factor (NGF) mRNA and synthesize and release ~-NGF. 2. Isoproterenol (IP), a fl-adrenergic agonist, stimulates NGF mRNA content in cortical astrocytes; this increase is blocked by the fl-adrenergic antagonist propranolol but not the a~-antagonist phenoxybenzamine. The ECs0 for the effect of IP is 5 nM. 3. IP increases astrocyte cyclic AMP as does forskolin, which directly activates adenylate cyclase and also increases NGF mRNA content. Cerebellar astrocytes contain about one-third as much NGF mRNA, which can also be increased by forskolin and cyclic AMP. 4. These results suggest that CNS astrocytes can serve as a source of NGF and that the NGF gene is one of the class of cyclic AMP regulated genes. INTRODUCTION fl-Nerve growth factor (NGF) has long been recognized as a neurotrophic factor for certain populations of peripheral neurons but only recently has there been evidence suggesting synthesis (Shelton and Reichardt, 1984; Korsching et al., 1Clinical Neuroscience Branch, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland20892. 2To whom correspondenceshould be addressed at CNB, NINDS, NIH, Building9--Room lWl15, Bethesda, Maryland20892. 447 0272-4340/90/0900-0447506.00/0© 1990 Plenum Publishing Corporation
448
Schwartzand Mishler
1985) as well as biological actions (Gnahn et al., 1983; Mobley et al., 1985) in the brain. The presence of NGF in the brain raised the question of which cells were responsible for its synthesis. Several recent publications demonstrate that • cultured astrocytes can secrete NGF (Tarris et al., 1986; Furukawa et al., 1986, 1987b; Yamakuni et al., 1987). However, all of these cultures were prepared from whole brain or a combination of brain regions and recent data demonstrate that certain properties of astrocytes are region specific (Kimelberg and Katz, 1986; Chamak et al., 1987; Lehmann and Hansson, 1987; Shinoda et al., 1989). In addition, morphological and surface marker characteristics have led to the separation of astrocytes into two types, polygonal (Type 1) and process-bearing or stellate (Type 2) (Raft et al., 1983; Wilkin et al., 1983). Since polygonal astrocytes express a high level of the fl-adrenergic receptor, whereas stellate astrocytes express almost none (Trimmer and McCarthy, 1986), and since we had previously demonstrated that the C6 glioma cell line (a CNS-derived glial tumor) expressed fl-adrenergic receptors which regulated expression of NGF (Schwartz et al., 1977; Schwartz and Costa, 1977; Schwartz, 1988), we have chosen to study regulation of NGF expression via the fl-adrenergic receptor in pure cultures of type 1 astrocytes from rat cerebral cortex. EXPERIMENTAL PROCEDURES
Cell Culture. Both cortical and cerebellar astrocytes were prepared according to the method of McCarthy and DeVellis (1980) from 3-day-old rats with only minor modifications (Noble and Murray, 1984; Dubois-Dalcq, 1987). After the third overnight shake, the cells were trypsinized and cultured for 24 hr in 10 IxM cytosine arabinoside, to kill off fibroblasts. Cultures used for experiments were 99% astrocytes (GFAP positive) (Fig. 1). Indirect immunofluorescence showed fewer than 1 per 100 cells to be microglia [MAC-1 positive (Springer and Ho, 1982)], type 2 astrocytes or neurons [A2B5-positive (Dubois-Dalcq, 1987)], oligodendrocytes [galactocerebroside positive (Raft et al., 1978)], or fibroblasts (fibronectin positive). Culture conditions were Dulbecco's modified Eagle's medium-10% fetal bovine serum (Whittaker-MA Bioproducts hybridoma serum), in a humidified atmosphere of 90% air-10% CO2. R N A Preparation and Analysis. Cells (0.5-1 x 106/100-mm petri dish) were isolated and total RNA was prepared and analyzed as described (Schwartz, 1988). RNA was extracted with 6 M guanidine HC1 and precipitated with ethanol twice, followed by proteinase K digestion. Three serial dilutions were spotted onto nitrocellulose using a slot-blot apparatus. Each blot was hybridized at 55°C in 40% formamide and washed at 60°C to 0.1 x SSC-0.1% SDS (1 x SSC is 0.15 M sodium chloride-0.017 M sodium citrate). Blots were hybridized with random hexamer-labeled probes for mouse NGF (Ullrich et al., 1984), followed by 1B15 (cyclophilin) (Milner and Sutcliffe, 1983), an internal standard which corrects for mRNA recovery and/or degradation. The area of the peak for the NGFhybridized band, obtained by densitometric scan of the autoradiogram, was divided by the area of the peak for 1B15--the value obtained is expressed as units of NGF mRNA.
Cyclic AMP Regulation of Nerve Growth Factor Synthesis
449
A
Fig. 1. GFAP immunohistochemistry of astrocyte cultures. Cortical (A) and cerebellar (B) cultures were stained using a rabbit polyclonal antibody for GFAP (1/500 for 30min), followed by a goat anti-rabbit antibody labeled with fluorescein. (1/250 for 30 min; Jackson Laboratories). Control cultures, to which normal rabbit serum was added instead of the first antibody, showed no cell staining, x200; reduced 20% for reproduction.
NGF Assay. N G F was analyzed using a two-site immunoassay based on a protocol provided by Boehringer-Mannheim. The first antibody was a rabbit polyclonal to mouse fl-NGF. The I g G fraction, purified by H P L C on a TSK-2000 S W G column, was used at a final dilution of 1130,000. The second antibody was a fl-galactosidose-linked monoclonal (Boehringer-Mannheim). The assay will detect 0.3 p g N G F . Cell samples were p r e p a r e d by homogenization in 5 0 m M Tris, p H 7 , 2 0 0 m M NaC1, 2 m M E D T A , 2 0 U / m l aprotonin (BoehringerMannheim), and 20/~M phenylmethylsulfonyl fluoride. A f t e r removal of an
450
Schwartz and Mishler
aliquot for protein assay (Lowry et al., 1951), BSA and Triton X-100 were added to final concentrations of 1.5-0.15%. Cell medium was concentrated approximately 10-fold using a Centricon-10 unit (Amicon) prior to assay. Recovery of added NGF was 90-95%. Cyclic A M P was assayed as previously described (Schwartz and Passonneau, 1974).
RESULTS
Type 1 rat cortical and cerebellar astrocytes express authentic fl-NGF mRNA, -1.3 kb in size, by Northern blot analysis (Fig. 2). Exposure of cortical astrocytes to the fl-adrenergic agonist isoproterenol (IP; 100 nM) resulted in a twofold stimulation of NGF mRNA content (Fig. 3B). The maximum increase occurred at 2 hr, at a time when the cyclic AMP content was returning toward control but still -150-fold elevated (Fig. 3A). That the effect of IP is mediated via a fl-adrenergic receptor is illustrated in Fig. 4: the fl-blocker propranolol (500 nM) completely blocked the stimulation of NGF mRNA produced by IP, whereas the a~-blocker phenoxybenzamine (500 nM) was ineffective. Neither blocker had any effect when added alone. The specific tr-agonist, phenylephrine, had no effect at the same dose (100 nM) as IP. The dose-response relationship for the effect of IP on NGF mRNA (Fig. 5) was comparable to that observed for fl-adrenergic receptors in other systems, with an ECs0 of 5 nM.
Cx Cb
Fig. 2. Northern blot analysis of NGF mRNA in cortical and cerebellar astrocytes. Total RNA, prepared as described under Experimental Procedures, was applied to a 1% agaroseformaldehyde gel (10 #g for cortical cells and 25 #g for cerebellar cells). The nitrocellulose blot was probed for both NGF mRNA, as shown, and for 1B15 (not shown): the corresponding units of NGF mRNA calculated are 2.9 (cortical astrocytes; Cx) and 1.0 (cerebellar; Cb). The arrows mark the position of 28 S (upper) and 18 S (lower) ribosomal RNA.
Cyclic AMP Regulation of Nerve Growth Factor Synthesis
451
A 8000
E 6000
Z E v
4000
2000 I
d U
0
. . . .
i
. . . .
i
30
. . . .
i
60
. . . .
90
i
. . . .
120
150
TIME WITH ISOPROTERENOL ( m i n u t e s )
B 2O
. 8-
.6v
< Z cr
E Z
4-
2 -
10
08
• 0
i 2
•
, 4
,
i
•
6
i 8
•
i
,
I0
i 12
•
J 14
,
i 16
18
TIME WITH ISOPROTERENOL (hours)
Fig. 3. Time course of isoproterenol stimulation of NGF mRNA (B) and cyclic AMP (A) in rat cortical astrocytes. Confluent cultures were exposed to 100 nM IP for the times indicated, after which they were extracted and analyzed for either NGF mRNA, expressed as units relative to 1B15 mRNA (B), or cyclic AMP (pmol/mg protein) (A). Values are means + SE: n = 3-4 for NGF mRNA; n = 4 for cyclic AMP.
< Z
2 ~e
Fig. 4. Effect of isoproterenol + 0:- and fladrenergic blockers on NGF mRNA content of cortical astrocytes. Confluent cultures were exposed to 100 nM IP + 500 nM propranolol (prop) or phenoxybenzamine (phen) for 3 hr prior to extraction and analysis of mRNA. Values are means + SE (n = 4). (*) P < 0.01 vs control; ( # ) P < 0 . 0 1 vs IP.
~ z
i
o
CON
IP
IP ÷ PROP
IP + PHEN
452
Schwartz and Mishler
z~ It.
L~ z
1.
I
I
I
1
0
-10
-9
-8
-7
-6
ISOPROTERENOL (M)
Fig. 5. Dose-response curve for IP stimulation of NGF mRNA. Confluent cultures were incubated with the indicated doses of IP for 3 hr prior to extraction and analysis of NGF mRNA. Values are means + SE (n = 3-6).
600 •
A
500"
E
400 •
E
o_ 300"
< 200" I u LD I00"
U
0 I
.
.
.
.
I
.
30
.
.
.
I
.
.
.
.
I
60
.
.
90
.
.
I
.
120
.
.
.
150
TIME WITH FORSKOLIN (minutes)
~ ,2.4
2.0
g
