(Received for publication, November 20, 1996, and in revised form, February 13, 1997)
From the Department of Neurology and
Medical
and Molecular Pharmacology, UCLA School of Medicine, Los Angeles,
California 90095 and the ¶ Department of Immunology, Berlex
Biosciences, Richmond, California 94804
The understanding of the induction and regulation of inducible nitric-oxide synthase (iNOS) in human cells may be important in developing therapeutic interventions for inflammatory diseases. In the present study, we not only demonstrated that human fetal mixed glial cultures, as well as enriched microglial cultures, synthesize iNOS and nitric oxide (NO) in response to cytokine stimulation, but also assessed the kinetics of iNOS and NO synthesis in human fetal mixed glial cultures. The iNOS mRNA was expressed within 2 h after stimulation and decreased to base line by 2 days. Significant levels of iNOS protein appeared within 24 h after stimulation and remained elevated during the culture period. A dramatic increase in NO production and NO-mediated events, such as the induction of cyclic guanosine monophosphate (cGMP), NADPH diaphorase activity, and nitrotyrosine occurred 3 days after stimulation, a delay of 48 h from the time of the first expression of iNOS enzyme. This delay of NO production was altered by the addition of tetrahydrobiopterin, but not by the addition of L-arginine, heme, flavin adenine dinucleotide (FAD), flavin mononucleotide (FMN), or NADPH. These findings suggest that a post-translational regulatory event might be involved in iNOS-mediated NO production in human glia.
Nitric oxide (NO) mediates functions as diverse as vasodilation (1-3), neurotransmission (4, 5), and immune-mediated cytotoxicity (6-8). NO production is catalyzed by NO synthases of which there are three isoforms (9-11). Multiple sclerosis (MS)1 is a central nervous system disorder with immune-mediated destruction of myelin and the myelin-producing cells, oligodendrocytes. The presence of inducible nitric-oxide synthase (iNOS or Type II NOS) and the "footprints" of NO in tissues of patients with MS and animals with experimental allergic encephalomyelitis, a model for MS, suggests that NO may play a role in this central nervous system autoimmune disease (12-15).
Glial cells including astrocytes, microglia, and oligodendrocytes are all involved in the lesion and plaque formation in MS and experimental allergic encephalomyelitis. Rodent astrocytes and microglia express high levels of Type II/iNOS and release significant NO within hours after lipopolysaccharide stimulation in vitro (8, 16, 17). In culture NO produces mitochondrial dysfunction, DNA damage, morphological changes, and necrotic cell death in rat oligodendrocytes (8, 18), while rat astrocytes and microglia are more resistant to NO-mediated damage (19, 20). If such glial cell-mediated NO-dependent damages were to occur in vivo in MS, it could contribute to the plaque formation and a loss of myelin.
Since MS is a human central nervous system disorder, insights into the role of NO in the immnopathology of MS require a better understanding of NO induction in human glial cells. However, the induction and the regulation of iNOS in human glial cells is still unclear and seems substantially different from that in rodents. First of all, the nature of inducing signals is different. Lipopolysaccharide, a potent NO inducer in rodent cells, does not induce NO in human glial cells when used alone or in combination with cytokines (21, 22). Second, the issue of whether or not human microglia synthesize iNOS and release NO is still unresolved (22-25). In contrast to what has been observed in rodent cultures (8, 16, 17), cultures of human astrocytes produce low amounts of NO, and its detection in culture supernatants is delayed from the time of stimulation (21, 22, 26, 27).
To understand the mechanism of iNOS production in human glial cells, we performed a kinetic analysis of iNOS mRNA expression, protein synthesis, enzyme activity, and NO production as well as an evaluation of NO-mediated events or "footprints" in enriched microglial and also in mixed cultures of astrocytes and microglia from human fetal brain tissue. We found that there is a temporal lag between iNOS enzyme synthesis and NO production. Insufficient intracellular levels of the iNOS cofactor tetrahydrobiopterin (BH4) may explain the deficiency of enzyme activity leading to delayed onset of NO production. These findings suggest a post-translational regulation of iNOS enzyme activity in human glial cells.
The human glial primary cultures were established from fetal brain tissues of fetuses 19.8 ± 1.8 weeks old (range 17.5-24.0 weeks). Entire brains of aborted fetuses were obtained from Advanced Bioscience Resources, Inc. (Alameda, CA). Cultures from a single brain were used for each experiment. A total of eight different brains were used to establish primary cultures throughout the course of these experiments.
Human brain tissue was prepared as described previously (27) and cultured in Iscove's modified Dulbecco's medium (Irvine Scientific, Santa Ana, CA) containing 10% non-heat-inactivated fetal calf serum (Gemini Bioproducts, Inc., Calabasas, CA) as well as L-glutamine (2 mM) and gentamicin sulfate (50 µg/ml), both purchased from Irvine Scientific. Once cultures were established (1-4 weeks), they were split at passage 1 and maintained in Iscove's medium containing microglial growth factors: macrophage colony-stimulating factor (M-CSF), granulocyte/macrophage colony-stimulating factor (GM-CSF), and interleukin 3 (IL-3) (each used at 2.8 ng/ml, Immunex Corp., Seattle, WA). Enriched microglial cells were harvested from attached mixed cultures by gentle shaking.
For phenotyping, viable or fixed glial cells were stained as described
previously (27). Astrocytes were detected by a polyclonal anti-glial
fibrillary acidic protein antibody (1:100; Boehringer Mannheim).
Microglia were labeled by monoclonal EBM-11 (anti-CD68, 1:50; DAKO
Corp., Carpinteria, CA), fluorescein-conjugated Ricinus communis agglutinin 1 (RCA-1) (1:50; Vector Laboratories, Inc., Burlingame, CA), and, for receptors for acetylated low density lipoprotein, 1 µg/ml 1,1-dioctade cyl-1-3,3,3
3
-tetramethyl indocarbocyanine perchlorate-conjugated LDL (Dil-
-LDL) (Biomedical Technologies, Inc., Stoughton, MA). Presence of neurons was examined by
staining with anti-neuronal specific enolase (1:50; DAKO), and
oligodendrocytes was detected by staining with a polyclonal antibody to
galactocerebroside (GalC 1:20, Boehringer Mannheim). Polyclonal Factor
VIII antibody (1:50; DAKO) was used for determining whether there were
endothelial cells in the cultures.
The cultures of
mixed human microglia and astrocytes (50:50) were either left
unstimulated or they were stimulated with 70 ng/ml human recombinant
interferon (IFN-
; Genentech, San Francisco, CA) and 5 ng/ml
human recombinant interleukin-1
(IL-1
; Immunex Corp, Seattle, WA)
for various time periods from 2 h to 7 days. The cell pellets were
harvested, and total RNA was isolated by guanidinium isothiocyanate
according to published methods (28). RNA concentration was measured in
a Beckman DU-65 spectrophotometer equipped with a 5-carat microcell.
RNA (50 µg/lane) was mixed with ethidium bromide (1-2 µg) and
loaded into a 1.2% agarose gel containing formaldehyde. After
separation by electrophoresis, RNA was transferred to a Nytran membrane
(Schleicher & Schuell) and cross-linked with ultraviolet irradiation.
The human hepatocyte iNOS cDNA fragment of 2.1 kb was digested from
pBluescript SK with EcoRI and BamHI (obtained
from Dr. D. Geller, University of Pittsburgh, Pittsburgh, PA), labeled
with [
-32P]dATP (10.0 mCi/ml, DuPont), and hybridized
with the membrane. A 0.7-kb BamHI cDNA fragment of rat
cyclophilin gene, isolated from pCD vector (obtained from Dr. J. G. Sutcliffe, The Scripps Research Institute, La Jolla, CA) was used as a
control for the amount of total RNA loaded in each lane. After 1 h
of prehybridization, hybridization was performed for 2 h at
65 °C using QuikHYB Rapid Hybridization solution (Stratagene, La
Jolla, CA). The blot was washed to 65 °C, and exposed to Kodak XAR-5
film at
70 °C. For iNOS, the exposure was 6-9 days, while, for
cyclophilin, exposure was overnight. For reprobing of the same membrane
with another cDNA, the membrane was stripped of the hybridized
probe with 0.1 × SSC, 0.1% SDS at boiling temperature for 20 min.
Human glial cells
were plated on four-well glass Lab-Tek chamber slides at 2 × 105 cell/ml. After an incubation period similar to that
described above, with or without cytokine stimulation, cultures were
rinsed with diethylpyrocarbonate-treated PBS and fixed for 10 min in 4% paraformaldehyde. After rinsing twice with
diethylpyrocarbonate-treated PBS, cells were incubated with 0.1 M triethanolamine and then 0.1 M
triethanolamine,0.25% acetic anhydride for 10 min each. After two
5-min rinses in 2 × SSC, cells were dehydrated through graded
alcohol up to 100%. After two 5-min incubations with chloroform, cells
went through 100% and 95% alcohol and were dried in air. All of the
above steps were performed at room temperature. 33P-Labeled
antisense and sense riboprobes of the human hepatocyte-iNOS were
synthesized by in vitro transcription. The pBluescript
plasmid containing 2.1-kb fragment of the human hepatocyte-iNOS
(inserted into EcoRI/BamHI site) was linearized
at NotI and HindIII sites for antisense and sense
probe, respectively. [-33P]UTP (10.0 mCi/ml, DuPont)
was incubated with cold nucleotides, DTT, RNasin, linearized template
plasmids, and RNA polymerase at 37 °C for 1 h. The template DNA
was then degraded by RNase-free DNase I at 37 °C for 15 min.
Finally, labeled probes were precipitated at
20 °C overnight. For
hybridization, probes were heat-denatured at 65 °C for 5 min and
quenched on ice, and then added into the hybridization buffer
containing 0.6 M NaCl, 0.04 M EDTA, 0.1% sodium pyrophosphate, 10% dextran, 0.2% SDS, 0.02% heparin, and 50%
formamide. The hybridization buffer was added to the cells on slides
which were then covered with cover slides and incubated at 53 °C
overnight. After a brief wash in 2 × SSC, the slides were
transferred to RNase A digestion buffer (10 mM Tris-HCl, pH
8.0, 5 mM EDTA, pH 8.0, 300 mM NaCl, 10 mM DTT, 30 mg/ml RNase A) for 30 min at room temperature.
Thereafter, slides were washed with gradually increased stringency from
2 × SSC at room temperature to 0.2 × SSC at 60 °C. The
slides were allowed to cool to room temperature, dehydrated, coated
with NTB-2 emulsion (Eastman Kodak Co., Rochester, NY), and exposed for
5 days. Slides were developed and counterstained with methylene blue
and examined for silver grains on a Zeiss Axioskop microscope.
The iNOS antibody used for iNOS immunocytochemical
staining of cells and on Western blots was a polyclonal antibody raised in rabbits against a specific human iNOS peptide sequence
(NESPQPLVETGK, encoding residues 54-65 of human iNOS) synthesized at
Berlex Biosciences, Richmond, CA. This sequence is not found in Type I
or III NOS. The specificity of the antibody was demonstrated by Western
blot analysis where a single band at 130 kDa was detected in the
astrocytoma line A-172 when stimulated with human recombinant IL-1
and IFN-
(Fig. 1). The band could be blocked by
preincubation of the antibody with the inducing peptide overnight at
4 °C (data not shown). The antibody did not stain neurons
(containing Type I NOS) or endothelial cells (containing type III NOS)
(data not shown).
Western Blot Analysis for iNOS Protein
Mixed cultures of
human fetal glia (both microglial and astrocytes 50:50) were stimulated
with IL-1/IFN-
or left unstimulated over the same time course as
above. Cell pellets were harvested and homogenized with 50 mM Tris buffer, pH 7.4, containing 0.1 mM EDTA,
0.1 mM EGTA, and 0.5 mM dithiothreitol at
0-4 °C. Homogenates were centrifuged at 20,000 × g
for 60 min at 4 °C. The protein concentration in the supernatant of
the cellular homogenates was determined by the Bradford Coomassie
Brilliant Blue method (Bio-Rad). Bovine serum albumin was used as the
standard. The cell lysates were diluted, electrophoresed on a 10%
polyacrylamide gel, and transferred overnight to a nitrocellulose blot.
The nitrocellulose blot was blocked with a solution containing 2%
bovine serum albumin and 5% nonfat milk for 1 h at room
temperature and washed with 0.2% Tween 20 in PBS three times for 5 min. The blot was probed with the polyclonal antibodies specific for
human iNOS peptide (1:500) overnight at 4 °C. The blot was washed
three times and then incubated with the secondary antibody, a goat
anti-rabbit horseradish peroxidase-conjugated IgG, for 1 h at room
temperature. The iNOS protein was visualized by chemiluminescence
(Amersham).
Mixed glia or enriched cultures of microglia or astrocytes were cultured in Lab-Tek glass chamber slides with or without cytokines as above for times ranging from 6 h to 7 days, after which cells were fixed in 4% paraformaldehyde at room temperature for 5 min. They were then rinsed in PBS, and permeabilized with 0.1% Triton in PBS. After three 5-min washes with PBS, cells were incubated with 1% normal goat serum in PBS for 30 min to block nonspecific binding. They were then incubated with the primary iNOS specific polyclonal antibody (see above) overnight in PBS at 4 °C. To look for one of the footprints of NO production, the nitration of tyrosines, cells were also incubated with polyclonal rabbit NT antibodies (Upstate Biotechnology, Inc., Lake Placid, NY) at 1:100. After PBS wash, cells were incubated at 4 °C with a biotinylated swine anti-rabbit IgG antibody (1:200 dilution) in PBS for 30 min, washed in PBS, and then incubated with an alkaline phosphatase-linked avidin (Vector) for 30 min at room temperature. Finally, cells were exposed to alkaline phosphatase substrate p-nitrophenyl phosphate (Vector) for 5 min. The slides were then washed in water and mounted with a glycerol/gelatin mounting medium (Sigma). As a control for nonspecific staining, either the primary antibody was omitted or normal rabbit IgG was used as the primary antibody.
Semi-quantitation of Histochemical ResultsTo semi-quantitate the histochemical staining results, we scored cells in a double-blind experiment after staining for iNOS and NT. According to the intensity, a 0-5 scale was set to describe the staining level and the number of positive cells. More than 200 cells in each group were examined, and the weighted average of the score for each group was derived as a measure of the presence of positive cells. For in situ hybridization results, cell images were captured into a Macintosh computer and scanned by the NIH Image Program. The mean optical density of each cell was recorded, and the average of 200 cells in each group was used as a semi-quantitative measurement.
Determination of iNOS Enzyme Activity by L-Citrulline LevelsHuman fetal glial iNOS activity was measured by
determining the conversion of L-[3H]arginine
to L-[3H]citrulline as described previously
(29). Briefly, cell pellets were harvested and homogenized with 50 mM TEA buffer, pH 7.4, containing 0.1 mM EDTA,
0.1 mM EGTA, 0.5 mM dithiothreitol, 1 µM pepstatin A, and 2 µM leupeptin at
0-4 °C. Homogenates were centrifuged at 20,000 × g
for 60 min at 4 °C, and the supernatant was used as the source of NO
synthase. Enzymatic reactions were conducted at 37 °C in 50 mM TEA, pH 7.4., containing 50 µM
L-arginine, 100 µM NADPH, 10 µM
BH4, 10 µM FMN, 10 µM FAD,
0.2-0.4 mg of supernatant protein, and approximately 200,000 dpm of
purified L-2,3,4,5-[3H]arginine HCl (77 Ci/mmol; Amersham). The reactions were terminated by addition of
ice-cold 20 mM sodium acetate buffer (pH 5.5) containing 2 mM EDTA, 0.2 mM EGTA, and 1 mM
citrulline. The samples were chromatographed on columns of Dowex AG
1-X8, OH form (prepared from the acetate form). The
eluate was collected, and counted in Beckman LS 3801 liquid
scintillation spectrometer.
The protein concentration in cellular homogenates was determined by the Bradford Coomassie Brilliant Blue method (Bio-Rad, Richmond, CA). Bovine serum albumin was used as the standard.
Determination of Total Nitric Oxide (NOxNOx was
determined by measuring the formation of both the stable oxidation
products of NO, namely nitrite (NO2
)
and nitrate (NO3
). NO in
oxygen-containing solutions is chemically unstable and undergoes rapid
oxidation to NO2
. The presence of
various biological tissue components catalyzes this oxidation and
promotes further oxidation of NO2
to
NO3
(10, 31). Therefore, it is
necessary to measure both NO2
and
NO3
to accurately determine the level
of total NO. The concentrations of NO2
can be determined by Griess reagent reaction as described by Tracey
(30). Nitrate in cell culture supernatants was first reduced to nitrite
by incubation the samples for 30 min with nitrate reductase (0.1 units/ml, Boehringer Mannheim) in the presence of 100 µM
NADPH and 10 µM FAD. Any remaining NADPH was oxidized with lactate dehydrogenase (10 units/ml) in the presence of 10 mM sodium pyruvate. The total nitrite concentration was
then determined by using the procedure based on the Griess reagent
reaction. Background nitrite and nitrate levels in the medium (about 8 µM) were subtracted from the experimental values.
For NADPH diaphorase staining,
fixed cells were incubated in 0.1 M phosphate buffer, pH
7.4, containing 1 mg/ml NADPH, 0.1 mg/ml nitro blue tetrazolium, and
0.3% Triton X-100 at 37 °C for 1 h and rinsed twice with
1 × PBS. The positive staining was also assessed in a
semi-quantitative way as for iNOS and NT staining.
The cGMP radioimmunoassay as a bioassay of NO
production was adapted from the procedure of Ishii et al.
(31) and modified by Simmons and Murphy (32). All the reagents for this
assay were purchased from Sigma. Briefly, cell cultures in 12-well
plates were washed twice with HBSS and then equilibrated in HBSS for 20 min. Buffer was then replaced with HBSS containing 0.6 mM
isobutylmethylxanthine, 20 units/ml superoxide dismutase, and
L-arginine (104 M) and/or
arginine analogues N
-nitro-L-arginine methyl
ester or
NG-monomethyl-L-arginine.
Incubation was for 15 min at 37 °C, after which buffer was replaced
by 150 µl of cold assay buffer (50 mM sodium
acetate with 0.1% sodium azide, pH 6.2), and cells were scraped
out with 300 µl of assay buffer and frozen at
20 °C until cGMP radioimmunoassay.
Cell pellets were generated by centrifugation at 14,000 × g for 10 min at 4 °C, and 100 µl of supernatants was
then assayed in duplicate using radioimmunoassay procedures adapted
from those of Steiner et al. (33). Rabbit antiserum against
cGMP was provided by Dr. S. Murphy (University of Iowa, Iowa City, IA).
It was prepared using limpet hemocyanin-conjugated cGMP as immunogen.
The samples were acetylated with 5 µl of triethylamine/acetic
anhydride (2:1 v/v) and then incubated with equal volumes of antiserum
(diluted 1:10,000) and 125I-cyclic GMP-tyrosine methyl
ester (20,000 cpm/reaction, DuPont) in assay buffer for 16-20 h at
4 °C. They were then incubated with 200 µl/reaction magnetic goat
anti-rabbit IgG (BioMag, Cambridge, MA) for 20 min at room temperature
and placed on a magnetic plate (BioMag) for 10 min to achieve
separation. The supernatant was decanted, and the pellets were counted
for radioactivity in a Beckman 4007 counter.
The
iNOS substrate L-arginine or NOS cofactors, including
BH4, FAD, FMN, and NADPH, were added separately or together
to the mixed microglia/astrocyte cultures. The total
NOx in the culture
supernatants was determined as above. The above substrates were used in
a range of concentrations from 10 µM to 500 µM.
The levels of
NOx, cGMP, and citrulline
produced by stimulation of glial cells were analyzed for significance
by means of a multivariate analysis of variance (ANOVA). Statistical significance was established at p < 0.05.
There were about 50% microglia and 50% astrocytes in
the mixed glial cultures, as determined by Dil--LDL and specific
microglial antibody staining and glial fibrillary acidic protein
staining, respectively. More than 98% of the cells in the enriched
microglial cultures were Dil-
-LDL positive. These cultures contained
CD4-negative parenchymal microglia, and not the
perivascular microglia (27). When cultures were stained with antibodies
to neuronal specific enolase, galactocerebroside, or Factor VIII, no
significant numbers of neurons, oligodendrocytes, or endothelial cells,
respectively, were found in either the mixed human fetal glial cultures
or the enriched microglial cultures (data not shown).
Northern blot analysis was used to determine iNOS
mRNA expression in human fetal mixed glial cultures. As indicated
in Fig. 2A, the iNOS probe identified a
single 4.1-kb mRNA in cultures treated by IL-1 and IFN-
. The
iNOS mRNA signal was detected within 2 h after cytokine
stimulation, and a significant level was maintained for at least 2 days. It gradually decreased after day 3, and was undetectable at day 5 and day 7. By a similar method, iNOS mRNA was also detected in
enriched human microglial cultures after cytokine stimulation as shown
in Fig. 2B.
To confirm the Northern blot results, in situ hybridization
was also used to identify iNOS mRNA in human mixed glial cells in vitro. Cells with both astrocytic and microglial cell
morphology demonstrated a more intense signal of silver grain
deposition with the antisense probe (Fig. 3A,
d) compared with the sense probe (Fig. 3A,
c). Unstimulated cells had a very low background of silver
grain deposition with either antisense (Fig. 3A,
b) or sense probe (Fig. 3A, a). The
silver grain deposition in the cells was semi-quantitatively analyzed
using a Macintosh computer and NIH Image Program. The results are shown
in Fig. 3B. As with the Northern blot analysis, the iNOS
mRNA was induced within 2 h after cytokine stimulation and
back to base line by 48 h.
The Kinetics of iNOS Protein Synthesis in Human Fetal Mixed Glial Cultures
Western blot analysis was used to detect iNOS protein
synthesis in mixed glial cultures. Using a polyclonal antibody directed against an iNOS-specific peptide, we detected a band of iNOS protein at
130 kDa in all lanes. The band was significantly increased in IL-1
and IFN-
-treated cultures (Fig. 4). A significant
level of iNOS protein was detected at day 1, reaching a maximal level at day 3 and decreasing by day 5. The faint band found in unstimulated control cultures suggested a very low level of iNOS protein synthesis in these unstimulated cells (Fig. 4).
The induction of iNOS protein was further assessed by
immunohistochemical staining. The unstimulated and
IL-1/IFN-
-stimulated cultures were stained with the polyclonal
antibody to the synthetic human iNOS peptide to assess the presence of
iNOS from day 1 to 7. iNOS staining of cultures was seen at 24 h
and was maintained through day 7 (Fig. 5A,
a and b; day 3 as an example). Some punctate staining was also found in a few unstimulated astrocytes or microglia (Fig. 5A, a and c). Normal rabbit IgG
or no primary antibody gave no staining (data not shown). iNOS protein
staining was also found in the enriched microglial cultures stimulated
with cytokines (Fig. 5A, c and d). The
staining of iNOS in the mixed glial cell cultures was semi-quantitated,
and the results are shown in Fig. 5B.
The Kinetics of iNOS Enzyme Activity and NO Production
As
determined by Griess reagent reaction, no significant increase of NO
was detected in supernatants of IL-1/IFN-
-stimulated cultures
until day 3. NO levels continued to accumulate over the remainder of
the 7 days, reaching the highest level of more than 40 µM
(Fig. 6A). The NO accumulation was diminished
52% by an arginine analogue
N
-nitro-L-arginine methyl ester treatment
(data not shown). Even though unstimulated cultures expressed low
levels of iNOS protein in the first 48 h, no NO above these
base-line levels in the medium (8 µM) was apparently
released into extracellular medium (Fig. 6A). Similar NO
production was also detected in the enriched microglial cultures (Fig.
6B). These results imply that NOx
production occurs
48-72 h after iNOS enzyme synthesis.
We extracted the cytosolic iNOS from cytokine-treated or untreated
cells and analyzed its efficiency in converting L-arginine to L-citrulline in the presence of substrate
L-arginine and the cofactors NADPH, BH4, FMN, and FAD at
non-limiting concentrations. With the addition of substrate and iNOS
cofactors, cytokine-stimulated cells showed a small to insignificant
iNOS enzyme activity by day 1, which was not detectable in controls.
The citrulline level peaked from day 3 to day 5, and returned to base
line at day 7. This suggested that the enzyme produced could be
functional if cofactors or substrate were not limiting (Fig.
7).
The Kinetics of NO-mediated Events
The 2-3-day lag in NO
production compared with iNOS production prompted us to assess iNOS
function in a separate assay. Since NADPH-d staining has been routinely
taken as a nonspecific histochemical indicator of NOS catalytic
activity, especially after formaldehyde fixation of cells or tissue, we
stained for NADPH-d in cytokine-stimulated and unstimulated mixed glial
cultures. There was no significant amount of NADPH-d staining over base
line detected between day 1 and 3. NADPH-d activity was eventually seen
3-4 days after stimulation (Fig. 8A,
a and b; day 3 is shown as an example). The
kinetics of NADPH-d were similar to NO and NO-mediated events, again
suggesting a delayed onset of iNOS catalytic activity (Fig.
8B, a).
The kinetics of NO-mediated events in human
fetal mixed glial cultures. Glial cells were incubated with 5 ng/ml human recombinant IL-1 plus 70 ng/ml human recombinant IFN-
from day 1 to day 7 in mixed glial cultures. The photographs of day 3 are shown as example. A: a and b,
NADPH-diaphorase staining; c and d, nitrotyrosine
staining; a and c, unstimulated cultures;
b and d, stimulated cultures. B,
semi-quantitative measurement of the stainings. a,
NADPH-diaphorase; b, nitrotyrosine. C,
radioimmunoassay of cGMP levels. Means ± S.D. are shown for eight different cultured fetal brains (*, p < 0.01, ANOVA). The bar
represents 50 µm.
NT residues in proteins are considered to be specific "footprints"
of peroxynitrite (ONOO), a powerful oxidant and cytotoxic
species formed by the rapid reaction between NO and superoxide anion
(O2
) (34). ONOO
and
other reactive nitrogen species would act as nitrosating agents to
modify intracellular tyrosine to form 3-NT and dityrosine (35-37).
Thus, NT has also been often used as a "footprint" for NO. We used
a polyclonal antibody against NT to detect NT immunoreactivity, and
found no significant increase of NT staining over the base line until 3 days after cytokine stimulation. The staining increased reaching the
highest level at 7 days post-stimulation (Fig. 8A, c and d; day 3 is shown as an example). No such
staining was seen in unstimulated cultures. The staining
intensity was semi-quantitated by double-blind examiners
(Fig. 8B, b).
Since NO has been reported to stimulate activity of soluble guanylate cyclase, and hence increase the production of cGMP (38, 39), we measured cGMP levels by radioimmunoassay. As with NO production and NADPH diaphorase and NT staining, significant levels of cGMP were not detected until 3 days after cytokine stimulation, reaching a day 7 peak (Fig. 8C).
Tetrahydrobiopterin Increases NO Production in Human Fetal Mixed Glial CulturesTo test the hypothesis that the delay of NO
production in human glia might be due to insufficient endogenous iNOS
substrate or cofactors, we added a range of concentrations of
L-arginine and FAD, FMN, NADPH, heme, and BH4
into the cultures alone or in combinations. The exogenous
BH4 increased NO levels in a dose-dependent (Fig. 9A) and time-dependent
(Fig. 9B) manner. As indicated by Fig. 9B, after
BH4 was added at an optimal concentration of 150 µM, significant NO levels were detected at 24 h and
elevated levels of NO were seen at all subsequent time points compared
with cultures not receiving BH4. Addition of substrate or any other
cofactors, either alone or in various combinations, had no effect upon
NO production in these cultures (data not shown). These results
demonstrate that BH4 levels may be limited in human glial
cells.
This is the first complete kinetic study comparing iNOS transcription, translation, enzyme activity, NO production, and NO-mediated events in both human astrocytes and microglia in response to cytokine stimulation as well as being the first study providing in situ hybridization and immunocytochemical staining evidence of iNOS expression in both cell types. By comparing the kinetics of these events, we found that iNOS was transcribed within 2 h after cytokine stimulation, followed by the translation of iNOS protein within 24 h. However, the iNOS catalytic activity, NO production, and NO-mediated events were not detected until 3 days after stimulation. Addition of iNOS cofactor BH4 significantly increased NO level in a dose- and time-dependent manner, suggesting that this cofactor works as a post-translational regulator of iNOS function and may be limiting in human glial cells at least in vitro. Moreover, we verified that, in addition to astrocytes, human microglia are capable of synthesizing iNOS and producing NO upon cytokine stimulation.
Many human cells such as hepatocytes, chondrocytes, endothelial cells,
and glioblastoma cells have been shown to express iNOS and release
large amounts of NO after stimulation with bacterial products and/or
cytokines, showing kinetic patterns similar to those of rodent glial
cells (40-43). The present study has demonstrated that normal human
glial cells display similar transcription and translation of iNOS as
other human cell types or rodent glia, but human glial iNOS is not
immediately functional in these cells; therefore, NO production has a
delayed kinetics. As with other human cells, iNOS mRNA in human
glial cells reaches a maximum level at 3-6 h after stimulation,
followed shortly thereafter by the translation into protein. As in
rodents, iNOS protein is stable in glia in vitro and can be
maintained in cells in culture for several days. This may be the result
of a lack of degradation of iNOS protein in vitro, an event
that can be mediated by cytokines like TGF (44). Nevertheless, while
NO synthesis occurs in rodent glial cells within 24 h
post-stimulation (8, 16, 45), significant levels of NO and NO-related
events in human glial cells were not detected until 3 days after
stimulation, which was at least 48 h after the appearance of iNOS
protein. In other human cells that have been studied in
vitro (40-43), levels of iNOS mRNA begin declining by 8 h. However, the steady state iNOS mRNA levels was prolonged out to
48 h in this study of human glial cells. Because NO (46) and other
factors like TGF
(44) suppress iNOS mRNA expression or inhibit
iNOS mRNA stability, and these cultures contained neither in the
first 2-3 days, the iNOS mRNA levels may have been maintained.
The lag in NO production following the appearance of iNOS suggested a block in functional iNOS enzyme activity. While NO release into culture supernatants became significant between day 3 and 4, the iNOS protein in homogenized cytosol displayed significantly more catalytic activity at earlier time points (days 2-3) in the presence of added substrate and cofactors. This suggested that a deficiency of either substrate or cofactors in glial cells might have led to the delay in functional enzyme. The addition of the substrate L-arginine into the cultures had no significant effect on NO levels. Recent studies have suggested that NO synthase monomers require dimerization for the enzyme to be functional (47, 48). The binding of heme and BH4 play a significant role in forming and stabilizing active dimeric NOS (48-51). Interestingly, other studies have also demonstrated that de novo synthesis of BH4 can be strongly stimulated by cytokines in human macrophages (52). BH4 synthesis was found to be required for iNOS-derived NO production in human endothelial cells (53). Consistent with these studies, we demonstrated here that addition of BH4, but not other cofactors, induced NO at time points when none was previously detected above background (days 1 and 2) and increased NO levels at other time points, shifting the kinetics of NO induction to time points more consistent with iNOS presence in cells. We are currently measuring BH4 levels and iNOS monomer/dimer ratios in these cells to determine if post-translational modification is the limiting step in NO production.
Unlike murine macrophages, which produce high levels of NO (54, 55),
human mononuclear phagocytes produce low levels of NO in
vitro after cytokine stimulation (56-58), although significant amounts of NO were found in patients with inflammatory diseases (59).
Post-translational regulation of iNOS by cofactors has been suggested
to be important in the human macrophage. BH4 production is
low in macrophages that lack 6-pyruvoyl-tetrahydropterin synthase; nevertheless, addition of exogenous BH4 does not enable
macrophages to produce an increased amount of NO, suggesting additional
regulatory events in these cells, at least in vitro (57). A
second issue critical in the induction of NO in human macrophages is
the stimulation used to induce iNOS. Cross-linking of CD23 or CD69 as
well as combinations of cytokines like GM-CSF, IL-4, IFN-, and
IL-1
may be optimal inducers in macrophages, thereby distinguishing them from glia.
There is accumulating in vitro and in vivo
evidence that, as we have shown here, human microglia are capable of
synthesizing iNOS and produce 10-20 µM NO in response to
various stimuli (23, 24, 46). However, recent studies by Liu et
al. (60), as well as previous reports from the same laboratory
(21, 22), failed to detect iNOS in human microglia. Moreover, in
contrast to the observations reported here, the Liu study using
purified astrocytes showed a delay in both iNOS mRNA and protein
expression, but no lag in nitrite production. IL-4 and TGF failed to
inhibit both iNOS expression and nitrite production by human astrocytes in the Liu study. However, we have shown that IL-4 and TGF
inhibit NO production by human glial cells (61). Both the culture conditions (we used IL-3, M-CSF, and GM-CSF to expand microglia) and the microglial cell phenotype were different in the two studies (27, 60).
Furthermore, in our studies, microglia were cultured with astrocytes
for up to 2 weeks before harvest, whereas in the Liu study, microglia
and astrocytes were separated from the beginning of the culture period
(60). Finally, different cytokine doses were used in the two studies.
Since the antibody reagents utilized in the two studies were also
different, it is difficult to assess the nature of the cell
localization of iNOS in their study compared with ours. Perivascular
and parenchymal microglia do not have completely identical phenotype.
Perivascular microglia are CD4-positive and are derived
from blood-borne macrophages, while the parenchymal microglia which we
have studied here are CD4-negative (27). The perivascular,
CD4-positive, microglia used in the Liu study may thus
explain their failure to show NO production (21, 22, 60).
We have consistently seen very low basal levels of iNOS mRNA and protein in unstimulated cultures. This may be the consequence of 1) activation by culture conditions, 2) stimulation of cells during the process of mechanical disruption into a single cell suspension which simulates injury, or 3) the fact that these cells came from an actively developing fetal brain, where iNOS may play a role. Nevertheless, there is also evidence suggesting that iNOS might be present in some normal adult and fetal tissues (62, 63).
In summary, we have demonstrated that human astrocytes and microglia are capable of synthesizing iNOS and NO in response to cytokine stimulation. The potential deficiency of the iNOS cofactor tetrahydrobiopterin in these human glial cells may have resulted in a block in the post-translational dimerization necessary for iNOS catalytic activity and thus a delay of NO production. Further studies are necessary to understand the role of this cofactor in regulating NO production in human glial cells. Such studies will be valuable in understanding the physiological and pathological role of NO in human neurological disorders.
We thank Genentech Inc., South San Francisco,
CA for the supply of human IFN-gamma] and Immunex Corp., Seattle, WA
for human IL-1.