Infectious Diseases Section, Department of Veterans Affairs Western New York Healthcare System, State University of New York at Buffalo, School of Medicine
Correspondence: Charles S. Berenson, Division of Infectious Diseases (151), VA Western NY Healthcare System, 3495 Bailey Avenue, Buffalo, NY 14215. E-mail: berenson{at}acsu.buffalo.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Conditions
were established for selective, reversible depletion of HMG with
D-threo-(R,R)-1-phenyl-2-decanoyl-amino-3-morpholine-1-propa-nol.
mAb25F4 had diminished recognition for ganglioside-depleted
macrophages, which was restored with regeneration of gangliosides.
Although desialylation of HMG did not impair mAb25F4 recognition,
enzymatic cleavage of ceramide abolished antibody binding. Antibody
recognition was specific for the ceramide fraction, with preferential
recognition for ceramide of HMG and murine macrophage gangliosides and
limited recognition for neural tissue ceramide and gangliosides. This
study underscores the importance of structurally distinct ceramide of
macrophage gangliosides as a critical domain for ganglioside-mediated
activation of human macrophages.
Key Words: glycosphingolipids • mononuclear phagocytes • cytokines • PDMP
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Conspicuous structural differences of gangliosides of macrophages have been recently appreciated and may define their immunologic properties. Macrophages of LPS-hyporesponsive mice are deficient in the sialidase-accessible form of GM1 (GM1b) compared with macrophages of their congenic, LPS-responsive counterparts [7 ]. The ceramide component of gangliosides of macrophages also differs structurally from ceramide of gangliosides of neural origin. Although ceramide species of both possess C18 sphingosine chains, the fatty acyl chain of ceramide of murine macrophage gangliosides (MMG) is predominantly comprised of C16 and C24 fatty acids, and ceramide of gangliosides of neural tissue contains mostly C18 fatty acids [7 , 8 ]. Recent studies confirmed that the same structurally distinct ceramide of gangliosides of murine-immune cells is also a constituent of gangliosides of human mononuclear cells [9 ]. These marked, structural distinctions may explain the potent, immunologic effects of macrophage gangliosides.
To better study the immunologic effects of macrophage gangliosides, we produced a monoclonal antibody (mAb25F4) to human macrophage gangliosides (HMG), which we have previously reported inhibits migration of human macrophages [10 ]. mAb25F4 further recognizes a cell surface epitope on human macrophages, which although present, is inaccessible to surface molecules in adults with advanced HIV infection [11 ]. Earlier studies indicated common recognition by mAb25F4 of the three major gangliosides of human macrophages, identified as GM3, sialosylparagloboside, and an extended monosialosylhexosylceramide [10 , 12 ]. Although this suggested that the critical molecular domain for 25F4 recognition was a molecular structure common to all three major HMG, the specific domain was not known. Further, the range of immunomodulatory properties of mAb25F4, beyond macrophage migration inhibition, was not known.
We theorized that ganglioside-mediated macrophage activation by mAb25F4 is not limited to migration inhibition and is mediated by specific recognition of a critical domain of the ganglioside molecule, and we performed studies to test this hypothesis.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Monoclonal antibodies 25F4 [immunoglobulin G (IgG)2a] and 21C8 were produced by immunization of Balb/c mice with purified HMG and purification by affinity chromatography, as previously detailed [10 ]. Hybridomas were products of fusion of splenocytes and SP-2 myeloma cells. mAb5F3, generously donated by Dr. Timothy Murphy (SUNY, Buffalo, NY), recognizes an epitope of the P6 outer membrane protein of nontypeable Haemophilus influenzae and served as an irrelevant IgG2a. All antibodies were tested for endotoxin content by Limulus amebocyte lysate assay (Sigma Chemical Co., St. Louis, MO), which detected <10 pg/ml.
Purification of human macrophages
Human mononuclear phagocytes were purified from buffy coat
suspensions, obtained from healthy, HIV-seronegative volunteers from
the Red Cross of Western New York. Cells for individual
experiments were from single donors and were not pooled unless
specified. Mononuclear cells were further purified by Ficoll-Hypaque
density centrifugation and seeded onto 100-mm glass petri dishes
(5x106 cells/ml) in RPMI 1640 supplemented with 10%
heat-inactivated human AB-positive serum. After incubation at 5%
CO2, 95% humidity, 37°C for 7 days, nonadherent cells
were removed with serial rinses of warm (37°C) phosphate-buffered
saline (PBS). Remaining monocyte-derived macrophages were incubated in
RPMI 1640 with 10% fetal calf serum and were consistently 98–100%
esterase-positive [10
].
Purification of murine macrophages
Murine peritoneal macrophages were obtained from 6- to
8-week-old female C3N/HeN mice, as previously described
[13
]. Mice were killed four days following
intraperitoneal injection of 1 ml 10% thioglycollate broth per mouse.
Cells were recovered by peritoneal lavage with Hanks’ balanced salt
solution. Cell suspensions were centrifuged at 10°C at 1000 RPM (200
g) for 10 min and resuspended. Approximately 25 x 106
cells in 10 ml were incubated on glass petri dishes for 90 min at
37°C with 5% CO2 and 95% humidity to permit adherence
of macrophages. Cellular lipids were extracted following removal of
nonadherent cells and rinse of adherent cells with 0.31 M
pentaerythritol to remove excess salts.
Purification of gangliosides
The ganglioside fraction was purified from the total lipid
extract as previously detailed [13
]. Lipid extracts were
filtered through a scintered glass filter funnel overlaid with a glass
fiber mat and applied to a column of diethylaminoethyl-Sephadex A-25
(Sigma Chemical Co.). Neutral lipids were eluted with
chloroform:methanol:water (30:60:8 v/v/v), and the
ganglioside-containing acidic lipid fraction was eluted with
chloroform:methanol:0.8 M sodium acetate (30:60:8 v/v/v). Gangliosides
were further purified by serial chromatographic steps as previously
described [14
].
Cytokine assays
Adherent human macrophages derived from 106
cells/well were treated with mAb25F4, mAb21C8, or mAb5F3 (25 µg/ml).
Cellular supernatants were assayed by sandwich-type, solid-phase
immunoassay for interleukin (IL)-1ß (R&D Systems, Inc., Minneapolis,
MN) or tumor necrosis factor (TNF-; Biosource International,
Inc., Camarillo, CA). Supernatants of treated cells were harvested at
appropriate timepoints, centrifuged, and removed from pelleted cells
before being frozen. Standards and samples (200 µl) were added to
individual wells precoated with mAb to IL-1ß and incubated at room
temperature for 2 h. After washing three times, 200 µl
polyclonal antibody to IL-1ß (or to TNF-), conjugated to
horseradish peroxidase, was incubated in each well for 20 min at room
temperature. Acidic "stop" solution was added, followed by
developer, and optical density (OD) of each well was measured at 450 nm
within 30 min by enzyme-linked immunosorbent assay (ELISA) reader. All
experiments included separate wells treated with Eschericha
coli K235 LPS (1 µg/ml; Sigma Chemical Co.) or with purified
SP-2 supernatants as additional controls and were repeated with
cells from three separate donors. All conditions were performed in
quadruplicate.
Ganglioside depletion
Depletion of cellular GSL was achieved with
D-threo-(R,R)-1-phenyl-2-decanoylamino-3-morpholine-1-propanol
(PDMP; Matreya, Inc., Pleasant Gap, PA), which competitively inhibits
synthesis of glycosylceramide from ceramide and UDP-glc
[15
, 16
]. For each analysis, 800–900 x 106 total peripheral blood mononuclear cells (PBMC) were
used for each condition. Of these, 5–10% are monocytes, which
differentiate into monocyte-derived macrophages [10
].
Time and dose kinetics for selective depletion of macrophage
gangliosides, without depletion of neutral lipids, were determined by
treating cells with 0–100 µM PDMP for 1–5 days. Neutral lipid and
ganglioside fractions were independently collected from purified
cellular lipids. Neutral lipids were placed on TLC, run in solvent of
chloroform:methanol:water (65:35:8 v/v/v) for 42 min, and visualized
with orcinol spray [17
]. Gangliosides were also placed
on chromatograms, run in solvent of chloroform:methanol:0.25% KCl
(50:45:10 v/v/v) for 45 min, and visualized by resorcinol spraying
[10
]. The major neutral lipid and the major ganglioside,
previously identified as paragloboside (nLc4) and GM3,
respectively [11
], were quantitated by scanning
densitometry (Molecular Dynamics, Sunnyvale CA). As previously
described, volumes of peaks were quantitatively determined by
densitometry, based on area and light absorbance. We have determined
that relative quantity maintains a linear relationship with OD at least
from 0 to 20 µg sialic acid [18
]. Although
GM3 and nLc4 were measured as representatives of their
respective lipid fractions, careful note was made of all minor GSL to
be certain of consistency. Results are expressed as (volume of
PDMP-treated GSL/volume of untreated GSL) x 100 and were repeated
with cells from three separate donors.
In selected experiments, PDMP-containing media was removed from treated macrophages and replaced with PDMP-free media. Cells were then maintained in culture for 3–5 days before harvesting lipids. Gangliosides of untreated cells, PDMP-treated cells, and cells cultured for an additional 3–5 days post-removal of PDMP were quantitated.
Immunofluorescent surface-labeling of PDMP-treated human
macrophages
Human PBMC (2x106/well) were grown in Lab-Tek
chamber slides (Nunc Inc., Naperville, IL) for 10 days. Adherent
macrophages were incubated with PDMP as described above. Cells were
then treated with 6% nonfat dry milk for 30 min at 37°C, rinsed, and
incubated with mAb25F4 (40 µg/ml) for 60 min at room temperature.
After removal of excess antibody with vigorous washes, cells were
incubated with fluorescein isothiocyanate-F(ab')2 fragments
of goat anti-mouse IgG (1:50; Zymed Laboratories, San Francisco, CA)
for 30 min at room temperature in a dark container. Cells were washed
and evaluated by immunofluorescent microscopy. In all experiments,
separate wells of the same donor’s cells, cultured on the same chamber
slide, were treated with protein G-purified SP-2 cell supernatant or
with an irrelevant IgG2a. Both negative controls have had identical
background immunofluorescence to one another in previous studies
[10
].
In selected experiments, PDMP-containing media was removed from treated macrophages and replaced with PDMP-free media. Cells were then maintained in culture for 3–5 days before immunostaining with mAb25F4, as described above. Gangliosides of untreated cells, PDMP-treated cells, and cells cultured for an additional 3–5 days post-removal of PDMP were evaluated by immunofluorescent microscopy.
Radioisotope labeling and detection of glycolipids of PDMP-treated
macrophages
Single-donor macrophages were isolated by density gradient
centrifugation and adherence, as described earlier. Equal numbers of
cells in culture were treated with 50 µM PDMP or with PDMP-free
medium for 5 days. On day 3 of PDMP incubation, all cells were
incubated with 14C-galactose (American Radiolabel
Chemicals, St. Louis, MO; specific activity, 55 mCi/mmol) at 5
µCi/20 x 106 cells for 48 h before extracting
cellular glycolipids [19
]. Purified neutral lipid and
ganglioside fractions were isolated. Five percent of each fraction was
analyzed by scintillation counting, and identical percentages of each
sample were plated on TLCs and run in chloroform:methanol:water
(65:35:8 v/v/v) for 42 min for neutral lipids and
chloroform:methanol:0.25% KCl (50:45:10 v/v/v) for 45 min for
gangliosides. 14C-galactose-labeled glycolipids were
visualized by autoradiography and created by exposure of chromatograms
to hypersensitized XAR-5 film (Eastman Kodak Co., Rochester, NY) for 1
week at -70°C. X-ray film was hypersensitized by exposure to 7%
hydrogen/93% nitrogen at 48°C for 16 h, as previously described
[20
]. Major glycolipids were localized by spraying TLC
plates with resorcinol for gangliosides and orcinol for neutral lipids.
Autoradiographic bands were then aligned by overlay with visible bands
on each TLC plate, which included known standards.
Sialidase treatment of macrophage gangliosides
HMG containing 5–7 µg sialic acid were incubated with
Arthrobacter ureafaciens sialidase (200 mU/ml; EY
Laboratories, San Mateo, CA) in 1 ml 100 mM sodium acetate buffer, pH
4.8, at 37°C for 18 h, following the method of Saito et al.
[21
]. A duplicate sample of equal quantity of
gangliosides was incubated in buffer alone. Reactions were terminated
with addition of 0.1 M NaOH and neutralized with 0.1 M HCl. Solutions
were desalted on SepPak (Waters Assoc., Milford, MA) columns, tested
for purity on TLCs, and run in chloroform:methanol:0.25% KCl
(50:45:10) by resorcinol-spraying. Volume was determined by
quantitation of resorcinol-positive intensity with scanning
densitometry, and the percent desialylation was expressed as 1 -
[volume of (resorcinol-positive) sialidase-treated sample/volume of
(resorcinol-positive) untreated sample] x 100.
Ceramide-glycanase treatment of macrophage gangliosides
HMG (5–7 µg) were incubated in 0.2 U Marobdella
decora ceramide glycanase [22
, 23
]
(V-Labs, Covington, LA) in 200 µl 50 mM sodium acetate buffer, pH
5.0, containing 0.75 µg/µl sodium cholate at 37°C for 16 h.
The reaction was terminated by addition of 5 vol chloroform:methanol
(2:1 v/v). Cleaved products were separated into lower
(ceramide-containing) organic and upper (carbohydrate-containing)
aqueous phases by gentle centrifugation. Each phase was dried
separately. After separation, the aqueous phase was additionally rinsed
with chloroform:methanol (2:1) to optimize purity.
Released oligosaccharide was detected by resorcinol spraying on TLC run in n-butanol:acetic acid:H2O (2:1:1, v/v/v) for 80 min. Ceramides were independently assayed on TLC run in chloroform:methanol (9:1 v/v) for 20 min and developed with Coomassie blue. To verify absence of uncleaved, sialylated substrate, the organic (ceramide-containing) phase was also run on TLC in chloroform:methanol (9:1) and developed with resorcinol spray.
ELISA
Each well of a 96-well microtiter plate (Immunlon-1, Dynatech
Laboratories, Inc., Chantilly, VA) was coated with substrate dissolved
in 100% ethanol and left to dry for 16 h. Substrates included
HMG; mixed neural-tissue ceramides; purified GM3,
GM1, GD3, GD1b, GD1a
all from neural tissue (Matreya, Inc.); sialidase-treated HMG; organic
and aqueous phases of ceramide glycanase-treated HMG; MBG; and
thioglycollate-elicited peritoneal MMG.
Glycolipid or oligosaccharide substrates were coated onto wells at 0.3 µg/well. Cleaved fractions of enzyme-treated HMG were coated in equimolar ratios of original product. Wells were rinsed with PBS/0.05% Tween 20 and incubated with 3% bovine serum albumin at room temperature for 1 h. After vigorous rinses with PBS/0.05% Tween 20, 50 µl purified mAb (10 µg/ml) was added and incubated for 3 h at 37°C. Wells were again rinsed, and 50 µl (1:1000) goat anti-mouse IgG peroxidase (Kierkegaard and Perry, Gaithersburg, MD) diluted in PBS was added for 1 h at room temperature. After rinsing, 50 µl 3,4,5-trimethoxybenzoic acid-developing buffer (0.1 µg/ml) was added. Reaction was stopped after 5–10 min with 4 N H2SO4 and read on a 96-well microplate reader (BioRad Instruments, Hercules, CA) at 450 nm. A negative control was performed and included substitution of SP-2 tissue culture supernatant for the primary antibody. Known positive controls were also run simultaneously. For all comparative ELISAs, results are expressed as [(OD substrate-OD background)/(OD macrophage gangliosides-OD background)] x 100. Background was determined by including empty wells with no substrate in every experiment.
To assess adherence of oligosaccharides and ceramides to ELISA wells and to determine if potential differential configurations would affect binding, TLC immunostaining was also done with mAb25F4 and 21C8 on the oligosaccharide and ceramide fractions derived from treatment of macrophage gangliosides with ceramide glycanase. Oligosaccharide fractions were run in butanol:acetic acid:H2O (2:1:1 v/v/v), and ceramides were run in chloroform:methanol (9:1 v/v) as detailed earlier. Immunostaining on TLCs was performed as described previously to confirm ELISA results [10 ].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
 View larger version (12K): [in a new window] |
Figure 1. mAb25F4-induced IL-1ß release from human macrophages. Peak responses
for antibody 25F4 ( ; 25 µg/ml) occurred at 24–48 h. Negligible
responses were seen with mAb21C8 ( ), mAb5F3 ( ), or with purified
SP-2 supernatant. Each time point represents means of three separate
measurements ± SE and was reproducible for cells from
three separate donors. To confirm macrophage reactivity, macrophages
from each donor were also treated with LPS ( ; 1 µg/ml) at each
time point.
|
production from human macrophages (Fig. 2
). Peak levels of 25F4-induced TNF- levels were seen at 24 h (1321.8±48.3 pg/ml) with cells from three donors. As with IL-1ß
production, measurements (mean±SE) were again taken from
quadruplicate samples of each donor. Treatment with mAb21C8, mAb5F3, or
SP-2 supernatant had a minimal effect on TNF- production.
Macrophages also produced TNF- in response to LPS (1 µg/ml) in all
experiments.
 View larger version (11K): [in a new window] |
Figure 2. mAb25F4-induced TNF- release from human macrophages. As with IL-1ß
production, peak TNF- release with 25F4 () occurred at 24 h.
Negligible responses were seen with mAb21C8 (), mAb5F3 (), or
with purified SP-2 supernatant. Each time point represents means of
three separate measurements ± SE and was reproducible
for cells from three separate donors. LPS-treated controls (1 µg/ml)
are also represented ().
|
 View larger version (28K): [in a new window] |
Figure 3. Selective depletion of macrophage gangliosides with PDMP. Depletion of
macrophage gangliosides (GM3) over time is shown (A) with
10 µM (), 50 µM (), and 100 µM PDMP (). Depletion of
GM3 with increasing concentrations of PDMP (C) is shown for
3 days () and 5 days () of incubation. Optimal depletion occurred
in macrophages incubated in 50 µM PDMP for 5 days. Ganglioside
depletion is represented by quantitation of the major component
GM3. Results are means ± SE for three
separate measurements. TLC results of one experiment are shown (B).
Blurring of bands (GM3) is a result of intentional
overloading of the TLC to permit maximal detection (B). Samples on TLC
were incubated with PDMP for 5 days. Lanes are bracketed in pairs based
on the concentration of PDMP (0, 10, and 50 µM) with which cells were
treated. Lanes: 1, 0 µM PDMP; 2, 0 µM PDMP plus 5 days incubation
post-removal; 3, 10 µM PDMP; 4, 10 µM PDMP plus 5 days incubation
post-removal; 5, 50 µM PDMP; 6, 50 µM PDMP plus 5 days incubation
post-removal. Position of ganglioside GM3 standard is shown
at right.
|
 View larger version (29K): [in a new window] |
Figure 4. Selective depletion of macrophage neutral GSL with PDMP. Depletion of
the neutral GSL fraction of human macrophages over time (A) is shown
with 10 µM (), 50 µM (), and 100 µM PDMP (). Neutral
lipid depletion with increasing concentrations of PDMP (C) is shown for
3 days () and 5 days () of incubation. Neutral lipid depletion is
represented by quantitation of the major component, nLc4. Maximal
depletion occurred in macrophages incubated in 10 µM PDMP for 5 days.
Results are means ± SE for three separate
measurements. TLC results of one experiment are shown (B). Blurring of
bands (nLc4) is a result of intentional overloading of the TLC to
permit maximal detection (B). Conditions of samples on lanes 1–6 are
identical to those in Figure 3
. Position of paragloboside (nLc4)
standard is shown at right.
|
 View larger version (16K): [in a new window] |
Figure 5. Restoration of macrophage gangliosides with removal of PDMP from cell
culture. Macrophages were depleted of gangliosides by incubation with
50 µM PDMP for 5 days and then incubated in PDMP-free medium for an
additional 3 (left) or 5 (right) days. Quantitation of GM3
by scanning densitometry of baseline, untreated sample (solid bars), of
PDMP-treated sample (dark-gray bars), and of PDMP-treated sample 5 days
post-removal of PDMP (light-gray bars) is shown. Results are expressed
as a quantitative percentage of ganglioside (GM3) of the
untreated sample. Each bar represents mean ± SEM of
samples from three separate donors. Samples treated for 5 days
correspond with TLC samples of Figure 3
. Results are given as a
quantitative percentage of ganglioside of untreated macrophages and
were reproducible in three separate experiments.
|
 View larger version (46K): [in a new window] |
Figure 6. Selective depletion of macrophage gangliosides with PDMP detected by
14C-galactose-labeled GSL. Selective depletion of
gangliosides relative to major neutral lipids was confirmed with GSL of
14C-galactose-labeled macrophages in culture for
autoradiographic comparison. Identical volumetric percentages of
neutral lipids (left) and of gangliosides (right) were taken of
macrophages from the same donor. Material in lane A of each panel was
purified from untreated macrophages; material in lane B of each panel
was purified from PDMP-treated (50 µM) macrophages. As in Figures 3
and 4
, a comparison was made between the major ganglioside
(GM3) and the major neutral lipid (nLc4) components.
|
Results confirm a modest increase of the major neutral lipid (nLc4) to 124.9% compared with depletion of the major ganglioside (GM3) to 14.4% of baseline quantity in PDMP-treated macrophages. Thus, an alternative method to measure glycolipid content confirmed that treatment with 50 µM PDMP for 5 days resulted in relative depletion of gangliosides compared with neutral lipids in human macrophages.
Ganglioside-depleted human macrophages have diminished surface
recognition by mAb25F4
To confirm relative specificity of mAb25F4 for gangliosides in
intact human macrophages, ganglioside-depleted cells were immunostained
with mAb25F4 (25 µg/ml) as previously described [10
].
In all experiments, concurrent immunostaining with an irrelevant IgG2a
was performed on same-donor cells to confirm absence of
nonspecific binding (not shown). Macrophage gangliosides and neutral
GSL from each donor’s cells used for these studies were simultaneously
analyzed by TLC (described above) to confirm selective depletion of
gangliosides. In all experiments, ganglioside depletion resulted
in diminished immunofluorescent surface-labeling with mAb25F4
(Fig. 7
). The presence of adherent cells was always confirmed by light
microscopy.
 View larger version (58K): [in a new window] |
Figure 7. Immunofluorescent surface-labeling of human macrophages with mAb25F4.
Macrophages of the same donor were immunostained with mAb25F4. Figure
includes macrophages that were untreated (A), ganglioside-depleted with
50 µM PDMP for 5 days (B), or had PDMP removed and demonstrated
regeneration of gangliosides (C), as described in Figures 3
and 5
.
Images correspond with quantitative ganglioside data shown in Figure 5
.
Results shown are from one donor and were reproducible with macrophages
of three separate donors.
|
Recognition of desialylated macrophage gangliosides by mAb25F4
To determine if sialic acid is a critical component for mAb25F4
recognition, HMG were treated with A. ureafaciens sialidase
[24
]. Before performing ELISA studies, it was crucial to
determine purity of desialylated preparations. Equimolar volumes of
sialidase-treated and untreated samples, prepared on TLC and sprayed
with resorcinol, confirmed that 98% of sialic acid residues were
effectively removed by sialidase treatment (Fig. 8
). Concomitant incubation of gangliosides with buffer alone did not
cause degradation. Treatment of control samples showed that
GM3 was no longer detectable with resorcinol spray, and
GM1a and GM2 remained resorcinol-positive.
 View larger version (117K): [in a new window] |
Figure 8. A. ureafaciens sialidase-treated HMG. Conditions were
established for removal of 98% of sialic acid residues, determined by
densitometric quantitation of resorcinol-sprayed TLC. A
sialidase-treated sample is compared with untreated sample of equal
concentration. Lanes: 1, Sialidase-treated sample; 2, buffer-treated
sample; 3, MBG standard. Ganglioside standards are indicated along the
right margin as GM3 (M3), GM1 (M1),
GD1a (D), and GT1b (T).
|
 View larger version (12K): [in a new window] |
Figure 9. ELISA recognition by mAb25F4 of desialylated macrophage gangliosides.
Desialylated macrophage gangliosides were produced as in Figure 8
.
Equimolar concentrations of desialylated macrophage gangliosides were
placed in ELISA wells and were tested for recognition by mAb25F4 (10
µg/ml). Results are expressed as a percentage of the
OD450 of HMG for each substrate and are given as means ± SE of three separate experiments. Each value was
calculated with controls from the same experiment.
|
Recognition of ceramide-glycanase degradation products of HMG by
mAb25F4
To independently determine if mAb25F4 has preferential recognition
of the ceramide portion or of the oligosaccharide portion, HMG were
treated with M. decora ceramide glycanase. Purity and
completeness of enzymatic cleavage of enzyme-treated fractions were
confirmed by TLC. The aqueous (carbohydrate-containing) fraction was
analyzed on resorcinol-sprayed TLCs and consisted predominantly of a
single band with identical chromatographic mobility to the
sialyllactose standard (Fig. 10
). Untreated (nondegraded) HMG were run on the same TLC to
determine chromatographic position and confirmed absence of any
residual, uncleaved substrate in the enzyme-treated samples.
 View larger version (85K): [in a new window] |
Figure 10. M. decora ceramide glycanase-treated HMG. Conditions were
established for optimal cleavage of macrophage gangliosides into
ceramide-containing (A) and carbohydrate-containing (B) fractions. TLC
of ceramide fraction (left) includes lane 1, ceramide (organic
fraction) of ceramide glycanase-treated HMG; 2, HMG incubated with
buffer; 3, HMG standard; and 4, ceramide of bovine neural tissue. TLC
of carbohydrate fraction (right) includes lane 5, carbohydrate fraction
of ceramide glycanase-treated HMG; 6, aqueous fraction of HMG incubated
with buffer alone; 7, HMG standard; and 8, sialyllactose standard.
Solvents for each are given in Materials and Methods. Labels to the
right indicate positions of GM3 and of sialyllactose (SL)
standards. HMG standards were included on all TLCs to determine
chromatographic position to verify absence of uncleaved substrate.
|
Equimolar amounts of each enzymatically cleaved fraction were then tested in ELISA for recognition by mAb25F4. Results indicated no binding by mAb25F4 to components of the carbohydrate-containing fraction. In contrast, ELISA recognition was exclusively retained by the ceramide-containing fraction (Fig. 11 ).
 View larger version (16K): [in a new window] |
Figure 11. ELISA recognition by mAb25F4 of ceramide glycanase-degraded fractions
of macrophage gangliosides. Enzymatically degraded fractions of HMG
were produced with ceramide glycanase, as shown in Figure 10
. Equimolar
concentrations of the organic (HMG Ceramide Fraction) and of the
aqueous carbohydrate-containing HMG CHO Fraction] components were
placed in ELISA wells and tested for recognition by mAb25F4 (10
µg/ml). Substrates also included ceramide of bovine neural tissue,
selected gangliosides of bovine neural origin, MBG, and MMG. Results
are expressed as a percentage of the OD450 of HMG for each
substrate and are given as mean ± SE of three
separate experiments. Each value was calculated with controls from the
same experiment.
|
As before, wells containing untreated HMG were run with each experiment, and results are expressed as a percentage of the OD450 for equimolar amounts of HMG run on the same ELISA plate. All results were reproducible in three separate experiments. In all ELISAs, empty wells containing no substrate were included to correct for nonspecific background binding.
mAb21C8 also recognized macrophage gangliosides but did not elicit
IL-1ß or TNF- (Figs. 1
and 2)
. To determine the molecular domain
of mAb21C8, ELISA was also performed with this mAb toward
oligosaccharide and ceramide fractions derived from ceramide-glycanase
cleavage of HMG. In contrast to mAb25F4, mAb21C8 did not show
recognition of the ceramide fraction, but displayed preferential
binding to the oligosaccharide-containing fraction (Fig. 12
). Further studies to more precisely localize the binding domain of
mAb21C8 are in progress.
 View larger version (13K): [in a new window] |
Figure 12. ELISA recognition by mAb21C8 of ceramide glycanase-degraded fractions
of macrophage gangliosides. Enzymatically degraded fractions of HMG
were produced with ceramide glycanase, as shown in Figure 10
. Equimolar
concentrations of the organic (HMG Ceramide Fraction), of the aqueous
(HMG CHO Fraction), and of intact HMG were placed in ELISA wells and
tested for recognition by mAb21C8 (30 µg/ml). Results are expressed
as the OD450 for each substrate and are given as means ± SE of three separate experiments. Each value was
calculated with controls from the same experiment.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We pursued this investigation by eliciting immunologic responses of human macrophages with mAb produced expressly to gangliosides of human macrophages [10 ]. Our early studies indicated that the 25F4 epitope was surface-accessible and was absent in several nonmacrophage cell lines. Moreover, mAb25F4 was capable of inducing antibody-specific migration inhibition of human macrophages. These findings were intriguing because collective results of several earlier studies suggested that the receptor for macrophage migration inhibitory factor (MIF), a known inhibitor of macrophage migration, is a macrophage ganglioside [27 , 28 ]. More recently, MIF has been recognized as having far broader immunologic functions than were originally recognized [29 , 30 ]. In the current study, binding of mAb25F4 to surface-accessible macrophage gangliosides also induced macrophage cytokine production. Specificity was demonstrated through lack of response to irrelevant mAb. Thus, mAb25F4 also has broader immunoregulatory properties than were originally recognized. Immunologic activation via glycolipid receptors was first demonstrated in T cells with antibodies to GD3 and O-acetyl GD3 and represents a novel means of induction [31 , 32 ]. Our results indicate that endogenous gangliosides of human macrophages also possess receptor properties and further indicate that ceramide of macrophage gangliosides is critical to this function.
Our previous data indicated that mAb25F4 recognized an epitope common to each of the major gangliosides of human macrophages [12 ]. Although antibody cross-recognition of diverse ganglioside structures that differ in core carbohydrate composition, such as GM3 and sialosylparagloboside, has been described for an antimelanoma mAb [33 ], the specific molecular domain of ganglioside molecules recognized by mAb25F4 was not known. Our current studies indicate that 25F4 not only recognizes a surface epitope, but preferentially also recognizes the ceramide moiety. Further, this antibody preferentially recognizes gangliosides of macrophages of humans and mice and discriminates between these and gangliosides of neural tissue. Our conclusions are supported by preferential recognition of ceramide of HMG, which is structurally distinct from ceramide of neural origin; failure to abolish recognition with desialylation; absence of recognition of several gangliosides of neural origin; and interspecies cross-recognition of MMG, which share the ceramide structure of HMG [8 , 9 ].
Ceramides are heterogeneous molecules comprised of sphingosine and fatty acid chains. Ceramides of MMG are predominantly comprised of C18 sphingosine and C16 and C24 fatty acids [7 ]. The latter finding is remarkable because gangliosides of neural tissues are predominantly comprised of ceramide with alternative fatty acid constituents. Furthermore, ceramide comprised of C16 and C24 fatty acids is a shared, structural attribute of macrophage gangliosides of mice and of humans [9 ]. The structurally distinct ceramide moieties, shared by gangliosides of macrophages of different species, might be the structural basis for some of the immunologic properties of macrophage gangliosides. It is intriguing that ceramide comprised of C16 and C24 fatty acids is also present in B lymphocyte gangliosides [34 ]. Although our findings support the hypothesis that structural distinctions might explain some of the biologic attributes of gangliosides of immune cells, the immunologic properties of macrophage gangliosides may also be reliant on discrete, molecular conformations of ceramide of macrophages, which have not yet been determined.
While the roles of carbohydrate moieties of GSL as molecular determinants of GSL-mediated biological activity are well described, increasing evidence implicates ceramide structures of gangliosides as important immunoregulatory components. As with our findings, Ladisch et al. [35 ] reported that fatty acyl ceramide length was also the principal determinant of immunosuppressive activity of neuroblastoma GD2 ganglioside. They further identified cross-species ceramide activity, dependent on fatty acyl chain length, in murine and human cellular gangliosides. The immunologic recognition of mAb25F4 is also directed at gangliosides of macrophages of human and murine origin. Verotoxin binding to globotriaosylceramide is similarly affected by heterogeneity in the ceramide fatty acyl chain length, and the greatest binding capacity is held by ceramides with longer fatty acyl chains [36 ]. Besides serving as lipid anchors for protruding alkyl chains, recent mechanistic paradigms also support direct interaction of ceramides with signaling proteins [37 ]. This may also occur in models of synthetic membranes in which long ceramide fatty acyl chains protruded further from the surface of the lipid bilayer, potentially availing themselves to surface molecules [38 ]. Surface access to relatively hydrophobic regions of gangliosides also occurs with ceramide glycanase and ceramidase in intact cells [39 40 41 ]. Enzymatic access to the lipid moiety by ceramide glycanase may be a result of the relative hydrophobicity of the enzyme itself, binding directly to more hydrophobic sphingosine moieties [42 ]. Finally, structural mimicry between gangliosides and bacterial lipooligosaccharides and between their respective ceramide and lipid A components has been previously appreciated [43 ]. Remarkably, lipid A of lipooligosaccharide of H. influenzae, also previously believed to be inaccessible to surface molecules, is clearly accessible to mAb in intact bacteria [44 ]. These results collectively suggest that the fluidity of cellular membranes permits greater surface accessibility of ganglioside-bound ceramide than has been traditionally considered.
Ceramide itself has numerous immunologic properties, including
regulation of intracellular pathways for induction of apoptosis
[45
]. Free ceramide can be generated through
sphingomyelinase hydrolysis of GSL and is involved in multiple
signaling pathways [46
]. Signal transduction through
ceramide has, among its downstream targets, IL-2 and IL-6 and can
therefore affect immune response directly or through cytokine induction
[47
]. Selected ceramide species also regulate
differentiation of HL-60 cells and act as second messengers in signal
transduction [48
, 49
]. Ceramide-mediated,
intracellular signaling can also regulate mRNA transcription of the
c-myc protooncogene, as well as nuclear factor-
B expression
[50
, 51
].
Besides being a constituent of gangliosides, ceramide is also a component of nonsialylated (neutral) GSL. Therefore, it was critical to our studies to establish appropriate time and dose kinetics for selective depletion of gangliosides with PDMP. Metabolic inhibition of glucosyl-ceramide synthase has been a valuable tool for studying the roles of GSL in many cell systems [15 ]. In previous studies, refinements have permitted selective depletion of individual GSL components. Depletion of globo series GSL demonstrated decreased cellular adherence of P-fimbriated E. coli [52 ]. In fact, PDMP-mediated, metabolic inhibition in one study was so selectively regulated as to permit synthesis of neuroblastoma gangliosides yet to impair extracellular shedding of gangliosides [53 ]. Metabolic depletion of gangliosides and other GSL by ceramide-synthase inhibition in mice resulted most markedly in cellular depletion of lymphoid tissues, underscoring the importance of GSL in the biology of the immune system [54 ], prompting us to apply metabolic depletion of GSL to the study of human macrophages. By establishing conditions for selective depletion of macrophage gangliosides, our ganglioside-depletion experiments further support the immunologic importance of ganglioside-bound ceramide. This does not discount the potential immunologic significance of nonganglioside-bound ceramide, including ceramide of neutral GSL of macrophages, which will require independent verification.
The mechanism by which intracellular signaling is initiated by binding to macrophage gangliosides remains speculative. GSL-mediated cell signaling may originate from detergent-insoluble lipid rafts, acting as localized membrane sites for initiating signal transduction by selected pathways. These domains are enriched, not only in gangliosides but also in phospholipids and cholesterol. In fact, T lymphocyte activation is enhanced by clustering signaling molecules in membrane lipid domains [55 ]. Lipid rafts may be potent substrate pools for production of ceramide, which may further form microdomains and regulate downstream signaling [56 ]. Indeed, 25F4-epitope signaling may originate from stabilization of membrane lipid rafts, an aspect that will be the focus of ongoing experiments of our laboratory.
Studies of immunologic functions of GSL have often relied on exogenous, commercially available compounds, often derived from neural tissue. It is reasonable to presume that had we attempted these studies with mAb raised to gangliosides of other tissues, we would not have arrived at the same results. This investigation therefore supports the use of endogenous GSL of immune cells for studies of GSL immunoregulation and highlights the importance of ceramide of HMG as a critical domain for ganglioside-mediated activation of human macrophages.
Received August 11, 2001; revised March 28, 2002; accepted April 17, 2002.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  C. S. Berenson, T. F. Murphy, C. T. Wrona, and S. Sethi Outer Membrane Protein P6 of Nontypeable Haemophilus influenzae Is a Potent and Selective Inducer of Human Macrophage Proinflammatory Cytokines Infect. Immun., May 1, 2005; 73(5): 2728 - 2735. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
