close

Вход

Забыли?

вход по аккаунту

?

Importance of the IgG isotype not the state of glycosylation in determining human rheumatoid factor binding.

код для вставкиСкачать
800
IMPORTANCE OF THE IgG ISOTYPE, NOT THE STATE
OF GLYCOSYLATION, IN DETERMINING
HUMAN RHEUMATOID FACTOR BINDING
MARIANNA M. NEWKIRK. ANTONINA LEMMO. and JOYCE RAUCH
We investigated the influence of carbohydrate on
the binding of human rheumatoid factors (RF) to the Fc
fragment of IgC. The monoclonal RF studied were
derived from the serum of patients with mixed cryoglobulinemia or from hybridomas generated from patients
with rheumatoid arthritis (RA) and systemic lupus
erythematosus. Polyclonal RF were derived from patients with RA. The carbohydrate located on the Fc
fragment, regardless of whether it contained different
amounts of mannose or reduced amounts of galactose,
or was removed, did not affect the binding of the RF. In
contrast, the isotype of the Fc was found to be critical.
Two groups of hybridoma RF could be delineated. One
group bound preferentially to IgCl and/or IgG2, and a
second group (primarily from patients with RA) bound
preferentially to IgG3 and/or IgG4. Our results indicate
that the isotype of the Fc fragment, and not the extent of
galactosylation, influences the binding of the RF.
Human rheumatoid factors (RF), a group of
antibodies to gamma globulins, are highly associated
with rheumatoid arthritis (RA) (1). These antibodies
have been demonstrated to bind region(s) in the Fc
fragment of gamma globulins, with good evidence to
suggest the Cy2-Cy3 cleft as a likely site for binding
From the Department of Medicine, McGill University.
Montreal General Hospital Research Institute. Montreal. Quebec,
Canada.
Supported in part by grants from the Arthritis Society of
Canada.
Marianna M.Newkirk, PhD; Antonina Lemmo. BSc; Joyce
Rauch. PhD.
Address reprint requests to Marianna M. Newkirk. PhD,
Montreal General Hospital Research Institute. 1650 Cedar Avenue,
Montreal, Quebec H3G IA4. Canada.
Submitted for publication September 20. 1989; accepted in
revised form January 1 I . 1990.
Arthritis and Rheumatism, Vol. 33, No. 6 (June 1990)
( 2 4 ) . A number of recent studies have shown reduced
activity of galactosyltransferase in patients with RA
(3,
which results in fewer galactose molecules on the
Carbohydrate chain found at position 297 in the Fcy
fragment. This change in glycosylation, however, has
also been associated with both aging and chronic
inflammatory disease (6-8). In contrast, it appears that
during pregnancy, the activity of this transferase is
increased, resulting in elevated amounts of galactose
present on the IgG molecule (9). This may be a
consequence of or an influence in the remission of the
disease that is often observed during pregnancy (10). It
has been postulated from these studies that the low
amounts of galactose present on the Fcy portion of Ig
in the sera of RA patients may make this molecule a
better target for the autoantibody response.
In order to address this hypothesis, we have
analyzed both human monoclonal RF, isolated either
from the serum of patients with mixed cryoglobulinemia or from human hybridomas generated from patients with R A or systemic lupus erythematosus
(SLE). and polyclonal RF isolated from the serum of
patients with RA, for their ability to bind to Fcy
fragments. Fc fragments from both monoclonal and
polyclonal IgG. which had differing carbohydrate compositions at position 297. were isolated and used in the
RF binding studies. Our results indicate that the
isotype of the Fcy fragment, and not the extent of
galactosylation, influences the binding of the RF.
PATIENTS AND METHODS
Patients. Blood samples were obtained from patients
attending outpatient clinics of the rheumatology and immunology departments at the Montreal General Hospital. All
IgG ISOTYPE AND RF BINDING
RA patients fulfilled the American Rheumatism Association
(ARA) criteria for classic RA (11) and were RF positive.
SLE patients fulfilled the A R A revised criteria for the
classification of SLE (12). N o SLE patient had detectable
circulating RF at the time of cell fusion. Normal controls
were healthy volunteers with no history of autoimmune
disease.
RF. Monoclonal IgM-RF (BOR, KAS, RIV, GLO.
SWZ, and KEZ) were isolated and purified from the washed
cryoprecipitates of patients who had mixed cryoglobulinemia as previously described (13) (the plasma was kindly
donated by R. Wistar, Department of the Navy, Bethesda,
MD, D. Posnett, Cornell Medical School, New York, NY.
and D. Danoff, Department of Immunology, Montreal General Hospital). The IgM was >95% pure as analyzed by
sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE). Monoclonal IgM-RF were generated from
either R A patients or SLE patients by hybridoma technology. Briefly, peripheral blood mononuclear cells (PBMC)
were isolated on Ficoll-H ypaque gradients from venous
blood. The PBMC were fused with the GM 4672 human
lymphoblastoid cell line obtained from the Cell Repository
Institute of Medical Research (Camden, NJ) at a cell ratio of
1 : 1 with 44.4% polyethylene glycol (PEG). as previously
described (14). Hybridoma clones that produce IgM-RF
were found to be monospecific (clones 5501, 5602, 5603,
8808. and 18141; all derived from patients with RA) or
polyreactive (clones 1859 and 1872; derived from patients
with RA and clones 103.4. 1400, 1414, 1700, 1900, and 3019;
derived from patients with SLE). The monoclonal antibodies
were precipitated from cell culture supernatants after a 16
hour incubation at 4°C with 10% PEG 3350 (Sigma, St.
Louis, MO), by centrifugation at 17,000g for 30 minutes. The
precipitate was resuspended in phosphate bugered saline
(PBS) and dialyzed against PBS prior to gel filtration (13).
Polyclonal RF (COY, KLI, BEL. and DAL) were precipitated from the serum or plasma of RA patients using 10%
PEG as described above. Polyclonal IgM was separated
from IgG by gel filtration (13).
Fc fragments of human IgC. Human IgG obtained
from either patients with R A (IAL and LYN). a normal
control (NEW), or patients with monoclonal gammopathies
or hybridoma-generated IgGl (FLO. McC, EVI-15, EV2.7,
GOO. and MAY), and IgG2 (KOP) purified by protein A
affinity chromatography (Bio-Rad, Richmond, CA) was digested with mercury1 (HG) papain (1 : 100 [weightlweight];
Worthington Biochemicals. Technicon, Montreal, Quebec,
Canada) for 30 minutes at 37°C (15). Fc fragments were
purified by affinity chromatography using protein A and by
ion exchange chromatography (Fast Q column, 35 x 2.5 cm;
Pharmacia. Baie D'Urfe. Quebec, Canada) in 50 mM Tris
buffer, pH 8.0, using a linear gradient of 0 - 0 . 3 M NaCI.
To analyze the effect of mannose (complexity and/or
amount) on RF binding, Fc fragments (monoclonal IgGI)
were separated by a n i t y chromatography according to
their ability to bind to concanavalin A (Con AkSepharose
(Sigma). After washing the column with 50 mM Tris, pH 7.0,
O.15M NaCI, I mM CaCI,. and I mM MgCI,, the bound Fc
fragments were eluted with methyl-a-D-mannopyranoside
(Sigma).
Carbohydrate was removed from Fc preparations by
80 1
incubating with 125 mU/ml of N-glycanase (Genzyme, Boston, MA) in 0.10M sodium acetate, pH 6.0. 10 mM phenanthroline hydrate for 16 hours at 37°C. Native Fc conformation was maintained by not using detergent during the
enzymatic digestion.
Electrophoresis and immunoblotting. Immunoglobulins or fragments were separated by SDS-PAGE (16).
Carbohydrate staining was performed using periodic acidSchiff reagent (PAS), following the protocol outlined by
Pharmacia.
To determine the nature of galactose and mannose at
position 297 of the Fc fragments, immunoblotting was performed using peanut agglutinin, which binds specifically to
galactose (D[ + ] galactose, methyl galactoside) when sialic
acid is absent, and Con A, which binds preferentially to
mannose al-6 (mannose al-3) (17). Fc fragments derived
from patients with RA or normal controls were transferred to
nitrocellulose (Nicron Separations Inc., Fisher Scientific,
Montreal, Quebec, Canada) using the conditions outlined by
Towbin et al (18). After washing 3 times with PBS, 1.5 mM
monobasic potassium phosphate, 8 mM dibasic sodium
phosphate, O.15M NaCl, 2.6 mM KCI. 0.05% Tween 20, pH
7.4 (PBS-Tween) (except for the biotinylated peanut agglutinin, where 10 mM HEPES, pH 7.5. O.15M NaCI, 0.1 mM
CaCI,, 0.05% Tween 20 [HEPES-Tween] was used throughout), the blots were incubated with horseradish peroxidase
(HRPkonjugated Con A (Sigma), HRP-conjugated goat
anti-Fcy (Cappel, Organon Teknika, Scarborough, Ontario,
Canada), HRP-conjugated peanut agglutinin (Sigma), or biotinylated peanut agglutinin (Vector. Burlingame, CA) for 16
hours at room temperature. For the biotinylated peanut
agglutinin blots, a I-hour incubation with HRP-conjugated
avidin (Vector) followed 3 washes with HEPES-Tween. AU
blots were subsequently washed 3 times in the appropriate
buffer and then incubated with the HRP substrate, 4chloro, I-naphthol. The blots were scanned by densitometry,
and the relative amounts of Con A or peanut agglutinin
binding to the Fc fragments were calculated as a percent of
the total IgG present in the blot, as determined by HRPconjugated anti-y binding.
IgC isotypic composition of polyclonal Fc preparations. Four polyclonal Fcy preparations, 2 from different
patients with RA (both RF positive) and 2 that represented
different peaks from ion exchange chromatography of Fc
from the same normal individual, were characterized for IgG
isotype composition. Since the polyclonal IgG preparations
were initially purified by affinity chromatography using protein A, little or no IgG3 was predicted to be in the preparations. An enzyme-linked immunosorbent assay (ELISA) was
used in which the Fc preparations were adsorbed to the solid
phase (hmulon-I plates: Dynatech. Alexandria, VA) following standard protocols (19). After an incubation for 1
hour at 37°C with mouse monoclonal antibodies specific for
the 4 different human IgG isotypes (anti-IgG 1, anti-IgG4, and
total anti-IgG; Southern Biotechnology. Birmingham, AL,
and anti-IgG2 and anti-IgG3 clones HP60014 and 6050; American Type Culture Collection, Rockville, MD), the plates
were washed and incubated for 2 hours at 37°C with HRPconjugated goat anti-mouse antisera (Cappel). After 3
washes, the plates were incubated with substrate (ophenylenediamine) and the optical densities (OD) were read
NEWKIRK ET AL
802
at 492 nm using an ELISA plate reader (SLT-Labinstmments, model EAR 400RT: Fisher Scientific, Montreal,
Quebec, Canada). Composition was estimated by comparing
the OD with each antiisotype reagent with the OD detected
using the anti-y (total).
RF assays. An ELISA was used to detect RF binding
activity. Briefly, Fc fragments (10 &ml) were adsorbed to
the Immulon-I plates for 16 hours at 4°C. After the plates
were washed 3 times with PBS-Tween, 100 pl of purified
IgM-RF (0.1-2 &ml in PBS) or culture fluid (undiluted or
diluted with PBS-Tween) was added to the wells and incubated for 4 hours at room temperature. The plates were
washed with PBS-Tween, 100 pl of HRP-conjugated F(ab'),
fragments of goat anti-human IgM (Cappel) was added, and
the plates were incubated for 2 hours at room temperature.
The plates were washed with PBS-Tween and o-phenylenediamine was added. Once sufficient color had developed
(-30 minutes). the reaction was stopped with 4M H,SO,.
The OD was measured as above.
For isotype binding studies, an ELISA using 2 monoclonal antibodies of each human IgG subclass was conducted
as previously described (13). Since some of the RF bound
poorly to monoclonal IgG, a more sensitive assay using
biotinylated F(ab'), fragments of goat anti-human IgM (Jackson ImmunoResearch Laboratories, Mississauga. Ontario,
Canada) with an incubation of 2 hours at room temperature,
followed by HRP-conjugated avidin with an incubation of I
hour at room temperature was also used to detect the RF
that bound to the different isotypes of IgG.
RESULTS
We analyzed the binding characteristics of 18
different monoclonal human R F and 4 preparations of
polyclonal RF. We monitored the capacity of these R F
to bind to either monoclonal or polyclonal Fc fragments and assessed the influence of both the nature of
the carbohydrate (present o r absent) and the isotype of
the IgG. In order to assess the influence of carbohydrate located on the Fc fragment on the binding of the
RF, we made use of the ability of specific lectins to
bind to mannose o r galactose, in a strategy depicted in
Figure 1.
Is RF binding to Fc influenced by mannose?
Lntact human IgG had been previously shown to bind
to Con A with some heterogeneity of binding (20). We
used Con A-Sepharose as a tool to separate different
populations of mannose-bearing monoclonal Fc fragments of IgG1 or IgG2. No dramatic differences in the
separated fragments could be visualized in SDS-PAGE
using either a protein stain o r a carbohydrate stain
(data not shown). However, in a quantitative Western
blot, which compared Con A binding with anti-IgG
(total) binding (Figure 2), 3-fold differences in the
amounts or complexity of mannose were clearly
HeuSAc
I
Gal
I
G1cHac
I
Han
GlcHac
Ha
I
Eeu5Ac
I
Gal 4 Peanut Agg
I
GlcHac
I
Han 4 Con A
/
-
GlcHac
I
GlcEac-Fncose
I
G In-Tpr-Asn-Ser-Thr
H-glpcanase
297
Figure 1. Simplified depiction of the branched carbohydrate located at position 297 of the IgG molecule. Branching can be
extensive, and the presence of sialic acid (NeuSAc). often absent in
IgG, N-acetylglucosamine (GlcNac), and fucose is highly variable.
Arrows denote lectin binding sites for peanut agglutinin (Peanut
Agg) (which binds to galactose when sialic acid is missing) and
concanavalin A (ConA). and the site of N-glycanase digestion. Gal
= galactose; Man = mannose (6.22); Gln = glutamine; Tyr =
tyrosine; Asn = asparagine; Ser = serine; Thr = threonine. For a
more detailed depiction of this structure. see ref. 22.
present (column void volume 15.5%, column eluate
65%). There were no appreciable differences detected
in the ability of either monoclonal R F (BOR, KAS,
KEZ, RIV, SWZ, 103.4, 1400, 1900, and 3019) or
polyclonal R F (COY, KLI, DAL, and BEL) to bind to
F c fragments which differed in the amounts or complexity of mannose present on either the IgG1 or IgG2
subclass. Representative binding profiles are presented in Figure 3. All of the R F binding assays were
done using the R F at concentrations below maximum
absolute binding, so that we could discern any differences. Indeed, when standard dilution curves of the
R F binding to these Fc fragments were compared, the
lines were superimposable (data not shown).
IgG ISOTYPE AND RF BINDING
Con ASepharom Void
A
Con Aaepharose Eluate
0
h
A id-
t
\
Figure 2. Densitometric scans of Western blots of Fc (FLO) fractions separated by the ability of mannose on the Fc fragments to
bind to concanavalin A (Con A). A and C, Fc that was in the void
volume fraction of the Con ASepharose column; B and D, Fc that
was eluted from the Con A-Sepharose column. Horseradish peroxidase (HRP)-conjugated Con A binding to blots is depicted in A and
C, and HRP-codugated anti-total IgG binding to the blots is
depicted in B and D. Arrows denote origin of the gel.
803
carbohydrate and isotype composition. To correct for
the amount of protein present, we compared the
percent of lectin binding to anti-y binding, thereby
permitting comparisons between experiments performed on different days. The little variation seen in
our immuno/lectin blots was not quantitated since the
study size was small. Since a normal Fc pool was run
concurrently with every RA pool of Fc, comparisons
between normal Fc and RA Fc, and thus, between
normal and reduced amounts of lectin-accessible galactose, were possible.
Densitometric scans of Western blots probed
with Con A or peanut agglutinin revealed that the Fc
fragments from patients with RA did have less galacA
0
t
Con A Eluate
Con A Void
1.2
Is RF binding to Fc influenced by the absence of
carbohydrate? Since a 3-fOld difference in the amount
or complexity of mannose present on the Fc molecule
did not appear to affect the binding of the RF, the
carbohydrate was removed from the Fc by enzymatic
degradation using N-glycanase. As depicted in Figure
4, Fc fragments were visibly smaller after treatment.
Three bands (-55 kd, 52 kd, and 48 kd) were consistently visualized after N-glycanase digestion, regardless of the source of Fc. Of these bands, the 55-kd and
52-kd bands stained with the carbohydrate stain
(PAS), but the 48-kd band stained poorly, if at all
(Figure 4). Thus, we were unable to remove all of the
carbohydrate by N-glycanase digestion. Lectin blot
analysis, however, showed that only a small percentage of the initial carbohydrate remained. RF binding
appeared to be similar regardless of whether the Fc
fragments were digested with N-glycanase, as seen in
Figure 5.
Is RF binding influenced by the source of p l y clonal Fc? Since the presence of differing amounts of
carbohydrate had no apparent effect (either enhancement or inhibition) on the ability of RF to bind to
monoclonal Fc preparations, we were interested in
determining whether polyclonal Fc fragments from
patients with classic RA would influence the binding in
a measurable way when compared with Fc from
normal controls. We characterized the polyclonal Fc
fragments both from the patients with RA and from the
normal controls to determine the differences. if any, in
0.8
"O1
0
B
1.4
n
1.2
1 .o
0.8
0.6
0.4
0.2
0.0
BOR
K U
3019
COY
DAI
RF
Figure 3. Rheumatoid factor (RF) binding to Fc fragments from
either IgGl (A) or lgG2 (B).which were separated by the binding of
mannose on the Fc fragments to a concanavalin A (Con A)Sepharose column. RF are from mixed cryoglobulin sources (as
represented by BOR and KEZ). a hybridoma-generated RF (3019).
and 2 polyclonal RF preparations (COY and DAL). O.D. = optical
density.
NEWKIRK ET AL
804
Table 1. Relative carbohydrate composition of polyclonal transblotted Fc preparations from rheumatoid arthritis patients and
controls measured by lectin binding. compared with total anti-y
an ti body binding'
70galactose
(concanavalin A )
HRP-peanut
agglutinin
Biotinylated
peanut agglutinin,
HRP-streptavidin
17.7
17.7
11.1
12.0
9.9
10.0
18.5
23.3
7.4
5.5
6.8
8.3
Q/r mannose
Preparation
Controls
A
B
Patients
I
2
Figure 4. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of undigested (lanes 2 and 4) and N-glycanase-digested (lanes 3
and 5 ) Fc fragments (IgGI, EV2.7; not reduced; S pg/lane). Lanes
1-3 are stained for protein. and lanes 4 and 5 are stained for
carbohydrate. Lane I shows molecular weight standards.
tose and appeared to have slightly more mannose
compared with normal controls (Table 1 and Figure 6).
Depending on the source of peanut agglutinin and the
type of assay, we obtained slightly different results. In
direct binding immunoblots (in which it was difficult to
obtain sufficient color for scanning) using HRPconjugated peanut agglutinin, the Fc isolated from
patients with RA had -60% of the levels of galactose
found on normal Fc. However, when biotinylated
peanut agglutinin was used, the Fc isolated from RA
*'O
1.6
* Values are the average percent of total anti-y antibody binding in
4-7 experiments. HKP = horseradish peroxidase.
patients appeared to have -77% of the levels found on
the Fc of normal controls. The reason for this discrepancy is not clear; it may, however, reflect the different
sources of peanut agglutinin, because this difference
was consistent when identical preparations were
probed at the same time.
Slight variations in the isotype composition of
the pools of polyclonal Fc were found in an ELISA
using specific anti-subclass antibodies. As predicted,
- N-Glycanase
+NGlycanase
1
1.6
1.4
1.2
d
1.0
0.0
0.6
0.4
0.2
0.0
BOA
KU
SIHZ
3013
103.4
COY
KLI
DAL
RF
Figure 5. Rheumatoid factor (RF) binding profiles to undigested
and N-glycanase-digested Fc (IgGI. FLO). RF are from mixed
cryoglobulin sources (BOR. KEZ. and SWZ). hybridoma generated
(3019 and 103.4), or from polyclonal preparations from patients with
rheumatoid arthritis (COY, KLI. and DAL). O.D. = optical density.
Figure 6. Western blotting analysis of sodium dodecyl sulfatepolyacrylamide gel electrophoresis-separated polyclonal Fc preparations (10 pg per lane) from a normal control (lane I ) or a patient
with rheumatoid arthritis (lane 2). A. lmmunoblotted with horseradish peroxidase (HRPkonjugated anti-y(total) ( I : 1.000 dilution). B,
Lectin blotted with HRP-conjugated concanavalin A ( I :So0 dilution), C, Lectin blotted with HRP-conjugated peanut agglutinin
(1 :200 dilution). Higher molecular weight bands are undigested IgG.
The 55-kd band (arrow) is the Fc fragment.
805
IgG ISOTYPE A N D R F BINDING
1.o
0.8
d
0
0.6
0.4
0.2
nn
V."
1
2
BOR
3019
3
1414
4
8808
RF
Figure 7. Rheumatoid factor (RF) binding to Fc preparations from
monoclonal 1gGl (McC). a polyclonal normal control (Poly Norm).
and a patient with rheumatoid arthritis (Poly RA). RF were from
mixed cryoglobulins (BOR). or from hybridoma-generated antibodies, which were polyreactive (3019 and 1414) or monospecific (8808).
O.D. = optical density.
no IgG3 was detected, since protein A was used in the
purification procedure. The 2 normal pools of Fc
represented different peaks from ion exchange chromatography of Fc from the same individual, with pool
A eluting at the same ionic strength as both of the RA
Fc pools (-O.05M NaCI), and normal pool B eluting at
a higher ionic strength (-0.15M NaCI). The isotypic
composition of pool A was 47% IgGl , 40% IgG2, and
23% IgG4, relative to total y binding. Pool B contained
a similar percentage of IgG1, but more IgG4 and less
IgG2 (53% and 39%. respectively). This pool, in fact,
had the highest percentage of IgG4 compared with all
of the pools. thus accounting for the higher ionic
strength required to elute the pool from the Fast Q
Sepharose column. The Fc preparations from the
patients with RA did not separate into 2 peaks. Both of
the RA Fc pools contained approximately half the
amount of IgG2 (22% and 10%) and twice the amount
of IgG4 (40% and 37%) when compared with the pool
A Fc from normal controls.
When RF binding assays were conducted with
the different pools of polyclonal Fc fragments, derived
either from RA patients or from normal controls as
above, and with monoclonal Fc preparations, it appeared that -50% of the 18 monoclonal RF bound
better to preparations of Fc from polyclonal sources
than from monoclonal IgG 1. Figure 7 presents some of
the different profiles seen for the monoclonal RF. All
of the RF from the mixed cryoglobulinemia source, as
well as the polyclonal RF, bound equally well to the
monoclonal and polyclonal Fc. The finding that some
monoclonal RF bound better to the polyclonal Fc than
to the monoclonal Fc suggested that different RF
might have binding specificities that were dependent
on the isotype of the IgG. This difference in binding
did not discriminate between the 2 sources of polyclonal Fc (RA or controls). The majority of the RF that
bound poorly to the monoclonal IgG1 Fc fragments
bound better to the normal Fc pool B than pool A,
even though the percentages of peanut agglutinin
binding (galactose) and Con A binding (mannose) were
the same in both pools. Pool B, however, was found to
have an increased percentage of IgG4, which likely
accounted for the difference in binding.
Does the isotype of the Fc influence the binding of
the RF? As depicted in Figure 8, RF derived from
mixed cryoglobulinemia plasma (BOR) showed preferential binding for IgG 1. The hybridoma monoclonal
antibody 3019. derived from a patient with SLE,
bound well to both IgGl and IgG2. However, the
difference in the binding ability of some hybridomdRF
(e.g., 1414 and 8808; Figure 8) was in fact reflected in
a preference for isotypes other than IgG1. This pattern
of binding (greater binding to IgG3 andor IgG4 than to
IgGl and/or IgG2) was found for 2 of 5 polyreactive
1.2 r
0.8
cj
0.6
0
0.4
0.2
0.0
BOR
3019
1414
BBo8
RF
Figure 8. Rheumatoid factor (RF) binding to the 4 subclasses of
IgG. Each bar represents the average binding to 2 representative
monoclonal IgGs of each subclass. R F are as outlined in Figure 7.
O.D. = optical density.
NEWKIRK ET AL
806
RF tested (1700 and 1414) and for 4 of 5 monospecific
RF (18141, 8808, 5603. and 5602). Notably, antibodies
18141 and 5602 bound to IgG4 twice as well as to IgGl
(data not shown). The majority of the hybridoma RF
that bound better to IgG3 and/or IgG4 were derived
from patients with RA.
DISCUSSION
We have confirmed the findings of a number of
investigators showing that patients with RA have
reduced amounts of galactose present on their IgG
(5-7). We determined more precisely that this defect is
on the isolated Fc portion of the molecule, as was
predicted. Although the lectin blot method of determining the relative quantities of galactose and mannose
is different from carbohydrate quantitation methods
used in previous studies, the results are comparable.
The earlier studies demonstrated that patients with R A
had -80% of the levels of galactose on the IgG
molecule when compared with age-matched controls.
We found that the percent of normal galactose on the
Fc of IgG from 2 patients with RA ranged from 60% to
77%, depending on the method of detection used.
The focus of these studies was to examine
whether the carbohydrate molecules located on the Fc
fragment can affect the binding of RF from a variety of
sources. There are several possible ways that one
could envision the carbohydrate as having an impact.
First, a smaller carbohydrate chain might cause a
conformational change in the Fc portion and thus
reveal new sites for RF binding. Alternatively, additional amino acids that contribute to an existing binding site might be revealed, resulting in a higher binding
affinity. Second, the presence of a large, branched
carbohydrate might inhibit the binding of RF. Third,
the presence of the carbohydrate, regardless of its size
or composition, might have no effect on the RF
binding, but might play a role in the clearance of IgG.
Although it has been suggested that the reduced
amount of galactose on the Fc fragment of IgG of
patients with RA may render this molecule more
immunogenic (5.6). a recent study (7) demonstrated
that polyclonal RF did not show a binding preference
for IgG derived from patients with RA when compared
with IgG from normal controls. Similarly, polyclonal
RF were found not to discriminate between polyclonal
Fc fractions with and without galactose (21). We have
extended these observations by characterizing the
binding of both monoclonal and polyclonal RF to
monoclonal and polyclonal Fc fragments, where the
composition of the carbohydrate in the Fc varied.
The carbohydrate chain that is located at position 297 of the Fc fragment of IgG has been shown to
be a complex structure, which can vary from one
myeloma IgG protein to another (22). Furthermore, it
is conceivable that the number and organization of
mannose sugars could vary greatly among identical
polypeptide chains. Indeed, we were able to distinguish as much as a 3-fold difference in mannose on
polypeptide-identical monoclonal Fc fragments for 3
different IgG molecules, after these molecules were
separated according to their ability to bind to Con A.
The interaction between Con A and mannose is complex and involves chain ends as well as the molecular
structure of the mannose itself (17,23,24). The Con A
binding studies reported here do not distinguish between varying concentrations of mannose (although
Con A does bind better to higher amounts of mannose), or the increased accessibility of Con A due to
the complexity of the branched carbohydrate. When
we examined the ability of monoclonal and polyclonal
RF to bind to monoclonal Fc fragments that varied in
mannose content or complexity but were similar in the
amount of galactose present (peanut agglutinin accessible; data not shown), we found no appreciable difference in the binding profiles. This indicates that
branched mannose molecules had neither positive nor
negative effects on RF binding when the polypeptide
backbone was identical.
To investigate whether the absence of the carbohydrate moiety from the Fc fragment (with identical
polypeptide chains) would cause an alteration,in the Fc
binding, we removed the carbohydrate by N-glycanase
digestion. When the RF (monoclonal and polyclonal)
were analyzed for their ability to bind to Fc with,intactor
greatly reduced carbohydrate, we could find no major
differences. It is unlikely that the small percentage of Fc
fragments with carbohydrate remaining after Nglycanase digestion could account for all of the RF
binding. The absence of reduced or increased binding to
the N-glycanasedigested Fc fragments suggests that
any conformational change that might have occurred
with the removal of the sugar neither destroyed nor
appreciably altered the binding site.
To further analyze the effects of galactose on
RF binding. we investigated whether polyclonal Fc
with less galactose. from patients with RA, or with
normal amounts of galactose, as present in normal
controls, would be recognized differently by the RF.
Theoretically, IgG in human serum is composed of
IgG ISOTYPE AND R F BINDING
65% IgG1, 23% IgG2, 8% IgG3, and 4% IgG4 (25).
Since these concentrations can vary between individuals and the sensitivity of the different isotypes to
papain digestion varies (26). we characterized the
isotypic components of each pool of Fc. The 2 polyclonal pools of Fc that originated from the normal
control had essentially the same carbohydrate characteristics, as measured by lectin binding, but varied in
isotypic content. The polyclonal Fc pools from the 2
RA patients differed from the normal pools, showing
different lectin binding profiles (carbohydrate composition or complexity), with less galactose, more mannose, and a slight difference in the composition of IgG
isot ypes.
We analyzed the RF for their ability to bind to
the different polyclonal pools of Fc fragments. Interestingly, we found that there were 2 different patterns
of monoclonal RF binding. One group of RF (all of the
mixed cryoglobulins and some of the hybridoma antibodies) bound equally well to all 4 of the polyclonal
and 3 of the monoclonal Fc preparations tested,
whereas the other group of RF (all hybridoma generated) bound poorly to the Fc preparations from monoclonal IgG1, but better to normal and RA-derived
polyclonal Fc preparations. The RF bound better to
the pool B Fc than to the pool A Fc from the normal
controls, despite the presence of similar amounts of
lectin-accessible galactose. The difference in binding
was likely due to the isotypic composition of the Fc
and not the carbohydrate concentration or structure,
since pool B had a higher percentage of IgG4 relative
to pool A.
These predictions were confirmed when the
isotypic binding profiles of the RF were analyzed.
Since the studies of Capra et al(27). it has been known
that RF show individual binding profiles for the 4
subclasses of IgG. However, most of the information
about the binding specificities of monoclonal human
RF has been based on mixed cryoglobulin-derived
RF. The majority of these molecules appear to bind
preferentially to IgGl andot IgG2, although there is a
small subset that binds equaIly well to IgG1, IgG2, and
IgG3 (4,13). We have now extended this observation
to monoclonal RF derived from patients with RA or
SLE. This is the first description of a group of human
monoclonal RF with a binding preference for IgG3
and/or IgG4 which is greater than that for IgG1 and/or
IgG2. One of the antibodies that showed a binding
preference for IgG3 and bound very poorly to IgGl
was hybridoma antibody 8808, a monospecific RF that
expresses many of the idiotypes characteristic of the
807
major Wa idiotypic family of human RF isolated from
patients with mixed cryoglobulinemia.
It is clear that we do not know which, if any, of
the RF are most associated with an active disease
process. Studies of the isotype specificity of polyclonal RF have found that, in contrast to serum RF,
which bind primarily to the IgGl isotype, polyclonal
RF produced by synovial cells bind preferentially to
the IgG3 subclass (28). Our study suggests that representatives of this synovial B cell population can also
be found in peripheral blood, since our hybridoma
fusions were done with peripheral blood lymphocytes
of patients with RA. A study by Hoffman et al (29)
suggests that synovial cells from RA patients with the
most active synovitis produced the most IgG3. Interestingly, the type 11 collagen-specific antibodies produced in the synovium are predominantly IgG3 (30).
Studies of the immunogenicity and clearance of
deglycosylated Fc indicate that the terminal sugar
(sialic acid) is most important (31.32). It is likely that
the imrnunogenicity is heightened by the tendency of
the deglycosylated Fc molecule to aggregate. However, desialated molecules are cleared more rapidly
from the circulation by specific galactose-receptor
sites on hepatocytes (33). In patients with RA, an
absence of galactose on the IgG molecule might be
expected to lead to inefficient clearance and a prolonged half-life for IgG. This, however, does not
appear to be the case. The catabolism of radiolabeled
autologous IgG is faster in patients with RA than in
normal controls (34). Thus, from the information available, the lack of the galactose on the Fc fragment does
not appear to prolong the half-life of the IgG molecule
in patients with RA.
We have clearly shown that the carbohydrate
located on the Fc fragment of IgG has no apparent
effect on RF binding and does not contribute to a
higher-aflinity binding of RF in patients with RA. The
amino acid sequence of the Fc, however, does appear
to be important in determining binding. Since each
monoclonal RF displays a fine specificity of Fc isotypic binding, this suggests that there may be multiple
binding sites on the Fc to which RF bind. Further
studies are required to determine whether IgG3- and/
or IgG4-specific RF are associated with a more active
disease process in RA than are RF that bind primarily
to IgGl andjor IgG2.
ACKNOWLEDGMENTS
We thank Cynthia Lussier and Joanne Wild for their
excellent technical assistance in the generation of the hybrid-
NEWKIRK ET AL
808
oma RF and Dr. John Esdaile for helpful comments regarding the manuscript. We regret the recent death of Dr. R.
Wistar, Jr., who supplied many of the myeloma antibodies
used in this study.
REFERENCES
1. Fye KH, Moutsopoulos HM, Talal N: Immunogenetics,
Basic and Clinical Immunology. Second edition. Edited
by HH Fudenberg. DP Stites, JV Wells. Los Altos, CA,
Lange Medical Publications, 1978
2. Jerreris R. Nik Jaafar MI, Steinitz M: Immunogenic and
antigenic epitopes of immunoglobulins. VIII. A human
monoclonal rheumatoid factor having specificity for a
discontinuous epitope determined by histidine/arginine
interchange at residue 435 of immunoglobulin G. Immuno1 Lett 7:191-194, 1984
3. Oppliger LR. Nardella FA, Stone GC, Mannik M: Human rheumatoid factors bear the internal image of the Fc
binding region of staphylococcal protein A. J Exp Med
166:702-7 10. 1987
4. Sasso EH, Barber CV, Nardella FA, Yount WJ, Mannik
M: Antigenic specificities of human monoclonal and
polyclonal IgM rheumatoid factors. J lmmunol 140:
309g3107, 1988
5. Axford JS. Mackenzie L. Lydyard PM, Hay FC. Isenberg DA, Roitt IM: Reduced B-cell galactosyltransferase activity in rheumatoid arthritis. Lancet 11: 14861488, 1987
6. Parekh RB. Dwek RA, Sutton BJ, Fernandes DL,
Leung A, Stanworth D, Rademacher TW. Mizuochi T.
Taniguchi T. Matsuta K, Takeuchi F, Nagano Y.
Miyamoto T , Kobata A: Association of rheumatoid
arthritis and primary osteoarthritis with changes in the
glycosylation pattern of total serum IgG. Nature 316:
452-457, 1985
7. Tomana M, Schrohenloher RE, Koopman WJ. Alarcon
GS, Paul WA: Abnormal glycosylation of serum IgG
from patients with chronic inflammatory diseases. Arthritis Rheum 31:333-338. 1988
8. Parekh R. Roitt I, Isenberg D. Dwek R, Rademacher T:
Age-related galactosylation of the N-linked oligosaccharides of human serum IgG. J Exp Med 167:1731-1736,
1988
9. Pekelharing JM, Hepp E, Kamerling JP, Gerwig GJ,
Leijnse B: Alterations in carbohydrate composition of
serum IgG from patients with rheumatoid arthritis and
from pregnant women. Ann Rheum Dis 47:91-95. 1988
10. Persellin RH, Cerere F: The interaction of pregnancy
and the rheumatoid diseases. Clin Rheum Dis 7:747-768,
1981
11. Ropes MW, Bennett GA, Cobb S, Jacox R. Jessar RA:
1958 revision of diagnostic criteria for rheumatoid arthritis. Bull Rheum Dis 9:175-176, 1958
12. Tan EM, Cohen AS, Fries JF. Masi AT, McShane DJ,
Rothfield NF. Schaller JG, Talal N, Winchester RJ: The
1982 revised criteria for the classification of systemic
lupus erythematosus. Arthritis Rheum 25: 1271-1277,
1982
13. Newkirk MM, Mageed RA, Jefferis R, Chen PP, Capra
JD: Complete amino acid sequences of variable regions
of two human IgM rheumatoid factors, BOR and KAS of
the Wa idiotypic family, reveal restricted use of heavy
and light chain variable and joining region gene segments. J Exp Med 166:550-564, 1987
14. Massicotte H, Rauch J. Shoenfeld Y, Tannenbaum H:
Influence of fusion cell ratio and cell plating density on
production of human-human hybridomas secreting antiDNA autoantibodies from patients with systemic lupus
erythematosus. Hybridoma 3:2 15-222, 1984
IS. Poljak RJ: Crystallization of immunoglobulins and their
fragments for X-ray diffraction studies. Methods Enzymol 116:190-200, 1985
16. Laemmli UK: Cleavage of structural proteins during the
assembly of the head bacteriophage T4. Nature 227:
680-685, 1970
17. Ohyama Y. Kasai K, Nomoto H, Inoue Y: Frontal
affinity chromatography of ovalbumin glycosasparagines
on a concanavalin A-sepharose column. J Biol Chem
260:6882-6887, 1985
18. Towbin H, Staehelin T. Gordon J: Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc
Natl Acad Sci USA 76:43504354, 1979
19. Voller A, Bidwell DE. Bartlett A: The Enzyme Linked
Immunosorbent Assay (ELISA). Alexandria. VA,
Dynatech, 1979
20. Heimer R, Klein G: The affinity of soluble immune
complexes for concanavalin A. Scand J Immunol 7:
315-320. 1978
21. Tsuchiya N. Endo T, Matsuta K, Yoshinoya S. Aiikawa
T , Kosuge E. Takeuchi F. Miyamoto T, Kobata A:
Effects of galactose depletion from oligosaccharide
chains on immunological activities of human IgG. J
Rheumatol 16:285-290, 1989
22. Mizuochi T, Taniguchi T, Shimizu A, Kobata A: Structural and numerical variations of the carbohydrate moiety of immunoglobulin G. J lmmunol 129:20162020,
1982
23. Goldstein 1J. Hollerman CE, Smith EE: Proteincarbohydrate interaction. 11. Inhibition studies on the
interaction of concanavalin A with polysaccharides.
Biochemistry 4:876-883. 1%5
24. Brewer CF, Bhattacharyya L: Specificity of concanavalin A binding to asparagine-linked glycopeptides. J
Biol Chem 261:730&7310, 1986
25. Roitt IM: Essential Immunology. Sixth edition. Oxford,
Blackwell Scientific Publications, 1988
26. Gergely J, Fudenberg HH, van Loghem E: The papain
IgG ISOTYPE AND R F BINDING
27.
28.
29.
30.
susceptibility of IgG myeloma proteins of different
heavy chain subclasses. lmmunochemistry 7: 1-6. 1970
Capra JD, Kehoe JM, Winchester RJ. Kunkel HG:
Structure-function relationships among anti-gamma
globulin antibodies. Ann N Y Acad Sci 190:371-381,
1971
Robbins DL. Benisek WF, Benjamini E, Wistar R Jr:
Differential reactivity of rheumatoid synovial cells and
serum rheumatoid factors to human immunoglobulin G
subclasses 1 and 3 and their CH3 domains in rheumatoid
arthritis. Arthritis Rheum 30:489-497, 1987
Hoffman WL, Goldberg MS. Smiley JD: Immunoglobulin G3 subclass production by rheumatoid synovial tissue cultures. J Clin Invest 69:13&144, 1982
Watson WC. Cremer MA, Wooley PH, Townes AS:
Assessment of the potential pathogenicity of type I1
collagen autoantibodies in patients with rheumatoid ar-
809
31.
32.
33.
34.
thritis: evidence of restricted IgG3 subclass expression
and activation of complement C5 to CSa. Arthritis
Rheum 29:131&1321. 1986
Duc Dodon M. Quash GA: The antigenicity of asialylated IgG: its relationship to rheumatoid factor. Immunology 42:401408. 1981
Murray JS. Brown JC: Hyperimmunization alters Fc
antigenicity: linkage to glycosylation. J Immunol I41 :
2668-2673, 1988
Kawasaki T, Ashwell G: Carbohydrate structure of
glycopeptides isolated from an hepatic membrancbinding protein specific for asialoglycoproteins. J Biol
Chem 2515292-5295, 1976
Catalan0 MA, Knck EH, deHeer DH, Nakamura RM,
Theofilopoulos AN, Vaughan JH: Metabolism of autologous and homologous IgG in rheumatoid arthritis. J
Clin Invest 60:313-322, 1977
Документ
Категория
Без категории
Просмотров
2
Размер файла
849 Кб
Теги
isotypes, factors, determinism, igg, glycosylation, state, importance, human, binding, rheumatoid
1/--страниц
Пожаловаться на содержимое документа