вход по аккаунту


Quantitation of retroviral gp70 antigen autoantibodies and immune complexes in extravascular space in arthritic mrl-lprlpr mice. use of a subcutaneously implanted tissue cage model

код для вставкиСкачать
Use of a Subcutaneously Implanted Tissue Cage Model
MRL-lpdlpr (MRL/I) mice spontaneously develop a disease that is characterized by glomerulonephritis, diffuse vasculitis, and arthritis associated with high
levels of autoantibodies that include IgG rheumatoid
factor (RF). To define the immunopathogenic mechanisms that lead to the development of extravascular
lesions such as arthritis, we implanted a tissue cage
subcutaneously in arthritic MRLA mice and compared
components of the tissue cage fluid, which resembles the
extravascular fluid, with those of sera. When compared
with those of sera, tissue cage fluids from arthritic
MRIA mice had similar levels of RF and one-third the
amount of Clq immune complexes. In contrast, antiDNA activities in tissue cage fluids corresponded to only
10% of the serum activities and, most strikingly,
nephritogenic retroviral gp70-anti-gp70 immune complexes were almost undetectable in tissue cage fluids.
This was also the case for another strain of autoimmune
mice, (New Zealand black X New Zealand white)F1
From the World Health Organization Immunology Research and Training Centre, Department of Pathology, University of
Geneva, and Centre de Transfusion, HBpital Cantonal Universitaire, Geneva, Switzerland.
Supported by grants 3.621.0.84 and 3.803.0.86 from the
Swiss National Foundation €or Scientific Research, by the World
Health Organization, by La Fondation Centre de Recherches
Mtdicales Carlos et Elsie de Reuter, and by the Swiss Confederation on the proposal of the Commission FtdCrale des Maladies
Eric Dayer, MD: Research Fellow; Haruyoshi Yoshida,
MD: Research Fellow; Shozo Izui, MD: ChargC de Recherche;
Paul-Henri Lambed, MD: Professor and Director of World Health
Organization Immunology Research and Training Centre.
Address reprint requests to Shozo Izui, MD, Department of
Pathology, University of Geneva, C.M.U., CH-1211 Geneva 4,
Submitted for publication October 3, 1986; accepted in
revised form April 1, 1987.
Arthritis and Rheumatism, Vol. 30, No. I1 (November 1987)
mice, although they did not produce RF. The absence of
gp70 immune complexes in tissue cage fluids could be
due to markedly limited diffusion of gp70 antigen in
these fluids. These results strongly suggest that serum
proteins, including autoantigen and autoantibodies, appear in extravascular fluid in a selective manner, depending on their size and charge. Their specific properties in sera or extravascular fluid could partly account
for the different manifestations of vascular and
extravascular lesions observed in autoimmune mice.
Several strains of inbred mice have provided
experimental models of systemic lupus erythematosus
and related rheumatic diseases in humans. These
mice, including (New Zealand black x New Zealand
white)F1([NZB x NZW]FI), BXSB/MpJ (BXSB), and
MRL-lpdlpr (MRL/I) express B cell hyperactivity associated with the production of various autoantibodies
and with the subsequent development of immune
complex (IC) lupus nephritis (1). Retroviral envelope
glycoprotein, gp70, has been implicated in the pathogenesis of murine lupus nephritis. Gp70 antigen is
found in IC deposits within the diseased glomeruli (2).
Gp70-anti-gp70 IC appear in the circulation shortly
after the onset of the disease and their concentrations
rise with the progression of the nephritis (3), providing
good evidence that gp70-anti-gp70 1C are a potential
source of renal injury. The participation of DNA-antiDNA IC was also demonstrated when significant concentrations of anti-DNA antibodies were found in
renal eluates of (NZB x NZW)F1 mice (2).
In addition, MRL/I mice spontaneously develop
arthritis (4), high levels of rheumatoid factors (RF),
and intermediate-sized IC consisting of IgG-RF (5,6)
that are similar to those found in patients with rheu-
matoid arthritis (RA). Such RF IC may play a significant role in the development of arthritis-like articular
changes that occur uniquely in adult MRLfl mice (4,6).
The analysis of various serologic markers in
synovial fluids from patients with RA has provided
valuable information about the immunopathogenesis
of RA (7). Although such studies are difficult to
perform in mice, it is possible to analyze the appearance: and kinetics of serologic parameters in the
extravascular fluid that accumulates in a space created
by tlhe subcutaneous implantation of a tissue cage on
the backs of the mice (8). In fact, this model has been
usedl to determine the distribution of antibiotics in
extrravascular fluids (9). Since the fluid accumulating in
the tissue cage resembles extravascular fluid (10) and
synovial fluid (1 l), the analysis of extravascular fluid
in the tissue cage should determine which parameters
may be most significantly expressed and involved in
the development of extravascular and, possibly, articular lesions. Using the tissue cage model, we have
compared the presence of various serologic markers
between extravascular fluids and sera from 2 different
autalimmune strains, arthritic MRL/I and nonarthritic
(NZB x NZW)F1 mice, to determine if there is any
selective expression of autoantibodies, IC, and/or
retroviral gp70 antigen in extravascular fluids. Nonautoimmune DBM2 mice were used as controls.
Tissue cage implantation. Female MRLA and DBN2
mice were purchased from Jackson Laboratories (Bar Harbor,
ME). (NZB x NZW)FI mice were obtained from Olac Laboratolies (Oxon, UK). The mice were bled from the retroorbital
plexus and the resulting sera were stored at - 20°C until use.
Tissue cages (length 3 cm, diameter 8 mm) (12) made of Teflon
tubing were prepared locally. Holes 0.2 mm in diameter were
perforated on the outer part (approximately 45% of the surface
was open) (Figure 1). At each end, a 2-mm hole was perforated,
through which liquid could be drawn. At 2 months of age,
15-20 mice were given general anesthesia (fentanyl [Hypnomi]; Duphar, Holland) (10 pg/20 gm body weight), and tissue
cages were implanted on their backs. The skin was incised on
one side. After disruption of subcutaneous tissues, the tissue
cage was slid to the other side of the back and the incision was
sewn. Four weeks later, when conditions were stabilized (i.e.,
the components in the tissue cage fluids became constant), the
infected tissue cages invariably were rejected. Infections occurred throughout the first 4 weeks in one-tenth of implanted
tissue cages. Tissue cage fluid was cultured for bacteria, with
nega.tive results. At each puncture, under sterile conditions,
50-100 pl of interstitial fluid could be drawn with minimal blood
Serologic assays. Total protein levels were measured
with a standard assay (Bio-Rad, Richmond, CA). Albumin
Figure 1. The tissue cage.
concentrations were measured by radial immunodiffusion
with a goat anti-mouse albumin antiserum (Cappel Laboratories, Cochranville, PA). Standard curves were established
using purified mouse serum albumin (Sigma, St. Louis, MO).
Serum and tissue cage fluid levels of total IgG, IgG
subclasses, and IgM were determined by a solid-phase
radioimmunoassay (RIA) using goat anti-mouse, y-chain and
p-chain-specific antibodies as described previously (13,141.
Standard curves were established for each assay with mouse
myeloma proteins of each immunoglobulin class and subclass (Litton Bionetics, Kensington, MD).
Antibodies to single-stranded DNA (ssDNA) were
measured with a modified Farr DNA binding RIA, as described previously (15). The results were expressed as a
percentage of 20 ng of '2SI-labeled DNA precipitated specifically after correction of nonspecific binding in a serum pool
from 2-month-old BALB/c mice. IgM antibodies to ssDNA
were inactivated by incubation of sera or tissue cage fluids
with 2-mercaptoethanol (2-ME; final concentration 0.1M)
(5). The treatment of tissue cage fluids with DNase
(deoxyribonuclease I; Worthington Biochemical, Freehold,
NJ) was performed as described previously ( 5 ) .
Concentrations of gp70 in sera or tissue cage fluids
were determined by their capacity to inhibit the binding of goat
antibody to feline leukemia virus to '251-labeled gp70 from
Rauscher murine leukemia virus (3,16). Anti-gp70 IC were
quantitated by the RIA combined with precipitation of sera or
tissue cage fluids with polyethylene glycol (PEG) (17).
The presence of anti-IgG activity was assessed by RIA
(14). Briefly, sera or tissue cage fluids were first treated with
acetate buffer, pH 3.5 (final concentration 0.25M), for 1 hour at
37°C and for 2 hours at room temperature, to dissociate
IgG-anti-IgG IC present in the samples. The '2sI-labeled mouse
IgG (Miles Laboratories, Elkhart, IN) was added just before
the adjustment of pH with Tris, and the mixture was incubated
overnight at 4°C. Labeled mouse IgG bound to IgM and IgG
antibodies was precipitated with 7% PEG. Results were expressed as a percentage of 100 ng of labeled mouse IgG
precipitated specifically by 1.5 p1 of serum.
Circulating IC were detected and quantified using a
modification of the C l q binding test (18). Sera or tissue cage
fluids were first treated with EDTA (final concentration
0.15M)for 30 minutes at 37°C and were then incubated with
Table 1. General characteristics of tissue cage fluid and blood in 3-
month-old MRL-lprllpr mice*
Red blood cells
White blood cells
Total protein
Tissue cage fluid
1.2 f 0.7
14.5 ? 3.1
1.7 k 1.5
64.0 2 7.0
43.0 f 10.0
24.0 2 5.0
89.0 f 18.0
* Values are the mean 2 1 SD of 5-7 mice.
t Tissue cage fluid:blood.
12SI-labeledC l q in the presence of PEG (final concentration
3.3%) for 1 hour at 4°C. The results are expressed as pg
equivalent of heat-aggregated mouse IgG per ml of serum or
tissue cage fluid.
Concentrations of haptoglobin in murine serum and
tissu'e cage fluid were measured by radial immunodiffusion in
agar with goat anti-human haptoglobin antisera (Cappel
Laboratories) (16). Results are expressed as a percentage of
the concentration in a pool of C57BL/6J male mice.
Sucrose density gradient ultracentrifugation.Twentyfivemicroliter samples were layered on 1040% (weight/
volume) linear sucrose gradient in veronal buffered saline,
pH 7.2, or 0.5M acetate buffer, pH 4 and were centrifuged at
60,OOO revolutions per minute for 5 hours at 4°C in an SW60
rotor in a Beckman 175 ultracentrifuge. The positions of IgG
and IgM were established by using radioactive markers.
Gradients were divided into 12-14 fractions. Acid gradient
fractions were neutralized by the addition of 4M of Tris.
Statistical analysis. Statistical analysis was performed
with the Wilcoxon 2-sample test. P values less than 0.05
were considered significant.
The mean protein concentration in tissue cage
fluid was 43 mg/ml, which corresponded to 67% of serum
levels. The major protein detected was albumin (33
mg/ml), which was 87% of serum levels. The quantity of
immunoglobulins measured in tissue cage fluids varied,
depending on their classes and subclasses. Mean concentrations of IgG and IgM in tissue cage fluids were
64% and 27% of their serum concentrations. When levels
of IgG subclasses were analyzed, IgG2a and IgG3 represented most of the IgG present in tissue cage fluids.
They amounted to 41% and 31%, respectively, of the
total IgG level in tissue cage fluids (45% and 32%,
respectively, in the corresponding sera). Mean concentrations of IgGl and IgG2b were 13% and 15% of the
total IgG concentration in tissue cage fluids (11% and
12%, respectively, in the corresponding sera).
3 mo
MRL/l 5 mo
General characteristics of tissue cage fluids and
sera of young MRLA mice. Blood and tissue cage fluids
were collected from each mouse once a month. In
3-month-old mice, the mean ? SD red blood cell
(RBC) count in the extravascular space created in the
tissue cage was 120,000 ? 70,000 cells/mm3(Table I),
which was 1.6% of the RBC count in the blood;
probably such cells passed into the extravascular fluid
following the minor injury of small vessels at time of
puncture. The mean 2 SD white blood cell count in
the tissue cage fluid was 1,700 ? 1,500 cells/mm3.
Their differential count showed a majority to be
polymorphonuclear cells (98 & I%), unlike the blood
differential count, which was composed mainly of
small lymphocytes. Similar results were obtained in
non-autoimmune DBN2 mice (data not shown).
Figure 2. DNA binding activities in diluted (l:lO, 1:30, 1:90, and
1:270) tissue cage fluids (0---0)
and sera (0-0)
of female
MRL-lprllpr (MRLII) and (New Zealand black x New Zealand
white)FI (NZB x W) mice, measured by '251-single-stranded DNA
(ss DNA) binding radioimmunoassay. Results are expressed as a
percentage of 20 ng of labeled DNA precipitated specifically.
- IC
5 15- 0
0 0
0 -
- IgG
- 8s DNA
0 0
Figure 3. Clq binding immune complexes (IC), anti-single-stranded DNA (ss DNA), and anti-IgG rheumatoid factor (RF) levels in sera
(0)and in tissue cage fluids (0)
from female 4-month-old MRL-lpr/lpr (MRUI), 8-month-old (New Zealand black x New Zealand white)F1
(NZB x W),and 8-month-oldDBN2 mice. Clq binding activities are expressed in pg equivalentsof heat-aggregated mouse IgG (AMG) per
ml of serum or tissue cage fluid. Anti-ss DNA activities are expressed as a percentage of 20 ng of labeled DNA precipitated specifically.
Anti-IgG-RF activities are expressed as a percentage of 100 ng of labeled mouse IgG precipitated specifically by 1.5 pI of serum. Note that,
in MRUI and NZB x W mice, levels of Clq IC and anti-ss DNA in sera and tissue cage fluids were significantlydifferent (P< 0.001), whereas
there was no significant difference in RF activity between sera and tissue cage fluids of MRUI mice (P> 0.1).
Serologic markers of autoimmune disease in tissue cage fluids and sera in aged MRLfl and (NZB x
NZW)F1 mice. As the MRLll mice aged, the IgG levels
in tissue cage fluids progressively increased. At age 5
months, their mean tissue cage fluid IgG concentration
was 6.6 mg/ml, which represented 46% of their serum
level. In parallel to the increase of serum anti-DNA
levels with age, the tissue cage fluids exhibited an
age-dependent increase in anti-DNA antibody activities (Figure 2). Almost all of the anti-ssDNA from sera
and tissue cage fluids of MRLA mice was resistant to
treatment with 2-ME, indicating its IgG nature. A
quantitative study using serially diluted tissue cage
fluids and sera showed that titers of anti-ssDNA
anlibodies in tissue cage fluids, measured by 1251DNA
binding RIA, were approximately 10 times less than
those of corresponding sera (Figure 2). Essentially
identical results were obtained by solid-phase antiDNA RIA (data not shown). The presence of free
DNA in tissue cage fluids, which may interfere with
the: DNA binding assays, was ruled out by the fact that
the fluids from MRL/I mice, as well as from nonautoimmune DBA/2 mice, did not inhibit anti-DNA
antibody binding to I2'I-DNA. Furthermore, none of
the tissue cage fluids from MRLA mice exhibited a
significant increase in DNA binding activities after
treatment with DNase (data not shown).
In contrast to the limited expression of antiDNA antibodies in tissue cage fluids, their R F activities were almost comparable with those of corresponding sera (P > 0.1) (Figure 3). Levels of IC detected by
the C l q binding test increased with age in MRL/I mice.
At 3 months, significant amounts of C l q binding
materials (mean 2 SD 27 ? 18 pg/ml) were already
detectable in tissue cage fluids; mean levels of immune
complexes reached 56 pg/ml at 4 months, which
corresponded to 34% of serum levels (Figure 3).
Similar analyses of tissue cage fluids and sera
were performed in another autoimmune, but nonarthritic, mouse strain: the (NZB x NZW)FI mouse
(Figures 2 and 3). Although significant levels of anti-
- IC
- 8s DNA
8 " "
' '
' ' 8 ' ' ' '12'
Gradient fractions
Figure 4. Ultracentrifugation analysis of Clq binding immune complexes (IC) and anti-singlestranded DNA (ss DNA) antibodies in sera (a and b) and in tissue cage fluids (c and d) under acidic
(0---0) and neutral ( L O ) conditions. Results are expressed as a percentage binding of labeled
Cl q and ssDNA. Arrows indicate the position of markers.
ssDNA antibodies were detected in tissue cage fluids,
their anti-ssDNA activities were again limited to onetenth of those of corresponding sera, as was found in
MRL/l mice. The mean level of Clq binding IC on the
(NZB x NZW)FI mice was 18 pg/ml of heat-aggregated
mouse IgG equivalent. This amounted to 32% of their
serum levels. However, no significant RF activities were
demonstrated in tissue cage fluids and sera from (NZB x
NZW)F, mice. Control non-autoimmune DBN2 mice
did not have significant levels of autoantibodies or 1C in
tissue cage fluids and sera (Figure 3).
Analysis of IC present in tissue cage fluids. The
sizes of the IC found in tissue cage fluids of 7 individual
MRLil mice were compared with those found in the
corresponding sera, by sucrose density gradient ultracentrifugation analysis. The presence of C lq binding
activities in tissue cage fluids and sera was always
demonstrated in the gradient fractions between 7 s and
19s. Under dissociating conditions, all the activity was
recovered at the 7 s fraction. Intermediate-sized antissDNA complexes, presumably associated with IgG-RF
(3,were present in the majority of tissue cage fluids, as
well as in sera. Again, such complexes were resistant to
DNase digestion and dissociated under acidic conditions
to form a single peak of anti-ssDNA antibodies at the 7 s
fraction. It should be noted that there was no significant
difference between tissue cage fluids and sera in the size
of Clq binding IC and anti-DNA complexes. Representative results of Clq-IC and anti-DNA are shown in
Figure 4.As shown previously (3), when the presence of
gp7O-anti-gp70 IC in gradient fractions was analyzed,
substantial amounts of gp70 IC, sedimenting between
7s and 19S, were found in sera. However, neither
complexed nor free-form gp70 were detected at a significant level in any ultracentrifugation gradient fractions of
tissue cage fluids from MRLh mice (data not shown).
Similar results were obtained in tissue cage fluids of
(NZB x NZW)FI mice.
Table 2. Distribution of total and complexed gp70 and gp70 immune complex (IC) in tissue cage
fluids and in sera of autoimmune murine strains*
3 mos. (7)
4 mos. (14)
5 mos. (8)
4 mos. (8)
8 mos. (8)
4 mos. (7)
Tissue cage fluid
Total gp70
gp70 IC
Total gp70
gp70 IC
36.7 ? 16.7
50.1 f 37.0
64.7 f 37.7
12.3 f 12.0
27.3 .+ 24.0
3.7 2 2.6
5.4 f 4.6
6.0 f 9.1
10.6 f 4.4
17.4 f 14.3
f 3.2
6.4 .+ 2.4
* Total gp70 refers to free and complexed gp70; gp70 IC refers to gp70 complexed with anti-gp70. The
differences between sera and tissue cage fluids were highly significant for both total gp70 levels and
gp70 IC levels (P < 0.001).Values are the mean pg/ml 2 1 SD. NT = not tested.
Retroviral gp70 antigen in sera and tissue cage
fluids. Since we could not detect significant amounts of
gp70 antigen in gradient fractions of tissue cage fluids,
levels of free gp70 and gp70 IC in tissue cage fluids
were compared with those in sera from MRL/I and
(NZB x NZW)F, mice (Table 2). Total levels of gp70
(free and complexed) in tissue cage fluids were extremely limited and averaged approximately 10% of
the levels in serum in mice at each age (P < 0.001).
Mean levels of gp70 IC circulating in sera were 12
pgliml and 27 pg/ml, respectively, in 4- and 5-monthold MRL/I mice and 11 pg/ml and 17 pglml, respectively, in 4- and 8-month-old (NZB x NZW)FI mice.
In contrast, gp70 IC were barely detectable in most
tissue cage fluids from both 5-month-old MRL/l and
8-m~onth-old(NZB x NZW)FI mice. It should be
notied that the concentration of free gp70 in tissue cage
fluids, which was quantitated by subtracting complexed gp70 from total gp70, was in the range of
1&20% of free gp70 serum levels (P < 0.001).
Since gp70 antigen behaves as an acute-phase
reactant (16), we examined whether gp70 antigen can be
transferred from the circulating blood to the tissue cage
fluid of (NZB x NZW)FI mice during the acute-phase
response induced by the injection of bacterial lipopolysaccharides (LPS) (Salmonella Minnesota Re595; 50 pg
intraperitoneally). Although gp70 levels in sera peaked
between 12 and 24 hours (twentyfold increase), there
was essentially no increase of gp70 in tissue cage fluids
during the period studied (60hours) (Figure 5). Levels of
another acute-phase glycoprotein, haptoglobin, incretased in sera approximately eightyfold and peaked
between 24 and 36 hours. Although haptoglobin levels in
tissue cage fluids were approximately one-tenth of serum
values, like gp70 levels under normal conditions,
haptoglobin levels in tissue cage fluids increased more
than fiftyfold during the acute-phase response, which
was in marked contrast with the lack of increase of gp70
levels in tissue cage fluids during this phase. In response
to the injection of LPS, serum levels of IgM increased,
with the highest level observed at day 9; in tissue cage
fluids a slight, but constant, increase was observed up to
day 30. The IgG3 response, which is the most markedly
enhanced by the LPS stimulation (19), peaked at the
same time in tissue cage fluid and sera (between day 9
and day 15).
To determine whether the limited expression of
gp70 in tissue cage fluids was associated with specific
enzymatic digestion, we performed experiments in
which sera containing high levels of gp70 were incubated with freshly drawn tissue cage fluids. No digestion of gp70, as determined by measuring the size of
gp70 by sucrose density gradient analysis, was found
in vitro after 1,2, and 3 hours of incubation at 37°C. In
addition, the clearance rate of gp70 from tissue cage
6.5 hours), which was determined following
the injection of serum gp70 into the tissue cage, was
comparable with the clearance rate of gp70 from sera
(T1/25.5 hours) (20). We tested other biologic fluids for
gp70 levels to exclude a peculiarity of our tissue cage
model. Again, gp70 was almost undetectable in the
ascites of aged, diseased MRL/I mice. The ascites
contained < I % of their serum gp70 concentration
(ascites 0.2 k 0.1 pglml; serum 30.0 ? 16.3 pg/ml),
although albumin represented 10% of the serum levels.
We have examined the appearance of known
serologic parameters in the extravascular fluid accumu-
- ___d----a
J +--d
0 3
rnQ/ml 4 -
lL, --..---.+'
0 3
After injection
Figure 5. Acute-phase response of a, gp70 antigen and c, haptoglobin in sera (0) and tissue cage fluids (0)
after intraperitoneal injection of Salmonella Minnesora Re595 lipopolysaccharides
into 4-month-old female (New Zealand black x New Zealand white)F,
mice. I) and d, Immunoglobulin levels in sera (A) and tissue cage fluids
up to 30 days after injection of lipopolysaccharides. Each point
represents the mean from 8-10 mice; bars represent 1 SD.
lated in tissue cages implanted subcutaneously in
autoimmune strains of mice. Tissue cage fluids from
arthritic MRLfl mice had similar levels of RF and onethird as much C lq IC when compared with the levels in
sera. ]Incontrast, anti-DNA activities in tissue cage fluids
corresponded to only 10% of serum activities, and most
strikingly, nephritogenic gp70-anti-gp70 IC were almost
undetectable in tissue cage fluids. The absence of gp70
IC in tissue cage fluids was due to the markedly limited
diffusion of gp70 antigen in tissue cage fluids. Apparently, there exists a selective diffusion of pathogenic
components from sera to extravascular fluids.
The tissue cage model used in the present study
was originally introduced for assessment of the nega-
tive pressure of interstitial fluids (21). It was demonstrated that fluids within the tissue cage were in direct
contact with the extracellular space. The resemblance
between tissue cage fluids and extravascular fluids was
documented by dynamic studies using radiolabeled
materials such as dextran, albumin, and sodium,
(10-12), and by the direct comparison of their components (9). Our analysis of cellular and protein compositions supported the notion of this resemblance between tissue cage fluid and extravascular fluid.
Our most striking observation was the almost
complete absence of retroviral gp70-anti-gp70 IC in
tissue cage fluids from autoimmune MRL/I and (NZB
x NZW)Fl mice. Notably, these mice have high levels
of gp70 IC in the circulating blood, and such 1C have
been demonstrated to be involved in the pathogenesis
of murine lupus nephritis (3,17). Clearly, the absence
of this IC in tissue cage fluids suggests that gp70 IC
may play little, or a very limited, role in the development of extravascular lesions and, possibly, articular
lesions. Of course, one should consider the possibility
that anti-gp7O antibodies may be locally produced and
may contribute to the development of extravascular
lesions, such as synovitis. In fact, the local production
of RF has been demonstrated in rheumatoid synovial
tissues (22). Such RF, following the formation of IC,
could be a potential source of tissue injury. However,
since the gp70 antigen is synthesized and secreted by
hepatic cells (16), the restricted diffusion of gp70 from
the circulating blood to extravascular spaces, as discussed below, would certainly limit the pathogenic
potential of anti-gp70 antibodies, compared with RF.
One may need to interpret the results of sucrose
density gradient analysis cautiously, since the cryoglobulin formation during centrifugation at 4°C may
cause loss of IC. In fact, MRL/I mice are known to
develop extremely large amounts of cryoglobulins (2).
However, such a loss could not account for the
absence of gp70 IC in tissue cage fluids. First, relatively high concentrations of gp70 IC were detectable
in sera under the same experimental conditions. Second, the direct quantitation of gp70 IC by PEG precipitation also failed to demonstrate the presence of
gp70 IC in tissue cage fluids, Since tissue cage fluids
did not contain more cryoprecipitable materials than
sera (data not shown), it is unlikely that cryoprecipitates were selectively lost from tissue cage fluids during
their collection. Finally, our recent study has demonstrated that levels of gp70 IC did not significantly differ
before or after the removal of cryoglobulins from sera
of h4RLII mice (Abdelmoula M, Izui S, Lambert P-H:
unpublished observations).
Instead, our results strongly suggest that the
absence of gp70-anti-gp70 IC in tissue cage fluids is a
result of markedly limited expression of serum retroviral gp70 antigen in tissue cage fluids. It is known that
the size of protein is an important factor limiting the
access to the extravascular fluid. In fact, we have
shown in the present study that relative concentrations
of tlhe major serum proteins such as albumin, IgG, and
IgMI in tissue cage fluids versus sera correlated well
with their molecular size. However, this was not true
for gp70. The molecular weight of gp70 (70 kd) is
similar to that of albumin, yet its concentration in
tissue cage fluids is only 10% of serum values, in
contrast to 87% for albumin. It is unlikely that gp70
antigen is degraded in tissue cages too rapidly to be
detected. First, tissue cage fluids did not exhibit
significant proteolytic activities for gp70. Second, kinetrics studies have shown that the disappearance rate
of gp70 injected into tissue cages was not significantly
different from its clearance rate in sera (20).
The relatively rapid rate of clearance of gp70 from
the circulating blood (TI,* 5.5 hours) could be a factor
limiting the access to extravascular fluid. However, the
lack of increase in gp70 in tissue cage fluids during the
acute-phase response, despite its twentyfold increase in
sera, is evidence against this possibility. Notably, levels
of another acute-phase protein, haptoglobin, were augmented to the same extent in tissue cage fluids as in sera
during the acute-phase response. Other biologic fluids
that are representative of the extravascular space, such
as ascites, contained <1% of the serum gp70 concentration; this fact excludes the possibility of a peculiarity of
the tissue cage fluid.
Most probably, the access of gp70 antigen to
the extravascular fluid is limited by its particular
physicochemical properties. This may be in part related to the fact that serum gp70 is highly negatively
charged and its isoelectric point is 4.6 (Yoshida H, lzui
S: unpublished observations), although serum albumin, with an isoelectric point that is not far from that
of gp70, has relatively good access to the tissue cage
fluid. It is, in fact, known that the charge of proteins is
an important factor influencing their diffusion through
vessel walls and particularly through the glomerular
basement membrane, which is heavily negatively
charged (23,24). It is possible that a similar mechanism
may explain the limited diffusion of gp70 antigen in
tissue cage fluids.
In contrast with gp70-anti-gp70 IC, substantial
amounts of intermediate-sized C lq binding IC and antiDNA complexes were detected in tissue cage fluids of
arthritic MRL/I mice. These anti-DNA antibodies
sedimenting in the intermediate position have been previously revealed to be complexed with IgG-RF without
the involvement of DNase-sensitive materials (3,thus
representing the IgG-IgG-RF complexes in MRLA mice.
Notably, such RF complexes were present neither in
sera nor in tissue cage fluids from lupus-prone (NZB x
NZW)F1 mice, which fail to develop arthritis-like joint
lesions. This, taken together with the fact that
nephritogenic gp70-anti-gp70 IC did not appear in
extravascular fluids, indicates that IgG-RF complexes
may play a significant role in the development of arthritic
lesions observed in MRLA mice.
It is significant that anti-DNA antibody concentrations in tissue cage fluids were relatively limited
when compared with concentrations of total IgG,
IgG-RF, and Clq IC. Since essentially all the antiDNA activities in sera from (NZB x NZW)FI mice
were found in the 7 s fraction of sucrose density
gradients (3,it is unlikely that the complex formation
of anti-DNA antibodies interferes with their transfer to
tissue cage fluids. Although we could not detect DNAlike antigen in tissue cage fluids, and the DNase treatment of tissue cage fluids did not increase their DNA
binding activities, one could not exclude the possible
release into the tissue cage fluid of DNA antigen that
adsorbed and eliminated anti-DNA activities from tissue
cage fluids. Alternatively, anti-DNA antibodies may
react with cross-reacting tissue antigens present in vascular basement membranes, and this may be partly
responsible for the relatively low concentration of antiDNA antibodies found in tissue cage fluids. Recent
studies of monoclonal anti-DNA antibodies have shown
the reactivity of anti-DNA antibodies with various nonDNA antigens (25-27). In this regard, it should be noted
that certain anti-DNA antibodies exhibit a strong crossreactivity with heparan sulfate present in glomerular
basement membranes (28).
Our findings strongly suggest that serum proteins including antigen, antibodies, and IC are present
in the extravascular fluids in a selective manner,
depending on their size, their charge, and possibly
their configuration. Clearly, selective diffusion of
pathogenic components such as RF IC, gp70-anti-gp70
IC, and anti-DNA antibodies from the circulating
blood to the extravascular fluids could account for the
different manifestations of vascular and extravascular
lesions observed in autoimmune mice.
We thank Professor Peter A. Miescher for his invaluable support during the course of investigation, and
Ghislaine Lange and Maya Bieri for excellent technical
1. Dixon FJ: Murine lupus: a model for human autoimmunity. Arthritis Rheum 28:1081-1088, 1985
2. Andrews BS, Eisenberg RA, Theofilopoulos AN, Izui S,
Wilson CB, McConahey PJ, Murphy ED, Roths JB,
Dixon FJ: Spontaneous murine lupus-like syndrome. J
Exp Med 148:1198-1207, 1978
3. Izui S, McConahey PJ, Theofilopoulos AN, Dixon FJ:
Association of circulating retroviral gp70-anti-gp70 immune complexes with murine systemic lupus erythernatosus. J Exp Med 149:1099-1105, 1979
4. Hang L, Theofilopoulos AN, Dixon FJ: A spontaneous
rheumatoid arthritis-like disease in MRL/I mice. J Exp
Med 155: 1690-1701, 1982
5. Izui S, Eisenberg RA: Circulating anti-DNA-rheumatoid
factor complexes in MRL/I mice. Clin Immunol Irnmunopathol 15536-551, 1980
6. Nardella FA, Teller DC, Izui S, Mannik M: Selfassociating IgG rheumatoid factors in MRLll autoimmune mice. Arthritis Rheum 27:1165-1173, 1984
7. Zvaifler NJ: The immunopathology ofjoint inflammation
in rheumatoid arthritis. Adv Immunol 13:265-298, 1973
8. Eickenberg HU: What is interstitial fluid? Biochemical
and physiological analysis of fluid obtained from tissue
cages. Scand J Infect Dis [Suppl] 14:166-170, 1978
9. Chisholm GD, Waterworth PM, Calnan JS, Garrod LP:
Concentration of antibacterial agents in interstitial tissue
fluid. Br Med J 1569-573, 1973
10. Gibson H, Gaar KA Jr: Dynamics of implanted capsules. Fed Proc 29:319-332, 1970
11. Hadler NM: The biology of the extracellular space. Clin
Rheum Dis 7:71-97, 1981
12. Calnan JS, Plug J J , Chisholm GD, Taylor LM:
Pathophysiology of tissue fluid. Proc R SOC Med 65:
715-722, 1972
13. Izui S, Eisenberg RA, Dixon FJ: IgM rheumatoid factors
in mice injected with bacterial lipopolysaccharides. J
Irnmunol 122:209&2101, 1979
14. Izui S, Kelley VE, Masuda M, Yoshida H, Roths JB,
Murphy ED: Induction of various autoantibodies by
mutant gene Ipr in several strains of mice. J Immunol
133~227-233, 1984
15. Izui S, Lambert PH, Miescher PA: Determination of
anti-DNA antibodies by a modified '*'I labeled DNA
binding test: elimination of non-specific binding of DNA
to non-immunoglobulin basic proteins by using an anionic detergent. Clin Exp Immunol 26:425-434, 1976
16. Hara I, Izui S, Dixon FJ: Murine serum glycoprotein
gp70 behaves as an acute phase reactant. J Exp Med
I55:345-357, 1982
17. Maruyama N, Furukawa F, Nakai Y,Sasaki Y , Ohta K ,
Ozaki S, Hirose S, Shirai T: Genetic studies of autoimmunity in New Zealand mice. IV. Contribution of NZB
and NZW genes to the spontaneous occurrence of
retroviral gp70 immune complexes in (NZB x NZW)FI
hybrid and the correlation to renal disease. J Immunol
130:740-746, 1983
18. Zubler RH, Lange G, Lambert PH, Miescher PA: Detection of immune complexes in unheated sera by a
modified ''1 C l q binding test: effect of heating on the
binding of Clq by immune complexes and application of
the test to systemic lupus erythematosus. J Immunol
116:232-235, 1976
19. Izui S, Eisenberg RA, Dixon FJ: Subclass-restricted IgG
polyclonal antibody production in mice injected with
lipid A-rich lipopolysaccharides. J Exp Med 153:
324-338, 1981
20. Lerner RA, Wilson CB, del Villano BC, McConahey PJ,
Dixon FJ: Endogenous oncornaviral gene expression in
adult and fetal mice: quantitative, histologic, and physiologic studies of the major viral glycoprotein, gp70. J
Exp Med 143:151-166, 1976
21. Guyton AC: A concept of negative interstitial pressure
based on pressures in implanted perforated capsules.
Circ Res 12:399-423, 1963
22. Munthe E, Natvig JB: Immunoglobulin classes, subclasses and complexes of IgG rheumatoid factor in
rheumatoid plasma cells. Clin Exp Immunol I2:55-70,
23. Brenner BM, Hostetter TH, Humes HD: The molecular
basis of proteinuria of glomerular origin. N Engl J Med
298:826-834, 1978
24. Gauthier VJ, Mannik M, Striker GE: Effect of cationized antibodies in preformed immune complexes on
deposition and persistence in renal glomeruli. J Exp Med
156:76&777, 1982
25. Carroll P, Stafford D, Schwartz RS, Stollar BD: Murine
monoclonal anti-DNA autoantibodies bind to endogenous bacteria. J Immunol 135:1086-1090, 1985
26. Naparstek Y, Duggan D, Schattner A, Madaio MP, Goni
F, Frangione B, Stollar BD, Kabat EA, Schwartz RS:
Imrnunochemical similarities between monoclonal antibacterial Waldenstrom's rnacroglobulins and monoclonal anti-DNA lupus autoantibodies. J Exp Med 161:
1525-1538, 1985
27. Eilat D: Cross-reactions of anti-DNA antibodies and the
central dogma of lupus nephritis. Immunol Today
61123-127, 1985
28. Faaber P, Rijke TP, van de Putte LV, Capel PJ, Berden
GH: Cross-reactivity of human and murine anti-DNA
antibodies with heparan sulfate: the major glycosaminoglycan in glomerular basement membranes. J
Clin Invest 77:1824-1830, 1986
Без категории
Размер файла
864 Кб
retroviral, mode, autoantibodies, immune, antigen, cage, complexes, tissue, gp70, quantitative, space, implanted, mrl, extravascular, lprlpr, mice, arthritis, use, subcutaneous
Пожаловаться на содержимое документа