вход по аккаунту


Characterization of immune complexes in progressive rubella panencephalitis.

код для вставкиСкачать
Characterization of Immune Complexes
in Progressive Rubella Panencephahtis
P. K. Coyle, MD," and Jerry S. Wolinsky, MD
I n two patients with the slow virus disease progressive rubella panencephalitis, high levels of immune complexes
were found in serum and, in one case, spinal fluid. These complexes contained immunoglobulin G. Serum complexes sedimented over a wide density range compared with the more homogeneous spinal fluid complexes. Complexes from both serum and spinal fluid contained antibody directed against rubella virus. Serum complexes were
also shown to contain rubella antigen (or antigens) that appeared to differ between the two patients.
Coyle PK, Wolinsky JS. Characterization of immune complexes in progressive rubella panencephalitis
Ann Neurol 9 557-562, 1981
Progressive rubella panencephalitis (PRP) is a rare
slow viral disease of the central nervous system
(CNS) [20]. The typical patient is male and has had
either congenital rubella or the naturally acquired,
self-limited childhood infection. After a decade of
normal or near normal development, neurological
signs appear consisting of intellectual deterioration
followed by prominent motor system abnormalities.
The disease has an ingravescent course [22].
It seems likely that this disorder is caused by reactivation of virus which has persisted since the initial
infection. Rubella virus has been recovered from
both brain biopsy tissue [G] and circulating mononuclear cells [2 31. Several diagnostically important
observations have been made. The cerebrospinal
fluid (CSF) immunoglobulin fraction is markedly
increased and shows an oligoclonal pattern by agarose gel electrophoresis; most of the immunoglobulin is reactive against rubella virus antigens [16].
The serum/CSF ratio of rubella antibody, measured
by either hemagglutination inhibition (HAI) or complement fixation, is markedly reduced, a finding
compatible with local production of antibody within
the CNS. Additional supportive data for immunoglobulin synthesis within the CNS include the
findings of homogeneous free light chains in CSF [ 161,
an increased percentage of RNA antibodies in CSF as
compared to serum [ 113, prominent plasma cell infiltration of brain, and measurements of 24-hour intrathecal IgG production [ 171.
Circulating immune complexes have been found in
sera from three patients with PRP. They have been
consistently demonstrated by both C Iq binding and
Raji cell assay in serial studies of two of these patients
[ 5 ] .In one, complexes were also found in CSF late in
the course of disease. This report details our attempts
t o characterize these complexes further.
Materials and Methods
Serum and CSF from two previously reported cases of PRP
were used in the present study. Patient 1 [15] had a history
of congenital rubella syndrome with progressive neurological dysfunction beginning at age 12% years. Patient 2
[21] had contracted naturally acquired rubella at age 7, and
neurological deterioration began at age 19. T h e test samples were all drawn late in the disease course, and aliquots
were stored frozen at -70°C o r lower until utilized in this
study; aliquots of these samples were used throughout. The
HA1 titers were: Patient 1 serum, 1 : 128; Patient 2 serum,
1 : 16; and Patient 2 CSF, 1 : 4 . Serum samples were negative for rheumatoid factor as determined by latex fixation
(Highland Diagnostics, Deerfield, IL). Control sera were
convalescent-stage human rubella antisera (Microbiological
Associates, Walkersville, MD), with HA1 titers of 1 : 128 or
greater, which lacked detectable immune complexes.
A modification of the Raji cell assay of Theofilopoulos e t
a1 [14] was used to detect and quantify immune complexes
in serum and CSF. A 25 p1aliquot of normal human serum
diluted 1 : 4 in Ca++- and Mg++-free phosphate-buffered
saline (PBS) was used as a complement source and added to
75 pl of either CSF or a 1 : 3 dilution of test serum in PBS.
After 30 minutes at 37"C, 25 pl portions of these preparations were then added to triplicate tubes containing 2 x 10"
Raji cells (HEM Laboratories, Rockville, MD) and incubated at 37°C for 45 minutes. Cells were washed three
times in PBS, and an optimum amount of radioactive
From the Department of Neurology, The Johns Hopkins University School of Medicine, Traylor Bldg, Room 72 1, Baltimore, M D
Received Sept 4 , 1980, and in revised form Oct 24. Accepted for
publication Nov 11, 1980.
Address reprint requests ro Dr Wolinsky.
'Present address: Department of Neurology, Health Sciences
Center, State Universiey of New York at Stony Brook, Stony
Brook, N Y 11794.
0364-5134/81/060557-06$01.25 @ 1980 by the American Neurological Association
iodine 12>-Iabeled, staphylococcus-derived protein A
(Amersham, Arlington Heights, IL) in PBS was used to
measure bound complexed IgG. The assay was quantitated
by plotting a curve of binding produced by known amounts
of heat-aggregated human globulin (AHG). Test samples
were recorded in microgram-equivalents of A H G per milliliter. With this method, as little as 18 pg/ml was detected
in serum and 6 pg/ml in CSF. In this laboratory, immune
complexes of unknown specificity are detectable in approximately 20T' of apparently normal donors (unpublished observations) and are infrequently detected in CSF
from patients with stable multiple sclerosis o r other neurological diseases 141. T h e test PRP samples had the following levels of immune complexes: 109 pg/ml (Patient 1
serum), 156 pglml (Patient 2 serum), and 40 pg/ml (Patient 2 spinal fluid). These results proved reproducible
even following a limited number of freeze-thaw cycles of
individual aliquots of the samples. Assay of other available
serum samples from these patients obtained over a threeyear interval late in their disease course always contained
immune complexes, which ranged from 20 to 294 pglml.
All other available CSF samples from these patients were
negative by Raji assay.
Sedimentation characteristics o f the complexes were
determined by ultraccntrifugation in isopycnic sucrose gradients. Samples of serum o r CSF (0.4 ml) were layered
onto 5 ml linear gradients consisting of 10 to 37% sucrose
(w/v) in 0.1 M PBS ( p H 7.4). The specimens were centrifuged at 150,000 g for 20 hours. Six-drop fractions (0.4
mlj were collected from the bottom of the gradient. Each
fraction was assayed for the presence of IgM and IgG using
radial immunodiffusion plates (Miles Laboratories, Elkhart,
IN) [ 101; these served as internal 1% and 7s reference
markers, respectively. To check for immune complex activity, duplicate 50 pl samples from the serum fractions o r
75 p l samples from the CSF fractions were brought up to
100 p1 with a 1: 4 dilution of normal human serum, and the
Raji cell assay was run as already outlined. Control serum
or CSF lacking complex activity from patients with other
neurological diseases was processed simultaneously. Consistent results were obtained on repeated analysis of test
sera and CSF on separate gradients.
T o bring about dissociation of immune complexes,
serum or CSF (0.4 ml) was incubated with 3 X 10' Raji
cells for 45 minutes at 37OC. Cells were washed three
times in PBS by centrifugation at 6 7 0 g for 5 minutes
each, and bound complexes were eluted and dissociated by
addition of 0.3 ml of 0.1 M sodium borate, p H 10, for a 10to 15-minute incubation at 37°C. The cells were then
sedimented at 1,500 g for 5 minutes, and the eluates were
collected and either examined directly or centrifuged to
equilibrium at 94,500g for 2 0 hours in a 10 to 25% linear
sucrose gradient in 0.1 M borate, pH 10. Ten-drop (0.7 ml)
fractions were then collected.
Antibody and antigen activity were determined by
radioimmunoassay, either directly in the eluates or on the
gradient-purified eluate fractions. Rubella antibody activity
was checked as follows. Commercial Rubelisa plates (Microbiological Associates) or polystyrene microtiter plates
(Dynatech Lab, Alexandria, VA) coated with purified Therien strain rubella virus were used as target antigens.
558 Annals of Neurology
Vol 9 No 6 June 1981
One-tenth milliliter of each test sample was plated per well.
After a 2-hour incubation at room temperature, the plates
were washed four times with PBS and 0.1 ml of an optimum amount of "'I protein A was added. The plates were
incubated for 1 hour at room temperature and were again
washed four times in PBS. Bound isotope was removed
from the commercial plates with two washes of 2 N sodium
hydroxide [ 3 ]and the washes counted in a gamma counter,
o r individual wells were cut from the flexible polystyrene
plates and individually counted. When commercial plates
were used, lZ"I activity from the control antigen well was
subtracted from the corresponding rubella antigen well.
To assay for rubella antigen, a sandwich radioimmunoassay was developed. This utilized 96-well flexible polystyrene microtiter plates coated with purified
human IgM having rubella antibody activity. Before plating, the serum was repeatedly passed over a column of
protein A bound to sepharose CL-4B (Pharmacia, Piscataway, NJ) to remove any protein A binding material. Peak
fractions containing IgM were shown to be free of detectable IgG activity by radial immunodiffusion (Highland
Diagnostics). This antirubella IgM (HA1 titer 1 :8 after removal of IgG) was diluted 1 : 50 in PBS, and 0.1 ml was
added to each of the microtiter wells. After an overnight
incubation at 37"C, 0.1 ml of 5% ovine serum albumin
(OSAj in PBS was added and the plates were incubated a
further 2 hours at 37°C. The wells were aspirated and the
test immediately carried out as follows. One-tenth milliliter
test samples were plated on duplicate wells for 2 hours at
37°C. The wells were then washed three times with 0.5%
OSA in PBS, and the samples were overlaid with 0.1 ml of
a 1 : 100 dilution of either a high-titer rubella antiserum or a
control serum lacking rubella antibody activity (Microbiological Associates). After a further hour-long incubation at 37"C,, the plates were again washed three times
and 0.1 ml of an appropriate concentration of ''I protein A
(1.6 x 10' cpm) was added for 1hour at room temperature.
The plates were then washed three times, individual wells
were cut from the plates, and ''>I binding activity was
counted. Control well counts (control sera) were subtracted from the corresponding rubella antiserum wells.
During standardization, the assay was shown capable of
detecting 0.1 hemagglutination unit of purified whole
rubella virus.
In addition, a monoclonal mouse antibody with activity
to the 62,000 dalton glycoprotein (gp62) hemagglutinin of
rubella virus (Wolinsky and Waxham, unpublished data)
was used in this system in place of the human rubella antisera; monoclonal antibody with activity against the gp78
hemagglutinin of mumps virus (kindly provided by D r
Alfred C. Server) served as the control.
Ultracentrifugation of test serum from Patient 1
showed that the immune complexes varied in size
from 7s to 1% o r greater over a density range of
1.15 to 1.26 gm/cm3 (Fig 1). Test serum complexes
from Patient 2 appeared more homogeneous, however, with a density range of 1.16 to 1.21 gm/cm3.
CSF complexes from Patient 2 were smaller than
1.15 .G
I I\
1.25 m1.20
1.15 -?
0 -
those found in serum, with a rather narrow density
range of 1.08 to 1.12 gm/cm3. Other selected sera
from both patients contained complexes that varied
in their sedimentation characteristics (data not
The Table outlines our results when dissociated
complexes eluted from Raji cells were directly assayed for rubella antibody activity. No activity was
found in control antisera. Serum complexes from Patient l were particularly high in rubella antibody
activity; eluates from both serum and spinal fluid
complexes from Patient 2 were positive for rubella
antibody activity as well.
To characterize this rubella antibody activity
further, the dissociated complexes from Patient 1
were studied on alkaline sucrose gradients. The activity appeared in a consistent portion of the gradient
(Fig 2). IgG was present by radial immunodiffusion in
those same fractions which showed rubella antibody.
The peak fraction was subjected to discontinuous
electrophoresis in a 5 % sodium dodecylsulfate
(SDS)-polyacrylamide gel under nonreducing conditions and was shown to have a calculated molecular
weight of approximately 150,000 daltons. Gradientpurified, dissociated complexes from serum and CSF
of Patient 2 (Fig 3 ) also showed similar rubella antibody peaks. Eluates of control samples, which included the high-titer rubella antisera as well as an
artificial complex ( AHG) that bound to the Raji cells
at a concentration of 160 &ml, showed no antibody
activity in fractions from paired gradients.
The Table shows the results obtained when eluates
containing dissociated cnmplexes were assayed for
rubella antigen using a human or monoclonal mouse
anticubella antibody in the sandwich radioimmuno-
Rubella Antibody and Antigen in Raji Cell Eluatesa
F i g 1 . Sedimentation characteristics of serum and CSF complexes: (A)Patient 1 serum; (Bi Patient 2 serum; (C) Patient
2 CSF. Serum and spinal fluid were fractionated by sucrose
density gradient ultracentrz$ugation, and individual fractions
were assayed for Raji binding activity. Complexes were found
in fractions from 7s t o 19s. The density of the complexes varied over a wide range zn serum; spinaljuid complexes were less
dense than those found in serum.
Patient 1
Patient 2
Patient 2
aRubella antibody activity was measured using a radioimmunoassay. Rubella antigen activity was measured in a sandwich
radioimmunoassay using either human antirubella IgG or monoclonal mouse antirubella I g G to bind antigen. Activity is expressed
as 1251 cpm binding; cpm was calculated as the difference in test
well minus control well binding.
ND = not done.
Coyle and Wolinsky: Immune Complexes in PRP 559
High Antibody Serum.-0
- 50
ot t o m
10 12
16 I8 2 0 22 2 4 26
F i g 2. Rubella antibody in gradient-purified, dissociated serum
complexes from Patient 1 . Serum complexes were bound t o Raji
cells, then eluted with borate bufhr at pH 10. This eluate was
fractionated by sucrose density gradient ultrarentrifugation,
and individual fractions were checked for rubella antibody
using a radioimmunoassay. The positiwfractions contained
IgG by radial immunodgfusion. An identically processed typical control serum with high rubella antibody activity but no
complex activitji is also shown.
11.05 E
assay. Antigen could not be detected in the relatively
low level CSF complexes. However, the dissociated
serum complexes from both patients were positive
when either human antirubella or monoclonal antigp62 antibody was used. The relative proportion
of antigen reacting with monoclonal antibody to
that reacting with polyvalent human antirubella antibody was greatly increased in Patient 1 as compared
to Patient 2. To characterize the antigen further, gradient-purified, dissociated complexes from Patient 1
were studied (Fig 4). Rubella antigen was detected
in a fraction distinct from those containing antibody activity. The density of this antigen was approximately 1.10 gm/cm3; purified whole rubella
virus has a modal density of 1.19 gm/cm3 in linear
sucrose gradients [9]. Figure 5 shows the results
for both the antibody and antigen assays carried out
on an identical gradient run on serum from Patient 1. Clearly, the complex components remained
dissociated during the procedure.
Immune complexes can be found in the serum of patients with a number of disorders [ l]. A central question is whether t h e complexes are related to the underlying disease, play a role in its pathogenesis, or are
an epiphenomenon of little importance [2, 191.
Complexes occur in normal persons who experience
a nonspecific insult such as an upper respiratory tract
560 Annals of Neurology
Vol 9
N o 6 June 1981
m +50/
F i g 3 , Rubella antibody in gradient-purijed, dissociated serum
(A)and C'SF ( B ) complexes from Patient 2. The eluates show
rubella antibody in the same portion of the gradient as for
Patient I .
infection [ 121, and this makes isolated positive
findings in groups of patients difficult to evaluate. It
becomes important, then, to develop techniques
whereby the component antigens and antibodies of
immune complexes can be identified. This is already
being done in experimental systems [13]. Among
human diseases, the ideal place to begin such attempts would he with an infectious disorder in which
the presumptive antibodies and antigens are already known. PRP offers such an opportunity. PRP is
a slow virus disease, caused by rubella, with high
levels of circulating immune complexes in serum and
CSF. In this study, rubella-specific IgG antibody
made up a large part of these complexes. Complexes
that contain partially degraded antibody are known to
persist longer in the circulation and are also more
likely to be deposited in tissues [ 7 , 81. However, the
antibody in our complexes appeared to be whole
IgG, as determined by SDS-polyacrylamide gel
electrophoresis. For a given individual, the size of the
complexes varied not only over time, but also within
the vascular and intrathecal compartments. The vari-
11.15 E
18 2 0 22 24 26
LBO t to m
F i g 4. Rubella antigen in gradient-purified, dissociated serum
complexesfrom Patient I . The eluate fractions uiere as.rayedfor
rubella antigen using a polyvalent human antiserum against
rubella in a sandwich radioimmunoassay. Rubella antigen activity was consistently found in an earlier part of the gradient
than rubella antibody activity. The density of this rubella
antigen was I . 10 gmlrm3.
+ 50C
Antibody Activity
Antigen Activity
0 - - 0
Fig 5. Rubella antibody and rubella antigen in gradientpurified, dissociated serum complexes from Patient 1 . Rubella
antibody was assayed on the identical fractions that bad been
analyzed for antigen in Figure 4 . The separation of antigen
and antibody is clearly displayed.
ability of complexes may relate to the proportions of
antigen o r antibody being produced in the host at a
given time and within a given body compartment.
The antigen component of the PRP complexes did
not appear to be whole rubella virus. These antigens
are probably heterogeneous, although in one patient
in our study the bulk of the complexes contained
antigens that reacted as the hemagglutinin of rubella.
The fact that complexes containing rubella antigen
could be found many years into the clinical course of
PRP suggests that rubella virus o r component anti-
gens must be produced continually. This production
could be occurring within brain, with release of virus
or viral antigens into the periphery, or at peripheral
sites, such as circulating mononuclear cells. Since
immune complexes are likely to be an indirect measure of continued complete or defective viral replication, they could provide an indicator for following
responses to drug therapy in PRP. If virus replication
could be eliminated, the complexes should then disappear.
We d o not yet know what pathogenetic role these
high levels of complexes may play in PRP. Perivascular glycoprotein and mineral-containing deposits
are present in the brains of patients with PRP and in
children with congenital rubella dying in the first
three years of life, but it is not known whether they
contain immune complex material. However, previous in vitro studies of PRP cases have shown inhibition of protein A-stimulated lymphocyte proliferation, of rubella virus-stimulated interferon
production [23], and of cell-mediated cytotoxicity to
rubella-infected cell cultures by unknown factors
present in serum [ 181. Immune complexes might be
capable of interfering with any of these responses.
Formation of high levels of complexes may interfere
with the effective immune response needed to clear
the virus.
The ability to recover antibodies and antigens
from these complexes offers a unique opportunity to
study the pathogenesis of this slow virus infection.
The affinity and functional ability of complexed antibody can be compared with free antibody found in
the same patients as well as with rubella antibody
from normal individuals. The antigens present in the
complexes at different times can be serially studied
with monoclonal antibody for evidence of antigenic
In summary, our study shows that a human slow
virus disease caused by rubella is associated with high
levels of circulating immune complexes in serum and
occasionally CSF. These complexes contain rubellaspecific IgG antibody and rubella antigen. The role of
the complexes and of their antibody and antigen in
the pathogenesis of PRP remains to be determined.
In addition, our system for dissection and analysis of
immune complexes is applicable to studies of other
human diseases in which such complexes are found.
Supported in part by US Public Health Service Grants NS-07000
and NS-15721 from the National Institute o f Neurological and
Communicative Disorders and Stroke, and aided by a Basic Research Grant from the National Foundation-March of Dimes
Dr Wolinsky is the recipient of Research Career Development
Award NS-00443 from the National Institute of Neurological and
Communicative Disorders and Stroke
Coyle and Wolinsky: Immune Complexes in PRP 561
1. Andrews BS, Penny R: The role of immune complexes in
the pathogenesis of disease. Aust NZ J Med 6:591-602,
2. Barnett EV, Knutson DW, Abrass CK, et al: Circulating immune complexes: their immunochemistry, detection, and importance. Ann Intern Med 91:430-440, 1979
3. Colombatti A, Hilgers J: A radioimmunoassay for virus antibody using binding of '251-labelled protein A. J Gen Virol
43:395-401, 1979
4. Coyle PK, Brooks BR, Hirsch RL, et al: Cerebrospinal fluid
lymphocyte populations and immune complexes in active
multiple sclerosis. Lancet 2:220-232, 1980
5. Coyle PK, Wolinsky JS, Griffin DE, et al: Circulating immune
complexes in progressive rubella panencephalitis contain
anti-rubella IgG (abstract). Neurology 30:427, 1980
6. Cremer NE, Oshira LS, Weil ML, et aI: Isolation of rubella
virus from brain in chronic progressive panencephalitis. J Gen
Virol 29:143-153, 1975
7. Haakenstad AO, Mannik M: The disappearance kinetics of
soluble immune complexes prepared with reduced and alkylated antibodies and with intact antibodies in mice. Lab Invest
35:283-292, 1976
8. Haakenstad AO, Striker GE, Mannik M: The glomerular deposition of soluble immune complexes prepared with reduced
and alkylated antibodies and with intact antibodies in mice.
Lab Invest 35:292-301, 1976
9. Liebhaber H , Gross PA: The structural proteins of rubella
virus. Virology 47:684-693, 1972
10. Mancini G , Carbonara AO, Heremans JF: Immunochemical
quantitation of antigens by single radial immunodiffusion.
Immunochemistry 2:235-254, 1965
11. Schuller EL, Delasnerie N, Allinquant B, et al: Intrathecal
rubella and RNA antibody synthesis in multiple sclerosis and
progressive rubella panencephalitis. Biomedicine 27: 139141, 1977
562 Annals of Neurology Vol 9 No 6 June 1981
12. Schwenk H U , Baenkler HW: Effect of gammaglobulin injection o n circulating immune complexes in various diseases. Eur
J Pediatr 131:43-48, 1979
13. Theofilopoulos AN, Eisenberg RA, Dixon FJ: Isolation of
circulating immune complexes using Raji cells. J Clin Invest
6111570-1581, 1978
14. Theofilopoulos AN, Wilson CB, Dixon FJ: The Raji cell
radioimmune assay for detecting immune complexes in
human sera. J Clin Invest 57:169-182, 1976
15. Townsend JJ, Baringer JR, Wolinsky JS, et al: Progressive
rubella panencephalitis. Late onset after congenital rubella. N
Engl J Med 292:990-993, 1975
16. Vandvik B, Weil ML, Grandien M, et al: Progressive rubella
panencephalitis: synthesis of oligoclonal virus-specific IgG
antibodies and homogeneous free light chains in the central
nervous system. Acta Neurol Scand 57:53-64, 1978
17. Weil ML, Itabashi H H , Cremer NE, et al: Chronic progressive panencephalitis due to rubella virus simulating subacute
sclerosing panencephalitis. N Engl J Med 292:994-998, 1975
18. Weil ML, Itabashi H H , Rola-Pleszcynski M, et al: Chronic
progressive panencephalitis due to rubella virus (abstract).
Arch Neurol 32:501-502, 1975
19. Wells JV: Immune complexes and disease. Aust N Z J Med
9:122-124, 1979
20. Wolinsky JS: Progressive rubella panencephalitis. In Vinken
PJ, Bruyn G W (eds): Handbook of Clinical Neurology.
Amsterdam, North-Holland, 1978, vol 34, pp 331-341
21. Wolinsky JS, Berg BO, Maitland CJ: Progressive rubella
panencephalitis. Arch Neurol 33:722-723, 1976
22. Wolinsky JS, Coyle P K Progressive rubella panencephalitis.
In Boese A (ed): Search for the Cause of Multiple Sclerosis
and Other Chronic Disease of CNS. Weinheim, Verlag
Chemie, 1980, pp 266-271
23. Wolinsky JS, Dau PC, Buimovici-Klein E, et al: Progressive
rubella panencephalitis: immunological studies and results of
isoprinosine therapy. Clin Exp Immunol 35:397-404, 1979
Без категории
Размер файла
559 Кб
progressive, immune, characterization, complexes, rubella, panencephalitis
Пожаловаться на содержимое документа