|
 For
many years Sydenham chorea (SC) has been considered the prototype of an
autoimmune disorder triggered by an infectious agent. Nevertheless, despite
its description more than 300 years ago by Thomas Sydenham, its defined
association with rheumatic fever in 1838, and linkage to a preceding group
A b-hemolytic streptococcal (GABHS) infection, there are numerous unanswered
questions about the underlying pathology and pathophysiology of SC. Recently,
clarification of the underlying mechanism in SC has assumed renewed importance
as the disorder has been proposed as a model for a spectrum of childhood
neurobehavioral disorders, including tics and obsessive-compulsive disorder,
termed PANDAS (pediatric autoimmune neuropsychiatric disorders associated
with streptococcal infection).
The
distinguishing clinical manifestation of SC is the presence of spontaneous,
involuntary choreiform movements affecting the face and extremities that
typically persist for several months before subsiding. Diagnosis is made
strictly on clinical observation. No confirmatory laboratory test for
SC is available, although an elevated titer against antistreptolysin O,
anti-deoxyribonuclease B, or anti-nicotinamide adenine dinucleotidase
is demonstrated in 80% of SC patients.
Onset
is usually between ages 5 and 15 years, and a female predominance is observed
in most studies. Associated neurological symptoms may include dysarthria,
hypometric saccades, hypotonia, weakness, hemiballismus, and gait disturbances
that correlate with severity of chorea. Affected individuals may present
with behavioral or emotional difficulties that predate the motoric abnormalities
by weeks to months. For example, most patients have concomitant psychological
dysfunction presenting as personality changes, emotional irritability,
distractibility, and age-regressed behaviors. Studies have also shown
a higher frequency of obsessive-compulsive symptoms in children with SC
or nonchoreic rheumatic fever. Motor or vocal tics and oculogyric crises
have also been reported in patients with SC. Despite a relatively common
occurrence of EEG abnormalities, seizures are uncommon. Rheumatic valvular
cardiac disease is seen in about one third of patients, although echocardiogram
screenings suggest a higher prevalence. Arthritis is uncommon.
The
outcome in SC is quite favorable. Most cases resolve in 1 to 6 months.
Some investigators have identified residual abnormal EEGs, minor lingering
motor abnormalities, persistent neuropsychiatric problems, and future
development of valvular heart disease. About 20% to 60% of cases have
recurrent episodes of chorea, usually within 1 to 2 years after the original
event. An individual may have multiple attacks with reactivation precipitated
by a GABHS infection or by another environmental stimulus. Pregnancy and
oral contraceptives may stimulate the reappearance of chorea after many
years of quiescence.
Magnetic
resonance imaging (MRI) and functional imaging studies in SC have localized
acute changes to the basal ganglia. In an MRI study evaluating the size
of the basal ganglia in 24 patients with SC and 48 matched controls, children
with SC had a 10% increase in size of the caudate and a 7% increase in
size of both putamen and globus pallidus. A limited number of postmortem
brain studies have demonstrated a variety of cerebral abnormalities, including
acute and chronic neuronal degenerative changes as well as vascular and
inflammatory lesions. Pathological changes within the cerebral cortex,
basal ganglia, and thalamus have included a reduction in neurons, cytoplasmic
and nuclear cell changes, gliosis, hyperemia, endothelial swelling, perivascular
round cell infiltration, and petechial hemorrhages. Primarily on the basis
of measurements of the dopamine metabolite homovanillic acid (HVA) in
cerebrospinal fluid, SC has been postulated to be associated with dopaminergic
dysfunction. Nevertheless, the results of studies have varied from showing
an elevated baseline HVA to reduced accumulation after administration
of probenecid.
The
greater prevalence of rheumatic fever among relatives compared with unrelated
subjects has led to suggestions of a genetic predisposition. The monoclonal
antibody D8/17 directed against a polymorphic protein on the surface of
B lymphocytes has been found to have an expanded expression in individuals
with rheumatic fever and is hypothesized to be a trait marker for susceptibility.
The monoclonal antibody reacts with epitopes present on a significantly
higher proportion of B lymphocytes in patients with acute rheumatic fever
or rheumatic heart disease (100%) than in controls (12%). A significantly
higher frequency of D8/17 positive cells has also been identified in children
with SC (89%) compared with healthy children (17%). How this marker relates
to the disease process is unknown, especially since it has been reported
in patients with other neuropsychiatric disorders of childhood onset who
have not had a diagnosis of rheumatic fever or SC. Researchers have also
failed to reproduce some of these findings with automated techniques,
rather than by manual counting of stained cells with a fluorescent microscope.
Studies
of human lymphocyte antigen (HLA) report conflicting results in detecting
associations between class I or class II major histocompatibility complex
antigens and acute rheumatic fever. Investigators have proposed that immunogenetic
susceptibility to rheumatic fever varies according to the major clinical
symptom manifested by the patient. Specific HLA alleles were overrepresented
in rheumatic carditis and arthritis subgroups, but were not increased
in frequency in a subgroup with SC only.
An
autoimmune process has been confirmed in rheumatic fever. For example,
in carditis, anti-heart antibodies cross-reactive with streptococcal antigens
have been well established. Myosin has been identified as the heart autoantigen,
and monoclonal anti-heart antibodies recognize streptococcal M protein
as well as streptococcal membranes.
In
contrast to the situation for the heart in nonchoreic rheumatic fever,
the case for autoimmunity in SC is less well defined. In SC, it is generally
hypothesized that antibodies against GABHS cross-react against CNS neurons
through the process of molecular mimicry. Support for the antineuronal
antibody proposal is based on results in three areas: the measurement
of serum antineuronal antibodies, proposed success of immunomodulating
therapies, and the establishment of an animal model (Fig.
1).
Husby
and colleagues were the first to describe antineuronal antibodies that
putatively arise in response to a GABHS infection. The investigators used
an immunofluorescent antibody staining technique to show specific cross-reactivity
of IgG to neuronal cytoplasmic antigens in caudate and subthalamic nuclei
in 47% (14/30) of acutely ill SC patients compared with 0% in child controls
(0/24) and 3% (1/31) in adult controls. In 1993, Swedo and colleagues,
using similar methods, showed that 91% (10/11) of SC patients (9 with
obsessive-compulsive behavior) tested positive for antineuronal antibodies,
but 50% (9/18) of healthy controls were also positive. More recently,
this immunofluorescence technique was used to study 40 individuals with
acute, chronic, or remote rheumatic chorea, compared with 40 controls;
antineuronal antibodies were found in 100%, 93%, and 44% of the SC patients,
respectively. In comparison, no control subjects had antineuronal antibodies.
A possible association between the antineuronal antibodies and GABHS is
supported by studies showing that antibodies to streptococcal M protein
can cross-react with human brain tissue.
Although
the results of the immunofluorescent antibody technique in SC are intriguing,
they nevertheless raise several concerns. The methodology uses a visually
graded estimate of immunofluorescent staining against tissue sections
from human caudate, putamen, and subthalamic nuclei. Normal brain itself,
however, especially in older adults, contains autofluorescing lipofuscins.
The degree of sensitivity and specificity of the assay raises further
concerns. Percent positive rates in SC cases have ranged from 47% to 100%,
while positive rates in control populations have ranged from 0% to 50%.
Moreover, there was a lack of correlation with clinical features. More
recently, other investigators have used enzyme-linked immunosorbent assay
(ELISA) and Western blot techniques to detect the presence of antineuronal
antibodies in the serum of patients with acute SC. ELISA titers against
subcellular fractions (synaptosome, mitochondria, and synaptic membrane)
from unfrozen postmortem human caudate and putamen showed no difference
between an SC cohort and control group (Singer et al., unpublished). The
presence of specific serum antibodies against a synaptosomal preparation
of putamen and caudate tissue was assessed by Western blotting and discriminant
analysis. Preliminary results show a separation in discriminant space
between the SC and control groupings in the putamen but not in the caudate.
Further analyses are in progress to determine the reproducibility of these
findings and to delineate specific epitopes that contribute to the changes.
The
responsiveness of SC symptoms to immunomodulatory treatment has been cited
as supporting evidence of an immune process. Swedo described a child with
SC who had improvement of choreic and obsessive-compulsive symptoms after
plasma exchange. Others have suggested that immunosuppressive measures
such as corticosteroids, intravenous immunoglobulin (IVIG), and plasmapheresis
are effective treatments of chorea in SC. The number of reported cases,
however, is small, and studies have been performed without appropriate
controls. Of interest, Voss and colleagues reported that IVIG treatment
failed to alter the natural history of acute rheumatic fever, with no
detectable difference in the clinical, laboratory, or echocardiographic
parameters of the disease process during the subsequent 12 months. Finally,
in an attempt to develop an animal model, Hallett and colleagues have
unilaterally infused SC serum into rodent subthalamic nucleus. After infusion,
apomorphine-stimulated ipsilateral circling was greater in rats infused
with SC sera than in animals receiving control sera. This model suggests
that sera from SC patients can cause immune-mediated, physiologically
meaningful damage to the rodent subthalamic nucleus. As this information
has yet to be published, further analysis of the model will be forthcoming.
The
mechanism by which autoimmunity, if present, may proceed in SC remains
unclear. Conventional wisdom holds that circulating antibodies do not
enter the CNS without a breach of the blood–brain barrier. Gadolinium-enhanced
MRI scans of two patients with SC suggest, however, that such a breach
of the blood–brain barrier may actually occur. Alternative mechanisms
in which the blood–brain barrier remains intact have been proposed.
For example, antigen-specific B cells can migrate into the CNS, possibly
as lymphoblasts, and after receiving antigen-dependent cytokine stimulation,
differentiate into antibody-producing plasma cells within the CNS. One
facet of rheumatic chorea unexplained by proposed mechanisms is its latent
onset, i.e., why is the onset of chorea delayed for 3 to 5 months after
GABHS infection, unlike the other features of rheumatic fever?
In
summary, an autoimmune hypothesis for SC has been built around the measurement
of serum antineuronal antibodies, results of immunomodulating therapy,
and limited studies in an animal model. Nevertheless, there are significant
concerns relating to each of these areas and further investigation into
the pathology of SC is required. The current wide acceptance of SC as
an “established” model for CNS autoimmunity highlights the need
for further studies of this hypothesis.
top of page
|
