|
 Sydenham
chorea (SC), a major manifestation of rheumatic fever (RF), is thought
to occur when antibodies directed against group A streptococcus (GAS)
cross-react with epitopes on neurons of the basal ganglia. In earlier
work with SC, Swedo and colleagues identified children who, in addition
to chorea, presented with obsessive-compulsive behavior. A precipitous
onset of childhood obsessive-compulsive disorder (OCD) after streptococcal
pharyngitis was subsequently described that shared many similarities to
SC but did not have chorea or clinical signs of RF such as arthritis and
carditis. Swedo and colleagues termed this subtype of childhood-onset
OCD, pediatric autoimmune neuropsychiatric disorders associated with
streptococcal infections (PANDAS).
Both
pathogen- and host-related factors appear to influence the risk of acquiring
RF, with only 2% to 3% of untreated individuals infected by GAS developing
RF. Susceptibility to RF is influenced by age, GAS serotypes, family history,
and environmental conditions. Children between the ages of 5 and 14 years
show the highest rate of this complication. The observation that RF is
more prevalent among relatives of the probands than unrelated controls
supports the hypothesis that susceptibility to RF is, in part, genetically
determined. Environmental influences, such as crowded living conditions,
may contribute to the risk of developing RF. Pathogen-mediated factors
play a role as well, with specific GAS serotypes conferring increased
susceptibility to RF, although genome-based analyses of GAS should lead
to identification of more specific virulence factors. In the absence of
carditis and arthritis, the diagnosis of SC is frequently a diagnosis
of exclusion. Elevated streptococcal titers at the time of presentation
suggest but do not prove a causative role. Similarly, the association
of streptococcal illness with children presenting with PANDAS may occur
coincidentally. Advances in identification of reliable clinical and/or
biological markers of these disorders could further our understanding
of pathophysiology and lead to increased specificity in diagnosis and
treatment.
Progress
toward identifying a molecular marker for RF began in the 1970s, when
Patarroyo and colleagues isolated an antibody from the serum of a multiparous
woman with several children who had developed RF. This alloantisera 883
reacted with B cells from 71% of patients with RF and 16% of controls.
Later, Zabriskie and colleagues produced two monoclonal antibodies by
immunizing mice with B cells from known 883-positive and 883-negative
RF patients, which, when used in combination, identified most patients
with RF. Later the monoclonal antibody D8/17 (mAb D8/17) was developed;
it reacts with epitopes expressed on expanded populations of B lymphocytes
from the majority of patients studied with documented RF, whether 883-negative
or 883-positive. This finding was remarkably consistent across the five
different geographic and ethnic populations tested.
However,
recent studies suggest that the discriminatory ability of D8/17 expression
may be reduced secondary to antigenic variation related to ethnicity.
Research on groups of patients with RF in northern India has found a lower
percentage of D8/17 positivity (62%–68%) and has found additional
B-cell epitopes that demonstrate more specificity. These differences may
be secondary to methodological, genetic, or environmental variations at
the different locations. Although some studies suggest that certain class
II MHC haplotypes are increased in patients with RF, D8/17 positivity
was not shown in one small study to associate with HLA classification.
Larger studies with improved technology would provide a more definitive
conclusion.
From
these early studies, mAb D8/17 was thought to recognize a constitutively
expressed antigen marker on antibody-positive B cells, and it was thought
that the level of D8/17 expression remains stable for years after subsidence
of active RF. However, levels of expression can be modified under certain
conditions; for example, corticosteroid treatment of RF patients has been
found to produce down-regulation. Moreover, transient increases above
an already elevated baseline occur during acute episodes of RF. No relationship
between D8/17 binding and disease characteristics, severity, or acuity
has been reported. One exception is a report of decreased percentage of
D8/17 positivity in patients with rheumatic chorea (~80%), compared with
those with other RF signs (~100%), leading the authors to conclude an
element of nosological uncertainty exists with SC.
How
the antigen D8/17 relates to the disease process is not understood. Research
suggests that it may serve as a trait marker for RF susceptibility. For
example, earlier studies with D8/17 suggest that increased binding is
specific to type of the poststreptococcal sequelae, as patients with poststreptococcal
glomerulonephritis did not demonstrate increased numbers of D8/17-positive
B cells. Several other immune-mediated diseases have not shown increased
binding. Further evidence to support that the cellular ligand to which
D8/17 binds may be constitutive is that increased binding is seen in 96%
of RF/rheumatic heart disease patients, 40.3% of unaffected siblings and
parents of RF relatives, and 6.7% of healthy controls. Other studies report
similar increases in relatives, especially female relatives. Antibodies
to these “rheumatic” antigens would suggest that these relatives
are at risk for subsequent development of RF; however, RF has been reported
in only one relative with a previously demonstrated increase in binding
to B cells. Although D8/17 expression appears increased in relatives,
its predictive value for risk of RF has not been determined.
The
molecular characteristics of the epitope to which mAb D8/17 binds and
its role in the pathophysiological process of RF are largely unknown.
Based on cross-reactivity experiments using the D8/17 antibody, studies
attempting to identify this B-cell antigen suggest homology to helical
coiled-coil molecules such as myosin, tropomyosin, and M6 protein of group
A hemolytic streptococcus. From this earlier work, many human organs and
tissues appear to express antigen reacting with D8/17. Strong binding
has been reported to smooth muscle, whereas weak fluorescence in the cytoplasm
of cortical and caudate neuronal cells has been detected. One interesting
study suggested that compartmentalization of B cells occurs with patients
with RF demonstrating the D8/17 antigen on peripheral B cells but not
tonsillar B cells, whereas the D8/17 antigen was present on the tonsillar
B cells from children without RF but not seen in peripheral B cells. No
other work has been reported on D8/17 expression within various areas
of the central nervous system or on the characterization of the putative
antigen.
Khanna
and colleagues conducted a study examining markers that indicate immune
activation and found no correlation with D8/17 binding. This finding led
them to conclude that the expansion in the number of B cells expressing
this antigen does not seem to reflect nonspecific B-cell activation in
response to infection. Many more cellular activation markers have been
identified in recent years, and comparison with D8/17 binding has yet
to be done. In our preliminary work, the intensity of D8/17 binding correlated
with duration of illness and was largely attributed to those patients
with concurrent comorbid OCD, tics, and attention-deficit/hyperactivity
disorder. Increased binding in those patients with a more chronic illness
may reflect longstanding effects of altered immune status or genetic up-regulation
of a cell surface protein, among many possibilities. A similar finding
was reported in which increases of D8/17 expression were observed with
increasing age up to the fifth decade. Of interest, a Russian study examining
mAb D8/17 binding to B cells in newborn relatives of patients with RF
found no binding; this finding suggests that up-regulation occurs sometime
postnatally. Another study related increased D8/17 binding to Na+/H+ antiporter
activity. This antiporter activity was increased in patients with RF and
rheumatic heart disease compared with healthy controls and patients with
atherosclerotic heart disease. The significance of how this relates to
pathogenesis is not known, but Na+/H+ antiporter is essential for the
maintenance and regulation of cell volume and intracellular pH.
As
the diagnosis of SC is often a diagnosis of exclusion, increased expression
of D8/17 has been proposed to help differentiate SC from other forms of
chorea. Subsequently, the possibility of an immune-mediated pathogenesis
of OCD/Tourette syndrome (TS) has generated interest in the potential
of monoclonal antibody D8/17 to identify patients having, or at risk for,
streptococcus-precipitated neuropsychiatric disorders. Of the studies
published to date, increased rates of binding of this monoclonal antibody
to B cells has been reported in patients meeting criteria for PANDAS,
childhood-onset OCD/TS, OCD (but not trichotillomania), anorexia, and
autism. The diagnostic specificity of this antibody and its relationship
to the pathophysiology of psychiatric disorders have yet to be established.
We are also aware of unpublished studies wherein group differences were
not obtained. Although D8/17 binding is specifically increased in patients
with RF when compared with other rheumatic illnesses, assessment among
several neuropsychiatric conditions is needed to confirm diagnostic specificity
of D8/17.
Most
of the D8/17 studies have used fluorescein isothiocyanate–conjugated
goat anti-mouse antibody to label mAb D8/17 and phycoerythrin-conjugated
murine anti-HLA-DR antibody to label B cells. In other words, all the
B cells label red and D8/17-positive B cells double-label green and red.
The cells are counted using a double-filter microscope, and the proportion
of green cells to red cells yields the level of D8/17 expression (typical
range 5%–40%). Most studies have defined 11% or 12% mAb D8/17 binding
as the threshold for positivity. This threshold ideally is based on the
upper 95% confidence limit of the healthy control population and is expected
to vary depending on such factors as the intrinsic differences of the
control population and assay technique unless standardized. Further studies
are also needed to determine sensitivity and specificity of the assay
as well as developing appropriate standardization of reagents and technique.
At this time, variability in the methodology limits the generalization
of conclusions.
MAb
D8/17 is an immunoglobulin M (IgM). Antibodies in this class have more
avidity or stickiness, and because of this, IgM antibodies are routinely
more difficult to use in assays than IgG monoclonal antibodies. This characteristic
could contribute to the possibility that the monoclonal antibody is binding
nonspecifically to a variety of cell surface proteins that may be up-regulated
for unrelated reasons. Identification of the antibody’s ligand should
provide further insights and improvements in technique. Recent modifications
in methodology have led most researchers to use flow cytometry and a more
specific antibody for B cells (CD19). Chapman and colleagues have published
a flow cytometric analysis comparing samples from patients with OCD/TS
to samples from controls and found increased binding of this antibody
to B cells in patients. Our experience with flow cytometry suggests that
distinct populations of D8/17-positive B cells are not present but binding
of mAb D8/17 to B cells is on a continuum. Figure
1 shows examples of flow cytometry dot plots from three subjects.
The percentage in the upper-right quadrant represents level of D8/17 positivity.
Additional improvements are needed such as appropriate positive and negative
controls to use for instrument calibration, a matched isotype control,
and standardization of instrument settings, antibody dilutions, and reagents.
Our recent work with flow cytometry methods and MAb D8/17 has found many
challenges with standardizing this assay, although group differences persist.
With
methodological refinements, further work with D8/17 may lead to more definitive
results, thereby advancing our understanding of whether some forms of
childhood-onset OCD/tics share susceptibility with RF. When we first embarked
upon this work, we assumed that an established body of research in rheumatology
could be applied to OCD and TS. Along the way we have discovered that
much of the early research needed to be reexamined using contemporary
immunological principles and procedures. At this time we would advise
caution against firm incorporation of monoclonal D8/17 data to establish
diagnosis or treatment trajectories in childhood neuropsychiatric disorders.
top of page
|
