Antibodies against fetal brain in sera of mothers with autistic children
Harvey S. Singera,⁎, Christina M. Morrisa, Colin D. Gausea, Pam K. Gillinb,
Stephen Crawfordc, Andrew W. Zimmermanb
aDepartment of Neurology, Johns Hopkins University School of Medicine, Baltimore, United States
bCenter for Autism and Related Disorders, Kennedy Krieger Institute, Baltimore, United States
cDepartment of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States
Received 25 September 2007; received in revised form 26 October 2007; accepted 14 November 2007
Serum antibodies in 100 mothers of children with autistic disorder (MCAD) were compared to 100 age-matched mothers with unaffected
children (MUC) using as antigenic substrates human and rodent fetal and adult brain tissues, GFAP, and MBP. MCAD had significantly more
individuals with Western immunoblot bands at 36 kDa in human fetal and rodent embryonic brain tissue. The density of bands was greater in fetal
brain at 61 kDa. MCAD plus developmental regression had greater reactivity against human fetal brain at 36 and 39 kDa. Data support a possible
complex association between genetic/metabolic/environmental factors and the placental transfer of maternal antibodies in autism.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Autoimmunity; Autistic disorder; Maternal antibodies; GFAP; Myelin basic protein; BDNF
Autism, with an estimated incidence of 1:150, is currently
recognized as the most common developmental disability among
children (CDC, 2007). Clinically, a period of apparent normal
neurodevelopment often precedes the identification of classical
deficits in areas of social interaction, communication and lan-
guage, and stereotypic behaviors (Rapin, 1997). At birth, autistic
brains are typically smaller than those of healthy infants, but
between 6 and 14 months of age undergo a period of accelerated
growth (Courchesne et al., 2003). Genetic, biochemical, and
environmental factors are most commonly mentioned as etio-
logical mechanisms in autism (Korvatska et al., 2002; Lawler
proposed (Cohly and Panja, 2005; Pardo et al., 2005).
To date, evidence that immune factors have a role in autism is
primarily circumstantial. Families with autism show clustering of
autoimmune disorders (Comi et al., 1999; Croen et al., 2005),
evidence of immune dysregulation (Gupta, 2000), and abnormal
levels of plasma immunoglobulins (Plioplys et al., 1994). Chil-
dren affected with autism have serum antibody reactivity against
epitopes located in both adult human and rodent cortical, sub-
cortical, and cerebellar brain regions (Cabanlit et al., 2007; Silva
et al., 2004; Singer et al., 2006), as well as against specific brain
proteins, e.g., glial fibrillary acidic protein (GFAP) and myelin
basic protein (MBP) (Singh et al., 1997, 1993). Based on these
findings, an acquired autoimmune abnormality that affects
dendritic fields and synaptogenesis has been proposed for some
cases of autism. A second autoimmune hypothesis, in contrast,
has suggested that the process begins in utero and is associated
with the placental transfer of maternal antibodies that, in turn,
interfere with fetal brain development. Evidence for the intra-
uterine immune hypothesis is limited. A small pilot study has
identified serum antibodies against embryonic rodent brain in
mothers with autistic offspring that were absent or reduced in
animal models have demonstrated that maternal antibrain
Journal of Neuroimmunology 194 (2008) 165–172
⁎Corresponding author. Johns Hopkins Hospital, David M. Rubenstein Child
Health Building, Suite2158, 200 N. WolfeStreet, Baltimore, MD 21287,United
States. Tel.: +1 410 955 7212; fax: +1 410 614 2297.
E-mail address: email@example.com (H.S. Singer).
0165-5728/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
antibodies are capable of crossing rodent placenta and causing
behavioral alterations in their offspring (Dalton et al., 2003;
Vincent et al., 2002).
The goal of this study is to expand our knowledge of serum
antibrain antibodies in mothers of children with autistic disorder
(MCAD). Western immunoblotting was performed with use of
These tissue samples were selected in order to identify serum
reactivity against fetal human epitopes, provide comparisons with
and evaluate differences in reactivity against developing and
this study were chosen on the basis of identified neuroanatomical
abnormalities in postmortem autism brains, magnetic resonance
imaging, and autoantibody studies in autistic patients (Bauman,
1991; Courchesne et al., 1988; Singer et al., 2006; Singh and
history of autoimmune disease, pregnancy, birth order, maternal
child with autism. We hypothesized that serum reactivity would
differ in mothers of children with autistic disorder and that results
2. Materials and methods
One hundred mothers of children with autistic disorder
(mean age 41±6 years; range 27–66) were recruited from the
Center for Autism and Related Disorders at the Kennedy
Krieger Institute. The study was approved by the Institutional
Review Board of the Johns Hopkins Medical Institutions.
Autism was diagnosed in children by the presence of abnor-
malities in social and communication development, marked
repetitive behavior, and limited imagination using the Diag-
nostic and Statistical Manual for Mental Disorders-IV (DSM-
IV) and Autism Diagnostic Observation Schedule-Generic
(Lord et al., 2000) or Childhood Autism Rating Scale
(CARS) (APA, 1994; Schopler et al., 1986). Children with
diagnoses of Asperger syndrome and Pervasive Developmental
Disorder-Not Otherwise Specified (PDD-NOS) were excluded.
All were judged clinically to have moderate to severe adaptive
deficits, or cognitive deficits (IQ, b70) by formal testing. All
affected children of mothers in the study had undergone medical
evaluations and were given the specific diagnosis of autistic
disorder by an expert in the field (AZ). Those with established
genetic or metabolic causes of autism were excluded.
Medical data on the 100 mothers of children with autistic
disorder participating in this study were obtained in personal
interviews by a masters level neonatal nurse (PG) with emphasis
placed on information pertaining to maternal and paternal age at
the child's birth,pregnancy histories (includingthe total number
of pregnancies before the birth of the proband), maternal,
paternal and familial (in first-degree relatives) autoimmune
disorders (AI) (including rheumatoid arthritis, autoimmune
thyroid disease, and psoriasis, among others; for further details,
see Comi et al. (1999) and Croen et al. (2005)) and the clinical
course of the affected offspring.
The control group consisted of 100 aged-matched mothers of
unaffected, non-autistic children (MUC, mean age 43±5 years;
range 27–66) from the local community. Criteria for selection
included no lifetime personal history of a DSM-IV diagnosis of
autism and no related disorders in their children or first-degree
relatives. No ethnic or racial groups were excluded from the
study. All clinical data were de-identified and entered into a
until use. Samples were coded and assayed by laboratory tech-
nicians who were blind to the diagnosis.
2.2. Serum IgG concentration measurement
Sandwich ELISA methods were used to measure serum IgG
concentrations in all 200 serum samples following the methodol-
ogy of Raux et al. (1999). Human IgG (Sigma) diluted in serum
generate a standard curve. Absorbance was measured at 450 nm
using a Model 680 Microplate Reader (Bio-Rad). Microplate
Manager software (Bio-Rad) normalized the data, generated a
2.3. Antineuronal antibody determinations
2.3.1. Tissue preparations
Fresh, unfixed, human fetal (17-week gestation, postmortem
interval less than 12 h) and adult brain tissues (postmortem
intervals less than 18 h) were obtained from the NICHD Brain
and Tissue Bank for Developmental Disorders at the University
of Maryland, Baltimore, MD. Fresh Brodmann's Area 9 (BA9)
and caudate were obtained from a 76-year-old male who died of
a cardiac disorder. Fresh cerebellum (CB) and cingulate gyrus
(CG) were obtained from a 36-year-old female who died of a
gunshot wound to the chest. Neither individual had evidence of
neurological disease. Rodent tissue included brains from adult
and embryonic rats (gestational day 18).
Brain tissues were homogenized in 0.9% NaCl (2.5 g of
tissue/10 ml of saline) containing protease inhibitors (1 μg/ml of
aprotinin, 10 μg/ml of leupeptin, 10 μg/ml of pepstatin, and
1 mM phenylmethylsulfonyl fluoride) in a Teflon-glass homo-
genizer on ice. A supernatant fraction was collected and aliquots
were storedat −80°C. Protein concentrationswere measuredby
the bicinchoninic acid (BCA) method (Pierce, Rockford, IL).
2.4. Western immunoblotting
2.4.1. Brain tissue
Serum samples from all 100 MCAD and 100 MUC subjects
BA9, caudate, cerebellum, and cingulate gyrus, embryonic rat
(E18), and adult rat brain.
166H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
Methodology used for immunoblotting was similar to that
previously described (Singer et al., 2006). A total of 30 μg of
brain tissue protein per sample was denatured, subjected to
electrophoresis, and transferred to 0.45-μm nitrocellulose. After
blocking, nitrocellulose was exposed to an individual's serum
diluted 1:500 for 90 min at room temperature. Secondary anti-
body was horseradish peroxidase-conjugated sheep anti-human
IgG (GE Biosciences, Pittsburgh, PA) diluted 1:3000. Mem-
branes were developed with GE ECL reagents according to the
vendor protocol, and blots were exposed to Denville Blue Bio
Films (Denville Scientific, Metuchen, NJ) for 60 s. Molecular
weights were estimated based on the distance migrated for
seven known molecular weight standards (Bio-Rad). Digital
image analysis and evaluation of Western blots were performed
by Quantity One (Bio-Rad), which creates quantitative den-
sitometric data of the blots showing the gray-intensity values (8-
bit gray values) vs. Rf values. For all bands on each blot,
Quantity One generates a peak intensity (pixel optical density)
for each band, assigns each peak a molecular weight, and
determines the peak height, defined as intensity for the entire
band. All peak height measurements were corrected for the
specific serum IgG content contained in each sample. The
density of an anti-IgG band, identified in human brain samples
at 50 kDa, was used to provide a gross measure of inter-gel
variation. Bands are localized at ±2–3 kDa to allow for impre-
cision in scanning and extrapolation of band locations.
2.4.2. Specific antigens (GFAP, MBP)
Serum from 20 MCAD (mean age 44±3 years; range 38–52)
and 20 MUC (mean age 44±4 years; range 38–51) were
randomly selected for evaluation of antibody reactivity against
GFAP and MBP. GFAP and MBP immunoblots were obtained
using 1 μg of human GFAP (American Research Products, Inc.,
Belmont, MA) and 3 µg of human brain MBP (US Biological,
in 15% acrylamide ready-gels (Bio-Rad). GFAP was transferred
to 0.45-μm nitrocellulose and MBP was transferred to 0.45-μm
Immobilon PVDF membranes. The membranes were blocked
to secondary antibody; HRP-conjugated sheep anti-human IgG
(GE Biosciences) diluted 1:3000. ECL detection was performed
and band presence/absence was determined visually.
To ensure the presence of specific antigens, blots were stripped
of serum antibodies using Restore stripping buffer (Pierce). Mem-
branes were then exposed to antigen-specific primary antibody;
mouse anti-human GFAP monoclonal antibody diluted 1:750
antibody diluted 1:3000 (Chemicon). Secondary antibody con-
sisted of either HRP-conjugated sheep anti-mouse IgG or HRP-
After washing,blotswere exposedtoDenville Blue BioFilmsand
band presence/absence was determined visually.
2.5. Serum levels of BDNF
In a post-hoc analysis, 25 MCAD (mean age 41±5 years;
range 33–53) and 25 MUC (mean age 43±5 years; range 31–
53) samples were selected for assay based on their Western
immunoblotting results showing high levels of antibodies
against human fetal brain at 61 kDa, 39 kDa, and 36 kDa.
plates using a sandwich enzyme immunoassay BDNF ELISA kit
(USBiological). Plates were pre-coated with rabbit anti-human
BDNF polyclonal antibody. Human serum was diluted 1:200 and
incubated overnight at 4 °C. Secondary antibody was mono-
clonal, biotinylated mouse anti-human BDNF diluted 1:1000.
After exposure to HRP-conjugated streptavidin diluted 1:1000
and washing, plates were developed with tetramethylbenzidine/
enhancer (TMB/E) solution. Sera BDNF concentrations were
measured by optical density at 450 nm on an automated Bio-Rad
Model 680 microplate reader. Standards were used to determine
the slope of OD (absorbances) vs. BDNF concentration. OD
sample before statistical analysis. All samples were assayed in
2.6. Statistical analysis
Statistical analyses were performed with SPSS 12.0 (SPSS
Inc., Chicago, IL) and Stata/SE 9.2 (Stata Corp LP, College
Station, TX), to test the hypothesis that anti-neuronal antibody
profiles differ among MCAD (cases) and MUC (controls).
BDNF assay calculated concentrations were analyzed as con-
tinuous variables. Western blot analysis provided information
about the presence of bands at each specific molecular weight
and band density as measured by corrected peak height OD
values. Band localization was assessed by comparing the pro-
portion of case and control mothers with each band using χ2
analysis. Fisher's exact test was used in place of the χ2test
when the numbers were small. Simple logistic regression was
Mother's age (mean±SD)
Maternal age at 1st pregnancy
Parity, mean and (range)
Number mothers with 1 AD child
2.6 (1–5)2.3 (1–6)
Number mothers with 1AD child
and 1 PDD/Asperger
Number mothers with 1 AD child
and 2 PDD/Asperger
Number mothers with 2 AD children
Number mothers with 2 AD children
and 1 PDD/Asperger
Mean interval between birth of child
with AD and blood draw
Number AD children with regression
History of autoimmune disease
8.3 (2–21 years)
AD = autistic disorder, PDD = pervasive developmental disorder.
167H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
also used to assess the association between the presence of each
band and case-control status. Band density was corrected for
serum IgG concentrations i.e., optical density (peak height)
measurement divided by the serum IgG concentration expressed
as mg/100 ml. Based on the presence of overlapping band
densities at 50 kDa, suggesting little inter-gel variation in re-
activity against human brain epitopes, no further corrections
were applied. Band density was treated as continuous and
quantified by calculating mean peak height. Analyses of cor-
rected band density were not performed unless there were at
least three samples in each group. The differences in band
density among groups were assessed using Student's t-test.
3.1. Clinical population
Information on maternal history, birth order, timing of blood
draw, and offspring regression are presented in Table 1. All
mothers were healthy at the time of blood draw. Eighty-seven
mothers had one child with autistic disorder, nine had two
affected children with autistic disorder, and four had one child
with autistic disorder plus one or more with Asperger syn-
drome or PDD-NOS. Of 101 offspring with autistic disorder,
representing one per family with the exception of two from a
mother with affected monozygotic twins, 48 had a positive
history for both social and language regression (Ozonoff et al.,
2005), based on parents' recall of decline in previously acquired
developmental milestones. MCAD had more than twice the
odds of having a personal history of autoimmune disease
(n=24) compared to MUC (n=12) (OR=2.32, p=0.028).
3.2. Serum IgG concentration
Measurements of serum IgG concentrations were similar in
3.3. Western blots for brain proteins
Numerous bands were identified in all subjects using fetal
and adult brain tissue epitopes from human and rodent sources
(Fig. 1). Analyses of band specificity (number within a clinical
group having a band at a specific molecular weight) as well as
corrected peak height OD values identified relatively few
significant differences ( p≤0.05) or trends ( p≤0.09). Compar-
isons of the total number of bands observed or the sum of all
corrected peak heights showed that for each brain tissue
evaluated there was no difference between MCAD and MUC
(data not presented). Quantitation of the IgG band at 50 kDa,
performed only with human epitopes, showed no mean dif-
ferences in values in BA9, caudate, cingulate gyrus, cerebellum
or fetal brain between MCAD and MUC groups. Preliminary
analyses identified peak heights to have a normal distribution
Fig. 1. Western blots from MCAD and MUC groups against human fetal and adult BA9 and rodent embryonic and adult brain tissues. Numerous bands (antigen–
antibody interactions) are observed in both mothers of children with autistic disorder (MCAD) and mothers of unaffected children (MUC). Arrows identify bands
showing differences between clinical groups.
168 H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
3.3.1. Human fetal brain
expressed as mean±SEM are presented in Table 2. Only results
showing statistical differences or trends are shown. At 36 kDa,
10% of MCAD had a band as compared to only 2% MUC.
MCAD, at 61 kDa, had a significantly larger corrected peak
heights (p=0.037) and at 39 kDa, corrected peak height was
greater (trend, p=0.085) than in MUC.
3.3.2. Human adult brain
Significant differences in band specificity were identified at
only two molecular weights: caudate at 155 kDa and BA9 at
63 kDa, with a trend in the cingulate gyrus at 129 kDa (Table 2).
Corrected peak height differences were identified in cingulate
gyrus at 91 kDa and cerebellum at 31 kDa, and trends in caudate
3.3.3. Rodent embryonic tissue
MCAD had greater band specificity at 36 kDa (MCAD,
n=48; MUC, n=31; p=0.010) and at 73 kDa (MCAD, n=47;
MUC, n=31; p=0.015). There were no differences in corrected
3.3.4. Rodent adult brain
Asignificant difference inbandspecificity was present atonly
27 kDa (MCAD, n=22; MUC, n=10; p=0.016). There was a
n=29, corrected OD 11.4±1.3; MUC, n=34, corrected OD 8.3±
3.4. Comparison of Western immunoblotting between fetal and
Changes identified in MCAD showed little overlap between
fetal and adult tissues in both rodent and human samples. The
corrected peak height against human fetal brain at 61 kDa and a
3.5. GFAP and MBP immunoblots
There were no significant differences between 20 MCAD
and 20 MUC regarding antigen reactivity: GFAP, MCAD=2,
MUC=3; MBP, MCAD=1, MUC=2.
3.6. BDNF serum concentration
There was no significant difference in the calculated concen-
tration of BDNF in serum from 25 MCAD (52.07 ng/ml±
17.15) versus 25 MUC (54.87 ng/ml±17.17; p=0.57).
3.7. Comparisons between clinical factors and fetal brain
Analyses on the subgroup of 47 maternal samples with
“regression in offspring” were performed to determine whether
there was an association between clinical regression and the pre-
sence of specific fetal epitopes. Results, presented in Fig. 2, show
that there is a significant association between a mother having
offspring with autistic disorder and their possessing a serum an-
Analyses performed to determine whether the presence of a re-
active band against fetal human brain at 36, 39, or 61 kDa was
13 multiplex mothers the serum of five had antibody reactivity to
humanfetalbrainat61kDa andone at39kDa.Thefive multiplex
mothers with reactive bands at 61 kDa had at least two children
with the diagnosis of autism and no unaffected offspring. A re-
active band at 36, 39, or 61 kDa did not predict outcome of future
offspring, since sixteen mothershad anunaffectedchild following
the birth of a proband with autistic disorder. No positive corre-
lations werefound between the presenceof reactivebands and the
the birth order of the child with autism.
Band specificity (number of subjects with band) and band density (corrected
peak height)a: Significant differences and trends between MCAD and MUC
with human tissue as the epitope
MCAD95% CI MUC95% CIp-
Adult human cingulate gyrus
0.012 9.7–20.3 6.1–11.1
0.072 8.6–15.3 6.5–10.2
0.088 6.6–14.1 5.2–8.7
Number with3927 0.049
Peak height values are expressed as mean±SEM. Only results with statistical
aAll peak height values were corrected for sample specific IgG content.
169H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
A variety of statistical analyses were performed to evaluate
the influence of a positive history for autoimmune disorders
(AI). No difference in reactivity to any specific fetal brain
epitope was identified in comparisons between MCAD with
(n=24) and without (n=76) AI. Forty-seven MCAD had chil-
dren with developmental regression; 13 of this group had a
positive history of AI and 34 lacked this history ( p=0.42).
Further analyses of MCAD with offspring having regression
plus the presence of specific anti-fetal brain reactivity (shown in
Fig. 2) demonstrated no significant contribution from AI. In
MCAD with a band at 36 kDa, 2/8 had AI; at 39 kDa, 4/10 had
AI; and at 61 kDa, 9/17 had AI.
No significant correlations were identified between maternal
or paternal age at the birth of the child with autistic disorder and
reactivity against fetal human brain. Mothers with a band at
36 kDa were 2.2 years older on average than mothers without
this band, but the association was not significant.
The majority of prior studies investigating an autoimmune
mechanism in autism have focused on the presence of abnormal
serum antibodies in subjects diagnosed with this disorder
(Cabanlit et al., 2007; Silva et al., 2004; Singer et al., 2006;
Singh and Rivas, 2004). Although providing important
contributions, these studies have not addressed whether the
immune trigger could be the result of prenatal environmental
factors, such as the maternal–fetal transfer of autoantibodies. In
order to answer this question, immunoblotting and analyses of
scanned blots were performed on serum from mothers of chil-
dren with autistic disorder using fresh postmortem human and
rodent, fetal and adult, brain tissues as antigenic substrates.
In this study, reactive bands were considered atypical if
present in more individuals in one group than another, or if the
optical density of the band (peak height), corrected for maternal
IgG content, was greater. Results show that sera from mothers
of children with autistic disorder contain antibodies that differ
from controls against prenatally expressed brain epitopes. Few
studies have been performed investigating maternal antibodies
against fetal epitopes in autism. In a study presented in abstract
form, sera from 61 mothers of children with autistic disorder
had more prevalent antibody reactivity against human fetal
brain tissue, with reactive bands at approximately 32, 37, 73,
and 100 kDa (Braunschweig et al., 2006). Maternal autore-
activity to the band at 37 kDa was found to confer the greatest
risk for autism. In a second small study, serum from mothers of
children with autistic spectrum disorder (n=11) had reactivity
against embryonic rat brain that differed from mothers of
healthy children, with extra bands identified in the 75–100 kDa
range (Zimmerman et al., 2007). Recognizing that molecular
weight determination is an estimate, usually ±2–3 kDa, results
in studies using human or rodent fetal brain as epitopes have
areas of potential similarity. Further conclusions, however, must
await the identification of the specific reactive brain epitopes.
Whether there are definite pathological consequences to the
fetus secondary to the presence of maternal antibodies remains
to be determined. Based on our results, however, maternal
sensitization alone likely does not account for autism. For
example, although mothers having autistic children with deve-
lopmental regression were more likely to have serum antibody
reactivity against human fetal brain at 36 and 39 kDa, mothers
with unaffected children also have reactivity to similar epitopes.
Additionally, despite documenting that five multiplex mothers
had serum antibodies at 61 kDa and solely affected offspring,
MCAD possessing similar antibodies had normal offspring
following the birth of an autistic child. Hence, rather than a
direct association, there is likely a complex relationship be-
tween maternal anti-fetal brain antibodies and genetic/meta-
bolic/environmental factors. We hypothesize that the fetal brain
tissue epitope(s) associated with developmental regression will
be identified as an essential developmental growth factor(s), but
are not GFAP, MBP, or BDNF as previously speculated
(Miyazaki et al., 2004; Nelson et al., 2001; Singh et al., 1997).
Circumstantial evidence for a maternal placental–fetal
transfer of IgG hypothesis in autism is expanding: HLA-DR4
has been found with increased frequency in MCAD and their
sons with autism (Lee et al., 2006; Rogers et al., 1999; Torres
et al., 2002); antibrain antibodies reactive against specific
epitopes in fetal brain tissue are present in mothers with autistic
offspring; and animal models have indicated that maternal anti-
bodies can cross the placenta,bind tofetal antigens,andresultin
behavioral changes (Dalton et al., 2003; Vincent et al., 2002).
Epidemiological studies, including this report, suggest an asso-
ciation between autoimmune disorders in mothers and the
Fig. 2. Mothers with offspring having developmental regression with and without specific band reactivity. Mothers with offspring having developmental regression
and reactivity to fetal human brain epitopes at 36, 39, and 61 kDa are presented as filled bar, and those without the band are shown as hatched bars. Significance is
determined by odds of having regression based on presence of band using the logistic regression analysis.
170H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
prevalence of offspring with autism (Comi et al., 1999; Croen
et al., 2005; Lee et al., 2006; Molloy et al., 2006; Sweeten et al.,
2003). Nevertheless, greater serum antibody reactivity was not
confirmed in this group, nor was there a correlation with off-
spring regression. Possible explanations include diagnoses
based on history and relatively small numbers of subjects. The
hygiene hypothesis, which claims a reduced incidence of in-
other agents (Bach, 2005; Strachan, 1989; Vercelli, 2006), has
been cited to explain an increased incidence of autoimmune
disorders (Fleming and Fabry, 2007) including autism (Becker,
2007). Bidirectional trafficking of cells and nucleic acids
between a pregnant woman and her fetus during pregnancy
has resulted in the long term persistence of fetal cells in the
mother (fetal microchimerism) (Adams and Nelson, 2004;
Bianchi, 2007). Whether this mechanism, however, could ex-
plain the presence of maternal anti-fetal brain antibodies that
persist for years after the birth of the affected child is unknown.
Investigators have previously identified serum antibody
reactivity against brain tissue in children with autism. Children
with autistic disorder have more serum antibody-induced reac-
tive bands at 100 kDa against caudate, putamen and prefrontal
cortex and more dense bands at 73 kDa in the cerebellum and
cingulate cortex (Singer et al., 2006), compared to controls.
Although there is no direct overlap between the reactivity pre-
viously reported in children with autism and autoantibody
reactivity identified in MCAD, to our knowledge, no study has
simultaneously evaluated the antibody status of mothers and
their affected and unaffected offspring.
This study has several weaknesses including the analysis of
antibody patterns years after the delivery of the affected offspring,
the selection of mothers skewed towards having offspring with
time point, the failure to correlate antibody alterations in maternal
sera to that in their offspring, and the exclusion of groups
containing mothers of autistic children with established genetic or
metabolic causes of autism. Additional techniques, such as
measurement of actin, would be beneficial in confirming the
accurate pipetting of brain proteins. We further emphasize that
confirmatory evidence for a placental-transfer autoimmune
mechanism in autism is lacking. More specifically, only one of
five criteria deemed necessary to establish a pathogenic role for
induction of symptoms with autoantigens, and passive transfer of
the disorder to animal models) (Archelos and Hartung, 2000) has
been established in autism. We await the results of future studies
that include antibody evaluations in children with autism, their
parents, and unaffected siblings, identification of specific protein
epitopes, and confirmation that specific maternal IgG can induce
an immune response in fetal brains, that in turn is associated with
postnatal developmental abnormalities and behavioral alterations.
This research was supported in part by a grant from National
Alliance for Autism Research. The authors also thank Shilpa
Vernekar, M.D. for her assistance in performing laboratory
Adams, K.M., Nelson, J.L., 2004. Microchimerism: an investigative frontier in
autoimmunity and transplantation. JAMA 291, 1127–1131.
Diagnostic and Statistical Manual of Mental Disorders: DSM-IV. American
Psychiatric Association, Washington, D.C.
Archelos, J.J., Hartung, H.P., 2000. Pathogenetic role of autoantibodies in
neurological diseases. Trends Neurosci. 23, 317–327.
Bauman, M.L., 1991. Microscopic neuroanatomic abnormalities in autism.
Pediatrics 87, 791–796.
Becker, K.G., 2007. Autism, asthma, inflammation, and the hygiene hypothesis.
Med. Hypotheses 69 (4), 731–740.
Bianchi, D.W., Robert, E., 2007. Gross lecture. Fetomaternal cell trafficking:
a story that begins with prenatal diagnosis and may end with stem cell
therapy. J. Pediatr. Surg. 42, 12–18.
Braunschweig, D.A., Ashwood, P., Van de Water, J.A., Hertz-Picciotto, I.,
Hansen, R., Croen, L., 2006. Increased Prevalence of Maternal Auto-
antibodies Against Fetal Brain in Autism. International Meeting for Autism
Research, Montreal, Canada.
Cabanlit, M., Wills, S., Goines, P., Ashwood, P., Van de Water, J., 2007. Brain-
specific autoantibodies in the plasma of subjects with autistic spectrum
disorder. Ann. N. Y. Acad. Sci. 1107, 92–103.
Centers for disease control and prevention. Autism and developmental
disabilities monitoring network surveillance year 2002 principal investiga-
tors. Prevalence of Autism Spectrum Disorders—Autism and Develop-
mental Disabilities Monitoring Network, 14 Sites, United States, 2002.
MMWR Surveill Summ, vol. 56, pp. 12–28.
Cohly, H.H., Panja, A., 2005. Immunological findings in autism. Int. Rev.
Neurobiol. 71, 317–341.
Comi, A.M., Zimmerman, A.W., Frye, V.H., Law, P.A., Peeden, J.N., 1999.
Familial clustering of autoimmune disorders and evaluation of medical risk
factors in autism. J. Child Neurol. 14, 388–394.
Courchesne, E., Carper, R., Akshoomoff, N., 2003. Evidence of brain
overgrowth in the first year of life in autism. Jama 290, 337–344.
Courchesne, E., Yeung-Courchesne, R., Press, G.A., Hesselink, J.R., Jernigan,
T.L., 1988. Hypoplasia of cerebellar vermal lobules VI and VII in autism.
N. Engl. J. Med. 318, 1349–1354.
Croen, L.A., Grether, J.K., Yoshida, C.K., Odouli, R., Van de Water, J., 2005.
Maternal autoimmune diseases, asthma, and allergies, and childhood autism
spectrum disorders. Arch. Pediatr. Adolesc. Med. 159, 151–157.
Dalton, P., Deacon, R., Blamire, A., Pike, M., McKinlay, I., Stein, J., Styles, P.,
Vincent, A., 2003. Maternal neuronal antibodies associated with autism and
a language disorder. Ann. Neurol. 53, 533–537.
Fleming,J., Fabry, Z., 2007.The hygienehypothesisand multiplesclerosis.Ann
Neurol 61, 85–89.
Gupta, S., 2000. Immunological treatments for autism. J. Autism Dev. Disord.
Korvatska, E., Van de Water, J., Anders, T.F., Gershwin, M.E., 2002. Genetic
and immunologic considerations in autism. Neurobiol. Dis. 9, 107–125.
Lawler, C.P., Croen, L.A., Grether, J.K., Van de Water, J., 2004. Identifying
environmental contributions to autism: provocative clues and false leads.
Ment. Retard. Dev. Disabil Res Rev 10, 292–302.
Lee, L.C., Zachary, A.A., Leffell, M.S., Newschaffer, C.J., Matteson, K.J.,
Tyler, J.D., Zimmerman, A.W., 2006. HLA-DR4 in families with autism.
Pediatr. Neurol. 35, 303–307.
Lord, C., Risi, S., Lambrecht, L., Cook, E.H.J., Leventhal, B.L., DiLavore, P.C.,
Pickles, A., Rutter, M., 2000. The autism diagnostic observation schedule-
generic: a standard measure of social and communication deficits associated
with the spectrum of autism. J. Autism Dev. Disord. 30, 205–223.
Miyazaki, K., et al., 2004. Serum neurotrophin concentratins in autism and
mental retardation: a pilot study. Brain Develop. 26, 292–295.
Hyman, S.L., McMahon, W.M., Goudie-Nice, J., Hepburn, S., Minshew, N.,
171 H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172
Rogers, S., Sigman, M., Spence, M.A., Tager-Flusberg, H., Volkmar, F.R., Download full-text
Lord, C., 2006. Familial autoimmune thyroid disease as a risk factor for
regression in children with Autism Spectrum Disorder: a CPEA Study.
J. Autism Dev. Disord. 36, 317–324.
Nelson, K.B., Grether, J.K., Croen, L.A., Dambrosia, J.M., Dickens, B.F., Jelliffe,
L.L., Hansen, R.L., Phillips, T.M., 2001. Neuropeptides and neurotrophins in
neonatal blood of children with autism or mental retardation. Ann. Neurol. 49,
Ozonoff, S., Williams, B.J., Landa, R., 2005. Parental report of the early
development of children with regressive autism: the delays-plus-regression
phenotype. Autism 9, 461–486.
Pardo, C.A., Vargas, D.L., Zimmerman, A.W., 2005. Immunity, neuroglia and
neuroinflammation in autism. Int. Rev. Psychiatry 17, 485–495.
Plioplys, A.V., Greaves, A., Kazemi, K., Silverman, E., 1994. Lymphocyte
function in autism and Rett syndrome. Neuropsychobiology 29, 12–16.
Rapin, I., 1997. Autism. N. Engl. J. Med. 337, 97–104.
Raux, M., Finkielsztejn, L., Salmon-Ceron, D., Bouchez, H., Excler, J.L.,
Dulioust, E., Grouin, J.M., Sicard, D., Blondeau, C., 1999. Development
and standardization of methods to evaluate the antibody response to an HIV-
1 candidate vaccine in secretions and sera of seronegative vaccine recipients.
J. Immunol. Methods 222, 111–124.
McMahon, W.M., Spiker, D., Lotspeich, L., Kraemer, H., McCague, P.,
autism. J. Autism Dev. Disord. 29, 195–201.
Schopler, E., Reichler, R.J., Renner, B.R., 1986. The Childhood Autism Rating
Scale (CARS): for Diagnostic Screening and Classification of Autism.
Irvington, New York, p. 63.
Silva, S.C., Correia, C., Fesel, C., Barreto, M., Coutinho, A.M., Marques, C.,
Miguel, T.S., Ataide, A., Bento, C., Borges, L., Olifeira, G., Vicente, A.M.,
2004. Autoantibody repertoires to brain tissue in autism nuclear families.
J. Neuroimmunol. 152, 176–182.
Singer, H.S., Morris, C.M., Williams, P.N., Yoon, D.Y., Hong, J.J., Zimmerman,
siblings. J. Neuroimmunol. 178, 149–155.
Singh, V.K., Rivas, W.H., 2004. Prevalence of serum antibodies to caudate
nucleus in autistic children. Neurosci. Lett. 355, 53–56.
Singh, V.K., Warren, R., Averett, R., Ghaziuddin, M., 1997. Circulating
Singh, V.K., Warren, R.P., Odell, J.D., Warren, W.L., Cole, P., 1993. Antibodies
to myelin basic protein in children with autistic behavior. Brain Behav.
Immun. 7, 97–103.
Strachan, D.P., 1989. Hay fever, hygiene, and household size. Br. Med. J. 299,
Sweeten, T.L., Bowyer, S.L., Posey, D.J., Halberstadt, G.M., McDougle, C.J.,
2003. Increased prevalence of familial autoimmunity in probands with
pervasive developmental disorders. Pediatrics 112, e420.
Torres, A.R., Maciulis, A., Stubbs, E.G., Cutler, A., Odell, D., 2002. The
transmission disequilibrium test suggests that HLA-DR4 and DR13 are
linked to autism spectrum disorder. Hum. Immunol. 63, 311–316.
Vercelli, D., 2006. Mechanisms of the hygiene hypothesis—molecular and
otherwise. Curr. Opin. Immunol. 18, 733–737.
Vincent, A., Deacon, R., Dalton, P., Salmond, C., Blamire, A.M., Pendlebury, S.,
Johansen-Berg, H., Rajogopalan, B., Styles, P., Stein, J., 2002. Maternal
magnetic resonance spectroscopy. J. Neuroimmunol. 130, 243–247.
Zilbovicius, M., Garreau, B., Samson, Y., Remy, P., Barthelemy, C., Syrota, A.,
Lelord, G., 1995. Delayed maturation of the frontal cortex in childhood
autism. Am. J. Psychiatry 152, 248–252.
Zimmerman, A.W., Connors, S.L., Matteson, K.J., Lee, L.C., Singer, H.S.,
Castaneda, J.A., Pearce, D.A., 2007. Maternal antibrain antibodies in
autism. Brain Behav. Immun. 21, 351–357.
172 H.S. Singer et al. / Journal of Neuroimmunology 194 (2008) 165–172