How Relevant are GFAP Autoantibodies in Autism and ... - Springer Link

J Autism Dev Disord (2008) 38:333–341 DOI 10.1007/s10803-007-0398-9

ORIGINAL PAPER

How Relevant are GFAP Autoantibodies in Autism and Tourette Syndrome? Nikki J. Kirkman Æ Jane E. Libbey Æ Thayne L. Sweeten Æ Hilary H. Coon Æ Judith N. Miller Æ Edward K. Stevenson Æ Janet E. Lainhart Æ William M. McMahon Æ Robert S. Fujinami

Received: 5 January 2007 / Accepted: 10 May 2007 / Published online: 20 June 2007 Springer Science+Business Media, LLC 2007

Abstract Controversy exists over the role of autoantibodies to central nervous system antigens in autism and Tourette Syndrome. We investigated plasma autoantibody titers to glial fibrillary acidic protein (GFAP) in children with classic onset (33) and regressive onset (26) autism, controls (25, healthy age- and gender-matched) and individuals with Tourette Syndrome (24) by enzyme-linked immunosorbent assays. We found a significant difference in autoantibody titers to GFAP, not accounted for by age, between the Tourette (significantly lower) and regressive autism groups. However, no differences were found between: classic/regressive; classic/controls; classic/Tourette; regressive/controls; or controls/Tourette. Autoantibody responses against GFAP are unlikely to play a pathogenic role in autism or Tourette Syndrome. Keywords Autism Tourette Syndrome Autoantibody Glial fibrillary acidic protein Immunoglobulin

Nikki J. Kirkman and Jane E. Libbey contributed equally to this manuscript. N. J. Kirkman J. E. Libbey T. L. Sweeten E. K. Stevenson R. S. Fujinami (&) Department of Neurology, University of Utah, 30 North 1900 East, 3R330 SOM, Salt Lake City, UT 84132-2305, USA e-mail: [email protected] H. H. Coon J. N. Miller J. E. Lainhart W. M. McMahon Department of Psychiatry, University of Utah, 30 North 1900 East, 5R110 SOM, Salt Lake City, UT 84132, USA

Introduction Autistic disorder (autism) is a neuropsychiatric developmental disorder manifesting with unusual social, communicative and behavioral development. The phenotypic heterogeneity is one of the most confounding aspects of this disorder. The etiology is unknown; however, both genetics and environmental factors play a role (Bailey et al., 1995; Coleman & Gillberg, 1985; Nelson, 1991). The prevalence rate of autism is currently estimated to be more than 10 per 10,000, and there is a biased male to female ratio of 3–4:1 (Fombonne, 2003; Yeargin-Allsopp et al., 2003). Over the last two decades, this disorder has been increasingly diagnosed throughout various countries (Chakrabarti & Fombonne, 2001; Gillberg & Wing, 1999; Yeargin-Allsopp et al., 2003). It is not known whether this increase is due to a combination of changing diagnostic practices, improved identification and availability of services, or an actual increase in incidence (Fombonne, 2003). Several lines of evidence suggest that autism may have autoimmunity as one of its many possible etiologies. Autoimmune disease has been found at an increased incidence in families with children with autism (Comi, Zimmerman, Frye, Law, & Peeden, 1999; Croen, Grether, Yoshida, Odouli, & Van de Water, 2005; Sweeten, Bowyer, Posey, Halberstadt, & McDougle, 2003). Involvement, in autism, of the human leukocyte-associated antigen (HLA) genes, which have been found to be important in disease susceptibility and outcome for some autoimmune diseases, has been shown by several studies (Lee et al., 2006; Torres, Maciulis, Stubbs, Cutler, & Odell, 2002; Torres et al., 2006; Warren et al., 1996). Anti-central nervous system (CNS) antibodies, such as autoantibodies to neurofilament protein (NFP), glial fibrillary acidic protein (GFAP), human endothelial cells, and an as yet

123

334

unidentified protein which is not myelin basic protein (MBP), have been detected in children with autism by some groups (Connolly et al., 1999; Silva et al., 2004; Singh, Warren, Averett, & Ghaziuddin, 1997), but not all (Todd, Hickok, Anderson, & Cohen, 1988). A recent Western-blot study found that the serum from more autism subjects than controls had reactivity to the caudate, putamen, prefrontal cortex, cerebellum and cingulate gyrus regions of the brain, however these findings could not be confirmed by ELISA (Singer et al., 2006). The presence of antibodies specific for CNS components in the circulation would suggest the possible involvement of an autoimmune process; however, it is not known whether the autoantibodies contribute to the development of the disorder or are a consequence of the disease. Tourette Syndrome is a childhood onset neurodevelopmental disorder of unknown etiology that is characterized by multiple motor and vocal tics as well as behavioral abnormalities (American Psychiatric Association, 2000; Jankovic, 2001). Similarities to autism include a heterogeneous phenotype, a male to female bias of 3–4:1 and strong genetic etiology (Jankovic, 2001). These children are of interest due to evidence that autoimmune mechanisms could be involved in the neuropathology of Tourette Syndrome (Kiessling, Marcotte, & Culpepper, 1993; Swedo et al., 1998; Yeh et al., 2006). In the study of Tourette Syndrome, two groups have demonstrated an antibody-mediated involvement by the induction of disease in rats after intracerebral injection of immunoglobulin (Ig) G or serum from patients with Tourette Syndrome (Hallett, Harling-Berg, Knopf, Stopa, & Kiessling, 2000; Taylor et al., 2002). Hallett et al. (Hallett et al., 2000) found that rats receiving intrastriatal microinfusions of sera or IgG from children with Tourette Syndrome produced more stereotypic behavior and episodic vocalizations than rats receiving similar infusions of sera or IgG from control children. Using immunohistochemical analysis, the presence of IgG selectively bound to striatal neurons was confirmed in the rats receiving serum infusions from children with Tourette Syndrome, but not from control children. Using similar methods Taylor et al. (Taylor et al., 2002) infused rats in the ventrolateral striatum with sera from Tourette Syndrome patients having high levels of anti-neural or anti-nuclear antibodies, sera from Tourette Syndrome patients with low levels of autoantibodies or sera from controls. Oral stereotypies significantly increased in rats infused with sera from Tourette Syndrome patients with high levels of autoantibodies compared to the controls. Taken together, these studies implicate a role for autoantibodies reacting to neurons of the basal ganglia in the pathophysiology of Tourette Syndrome; however, this is still controversial as two more recent studies have found no association between the level of anti-neuronal

123

J Autism Dev Disord (2008) 38:333–341

antibodies in the sera of children with Tourette Syndrome and the induction of behavioral abnormalities in rats subsequent to infusion of the sera (Loiselle, Lee, Moran, & Singer, 2004; Singer et al., 2005). Additionally, increased levels of anti-neuronal antibodies against putamen as determined by ELISA were seen in 41 patients with Tourette Syndrome compared with 39 controls (Singer et al., 1998). However, no association was shown between the level of such antibodies and the severity of tics, the age of onset, or the presence of coexisting disorders (Singer et al., 1998). In order for serum autoantibodies against CNS antigens to be relevant in disease, the antibodies must gain access to the target antigens in the CNS. The blood-brain barrier (BBB) provides protection to the CNS under normal circumstances [reviewed in (Hickey, 2001)]. Under disease conditions, the BBB can become permeable and allow entry to activated immune molecules such as autoantibodies. However, even in a state of health, there is movement of cells, including B lymphocytes, and proteins to and from the CNS. The cerebrospinal fluid (CSF) and CNS fluid have an inherent level of IgG due to the partitioning of Igs across the BBB based on their charge and molecular weight. Various cytokines can influence the permeability of the BBB [reviewed in (Hickey, 2001)]. A defining characteristic of several inflammatory CNS diseases, such as multiple sclerosis, is the presence of strong, narrowly focused families of IgGs, called oligoclonal bands, detected by protein electrophoresis of the CSF (Hickey, 2001; Meinl, Krumbholz, & Hohlfeld, 2006). These oligoclonal bands are generated by the entry of B lymphocytes into the CNS followed by clonal expansion, differentiation into plasma cells and secretion of antibody. These high concentrations of antibodies may or may not be proportionally represented in the serum (Hickey, 2001). For this reason, it would invariably be better to analyze CSF; however, it is difficult to obtain CSF from children, especially healthy control children. So as the next best thing, serum antibodies are generally analyzed. To investigate the role of autoantibodies in autism and Tourette Syndrome, we measured plasma antibody levels, by enzyme-linked immunosorbent assay (ELISA), to the CNS astrocyte protein, GFAP, in children with autism, both the classic onset (33) and regressive onset (26) forms, and children with Tourette Syndrome (24), compared to control children (25). We tested for differences between children with classic versus regressive autism to investigate the hypothesis of possible different immune/autoimmune characteristics in these groups. Regressive autism was characterized for this study as a period of apparently normal development followed by the loss of abilities in social interaction, imaginary play and/or language. Our definition of regression was very inclusive, as regression may involve


a wide range of presentations and occur at a variety of ages (Goldberg et al., 2003; Luyster et al., 2005). There is an onset of repetitive and restrictive behaviors, and loss of the previous ability to communicate either through non-verbal signals or speech with at least five coherent words. We also tested the autism groups against controls and children with Tourette Syndrome. The Tourette group was included in order to determine how specific results were for autism. If autism and Tourette subjects show positive results in the same direction, then the assumption is that the findings may be involved with phenotypes that are shared by the two diagnoses. In our current study, we found a significant difference in autoantibody titers to GFAP, which could not be accounted for by age, between the Tourette and regressive autism groups, as determined by analysis of variance (ANOVA). Autoantibodies to GFAP were significantly lower in children with Tourette Syndrome, compared to children with regressive autism. Our Western-blotting analysis testing for the presence of antibodies against GFAP in the plasma showed that controls were more likely to have these antibodies than children with autism (classic and regressive combined), though the difference was not significant.

Methods Subjects The study was approved by the University of Utah Institutional Review Board. Written informed consent was obtained from each subject or each subject’s parent or legal guardian after a complete description of the study. Affected participants were recruited from the Child and Adolescent Psychiatry Clinics at the University of Utah, School of Medicine and from the surrounding community. Control subjects were recruited from the surrounding community. All subjects met the Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition) (DSM-IV) criteria for autism or Tourette Syndrome as determined by a boardcertified child and adolescent psychiatrist or psychologist. To verify correct diagnosis, autism subjects, as well as controls, were administered the Autism Diagnostic Observation Schedule-Generic (ADOS-G) (Lord et al., 2000). Control subjects were excluded from the study if they showed any significant communicative or social concerns on the ADOS-G. Subjects with autism were subsequently administered the Autism Diagnostic Interview-Revised (ADI-R) (Lord, Rutter, & Le Couteur, 1994). ADI-R questions about possible loss in language or other abilities were used as regression screening questions (Goldberg et al., 2003; Luyster et al., 2005). If affirmative answers were obtained to any of these ADI-R regression screening

335

questions, then details concerning the onset, duration and types of skills lost were obtained. The regression group included those with a history of loss of verbal language with or without concurrent loss in non-language areas and those with a history of loss of social or play skills but no language loss (to include children with fewer than five words in their repertoire). To verify correct diagnosis, Tourette syndrome subjects were administered interviews utilizing the Family SelfReport Questionnaire For Tics, Obsessive-Compulsiveness, Attentional Difficulties, Impulsivity, and Motor Hyperactivity [Tourette Syndrome Association (TSA) International Consortium for Genetics, 1995, revised April 2000] that includes the Conners’ Parent Rating Scale-Revised (L), the semi-structured diagnostic interview Kiddie Schedule for Affective Disorders and Schizophrenia for school-age children, Present and Lifetime version (K-SADS PL) (Kaufman et al., 1997) and an interview based questionnaire concerning dates and durations of various physical symptoms, strep infections, and general behavioral changes. Subjects were primarily of Caucasian descent and were relatively healthy (colds—flu-like symptoms, ear infections, asthma and allergies were accepted) with no major diseases other than autism or Tourette Syndrome. Subjects were not on either SSRI’s (selective serotonin reuptake inhibitors) or risperidone (Risperdal, atypical antipsychotic), which is both a serotonin and dopamine receptor antagonist (Schotte, Janssen, Megens, & Leysen, 1993). Medications were coded, for the purpose of regression analysis, into five groups as follows: psychiatric medicines (antipsychotics, anticonvulsants, benzodiazepines, antidepressants), stimulants, anti-inflammatory medicines (ibuprofen, aspirin, Tylenol, steroids, antihistamines, allergy medicines), antibiotics and other drugs (anything not in the other four categories, except multivitamins which were not coded). Design Blood was drawn into BD Vacutainer CPTTM Cell Preparation tubes with sodium citrate (8 ml draw capacity, BD Vacutainer Systems, Franklin Lakes, NJ) and centrifuged to obtain plasma prior to freezing at –70C. Plasma from children with autism, healthy age- and gender-matched controls and children with Tourette Syndrome were compared for antibodies to GFAP. The subject groups were of similar age and gender composition, with the exception of the Tourette group which was significantly older. Antibody levels were determined by an ELISA. GFAP levels were adjusted for significant effects of age prior to further analysis. Other covariates (gender and medication status) did not have significant main or interaction effects on

123

336

GFAP levels, and were dropped. Additionally, the presence of antibodies against GFAP, in the plasma of 10 controls, 10 classic autism and 10 regressive autism subjects, was analyzed by Western blot. All individuals performing the ELISA and Western-blot assays were blinded as to the diagnosis of the individual samples tested. GFAP Antigen Preparation Human GFAP was purchased from Biodesign International (Saco, ME). Bovine GFAP was purified according to the method of Dahl et al. (Dahl, Crosby, Gardner, & Bignami, 1982). Briefly, 25 g of frozen bovine gray matter was minced and homogenized in 1 mM sodium phosphate, 5 mM EDTA, 1 mM b-mercaptoethanol, pH 8 (w/v 1:2). After stirring for 1 h, the solution was centrifuged for 30 min at 100,000 g using a Beckman 28.1 rotor. The pellet was re-suspended in 50 ml of the previously described phosphate buffer containing 2 M urea, incubated for another hour, and centrifuged again for 30 min at 100,000 g. The cleared supernatant was removed and diluted with the appropriate buffer to a final concentration of 6 M urea, 0.01 M sodium phosphate, 5 mM EDTA, 0.1% b-mercaptoethanol, pH 7.5. This extract was applied to a DEAE-cellulose (Whatman DE-52, Whatman, Florham Park, NJ) column equilibrated with the same buffer. A linear salt gradient of 25–200 mM NaCl was used to elute protein. Prior to this, the column was washed extensively with column buffer containing 25 mM NaCl. Elution fractions of 5 ml were collected and monitored by spectrophotometry (Milton Roy Company, Rochester, NY) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS–PAGE) (Sambrook, Fritsch, & Maniatis, 1989). Fractions containing purified GFAP were pooled and dialyzed against phosphate-buffered saline (PBS) and cleared by centrifugation at 150,000 g. Protein concentrations were determined using the Bio–Rad Protein Assay (Bio–Rad Laboratories, Hercules, CA). All procedures were performed at 4C. ELISA Flat-bottomed 96-well Nunc-Immuno plates, MaxiSorp surface (Nalge Nunc Int., Rochester, NY) were coated with 10 lg/ml GFAP solution in PBS overnight at 4C. Plates were washed with PBS containing 0.2% Tween-20 prior to blocking with PBS containing 0.2% Tween-20 plus 2% bovine serum albumin (BSA) (blocking buffer) for 60 min. The blocking buffer was used as the diluent throughout the assays. Plasma were serially diluted 2-fold across the plates, starting at concentrations as low as 1:64 (26), and incubated at room temperature for 90 min. After washing, 50 ll of horse-radish peroxidase-conjugated goat anti-human IgG

123


(Jackson ImmunoResearch, West Grove, PA) was added to each well at a dilution of 1:5,000. After a final wash, immunoreactive products were visualized by adding 50 ll/ well of substrate containing o-phenylenediamine dihydrochloride (Sigma, St. Louis, MO). The color was allowed to develop for 30 min in the dark and stopped by addition of 50 ll/well 1 M HCl. Wells were quantitated as absorbance at 492 nm on a Titertek Multiscan Plus microplate reader (Titertek Instruments, Huntsville, AL). All measurements were performed in duplicate or triplicate. Control wells that received no plasma were used as background values. All data was corrected by subtraction of these background values. Intra-assay coefficients of variation were determined by running a control sample multiple times within a plate. A control sample was repeatedly run on various plates throughout each assay for use in determining inter-assay coefficients of variation. Optical density values (reported as: optical density · 1,000) were reported at plasma dilutions where standard curve readings were linear and the coefficients of variation were optimal. The inter-assay coefficient of variation was 0.19 at the 1:512 dilution. The subjects with the two highest and two lowest results from the mean were subjected to duplicate assays, and no significant differences between the repeat assays were seen by a paired t-test analysis. Therefore, no samples were excluded from further analysis. Western Blot The presence of antibodies against GFAP, from 10 controls, 10 classic autism and 10 regressive autism subjects, randomly selected from within each group, was analyzed by Western blot (Fujinami, Zurbriggen, & Powell, 1988; Tolley, Tsunoda, & Fujinami, 1999; Yamada, Zurbriggen, Oldstone, & Fujinami, 1991). Histone H1 was included as a control. Human GFAP was 43–49 kDa with two major bands at 46 kDa and histone H1 was 32 kDa in size when separated on a 12% SDS–PAGE gel and stained with Coomassie blue (Sambrook et al., 1989). Human GFAP protein at a concentration of 1 lg per well was loaded into the appropriate number of wells of 12% SDS–PAGE gels. In addition the Blue Ranger Prestained marker (Pierce, Rockford, IL) was run on every gel. The proteins were electrophoretically separated on the gels and then transferred onto nitrocellulose paper using standard procedures (Sambrook et al., 1989). Blots were washed in TBST buffer (50 mM Tris–HCl, pH 7.5, 0.5 M NaCl, 0.05% Tween-20) and blocked with TBST buffer containing 5% (w/v) nonfat powdered milk. Plasma was diluted 1:400 in TBST buffer and applied to the blots in glass test tubes at room temperature (22C) for 90 min with shaking. Monoclonal mouse anti-GFAP antibody (IgG1 isotype, Sigma) diluted 1:1,000 was used as a positive control. The


337

autism, controls and children with Tourette Syndrome (Table 1, p = .2351). There was a significant difference in age found by ANOVA between the groups (Table 1, p < .001). Using the Bonferroni/Dunn post-hoc test, the difference was found to be significant between the classic autism versus Tourette groups (p < .001) and regressive autism versus Tourette groups (p < .001). This difference in age can be accounted for by the fact that children with Tourette Syndrome are usually diagnosed at a later age than children with autism. Antibody titers to GFAP from children with autism (both classic and regressive), controls and Tourette Syndrome were determined by ELISA. To investigate the effects of covariates on antibody level, we ran the SAS GLM procedure, predicting GFAP level from age, gender and medication status. There was a significant age effect (F = 4.76, p < .05, DF = 1, 106) on antibody titers to GFAP: as age increased, antibody titers to GFAP decreased. Gender and medication status had no significant effect on antibody titers to GFAP. No significance on antiGFAP level was found either using individual medication classes, or combining medications into a single dichotomous medicated-unmedicated variable (Table 2). No significant interaction effects among covariates were found. Using the residuals from the regression analysis controlling for the effects of age, the autism groups (classic and regressive), the Tourette group and the controls were compared by ANOVA. A significant difference in antibody titers to GFAP was found (Table 3, p < .05). Using the Fisher’s PLSD post-hoc test, the only significant group differences in autoantibody titers to GFAP were between Tourette Syndrome and regressive autism (p < .01). This effect is independent of age, as the residuals were used in this test. Autoantibodies to GFAP were significantly lower in children with Tourette Syndrome, compared to children with regressive autism (Fig. 1). The ELISA data for the classic onset and regressive onset autism groups show higher variability than the control group. In order to achieve better homogeneity, the ELISA data was log10 transformed and the ANOVA analysis was repeated. A statistically significant difference was found by ANOVA (F = 2.97, p < .05, DF = 3,104) which was determined to be between the regressive onset autism group and the Tourette Syndrome group (p < .01)

blots were washed and secondary goat anti-human peroxidase-conjugated IgG [heavy + light (H + L), Jackson ImmunoResearch] diluted 1:2,500 in TBST buffer, or goat anti-mouse peroxidase-conjugated IgG (H + L, Invitrogen, San Diego, CA) diluted 1:5,000 in TBST buffer, was applied respectively for 90 min at room temperature. The blots were washed and protein was visualized by incubation in fresh 4-chloro-1-napthol (Sigma) in methanol plus 30% hydrogen peroxide (Sigma) in TBST buffer for 30 min. The blots were washed with dH2O overnight and then dried. All subjects were repeated at least twice. Dried blots were scanned using an HP Scanjet 8200 and the HP Photo and Imaging Gallery software. Positive and negative scores were assigned based on the presence or absence of a band at the correct size on the blot. Band intensity varied from sample to sample so there was some subjectivity to the assignment of a positive score. The groups were compared using chi square analysis. Statistical Analyses All statistical analyses were performed using the StatView or SAS programs (SAS Institute Inc., Cary, NC). ANOVA, followed when indicated by the Fisher’s Protected Least Significant Difference (Fisher’s PLSD) or the Bonferroni/ Dunn post-hoc tests, were used to determine group differences for continuous data. For nominal data, such as gender, the chi-square test was utilized. After checking for basic group differences, the SAS General Linear Models (GLM) procedure was used to create residual ELISA scores adjusting for age, gender and medication status. These residual scores were used in subsequent analyses. We tested for differential effects of covariates within diagnostic group and within gender, but did not find significance. Covariate adjustment was therefore done using the entire sample, not within gender or within diagnosis. Statistical power was determined using the Power and Precision program (Borenstein, Rothstein, & Cohen, 2000).

Results Our initial tests focused on effects of individual covariates. We found that there were no significant differences in the gender ratio for the groups of classic autism, regressive Table 1 Age and gender

DF = Degrees of freedom; SD = standard deviation; F = female; M = male

Classic

Regressive

Control

Tourette

Number

33

26

25

24

Mean Age (SD)

7.3 (3.0)

6.7 (2.7)

8.9 (3.4)

10.0 (2.6)

Range (years)

3–15

3–12

3–13

4–14

Gender

2F, 31M

6F, 20M

3F, 22M

5F, 19M

Statistic

p value

DF

F = 6.729

0.0003

3, 104

v2 = 4.256

0.2351

3

123

338

J Autism Dev Disord (2008) 38:333–341 600

Table 2 Medications Regressive

Control

Tourette

500 400

Medicated–Unmedicated

12–21

15–11

5–20

14–10

Anti-Inflammatory

6

4

4

9

Antibiotics

2

1

1

3

Psychiatric

1

4

0

6

Stimulants

3

5

0

2

Other

3

12

0

0

OD*

Classic

300

**

200

**

100 0

Many subjects were taking multiple medications

Control

by Fisher’s PLSD analysis. The transformed data gave the same result as the previous analysis. In addition to the above ELISAs, the presence of antibody against GFAP in the plasma of randomly selected children with autism, 10 classic and 10 regressive, and 10 controls were assessed by Western blotting. Human GFAP has multiple bands when run on a 12% SDS–PAGE gel (Fig. 2A). The band intensity on the Western blots varied from sample to sample and the banding pattern was not always consistent (Fig. 2B). No significant differences were found in reactivity between the classic and regressive autism groups and no significant differences were found between the classic and regressive combined autism group and controls.

Discussion GFAP is the principal 8–9 nm intermediate filament protein present in mature astrocytes of the CNS [reviewed in (Eng, Ghirnikar, & Lee, 2000)]. It plays a role in modulating astrocyte shape and motility by providing structural stability to the astrocyte processes. Studies suggest a role for GFAP in normal white matter architecture and BBB integrity [reviewed in (Eng et al., 2000)]. The amount of GFAP in astrocytes increases gradually with age. Astrogliosis is the response of astrocytes adjacent to damage in the CNS and is characterized by the rapid synthesis of GFAP [reviewed in (Eng et al., 2000)]. We found that there is a significant difference in autoantibody titers to GFAP, which can not be accounted for by

Classic

Regressive

Tourette's

Fig. 1 Comparison of plasma autoantibody titers against glial fibrillary acidic protein (GFAP) between control, Tourette, classic onset autism and regressive onset autism subjects. *, OD = OD · 1,000, adjusted for age. Horizontal lines represent the mean optical densities (OD) (·1,000, adjusted for age) for each group. Tourette subjects had significantly lower plasma titers of autoantibodies to GFAP compared to regressive autism subjects (**, p < .01), as determined by enzyme linked immunosorbent assays (ELISA)

age, between Tourette Syndrome and regressive autism, as determined by ANOVA and the Fisher’s PLSD post-hoc test. Autoantibodies to GFAP were significantly lower in children with Tourette Syndrome, compared to children with regressive autism. However, there was no difference between the two autism groups or the Tourette group and controls for these measurements. Additionally, the classic onset and regressive onset autism groups were combined in order to compare them to the control group via a t test. The difference was not significant for either the data as presented (t = 1.34, p = .184, DF = 82) or the transformed (t = 1.09, p = .28, DF = 82) ELISA data. It is not surprising that the t test did not show a significant difference between the autism group and control, as the ANOVA also did not show a significant difference between classic onset autism and control, and regressive onset autism and control. If these autoantibodies were directly involved in an autoimmune process, one would expect titers to be increased, as previous ELISA data reported in autism for various neuron-specific antigens, such as MBP and myelin oligodendrocyte glycoprotein (MOG) (GFAP was not tested) (Vojdani et al., 2002). This previous study reported a significant increase (large effect) in the mean optical

Table 3 ANOVA statistical results for antibody to GFAP

GFAP

Classic

Regressive

Control

Tourette

F value

p value

DF

169 (142)

208 (149)

145 (78)

108 (60)

3.23

.03

3, 104

DF = Degrees of freedom; GFAP = glial fibrillary acidic protein Mean optical densities · 1,000, adjusted for age (standard deviation) Serum dilution was 1:512

123

J Autism Dev Disord (2008) 38:333–341 M

339

1

A

M

1

2

3

4

5

B

Fig. 2 Western blotting. (A) A Coomassie stained 12% SDS–PAGE gel showing the purity, size and banding pattern of the human glial fibrillary acidic protein (GFAP) used for the Western blots. M: Blue Ranger Prestained marker, from top to bottom: 206, 113, 82, 49, 32, 26, 17.8 kDa. Lane 1: 4 lg human GFAP. (B) Representative Western blots showing reactivity against 1 lg human GFAP. Lane 1: positive control monoclonal mouse anti-GFAP antibody; Lane 2: classic autism plasma scored positive for GFAP; Lane 3: regressive autism plasma scored positive for GFAP; Lane 4: control plasma scored positive for GFAP; Lane 5: control plasma scored negative for GFAP

density values, at a serum dilution of 1:100, for IgG, IgM and IgA to the nine neuron-specific antigens tested in the sera of autism subjects, compared to controls (Vojdani et al., 2002). Our sample size was comparable to the sample size used in this previous ELISA study which included 40 autism (classic and regressive not distinguished) and 40 control subjects. With our sample sizes, we should have been able to detect differences as large as those seen by Vojdani et al. (Vojdani et al., 2002) with 80% power (Borenstein et al., 2000) even if the observed standard deviation was twice what was observed in that study. However, we found no evidence of an increase in the level of antibodies to GFAP in autism subjects compared to our controls. Singh et al. (Singh et al., 1997) previously investigated autoantibodies to GFAP via Western blotting. They reported a significant increase in the incidence of autoantibodies to GFAP in autism subjects (17/53, 32% positive),

compared to control subjects (5/58, 9% positive). This study did not distinguish between classic onset autism and regressive onset autism. Although we performed Western blotting and plasma screening for the presence of autoantibodies to GFAP using the same methods as Singh’s group, we were unable to replicate their findings. In our study, no significant differences were found in reactivity between the classic and regressive autism groups and no significant differences were found between the classic and regressive combined autism group and controls. With our relatively small sample size of 20 autism (classic and regressive combined) and 10 control, and the significant heterogeneity in the clinical manifestations of autism and serum autoantibody levels, the lack of group differences in the Western-blotting experiment may be accounted for by Type II error. There may indeed have been a difference if we were able to test more samples. Previously, in autism, some groups have reported the presence of autoantibodies to neuronal tissues, including MBP, MOG, ganglioside, NFP, GFAP, temporal lobe cortex, cerebellum and caudate nucleus, while other investigators could not reproduce the results (Connolly et al., 1999; Plioplys, Greaves, Kazemi, & Silverman, 1994; Singer et al., 2006; Singh & Rivas, 2004; Singh et al., 1997; Singh, Warren, Odell, Warren, & Cole, 1993; Todd et al., 1988; Vojdani et al., 2002). Vargas et al. (Vargas, Nascimbene, Krishnan, Zimmerman, & Pardo, 2005) found no Ig deposition in the brain of subjects with autism using immunohistochemical analyses. The nonreplication results and lack of Ig deposition are in accord with our current negative ELISA and Western-blotting results. Additionally, although atypical prenatal maternal immune responses may be linked to the pathogenesis of autism, an immunoblotting study determined that neither GFAP nor MBP were specific targets of the maternal antibodies (Zimmerman et al., 2007). Therefore, we concluded that antibody responses against GFAP are unlikely to play a pathogenic role in autism in our sample group. With regards to Tourette Syndrome, serum anti-neuronal antibodies reacting to the basal ganglia have been described (Singer et al., 1998). With our sample sizes (Tourette = 24, controls = 25), we had 93% power (Borenstein et al., 2000) to detect a difference in means of the magnitude reported for antibody level against putamen as determined by ELISA (Singer et al., 1998). However, our study found no difference in antibody level against GFAP between the Tourette group and controls. Therefore, we concluded that antibody responses against GFAP are unlikely to play a pathogenic role in Tourette Syndrome in our sample group. Acknowledgments We wish to acknowledge Sarah E. Doyle, BS, M. Michal Peterson, BS and Nancy K. Burgess, BS, for excellent

123

340 technical assistance, Ikuo Tsunoda, MD, PhD, for many helpful discussions and Kathleen Borick for the outstanding preparation of the manuscript. This work was supported by NIH U19 D/DC35476, which is part of the NICHD/NIDCD Collaborative Programs for Excellence in Autism.

References American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed.). Text Revision. Washington, DC. Bailey, A., Le Couteur, A., Gottesman, I., Bolton, P., Simonoff, E., Yuzda, E., & Rutter, M. (1995). Autism as a strongly genetic disorder: Evidence from a British twin study. Psychology of Medicine, 25, 63–77. Borenstein, M., Rothstein, H., & Cohen, J. (2000). Power and Precision, version 2.0. www.PowerAndPrecision.com. Chakrabarti, S., & Fombonne, E. (2001). Pervasive developmental disorders in preschool children. Journal of the American Medical Association, 285, 3093–3099. Coleman, M., & Gillberg, C. (1985). The biology of the autistic syndromes. New York: Praeger Publishers. 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. Journal of Child Neurology, 14, 388–394. Connolly, A. M., Chez, M. G., Pestronk, A., Arnold, S. T., Mehta, S., & Deuel, R. K. (1999). Serum autoantibodies to brain in LandauKleffner variant, autism, and other neurologic disorders. Journal of Pediatrics, 134, 607–613. 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: A casecontrol study. Archives of Pediatrics, & Adolescent Medicine, 159, 151–157. Dahl, D., Crosby, C. J., Gardner, E. E., & Bignami, A. (1982). Purification of the glial fibrillary acidic protein by anionexchange chromatography. Analytical Biochemistry, 126, 165– 169. Eng, L. F., Ghirnikar, R. S., & Lee, Y. L. (2000). Glial fibrillary acidic protein: GFAP-thirty-one years (1969–2000). Neurochemical Research, 25, 1439–1451. Fombonne, E. (2003). The prevalence of autism. Journal of the American Medical Association, 289, 87–89. Fujinami, R. S., Zurbriggen, A., & Powell, H. C. (1988). Monoclonal antibody defines determinant between Theiler’s virus and lipidlike structures. Journal of Neuroimmunology, 20, 25–32. Gillberg, C., & Wing, L. (1999). Autism: Not an extremely rare disorder. Acta Psychiatra Scandinavica, 99, 399–406. Goldberg, W. A., Osann, K., Filipek, P. A., Laulhere, T., Jarvis, K., Modahl, C., Flodman, P., & Spence, M. A. (2003). Language and other regression: assessment and timing. Journal of Autism and Developmental Disorders, 33, 607–616. Hallett, J. J., Harling-Berg, C. J., Knopf, P. M., Stopa, E. G., & Kiessling, L. S. (2000). Anti-striatal antibodies in Tourette syndrome cause neuronal dysfunction. Journal of Neuroimmunology, 111, 195–202. Hickey, W. F. (2001). Basic principles of immunological surveillance of the normal central nervous system. Glia, 36, 118–124. Jankovic, J. (2001). Tourette’s syndrome. New England Journal of Medicine, 345, 1184–1192. Kaufman, J., Birmaher, B., Brent, D., Rao, U., Flynn, C., Moreci, P., Williamson, D., & Ryan, N. (1997). Schedule for affective disorders and schizophrenia for school-age children-present and

123

J Autism Dev Disord (2008) 38:333–341 lifetime version (K-SADS-PL): initial reliability and validity data. Journal of the American Academy of Child & Adolescent Psychiatry, 36, 980–988. Kiessling, L. S., Marcotte, A. C., & Culpepper, L. (1993). Antineuronal antibodies in movement disorders. Pediatrics, 92, 39–43. 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. Pediatric Neurology, 35, 303–307. Loiselle, C. R., Lee, O., Moran, T. H., & Singer, H. S. (2004). Striatal microinfusion of Tourette syndrome and PANDAS sera: failure to induce behavioral changes. Movement Disorders, 19, 390– 396. Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Jr., 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. Journal of Autism and Developmental Disorders, 30, 205–223. Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnostic interview-revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24, 659–685. Luyster, R., Richler, J., Risi, S., Hsu, W.-L., Dawson, G., Bernier, R., Dunn, M., Hepburn, S., Hyman, S. L., McMahon, W. M., Goudie-Nice, J., Minshew, N., Rogers, S., Sigman, M., Spence, M. A., Goldberg, W. A., Tager-Flusberg, H., Volkmar, F. R., & Lord, C. (2005). Early regression in social communication in autism spectrum disorders: A CPEA Study. Developmental Neuropsychology, 27, 311–336. Meinl, E., Krumbholz, M., & Hohlfeld, R. (2006). B lineage cells in the inflammatory central nervous system environment: migration, maintenance, local antibody production, and therapeutic modulation. Annals of Neurology, 59, 880–892. Nelson, K. B. (1991). Prenatal and perinatal factors in the etiology of autism. Pediatrics, 87, 761–766. Plioplys, A. V., Greaves, A., Kazemi, K., & Silverman, E. (1994). Immunoglobin reactivity in autism and Rett’s syndrome. Developmental Brain Dysfunction, 7, 12–16. Sambrook, J., Fritsch, E. F., & Maniatis, T. (1989). Molecular cloning: A laboratory manual. New York: Cold Spring Harbor Laboratory. Schotte, A., Janssen, P. F., Megens, A. A., & Leysen, J. E. (1993). Occupancy of central neurotransmitter receptors by risperidone, clozapine and haloperidol, measured ex vivo by quantitative autoradiography. Brain Research, 631, 191–202. Silva, S. C., Correia, C., Fesel, C., Barreto, M., Coutinho, A. M., Marques, C., Miguel, T. S., Ataide, A., Bento, C., Borges, L., Oliveira, G., & Vicente, A. M. (2004). Autoantibody repertoires to brain tissue in autism nuclear families. Journal of Neuroimmunology, 152, 176–182. Singer, H. S., Giuliano, J. D., Hansen, B. H., Hallett, J. J., Laurino, J. P., Benson, M., & Kiessling, L. S. (1998). Antibodies against human putamen in children with Tourette syndrome. Neurology, 50, 1618–1624. Singer, H. S., Mink, J. W., Loiselle, C. R., Burke, K. A., Ruchkina, I., Morshed, S., Parveen, S., Leckman, J. F., Hallett, J. J., & Lombroso, P. J. (2005). Microinfusion of antineuronal antibodies into rodent striatum: failure to differentiate between elevated and low titers. Journal of Neuroimmunology, 163, 8–14. Singer, H. S., Morris, C. M., Williams, P. N., Yoon, D. Y., Hong, J. J., & Zimmerman, A. W. (2006). Antibrain antibodies in children with autism and their unaffected siblings. Journal of Neuroimmunology, 178, 149–155.

J Autism Dev Disord (2008) 38:333–341 Singh, V. K., & Rivas, W. H. (2004). Prevalence of serum antibodies to caudate nucleus in autistic children. Neuroscience Letters, 355, 53–56. Singh, V. K., Warren, R., Averett, R., & Ghaziuddin, M. (1997). Circulating autoantibodies to neuronal and glial filament proteins in autism. Pediatric Neurology, 17, 88–90. 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 Behavior and Immunity, 7, 97–103. Swedo, S. E., Leonard, H. L., Garvey, M., Mittleman, B., Allen, A. J., Perlmutter, S., Lougee, L., Dow, S., Zamkoff, J., & Dubbert, B. K. (1998). Pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections: Clinical description of the first 50 cases. American Journal of Psychiatry, 155, 264–271. 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. Taylor, J. R., Morshed, S. A., Parveen, S., Mercadante, M. T., Scahill, L., Peterson, B. S., King, R. A., Leckman, J. F., & Lombroso, P. J. (2002). An animal model of Tourette’s syndrome. American Journal of Psychiatry, 159, 657–660. Todd, R. D., Hickok, J. M., Anderson, G. M., & Cohen, D. J. (1988). Antibrain antibodies in infantile autism. Biological Psychiatry, 23, 644–647. Tolley, N. D., Tsunoda, I., & Fujinami, R. S. (1999). DNA vaccination against Theiler’s murine encephalomyelitis virus leads to alterations in demyelinating disease. Journal of Virology, 73, 993–1000. Torres, A. R., Maciulis, A., Stubbs, E. G., Cutler, A., & Odell, D. (2002). The transmission disequilibrium test suggests that HLADR4 and DR13 are linked to autism spectrum disorder. Human Immunology, 63, 311–316.

341 Torres, A. R., Sweeten, T. L., Cutler, A., Bedke, B. J., Fillmore, M., Stubbs, E. G., & Odell, D. (2006). The association and linkage of the HLA-A2 class I allele with autism. Human Immunology, 67, 346–351. Vargas, D. L., Nascimbene, C., Krishnan, C., Zimmerman, A. W., & Pardo, C. A. (2005). Neuroglial activation and neuroinflammation in the brain of patients with autism. Annals of Neurology, 57, 67–81. Vojdani, A., Campbell, A. W., Anyanwu, E., Kashanian, A., Bock, K., & Vojdani, E. (2002). Antibodies to neuron-specific antigens in children with autism: possible cross-reaction with encephalitogenic proteins from milk, Chlamydia pneumoniae and Streptococcus group A. Journal of Neuroimmunology, 129, 168–177. Warren, R. P., Odell, J. D., Warren, W. L., Burger, R. A., Maciulis, A., Daniels, W. W., & Torres, A. R. (1996). Strong association of the third hypervariable region of HLA-DR b1 with autism. Journal of Neuroimmunology, 67, 97–102. Yamada, M., Zurbriggen, A., Oldstone, M. B. A., & Fujinami, R. S. (1991). Common immunologic determinant between human immunodeficiency virus type 1 gp41 and astrocytes. Journal of Virology, 65, 1370–1376. Yeargin-Allsopp, M., Rice, C., Karapurkar, T., Doernberg, N., Boyle, C., & Murphy, C. (2003). Prevalence of autism in a US metropolitan area. Journal of the American Medical Association, 289, 49–55. Yeh, C. -B., Wu, C. -H., Tsung, H. -C., Chen, C. -W., Shyu, J. -F., & Leckman, J. F. (2006). Antineural antibody in patients with Tourette’s syndrome and their family members. Journal of Biomedical Science, 13, 101–112. 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 Behavior and Immunity, 21, 351–357.

123