Childhood Serum Anti-Fetal Brain Antibodies Do Not Predict Autism

18 Sep

One of the big new concepts in autism research in the last two years is the idea that maternal antibodies might be involved in developing autism.

Two groups, one from Johns Hopkins, the other from California showed that in some mothers with autistic children, their blood sera had antibodies against fetal brain tissue.

This raised the question of whether maternal antibodies, transferred during pregnancy, could influence the risk of autism.

One big question raised was the possibility that this could lead to regressive autism. Could something prenatal be linked to regressions in children who were age 1 or 2?

One big question raised (and unanswered) by these studies was the question of antibodies in the autistic children themselves. Do they have antibodies against fetal brain tissue? Could this be a biomarker? And, if so, could this be involved in developing autism in some cases?

The Hopkins team has just published a paper looking into this question.

Autoimmune hypotheses for autism include in utero transplacental exposure to maternal antibodies and acquired postnatal insults. Previous work demonstrated that some mothers of children with autistic disorder have specific antibodies against human fetal brain that differentiate them from mothers with typical children. In the present study, Western immunoblotting was used to determine whether children with autistic spectrum disorders (n = 29) have serum reactivity against human fetal brain that differs from that of controls (n = 14). There was no significant difference in reactivity, corrected for serum immunoglobulin G content and brain actin content and with special attention to reactive bands at 36, 39, 61, and 73 kDa, between autistic children and normal control subjects. Thus, in contrast to mothers, antibody reactivity against human fetal brain as measured in children ages 3-12 years does not appear to be a useful biomarker for autism.

The paper is fairly brief as they don’t find any evidence that the children’s blood sera reacted with fetal brain tissues. Here is a quote from the discussion section:

The present data indicate that the measurement of serum antibody reactivity against human fetal brain in children with autistic spectrum disorders does not predict an autistic diagnosis, either for idiopathic autistic disorder or for those individuals with identified etiologies. These data differ from previous findings in mothers of autistic children, which indicated that anti-human fetal brain antibodies reactive against proteins at 36, 39, 61, or 73 kDa appear to be a biological marker for the disorder in some individuals

Before anyone jumps on me for denying the possibility of any immunological effects in risk of autism, here is another quote from the paper:

The present study does not rule out the possibility that other anti-brain antibodies in the children may be important for postnatal effects. For example, children with autism have been shown to possess antibodies against various central nervous system self-components such as glial fibrillary acidic protein, myelin basic protein, neurofilament proteins, cerebellar neurons, and brain endothelial cell proteins [6-9,23,24]. We hypothesize that, rather than a direct association, there is a complex relationship between maternal anti-fetal brain antibodies and various intrauterine genetic, metabolic, and environmental factors.

Papers that don’t show a connection, show a “null” or lack of an effect, are somewhat unsatisfying. They are certainly not unimportant. I look forward to more research from Dr. Zimmerman’s group at Johns Hopkins,

11 Responses to “Childhood Serum Anti-Fetal Brain Antibodies Do Not Predict Autism”

  1. RAJ September 19, 2009 at 00:55 #

    The Johns Hopkins group was hoping that the study might identify a biomedical marker present in some mothers that would be present in the children and could predict a later high risk for a diagnosis of autism later in development. The study has to be disappointing.

    There is a multi center project involving many of the leading research groups devoted to autism research called the High Risk Baby Sib Project. Millions of dollars have been allocated to these rsearch centers by Autism Speaks and the NIH.

    The purpose is quite worthy, finding biomedical markers or behaviors and social interaction problems in high risk babies would yield a ‘test’ that could be universally applied to all newborns that would be predictive of a possible later diagnosis of autism. Very early identification would lead to earlier interventions and better outcomes if such a test could be proven to be reliable.

    An even more disappointing study was just published. The authors looked at scores of ‘eye’ gaze and early social behaviors that might identify a group of six months old children at high risk for autism especially given that the infants all had older siblings diagnosed with an ASD.

    The results were puzzling to the authors as the babies, first tested at six months of age and re-eximened a number of times up to 24 months of age produced results completly opposite of what they had hoped to find.

    The study looked at eye gaze and eye aversion as well as the babies social interactions with their mothers. None of the children who scored lower on eye gaze measures or social interation measures showed any signs of autism at the 24th month followup.

    None of the three infants who were diagnosed with autism at 24 months showed any evidence of gaze aversion or impairments in social interactions with mothers at six months of age.

    http://www.ncbi.nlm.nih.gov/pubmed/19702771?

  2. MJ September 19, 2009 at 01:22 #

    I think you are confusing what the the study is saying.

    The California group looked at antibodies in the mother’s blood during pregnancy and found that, in some cases, the mothers had circulating antibodies that would react to fetal brain proteins. The theory was that these antibodies would cross the placenta and damage the fetus while it was still in the womb thus setting the stage for autism.

    The John Hopkins group was looking at the children three to twelve years after they were born to see if there were any similar antibodies still present in the children and found none.

    So the John Hopkins work does not in any way say that the first finding was somehow incorrect. It was only an attempt to extrapolate on the earlier work to see if it could be used as a biomarker for autism.

    • Sullivan September 19, 2009 at 03:36 #

      So the John Hopkins work does not in any way say that the first finding was somehow incorrect. It was only an attempt to extrapolate on the earlier work to see if it could be used as a biomarker for autism.

      Did anyone say this study shows the maternal antibodies studies were somehow incorrect?

  3. Joseph September 19, 2009 at 02:16 #

    Very early identification would lead to earlier interventions and better outcomes if such a test could be proven to be reliable.

    Which makes sense, intuitively, but no one has ever produced convincing evidence that really early interventions lead to better outcomes (much less better adult outcomes.)

    If you think about it, it’s amazing that no one has ever demonstrated the existence of a method to prevent autism in the siblings of autistic children.

  4. MJ September 19, 2009 at 04:02 #

    “Which makes sense, intuitively, but no one has ever produced convincing evidence that really early interventions lead to better outcomes (much less better adult outcomes.)”

    The funny thing about convincing evidence is that you have to be willing to consider it. If you don’t believe that early interventions lead to positive results you haven’t been paying attention.

    “Did anyone say this study shows the maternal antibodies studies were somehow incorrect?”

    Then what exactly were you trying to say?

  5. dr treg September 19, 2009 at 17:32 #

    Perhaps CSF brain antibody measurement would have a yield but elective lumbar puncture is unethical.
    n.b. CSF TNF may be related to the severity of autism.
    http://linkinghub.elsevier.com/retrieve/pii/S0887899407000628

  6. passionlessDrone September 21, 2009 at 13:33 #

    Hello friends –

    A short, but interesting study. It may be that what we are observing here is that having antigens at these specific molecular weights is not as important so much as when you are exposed to these antigens. In other words, developing antigens to fetal brain tissue once you are a child isn’t the same as being exposed to them in utereo. Also, it is very likely that the mothers in the original studies didn’t have antibodies that were originally directed towards fetal brain tissue so much as antigens directed towards other, actual pathogens whose epitope was sufficiently similar to proteins expressed in fetal brain tissue that the immune system became confused; and attacked brain tissue thinking it was attacking a pathogen.

    For example, some strains of H1N1 appear to be able to elicit generation of antibodies to brain tissue.

    http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=286833

    Immunization of rabbits with certain H1N1 influenza viruses, including the neurotropic strains NWS/33 and WSN/33 and the New Jersey/76 strain, resulted in the production of autoantibodies to a brain-specific protein of 37 kDa that is present in various species, including humans. Autoantibodies were produced to brain only; various other tissues tested were negative. These antibodies were not elicited by other influenza A or B viruses, including closely related recombinant strains, but were elicited by the isolated hemagglutinin of A/Bellamy/42 strain and by formaldehyde-fixed WSN virus–demonstrating that infection was not essential for the induction of autoantibodies. In histological studies, reaction with anti-viral antisera was specific to gray matter and was confined to sera that recognized the 37-kDa protein. Antibody binding was prominent in regions comprised of neuronal cell bodies in cellular layers of the dentate gyrus, hippocampus, cerebral cortex, and cerebellum and was undetectable in myelin-rich regions, such as the corpus callosum. The 37-kDa protein, therefore, appears to be a neuronal antigen. Antibodies directed against this protein may be involved in the pathogenesis of one or more of the neuropsychiatric disorders that occur after infection with influenza.

    Of particular interest to autism, one study not mentioned by this paper found antigens in the 37kda range.

    http://cat.inist.fr/?aModele=afficheN&cpsidt=20200923

    We observed reactivity to two protein bands at approximately 73 and 37 kDa in plasma from 7 of 61 (11.5%) mothers of children with autism (AU) against fetal but not adult brain, which was not noted in either control group (TD; 0/62 p = 0.0061 and DD; 0/40 p = 0.0401).

    It looks like the Zimmerman paper looked for bands at 36kda, but not 37kda. I’m not sure if this is a precision issue, or they just didn’t study for it, or what.

    On a related note, an agency in California is now offering antibody tests to women to determine if they have some of the known antigens as a guidepost towards an increased risk of autism in children. It seems likely this was started by the MIND guys. The test is not available to women who are already pregnant.

    – pD

  7. Joseph September 21, 2009 at 14:55 #

    The funny thing about convincing evidence is that you have to be willing to consider it. If you don’t believe that early interventions lead to positive results you haven’t been paying attention.

    @MJ: I don’t believe there’s any reason to suppose I’m not willing to consider convincing evidence, so try me. Convincing evidence, BTW, is not one study; it’s a body of widely replicated results with methodologies of increasing quality that address the methodological problems of prior work.

    I also don’t believe that trials of interventions with younger children demonstrate better outcomes than those with older children.

    Additionally, there are observational studies that follow children first diagnosed when they were very young (age 2). Obviously, these studies look at variables that might predict outcome. Rarely do they report intervention as a variable predictive of outcome, even though age might be. (Actually, age appears to be predictive of diagnostic non-stability.)

  8. RAJ September 21, 2009 at 15:13 #

    “If you think about it, it’s amazing that no one has ever demonstrated the existence of a method to prevent autism in the siblings of autistic children”

    1. The development of an effective Rubella Vaccine.

    2. The removal of Thalidomide as a treatment for morning sickness in pregnant women.

    3. All newborns in developed countries are tested for the presence of Phenylketonuria. A positive test results in implementation of a phenylaline-free diet.

    All these implementations are associated with environmental insults associated with a significantly increased risk for ASD and prevention of autism and other develomental disorders.

  9. Joseph September 21, 2009 at 16:12 #

    All these implementations are associated with environmental insults associated with a significantly increased risk for ASD and prevention of autism and other develomental disorders.

    That’s not what I was referring to, and in any case, I doubt these measures have reduced the incidence of autism by even 1%.

    What I was referring to is that, if early intervention is good, super-early intervention must be super-good. Surely, if the siblings of autistic children are put through super-early intervention (starting when they are babies), familial autism should be largely preventable in theory. In reality, I don’t think it is.

    If early intensive intervention is effective and harmless for autistic children, I see no reason why it wouldn’t be effective and harmless for children in general, so there should be no problem (ethically) in having any and all siblings of autistic children be subjected to it.

  10. Cheryl Howard November 9, 2009 at 13:13 #

    Although autism is the result of a neurological abnormality, the cause of these fetal brain damage with the nervous system is unknown in most cases. Research findings indicate a strong genetic component. Most likely, environmental, immunologic, and metabolic factors also influence the development of the disorder.

Leave a Reply to Sullivan Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: