Maternal Antibodies Linked to Autism

Some children with autism are born to mothers carrying antibodies that bind to proteins involved in brain development.

By Ed Yong | July 9, 2013

WIKIPEDIA, KEN HAMMOND (USDA)In 2008, Judy van de Water from the University of California, Davis, discovered a group of autoantibodies—those that trigger immune responses against the body’s own molecules—that are especially common in mothers of children with autism. Now, her team has identified what these antibodies bind to: six proteins involved in varied aspects of brain development. By crossing the placenta and affecting these proteins in a fetus’s brain, the maternal antibodies could increase the risk of developmental problems in some cases of autism, according to the new research, published today (July 9) in Translational Psychiatry.

“I cannot laud these authors enough,” said Andrew Zimmerman, a neurologist from the Kennedy Krieger Institute, who has also been studying maternal antibodies but was not involved in this study. “Given that, at present, only between 15 and 20 percent of children with autism have known causes—mainly genetic and infectious mechanisms—this will be a major advance.”

Van de Water’s team, led by graduate student Dan Braunschweig, is now using their discovery to develop a test that predicts a child’s risk of developing autism spectrum disorders based on the mother’s antibodies. “It would allow mothers to plan,” said van de Water, by enrolling their children in educational programs that promote social skills from an early age.

The antibody hypothesis would only apply to a quarter of autism cases at most, but van de Water said that it is valuable for affected parents to get some clues about the biology behind their children’s condition. “It provides some answers,” she said. “They couldn’t have done anything about this—it’s not like they did anything to cause the antibodies. But as a parent, you just want to know what happened so you can move forward.”

The proteins that the team identified have a wide variety of roles. STIP1 influences the creation of new neurons, for example, while cypin affects the number of branches they have. CRMP1 and CRMP2 stop neurons from growing and determine their length. YBX1 is involved in gene transcription, as well as neural migration during development. Finally, LDH is the most mysterious of the sextet but is also the most strongly linked to autism. Earlier studies suggest that it may play a role in metabolism or in responses to viruses or toxins.

All six are highly expressed in the fetal brain. Of 246 mothers with children living with autism, 23 percent had antibodies that recognized two or more of these proteins, compared to just 1 percent of 149 mothers with normally developing children. The antibodies have more than 99 percent specificity for autism risk, which means that they have less than a 1 in a 100 chance of finding a false positive.

Meanwhile, the team’s colleagues Melissa Bauman and David Amaral, also from UC Davis, injected eight pregnant rhesus monkeys with antibodies purified from mothers with autistic children. These monkeys were more protective towards their young during their first 6 months, compared to those that were injected with antibodies from women with neuro-typical children. As the young monkeys grew up, they showed unusual social behavior: compared to typical macaques, they were more likely to approach both familiar peers and strangers, even when their advances weren’t rewarded with sustained social interactions.

 “Moving this to monkeys is a big step,” said Paul Patterson, a neuroimmunologist from the California Institute of Technology, who was not involved in the work. “This very careful behavioral study shows that at least some of the antibodies do have an effect on fetal brain development.”

Betty Diamond, an immunologist at the Feinstein Institute for Medical Research, agrees the studies represent “an important step forward.” However, she noted that antibodies often bind to many possible targets, and the proteins that the team identified may not be the relevant ones. She also said that some of the alleged target proteins are found within cells, “and it is not clear how or whether the antibodies can penetrate developing neurons.”

Zimmerman added, “Much work remains to be done to show how these antibodies are relevant, how they affect fetal brain development, and what factors lead some mothers to develop these antibodies.”

The team is now working to address these issues, trying to identify the specific parts of the six proteins that the antibodies stick to, determine how they affect the developing brain, and understand how they might be used to predict autism risk. Van de Water and Amaral are consulting for Pediatric Bioscience, which is creating a predictive test based on the results.

“The next step is to come up with a therapeutic to block the antibodies—not just to pick them up, but to do something about it,” said van der Water. Although the concept of preventing autism can be controversial, she points out that her panel of antibodies seem to correlate with the most severe symptoms and language problems.

Still, she is treading cautiously. “The parents have been surprisingly supportive,” she said. “But the autism field has been fraught with false alarms, so we want to be really careful.”

D. Braunschweig et al., “Autism-specific maternal autoantibodies recognize critical proteins in developing brain,” Translational Psychiatry, 3:e277, 2013.

M.D. Bauman et al., “Maternal antibodies from mothers of children with autism alter brain growth and social behaviour development in the rhesus monkey,” Translational Psychiatry, 3:e278, 2013.

Clarification (July 10): This story has been updated from its original version, which included this quote in relation to a potential test: "If it’s positive, their risk is virtually 100 percent". With a 99 percent specificity for autism risk, such a test would still return false positives for 1 percent of the non-autistic population.

Add a Comment

Avatar of: You



Sign In with your LabX Media Group Passport to leave a comment

Not a member? Register Now!

LabX Media Group Passport Logo


Avatar of: Baxter Zappa

Baxter Zappa

Posts: 11

July 10, 2013

This is fascinating! Clearly autism had nothing to do with vaccinating the children, as we already knew. But, if autism has increased, and this is partly due to more mothers having these autoantibodies, what are some likely causes of *that*?

I'd like to see what medical history / exposures / genetics the autoimmune mothers have in common.

Avatar of: Jon Brock

Jon Brock

Posts: 1

July 10, 2013


Great write up, Ed, and a fascinating pair of studies. I'd just like to point out an issue that I think the authors get wrong in your article.

A negative result on this test would not rule out a risk, “but if it’s positive, their risk is virtually 100 percent,” said van de Water. “It would allow mothers to plan,” she added, by enrolling their children in educational programs that promote social skills from an early age.

This mistakes the specificity of the test (99%) for the positive predictive value, which is, by my back-of-the-envelope calculations, around 19%, based on the figures quoted. That is, if a mum tested positive, she'd have a 19% risk of having a kid with autism, not a "virtually 100 percent" risk.

Say you had 10,000 mums drawn from the general population.

Based on current estimates, you'd expect roughly 1% (100) of these mums to have kids to have autism (ignoring for now the fact that some mums have more than one kid with autism).  The study suggests that 23% of mothers of kids with autism have a "risky" combination of antibodies, so we'd expect roughly 23 of these mums to test positive.

The specificity is 99%. So of the remaining 9,900 mums in our sample, we'd expect 1% (99) to test positive, in this case falsely because their child is not autistic.

Altogether, we'd expect 23+99=122 mums to test positive. Of these, only 23 (19%) will actually have an autistic child.

This 19% figure is a very rough estimate. It's clearly much bigger than 1% (the population risk for autism) so the antibodies do seem to play a role. But it's also a lot less than 100%.

This would obviously have major implications for counselling mums who test positive. The quote is wrong. Chances are, they won't have an autistic kid (although they should perhaps be extra vigilant for other signs).

It also impacts on how we think the antibodies (or something associated with the antibodies) might lead to autism. If "virtually 100%" of mums who test positive go on to have an autistic child, this would indicate that the antibodies can cause autism all by themselves. However, if only a relatively small proportion of mums who test positive end up having an autistic child, we need to look at how the antibodies interact with other risk factors to cause autism in some cases but not in others.


Avatar of: Frederik Aagaard

Frederik Aagaard

Posts: 1

July 10, 2013


I find this to be a very interesting article, but two questions still lie in my mind:

1) Would it be possible, without checking the mothers, to discern the autistic children with mothers with the antibodies from the autistic children with mothers without the antibodies? If that is the case then we will know that autism is indeed an umbrella for several individual diagnoses stemming from individual causes (we might already know that - I'm not certain).

2) What is the ratio between boys and girls within the group of autistic children with mothers with the antibodies? As far as I am aware more boys are diagnosed with autism than girls. There seams to be general disagreement on the internet as to the actual ratio, but they generally lie between 3:1 and 5:1. In this though experiment we will assume the that the ratio of 4:1 is true, which means that 20 % of all autistic children are girls.

If the ratio between autistic boys and girls with mothers with the antibodies is 1:1, then we must conclude that 11,5 % of all autistic children are female and have mothers with the antibodies. That leaves 8,5 % autistic girls within the group of 77 % percent autistic children with mothers without the antibodies, which means that only 11,04 % of autistic children with mothers without the antibodies are girls.

If the ratio between autistic boys and girls with mothers with the antibodies is not 1:1, then we must conclude that either the antibodies or the whatever causes their presence is swifting the balance of the sexes which, at least for my part, is unheard of.

I might be asking stupid questions or making illogical conclusions, though I certainly hope not; I'm simply a boy of 16 years sitting at my desk in Danmark being generally curious.

P.S. I have throughout my comment have assumed that the article is correct when it states that mothers with the antibodies have a 99 % chance of giving birth to autistic children, rather than Jon Brocks 19 %.

P.P.S. When checking my facts a second time, I now discover that recent studies have indicated that rather than more boys having autism than girls we are simply better at diagnosing them. If this the case, then my second question is irrelevant.

Avatar of: Michael Frost

Michael Frost

Posts: 1

July 11, 2013

We seem to be getting closer to some real answers to autism. The immune system has been the focus of a number of researchers for a few years now and this seems to tie some of it together.

Paul Patterson (see above) has been doing some great animal model research and prenatal viral infection and it is going to be interesting whether this complements.

The gut, allergy, autoimmune / autoinflammatory disease do they all come together or is this just one part of the autisms ?

Congratulations to all the researchers, mothers and children (all the staff at MIND) these types of breakthroughs only happen through long long hours, patience and fortitiude.

Avatar of: spoxlady


Posts: 2

July 11, 2013

So are women with autoimmune disorders at greater risk for this?

Avatar of: spoxlady


Posts: 2

July 11, 2013

So are women with autoimmune disorders at greater risk for this?

Avatar of: Samlyn


Posts: 1

June 11, 2014

This is very interesting.  My first pregnancy was normal and yielded a healthy, developmentally "normal" child.  I began to feel ill with a series of symptoms and at ten months postpartum, I was diagnosed with lupus.  I had been  completely healthy until this, not even so much as an allergy.  The diagnosing rheumatologist told me that if I wanted to  have another child to do so right away.  I got pregnant immediately.  This pregnancy wasn't so easy.  Lots of fetal monitoring.  Everything looked good.  Went into anaphylactic reaction to IV ampicillin during my induction. I was close to being intubated. Then proceeded to have a difficult birth to my beautiful son.  He was diagnosed with autism on his second birthday.  I have continued to suffer with my autoimmune diseases, recently adding rheumatoid arthritis to the list. Extreme fatigue has recently sent me in to see an endocrinologist who's lab orders have  shown that I have antithyroid antibodies.  To my knowledge I have never been tested for them before.  So I am likely developing Hoshimotos. None of these autoimmune diseases were present with the pregnancy of my nonautistic child.  I probably did have the antibodies during the pregnancy of my autistic child.   Also of note, my antithyroid level was 64 

(end normal range is 60) and my son is only "mildly autistic". His actual diagnosis was PDD NOS which was converted to ASD  with the DSM manual change two weeks later. 

Popular Now

  1. How to Separate the Science From the (Jerk) Scientist
  2. RNA Moves a Memory From One Snail to Another
  3. Could a Dose of Sunshine Make You Smarter?
  4. Sweden Cancels Agreement With Elsevier Over Open Access