News The latest developments in autism research.

Brain imaging study points to microglia as autism biomarker

by  /  10 January 2013

Rainbow bright: The brain of an individual with autism (bottom) shows more activated microglia than a control brain does (top).

Microglia, brain cells that are part of the immune system, are more activated in young men with autism than in controls, according to a brain imaging study published 26 November in the Archives of General Psychiatry1.

Postmortem studies have shown that microglia are altered in autism, but the new study marks the first time that researchers have tracked the cells in living people with the disorder.

Microglia are known to rapidly transform from a spider-shaped resting state into a bulbous active state when they’re fighting off infection or damage. But the role of activated microglia in autism and related disorders is complex and largely mysterious. For example, studies in the past couple of years have shown that active microglia are important not only for immunity, but for the development of a healthy brain.

It’s not yet clear what the findings from the new study mean. But experts say that tracking microglia activity in live brains could be used as a biomarker of how the brain changes over time, such as before and after a new treatment.

“Doing this in vivo characterization gives a more dynamic perspective of what these cells are doing in the brain,” notes Carlos Pardo, associate professor of neurology and neuropathology at Johns Hopkins University in Baltimore, who was not involved in the work.

In the new study, researchers from Japan used a type of brain scanning called positron emission tomography (PET) to track the distribution of microglia in the brain. Their scans point to an abundance of activated microglia in men with autism, particularly in the cerebellum — a region known to process sensory information, movement and learning, and which has also been linked to autism.

The data echo what’s been seen in postmortem tissue. For example, in a landmark study in 2005, Pardo’s team showed that cerebellar tissue from individuals with autism shows an excess of activated microglia and contains certain chemicals, called cytokines, that are involved in inflammatory responses2.

Tricky tracer:

In PET scanning, radioactive tracers injected into participants’ blood make their way into the brain, bind to specific receptors, and can then be seen with a brain scanner.

In this case, the researchers used a tracer dubbed [11C](R)-PK11195, which binds to activated microglia. The tracer has been used for decades in people with various neurological disorders. It had not been used in people with autism, however, because the link between microglia and the disorder is relatively new.

The researchers injected the tracer into 20 men with high-functioning autism and 20 controls matched by sex, age and intelligence quotient. They found that the tracer binds more strongly in the brains of the participants with autism, presumably because they have more activated microglia.

This was true across all brain regions tested, says lead investigator Kazuhiko Nakamura, associate professor of psychiatry and neurology at Hamamatsu University School of Medicine in Shizuoka, Japan. “But the most prominent increase was evident in the cerebellum.”

No one knows what activated microglia may be doing in the brains of people with autism, some experts note. Others are concerned by some of the details of the study’s design.

For example, the tracer isn’t specific to activated microglia. It can also bind other brain cells, including astrocytes, star-shaped cells whose long projections help support synapses, the junctions between neurons.

Less frequently, the tracer also binds resting microglia and neurons, notes Robert Innis, chief of the molecular imaging branch of the National Institute of Mental Health, who was not involved in the new study.

Innis is using tracers that are more specific to activated microglia in people with autism. Nakamura says his team, too, is developing more specific tracers.

The new study’s design is also not ideal, according to Innis, because the researchers didn’t simultaneously look at the tracer in both the blood and brain. This comparison can rule out the possibility that people with autism simply process the tracer differently.

“Maybe the patients are just not metabolizing this drug as fast, and therefore more of it is getting into the brain,” Innis says.

Nakamura agrees that this is a possibility, but argues that metabolic abnormalities are unlikely because the scans indicated that there were no differences in blood flow in the brains of people with autism and controls.

Because it requires exposure to radiation, PET scanning is usually not done on children, especially for studies that require a healthy control group.

But newer microglia markers that require less radiation than conventional markers could conceivably be used to track the effectiveness of a therapy. “I could see it being extended into research studies of anti-inflammatory therapies, where the child acts as their own control,” Innis says.

Methodological issues aside, other researchers question the meaning of the data. Several big questions remain unanswered, such as whether activated microglia are a cause or consequence of autism, notes Jonathan Kipnis, professor of neuroscience at the University of Virginia in Charlottesville.

Last year, Kipnis’ team reported that replacing the microglia in mice that model Rett syndrome alleviates some symptoms and extends the mice’s lives. Intriguingly, these mutant animals seem to have under-active microglia, he says, illustrating that the cells’ role in neurodevelopmental disorders is not straightforward.

“Just looking at the number of microglia, and whether they’re activated or not — I’m not saying it’s meaningless,” Kipnis says, “but it’s not meaningful enough.”

  1. Suzuki K. et al. JAMA Psychiatry Epub ahead of print (2012) Abstract
  2. Vargas D.L. et al. Ann. Neurol. 57, 67-81 (2005) PubMed

8 responses to “Brain imaging study points to microglia as autism biomarker”

  1. Matt Carey says:

    The basic idea of this study is excellent–a method to investigate in-vivo the microglia. The discussion above does indicate that there is at least another step forward to go to make this a useful tool. I read this paper when it came out and was impressed, but I did wonder if the tracer was as specific to microglia as claimed.

    Vargas (2005) is one of the stand out papers of the past decade in autism research. It is very frustrating that 8 years later we can’t answer some basic questions.

    For one–are the microglia themselves different in autistic brains, or are they responding to a different stimulus? The fact that the larger number of activated microglia are not uniformly distributed indicates to this lay person that they are responding to a different stimulus.

    Second, again from a lay person’s view, what are the microglia doing? If they are pruning synapses, then either old autistics should have much fewer in these regions, or there is a process whereby they are creating new synapses faster.

    Third–what does this tell us, if anything, about treatments? Some practitioners jumped on Vargas (2005) to promote untested treatments on autistics. Even though the Hopkins FAQ says there is no indication for such treatments, prednisone, actos and other drugs have been and continue to be used. I saw a spike of discussion of anti-inflammatory discussions on autism parent online groups with the publication of this study. If this is a pathway to treatment, we should know and move forward. If reducing inflammation is bad, we should know and stop the off-label treatments that are ongoing.

    • passionlessDrone says:

      Hi Matt Carey –

      For one–are the microglia themselves different in autistic brains, or are they responding to a different stimulus? The fact that the larger number of activated microglia are not uniformly distributed indicates to this lay person that they are responding to a different stimulus.

      The morphological changes, which I’d think constitute different, are a good place to start answering that question, so from that perspective I’d say they are different. But there might be other ways too.

      There is another possibility that you may want to consider; perhaps the spatial distributions you refer to are the consequence of an altered developmental colonization pattern of the microglia into the brain?

      Animal studies are exposing a very delicate, temporally and spatially sensitive migration of microglia through the developing brain, a process that starts in the womb and continues in infancy. There is also experimental evidence to believe that disturbances in this process, in some cases by risk factors known to be associated with autism, can cause long term alterations in the pattern and morphological characteristics of the adult animal.

      A recent review, Microglia in the developing brain: A potential target with lifetime effects speaks towards this possiblity.

      The slow turn-over rate for mature microglia raises an issue related to changes that may occur in this critical neural cell population. While this has not been a primary issue of investigation there is limited data suggesting that microglia maintain a history of previous events. Thus, if this history alters the appropriate functioning of microglia then the effects could be long lasting. Additionally, a simple change in the number of microglia colonizing the brain during development, either too many or too few, could have a significant impact not on only the establishment of the nervous system network but also on critical cell specific processes later in life.

      Second, again from a lay person’s view, what are the microglia doing?

      They may be stuck in a something of a rut. For a very recent example, Neonatal lipopolysaccharide exposure induces long-lasting learning impairment, less anxiety-like response and hippocampal injury in adult rats came out this week, and researchers found that a *very direct* induction of the CNS immune system during the neonatal period and reported Neonatal LPS exposure also resulted in sustained inflammatory responses in the P71 rat hippocampus, as indicated by an increased number of activated microglia and elevation of interleukin-1ß content in the rat hippocampus. ,

      There was no pathogen and no further insult, but the effect on microglial morphology and the neuroimmune evnironment was persistent. There are many similar studies, including some on seizures, but they seem to implicate the early stages of life as critical timeframes for setting the new baseline.

      For an indirect, but interesting look at this possibility in the autism population, Immune transcriptome alterations in the temporal cortex of subjects with autism found that the genetic expression was consistent with an impaired regulatory ability. The deregulation of these gene pathways might indicate that the profound molecular differences observed in the temporal cortex of autistic subjects possibly originate from an inability to attenuate a cytokine activation signal..

      If they are pruning synapses, then either old autistics should have much fewer in these regions, or there is a process whereby they are creating new synapses faster.

      Well, some of the pruning studied seem to indicate that microglial participation in pruning is happening during a state where they are not ‘activated’; i.e., they aren’t in the different morphological shape. I’m not sure that we should expect to see ongoing pruning differences past infancy, at least not at the scale that is supposed to be happening early on. It seems likely (to me?) that the synaptic maintenance operation has lots of players, and if the other right signals aren’t in place, it may be an opportunity lost for good.

      Third–what does this tell us, if anything, about treatments?

      I think that, if anything, potential treatments could be a mixed bag for some kids. There is evidence of differential microglial participation at the synapse at sensory management time, something I’d bet a lot of people with autism would like to have understood very keenly, yet, there may be no mechanism to attenuate the effects of the lack of neural optimization during the gestastional and neonatal period.

      I do agree with much of your sentiments.

      – pD

    • Katie Wright says:

      I agree with you. The pace of this research is inexcusable.

  2. Sarah says:

    Mice deficient in fractalkine are less susceptible to cerebral ischemia-reperfusion injury

    Sulpicio G. Soriano, a, , Lakshmi S. Amaravadib, Yanming F. Wanga, Hong Zhoub, Gary X. Yub, James R. Tonrab, Victoria Fairchild-Huntressb, Qing Fangb, Judy H. Dunmoreb, Dennis Huszarb, Yang Panb, 1
    a Department of Anesthesia, Children’s Hospital, Boston, and Harvard Medical School, 300 Longwood Avenue, Boston, MA 02115, USA
    b Millennium Pharmaceutical Incorporated, 640 Memorial Drive, Cambridge, MA 02114, USA
    Corresponding author. Tel.: +1-617-355-6457; fax: +1-617-355-7887
    Received 29 October 2001Revised 8 January 2002Accepted 23 January 2002Available online 14 February 2002


    Fractalkine (FKN), also known as neurotactin, is a CX3C chemokine that exists in both secreted and neuronal membrane-bound forms and is upregulated during brain inflammation. There is accumulating evidence that FKN induces chemotaxis by binding to its receptor CX3CR1 on leukocytes and microglia. We generated FKN-deficient mice to study the role of FKN in postischemic brain injury. After transient focal cerebral ischemia, FKN-deficient mice had a 28% reduction in infarction size and lower mortality rate, when compared to wild-type littermates. The findings of this study indicate a possible role for FKN in augmenting postischemic injury and mortality after transient focal cerebral ischemia.

    Adhesion molecules; Cerebral ischemia; Fractalkine; Stroke


    • passionlessDrone says:

      Hi Sarah –

      There is also a fractalkine study modeling Alzheimer’s where knock out mice showed decreased synapse loss, another way of looking at the question.

      Microglial Cx3cr1 knockout prevents neuron loss in a mouse model of Alzheimer’s disease

      • Sarah says:

        I wonder if they could inhibit fracktalkine in children with autism, whether the behaviors would improve. Not sure what the mechanism would be but hope they research fractalkine further as something to target.

  3. Sarah says:

    Here is another excellent article on microglia…

    “The brain treats unwanted synapses like invading microbes – Housekeeping cells called microglia engulf unwanted neuronal connections in the developing brain” – The Guardian, August 2011

    Here an excerpt:

    “Next, the researchers created a strain of mutant mice lacking the gene encoding the fractalkine receptor. Fractalkine is a small signalling molecule which is known to play many roles in the immune system. Neurons in the brain ramp up fractalkine production when they are forming synapses; the fractalkine receptor is synthesized exclusively by microglia and is essential for their survival and migration.

    The mutants had significantly greater numbers of synapses at two and three weeks of age than their normal litter-mates, leading to an increase in the frequency of spontaneous electrical impulses. Consequently, long-term depression, a form of synaptic plasticity in which synaptic connections are weakened, was enhanced, but this effect was reversible and had disappeared by the time the mice were 40 days old. The three-week-old mice were also less susceptible to drug-induced epileptic seizures but this, too, was not seen in the adults.

    This suggests that synapse pruning in the mutants’ brains was diminished, causing the connections to mature more slowly than they normally would, leading to a delay in the development of brain circuitry.

    Thus, the developing brain treats unwanted synapses as if they were unwanted invaders. It dispatches microglial cells to survey the state of synapses in their surroundings and to dispose of the ones that are wired incorrectly or superfluous. Abnormal neural connectivity has been implicated in neurodevelopmental disorders such as autism, so deficiencies in microglial surveillance may contribute to such conditions.

    “We are very interested to see if there are long-lasting behavioural changes due to the deficient pruning in the mutant mice,” says senior author Cornelius Gross. “In particular, we are interested to see if autism-like behavioural phenotypes are revealed in these mutants. The second question is: What is the local ‘eat me’ signal from synapses that indicate they are ready to be pruned? We think it might be fractalkine [and] we also suspect that the complement cascade proteins are involved.”

    The making and breaking of synapses also occurs throughout life, and is essential for learning and memory. It is, therefore, intriguing to speculate that microglia could also be involved in these processes. “It is certainly possible,” Gross says, “but so far, there is no evidence that microglia are required for the loss of synapses under these circumstances. Synapse loss and gain can be seen in cultured cells, so microglia are not absolutely essential.”

  4. ASD Dad says:

    The work above thought provoking and very positive research that leads us down one of the most exciting areas of autism research that has the most promise for a real treatment regimes.

    The immune system and how it interacts in autism is the ‘go to’ area in not just neurology but physiology, genetics and treatment.

Leave a Reply

Your email address will not be published. Required fields are marked *