Early cerebellum injury hinders neural development, possible root of autism, theory suggests

A brain region largely known for coordinating motor control has a largely overlooked role in childhood development that could reveal information crucial to understanding the onset of autism, according to Princeton University researchers.

The cerebellum — an area located in the lower rear of the brain — is known to process external and internal information such as sensory cues that influence the development of other brain regions, the researchers report in the journal Neuron. Based on a review of existing research, the researchers offer a new theory that an injury to the cerebellum during early life potentially disrupts this process and leads to what they call “developmental diaschisis,” which is when a loss of function in one part of the brain leads to problems in another region.

The researchers specifically apply their theory to autism, though they note that it could help understand other childhood neurological conditions. Conditions within the autism spectrum present “longstanding puzzles” related to cognitive and behavioral disruptions that their ideas could help resolve, they wrote. Under their theory, cerebellar injury causes disruptions in how other areas of the brain develop an ability to interpret external stimuli and organize internal processes, explained first author Sam Wang, an associate professor of molecular biology and the Princeton Neuroscience Institute (PNI).

"It is well known that the cerebellum is an information processor. Our neocortex [the largest part of the brain, responsible for much higher processing] does not receive information unfiltered. There are critical steps that have to happen between when external information is detected by our brain and when it reaches the neural cortex," said Wang, who worked with doctoral student Alexander Kloth and postdoctoral research associate Aleksandra Badura, both in PNI.

"At some point, you learn that smiling is nice because Mom smiles at you. We have all these associations we make in early life because we don’t arrive knowing that a smile is nice," Wang said. "In autism, something in that process goes wrong and one thing could be that sensory information is not processed correctly in the cerebellum."

Mustafa Sahin, a neurologist at Boston’s Children Hospital and associate professor of neurology at Harvard Medical School, said that Wang and his co-authors build upon known links between cerebellar damage and autism to suggest that the cerebellum is essential to healthy neural development. Numerous studies — including from his own lab — support their theory, said Sahin, who is familiar with the work but was not involved in it.

"The association between cerebellar deficits and autism has been around for a while," Sahin said. "What Sam Wang and colleagues do in this perspective article is to synthesize these two themes and hypothesize that in a critical period of development, cerebellar dysfunction may disrupt the maturation of distant neocortical circuits, leading to cognitive and behavioral symptoms including autism."

Traditionally, the cerebellum has been studied in relation to motor movement and coordination in adults. Recent studies, however, strongly suggest that it also influences childhood cognition, Wang said. Several studies also have found a correlation between cerebellar injury and the development of a disorder in the autism spectrum, the researchers report. For instance, the researchers cite a 2007 paper in the journal Pediatrics that found that individuals who experienced cerebellum damage at birth were 40 times more likely to score highly on autism screening tests. They also reference studies in 2004 and 2005 that found that the cerebellum is the most frequently disrupted brain region in people with autism.

"What we realized from looking at the literature is that these two problems — autism and cerebellar injury — might be related to each other" via the cerebellum’s influence on wider neural development, Wang said. "We hope to get people and scientists thinking differently about the cerebellum or about autism so that the whole field can move forward."

The researchers conclude by suggesting methods for testing their theory. First, by inactivating brain-cell electrical activity, it should be possible to pinpoint the developmental stage in which injury to one part of the brain affects the maturation of another. A second, more advanced method is to reconstruct the neural connections between the cerebellum and other brain regions; the federal BRAIN Initiative announced in 2013 aims to map the activity of all the brain’s neurons. Finally, mouse brains can be used to disable and restore brain-region function to observe the “upstream” effect in other areas.

Are Three Brain Imaging Techniques Better than One?

Many recent imaging studies have shown that in children with autism, different parts of the brain do not connect with each other in typical ways. Initially, most researchers thought that the autistic brain has fewer connections between key regions. The most recent studies, however, point to an opposite conclusion: The brains of people with autism exhibit overconnectivity.

To date, almost all studies of autism in children have used a single imaging technique to explore connectivity. None has been able to capture a robust picture of the brain abnormalities associated with autism—until now.

Two new grants from the National Institute of Mental Health (NIMH) will allow San Diego State University Psychology Professor Ralph-Axel Müller to combine three imaging techniques and harness the best of each one in his study of autism.

Techniques in tandem

Although the term “brain imaging” gets thrown around a lot when describing the latest advances in neuroscience and psychology, there are dozens of different brain imaging techniques. Each gives scientists a different view of the inner workings of the brain, and each comes with its own strengths and limitations.

For example, the frequently cited technique of fMRI, or functional magnetic resonance imaging, measures blood flow in different areas of the brain at specific snapshots in time, based on the knowledge that increased blood flow indicates increased activity of nerve cells in that area of the brain. The technique is powerful, but has limitations when it comes to detecting dynamic changes in brain activity that occur very fast, within milliseconds.

EEG (electroencephalography), a much older technique, is actually better at detecting such dynamic changes, although it cannot pinpoint exactly where in the brain the activity occurs. A powerful and more recent technique is MEG, or magnetoencephalography, which can detect dynamic changes in brain activity that happen within a few milliseconds.

Müller looks for disorganized patterns of brain activity that could be responsible for some of the telltale characteristics of autism spectrum disorder, such as inattention to social cues and repetitive and obsessive behaviors. For example, last year, Müller and his colleagues discovered that in children with autism, connectivity was impaired between the cerebral cortex and the thalamus, a deep brain structure that is important for sensorimotor functions and attention.

With $4.2 million in new funding from NIH, Müller—together with collaborators Ksenija Marinkovic at SDSU and Thomas Liu at the University of California, San Diego—will apply fMRI, EEG, and MEG to study both autistic and non-autistic, or typically-developing, children and adolescents during a variety of tests, including language tests designed to tease out activity in various parts of the brain.

Defining the differences

One component of the project will concern the visual system. Previous research has shown that people with autism rely on their visual cortex more than typically- developing people during thought processes, for example, when making a semantic distinction, such as deciding whether a truck is a vehicle. Using the one-two punch of fMRI and MEG together, Müller and his team will be able to determine the dynamic processes in how brain regions work together to come up with a response, and how these processes differ in autism.

The study will also examine brain function during its resting state in order to identify abnormalities in brain network organization. The combined use of EEG and MEG, together with fMRI techniques that reveal brain anatomy, will produce a much more complete picture of abnormal brain organization in autism.

Ultimately, Müller and his colleagues hope to identify biomarkers in the brain that can reliably indicate whether the participant falls on the autism spectrum.

“Autism is a brain-based disorder, but its diagnosis is still based entirely on behavioral observation,” Müller said. “This is inadequate. We need to find brain biomarkers for autism.”

Another goal of the researchers is to find brain biomarkers that can distinguish different subtypes of autism. It is generally suspected that the term “autism” actually covers several different disorders, each of which may be caused by different genetic and environmental risk factors. Eventually, brain biomarkers might be tied to genetic data, giving scientists a better understanding of the origins of autism, as well as new leads for treatment.

“For decades, research teams studying autism have specialized in one or another scientific technique, often without understanding well what other techniques can reveal. Our study combining several of the major imaging techniques will be one step toward a more comprehensive account of how the autistic brain differs from the typically developing one – and what may be done about it,” Müller said.

(Image caption: In a study of brains from children with autism, neurons in brains from autistic patients did not undergo normal pruning during childhood and adolescence. The images show representative neurons from unaffected brains (left) and brains from autistic patients (right); the spines on the neurons indicate the location of synapses. Credit: Guomei Tang, PhD and Mark S. Sonders, PhD/Columbia University Medical Center)

Children with Autism Have Extra Synapses in Brain

Children and adolescents with autism have a surplus of synapses in the brain, and this excess is due to a slowdown in a normal brain “pruning” process during development, according to a study by neuroscientists at Columbia University Medical Center (CUMC). Because synapses are the points where neurons connect and communicate with each other, the excessive synapses may have profound effects on how the brain functions. The study was published in the August 21 online issue of the journal Neuron.

A drug that restores normal synaptic pruning can improve autistic-like behaviors in mice, the researchers found, even when the drug is given after the behaviors have appeared.

“This is an important finding that could lead to a novel and much-needed therapeutic strategy for autism,” said Jeffrey Lieberman, MD, Lawrence C. Kolb Professor and Chair of Psychiatry at CUMC and director of New York State Psychiatric Institute, who was not involved in the study.

Although the drug, rapamycin, has side effects that may preclude its use in people with autism, “the fact that we can see changes in behavior suggests that autism may still be treatable after a child is diagnosed, if we can find a better drug,” said the study’s senior investigator, David Sulzer, PhD, professor of neurobiology in the Departments of Psychiatry, Neurology, and Pharmacology at CUMC.

During normal brain development, a burst of synapse formation occurs in infancy, particularly in the cortex, a region involved in autistic behaviors; pruning eliminates about half of these cortical synapses by late adolescence. Synapses are known to be affected by many genes linked to autism, and some researchers have hypothesized that people with autism may have more synapses.

To test this hypothesis, co-author Guomei Tang, PhD, assistant professor of neurology at CUMC, examined brains from children with autism who had died from other causes. Thirteen brains came from children ages two to 9, and thirteen brains came from children ages 13 to 20. Twenty-two brains from children without autism were also examined for comparison.

Dr. Tang measured synapse density in a small section of tissue in each brain by counting the number of tiny spines that branch from these cortical neurons; each spine connects with another neuron via a synapse.

By late childhood, she found, spine density had dropped by about half in the control brains, but by only 16 percent in the brains from autism patients.

“It’s the first time that anyone has looked for, and seen, a lack of pruning during development of children with autism,” Dr. Sulzer said, “although lower numbers of synapses in some brain areas have been detected in brains from older patients and in mice with autistic-like behaviors.”

Clues to what caused the pruning defect were also found in the patients’ brains; the autistic children’s brain cells were filled with old and damaged parts and were very deficient in a degradation pathway known as “autophagy.” Cells use autophagy (a term from the Greek for self-eating) to degrade their own components.

Using mouse models of autism, the researchers traced the pruning defect to a protein called mTOR. When mTOR is overactive, they found, brain cells lose much of their “self-eating” ability. And without this ability, the brains of the mice were pruned poorly and contained excess synapses. “While people usually think of learning as requiring formation of new synapses, “Dr. Sulzer says, “the removal of inappropriate synapses may be just as important.”

The researchers could restore normal autophagy and synaptic pruning—and reverse autistic-like behaviors in the mice—by administering rapamycin, a drug that inhibits mTOR. The drug was effective even when administered to the mice after they developed the behaviors, suggesting that such an approach may be used to treat patients even after the disorder has been diagnosed.

Because large amounts of overactive mTOR were also found in almost all of the brains of the autism patients, the same processes may occur in children with autism.

“What’s remarkable about the findings,” said Dr. Sulzer, “is that hundreds of genes have been linked to autism, but almost all of our human subjects had overactive mTOR and decreased autophagy, and all appear to have a lack of normal synaptic pruning. This says that many, perhaps the majority, of genes may converge onto this mTOR/autophagy pathway, the same way that many tributaries all lead into the Mississippi River. Overactive mTOR and reduced autophagy, by blocking normal synaptic pruning that may underlie learning appropriate behavior, may be a unifying feature of autism.”

Alan Packer, PhD, senior scientist at the Simons Foundation, which funded the research, said the study is an important step forward in understanding what’s happening in the brains of people with autism.

“The current view is that autism is heterogeneous, with potentially hundreds of genes that can contribute. That’s a very wide spectrum, so the goal now is to understand how those hundreds of genes cluster together into a smaller number of pathways; that will give us better clues to potential treatments,” he said.

“The mTOR pathway certainly looks like one of these pathways. It is possible that screening for mTOR and autophagic activity will provide a means to diagnose some features of autism, and normalizing these pathways might help to treat synaptic dysfunction and treat the disease.”

(Image credit: The insular cortex of an autism mouse model is already so strongly activated by a single sensory modality (here a sound), that it is unable to perform its role in integrating information from multiple sources. Credit: © MPI of Neurobiology / Gogolla)

Insular cortex alterations in mouse models of autism

The insular cortex is an integral “hub”, combining sensory, emotional and cognitive content. Not surprisingly, alterations in insular structure and function have been reported in many psychiatric disorders, such as anxiety disorders, depression, addiction and autism spectrum disorders (ASD). Scientists from Harvard University and the Max-Planck Institute of Neurobiology in Martinsried now describe consistent alterations in integrative processing of the insular cortex across autism mouse models of diverse etiologies. In particular, the delicate balance between excitation and inhibition in the autistic brains was disturbed, but could be pharmacologically re-adjusted. The results could help the development of novel diagnostic and therapeutic strategies.

Autism is a neurodevelopmental disorder characterized by impaired social interaction, verbal and non-verbal communication, and by restricted and repetitive behaviors. Diagnosis is solely based on behavioral analysis as biological markers and neurological underpinnings remain unknown. This makes the development of novel therapeutic strategies extremely difficult. 

As the cellular basis of autism spectrum disorders cannot be addressed in human patients, scientists have developed a number of mouse models for the disease. Similar to humans, mice are social animals and communicate through species-specific vocalizations. The mouse models harbor all diagnostic hallmark criteria of autism, such as repetitive, stereotypic behaviors and deficits in social interactions and communication.

Nadine Gogolla and her colleagues in the laboratory of Takao Hensch at Harvard University have now searched for common neural circuit alterations in mouse models of autism. They concentrated on the insular cortex, a brain structure that contributes to social, emotional and cognitive functions. ‘We wanted to know whether we can detect differences in the way the insular cortex processes information in healthy or autism-like mice’, says Nadine Gogolla, who was recently appointed Leader of a Research Group at the Max Planck Institute of Neurobiology.

As the researchers now report, the insular cortex of healthy mice integrates stimuli from different sensory modalities and reacts more strongly when two different stimuli are presented concomitantly (e.g. a sound and a touch). ‘We recognize a rose more easily when we smell and see it rather than when we just see or smell it’ says Nadine Gogolla. This capacity of combining sensory stimuli was consistently affected in all autism models the researchers looked at. Interestingly, often one sense alone elicited such a strong response that adding a second modality did not add further information. This is very reminiscent of the sensory hyper-responsiveness experienced by many autistic patients. The scientist further discovered that the insular cortex of adult autism-model mice resembled the activation patterns observed in very young control mice. ‘It seemed as if the insular cortex of the autism-models did not mature properly after birth’, says Gogolla.

For proper brain function, excitation and inhibition have to be in equilibrium. In the now identified part of the insular cortex, the scientists found that this equilibrium was disturbed. In one of the mouse models, inhibitory contacts between nerve cells were strongly reduced.

To test the influence of this reduction on sensory processing, the researchers gave mice the drug Diazepam, which is also known under the trade name Valium, to boost inhibitory transmission in the brain. Indeed, this treatment transiently rescued the capacity of the insular cortex to combine stimuli of different sensory modalities. The balance between excitation and inhibition in the brain is established after birth. The scientists thus treated young animals over several days with Diazepam. This treatment was efficient in reestablishing the insular cortex capacity for sensory integration permanently, even in adult mice that did not received any further treatment. Interestingly, also the stereotypic grooming of the animals was significantly reduced.

All autism models investigated showed alterations in inhibitory molecules. However, the alterations were very diverse. While in some models certain molecules were reduced, the opposite was true in another model. These results suggest that the disequilibrium between excitation and inhibition may be an important factor in the neuropathology of autism. However, future therapies will need to be carefully tailored to each particular subgroup of autism. For instance, an artificial boost of inhibition through a drug like Diazepam in healthy mice can throw the delicate equilibrium off and create changes in the insular cortex similar to those seen in the autism models. Whether a therapeutic strategy aimed on keeping the brain’s equilibrium between excitation and inhibition could be useful and if so, how to test the individuals’ status of the excitation/inhibition balance and how to implement individually tailored treatments, would need to be established through further studies and pre-clinical tests.