Glia, the non-neuronal cells that reside in the brain, were once thought to play a secondary role to neurons. But studies in the past few years have shown that, in fact, they play critical roles and may be involved in many brain disorders, including autism.

On 7 December, the Simons Foundation Autism Research Initiative, SFARI.org’s parent organization, along with the Rett Syndrome Research Trust, organized a meeting of top scientists working on the intersection of glia with autism and Rett syndrome.

Most autism research focuses on neurons. However, it is becoming increasingly clear that glia, named after the Greek word for ‘glue,’ are essential for brain development and function. They may also be potential targets for autism treatments.

The first scientists studying glia were able to differentiate them based on their size and shape. For example, astrocytes were named for their star-like shape, and microglia for their relatively small size.

Subtypes of glia do not all originate from the same source. Some of the earliest cells involved in brain growth are radial glia, which serve as both structural scaffolds and as a source of new cells for the developing organ.

The majority of cell types in the brain arise from these radial glia. They include neurons and two of the most critical types of glia, astrocytes and oligodendrocytes, without which neurons cannot properly function. The mature brain also contains adult neural stem cells that can become new neurons or glia.

Microglia, on the other hand, are unique because they derive in the yolk sac from a type of immune cell called primitive macrophages. These macrophages develop before the cells that later give rise to the majority of the immune system. Microglia derive from macrophages before blood cells. In mice, they enter the developing brain after about ten days of gestation.

Microglia are not only needed for their protective immune role in the adult brain, but for normal brain development and maintenance.

Early studies of glia focused on their role in injury and infection, because they tend to respond dramatically when the brain’s integrity is compromised. But sensitive new technology has made it possible to detect, and even see, the key functions glia play in the normal brain. This technology has also detected differences in glia in the brains of individuals with autism.

Several teams have measured significant differences in the number of microglia in the brains of individuals with autism compared with controls. There is evidence of innate immune cell and microglial activation in autism, according to Carlos Pardo, assistant professor of neurology and pathology at Johns Hopkins University in Baltimore, Maryland.

Data from his and other labs suggest that astrocytes and microglia are activated in the brains of individuals with autism, and that this may be what causes the symptoms of the disorder. Autism brains also have elevated levels of cytokines, chemicals that relay signals between immune cells. These cytokines include the interleukins 6 and 10 and transforming growth factor-beta, all of which are potent regulators of the inflammatory immune response.

In particular, glia are hyperactivated in autism brains in the cerebellum, a part of the brain involved in balance, movement coordination, learning and emotion processing, according to Pardo.

The cerebellum contains about 70 percent of all the neurons in the brain, with about 100 billion granule cells, which integrate signals from throughout the central nervous system. They also contain about 25 million Purkinje neurons, which inhibit brain signals.

Purkinje cells are highly active and so burn a lot of energy. This makes them more vulnerable to stress and damage than other parts of the brain, says John Allman, professor of neurobiology at the California Institute of Technology in Pasadena, California.

Some autism brains may have about 20 percent more microglia than control brains in certain regions, and 120 percent more in a layer of the cerebellum in which Purkinje neurons receive signals. They also have about one-third fewer Purkinje neurons than controls do.

Because Purkinje neurons are critical for integrating, receiving and sending signals to and from the cerebellum, their loss may cause serious problems in coordination and balance, says Allman.

Allman’s group has used RNA-Seq — a technique that measures the abundance of messenger RNA, the intermediate between DNA and protein — to look at differences between the cerebellum in autism brains and control brains. The researchers combined this approach with a technique called laser microdissection to isolate individual Purkinje cells from postmortem brains.

They found that levels of TMSB4X (beta-thymosin), which may protect neurons, and CDR1, or cerebellum damage response 1, are both highly elevated in Purkinje cells in autism brains.

CDR1 levels are also elevated in people who have cerebellar neoplastic syndrome. In this syndrome, antibodies against cancer cells elsewhere in the body cross the blood-brain barrier and activate an autoimmune response against CDR1, leading to Purkinje cell damage.

Immune activity outside of the brain may lead to the elevated levels of CDR1 in Purkinje cells in autism brains, and enhanced microglia activity may be a response to the resulting changes in Purkinje cells, says Allman.

Although it is not clear what elevated CDR1 levels mean for autism brains, it may indicate that cerebellar damage related to microglia and immune responses is a feature of autism.

Individuals with autism have elevated levels of three immune molecules — CX3CL1, IL-8 and MCP1 — in their cerebrospinal fluid, according to Pardo. This may indicate inflammation. It is likely that astrocytes produce MCP1 and induce brain inflammation by activating both microglia and monocytes.

However, neurons under stress may release CX₃CL1, says Richard Ransohoff, professor of neuroscience at the Cleveland Clinic. Because there are multiple pathways by which the activity of these molecules can be enhanced, it is unclear whether these inflammatory changes are the cause of, or response to, brain pathology in autism.

Based on these data, Pardo led a clinical trial to see whether minocycline — which may lessen glial activation — can alleviate autism symptoms. The trial showed that minocycline has a negligible effect on autism symptoms, and may not be specific enough to target immune activation in autism brains.

A better understanding of the role inflammation plays in autism may lead to more targeted immune-based therapies, the workshop participants said.

Activated glia are classically associated with inflammation-inducing events such as brain damage and infection. In contrast, the autism brain may activate glia to improve brain function. In this case, decreasing glial activation would be the wrong approach and researchers should instead seek to enhance or increase glial activity.

In the past few years, it has become increasingly clear that the immune system interacts intimately with the nervous system to maintain brain function. The primary immune cells within the brain are microglia, which populate the brain early in development.

Astrocytes develop from the same progenitors as neurons do, but are capable of releasing immune molecules. They also have some ability to perform phagocytosis — a process by which cells eat pathogens, debris, pieces of other cells and dead cells.

Ransohoff studies the difference between microglia and monocyte-derived macrophages in the brain. Monocytes are immune cells that are made in the bone marrow and travel within the blood to circulate in the body. They are capable of secreting immune molecules and differentiate into macrophages when they leave the bloodstream to enter other tissues.

Monocyte-derived inflammatory macrophages and microglia play unique roles in a mouse model of multiple sclerosis, called experimental autoimmune encephalitis.

In these mice, astrocytes secrete CCL2, a chemokine that attracts monocytes to the brain via CCR2, which is absent from microglia. Ransohoff’s team used CCR2 as a marker to distinguish cells derived from monocytes and those from microglia. Once in the brain, monocytes become inflammatory macrophages and strip myelin, an insulating protein, off neuronal projections, which impairs neuronal signaling. Microglia then clean up the myelin debris left by the inflammatory macrophages.

Microglia and monocytes are distinct cell types with different functions, and researchers should make an effort to distinguish them, Ransohoff says.

Beth Stevens, assistant professor of neurology at Children’s Hospital Boston, also studies a potentially beneficial role of microglia — removing unneeded neuronal synapses, or junctions.

During brain development, neurons form more synapses than are necessary. Microglia preferentially phagocytose, or eat, the synapses that have weak signals. In this way, microglia facilitate the formation of an efficient, strong neuronal network.

Complement proteins, traditionally thought to be immune molecules that fight infection, are involved in this process. Complement proteins tag synapses for elimination, and microglia sense these tags via receptors on their surface.

Although researchers originally assumed that microglia are quiescent until an inflammatory event occurs, they are in fact highly active.

Wenbiao Gan, professor of physiology and neuroscience at the Skirball Institute at New York University, has shown using microscopy in live mice that microglial processes are dynamic. Microglial processes also respond rapidly to injury. Remodeling of the signal-receiving branches of neurons, a process that underlies learning and memory, is less active in the absence of microglia. This further supports the idea that microglia are critical for the proper regulation of synapses.

Studies suggest that fetal exposure to immune molecules may lead to autism. Work by Stevens, Gan and others suggests that inflammation can lead to elevated levels of microglia, which would influence processes that regulate synapses. Fine-tuning of synapses underlies learning and memory and is a key factor in brain development, so these changes may lead to problems later in life.

Philip Haydon, professor of neuroscience at Tufts University in Boston, is interested in the role of astrocytes in the synapse. Over the past few years, his group has shown a critical role for astrocytes in regulating the sleep cycle. Astrocytes interact in a synapse between two communicating neurons. This trio is together called the tripartite synapse.

In the tripartite synapse, astrocytes modulate chemical messengers that signal between neurons. They regulate sleep by releasing the energy molecule adenosine triphosphate (ATP), which is converted to adenosine and then reabsorbed by the astrocyte.

When people are awake, adenosine builds up at synapses As adenosine levels increase, so does the desire to sleep. During sleep, astrocytes remove adenosine and thereby reset the system.

Dysfunction in this system affects several diseases, and alcoholism can perturb adenosine balance and cause sleep fragmentation. However, a night of sleep deprivation may actually have beneficial effects for those suffering from depression. As sleep problems are common in autism, it will be important to investigate if this sleep regulatory system is disrupted.

Rett syndrome primarily affects girls and has a prevalence of 1 in 10,000 female births. It is caused by mutations in MeCP2, which regulates the expression of thousands of other genes.

Males with a mutation in MeCP2 typically either do not survive gestation or die young. Rett syndrome is usually diagnosed around 18 months of age when symptoms become apparent. People with Rett syndrome have many features of autism, such as communication deficits, but also have growth and motor deficits, seizures and breathing dysfunction — all of which may lead to early death.

A 2011 study showed that astrocytes play a key role in Rett syndrome, because restoring healthy astrocytes in a mouse model of Rett syndrome fixed many aspects of the disorder. Genetic rescue is not currently possible in people.

However, last year Jonathan Kipnis, professor of neuroscience at the University of Virginia in Charlottesville, showed that transplanting healthy bone marrow — which can repopulate healthy microglia in the brain — into mice with Rett syndrome arrests many Rett symptoms.

Preventing the migration of new transplanted microglia into the brain interferes with the beneficial effects of the transplant, suggesting that microglia are critical for the rescue. Also, microglia from mice that model Rett syndrome are unable to properly perform phagocytosis, which cleans up dead cells and prunes synapses during brain development.

As Rett syndrome has many features of autism, these findings suggest that microglial defects may also be involved in autism.

In fact, a 2012 study found that bone marrow transplantation ameliorates autism-like symptoms in mice that were exposed to elevated immune activity during gestation. The same technique, if it proves safe and beneficial, could be used to treat people with the disorder.

Mice lacking microglia are unable to properly remodel dendritic spines, according to Gan. And conversely, dendritic spine remodeling is enhanced in a Rett syndrome model, which may have activated microglia. This effect may develop with age, as 20-day-old mice behave normally, whereas 30-day-old mice are significantly different from controls, said Gan.

The fact that these defects appear to develop with age in mice gives some clues as to why symptoms do not become apparent until about 18 months in girls with Rett syndrome. Why this is the case, and whether this is driven by alterations to microglia, is still under investigation.

Overall, data presented at the workshop suggest that there is elevated inflammation and high numbers of microglia in autism brains, and indicate a critical role for glia in healthy brain development and function. Together these findings and future studies may lead to therapies that alter or replace glia for the treatment of autism and related disorders.

However, in the final discussion, the participants raised an important question: Are the elevated numbers of microglia in autism brains causing symptoms of the disorder, or are they the reaction to an unhealthy environment?

The answer to this question may significantly change the course of both research and the development of treatments, and should be considered in research addressing the role of glia in autism.