Sitting in a darkened lab at the National Institutes of Health in 1999, my colleague Beth Stevens and I prepared to send a mild electric current through fetal mouse neurons in a cell culture. We were using a new microscope technique that would let us see electrical activity as a bright fluorescence emitted from a dye we had added to the culture, and we were hoping to find out if another kind of cell common in the nervous system would react in some way—Schwann cells, odd-looking cells that fabricate insulation around neurons. We didn’t really expect them to; Schwann cells cannot communicate electrically. I flipped the switch. The neurons immediately glowed. But then the Schwann cells began to glow as well. It was as if they were talking back.
The most mysterious substance on earth is the stuff between your ears, and much of the intrigue exists because many long-held beliefs about how the brain works have turned out to be wrong. Like medieval astronomers who were shocked to learn that the earth is not the center of the universe, neuroscientists today are facing a similar revelation about neurons.
Until recently, our understanding of the brain was based on a century-old idea called the neuron doctrine. This theory holds that all information in the nervous system is transmitted by electrical impulses over networks of neurons linked through synaptic connections. But this bedrock theorem is deeply flawed. New research proves that some information bypasses the neurons completely, flowing without electricity through networks of cells called glia. The studies are upending our understanding of every aspect of brain function in health and disease, bringing answers to long-standing riddles about how we remember and learn.
Glial cells interact with neurons, control them, work alongside them—and the functions of these strange-looking cells are myriad. Star-shaped astrocytes ferry neurotransmitters, food and waste. Cephalopodlike oligodendrocytes and sausage-shaped Schwann cells wrap themselves around neurons like sheaths, speeding their electrical transmissions and helping control muscle contractions throughout the body. Microglia, ranging in form from multibranch to ameboid, are the brain’s first responders to injury and disease, killing invading germ cells and beginning the process of repair.
Especially exciting is new research showing the central role of glia in information processing, neurological disorders and psychiatric illness. Some glial cells speed information between distant regions of the brain, helping us master complex cognitive processes. Others break down as they age and in their failure bring dementia. This research has great implications not only for understanding how the brain works but also for developing new treatments for neurological and psychological illnesses.
And all this comes down to a class of brain cells dismissed for 100 years as mere putty. In the 19th century, when pioneering scientists first trained microscopes on gray matter, they were amazed to find a cell unlike any other in the body: the neuron. At one end of this dazzling cell was a long, wirelike structure called the axon that carried electrical impulses to a cluster of transmission terminals. At the opposite end, the neuron sprouted busy, rootlike dendrites that received signals from the axons of other neurons, ferried across the space that separated them—the synapse—by tailor-made chemicals. Neurons were scattered sparsely throughout the brain like juicy raisins, but few cared to examine the seemingly bland dough in which they were embedded.
But, as Sherlock Holmes observed, “There is nothing more deceptive than an obvious fact,” and the fact that scientists were ignoring is that neurons make up only 15 percent of our brain cells; the other 85 percent were considered little more than packing material. Indeed, 19th-century German pathologist Rudolf Virchow, one of the first to study glia, likened this brain matter to connective tissue and dubbed it nervenkitt, meaning nerve putty or cement, which in English became “neuroglia,” from the Greek root for glue.
Few scientists are drawn to brain research to study glue. We still have no singular noun equivalent to neuron when we speak about an individual glial cell. Virchow barely distinguished between the different sorts of glia. And none of this mishmash of bizarre-looking cells had any of the telltale features essential for neuronal communication, such as axons, dendrites or synapses, so scientists had no reason to suspect that glia might be communicating in secret and doing so in an unexpected way.
A Language of Their Own
Neurons use both electricity and chemistry to convey information, with electricity transmitting impulses along the wirelike axon and chemicals carrying those signals across the synapse to another neuron. The recipient neuron then fires an electrical impulse and relays the signal to the next neuron in the chain.
Only in the past few years have scientists come to realize that the glial cells called astrocytes can control synaptic communication. So named because early anatomists thought they resembled stars, astrocytes were at first thought to be responsible only for housekeeping functions such as transporting nutrients from the bloodstream to the neurons and carrying waste in the opposite direction. These functions were surmised from the way many astrocytes cling to blood vessels with some of their arms and reach deep into brain tissue with others, tightly grasping neurons and their synapses. Only later did scientists come to see that neurons are utterly dependent on glia to fire their electrical impulses and to pass messages to one another across synapses. A clue that this dependency might be the case was the discovery of the same neurotransmitter receptors on glia as on neurons. As it happens, glia were listening to neurons and talking among themselves without using electricity at all.
This discovery awaited the invention of new tools allowing electrical activity to be seen as flashes of light. The microelectrodes that neuroscientists typically use to probe neuronal function are deaf to glial communication. But video and laser-illuminated microscopes developed in the 1980s and 1990s let researchers monitor neuronal firing by adding tracer dyes to the cells. Like the fluorescent fluid in a glow stick, these dyes shone when ions such as calcium entered neurons as their axons carried a signal, causing the dye to generate light. Very quickly those of us using these new methods saw that when we stimulated a neuron to fire an impulse, the neuroglia, hidden in plain sight, flashed back. Glia had sensed the electrical activity in neurons, and somehow calcium ions had flooded into them as well, producing the same green glow.
The new technique also revealed that glia communicate with one another in the same way. Scientists observed that when neurotransmitters released by neurons stimulated receptors on glia, the glia released neurotransmitters as well. And the release stimulated a chain reaction as the message was passed to other glia. The glial communication is stunningly evident as a wave of fluorescent light sweeping from one glial cell to the next after a neuron has fired and released a neurotransmitter.
This finding led to a bigger question: whether glial networks use the information gleaned about neuronal communication at a synapse to manage neuronal signaling at synapses in distant parts of the brain. If so, glia might have a central part in information processing itself.
Recent research provides tantalizing evidence of such a role. Using a laser to stimulate a calcium wave in an astrocyte next to an axon, a team led by neurobiologist Norio Matsuki of the University of Tokyo reported earlier this year that neurotransmitters released from the astrocyte boosted the strength of an electrical impulse in the axon. A 2005 study led by neurobiologist Philip Haydon, now at Tufts University, showed that astrocytes provide a nonelectrical pathway for communication between synapses in a brain area governing memory, the hippocampus. After responding to the neurotransmitter glutamate released from one synapse, astrocytes released a different neurotransmitter, adenosine, affecting the strength not only of its neuronal neighbor but of distant synapses as well. By controlling data processing at synapses, glia participate in aspects of vision, memory, muscle contraction and unconscious brain functions such as sleep and thirst.
The pace and breadth of glial communication provide another bit of evidence that glia play a part in information processing. Unlike neurons, which communicate serially across chains of synapses, glia broadcast their signals widely, like cell phones. Neurons’ electrical communication is quite rapid, zipping through neural networks in mere thousandths of a second, but the chemical communication of glia is very slow, spreading as a tidal wave through neural tissue at a pace of seconds or tens of seconds. Rapid response is critical for certain functions—reflexive recoil from a pain stimulus, for example—but many important processes in the brain occur over longer periods.
Not the least important of these is learning. New human brain-imaging techniques have revealed that after learning to play a musical instrument or to read or to juggle, structural changes occur in brain areas that control these cognitive functions. Remarkably the changes are seen in regions where there are no complete neurons: the “white matter” areas, formed from bundles of axons coated with myelin, a white electrical insulator. Previously all theories of learning held that we incorporate new information solely by strengthening synaptic connections, but there are few synapses in white matter. Clearly, something else is happening.
Findings from my lab in the past 10 years concern two different types of glial cells that cling to axons and coat them with myelin insulation—oligodendrocytes in the brain and Schwann cells in the body. Like an octopus, each cellular tentacle of an oligodendrocyte cell grips an individual segment of an axon and wraps up to 150 layers of compacted cell membrane around it in the way an electrician wraps tape around a wire. This insulation changes how impulses travel through axons, increasing the transmission speed by up to 50 times.
And much like astrocytes at synapses, these myelin-forming glia could sense the impulses transmitted through axons. This capability was a puzzle at first, because such glia are far from the synapses where neurotransmitters are released. But my lab recently discovered that axons also release neurotransmitters through channels in their membrane that open when the axon fires. I was able to see the release of one such neurotransmitter—adenosine triphosphate, or ATP—by fitting my microscope with an extremely high-gain night-scope image intensifier that can detect single photons. For my experiment, I exploited the chemical reaction that produces a firefly’s telltale green flash. I took the protein and enzyme from the tail of a firefly and added them to cultures containing mouse neurons. The firefly proteins require one more ingredient before they can glow: ATP, normally supplied by firefly cells. When I stimulated the mouse axons with a mild electric shock, they released ATP, eliciting a burst of photons.
The formation of myelin in response to stimuli likely means that early life experience plays a big role in brain development. By increasing the speed of information transfer between parts of the brain involved in mastering complex cognitive tasks, these glial cells are essential to learning, too.
How the Brain Goes Awry
Glial cells have also emerged as major actors in a host of neurological and psychological illnesses ranging from epilepsy to chronic pain to depression. Indeed, recent research has found that many neurological disorders are in fact disorders of the glia, in particular a class of cells called the microglia, which serve as the brain’s defense against disease. These specialists seek out and kill invading germs and promote recovery from injury, clearing away diseased tissue and releasing powerful compounds that stimulate repair. And their function is a factor in every aspect of neurological illness.
New research suggests to some scientists that the dementia of Alzheimer’s disease could be a direct outcome of microglia that have lost the ability to clear waste. Alois Alzheimer first noted that microglia surround the amyloid plaques that are the hallmark of the disease. Normally microglia digest the toxic proteins that form these plaques. But recent studies led by neuroscientist Wolfgang J. Streit of the University of Florida College of Medicine and others suggest that microglia become weaker with age and begin to degenerate. The atrophy is visible under a microscope. Senescent microglia in aged brain tissue become fragmented, losing many of their cellular branches.
The way Alzheimer’s courses through the brain is one more sign of microglial involvement. Tissue damage spreads in a predetermined manner, beginning near the hippocampus and eventually reaching the frontal cortex. Streit’s observations show that microglial degeneration follows the same pattern—and in advance of neuronal degeneration, suggesting that senescence of microglia is a cause of Alzheimer’s dementia and not a response to neuron damage, as Alzheimer and most experts had presumed. This discovery may lead to new treatments for dementia, once researchers determine why microglia become senescent with age in some people but not in others.
The functions of the glial cells also account for why some people develop horrible chronic pain that does not relent after an injury has healed and sometimes even worsens. Doctors must use powerful narcotics such as morphine and other opiates to blunt the unrelenting pain in such patients. These drugs lose their strength over time, necessitating higher doses for the same effects, which can lead to drug dependence [see “When Pain Lingers,” by Frank Porreca and Theodore Price; Scientific American Mind, September/October 2009].
We now know that malfunctions of glial cells may account for both persistent pain and the diminishing power of some pain-relieving drugs. Research by Linda Watkins of the University of Colorado at Boulder, Kazuhide Inoue of Kyushu University in Fukuoka, Japan, and Joyce DeLeo of Dartmouth Medical School, among many others, reveals that microglia and astrocytes respond to the hyperactivity in pain circuits after injury by releasing compounds that initiate the healing process. These substances also stimulate neurons. Initially this heightened sensitivity is beneficial, because the pain forces us to protect the injury from further damage. With chronic pain, microglia do not stop releasing these substances even when healing is complete. But in recent studies pain in experimental animals was sharply reduced when the researchers blocked either the signals from neurons to glia or the signals that glia release. Scientists are now developing painkillers that target glia rather than neurons.
Glial cells also account for the ancient mystery of why spinal cord injury results in permanent paralysis. Martin Schwab of the University of Zurich and others have found that proteins in the myelin insulation that oligodendrocytes wrap around axons stop injured axons from sprouting and repairing damaged circuits. Blocking these proteins allows damaged axons to regrow in experimental animals. Clinical trials on patients with spinal cord injury are now under way.
That glia would play a central role in neurological illness is easy to understand because astrocytes and microglia are the first responders to disease. We have also long known that demyelinating disorders such as multiple sclerosis, which strips the myelin insulation from axons, cause severe disability. But it came as a recent surprise to find glia implicated in psychiatric illness. Recent work has linked chemicals called cytokines, which are released by immune system cells and microglia, to obsessive-compulsive disorder. In 2002 molecular geneticist Mario Capecchi and his colleagues in the department of human genetics at the University of Utah reported that mice with a mutation in the Hoxb8 gene exhibited compulsive grooming and hair removal behavior similar to humans with obsessive-compulsive disorder. The only cells in the brain that have this gene are microglia. Then, in a 2010 study, the researchers harvested immature immune cells that will develop into microglia from normal mice and transplanted them into the mutants. The mice were cured of their compulsive grooming behavior. Presumably cytokines released from microglia excite brain circuits responsible for habit formation. [For more about habits, see “Obsessions Revisited,” by Melinda Wenner Moyer.]
Analysis of postmortem brain tissue has also linked oligodendrocytes and astrocytes to depression and schizophrenia by revealing reduced numbers of these cells. So have MRI examinations of people with schizophrenia, which show anomalies in subcortical white matter regions of the brain. Although psychiatric illnesses are likely to have many different causes, schizophrenia and several other mental illnesses have a strong genetic basis. If an identical twin develops schizophrenia, there is a 50–50 chance that the sibling will as well.
Some of the genes implicated in these mental illnesses are found only in oligodendrocytes; others control development of these myelin-forming glia. An analysis of 6,000 genes in tissue from the prefrontal cortex of people with schizophrenia by Yaron Hakak, then at the Genomics Institute of the Novartis Research Foundation in San Diego, revealed that 89 genes were abnormal; remarkably 35 of them are involved in myelination. Presumably these genetic abnormalities upset such processes as synaptic function and myelin insulation, which in turn could disrupt information transmission in the higher-level cognitive circuits affected in psychiatric illnesses.
Roots of Mental Illness
Investigators have set out to learn why glial cells would cause these synaptic snafus. Consider that the biological basis for most mental illness is an imbalance in neurotransmitter chemicals in circuits controlling perception, emotion and thought. All drugs used to treat mental illness and most neurological diseases work by regulating the balance of neurotransmitters. The selective serotonin reuptake inhibitors (SSRIs) used to treat chronic depression and many other psychiatric conditions work by impairing removal of serotonin and dopamine from synapses, allowing these neurotransmitters to build up and in effect boosting the signal. In a similar way, all hallucinogenic drugs, from LSD to PCP, produce their mind-bending effects by altering the levels of neurotransmitters in specific neurological circuits. Regulating neurotransmitter levels at synapses is precisely what astrocytes do.
In theory, then, astrocytes are in a position to control the balance between mental health and madness. In a strange and largely forgotten coincidence, glia were the inspiration for the revolutionary idea that mental illness could have a biological cause and that psychiatric illness could be corrected with medical treatment, albeit a very peculiar one. In the 1930s Hungarian psychopathologist Ladislas von Meduna noticed during autopsies that the number of astrocytes was abnormally low in the cerebral cortex of people who had suffered from chronic depression and schizophrenia. Von Meduna and other pathologists also knew from examination of brain tissue obtained by biopsy that the number of astrocytes increases after epilepsy, presumably to regulate electrical activity when it spins wildly out of control.
Von Meduna observed as well that people with epilepsy rarely suffered schizophrenia. He surmised that a deficiency in astrocytes was the biological reason for schizophrenia and chronic depression. By inducing a seizure in such people, he could correct the imbalance in astrocytes and cure patients suffering from these illnesses. He later wrote in his autobiography: “I published this work in 1932 without knowing that this would become the origin of shock treatment.” How it works is still unclear, but electroshock therapy remains the most effective treatment for chronic depression in people who are not responsive to drugs.
The new awareness of glia in brain function suggests that drugs targeting glia might help treat mental and neurological illnesses. “Epilepsy is a prime candidate for glial-based therapeutics,” says Haydon of Tufts. Recent studies by Haydon, Maiken Nedergaard of the University of Rochester Medical Center, Giorgio Carmignoto of the University of Padua in Italy, and many others are using calcium imaging and electrophysiology to show that when neuronal activity is heightened, glia release neurotransmitters that can either contribute to seizure activity or suppress it. New research also implicates glia in sleep disorders, a component of many mental illnesses. Haydon demonstrated the link in experiments on mice genetically altered to prevent their astrocytes from releasing neurotransmitters, disrupting sleep regulation.
Transformational moments are legendary in scientific history, but it is rare to witness one. Until quite recently, we neuroscientists had dismissed more than half of the brain as uninteresting—a humbling realization. We see only now that the glial and neuronal brains work differently, and it is their intimate association that accounts for the astonishing abilities of the brain. Neurons are elegant cells, the brain’s information specialists. But the workhorses? Those are the glia.