ABOVE: Modified from © istock.com, Viki, yodiyim

Susannah Cahalan was 24 years old when her world turned upside down.

Cahalan was living a busy life as a news reporter at the New York Post when she suddenly began experiencing sensitivity to light, numbness in her limbs, and an unsettling feeling that something was not quite right in her body and her brain. One day at work, she found herself inexplicably going from crying hysterically to skipping giddily down a hall. After a seizure landed her in the hospital, her condition rapidly worsened. She started having delusions and hallucinations, believing that her father was a murderer, that she was being secretly recorded, and that she could age people using her mind. In a matter of weeks, walking, speaking, and swallowing became difficult. She eventually became immobile and unresponsive, lying in her hospital bed in a catatonic state. 

 Despite her worsening condition, dozens of specialists from various fields—psychiatry, neurology, internal medicine—couldn’t figure out what was wrong. Numerous blood tests and brain scans failed to generate answers. To many who saw her, Cahalan’s condition looked indistinguishable from mental illnesses such as bipolar disorder or schizophrenia, in which people can experience delusions and hallucinations that make it difficult for them to distinguish what’s real and what’s not. It wasn’t until a neurologist asked Cahalan to draw a clock that the problem became clear. Cahalan had drawn all the numbers on just one side of the clock face, indicating that there was a problem in the functioning of one half of her brain. 

A brain biopsy confirmed what the doctor had suspected. Cahalan had anti-NMDAR encephalitis, a rare autoimmune disease in which the body produces antibodies that attack the NMDA receptor, a protein found throughout the brain. The condition had only been discovered in the early 2000s, just a few years prior to Cahalan’s
diagnosis, by neurologist Josep Dalmau, then at the University of Pennsylvania. This diagnosis was much-needed good news for sufferers of the mysterious condition—their disease was treatable. After receiving immunotherapy, Cahalan was able to fully recover.

Cases like Cahalan’s, which she details in her memoir Brain on Fire, are rare. The estimated occurrence of anti-NMDAR encephalitis is just 1.5 cases per 1 million people each year. Despite the disease’s rarity, the discovery of anti-NMDAR encephalitis was a game changer. Not only did it lead researchers to uncover more than a dozen other autoimmune diseases of the brain, but the uncanny similarities between the psychotic symptoms of autoimmune encephalitis and schizophrenia generated a flurry of interest among scientists seeking an answer to an age-old question: What role does the immune system play in schizophrenia? 

There were records of influenza pandemics whereby people acutely affected with influenza acted in ways consistent with what we now understand to be psychosis.

—Thomas Pollak, King’s College London

The neurological underpinnings of schizophrenia have long eluded scientists. In some instances, people diagnosed with schizophrenia were later found to have other ailments, including anti-NMDAR encephalitis. The majority of schizophrenia patients still await answers. Evidence from many lines of research, including epidemiology and genetics, point to the possibility that in at least a subset of schizophrenia cases, the immune system is the culprit. Many research groups continue to hunt for clues, and as part of this search, several are assessing whether therapies that target the immune system might also benefit those with schizophrenia and other psychiatric illnesses. 

Some researchers remain skeptical that the immune system plays a meaningful part in mental illnesses that are not clear-cut cases of autoimmunity. But others are optimistic. “I’m hopeful that this is a new era for schizophrenia research,” says Belinda Lennox, a professor of psychiatry at the University of Oxford in the UK. “I think getting people excited and interested in the immune system in schizophrenia can only be for the benefit of patients, because we’re so desperately in need of new treatments and for a better understanding of the causes.” 

The infection connection

Following large outbreaks of influenza in the 1800s and 1900s, including the 1918 flu pandemic, physicians took note of an uptick in patients exhibiting mental disturbances, particularly delusions and hallucinations. Such cases were so common that medical professionals at the time even coined the term “psychoses of influenza” to describe this condition. “There were records of influenza pandemics whereby people acutely affected with influenza acted in ways consistent with what we now understand to be psychosis,” says Thomas Pollak, a neuropsychiatrist and clinical lecturer at King’s College London in the UK. 

This type of infection-triggered psychosis was not unique to influenza. Many other instances of psychiatric symptoms developing after encountering a pathogen have been documented throughout history. Neurosyphilis, for example, occurs when the bacterium Treponema pallidum invades the central nervous system, causing psychosis and depression, among other schizophrenia-like symptoms. When the link with the syphilis-causing microbe was established, some mental institutions reported that the symptoms in around a third of their patients could be attributed to the infectious disease. Other examples include Toxoplasma gondii, a parasite that commonly infects cats: anti-Toxoplasma antibodies have been observed in the blood of some patients, suggesting that exposure to the parasite might trigger schizophrenia. And during the COVID-19 pandemic, cases of psychosis following a SARS-CoV-2 infection among people with no history of mental illness have emerged all around the globe. 

Some data suggest that maternal exposure to infections may heighten the risk of offspring developing conditions such as schizophrenia. Numerous studies have found that children who are born in the winter or spring—when infectious agents tend to circulate—are at a greater risk for schizophrenia and other mental illnesses than those born during another season. There are also reports of maternal influenza infections increasing the risk of psychosis, although this link has not been consistently found. 

More-recent evidence has also emerged from large epidemiological studies in Scandinavia, where researchers have access to nationwide registries containing medical data and other personal information. A series of studies using such data in Denmark from a group led by Michael Benros, a professor of immunopsychiatry at the University of Copenhagen, revealed that as an individual’s number of infections increased, the subsequent risk of developing a schizophrenia spectrum disorder—a cluster of mental illnesses that share similar features, including psychosis—increased in a dose-dependent manner. The researchers found that more severe infections, such as those requiring hospitalization, led to a higher risk of developing schizophrenia later in life. These assessments also revealed that there was a higher prevalence of autoimmune diseases among people with schizophrenia, and vice versa. 

Despite the centuries-old link between infections and psychosis, researchers’ interest in this connection waned until recently, according to Benros. One of the reasons for this, he explains, was that many scientists believed the brain to be immune privileged, with the blood-brain barrier keeping it safe from the collateral damage sometimes caused by the body’s defensive maneuvers. Over the last decade, however, several findings have overturned this notion. Researchers have recognized that the blood-brain barrier can become leaky, compromising its ability to act as an impermeable shield, and that the brain possesses resident immune cells, microglia, as well as its own lymphatic system, a network of vessels that shuttles immune cells around the body. Large-scale genomic studies in people with schizophrenia turned up genetic sequences associated with major histocompatibility complex (MHC) proteins, molecules on the surface of immune cells that present pathogen-derived peptides, as key regions implicated in the illness.

“A lot of research—and particularly the discovery of anti-NMDAR antibodies—brought back the interest into this field,” Benros says. 

Autoimmune Underpinnings of Psychosis 

One of the most well-known autoantibodies with targets in the brain is the anti-NMDAR antibody, which targets the NMDA receptor (NMDAR) that is found on excitatory neurons in the brain. When present, this autoantibody prompts neurons to engulf NMDARs and reduces these receptors’ numbers at the synapse. This dearth of NMDARs, in turn, causes problems in synaptic transmission that underlie a range of neuropsychiatric symptoms such as hallucinations, delusions, seizures, and movement abnormalities. Researchers have pinpointed more than two dozen other brain-targeting antibodies, most of which are found in patients with autoimmune disease of the central nervous system. The role these antibodies play in psychiatric illnesses such as schizophrenia is the subject of active investigation. 

Illustration comparing healthy synapse and anti-NMDAR disease
See full infographic: WEB | PDF

Searching for signs

Though tantalizing, the many studies reporting a link between infection and neuropsychiatric symptoms like psychosis merely illustrate an association, not a causative relationship, between the immune system and schizophrenia. The discovery of anti-NMDAR encephalitis, however, provided evidence that the body’s own antibodies could attack the brain and trigger symptoms that resembled those of schizophrenia. Not only that, the findings fit nicely with one of the leading hypotheses of schizophrenia: that the illness is caused by problems in the synapses of glutamate neurons, where NMDA receptors are plentiful. 

The “groundbreaking discovery” was finding antibodies aimed at neuronal cell surface targets that caused a catastrophic but treatable brain illness, Lennox says. “That was the dramatic breakthrough, that you could remove antibodies from the brain relatively easily and reverse a clinical presentation—treat people and actually quite often completely get rid of the underlying problem.” For Lennox, what was striking was how patients with autoimmune encephalitis would often present symptoms nearly identical, at least in the early stages of their illness, to those she was seeing with first-episode psychosis—new-onset psychotic symptoms that, in some cases, mark the beginning of schizophrenia. This got her thinking: What if the first-episode patients also had antibodies? And could giving immune-modulating treatments benefit these patients, too? “That was sort of the start of my research journey, and I thought it would be quick and straightforward,” Lennox notes. “But of course, it hasn’t been.” 

In the years since the discovery of anti-NMDAR encephalitis, scientists have discovered self-directed antibodies, or autoantibodies, against more than a dozen other neuronal proteins. Like anti-NMDAR antibodies, many of these antibodies can cause confusion and hallucinations, among other neuropsychiatric symptoms. But as research groups, including Lennox’s, have begun looking more broadly for neuronal autoantibodies in patients with first-episode psychosis, schizophrenia, or other psychiatric disorders, they’ve generated highly variable results. 

Researchers doing this work typically test a patient’s blood for such antibodies, then conduct further lab experiments to see how strongly those antibodies bind to neuronal proteins such as NMDAR. To diagnose an autoimmune encephalitis, clinicians use a spinal tap to take cerebrospinal fluid (CSF), which unlike blood is in direct contact with the brain. But spinal taps are not always available—largely because they are not often conducted in psychiatric clinics—so many researchers use blood as a proxy. One 2014 meta-analysis by Pollak and his colleagues found that across seven studies with 1,441 patients, approximately 8 percent of people with first-episode psychosis had neuronal autoantibodies in their blood; the prevalence rates were mixed, however—some of the studies had found no antibodies at all. Researchers have also found such autoantibodies in the blood of healthy people, raising the question of whether the presence of these molecules has any clinical significance.

I’m hopeful that this is a new era for schizophrenia research.

—Belinda Lennox, University of Oxford

Some, including Dalmau, now at the August Pi i Sunyer Biomedical Research Institute (IDIBAPS) in Spain, are unconvinced that antibodies against NMDAR or other neuronal targets found in the blood are responsible for symptoms in primary psychiatric disease—such antibodies must be found in the cerebrospinal fluid to definitively diagnose them with an autoimmune encephalitis, he says. According to Dalmau, many of these studies have been plagued by methodological problems, such as mixing patients with different disorders, a lack of control groups with healthy individuals, and an absence of confirmatory results using other methods. If you look at the literature over the past 10 to 15 years for clinically meaningful immune markers in people who only have a psychiatric disorder, Dalmau says, “there is no evidence whatsoever.” 

Pilar Martinez Martinez, a professor of neuroinflammation in neuropsychiatric disorders at Maastricht University in the Netherlands, also notes that based on her research, it appears that people with autoimmune diseases such as anti-NMDAR encephalitis usually show neurological symptoms such as seizures or movement abnormalities in addition to psychiatric symptoms such as hallucinations and delusions. “I think in the unlikely scenario that there are only psychiatric symptoms, these are very, very, very rare cases, which I have not been able to identify myself,” Martinez Martinez says. However, she adds, “I do think that in the future we might be able to identify other antibodies that may [cause] symptoms that at the moment we cannot separate from [psychiatric disorders].”

Other groups have investigated immune system dysfunction by focusing on molecules such as cytokines, which, unlike antibodies, don’t have a specific target. Rather, they function as part of more-generic immune responses like inflammation. Sophie Erhardt, a professor of experimental psychiatry at the Karolinska Institutet in Sweden, says that the benefit of targeting cytokines is that they provide a “snapshot of what the immune activation is right now.” The downside, however, is that it’s not easy to interpret why, exactly, cytokines are elevated, Erhardt adds, as levels of these molecules can rise in response to various insults, from sleep deprivation to infection. 

The search continues, with researchers conducting experiments and analyses that could uncover more-robust evidence of the immune system’s contributions to schizophrenia and other psychiatric illnesses. Benros’s team, for example, is currently conducting a large study in Denmark that is looking in the CSF, blood, and microbiome for signs of immune dysfunction in patients with a new diagnosis of a psychotic disorder that are not present in healthy individuals. Currently, only about 1 percent of people with psychosis or other psychiatric symptoms have clear neuroimmunological alterations detectable in the CSF. “We’re trying to see if we can broaden the percentage of people for whom we can say there’s a direct immune cause to psychiatric symptoms,” Benros says. 

Cytokine Contributions to Psychosis 

Some researchers have found higher-than-normal levels of cytokines, chemical messengers of the immune system, in people with psychosis. The mechanism through which these molecules might contribute to psychiatric symptoms remains an open question, but at least two potential pathways have been proposed: one in which cytokines act via connections with the peripheral nervous system (A) and another in which the molecules enter the brain by penetrating the blood-brain barrier (B)

Illustration comparing neural pathway and the humoral pathway
© lisa clark

A) Neural pathway 

Cytokines in the body’s periphery may activate the vagus nerve, a large, multi-branched cluster of neurons that carries signals from the brain to various organs and vice versa. This may, in turn, trigger immune cells and chemical messengers in the brain that alter neurotransmission.

B) Humoral pathway

Inflammation may cause the blood-brain barrier to become “leaky” and allow immune cells and molecules such as cytokines to enter the surrounding brain tissue.

See full infographic: WEB | PDF

Seeking solutions 

Several clinical trials have tested whether various immunomodulatory drugs can improve the symptoms of schizophrenia. In one recently published study, researchers reported that patients who received methotrexate, a powerful immunosuppressive drug used for chemotherapy, exhibited a slight reduction in symptoms such as hallucinations and delusions compared to those who received a placebo. Most other trials, however, have found little or no clear benefit or such treatments. 

The trouble with the studies that have been conducted to date, Lennox says, is that they did not select patients who show signs of immune dysfunction. The immune system may only be involved in a subset of patients with schizophrenia or other psychosis-related disorders, she explains, so it is unlikely that everyone with these conditions would see improvements on immune-suppressing drugs. To address this limitation, Lennox and her colleagues are conducting a double-blind, placebo-controlled trial to test whether an immunotherapy—intravenously injected antibodies from healthy donors combined with rituximab, an antibody currently used to treat autoimmune diseases such as arthritis—can benefit patients with psychosis who have detectable brain-targeting antibodies in their blood. 

For now, both skeptics and proponents of the idea that the immune system may be involved in schizophrenia agree on at least one thing: clinicians seeing patients with psychotic symptoms need to more vigilant about identifying possible auto-immune causes of their illness. Some, like Ludger Tebartz van Elst, a professor and psychiatrist at the University of Freiberg and its associated hospital in Germany, are now administering spinal taps to all patients with psychotic symptoms. In most countries, however, such a practice is difficult to implement, Tebartz van Elst notes, largely due to the fact that most psychiatrists don’t conduct spinal taps. (Germany is unusual in that psychiatrists receive a year of neurology training, and neurologists of psychiatry training, making the two practices much more closely intertwined than in other parts of the world.) 

For this reason, Tebartz von Elst, Pollak, and several of their colleagues recently published a paper outlining several red flags that clinicians could use to help diagnose their patients when a spinal tap is not an option. These suggestions were met with some criticism, including from Martinez Martinez, who noted that there was no evidence that patients without autoantibodies in their CSF would benefit from immunotherapy. On top of that, Dalmau notes that his team attempted to validate these red flags in a large cohort of patients with first episode psychosis and found that those criteria were unable to identify individuals with anti-NMDAR encephalitis.

Pollak notes that a spinal tap should be sought whenever a patient has these red flags—and that there could be downsides to screening every single patient with psychosis for signs of auto-immunity, such as uncovering incidental findings that might lead to misdiagnoses. “De-diagnosing someone is a very difficult thing to do,” Pollak says. “For whatever reason, we’re in this slightly unusual position, where psychiatric diagnoses are so stigmatized . . . that in some cases an inflammatory diagnosis feels more real.” Ultimately, he adds, “until investigation in psychiatry is given parity with investigation of physical health problems, this is going to be an ongoing problem.”