The first time Kees van Heeringen met Valerie, the 16-year-old girl had just jumped from a bridge. It was the 1980s and van Heeringen was working as a trainee psychiatrist at the physical rehabilitation unit at Ghent University Hospital in Belgium. As he got to know Valerie, who’d lost both legs in the jump and spent several months at the hospital, he pieced together the events leading up to the moment the teenager tried to end her life, including stressful interactions with people around her and a steady accumulation of depression symptoms.
Van Heeringen, who would later describe the experience in his 2018 book The Neuroscience of Suicidal Behavior, says Valerie’s story left a permanent impression on him. “I found it very difficult to understand,” he tells The Scientist. He asked himself why anyone would do “such a horrible thing,” he recalls. “It was the first stimulus for me to start studying suicidal behavior.”
In 1996, van Heeringen founded the Ghent University Unit for Suicide Research. He’s been its director ever since, helping to drive scientific research into the many questions he and others have about suicide. Many of the answers remain as elusive as they seemed that day in the rehabilitation unit. Suicide rates are currently climbing in the US and many other countries, and suicide is now the second leading cause of death among young people globally, after traffic accidents. The World Health Organization recently estimated that, worldwide, one person ends their own life every 40 seconds.
Suicide is as complicated as it is tragic. Suicidal behaviors come in many varieties, ranging from suicidal thinking, or ideation, to suicide attempt and completion, all of which may be associated with various levels of violence or intent. The behaviors themselves differ in incidence among genders, ethnicities, and other demographic categories, and almost always occur against a background of depression or some other mood disorder—although only a fraction of people with mood disorders become suicidal.
No field of scientific inquiry can single-handedly untangle a phenomenon as complex as suicide. But van Heeringen and many other scientists are hoping to shed light on the problem by digging into the neurobiological processes underlying thoughts about ending one’s own life and attempts to do so. This work is building support for the idea that suicide is tied to specific biochemical changes that can be measured and targeted independently of, and possibly in parallel with, the mental health disorders they often accompany. Findings from this work, researchers hope, could help reveal new treatments, and perhaps even opportunities to identify the people most at risk in time to intervene.
“The knowledge we have today is way larger than what we had twenty years ago,” says Gustavo Turecki, a psychiatrist at McGill University and the director of the McGill Group for Suicide Studies at the Douglas Research Centre in Montreal. “[We’ve] made tremendous advances . . . in terms of understanding the complexity of the problem, understanding the neurobiology, understanding the causes.”
NEUROBIOLOGICAL PATHWAYS LINKED TO SUICIDE RISK
Scientists have identified several key neurobiological pathways with ties to suicidal behaviors. Research in the field addresses only a fraction of the complexity of this serious public health problem, and the literature on the topic is complicated by variation in study design, but the clues point to several interacting moderators of suicide risk. Three of the systems best-studied in relation to suicide are depicted below.
© LISA CLARK
The role of the brain’s stress pathways in suicide
Valerie’s account shared elements with the stories of many other people who have attempted to end their lives. She showed signs of depression and social stress, and, as van Heeringen later discovered, she had a family history of suicide—a known risk factor for suicidal behaviors, independent of any psychiatric disorders.
Scientists now think about suicide risk in terms of stress-diathesis models, which treat suicide as a product of both so-called precipitating factors such as elevated stress or mood disorders and predisposing factors—the “diathesis”—such as family history, particular genetic variants, or early-life adversity such as abuse or neglect. “Suicide is more than . . . being very depressed,” explains Columbia University’s John Mann, a psychiatrist and translational neuroscientist who helped develop the conceptual framework with Columbia neurobiologist Victoria Arango.
This framework has helped focus research on biochemical pathways that regulate the brain’s response to stress, and how those pathways could be altered in people who become suicidal. The brain has multiple stress responses, but the best-studied in relation to suicide is the hypothalamic-pituitary-adrenal (HPA) axis, which controls the release of the stress hormone cortisol and is known to be upregulated in clinical depression.
Early clues regarding the link between the HPA axis and suicide include findings of higher concentrations of corticotropin-releasing hormone (CRH), which triggers the synthesis of cortisol and other glucocorticoids involved in stress signaling, in postmortem brain samples from people who died by suicide than in samples from people who died by other means. Other research has hinted that people who died by suicide have enlarged adrenal glands—sites of cortisol production. Due to the high incidence of depression and other mood disorders among people who end their own lives, however, studies such as these didn’t attempt to determine whether the observed effects were specific to suicide or to mood disorders more generally.
More recently, the case for a central role for the HPA axis in suicide has gained support from work led by Turecki and others revealing that early-life adversity, one of the strongest risk factors for suicide even when psychiatric disorders are controlled for, can have long-term effects on HPA axis function. In the mid-2000s, Turecki teamed up with McGill University geneticist Moshe Szyf, who had shown that rats neglected by their mothers exhibit altered epigenomes in the hippocampus—a brain region involved in stress, learning, and memory—and dysfunctional HPA responses to stress. In the hippocampi of people who have died by suicide and had a history of childhood abuse, Turecki, Szyf, and their colleagues found evidence of hypermethylation and reduced expression of the gene coding for NR3C1, a glucocorticoid receptor that helps dampen cortisol signaling, compared with healthy controls or people who died by suicide but hadn’t experienced abuse.
Research since then has linked suicidal behaviors to methylation abnormalities in other HPA-related genes. One 2018 assessment of nearly 90 people who had attempted suicide identified reduced methylation at the CRH gene in blood samples from some of the study’s subjects—specifically, those who made attempts that were more violent or more likely to result in death. And several studies have identified hypermethylation and reduced expression of SKA2, which codes for a protein that interacts with NR3C1, in people who died by suicide compared with healthy controls and with nonsuicidal patients with depression, schizophrenia, or other psychiatric disorders.
The relationship between the HPA axis and suicidal behavior is complicated. For example, while some studies imply that the HPA axis overreacts to stress in people who die by suicide, others indicate that people who attempt suicide have lower baseline cortisol levels and/or blunted HPA reactivity to stress compared with controls. “It is a confusing literature,” says Nadine Melhem, a psychiatric genetic epidemiologist at the University of Pittsburgh School of Medicine who found a few years ago that, among around 200 people whose parents had mood disorders, those who attempted suicide had overall lower HPA activity. “Almost every [possible] finding has been reported.”
Part of this inconsistency likely stems from small study samples and variations in experimental design, Melhem notes. But variability may also come from differences in the drivers of suicidal behavior in different groups of people. Mann’s group reported last year that, of 35 people who attempted suicide, only those who scored highly for impulsive aggression in personality tests had significantly elevated cortisol responses to stress compared with nonsuicidal controls. And one meta-analysis published a few years ago found a positive correlation between cortisol levels and risk of suicide attempt in studies of people under 40 years old, but a negative correlation in studies of older people.
Until now, “we have not been able in the literature to capture the dynamic nature of these pathways in relation to suicide risk,” says Melhem, adding that she and colleagues are beginning to build longitudinal datasets to address this gap. “It needs a lot more work.”
Many studies have linked suicidal behaviors to dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis and other mediators of the body’s responses to stress.
Corticotropin-releasing hormone (CRH) has been found in higher concentrations in the brains of people who die by suicide.
People who die by suicide, and particularly those who die by violent means, may have enlarged adrenal glands.
Basal cortisol levels have been found to be both higher and lower than normal in people who have attempted suicide. The reactivity of cortisol to stress may also be dysfunctional in people with suicidal behaviors.
NR3C1, also known as the glucocorticoid receptor, may be in lower abundance in people who die by suicide, particularly those with a history of childhood abuse.
The effect of serotonin and other neurotransmitters
Mann first became interested in the neurobiology of suicide while studying a rather different aspect of brain chemistry. Throughout the 1980s and ’90s, he and others found deficits in serotonin (5-hydroxytryptamine, or 5-HT) signaling and in the neurotransmitter’s main metabolite, 5-hydroxyindoleacetic acid (5-HIAA), in the brains of people who died by suicide, regardless of psychiatric diagnosis, compared with brains of people with or without psychiatric disorders who had died by other means. The findings were key in the realization that there might be biochemical changes specific to suicide, Mann says. Since then, the serotonergic system has become one of several neurotransmitter systems being probed for clues about suicidality.
Like the HPA axis, serotonin signaling appears to be modulated by early-life adversity. For example, methylation of HTR2A, which codes for a serotonin receptor known as 5-HT2A, is altered in children who have suffered early-life adversity—although it’s not yet clear how those methylation changes affect HTR2A expression. A 2016 study of twins in the UK revealed that children who were bullied had hypermethylation at SERT—a gene coding for a protein that helps transport serotonin from the synapse back into the presynaptic neuron—compared with children who weren’t. Bullied children also showed blunted cortisol responses to stress, hinting at a link between the serotonergic system and HPA functioning.
How such physiological changes might influence suicidal behavior remains to be seen, but groups such as Mann’s are working to disentangle some of the details. For example, he and his colleagues recently published a more concrete link between serotonin and HPA-axis activity: even when psychiatric diagnoses are controlled for, levels of the serotonin receptor 5-HT1A are correlated with cortisol reactivity to stress. The team has also explored levels of serotonin receptors in depressed and nondepressed people exhibiting suicidal behaviors, and found that levels of 5-HT1A in some regions of the cortex are higher in people who attempt or die by suicide, regardless of psychiatric diagnosis, than in controls.
Somewhat counterintuitively, higher levels of 5-HT1A could contribute to a deficit in serotonin signaling, Mann explains, because the receptor is part of a neural feedback response that inhibits further serotonin release into synapses. Accordingly, it seems that in people who are suicidal, “the problem is not the capacity to make serotonin but . . . the capacity to use that serotonin,” he says. This role for 5-HT1A could also help explain why selective serotonin-reuptake inhibitors (SSRIs) do a better job of dampening suicidal thoughts and behaviors than some other antidepressants, he adds: among other effects, SSRIs reduce the number and responsiveness of 5-HT1A receptors and may thereby quiet the negative feedback loop that suppresses serotonin signaling.
Neurotransmitters besides serotonin, including glutamate, GABA, and dopamine, have also been investigated in the context of suicidal behavior—particularly following recent findings that drugs such as ketamine and esketamine, which interact with the glutamate receptor NMDAR, reduce suicide risk in patients with clinical depression. However, the literature on these neuro-transmitters is fairly inconsistent, spurring researchers to continue looking for new mechanisms to explain suicidal behaviors.
Neural communication via serotonin and other neurotransmitters such as glutamate often shows signs of dysregulation in people who die by suicide.
© LISA CLARK
Disruption of serotonin signaling has repeatedly been found in the brains of people who die by suicide.
Levels of the serotonin transporter SERT, which shuttles serotonin back into the presynaptic neuron, may be lower in people who die by suicide.
Levels of the serotonin receptors 5-HT1A and 5-HT2A may be higher in people who attempt or die by suicide.
A potential link to neuroinflammation
A couple of years ago, researchers in Denmark reported a link between suicide and infectious disease. Analyzing three decades’ worth of health records from more than 7 million people, the team found that being hospitalized with an infection was associated with more than a 40 percent greater probability of suicide. Spending more than three months in the hospital was linked to a more than doubled suicide incidence. While acknowledging that such observational data can’t demonstrate causation, the team calculated that the statistical risk associated with hospitalization for infections could account for about 10 percent of the Denmark’s suicides.
There are many possible explanations for this finding—one being that treatment of infections with antibiotics or other hospital medications influences mental health. But van Heeringen and others point out that the study ties into another hypothesis about suicidal behavior, one that involves a role for inflammation.
Elevated suicide risk has previously been reported in people with autoimmune disorders and traumatic brain injury—conditions that, like infections, typically involve inflammation. Further clues come from epidemiological studies of Toxoplasma gondii—a parasite that causes chronic, low-level neuroinflammation in humans. A 2018 study of nearly 300 people in Korea found that 14 percent of people who attempted suicide tested positive for the parasite, compared to just 6 percent of healthy controls—mirroring a correlation found in several US cohorts. Together, the findings paint a compelling picture that neuroinflammation “is part of the story,” says Melhem.
While depression is not thought of as an inflammatory disease, signs of neuroinflammation in the brain have been repeatedly documented in people who suffer from depression, and a number of anti-inflammatory drugs show antidepressant effects. Microglia, the central nervous system’s primary immune cells and mediators of inflammation, tend to show increased activation in the brains of people who die by suicide, Melhem adds, and several studies have identified elevated concentrations of inflammatory cytokines such as interleukins IL-2, IL-6, and IL-8 in people with fatal and nonfatal suicidal behaviors. One 2019 analysis of nearly 2,000 Mexican-Americans, for example, found that blood levels of IL-8 were elevated in depressed and nondepressed women who had attempted suicide.
How exactly neuroinflammation might contribute to suicidal behavior is still unclear, and some recent epidemiological studies have raised doubts about whether the association exists independently from depression. One route that researchers are exploring is neuroinflammation's interaction with the serotonergic system. In a process thought to be mediated by microglia, neuroinflammation triggers a shift in the metabolism of serotonin’s molecular precursor, tryptophan, away from the production of serotonin and towards other chemical pathways—potentially reducing serotonin signaling and triggering other suicide-related changes in the brain.
That’s just one hypothesis, says Melhem, who recently won a grant with Mann to investigate the suicide-inflammation link. These pathways haven’t “been interrogated enough,” Melhem says. “We’ll be looking more into that in future work.”
People who die by suicide show signs of increased inflammation in the brain while epidemiological data reveal that some inflammation-related health conditions are associated with higher suicide risk.
© LISA CLARK
The brains of people who die by suicide show higher levels of microglia activation.
Blood levels of inflammatory cytokines, particularly some types of interleukins, have been found at higher levels in people who attempt suicide.
Tools to predict and prevent suicide
One of the defining moments in psychiatrist David Brent’s career happened during his medical residency some 40 years ago. Brent had been assigned to work with young people admitted for intentional drug overdoses at the University of Pittsburgh Medical Center Children’s Hospital. He had to determine who should be referred to a psychiatric ward and who could safely go home. “I found that I really didn’t have a very good way of making that determination,” says Brent, now a professor at Pitt. As he learned more about how other clinicians made such decisions, “I realized I was in good company—that nobody really knew what they were doing.”
It’s still a dilemma facing anyone attempting to provide care for people at risk of suicide. Today’s clinicians often rely on patients to report their intentions in order to decide on appropriate interventions. But the approach has limitations. One 2019 meta-analysis of studies on suicidal ideation found that around 60 percent of people who ended their lives had denied having suicidal thoughts when asked by a clinician or doctor in the weeks or months before their death.
This problem has led some researchers to look for ways to translate findings from neurobiology, however preliminary, into the identification of biomarkers to predict the onset of suicidal behaviors. Given its strong association with suicide, the HPA axis has long been a focus of this work, and there’s some evidence that abnormal cortisol levels—higher or lower than normal—in blood or saliva could hold promise as a biomarker. A few months ago, for example, Melhem, Brent, and colleagues published findings from a long-term study of teenagers that suggested a person’s baseline cortisol levels might be used to predict future suicidal thinking, with higher cortisol associated with increased ideation within the next couple years.
Cortisol tests may help provide predictive power to other measures of suicidality, such as questionnaires about social and academic stress. One recent analysis showed that while survey data were good predictors of who among 220 teenage girls with mental health concerns would be thinking about suicide within the next few months, they were poor predictors of who would attempt suicide during that period. But when the researchers focused only on girls who had shown blunted cortisol responses in lab tests, the questionnaire data predicted suicide attempts much better.
Looking beyond stress responses, other groups have attempted to identify biomarkers related to neurotransmission. A few years ago, Mann’s group used positron emission tomographic (PET) imaging to assess levels of 5-HT1A serotonin receptors in the midbrains of 100 patients with major depressive disorder. The scientists found that higher 5-HT1A levels predicted greater suicidal ideation and more-lethal suicidal behavior within the next two years. Last summer, a team led by Yale University neuropsychologist Irina Esterlis reported that levels of glutamate receptor mGluR5, as measured by PET, was linked to current suicidal ideation in patients with post-traumatic stress disorder—though the results didn’t hold for patients with major depressive disorder.
Opinions differ among researchers about the potential of such biochemical signatures to assess suicide risk. Greg Ordway, a pharmacologist studying depression at East Tennessee State University, says that while biology might identify people predisposed to suicidal behavior, it’s unlikely to produce one or a handful of biomarkers that reliably reveal whether a person is about to end their life. “Suicide is extremely difficult to predict,” he says. “People are always trying to do it—people like me are looking for markers. But in reality, I don’t think we’ll probably ever find that.”
Some of the most promising tools for assessing immediate risk might instead come from other areas of neuroscience that measure more-complex, emotional signals in the brain as opposed to biochemical signatures. In 2017, Brent, along with Carnegie Mellon University neuroscientist Marcel Just and other colleagues, used functional MRI to image the brains of 34 people as they contemplated words such as “death,” “trouble,” and “carefree.” Using machine learning algorithms to process the data, the team could distinguish between people who were thinking about suicide, as self-reported during the study, and those who weren’t with 91 percent accuracy. Among those who were, the team identified people who’d already attempted suicide with 94 percent accuracy.
The researchers recently received $3.8 million from the National Institute of Mental Health to scale up the project and are planning long-term monitoring of people with and without various types of mood disorders. As part of the study, the researchers hope to extend their tool to identify people who might attempt suicide in the future, not just those who are thinking about it at the time of the scan or who have attempted it in the past. Just tells The Scientist that the team also plans to adapt the technique to a cheaper, more clinic-friendly technology than MRI, such as electroencephalography (EEG).
Melhem says she’s hopeful that combining techniques will improve predictive approaches in the coming years. In 2019, she and colleagues published a model that improved on the accuracy and performance of existing models to predict suicide attempts based on factors such as the severity and variability of a person’s depression symptoms over time. Integrating this sort of easy-to-collect clinical data with biological information from brain scans or other diagnostic tests should lead to more-accurate predictions, she says.
The search for such tests has important consequences for suicide prevention even beyond their potential to assess risk. “When we introduce biological markers, just like [for] any other area of medicine, then stigma will be reduced at the level of the patient,” Melhem says. Patients are often surprised to hear that researchers are studying the biology underlying suicide “because they’ve been thinking that this is a behavioral flaw in their character, and they feel guilty about it. That’s part of the stigma that we want to break.”
Medical professionals consider suicide a preventable public health problem. In the US, agencies such as the Centers for Disease Control and Prevention and the Substance Abuse and Mental Health Services Administration oversee initiatives designed to help assess and respond to suicide risk in the general population, and particularly in communities considered to be at high risk, including among people with mood disorders, substance abuse problems, or a family history of suicide.
Many nonprofit organizations also work to raise awareness of the problem, fund research on suicide, and provide resources for people affected by suicide. Find information about suicide warning signs, treatment, and other resources at the American Foundation for Suicide Prevention website. For help, call the confidential, free 24/7 National Suicide Prevention Lifeline at 1-800-273-8255.
Catherine Offord is an associate editor at The Scientist. Email her at email@example.com.