Schizophrenia’s Unyielding Mysteries

Last year, when researchers in Cambridge, Mass., announced that they had found a gene strongly linked to a higher risk of schizophrenia, the news media reacted with over-zealous enthusiasm. A “landmark study,” declared both the New York Times and the Washington Post. “Ground-breaking,” trumpeted CNN. Even the Economist dropped its normal reserve: “Genetics throws open a window on a perplexing disorder.”

The hype was somewhat understandable. Historically, schizophrenia research has left a trail of disappointment. The biological basis of the illness, one of the most puzzling and complex mental disorders, has long been an enigma. The toll, however, has always been clear. In the U.S. alone, estimates place the total cost of caring for patients at more than $60 billion a year, a figure that includes both direct health care costs and indirect economic losses from unemployment and early death. Any breakthrough in understanding the causes of the illness would be a major medical advance.

Since the advent of large-scale genetic studies just more than a decade ago, hopes have risen that new insights and therapies were on the way. They are much needed. Existing antipsychotic drugs dampen only the most overt symptoms, such as delusions and hallucinations. They often cause serious side effects and do little or nothing for chronic symptoms such as social withdrawal and cognitive deficits.

On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

But genetic studies have yet to deliver on this promise. Gargantuan gene studies for schizophrenia, as well as depression and obsessive-compulsive and bipolar disorders, have driven home the message that most likely no single gene will lead to new treatments. The study behind last year's exuberant headlines was no exception. If nothing else, though, that research provides an inside look at the immense difficulties in understanding the mental processes that veer off course in schizophrenia.

The 1 Percent

Scientists who study psychiatric disorders had solid reasons to think that genetic clues might help overcome the field's stagnation. Decades of family and twin research suggest a strong genetic component to schizophrenia risk—one underlined by the steady rate at which the disorder occurs. Its prevalence is estimated to be about 1 percent throughout the world, notwithstanding vast environmental and socioeconomic differences across societies. Geneticists also knew that the hunt would not be straightforward. Individual genes powerful enough to generate a high risk of schizophrenia were likely to be very rare in the overall population and thus relevant to only a small percentage of schizophrenia cases. More common genes, on the other hand, would have much smaller effects in triggering schizophrenia and thus be much harder to detect. To find them would require greater statistical power, which would mean working with big sample sizes—tens of thousands of cases and control subjects. Acknowledging the challenges at hand, scientists in 2007 launched the Psychiatric Genomics Consortium (PGC) to study schizophrenia and other mental disorders. At present, the PGC has more than 800 collaborators from 38 countries and samples from more than 900,000 subjects.

Michael O'Donovan, a psychiatric geneticist at Cardiff University in Wales and chair of the PGC's schizophrenia working group, says a global approach was essential to assembling the “truly enormous sample sizes” needed to do the job in what is known as a genome-wide association study (GWAS). A big splash came in July 2014, when the group reported a GWAS involving about 37,000 schizophrenia cases and 113,000 control subjects. The study identified 108 genes (genetic regions) linked to schizophrenia, including a number that code for brain-signaling systems, the main targets for current antipsychotic drugs. These correlations were a sign that researchers might be on the right track.

The genetic region that showed the strongest link to schizophrenia codes for proteins of the major histocompatibility complex (MHC), which is intimately involved in recognizing molecules alien to the body and alerting the immune system. That discovery led Steven McCarroll, a geneticist at the Broad Institute of Harvard University and the Massachusetts Institute of Technology, to think that the MHC region might be a good target for additional study. When McCarroll's team probed further, it turned up a variant of C4, an MHC gene, that elevated schizophrenia risk from about 1 to 1.27 percent in the populations studied.

Although that is a relatively small increase, the researchers suggested in their report in Nature that it could hint at how some cases of schizophrenia arise. The C4 results were important for other reasons as well. Variations in human C4 consist not only of differences in the gene's DNA sequence but also of disparities in its length and how many copies of that gene an individual has.

From previous studies, scientists suspected that relatively rare copy number variations (CNVs) played important roles in schizophrenia—and they continue to debate whether key schizophrenia genes are likely to be uncommon variants that raise risk dramatically or common versions that increase risk only slightly. The new study provided strong confirmation of CNVs' tie to schizophrenia. And when the team compared the brains of both living and deceased schizophrenia patients with those of control subjects, it found that markedly more of the C4 protein was produced in the patients' brains, which was associated with the presence of additional copies of the gene.

To look more closely at what C4 does at the molecular level, the researchers turned to mouse brains. Beth Stevens of the Broad Institute, who spearheaded this part of the study, found that the protein assisted in brain development by “pruning” neural connections, called synapses, when they are no longer needed. Synaptic pruning is a normal part of brain maturation. But if this process is overactive and pares back too many synapses, it could perhaps elucidate some of the features of schizophrenia. It might explain why affected patients tend to have thinner cerebral cortexes and fewer synapses. And schizophrenia, along with other forms of psychosis, is usually first diagnosed in people in their late teens or early adulthood, when brain maturation reaches its final stages.

For some scientists, the finding was a vindication for GWAS as a relatively new way to hunt down disease-associated genes. GWAS has triggered an “amazingly positive and unprecedented explosion of new knowledge” about mental disorders, says Patrick Sullivan, a psychiatric geneticist at the University of North Carolina at Chapel Hill School of Medicine. As for the C4 study, David Goldstein, director of Columbia University's Institute for Genomic Medicine—who has long been a skeptic of GWAS's potential—says that by pointing the way to a possible biological pathway for schizophrenia, the new finding represents “the first time we have gotten what we wanted out of a GWAS.” Others, including some leading geneticists, are less certain, however. “GWAS will have no impact on resolving the biology of schizophrenia,” says Mary-Claire King of the University of Washington, who in 1990 identified BRCA1 as a major risk gene for breast cancer.

In scientific parlance, most cases of schizophrenia appear to be highly “polygenic”—hundreds or perhaps thousands of genes are involved. “GWAS shows that schizophrenia is so highly, radically polygenic that there may well be nothing to find, just a general unspecifiable genetic background,” says Eric Turkheimer, a behavioral geneticist at the University of Virginia.

Indeed, it might be argued that one of GWAS's most important contributions—and the C4 study was no exception—has been to disabuse researchers of simplistic notions about psychiatric genetics. The new findings so far have dashed hopes that schizophrenia can be pinned on just one or even a few genetic mutations. The skepticism stems from the realization that each of the 108 genetic locations linked to schizophrenia so far confers only a tiny risk for the disorder. And the few genes that confer a high risk—in the case of copy number variants and other rare mutations—account for only a small percentage of schizophrenia cases. That makes it less likely that the new findings will lead to therapies anytime soon. It also poses obstacles for neuroscientists and psychiatrists who hoped to find genetic clues for the underlying roots of the disorder. “It would have been way better if there were one single gene,” says Kenneth Kendler, a psychiatric researcher at the Virginia Commonwealth University's School of Medicine. “Then all of our research could have gone into that area.”

In the case of C4, a recognition of these limitations has led to questions about just how relevant the gene will be to understanding schizophrenia or developing new therapies. Whereas about 27 percent of the nearly 29,000 schizophrenia patients in the study had the highest-risk C4 variant, roughly 22 percent of the 36,000 healthy control subjects also carry it, according to McCarroll. “Even if the C4 story is right, it accounts for only a trivial amount of schizophrenia,” says Kenneth Weiss, an evolutionary geneticist at Pennsylvania State University. “How useful that will be is debatable.” And the study does not prove a direct relation between synaptic pruning and schizophrenia, McCarroll and others concede. Its importance seems to lie more in its potential to help pinpoint what kinds of biological pathways might be involved.

Still other problems beset GWAS. To procure huge samples, geneticists usually distinguish between cases and controls depending on whether a person has received a formal schizophrenia diagnosis or not. But the criteria are very broad. In the U.S., the diagnostic rules are dictated by the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders, whereas many psychiatrists in other countries rely on the World Health Organization's International Classification of Diseases. In the criteria set out in both volumes, patients can have markedly different symptoms, ranging from delusions to hallucinations to cognitive defects, and still be diagnosed with a case of schizophrenia.

Hannelore Ehrenreich, a neuroscientist at the Max Planck Institute of Experimental Medicine in Göttingen, Germany, describes schizophrenia as “an umbrella diagnosis” rather than a distinct disease: “We are focusing on people who are on the extreme end of human experience, who are part of a continuum and not a separate category.” William Carpenter, a psychiatrist at the University of Maryland School of Medicine and editor in chief of the flagship journal Schizophrenia Bulletin, does not go that far, but he acknowledges that schizophrenia is a group of disorders or symptoms and not a distinct disease. “That makes it a weak target for gene discovery,” he says.

Goldstein, who thinks the C4 findings “are the best case we've got” for understanding how a schizophrenia risk gene might exert its effects, still calls for researchers to express “a whole lot more humility” about GWAS results. “People working in the schizophrenia genetics field have greatly overinterpreted their results.”

Some of the strongest skepticism about the search for schizophrenia genes comes from psychiatrists, patient advocates and former patients themselves. The GWAS approach focuses on finding new drugs to lessen symptoms of the disorder. But patients often look askance at this goal. “This obsession with symptom reduction does not entirely correspond with the viewpoint of the patients,” says Jim van Os, a psychiatrist at the Maastricht University Medical Center in the Netherlands. Rather, van Os says, patients want to be able to live productive lives and function in society—and doing so does not necessarily correspond with being more medicated.

Van Os and a growing number of patient advocates argue that the term “schizophrenia” itself is part of the problem because it stigmatizes and dehumanizes patients without adequately describing what is wrong with them. Jim Geekie, a clinical psychologist who works at a National Health Service inpatient unit just outside London, says that “knowing somebody's diagnosis tells me next to nothing about them.”

Indeed, a number of countries and regions in Asia, including Japan, South Korea, Hong Kong and Singapore, have eliminated the classification altogether. The Japanese term “mind-split disease,” used to describe a person with schizophrenia, has been changed to “integration disorder,” and a similar term in Korean has been changed to “attunement disorder.”

For many researchers and advocates, the main problem with the nomenclature—and with the gene search itself—is the lingering implication that patients are suffering from a form of brain disease. “If there are genetic variations that mean some people are prone to having these experiences, then we need to make sure people's environments don't switch these things on,” says Jacqui Dillon, chair of the U.K.'s Hearing Voices Network. Dillon, who was told as a young woman that she had schizophrenia and still hears voices today, adds that understanding schizophrenia genetics “doesn't change what we need to do to keep people from going mad.”

A Deep Flaw

Some researchers insist that the search for genes is misguided because it largely ignores the environmental context, as well as the personal and family circumstances, that contributes to schizophrenia risk. “The whole enterprise is deeply flawed,” says University of Liverpool psychologist Richard Bentall. This view is especially strong among clinicians, such as Bentall, who directly treat schizophrenia patients. They argue for increased funding for pragmatic, nonbiological approaches, ranging from family therapy to cognitive-behavioral therapy (CBT).

At times, questions also arise about the fundamental idea, derived largely from family and twin studies, that schizophrenia has a high “heritability.” This term is often assumed, even by many scientists, to mean that genetic factors play a major role. Yet the concept of heritability is complex and not a direct measure of how “genetic” a particular trait—such as a formal schizophrenia diagnosis—actually is [see "Heritability: Missing or Just Hiding?" below].

In fact, environmental and social factors, some researchers insist, confer a greater schizophrenia risk than most genes identified so far. Epidemiological studies have shown that risk factors range from living in an urban environment or being an immigrant to experiencing poverty and emotional and sexual abuse.

Just how such factors contribute to schizophrenia risk is not well understood, aside from speculations that they are sources of emotional stress. Recently, for example, an Israeli team found that Holocaust survivors suffered higher rates of schizophrenia. Another group found increased risk among people who had lived through the violent “Troubles” in Northern Ireland.

There is growing evidence that progress can be made only if researchers consider a spectrum of risk factors. Whereas genetics may make some people more vulnerable to mental disorders, influences from family or a social circle may push a susceptible individual across a threshold that results in a first psychotic episode. The key task is to figure out how genetic and environmental factors interact to produce schizophrenia.

Even diehard gene jockeys admit that environmental influences must be playing some kind of role. “Genes are not destiny,” McCarroll agrees. He points out that when one member of a pair of identical twins is diagnosed with schizophrenia, the other twin is affected by the disorder only about half of the time—a clear indication that nongenetic factors must be important.

Environmental Roots

Frustrations in the hunt for schizophrenia genes have forced the field to reassess how to move forward. Genetics is still considered important to understanding the biological underpinnings of the disorder and coming up with new drugs. But most researchers and clinicians now agree that a broader strategy that supplements genomic approaches is needed, one that builds on expertise gained from experts in sociology, psychotherapy and even prenatal health.

Over the past several years psychologists, psychiatrists, epidemiologists and social workers have accumulated a deeper understanding of the environmental and social factors underlying the disorder. Many new studies are now focusing on “childhood adversity,” an umbrella term that includes sexual, physical and emotional abuse, neglect, bullying, and the loss of one or more parents.

One of the most widely cited of these studies, a meta-analysis by van Os and his colleagues, published in 2012 in Schizophrenia Bulletin, combined results from several studies to increase statistical power and found that patients suffering psychotic symptoms were nearly three times as likely to have been the victims of adversity, far greater than the risk of any gene identified so far in a GWAS. “We need a stronger focus on changing the environment so we can prevent schizophrenia,” says Roar Fosse, a neuroscientist at the Vestre Viken Hospital Trust in Norway. “We need to give children better childhoods and better chances to avoid extreme stress.”

And in a 2014 paper in the Lancet, Ehrenreich and her colleagues demonstrated how studies that combine genetic and environmental data can provide new insights. The team reported on 750 male schizophrenia patients in Germany for whom—unusually—both GWAS and detailed environmental and social risk data were available. The team looked at the age of schizophrenia onset in these patients, a key indicator of how well they are likely to do over the long run: the earlier the age of onset, the worse the eventual outcome. It found that environmental factors, including early brain damage, childhood trauma, living in an urban environment, coming from an immigrant family, and especially cannabis use, were significantly associated with earlier onset. The average age of onset was nearly 10 years earlier for patients who had four or more environmental risk factors than for those who had none. On the other hand, so-called polygenic risk scores calculated from the GWAS data had no detectable effect on age of onset.

Ehrenreich does not interpret these results to mean that genes are irrelevant. It is more likely, she says, that “the genetic factors are so different from one individual to the next that each person has a different reason for having the disorder.” Other researchers, meanwhile, are looking at how environmental stresses, at home or school or through exposure to certain chemicals, might turn genes off and on—a pursuit known as epigenetics.

Ehrenreich and others urge GWAS researchers to begin incorporating environmental data into their studies whenever possible so they can derive a statistical model of how genes and environment interact to make people sick. “It is a shame that researchers neglect assessing environmental information in some of the most expensive and technologically advanced genetic studies,” says Rudolf Uher, a psychiatric researcher at Dalhousie University in Nova Scotia.

Unfortunately, combining epidemiology with genetics may be a tall order. “The cost of gathering environmental data is enormous, and there is considerable disagreement about how to define these environmental variables,” Cardiff's O'Donovan comments. Even so, in 2010 the European Union funded a five-year pilot program to do just that, led by O'Donovan, van Os and others—and researchers have now begun analyzing the data generated.

The big question, of course, is whether the search for genes, even in the context of environmental influences, will eventually lead to new therapies. Most scientists agree that it will take many more years for this research to pay off in new drugs or other interventions. Genetics “has provided the first hard biological leads in understanding schizophrenia,” says Peter Visscher, a geneticist at the University of Queensland in Brisbane, Australia. “It is too early to say whether these discoveries will lead to new therapies, but there is no reason why they could not.” Psychiatric researcher John McGrath, also at Queensland, agrees: “The science is hard, and the brain is hard to understand. But there is no need to throw our hands up in despair.”

Meanwhile, in parallel with the genetic studies, schizophrenia researchers are pursuing numerous other lines of inquiry. They have begun looking for biomarkers—telltale molecules in blood or brain anomalies from neuroimaging that might help them identify people at high risk for the disorder. This could lead to earlier treatment, which numerous studies demonstrate can lead to a better long-term prognosis. Prompted by studies suggesting that the children of women who come down with infectious diseases during pregnancy might be at higher risk for schizophrenia—possibly because of immune responses harmful to the brain of the fetus—other teams are testing anti-inflammatory compounds to see if they might reduce symptoms.

A number of recent clinical trials, meanwhile, suggest that psychosocial therapies, especially CBT, can help lessen both symptoms and suffering in schizophrenia patients. While this research is controversial and the effects are only modest so far, advocates of such approaches are gaining traction in both Europe and the U.S. In the U.K., for example, CBT is now recommended by government health authorities for all first-episode cases of psychosis. “The imbalance in funding between genetic and pharmacological research and psychosocial research needs to be addressed and corrected,” says Brian Koehler, a neuroscientist at New York University who also treats schizophrenia patients in private practice.

The intricacies of schizophrenia mean that comprehensive new treatments are still speculative. Researchers hope that one day brain imaging or other diagnostic tests may help spot a youngster at risk either before or during adolescence. If so, new medications and psychological counseling may be able to delay or prevent a first psychotic break. To achieve that goal, biologists and social scientists must continue to merge their expertise to piece together a composite profile of one of the most complex of all psychiatric illnesses.

Heritability: Missing or Just Hiding?

A concept that seems obvious is not

Researchers have been looking for schizophrenia-related genes for at least 50 years. What makes them think they will find them? The rationale is spelled out in the introduction to nearly every scientific paper on schizophrenia genetics: The disorder has a high heritability. This term is often interpreted as a measure of the relative role played by genes. Heritability is usually expressed as a percentage between 0 and 100 percent.

Scientists have estimated the heritability of schizophrenia using several approaches, including studies of twins. Most estimates hover around 80 percent. Many researchers argue that heritability estimates for schizophrenia can be very misleading, however. They question key suppositions, including the so-called equal environment assumption (EEA), which considers both identical and fraternal twins to be subject to the same environmental influences.

“These basic assumptions are wrong,” says Roar Fosse, a neuroscientist at the Vestre Viken Hospital Trust in Norway, who led a recent critical assessment of the EEA. But twin researchers have mounted a vigorous defense of the approach. “I don't think it's likely that current heritability numbers are substantially overestimated,” says Kenneth Kendler, a psychiatrist at the Virginia Commonwealth University's School of Medicine.

Some researchers have an even more profound critique of heritability. They argue that the technical calculations of the term do not account for the relative role of genes and environment. Heritability, rather, measures only how much the variation of a trait in a particular population—whether height, IQ or being diagnosed with schizophrenia—reflects genetic differences in that group.

As an example of how misleading heritability estimates can be, Eric Turkheimer, a geneticist at the University of Virginia, points to the human trait of having two arms. Nearly everyone in a given population has two of them, and there is normally no difference in the number of arms between identical twins—who share nearly 100 percent of their DNA sequence—and fraternal twins, who are assumed to share 50 percent of their genes on average. Thus, when heritability for arm number is calculated using standard heritability equations, it comes out to 0. And yet we know that having two arms is almost entirely genetically determined.

Figuring out what heritability for schizophrenia actually means is key, researchers say, because even the most high-powered genetic studies have identified only about a third of the predicted genetic component. Will this so-called missing heritability eventually show up in more sophisticated studies—or will it turn out that genes are not playing as big a role as heritability estimates have long predicted? The jury is still out. —M.B.