Divided Selves

To get to the bottom of schizophrenia, two scientific rivals are seeking help from some unusual twins--twins who are identical in all respects but one.

By Tony Dajer|Tuesday, September 01, 1992
Fuller Torrey has one goal in life: to prove Irving Gottesman wrong. Gottesman, for his part, is just as determined to nail Torrey. And either would be delighted to see the other win.

Torrey, 55, senior psychiatrist at St. Elizabeths Hospital in Washington, D.C., is a world expert on the viral theory of schizophrenia. Gottesman, 61, Commonwealth Professor of Psychology at the University of Virginia, is one of the best gene sleuths around. The question at hand: whether the baffling illness of schizophrenia springs primarily from an infection or is caused by defective genes. To settle the issue, the friendly archrivals have embarked on an unusual venture designed to leave one of them in the dust. Irving, says Torrey, is a lively and honest researcher with whom I can disagree with pleasure.

The whole enterprise hinges on a medical quirk. Schizophrenia is a common disease, affecting over a million people in this country alone. Given such large numbers it’s possible to find occasional cases of people with schizophrenia who have an identical twin, a sibling who shares their exact genetic heritage. About half the time the twin is also schizophrenic, but the rest of the time the twin is normal (though some display borderline schizophrenic traits that label them as slightly eccentric). These discordant twin pairs--one ill, one well--constitute a potentially powerful means to tease out schizophrenia’s secrets.

Despite the mental devastation it creates, schizophrenia leaves maddeningly few traces. You can’t point to a definite cause like a virus or a bacterium or a defective gene, as you can with many other brain diseases. Nor can you see glaring damage like holes or scars when you autopsy a schizophrenic’s brain. And although researchers have tried scanning the brains of hundreds of schizophrenics and healthy volunteers, they’ve been unsure if the differences they saw were due to schizophrenia or to individual brain variation.

That’s why twins are so inordinately useful for this kind of study. If you could compare a schizophrenic with his or her genetically identical yet normal twin, any differences would very likely be due to the disease process. It would be like superimposing, in the same person, the cardboard cutout for disease on the one for health. If you found discrepancies between the two, you could conclude that that’s where schizophrenia probably lurks and search for its cause.

That, in short, was Torrey’s reasoning. So in 1986, after seeking out Gottesman’s involvement as a respected counterpoint to my own bias, he assembled a sample of willing twins and a network of psychologists, geneticists, virologists, biochemists, statisticians, and brain scanners to launch an unprecedented assault on the roots of madness.

The first descriptions of what a modern psychiatrist would call schizophrenia were written in 1809, but it wasn’t until a century later that Eugen Bleuler, a Swiss psychiatrist, gave the disease a name. Schizophrenia literally means split mind, but popular myth notwithstanding, it has nothing to do with multiple personalities. Bleuler was referring to an odd ungluing of the mind, to the kind of striking dissociation of reason and emotion that makes patients laugh during funerals or imbue a mundane object or gesture with some spectacularly inappropriate significance. Schizophrenics, who are typically diagnosed when they are adolescents or young adults, become prey to fantastic hallucinations and hear voices conversing in their head. They may develop paranoid or grandiose delusions--a fear that the CIA can control or read their thoughts, for example, or the conviction that their destiny is to fulfill some exalted messianic mission. Yet they can also become intensely withdrawn, apparently unfeeling, or overcome by an apathetic stupor. So profoundly does schizophrenia unhinge the mind that many victims never make their way back to reality.

Torrey vividly remembers his first encounter with the illness. He was 19 and a premed student at the time. It was the summer before Rhoda, my seventeen-year-old sister, was supposed to start college, he recounts. She began having delusions that the British redcoats were attacking America. My mother told me she would discuss the Revolutionary War at dinner, the kind of stuff you learn in history class. At first my mother thought Rhoda was kidding. But then one day she found her lying on the front lawn talking to imaginary voices about the British attacks. When I got home from college, my sister looked physically, neurologically sick to me. In just weeks she’d gone from normalcy--the sister I’d grown up with-- to full-blown psychosis.

By the end of medical school Torrey had decided to specialize in psychiatry. After completing his residency at Stanford Medical School and doing a stint at the National Institute of Mental Health, he was put in charge of a ward at St. Elizabeths. There the young psychiatrist quickly found himself immersed in the mystery of his patients’ insanity.

A major influence on him at the time was the physician and virologist Carleton Gajdusek. In 1972 Gajdusek jolted the medical world with his discovery that so-called slow viruses could linger in the brain for 20 years or more before causing symptoms. The classic disease of this kind was kuru, which started as clumsiness and ended as mind-obliterating dementia and which afflicted only tribesmen in the New Guinea highlands who ate the brains of the deceased during their funeral rituals. To contract kuru, Gajdusek found, you had to consume the brain of someone already infected. Gajdusek (who won the 1976 Nobel Prize for medicine) proved his point by injecting infected brain tissue into chimpanzees: not only did they get the disease, but their damaged brains developed the same peculiar Swiss-cheese appearance as those of human kuru victims. When New Guinea highlanders stopped eating brains, Torrey says, they stopped getting kuru.

As a psychiatrist, though, Torrey was swimming against the tide of his profession by even considering biological explanations for mental illness. The rise of psychoanalysis at the turn of the century--and a paucity of biological findings--had given psychological explanations like upbringing and bad parenting a stranglehold on the debate over the causes of schizophrenia.

When I met Gajdusek, he recalls, the first thing he did was blast me: You psychiatrists have gotten so hung up on Freud, you’ve forgotten how to be scientists! And he was right; we had stopped treating schizophrenia like a physical, measurable disease. Goaded by Gajdusek, Torrey began acting like a microbiologist. He asked permission from patients to perform spinal taps and analyze their cerebrospinal fluid, the fluid that bathes the brain and spinal cord, to look for the footprints of viral infection. He didn’t find much, but back then, no one did. (Gajdusek tried injecting chimps with brain tissue taken from autopsied schizophrenics, but the experiment failed to work.) When I look back, says Torrey, I shudder to think how primitive our methods were. In the 1970s viral research was barely in its infancy.

In the meantime, however, the advocates of a biological explanation were getting reinforcements from a very different quarter. Although it was known that schizophrenia can run in families, the blame had usually been ascribed to nurture (the home environment) rather than to nature (the patients’ genes). But in the late 1960s and early 1970s studies began to show that genes far outweighed upbringing as a risk factor for schizophrenia. Children of schizophrenics adopted by normal families, for example, had the same risk of developing the illness as children raised by schizophrenic parents. What’s more, an identical twin of a schizophrenic was four times as likely to develop the illness as a nonidentical twin.

One of the young researchers who was helping kick the door open for biology was Gottesman. In 1971 he had joined a pioneering team of Danish geneticists that was studying the children of identical twins who were discordant (one ill, one well) for schizophrenia. After following the families for 18 years, Gottesman confirmed the startling finding that the risk of schizophrenia in the children of either twin was exactly the same: 17 percent. That meant that even if schizogenes were not activated in one generation, they could be passed on to the next and then make mischief.

But while studies like these showed that schizophrenia had to have a genetic component, they left a thorny question in their wake: If schizophrenia were due only to genes, why didn’t 100 percent of the identical twins--not the observed 50 percent--share the disease? The most likely explanation, ventures Gottesman, is that the right combination of genes--probably four or five--plus some as yet undefined environmental stressors must be thrown together to trigger schizophrenia. But before we can figure out what activates the genes in the twins who become ill, we must first find the genes themselves. The trick now is to hunt those genes down to their chromosomes and map them--a trick that he hopes some of Torrey’s twins will help him pull off.

While Gottesman was making a name for himself studying Danish twins, Torrey continued to collect schizophrenics’ cerebrospinal fluid and blood in pursuit of a schizovirus. It was a monumental wild-goose chase, but a number of clues sustained him in his belief.

For example, viral infections of the temporal lobes--notably herpes simplex type I, the common cold-sore virus--can produce hallucinations and bizarre behavior bearing an uncanny resemblance to schizophrenia. In fact, physicians often mistake the one for the other. Moreover, unlike any other mental illness, schizophrenia is more common among those born in winter, when viral infections abound. In a Scandinavian study of children with a strong family history of schizophrenia, an increase of 70 percent in the rate of schizophrenia was found among those whose mothers had contracted influenza during the second trimester of pregnancy.

Yet despite such circumstantial evidence, Torrey’s cerebrospinal- fluid and blood analyses failed to turn up solid viral suspects. And even if his schizovirus existed, those tests couldn’t tell him where in the brain it might be doing its work. To identify the virus he would have to locate its base of operations.

A possible approach turned up in the 1980s with the introduction of a brain-scanning technique called magnetic resonance imaging (MRI). Compared with existing technologies such as CT scans, the new tool produced dazzling brain pictures with an astonishing amount of anatomical detail. Even so, it wasn’t initially all that helpful for schizophrenia. It found variations in schizophrenics’ brains--but they were subtle and not peculiar enough to schizophrenia to distinguish it from other brain diseases or even from normal variation. Studies showed, for example, that the ventricles, a pair of fluid-filled structures that curl around the brain’s inner pith like ram’s horns, were often unusually large in schizophrenics. But enlarged ventricles were also seen in Alzheimer’s and Parkinson’s patients, and even in normal old people.

By 1986 it was clear to Torrey that only identical twins who were discordant for schizophrenia could show up the small discrepancies he needed to flush out his prey. Through the National Alliance for the Mentally Ill, he gathered twins that fit the bill. So far he has found 30 pairs of clearly discordant twins--and, for comparison, 30 more pairs who are either both schizophrenic, both normal, or somewhere in between (one schizophrenic twin and one ostensibly normal twin with schizophrenic tendencies).

The studies haven’t been easy to do, however. MRI scans are obtained by submitting the brain to strong magnetic fields and then measuring the signals from the different tissues inside it. Taking these images requires patients to lie without moving a muscle inside a dark, clanking, tubelike chamber, a process that terrifies even normal patients who have a touch of claustrophobia. For patients whose grip on reality is already fragile, the procedure required immense courage. With one schizophrenic twin, Torrey recalls, I had to promise to buy her a skirt and blouse if she went through the whole thing. I held her hand the whole time. She was very brave. Altogether, 15 pairs of discordant twins were examined for Torrey’s initial MRI study.

The results, published in March 1990 in the New England Journal of Medicine, caused quite a stir. Like earlier MRI studies, this one showed that the schizophrenic twins had enlarged ventricles, but it also revealed a striking change in a crucial brain structure called the hippocampus. The hippocampus (the name derives from the Greek for sea horse, which it’s said to resemble) clings to the inner surface of the temporal lobes, which are behind the temples, on either side of the brain. The hippocampus is apparently where input from the senses is hammered into new memories and where the components of old memories are reassembled for recall. And lo and behold, in Torrey’s schizophrenics, the hippocampus, especially the left half of the hippocampus, was noticeably smaller than in the normal twins. Indeed, his scanning team thinks that the ventricles may become enlarged in schizophrenics because of a loss of surrounding tissue, including shrinkage of the hippocampus.

That evidence seemed to tie in nicely with autopsy studies begun in the mid-1980s that showed signs of cell loss and disarray in the left hippocampus and its anatomic neighbors in the limbic system. The limbic system controls our emotional response--another function thrown out of whack by schizophrenia. And it fit with another finding: that schizophrenics did poorly on certain memory tests, suggesting that impaired memory was perhaps a component of schizophrenia.

Classically, schizophrenia was considered a disease of the frontal lobes, the seat of abstract, higher thought. (This idea had unfortunate repercussions: in the 1940s, it served as the rationale for the frontal lobotomies that were performed on thousands of schizophrenics.) But the left hippocampus, linchpin of memory, is part of the left temporal lobe. Could schizophrenia be a problem of the left temporal lobe instead?

Another newer school of thought implicated the entire left side of the brain--the hemisphere that generates language and thus defines the interpretation of words and symbols. One thing that makes the world so terrifying for schizophrenics is the destruction of those defining limits. Thus a misplaced coffee cup can become imbued with peculiar meaning; a stranger’s gesture can signal the arrival of the redcoats or the CIA. The result is a paralyzing paranoia.

Torrey, for his part, suspected that the temporal lobe was to blame. But for him the hallmark of schizophrenia is hearing voices. No other symptom, he argues, is as specific to schizophrenia: 75 percent of all patients hear voices--voices that command you to kill yourself, voices from outer space, two voices carrying on a conversation, even the voice of God. To us scientists, they seem to be saying, Pay attention, there may be a big clue here.

The temporal lobe, it turns out, is home not only to the hippocampus and other limbic structures but to the nerve fibers that carry input from the ears. If an infection caused the limbic-system damage seen in the autopsy studies, Torrey argues, then maybe it could damage these fibers as well, and one might hear voices. Autopsies showing cell disarray in these areas, he says, appear to fit with what we’ve found on our MRIs.

Still, good as the MRIs have been at locating abnormalities in the brain, they could say nothing about when the damage occurred. That’s the question Stefan Bracha, a psychiatrist at the University of Arkansas Medical School in Little Rock, set out to answer. One of schizophrenia’s peculiarities is its predictable age of onset: 18 for men, 23 for women, on average. Does that mean that young adults are more susceptible than other age groups to certain viruses or genetic malfunctions? Or is schizophrenia a delayed reaction to damage that occurred years before? (Kuru, recall, took up to 20 years to manifest itself in New Guinea highlanders who had eaten infected brains.)

Pathological studies of schizophrenic brains have never shown signs of the scarring one would expect from viral infections. However, in the special case of damage caused to a fetus in its mother’s womb, the brain doesn’t form scar tissue. Instead it ends up with just the kind of cell disorganization described in the autopsies. At the very least, then, these autopsy findings were consistent with the idea that the damage had occurred prenatally. Adding weight to the notion was the Scandinavian study suggesting that influenza infection in the second trimester of pregnancy-- months four, five, and six--boosted the risk of schizophrenia.

Bracha knew that the time-tested way of looking for evidence of prenatal infection was dermatoglyphics, the study of fingerprints and finger structure. But it was old technology, so no one else wanted to do it, he recalls wryly. He also knew that the second trimester of pregnancy is a time of major brain-cell reorganization. If a virus hits at that particular time, it might also affect the hands, which are simultaneously undergoing finishing touches.

Unlikely as it sounds, it is possible for a virus to infect one identical twin fetus and not the other. So Bracha examined 24 pairs of Torrey’s twins for the odd fingerprint whorls and stunted digits that are the fossilized evidence, as he puts it, of infections inside the womb. His study, published in November 1991, found that schizophrenics had four times as many abnormalities as their healthy twins, who showed Bracha what the schizophrenics’ hands should have looked like.

These divergent fingerprints suggested that the seeds of schizophrenia are indeed sown very early in life. In the same vein, Torrey has recently completed another study, which uses family interviews and school records to show that the twins start diverging before the age of five. By putting a ceiling on the time of infection, he can focus his search on agents that act in infancy or before.

Meanwhile, back in the geneticists’ camp, Gottesman is aware that Torrey has stacked the deck a bit by focusing on clearly discordant twin pairs. This group most likely has a low genetic predisposition to schizophrenia, requiring a big environmental jolt to make the twins so different. But Torrey, remember, has also recruited less clear-cut pairs, where one twin is schizophrenic and the other, apparently healthy one is only mildly schizoid. From Gottesman’s point of view, these twins may prove the most useful of all.

In these pairs, the healthy twin has some, but not all, of the symptoms of schizophrenia, he says. So we assume they carry some predisposing genes for the disease, even if they’re not fully turned on. Moreover, these healthy twins often display subtle physical signs like abnormal eye tracking or easy distractibility. If we’re lucky, these abnormalities may lead us to DNA ‘markers,’ which are like visible flags inherited along with the actual disease genes. Then we can look for each marker to see if it turns up consistently among families of schizophrenics. If the markers and genes are truly linked--that is, are situated close by on the same stretch of DNA--we may be able to track down some of the genes we think are involved.

In practice, this approach works well for single-gene diseases such as Huntington’s chorea. But when four or five genes are involved, as is thought to be the case with schizophrenia, the task explodes into mind- numbing complexity. Yet if the genes can be found, the rewards will be enormous: researchers will be able to tell very quickly what proteins they make and, eventually, what the proteins do (or fail to do) to cause schizophrenia.

While Gottesman pursues the genes that make the schizophrenic brain malfunction, Torrey is beginning to look at how and where it malfunctions. The procedures he and his colleagues are using can literally see the brain in action. Cerebral blood flow studies and PET (positron emission tomography) scans offer a color-coded glimpse into the brain’s workings as eerie as MRI’s sharp dissection of its living anatomy. For the cerebral blood flow studies, xenon gas is inhaled into the lungs to make the patient’s blood briefly radioactive; the harder the cells in a particular brain region work, the more blood flow they get and the more detectable radiation they emit. PET uses radioactive glucose or oxygen to similarly light up areas of hot metabolic activity in the brain.

Initially Torrey was reluctant to subject his twins to PET scans. The scans are tough. I take all the tests the patients do, and I found this one uncomfortable. They have to lie in a ring of sensors with their head pinned by a form-fitting mask and IV lines in their arms. And besides, you’re asking to read the thoughts of someone who’s already paranoid about his thoughts being read. What changed his mind was a study by Susan Resnick, another of his extended network of researchers, at the University of Pennsylvania. PET scans of seven schizophrenics and their normal twins showed a consistent difference in the basal ganglia, acorn-size cell clusters lying beneath the ventricles at the brain’s center. The ganglia are action integrators--if you’re a catcher, they help you get the mitt between the fastball and your groin--but in seven out of seven schizophrenics (in contrast to their well twins) the basal ganglia mysteriously lit up, even when they were resting.

Using such scans, Torrey’s Washington team made another intriguing discovery. When they challenged twins to sort playing cards by suit, number, or color, they found that the schizophrenics not only did worse than the normals but also failed to activate a region--technically known as the dorsolateral prefrontal cortex, or DLPFC--within the frontal lobe. This region, says Daniel Weinberger, a psychiatrist collaborating with Torrey, is critical to performing complicated tasks and thinking well- ordered thoughts: It’s highly evolved and serves perhaps more than anything else as a hallmark of the human brain. What’s more, this year, by comparing the blood-flow scans with MRI data, the team showed that poor DLPFC activation correlated with small hippocampus size. For the first time in the history of schizophrenia research, a functional deficit in one brain area was linked to an anatomical defect in another.

That seems to imply, says Torrey, that schizophrenia is not simply a disorder of particular areas in the brain, but a breakdown--most likely on the brain’s left side--in the connections between them. And it so happens that the DLPFC, the hippocampus, the emotion-regulating limbic system, and the basal ganglia are all part of what’s known as the brain’s dopamine network. Dopamine is a neurotransmitter, and almost all the drugs for treating schizophrenia’s symptoms are now known to block dopamine receptors. In the dopamine network, too little activity in one area could lead to overactivity in another. A DLPFC that’s too quiet, Weinberger speculates, may disinhibit the limbic system and lead to the florid, inappropriate emotionality seen in many schizophrenics. In other words, if the Speaker of the House falls asleep, overemotional congressmen may soon get out of control.

Of course, that still doesn’t settle the question that launched Torrey and Gottesman’s original bet: Are genes or a virus the main actor in schizophrenia? Although there’s evidence for both, neither researcher is ready to call it a draw. The point is, says Gottesman, to what degree is each important? Could schizophrenia occur without a viral insult? Could there even be nongenetic cases?

We must each push our theory as far as it will go, Gottesman insists. If we don’t try our darnedest to prove the other wrong, we’ll never prove anything right. Is schizophrenia nine-tenths genetic, or one- twentieth? That’s what we need to find out.
Comment on this article