Molecules on the Mind

By Jason Socrates Bardi

O, let me not be mad, not mad, sweet heaven
Keep me in temper: I would not be mad!

——William Shakespeare, King Lear, 1605

The irony of schizophrenia and bipolar disorder is that we understand how to treat these diseases a lot better than we understand how they actually work.

We do know that they are both chronic and often devastating brain disorders that affect a significant portion of the population. According to the National Institute of Mental Health, more than two million American adults have bipolar disorder in any given year, and approximately one percent of our population will suffer from schizophrenia in their lifetimes.

We know that schizophrenia is a very heterogeneous and complicated disorder involving many genes and manifesting itself through many different symptoms. There are whole bodies of literature describing the symptoms people with these diseases suffer, and there are established guidelines for diagnosing both conditions. And while we cannot treat either disease perfectly, there are numerous anti-psychotics available for treating schizophrenia—some of which have been around since the mid-1950s. Likewise, there are numerous anti-depressants and mood stabilizing drugs available for treating bipolar disorder.

Therapeutic approaches that combine various drugs with individual and group therapy have helped countless people lead lives free from the sometimes-debilitating symptoms of schizophrenia and bipolar disorder.

And yet what is missing is a truly reductionist understanding—a molecular understanding—of schizophrenia and bipolar disorder.

We do not know what specific genes predispose people for these disorders. We do not understand what environmental factors determine why identical twins often, but not always, have the same illness. We do not know the factors that determine whether a schizophrenic patient will have mild symptoms or full-blown hallucinations. We do not know why some people with bipolar disorder suffer shifts between depression and mania a few times a year while others "cycle" more rapidly—several times a week.

Nor do we have a blood test, a brain scan, or some easy way to identify whether somebody is afflicted with schizophrenia or bipolar disorder. Nor can we predict, based on somebody's symptoms, whether they will be among the 10 percent of people with schizophrenia who end their lives in suicide.

Towards a Molecular Understanding

These are all questions that motivate scientists at The Scripps Research Institute (TSRI) who study the molecular aspects of diseases like schizophrenia and bipolar disorder so that they can be better diagnosed and treated.

"Everybody here," says Professor J. Gregor Sutcliffe, speaking of his laboratory in the Department of Molecular Biology, "has some interest in the molecular biological aspects of central nervous system problems."

His own interest in such problems goes back to when he first began to collaborate with TSRI Department of Neuropharmacology Chair Floyd Bloom, who was then at the Salk Institute for Biological Studies. Sutcliffe and Bloom decided to make cDNA clones of various RNAs in the brain and then characterize the proteins the RNAs made by using recombinant DNA technology to identify unknown gene products expressed selectively in the brain. This was accomplished by isolating thousands of mRNAs from brain tissue and cloning and sequencing the ones that were not detectable in the liver or kidney.

After identifying some 20,000 to 40,000 sequences in the mid-1980's, Sutcliffe decided to concentrate his efforts on only a few of the genes that were expressed. In particular, he was interested in identifying and characterizing RNAs that are expressed in limited fashion—in small numbers or in selected brain tissues—and that has been his guiding principle for research ever since.

"We thought that if we could find RNAs that were highly selective in their expression to particular regions," he says, "then those would give us more insight into the functionality of those particular regions."

This approach is not unique—it is a way of looking at nature that is common to all fields of science. And despite how much "reductionism" has today come to resemble a four-letter word, this is exactly what is needed to elucidate the molecular basis of diseases like schizophrenia and bipolar disorder.

The problem is that our definitions and diagnoses of schizophrenia and bipolar disorder depend upon the symptoms, but the symptoms can range from mild to severe in both diseases.

"These symptoms may actually represent distinct subtypes within themselves," says Assistant Professor Elizabeth Thomas, who works closely with Sutcliffe.

And to even further complicate the situation, the two diseases can be so similar to each other that someone with one can be misdiagnosed as having the other. For instance, the flat emotional affect and social withdrawal that are common among schizophrenics are similar to symptoms that bipolar people display during depressive episodes. Likewise, psychotic behaviors also manifest in both schizophrenic and bipolar patients.

"What we're [closing in on] in the genomic era is what genes are distinguishing these two disorders," says Thomas, "If we can identify which genes are making up the sub-types of these disorders, it would be a huge breakthrough in understanding. Hopefully, this would help us find cures that would be specifically tailored towards one [disorder] or the other."

Withdrawals from the Brain Bank

In a recent study, Sutcliffe and Thomas compared the expression levels of one protein in post-mortem tissue samples taken from various brain regions in schizophrenic and bipolar subjects to control subjects matched for age and other variables.

All of the tissue samples came from a "brain bank" maintained by Thomas and Sutcliffe's collaborators at the Rebecca Cooper Research Laboratories of the Mental Health Institute in Parkville, Victoria, Australia. This brain bank maintains the brains in freezers and has records of the patients so that tissue samples can be well-matched with controls for such variables as the age and sex of the patients.

The recent study was actually the third in a series of studies by Thomas and Sutcliffe in which they addressed the question of which genes are involved in these diseases.

In the first study, they looked at the effect of "chronic" dosing of the schizophrenia drug clozapine on gene expression in the brain. The brain responds to a large dose of the drug by increasing levels of expression of particular genes. One thing they noticed right away from this study is that the levels of expression of a gene coding for a 29,000 Dalton protein called apolipoprotein D (apoD) increased. This led them to hypothesize that the protein plays an important role in the pathology of the disease.

In their second study, they concentrated on the elevated expression levels of apoD in one particular region in the brain, the prefrontal cortex, which is the key brain structure involved in a range of cognitive functions, such as planning, judgement, decision making, and the mediation of working memory. Impairment of all of these cognitive functions is a characteristic of both schizophrenia and bipolar disorder, and the prefrontal cortex was previously known to be involved in both of these disorders.

The significance of this second study is that it correlated elevated apoD expression levels in the prefrontal cortex with schizophrenia and bipolar disorder. While the prefrontal cortex was already known to be involved, the apoD gene is one that perhaps people would not have looked at before. And while the physiological role apoD is playing in these psychiatric disorders is not yet clear, its presence in these areas of the brain implicates its involvement.

"The elevated expression of apoD," says Sutcliffe, simply, "is a signature of the pathology."

Guilt by Association

In their latest study, to be published this month in the journal Molecular Psychiatry, Sutcliffe, Thomas, and their colleagues looked at apoD expression in several other brain regions of people with schizophrenia or bipolar disorder as compared to the well-matched control brains. They examined 12 regions of the brain and looked at the apoD expression in these regions to gauge the differences.

Some of the brain regions Sutcliffe, Thomas, and their colleagues studied were unaffected and some were very affected—showing dramatic increases in apoD expression in the disease brains as compared to the controls.

Significantly, some regions had highly overlapping expression of apoD in the brains of patients with schizophrenia and bipolar disorder—the prefrontal cortex, for instance—which suggests that pathological similarities between the diseases may account for some of their symptomatic similarities.

But the real strength of the work is that it teased out some previously unknown differences between the two disorders. "There are regions that distinguish between the two," says Thomas.

For example, in the schizophrenic patients, the researchers found increases of apoD expression in the amygdala, a small brain region associated with certain types of emotional behavior. In the bipolar patients, there was no such increase.

However, in the bipolar patients, they found several cortical regions that were expressing higher levels of apoD than in the schizophrenic patients, which suggests that the cyclic nature of bipolar disorder could be due to a cortical imbalance.

These results do not give a complete picture of the mechanisms of the two diseases, answer all the unanswered questions as to how they differ, nor identify all the genes involved, but the data do give solid physiological evidence that the two diseases differ at the cellular and molecular level.

The study also demonstrates the power of high-throughput genomic methods to address questions about the fundamental nature of diseases like bipolar disorder and schizophrenia.

The TOGA® Technology

For these studies, Sutcliffe and Thomas used a PCR-based method called total gene expression analysis (TOGA®) that Sutcliffe invented a few years ago. TOGA® is currently licensed to San Diego biotech company Digital Gene Technologies, which analyzed the samples in their fully-automated Torrey Pines facility.

The technology basically divides all the RNA in a tissue sample into 256 pools, and accounts for all the RNA in each pool by using polymerase chain reaction (PCR) to amplify them.

First the mRNA in a sample is purified and then an enzyme is used to create "complimentary" cDNA from the RNA strips, which is necessary in order to do the PCR.

The cDNAs are then primed—molecules are added that anneal to the "polyA" repeating track of A nucleotides at the 5' end (the beginning) of the cDNA. Also at this end is attached a biotin fragment, which is like a piece of molecular velcro that allows the cDNA to be fished out later.

The primed cDNAs are then cut with enzymes that recognize four specific nucleotide bases, and the pieces of cDNA with the biotin attached are fished out and separated according to the sequence of four nucleotides adjacent to the cleavage site. This sounds complicated, but the basic thing to keep in mind is that this allows the RNA to be divided into 256 pools (4* 4*4* 4 = 256), and identified individually.

Each pool of RNA is then amplified with PCR and the PCR products are then subjected to capillary electrophoresis, a technique that essentially separates the pieces based on their length—the length from the 3' cleavage site to the poly A tail—and detects them through their fluorescence.

When each lot of cDNA passes by the laser in the capillary electrophoresis apparatus, a "peak" of fluorescence emission is detected. The timing of this peak appears on the length of the original RNA, and it is actually predictable. By counting the number of bases between the poly A tail and the cleavage site, and by taking into account the 4 bases adjacent to the cleavage site, it is possible to know where to expect it.

"For every RNA of known sequence, we know which of the 256 pools will contain that RNA and how long the product will be," says Sutcliffe.

So when the computers collect an array of bands, these data can be compared to a list of known sequences of RNA, and candidate genes can be assigned to them. This is all done automatically.

The technique, then, takes a piece of tissue and returns a set of data representing which genes are being expressed in the tissue. Rather than looking for one gene in particular, the computer provides a range of genes that are active.

The Hypocretins and Narcolepsy

This technology has proven invaluable in other studies that have originated from Sutcliffe's laboratory. A few years ago, he and his colleagues found two excitatory neuropeptides expressed by only about 3,000 neurons in the hypothalamus, the brain center that governs most aspects of autonomic regulation—such as aspects of energy metabolism, cardiovascular function, hormone homeostasis, and sleep-wake behaviors.

These two peptides, now called the hypocretins, are expressed in neurons with connections to many parts of the brain, from the cerebral cortex to the base of the spinal cord. Electron microscopy studies showed these neurons packaging the hypocretins in vesicles, and the vesicles accumulating at the synapses, so Sutcliffe and his colleagues arrived at the hypothesis that the hypocretins are neurotransmitters—neurons fire action potentials as a result of the release of the peptides, and these action potentials cause humans and other animals to wake up.

The name hypocretin is a shortened name for hypothalamus peptide with a sequence that is related to secretin. It is a name that is actually attached to two separate, closely-related peptides that are concentrated in an area of the hypothalamus that is implicated in arousal, feeding, blood pressure, and the release of hormones. Not surprisingly, the hypocretins are important modulators of all of these—especially the sleep/wake cycle.

In fact, the hypocretins are responsible for narcolepsy. Narcoleptics suffer hallucinations, loss of muscle control, and, most notably, frequent sleep "attacks" throughout the day, even if they are fully rested.

Narcolepsy- is a disease that results from not having hypocretins or hypocretin receptors. Animals with either hypocretin or the hypocretin receptor knocked out display signs of narcolepsy, and humans with narcolepsy have no detectable hypocretins in their cerebrospinal fluid—there is normally a fixed amount. The cause of this loss, developed in adolescence, is not well understood

What is known is that narcolepsy is one of the most common neurodegenerative disorders—affecting some 250,000 Americans according to the National Institute of Neurological Disorders and Stroke. And by continuing to identify, describe, and study the hypocretins that cause the disease, Sutcliffe hopes to elucidate the physiological mechanism of narcolepsy and, hopefully, contribute research that will lead to better treatments.


Go back to News & Views Index

Professor J. Gregor Sutcliffe (right) and Assistant Professor Elizabeth Thomas, both investigators in the Department of Molecular Biology, are interested in teasing apart the differences between schizophrenia and biopolar disorder. Photo by Jason S. Bardi.