Gene therapy, as its very name implies, aims to repair orreplace a missing or mutated gene that underlies aspecific syndrome or disease.

Schizophrenia is something else again. Its name used tomean "split personality." Now that split is recognized asmultiple rather than merely dual. Hence, its root cause ismore likely to be polygenic than monogenic. Except thatin schizophrenia, things, including gene or genes, areseldom what they seem to be.

For one thing, schizophrenias (they're plural) take manyforms _ from ambulatory to hebephrenic to paranoid toreactive to residual, for example. For another, theirdiagnosis, often equivocal, depends on a multiplicity ofpossible psychopathological signs and symptoms _ fromdelusions and hallucinations, often warrantinghospitalization, to disabling depression, withdrawal andapathy. Alcohol intoxication, narcotic highs and braintumors may masquerade as schizophrenia.

So seeking for the genetic patterns of inheritance thatschizophrenics may, or may not, have in common handspsychiatric geneticists a tougher task than, say,discovering the gene for cystic fibrosis or Huntington'sdisease ever was.

Today's Proceedings of the National Academy ofSciences (PNAS) reports one modest step on the journeytoward a genomic basis for the schizophrenias. Its title:"Schizophrenia susceptibility associated with interstitialdeletions of chromosome 22q11."

Psychiatrist/molecular geneticist Maria Karayiorgou, aschizophrenia-focused senior staff scientist at the FredHutchinson Cancer Research Center in Seattle, is thePNAS paper's first author.

A gene is known by the company it keeps. And it'sdiscovered by the pattern of polymorphic variants(alleles) in its chromosomal neighborhood. SoKarayiorgou and her co-authors took as their point ofdeparture a different disease, velo-cardiofacial syndrome(VCFS) of which the polymorphic configuration wasknown to exist on human chromosome 22.

"We thought this VCFS region was interesting,"Karayiorgou told BioWorld Today, "because we had donestudies showing that a higher percentage of VCFS peoplelinked to chromosome 22 develop schizophrenia."

In the general population, the lifetime risk ofschizophrenia is one in 100.

One in 5,000 babies are born with VCFS, and 78 percentof these are missing small stretches on one of their twoinherited chromosome 22s .

VCFS people suffer cleft palate, heart defects, acharacteristic facial appearance (notably low-set ears),and long, slender fingers and toes. Forty percent of suchchildren are mentally retarded, and 10 percent go on todevelop chronic paranoid schizophrenia in their teens ortwenties.

"We did systematic studies of 18 patients," Karayiorgousaid, "showing that a higher percentage of VCFS peoplewith genomes linked specifically to the long arm's 22q11region _ 30 percent _ do go on to developschizophrenia." ("q" is shorthand for long arm; "11"measures the region's distance along the chromosome.)

"If schizophrenia were a single-gene, totally Mendelian,inherited disease," she explained, "then we would be ableby polymorphic linkage to pinpoint exactly the marker ofinvolvement, and the region on 22q11. But since that wasnot the case," she continued, "the whole 22q could not beexcluded. So we had to look up and down thechromosome for candidate schizophrenia regions."

In the first of two studies reported in PNAS, the co-authors chose at random 100 of 695 firmly diagnosedschizophrenia patients, and discovered in two of themmicrodeletions 1.5 to 2 megabases long in the suspected22q11 region. "The exact risk for schizophreniaassociated with 22q11 deletions," Karayiorgou's paperobserved, "will be determined only after more extendedstudies involving larger numbers of patients and controlsubjects."

Secondly, they examined 18 VCFS patients, of whomfour _ 22 percent _ had schizophrenia _ much higherthan the prevalence of schizophrenia in the generalpopulation." But all 18 averaged 22q11 deletions of thesame sizable length, an overlap between both diseases.

Mutation Analysis Under Way To Pinpont Gene

"That is where things stand," Karayiorgou said. So nowthey are taking two approaches:

"One is to screen additional schizophrenia patients inhopes of finding some with smaller deletions, which willhelp pinpoint and define the region better. The other: Wehave already started doing mutation analysis on genesfrom the region, in order to find the schizophrenia gene."

One candidate gene is long known to inhabit thatchromosome 22 region. It encodes catechol O-methyltransferase, a.k.a. COMT, an enzyme that breaksdown the catecholamines, which include such moleculesof neurologic interest as epinephrine and dopamine.

"Catecholamine build-up," Karayiorgou noted, "has beenconsidered involved in psychoses. Most of the anti-psychotic drugs used in treatment help break them down.So one would suppose," she added, "that if one copy ofthe COMT gene is missing, you don't make as much ofthis enzyme, so you don't have all your natural defensesagainst those catecholamines."

She is "not at all sure" that COMT will prove to be thegene for schizophrenia. n

-- David N. Leff Science Editor

(c) 1997 American Health Consultants. All rights reserved.