Psychiatric Disorders and Mutations at the WolframSyndrome Locus

Michael Swift and Ronnie Gorman Swift

Identifying genetic loci at which mutations predisposeindividuals to common psychiatric illnesses will havemajor impact on the diagnosis and treatment of mentalillness. The available evidence indicates that mutations atthe Wolfram syndrome locus contribute substantially tothe prevalence of psychiatric illness in the generalpopulation.

Patients with mutations at this locus on both parentalchromosomes, called Wolfram syndrome homozygotes,have a distinctive and rare autosomal recessive syndromecharacterized by juvenile onset diabetes mellitus andbilateral progressive optic atrophy. Diverse and seriouspsychiatric manifestations frequently have been observedin Wolfram syndrome patients; however, the populationburden of mental illness attributable to mutations at thislocus is almost entirely from individuals who carry asingle mutation, called Wolfram syndrome heterozygotes,who have no distinguishing physical characteristics butconstitute approximately 1% of the population.

Molecular genotyping of blood relatives of Wolfram syn-drome patients has shown that Wolfram syndrome het-erozygotes are 26-fold more likely than noncarriers tohave a psychiatric hospitalization. Severe depression wasthe predominant finding in the test group studied. Theprediction that approximately 25% of all patients hospi-talized for depression are Wolfram syndrome heterozy-gotes now can be tested by mutation screening of hospi-talized patients from the general population. Many otherbehavioral and cognitive difficulties also have been ob-served in Wolfram syndrome families. For each specificpsychiatric abnormality, a “test group” of blood relativeswithin Wolfram syndrome families with that abnormalitycan be formed. By comparing the number of Wolframsyndrome heterozygotes found in each test group bymolecular genotyping with the number expected under thenull hypothesis, the index-test method can determine

which clinical phenotypes result from mutations at theWolfram syndrome locus. This method can be utilized toidentify other loci at which mutations predispose individ-uals to psychiatric illnesses.Biol Psychiatry 2000;47:787–793 ©2000 Society of Biological Psychiatry

Key Words: Wolfram syndrome, genetic predisposition,heterozygotes, psychiatric illness, depression, suicide

Introduction

Many improvements in the diagnosis, prevention, andtreatment of mental illness will result from identi-

fying the specific genes involved in common psychiatricdisorders. A gene is said to predispose to psychiatricillness when mutations of that gene often lead to suchillness. The challenge is to identify the most importantpredisposing loci because more than 10,000 differentgenes are thought to determine the structure and functionof the brain.

The Wolfram syndrome locus became a “candidatelocus” for common psychiatric disorders through clinicalobservations of families and patients with this autosomalrecessive syndrome, which is defined by the occurrence ofjuvenile onset diabetes mellitus and bilateral progressiveoptic atrophy (Kinsley et al 1995; Wolfram 1938). Theassociation between mutations at that locus and suchdisorders was established through systematic studies(Swift RG et al 1990, 1991) and then confirmed andquantified using a statistically powerful, unbiased methodemploying molecular genotyping (Swift et al 1998).

Homozygotes and Heterozygotes

We began collecting clinical data for Wolfram syndromehomozygotes and their close blood relatives to test thehypothesis that Wolfram syndrome heterozygotes are pre-disposed to diabetes mellitus. This hypothesis is importantbecause heterozygotes, unlike the rare homozygotes withthe distinctive syndrome, constitute about 1% of thegeneral population (Swift et al 1991).

From the Departments of Medicine and Psychiatry and The Institute for the GeneticAnalysis of Common Diseases, New York Medical College, Hawthorne, NewYork.

Address reprint requests to Michael Swift, M.D., New York Medical College, TheInstitute for the Genetic Analysis of Common Diseases, 4 Skyline Dr,Hawthorne NY 10532.

Received September 27, 1999; revised December 23, 1999; accepted January 6,2000.

© 2000 Society of Biological Psychiatry 0006-3223/00/$20.00PII S0006-3223(00)00244-4

Psychiatric Manifestations in WolframSyndrome Homozygotes

In all our studies of autosomal recessive syndromes (e.g.,Swift et al 1987), the diagnoses are always confirmed forthe index cases from medical records before collectingclinical data from the family. The impression that therewas a high prevalence of psychiatric difficulties in Wol-fram syndrome homozygotes led to a systematic recordreview. We found that of the 68 Wolfram syndromepatients located by writing to endocrinologists and oph-thalmologists at major medical centers and schools andinstitutes for the blind throughout the United States, 41(60%) had episodes of severe depression, psychosis, or-ganic brain syndrome, or impulsive verbal and physicalaggression (Swift RG et al 1990). The psychiatric mani-festations were so severe that 17 patients (25%) requiredadmission to a psychiatric hospital, had attempted suicide,or both.

In four cases, there was striking evidence that theseriousness of the patient’s psychiatric condition was notappreciated early enough. One patient made multiplesuicide attempts before appropriate treatment was ob-tained. Another presented in ketoacidosis because he wasno longer capable of correctly assessing his urine sugars.Two other patients developed severe water intoxicationbecause of their inability to comply with attempts at waterrestriction. In several patients, antidepressants alone or attimes in combination with antipsychotics were effective incontrolling symptoms. Benzodiazepines decreased the ag-gression of some patients, but not all.

Case reports describe the following psychiatric difficul-ties in Wolfram syndrome patients: “unstable . . . way-ward, difficult to manage . . . would fly into a tantrum”(Tyrer 1943); “emotionally unstable” (Tunbridge andPaley 1956); “admitted to a psychiatric department” (Roseet al 1966); “agitated depression” (Bretz et al 1970); “along psychiatric history of depression . . . electroshocktherapy . . . receiving chlorpromazine, thioridiazine, diaz-epam, and doxepin at various times” (Gossain et al 1975);“alteration of personality and mental function” (Kehl andKeller 1982); “multiple psychiatric disturbances, includingbulimia, depression, and a borderline personality disorder”(Rando et al 1992); “delirious state with ideas of referenceand persecution during manic episodes” (Nanko et al1992); “The father noted that the patient’s emotion hadbecome very labile and his moods changed rapidly”(Chuah et al 1993); “moody and depressed” (Bitoun1994); “impaired impulse control, affective incontinence,deficient judgement, and reduced initiative” (Kellner et al1994); “depression; mental retardation” (Okten et al1995); “memory loss, together with personality changes”(Genis et al 1997). The failure of a group in the United

Kingdom (Barrett et al 1995) to detect psychiatric mani-festations may reflect the limited clinical data that werecollected.

Wolfram syndrome homozygotes have diverse neuro-logical abnormalities in addition to symptoms character-ized as “psychiatric,” including mental retardation, sei-zures, central diabetes insipidus, nerve deafness, ataxia,anosmia, pupillary abnormalities, peripheral neuropathy,and bulbar dysfunction (Kinsley 1995). Neuropathologicchanges in Wolfram syndrome patients are widespread(Genis et al 1997; Kinsley and Firth 1992; Rando et al1992; Shannon et al 1999). In many areas of the brain,axonal degeneration is more striking than neuronal loss.

Psychiatric Manifestations Due to Mutationsat a Single Locus

Psychiatric findings observed in Wolfram syndrome ho-mozygotes are strikingly diverse, even though all of thesemanifestations are due to mutations at a single geneticlocus. Violent or assaultive behavior was seen in somecases, but certainly not in all. In others, severe depressionwas the only psychiatric symptom. Others showed nosymptoms until they developed cognitive and memorydifficulties, typically late in the course of the illness but ata much younger age than members of the general popu-lation. Changing characteristics of gene-associated psychi-atric illness over the patient’s lifetime have been observedin other single-gene syndromes, such as Huntington dis-ease, that affect the central nervous system. Thus, nosingle diagnostic category will fit the symptoms caused bymutations at any specific genetic locus.

Wolfram Syndrome Heterozygotes

Although the psychopathology of Wolfram syndromehomozygotes is striking, their illnesses constitute only atiny fraction of all psychiatric illnesses because thesehomozygotes are rare. On the other hand, individuals whocarry a single Wolfram syndrome mutation (heterozy-gotes) are estimated to be approximately 1% of the generalpopulation. If a single mutation at the Wolfram syndromelocus predisposes individuals to psychiatric illnesses, thenthis locus is quantitatively important in contributing to thepublic health burden of mental illness.

Wolfram syndrome heterozygotes cannot be distin-guished in the general population by conventional physicalexamination or laboratory testing. In particular, no focal ordistinctive neurological abnormalities have been describedin Wolfram syndrome heterozygotes. Until recently, Wol-fram syndrome heterozygotes were identified only when acouple had a child with the clinical Wolfram syndrome.From the beginning of our study of Wolfram syndrome

788 M. Swift and R.G. SwiftBIOL PSYCHIATRY2000;47:787–793

families, there were indications that heterozygotes mighthave psychiatric disturbances.

Physicians’ notes for some Wolfram syndrome patientscommented on “difficult family relationships.” In addi-tion, physicians often remarked that the family was lost tofollow-up. This itself was remarkable because both diabe-tes mellitus and distinctive genetic disorders require con-tinuing specialized care. As we began to collect clinicaldata from the Wolfram syndrome families, there was asubstantial proportion of relatives who were difficult tocontact and reluctant to participate. In some families, thenuclear family of the index case did not know theaddresses or telephone numbers of their closest relatives.

Comparison of Blood Relatives and SpouseControl Subjects

We collected clinical data uniformly for blood relativesand spouse control subjects (Swift et al 1991). The bloodrelatives contain a high proportion—typically 50% ormore—of individuals carrying a single Wolfram syn-drome mutation, whereas the proportion among spousecontrol subjects is expected to be the population rate of1%.

To test the hypothesis that Wolfram syndrome heterozy-gotes had an excess of psychiatric illness, the completemedical records of each person for whom the itemsrelating to mental illness or chronic nervous trouble wereanswered “yes” on a questionnaire or who had an ICD-8code between 290.0 and 309.9 (psychoses, other mentaldisorders) were retrieved and reviewed blind as to bloodrelative or spousal status. Study subjects were counted ashaving a psychiatric hospitalization only if their recordswere identified through these computer searches andverified by the medical records from the hospitalization orif a medical record explicitly referred to the psychiatrichospitalization. A relative was considered to have com-mitted suicide only if his or her death certificate listedsuicide as the cause of death.

A significantly larger proportion (0.019) of the 543blood relatives age 20 or older who were married orwidowed had a psychiatric hospitalization or had commit-ted suicide than the 365 spouses (0.004) in the Wolframsyndrome families. The spouse control subjects were wellmatched to the Wolfram syndrome blood relatives in age,socioeconomic class, ethnic origins, and adult environ-mental exposures; they were the best available comparisonpopulation. Nonetheless, the use of spouse control subjectsmay have introduced some bias toward the null hypothe-sis. The prevalence of psychiatric disorders in thesecontrol subjects may have been higher than that inmatched unrelated individuals because the prevalence ofsuch disorders was high in the blood relatives and there is

assortative mating for psychological characteristics. Com-paring disease incidence in close blood relatives to that inspouse control subjects was standard in our family studiesat that time.

Positional Cloning

These data made it clear that it was important to clone theWolfram syndrome gene. DNA was collected from Wol-fram syndrome families for linkage analysis directedtoward positional cloning. In 1994, the Wolfram syndromelocus was mapped to chromosome 4p (Polymeropoulos etal 1994).

Rigorous Hypothesis Testing through theIndex-Test Method

This successful mapping led to a rigorous test of thehypothesis that Wolfram syndrome heterozygotes are pre-disposed to psychiatric illness. It became possible todetermine accurately which blood relatives in Wolframsyndrome families carried a Wolfram syndrome mutationby analyzing their haplotypes of the highly polymorphicmarkers that flanked the Wolfram syndrome locus. Thesegenotypes (carrier or noncarrier) were the basis for hy-pothesis testing using the index-test method (Swift M1990).

This method is simple to use and statistically powerfulin detecting true associations. The first step was to selecta “test group” of blood relatives in Wolfram syndromefamilies. Each relative in this group had had a psychiatrichospitalization and each had a prior probability of 0.5 or0.25 of carrying a Wolfram syndrome mutation. Theparents of an index case cannot be used in this methodbecause they are obligate heterozygotes for a mutation.

For each member of the test group, the next step was todetermine if he or she carried the Wolfram syndromemutation transmitted in his or her family. This was donethrough DNA analysis of marker haplotypes (Swift RG etal 1998). The results were remarkable; 10 of the 11genotyped test subjects were mutation carriers, althoughwe expected only 2.75 to be carriers if there were noassociation (the null hypothesis). These data led to anestimate that Wolfram syndrome heterozygotes have a26-fold relative risk of hospitalization for psychiatricillness.

Psychiatric Phenotype

The choice of a robust phenotype for psychiatric illnesswas crucial to the success of these studies. Psychiatricdiagnoses may vary in the same patient over time (Hyde etal 1992), may differ between observers, and may not

Wolfram Syndrome Locus 789BIOL PSYCHIATRY2000;47:787–793

correspond to genetic loci in a one-to-one fashion. Forthese reasons, we began by studying the prevalence ofpsychiatric hospitalizations and suicides, which provideobjective evidence of a severe behavioral disorder. Thesedata can be scored “yes” or “no,” without ambiguity.There was no chance of interobserver variability, and theconclusions of the study did not depend on arbitrarilyselected diagnostic criteria, which might or might not fitthe psychiatric disorders experienced by Wolfram syn-drome heterozygotes. There was no need to specify inadvance a specific DSM or other diagnostic category.

Refinement of Gene-Associated ClinicalPhenotype

One of the most valuable features of the index-test methodis that clinical psychiatric phenotypes can be reliablyassociated with mutations at specific locus through itera-tive analysis of empirical data. Test groups of bloodrelatives can be formed to test different hypotheses. Forexample, the analysis of clinical phenotypes associatedwith Wolfram syndrome mutations began with a test groupdefined by the objective criterion of having a psychiatrichospitalization. Members of this group had to have suffi-cient functional capacity to return a questionnaire, signmedical release forms, and donate a blood sample.

Association with Depression and SuicideAttempts

Once the overall association of Wolfram syndrome muta-tions with psychiatric hospitalizations was confirmed, asdescribed above, the clinical records of the Wolframsyndrome mutation carriers in the test group were re-viewed. All carrier blood relatives in the test group haddepression as their principal psychiatric manifestation.Several had made suicide attempts. This suggests a newhypothesis with a specific clinical phenotype: Wolframsyndrome heterozygotes are predisposed to depression,regardless of whether it leads to hospitalization. A new testgroup of Wolfram syndrome blood relatives with docu-mented depression can be used to test this new hypothesis.

Other Clinical Phenotypes Associated withWolfram Syndrome Heterozygosity

Clinical data from Wolfram syndrome families can gen-erate additional hypotheses about other specific behavioralor cognitive difficulties associated with Wolfram syn-drome locus mutations. Questionnaires received fromWolfram syndrome families reported several differenttypes of behavioral difficulties (Table 1). Many subjectsresponded to more than one item, so they are included in

several rows. There was a constellation of related findingsin approximately 40 of the 282 who replied. These personstypically had school difficulties demonstrated by “beingleft back a grade,” suspension or expulsion, alcohol ordrug abuse, depression or other psychiatric symptoms,arrests or convictions, marital difficulties, and low occu-pational status.

The index-test method applied within Wolfram syn-drome families is the most reliable way to determine ifmutations at the Wolfram syndrome locus predispose toeach of these manifestations. Genotyping test groups ofWolfram syndrome blood relatives with distinctive behav-ioral abnormalities will provide a clear answer. Theprevalence of an aberrant behavior in Wolfram syndromefamilies is not sufficient in itself to prove an associationwith mutations at the Wolfram syndrome locus becausethis behavior may be equally prevalent in carriers andnoncarriers in these families. Once an association isestablished, precise psychiatric phenotypes associatedwith Wolfram syndrome mutations can be defined byiterative applications of the method, as described above.

Wolfram Syndrome Mutation Screening inClinical Diagnosis

Testing individual psychiatric patients for Wolfram syn-drome mutations (Inoue et al 1998; Strom et al 1998) canidentify psychiatric patients whose illnesses are caused bya mutation at the Wolfram syndrome locus. For example,suppose it is learned that children who carry a mutatedWolfram syndrome gene have difficulties with learningand school. Children in the general clinical populationwith such problems can then be tested for a Wolfram

Table 1. Wolfram Syndrome Blood Relativesa IndicatingCognitive or Behavioral Problems, from 284 QuestionnairesReceived out of 602 Sent to 25 Wolfram Syndrome Families

CriterionNumber ofrespondents

Checked that they had a psychiatric hospitalization 14Problem with drugs and alcohol 37Arrested, convicted, or both 29“Dementia” or periods of confusion 31Felt people were controlling their thoughts or had

hallucinations24

Checked “manic-depressive illness” 8Postpartum depression 11“Panic disorder” 14Depression lasting at least 2 weeks 51Attempted suicide 10Adult emotional outburst: hit someone or

destroyed property38

Expelled or suspended from school 20Diabetes mellitus 45

a Excluding homozygotes and obligate heterozygotes.

790 M. Swift and R.G. SwiftBIOL PSYCHIATRY2000;47:787–793

syndrome mutation. If such a mutation is found, therapycan be directed at this specific metabolic problem. Exist-ing drugs that have worked for other Wolfram syndromeheterozygotes can be utilized until new targeted drugsbecome available. There will be no need to blame “poorschooling” or “too much candy” if a specific genetic causeis identified.

Structure and Function of the WolframSyndrome Gene

The Wolfram syndrome locus has been identified throughstandard positional cloning (Inoue et al 1998; Strom et al1998). The gene consists of eight exons spanning 33.4kilobase of genomic DNA. The cDNA sequence predicts apolypeptide chain of 89 amino acids with a molecularmass of 100.29 kd, and this protein product has beendesignated “wolframin.” Inoue et al (1998) reported amatch, in the Prosite database, between its amino acidsequence and the prenyltransferasea-subunit repeat struc-ture, but the function of wolframin is not yet known. Thereare three structural domains: 1) a hydrophilic N-terminalregion of approximately 300 residues, 2) a hydrophilicC-terminal region of approximately 240 residues, and 3) acentral hydrophobic core of approximately 350 residues.The hydrophobic core appears to contain approximately10 transmembrane segments.

Determining the function of wolframin in normal andmutated nerve cells is a top priority. There are manyapproaches that can be tried, including histological localiza-tion, structure-function analysis, expression in cell culture,and the creation of transgenic animals with a mutation at thislocus. Understanding how mutations at the Wolfram syn-drome locus affect neuronal function will be the basis forrational design of new drugs to prevent or treat the psychi-atric symptoms of Wolfram syndrome heterozygotes.

Measuring the Contribution of WolframSyndrome Mutations to Psychiatric Illnessin the General Population

It is now possible to measure the public health burdenattributable to Wolfram syndrome mutations by screeningDNA from psychiatric patients for mutations at this locus.For example, based on the estimated relative risk and anestimated population heterozygote frequency of 1%, weestimated that 25% of all people hospitalized for similarillnesses with depression, suicide attempts, or both areWolfram syndrome heterozygotes. The population fre-quency of Wolfram syndrome heterozygotes among pa-tients who have had severe depression or attemptedsuicide now can be measured directly. New techniquesmake it practical to screen DNA from several hundred

subjects out of the general population to determine theproportion that carry a Wolfram syndrome mutation.

It also will be possible to measure the frequency ofWolfram syndrome heterozygotes in the healthy popula-tion. Comparing these two proportions is not a rigoroustest of the association of such heterozygosity with diseasebecause it is difficult to match the healthy and ill popula-tions for important risk factors that may actually beconfounders. Still, the measurement of these populationfrequencies will be invaluable in demonstrating how muchWolfram syndrome mutations contribute to the publichealth burden of psychiatric illness.

Furlong et al (1999) provide the first, limited, popula-tion data supporting the association of wolframin muta-tions with psychiatric disease, with the finding that aspecific mutation, Ala559Thr, was present in four of 314patients with affective disorders and in none of 382 controlsubjects (p 5 .065; Fisher’s exact test, one sided). Thismutation was found as a byproduct of an unsuccessfulattempt to test this gene–disease association through aneutral nonpathogenic single nucleotide polymorphism(SNP). Neutral SNPs are unlikely to detect importantgene–disease associations because their distribution re-flects primarily the evolutionary, migration, and intermar-riage history of ethnic groups (Li et al 1999). To estimatethe contribution of mutations at this locus to affectivedisorders it would be desirable to screen both the diseaseand control groups for mutations anywhere in the wol-framin gene; any single mutation is only a small part of thestory.

General Applications of the Index-TestMethod

Enormous resources, both federal and commercial, havebeen dedicated to finding the genes responsible for psy-chiatric and other chronic diseases. Before investing fur-ther resources in the study of a specific gene, we need tobe certain this gene is truly clinically important. Theindex-test method has such statistical power and freedomfrom bias that genetic predisposition identified throughthis method is certain to be clinically important. Today’spsychiatric drugs, although better than those used in thepast, still have problematic side effects that may lead tononcompliance. Drugs developed to target the metabolicrole of a gene proven to predispose to psychiatric illnessare likely to be more effective and have fewer side effects.

Advantages of the Index-Test Method

The index-test method (Swift M et al 1990) has severaladvantages in testing the association of mutations at aspecific locus with psychiatric disease. The results cannot

Wolfram Syndrome Locus 791BIOL PSYCHIATRY2000;47:787–793

be explained away by undetected confounders or unin-tended bias. No confounder can affect the result of theindex-test method because each individual’s genotype isfixed at the time of conception. Unintended bias isunlikely because the cases are selected before genotypingby explicit objective criteria, such as having hospitalrecord confirmation of a psychiatric hospitalization. Con-clusions from the index-test method are robust becausethey are unaffected by the other genetic and environmentalfactors responsible for many cases of psychiatric illness inthe general population.

The index-test method can be used to test any hypoth-esized association between mutations at a specific locusand psychiatric disease. The candidate locus could, butneed not, be identified because homozygotes have adistinctive autosomal recessive syndrome. When they do,as in the Wolfram syndrome, the “index” individuals arethe homozygous patients because they identify families inwhich mutations at the candidate locus are segregating.

The index-test method applies equally well to candidateloci where mutations do not produce a distinctive autoso-mal recessive syndrome in homozygotes. There are manyloci that are candidates for predisposition to psychiatricillness because the function of the gene product affectsbrain function. The genes coding for neurotransmitterreceptors are obvious examples.

For each such candidate locus, the challenge is to findfamilies in which mutations at that locus are segregating.Methods for doing this are described in detail in M. Swiftet al (1990) and United States Patent 5,464,742 (Swift et al1995). From this point forward, the index-test method isapplied as it was for the Wolfram syndrome. Test groupsof psychiatrically ill blood relatives are formed, and eachmember of this test group is tested for the specificmutation in that relative’s family. The observed number ofmutation carriers is then compared with the numberexpected in the test group under the null hypothesis of noassociation.

Comparison to Linkage Analysis

No gene for any common psychiatric illness has beenmapped unequivocally through linkage to DNA markers(e.g., Egeland et al 1987; Kelsoe et al 1989) because 1)most families in which psychiatric disorders are prevalentdo not show classic Mendelian inheritance patterns; 2) it isuncertain who in a family should be considered “affected,”given that the established diagnostic categories may notrespond to genetic entities in a one-to-one fashion; and 3)genes at different loci may predispose individuals toclinically indistinguishable illnesses. These factors do notaffect the outcome of the index-test method, for which

success already has been demonstrated by the studies ofthe Wolfram syndrome locus.

ReferencesBarrett TG, Bundey SE, Macleod AF (1995): Neurodegeneration

and diabetes: UK nationwide study of Wolfram (DIDMOAD)syndrome.Lancet346:1458–1463.

Bitoun P (1994): Wolfram syndrome. A report of four cases andreview of the literature.Ophthalmic Genet15:77–85.

Bretz GW, Baghdassarian A, Graber JD, Zacherle BJ, NorumRA, Blizzard RM (1970): Coexistence of diabetes mellitusand insipidus and optic atrophy in two male siblings.Am JMed 48:398–403.

Chuah G, Seah S, Chee SP (1993): Progressive optic atrophyassociated with juvenile diabetes mellitis: Report of two casesamong first cousins.Singapore Med J34:343–345.

Egeland JA, Gerhard DS, Pauls DL, Susex JN, Kidd KK, AllenCR, et al (1987): Bipolar affective disorders linked to DNAmarkers on chromosome 11.Nature325:783–787.

Furlong RA, Ho LW, Rubinsztein JS, Michael A, Walsh C,Paykel ES, Rubinsztein DC (1999): A rare coding variantwithin the wolframin gene in bipolar and unipolar affectivedisorder cases.Neurosci Lett277:123–126.

Genıs D, Davalos A, Molins A, Ferrer I (1997): Wolframsyndrome: A neuropathological study.Acta Neuropathol(Berl) 93:426–429.

Gossain VV, Sugawara M, Hagen GA (1975): Co-existentdiabetes mellitus and diabetes insipidus, a familial disease.J Clin Endocrinol Metab41:1020–1024.

Hyde TM, Ziegler JC, Weinberger DR (1992): Psychiatricdisturbances in metachromatic leukodystrophy. Insights intothe neurobiology of psychosis.Arch Neurol49:401–406.

Inoue H, Tanizawa Y, Wasson J, Behn P, Kalidas K, Bernal-Mizrachi E, et al (1998): A gene encoding a transmembraneprotein is mutated in patients with diabetes mellitus and opticatrophy (Wolfram syndrome).Nat Genet20:143–148.

Kehl O, Keller U (1982): DIDMOAD syndrome (diabetes insipidus,diabetes mellitis, optic atrophy, deafness, mit zerebello-pontineratrophie.Schweiz Med Wochenschr112:348–352.

Kellner M, Strian F, Fassbender K, Kennerknecht I, Klein R(1994): DIDMOAD (Wolfram) syndrome (letter).SchweizMed Wochenschr164:132.

Kelsoe JR, Ginns EI, Egeland JA, Gerhard DS, Goldstein AM, BaleSJ, et al (1989): Re-evaluation of the linkage relationshipbetween chromosome 11p loci and the gene for bipolar affectivedisorder in the Old Order Amish.Nature342:238–243.

Kinsley BT, Firth RGF (1992): The Wolfram syndrome: Aprimary neurodegenerative disorder with lethal potential.IrMed J85:34–36.

Kinsley BT, Swift M, Dumont RH, Swift RG (1995): Morbidityand mortality in the Wolfram syndrome.Diabetes Care18:1566–1570.

Li A, Huang Y, Swift M (1999): Neutral sequence variants andhaplotypes at the 150 kb ataxia-telangiectasia locus.Am JMed Genet86:140–144.

792 M. Swift and R.G. SwiftBIOL PSYCHIATRY2000;47:787–793

Nanko S, Yokoyama H, Hoshino Y, Kumashiro H, Mikuni M(1992): Organic mood syndrome in two siblings with Wol-fram syndrome (letter).Br J Psychiatry161:282.

Okten A, Gedik Y, Demirci A, Mocan H, Erduran E, Aslan Y(1995): Various clinical aspects of DIDMOAD (Wolfram)Syndrome.Turk J Pediatr37:235–240.

Polymeropoulos MH, Swift RG, Swift M (1994): Linkage of theWolfram syndrome disease gene with markers on the shortarm of chromosome 4.Nat Genet8:95–97.

Rando TA, Horton JC, Layzer RB (1992): Wolfram syndrome:Evidence of a diffuse neurodegenerative disease by magneticresonance imaging.Neurology42:1220–1224.

Rose FC, Fraser GR, Friedmann AI, Kohner EM (1966): Theassociation of juvenile diabetes mellitus and optic atrophy:Clinical and genetical aspects.QJM 35:385–405.

Shannon P, Becker L, Deck J (1999): Evidence of widespreadaxonal pathology in Wolfram syndrome.Acta Neuropathol(Berl) 98:304–308.

Strom TM, Hortnagel K, Gekeler F, Scharfe C, Rabl W,Gerbitz KD, et al (1998): Diabetes insipidus, diabetesmellitus, optic atrophy and deafness (DIDMOAD) causedby mutations in a novel gene (wolframin) coding for apredicted tramsmembrane protein.Hum Mol Genet7:2021–2028.

Swift M, Kupper LL, Chase CL (1990): Effective testing ofgene-disease associations.Am J Hum Genet47:266–274.

Swift MR, Kupper LL, Chase CL (1995): Process for testinggene-disease associations. U.S. Patent 5,464,742, issued No-vember 7, 1995.

Swift M, Reitnauer PJ, Morrell D, Chase CL (1987): Breast andother cancers in families with ataxia-telangiectasia.N EnglJ Med316:1289–1294.

Swift RG, Perkins DO, Chase CL, Sadler DB, Swift M (1991):Psychiatric disorders in 36 Wolfram syndrome families.Am JPsychiatry148:775–779.

Swift RG, Polymeropoulos M, Torres R, Swift M (1998):Predisposition of Wolfram syndrome heterozygotes to psy-chiatric illness.Mol Psychiatry3:86–91.

Swift RG, Sadler DB, Swift M (1990): Psychiatric findings inWolfram syndrome homozygotes.Lancet336:667–669.

Tunbridge RE, Paley RG (1956): Primary optic atrophy indiabetes mellitus.Diabetes5:295–296.

Tyrer J (1943): A case of infantilism with goitre, diabetesmellitis, mental defect and bilateral primary optic atrophy.Med J Aust2:398–401.

Wolfram DJ (1938): Diabetes mellitus and simple optic atrophyamong siblings: Report of four cases.Staff Proc Mayo Clin13:715–718.

Wolfram Syndrome Locus 793BIOL PSYCHIATRY2000;47:787–793