GENETIC FACTORS IN AFFECTIVE ILLNESS zyxwvutsrqponmlkjihgfedcb WADE H. BERRETTINI, LYNN R. GOLDIN, JOHN I. NURNBERGER, JR. and ELLIOT S. GERSHON Sectionon Psychogenetics,BiologicalPsychiatryBranch, National Institute Building 10, Room 3N220, Bethesda, MD 20205, U.S.A. of Mental Health, Summary-Research strategies for determining genetic vulnerability markers in affective illness are delineated. Using these strategies, recent developments in the biology of manic-depressive illness are discussed, including results from association and linkage studies, pharmacologic challenge protocols, and cerebrospinal fluid (CSF) data. Several lines of evidence suggest that one genetically determined vulnerability to affective disorder may be a cholinergic supersensitivity, possibly mediated through increased numbers of cholinergic receptors. INTRODUCTION THIS paper will attempt to review some recent developments in the search for genetic vulnerability factors in affective illness. Research strategies used in this search will be delineated. Recent studies of association and linkage of affective illness to chromosomal markers will be reviewed. Finally, some clinical neurobiological studies will be presented to suggest promising putative genetic markers for vulnerability to affective illness. For the purposes of this paper, it is assumed that some forms of affective illness are associated’with genetic vulnerability to the disease. The issue of genetic transmission in affective illness has been reviewed recently in detail elsewhere (NURNBERGER and zyxwvutsrqponmlkjihgfedc GERSHON, 1982). Briefly, the evidence in favor of genetic transmission of vulnerability factors in affective illness is quite strong. RESEARCH STRATEGIES In any periodic illness, genetic vulnerability implies that episodically ill individuals possess at least one continuously heritable (constitutional) factor which increases the risk of developing the illness in question. A genetic vulnerability factor for affective illness should have the following characteristics (RIBDER and GERSHON, 1978): (1) stable and heritable; (2) state independent (present regardless of whether the individual is currently in an episode of illness); (3) differentiates those with affective illness from the general population; and (4) within a pedigree in which the proband has the marker, the marker should differentiate relatives with the illness from those without. This is not an absolute differentiation because of decreased penetrance and possible genetic heterogeneity of the illness. In the progress of research in affective disorders, several studies must be done before a biological parameter can be considered as a putative genetic vulnerability marker. The parameter should be stable in individuals over time. If the parameter is highly concordant in monozygotic twins and less concordant in dizygotic twins, then the parameter is most likely under some degree of genetic control. That is, it is heritable. The mean value of the 329 330 WADE zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHG H. BERRETTINI et zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP al. parameter for euthymic (well-state) patients must be significantly different from controls. Finally, and most importantly, the putative marker must be associated with illness within a family in which the ill proband has the marker. If some biological parameter fulfils these criteria then it may be considered a genetic vulnerability marker for affective illness. Two different research strategies have been employed in the search for genetic vulnerability markers in affective illness. One can study biological parameters of theoretical interest in affective illness (such as receptor function, enzyme activity, etc.) to determine whether the parameter satisfies criteria for a genetic vulnerability marker. Alternatively, one can study known chromosomal markers, such as ABO blood types or human leukocyte antigens (HLA) in individuals with the illness. This latter strategy refers to association and linkage studies. The object of these studies is to determine whether a particular allele at a chromosomal locus is associated with illness in the general population of ill individuals, or chromosomally linked to a susceptibility gene for development of illness within given pedigrees. In the first strategy one commonly studies biological parameters, which reportedly differentiate affectively ill from well populations (that is, the means for ill and well groups are significantly different), in an attempt to prove that the parameter is heritable and is associated with a greater risk for illness within certain pedigrees. In the second strategy one studies parameters which are known to be transmitted by a single major locus (ABO blood groups for example) in an attempt to determine whether the marker locus is related to susceptibility for illness. If ill populations have a significantly higher incidence of an allele at a marker locus than do control populations, this suggests that the allele plays some role in disease susceptibility. However, if alleles at a marker locus co-segregate with the illness in families, this is evidence that the disease locus is chromosomally linked to the marker locus. Linkage will generally not lead to the association of illness with particular alleles at the marker locus at the population level. However, in some cases, (such as loci in the HLA region) associations of alleles at closely linked loci are found, presumably due to disequilibrium. Separate methodologies exist for studying linkage and association, depending on the type of data obtained. The association of a genetic marker trait (such as an HLA antigen or an ABO blood type) with a disease is tested by comparing the frequency of the trait in populations of patients and controls. This is illustrated in the 2 x 2 table below. Marker Disease Trait Present a Present Absent C Absent b d In this table, a, 6, c and d are the numbers of individuals falling into each category. The most commonly used measure of association is the relative risk (R) given below, which describes the increased risk for an individual to have the disease when he has the marker trait: R= ad -. bc GENETICS AND AFFECTIVE ILLNESS 331 Other measures of association and techniques for combining results across studies are described in detail in THOMSON(1980). The significance of an association can be determined by a chi-square statistic from the 2x2 table. However, if more than one marker trait is tested (e.g. large series of HLA antigens) then the significance level must be multiplied by the number of tests. SMOUSE and WILLIAMS (1982) have developed an alternative method for the analysis of HLA antigens which yields a single chi-square test for each locus and an overall test for the whole complex. This takes into account allelic correlations within and between HLA loci. Two methods of detecting linkage can be used depending on the type of data available and the assumptions that can be reasonably made. If the data consist of complete nuclear families or larger pedigrees, then it is possible to apply the lod score method of MORTON (1955). Methods have also been developed to detect linkage in samples of affected sibpairs. These two methods will be discussed in terms of the underlying assumptions and relative advantages and disadvantages. Lod score method The lod score method developed by MORTON (1955) is a means of testing the hypothesis of linkage between two loci when the mode of transmission of each locus is known. The underlying assumptions are that: (1) the parameters (gene frequency and genotypic penetrances) for the disease locus and marker locus are known; (2) there is no population association between the disease locus and marker locus; (3) mating is random. Under these assumptions one compares the probability of observing the pattern of segregation of the two traits in a family if there is linkage to the probability of observing the family, if there is no linkage. The probability of linkage is expressed as a function of the recombination fraction (0) where Clis some value between 0 and l/2. The probability of no linkage is the probability that the two loci are segregating independently (i.e. 8 = l/2). This odds ratio t3 is expressed by a statistic called the lod score (or log of the odds ratio) and is defined as follows: probability of observing a family for 8 < l/2 lod score = log,, probability of observing a family for t3= l/2’ A lod score of 1.0 means that linkage is 10 times more likely than no linkage. The lod scores for small families can be done by hand (LEVITAN and MONTAGU, 1971) or using tables in simple cases. For larger or more complex families, a widely available computer program LIPED (OTT, 1974) performs the calculations. In practice, for each family the lod score is evaluated for several values of 8. Since the scores represent independent probability ratios (the scores are in log,, units), they can be summed across families. If linkage is true, the best estimate of 0 is that value of 8 which results in the highest lod score. Data can be evaluated sequentially after each sample of pedigrees is collected. MORTON (1955) originally proposed lod scores of 3.0 being the cut-off value for acceptance of linkage and -2.0 being the cut-off value for rejection of linkage. If the score is intermediate, then more families should be collected until the hypothesis can either be accepted or rejected. This convention has generally been followed in linkage studies. However, use of these absolute cut-off points may not always be appropriate; when tests of linkage of a 332 disease locus to many marker in a single study. WADE H. BERRETTINI et al. loci are made, significant lod scores may occur by chance Sib-pair method The use of samples of affected sib-pairs for linkage analysis may be desirable because fewer assumptions are required. The method used currently derives from Penrose’s sib-pair method (PENROSE, 1935, 1953). The idea behind the method is that if a marker locus is linked to a disease locus, then affected pairs of siblings will have the same phenotype at the marker locus more often than expected by chance. Since only affected sibs are used, this method is especially useful for disease susceptibility loci that are thought to have low penetrance. Although general in theory, this method has been mainly developed to apply to problems of detecting linkage to the HLA loci. Because there is so much polymorphism in the HLA region, each parental chromosome has a different set of HLA alleles (or haplotype). Thus, it is usually possible to determine whether affected sib-pairs share exactly 2, 1 or 0 haplotypes identical by descent (IBD) at the marker locus. If there is no linkage, then the proportion of affected sib-pairs sharing 2, 1, and 0 haplotypes is l/4, l/2, l/4, respectively. If linkage is present, then this distribution is skewed so that more than 25% of affected sib-pairs have identical haplotypes. The simple hypothesis of linkage can be tested by comparing the observed IBD distribution in a sample of independent affected sib-pairs with that expected when there is no linkage. The affected sib-pair method has been further developed by several investigators (SUAREZ et al., 1978; SUAREZ, 1978; DAY and SIMONS, 1976; THOMSON and BODMER, 1977a, b; GREEN and WOODROW, 1977) in order to make some inferences about the mode of inheritance of the disease trait and the recombination fraction between the disease and marker loci. For example, THOMSON and BODMER (1977a, b) have derived tables of the expected IBD distribution for simple dominant and recessive models with varying gene frequency. In addition, methods have been developed (GREEN and WOODROW, 1977; SUAREZ and CROUGHAN, 1982) to evaluate the degree of haplotype sharing when more than two affected sibs in a sibship are available for study. In theory, it is also possible to use unaffected pairs and unaffectedaffected pairs of siblings for this test; however, SUAREZ et al. (1978) have shown that these types of sib-pairs do not contribute very much information about linkage. It should be noted that although the sib-pair method has been developed and applied largely to the HLA locus, it is general and can be applied to other marker loci (e.g. ABO, Rh, etc.). However, because most other loci are not as polymorphic as HLA, not all sib-pairs will be informative and a larger sample size will be needed in order to detect linkage. This method has been widely applied to testing of linkage to HLA in samples of affected sib-pairs with juvenile diabetes (see review in CLERGET-DARPOUX, in press) and multiple sclerosis. Later we will review sib-pair studies of HLA and major affective disorder. Comparison of methodoiogies of linkage analysis The choice of methods to use for linkage analysis will depend on several factors. For some diseases, it may be easier to collect data on a sample of affected sib-pairs than on complete families. While the lod score method is easily applied to small nuclear families, it is most powerful for larger multigenerational families (which are generally more difficult to study). On the other hand, as stated previously, the affected sib-pair method is most GENETICS ANDzyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHG AFFECTIVEIILNES 333 efficient for highly polymorphic loci such as HLA and will require larger sample sizes in the case of other loci. Most importantly, the underlying assumptions required for these methods are different. For the lod-score method, the underlying genetic parameters (gene frequency, penetrances) must be specified. Unfortunately, for many diseases studied, the mode of transmission is not known and is often guessed in order to calculate lod scores. However, incorrect specification of the model can give the wrong results (SUAREZand VAN EERDEWEGH,1981). MORTONand KIDD (1980) and KRUGERetzyxwvutsrqponmlkjihgfedcbaZYXW al. (1982) have also shown how variation in unknown model parameters can affect the results of linkage analysis using real data. As stressed previously, no such assumptions are needed to test the hypothesis of linkage in affected sib-pairs, and this method is therefore most appropriate when the exact mode of transmission of the disease locus is unknown. If preliminary evidence for linkage is obtained from sib-pairs, it is possible to estimate genetic parameters under certain models and thereby make some inferences about the mode of transmission of the disease. Another problem that needs to be considered is possible genetic heterogeneity of the disease under study. Since families are pooled for both methods, it is assumed that the same disease locus is segregating in each family. If there is heterogeneity then such pooling of families will lead to erroneous results. It is possible to test for heterogeneity of linkage (MORTON,1956); however, large families are generally needed to be able to detect significant heterogeneity. The presence of association of alleles at the disease and marker locus can bias the results of linkage. CLERGET-DARPOUX (1982) has shown that ignoring associations in a lod score analysis biases the estimate of the recombination fraction (slightly upward) and decreases the lod score. However, in theory, it is possible to calculate lod scores taking association of phenotypes into account. These deviations from simplifying assumptions must be taken into account when performing linkage analyses (see SPENCE, 1980). This is especially true for psychiatric disorders where the validity of most of the simplifying assumptions are not known, i.e. the mode of transmission of illness is unknown, the illnesses may be genetically heterogenous and assortative mating may exist (see GERSHONet al., 1973; DUNNERet al., 1976; NEGRI et al., 1981). In addition, the “ill” phenotype cannot be exactly defined since there are numerous theories with regard to which diagnoses should be included in a particular “spectrum”. Association and linkage studies in affective disorder Association studies ABO locus. Several studies have reported the frequencies of ABO types in patients with affective disorder (or sub-groups of patients) as compared to control populations. Results from most of these studies are conflicting. A few studies (e.g. PARKER et al., 1961; MENDLEWICZet al., 1974; R~NIERISet al., 1979) have found a higher frequency of blood type 0 in manic-depressive patients than in controis. In contrast, FLEMENBAUM and LARSON(1976) found type A to have a higher frequency in patients and BECKMANet al. (1978a) found the frequency of type B to be increased in unipolar (UP) and bipolar (BP) patients. Other studies have found no differences between patients and controls (TANNA WADEH. BERRETTINI et al. 334 and zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA WINOKUR, 1968; JAMES et al., 1979). SHAPIRO et al. (1977a) found no overall difference in patients but found bipolars to have a higher frequency of type 0 and unipolars to have a higher frequency of type A. Clearly, there is no consistency in these reports. FLEMENBAUMand LARSON (1976) suggest that many of the significant results reported in the literature, including their own, are artifacts due to skewed patient population samples. In their own study, they found that the entire population of psychiatric patients had different ABO frequencies than the control population. If manic-depressives were compared to all other psychiatric patients, there were no significant ABO differences. They present evidence to suggest that this is true in some of the other published studies of ABO types in psychiatric disorders and conclude that psychiatric patients may be different from the overall population with respect to ABO types. This demonstrates the potential effect of population stratifications on association studies. HLA loci. Ten studies of HLA types in patients with affective disorders and controls are summarized in Table 1. This table presents the results of whole samples of patients and only the latest study of each group of investigators. One of the first studies by SHAPIRO et al. (1976, 1977b) reported that HLA BW16 had a significantly higher frequency in a sample of Danish patients than in controls. The difference was significant @=0.046) after statistical correction for multiple tests and was more pronounced (although not significant) in patients without a family history of affective disorder. This finding generated some excitement; unfortunately, subsequent studies were not able to replicate the association. Associations with other antigens were reported but none were statistically significant after correction for the number of antigens tested. The fact that HLA BW16 was found to be higher in the Danish study is interesting since this antigen is found with much higher frequency in Ashkenazi Jewish populations than in other European Caucasian populations. zyxwvutsrqponm TABLE 1. ASSOCIATION STUDIES:HLA AND MAJORAFFECTIVE DISORDER No. Study SHAPIKO et al. (1977b) of patient!, 107 STEMBERand FEVE (1977) 50 Results f BWl6 t BS, 813, i BlZ Significance @ < 0.05) after statistical correction t BW16 BW35 t B13 t B5 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA No difference 118 GOVAERTSei al. (1977) 1 B1.S SMERALDIet al. (1978~1) 91 BECKMANet al. (1978b) 168 JOHNSON (1978) TEMPLF et al. (1979) TARGUM eI a/. (1979) .IAMES et al. (1980) M~NDLE WICZ et al. ( 198 I) ~~--~--_-_ t A29, BW22 1 AlO, A30 No difference t A10 1 B7 No difference 36 No difference 38 t BW22 C AY, CW3 1 cw3 92 t BW14 ? BW27 No difference I14 No difference 47 No diffrrznce GENETICSANDAFFE~TIVE ILLNESS 335 Ashkenazi Jews also have a higher frequency of manic-depressive illness than other European populations (GERSHON and LIEBOWITZ, 1979). Two studies (TEMPLE et al., 1979; TARGUM et al., 1979) compared HLA types in Ashkenazi Jewish patients to ethnically matched controls and found no differences. In most of the HLA studies reported, BW16 is slightly higher in patients than in controls. This may be another example of population stratification if the samples of patients are enriched with individuals of Ashkenazi Jewish origin. TARGUM et al. (1979) have shown how a spurious association can be found if comparisons are made between heterogeneous populations. Linkage studies Studies of linkage of affective disorder to genetic markers have included loci both on the X-chromosome and the autosomes. The X-chromosome studies were motivated by the findings of WINOKUR et al. (1969) that in BP families there was a high rate of motherson transmission of illness, virtually no father-son transmission and a higher proportion of females affected than males. These findings are consistent with the pattern of an X-linked dominant gene. Other family studies (GOETZL et al., 1974; GERSHON et al., 1975) have not found these differences in transmission but the results of linkage studies with X-chromosome markers have stirred up considerable controversy. Other autosomal marker loci have been studied with special emphasis on the HLA loci. In the next two sections, we shall review studies of linkage to X-chromosome and autosomal markers. X-chromosome marker studies. WINOKUR et al. (1969) were the first to present evidence to support the X-chromosome transmission hypothesis by showing that the X-linked colorblindness (CB) locus segregated with manic-depressive illness in a single large pedigree. This finding was supported in a large series reported by MENDLEWICZ et al. (1972), in a report by MENDLEWICZ and FLEISS (1974), and in a series of families with schizoaffective probands (MENDLEWICZ, 1977). FIEVE et al. (1973) also reported positive evidence for linkage of BP illness to the red cell Xg locus; however, the CB and Xg loci are on different arms of the X-chromosome and linkage of a gene for BP illness to both loci is incompatible. More recently, MENDLEWICZ et al. (1980) reported linkage between the G6PD locus (near the CB locus) in a single multigenerational pedigree of a BP proband. BARON (1977) has also reported a single pedigree containing SA and BP individuals which supports linkage of these disorders to the CB locus. Data on 6 informative families from our own large series of BP pedigrees showed no evidence for linkage of affective disorder and CB (GERSHON et al., 1979) or Xg (LECKMAN et al., 1979). These families were also reported as part of a WHO collaborative study (GERSHON et al., 1980). The pooled series of 16 families was found to be heterogeneous with one family favoring linkage (lod score > 2.0) and the remainder being indeterminate (lod scores between -1 and + 1). In a preliminary report, EGELAND (1980) found no evidence for linkage to CB in a large BP pedigree from the Old Order Amish population. It should be noted that there are other methodological issues with regard to the CB linkage studies. In the early reports, lod scores were calculated with no correction for variable age of onset of illness. MORTON and KIDD (1980) have demonstrated how age correction can affect the results of linkage analyses. As discussed by GERSHON and BUNNEY (1977), affective disorder seemed to be associated with colorblindness in the 336 WADEHzyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJI . BERRETTINI et al. MENDLEWICZ et al. (1972) series. Such an association is not expected if there is actual linkage. This raises the possibility of a bias in ascertainment of families. Recently RISCH and BARON (1982) have reanalyzed all of the published X-chromosome data incorporating age of onset and concluded that there is heterogeneity of the illness so that a proportion of families are segregating for a gene linked to the CB locus of the X-chromosome. However, this interpretation is controversial since most of the families favoring linkage are from one group of investigators and most of the families favoring non-linkage are from another group of investigators. This issue may be resolved in the near future with the use of DNA polymorphisms in the CB region of the X-chromosome (SINISCALCO et al., 1982). If enough polymorphisms can be identified, all families will be informative for X-chromosome linkage. Autosomal markers. Most studies of linkage of affective disorder to autosomal markers have been of the HLA region of chromosome 6. However, a few studies have examined linkage to other polymorphic loci. Tanna and co-workers found positive evidence for linkage of “depression spectrum disease” to the Haptoglobin locus first in sib-pairs (TANNA et al., 1976) and then by calculating lod scores in families (TANNA et al., 1979). [Depression spectrum disease is defined as unipolar illness in a proband who has at least one first degree relative with alcoholism or sociopathy (WINOKUR, 1972).] However, the maximum lod score in 14 families was 1.03 (i.e. linkage 12 times more likely than non-linkage). In the absence of further evidence, a score of approximately 1.0 is likely to be due to chance. A study of 30 markers in a single large UP pedigree by CROWE et al. (1981) and a study of 29 markers in a single large UP pedigree (WEITKAMP et al., 1980) did not suggest any linkages or provide any support for linkage to the Haptoglobin locus. Families with BP, UP and SA illness from our own laboratory have recently been analyzed for linkage to 21 autosomal loci (GOLDIN et al., 1983). There were no positive linkages, the highest lod score being 1.39 (0 = 20%) for the MNS locus in 30 informative families. Recently there has been considerable interest in the relationship of a possible susceptibility gene for affective disorders to the HLA region. Smeraldi and coworkers in Italy (SMERALDI et al., 1978b; SMERALDIand BELLODI, 1981) found that pairs of siblings, both of whom had a major affective disorder (UP or BP), shared HLA haplotypes more often than would be expected by chance. This suggests that a susceptibility locus for affecrive disorder is linked to the HLA locus, but the finding had only borderline statistical significance in their sample of 26 sib-pairs. WEITKAMP et al. (1981) reported data from 21 sibships (6 pairs were part of a single large multigenerational pedigree) from UP and BP families. The distribution of HLA haplotypes in the overall sample did not deviate from random. However, a subset of sib-pairs, where there were only two affected sibs in the sibship (as opposed to three or more affected) did share more HLA types than expected from chance. The authors hypothesized that the significant deviation in this particular subset of families was expected on the basis of an assumption that sibships with only two affected sibs represented families where parents had “fewer” susceptibility genes than in sibships with more than two affected. With “fewer” parental susceptibility genes, the increase sharing of HLA types would be more easily detectable. GOLDIN et al. (1982) have observed that this GENETI~~.~NDATFEcTIvE ILLNESS 337 subdivision of the data is not theoretically justified because it does not correspond to any specific genetic hypothesis. A study by TKJRNER and KING (1981) showed positive evidence for linkage of HLA to affective disorder in several BP pedigrees that were consistent with autosomal dominant transmission of illness. Two of these families were recently reanalyzed by KRUGER et al. (1982) who found that the likelihood of linkage depended to a large extent on the underlying assumptions about the mode of transmission of affective disorder, age of onset correction, marker allele frequencies, and diagnostic criteria. The maximum likelihood of linkage to HLA was found when loose criteria for affective disorders (including UP, BP and several personality disorders) were applied. However, when the GLO locus (a marker closely linked to HLA) was analyzed, the maximum likelihood of linkage was found when strict criteria for affective disorder (including only UP and BP) were applied. These two results appear to be inconsistent. We and others have not been able to demonstrate a relationship of HLA to major affective disorder, Data from our own laboratory (TARGUM et al., 1979; GOLDIN et al., 1982) have been analyzed by two methods. Both a sample of 19 pedigrees (UP and BP) analyzed for linkage of HLA to affective disorders using standard methods of linkage analysis and a sample of 21 affected sib-pairs (mostly from the same families) failed to show evidence for an HLA role in this disorder. A recent analysis of affected sib-pairs in 10 families by SUAREZ and CROUGHAN (1982) has also failed to provide evidence for a susceptibility locus in the HLA region. In addition, both our data and that of Suarez and Croughan are inconsistent with Weitkamp’s predictions of increased haplotype sharing in sibships with only two affected sibs. In summary, there are conflicting reports regarding association and linkage of affective disorders to these known chromosomal markers. Our data do not support linkage to any of these markers. In view of the conflicting reports, one might conclude conservatively that the affective disorders are not closely linked to any of these chromosomal markers. For each marker, there are several negative studies. If linkage to affective illness exists for any of these chromosomal markers, it is weak and not of primary importance in the development of the disease. Clinical neurobiological studies The search for neurobiological genetic vulnerability markers in affective illness is an important research strategy that involves two independent studies. In one, it must be shown that the neurobiological variable is heritable. In the second, it must be demonstrated that, within a family in which the ill proband has the neurobiological variable, the variable should differentiate well from ill relatives (RIEDER and GERSHON, 1978). In this section several clinical neurobiological studies, performed as part of a search for vulnerability, will be discussed. CSF studies Few studies of CSF biochemistry have been conducted to establish CSF parameters as genetic vulnerability markers. However, several lines of indirect evidence are compatible with the hypothesis that CSF 5-hydroxyindoleacetic acid (S-HIAA) levels may be stateindependent and genetically determined. SEDVALL and OXENSTIERNA (1981) found a 33x WADE E. BERRETTINI era/. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA significant intraclass correlation among monozygotic twins but did not demonstrate that this concordance is higher than that found in dizygotic twins. Additionally, the studies of SWANN et al. (1983) and KOSLOW et al. (1983) found elevated CSF 5-HIAA values in 14 manic and 43 depressed women compared to the 29 female normal volunteers. The data of POST et al. (1980) are also compatible with 5-HIAA as a state independent variable. However, VESTERGAARD et al. (1978) found that recovered patients had significantly higher @ < 0.02, paired ?-test) CSF 5-HIAA levels compared to depressives. This suggests that 5-HIAA may be state dependent. Additional studies are required to determine whether CSF 5-HIAA levels are independent of mood state and genetically determined. SEDVALLet ul. (1980) found that 28 psychiatrically healthy individuals with a positive family history of severe psychiatric illness showed greater variation in CSF 5-HIAA than did 32 normals without the positive family history. Additionally, they reported that those with abnormally low CSF 5-HIAA tended to have a family history of depression while those with high 5-HIAA tended to have a family history of schizophrenia. There was a significant difference between levels of CSF 5-HIAA in the realtives of schizophrenics compared to the relatives of depressives. These data are compatible with, but are not strong evidence for, the hypothesis that low CSF 5-HIAA is a familial, state-independent vulnerability factor for affective illness. The studies of VAN PRAAG and DE HAAN (1980) are also compatible with this hypothesis. They found that 33 of 54 depressed individuals with low post-probenecid accumulation of CSF 5-HIAA still showed this low accumulation after recovery from the episode of depression. These individuals (VAN PRAAG, 1980) tended to respond well to oral 5-hydroxytryptophan treatment (only one of 13 relapsed) compared to those individuals with normal post-probenecid 5-HIAA accumulation (five of seven relapsed). Additionally, VAN PRAAC and DE HAAN (1980) found a greatet number of hospital admissions for depression among the low 5-HIAA groups and among their relatives compared to the normal 5-HIAA group. However, most importantly, there was no report of higher incidence of illness among the relatives of the low 5-HIAA group. At present, we cannot exclude the possibility that some affective disorders are associated with a state-independent low CSF 5-HIAA after probenecid. We performed lumbar punctures on 25 lithium treated euthymic bipolar patients (12 of whom had repeat punctures after a minimum two weeks off lithium) and 25 carefully screened normal volunteers to determine whether various abnormalities reported in the ill state could be detected in the well state. That is, we attempted to determine whether examination of CSF biochemistry would yield state-independent trait markers for affective illness. Results are summarized in Table 2. Monoamines and their metabolites (NURNBERGER et al., 1983a) were measured by high-performance liquid chromatography with electrochemical detection. No significant differences (defined asp 5 0.01) were found between patients and controls for norepinephrine, 4-hydroxy-3-methoxyphenylglycol (MHPG) (a norepinephrine metabolite), homovanillic acid (HVA) and 3,4_dihydroxyphenylacetic acid (DOPAC) (two dopamine metabolites), or 5-HIAA. A strong correlation was found between CSF HVA and 5-HIAA for patients (unmedicated and lithium-treated) and controls (see Fig. !). A similar strong correlation was found by AGREN (1982) between CSF HVA and 5-HIAA (n = 97, r= 0.79, p < 0.0001). Patients (n=6) with a history of suicide attempts did not have lower CSF 5-HIAA than other patients or controls. GENETICS 339 ANDAFFECTIVE ILLNESS zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP TABLE~.CSFBIOCHEMISTRY INEUTHYMICBIPOLARS Unmedicated bipolars n=lO Controls II=20 Monoamines (pmol/ml) NE MHPG HVA’ DOPAC’ 5-HIAA+ Peptides/hormones (pg/ml) NTF /I-LPH a-MSH Cortisol VIP SRIF* AVP ACTHt /%EndorphinTrlz GABAS (pmol/ml) 0.32 45.3 180.0 2.6 75.0 427 77.0 8.7 5800.0 17.2 33.1 4.7 22.8 40.2 + + f + * 0.20 7.9 54.0 0.9 20.0 f 194.0 * 25.0 f 4.2 f 2000.0 k 8.4 * 14.0 zk 1.o f 5.8 f 6.9 127.0 f ANDCONTROLS 50.0 0.34 44.2 178.0 2.3 92.3 457.0 87.0 8.8 6700.0 20.2 43.8 4.7 22.3 39.2 + + f + k 0.21 7.0 72.0 0.9 41.4 -t 228.0 k 36.0 k 3.2 k 1600.0 k 7.6 k 18.0 k 2.1 f 5.4 k 6.8 129.0 k 36.0 Lithium-treated bipolars n=20 0.59 51.3 183.0 2.6 94.7 -c + -t + * 428.0 86.0 9.8 7400 15.0 29.6 4.5 21.4 37.8 k 217.0 ? 27.0 ? 3.0 ~3100.0 k 6.6 k 10.0 k 0.9 f 5.8 f 7.5 134.0 f 0.5 1 25.6 58.0 1.0 31.7 32.0 Values are mean -CSD. No significant (p < 0.01) group differences were found. *Corrected for height and weight. *Corrected for sex. *Corrected for age. For abbreviations. see text. Reports of decreased CSF vasopressin [AVP (GOLD et al., 1981)], CSF vasoactive [SRIF (RUBINOW intestinal peptide [VIP (GJERRIS et al., 1982)] and CSF somatostatin et al., 1982)] in depressed patients prompted us (Berrettini, W. H., Nurnberger, J. I., Reichlin, S., Zerbe, R. L., Simmons-Alling, S., Gershon, E.S., unpublished observation) to measure these peptides by RIA in euthymic patients. None of these substances differentiated euthymic patients from controls, suggesting that these reported abnormalities may be state dependent. GJERIUS et al. (1982) reported decreased CSF VIP concentration in a subgroup of atypical depressed patients, while endogenously depressed patients had values similar to controls. We found no differences between euthymic unmedicated bipolars and controls (Table 2). Differences in patient population may explain these results. A strong correlation was found among the patients and controls between CSF SRIF and CSF VIP (see Fig. 2). Lithium treatment seemed to decrease both CSF and plasma VIP (see Figs 3 and 4). A series of CSF substances associated with hypothalamic-pituitary-adrenal (HPA) axis function were measured, because a dysregulation of this axis (abnormal dexamethasone suppression test [DST]) commonly occurs in depression (CARROLL et al., 1981). This 340 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA WADE H.BERRETTINI~~~L zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFED 180 r . .. z 120- 1 E” . Sloe- 2 . . . y EOIn 60 . - l - . . l .. . . .:* l. l. 40 . :. . -. . ” . . . . . : l . . 1,,,1,,11 20-1 50 ,,,,,,,,,,j 90 130 170 210 250 WA FIG. 1. Correlation of CSF S-HIAA 290 330 370 410 450 (pmolimli and CSF HVA in normal volunteers and bipolar patients. 90 80 r 69 p- Ocol 70 n = 42 . 60 . 550 . z . 40 . . . . . . 30 20 . . . . . . . . . . . . . . . . . .. . . 10 . . 0 5 I I r I 6 7 8 9 I1 1 10 11 12 I I 13 14 I I , 15 16 17 , 18 1, 19 I 20 I I, 21 22 23 I, 24 25 26 VIP Fm. 2. Correlation of CSF SRIF and CSF VIP. , 1 27 28 I , 29 30 I I 31 32 , I 33 34 I, 35 36 341 Unmedicated Lrthrum-Treated p= 0.01 (pafred t test) FIO. 3. Lithium decreases plasma VIP, Unmed i cat ed Lithium-Treated p = (paired 0.0005 t test) Ra. 4. Lithium decreases CSF VIP. WADEHzyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJ . BERRETTINI e t a zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ l. 342 abnormality is thought to be state-dependent. It was hypothesized that the abnormal DST might be the result of an underlying genetic vulnerability marker for affective illness. The polypeptide pro-opiomelanocortin (POMC) is synthesized, released and metabolized in response to corticotropin releasing factor (CRF). POMC is cleaved to yield an N-terminal fragment (NTF), corticotropin (ACTH), beta-lipotropin (beta-LPH), alpha-melanocyte stimulating hormone (alpha-MSH), beta-endorphin and related peptides (CHR~TIEN et al., 1979). Plasma and CSF levels (as detemined by RIA) of substances produced by the HPA axis (NTF, beta-LPH, alpha-MSH, ACTH and beta-endorphin) were not different between patients and controls (see Table 2; BERRETTINI et al., 1984). These data suggest that the abnormal DST result in affectively ill patients is not associated with an underlying state-independent abnormality in HPA axis function. However, more work needs to be done to strengthen this argument. GABA studies Several measures of gamma-aminobutyric acid (GABA) metabolism were determined in this patient population. GOLD et al. (1980), GERNER and HARE (1981), and KASA et al. (1982) found low CSF GABA in unmedicated depressed patients. PETTY and SCHLESSER (1981) found low plasma GABA in some depressed patients on various medications. Lithium - Tre a te d Unm e dic a te d I \ p = ,003 \ + x= S.D. N= f p= 36. i i’ ,001 / L x= 1 197. S.D. 45. N x S.D. N FIG. C o ntro ls = f = 197. 30. 56. 171. Z!C = 34. 31. 5. Plasma GABA (pmoVm1) in euthymic bipolars. These studies prompted measurements of CSF GABA, which did not differentiate euthymic bipolars from controls (see Table 2). Platelet GABA-transaminase [GABA-T (BERRETTINI et al., 1982b)] and plasma GABA (BERRETTINI et al., 1982a, 1983) were found to be lower in euthymic bipolars (Figs 5 and 6). Both these measures were con- GENETI~SANDAFFECTIVE Lithium-Treated F? = SD. f N = lpmol/min/mg ILLNESS 343 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ Unmed i cot ed 11.8 ii = 6.0 S.D. f 28. N = 10.6 4.1 27. 5? = S.D. f N = 13.5 6.2 56. protein FIG. 6. Platelet GABA-T+ in euthymic bipolars and controls. cordant in monozygotic twins (Fig. 7, data for GABA-T not shown). This suggests that low plasma GABA may be a genetic vulnerability factor. Interestingly, we found that lithium treatment tends to increase plasma GABA (BERRETTINI et al., 1983), which may explain the contradiction between our data and those of PETTY and SHERMAN (1982), who found increased plasma GABA in lithium treated remitted bipolars, but lower plasma GABA in unipolar depressed patients who were not taking lithium. Family studies are required to determine whether plasma GABA differentiates ill from well relatives in pedigrees in which the proband has low plasma GABA. The origin or physiologic significance of plasma GABA is unknown. No significant correlation between plasma and CSF GABA was found, although pharmacologic blockade of GABA metabolism leads to increases in brain, CSF and plasma GABA (LOSCHER, 1979). ‘H-Imipramine binding to platelets Recently several studies of 3H-imipramine binding to platelets from depressed subjects have reported decreased maximal binding (B,~ without any change in affinity (RAISMAN et al., 1981; PAUL et al., 1982; ASARCH et al., 1980). USMAN et al. (1981) reported that the B,, remained low after recovery in a group treated with tricyclics. However, BERRETTINI et al. (1982~) and SURANYI-CADOTTE et al. (1982) found no difference between controls and euthymic patients who had been recovered for 2 months or more. Thus, it is controversial whether low B,, is a stateindependent marker. Most optimistically it may remain low in a subgroup of recovered patients, although all 10 depressed patients, studied by SURANYI-CADOTTE et al. (1982), showed robust increases in B,,, 2 months after recovery. This increase seemed to be independent of tricyclic medication. 344 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA WADE H. BERRETTINIet al. _B_ INTRACLASS I= 72 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQP P= 007 FIG. 7. Plasma GABA in monozygotic twins. Melatonin studies Melatonin is secreted by the pineal gland at high levels during dark hours. Levels are very low during daylight hours. Bright room lighting (2500 lux) at night will suppress melatonin secretion (LEWY et al., 1980). Normal room lighting at night (less than 500 lux) will not suppress melatonin secretion in most individuals, but LEWY et al. (1981) found suppression with 500 lux in two manic and two depressed patients. This led them to speculate that affectively ill patients are supersensitive to light’s effect on melatonin secretion. NURNBER~ER et al. (1983b) showed that this supersensitivity was detectable in 11 euthymic unmedicated patients, compared with 24 controls. Suppression of melatonin secretion was greater in the patients (p < 0.005). The neuronal pathway from retina to hypothalamus mediates this response. There is evidence that acetycholine may be the pertinent neurotransmitter in this tract (ZATZ and BROWNSTEIN, 1979). If the melatonin response is dependent on a cholinergic synapse, then this result is compatible with a state-independent cholinergic supersensitivity in affective illness, although this may be a nicotinic (not muscarinic) cholinergic response. Cholinergic REM induction A second series of studies suggests that cholinergic supersensitivity may be a genetic vulnerability marker in affective illness. SITARAM et al. (1980, 1982) showed that euthymic unmedicated patients entered their second REM sleep period significantly more rapidly @ < 0.001) than did controls following an intravenous 0.5 mg dose of arecoline (a cholinergic agonist), given 25 min after the end of the first REM period. Interestingly, this difference disappeared when a 1.0 mg dose of arecoline was used. At this dose controls entered the second REM period as fast as patients. NURNBERGER et al. (1983~) studied cholinergic REM induction in seven monozygotic twin pairs. Included in this sample were four twin pairs discordant for minor (n = 3) or major (n = 1) depression (criteria of MAZURE and GERSHON, 1979). The REM induction time following arecoline was concordant in these seven twin pairs (p < 0.02). One twin pair showed a rapid induction, but did not have any history of minor or major depression. In contrast to the arecoline induced second REM period, the endogenous time from first to second REM period was not concordant 0, = 0.44). These studies suggest that euthymic patients are supersensitive to the cholinergic GENETICSANDAFFE~TNEILLNESS 345 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO induction of REM sleep, and that this phenomenon is heritable. It remains to be determined whether the cholinergic REM induction test will differentiate ill from well relatives in pedigrees in which the probands show the cholinergic supersensitivity. A more recent observation of cholinergic supersensitivity was made by NADI et al. (1984) who showed that fibroblasts possess a muscarinic binding site. The number of binding sites on fibroblasts from bipolar patients and their ill relatives is significantly greater than the number of sites on fibroblasts from the well relatives and controls. Tyramine conjugation SANDLER et al. (1975) first reported that depressed patients excrete a decreased amount of conjugated tyramine in urine after an oral dose of 100 mg of tyramine. Subsequently, they reported that this phenomenon was detectable in recovered patients as well (BONHAM CARTER et al., 1978a). Their finding have been recently replicated in endogenously depressed outpatients before and after recovery (HARRISON et al., 1983). Interestingly, a much higher incidence of affective disorder (73.3%) was found among the lowest tyramine excreters compared to the highest excreters (26.7%) from among 77 pregnant women screened for tyramine excretion after oral challenge (BONHAM CARTER et cd., 1980). Platelet levels of the conjugating enzyme, phenolsulfotransferase, do not explain the difference (SANDLER et al.,1983), nor do changes in gastrointestinal motility seem to be responsible (BONHAM CARTER et al., 1978b). This deficit may be a trait marker for the illness, but demonstration of genetic control and pedigree studies are required. DISCUSSION JANOWSKY et al. (1972) proposed that mania and depression might be associated with an imbalance between cholinergic and adrenergic systems in the brain. Both the cholinergic REM induction test reports and the melatonin studies are compatible with a stateindependent genetically-determined cholinergic supersensitivity in affective illness. The hypothesis of cholinergic supersensitivity in affective illness has been given further support by the recent discovery that fibroblasts of bipolar patients and their ill relatives have significantly greater numbers of muscarinic cholinergic receptors than do controls (NADI et al., 1984). Thus, several lines of evidence are compatible with a cholinergic supersensitivity in affective illness. This represents a promising area of investigation in the genetics of affective disorders. A second promising area may be the study of euthymic patients who have low CSF 5-HIAA. Here again several lines of evidence support the possibility that low CSF 5-HIAA may be a genetic marker for vulnerability to affective illness. There are studies indicating that CSF 5-HIAA is heritable, that it remains low in some recovered patients and that it may be associated with violent suicide, a relatively common occurrence in depression. The arduous task of determining CSF 5-HIAA in pedigrees remains to be done. It would be quite interesting to study these individuals with other serotonergic probes, such as platelet serotonin uptake and ‘H-imipramine binding. Perhaps these are the individuals in whom RAWUN et al. (1981) find low numbers of ‘H-imipramine binding sites even after recovery. 346 WADE H. BERRETTINIet al. Other promising areas of investigation include plasma GABA and tyramine conjugation. Pedigree studies are needed for both these parameters. This paper has presented some promising areas of investigation in genetics of affective illness. This review is not intended to be comprehensive; other areas of intensive research in genetics of affective illness, such as lithium ratio (DORUS et al., 1983; PANDEY et al., 1979) have not been discussed. REFERENCES zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONML neurotransmission-state, traits and interactions. In Origins of (Edited by ANGST, J.). Springer, Berlin. ASARCH, K. B., SHIH, J. C. and KUBESAR, A. (1980) Decreased ‘H-imipramine binding in depressed males and females. Commun. Psychophormac. 4,425-430. BARON, M. (1977) Linkage between an X-chromosome marker (deutan color-blindness) and bipolar affective illness. Archsgen. Psychiat. 34721-725. BECKMAN, L., CEDEROREN, B., PERRIS, C. and STRANDMAN, E. (I978a) Blood groups and affective disorders. AOREN, H. (1982) Depression: Depression Current Concepts and altered ond Approaches Hum. Hered. 28,48-55. BECKMAN, L., PERRIS, C., STRANDMAN, E. and WAHLBY, L. (1978b) HLA antigens and affective disorders. Hum. Hered. 28,96-99. BERRETTPII, W. H., NURNBEROER, J. I., JR., HARE, T., GERSHON, E. S. and POST, R. M. (1982a) Plasma and CSF GABA in affective illness. Br. J. Psych&. 141,483-487. BERRETTINI, W. H., UMBERKOMAN-WIITA, B., NURNBEROER, J. I., JR., VOGEL, W. H., GERSHON, E. S. and POST, R. M. (1982b) Platelet GABA-transaminase in affective illness. Psychiot. Res. 7,255-260. BERRETTINI, W. H., NURNBEROER, J. I., JR., POST, R. M. and GERSHON, E. S. (1982~) Platelet ‘H-imipramine binding in euthymic bipolar patients. Psych&. Res. 7,215-219. BERRETTINI, W. H., NURNBEROER, J. I., JR., HARE, T., SIMMONS-ALLINO, S., GERSHON, E. S. and POST, R. M. (1983) Reduced plasma and CSF GABA in affective illness: effect of lithium carbonate. &of. Psych&. 18, 185-194. BERRETTINI, W. H., NURNBEROER, J. I., JR., CHROUSOS, G., SEIDAH, N. G., GOLD, P. W., CHR~TIEN, M., CHAN, J., SIMMONS-ALLINO, S. and GERSHON, E. S. (1984) Proopiomelanocortin-related peptides in CSF. Presented at the 1984 C.I.N.P. Meeting, Florence, 1984. BONHAM CARTER, S. M., REVELEY, M. A., SANDLER, M., DEWHURST, J., LITTLE, B. C., HAYWORTH, J. and PRIEST, R. G. (1980) Decreased urinary output of conjugated tyramine is associated with lifetime vulnerability to depressive illness. Psych&. Res. 3, 13-21. BONHAM CARTER, S., SANDLER, M., GOODWIN, B. L., SEPPINO, P. and BRIDGES, P. K. (1978a) Decreased urinary output of tyramine and its metabolites in depression. Br. J. Psych&. 132, 125-132. BONHAM CARTER, S., SANDLER, M., SEPPING, G. and BRIDGES, P. K. (1978b) Decreased conjugated tyramine outputs in depression: gastrointestinal factors. Br. J. c/in. Phormac. 5,269-271. CARROLL, B. J., FEINBERO, M., GREDEN, J. F., TARIKA, J. S., ALBALA, A. A., HASKELL, R. F., JAMES, N. MCI., KRONFOL, Z., LOHR, N., STEINER, M., DEVIONE, J. P. ard YOUNG, E. (1981) A specific laboratory test for the diagnosis of melancholia: standardization, validation and clinical utility. Archs gen. Psychiot. 38, 15-22. CH~~TIEN, M., BENJANNET, S., GOSSARD, G., GIANORLAKIS, C., CRINE, P., LIS, M. and SEIDAH, N. G. (1979) From beta-lipotropin to beta-endorphin and pro-opio-melanocortin. Can. J. Biochem. 57, 111 I-l 121. CLEROET-DARPOUX, F. (1982) Bias of the estimated recombination fraction and lod score due to an association between a disease gene and a marker gene. Ann. hum. Genet. 46,363-372. CLERGET-DARPOUX, F. (In press) Analysis for HLA-association with diseases: terminology, methodology and interpretation. In Compendium of Immunology, Vol. 4 (Edited by SCHWARTZ, L.). Van Nostrand, Reinhold, New York. CKOWE, R. R., NAMBOODIRI, K. K., ASHBY, H. B. and ELSTON, R. C. (1981) Segregation and linkage analysis of a large kindred of unipolar depression. Neuropsychobiology 7,20-25. DAY, N. E. and SIMONS, M. J. (1976) Disease susceptibility genes-their identification by multiple case family studies. Tissue Antigens 8, 109-l 19. DUNNER, D. L., FLEISS, J. L., ADDONIZIO, G. and FIEVE, R. R. (1976) Assortative mating in primary affective disorders. Biol. Psychiaf. 11,43-51. EGELAND, J. (1980) Affective disorders among the Amish: 1976-1980. Presented at the Annual Meeting of the American College of Neuropsychopharmacology, Puerto Rico, December. FIEVE, R. R., MENDLEWICZ, J. and FLEISS, J. L. (1973) Manic-depressive illness: linkage with the Xg blood group. Am. J. Psychiut. 130, 1355-1359. GENETICS AND AFFECTIVE ILLNESS FLEMENBAUM,A. and LARSEN, J. W. (1976) ABO-Rh blood 347 groups and psychiatric diagnosis: a critical review. Dis. nerv. Syst. 31,581-583. GERNER, R. H. and HARE, T. A. (1981) CSF GABA in normals, depression, schizophrenia, mania and anorexia nervosa. Am. J. Psychiat. 138,1098-l 101. GERSHON, E. S., DUNNER, D. L., STURT, L. and GOODWIN, F. K. (1973) Assortative mating in the affective disorders. Biol. Psychiat. 1,63-74. GERSHON, E. S., MARK, A., COHEN, N., BELIZON, N., BARON, M. and KNOBE, K. E. (1975) Transmitted factors in the morbidity of affective disorders: a controlled study. J. psychiat. Res. 12,283-299. GERSHON, E. S. and BUNNEY, W. E., JR. (1977) The question of X-linkage in bipolar manic-depressive illness. J. psych&. Res. 12.301-317. GERSHON, E. S., TARGUM, S. D., MATTHYSSE, S. and BTJNNEY, W. E., JR. (1979) Color blindness not closely linked to bipolar illness. Archsgen. Psychiat. 36,1423-1434. GERSHON, E. S. and LIEBO~ITZ, J. H. (1979) Affective illness. In Genetic Diseases Among Ashkenazi Jews (Edited by GOODM A N,R. M . and MOTULSKY, A. G.), pp. 343-352. Raven Press, New York. GERSHON, E. S., MENDLEWICZ, J., GASTPAR, M., BECH, P., GOLDIN, L. R., KIELHOLZ, P., RAFAELSEN, 0. J., VARTANIAN, F. and BUNNEY, W. E., JR. (1980) A collaborative study of genetic linkage of bipolar manicdepressive illness and red/green colorblindness. Actapsychiat. stand. 61,3 19-338. GJERRIS, A ., FAHRENKRUG, J., HAMMER, M., SCHAFFALITZKYDE MUCKADELL, 0. P., BOJHOLM, S., VENDSBORG, P. and RAFAELSEN, 0. J. (1982) In Proceedings of the 3rd World Congress of Biological Psychiatry (Edited by JANSSON, B., PERRIS, C. and STREW, G.), Abstract 237. Elsevier, Amsterdam. GOETZL, V., GREEN, R., WHYBROW, P. and JACKSON, L. (1974) X-linkage revisited: a further family study of manic-depressive. Archsgen. Psychiat. 31,665-672. GOLD, B. I., BOWERS, M. B., JR., ROTH, R. H. and SWEENEY, D. W. (1980) GABA levels in CSF of patients with psychiatric disorders. Am. J. Psychiat. 137,362-364. GOLD, P. W., GOODWIN, F. K., POST, R. M. and ROBERTSON, G. L. (1981) Vasopressin function in depression and mania. Psychopharmac. Bull. 17,7-10. GOLDIN, L. R., CLER~ET-DARPOUX, F. and GERSHON, E. S. (1982) Relationship of HLA to major affective disorder not supported. Psychiat. Res. 7,29-45. GOLDIN, L. R., GERSHON, E. S., TARGUM, S. D., SPARKES, R. S. and MCGINNIS, M. (1983) Segregation and linkage analyses in families of patients with bipolar, unipolar, and schizoaffective mood disorders. Am. J. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA hum. Genet. 35,274-287. GOVAERTS, A., MENDLEWICZ, J. and VERBANK, P. (1977) Manic-depressive illness and HLA. Tissue Antigens 10,6&62. GREEN, J. R. and WOODROW, J. C. (1977) Sibling method for detecting HLA-linked genes in disease. Tissue Antigens9,31-35. HARRISON, W. M., COOPER, T. B., QUITKIN, F. M., LEIBOWITZ, M. R., MCGRATH, P. J., STEWART, J. W. and KLEIN, D. F. (1983) Tyramine sulfate excretion test in depressive illness. Psychopharmac. Bull. 19, 503-504. JAMES, N. M., CARROLL, B. J., HAINES, R. F. and SMOUSE, P. E. (1979) Affective disorders: ABO and HLA. In Genetic Aspects of Affective Zlbzess (Edited by MENDLEWICZ, J. and SHOPSIN, B.), pp. 35-44. Spectrum Publications, New York. JAM ES, N. M ., SM OUSE, P. E., CARROLL, B. J. and HAINES, R. F. (1980) Affective illness and HLA frequencies: no compellling association. Neuropsychobiology 6,208-216. JANOW SKY, D. S., EL-YOUSEF, M. K. and DAVIS, J. M. (1972) Cholinergic-adrenergic hypothesis of mania and depression. Lancet ii,632. JOHNSON. G. F. S. (1978) HLA antigens and manic-depressive disorders. Biol. Psychiat. 13, w-412. KASA, K.‘, SA~JRO; ~.,‘MITSUTOSH~, Y., MITSUMOTO; S., KURODA, H. and O&WA, N. (1982) CSF GABA and HVA in depressive disorders. Biol. Psychiat. 17,877-884. KOSLOW, S. H., MAAS, J. W., BOWDEN, C. L., DAVIS, J. M., HANIN, I. and JAVA~D, J. (1983) CSF and urinary biogenic amines and metabolites in depression and mania: a controlled univariate analysis. Archs gen. Psychiat. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 40,999. KRUGER, S. D., TURNER, W. J. and KIDD, K. K. (1982) The effects of requisite assumptions on linkage analyses of manic-depressive illness with HLA. Biol. Psychiat. 17, 1081-1099. LECKMAN, J. F., GERSHON, E. S., MCGINNISS, M. H., TARGUM, S. D. and DIBBLE, E. D. (1979) New data do not suggest linkage between the Xg blood group and bipolar illness. Archsgen. Psychiat. 36,1435-1441. LEVITAN, M. and MONTAGU, A. (1971) A Textbook of Human Genetics. Oxford University Press, London. LEWY, A. J., WEHR, T. A. and GOODWIN, F. K. (1980) Light suppresses melatonin secretion in humans. Science210,1267-1270. WADE H. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF BERRETTINIet al. 348 LEV~Y, A. J., WEHR, T. A., GOODWIN, F. K., NEUSOME, D. A. and ROSENTHAL, N. E. (1981) Manic-depressive patients may be supersensitive to light. Luncet ii, 303-304. L&CHER, W. (1979) GABA in plasma and CSF of different species. Effects of y-acetylenic GABA, y-vinyl GABA, and sodium valproate. J. Neurochem. 32,1587-1891. MAZURE, C. and GERSHON, E. S. (1979) Blindness and reliability in lifetime psychiatric diagnosis. Archs gen. P&riot. 127, l-10. MENDLE~~CZ, J., FLEISS, J. L., FIEVE, R. R. (1972) Evidence for X-linkage in the transmission of manic-depressive illness. J. Am. med. Ass. 222.1624-1627. MENDLE\KICZ, J., MA~SART-GUIOT, T., W~LMOTTE, J. and FLEISS, J. L. (1974) Blood groups in manic-depressive illness and schizophrenia. Dis. nerv. Syst. 35,39-41. MENDLEWICZ, J., FLEtss, J. L. (1974) Linkage studies with X-chromosome markers in bipolar (manic-depressive) and unipolar (depressive) illness. Biol. Psychiut. 9,261-264. MENDLEWICZ, J. (1977) Genetic studies in schizoaffective illness. In The Impact of Biology on Modern Psychiutry (Edited by GERSHON, E. S., BELUKER, R. H., KETY, S. S. and ROSENBAUM, M. R.), pp. 229-240. Plenum Press, New York. MENDLEU~ICZ, J., L~KOWSKI, P. and WUMOTTE, J. (1980) Linkage between glucose 6-phosphate dehydrogenase deficiency and manic-depressive psychosis. Br. J. Psychiut. 134,337-342. MENDLE\KICZ, J., VERBANCK, P., LINKOWSKI, P. and G~VAERTS, A. (1981) HLA antigens in affective disorders and schizophrenia. J. &feet. Dis. 3, 17-24. MORTON, L. A. and KIDD, K. K. (1980) The effects of variable age of onset and diagnostic criteria on the estimates of linkage. An example using manic-depressive illness and color blindness. Sot. Biol. 27, l-10. MORTON, N. E. (1955) Sequential tests for the detection of linkage. Am. J. hum. Genet. 7,277-318. MORTON, N. E. (1956) The detection and estimation of linkage between the genes for elliptocytosis and the Rh blood group. Am. J. hum. Genet. 8,80-96. NADI, N. S., NURNBEROER, J. I., JR. and GERSHON, E. S. (1984) Muscarinic cholinergic receptors on skin fibroblasts in familial affective disorders. New Engl. J. Med., in press. NEORI, F., MELLICA, A. M., ZULICPTI, R., GASPERINI, M., MARCIARDI, F. and SMERUDI, E. (1981) Genetic implication in assortative mating of affective disorders. Br. J. Psychiut. 138,236-239. NURNBEROER, J. I., JR. and GERSHON, E. S. (1982) Genetics of affective disorders. In Hundbook of Mfective Dborders (Edited by PA~~~EL, E.), pp. 126-145. Churchill Livingstone, London. NURNBEROER, J. L., JR., BERRETTINI, W. H., SCHEININ, M., SIMMONS-ALLINO, S., LINNOILA, M. and GERSHON, E. S. (1983a) CSF monoamine metabolites in “well state” bipolar patients and normal volunteers. Abstracts of the 1983 Annual Meeting of the Society for Biological Psychiatry, No. 58, p. 92. NURNBEROER, J. I., JR., LEWY, A. J., PACK, D., BECKER, L., WEHR, T. A., POHZLL, R. L., SIMMONS-ALLINO, S. and GERSHON, E. S. (1983b) Bipolar patients are supersensitive to light. Abstracts of the 1983 Annual Meeting of the American Psychiatric Association. NURNBEROER. J. I.. JR.. SIT&AM, N.. GERSHON, E. S. and GILLIN. J. C. (1983~) A twin study of cholinernic REM induction. biol. Psychiut. i8, i 161-1173.. OTT, J. (1974) Estimation of the recombination fraction in human pedigree: efficient computation of the likelihood for human linkage. Am. J. hum. Genet. x.588-597. PANDEY, G. N., TORUS, E., DALLY,J. M. and TOSTESON, D. C. (1979) Lithium transport in human red blood cells. Genetic and clinical aspects. Archsgm. Psychiut. 36.902-908. PARKER, J. B., THEILLIE, A. and SPIELBEROER, C. D. (1961) Frequency of blood types in an homogeneous group of manic-depressive patients. J. ment. Sci. 107,936-942. PENROSE, L. S. (1935) The detection of autosomal linkage in data which consists of pairs of brothers and sisters of unspecified parentage. Ann. Eugen. 6,133-138. PENROSE, L. S. (1953) The general purpose sib-pair linkage test. Ann. Eugen. 18, 120-124. PETTY, F. and SCHLESSER, M. A. (1981) Plasma GABA in affective illness. J. clffect. Dis. 3,334-343. PETTY, F. and SHERMAN, A. D. (1982). Plasma GABA: a blood test for bipolar affective disorder trait? Res. Commun. Psychol. Psychiut. Behuv. 7,431-W. PAUL, S. M., REHAVI, M., SKOLNICK, P., BALLENOER, J. C. and GOODWIN, F. K. (1982) Depressed patients have decreased binding of tritiated imipramine to platelet serotonin “transporter”. Archs gen. Psychiut. 38, 1315-1320. RAISMAN,R., SECHTER, D., BRILEY, M. S., ZARIRAN, E. and LANCER, S. Z. (1981) High-affinity binding in platelets from untreated and treated depressed patients compared to healthy 3H-imipramine volunteers. Psycho- phurmucology 75,368-372. RIEDER, R. 0. and GERSHON, E. S. (1978) Genetic strategies in biological psychiatry. Archs gen. Psych&. 35, 866-873. P. M., STEFANIS, C. N., LYKOURAS, E. P. and VARSOU, E. K. (1979) Affective blood types. Actu Psychiut. Scund. 60,272-278. RINIERIS, disorders and ABO GENETICS ANDAFFECTIVE ILLNESS 349 Rrscrr, N. and BARON,M. (1982) X-linkage and genetic heterogeneity in bipolar-related major affective illness: reanalysis of linkage data. Ann. hum. Genet. 46,153-166. Runprow, D. R., GOLD, P. W., POST, R. M., BALLENOER, J. C., COWDRY,R., B~LLINOER,J. and REICHLIN,S. (1983) CSF somatostatin in affective illness. Archsgen. Psychiat. 40.409-412. SANDLER,M., Bow CARTER,S., CUTHBERT,M. F. and PARE, C. M. B. (1975) Is there a decrease in monoamineoxidase activity in depressive illness? Loncet 1,1045-1048. SANDLER,M., B~NWM CARTER,S. M. and WALKER,P. L. (1983) Tyramine conjugation deficit as a trait-marker in depression. Psychopharmac. Bull. 19,501-502. SEDVAU,G., Fvao, B., GUL.LBERO, B., MYBACK, H., WIESEL, F.-A. and WODE-HELOODT, B. (1980) Relationships in healthy volunteers between concentrations of monoamine metabolites in cerebrospinal fluid and family history of psychiatric morbidity. Br. J. Psychiat. l&366-374. SEDVALL, G. and OWNSTIERNA,G. (1981) Genetic and environmental influences on central monoaminergic mechanisms in man. Abstract No. 5344, Third World Congress of Biological Psychiatry, Stockholm. SHAPIRO,R. W., BOCK, E., RAFAELSEN, 0. J., RYDER,L. P. and A. (1976) Archsgen. R. W., RAFAELSEN, 0. J., L. P., A. and SORENSEN, H. (1977a) in unipolar J. Psychiut. R. W., L. P., A. and 0. of SINISCALCO, M., SZABO,P., FILIPPI, F. and RnULDI, A. (1982) Combination of old and new strategies for the molecular mapping of the human X-chromosome. In Human Genetics, Part A: The Unfolding Genome (Edited by B~NNE-TA~QI~, B.), pp. 103-124. Alan R. Liss, New York. SITARAM,N., NURNBEROER, J. I., JR., GERSHON, E. S. and GALLON, J. C. (1980) Faster cholinergic REM sleep induction in euthymic patients with primary affective illness. Science 208,200-202. SITARAM,N., N~J~NBEROER, J. I., JR., GERSHON,E. S. and GILLIN,J. C. (1982) Cholinergic regulation of mood and REM sleep: potential model and marker of vulnerability to affective disorder. Am. J. Psychiat. 139, 571-576. SMERUDI, E., NEO~U,F., MELICA,A. M., SCORW-SMEULDI,R., FABIO,G., BONARA, P., BBLLODI, L., SACCHE~, E., S~BA~INI-VILLA,M. G., CU~JLLO, C. L. and ZANUSSI,C. (1978a) HLA typing and affective disorders: a study in the Italian population. Neuropsychobiology 4,344352. SME~ULDI, E., NEORI,F., MELICA,A. M. and SCORW-SMERUDI,R. (1978b) HLA system and affective disorders: a sibship genetic study. Tissue Antigens 12,270-274. SMEULDI, E. and BELLODI,L. (1981) Possible linkage between primary affective disorder susceptibility locus and HLA haplotypes. Am. J. Psychiat. 138.1232-1234. SMOUSE,P. E. and W~L~AMS,R. C. (1982) Multivariate analysis of HLA-disease associations. Biometrics 38, 757-768. SPENCE,M. A. (1980) Linkage methods in psychiatric disorders. In Genetic Strategies for Psychobiology and Psychiatry (Edited by GERSHON, E. S., MATTHYSSE,S., BREAKEFIELD,X. 0. and CLUUNELLO,R. D.), pp. 295-304. Boxwood Press, Pacific Grove, CA. STEMBER,R. H. and Fmva, R. R. (1977) Histocompatibility antigens in affective disorders. C/in. Immunol. Immunopath. 7,10-14. SUAREZ,B. K. (1978). The affected sib-pair IBD distribution for HLA linked disease suspectibility loci. Tissue Antigens 12.87-93. SUAREZ, B. K., RICE, J. and REICH, T. (1978) The generalized sib-pair IBD distribution: its use in the detection of linkage. Ann. hum. Genet. 42,87-94. SUAREZ,B. K. and CROUOHAN,J. (1982) Is the major histocompatibility complex linked to aenes that increase susceptibility to affective disorders? A critical appraisal. Psychiat. Res: 7,19127. SUAREZ,B. K. and VAN EERDEWEOH.P. (1981) Tvue I insulin-deuendent diabetic mellitus. Diabetoloaia _ 20.I 524-529. SUIUNYI-CADOTTE, B. E., WOOD, P. L., NAIR, N. P. V. and SCHWARTZ,0. (1982) Normalization of platelet 3H-imipramine binding in depressed patients during remission. Eur. J. Pharmac. 85,357-358. SWANN,A. C., SECUNDA,S., DAMPS,J. M., ROBINS,E., HANIN,I., KOSLOW,S. H. and MMS, J. W. (1983) CSF monoamine metabolites in mania. Am. J. Psychiot. 140,396. TANNA,V. L. and WINOKUR, 0. (1968) A study of association and linkage of ABO blood types and primary affective disorders. Br. J. Psychiot. 114, 1175-l 181. TANNA,V. L., WINOKUR,G., ELSTON,R. C. and Go, R. C. P. (1976) A linkage study of depression spectrum disease. The use of the sib-pair method. Neuropsychobiology 2,52-62. TANNA,V. L., Go, R. C. P., WINOKUR,G. and ELSTON,R. C. (1979) Possible linkage between a-haptoglobin and depression spectrum disease. Neuropsychobiology 5,102-l 13. 350 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA WADE H . BERRETTINIet al. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA TARGUM, S. D., GERSHON, E. S., VAN EERDEWEGH, M. and ROOENTINE, N. (1979) Human leukocyte antigen system not closely linked to or associated with bipolar manic-depressive illness. Biol. Psychiat. 14,615-636. TEMPLE, H., DUPONT, B. and SHOPSIN, B. (1979) HLA antigen and affective disorders: a report and critical assessment of histocompatibility studies. Neuropsychobiology $50-58. THOMSON, G. (1980) A review of theoretical aspects of HLA and disease associations. Theoret. Population Biol. 20,168-208. THOMSON, Cl. and BODMER, W. F. (1977a) The genetics of HLA and disease associations. In Measuring Selection in Natural Popuhtions (Edited by CHRISTIANSEN, F. B., FENCHAL, T. and BARNDORFF-NIELSON, 0.). pp. 545-564. Springer, Berlin. THOMSON, G. and BODMER, W. F. (1977b) The genetic analysis of HLA and disease associations. In HLA and Disease (Edited by DAUSSET, J. and SVEJGAARD, A.), pp. 84-93. Williams&Wilkins, Baltimore. TRASKMAN, L., ASBERG, M., BERTILSSON, L. and SJOSTRAND, L. (1981) Monoamine metabolites in CSF and suicidal behavior. Archsgen. Psychiat. 38,631-636. TURNER, W. J. and KING, S. (1981) Two genetically distinct forms of bipolar affective disorder? Biol. Psychiat. 16,417-439. VAN PRAAG, H. M. (1980) Central monoamine Compreh. Psychiat. 21,30-43. metabolism in despresions--I. VAN PRAAG, H. M. and DE HAAN, S. (1980) Central serotonin vulnerability. Actapsychiat. stand. Suppl. 280, pp. 89-103. deficiency-a Serotonin factor and related compounds. which increases depression VESTERGAARD, P., SORENSEN, T., HOPPE, E., RAPHAELSEN, 0. J., YATES, C. M. and NICOLAOU, N. (1978) Biogenic amine metabolites in cerebrospinal fluid of patients with affective disorders. Actapsychiatscand. 58, 88-96. WEITKAMP, L. R., PARDUE, L. H. and HUNTZINGER, R. S. (1980) Genetic marker studies in a family with unipolar depression. Archsgen. Psychiat. 37, 1187-1192. WEITKAMP, L. R., STANCER, H. C., PERSAD, E., FLOOD, C. and GUTTORMSEN, S. (1981) Depressive disorders and HLA: a gene on chromosome 6 that can affect behavior. NewEngl. J. Med. 305,1301-1306. WINOKUR, G., CLAYTON, P. J. and REICH, T. (1969) ManicDepressive Illness. Mosby, St. Louis. WINOKUR, G. (1972) Depression spectrum disease: description and family study. Compreh. Psychiat. 13, 3-8. ZATZ, M. and BROWNSTEIN, M. J. (1979) Intraventricular carbachol mimics the effects of light on the circadian rhythm in the rat pineal gland. Science 203,358-360.