- Email: [email protected]
Close linkage of bipolar disorder to chromosome 11 markers is excluded in two large Australian pedigrees
23
Journal of A/fective Disorders, 21 (1991) 23-32 0 1991 Elsevier Science Publishers B.V. (Biomedical Division) 01650327/91/$03.50 ADONIS 016503279100054Q
JAD 00767
Close linkage of bipolar disorder to chromosome 11 markers is excluded in two large Australian pedigrees Philip Mitchell ’ School of Psychiatv,
‘, Brent Waters ‘, Nigel Morrison *, John Shine *, Jennifer and John Eisman *
Donald
3
University of New South Wales, Sydney, Australia, ’ Garvan Institute of Medical Research, St. Vincent’s Hospital, Sydney, Australia and ’ School of Biological Sciences, Macquarie University, Sydney, Australia (Received 4 July 1990) (Revision received 29 August 1990) (Accepted 24 September 1990)
Summary The relationship between bipolar disorder and chromosome 11 markers remains uncertain. Whilst re-analysis of the Amish pedigree weakened previous evidence for close linkage (but could not exclude the possibility of genetic heterogeneity), a recent French study has found a significant association between this condition and tyrosine hydroxylase polymorphisms. We aimed to determine if bipolar disorder in two large Australian pedigrees (of Irish and English extraction respectively) was linked to these markers. Of the 84 family members available for testing, nine were diagnosed as bipolar I, one as bipolar II and six had recurrent unipolar depression. Linkage of bipolar disorder and recurrent depression to the chromosome 11~15 markers c-Harvey ras, insulin and tyrosine hydroxylase was tested using a series of genetic models with varying penetrance levels. Additionally, linkage was examined using a series of levels of definitions of affective status (ranging from bipolar I alone to all affective illnesses). Close linkage to these markers was strongly excluded using each model and definition. The findings also persisted when a wide range of rates of ‘sporadic’ (non-genetic) presentations of illness were incorporated in the analysis. These results are consistent with other recent studies indicating that bipolar disorder is not linked to chromosomal region llp15.
Key words: Bipolar disorder; Chromosome 11; Linkage
Introduction Address for correspondence: Dr. Philip Mitchell, Mood’ Disorders Unit, Division of Psychiatry, Prince Henry Hospital, Little Bay, NSW 2036, Australia.
Accumulated evidence from family, twin and adoption studies (Gershon, 1990) has confirmed
24
the clinical observations of 19th- and early 20thcentury psychiatrists (e.g., Kraepelin, 1921) that there is significant genetic transmission of both bipolar disorder and recurrent depressive illness. Although there is considerable evidence for monogenie transmission, the mode of transmission is still controversial (Rice et al., 1987; Cox et al., 1989; Gershon and Goldin, 1989). Linkage studies (which assume monogenic transmission) have been used to identify the chromosomal location of the putative responsible gene(s). The development of recombinant DNA techniques, and in particular the demonstration of restriction fragment length polymorphisms (RFLPs) (Kan and Dozy, 1978), has enabled the development of a detailed chromosomal map, facilitating testing for linkage of disease genes to specific genes or chromosomal regions (Botstein et al., 1980). The present situation concerning molecular studies in bipolar disorder must be regarded as uncertain (Owen and Mullan, 1990). In 1987, Egeland et al. reported tight linkage (a lod score of 4.08 at a recombination fraction of 0.00) of bipolar affective disorder to the c-Harvey-ras oncogene (HRAS), and loose linkage to the insulin gene (INS) - genes sited on the short arm of chromosome 11 (region 11~15) - in a large U.S. Amish pedigree. Hodgkinson et al. (1987) and Detera-Wadleigh et al. (1987) however, reported the exclusion of close linkage in several Icelandic and North American bipolar pedigrees respectively. A subsequent short report by Gill et al. (1988) similarly excluded close linkage in a large Irish pedigree. Wesner et al. (1990) also found no evidence for linkage in a bipolar pedigree. Their family, however, only included one bipolar I subject, and the results were highly dependent upon assumed penetrance levels. A reanalysis of the Amish pedigree by Kelsoe et al. (1989) however, reduced the likelihood of linkage of bipolar disorder to the chromosome 11 markers. Firstly, addition of genotypes from some previously untyped members and inclusion of two new presentations of illness in the original ‘core’ pedigree reduced the statistical probability of linkage. Secondly, inclusion of two lateral extensions of the ‘core’ pedigree led to exclusion of close linkage to this region. Although the authors of the re-analysis considered it most likely that the find-
ing of linkage in the original ‘core’ pedigree was due to chance, they could not discount the possibility of genetic heterogeneity (with regard to bipolar disorder) in the Amish population (Kelsoe et al., 1989). The latter explanation is possible even in a community as genetically isolated as the Amish, as firstly several progenitor couples had psychiatric illness among their first- and seconddegree descendants (Kelsoe et al., 1989) and secondly a Mennonite individual married into one of the earlier generations of the particular family involved in these studies (Kelsoe et al., 1990). The possibility of genetic heterogeneity has been highlighted by a recent population association (nor linkage) study of 50 unrelated French bipolar patients and 50 matched controls (Leboyer et al., 1990) . That investigation found a significant association between tyrosine hydroxylase (TH) RFLPs and bipolar disorder (P < 0.01) suggesting that liability to bipolar disorder may be dependent upon a defect within either the TH gene, or its regulatory domain. (The TH gene is in chromosomal region 11~15 near the HRAS and INS genes.) Todd and O’Malley (1989) had previously reported no association between TH RFLP alleles and bipolar disorder, but they only investigated 18 subjects. In the light of these discrepant findings we report a study which aimed to examine w,hether linkage exists between bipolar affective disorder in two large Australian pedigrees and the 11~15 markers.
Methods Case ascertainment and diagnosis We sought families with a large number of relatives, a single disease source, illness over two or three generations, a large number of affected family members and large sibships. Families were identified through the Mood Disorders Unit, Prince Henry Hospital, Sydney: the New South Wales Manic Depressive Self-Help Group; and local psychiatric units. In the 35 bipolar families identified, details of pedigrees were recorded using the method described by Thompson (1979). Families were assessed using the Family History-RDC (FH-RDC)
0:: P::
9:: 1:: ::
26
method of Andreasen et al. (1977) with interviews performed with both the proband and another informative family member. From the families contacted, we identified two large families with three generations of illness and a large number of affected individuals. All available relatives were interviewed by P.M. using the Schedule of Affective Disorders and Schizophrenia-Lifetime Version (SADS-L). Those interviews were used to derive Research Diagnostic Criteria (RDC) diagnoses using the strict NIMH criteria for depression which require at least 1 month’s duration of symptoms, and functional impairment or incapacity. In conjunction with the SADS-L interview, all marrying-in individuals were routinely questioned about any family history of psychiatric disorder, to ensure unilateral descent of illness in the pedigrees. Hospital and other medical records were obtained. Final diagnoses were made using the Yale-NIMH ‘Best Estimate’ Diagnosis Consensus guidelines which are assigned from FH-RDC and SADS-L derived RDC diagnoses, and medical records. ‘Best estimate’ diagnoses were also made independently by two other psychiatrists (B.W., Ian Hi&e) after assessment of all available interviews and records. These psychiatrists were aware of the disorder
under study. Diagnoses were made blind to laboratory investigations. Family one was a large family of Irish descent. Sixty-eight of the 78 members at risk were interviewed with SADS-L and gave blood for DNA testing. At the time of assessment there were 11 living affected individuals (one dying subsequently), five of whom were diagnosed as bipolar (four bipolar I, one bipolar II), and six had recurrent unipolar depression (NIMH criteria), each of whom fulfilled RDC for endogenous depression. The age of onset ranged from 12 to 47 years. Five of the individuals in this pedigree (bipolar and unipolar) had a late-onset presentation of illness (> 40 years). Ten other individuals were diagnosed as having either single episodes of major depression or minor depression. Family two was a family of English descent with seven affected individuals. Sixteen of the 20 members at risk were assessed. There were seven affected individuals (five available for testing), each with bipolar I disorder. There were no individuals with recurrent unipolar depression in this particular family. The age of onset ranged from 19 to 40 years. For the linkage analyses only subjects of at least 15 years of age were included. These pedigrees are outlined in Figs. 1 and 2.
BD
BD
CD
CD
XY
XY
XY
YZ
-0
El--
Tici DD XY
6 <4 CD XY
00 DD YY
DD XY
DD YZ
cr’: DD XY
DD xx
DD xx
Fig. 2. Pedigree 2. Circles = females; squares = males; diamonds = males or females. Full hatch = bipolar disorder: half hatch = recurrent major depressive disorder. Alleles are in the following order: HRAS, TH. Sex and birth order have been disguised to protect anonymity.
27
Genetic markers DNA extraction. Blood (60 ml) was collected into sodium heparin tubes (Becton Dickinson) and immediately centrifuged in a clinical centrifuge to separate the buffy coat. The buffy coat was further separated by repeated rounds of centrifugation through normal saline until the leukocytes were largely free of red blood cells. The cellular pellet was then resuspended in red blood cell lysis buffer for 20 min at room temperature. After red cell lysis had occurred, the leukocytes were then collected by centrifugation and resuspended evenly in 3 ml of normal saline. DNA was then liberated from the cells by lysis after addition of an equal volume of leukocyte lysis buffer (10 mM Tris-HCl, pH 7.4; 1 mM EDTA; 0.2% sodium dodecyl sulfate (Biorad Inc.)). The lysate was then treated with proteinase K (International Biotechnologies Inc.) at 50 pg/ml overnight at 37” C or for 4 h at 65 o C. DNA was then purified by repeated solvent extraction using phenol-chloroform and chloroform, essentially according to Maniatis et al. (1980). DNA was precipitated without the addition of further salts or low temperature treatment by the addition of two volumes of ethanol. Precipitated high molecular weight DNA was either removed by pipette or centrifuged at 5000 X g (Beckman G6-M) for 20 min. DNA pellets were washed in 70% ethanol followed by 100% ethanol and vacuum dried. DNA was redissolved in 1.0 ml TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and quantitated by spectroscopy at 260 nm wavelength. DNA was digested with various restriction endonucleases (all from Pharmacia) according to the manufacturer’s instructions. Southern hybridization analysis. Restricted DNA was resolved on 1.0% agarose (International Biotechnology Inc.) gels in TPE buffer essentially as described by Maniatis et al. (1980). DNA was transferred to Biorad Zetaprobe nylon membranes using the alkaline denaturation protocol recommended by the manufacturer. Plasmid DNA for probe preparation was made by alkaline lysis followed by cesium chloride ultracentrifugation (40 krpm for 40 h in a Beckman Ti-70 rotor) as described by Maniatis et al. (1980). Insert DNA from plasmid was prepared by elution from agarose gel chromatographically separated frag-
ments of restriction digested plasmid DNA. Radioactive insert probes were made by standard nick translation using a kit (BRL, Bethesda, MD). Incorporated radioactivity was separated from unincorporated by chromatography over a Pharmacia nick column. Probes were labelled to a specific activity between 2 and 5 x lo9 dpm/pg. Probes were hybridized to 3 x Cot l/2 values, calculated based on the complexity of the probe, at a concentration of 10 ng/ml using 0.5 ml of hybridization fluid per cm* of Zetaprobe. Hybridization and washing protocols were as described by Biorad for the Zetaprobe membrane without the use of dextran sulfate. Filters were exposed to Kodak XARS film with two DuPont Cronex Lightningplus intensifying screens for between 2 and 5 days. The probes used to detect RFLPs were the pEJ6.6 probe for HRAS (Krontiris et al., 1985) phins310 for INS (Bell et al., 1982) and a 1.5 kb EcoRI fragment of a TH cDNA clone (pCB-05) for TH (Tinklenberg et al., 1988). Multiple alleles were observed for each locus after restriction enzyme digestion. In accordance with previous studies, we observed: four alleles for HRAS with Sac1 (as described by Egeland et al., 1987) i.e., A (- 6.8 kb), B (- 6.2 kb), C (- 5.7 kb), D (- 5.2 kb); six alleles for INS with PuuII: 1 (- 2.4 kb), 2 (- 2.1 kb), 3 ( - 0.85 kb), 4 ( - 0.80 kb), 5 ( - 0.74 kb), 6 (- 0.68 kb) (Willard et al., 1985); and three for TH with BgfII: X (- 8.4 kb), Y (- 6.9 kb) and Z (- 8.0 kb) (Moss et al., 1986; Korner et al., 1988). The genotypes of family members for these markers are shown in Figs. 1 and 2. For technical reasons insulin RFLP data are not available for family 2. Immortalized cell lines for pedigree numbers are currently being established. Linkage analysis Linkage analyses were undertaken using the computer program LIPED (Ott, 1974). Lod scores > 3 are considered to be indicative of linkage and scores < -2 indicate rejection of linkage at that recombination fraction. (The term ‘lod’ is an acronym for logarithm (base 10) of the odds favoring linkage in a particular set of pedigree data. Although a lod score of 3 may suggest a probability of 0.001, this score in fact equates with a probability of 0.05 for two genes for which there is no a priori knowledge that they are linked
28 TABLE
1
PARAMETERS Model
OF THE FOUR Age-specific
Age: I” II * Age: III
b
IV c
GENETIC
MODELS
penetrance
15-19
20-29
30-39
40+
0.12 0.20
0.30 0.50
0.45 0.75
0.60 1.00
15-19
20-24
25-29
29+
0.10 0.18
0.27 0.42
0.56 0.73
0.63 0.85
’ Age of onset curve derived from combined ages of onset for families b Age of onset curve of total Amish sample. ’ Age of onset curve of Amish pedigree 110.
(Gershon, 1990)). Analyses were performed asdominant transmission suming an autosomal model with varying levels of penetrance. We tested for linkage using four models (Table 1). For model I a maximum penetrance of 0.60 was used, as a conservative representation of the usual penetrance of this disorder. With model II a maximum penetrance of 1.0 was chosen to specifically reflect the high density of illness in the two families under study. For these first two models the age of onset curve was obtained using the combined ages of onset for the two study families. Curves were derived by dividing the total number of affected cases by the number who were affected before the ages of 20, 30, 40 and 50 years respectively. For example, when the maximum assumed penetrance was 1.0, the penetrance values were 0.20, 0.50, 0.75, and 1.00. In models I and II the frequency of the disease allele was taken as 0.035 (Detera-Wadleigh et al., 1987), and the penetrance for the non-susceptible genotype (the ‘sporadic rate’) was set at 0.005 (Detera-Wadleigh et al., 1987). In models III and IV the penetrances were set equal to those found in the Amish population at ages 15-19. 20-24, 25-29 and > 30 (Egeland et al., 1987). In model III, using the maximum penetrance of 0.63 found in the total Amish sample, penetrance values were 0.10, 0.27. 0.56, and 0.63 in the above age brackets. The disease allele frequency was 0.015 and the maximum penetrance
Disease allele frequency
Sporadic rate
0.035 0.035
0.005 0.005
0.015 0.021
0.0001 0.001
1 and 2.
for the non-susceptible genotype 0.0001. In model IV a penetrance of 0.85 (as found in Amish pedigree number 110) was employed. Penetrance values were therefore 0.18, 0.42, 0.73 and 0.85 in the above age brackets. A disease allele frequency of 0.021 and a maximum sporadic rate of 0.001 were used. Published allele frequencies for the marker RFLPs were used in the linkage analyses (i.e., Egeland et al., 1987: Moss et al., 1986; Komer et al., 1988). Alleles at the INS locus were combined into an artificial ‘four-allele’ system for the analysis (as described by Ott, 1985).
Results
Linkage of bipolar disorder and recurrent unipolar depression to HRAS, INS and TH Combined lod scores for families 1 and 2, using each of the four models, are outlined in Table 2. (Family 2 was not informative for HRAS.) With both models II and IV close linkage to each of the three markers was excluded at a 8 (recombination fraction) of 0.05 or 0.10, and with models I and III close linkage was excluded at 8 = 0.00. Fig. 3 displays the lod scores for linkage between bipolar disorder (bipolar I and II and recurrent major depression) and HRAS in family 1, as at a a function of recombination fraction,
29
Linkage thresh;&
to HRAS, INS and TH using other of diagnosis for affective disorder (Table
3)
-3
/ 0.0
0.1
0.2
0.4
0.3
Recombination
0.5
traction
Fig. 3. Graph of lod scores for linkage of HRAS to bipolar disorder (bipolar disorder plus recurrent major depressive disorder) in family 1. Age dependent penetrance to maximum of 0.60 (model I).
penetrance of 0.60 (model I). Different gene frequencies of both the disease allele (from 0.015 to 0.035) and marker allele were tested with each model, but these had no significant effect on the lod score results.
TABLE
Linkage was tested with each of the markers to groups defined by the following thresholds of illness: A - bipolar I disorder; B - bipolar I and II; C - bipolar I, II and recurrent unipolar depression; and D - bipolar disorder, recurrent unipolar depression, single episodes of major depression and minor depression. (When a narrow definition of the disorder was used, ill subjects who were not defined as affected were defined as unaffected.) Linkage of these varying diagnostic thresholds to each gene marker was rejected. For example, in Table 2 linkage findings with HRAS in family 1 are presented using a penetrance of 1.0 (model II). These results were consistent, irrespective of whether the narrow bipolar I group or broader groups with depressed subjects were included. The effect of variable rates of sporadic (nongenetic) presentations of illness, particularly depression, has not been closely examined in previous linkage studies of bipolar disorder. To examine for the effect of this, particularly at threshold D, we tested a range of sporadic rates varying from 0.001 to 0.1. Varying the sporadic rate did not alter the conclusions of absence of linkage. At a sporadic rate of 0.001 close linkage
2
COMBINED
LOD SCORES
FOR FAMILIES
Recombination
1 AND 2
fraction
0.00
0.05
0.10
0.20
0.30
0.40
1 II III IV
- 2.70 - 22.08 - 4.75 - 9.07
-1.15 - 4.91 - 1.71 - 3.66
- 0.64 - 2.93 - 1.00 -2.30
-0.16 -1.12 -0.33 - 0.91
0.00 - 0.33 - 0.07 - 0.27
+ 0.02 - 0.04 0.00 - 0.03
I II III IV
- 2.39 - 23.75 - 3.89 - 8.04
-
0.76 5.32 0.98 2.78
- 0.29 - 2.97 -0.39 - 1.59
+ -
+0.13 - 0.22 +0.15 0.00
+ + + +
0.07 0.01 0.07 0.06
- 2.35 - 20.56 - 3.34 - 6.51
-
1.89 4.58 2.41 3.66
-
- 0.64 -1.11 - 0.61 - 0.95
-
-
0.03 0.06 0.03 0.05
c-Harvey T(LF Model:
Insulin Model:
0.07 0.97 0.07 0.43
Tyrosine hydroxylase Model:
I II III IV
1.37 2.67 1.61 2.29
0.23 0.38 0.25 0.33
TABLE
3
LOD SCORES
FOR FAMILY
Diagnostic threshold A. Bipolar
B. Bipolar and II
1 AT VARIOUS
DIAGNOSTIC
Recombination 0.00 I
1
C. Bipolar disorder and recurrent MMD D. Bipolar disorder, recurrent MDD, singie
THRESHOLDS
a - HRAS
fraction 0.05
0.10
0.20
0.30
0.40
M F
- 6.41 - 3.51
- 1.83 - 1.07
- 0.99 - 0.67
- 0.28 -0.31
- 0.03 -0.12
+ 0.02 - 0.03
C
- 9.92
- 2.84
-1.66
- 0.59
-0.15
-0.01
M F
- 8.70 -1.93
- 1.82 -0.97
-0.98 - 0.65
- 0.27 - 0.31
- 0.02 -0.12
+ 0.02 - 0.03
C
- 10.63
- 2.79
- 1.63
- 0.53
-0.14
- 0.01
M F
- 16.94 - 5.14
-2.11 - 2.80
-1.10 - 1.83
- 0.29 - 0.83
- 0.02 - 0.31
+ 0.02 - 0.06
C
- 22.08
-4.91
- 2.93
- 1.12
-0.33
- 0.04
M F
- 19.06 - 3.99
-1.43 -1.55
-0.52 -0.98
+ 0.09 - 0.44
+0.16 -0.18
+ 0.07 - 0.04
C
- 23.05
- 2.98
- 1.50
- 0.35
- 0.02
+ 0.03
episode MMD. minor depression
’ Model II (maximum penetrance families 1 and 2 combined). M, male; F. female; C, combined
of 1.0; disease allele frequency male and female:
MDD.
of 0.035; sporadic
major depressive
was excluded at 0 = 0.10 (lod score of -5.12). whilst at 0.1 close linkage was excluded at 0 = 0.05 (lod score - 2.18). These findings indicate that the clear exclusion of linkage was not due potentially to a significant rate of sporadic presentations of illness. Discussion
Our finding of a clear exclusion of close linkage of the chromosome 11 markers to bipolar disorder is consistent with the re-analysis of the Amish pedigree data (Kelsoe et al., 1989) and previous negative reports (Hodgkinson et al., 1987; Detera-Wadleigh et al., 1987; Gill et al., 1988; Wesner et al., 1990). It is still possible, however, that our results are compatible with the association study findings of Leboyer et al. (1990) as allele variation at the TH locus may potentially confer some susceptibility to bipolar disorder without the effect being sufficiently ‘major’ to be
rate of 0.005; age of onset curve obtained
from
disorder.
demonstrated by linkage analysis. Overall, though, our result suggests that if there is a responsible gene on the short arm of chromosome 11, then this is a rare cause of bipolar disorder. Although our findings are negative, they are important because of the likely genetic heterogeneity of this condition. The validity of our results was strengthened by three particular findings. Firstly, close linkage was excluded at a series of different thresholds (levels of definition) of affective status. as indicated in Table 3. The importance of testing for linkage at these various levels was stated in the report of the MacArthur Foundation Workshop on linkage and clinical features in affective disorders (Merikangas et al., 1989). Although it is now generally accepted that the ‘bipolar diagnostic spectrum’ (Gershon, 1990) includes schizo-affective, bipolar I and II. unipolar and cyclothymic disorders, testing at various levels is necessary because of both the diagnostic instability of bipolar II diagnoses and
31
the possibility of non-genetic or sporadic presentations of depression in these families. In view of this latter problem we only included unipolar depressed subjects who fulfilled the strict YaleNIMH criteria for unipolar depression (Gershon and Goldin, 1989). Misclassification of the nongenetic forms of affective illness significantly reduces ability to demonstrate linkage (Martinez et al., 1989). Secondly, the finding of exclusion of close linkage was consistent at a range of penetrances (60100%) with the different models (I-IV). The negative finding at a penetrance level of 60% is particularly important as this represents a conservative approximation for this disorder. Other groups (e.g., Hodgkinson et al., 1987) have used a penetrance level of 100%. Thirdly, we tested for various rates of ‘sporadic’ presentations of illness. Varying sporadic rates did not affect our conclusion of exclusion of close linkage, an important finding in the light of the findings of Martinez et al. (1989) discussed above. Previous assumptions for sporadic rates of 0.0001 (Egeland et al., 1987) and 0.005 (Detera-Wadleigh et al., 1987) may not have adequately accounted for the frequency of non-genetic presentations (particularly depression) in bipolar pedigrees. Sporadic rates of at least 0.01 may more accurately reflect the clinical situation. Although we used a commonly employed method for excluding families in which more than one disease gene is segregating, it is still possible that this technique was not sufficiently rigorous to exclude the possibility of genetic heterogeneity, which would of course weaken our claim for exclusion of linkage. Future studies will need to systematically check the family history of both first- and second-degree relatives of marrying-in individuals to further reduce the likelihood of more than one gene segregating in a pedigree. In summary, our findings in two Australian pedigrees are consistent with other studies indicating that bipolar disorder is not related to these chromosome 11~15 loci. A recent study has also excluded linkage to the X-chromosomal region Xq28 (Berrettini et al., 1990). Despite these negative findings, there is no doubt concerning the potential power of molecular genetic linkage studies to elucidate the basic pathology of bipolar
disorder. Future studies, however, will need to pay closer attention to the specific clinical features of this disorder. Acknowledgements This study was supported by the Prince Henry Hospital Coast Centenary Fund, the N.S.W. Institute of Psychiatry, the Australian National Health and Medical Research Council and the Burleigh Foundation. We are grateful to Mr. Mark Graham for performing the DNA extraction and Southern hybridization analyses; Dr. Ian Hickie of the Mood Disorders Unit, Prince Henry Hospital, for assistance in reviewing diagnoses; Ms. Anna Zoumazi for technical assistance; and Mrs. Zora Vuckovic and Mrs. Femande Ahtime for preparation of the manuscript. References Andreasen. N.C., Endicott, J., Spitzer, R.L. and Winokur, G. (1977) The family history method using diagnostic criteria: reliability and validity. Arch. Gen. Psychiatry 34, 12291235. Bell, G.I., Selby, M.J. and Rutter. W.J. (1982) The highly polymorphic region near the human insulin gene is composed of simple tandemly repeating sequences. Nature 295, 31-35. Berretini. W.H., Goldin, L-R., Gelemter, J., Gejman, P.V.. Gershon, E.S. and Detera-Wadleigh. S. (1990) X-chromosome markers and manic-depressive illness: rejection of linkage to Xq28 in nine bipolar pedigrees. Arch. Gen. Psychiatry 47, 366-376. Botstein, D. (1980) Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32. 314. Cox, N., Reich, T.. Rice. J.. Elston. R., Schober, J. and Keats, B. (1989) Segregation and linkage analyses of bipolar and major depressive illness in multigenerational pedigrees. J. Psychiatr. Res. 23, 109-123. Detera-Wadleigh. S.D., Berrettini. W.H. and Goldein, R. (1987) Close linkage of c-harvey-ras-1 and the insulin gene to affective disorder is ruled out in three North American pedigrees. Nature 325, 806-808. Egeland, J.A., Gerhard, D.S. and Pauls, D.L. (1987) Bipolar affective disorders linked to DNA markers on chromosome 11. Nature 325, 783-787. Gershon, E. (1990) Genetics. In: F.K. Goodwin and K.R. Jamison (Eds.), Manic-Depressive Illness. Oxford University Press, New York, NY. Gershon, E. and Goldin, L. (1989) Linkage data on the affective disorders in an epidemiologic context. Genet. Epidemiol. 6. 201-209.
32 Gill, M., McKeon, P. and Hurnphries, P. (1988) Linkage analysis of manic depression in an Irish family using H-ras 1 and INS DNA markers. J. Med. Genet. 25. 634-637. Hodgkinson, S., Sherrington, R. and Gurling, M. (1987) Molecular genetic evidence for heterogeneity in manic depression. Nature 325, 805-806. Kan, Y.W. and Dozy, A.M. (1978) Polymorphism of DNA sequence to human beta-globin structural gene: relationship to sickle mutation. Proc. Natl. Acad. Sci. U.S.A. 75, 5631-5635. Kelsoe, J.R., Ginns, E.I., Egeland, J.A., Gerhard, D.S., Goldstein, A.M., Bale, S.J., Pauls, D.L., Long, R.T., Kidd, K.K., Conte, G., Housman, D.E. and Paul, SM. (1989) Reevaluation of the linkage relationship between chromosome Ilp loci and the gene for bipolar affective disorder in the old order Amish. Nature 342, 238-243. Kelsoe, J.R., Ginns, E.l., Egeland, J.A., Goldstein, A.M., Bale, S.J., Pauls, D.L., Long, R.T., Conte, G., Gerhard, D.S., Housman, D.E. and Paul, S.M. (1990) Bipolar affective disorder in the Older Order Amish. APA 143rd Annual Meeting, New York, May (Abstract). Korner, J.. UhIhaas, S. and Mallett, J. (1988) Propping, P, Gal, A. Further RFLPs at the human tyrosine hydroxylase locus. Nucleic Acids Res. 16, 9078. Kraepehn, E. (1921) Manic Depressive Insanity and Paranoia. (Transl. M. Barclay). Churchill Livingstone, Edinburgh. Krontiris, T.G., DiMartino, N.A., Kolb, M. and Parkinson, D.R. (1985) Unique allelic restriction fragments of the human Ha-ras locus in leukocyte and tumour DNAs of cancer patients. Nature 313, 369-374. Leboyer, M., Malafosse, A., Boularand, S., Campion, D., Gheysen, F., Samolyk, D., Henriksson, B., Denise, R.E., des Lauriers, A., Lepine, J.-P., Zarifian, E., ClergetDarpoux, F. and Mallet, J. (1990) Tyrosine hydroxylase polymorphisms associated with manic-depressive illness. Lancet 335, 1219. Maniatis, T., Fritsch, E.F. and Sambrook. J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Martinez, M., Khlat, M., Leboyer, M. and Clerget-Darpoux, F. (1989) Performance of linkage analyses under misclassifica-
tion error Epidemiol.
when the genetic 6, 253-258.
model
is unknown.
Genet.
Merikangas, K.R., Spence, A. and Kupfer, D.J. (1989) Linkage studies of bipolar disorder: methodologic and analytic issues. Arch. Gen. Psychiatry 46, 1137-1141. Moss, P.A.H., Davies, K.E., Boni, C., Mallett, J. and Reeders, S.T. (1986) Linkage of tyrosine hydroxylase to four other markers on the short arm of chromosome 11. Nucleic Acids Res. 14, 9927-9932. Ott, J. (1974) Estimation of the recombination fraction in human pedigrees: efficient computation of the likelihood for human linkage studies. Am. J. Hum. Genet. 28.525-529. Ott, J. (1985) Analysis of Human Genetic Linkage. Johns Hopkins University Press, Baltimore, MD. Owen, M.J. and Mullan, M.J. (1990) Molecular genetic studies of manic-depression and schizophrenia. Trends Neurol. Sci. 13, 29-31. Rice, J., Reich, T., Andreasen, N.C., Endicott, J., Van Eerdewegh, M., Fishman, R., Hirschfeld, R.M.A. and Klerman, G.L. (1987) The familial transmission of bipolar illness. Arch. Gen. Psychiatry 44, 441-450. Thompson, W.D., Kidd, J.R. and Weissman, M.M. (1979) A procedure for the efficient collection and processing of pedigree data suitable for genetic analysis. J. Psychiatr. Res. 15, 291-303. Tinklenberg, K.J., Masharani, V., Coleman, R.T., Taylor, J.. Cawthom, R. and Frossard, P. (1988) DraI and ScaI RFLPs at human tyrosine hydroxylase (TH) gene locus. Nucleic Acids Res. 16, 10948. Todd. R.D. and O’MalIey, K.L. (1989) Population frequencies of tyrosine hydroxylase restriction fragment length polymorphisms in bipolar affective disorder. Biol. Psychiatry 25, 626-630. Wesner, R.B., Scheffner, W., Palmer, P.J., Crowe, P.R. and Winokur, G. (1990) The effect of co-morbidity and penetrance on the outcome of linkage analysis in a bipolar family. Biol. Psychiatry 27, 241-244. Willard, H.F.. Skolnick, M.H., Pearson, P.L. and Mandel, J.L. (1985) Report of the committee on human gene mapping by recombinant DNA techniques. Cytogenet. Cell Genet. 40, 360-489.
Journal of A/fective Disorders, 21 (1991) 23-32 0 1991 Elsevier Science Publishers B.V. (Biomedical Division) 01650327/91/$03.50 ADONIS 016503279100054Q
JAD 00767
Close linkage of bipolar disorder to chromosome 11 markers is excluded in two large Australian pedigrees Philip Mitchell ’ School of Psychiatv,
‘, Brent Waters ‘, Nigel Morrison *, John Shine *, Jennifer and John Eisman *
Donald
3
University of New South Wales, Sydney, Australia, ’ Garvan Institute of Medical Research, St. Vincent’s Hospital, Sydney, Australia and ’ School of Biological Sciences, Macquarie University, Sydney, Australia (Received 4 July 1990) (Revision received 29 August 1990) (Accepted 24 September 1990)
Summary The relationship between bipolar disorder and chromosome 11 markers remains uncertain. Whilst re-analysis of the Amish pedigree weakened previous evidence for close linkage (but could not exclude the possibility of genetic heterogeneity), a recent French study has found a significant association between this condition and tyrosine hydroxylase polymorphisms. We aimed to determine if bipolar disorder in two large Australian pedigrees (of Irish and English extraction respectively) was linked to these markers. Of the 84 family members available for testing, nine were diagnosed as bipolar I, one as bipolar II and six had recurrent unipolar depression. Linkage of bipolar disorder and recurrent depression to the chromosome 11~15 markers c-Harvey ras, insulin and tyrosine hydroxylase was tested using a series of genetic models with varying penetrance levels. Additionally, linkage was examined using a series of levels of definitions of affective status (ranging from bipolar I alone to all affective illnesses). Close linkage to these markers was strongly excluded using each model and definition. The findings also persisted when a wide range of rates of ‘sporadic’ (non-genetic) presentations of illness were incorporated in the analysis. These results are consistent with other recent studies indicating that bipolar disorder is not linked to chromosomal region llp15.
Key words: Bipolar disorder; Chromosome 11; Linkage
Introduction Address for correspondence: Dr. Philip Mitchell, Mood’ Disorders Unit, Division of Psychiatry, Prince Henry Hospital, Little Bay, NSW 2036, Australia.
Accumulated evidence from family, twin and adoption studies (Gershon, 1990) has confirmed
24
the clinical observations of 19th- and early 20thcentury psychiatrists (e.g., Kraepelin, 1921) that there is significant genetic transmission of both bipolar disorder and recurrent depressive illness. Although there is considerable evidence for monogenie transmission, the mode of transmission is still controversial (Rice et al., 1987; Cox et al., 1989; Gershon and Goldin, 1989). Linkage studies (which assume monogenic transmission) have been used to identify the chromosomal location of the putative responsible gene(s). The development of recombinant DNA techniques, and in particular the demonstration of restriction fragment length polymorphisms (RFLPs) (Kan and Dozy, 1978), has enabled the development of a detailed chromosomal map, facilitating testing for linkage of disease genes to specific genes or chromosomal regions (Botstein et al., 1980). The present situation concerning molecular studies in bipolar disorder must be regarded as uncertain (Owen and Mullan, 1990). In 1987, Egeland et al. reported tight linkage (a lod score of 4.08 at a recombination fraction of 0.00) of bipolar affective disorder to the c-Harvey-ras oncogene (HRAS), and loose linkage to the insulin gene (INS) - genes sited on the short arm of chromosome 11 (region 11~15) - in a large U.S. Amish pedigree. Hodgkinson et al. (1987) and Detera-Wadleigh et al. (1987) however, reported the exclusion of close linkage in several Icelandic and North American bipolar pedigrees respectively. A subsequent short report by Gill et al. (1988) similarly excluded close linkage in a large Irish pedigree. Wesner et al. (1990) also found no evidence for linkage in a bipolar pedigree. Their family, however, only included one bipolar I subject, and the results were highly dependent upon assumed penetrance levels. A reanalysis of the Amish pedigree by Kelsoe et al. (1989) however, reduced the likelihood of linkage of bipolar disorder to the chromosome 11 markers. Firstly, addition of genotypes from some previously untyped members and inclusion of two new presentations of illness in the original ‘core’ pedigree reduced the statistical probability of linkage. Secondly, inclusion of two lateral extensions of the ‘core’ pedigree led to exclusion of close linkage to this region. Although the authors of the re-analysis considered it most likely that the find-
ing of linkage in the original ‘core’ pedigree was due to chance, they could not discount the possibility of genetic heterogeneity (with regard to bipolar disorder) in the Amish population (Kelsoe et al., 1989). The latter explanation is possible even in a community as genetically isolated as the Amish, as firstly several progenitor couples had psychiatric illness among their first- and seconddegree descendants (Kelsoe et al., 1989) and secondly a Mennonite individual married into one of the earlier generations of the particular family involved in these studies (Kelsoe et al., 1990). The possibility of genetic heterogeneity has been highlighted by a recent population association (nor linkage) study of 50 unrelated French bipolar patients and 50 matched controls (Leboyer et al., 1990) . That investigation found a significant association between tyrosine hydroxylase (TH) RFLPs and bipolar disorder (P < 0.01) suggesting that liability to bipolar disorder may be dependent upon a defect within either the TH gene, or its regulatory domain. (The TH gene is in chromosomal region 11~15 near the HRAS and INS genes.) Todd and O’Malley (1989) had previously reported no association between TH RFLP alleles and bipolar disorder, but they only investigated 18 subjects. In the light of these discrepant findings we report a study which aimed to examine w,hether linkage exists between bipolar affective disorder in two large Australian pedigrees and the 11~15 markers.
Methods Case ascertainment and diagnosis We sought families with a large number of relatives, a single disease source, illness over two or three generations, a large number of affected family members and large sibships. Families were identified through the Mood Disorders Unit, Prince Henry Hospital, Sydney: the New South Wales Manic Depressive Self-Help Group; and local psychiatric units. In the 35 bipolar families identified, details of pedigrees were recorded using the method described by Thompson (1979). Families were assessed using the Family History-RDC (FH-RDC)
0:: P::
9:: 1:: ::
26
method of Andreasen et al. (1977) with interviews performed with both the proband and another informative family member. From the families contacted, we identified two large families with three generations of illness and a large number of affected individuals. All available relatives were interviewed by P.M. using the Schedule of Affective Disorders and Schizophrenia-Lifetime Version (SADS-L). Those interviews were used to derive Research Diagnostic Criteria (RDC) diagnoses using the strict NIMH criteria for depression which require at least 1 month’s duration of symptoms, and functional impairment or incapacity. In conjunction with the SADS-L interview, all marrying-in individuals were routinely questioned about any family history of psychiatric disorder, to ensure unilateral descent of illness in the pedigrees. Hospital and other medical records were obtained. Final diagnoses were made using the Yale-NIMH ‘Best Estimate’ Diagnosis Consensus guidelines which are assigned from FH-RDC and SADS-L derived RDC diagnoses, and medical records. ‘Best estimate’ diagnoses were also made independently by two other psychiatrists (B.W., Ian Hi&e) after assessment of all available interviews and records. These psychiatrists were aware of the disorder
under study. Diagnoses were made blind to laboratory investigations. Family one was a large family of Irish descent. Sixty-eight of the 78 members at risk were interviewed with SADS-L and gave blood for DNA testing. At the time of assessment there were 11 living affected individuals (one dying subsequently), five of whom were diagnosed as bipolar (four bipolar I, one bipolar II), and six had recurrent unipolar depression (NIMH criteria), each of whom fulfilled RDC for endogenous depression. The age of onset ranged from 12 to 47 years. Five of the individuals in this pedigree (bipolar and unipolar) had a late-onset presentation of illness (> 40 years). Ten other individuals were diagnosed as having either single episodes of major depression or minor depression. Family two was a family of English descent with seven affected individuals. Sixteen of the 20 members at risk were assessed. There were seven affected individuals (five available for testing), each with bipolar I disorder. There were no individuals with recurrent unipolar depression in this particular family. The age of onset ranged from 19 to 40 years. For the linkage analyses only subjects of at least 15 years of age were included. These pedigrees are outlined in Figs. 1 and 2.
BD
BD
CD
CD
XY
XY
XY
YZ
-0
El--
Tici DD XY
6 <4 CD XY
00 DD YY
DD XY
DD YZ
cr’: DD XY
DD xx
DD xx
Fig. 2. Pedigree 2. Circles = females; squares = males; diamonds = males or females. Full hatch = bipolar disorder: half hatch = recurrent major depressive disorder. Alleles are in the following order: HRAS, TH. Sex and birth order have been disguised to protect anonymity.
27
Genetic markers DNA extraction. Blood (60 ml) was collected into sodium heparin tubes (Becton Dickinson) and immediately centrifuged in a clinical centrifuge to separate the buffy coat. The buffy coat was further separated by repeated rounds of centrifugation through normal saline until the leukocytes were largely free of red blood cells. The cellular pellet was then resuspended in red blood cell lysis buffer for 20 min at room temperature. After red cell lysis had occurred, the leukocytes were then collected by centrifugation and resuspended evenly in 3 ml of normal saline. DNA was then liberated from the cells by lysis after addition of an equal volume of leukocyte lysis buffer (10 mM Tris-HCl, pH 7.4; 1 mM EDTA; 0.2% sodium dodecyl sulfate (Biorad Inc.)). The lysate was then treated with proteinase K (International Biotechnologies Inc.) at 50 pg/ml overnight at 37” C or for 4 h at 65 o C. DNA was then purified by repeated solvent extraction using phenol-chloroform and chloroform, essentially according to Maniatis et al. (1980). DNA was precipitated without the addition of further salts or low temperature treatment by the addition of two volumes of ethanol. Precipitated high molecular weight DNA was either removed by pipette or centrifuged at 5000 X g (Beckman G6-M) for 20 min. DNA pellets were washed in 70% ethanol followed by 100% ethanol and vacuum dried. DNA was redissolved in 1.0 ml TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and quantitated by spectroscopy at 260 nm wavelength. DNA was digested with various restriction endonucleases (all from Pharmacia) according to the manufacturer’s instructions. Southern hybridization analysis. Restricted DNA was resolved on 1.0% agarose (International Biotechnology Inc.) gels in TPE buffer essentially as described by Maniatis et al. (1980). DNA was transferred to Biorad Zetaprobe nylon membranes using the alkaline denaturation protocol recommended by the manufacturer. Plasmid DNA for probe preparation was made by alkaline lysis followed by cesium chloride ultracentrifugation (40 krpm for 40 h in a Beckman Ti-70 rotor) as described by Maniatis et al. (1980). Insert DNA from plasmid was prepared by elution from agarose gel chromatographically separated frag-
ments of restriction digested plasmid DNA. Radioactive insert probes were made by standard nick translation using a kit (BRL, Bethesda, MD). Incorporated radioactivity was separated from unincorporated by chromatography over a Pharmacia nick column. Probes were labelled to a specific activity between 2 and 5 x lo9 dpm/pg. Probes were hybridized to 3 x Cot l/2 values, calculated based on the complexity of the probe, at a concentration of 10 ng/ml using 0.5 ml of hybridization fluid per cm* of Zetaprobe. Hybridization and washing protocols were as described by Biorad for the Zetaprobe membrane without the use of dextran sulfate. Filters were exposed to Kodak XARS film with two DuPont Cronex Lightningplus intensifying screens for between 2 and 5 days. The probes used to detect RFLPs were the pEJ6.6 probe for HRAS (Krontiris et al., 1985) phins310 for INS (Bell et al., 1982) and a 1.5 kb EcoRI fragment of a TH cDNA clone (pCB-05) for TH (Tinklenberg et al., 1988). Multiple alleles were observed for each locus after restriction enzyme digestion. In accordance with previous studies, we observed: four alleles for HRAS with Sac1 (as described by Egeland et al., 1987) i.e., A (- 6.8 kb), B (- 6.2 kb), C (- 5.7 kb), D (- 5.2 kb); six alleles for INS with PuuII: 1 (- 2.4 kb), 2 (- 2.1 kb), 3 ( - 0.85 kb), 4 ( - 0.80 kb), 5 ( - 0.74 kb), 6 (- 0.68 kb) (Willard et al., 1985); and three for TH with BgfII: X (- 8.4 kb), Y (- 6.9 kb) and Z (- 8.0 kb) (Moss et al., 1986; Korner et al., 1988). The genotypes of family members for these markers are shown in Figs. 1 and 2. For technical reasons insulin RFLP data are not available for family 2. Immortalized cell lines for pedigree numbers are currently being established. Linkage analysis Linkage analyses were undertaken using the computer program LIPED (Ott, 1974). Lod scores > 3 are considered to be indicative of linkage and scores < -2 indicate rejection of linkage at that recombination fraction. (The term ‘lod’ is an acronym for logarithm (base 10) of the odds favoring linkage in a particular set of pedigree data. Although a lod score of 3 may suggest a probability of 0.001, this score in fact equates with a probability of 0.05 for two genes for which there is no a priori knowledge that they are linked
28 TABLE
1
PARAMETERS Model
OF THE FOUR Age-specific
Age: I” II * Age: III
b
IV c
GENETIC
MODELS
penetrance
15-19
20-29
30-39
40+
0.12 0.20
0.30 0.50
0.45 0.75
0.60 1.00
15-19
20-24
25-29
29+
0.10 0.18
0.27 0.42
0.56 0.73
0.63 0.85
’ Age of onset curve derived from combined ages of onset for families b Age of onset curve of total Amish sample. ’ Age of onset curve of Amish pedigree 110.
(Gershon, 1990)). Analyses were performed asdominant transmission suming an autosomal model with varying levels of penetrance. We tested for linkage using four models (Table 1). For model I a maximum penetrance of 0.60 was used, as a conservative representation of the usual penetrance of this disorder. With model II a maximum penetrance of 1.0 was chosen to specifically reflect the high density of illness in the two families under study. For these first two models the age of onset curve was obtained using the combined ages of onset for the two study families. Curves were derived by dividing the total number of affected cases by the number who were affected before the ages of 20, 30, 40 and 50 years respectively. For example, when the maximum assumed penetrance was 1.0, the penetrance values were 0.20, 0.50, 0.75, and 1.00. In models I and II the frequency of the disease allele was taken as 0.035 (Detera-Wadleigh et al., 1987), and the penetrance for the non-susceptible genotype (the ‘sporadic rate’) was set at 0.005 (Detera-Wadleigh et al., 1987). In models III and IV the penetrances were set equal to those found in the Amish population at ages 15-19. 20-24, 25-29 and > 30 (Egeland et al., 1987). In model III, using the maximum penetrance of 0.63 found in the total Amish sample, penetrance values were 0.10, 0.27. 0.56, and 0.63 in the above age brackets. The disease allele frequency was 0.015 and the maximum penetrance
Disease allele frequency
Sporadic rate
0.035 0.035
0.005 0.005
0.015 0.021
0.0001 0.001
1 and 2.
for the non-susceptible genotype 0.0001. In model IV a penetrance of 0.85 (as found in Amish pedigree number 110) was employed. Penetrance values were therefore 0.18, 0.42, 0.73 and 0.85 in the above age brackets. A disease allele frequency of 0.021 and a maximum sporadic rate of 0.001 were used. Published allele frequencies for the marker RFLPs were used in the linkage analyses (i.e., Egeland et al., 1987: Moss et al., 1986; Komer et al., 1988). Alleles at the INS locus were combined into an artificial ‘four-allele’ system for the analysis (as described by Ott, 1985).
Results
Linkage of bipolar disorder and recurrent unipolar depression to HRAS, INS and TH Combined lod scores for families 1 and 2, using each of the four models, are outlined in Table 2. (Family 2 was not informative for HRAS.) With both models II and IV close linkage to each of the three markers was excluded at a 8 (recombination fraction) of 0.05 or 0.10, and with models I and III close linkage was excluded at 8 = 0.00. Fig. 3 displays the lod scores for linkage between bipolar disorder (bipolar I and II and recurrent major depression) and HRAS in family 1, as at a a function of recombination fraction,
29
Linkage thresh;&
to HRAS, INS and TH using other of diagnosis for affective disorder (Table
3)
-3
/ 0.0
0.1
0.2
0.4
0.3
Recombination
0.5
traction
Fig. 3. Graph of lod scores for linkage of HRAS to bipolar disorder (bipolar disorder plus recurrent major depressive disorder) in family 1. Age dependent penetrance to maximum of 0.60 (model I).
penetrance of 0.60 (model I). Different gene frequencies of both the disease allele (from 0.015 to 0.035) and marker allele were tested with each model, but these had no significant effect on the lod score results.
TABLE
Linkage was tested with each of the markers to groups defined by the following thresholds of illness: A - bipolar I disorder; B - bipolar I and II; C - bipolar I, II and recurrent unipolar depression; and D - bipolar disorder, recurrent unipolar depression, single episodes of major depression and minor depression. (When a narrow definition of the disorder was used, ill subjects who were not defined as affected were defined as unaffected.) Linkage of these varying diagnostic thresholds to each gene marker was rejected. For example, in Table 2 linkage findings with HRAS in family 1 are presented using a penetrance of 1.0 (model II). These results were consistent, irrespective of whether the narrow bipolar I group or broader groups with depressed subjects were included. The effect of variable rates of sporadic (nongenetic) presentations of illness, particularly depression, has not been closely examined in previous linkage studies of bipolar disorder. To examine for the effect of this, particularly at threshold D, we tested a range of sporadic rates varying from 0.001 to 0.1. Varying the sporadic rate did not alter the conclusions of absence of linkage. At a sporadic rate of 0.001 close linkage
2
COMBINED
LOD SCORES
FOR FAMILIES
Recombination
1 AND 2
fraction
0.00
0.05
0.10
0.20
0.30
0.40
1 II III IV
- 2.70 - 22.08 - 4.75 - 9.07
-1.15 - 4.91 - 1.71 - 3.66
- 0.64 - 2.93 - 1.00 -2.30
-0.16 -1.12 -0.33 - 0.91
0.00 - 0.33 - 0.07 - 0.27
+ 0.02 - 0.04 0.00 - 0.03
I II III IV
- 2.39 - 23.75 - 3.89 - 8.04
-
0.76 5.32 0.98 2.78
- 0.29 - 2.97 -0.39 - 1.59
+ -
+0.13 - 0.22 +0.15 0.00
+ + + +
0.07 0.01 0.07 0.06
- 2.35 - 20.56 - 3.34 - 6.51
-
1.89 4.58 2.41 3.66
-
- 0.64 -1.11 - 0.61 - 0.95
-
-
0.03 0.06 0.03 0.05
c-Harvey T(LF Model:
Insulin Model:
0.07 0.97 0.07 0.43
Tyrosine hydroxylase Model:
I II III IV
1.37 2.67 1.61 2.29
0.23 0.38 0.25 0.33
TABLE
3
LOD SCORES
FOR FAMILY
Diagnostic threshold A. Bipolar
B. Bipolar and II
1 AT VARIOUS
DIAGNOSTIC
Recombination 0.00 I
1
C. Bipolar disorder and recurrent MMD D. Bipolar disorder, recurrent MDD, singie
THRESHOLDS
a - HRAS
fraction 0.05
0.10
0.20
0.30
0.40
M F
- 6.41 - 3.51
- 1.83 - 1.07
- 0.99 - 0.67
- 0.28 -0.31
- 0.03 -0.12
+ 0.02 - 0.03
C
- 9.92
- 2.84
-1.66
- 0.59
-0.15
-0.01
M F
- 8.70 -1.93
- 1.82 -0.97
-0.98 - 0.65
- 0.27 - 0.31
- 0.02 -0.12
+ 0.02 - 0.03
C
- 10.63
- 2.79
- 1.63
- 0.53
-0.14
- 0.01
M F
- 16.94 - 5.14
-2.11 - 2.80
-1.10 - 1.83
- 0.29 - 0.83
- 0.02 - 0.31
+ 0.02 - 0.06
C
- 22.08
-4.91
- 2.93
- 1.12
-0.33
- 0.04
M F
- 19.06 - 3.99
-1.43 -1.55
-0.52 -0.98
+ 0.09 - 0.44
+0.16 -0.18
+ 0.07 - 0.04
C
- 23.05
- 2.98
- 1.50
- 0.35
- 0.02
+ 0.03
episode MMD. minor depression
’ Model II (maximum penetrance families 1 and 2 combined). M, male; F. female; C, combined
of 1.0; disease allele frequency male and female:
MDD.
of 0.035; sporadic
major depressive
was excluded at 0 = 0.10 (lod score of -5.12). whilst at 0.1 close linkage was excluded at 0 = 0.05 (lod score - 2.18). These findings indicate that the clear exclusion of linkage was not due potentially to a significant rate of sporadic presentations of illness. Discussion
Our finding of a clear exclusion of close linkage of the chromosome 11 markers to bipolar disorder is consistent with the re-analysis of the Amish pedigree data (Kelsoe et al., 1989) and previous negative reports (Hodgkinson et al., 1987; Detera-Wadleigh et al., 1987; Gill et al., 1988; Wesner et al., 1990). It is still possible, however, that our results are compatible with the association study findings of Leboyer et al. (1990) as allele variation at the TH locus may potentially confer some susceptibility to bipolar disorder without the effect being sufficiently ‘major’ to be
rate of 0.005; age of onset curve obtained
from
disorder.
demonstrated by linkage analysis. Overall, though, our result suggests that if there is a responsible gene on the short arm of chromosome 11, then this is a rare cause of bipolar disorder. Although our findings are negative, they are important because of the likely genetic heterogeneity of this condition. The validity of our results was strengthened by three particular findings. Firstly, close linkage was excluded at a series of different thresholds (levels of definition) of affective status. as indicated in Table 3. The importance of testing for linkage at these various levels was stated in the report of the MacArthur Foundation Workshop on linkage and clinical features in affective disorders (Merikangas et al., 1989). Although it is now generally accepted that the ‘bipolar diagnostic spectrum’ (Gershon, 1990) includes schizo-affective, bipolar I and II. unipolar and cyclothymic disorders, testing at various levels is necessary because of both the diagnostic instability of bipolar II diagnoses and
31
the possibility of non-genetic or sporadic presentations of depression in these families. In view of this latter problem we only included unipolar depressed subjects who fulfilled the strict YaleNIMH criteria for unipolar depression (Gershon and Goldin, 1989). Misclassification of the nongenetic forms of affective illness significantly reduces ability to demonstrate linkage (Martinez et al., 1989). Secondly, the finding of exclusion of close linkage was consistent at a range of penetrances (60100%) with the different models (I-IV). The negative finding at a penetrance level of 60% is particularly important as this represents a conservative approximation for this disorder. Other groups (e.g., Hodgkinson et al., 1987) have used a penetrance level of 100%. Thirdly, we tested for various rates of ‘sporadic’ presentations of illness. Varying sporadic rates did not affect our conclusion of exclusion of close linkage, an important finding in the light of the findings of Martinez et al. (1989) discussed above. Previous assumptions for sporadic rates of 0.0001 (Egeland et al., 1987) and 0.005 (Detera-Wadleigh et al., 1987) may not have adequately accounted for the frequency of non-genetic presentations (particularly depression) in bipolar pedigrees. Sporadic rates of at least 0.01 may more accurately reflect the clinical situation. Although we used a commonly employed method for excluding families in which more than one disease gene is segregating, it is still possible that this technique was not sufficiently rigorous to exclude the possibility of genetic heterogeneity, which would of course weaken our claim for exclusion of linkage. Future studies will need to systematically check the family history of both first- and second-degree relatives of marrying-in individuals to further reduce the likelihood of more than one gene segregating in a pedigree. In summary, our findings in two Australian pedigrees are consistent with other studies indicating that bipolar disorder is not related to these chromosome 11~15 loci. A recent study has also excluded linkage to the X-chromosomal region Xq28 (Berrettini et al., 1990). Despite these negative findings, there is no doubt concerning the potential power of molecular genetic linkage studies to elucidate the basic pathology of bipolar
disorder. Future studies, however, will need to pay closer attention to the specific clinical features of this disorder. Acknowledgements This study was supported by the Prince Henry Hospital Coast Centenary Fund, the N.S.W. Institute of Psychiatry, the Australian National Health and Medical Research Council and the Burleigh Foundation. We are grateful to Mr. Mark Graham for performing the DNA extraction and Southern hybridization analyses; Dr. Ian Hickie of the Mood Disorders Unit, Prince Henry Hospital, for assistance in reviewing diagnoses; Ms. Anna Zoumazi for technical assistance; and Mrs. Zora Vuckovic and Mrs. Femande Ahtime for preparation of the manuscript. References Andreasen. N.C., Endicott, J., Spitzer, R.L. and Winokur, G. (1977) The family history method using diagnostic criteria: reliability and validity. Arch. Gen. Psychiatry 34, 12291235. Bell, G.I., Selby, M.J. and Rutter. W.J. (1982) The highly polymorphic region near the human insulin gene is composed of simple tandemly repeating sequences. Nature 295, 31-35. Berretini. W.H., Goldin, L-R., Gelemter, J., Gejman, P.V.. Gershon, E.S. and Detera-Wadleigh. S. (1990) X-chromosome markers and manic-depressive illness: rejection of linkage to Xq28 in nine bipolar pedigrees. Arch. Gen. Psychiatry 47, 366-376. Botstein, D. (1980) Construction of a genetic linkage map in man using restriction fragment length polymorphisms. Am. J. Hum. Genet. 32. 314. Cox, N., Reich, T.. Rice. J.. Elston. R., Schober, J. and Keats, B. (1989) Segregation and linkage analyses of bipolar and major depressive illness in multigenerational pedigrees. J. Psychiatr. Res. 23, 109-123. Detera-Wadleigh. S.D., Berrettini. W.H. and Goldein, R. (1987) Close linkage of c-harvey-ras-1 and the insulin gene to affective disorder is ruled out in three North American pedigrees. Nature 325, 806-808. Egeland, J.A., Gerhard, D.S. and Pauls, D.L. (1987) Bipolar affective disorders linked to DNA markers on chromosome 11. Nature 325, 783-787. Gershon, E. (1990) Genetics. In: F.K. Goodwin and K.R. Jamison (Eds.), Manic-Depressive Illness. Oxford University Press, New York, NY. Gershon, E. and Goldin, L. (1989) Linkage data on the affective disorders in an epidemiologic context. Genet. Epidemiol. 6. 201-209.
32 Gill, M., McKeon, P. and Hurnphries, P. (1988) Linkage analysis of manic depression in an Irish family using H-ras 1 and INS DNA markers. J. Med. Genet. 25. 634-637. Hodgkinson, S., Sherrington, R. and Gurling, M. (1987) Molecular genetic evidence for heterogeneity in manic depression. Nature 325, 805-806. Kan, Y.W. and Dozy, A.M. (1978) Polymorphism of DNA sequence to human beta-globin structural gene: relationship to sickle mutation. Proc. Natl. Acad. Sci. U.S.A. 75, 5631-5635. Kelsoe, J.R., Ginns, E.I., Egeland, J.A., Gerhard, D.S., Goldstein, A.M., Bale, S.J., Pauls, D.L., Long, R.T., Kidd, K.K., Conte, G., Housman, D.E. and Paul, SM. (1989) Reevaluation of the linkage relationship between chromosome Ilp loci and the gene for bipolar affective disorder in the old order Amish. Nature 342, 238-243. Kelsoe, J.R., Ginns, E.l., Egeland, J.A., Goldstein, A.M., Bale, S.J., Pauls, D.L., Long, R.T., Conte, G., Gerhard, D.S., Housman, D.E. and Paul, S.M. (1990) Bipolar affective disorder in the Older Order Amish. APA 143rd Annual Meeting, New York, May (Abstract). Korner, J.. UhIhaas, S. and Mallett, J. (1988) Propping, P, Gal, A. Further RFLPs at the human tyrosine hydroxylase locus. Nucleic Acids Res. 16, 9078. Kraepehn, E. (1921) Manic Depressive Insanity and Paranoia. (Transl. M. Barclay). Churchill Livingstone, Edinburgh. Krontiris, T.G., DiMartino, N.A., Kolb, M. and Parkinson, D.R. (1985) Unique allelic restriction fragments of the human Ha-ras locus in leukocyte and tumour DNAs of cancer patients. Nature 313, 369-374. Leboyer, M., Malafosse, A., Boularand, S., Campion, D., Gheysen, F., Samolyk, D., Henriksson, B., Denise, R.E., des Lauriers, A., Lepine, J.-P., Zarifian, E., ClergetDarpoux, F. and Mallet, J. (1990) Tyrosine hydroxylase polymorphisms associated with manic-depressive illness. Lancet 335, 1219. Maniatis, T., Fritsch, E.F. and Sambrook. J. (1982) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Martinez, M., Khlat, M., Leboyer, M. and Clerget-Darpoux, F. (1989) Performance of linkage analyses under misclassifica-
tion error Epidemiol.
when the genetic 6, 253-258.
model
is unknown.
Genet.
Merikangas, K.R., Spence, A. and Kupfer, D.J. (1989) Linkage studies of bipolar disorder: methodologic and analytic issues. Arch. Gen. Psychiatry 46, 1137-1141. Moss, P.A.H., Davies, K.E., Boni, C., Mallett, J. and Reeders, S.T. (1986) Linkage of tyrosine hydroxylase to four other markers on the short arm of chromosome 11. Nucleic Acids Res. 14, 9927-9932. Ott, J. (1974) Estimation of the recombination fraction in human pedigrees: efficient computation of the likelihood for human linkage studies. Am. J. Hum. Genet. 28.525-529. Ott, J. (1985) Analysis of Human Genetic Linkage. Johns Hopkins University Press, Baltimore, MD. Owen, M.J. and Mullan, M.J. (1990) Molecular genetic studies of manic-depression and schizophrenia. Trends Neurol. Sci. 13, 29-31. Rice, J., Reich, T., Andreasen, N.C., Endicott, J., Van Eerdewegh, M., Fishman, R., Hirschfeld, R.M.A. and Klerman, G.L. (1987) The familial transmission of bipolar illness. Arch. Gen. Psychiatry 44, 441-450. Thompson, W.D., Kidd, J.R. and Weissman, M.M. (1979) A procedure for the efficient collection and processing of pedigree data suitable for genetic analysis. J. Psychiatr. Res. 15, 291-303. Tinklenberg, K.J., Masharani, V., Coleman, R.T., Taylor, J.. Cawthom, R. and Frossard, P. (1988) DraI and ScaI RFLPs at human tyrosine hydroxylase (TH) gene locus. Nucleic Acids Res. 16, 10948. Todd. R.D. and O’MalIey, K.L. (1989) Population frequencies of tyrosine hydroxylase restriction fragment length polymorphisms in bipolar affective disorder. Biol. Psychiatry 25, 626-630. Wesner, R.B., Scheffner, W., Palmer, P.J., Crowe, P.R. and Winokur, G. (1990) The effect of co-morbidity and penetrance on the outcome of linkage analysis in a bipolar family. Biol. Psychiatry 27, 241-244. Willard, H.F.. Skolnick, M.H., Pearson, P.L. and Mandel, J.L. (1985) Report of the committee on human gene mapping by recombinant DNA techniques. Cytogenet. Cell Genet. 40, 360-489.