Association and Linkage Studies of CRH and PENKGenes in Bipolar Disorder: A CollaborativeIGSLI Study

Martin Alda,1,2* Gustavo Turecki,3 Paul Grof,2 Patrizia Cavazzoni,2 Anne Duffy,1 Eva Grof,2Bernd Ahrens,4 Anne Berghofer,4 Bruno Muller-Oerlinghausen,4 Marta Dvorakova,5Eva Libigerova,5 Milos Vojtechovsky,5 Petr Zvolsky,5 Ridha Joober,3 Agneta Nilsson,6Helena Prochazka,6 Rasmus W. Licht,7 Niels A. Rasmussen,7 Mogens Schou,7 Per Vestergaard,7Anita Holzinger,8 Claudia Schumann,8 Kenneth Thau,8 and Guy A. Rouleau3

1Department of Psychiatry, Dalhousie University, Halifax, Nova Scotia, Canada2Department of Psychiatry, University of Ottawa,Ottawa, Ontario, Canada3Centre for Research in Neuroscience, The Montreal General Hospital, McGill University, Montreal, Quebec Canada4Department of Psychiatry, Free University, Berlin, Germany5Department of Psychiatry, Charles University, Prague and Hradec Kralove, Czech Republic6Karsudden Hospital, Katrineholm, Sweden7Psychiatric Hospital, University of Århus, Risskov, Denmark8Department of Psychiatry, University Clinic of Vienna, Vienna, Austria

Corticotropin-releasing hormone (CRH)and proenkephalin (PENK) are hypotha-lamic peptides involved in the stress re-sponse and hypothalamic-pituitary axisregulation. Previous research has impli-cated these peptides in the pathogenesis ofaffective disorders. In this study we investi-gated two polymorphisms located in thegenes that code for CRH and PENK bymeans of association and linkage analyses.A total of 138 bipolar patients and 108 con-trols were included in the association study.In addition, 24 families were available forlinkage analysis, including six families ofprobands with documented periodic posi-tivity of dexamethasone suppression tests(DST) during remission. We found no asso-ciation of bipolar disorder with either gene.Similarly, we did not find any evidence oflinkage (P = 0.56 for CRH and 0.52 for PENK)in the entire sample or in the subsample offamilies of DST positive probands. In con-clusion, our study does not support the hy-pothesis that genes coding for CRH or PENKcontribute to the genetic susceptibility to

bipolar disorder. Am. J. Med. Genet. (Neuro-psychiatr. Genet.) 96:178–181, 2000.© 2000 Wiley-Liss, Inc.

KEY WORDS: bipolar disorder; lithium; ge-netics; CRH; PENK; linkagestudy

INTRODUCTION

Bipolar disorder (BD) and depression are frequentlyassociated with a dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis. Many bipolar or de-pressed patients show an abnormality in the dexa-methasone suppression test (DST) during episodes oftheir illness with subsequent normalization of the testin remission [Holsboer et al., 1982]. HPA axis dysregu-lation has been reported in healthy relatives of pro-bands with affective disorders, suggesting that thiscould be a familial trait [Holsboer et al., 1995]. Fur-thermore, CSF levels of corticotropin-releasing hor-mone (CRH) have been observed to be elevated in de-pressed patients [Nemeroff et al., 1984]. Similarly,CRH receptors in the frontal cortex have been shown tobe down-regulated in suicide victims [Nemeroff et al.,1988], and the number of CRH producing neurons in-creased in depressed subjects [Raadsheer et al., 1994].Whereas most studies of HPA axis dysregulation havebeen carried out in patients with unipolar depression,there is good evidence that bipolar patients share manysimilarities [Stokes et al., 1984; Linkowski et al., 1994].Also, Vieta et al. [1997] have shown that a bluntedACTH response to CRH predicts relapse in bipolar sub-

Contract grant sponsor: Medical Research Council of Canada;Contract grant sponsor: Department of Health of Province ofNova Scotia.

*Correspondence to: Martin Alda, MD, Department of Psychia-try, Dalhousie University, Abbie J. Lane Building Room 3094,5909 Jubilee Road, Halifax, Nova Scotia B3H 2E2, Canada.E-mail: malda@is.dal.ca

Received 15 June 1999; Accepted 2 November 1999

American Journal of Medical Genetics (Neuropsychiatric Genetics) 96:178–181 (2000)

© 2000 Wiley-Liss, Inc.

jects. We have reported recently on a group of bipolarpatients successfully treated with lithium in whom weobserved an intermittent DST positivity despite fullclinical stabilization [Deshauer et al., 1999].

Taken together, the findings in the literature arecompatible with CRH overactivity persistent duringacute episodes of mood disorders, and intermittent dur-ing remissions. However, it is not clear whether thisoveractivity is due to changes in the CRH gene or sec-ondary to a dysfunction of some other factor influenc-ing CRH secretion. CRH is found in highest concentra-tions in the paraventricular nucleus of the hypothalamusand its gene is known to be regulated by a number offactors including: stimulation by norepinephrinethrough cAMP pathways, inhibition by hippocampalafferents mediated by mineralocorticoid and glucocor-ticoid receptors, and inhibition mediated by hypotha-lamic glucocorticoid receptors [Holsboer et al., 1992].Conversely, the presence of CRH in the raphe nucleiand locus coeruleus may indicate the role of CRH inmodulating serotonergic and noradrenergic systems[Arborelius et al., 1999]. Furthermore, CRH and otherhypothalamic hormones are regulated by proenkepha-lin (PENK), which is also expressed in the paraven-tricular nucleus and has been shown to play a role inregulating stress responses.

In this study, we present results of linkage and as-sociation studies with CRH and PENK genes in bipolardisorder. Genetic factors have been recognized as im-portant in the etiology of mood disorders. Several strat-egies have been proposed to identify susceptibility lociincluding studies of candidate genes. Whereas linkagemethodology can be helpful in detecting loci of majoreffect, association studies can detect effects of modify-ing (susceptibility) genes. There has been one negativelinkage study with the CRH gene published to date[Stratakis et al., 1997]. This is in agreement with themajority of genome scans in bipolar disorder, whichhave failed to detect significant linkage on chromosome8 [see Wildenauer et al., 1999 for review]. However,two recent studies suggest the possibility of linkage ofbipolar disorder to markers in the 8q24 [McInnis et al.,1999] and 8q [Liu et al., 1999] regions. The interpreta-tion of the linkage results may be hampered by the factthat such studies have limited power to detect genesthat contribute only a small portion of the total geneticvariability. Therefore, in addition to linkage, we usedan association design. In order to increase the samplehomogeneity, we included only patients with a typicalbipolar disorder (as confirmed by their clear-cut re-sponse to lithium prophylaxis). In the linkage study weseparately analyzed data from families of probandswho have shown abnormal positivity of DST in a long-term study. The association study was conducted on alarge, carefully assessed population followed by the In-ternational Group for the Study of Lithium (IGSLI).

MATERIALS AND METHODSSubjects

The subjects for this study were 138 bipolar patientsfollowed by the International Group for the Study ofLithium (IGSLI) in research clinics in Austria, Canada,

Czech Republic, Denmark, Germany, and Sweden.Their mean age was 50.0 ± 14.4 years, and the sex ratio(M/F) was 0.87. The average age of onset of BD was27.6 ± 9.9 years. All subjects were interviewed usingSADS-L semistructured interviews [Endicott andSpitzer, 1978] and diagnosed according to Research Di-agnostic Criteria (RDC) [Spitzer et al., 1978]. In addi-tion to a diagnosis of bipolar illness, all subjects metcriteria for definite prophylactic response to lithium asdescribed previously [Turecki et al., 1998]. The pro-bands had a high recurrence risk as evidenced by ahigh number of illness episodes prior to lithium treat-ment (8.2 ± 10.1) and they had been fully stabilized onlithium monotherapy for 14.4 ± 6.8 years. To furtherensure reliability of the diagnoses and uniform deter-mination of the lithium response, all case histories andrecords have been reviewed by a single senior clinicalinvestigator (PG) who also personally re-interviewedall subjects in Canada, Czech Republic, and Germany.

The comparison group consisted of 108 psychiatri-cally healthy subjects of similar age (mean 51.5 ± 14.8),sex ratio (0.86), and ethnic background. They weremarried-in family members in the linkage sample orhealthy volunteers.

Twenty-four probands had families available for alinkage study. A total of 171 subjects in these familieswere interviewed and genotyped, of these 72 were af-fected with a bipolar-spectrum disorder (bipolar I orbipolar II disorder, recurrent schizoaffective disorder,or recurrent unipolar depression). Diagnoses in rela-tives were also based on SADS-L interviews and RDCcriteria. The interviews were conducted by two psy-chiatrists blind to familial affiliation and group andfinal diagnosis was made by a consensus panel of atleast two additional psychiatrists using all availabledata including the diagnostic interview and collateralinformation in a blind fashion.

Six of the probands in the linkage study have beeninvolved in a longitudinal investigation of the HPA axisusing sequential dexamethasone suppression tests.They all showed intermittent positivity of the testwhile fully stabilized on lithium [Deshauer et al.,1999]. Families of these subjects were analyzed sepa-rately for linkage, as well as being included in theanalysis of the entire sample.

Laboratory Analysis

Genomic DNA was extracted by a standard method[Sambrook et al., 1989] from venous blood samples. Wetested two polymorphisms: CRH (GDB:162561, hetero-zygosity 0.72) on 8q13 and PENK (GDB:156549, het-erozygosity 0.53) on 8q11.23-8q12 (8q23-8q24 accord-ing to OMIM, 1996). Both markers are dinucleotiderepeats and are intragenic. The polymerase chain re-action was carried out in a total volume of 12.5 ml con-taining 40 ng of genomic DNA, 125 ng of the specificprimer, 200 mM each of dGTP, dCTP, and dTTP, 25 mMdATP, 1.5 mCi [35S]DATP, 0.5 units of Taq DNA poly-merase (Bio/Can Scientific), and 2.0 ml of 10× buffer(Bio/Can Scientific) with MgCl2 included in the finalconcentration of 1.5 mM. Samples were over laid withmineral oil and processed throughout 35 cycles of de-

CRH and PENK Genes in Bipolar Disorder 179

naturation at 94°C, annealing at 55°C, and elongationat 72°C, followed by a final elongation period of 72°C.Polymerase chain reaction products were analyzed on a6% denaturating polyacrylamide gel (38:2 acrylamide/bisacrylamide). Samples were run for a 2-hour periodin a vertical electrophoresis gel apparatus (Life Tech-nologies). Gels were dried and exposed to x-ray filmsfor 24 hours at room temperature. All marker determi-nations were made blind to clinical diagnoses.

Statistical Analysis

Association data were analyzed by x2 test with P val-ues determined empirically because of the low expectedcell frequencies. The method described by Zaykin andPudovkin [1993] was used both for comparison of allelefrequencies and for testing whether the genotype fre-quencies were in Hardy-Weinberg equilibrium.

The linkage data were analyzed both by nonpara-metric GENEHUNTER [Kruglyak et al., 1996] andparametric FASTLINK [Cottingham et al., 1993] meth-ods. The models for parametric analysis included re-cessive, dominant, and intermediate models with lowor high penetrance of the susceptibility genotype andwith or without phenocopies. The parameters of themodels were set so that the population frequency of thephenotype was 0.01 for males and 0.017 for femalesand the proportion of sporadic cases was between 0.45and 0.5 for models with phenocopies.

RESULTSAssociation Study

Neither CRH nor PENK polymorphisms were asso-ciated with BD in our sample (Tables I and II). Thegenotype frequencies of these markers in both patientand control groups conformed to Hardy-Weinberg equi-librium.

Linkage Study

In parametric linkage analysis, there was no evi-dence of linkage of either the PENK or CRH polymor-phism to BD under any of the models explored. Maxi-mum values of lod scores were obtained under a low-penetrance recessive model for CRH (Z 4 0.2 at u 40.3) and under a low-penetrance intermediate modelfor PENK (Z 4 0.18 at u 4 0.2). The results of non-parametric linkage analysis were also negative for bothCRH (NPL 4 −0.18; P 4 0.56) and PENK (NPL 4−0.07; P 4 0.52).

In the subset of six families of probands with docu-mented positivity of the DST, the lod score analysis didnot support linkage to any of the loci. The maximumlod score for PENK was 0.40 at u 4 0.1 under thedominant model with low penetrance. For CRH the

maximum lod score was 0.15 at u 4 0.4 also under thedominant model, with all lod scores negative for u < 0.2.

DISCUSSION

Our results argue against a significant role of theCRH and PENK genes in the susceptibility to bipolarillness. The findings were negative even in a subpopu-lation of individuals who exhibited a distinct HPA axisdysregulation. CRH represents only one element of thesystem regulating cortisol production. As the regula-tion involves multiple feedback loops, it is not easy todetermine whether causes upstream or downstreamfrom CRH secretion are responsible for the abnormalcortisol responses seen in mood disordered patients.For instance, it has been shown using a transgenicmouse model that HPA axis dysregulation could resultfrom abnormal glucocorticoid receptor gene function.This abnormality would manifest in an abnormal DST,which could be normalized following administration ofantidepressants [Barden, 1999].

As in other similar studies of candidate genes, thestatistical power is a significant factor limiting the in-terpretation of findings. Post hoc analysis shows thatour sample had approximately 75% power to detect acommon susceptibility gene associated with an oddsratio of 2 and more than 95% power for an odds ratio of3 or more. However, it is conceivable that for complextraits the relative risk attributable to a single locus issmaller than we had the power to detect in this study.Furthermore, our association study findings could befalse negative in an absence of linkage disequilibriumbetween the markers tested and functional alleles ofthe respective genes.

Our findings indicate that the HPA axis dysregula-tion, so common in bipolar mood disorders, is mostlikely due to factors other than an abnormality in CRHand PENK genes.

ACKNOWLEDGMENTS

We thank Carrie Robertson and Josh Wilson fortechnical assistance and help with manuscript prepa-ration.

TABLE I. CRH Allele Frequencies in BD Subjects and Controls (x2 4 13.89; df 4 8; Empirical P 4 0.055)

Group

Allele

156 154 152 150 148 146 144 138 128

BD (n 4 133) 0.000 0.015 0.000 0.523 0.365 0.068 0.023 0.004 0.004Controls (n 4 90) 0.006 0.000 0.022 0.461 0.378 0.100 0.022 0.000 0.011

Allele sizes in base pairs.

TABLE II. PENK Allele Frequencies in BD Subjects andControls (x2 4 6.93; df 4 3; Empirical P 4 0.063)

Group

Allele

83 81 79 77

BD (n 4 138) 0.007 0.496 0.478 0.018Controls (n 4 102) 0.005 0.525 0.412 0.059

Allele sizes in base pairs.

180 Alda et al.

REFERENCESArborelius L, Owens MJ, Plotsky PM, Nemeroff CB. 1999. The role of

corticotropin-releasing factor in depression and anxiety disorders. JEndocrinol 160:1–12.

Barden N. 1999. Regulation of corticosteroid receptor gene expression indepression and antidepressant action. J Psychiatry Neurosci 24:25–39.

Cottingham RWJ, Idury RM, Schaffer AA. 1993. Faster sequential geneticlinkage computations. Am J Hum Genet 53:252.

Deshauer D, Grof E, Alda M, Grof P. 1999. Patterns of DST positivity inremitted affective disorders. Biol Psychiatry 45:1023–1029.

Endicott J, Spitzer RL. 1978. A diagnostic interview: the schedule for af-fective disorders and schizophrenia. Arch Gen Psychiatry 35:773–782.

Holsboer F, Liebl R, Hofschuster E. 1982. Repeated dexamethasone sup-pression test during depressive illness: normalization of test resultscompared with clinical improvement. J Affect Disord 4:93–101.

Holsboer F, Spengler D, Heuser I. 1992. The role of corticotropin-releasinghormone in the pathogenesis of Cushing’s disease, anorexia nervosa,alcoholism, affective disorders, and dementia. Prog Brain Res 93:385–417.

Holsboer F, Lauer CJ, Schreiber W, Krieg J-C. 1995. Altered hypotha-lamic-pituitary-adrenocortical regulation in healthy subjects at highfamilial risk for affective disorders. Neuroendocrinology 62:340–347.

Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. 1996. Parametric andnonparametric linkage analysis: a unified multipoint approach. Am JHum Genet 58:1347–1363.

Linkowski P, Kerkhofs M, Van Onderbergen A, Hubain P, Copinschi G,L’Hermite-Baleriaux M, Leclercq R, Brasseur M, Mendlewicz J, VanCauter E. 1994. The 24-hour profiles of cortisol, prolactin, and growthhormone secretion in mania. Arch Gen Psychiatry 51:616–624.

Liu J, Aita VM, Knowles JA, Terwilliger JD, Baltazar R, Grunn A, Loth JE,Kanyas K, Lerer B, Endicott J, Wang Z, Penchaszadeh G, Gilliam TC,Baron M. 1999. Search for susceptibility loci in extended pedigreeswith bipolar affective disorder. Mol Psychiatry 4:S21.

McInnis MG, Mahoney AT, Chellis J, McMahon FJ, MacKinnon DF, Pot-ash J, Simpson SG, Lan TH, Beaty T, DePaulo JR. 1999. Genome widescan of bipolar disorder and evidence of parent of origin effect. MolPsychiatry 4:S20.

Nemeroff CB, Widerlov E, Bissette G, Walleus H, Karlsson L, Eklund K,Kilts CD, Loosen PT, Vale W. 1984. Elevated concentrations of CSFcorticotropin-releasing factor-like immunoreactivity in depressed pa-tients. Science 226:1342–1344.

Nemeroff CB, Owens MJ, Bissette G, Andorn AC, Stanley M. 1988. Re-duced corticotropin-releasing factor binding sites in the frontal cortexof suicide victims. Arch Gen Psychiatry 45:577–579.

Online Mendelian Inheritance in Man, OMIM (TM). 1996. Johns HopkinsUniversity, Baltimore, MD. MIM Number: 131330: World Wide WebURL: http://www.ncbi.nlm.nih.gov/omim/ (October 9, 1996).

Raadsheer FC, Hoogendijk WJG, Stam FC, Tilders FJH, Swaab DF. 1994.Increased number of corticotropin-releasing hormone expressing neu-rons in the hypothalamic paraventricular nucleus of depressed pa-tients. Clin Neuroendocrinol 60:436–444.

Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a laboratorymanual. Cold Spring Harbor, NY: Cold Spring Harbor LaboratoryPress.

Spitzer RL, Endicott J, Robins E. 1978. Research diagnostic criteria: ra-tionale and reliability. Arch Gen Psychiatry 35:773–782.

Stokes PE, Stoll PM, Koslow SH, Maas JW, Davis JM, Swann AC, RobinsE. 1984. Pretreatment DST and hypothalamic-pituitary-adrenocorticalfunction in depressed patients and comparison groups. Arch Gen Psy-chiatry 41:257–267.

Stratakis CA, Sarlis NJ, Berrettini WH, Badner JA, Chrousos GP, GershonES, Detera-Wadleigh SD. 1997. Lack of linkage between the cortico-tropin-releasing hormone (CRH) gene and bipolar affective disorder.Mol Psychiatry 2:483–485.

Turecki G, Grof P, Cavazzoni P, Duffy A, Grof E, Ahrens B, Berghofer A,Muller-Oerlinghausen B, Dvorakova M, Libigerova E, Vojtechovsky M,Zvolsky P, Joober R, Nilsson A, Prochazka H, Licht RW, RasmussenNA, Schou M, Vestergaard P, Holzinger A, Schumann C, Thau K, Rou-leau GA, Alda M. 1998. Preliminary evidence for a role of phospholi-pase C-gamma 1 in the pathogenesis of bipolar disorder. Mol Psychia-try 3:534–538.

Vieta E, Gasto C, Martinez de Osaba MJ, Nieto E, Canto TJ, Otero A,Vallejo J. 1997. Prediction of depressive relapse in remitted bipolarpatients using corticotrophin-releasing hormone challenge test. ActaPsychiatr Scand 95:205–211.

Wildenauer DB, Schwab SG, Blaveri A, Culverhouse R, Curtis D, EdwardsJ, Foroud T, Gershon ES, Kelsoe J, Lichtermann D, McMahon F, Nurn-berger J, Owen M, Todd R. 1999. Chromosomes 8 and 10 workshop. AmJ Med Genet (Neuropsychiatr Genet) 88:239–243.

Zaykin DV, Pudovkin AI. 1993. Two programs to estimate significance of x2

values using pseudo-probability tests. J Hered 84:152.

CRH and PENK Genes in Bipolar Disorder 181