Advances towards the understanding of the etiological mechanisms involved in mood disorders provide interesting yet diverse hypotheses and promising models. In this context, molecular genetics has now been widely incorporated into genetic epidemiological research in psychiatry. Affective disorders and, in particular, bipolar affective disorder (BPAD) have been examined in many molecular genetic studies which have covered a large part of the genome, specific hypotheses such as mutations have also been studied. Most recent studies indicate that several chromosomal regions may be involved in the aetiology of BPAD. Other studies have reported the presence of anticipation in BPAD and in unipolar affective disorder (UPAD).^{(1, 2 and 3)} In parallel to these new developments in molecular genetics, the classical genetic epidemiology, represented by twin, adoption and family studies, provided additional evidence in favour of the genetic hypothesis in mood disorders. Moreover, these methods have been improved through models to test the gene-environment interactions.

In addition to genetic approaches, psychiatric research has focused on the role of psychosocial factors in the emergence of mood disorders. In this approach, psychosocial factors refer to the patient’s social life context as well as to personality dimensions. Abnormalities in the social behavior such as impairment in social relationships have been observed during episode of affective disorders, and implicated in the etiology of affective disorders. Further, gender and socio-economic status also emerged as having a possible impact on the development of affective disorders. Finally, the onset and outcome of affective disorders could also be explained by interactions between the social life context and the individual’s temperament and personality. The importance of temperament and personality characteristics in the etiology of depression has been emphasized in various theories, although disagreement exists with regard to terminology and the etiology.

While significant advances have been done in these two major fields of research, it appears that integrative models, taking into account the interactions between biological (genetic) factors and social (psychosocial environment) variables offer the most reliable way to approach the complex mechanisms involved in the etiology and outcome of mood disorders. This chapter will review some of the most promising genetic and psychosocial hypotheses in mood disorders that can be integrated in interactive models.

The various strategies available to investigate genetic risk factors in psychiatric disorders belong to the wider discipline of genetic epidemiology. This combines both epidemiological and genetic investigations and has the primary objective of identifying the genetic and non-genetic (environmental) causes of a disease. Genetic epidemiological data in affective disorders has come mostly
from family, twin, adoption and segregation (within families) studies. Family, twin and adoption studies are the mainstay in establishing the genetic basis of affective disorders. These methods firstly demonstrated that genetic factors are involved in the aetiology of these disorders.⁽⁴⁾ Twin and adoption data may also be used to investigate the relative contributions of genetic and environmental factors to the aetiology of a disease.⁽⁵⁾ The exact contribution of these factors is not yet firmly understood for affective disorders but some studies provide contributing findings. The study of adoptees who are separated from their biological parents has consistently favoured the gene-environment hypothesis in the aetiology of diverse psychiatric disorders.⁽⁶⁾ In adoption studies, both UPAD and BPAD have been described to be more frequent in biological relatives of the adopted subject, suffering from affective disorders than in adopted relatives.⁽⁷⁾

The diagnostic validation and the structure of the genetic and environmental risk factors in mood disorders are also approached in twin studies.⁽⁸⁾ From the landmark study of Rosanoff et al. in 1934 to more recent works from Mc Guffin et al.⁽⁹⁾ and Kieseppa et al.,⁽¹⁰⁾ concordance rates for BPAD are higher in monozygotic twins (20–100 per cent) than in dizygotic twins (0–38 per cent).

The rapid advance in molecular genetic techniques over the last decade has generated a large database of DNA markers across the whole human genome and has enabled chromosomal regions throughout almost the entire genome to be studied in affective disorders. These studies have been performed mainly using linkage and association methodologies. Current linkage and association methods investigate heritable factors at a molecular genetic level, and enable genes to be mapped.⁽¹¹⁾ These approaches are mostly applied to BPAD, which is considered to be the ‘core’ phenotype in affective disorders. Linkage analysis tests the hypothesis that a linkage relationship exists between a known genetic marker and a trait which is known to be genetically determined but has not yet been mapped on a chromosome.⁽¹²⁾ Two genetic loci are linked if they are located closely together on a chromosome. In linkage analysis, the distance between a marker locus and the gene under investigation is used for gene mapping. This method was originally designed to explore a major single genetic transmission and to evaluate the extent of co-segregation between genetic markers and the phenotype investigated in pedigrees. The major problems which linkage methodology face when applied to affective disorders are the complex aetiology and inheritance patterns. More than one locus are probably involved in susceptibility to these disorders, and the exact mode of transmission is not known. Mis-specification of the genetic parameters of the phenotype may lead to errors in linkage studies.⁽¹²⁾ Furthermore, the linkage approach fails to detect minor gene effects which contribute to genetic susceptibility to the disorder.⁽¹³⁾ More recently, genome-wide linkage studies have been performed on samples of families with multiply affected members.⁽¹⁴⁾

The association method offers an alternative strategy of studying genetic factors involved in complex diseases in which the mode of transmission is not known.⁽¹⁵⁾ The association strategy does not require genetic parameters to be known (non-parametric method). The purpose of association studies is to compare frequencies of genetic marker alleles in patient and control populations in order to detect linkage disequilibrium. Linkage disequilibrium between the disease locus and the marker tested is defined as a level of concordance between the two loci which is higher than would be expected by chance. The major reason for this is their proximity on a chromosome. The major advantage of association studies is that they can detect genes with minor effects other than a single major locus (SML). The major limitation of this approach is that spurious associations between a genetic marker and a disorder may result from variations in allele frequency between cases and controls observed if the two populations are ethnically different (population stratification). It is important in this case to compare populations which are homogenous in their ethnic background. A further major difficulty in association studies is the interpretation of the precise meaning of the association observed.⁽¹⁶⁾ The result may be interpreted as linkage disequilibrium between the disease locus and the associated marker allele(s). Alternatively, the associated marker may be interpreted as a susceptibility factor which is directly involved in the disease. The candidate gene approach in association studies is a useful method to investigate linkage between markers and diseases. A candidate gene refers to a region of the chromosome which is potentially implicated in the aetiology of the disorder concerned. The possibility of false positive results must be taken into account, as a very large number of candidate genes now exist. The probability that each of these genes is involved in the aetiology of the disorder is relatively low.

From more than two decades of linkage studies, it seems that several chromosomal locations have been associated with affective disorders, sometimes with conflicting results. Mendlewicz et al⁽¹⁸⁾ first reported possible genetic linkage between manic depression and coagulation Factor IX (F9) at Xq27 in 11 pedigrees. Another region of interest seems to be the chromosome 18 where the pericentromeric region was suggested to carry susceptibility genes. The chromosome 11 has been thoroughly investigated in AD but showed contradictory results. Chromosomes 4, 6, and 10 were also investigated with conflicting and/or unreplicated results. Darier’s disease (keratosis follicularis), a rare autosomal dominant skin disorder associated with increased prevalence of epilepsy and mental retardation, whose gene was mapped on chromosome 12 (12q23–24.1), was found to cosegregate with BPAD in one pedigree. This result was replicated in several family studies. Genome-wide linkage analyses provide an accurate tool to study regions of interest. In BPAD, early positive and promising results were contradicted by further analyses. This fact is not surprising, since these studies were performed on small samples sizes, insufficient to replicate modest linkage signals.⁽¹⁹⁾ Meta-analyses were thus performed on BPAD to increase the power to detect modest linkage signals.⁽¹⁴⁾ Bipolar loci with evidence of linkage were found on the following arms: 4p, 6p, 6q, 9p, 10q, 12q, 13 q, 14q, 17q, 18p–q, 21q, 22q.^(14,20) McQueen et al.⁽²¹⁾ found susceptibility loci on chromosomes 6q and 8q by using a combined analysis of eleven linkage studies. A recent study from Schumacher et al.,⁽²²⁾ in four European samples, confirmed previously reported loci, 4q31 and 6q24, and provided evidence for a new linkage locus, 1p35-36.

One of the main difficulties in clinical practice is the inability to know a priori which psychotropic drug will be best suited for each case. Therefore, several groups worldwide try to overcome that obstacle by searching genetic markers that might be predictive of treatment response.

The first studies on the relationship between response to lithium and family history have been published in the 1970s, supporting an association between a family history of BPAD and satisfactory response to treatment. Mendlewicz et al.⁽⁵⁰⁾ first reported a study of 36 patients through a double blind study of lithium prophylaxis. They found that 66 per cent of the responders to lithium had at least one first-degree relative with BPAD, and that only 2–1 per cent of the lithium nonresponders had a first-degree relative with BPAD. Lipp et al.⁽⁵¹⁾ first reported an association between DRD2 and nonresponse to lithium. Several studies have followed (See⁽⁵²⁾ for review). Positive results were found in TPH and 5HTT.^(53,54)