by John R. Kelsoe, M.D.

The highs and lows of bipolar disorder can be considered a metaphor for the
exciting and sometimes frustrating search for genes for this complex
disorder. Despite some early starts and stops, this field is now poised to
make significant advances regarding the biology of bipolar disorder. This
review will summarize the background, technology and recent history of the
search for genes and describe the exciting promise of its future impact on
clinical practice.

Over many decades, numerous epidemiological studies have pointed to the
likely role of genetic factors in the predisposition to bipolar disorder
(Tsuang and Faraone). These data come primarily from three types of studies:
family, twin and adoption. In a family study, the familial nature of the
disorder is explored by determining the rate of illness among the
first-degree relatives of a bipolar proband. On average, these studies
indicate an approximately 7% risk to family members. Compared to a 1%
lifetime prevalence rate in the general population, this represents a
substantial increase in risk within families. Though these data indicate
that bipolar disorder is familial, they do not prove that it is genetic. To
separate nature from nurture, twin studies and adoption studies have been
employed.

In twin studies, both monozygotic (MZ) and same-sex dizygotic (DZ) twin
pairs are identified in which at least one twin has bipolar disorder. The
portion of twin pairs in which the co-twin is also affected is referred to
as the concordance rate. In the most common strategy, the twin pairs have
been raised together, so that presumably environmental factors are the same.
The role of genes is indicated by a higher concordance rate among MZ twins,
who share all their genes, compared to DZ twins. Though these studies have
varied greatly in methodology, on average they indicate an approximately 65%
concordance for MZ twins versus a 15% concordance for DZ twins. This
represents an over four-fold increase in risk and is the strongest evidence
for the role of genes.

Adoption studies separate nature from nurture by examining the rates of
illness in the biological parents of probands with bipolar disorder who were
adopted away at birth. Though fewer data are available for this more
difficult strategy, these studies also provide some support for the role of
genetics.

Though the epidemiological data indicating the role of genes have been
available for some time, only recently have methods been developed to
identify the specific genes involved. This process, referred to as
positional cloning, includes two steps: genetic mapping and physical
mapping. In the genetic mapping or linkage mapping stage, the approximate
chromosomal location of a disease susceptibility gene is determined by
identifying a DNA marker whose alleles are transmitted along with the
illness to the affected members in a family. It is the development of DNA
marker technology that made genetic mapping possible. These markers are
highly variable between individuals and have been mapped to precisely known
chromosomal locations. Thousands of such markers have now been developed for
this purpose, making it possible to systematically search the genome for
disease genes. Once a disease locus has been mapped by genetic methods, its
location has been narrowed from the 3 billion base pairs of DNA which make
up the entire genome to a specific region on a chromosome typically 5 to 10
million base pairs in size. Physical mapping methods, such as cloning and
sequencing large regions of DNA, are then required to identify the disease
gene. Ultimately, one or more mutations are found in the disease gene, which
cause it to malfunction, resulting in the disorder.

The first reported success using these molecular genetic methods in
psychiatry was a study by Egeland and others of bipolar disorder in a large
Old Order Amish kindred. These authors examined two markers on the end of
chromosome 11 and found strong evidence of genetic linkage. The story was
made more enticing by the recent mapping of the gene for tyrosine
hydroxylase, which controls catecholamine biosynthesis, to the same
chromosomal region (Meloni and others). This finding resulted in a great
deal of excitement about the prospects of applying positional cloning
methods to psychiatric disorders. Unfortunately, a number of subsequent
studies in numerous other populations failed to replicate the finding.
Furthermore, a subsequent study of an expanded version of the same Amish
pedigree failed to find significant evidence of linkage (Kelsoe and others
1989).

A similar course of events occurred at about the same time regarding a
report of linkage on the X chromosome. A variety of epidemiological data had
for many years suggested the possibility of a gene for bipolar disorder on
the X chromosome. For example, some studies indicated a lower rate of
father-to-son transmissions of illness than would be expected. This is
consistent with an X-linked gene. In 1987, Baron and colleagues examined
several conventional genetic markers including color blindness and
glucose-6-phosphate hydrogenase deficiency in a set of Israeli families. The
researchers found strong evidence of linkage to the Xq28 region. In similar
fashion to the report in the Amish, this finding caused considerable
excitement and stimulated several efforts to replicate the finding in other
populations. Unfortunately, no other study could replicate the linkage
result. Subsequently, a reexamination of the same families using molecular
genetic markers resulted in significantly diminished evidence for linkage
(Baron and others 1993).

It is in those two findings that the course of genetic research seems to
mirror the highs and lows of the illness itself. Investigators in psychiatry
and genetics were left disappointed and puzzled by those events. They were
also left with the question of how those reversals could occur given the
strength of the original findings. Though the answer to this question is
still not clear, a likely culprit is the genetic heterogeneity of bipolar
disorder. Epidemiological data suggest that many genes are involved in the
genetic transmission of bipolar disorder. They may be a complex combination
of genes of large effect and small effect, as well as autosomal dominant,
recessive and X-linked transmission.

Multiple Genes?

In many families the susceptibility to illness may require the interaction
of multiple genes. Other factors also complicate the problem. The twin
studies indicate that some individuals may inherit genes for the disorder,
yet never display symptoms. This reduced penetrance of the genes makes it
difficult to assess the status of unaffected individuals in families.
Further problems are added by the difficulties in diagnosis, the range of
manifestations of the illness and the resulting uncertainties in the optimum
definition of illness. The initial response to these dramatic reversals was
a conservative one. Many investigators assumed that false positives would be
the most serious problem confronting such studies. Many, therefore, proposed
that much higher statistical thresholds (lod scores) must be achieved before
linkage could be confidently declared.

Subsequent experience seemed to defy this expectation. As more investigators
surveyed more of the genome in more families, they began to fear they would
reach the end of the genome with no linkages rather than finding frequent
high lod scores. The field, therefore, went through a discouraging period of
several years when no positive results were reported meeting the high
threshold assumed necessary for statistical significance.

After this period of scientific depression, the outlook has recently begun
to improve based on several new developments. Stimulated at least partly by
the discouraging results in psychiatry, geneticists developed alternative,
more sophisticated statistical methods for complex heterogeneous disorders.
Some of these methods (termed nonparametric methods) avoided the problems
resulting from assuming a given mendelian mode of transmission and instead
focused only on the sharing of marker alleles between affected individuals.
Other approaches have involved developing more complex models of the
interaction of multiple genes in the same families. Also, methods were
developed to determine how sensitive a given linkage result is to the
diagnosis in a given individual. This would prevent the dramatic change in
linkage statistics resulting from a new onset of illness, as occurred in the
original Amish chromosome 11 finding. Another important change in the field
has been a move to greater caution regarding both false positives and false
negatives. In other words, strong linkage results should not be too readily
embraced, and mildly suggestive results not too readily rejected. Rather,
support for a finding should come from multiple families, multiple markers,
a variety of statistical methods and ultimately independent replication in
different family collections.

Chromosome 18

Such support seems to be developing for several different regions of the
genome. One of the most promising is the peri-centromeric region of
chromosome 18. This finding was originally reported by Berrettini and
colleagues, who found suggestive evidence of linkage using conventional
model-based methods in several families at several markers on chromosome 18.
However, using nonparametric methods, such as the affected sibling pair
method, they found highly significant evidence for linkage. Subsequently,
several other groups have reported independent replication of these results,
though some found the strongest evidence for linkage at markers somewhat
distant from those originally reported. In their replication of the
chromosome 18 linkage, Stine and others found the strongest evidence of
linkage in families in which illness was transmitted through the fathers
rather than the mothers. Such paternal transmission suggests a
parent-of-origin effect as seen in other genetic disorders. It may indicate
genetic imprinting in the transmission of bipolar disorder.

Chromosome 21

Another promising region is chromosome 21. Straub and colleagues have
reported and Detera-Wadleigh and coworkers confirmed evidence of linkage to
several markers on chromosome 21 (21q22.3). Though the strongest evidence
for linkage was found in one large American family, nonparametric
statistical methods indicated strong evidence for linkage in their entire
family collection. This linkage has been replicated recently by British
investigators who found evidence for linkage to both chromosome 21 and the
chromosome 11 region originally reported in the study of the Amish (Gurling
and others). They employed a statistical method that examined the effect of
both these loci together in their family collection. Further support for the
original chromosome 11 region comes from an association study of the
tyrosine hydroxylase gene (Meloni and others; however, other investigators
have not replicated these results.) Several other regions of the genome
appear as promising spots for bipolar loci. Dawson and others observed
co-segregation of bipolar disorder and a rare skin disease, Darier's
disease, in one Welsh family. This autosomal dominant skin disorder has been
mapped to 12q23-q24.1. The researchers then examined markers near the
Darier's locus in several bipolar families not affected with Darier's
disease and found suggestive evidence of linkage to bipolar disorder.

The X Chromosome

A decade ago, Mendlewicz and colleagues focused attention on X chromosomal
markers for manic depression. Now, a recent report by a Finnish group of
linkage to X chromosome markers brings renewed interest to this chromosome
(Pekkarinen and others), as has the work of a French group (Lucotte and
coworkers). This writer's laboratory has recently reported suggestive
evidence of linkage to the locus for the dopamine transporter gene on
chromosome 5 (Kelsoe and others. In press.). As the site of action of
amphetamine and cocaine, this is a very interesting candidate gene for
bipolar disorder. Another group of investigators have also recently reported
suggestive evidence for linkage to loci on chromosome 16 (Ewald and others).

What lies ahead in the search for bipolar genes? These promising results
suggest that the approximate chromosomal location of several genes for
bipolar disorder may have already been identified. For those loci already
replicated, further work needs to be done to both more securely confirm the
result, and to more finely map the implicated chromosomal region. Of the
more preliminary results, some will likely be confirmed through replication,
while others may prove to be false positives.

Soon, it will be very exciting to see some of these loci move to the next
stage of the process, physical mapping. Here a task equally arduous to that
of genetic mapping will be faced. It will likely produce its share of
frustration and false positives, but ultimately will lead to the
identification of specific susceptibility genes and the mutations affecting
them.

The benefit of identifying such specific genes will be great and manifold.
First and most importantly will be elucidation of the basic brain
pathophysiology of the disorder. As many of these genes will likely be
previously unknown, they may point to errors in previously unknown neuronal
physiology or new aspects of known neurotransmitter systems. This may lead
to a better understanding of the mechanism of action of current therapeutic
agents, and may identify targets for the development of new drugs. Such new
drugs may potentially be much more specific and efficacious in their action.
Having the specific genes in hand may make possible new models of illness at
the protein, cellular and, through transgenic technology, organismic levels.

A transgenic "manic-depressive" mouse would be an invaluable model for study
of disease mechanisms. A genetically based knowledge of pathophysiology may
also lead to a very different diagnostic system based on different disease
mechanisms. Such a different classification system might better predict both
prognosis and treatment response.

Gene Therapy

Genetic testing might also become available to identify children or adults
at risk for the disorder. This information could be invaluable for early
intervention, but also raises complex ethical and confidentiality issues.
Lastly, the major new therapeutic modality of the 21st century may likely be
gene therapy. This technology is now in early developmental stages, but
promises to allow the treatment or cure of genetic disorders by the
administration of DNA itself as a therapeutic agent. By delivering a normal
copy of the disease gene or otherwise directly manipulating gene function,
powerful treatments may become available for previously incurable genetic
disorders. Delivery of gene therapy agents to the brain and methods to
address psychiatric disorders are formidable problems. However, it is likely
that these methods may one day be applied to severe psychiatric disorders.
In short, the hope is that current genetic research may be the stepping
stone to a new era of biological psychiatry based on much more specific
knowledge and treatment of this debilitating illness.

Dr. Kelsoe is a psychiatrist who was trained at the University of Alabama,
the University of California, San Diego, and the National Institute of
Mental Health. He has spent the last 10 years hunting for bipolar genes, and
is now an assistant professor in the department of psychiatry at the
University of California, San Diego.

References

1. Baron M, Freimer NB, Risch N, et al. Diminished support for linkage
between manic-depressive illness and X-chromosome markers in three Israeli
pedigrees. Nat Genet. 1993;3(11)49-55.

2. Baron M, Risch N, Hamburger R, et al. Genetic linkage between
X-chromosome markers and bipolar affective illness. Nature.
1987;326(6110):289-292.

3. Berrettini WH, Ferraro TN, Goldin LR, et al. Chromosome 18 DNA markers
and manic-depressive illness: evidence for a susceptibility gene. Proc Natl
Acad Sci USA. 1994 (June);91:5918-5921.

4. Dawson E, Parfitt E, Roberts Q, et al. Linkage studies of bipolar
disorder in the region of the Darier's disease gene on chromosome
12q23-24.1. Am J Med Genet. 1995;60(2):94-102.

5. Detera-Wadleigh SD, Badner JA, Goldin LR, et al. Analysis of linkage to
bipolar illness on chromosome 21q. Am J Hum Genet. 1996;58:1279-1285.

6. Egeland JA, Gerhard DS, Pauls DL, et al. Bipolar affective disorders
linked to DNA markers on chromosome 11. Nature. 1987;325(6107):783-787.

7. Ewald H, Mors O, Flint T, et al. A possible locus for manic-depressive
illness on chromosome 16p13. Psychiatr Genet. 1995;5(2):71-81.

8. Gurling H, Smyth C, Kalsi G, et al. Linkage findings in bipolar disorder.
Nat Genet. 1995;10(1):8-9. Letter, comment.

9. Kelsoe JR, Ginns EI, Egeland JA, et al. Re-evaluation of the linkage
relationship between chromosome 11p loci and the gene for bipolar affective
disorder in the Old Order Amish. Nature. 1989;342(6247):238-243.

10. Kelsoe JR, Sadovnick AD, Kristbjarnarson H, et al. A possible locus for
bipolar disorder near the dopamine transporter on chromosome 5. Am J Med
Genet. 1996. In press.

11. Lucotte G, Landoulsi A, Berriche S, et al. Manic depressive illness
linked to factor IX in a French pedigree. Ann Genet. 1992;35(2):93-95.

12. Meloni R, Leboyer M, Bellivier F, et al. Association of manic-depressive
illness with tyrosine hydroxylase microsatellite marker. Lancet.
1995;345(8954):932. Letter. See comments.

13. Mendlewizc J, Simon P, Sevy S, et al. Polymorphic DNA marker on
X-chromosome and manic depression. Lancet. 1987;1(8544):1230-1232.

14. Pekkarinen P, Terwilliger JD, Bredbacka PE, et al. Evidence of a
predisposing locus to bipolar disorder on Xq24-q27.1 in an extended Finnish
pedigree. Genome Res. 1995;5:105-115.

15. Stine OC, Xu JF, Koskela R, et al. Evidence for linkage of bipolar
disorder to chromosome 18 with a parent-of-origin effect. Am J Hum Genet.
1995;57(6):1384-1394.

16. Straub RE, Lehner T, LuoY, et al. A possible vulnerability locus for
bipolar affective disorder on chromosome 21q22.3. Nat Genet. 1994;
8(3):291-296. See comments.

17. Tsuang MT, Faraone SV. The Genetics of Mood Disorders. Baltimore: Johns
Hopkins University Press; 1990.