To the Editor: We thank Dr. Kato for his kind letter commenting on our report of a nominal association between a mitochondrial DNA variant at nucleotide position 10398 and bipolar disorder. We are intrigued by his finding that this same variant may also be associated with bipolar disorder in a Japanese population. The association of the same mitochondrial DNA variant with bipolar disorder in two ethnically distinct populations increases the likelihood that the association is genuine rather than due to the ethnic particulars of a single study population. But the story is not straightforward.

The A/G polymorphism at nucleotide position 10398 occurs within the gene for NADH dehydrogenase subunit 3, a major component of cellular respiration. This single nucleotide polymorphism changes a threonine to an alanine and introduces a DdeI restriction site, making for easy detection in the laboratory. (Subjects who lack this restriction site may or may not carry the A allele; this can be established definitively only by sequencing.) A GENBANK search showed that both threonine and the alanine residues have been observed among primates, arguing against a significant physiological effect.

The two alleles at nucleotide position 10398 subdivide the mitochondrial DNA superhaplogroup L3, which encompasses all mitochondrial DNA seen in Europe and Asia, into two major subsets (1). The subset defined by the A allele is quite common among Europeans, occurring in 74% of those studied by Torroni et al. (2). Thus, it appears that the nucelotide position 10398 polymorphism arose early in human history, before the divergence of the European and Asian populations (2).

We compared the rates of the A allele in the two groups of probands and in comparison populations. In the Kato (Japanese) clinical group, 22% of the comparison subjects and 33% of the bipolar probands had the A allele. In our (European American) clinical group, 64% of the comparison subjects and 78% of the bipolar probands had the A allele. In the Torroni et al. population sample (2), 34% of the Asian subjects and 74% of the Caucasian subjects had the A allele.

These data show that the differences between probands with bipolar disorder and comparison subjects observed within Dr. Kato’s and our studies are small compared to the differences between the populations from which the subjects were drawn. Furthermore, the A allele occurred less frequently in both groups of clinical comparison subjects than in the corresponding population sample for each ethnicity. Although the population samples presumably included some people carrying susceptibility alleles for bipolar disorder, the meaning of these differences is not immediately apparent. Could this reflect a common susceptibility allele of low penetration? A recently published study of a British population (3) also failed to show an association between bipolar disorder and the nucleotide position 10398 polymorphism. On the basis of these data, we stand by our original conclusion that a pathogenic role of the nucleotide position 10398 polymorphism in bipolar disorder appears unlikely, but we agree that it cannot be entirely ruled out, especially in light of Dr. Kato’s new results. More data from larger groups of Europeans and Asians might clarify the situation.