Drs. DeLisi and Crow Reply

To the Editor: We are grateful that the manuscript we co-authored has provoked some response and thought from others. In the face of the inconsistencies of linkage studies, Dr. Pittelli draws attention to the susceptibility of genome scans to false positive findings and suggests that schizophrenia and other mental disorders, even though shown to be heritable in twin studies, are not actually genetic and says that “the mind is a complex thing.” While we think that the twin and adoption study evidence cannot be discarded, we have some sympathy with this opinion if it refers to the gene sequence rather than its expression. In this case, the proposal is consistent with our suggestion that the variation is “epigenetic,” i.e., related to methylation of the gene sequence, a suggestion that is similar to Dr. Peedicayil’s concept of an epimutation. The problem in testing this possibility is that of identifying a gene sequence or sequences (in the absence of linkage) the epigenetic status of which can be investigated.

However, we see no merit in Dr. Brown’s proposal that psychosis represents an anomaly of expression of a protein that protects against infection and other stresses (a hypothesis that resembles that of Huxley et al. [1]) because there is no evidence that patients with psychosis have this advantage, and we see no support in the literature for the particular genetic candidate that Dr. Brown suggests.

Over the past decade or more, the use of new molecular genetic linkage strategies have resulted in many positive reports emerging in the search for genes for schizophrenia and with results on only very small numbers of families. Sherrington et al. (2) in 1988 claimed a locus on proximal chromosome 5q that was never replicated. Pulver and colleagues (3–5) reported at various times positive linkages of schizophrenia to chromosomes 3, 8, 13, and 22. Still others reported additional linkages on chromosomes 1q, 2p, 2q, 4p, 6p, 6q, 10p, 15q, 18p, and X (reviewed in reference 6). In brief, almost every chromosomal arm has been reported to have a linkage to schizophrenia. Thus, if we place our study in context with past literature, although we find a significantly positive lod score on both chromosomes 2p and 10p, we fail to find the several other “positives” previously reported in our group of 382 sibling pairs. Thus, when Dr. Brown writes that our linkage on chromosome 2 is consistent with a gene for resistance to infection, his argument is clearly very weak from this angle as well. Moreover, Dr. Brown indicates that its location on chromosome 2q35 suggests this relationship is based on our linkage to chromosome 2p-q. However, chromosome 2 is one of the largest chromosomes, and the linkage we report is not at all linked to the region for the NRAMP1 gene but is several centimorgan map distances away. More important is the fact that Dr. Brown’s hypothesis does not explain any of the facts we know about schizophrenia today—i.e., that it is likely a developmental structural brain disorder, that it has an onset in early adulthood and continues through much of the lifetime of the affected individual thereafter, and that it has sex differences in age at onset, outcome, and symptoms. These clues need to be considered in the formulation of any hypothesis about the pathophysiology of this disorder.

While reports exist now of candidate genes that resulted from reports of linkage on chromosomes 6p (7), 8p (8), and 13q (9), these genes have not yet been determined to show specific modification in multiple members with schizophrenia within families. Another possible gene (protocadherin X-Y) proposed by one of us (10) has the unique advantage of explaining several of the facts just mentioned about schizophrenia. This gene is located on a recently evolved region of the X and Y chromosomes. On the basis of a combination of evidence from individuals with sex chromosome anomalies and the uniquely human aspects of schizophrenia, it has been proposed as a candidate gene. Since this gene is located in a region that has homologous genes on both X and Y chromosomes, even with an assumed sequence variation in the gene, linkage studies that rely on assumptions of standard autosomal or X-lined inheritance may not detect it. If epigenetic modification of the gene is the crucial event leading to schizophrenia, it is also possible that linkage studies will not detect it. This region thus is being further pursued closely in our families with schizophrenia to determine whether either or both situations are present, i.e., a sequence variation or abnormal methylation in one of the exons or the promotor of the gene.

Reprints are not available; however, Letters to the Editor can be downloaded at http://ajp.psychiatryonline.org.