Celiac Disease and Schizophrenia

To the Editor: In the June issue, Karlsson et al. (1) report an association between high levels of antigliadin immunoglobulin (Ig) G in maternal circulation and elevated risk of nonaffective psychosis in offspring. The authors propose several possible mechanisms to explain this association, and I wish to suggest an alternative explanation. Because gliadin must cross the intestinal epithelium to evoke IgG antigliadin antibodies, greater intestinal permeability, both paracellular and transcellular, has been hypothesized as an early event in the development of celiac disease (2). Estrogens are described as playing a key role in the development and maintenance of the intestinal barrier (3). Bisphenol-A (BPA), an estrogenic endocrine disruptor, can prematurely and permanently close the barrier in perinatally exposed female but not male rats (3). As some degree of permeability is required for maturation of the immune system through the development of tolerance, the perinatally BPA-exposed female rat, lacking immunological tolerance, develops enhanced colonic inflammatory responses in adulthood (3). This would set the stage so that when the perinatally BPA-exposed female rat becomes pregnant, the pregnancy may be marked by enhanced inflammation. Paradoxically, estrogenic exposure may have anti-inflammatory effects in the exposed adult, but inappropriate estrogen exposure may have pro-inflammatory effects in the perinatally exposed offspring. These effects were observed at levels of BPA exposure previously believed to be too low for observed adverse effects in humans (3).

I have proposed elsewhere an estrogenic endocrine disruption theory of schizophrenia, in which inappropriate dosage, timing, or duration of prenatal estrogen exposure causes schizophrenia (4, 5). Within this theoretical framework, inappropriate estrogen exposure occurring in the brain could also be occurring in the colon so that an association of celiac disease or some other inflammation and schizophrenia may be observable not from a genetic link per se but rather a transgenerational effect of prenatal estrogen exposure.