Environmental factors, especially the diet, play a prominent role in the epidemic of prostate cancer (PCA), in the United States. Many candidate dietary components have been proposed to influence human prostatic carcinogenesis, including fat, calories, fruits and vegetables, anti-oxidants, and various micronutrients, but the specific roles dietary agents play in promoting or preventing PCA remain controversial. We have collected evidence to suggest that GSTP1, the gene encoding the pi-class glutathione S-transferase (GST), may serve a "caretaker" function for prostatic cells. Although GSTP1 can be detected in normal prostatic epithelium, in almost all PCA cases, PCA cells fail to express GSTP1 polypeptides, and lack of GSTP1 expression most often appears to be the result of somatic "CpG island" DNA methylation changes. Loss of GSTP1 function also appears to be characteristic of prostatic epithelial neoplasia (PIN) lesions, thought to represent PCA precursors. We have recently learned that a new candidate early PCA precursor lesion, proliferative inflammatory atrophy (PIA), characterized by proliferating prostatic cells juxtaposed to inflammatory cells, contains epithelial cells that express high levels of GSTP1. These findings have formed the basis for a new model of prostatic carcinogenesis, in which prostatic cells in PIA lesions, subjected to a barrage of inflammatory oxidants, induce GSTP1 expression as a defense against oxidative genome damage. When cells with defective GSTP1 genes appear amongst the PIA cells, such cells become vulnerable to oxidants and electrophiles that inflict genome damage that tends to promote neoplastic transformation to PIN and PCA cells. Subsequently, PIN and PCA cells with defective GSTPI genes remain vulnerable to similar stresses tending to promote malignant progression. This new model for prostatic carcinogenesis has implications for the design of new prostate cancer prevention strategies. Rational prevention approaches might include: (i) restoration of GSTPI expression via treatment with inhibitors of CpG methylation, (ii) compensation for inadequate GSTPI activity via treatment with inducers of general GST activity, and (iii) abrogation of genome-damaging stresses via avoidance of exogenous carcinogens and/or reduction of endogenous carcinogenic (particularly oxidant) stresses.