The naturally occurring food constituent d-limonene has been found to cause tumors at high doses only in the kidney of the male rat in association with the development of hyaline droplet nephropathy. In contrast, neither kidney tumors nor the associated nephropathy have been found in female rats or mice at much higher doses. Adult male rats produce large quantities of a specific low-molecular-weight protein in the liver, which is known as alpha 2U-globulin (alpha 2U-g). With administration of sufficient doses of d-limonene to male rats, this protein has been found to accumulate excessively in the P2 segment cells of renal proximal tubules, resulting in hyaline droplet formation as a manifestation of protein overload. Hyaline droplet accumulation is the first stage in a unique sequence of nephropathic lesions (also known as alpha 2U-g nephropathy), including granular casts in the outer medulla and linear mineralization in the papilla. The mechanism underlying protein accumulation appears to be the reversible binding of chemical to alpha 2U-g with subsequent prolongation of its half-life in the tubule cell. In the case of d-limonene, the minor metabolite d-limonene-1,2-oxide has been shown to be the primary chemical species that binds reversibly to alpha 2U-g, impeding the normal process of lysosomal proteinase degradation of alpha 2U-g. The ensuing nephropathy is associated with a sustained increase in compensatory renal tubule cell proliferation, which provides the putative mechanistic link with renal tumor formation possibly through tumor promotion of spontaneously initiated cells or enhanced spontaneous mutagenesis. This proposed mechanism has been supported by additional information, including negative genotoxicity tests for d-limonene and its oxide metabolites, experimentally verified tumor promotion, and enhanced cell proliferation primarily in P2 segment tubule cells in male F344 rats, but no such effects in the alpha 2U-g-deficient NBR rat. The mechanism of d-limonene tumor development does not appear to be possible in humans since neither the quantity nor the type of protein that binds d-limonene or d-limonene-1,2-oxide is present. The deduction that the renal tumors induced in male rats are not relevant to human carcinogenicity in the hazard evaluation step of risk assessment completes the evaluation of human risk for d-limonene. Consequently, it can be concluded that d-limonene does not pose any carcinogenic or nephrotoxic risk to humans.(ABSTRACT TRUNCATED AT 400 WORDS)