Peroxisomes are subcellular organelles found in all eukaryotic cells. In the liver they are usually round and measure about 0.5-1.0 microns; in rodents they contain a prominent crystalloid core, but this may be absent in newly formed rodent peroxisomes as well as in human peroxisomes. A major role of the peroxisomes is the breakdown of long-chain fatty acids, thereby complementing mitochondrial fatty-acid metabolism. Many chemicals are known to increase the number of peroxisomes in rat and mouse hepatocytes. This peroxisome proliferation is accompanied by replicative DNA synthesis and liver growth. No clear structure-activity relationships are apparent. Many of these peroxisome proliferators contain acid functions that can modulate fatty acid metabolism. Two mechanisms have been proposed for the induction of peroxisome proliferation. One is based on the existence of one or several specific cytosolic receptors that bind the peroxisome proliferator, facilitating its translocation to the cell nucleus and the activation of the expression of specific genes. The second, perhaps more general, hypothesis involves chemically mediated perturbation of lipid metabolism. These two hypotheses are not mutually exclusive. Many peroxisome proliferators have been shown to induce hepatocellular tumours, despite being uniformly non-genotoxic, when administered at high dose levels to rats and mice for long periods. Three mechanisms have been proposed to explain the induction of tumours. One is based on increased production of active oxygen species due to imbalanced production of peroxisomal enzymes; it has been proposed that these reactive oxygen species cause indirect DNA damage with subsequent tumour formation. In rodents, an alternative mechanism is the promotion of endogenous lesions by sustained DNA synthesis and hyperplasia. Thirdly, it is conceivable that sustained growth stimulation may be sufficient for tumour formation. Marked species differences are apparent in response to peroxisome proliferations. Rats and mice are extremely sensitive, and hamsters show an intermediate response while guinea pigs, monkeys and humans appear to be relatively insensitive or non-responsive at dose levels that produce a marked response in rodents. These species differences may be reproduced in vitro using primary culture hepatocytes isolated from a variety of species including humans. The available experimental evidence suggests a strong association and a probable casual link between peroxisome-proliferator-elicited liver growth and the subsequent development of liver tumours in rats and mice. Since humans are insensitive or unresponsive, at therapeutic dose levels, to peroxisome-proliferator-induced hepatic effects, it is reasonable to conclude that the encountered levels of exposure to these non-genotoxic agents do not present a hepatocarcinogenic hazard to humans.(ABSTRACT TRUNCATED AT 400 WORDS)