Cancer chemoprevention by dietary chlorophyllin (CHL) was investigated in a rainbow trout multi-organ tumor model. In study 1, duplicate groups of 130 juvenile trout were treated for 2 weeks with control diet, 500 p.p.m. dibenzo[a,l]pyrene (DB[a,l]P) or 500 p.p.m. DB[a,l]P + 2052 p.p.m. CHL, then returned to control diet. DB[a,l]P alone proved somewhat toxic but induced high tumor incidences in liver (61%), stomach (91%) and swimbladder (53%) 11 months after initiation. CHL co-feeding abrogated DB[a,l]P acute toxicity and reduced tumor incidences to 18% in liver, 34% in stomach and 3% in swimbladder (P </= 0.01). A second tumor and DNA adduct study using a non-toxic initiation protocol (200 p.p.m. DB[a,l]P +/- 4000 p.p.m. CHL for 4 weeks) confirmed these results. Potential CHL inhibitory mechanisms were investigated. Dietary CHL inhibited hepatic DB[a, l]P-DNA adducts in the two tumor studies by 89 and 76%, respectively. CHL was shown to complex strongly with DB[a,l]P (K(d1,2) = 1.59 +/- 0.01 microM, stoichiometry 2CHL:DB[a,l]P) and strongly inhibited DB[a,l]P mutagenesis in the Salmonella assay. Significant inhibition occurred at CHL concentrations substantially less than stoichiometric with DB[a,l]P and thus not reflecting simple DB[a,l]P sequestration via complexation. These initial findings suggest that CHL chemoprevention reflects complexation that might limit DB[a,l]P uptake in vivo, antimutagenic mechanisms such as catalytic degradation of the proximate electrophile in target cells, or both. These results demonstrate that dietary CHL is a reproducibly effective chemopreventive agent for DB[a,l]P multi-organ tumorigenesis in trout and suggest that reduced DB[a,l]P-DNA adducts may be predictive biomarkers of CHL reduction of DB[a,l]P-initiated hepatic tumors.