We have evaluated senescence related changes in canine red blood cells (RBCs) using the biotinylation system, where RBCs are labeled in vivo with biotin at the beginning of their life span, and retrieved from circulation on immobilized avidin at the end of their life span. This approach avoids the controversial use of density gradient centrifugation to collect presumably old RBCs. Furthermore, the dog is an appropriate model for human RBC senescence because it has a low degree of random RBC loss and a similarly long RBC life span (approximately 110 days). Two dogs had 97% to 100% of their circulating RBCs biotinylated by infusion of N-hydroxysuccinimido biotin (Clontech, Palo Alto, CA; Calbiochem, La Jolla, CA) dissolved in dimethyl sulfoxide. At postbiotinylation days 104 and 107 for one dog and day 110 for the other dog, biotinylated RBCs were isolated by magnetic cell sorting and analyzed for the presence of autologous IgG using 125I- labeled sheep-antidog IgG (SAD IgG). On all 3 days, there were at least three times more SAD IgG molecules per RBC on senescent biotinylated RBCs than on control (unfractionated) RBCs (day 104: 11,677 v 3,399; day 107: 6,710 v 2,115; day 110: 6,042 v 1,838 molecules of SAD IgG per senescent v control RBC). Furthermore, it is unlikely that an immune response to the conjugated biotin had been elicited, because fresh in vitro biotinylated RBCs that were incubated in autologous plasma (taken after exposure to circulating biotinylated RBCs for 113 days) and then exposed to the SAD IgG showed no increase in antibody binding over control (non-biotinylated) RBCs (1,431 v 1,378 cpm/10(8) biotinylated v control RBCs; P > .20). These results suggest that senescence of canine biotinylated RBCs is characterized by binding of autologous IgG and that antibiotin antibodies do not contribute to this process.