Evaluation of biochemical changes during in vivo erythrocyte senescence inthe dog

One hypothesis to explain the age-dependent clearance of red blood cells (R BCs) from circulation proposes that denatured/oxidized hemoglobin (hemichro mes) arising late during an RBC's life span induces clustering of the integ ral membrane protein, band 3. In turn, band 3 clustering generates an epito pe on the senescent cell surface leading to autologous IgG binding and cons equent phagocytosis. Because dog RBCs have survival characteristics that cl osely resemble those of human RBCs (ie, low random RBC loss, approximate to 115-day life span), we decided to test several aspects of the above hypoth esis in the canine model, where in vivo aged cells of defined age could be evaluated for biochemical changes. For this purpose, dog RBCs were biotinyl ated in vivo and retrieved for biochemical analysis at various later dates using avidin-coated magnetic beads. Consistent with the above hypothesis, s enescent dog RBCs were found to contain measurably elevated membrane-bound (denatured) globin and a sevenfold enhancement of surface-associated autolo gous IgG. Interestingly, dog RBCs that were allowed to senesce for 115 days in vivo also suffered from compromised intracellular reducing power, conta ining only 30% of the reduced glutathione found in unfractionated cells. Al though the small quantity of cells of age greater than or equal to 110 days did not allow direct quantitation of band 3 clustering, it was nevertheles s possible to exploit single-cell microdeformation methods to evaluate the fraction of band 3 molecules that had lost their normal skeletal linkages a nd were free to cluster in response to hemichrome binding. Importantly, ban d 3 in RBCs greater than or equal to 112 days old was found to be 25% less restrained by skeletal interactions than band 3 in control cells, indicatin g that the normal linkages between band 3 and the membrane skeleton had bee n substantially disrupted. Interestingly, the protein 4.1a/protein 4.1b rat io, commonly assumed to reflect RBC age, was found to be maximal in RBCs is olated only 58 days after labeling, implying that while this marker is usef ul for identifying very young populations of RBCs, it is not a very sensiti ve marker for canine senescent RBCs. Taken together, these data argue that several of the readily testable elements of the above hypothesis implicatin g band 3 in human RBC senescence can be validated in an appropriate canine model. (C) 1999 by The American Society of Hematology.