Replication-deficient adenovirus (Ad) vectors are effective in transferring genes in vivo, but their use is associated with significant variation in the extent and/or duration of expression observed among different strains of experimental animals and different routes of administration of the vector. We have minimized the variables of the heterologous transgene and animal-to-animal variation by constructing an Ad vector encoding murine thrombopoietin (mTPO, AdmTPO), a homologous protein that induces a physiologic response (elevation of blood platelet levels) that can be followed sequentially over time in the same animal. Using the C57BL/6 and BALB/c stains, liver administration was accomplished by intravenous administration and skeletal muscle administration by direct injection. Despite the use of a homologous cDNA as a transgene, the Ad genome was rapidly lost from the liver after intravenous administration over the first 1 to 2 weeks, with no difference in pattern of decline between the C57BL/6 and BALB/c strains. Both strains exhibited a cytotoxic T lymphocyte (CTL) response directed against the AdmTPO vector. Consistent with the decline in vector genome over time, the initial high levels of mTPO mRNA in the liver declined to an undetectable level within 2 weeks. Platelet counts peaked at 8- to 10-fold above baseline within the first 2 weeks, and then gradually declined, returning to normal level by 50 to 60 days. Intravenous administration of the AdmTPO vector to beta2-microglobulin-deficient mice resulted in a longer persistence of elevated platelets levels, although the eventual return of platelet levels to normal in these mice suggests the elimination of the Ad vector cannot be explained solely by CTL response. Although the intramuscular administration of the AdmTPO vector resulted in platelet levels with a lower peak and minor differences over time compared with the intravenous route, the C57BL/6 and BALB/c strains demonstrated the same rapid loss of Ad genome and mTPO mRNA levels in the muscle as in the liver. Together, these observations suggest that simplifying the experimental design by eliminating the variable of host response to a heterologous transgene, and by following the consequences of gene transfer in the same animals over time, there can be remarkable similarity in strain- and route-dependent responses to an Ad vector.