Pre-existent humoral antibody to adenovirus potentially confounds human clinical trials involving intravascular administration of adenovirus. Using the LNCaP prostate cancer xenograft model in BALB/c nu/nu mice and the prostate-specific attenuated replication-competent adenovirus (ARCATM) CN706, we developed an animal model that systematically controls both the dose of intravascularly administered adenovirus and the titer of the pre-existent anti-Ad5 antibody, and then measures the virus-induced toxicity as well as antitumor activity. We prepared hyperimmune sera to adenovirus in rabbits, passively injected the purified rabbit anti-Ad5 antibody into tumor-bearing mice, and established measurable humoral anti-Ad5 antibody titers. CN706 was intravenously injected into the tail vein of animals 24 hr after passive anti-Ad5 antibody administration. In the absence of pre-existent antibody, the lethal dose (LD100) for BALB/c nu/nu mice was 2.5x10(11) CN706 particles, whereas 1x10(11) CN706 particles was not lethal. However, in the presence of a 1:80 pre-existent titer of Ad5 neutralizing antibody (NAb), intravenous injection of 5x10(11) CN706 particles was no longer lethal. In addition, pre-existent antibody also prevented antitumor activity in a dose-dependent manner: 1x 10(11) CN706 particles prevented LNCaP xenograft tumor progression, but antitumor activity was eliminated by a pre-existent 1:80 NAb titer. These results led us to propose transient removal of pre-existent adenovirus antibody by immunoapheresis. An affinity column of cloned virus capsid proteins was constructed that was able to specifically remove adenovirus antibody from human clinical serum samples. A 5-min disposable immunoassay was also developed to monitor the level of pre-existent antibody in sera before and after immunoapheresis. Clinically, this approach may enable controlled clinical studies of intravenously administered adenovirus in patients with pre-existent anti-adenovirus antibody.