Defects in the genes encoding the human coagulation factor VIII (hFVIII) and IX (hFIX) result in life-threatening haemorrhages and severe arthropathies. While haemophiliacs are currently treated by blood-derived factors or recombinant hFVIII and hFIX, a number of recent technical advances make the prospect of using gene therapy to treat such genetic diseases a realistic goal. Several gene therapy strategies have therefore been developed and evaluated in recent years. Most of the initial protocols were ex vivo gene transfer approaches in which the target cells (fibroblasts, keratinocytes, myoblasts, ...) were expanded and genetically-engineered in the laboratory and then implanted in the host. However, the complexity of most ex vivo gene therapy strategies, together with the disappointing results obtained in various animal models stimulated the development of more direct in vivo gene therapy protocols. In aiming to establish such an in vivo gene transfer protocol for haemophilia B, we constructed and tested in vitro and in vivo various recombinant adenovirus vectors expressing human FIX. Intravenous administration of this vector into various strains of immunocompetent and immunodeficient mice led to an efficient hFIX gene transfer in liver and lung. As a consequence, the hFIX protein was correctly produced and secreted at high levels in the blood of the treated animals. However, expression was transient in all immunocompetent mice, except surprisingly in C57B1/6 animals. A systematic molecular and immunological analysis allowed us to identify the parameters that prevent the long-term in vivo expression of the human molecule and to improve the current adenovirus vectors.