BACKGROUND: Visceral leishmaniasis (VL) is a severe vector-born disease of humans and dogs caused by Leishmania donovani complex parasites. Approximately 0.2 to 0.4 million new human VL cases occur annually worldwide. In the new world, these alarming numbers are primarily due to the impracticality of current control methods based on vector reduction and dog euthanasia. Thus, a prophylactic vaccine appears to be essential for VL control. The current efforts to develop an efficacious vaccine include the use of animal models that are as close to human VL. We have previously reported a L. infantum-macaque infection model that is reliable to determine which vaccine candidates are most worthy for further development. Among the few amastigote antigens tested so far, one of specific interest is the recombinant A2 (rA2) protein that protects against experimental L. infantum infections in mice and dogs. METHODOLOGY/PRINCIPAL FINDINGS: Primates were vaccinated using three rA2-based prime-boost immunization regimes: three doses of rA2 plus recombinant human interleukin-12 (rhIL-12) adsorbed in alum (rA2/rhIL-12/alum); two doses of non-replicative adenovirus recombinant vector encoding A2 (Ad5-A2) followed by two boosts with rA2/rhIL-12/alum (Ad5-A2+rA2/rhIL12/alum); and plasmid DNA encoding A2 gene (DNA-A2) boosted with two doses of Ad5-A2 (DNA-A2+Ad5-A2). Primates received a subsequent infectious challenge with L. infantum. Vaccines, apart from being safe, were immunogenic as animals responded with increased pre-challenge production of anti-A2-specific IgG antibodies, though with some variability in the response, depending on the vaccine formulation/protocol. The relative parasite load in the liver was significantly lower in immunized macaques as compared to controls. Protection correlated with hepatic granuloma resolution, and reduction of clinical symptoms, particularly when primates were vaccinated with the Ad5-A2+rA2/rhIL12/alum protocol. CONCLUSIONS/SIGNIFICANCE: The remarkable clinical protection induced by A2 in an animal model that is evolutionary close to humans qualifies this antigen as a suitable vaccine candidate against human VL.