Influenza virus is characterized by frequent and unpredictable changes of the surface glycoproteins which enable the virus to escape the immune system. Approved vaccines which consist of the whole virus or the surface glycoproteins fail to induce broad specificity protection. We have previously reported that a peptide-based experimental recombinant vaccine which includes conserved epitopes of B and T lymphocytes was efficient in mice, leading to cross-strain, long-term protection. In the present study, this approach was adapted for the design of a human vaccine, based on epitopes recognized by the prevalent HLAs. These epitopes were expressed in Salmonella flagellin and tested for their efficacy in human/mouse radiation chimera in which human peripheral blood mononuclear cells (PBMC) are functionally engrafted. The vaccinated mice demonstrated clearance of the virus after challenge and resistance to lethal infection. The production of virus-specific human antibodies was also higher in this group. Control groups of either non-vaccinated, or vaccinated mice which had not been engrafted with the human PBMC, did not exhibit the protective immune response. FACS analysis showed that most human cells in the transplanted mice are CD8(+) and CD4(+). Hence, it may be concluded: (i) that the protection involves cellular mechanisms, but is most probably accomplished without direct lysis of influenza-infected pulmonary cells by cytotoxic T lymphocytes, but rather via a cytokine-mediated mechanism, (ii) that the human/mouse radiation chimera model may be of some value in the investigation of new vaccines, as an additional tool prior to clinical trials, and (iii) that the synthetic recombinant vaccine can induce a response in the human immune system and confers protection against influenza infection. Further investigation is needed to establish the efficacy of such a peptide vaccine in human subjects.