At present, virus replication is quite efficiently blocked by highly active antiretroviral therapy (HAART). However current HAART therapies still have important limitations that compromise in the long term the health of the patient. This failure is due to the existence of viral sanctuaries of replication-competent HIV-1, which can persists for the whole life. Two main strategies of HIV-1 gene therapy have been tried. One is based in the rationale of killing the infected cells before the virus can produce infective particles and the other strategy is to protect the cell from the virus. The first strategy uses lethal genes that are expressed in response to the expression of a HIV-1 provirus. Although technical limitations and safety concerns have limited the use of this approach in the last years, recent development of efficient lentiviral vectors and the need of the elimination of the latently infected cells are reasons to reconsider and come back to use this approach. Elimination of latently infected cells would also represent a perfect complement of the HAART therapy and may represent a hope in the long term for the development of a real cure of this infection.