Eventually, gene therapy may be a valid option for chronic viral infections, including retroviral infections. Human retroviral diseases fit two categories: (1) those that result from a monoclonal outgrowth of a human T-cell leukemia virus type I (HTLV-I)-infected cell, as in the case of adult T cell leukemia (ATL); and (2) those that appear to result directly from virus load rather than monoclonal outgrowth--such as tropical spastic paraparesis/HTLV-I associated myelopathy (TSP/HAM) and human immunodeficiency virus (HIV)-associated acquired immune deficiency syndrome (AIDS). For ATL gene therapy, corrective mechanisms directed at regulatory sequences rather than viral sequences may be most important, though perhaps anti-tax therapy would be useful. For TSP/HAM and AIDS, gene therapy directed to control virus replication may be most useful. For anti-retroviral therapy, one may use dominant negative mutants and a variety of other approaches that direct toxins or compete out viral regulatory gene signal sequences. For maximum benefit, such therapy should be directed to different essential genes (eg gag, pol, env, tat or rev) involved in the virus replication cycle and utilize different toxic approaches. A major impediment to the use of gene therapy for AIDS is our inability to transfect a significant fraction of target cells in vivo. Except for reconstituted mice, retroviral systems of animals have been under-utilized as models for gene therapy. Naturally occurring retroviral diseases of cats, goats, horses, and other species provide models for future development.