Human monoclonal antibodies are desirable, especially as therapeutic agents, but the best means of producing them is still a matter of investigation. It is clear that human antibodies of predicted specificity from patients with autoimmune disease can be derived, and this may help unlock some of the mysteries of these illnesses. Human monoclonal antibodies against tumor-specific antigens for use in in vivo diagnosis and therapy remain desirable goals. Problems involved in their routine development include the lack of available, adequately immunized, and differentiated lymphocytes and the nature and paucity of the available human "myeloma" cell lines. These lines have been compared now by a number of authors who have reached similar conclusions. Our study directly compared the greatest number of cell lines and found UC729-6 and HF2 to be the best; on the other hand, our success in developing IgG-secreting hybridomas from U-266, using hyperimmunized lymphocytes, suggests that this line may only be capable of secretion with the more differentiated cell, the human equivalent of those hyperimmunized murine spleens. Hence both sides of the fusion equation must be made optimal. Two new approaches to circumvent this problem involve the use of either a human-murine myeloma chimera as the parental myeloma line or, more recently, genetic engineering techniques to substitute human constant regions for the murine while retaining the murine hypervariable region, preserving the binding specificity of the murine antibody.