IFN-y occupies a central role as a potent modulator of both the affector and the effector limbs of the immune response . While promising as a therapeutic agent for immunodeficiency states, infectious and neoplastic diseases, and autoimmunity (1-3), IFN-y has perhaps more often been evoked in the pathogenesis of such disorders (4-9). In the latter regard particularly, recent evidence from transgenic mouse models has clearly demonstrated that overexpression of IFN-y can result in autoimmune destruction of tissues by inflammatory mechanisms (10). In the course of studying the cell and molecular basis for IFN-y gene regulation (11), we observed an IFN response consensus sequence in the most proximal 5' region of the human IFN-y gene very near a T cell-specific, inducible DNAse I hypersensitive site 250 by from that gene's cap site (12). Given the role of such response elements in upregulating IFN responsive genes (13, 14), we queried whether the product ofthe human IFN-y gene locus might somehow affect its own expression. In this article, definitive evidence is presented using both total human PBMCs and specific subsets ofthose cells, demonstrating a strong IFN-y autosuperinduction response to either IFN-y "priming or costimulation .Volume 170 September 1989170 September 1021170 September -1026 Materials and Methods
Brief Definitive ReportFor each experiment, human PBMC were isolated from 1 U of heparin-treated, random donor buffy coats (GulfCoast Regional Blood Center, Houston, TX) . All samples used were negative by HIV and hepatitis serologies. PBMC were isolated by isoïymph gradient centrifugation . Mixed leukocytes were washed in HBSS and suspended in RPMI 1640, 10% FCS, penicillin/streptomycin at 5 x 106 cells/ml before stimulation. PBMC separated into E+ and E-fractions (15) were first panned in plastic petri dishes at 37°C for 45 min to remove adherent cells . For reconstitution experiments, adherent monocytes from 3 x 107 cells were added back to equal numbers of rosetted cells before stimulation. For co-stimulati experiments, human rIFN-y (Genzyme Corp., Boston, MA) and PHA (final 1.0 g/ml) were added simultaneously, the mixture was incubated at 37°C for 6 h, and nonadherent cells were removed, washed three times in cold PBS, and used directly for total RNA isolation . In preincubation (primed) experiments, cells as above were first incubated with rIFN (or media alone) for 6 h at 37°C before addition of PHA . Preincubation of monocytes alone with IFN-y had no effect on subsequent upregulation of IFN-y transcripts . IFN-y-primed