Blue Dextran (Cibacron Blue F3GA-dextran) was immobilized on cyanogen bromide activated agarose and used as a ligand for human fibroblast and leukocyte interferons in a solvent of phosphate-buffered (pH 7.4), physiological saline (0.15 M NaCl). Fibroblast interferon binds completely and is not displaced from the column by an increase in ionic strength of the solvent (1.0 M NaCl); it can be, however, recovered with ethylene glycol, indicating the hydrophobic nature of interaction. Leukocyte interferon also binds to Blue Dextran-agarose but it can be recovered simply by an increase in the ionic strength of the solvent, indicating primarily the electrostatic nature of binding. Attempts to displace both interferons selectively with nucleosides and aromatic amino acids were unsuccessful. When Cibacron Blue F3GA is immobilized directly to agarose matrix or via molecular arm, the strength of binding of fibroblast interfern is significantly decreased, although ethylene glycol is still required for its displacement from the column. Leukocyte interferon, by contrast, does not bind at all under the same solvent conditions; it does bind when the pH value of the solvent is in the range 3-5 i.e., below its isoelectric point. Human fibroblast interferon binds completely to: aminobenzene, aminonaphthalene, and aminoanthracene, all immobilized on agarose, and it can be recovered with ethylene glycol. In contrast, human leukocyte interferon does not bind to benzene-agarose; it is retarded on naphthalene-agarose and completely retained on an anthracene-agarose column. All data point to a higher intrinsic hydrophobicity of human fibroblast interferon vis-á-vis human leukocyte interferon. Selective binding of human fibroblast interferon of Cibacron Blue F3GA-agarose results in a significant purification, about 800-fold.

Nonparenchymal liver cells from untreated C3HeB/FeJ mice, when incubated in medium containing-10% fetal bovine serum or portal serum, produced significant amounts of interferon alpha/beta (IFN alpha/beta). In contrast, other cell populations (spleen, mononuclear blood cells and peritoneal cells) from C3HeB/FeJ mice or nonparenchymal liver cells from other strains of mice (C3H/HeJ, germ-free C3H/HeN and C57Bl/6J) produced little or no detectable IFN in fetal bovine serum under the same culture conditions. The cells in the nonparenchymal liver cell population responsible for IFN alpha/beta production were adherent, phagocytic, silica-sensitive, carbonyl-iron-sensitive, and Thy1.2-, presumably Kupffer cells or resident liver macrophages. IFN alpha/beta production by cultured Kupffer cells was not observed if medium containing fetal bovine serum or portal serum was treated with polymyxin B or if Kupffer cells were cultured in serum-free medium. This suggested that small amounts of endotoxin in fetal bovine or portal serum stimulated Kupffer cells to produce IFN alpha/beta. Possibly, Kupffer cells are in a different state of activation/maturation than peritoneal and splenic macrophages since the sensitivity of resident Kupffer cells from C3HeB/FeJ mice to the stimulatory effects of endotoxin. The endogenous production of IFN alpha/beta by Kupffer cells from C3HeB/FeJ mice can augment liver-associated natural killer (NK) activity against YAC-1 cells (4h) and induce liver-associated cytotoxic activity, not restricted by the major histocompatibility complex, against NK resistant P815 mastocytoma cells (18 h).