W e have examined the ability of a highly purified 38-kD phospholipase-inhibitory protein (p38) isolated from human placental membranes that is also a preferred substrate for the epidermal growth factor-urogastrone (EGF-URO) receptor/kinase, to block the release of arachidonate from zymosan-stimulated murine peritoneal macrophages in vitro and to exhibit antiinflammatory activity in a carrageenin rat paw edema test in vivo. The ability of glucocorticoids to increase the amounts of this protein in macrophage cultures was also examined. p38 represents the naturally occurring, intact, NH2-terminally blocked human placental form of the protein termed calpactin II (or lipocortin I), for which partial amino acid sequence data and a complete amino acid sequence deduced from cDNA analysis have been reported. Our data demonstrated that, whereas p38 was an effective inhibitor of pancreatic phospholipase A2 in vitro, it was unable to inhibit either the release of arachidonate from cultured zymosan-stimulated mouse peritoneal macrophages or inflammation in a rat paw edema test. At comparatively high protein concentrations, p38 enhanced either arachidonate release from intact macrophages in vitro (0.5-10 micrograms/ml) or carrageenin-induced paw swelling in vivo (2.5 or 25 micrograms per injection). Furthermore, we were unable to detect induced amounts of p38 in cultures of glucocorticoid-treated peritoneal macrophages obtained from either mice or rats. Our data indicate that the antiphospholipase activity of p38 in vitro and the ability of p38 to serve as a receptor/kinase substrate may in no way relate to the putative ability of the protein to modify eicosanoid release from macrophages in vivo, so as to modulate the inflammatory process. Our data also raise the possibility that p38 (calpactin II) may not be a true representative of the lipocortin family of glucocorticoid-inducible antiinflammatory proteins, despite its ability to inhibit phospholipase A2 in vitro.