Parthenolide, a sesquiterpene lactone, inhibited lipopolysaccharide (LPS) stimulated nuclear factor (NF)-kappabeta and cytokine production in vitro and in rats, and improved survival in LPS challenged Swiss albino mice. We investigated whether increased survival with parthenolide was associated directly with inhibition of NF-kappabeta and cytokines in LPS challenged C57BL/6J mice. In RAW 264.7 cells, parthenolide inhibited LPS-stimulated NF-kappabeta and cytokines (interleukin [IL]-1alpha, -1beta, -2, -4, -6, and -10, interferon-gamma, tumor necrosis factor-alpha, granulocyte macrophage-colony stimulating factor, migratory inhibitory protein-1 and -2alpha, JE, and RANTES). In mice (n = 366) receiving lethal intraperitoneal (i.p.) LPS (40 mg/kg), compared to placebo, each of 5 parthenolide doses (0.25 to 4 mg/kg i.p. following LPS) reduced survival at 168h and overall worsened the hazards ratio of survival (mean +/- S.E.M.) (1.29 +/- 0.12, p = 0.04). In other mice (241), compared to saline challenge, nonlethal LPS (2.5 mg/kg) increased NF-kappabeta in lung and kidney combined and 12 of 13 plasma cytokines early (1 and 3 h) and late (6, 9 and 12 h) (p < or = 0.002 for each). Compared to nonlethal LPS, lethal LPS increased NF-kappabeta and 12 of 13 cytokines early but not significantly and late significantly (p < or = 0.05 for each). With lethal LPS, compared to placebo, parthenolide (1 mg/kg) decreased NF-kappabeta and 10 of 13 cytokines early and increased NF-kappabeta and 11 of 13 cytokines late (p < or = 0.02 for early vs. late). Although parthenolide inhibits NF-kappabeta and cytokines in vitro, its effects on these mediators and survival in animal sepsis models vary. Theses differences must be understood before parthenolide or related agents are applied clinically for sepsis.