Passive dual immunization against tumour necrosis factor-alpha (TNF-alpha) and IL-1 beta maximally ameliorates acute aminonucleoside nephrosis.

Rats receiving a single dose (10 mg/100 g) of aminonucleoside of puromycin (PAN) develop heavy proteinuria and acute interstitial nephritis (AIN). Whole isolated glomeruli from rats injected with PAN secreted both TNF-alpha and IL-1 beta cytokines. TNF-alpha secretion was first and maximally detected on day 3, whereas IL-beta activity was found on day 7, when rats were heavily proteinuric and AIN developed. In vivo treatment with either anti-TNF-alpha or anti-IL-1 beta antibodies produced a drastic and simultaneous reduction in both levels of proteinuria and intensity of interstitial cell infiltrate. These effects improved when both antibodies were administered together. Our studies demonstrate the effectiveness of immunosuppressive therapy against these two cytokines in rats with PAN-induced nephrosis.     

Interaction of IL-1 beta, IL-6 and tumour necrosis factor-alpha (TNF-alpha) in human T cells activated by murine antigens.

We used a mixed leucocyte culture between human T cells and irradiated murine splenocytes which allowed us to distinguish between cytokine production from the responder and stimulator cells by the use of species-specific assays for mRNA up-regulation. Using this model of T cell activation by antigen, we studied the effects of human antigen-presenting cell-derived cytokines IL-1 beta, IL-6 and TNF-alpha on the activation of human T cell subsets. We show in this system that exogenously added IL-1 beta, IL-6 and TNF-alpha induces IL-2 receptor (R) up-regulation and IL-2 production, and proliferation by both CD4+ and CD8+ cells. The addition of IL-1 beta induces IL-6 mRNA, and anti-IL-1 antibodies or an IL-1R antagonist protein completely suppresses IL-6 and TNF-alpha supported proliferation. Similarly, addition of IL-6 or TNF-alpha induces up-regulation of IL-1 beta mRNA. However, anti-IL-6 and anti-IL-6R antibodies only partially block proliferation supported by IL-1 beta. These findings suggest that IL-6 and TNF-alpha will induce IL-2R up-regulation/IL-2 secretion via the induction of IL-1 beta production.     

Evidence for control of tumour necrosis factor-alpha (TNF-alpha) activity by TNF receptors in patients with proliferative diabetic retinopathy

TNF-alpha has been implicated in the pathogenesis of insulin- dependent diabetes mellitus (IDDM). At present there are no studies linking serum levels of soluble TNF receptors (sTNF-R) to the development of diabetic microvascular complications such as proliferative diabetic retinopathy (PDR), or to the production of TNF-alpha in these patients. We investigated serum levels of sTNF receptors (sTNF-RI and sTNF-RII) in IDDM patients with or without PDR, and related these to the in vitro production of TNF-alpha upon activation of whole blood and isolated mononuclear cells (MNC). We observed higher serum levels of sTNF-RI in IDDM patients with active (range 945-6630 pg/ml; P = 0.029) or quiescent PDR (range 1675-4970 pg/ml; P = 0.00092) than in individuals with IDDM without retinopathy (range 657-2617 pg/ml) or healthy controls (range 710-1819 pg/ml; P = 0.0092 and 0.0023, respectively). Increased serum levels of sTNF-RII were also seen in IDDM patients with active PDR (range 1749-5218 pg/ml; P = 0.034) or quiescent PDR (range 1494-5249 pg/ml; P = 0.0084) when compared with disease controls (range 1259-4210 pg/ml) or healthy subjects (range 1237-4283 pg/ml). Whole blood production of biologically active TNF-alpha was lower in PDR patients than in disease (P = 0.04) and healthy controls (P < 0.005), contrasting with a higher production of TNF-alpha by lipopolysaccharide (LPS)-activated MNC from PDR patients (P = 0.013). Inhibition of TNF-alpha by TNF-R in plasma supernatants of activated blood from PDR patients was demonstrated by increase of TNF-alpha activity in the presence of anti-TNF-RI and anti-TNF-RII antibodies. These observations suggest that abnormalities in TNF-alpha production and control may operate during the development of microvascular complications of diabetes mellitus.     

Tumor necrosis factor-alpha (TNF-alpha)-induced and interleukin-1 beta (IL-1 beta)-induced shedding of TNF receptors from gingival fibroblasts.

Tumor necrosis factor-alpha (TNF-alpha) exerts its functions by binding two different receptors (TNFR55 and TNFR75). Both TNFR55 and TNFR75 exist in cell-associated and soluble forms. Soluble TNF receptors (sTNFR), sTNFR55 and sTNFR75, are proteolytically shed upon inflammatory stimuli and then modulate various TNF-alpha bioactivities. As human gingival fibroblasts (HGF) can be potential targets for TNF-alpha in inflamed gingiva, we hypothesized that HGF partially modulate the cellular responses to TNF-alpha by regulating their own TNFR. In this study, the kinetics of expression of cell-associated and soluble forms of both receptors from cultured HGF in response to proinflammatory cytokines TNF-alpha and interleukin-1 beta (IL-1 beta) were investigated in vitro. Both TNF-alpha and IL-1 beta upregulated the gene expression of TNFR75 and did not affect that of TNFR55. TNF-alpha and IL-1 beta decreased binding of [(125)I]TNF-alpha to HGF. Moreover, TNF-alpha and IL-1 beta upregulated the release of sTNFR75 from HGF but not that of sTNFR55. These results suggest that HGF under inflammatory conditions may contribute to the inactivation of circulating TNF-alpha through the preferential induction and shedding of TNFR75.     

Differential effects of pentoxifylline on the production of tumour necrosis factor-alpha (TNF-alpha) and interleukin-6 (IL-6) by monocytes and T cells.

Pentoxifylline (PTX) is a methylxanthine compound known to inhibit the production of tumour necrosis factor-alpha (TNF-alpha) by monocytic cells. In this study, we found that PTX differentially regulates the production of TNF-alpha and interleukin-6 (IL-6). Indeed, PTX at high concentrations triggers the production of IL-6 but not of TNF-alpha by peripheral blood mononuclear cells (PBMC). Further experiments indicated that monocytes are responsible for this PTX-induced IL-6 production. When PBMC were stimulated with LPS, PTX was found to inhibit the secretion of TNF-alpha as well as the accumulation of TNF-alpha messenger RNA (mRNA). In contrast, no inhibitory effect was observed on the induction of IL-6. Similar results were obtained when PBMC were stimulated with OKT3 monoclonal antibody (mAb). In addition, the in vivo administration of PTX in transplant patients receiving the first dose of OKT3 allowed to decrease the systemic release of TNF-alpha but not of IL-6. Since monocytes represent a major source of TNF-alpha and IL-6 in these settings, additional experiments were performed in vitro on purified T cells stimulated with the CLB-T3/3, an anti-CD3 mAb which does not require the presence of accessory cells to activate T cells. In this system, PTX was found to inhibit the secretion of both TNF-alpha and IL-6 by T cells. We suggest that cAMP could be involved in these differential effects of PTX on production of TNF-alpha and of IL-6.     

Impaired tumour necrosis factor-alpha (TNF-alpha) production and abnormal B cell response to TNF-alpha in patients with systemic lupus erythematosus (SLE).

We examined the TNF-alpha activity in culture supernatants of monocytes isolated from the peripheral blood of patients with SLE and of normal individuals. The monocytes from patients with SLE stimulated with silica particles, lipopolysaccharide or Staphylococcus aureus Cowan 1 secreted significantly lower amounts of TNF-alpha than did normal monocytes. A decreased TNF mRNA expression was observed in peripheral blood mononuclear cells stimulated by mitogens from patients with SLE. Furthermore, we examined the effect of recombinant TNF-alpha (rTNF-alpha) on the B cell function in SLE patients. rTNF-alpha inhibited the spontaneous B cell proliferation of SLE, but tended to enhance the normal B cell proliferation. Spontaneous IgM production from SLE B cells was inhibited by rTNF-alpha, but that from normal B cells was not. Spontaneous IgG production was unaffected by rTNF-alpha. Also, rTNF-alpha did not affect the viability of B cells. These findings suggest that an impaired TNF-alpha production and an abnormal B cell response to TNF-alpha play a role in the immunological dysfunction in patients with SLE.     

Exposure to hyperbaric oxygen induces tumour necrosis factor-alpha (TNF-alpha) secretion from rat macrophages.

We investigated the secretion of TNF-alpha by monocytes and macrophages derived from the peripheral blood, spleen, and lungs after a single exposure to a therapeutic profile of hyperbaric oxygen (HBO). Rats were exposed for 90 min to either 100% oxygen at 0.28 MPa (2.8 atmospheres absolute) or air. Immediately after exposure, mononuclear cells were isolated from blood, spleen, and lungs and cultured for 18 h. The secretion of TNF-alpha from the cultured monocytes/macrophages was determined with and without stimulation with lipopolysaccharide (LPS). Exposure to hyperbaric oxygen induced a significant increase in the spontaneous ex vivo secretion of TNF-alpha (without LPS) by mononuclear cells from the blood, spleen, and lung (P < 0.05 from air controls). Stimulation with LPS after exposure to HBO induced a significant increase in TNF-alpha secretion by lung and spleen macrophages compared with air controls (P < 0.05). However, absolute TNF-alpha levels were not significantly higher than those achieved 'spontaneously' in macrophages exposed to HBO without LPS. Stimulation with LPS induced a marked increase in secretion of TNF-alpha from blood monocytes after exposure to air, but not after exposure to HBO. These results provide evidence in support of a role played by TNF-alpha in mediating HBO effects on different tissues and their immune responses.     

Alterations in mononuclear cell tumour necrosis factor-alpha (TNF-alpha) response in patients on long term cuprophane haemodialysis.

We have investigated TNF-alpha secretory response of peripheral blood mononuclear cells (PBMC) from 13 uraemic patients undergoing regular haemodialysis with cuprophane membrane (CM). Sixteen healthy subjects and five uraemic patients under conservative therapy were also studied as controls. Cells of haemodialysis patients exhibited increased TNF-alpha release in vitro in the absence of activating stimuli other than culture conditions, as compared with normal and uraemic controls. In contrast to normal cells, this spontaneous secretion of TNF-alpha from dialysis PBMC could not be significantly reduced by addition of polymyxin B to culture medium, thus indicating its independence of trace amount of lipopolysaccharide (LPS) present in the medium as contaminant. Furthermore, predialysis PBMC were considerably more sensitive to stimulation with 10(7) pg/ml of LPS under in vitro culture conditions than normal and uraemic controls. To elucidate a role of direct contact with CM in stimulation of TNF-alpha release from monocytes, PBMC were cultured on CM in vitro. Contact with CM stimulated TNF-alpha secretion from PBMC above the level of cells cultured on tissue culture plastic. This response persisted with time in culture in contrast to a transient LPS-induced TNF-alpha release. Furthermore, PBMC stimulated by contact with CM for 2 days did not lose the capacity to secrete TNF-alpha in response to a subsequent LPS stimulation, while a 2-day treatment of cells with LPS was followed by LPS refractory state. Therefore, direct contact with CM induces in PBMC a long-lasting TNF-alpha response which is not down-regulated by the acquisition of refractoriness in a manner similar to that which occurs in the case of LPS stimulation. These in vitro findings provide a possible explanation of the observation that predialysis PBMC exhibit elevated TNF-alpha secretory capacity.     

Occlusion of pulmonary vessels by megakaryocytes after treatment with tumour necrosis factor-alpha (TNF-alpha)

Pulmonary thrombosis in the course of shock remains life-threatening, despite advances in diagnosis, prophylaxis and therapy of the disease. Tumour necrosis factor-alpha (TNF-alpha) is an important mediator of shock. The aim of this study was to analyze the morphological changes in the pulmonary capillary bed in rats after intraperitoneal administration of multiple doses of TNF-alpha (10 microg TNF-alpha/24 h for 5 days; biological activity of 2-4x10(7)U/mg of protein). Morphological investigations were undertaken by light and transmission electron microscopy with emphasis on pulmonary thrombopoiesis. The study confirmed that the lungs may be an important site of extramedullary thrombopoiesis in the course of shock. The observations also suggested that megakaryocytes shed large fragments of cytoplasm within the pulmonary capillary bed and that megakaryocytes with copious cytoplasm occlude pulmonary vessels.     

Beneficial effects of tumour necrosis factor-alpha (TNF-alpha) blockade in rheumatoid arthritis (RA)

The biological properties of TNF-alpha make it a candidate therapeutic target in RA. Our studies have demonstrated that TNF-alpha and its receptors are up-regulated and co-expressed in the synovium and cartilage-pannus junction of RA joints. Neutralizing TNF-alpha antibodies reduce the production of the many pro-inflammatory cytokines, including IL-1 and granulocyte-macrophage colony-stimulating factor (GM-CSF), produced by mononuclear cells from RA in culture. When injected into DBA/1 mice with collagen-induced arthritis and TNF-alpha transgenic mice with arthritis, anti-TNF MoAbs decrease inflammatory damage of joints. Clinical trials employing cA2, a chimaeric anti-TNF-alpha MoAb, in open-label and randomized placebo-controlled studies have demonstrated a dose-dependent efficacy with impressive improvement in disease activity and acute-phase responses lasting several weeks. We conclude that TNF-alpha is a critical mediator of inflammation in RA, and is an important therapeutic target in this disease.     

Polymorphism of tumour necrosis factor-alpha (TNF-alpha) at position -308 in relation to ankylosing spondylitis.

In addition to HLA-B27, other genetic factors are thought to be involved in the pathogenesis of ankylosing spondylitis (AS). Because of the localization, in the proximity of the HLA-B locus, and the biological activities of TNF-alpha, we investigated the association between AS and a single base polymorphism located at position -308 of the TNF-alpha gene. An allele-specific polymerase chain reaction was developed to monitor this polymorphism. The frequency of the TNF-alpha alleles was determined in 66 AS patients and 37 healthy controls. The TNF-alpha allele frequency was not significantly different between AS patients and controls.     

Human tumour necrosis factor-alpha (TNF-alpha) directly stimulates arachidonic acid release in human neutrophils.

The ability of tumour necrosis factor-alpha (TNF-alpha) to directly stimulate phospholipid turnover from human neutrophils was studied. Stimulation with recombinant human (rH) TNF-alpha induced the release of significant amounts of radioactivity from [3H]arachidonic acid-labelled neutrophils. This stimulation was equipotent to that induced by the bacterial tripeptide formyl-methionyl-leucylphenylalanine (FMLP). The time of maximum stimulated release varied between donors, with the most common maximal stimulation being 45 min. Dose-response experiments indicated that 100-1000 U/ml rH TNF-alpha were required for the maximum stimulatory effect. High-performance liquid chromatography analysis of the supernatants revealed that the radioactivity was associated with arachidonic acid, but not with its metabolites, indicating that TNF-alpha stimulates the release of arachidonic acid from cellular phospholipids but does not stimulate its metabolism. A comparison of TNF-alpha with other cytokines indicated that stimulation of arachidonic acid release paralleled the 'priming' of neutrophils for enhanced superoxide production, raising the possibility that phospholipid turnover and priming of neutrophils are causally related.     

Platelet expression of tumour necrosis factor-alpha (TNF-alpha), TNF receptors and intercellular adhesion molecule-1 (ICAM-1) in patients with proliferative diabetic retinopathy.

Microvascular complications of insulin-dependent diabetes mellitus (IDDM) have been strongly associated with platelet abnormalities, whilst TNF-alpha has been implicated in the pathogenesis of this condition. However, at present it is not clear whether human circulating platelets express TNF-alpha or TNF receptors (TNF-R) or whether impaired expression of these molecules and of the TNF-reactive adhesion molecule ICAM-1 may be associated with platelet abnormalities in patients with IDDM. On this basis we investigated the platelet expression of these molecules in patients with IDDM complicated or uncomplicated by proliferative diabetic retinopathy (PDR) and in healthy subjects. We observed that the proportion of platelets staining for TNF-alpha was significantly higher in IDDM patients with active PDR than in patients without microvascular complications (P = 0.0078), quiescent PDR (P = 0.003) or healthy subjects (P = 0.0013). Patients with active PDR also showed a higher proportion of platelets expressing TNF-RI (P = 0. 0052) and TNF-RII (P = 0.015) than healthy controls or patients with quiescent PDR (P = 0.009 and 0.0006, respectively). In addition, the percentage of ICAM-1+ platelets was significantly higher in patients with active PDR than in patients with quiescent PDR (P = 0.0065) or normal subjects (P = 0.013). There was a direct correlation between platelet expression of TNF-alpha and that of TNF-R in PDR patients, indicating that platelet staining for TNF-alpha may be due to binding of this cytokine to its receptors. The results suggest that increased platelet expression of TNF-alpha, TNF-R and ICAM-1 in IDDM patients may constitute important markers of thrombocyte abnormalities during the development of microvascular complications of diabetes mellitus.     

Passive immunization against tumour necrosis factor-alpha (TNF-alpha) and IL-1 beta protects from LPS enhancing glomerular injury in nephrotoxic nephritis in rats.

Glomerular injury caused by injection of heterologous anti-glomerular basement membrane antibodies (anti-GBM Ab) is increased in rats pretreated with small doses of bacterial lipopolysaccharide (LPS). We have investigated the involvement of tumour necrosis factor-alpha (TNF-alpha), IL-1 alpha and IL-1 beta in this phenomenon by passive immunization against these cytokines. Anti-TNF-alpha or anti-IL-1 beta antibodies given 1.5 h before the induction of nephritis significantly decreased injury in this model, whether assessed by the magnitude of albuminuria, the prevalence of glomerular capillary thrombi or the intensity of glomerular neutrophil infiltrate. Albuminuria in anti-GBM Ab alone was 11 +/- 3, LPS/anti-GBM Ab 87 +/- 22, and anti-TNF-alpha antibodies/LPS/anti-GBM Ab 21 +/- 6 mg/24 h (mean +/- s.e.) P < 0.05. Passive immunization with antibodies to IL-1 beta had a similar effect (anti-GBM Ab, 0.6 +/- 0.1, LPS/anti-GBM Ab, 92 +/- 19, anti-IL-1 beta antibodies/LPS/anti-GBM Ab 39 +/- 8 mg/24 h, P < 0.05). The prevalence of glomerular capillary thrombi was also reduced significantly by these treatments; from 22 +/- 5% to 4 +/- 1% in the case of anti-TNF-alpha antibodies and 28 +/- 5% to 13 +/- 4% with anti-IL-1 beta antibodies. Similarly, the glomerular neutrophil infiltrate was also reduced by these treatments; from 42 +/- 3 to 25 +/- 1 in the case of anti-TNF-alpha and 47 +/- 2 to 30 +/- 1 with anti-IL-1 beta antibodies. In contrast, passive immunization against IL-1 alpha had no effect on either albumin excretion (4 +/- 3, 83 +/- 22 and 77 +/- 24 mg/24 h), glomerular capillary thrombi (2 +/- 1; 19 +/- 5 and 16 +/- 3) or glomerular neutrophil infiltrate (22 +/- 3; 47 +/- 5 and 48 +/- 5 from the three groups respectively). These results demonstrate that enhanced antibody mediated injury in the kidney is modulated by TNF-alpha and IL-1 beta but not by IL-1 alpha.     

Endothelial cell activation by tumour necrosis factor-alpha (TNF-alpha) and the development of pre-eclampsia.

Pre-eclampsia may develop as a result of an endothelial activation. Tumour necrosis factor-alpha (TNF-alpha) activates endothelial cells which release soluble E-selectin, a putative circulating marker specific for endothelial damage. A retrospective longitudinal study of maternal blood samples, collected at different gestational ages in pregnancy, was undertaken to determine whether the development of pre-eclampsia is associated with TNF-alpha-mediated endothelial activation. This study included 19 women who developed pre-eclampsia and 22 women whose pregnancy outcome was normal. Ten women had blood samples taken before pre-eclampsia was clinically detected and, in all these, TNF-alpha was below the immunoassay limit of detection (< 80 pg/ml). Five had further samples taken after pre-eclampsia was clinically diagnosed and, initially, TNF-alpha was still below the lower limit of detection in all five pregnancies, but rose later in three (80, 156 and 250 pg/ml). In nine other patients with diagnosed pre-eclampsia, TNF-alpha was detected in only two (80 and 650 pg/ml). TNF-alpha was identified in only one of the 22 normal pregnancies (80 pg/ml), this being at term. There was no statistical difference in soluble E-selectin levels between normal and pre-eclamptic pregnancies, neither before nor after pre-eclampsia was diagnosed. Hence, blood TNF-alpha levels measured by immunoassay can be elevated in approximately 36% of cases of established pre-eclampsia, but this rise occurs only after the syndrome is detected clinically. Blood concentrations of TNF-alpha and soluble E-selectin are not related to severity of the disorder. These findings suggest that circulating TNF-alpha does not contribute to the initiation of endothelial cell activation that may be associated with the development of pre-eclampsia, but may rise as a consequence of the pathological processes of this disorder.     

Effects of granulocyte-macrophage colony-stimulating factor (GM-CSF), IL-2, interferon-gamma (IFN-gamma), tumour necrosis factor-alpha (TNF-alpha) and IL-6 on the production of immunoreactive IL-1 and TNF-alpha by human monocytes.

The effects of GM-CSF, IL-2, IFN-gamma, TNF-alpha and IL-6 on the production of IL-1 (both secreted and cell associated) and TNF-alpha by peripheral blood monocytes were studied. Monocytes were cultured for 20 h in suspension and in serum-free conditions which minimized background stimulation of monokine production. GM-CSF, IL-2 and TNF-alpha directly induced the production of cell-associated IL-1 but little or no IL-1 or TNF-alpha secretion. Combination of GM-CSF with IFN-gamma, IL-2 or TNF-alpha synergistically enhanced IL-1 secretion and had an additive effect on cell-associated IL-1 production. Combination of IL-2 with IFN-gamma or TNF-alpha also synergistically enhanced IL-1 secretion but the effect on cell-associated IL-1 production was less than additive. GM-CSF synergistically enhanced TNF-alpha secretion induced by IFN-gamma but not by lipopolysaccharide. GM-CSF did not enhance TNF-alpha secretion induced by IL-2 or TNF-alpha. In contrast, IL-2 synergistically enhanced TNF-alpha secretion induced by IFN-gamma. These results are discussed in relation to cytokine involvement in rheumatoid arthritis.     

Multiple sclerosis is associated with an imbalance between tumour necrosis factor-alpha (TNF-alpha)- and IL-10-secreting blood cells that is corrected by interferon-beta (IFN-beta) treatment

The up-regulated B cell responses detectable in cerebrospinal fluid (CSF) and the augmented myelin antigen-specific T cell responses observed in the CSF as well as systematically in patients with multiple sclerosis (MS) suggest the involvement of cytokines in disease development and perpetuation. Here we report on the parallel involvement of TNF-alpha, IL-6, IFN-gamma and IL-10 in MS and controls, using enzyme-linked immunospot (ELISPOT) assays to detect and enumerate cytokine-secreting mononuclear cells (MNC) prepared from blood and, for IL-6 and IL-10, from CSF without in vitro stimulation. MS is associated with elevated levels of TNF-alpha-secreting blood MNC when compared with levels in groups of control patients with myasthenia gravis (MG) and other neurological diseases (OND) or healthy subjects. This elevation was confined to patients with untreated MS and not present in those examined during ongoing treatment with IFN-beta. Untreated patients with MS had lower numbers of IL-10-secreting blood MNC compared with the three control groups. In patients undergoing treatment with IFN-beta, numbers of IL-10-secreting cells were in the same range as in controls. Normalization of TNF-alpha from elevated, and of IL-10 from decreased levels could be one reason for the beneficial effects of IFN-beta in MS, although it remains to be shown whether these changes reflect phenomena primarily involved in MS pathogenesis or secondary changes. In CSF, levels of IL-10-secreting cells were higher than in blood in both MS and OND, with no difference between these groups. Systemic aberrations of IL-6 and IFN-gamma and of IL-6 in CSF in MS versus controls were only minor, irrespective of treatment with IFN-beta.     

Cytokine regulation of human monocyte interleukin-1 (IL-1) production in vitro. Enhancement of IL-1 production by interferon (IFN) gamma, tumour necrosis factor-alpha, IL-2 and IL-1, and inhibition by IFN-alpha.

IL-1 production (secreted and cell-associated) was measured in monocyte cultures stimulated by a variety of agents in vitro. Monocytes either adherent to conventional plastic culture plates in serum-free conditions, or in suspension in culture medium containing serum were stimulated to produce IL-1 during culture. In non-adherent, serum-free conditions, monocytes produced very low or undetectable amounts of IL-1 during 20 h of culture. Lipopolysaccharide (LPS) induced equivalent amounts of secreted and cell-associated IL-1, although at very low concentrations more cell-associated IL-1 was produced. IL-1 production in response to LPS could be augmented by crude lymphokine, IFN-gamma, or tumour necrosis factor (TNF) alpha. TNF-alpha preferentially augmented the production of cell-associated IL-1 in LPS-stimulated cultures. TNF-alpha induced a significant amount of IL-1 (mainly cell-associated) directly but could also induce IL-1 secretion when combined with IL-2 or IFN-gamma, or when in the presence of serum. IL-2 acted synergistically with low concentrations of IFN-gamma or IL-1 to induce significant levels of IL-1 production. IFN-alpha did not induce any IL-1 production, but was a potent inhibitor of IL-1 production induced by a variety of stimuli. These results suggest that IL-1 production may be enhanced or reduced by different cytokines at concentrations likely to be found in chronic inflammatory lesions.