Search

An important role of TNF interacting with TNFR2 has been shown in different models of ischemic, nephrotoxic and immune-mediated renal injury. To systematically evaluate the expression of TNFR2 in renal allograft rejection, we investigated human renal allograft biopsies and, in addition, established an experimental transplantation model in rats to verify the human data under standardized conditions. The expression of TNFR2 was analyzed in 96 human renal allograft biopsies with different disease entities. In a 6-day and a 28-day experimental protocol, TNFR2 was examined in kidney specimens and in the urine of control, uni-nephrectomized and transplanted rats ± cyclosporine treatment (n = 114). In human biopsies and in rat allografts on day 6 with acute allograft rejection, significantly elevated expression of TNFR2 was observed in tubular epithelial cells, podocytes, B cells and monocytes/macrophages. The expression level was associated with renal function. The TNFR2 expression level at day 28 was significantly lower compared to day 6. TNFR2 is markedly upregulated both in human and experimental acute renal allograft rejection. Our data are robust and consistent between different species, suggesting a role for TNFR2 in the early course of rejection.

Several lines of evidence point to an activation mechanism of monocytes/macrophages by tumor cells. In this study we present data for distinct surface structures on K562 and Jurkat cells to directly induce TNF-mRNA expression and TNF production by human peripheral blood monocytes. Northern analysis showed that incubation of monocytes with either K562 or Jurkat cells led to a significant increase in TNF-mRNA expression. In addition, enhanced TNF production was detected in supernatants of monocyte cultures activated by Jurkat cells. Not only viable tumor cells but also metabolically inactivated tumor cells, cytoblasts, and membrane preparations from Jurkat and K562 cells induced TNF-mRNA expression. We identified two different membrane protein fractions with relative molecular mass of 32 to 38 kDa for Jurkat cells and 46 to 54 kDa for K562 cells that were responsible for monocyte activation.

The mechanisms involved in the antimetastatic effect of CpG-containing DNA were investigated in a mouse model of experimental metastasis. Tumor cell colony formation in lungs or livers of mice after i.v. inoculation with syngeneic fibrosarcoma or thymoma cells was determined. The i.v. injection of plasmid DNA or synthetic oligodeoxynucleotides (ODNs) containing unmethylated CpG motifs before tumor cell application strongly inhibited metastasis. Because synthetic CpG-ODN was not directly tumor cytotoxic, the target cells for this CpG-ODN effect were determined. The cytotoxic activity on standard natural killer (NK) targets as well as on fibrosarcoma cells of splenic NK cells and NKT cell-containing liver mononuclear cells derived from CpG-ODN-treated mice was strongly enhanced. Participation of NK/NKT cells in the CpG-induced antimetastatic effect was demonstrated by reduction of the antimetastatic effect in mice depleted of NK/NKT cells and beta2-microglobulin-deficient mice. Neutralization of interleukin 12, interleukin 18, or IFN-gamma did not interfere with the CpG-induced antimetastatic effect. However, in sera of CpG-ODN-treated mice, high levels of IFN-alpha were detected, and in IFN-alpha/beta receptor-deficient mice, the CpG-ODN-induced antimetastatic effect was strongly reduced. These data indicate that CpG-ODNs activate NK/NKT cells for antimetastatic activity indirectly via IFN-alpha/beta receptor activation. The exploitation of the stimulatory activity of CpG-ODN for the innate immune system might be a useful strategy for antimetastatic therapy.

The lymphotoxin beta-receptor (LT beta R), a member of the tumor necrosis factor (TNF) receptor family, plays a crucial role in lymphoid organogenesis by signaling through its functional ligand LT alpha(1)beta(2). While the receptor is expressed on a wide range of cell types e.g. fibroblasts and monocytes, the ligand is expressed only on activated T, B and NK cells. Remarkably, no cell type has been identified so far that expresses both the receptor and the ligand. In order to characterize the mouse LT beta R gene expression on a molecular level, we isolated about 1 kb of the 5' flanking region of the LT beta R gene. Primer extension analysis revealed one transcriptional start site located at - 60 upstream of the ATG-containing first exon. Northern blot analysis showed that the LT beta R is abundantly expressed in the mouse fibroblast cell line NIH 3T3, and to a lesser extent, in the mouse macrophage-like cell line RAW 264.7. To determine whether the 5' flanking region exerts functional promoter activity, we generated deletion mutants fused to the luciferase reporter gene. Transfection experiments using these reporter gene constructs showed that the isolated 5' flanking region is transcriptionally active in NIH 3T3 and RAW 264.7 cells, and determined a minimum length required for the transcriptional activity of the LT beta R promoter in these cells. Further sequence analysis of the isolated 5' flanking region identified a number of putative DNA-binding sites for transcription factors. Interestingly, incubation of NIH 3T3 cells with dexamethasone resulted in an elevated mRNA level of the LT beta R gene. This effect was abolished by using the specific glucocorticoid receptor inhibitor RU486, indicating an increased transcriptional activity of the LT beta R promoter after glucocorticoid stimulation.

We investigated the role of the TNF receptors, type I (p55TNFR) and type II (p75TNFR), in a mouse model of contact hypersensitivity, i.e., a model of a delayed type hypersensitivity (DTH) allergic reaction. Mice deficient for p55TNFR or p75TNFR were used to investigate the functions of these receptors in development of the DTH reaction. We show that both TNF receptors have a strong influence on the overall outcome of the DTH reaction, with the two TNF receptors exerting distinct functions. Dendritic cells of mice lacking p55TNFR had a defect in allergen uptake but showed normal migration into regional lymph nodes. In contrast, dendritic cells of p75TNFR-deficient mice showed diminished migration into regional lymph nodes after allergen contact, whereas the allergen uptake was independent of the p75TNFR. Thus, both TNF receptors are required for the development of a complete DTH reaction.

It is believed that an inflammation-induced activation of the CNS leads to an inhibition of overshooting immune responses to prevent extensive local cytokine secretion. However, immunosuppression by the sympathetic nervous system may be unfavorable when bacteria are present locally and when TNF-alpha is necessary to overcome infection. We now report in a superfusion model, using mouse spleen slices, that although local Pseudomonas aeruginosa increased splenic TNF-alpha and IL-6 secretion severalfold over basal levels, electrically released neurotransmitters attenuated cytokine secretion to similar basal level as under bacteria-free conditions. Bacteria reversed noradrenergic inhibitory effector mechanisms: Under bacteria-free conditions, TNF-alpha secretion was very low and IL-6 secretion was mainly inhibited by alpha2-adrenoreceptor ligation. In the presence of bacteria, TNF-alpha and IL-6 secretion were high and IL-6 secretion was mainly inhibited by beta-adrenoreceptor ligation. The alpha- to beta-adrenoswitch of IL-6 inhibition in the presence of bacteria was mediated by the prior adrenergic regulation of TNF-alpha. In vivo, chemical abrogation of sympathetic inhibition reduced accumulation of bacteria in the spleen, which depended at least in part on TNF-alpha. This suggests that activation of the sympathetic nervous system may be a forerunner for accumulation of bacteria in tissue and consecutively sepsis due to intensified inhibition of TNF-alpha secretion.

Mechanisms involved in the antimetastatic effect of IL-12 were analyzed in a mouse model of experimental metastasis with either syngeneic fibrosarcoma cells colonizing the lungs or syngeneic B cell lymphoma cells colonizing the liver. IL-12 pretreatment effectively reduced the number of tumor colonies in both systems. This effect was already manifest 24 hours after tumor cell injection, indicating a T and B cell-independent mechanism. Therefore, the involvement of NK and alphabetaNKT cells was investigated using mice with defective NK and alphabetaNKT cell functions. Mice with impaired NK functions due to NK cell depletion, were less responsive to the antimetastatic IL-12 effect. IL-12 treatment failed to inhibit metastasis in beta2-microglobulin-deficient mice which lack alphabetaNKT cells in addition to having impaired NK cell activity, thus, demonstrating the functional importance of IL-12-activated NK and alphabetaNKT cells. While the IL-12-induced antimetastatic effect of NK cells was dependent on IFN-gamma action, IL-12 activation of alphabetaNKT cells did not involve IFN-gamma. The neutralization of IFN-gamma or the use of IFN-gamma receptor-deficient mice did not alter the IL-12-induced effect in the absence of NK cells. Activation of effector cells of the innate immune system, such as NK and alphabetaNKT cells, seems to be the main mechanism for the antimetastatic effect of IL-12.

Shock is frequently accompanied by thrombocytopenia. To investigate the pathogenic role of platelets in shock, we examined the in vivo effects of monoclonal antibodies (MoAbs) against mouse platelet membrane proteins. Injection of the platelet-specific MoAb MWReg30 to the fibrinogen receptor (gpIIb/IIIa) rendered mice severely hypothermic within minutes. Isotype-matched control antibodies, even if they also recognized platelet surface antigens, did not induce comparable signs. MWReg30 induced early signs of acute lung injury with increased cellularity in the lung interstitium and rapid engorgement of alveolar septal vessels. Despite this in vivo activity, MWReg30 inhibited rather than stimulated platelet aggregation in vitro. MWReg30-binding to platelets led to phosphorylation of gpIIIa, but did not induce morphological signs of platelet activation. The MWReg30-induced reaction was abolished after treatment with MoAbs 2.4G2 to FcgammaRII/III and was absent in FcgammaRIII-deficient mice, clearly demonstrating the requirement for FcgammaRIII on involved leukocytes. Simultaneous administration of tumor necrosis factor exacerbated, whereas a tolerizing regimen of tumor necrosis factor or bacterial lipopolysaccharide completely prevented the reaction. These data suggest that platelet surface-deposited MWReg30-immune complexes lead to an acute Fc-mediated reaction with pulmonary congestion and life-threatening potential that could serve as an in vivo model of acute lung injury.

Natural killer (NK) cells provide effective antitumoral activity in the blood stream of mice, leading to reduced metastasis. There are, however, tumor cells that metastasize despite the presence of an intact NK system. The capability of tumor cells to induce platelet aggregation, on the other hand, correlates with their enhanced metastatic potential. A counteractive role of platelets for the NK function in metastasis has never been conceived. Here we demonstrate for the first time that platelets directly protect tumor cells from NK lysis in vitro as well as in vivo. Using three different tumor cell lines in a mouse model of experimental metastasis, tumor seeding in the target organs was reduced when the host was platelet depleted, but only if the tumor cells were NK sensitive. Aggregation of platelets around tumor cells also inhibited in vitro NK tumorilytic activity. This protection of tumor cells by platelets was mouse strain independent and was equally observed with platelets from beta2-microglobulin-deficient mice, excluding a NK inhibitory function of MHC class I on platelets. Thus, even if tumor cells are NK susceptible and cytotoxic NK cells threaten their survival in the blood, platelets are capable of protecting them from cytolysis, thereby promoting metastasis. Surface shielding by platelet aggregates seems to be the main mechanism of this protection.

Monocytes and macrophages can distinguish between tumour cells and non-malignant cells of the same cell type. The surface structures mediating tumour cell recognition and tumour cell-induced cytokine production by monocytes or macrophages are still poorly characterized. The authors previously described N-linked sialic acid-containing carbohydrates associated with CD2 of the CD4+ tumour cell line Jurkat which induced tumour necrosis factor (TNF) secretion by human monocytes. In this report the authors demonstrate a role for monocytic L-selectin in this process. Mannose-6-phosphate, an inhibitor of L-selectin binding, blocked CD2-induced TNF production whereas other ligands for L-selectin (the monomeric monoclonal antibody to L-selectin, DREG-200; inositol hexaphosphoric acid; heparin) amplified CD2-induced secretion of TNF. Polyvalent L-selectin ligands as soluble monoclonal antibody (MoAb) DREG-56, the cross-linked MoAb DREG-200, and the marine algal polysaccharide fucoidin induced TNF production without addition of CD2. Jurkat-CD2 is associated with sialyl Lewis X-related carbohydrates which differ in their expression or composition from that of the non-stimulatory CD2 from non-malignant T cells. Taken together the data presented in this report suggest that-at least in the case of Jurkat-CD2-L-selectin is involved in monocyte activation by tumour typical carbohydrate structures.

Inflammatory mediators such as tumor necrosis factor (TNF) and interleukin-1 enhance tumor colony formation in different models of experimental and spontaneous metastasis. The involvement of the natural killer (NK) cell system in this process was investigated. Tumor necrosis factor does not appear to act directly on tumor cells by reducing their susceptibility to the cytotoxic action of NK cells but rather impairs NK activity in tumor-bearing mice. Such impairment of the natural killer system might be one means by which TNF supports tumor colony formation. Even though the metastasis- enhancing effect of TNF remained detectable in mice which have a greatly reduced NK cell cytotoxic activity due to a defect in the bg locus, normal mice which were depleted of NK cells by antibody treatment did not show enhanced metastasis after TNF injection. Therefore, the TNF-enhanced metastasis can only be seen as long as some NK cell function is operating in the animals.

Requirement for endogenous TNF for survival of experimental septic peritonitis has been demonstrated in a mouse model of cecal ligation and puncture (CLP). Induction of endogenous TNF production before CLP or administration of TNF before or after CLP confered protection from death. Interaction of TNF with the p55TNF receptor, formation of fibrin deposits, and granulocyte function was necessary to survive CLP. The mast cell seems to be an important cell type to provide the TNF required for protection in this model.

Mechanisms for TNF-enhanced adhesion of tumor cells to endothelial cells were investigated for their in vivo relevance in a model of experimental metastasis. Mouse fibrosarcoma and thymoma cells were used to analyze TNF-modified adherence to three different mouse endothelioma cell lines and the results were compared to the in vivo colonization behavior of the tumor cells. TNF enhanced tumor cell adhesion in vitro and extravasation in vivo with similar characteristics. The role of different adhesion molecules in these experimental systems was tested. Blocking of ICAM-1, LFA-1, VCAM-1, E-selectin, and P-selectin did not reduce TNF-enhanced metastasis even though tumor cell adhesion in vitro was reduced. However, the correlation between inhibition of integrin binding and inhibition of metastasis achieved with competing peptides indicated an important role for extracellular matrix components in tumor cell attachment. Platelets play a dual role: although in vitro platelets prevented tumor cell adhesion to endothelial cells, in vivo platelet-depletion of mice reduced metastasis.

Monocytes/macrophages can kill tumour cells and mediate tumour-destructive host responses e.g. by releasing tumour necrosis factor-alpha (TNF-alpha). The underlying mechanisms of tumour cell recognition, however, are not clear. Previous work in our laboratory suggested that carbohydrate moieties associated with the T cell adhesion molecule CD2 of Jurkat cells induce TNF-alpha secretion by human monocytes. In this study we present data indicating that the stimulatory capacity for TNF-alpha secretion is confined to carbohydrate moieties of tumour cell CD2. Irradiated resting peripheral T cells did not display stimulatory activity in contrast to irradiated Jurkat cells although surface expression of CD2 was similar. Activated T cells, however, induced TNF-alpha production by monocytes via a CD2-independent mechanism. Only affinity purified CD2 prepared from Jurkat cells but not from non-transformed T cells activated monocytes to secrete TNF-alpha. This activation process was blocked by anti-CD2 antibodies. Neuraminidase and PNGaseF treatment of isolated CD2 inhibited the stimulatory capacity whereas pronase treatment did not. These data suggest that carbohydrate moieties containing sialic acid mediate stimulation of monocytes. Taken together, these results indicate a role for glycosylation patterns typical of tumour cells in the recognition process of tumour cells by monocytes/macrophages.

The mechanism of action of the pleiotropic cytokine interleukin-1 (IL-1) is only incompletely understood. A unique feature among cytokines is its internalization and translocation to the nuclear area in nondegraded form, suggesting intracellular activities of the molecule. To define activities of that kind, a pair of IL-1 receptor type I (IL-1RI)-positive EL4 thymoma cells with differential receptor functionality was transfected with plasmids which caused intracellular expression of FLAGIL-1 alpha fusion peptides. Intracellular delivery of IL-1 costimulated expression of IL-2 mRNA and production of IL-2 protein. This effect was not mediated by the plasma membrane IL-1RI. The cells were permanently activated, and in cells with functional IL-1RI, appearance of membrane IL-1RI was abrogated. Thus, intracellularly delivered IL-1 can bypass and replace the plasma membrane IL-1RI, possibly via an as yet undefined intracellular receptor. This is a new modality of IL-1 action and suggests a role for the intracellular IL-1R antagonists (icIL-IRa).

The enhancement of tumor metastasis by concurrent inflammatory processes is mainly due to the cytokines TNF and IL-1. In the case of TNF this effect is not restricted to metastasis models as measured by in vivo colony formation but also found in experimental model systems of spontaneous metastasis. Direct effects on the tumor cells or interference with the host NK cell system did not seem to account for the observed TNF effect. Experimental evidence from different test systems rather points to TNF- or IL-1-induced enhanced adhesion of tumor cells to the endothelial cell layer as the underlying mechanism. Blocking of integrin-matrix interactions with monoclonal antibodies or competing peptides inhibited tumor cell adhesion to endothelioma cells in vitro and lung colony formation of tumor cells in vivo.

By intrasplenic immunization we raised a rat mAb (mAb V1q; IgG2a, kappa) with a potent neutralizing activity against natural mouse TNF (1 microgram/ml mAb V1q/100 U/ml TNF). mAb V1q was used to study the role of endogenous TNF in experimental peritonitis induced by sublethal cecal ligation and puncture. mAb V1q persisted for over 5 days in the serum of mice injected with 100 micrograms of the antibody and, therefore, proved useful for in vivo experiments. As little as 20 micrograms mAb V1q/mouse prevented lethal shock of the animals by 400 micrograms LPS/mouse. In sublethal cecal ligation and puncture i.p. injection of mAb V1q directly and up to 8 h after induction of experimental peritonitis lead to death of the animals within 1 to 3 days. The lethal effect of mAb V1q was compensated by injection of recombinant mouse TNF. Similar mAb V1q effects as in immunocompetent mice were shown in severe combined immune deficiency mice deficient of mature functional B and T cells. Taken together, these data suggest that during the early phase of peritonitis endogenous TNF may stimulate nonlymphoid cells such as granulocytes, macrophages, platelets, and fibroblasts to ingest bacteria and to localize inflammation, respectively. These beneficial effects of TNF may determine survival. Thus, our data may have implications for the therapeutic management of a beginning peritonitis.

Deoxycoformycin (DCF) has been reported to cause immediate reduction and dysfunction of T lymphocytes, but the long-term effects on immune functions are still not known. As cytokine production is regulated by T helper-inducer lymphocytes and might represent a parameter for functional integrity of immunocompetent cells, we have measured the production of interleukin-2 (IL-2), tumor necrosis factor (TNF), and interferons (IFN) by peripheral mononuclear cells (PMNC) from 10 patients with hairy cell leukemia 11-24 months after end of therapy with DCF. The patients were in continuous remission at the time of study. Despite an absolute reduction in CD3+ and CD4+ lymphocytes. there were no significant differences in IL-2 or TNF release between patients and controls. Except for a significant reduction in IFN-alpha release stimulated by Newcastle disease virus (NDV), IFN productions induced by other mitogens (phytohemagglutinin, PHA; Concanavalin A, ConA; pokeweed mitogen, PWM) and viral antigens were within normal range. There was also a decrease in proliferative responsiveness to PHA, but responses to ConA, PWM, and other viral antigens were normal. In five of the patients, we have monitored closely the changes in IL-2, TNF, and IFN before, during, and after treatment and could demonstrate a rapid normalization of initially decreased IL-2 release in all cases and also of TNF if the initial production was reduced. This study shows that, even though the absolute number of T lymphocytes and helper cells are reduced in the long-term observation after DCF treatment, the capacity to produce IL-2, TNF, and IFN-gamma was within normal range. Parallel to this observation, no opportunistic infections or frequency of infectious complications occurred in these patients.

Several lines of evidence point to an activation mechanism of monocytes/macrophages by tumor cells. In this study we present data for distinct surface structures on K562 and Jurkat cells to directly induce TNF-mRNA expression and TNF production by human peripheral blood monocytes. Northern analysis showed that incubation of monocytes with either K562 or Jurkat cells led to a significant increase in TNF-mRNA expression. In addition, enhanced TNF production was detected in supernatants of monocyte cultures activated by Jurkat cells. Not only viable tumor cells but also metabolically inactivated tumor cells, cytoblasts, and membrane preparations from Jurkat and K562 cells induced TNF-mRNA expression. We identified two different membrane protein fractions with relative molecular mass of 32 to 38 kDa for Jurkat cells and 46 to 54 kDa for K562 cells that were responsible for monocyte activation.

Tumor-associated tumor necrosis factor (TNF) production in patients as well as a TNF-inducing membrane constituent of tumor cells have been reported. In a murine fibrosarcoma model we analyzed TNF production during growth of a tumor transplant. In situ hybridization showed that a gradually increasing number of cells within the tumor tissue became positive for TNFmRNA. Also, in spleen cells of tumor-bearing mice TNFmRNA became more abundant in later stages of tumor growth compared to early stages. In plasma of these animals, however, TNF activity was not detected at any time even after stimulation with bacterial endotoxin. Neutralization with monoclonal antibodies of endogenous TNF during tumor growth did not affect the growth rate of the tumor, indicating that either the antibodies did not reach the relevant TNF production and action sites or that endogenously produced TNF did not play a significant role in this tumor model.

The dependency of induction of T cell cytotoxicity on lymphokines was studied. 1 X 10(5) nylon wool-purified thymic lymphocytes or 10(4) spleen cells were cultured with TNP-haptenated syngeneic UV-irradiated spleen cells in the presence of a variety of lymphokine preparations. Concanavalin A-induced spleen cell supernatants mediated strong cytotoxic responses in this system. Three other preparations, namely, a partially purified IL 2 preparation from PMA-stimulated EL-4 thymoma cells, a Con A-induced spleen cell supernatant that was absorbed with an IL 2-dependent cell line, and a Con A-induced supernatant that was dialyzed at pH 2 were all ineffective in mediating a cytotoxic response. In reconstitution experiments, cytotoxic responses were only obtained when either the absorbed preparation or the pH 2-treated preparation was mixed with the IL 2 preparation from EL-4 cells. No reconstitution occurred after mixing of the absorbed with the pH 2-treated preparation. pH 2 treatment of the absorbed preparation did not abolish its synergistic effect when added to the IL 2 preparation from EL-4 cells. These results led to the conclusion that activation of cytotoxic lymphocyte precursors requires at least two other lymphokines in addition to IL 2. One T cell cytotoxicity-inducing factor (TCF1) remained in Con A-induced supernatants after absorption with IL 2 receptor-bearing T cell line cells. It was pH 2-resistant and was not found in EL-4 supernatants. A second T cell cytotoxicity-inducing factor (TCF2) was pH 2-sensitive and was found in Con A-induced spleen cell supernatants as well as in interferon-free supernatants of PMA-stimulated EL-4 cells. This activity co-purified with IL 2. It was absorbed by the IL 2-dependent T cell line together with IL 2. IL 2 differs from TCF2 since it is pH 2-resistant.

Thymocyte cultures from C3H/HeJ mice were stimulated for proliferative responses with purified preparations of interleukin 1 (IL 1) and interleukin 2 (IL 2). Synergistic responses were obtained in the absence of mitogen. In the presence of excess IL 2, the thymocyte proliferation response was strictly dependent on the amount of IL 1 in the cultures. Antibodies to IL 1 inhibited the response in a dose-dependent manner. The combination of IL 1 plus IL 2 induced the appearance of IL 2 receptors on murine thymocytes as detected with a monoclonal antibody directed against the IL 2 receptor. Neither IL 1 nor IL 2 alone had this effect. The thymic subpopulation found to become IL 2 responsive upon IL 1 stimulus was the peanut agglutinin-negative (PNA-) medullary fraction.

Human recombinant tumor necrosis factor, TNF, was used to assess its ability to stimulate granulopoiesis and to protect mice against lethal irradiation, effects known to be inducable with TNF-rich postendotoxin serum from BCG infected mice (BCG/ET serum). Although the endotoxin contamination of this TNF preparation is extremely low its effects were compared in endotoxin low responder C3H/HeJ mice and susceptible NMRI mice. TNF is a potent inducer of serum colony stimulating activity, CSA, in both mouse strains. In peripheral blood a marked granulocytosis with a concomitant decrease in lymphocytes and monocytopenia occurs at 2 hours after injection of TNF. Moreover, TNF induces an increase in the number of splenic myelopoietic committed stem cells (GM-CFC, granulocyte-macrophage colony forming cells) determined five days after injection. The lethality rate, registered over 30 days after exposure to 660 cGy whole body X-irradiation is reduced to 40% in C3H/HeJ mice as compared to 75% in control animals. The reduction in lethality is observed both, when TNF was injected 24 hours before or after irradiation. In vitro, TNF significantly increases the number of colonies in the presence of CSA in bone marrow cultures. TNF per se does not effect colony growth. The studies reported here demonstrate that TNF is a myelopoiesis stimulating factor in mice which may be related to the reduction in lethality following whole body irradiation.

We investigated the role of cytokines for the growth of CD4-8-thymocytes (double negative thymocytes) (DNT) in vitro and found that IL-1-induced IL-2-dependent proliferation of only the IL-2R-positive DNT subpopulation. The presence of IL-1 during the first 18 h of culture was sufficient for an optimal response and suggested that IL-1 induced DNT differentiation. We could indeed show by RNA dot blot analysis that IL-1 stimulated de novo expression of the p55 chain of the IL-2R thus initiating high affinity IL-2 binding and a proliferative response. Because macrophages and epithelial cells in the thymus produce IL-1 we propose that IL-1 is involved in early events during maturation of immature thymocytes.

We investigated the role of T cell cytotoxicity inducing factor 1 (TCF1) in the induction of a cytotoxic T cell response. We found that help-deficient thymocyte cultures supplied with saturating amounts of purified IL 2 did not develop CTL in a 5-day culture. The expression of cytotoxicity was dependent on the addition of TCF1 derived from the T cell hybridoma K15. TCF1 also induced proliferation of thymocytes in the presence of IL 2. Only the PNA- thymocyte subpopulation responded to TCF1 with proliferation and cytotoxicity in the presence of IL 2. The monokine IL 1 also induced proliferation in this subpopulation but failed to induce cytotoxicity. IL 1 was further distinguished from TCF1 by inhibition of IL 1-induced but not TCF1-induced proliferation by anti-IL 1 antibodies. In addition, using anti-IL 2 receptor antibodies (AMT 13), we showed that TCF1 in the presence of IL 2 substantially increased IL 2 receptor expression in thymocytes. IL 1 had the same effect on induction of IL 2 receptor expression as TCF1. Because some effects of IL 1 and TCF1 are distinct and some overlap, we discuss whether IL 1 and TCF1 induce different subsets of PNA- thymocytes.

We have employed a human promyelocytic leukemic cell line, HL-60, to investigate the conditions that regulate TNF-alpha gene expression in vitro. Using a cloned TNF-alpha specific cDNA probe, we show by Northern blot analysis that TNF-alpha mRNA rapidly accumulates in HL-60 cells following treatment with phorbol myristate acetate (PMA). TNF-alpha mRNA levels peak at 1-2 hours and then decline to base-line levels within 8 hours. TNF-alpha protein levels as detected in the supernatants of PMA-stimulated HL-60 cells peak at 2 hours and decline within 24 hours. Cytokines of recombinant source like interferon-alpha and -gamma, and, more intruigingly, TNF-beta as well as TNF-alpha itself are also able to induce transient TNF-alpha mRNA expression in HL-60 cells. In the presence of a protein kinase inhibitor, neither PMA nor any of the cytokines used are able to induce TNF-alpha mRNA accumulation, indicating that protein kinases may be crucially involved in signal transduction leading to activation of the TNF-alpha gene. The data presented provide new insights into the control of TNF-alpha gene expression suggesting regulatory functions of T- and B-cell derived lymphokines as well as a TNF-alpha-mediated positive feedback mechanism.

Thymocyte cultures from C3H/HeJ mice were stimulated for proliferative responses with purified preparations of interleukin 1 (IL 1) and interleukin 2 (IL 2). Synergistic responses were obtained in the absence of mitogen. In the presence of excess IL 2, the thymocyte proliferation response was strictly dependent on the amount of IL 1 in the cultures. Antibodies to IL 1 inhibited the response in a dose-dependent manner. The combination of IL 1 plus IL 2 induced the appearance of IL 2 receptors on murine thymocytes as detected with a monoclonal antibody directed against the IL 2 receptor. Neither IL 1 nor IL 2 alone had this effect. The thymic subpopulation found to become IL 2 responsive upon IL 1 stimulus was the peanut agglutinin-negative (PNA-) medullary fraction.

Recombinant tumor necrosis factor (TNF) had hypothermic activity in vivo. Intravenous injection of TNF resulted in a hypothermic reaction of mice within 3 to 6 hours. This reaction was not the result of interleukin 1 (IL1) release from macrophages/monocytes. Peritoneal exudate cell cultures from endotoxin low responder mice or human peripheral mononuclear leukocyte cultures did not generate IL1 activity in the supernatant after exposure to TNF. The addition of interferon-gamma (IFN-gamma) or preexposure to IFN-gamma and then stimulation with TNF did also not result in IL1 secretion. No IL1 inhibitor was generated and TNF did not interfere with the IL1 test systems. Therefore, we conclude that the hypothermic activity of TNF is not mediated via the induction of IL1 production by mononuclear phagocytes.

Cells of the proerythromyeloid cell line K562 and of the T cell line Jurkat were able to induce an increase in TNF-, IL1 alpha-, and IL1 beta-mRNA expression in human peripheral blood derived monocytes in vitro. The activating principle of these tumor cells was associated with fractions of membrane preparations of distinct molecular mass, 32-38 kD for Jurkat cells and 46-54 kD for K562 cells, respectively. At least part of the activating constituent seemed to be protein in nature. Isolated membrane preparations of both cell types induced production and secretion of TNF. Stimulation of monocytes with the viable tumor cells led to TNF release only when Jurkat cells were used. Viable K562 cells induced enhanced TNFmRNA expression but seemed to absorb soluble TNF from the supernatant.

Tumor necrosis factor (TNF), a protein predominantly produced by activated macrophages/monocytes, is presently available in recombinant, purified form for clinical trials. Intensive studies in many laboratories have shown that besides the tumorcytotoxic effects, TNF acts on a large array of different cells and has potent immunomodulatory activities. In a clinical phase I study, some immunologic functional parameters of blood cells from patients who received 24-hour infusions of recombinant human TNF (rhTNF) were analyzed. Natural killer (NK) cell activity, TNF production, interleukin-1 (IL-1) production and mitogen-induced proliferation were measured either in whole blood samples or in cultures of peripheral mononuclear leukocytes of the patients directly before and after rhTNF infusion. NK cell activity, TNF and IL-1 production capacity and proliferative responses to concanavalin A (Con A) were significantly reduced after rhTNF application. We conclude from these observations that rhTNF in vivo acts directly or indirectly on NK cells and monocytes by either inactivating their functional capacity or by absorbing the relevant cells to the endothelial cell layer, thus removing them from circulation.

C3H/HeN mice administered BCG followed by lipopolysaccharide 14 days later released into their serum a cytotoxic factor for tumor cells and a factor that restored the anti-SRBC plaque-forming cell response of nude mouse spleen cells (helper activity). Gel filtration of serum containing the cytotoxic and the helper activities indicated that both factors exhibited an apparent m.w. of 125,000 to 150,000. The helper activity was also found at lower m.w. (60,000 and 13,000) suggesting the possibility that this factor existed in aggregated forms. Gel filtration of ammonium sulfate (40 to 60% saturation) precipitated serum in a high ionic strength buffer (1.6 M NaCl) resulted in shifts in the apparent m.w. of both factors. The cytotoxic factor now exhibited a m.w. of 55,000. The helper activity eluted with an apparent m.w. of 13,000, and thus was clearly separated from the cytotoxic factor. The helper activity was further shown to co-elute with macrophage-derived lymphocyte activating factor (LAF). This as well as other data represent the first demonstration of in vivo produced LAF.