Inhibition of cytotoxic T cell development by transforming growth factor beta and reversal by recombinant tumor necrosis factor alpha.

The immunoregulatory effects of transforming growth factor beta (TGF-beta) and recombinant murine tumor necrosis factor alpha (rMuTNF-alpha) on CTL generation and activity were examined. The results demonstrate that TGF-beta, in a dose-dependent manner, inhibited CTL generation but not CTL activity. The inhibitory effects were detected only when TGF-beta was added within the first 48 h of the MLC. Little activity was seen when it was added thereafter, including the addition of TGF-beta to the cytotoxicity assay. The production of TNF-alpha, which occurs during early phases of the MLC and which is inhibited in the presence of TGF-beta, appears to have an important regulatory role, as altering the levels of TNF-alpha in an MLC can significantly influence CTL development. The inhibitory effects of TGF-beta on the MLC can be significantly reversed by the addition of rMuTNF-alpha to the cultures. These results demonstrate that TGF-beta can inhibit MLC and subsequent CTL generation at early stages of the reaction, and such inhibition may involve the suppression of TNF-alpha production.

Immunoregulatory role of transforming growth factor beta (TGF-beta) in development of killer cells: comparison of active and latent TGF-beta 1.

Using recombinant DNA technology, we have generated Chinese hamster ovary (CHO) cell lines that synthesize latent transforming growth factor beta 1 (TGF-beta 1) to study immune regulation by TGF-beta 1. In vitro, latent TGF-beta 1 synthesized by transfectants or added exogenously as a purified complex after activation inhibited CTL generation to a similar extent as seen with acid-activated recombinant human (rHu) TGF-beta 1. In vivo, serum from nu/nu mice bearing CHO/TGF-beta 1 tumors contained significant levels of latent TGF-beta 1 in addition to depressed natural killer (NK) activity in spleens which paralleled that seen in C3H/HeJ mice treated with acid-activated rHuTGF-beta 1. rHuTGF-beta 1 treatment of mice receiving heart allografts resulted in significant enhancement of organ graft survival. Because of possible regulated tissue-specific activation, administration of latent rather than active TGF-beta may provide a better route to deliver this powerful immunosuppressive agent in vivo.

Inhibition of cytokine production by cyclosporin A and transforming growth factor beta.

We investigated the ability of cyclosporin A (CsA) and transforming growth factor beta (TGF-beta) to modulate the production of TNF-alpha and TNF-beta and IFN-gamma by unseparated, nonadherent, and adherent PBMC. Treatment of unseparated PBMC with CsA resulted in a significant dose-dependent inhibition of all three cytokines ranging from greater than 90% inhibition for IFN-gamma and TNF-beta, to approximately 70% for TNF-alpha. Pretreatment of unseparated or nonadherent PBMC with TGF-beta inhibited the production of IFN-gamma by 60-70%. However, the inhibition of TNF-alpha and TNF-beta production by these cells was only minimally affected, and at 0.1-1 ng/ml TGF-beta could enhance TNF-alpha production by unseparated PBMC. In contrast, pretreatment of adherent PBMC with TGF-beta inhibited the production of TNF-alpha by approximately 60%. TGF-beta also inhibited both TNF-alpha production and tumor cell cytotoxicity mediated by murine peritoneal-derived macrophages. These observations indicate that the biological effects of CsA and TGF-beta on immune functions are of a wider range than previously reported.

Tumor necrosis factor (cachectin) is an endogenous pyrogen and induces production of interleukin 1.

Recombinant human tumor necrosis factor (rTNF alpha) injected intravenously into rabbits produces a rapid-onset, monophasic fever indistinguishable from the fever produced by rIL-1. On a weight basis (1 microgram/kg) rTNF alpha and rIL-1 produce the same amount of fever and induce comparable levels of PGE2 in rabbit hypothalamic cells in vitro; like IL-1, TNF fever is blocked by drugs that inhibit cyclooxygenase. At higher doses (10 micrograms/kg) rTNF alpha produces biphasic fevers. The first fever reaches peak elevation 45-55 min after bolus injection and likely represents a direct action on the thermoregulatory center. During the second fever peak (3 h later), a circulating endogenous pyrogen can be shown present using passive transfer of plasma into fresh rabbits. This likely represents the in vivo induction of IL-1. In vitro, rTNF alpha induces the release of IL-1 activity from human mononuclear cells with maximal production observed at 50-100 ng/ml of rTNF alpha. In addition, rTNF alpha and rIFN-gamma have a synergistic effect on IL-1 production. The biological activity of rTNF alpha could be distinguished from IL-1 in three ways: the monophasic pyrogenic activity of rIL-1 was destroyed at 70 degrees C, whereas rTNF alpha remained active; anti-IL-1 neutralized IL-1 but did recognize rTNF alpha or natural cachectin nor neutralize its cytotoxic effect; and unlike IL-1, rTNF alpha was not active in the mitogen-stimulated T cell proliferation assay. The possibility that endotoxin was responsible for rTNF alpha fever and/or the induction of IL-1 was ruled-out in several studies: rTNF alpha produced fever in the endotoxin-resistant C3H/HeJ mice; the IL-1-inducing property of rTNF alpha was destroyed either by heat (70 degrees C) or trypsinization, and was unaffected by polymyxin B; pyrogenic tolerance to daily injections of rTNF alpha did not occur; levels of endotoxin, as determined in the Limulus amebocyte lysate, were below the minimum rabbit pyrogen dose; and these levels of endotoxin were confirmed by gas chromatography/mass spectrometry analysis for the presence of beta-hydroxymyristic acid. Although rTNF alpha is not active in T cell proliferation assays, it may mimic IL-1 in a T cell assay, since high concentrations of rTNF alpha induced IL-1 from epithelial or macrophagic cells in the thymocyte preparations. These studies show that TNF (cachectin) is another endogenous pyrogen which, like IL-1 and IFN-alpha, directly stimulate hypothalamic PGE2 synthesis. In addition, rTNF alpha is an endogenous inducer of IL-1.(ABSTRACT TRUNCATED AT 400 WORDS)

Multiple defects of immune cell function in mice with disrupted interferon-gamma genes.

Interferon-gamma (IFN-gamma) is a pleiotrophic cytokine with immunomodulatory effects on a variety of immune cells. Mice with a targeted disruption of the IFN-gamma gene were generated. These mice developed normally and were healthy in the absence of pathogens. However, mice deficient in IFN-gamma had impaired production of macrophage antimicrobial products and reduced expression of macrophage major histocompatibility complex class II antigens. IFN-gamma-deficient mice were killed by a sublethal dose of the intracellular pathogen Mycobacterium bovis. Splenocytes exhibited uncontrolled proliferation in response to mitogen and alloantigen. After a mixed lymphocyte reaction, T cell cytolytic activity was enhanced against allogeneic target cells. Resting splenic natural killer cell activity was reduced in IFN-gamma-deficient mice. Thus, IFN-gamma is essential for the function of several cell types of the murine immune system.

Recombinant human tumor necrosis factor-alpha: effects on proliferation of normal and transformed cells in vitro.

Modulation of the growth of human and murine cell lines in vitro by recombinant human tumor necrosis factor-alpha (rTNF-alpha) and recombinant human interferon-gamma (rIFN-gamma) was investigated. rTNF-alpha had cytostatic or cytolytic effects on only some tumor cell lines. When administered together with rIFN-gamma, rTNF-alpha showed enhanced antiproliferative effects on a subset of the cell lines tested. In contrast to its effects on sensitive tumor cells, rTNF-alpha augmented the growth of normal diploid fibroblasts. Variations in the proliferative response induced by rTNF-alpha were apparently not due to differences in either the number of binding sites per cell or their affinity for rTNF-alpha. These observations indicate that the effects of rTNF-alpha on cell growth are not limited to tumor cells, but rather that this protein may have a broad spectrum of activities in vivo.

Characterization of the in vitro and in vivo species preference of human and murine tumor necrosis factor-alpha.

The species preference of human and murine tumor necrosis factor-alpha (TNF-alpha) was evaluated in human and murine systems for cytotoxic/cytostatic effects and receptor binding in vitro and murine systems for toxicity and antitumor activity in vivo. The in vitro cytotoxic/cytostatic effects of both species TNF-alpha on human and murine cell lines as well as the receptor binding studies using 125I-labeled recombinant human TNF-alpha demonstrated homologous species preferences. Species preference of TNF-alpha was also apparent in toxicity studies with BALB/c nu/nu and CB6F1 mice, and antitumor responses of CB6F1 mice to s.c. Meth A sarcoma implants. Moreover the growth of Meth A sarcoma implanted i.p. was not inhibited by either human or murine TNF-alpha. These results are discussed in view of the potential for underestimation of the biological potency of TNF-alpha from heterologous sources.

Regulation of neutrophil migration and superoxide production by recombinant tumor necrosis factors-alpha and -beta: comparison to recombinant interferon-gamma and interleukin-1 alpha.

We compared the ability of recombinant human tumor necrosis factor-alpha (rHuTNF-alpha) and tumor necrosis factor-beta (rHuTNF-beta) to stimulate polymorphonuclear neutrophil (PMN) migration and superoxide production. Significant PMN migration occurred across polycarbonate filters after stimulation with rHuTNF-alpha at concentrations ranging from 10(-7) to 10(-10) mol/L and at 10(-7) to 10(-8) mol/L for rHuTNF-beta and N-formylmethionyl-leucyl phenylalanine (FMLP), whereas recombinant human interferon-gamma was only minimally active at 10(-7) mol/L and recombinant human interleukin-1 alpha was inactive at the doses tested. In addition, antibodies to rHuTNF-alpha completely inhibited rHuTNF-alpha but not rHuTNF-beta or FMLP-induced PMN migration. Combinations of rHuTNF-alpha and rHuTNF-beta (at similar molar concentrations) stimulated PMN migration levels comparable to that obtained with rHuTNF-alpha alone. Checkerboard analyses performed by placing different concentrations of rHuTNF-alpha and rHuTNF-beta above and below polycarbonate filters of microchemotaxis chambers demonstrated that rHuTNF-alpha and rHuTNF-beta stimulated both chemotactic and chemokinetic responses by PMN. Additional studies demonstrated that 1 X 10(-8) mol/L rHuTNF-alpha and 3 X 10(-9) mol/L rHuTNF-beta (which represents 10(4) U/mL of each cytokine) were similar in their ability to induce superoxide production by PMNs; however, at ten- to 100-fold lower molar concentrations (10(3) and 10(2) units), rHuTNF-alpha was significantly more active than rHuTNF-beta. At the doses tested, both cytokines were less active than phorbol myristate acetate at stimulating O2- release. The results demonstrate that rHuTNF-alpha and rHuTNF-beta differ quantitatively but not qualitatively in their effects on PMN functions in vitro and suggest that rHuTNF-beta may be less toxic than rHuTNF-alpha in vivo.

Transforming growth factor-beta 1 (TGF-beta 1) and recombinant human tumor necrosis factor-alpha reciprocally regulate the generation of lymphokine-activated killer cell activity. Comparison between natural porcine platelet-derived TGF-beta 1 and TGF-beta 2, and recombinant human TGF-beta 1.

We have investigated the ability of porcine-platelet-derived transforming growth factor-beta 1 (TGF-beta 1) to inhibit the generation of lymphokine-activated killer (LAK) cells by human rIL-2. The results demonstrate that TGF-beta 1, in a dose-related manner, significantly inhibits rIL-2-induced LAK cell activity against Daudi and COLO target cells and, to a lesser degree, against K-562 cells. Maximal inhibition was obtained by the addition of TGF-beta 1 at the time of culture initiation and, to a lesser degree, on day 1. Only minimal inhibition was obtained when TGF-beta 1 addition was delayed until day 2 of culture or when added directly into the LAK cell assay. Additional studies demonstrated that porcine platelet-derived TGF-beta 2 and human rTGF-beta 1 inhibited LAK cell generation similar to that obtained with TGF-beta 1. The inhibition of LAK cell activity by TGF-beta 1 was reversed by the addition of human rTNF-alpha at the initiation of culture. In addition, rTNF-alpha synergized with suboptimal levels of rIL-2 in the generation of LAK activity. After stimulation with rIL-2, LAK cells produced significant levels of IFN-gamma, TNF-alpha, and TNF-beta. TGF-beta 1 inhibited the production of these cytokines in a dose-related manner. The results extend the previous known activities for human rTNF-alpha and TGF-beta 1 and further demonstrate the reciprocal relationship between these two molecules in the regulation of certain immune functions.

Possible relationships between in vivo antitumour activity and toxicity of tumour necrosis factor-alpha.

The discoveries of a tumour necrosis-inducing substance in sera of experimental mice and of cytotoxic factor(s) in cultures of stimulated lymphoid cells triggered intense research efforts which have culminated in the production of two distinct but related recombinant materials, human tumour necrosis factors TNF-alpha (cachectin) and TNF-beta (lymphotoxin). The necrosis of tumours by TNF is but one feature of high doses of these immune system hormones that possess numerous biological activities. Apart from their direct cytotoxic/cytostatic activities against tumours in vitro and in vivo, the in vivo antitumour activities of TNF-alpha or TNF-beta may involve the following biological activities: the induction of interleukin 1 production; activation of polymorphonuclear neutrophil functions; modulation of endothelial cell functions; and augmentation of specific immune functions. Many of these activities are associated with an irreversible acute inflammation which appears to be the immediate lethal effect of TNF on transplantable tumours in mice. This inflammation leads to thrombosis, disruption of the tumour's blood supply and, finally, tumour death. Inflammatory effects of high doses of TNF are also seen in the rodent gastrointestinal tract but here the inflammation seems to be reversible.

The two different receptors for tumor necrosis factor mediate distinct cellular responses.

The individual roles of the murine type 1 and type 2 tumor necrosis factor (TNF) receptors (TNF-R1 and TNF-R2) were investigated utilizing (i) the strong species specificity of TNF-R2 for murine TNF compared to human TNF and (ii) agonistic rabbit polyclonal antibodies directed against the individual TNF receptors. Proliferation of mouse thymocytes and the murine cytotoxic T-cell line CT-6 is stimulated by murine TNF but not by human TNF. Consistent with this observation, polyclonal antibodies directed against TNF-R2 induced proliferation in both of these cell types, whereas polyclonal antibodies directed against TNF-R1 had no effect. In contrast, cytotoxicity in murine LM cells (which are sensitive to murine and human TNF) was induced by antibodies against TNF-R1 but not by antibodies against TNF-R2. Also, the steady-state level of manganous superoxide dismutase mRNA in the murine NIH 3T3 cell line was induced by murine TNF, human TNF, and anti-TNF-R1 but not by anti-TNF-R2. These results suggest that TNF-R2 initiates signals for the proliferation of thymocytes and cytotoxic T cells, whereas TNF-R1 initiates signals for cytotoxicity and the induction of the protective activity, manganous superoxide dismutase. The nonredundant signaling observed for the two TNF receptors cannot be explained simply by the differential expression of the two TNF receptors in the various cell types, because LM cells express on their surface higher levels of TNF-R2 than TNF-R1, and LM cells, NIH 3T3 cells, and thymus cells all express mRNA corresponding to both receptor types. It is therefore likely that the two receptors initiate distinct signaling pathways that result in the induction of different cellular responses.

Stimulation of human T-cell proliferation by specific activation of the 75-kDa tumor necrosis factor receptor.

TNF-alpha can enhance the proliferation of human thymocytes stimulated by the comitogen Con A. To determine which of the two different TNF receptors is responsible for signaling this cellular response, we investigated the proliferation of human thymocytes in response to agonistic antibodies specific for the two TNF receptor types. In contrast to previously examined TNF activities in human cells, thymocyte proliferation was stimulated in response to rabbit polyclonal antibodies directed against the 75-kDa TNF receptor (TNF-R2), but not those directed against the 55-kDa TNF receptor (TNF-R1). Lymphotoxin (TNF-beta) was also shown to stimulate human thymocyte proliferation, demonstrating that TNF-beta can initiate a biologic response that is mediated by TNF-R2. TNF-R2-mediated T-cell proliferation was not restricted to the immature T cells within the thymus, as the anti-TNF-R2 antibodies also stimulated the proliferation of peripheral T cells. As a first step toward identifying a specific agonist of TNF-R2 with therapeutic potential, 10 anti-TNF-R2 mAb were examined for potential agonist activity. Nine of these significantly stimulated human thymocyte proliferation with maximal responses ranging from twofold to significantly greater than that obtained with TNF-alpha by itself.

The involvement of human tumor necrosis factors-alpha and -beta in the mixed lymphocyte reaction.

The influence of recombinant human TNF-alpha and -beta (rHuTNF-alpha and -beta) in a human mixed lymphocyte reaction (MLR) was investigated. The addition of 1000 U/ml of either cytokine at the initiation of culture caused up to a sixfold increase in [3H]thymidine incorporation by responder cells. Furthermore, it was found that endogenous HuTNF-alpha is produced after allogeneic cell interaction and can be detected in the MLR supernatant within 1 h of culture initiation. The results also show that, in the absence of exogenous HuTNF-alpha, antibodies to rHuTNF-alpha can cause a significant inhibition of the MLR. These observations indicate the importance of TNF-alpha in allogeneic cell interaction and raise considerations for the use of antibodies, or other antagonists, to TNF-alpha as regulators of disease states associated with cell-mediated immune reactions.

The autocrine production of transforming growth factor-beta 1 during lymphocyte activation. A study with a monoclonal antibody-based ELISA.

We describe the production and characterization of three mAb to transforming growth factor-beta (TGF-beta) and the use of two of them for the development of a TGF-beta 1-specific ELISA and for the study of the regulation of immune function in vitro. All three mAb bound recombinant human TGF-beta 1 (rHuTGF-beta 1) with high affinity and recognized the dimer form of this molecule in immunoblots. mAb 2G7 immunoprecipitated rHuTGF-beta 1, TGF-beta 2, and rHuTGF-beta 3 and neutralized the growth inhibitory activity of all three molecules in vitro on mink lung epithelial-like cells, Mv1Lu, indicating a shared neutralization epitope. mAb 4A11 neutralized and immunoprecipitated only rHuTGF-beta 1, and mAb 12H5 immunoprecipitated rHuTGF-beta 1 but had no effect on the bioactivity of either rHuTGF-beta 1, TGF-beta 2, or rHuTGF-beta 3. These results suggest that a second neutralization epitope may be unique to TGF-beta 1. The ELISA was developed with mAb 4A11 and 12H5, with a range of 0.63 to 40 ng/ml, i.e., a sensitivity of 0.63 ng/ml or 63 pg/sample. The assay is accurate, precise, and specific for the active but not the inactive or latent TGF-beta 1 complex and fails to react with TGF-beta 2, rHuTGF-beta 3, inhibin A, and activin A. Supernatants obtained from serum-free cultures of human PBMC from multiple donors contained significant quantities of TGF-beta 1 (3 to 15 ng/ml), which was detected in the ELISA only after pH 2 treatment to convert latent TGF-beta to the active form. Treatment of the PBMC with either recombinant human IL-2 (rHuIL-2) or PHA-P/PMA enhanced the production of latent TGF-beta 1. mAb 4A11 and 2G7, but not mAb 12H5 enhanced both the proliferative response of PBMC to rHuIL-2/rHuTNF-alpha and PHA-P and the development of the rHuIL-2/rHuTNF-alpha treated PBMC into LAK cells with cytotoxic activity against COLO target cells. These findings suggest that although PBMC secrete latent TGF-beta 1, mechanisms that convert the latent TGF-beta complex into an active form exist in vitro and that the endogenously produced TGF-beta can regulate immune functions in an autocrine fashion.

Protection against endotoxic shock by a tumor necrosis factor receptor immunoadhesin.

Tumor necrosis factors (TNF) alpha and beta are structurally related cytokines that mediate a wide range of immunological, inflammatory, and cytotoxic effects. During bacterial infection of the bloodstream (sepsis), TNF-alpha induction by bacterial endotoxin is thought to be a major factor contributing to the cardiovascular collapse and critical organ failure that can develop. Despite antibiotic therapy, these consequences of sepsis continue to have a high mortality rate in humans. Here we describe a potent TNF antagonist, a TNF receptor (TNFR) immunoadhesin, constructed by gene fusion of the extracellular portion of human type 1 TNFR with the constant domains of human IgG heavy chain (TNFR-IgG). When expressed in transfected human cells, TNFR-IgG is secreted as a disulfide-bonded homodimer. Purified TNFR-IgG binds to both TNF-alpha and TNF-beta and exhibits 6- to 8-fold higher affinity for TNF-alpha than cell surface or soluble TNF receptors. In vitro, TNFR-IgG blocks completely the cytolytic effect of TNF-alpha or TNF-beta on actinomycin D-treated cells and is markedly more efficient than soluble TNFR (24-fold) or monoclonal anti-TNF-alpha antibodies (4-fold) in inhibiting TNF-alpha. In vitro, TNFR-IgG prevents endotoxin-induced lethality in mice when given 0.5 hr prior to endotoxin and provides significant protection when given up to 1 hr after endotoxin challenge. These results confirm the importance of TNF-alpha in the pathogenesis of septic shock and suggest a clinical potential for TNFR-IgG as a preventive and therapeutic treatment in sepsis.

Characterization of the antitumor activities of human tumor necrosis factor-alpha and the comparison with other cytokines: induction of tumor-specific immunity.

We have investigated the in vitro and in vivo antitumor activities of recombinant human tumor necrosis factor-alpha (rHuTNF-alpha) against Meth A sarcoma. Meth A sarcoma cells were found to a) be relatively insensitive in vitro to rHuTNF-alpha, and b) express low numbers of TNF-alpha receptors. Intraperitoneally implanted Meth A sarcoma was insensitive to the antitumor effects of rHuTNF-alpha. In contrast, rHuTNF-alpha was highly efficacious against subcutaneously implanted Meth A sarcoma. Biodistribution studies with 125I- or 3H-labeled rHuTNF-alpha demonstrated that, after intravenous administration, the majority of the labeled rHuTNF-alpha localized in the kidney, lungs, and liver. Only low levels of radiolabel were found in subcutaneous Meth A implants. These results support the in vitro data on the low number of TNF-alpha receptors on Meth A sarcoma cells. The ability of rHuTNF-alpha to induce regression of established (7 days) subcutaneous Meth A implants, positively correlated with the degree of both macroscopic and microscopic tumor necrosis. In addition, recombinant human tumor necrosis factor-beta (lymphotoxin) and recombinant murine tumor necrosis factor-alpha induced similar levels of necrosis. Other lymphokines with known antitumor activities, recombinant human interferon-gamma, murine interferon-gamma, and human interleukin 1 alpha, failed to induce detectable necrosis of Meth A sarcoma. Mice which had rejected subcutaneous Meth A sarcoma implants after rHuTNF-alpha treatment and which were later challenged subcutaneously with Meth A sarcoma or other noncross-reacting chemically induced sarcomas were found to be specifically immune to Meth A sarcoma. In addition, low levels of cytotoxic antibodies reactive to Meth A sarcoma were detected in the sera of 21 of 30 Meth A immune mice. Histological evaluation of the hemorrhagic tumor necrosis induced by rHuTNF-alpha suggests that the primary lesion is vascular, possibly directly on the endothelial cells. The mechanisms involved in the generation of specific cell-mediated antitumor immunity in this model are at present unknown.