Induction of tumor necrosis factor (TNF) by OK-432

The induction of tumor necrosis factor (TNF) by OK-432 was investigated. The results were as follows: Serum from OK-432 injected into NZW rabbits treated with LPS was cytotoxic to tumor cells in vitro. With respect to the target cell spectrum this factor was cytotoxic to L929, Lewis, M7609 and K562, but not to Chang liver cells, HMV cells or HeLa cells. The cytotoxic activity of this factor was found to be M.W. 48,000 on H.P.L.C., and the pI was revealed to be 5.4 by isoelectric focusing. These results were consistent with previously reported findings on TNF.     

Two discrete types of tumor necrosis factor-resistant cells derived from the same cell line.

From the murine fibrosarcoma cell line L929s, which is sensitive to tumor necrosis factor (TNF)-mediated cell lysis, two discrete types of TNF-resistant variants were derived by TNF selection. Cells of the first type (named L929r1) were not sensitized to TNF cytotoxicity by cotreatment with either inhibitors of protein or RNA synthesis, or gamma-interferon, despite the presence of a functional gamma-interferon response. L929r1 constitutively produced TNF in the supernatant and expressed membrane-bound TNF, which was not bound to the TNF receptor. In fact, TNF receptors could not be demonstrated on L929r1 cells, not even after low pH treatment and/or incubation with antiserum to TNF. L929r1 exhibited a stable TNF-resistant phenotype in the absence of further TNF selection. No evidence could be obtained that TNF acted as an autocrine growth factor for these cells. L929r2, the second type of TNF-resistant L929 cells, became sensitive to TNF lysis in the presence of RNA or protein synthesis inhibitors, or in the presence of gamma-interferon. TNF induced the secretion of interleukin 6 in these cells, additionally showing that functional TNF signaling in these cells indeed takes place, but does not lead to cell lysis under normal conditions. L929r2 did not produce TNF, also not upon stimulation with exogenous TNF. The number and binding affinity of TNF receptors were not consistently different between L929s and L929r2 cells. In the absence of further TNF selection, L929r2 gradually reverted to TNF sensitivity. This sensitivity was not reversible to TNF resistance by the gene-regulatory agents 5-azacytidine or sodium butyrate. Treatment with these agents also did not affect the TNF sensitivity of L929s cells nor the TNF resistance of L929r1 and L929r2 cells. In summary, our results suggest the existence among cells of the same cell line of discrete mechanisms for acquisition of resistance to TNF-mediated cell lysis.     

Association of lysosomal activity with sensitivity and resistance to tumor necrosis factor in murine L929 cells

The cytotoxic mechanism of action of tumor necrosis factor (TNF) was examined using murine L929 fibrosarcoma cells in vitro. Two cell lines were evaluated: parental TNF sensitive (L929S) (50% cytotoxic concentration, 2-6 ng/ml); and TNF resistant (L929R) (50% cytotoxic concentration, greater than 10,000 ng/ml). The latter resistant cell line was developed by serial passage in increasing concentrations of recombinant human TNF. Sensitive cells demonstrated cytolytic and cytostatic effects at TNF concentrations between 2 and 6 ng/ml, respectively. However, TNF failed to show any selective depression of RNA, DNA, or protein synthesis or ATP content in these cells until general cell death was apparent, as defined by the cell rounding and lifting off the plastic surface. The cytokine also failed to cause DNA single-strand breaks, as detected by alkaline elution techniques. TNF was also found to be no more active in glutathione-depleted cells than in target cells containing normal glutathione levels. In contrast, various nonspecific lysosomotropic agents such as ammonium chloride and D-saccharic acid lactone led to a marked inhibition of the cytotoxic action of TNF in vitro. Furthermore, significant differences in lysosomal enzyme activity were noted between L929S and L929R cells. The changes in L929R cells involved a 50% reduction in total lysosomal protein levels and a marked depression of beta-glucuronidase activity. In contrast, L929R lysosomal hexosaminidase activity was significantly elevated over the L929S cells. From these studies it is concluded that the antitumor activity of TNF does not involve specific inhibition of macromolecular synthesis, ATP production, or the level of reduced thiols. Instead, TNF cytotoxicity appears to require functional lysosomes, which are altered when TNF resistance develops in vitro.     

Distinction of partially purified human natural killer cytotoxic factor from recombinant human tumor necrosis factor and recombinant human lymphotoxin

Natural killer cell cytotoxic factor (NKCF), a cytotoxic factor contributing to human natural killer cell-mediated cytotoxicity, was generated from lymphocyte-conditioned medium using various stimuli. Crude NKCF activity was concentrated, and partially purified by ammonium sulfate precipitation and gel filtration. NKCF activities eluted as two molecular weight peaks, corresponding to Mr 33,000-43,000 (pool I) and approximately Mr 5,000 (pool II). The cytotoxic activity and target specificity of the partially purified NKCFs were found to be different from both recombinant human TNF and recombinant human lymphotoxin. In the NKCF assay, up to 10(6) units/ml of TNF and lymphotoxin had virtually no effect, whereas both NKCFs lysed 22% (range 17-33%) of the NK-sensitive target K562. In contrast, TNF and lymphotoxin were active in a standard assay against the sensitive murine L929 fibroblast cell line in all concentrations tested (10(-1)-10(6) units/ml). In addition, the effect of these cytotoxic factors in a short-term (4-h) chromium-release assay using peripheral blood mononuclear cells as effector cells was tested: only NKCF (pool I), but not TNF, lymphotoxin, or low molecular weight NKCF (pool II), enhanced NK and lymphokine-activated killer cell cytolysis, both against the NK-sensitive target K562 and the NK-resistant melanoma cell line SK-MEL 30. Results were not affected in the presence of neutralizing antibodies against TNF. NKCF could, therefore, be distinguished from TNF and lymphotoxin with respect to their biological activities.     

Potentiation of topoisomerase inhibitor-induced DNA strand breakage and cytotoxicity by tumor necrosis factor: enhancement of topoisomerase activity as a mechanism of potentiation

A combination of tumor necrosis factor (TNF) and the topoisomerase I inhibitor, camptothecin, or the topoisomerase II inhibitors, teniposide and amsacrine, produced dose-dependent synergistic cytotoxicity against the murine L929 fibrosarcoma cells. Similar synergy was not observed with a combination of TNF and bleomycin. To define the role of TNF in the augmentation of tumor cell killing by topoisomerase I or II inhibitors, the effect of TNF on the production of enzyme-linked DNA strand breaks induced in cells by topoisomerase inhibitors was investigated. L929 cells incubated for 1 h with the topoisomerase inhibitors contained protein-linked strand breaks. In contrast, TNF alone did not induce DNA strand breakage. However, when cells were incubated simultaneously with TNF and camptothecin, amsacrine, Adriamycin, actinomycin D, teniposide, or etoposide, increased numbers of strand breaks were produced. Preincubation of the cells with TNF for 30 min or 3 h before the addition of camptothecin or etoposide resulted in no more strand breaks than that observed in cells incubated with the drugs alone. TNF treatment of L929 cells produced a rapid and transient increase in specific activity of extractable topoisomerases I and II. These increases were maximum at 2-5 min of TNF treatment and by 30 min the activities of extractable enzymes were equal to or less than those detected in extracts from untreated cell controls. The transient nature of the increase in extractable topoisomerase activity may explain the kinetics and significance of the order of addition of TNF and inhibitors for maximal synergistic activity. These data are consistent also with a role for topoisomerase-linked DNA lesions in the TNF-mediated potentiation of killing of L929 cells by topoisomerase inhibitors.     

Expression of the tumor necrosis factor gene in tumor cells correlates with reduced tumorigenicity and reduced invasiveness in vivo

We investigated whether a constitutive production of low amounts of tumor necrosis factor (TNF) by neoplastic cells affects their in vivo tumorigenicity. TNF-resistant derivatives were isolated from the TNF-sensitive murine fibrosarcoma cell lines L929s and WEHI164cl13s, L929r1-type TNF-resistant subclones were found to constitutively produce TNF in vitro, in contrast to non-TNF-producing but TNF-resistant L929r2 and WEHI164cl13r2 cell clones. The TNF-sensitive parental cell lines as well as the L929r2 and WEHI164cl13r2 cell lines similarly induced fast-growing tumors upon s.c. inoculation into nude mice (Swiss-nu/nu). In contrast, the TNF-producing L929r1-type cells showed reduced tumorigenicity and in vivo growth rate, which both inversely correlated with the level of in vitro-produced TNF. Tumor take incidence but not the in vivo growth rate of L929r1-type cells was greatly facilitated by prior whole body gamma-irradiation (350 rads) of the recipient, implying the involvement of host mechanisms at least in the lower take incidence of L929r1 tumors. These host mechanisms, possibly activated by tumor-produced TNF, acted only locally, inasmuch as the growth of an inoculum of L929s cells was not influenced either by a simultaneous distant inoculum of L929r1 cells or by established, distant L929r1 tumors. Efforts to eliminate these host mechanisms by prior local UV irradiation of the skin were unsuccessful. All L929 cell types were found to be similarly susceptible to killing by host cytotoxic effector cells (macrophages and natural and lymphokine-activated killer cells). Histological investigation did not reveal clear differences in tumor-associated inflammatory cells but revealed that tumors induced by L929r1-type cells, in contrast to L929s and L929r2 tumors, did not show invasiveness in host tissues. Moreover, L929r1 tumors were frequently encapsulated, which was never observed for tumors induced by L929s and L929r2 cells. Taken together, our results suggest that tumor-derived TNF locally activates host antitumor activities. Possible effector mechanisms are discussed.     

Serum protease inhibitor abrogation of Newcastle disease virus enhancement of cytolysis by recombinant tumor necrosis factors alpha and beta

Newcastle disease virus (NDV) has been used to induce regression of tumors in human cancer patients. We recently demonstrated that human malignant melanoma cells resistant to the lytic effects of tumor necrosis factor-alpha (TNF-alpha) become susceptible after treatment with NDV. We examined the effects of a serine protease inhibitor, N-1-tosylamide-2-phenyl-ethyl-chloromethyl ketone (TPCK), on viral enhancement of TNF cytotoxicity. Virulent NDV (but neither heat- nor UV-inactivated NDV) induced a 100-fold increase in the sensitivity of murine fibroblast L929 cells to recombinant human TNF-alpha (rHuTNF-alpha), rHuTNF-beta, and recombinant murine TNF-alpha (rMuTNF-alpha). TPCK, which is an inhibitor of chymotrypsin-like proteases, blocked between 42% and 93% of the cytolytic activity of rMuTNF-alpha, rHuTNF-alpha, and rHuTNF-beta toward NDV-treated L929 cells. Similarly, TPCK abrogated 62% of the cytotoxicity of rMuTNF-alpha toward dactinomycin-treated L929 cells. In contrast, TPCK had no effect on WEHI 164 clone 13 cells, a murine fibrosarcoma cell line that is much more sensitive to the lytic effects of TNF and does not show enhanced sensitivity to TNF after treatment with either NDV or dactinomycin. These results suggest a role for a cellular protease in the mechanism by which some viruses sensitize tumor cells to the cytolytic activity of TNF.     

Relationship between tumour cell morphology, gap junctions and susceptibility to cytolysis by tumour necrosis factor

Tumour necrosis factor (TNF) is directly cytolytic to certain tumour cell lines in vitro, although TNF-resistant variants can be selected from these susceptible lines by exposure to TNF. While studying TNF-susceptible L929 cells and their resistant variant, L929/R, we noted that within L929 colonies the cells were widely spaced whereas they were closely packed in L929/R colonies. L929/R cells also adhered more strongly to plastic and differed from L929 in cell shape. Similar observations were made with TNF susceptible and resistant variants of two other cell lines (RK13 and a plastic adherent U937 subline). The tendency of resistant cells to grow closely together suggests the possibility of inter-cell communication for the TNF resistant state. However, like L929 and U937, L929/R and U937/R did not communicate by gap junctions and we could find no evidence of extracellular mediators of TNF resistance. Rather the differences in colonial morphology, cell shape and plastic adherence may be secondary to an underlying mechanism which defines TNF susceptibility/resistance.     

Caffeine potentiates the lethality of tumour necrosis factor in cancer cells.

In this study we have investigated the interaction of caffeine, a prototypic methylxanthine, and TNF on the induction of cell death in mouse and human cell lines during progression from G1 to successive phases of the cell cycle. Exposure of cells to TNF (0.1-100 ng ml-1) as single agent for 48 h caused low or no lethality. The rates of cell death increased significantly when cells cultured with TNF for 24 h were exposed to caffeine (2.5-20 mM). The magnitude of the enhancement by caffeine was TNF and caffeine dose-dependent. The most effective response to this combination was observed in the mouse cell lines, WEHI and L929, followed by the human cell lines, HeLa, A375 and MCF-7, respectively. In L929 cells, TNF treatment did not inhibit DNA synthesis during the first S phase of the cell cycle (20-24 h), but it did block the progress toward a second S phase, indicating the cells were arrested at G2 phase or mitosis. Caffeine had great enhancer effect on L929 cells exposed to TNF for 24 h, but the effect was reduced in cells with either less than 24 h or greater than 28 h of exposure. L929 cells stimulated with TNF died via apoptosis, as judged by both morphological criteria and the occurrence of internucleosomal DNA cleavage. Exposure of TNF-treated cells to caffeine caused a greater increase in the proportion of apoptotic cells as well as the extent of internucleosomal DNA fragmentation.     

Antitumor Activities of IKP-104, a 4(1H)-Pyrizinone Derivative, on Cultured and Implanted Tumors

Antitumor activities of IKP-104, a 4(1 H )-pyrizinone derivative, were investigated with cultured tumor cell lines and implanted tumors in mice. IKP-104 inhibited the growth of cultured murine tumor cell lines (L1210 leukemia, Lewis lung carcinoma and B16 melanoma) and human tumor cell lines (K562 leukemia and HeLa cervical carcinoma). It also had antitumor effects on implanted murine ascitic tumors (L1210 leukemia and sarcoma 180) and a murine solid tumor (Lewis lung carcinoma). IKP-104 could be classified as a phase-dependent cytostatic drug based on the mode of growth inhibition of cultured B16 melanoma cells compared with those of several other antitumor agents. The effect of IKP-104 on the cell cycle traverse of cultured B16 melanoma cells was estimated by morphological and flow cytometric analyses. Cells accumulated in the mitotic phase, and abortive mitosis or polyploidy or multinucleation was induced from 6 h after exposure to IKP-104. Based on these results, IKP-104 is expected to be useful for the treatment of tumors, and its mode of action seemed to be similar to that of metaphase arrestants such as colchicine or vinca alkaloids.     

Antitumor activities of IKP-104, a 4(1H)-pyrizinone derivative, on cultured and implanted tumors

Antitumor activities of IKP-104, a 4(1H)-pyrizinone derivative, were investigated with cultured tumor cell lines and implanted tumors in mice. IKP-104 inhibited the growth of cultured murine tumor cell lines (L1210 leukemia, Lewis lung carcinoma and B16 melanoma) and human tumor cell lines (K562 leukemia and HeLa cervical carcinoma). It also had antitumor effects on implanted murine ascitic tumors (L1210 leukemia and sarcoma 180) and a murine solid tumor (Lewis lung carcinoma). IKP-104 could be classified as a phase-dependent cytostatic drug based on the mode of growth inhibition of cultured B16 melanoma cells compared with those of several other antitumor agents. The effect of IKP-104 on the cell cycle traverse of cultured B16 melanoma cells was estimated by morphological and flow cytometric analyses. Cells accumulated in the mitotic phase, and abortive mitosis or polyploidy or multinucleation was induced from 6 h after exposure to IKP-104. Based on these results, IKP-104 is expected to be useful for the treatment of tumors, and its mode of action seemed to be similar to that of metaphase arrestants such as colchicine or vinca alkaloids.     

Lymphokine-independent, leukemia cell-mediated induction of tumor necrosis factor in human monocytes

Unactivated human blood monocytes and monocytic THP-1 cells were found to respond to some leukemia cells by tumor necrosis factor (TNF) production. The TNF production by THP-1 cells in response to K562 cells was preceded by a rapid rise in [Ca2+]i, initiated within 1 h and terminated within 4 h as a refractory state took over. Neither the amount nor the duration of TNF production was enhanced by gamma-interferon. The P32/ISH cells did not induce a significant [Ca2+]i change of TNF production, while MOLT-4 cells failed to induce TNF despite their capacity to mobilize Ca2+ in THP-1 cells. The failure of P32/ISH or MOLT-4 to induce TNF was attributed primarily to a lack of stimulatory membrane molecules rather than to suppression by an inhibitory component, since liposomes carrying membrane components of K562 and MOLT-4 or P32/ISH in varying proportions elicited TNF production that precisely reflected the K562 proportion. The ability of K562 to induce TNF was selectively impaired by trypsin, whereas the ability to mobilize [Ca2+]i was more sensitive to glutaraldehyde, although once the latter activity was extinguished, the K562 cell could no longer induce TNF. These results suggest that some leukemia cells are equipped with two or more signaling membrane moieties which together stimulate monocytes for transient tumoricidal expression in the preimmune stage.     

Oxidative damage in murine tumor cells treated in vitro by recombinant human tumor necrosis factor

Treatment of three murine tumor cell lines, L929, P388, and Pan-02, in vitro with recombinant human tumor necrosis factor (rhTNF) produced evidence of oxidative damage as measured by (a) increases in intracellular glutathione levels, (b) the formation of intracellular oxidized glutathione and (c) the formation of thymine glycols in DNA. L929, the most sensitive of the three cell lines to the cytotoxic activity of rhTNF, had the lowest total glutathione content and was observed to have the highest levels of oxidized glutathione and thymine glycol formation. In addition, the radical buffering capacity of these cells was significantly compromised within 7 h of treatment with rhTNF. The P388 and Pan-02 cell lines, with total glutathione levels about 50-fold higher than L929, also showed evidence of oxidative attack, although to a lesser extent than L929. The radical buffering capacity of these cell lines was not altered by rhTNF treatment. A rhTNF-resistant subline of L929 (L929r), produced by successive passaging in vitro in the presence of TNF, increased its glutathione and oxidized glutathione levels in response to a subsequent rhTNF challenge. Meth A, a cell line resistant to rhTNF in vitro but not in vivo, showed no evidence of oxidative damage following rhTNF treatment, despite having a low radical scavenging capacity and a sensitivity to H2O2. The results with Meth A suggest that the interaction of rhTNF with this cell line does not occur in the same manner as the other cell lines, perhaps due to receptor differences or to some type of "uncoupling" of the signal-response network between the TNF receptor and a putative secondary messenger(s). These results are consistent with the hypothesis that: (a) the mechanism of action of rhTNF involves the production of oxidative damage, including damage to the DNA; (b) the sensitivity to rhTNF in vitro is related to the radical scavenging capacity of the cell; and (c) cells can respond to rhTNF challenge by increasing their free radical scavenging capacity.     

K562 erythroleukemic cells are equipped with multiple mechanisms of resistance to lysis by complement

Resistance of tumor cells to lysis by complement is generally attributed to several protective mechanisms. The relative impact of these mechanisms in the same tumor cell, however, has not been assessed yet. We have analyzed the interaction of the human erythroleukemia tumor cell line K562 with human complement. K562 cells express the membrane complement regulatory proteins CD59, CD55 and CD46. As shown here for the first time, K562 also spontaneously release the soluble regulators C1 inhibitor, factor H, and soluble CD59. Complement resistance of K562 cells is augmented upon treatment with PMA, TNF or even with sublytic complement. Unlike TNF and sublytic complement, PMA enhanced the expression of membrane-bound CD55 and CD59 and led to increased secretion of soluble CD59. In addition, we show that complement-resistant K562 cells express a membrane-associated proteolytic activity, higher than the complement-sensitive K562/S cells. Treatment of complement-resistant K562 cells with serine protease inhibitors enhance their sensitivity to complement-mediated lysis. Inhibitors of protein kinase C (PKC) also sensitize K562 cells to complement lysis, implicating PKC-mediated signaling in cell resistance to complement. Neutralization of the CD55 and CD59 but not of CD46 regulatory activity with specific antibodies significantly increases complement-mediated K562 cell lysis. Treatment of K562 cells with a mixture of inhibitory reagents results in a significant additive enhancing effect on complement-mediated lysis of K562. In conclusion, K562 cells resist a complement attack by concomitantly using multiple molecular evasion strategies. Future attempts in antibody-based tumor therapy should include strategies to interfere with those resistance mechanisms.     

Characterization of a Novel Human Tumor Necrosis Factor-α Mutant with Increased Cytotoxic Activity

Various novel recombinant human tumor necrosis factor-α (TNF) mutants were prepared using protein engineering techniques, and their cytotoxic activity was compared with that of the intact form of TNF (intact TNF). Mutant 471 (a TNF mutant molecule with the deletion of 7 amino acids at the amino-terminal and the substitution of Pro⁸Ser⁹Asp¹⁰ by ArgLysArg) had a 6-fold higher cytotoxic activity against murine L929 cells. The mutant TNF had an increased ability to bind to TNF receptor on murine L929 fibroblasts cells. A cross-linking study revealed that mutant 471 had an increased ability to form an active trimer. Mutant 471 also showed higher cytotoxic activity against human KYM myosarcoma cells and human MIA PaCa-2 pancreatic carcinoma cells. The possible cachectin activity of the mutant was almost the same as that of intact TNF. These results suggest that mutant 471 might be a more promising candidate as an anticancer agent than intact TNF.     

Tumor degeneration by human embryonic fibroblasts and its enhancement by interferon

KB cells from a human nasopharyngeal tumor were cocultivated with human embryonic fibroblasts (HF 8101 cells); 7 to 14 days after incubation, "spongy degeneration"-like changes developed in the target cell-growing area. These changes developed in other target cells [HeLa cells from human cervical cancer, human hepatoma cells (PLC/PRF/5), and human amnion FL cells] cocultured with several kinds of human embryonic fibroblasts [HF 8101 cells, HF 8103 cells, and HEL cells]; however, HF 8101 cells did not cause degenerative changes in murine L929 cells. The degenerative changes were enhanced by treatment with human leukocyte interferon or human fibroblast interferon at a dose of 1,000 or 10,000 IU/ml, but there was no significant difference in the enhancing effect between human leukocyte and human fibroblast interferons. Mouse L929 interferon did not enhance the degenerative changes in KB cells caused by HF 8101 cells. It was concluded that human fibroblasts caused the degenerative changes in the human tumor cells and the continuous cell line and that the changes were enhanced by treatment with either human leukocyte interferon or human fibroblast interferon.     

Suppressive effects of intracellular glutathione on hydroxyl radical production induced by tumor necrosis factor

Protective effects of intracellular glutathione (GSH) against the cytotoxicity of human recombinant tumor necrosis factor (TNF) were investigated. Three tumor cell lines (L-M, B-16, HeLa) were used as target cells. Exposure of these cells to buthionine sulfoximine (BSO) or diethyl maleate (DEM) resulted in the depletion of intracellular GSH content to 5.2-43.0% of control values and enhancement of their susceptibility to TNF cytotoxicity. The hydroxyl radical production in L-M cells stimulated by TNF was increased by treatment with BSO or DEM. These results are consistent with the suggestion that intracellular GSH exerts its protective function against the cytocidal effect of TNF by inhibiting the hydroxyl radical production stimulated by TNF.     

Changes in cellular lipid synthesis of normal and neoplastic cells during cytolysis induced by alkyl lysophospholipid analogues

Susceptibility of eight different cell types of murine or human origins to alkyl lysophospholipid analogue (ALP)-induced cytolysis correlated well with a selective, dose-dependent inhibition of radiolabeled oleic acid incorporation into phosphatidylcholine (PC) and a concomitant stimulation of incorporation into neutral lipids (NL), mainly triacylglycerols. In resistant cells (murine macrophages, L929S, K562, and rMeth A) a counts per minute NL/counts per minute PC ratio of 0.8-1.0 was observed with 30 micrograms ALP/ml; in sensitive tumor targets (Meth A, HL60, YAC, and ABLS-8.1) values greater than 2.7 were found with 5-10 micrograms ALP/ml. Changes in lipid metabolism preceded cytolysis in Meth A fibrosarcoma cells. In degradation experiments the percentage of total lipid radioactivity in PC was reduced after 24 hours to 47% compared to that in controls in sensitive Meth A with 10 micrograms ALP/ml. The macrophage-PC was unaffected at the same concentration. Sensitivity to ALP was independent of cell proliferation. Resistance was not restricted to normal cells and was inducible in Meth A (and rMeth A).     

Characterization of human tumor necrosis factor produced by peripheral blood monocytes and its separation from lymphotoxin

Cultures of human peripheral blood leukocytes (PBL) induced with phytohemagglutinin (PHA) and the phorbol ester 12-0-tetradecanoylphorbol 13-acetate (TPA) produced two types of cytotoxic proteins, indistinguishable in the in vitro assay employing murine L 929 cells as targets. One of these proteins had the antigenic and physicochemical properties of lymphotoxin (LT). We have identified the other cytotoxin as tumor necrosis factor (TNF), mainly on the basis of antigenic cross-reactivity demonstrated with antiserum to TNF, and also by its characteristic physicochemical properties and cell source. Unlike LT, PBL-derived TNF did not bind to Concanavalin A-Sepharose or to several other agglutinin-Sepharose columns specific for carbohydrate moieties common in glycoproteins. The molecular weight of native TNF determined by gel filtration was approximately 40,000 while SDS-PAGE revealed a single sharp peak of 16,500 +/- 500. When cultures of monocytes and lymphocytes separated by elutriation were stimulated with PHA and/or TPA, monocytes were the major source of TNF. In contrast, only lymphocytes produced LT. A mixture of antisera to TNF and LF neutralized all cytotoxicity of crude human lymphokine preparations for L 929 cells, suggesting that TNF and LT are either the only, or the major, cytotoxic proteins present in such crude lymphokine preparations demonstrable in this assay.     

Natural cytotoxicity of human blood lymphocytes and monocytes and their cytotoxic factors: effect of interferon on target cell susceptibility

The effect of interferon (IFN) on target cell susceptibility to human natural killer (NK) cells and monocytes was analyzed in direct cell-mediated and their cytotoxic factor-mediated cytotoxicity assays. Treatment of K562 cells with IFN resulted in a decrease in their sensitivity to lysis by nonadherent lymphocytes and Percoll-purified large granular lymphocytes (LGL) when tested in a 4-hour 51Cr release assay. In contrast, the treatment did not affect the target susceptibility to monocytes purified by adherence to autologous serum-coated plastic surfaces. In the target-binding assay with LGL or monocytes the number of conjugates was not altered after IFN treatment of K562. Lymphocytes and monocytes were induced to release soluble cytotoxic factors, termed "natural killer cytotoxic factors (NKCF) and monocyte cytotoxic factors (MCF)," respectively, when co-cultured with K562. Both NKCF and MCF lysed K562 in a 48-hour microcytotoxicity assay or in an 18-hour 51Cr release assay in the presence of dactinomycin. IFN-treated K562 reduced or completely lost their ability to stimulate the release of NKCF, whereas they triggered MCF secretion as effectively as did the untreated K562. When lymphocytes or monocytes were pretreated with IFN, they released NKCF or MCF with augmented lytic activity. In contrast to the sensitivity to NK cell-mediated lysis, IFN pretreatment of K562 induced no change in their susceptibility to NKCF and MCF. When IFN was added to NKCF and MCF assays, the cytotoxicity was enhanced. The addition of IFN to K562 that had been pretreated with NKCF or MCF and washed resulted in no increase in lysis. The capacity of K562 to absorb the lytic activity of NKCF and MCF was not altered by IFN. These results indicate that IFN treatment of target cells can be used to distinguish the two distinct types of blood mononuclear cells with natural cytotoxicity, NK cells and monocytes, and that each effector cell type is stimulated to release cytotoxic factors by the different target determinant after the initial effector-target cell binding.