Participation of polymorphonuclear leukocyte-derived factor in murine tumour cell killing.

Previous studies showed that murine polymorphonuclear leukocytes (PMNs) lyze tumour cells in the presence of wheat germ agglutinin or actinomycin D. This paper reports studies on whether a soluble factor participates in PMN-mediated cytolysis dependent on lectin or a chemotherapeutic drug. Tumour lysis was observed with supernatants from PMN cocultured with wheat germ agglutinin or actinomycin D. The supernatant from cultures of PMNs alone was not cytotoxic, but addition of these agents to the supernatant induced tumour lysis. PMNs released a soluble factor spontaneously into the medium and cytolysis was induced by a combination of this factor and wheat germ agglutinin or actinomycin D. This factor was not an oxygen metabolite, but a protein with a molecular weight of approximately 100 K daltons. These results suggest that a soluble factor(s) from PMNs participates in tumour killing in cooperation with appropriate reagents.

Previously, we showed that PMNs from the peritoneal cavity of mice could kill murine tumour cells in vitro on addition of appropriate mediators; viz plant lectins , animal lectins  antitumour antibody , anticancer chemotherapeutic drugs (Ikenami et al., 1985) and immunomodulators (Morikawa et al., 1985). Reactive oxygen species produced by PMNs are important in the lytic process by immunomodulators (Morikawa et al., 1985), but the mechanisms of other types of killing are unknown. In this work, we investigated the mechanisms of PMN-mediated cytolysis dependent on lectin and chemotherapeutic drug, by studies on whether soluble factor from PMNs participates in tumour killing in vitro. We found that a PMN-derived factor can lyse tumour cells in cooperation with wheat germ agglutinin or actinomycin D and that this factor is a protein of high molecular weight. from Shizuoka Experimental Animal Farm (Shizuoka, Japan). Mice were used at 8-11 weeks of age.

Materials and methods
Tumour cells MM46, a transplatable ascites tumour from a spontaneous mammary adenocarcinoma in a C3H/He mouse, was mainly used as a target cell. MM48 mammary adenocarcinoma and MH 134 hepatoma cells were also used as target cells. L929 cells were harvested from in vitro culture.
Polymorphonuclear leukocytes (PMNs) Cells were prepared as described previously . Briefly, 2 ml of 12% casein solution was injected into the peritoneal cavity of mice and the peritoneal exudate was harvested 6h later, passed through nylon mesh and centrifuged at 300g for 5min. The precipitated cells were washed twice with RPMI-1640 medium (Nissui Seiyaku Co., Tokyo) supplemented with 10OUml-m of penicillin (Banyu Pharmaceutical Co., Tokyo) and 100gml-l of streptomycin (Meiji Seika Co., Tokyo). Usually, the peritoneal cells were suspended in RPMI-1640 medium containing 5% heat-inactivated foetal calf serum (Gibco, Grand Island, NY; called medium hereafter). They were stained with Giemsa stain and the proportions of PMNs were determined by morphological observation. These peritoneal cells, containing 93-98% of polymorphonuclear leukocytes, were used as the polymorphonuclear leukocyte preparation. About 108 PMNs were obtained from a C3H/He mouse.
Cytolytic assay Cytolysis of MM46 tumour cells was assayed as described previously (Yamazaki et al., 1975). Briefly, PMN-culture supernatants and 5 Cr-labelled MM46 tumour cells (5 x 103 cells) were mixed in wells (7mm diameter) of flat-bottomed microplates. The mixture were incubated in 0.2 ml of medium for 18-24 h at 37°C under 5% CO2 in air and the radioactivity of the supernatant was measured. Cytolytic activity was calculated as follows: Cytolysis (%) = Experimental count -control count Maximum releasable count-control count Maximum release of 51Cr was measured by freezing-thawing labelled tumour cells 3 times. The control count was measured as the radioactivity released from labelled cells in the presence of wheat germ agglutinin or actinomycin D without culture supernatant. The control count was usually equivalent to the count released spontaneously from labelled cells alone.
Cytolysis of L929 cells was measured by the method of Ruff & Gifford (1980). Briefly, L929 cells (8 x 104 cells) and PMN-culture supernatants were mixed in the wells (7mm diameter) of flatbottomed microplates, and incubated in a CO2 incubator for 18 h. Then, the medium was removed and residual cells were stained for 15 min with crystal violet. After addition of 0.1 ml of sodium dodecyl sulfate (0.5%), absorbance at 590 nm of the supernatant was measured in a photometer (Myreader 7, Sanko Junyaku Co., Ltd., Tokyo). Cytolytic activity was calculated as follows: Cytolysis (%) = 1 Experimental absorbance 0 Control absorbance Cytolysis in Marbrook vessels Marbrook vessels with 2 chambers separated by a nuclepore membrane (Pore size, 0.4pm; thickness lOpm; Nuclepore Co., Pleasanton, CA) were prepared. PMNs (1.2 x 107) were introduced into the outer chamber, and 3 x 104 51Cr-labelled MM46 tumour cells were into the inner chamber. The cells were incubated in 2ml of medium with or without wheat germ agglutinin (30 pg ml -') at 370C for 24 h. As controls, 5'Cr-labelled MM46 tumour cells were placed in the outer chamber with PMNs. After incubation, the radioactivities of the supernatants of the inner and outer chambers were measured.

Reagents
Wheat germ agglutinin was purchased from E-Y Laboratories (San Mateo, CA). Actinomycin D, catalase, arginine, trypsin and soybean trypsin inhibitor were obtained from Sigma Chemical Co., (St. Louis, MO). Leupeptin and bestatin were gifts from Dr T. Takeuchi (Institute of Microbiol Chemistry, Tokyo).

Gel filtration
The PMN-culture supernatant was applied to a column of Sephacryl S-300 (Pharmacia Fine Chemicals, Sweden) previously equilibrated with phosphate-buffered saline (pH 7.4) and material was eluted with the same buffer. Fractions of 10ml were collected and their cytolytic activity was tested in the presence of wheat germ agglutinin or actinomycin D.

Cytolytic activity of supernatants of cultures of PMNs
We used Marbrook vessels to study the role cell-tocell interaction in cytolysis, i.e. whether effectortarget cell interaction is necessary for cytolysis. When effector PMNs and target tumour cells were incubated separately in different chambers, cytolysis was observed in the presence of wheat germ agglutinin just as when both types of cells were incubated in the same chamber ( Figure 1) unlabelled tumour cells were killed by PMNs in th outer chamber in the presence of wheat gern agglutinin, 51Cr-labelled tumour cells in the inne chamber were also killed. Similar results were obtained in actinomycin dependent PMN-mediated cytolysis: cell-fre supernatants from cocultures of PMNs wit] actinomycin D lysed tumour cells (Figure 2).
These results suggested that direct contac between effector PMNs and target tumour cells wa not essential for tumour lysis and that the lysi involved a soluble factor(s) released into th medium.  Figure 3 shows the dose-response curves of culture supernatants from PMNs with wheat germ agglutinin and actinomycin D respectively. Cytolytic activities were detected with up to 4-fold dilutions of both supernatants.

Characterization of supernatants from cultures of PMNs Is
The specificity of cytolysis was examined with Is several kinds of target cells. Table I shows that 3  e other syngeneic tumour cells, MM48, MH134 and L929 cells, were lyzed by the supernatant of cultures of PMNs with wheat germ agglutinin. However, no cytolysis of normal spleen cells was observed.
Next, we examined whether proteases, arginase and oxygen metabolites act as lytic substances in cytolysis by the supernatants. For examination of the effects of proteases the following inhibitors were used: soybean trypsin inhibitor against trypsin, leupeptin against plasmin, trypsin and papain, and bestatin against aminopeptidase B and leucine aminopeptidase. For examination of the effects of oxygen metabolities, such as H202, catalase was used as a scavenger of these substances. As shown in Table II,   'Supematants were obtained after 5 h cultures of PMNs with or without wheat germ agglutinin (50gml-1); bCytolysis of tumour cells was assayed by 51Cr release method. Cytolysis of spleen cells was determined by the dye exclusion test. Cytolysis (35%) by medium alone was subtracted as a background value. Mean + s.d. (n = 3). 'Various inhibitors were added to the culture supernatant before its cytolytic activity was assayed; bSupernatants were obtained after 5 h-cocultures of PMNs with actinomycin D (0.5pgmln l) or wheat germ agglutinin (100 gml-1); * =significant difference (P< 0.05).
supernatants. Table II shows that the cytolytic activity was lost on heating at 700 for 1 h and on trypsin treatment.
Spontaneous release of the soluble factor from PMNs As described above, supernatants from PMNs cocultured with wheat germ agglutinin or actinomycin D were cytolytic, but supernatants from PMNs cultured alone were not. Next, we examined whether these reagents were required to induce a factor from PMNs i.e., whether addition of these reagents to supernatants from PMNs cultured alone induced tumour lysis as well as supernatants from PMNs cocultured with these reagents.
As shown in Table III, tumour lysis was induced by addition of wheat germ agglutinin or actinomycin D to supernatants of PMN cultures. The supernatant alone was not cytolytic to tumour cells. Thus, cytolysis seemed to be induced by combination of a factor from PMNs and these reagents, and PMNs seemed to release this factor spontaneously into the medium.
The kinetics of release of this factor is shown in Figure 4. Maximum release was observed within 5 h, and no cytolytic activity was detected in supernatants of overnight cultures. Therefore, this factor may be released spontaneously from fresh PMNs but not from damaged or dead PMNs.
Next, the nature of the factor that induced cytolysis in cooperation with wheat germ agglutinin or actinomycin D was examined by subjecting the  Figure 4 Time course of spontaneous release of PMN-factor. PMNs (2 x 10' cells) were cultured alone in PBS (lml) and supernatants were obtained at the indicated times. Medium containing 50% of PMNculture supernatant and 5"Cr-labelled MM46 tumour cells (5 x 103 cells) was incubated for 24h with actinomycin D (0.3 ygml-1) (A), or wheat germ agglutinin (50 yg ml-) (0). Bars indicate sd (n= 3). supernatants from serum free-cultures of PMNs to gel filtration on Sephacryl S-300. As shown in Figure 5, the cytolytic activity was recovered in a fraction corresponding to a mol. wt of about -100 K daltons.

Discussion
Previously we observed mediator-dependent cytolysis when murine syngeneic tumour cells were lyzed by casein-induced peritoneal PMNs in the Fraction number (10 ml/fr.) Figure 5 Gel filtration of PMN-culture supernatants. PMNs (2 x 107 cellsml-1) were cultured for 5h in 23 ml PBS. Concentrated supernatants (3 ml) were applied to a Sephacryl S-300 column (1.8 x 100cm) and fraction of 10ml were collected. Medium containing 50% of eluant and L929 cells ( presence of lectins Yamazaki et al., 1983), anti-cancer chemotherapeutic drugs (Ikenami et al., 1985) or immunomodulators (Morikawa et al., 1985a). In the present work, we examined the mec4anisms of lectin-drugs-dependent PMNmediated cytolysis. We found that supernatants from PMNs cocultured with wheat germ agglutinin or actinomycin D could lyse tumour cells (Figure 1 and 2). As far as we know, tumour lysis by the supernatant of PMN cultures has not been reported previously. We also observed tumour lysis on addition of wheat germ agglutinin or actinomycin D to the supernatant of PMNs cultured alone (Table III). However, the factor itself was not cytolytic to tumour cells. Thus, cytolysis seemed to be induced by the action of the factor with these reagents.
Moreover, this factor was released spontaneously from glass-nonadherent cells, but not from glassadheremt macrophage-rich cells, and PMNs could not lyse tumour cells in the presence of lipopolysaccharide, which is known to activate macrophages (data not shown). Therefore, the PMN-mediated cytolysis was not due to contaminating macrophages.
Most previous studies on PMN cytotoxicity have focussed on oxygen metabolities. PMNs have been shown to kill tumour cells through oxygendependent pathways (Clark & Klebanoff, 1979: Nathan et al., 1979Dallegri et al., 1983). We also reported that hydrogen peroxide was an effector molecule in immunomodulator-dependent PMNmediated tumour lysis (Morikawa et al., 1985b). However, the present factor was different from oxygen metabolites such as hydrogen peroxide (Table II). This factor seemed to have a mol. wt of about -100 K daltons ( Figure 5), and to be a protein, since it was heat-labile and inactivated by trypsin (Table II). The activity was not inhibited by protease inhibitors and arginine. These data suggest that the factor is neither protease nor arginase, which are known to be effector molecules of macrophages. A lysosomal cationic protein from PMNs that induces cytolysis has been reported (Thorne et al., 1984), but this seemed to differ from our factor in mol. wt and its direct cytotoxicity on target cells.
This factor was released spontaneously from fresh PMNs (Figure 4). From this finding and the fact that PMN-mediated cytolysis was inhibited by an inhibitor of protein synthesis, cycloheximide (data not shown), we conclude that damaged or dead PMNs do not release the factor, but rather that PMNs may die after production of this factor. Spontaneous productions of cytotoxins by alveolar macrophages (Sone et al., 1984) and a macrophage-like cell line (Kull & Cuatrecases, 1984) have been reported. Our results suggest that this may also be the case with PMNs: exudate PMNs may secret a factor that participates in target killing at an inflammatory site in vivo. In fact, recently we found a similar soluble factor to that reported here in inflammatory ascites containing many PMNs (manuscript in preparation). At present the primary action of this factor on target cells and its synergistic actions with wheat germ agglutinin and actinomycin D are not clear. We are now purifying and characterizing the factor further.