Immunology 1990 70 116-120
Indirect inhibition of generation of murine lymphokine-activated killer cell activity in splenocyte cultures by interferon-gamma T.-Y. CHAO, H. OHNISHI & T. M. CHU Department of Diagnostic Immunology Research and Biochemistry, Roswell Park Memorial Institute, New York State Department of Health, Buffalo, New York, U.S.A.
Acceptedfor publication 5 January 1990
SUMMARY (rIFN-y) on murine lymphokine-activated killer interferon-gamma recombinant The effect of mouse (LAK) cell activity was investigated using natural killer (NK)-resistant, spontaneously developed, weakly immunogenic and highly tumourigenic, syngeneic murine mammary adenocarcinoma, JC, mimicking that of human disease, as the target. Murine YAC-1 also was used as a target cell line. rIFN-y, when used in combination with recombinant interleukin-2 (rIL-2), was shown to exhibit a suppressed effect on LAK cell activity generated from BALB/c mouse splenocytes, compared to that with rIL-2 alone. The decrease in LAK cell activity was rIFN-y dose-dependent. Kinetic study revealed that this inhibitory effect was demonstrated only when rIFN-y was added to the medium at the early phase of rIL-2 culture. The inhibitory effect on LAK cell generation by rIFN-y was completely abrogated when the nylon-wool-treated non-adherent 'macrophage-free' splenocytes were incubated with rIL-2 and rIFN-y. These results indicated that the LAK cell activity generated from murine splenocytes cultured with rIL-2 could be depressed by rIFN-y, and that the macrophages may be involved as mediators. INTRODUCTION The induction of lymphokine-activated killer (LAK) cell activity by recombinant interleukin-2 (rIL-2) from lymphoid cells exhibiting cytotoxicity against a broad spectrum of fresh tumour cells and cultured tumour cell lines has been well demonstrated (Rosenstein et al., 1984; Ramsdell & Golub, 1987; Vujanovic, Herberman & Hiserodt, 1988). IL-2-induced LAK cells exhibiting such a non-major histocompatibility complex (MHC)-restricted cytotoxicity have been used successfully in experimental adoptive immunotherapy against animal tumours (Mazumder & Rosenberg, 1984; Ottow et al., 1987; Takai et al., 1988). Although LAK cell activity was generated primarily from natural killer (NK) cells and T cells (Suzuki et al., 1983; Yang, Mule & Rosenberg, 1986; Kalland et al., 1987), other immune cells such as monocytes and macrophages have been reported to exhibit regulatory function on LAK cell activity (Ibayashi, Hoon & Golub, 1987; Nii et al., 1988; Silvennoinen, Vakkila & Hurme, 1988). An alteration of the generation of LAK cell activity is thus possible through the modulation of the interactions among these cells, which may provide insight into the Abbreviations: JC, a spontaneously developed syngeneic murine adenocarcinoma; LAK, lymphokine-activated killer; NK, natural killer; rIFN-y, mouse recombinant interferon-gamma; rIL-2, human recombinant interleukin-2. Correspondence: Dr T. M. Chu, Dept. of Diagnostic Immunology Research and Biochemistry, Roswell Park Memorial Institute, 666 Elm Street, Buffalo, NY 14263, U.S.A.
mammary
mechanisms of action by LAK cells, with the use of other agents. Interferon-gamma (rIFN-y), a potent lymphokine modulating some immune response, has been reported to be able to augment NK activity and to exhibit synergistic effects with rIL-2 on the generation of human LAK cell activity (Svedersky et al., 1984; Itoh et al., 1985). However, a negative effect by rIFN-y on LAK cell activity has also been reported (Hoyer et al., 1986; Sayers, Mason & Ortaldo, 1986; Giovarelli et al., 1988). A spontaneously developed syngeneic murine mammary adenocarcinoma cell line, designated JC, has been established that is resistant to NK cell but susceptible to LAK cell lysis (Capone, Kadohama & Chu, 1987; Chao & Chu, 1989). Because it is weakly immunogenic and highly tumourigenic, the major biologic characteristics mimicking that ofhuman disease, JC is a more suitable model for experimental immunotherapy than other chemically or virally induced mammary tumour models. Using JC and YAC-1 as the targets, the effect of rIFN-y on murine LAK cell activity in vitro has been investigated. The results indicated that rIFN-y was capable of exerting an inhibitory effect on the generation of rIL-2-induced LAK cell activity from BALB/c mouse splenocytes. The suppressed effect was completely abrogated by treatment of whole splenocytes with nylon-wool, suggesting a potential role played by the macrophages in the decrease in LAK cell activity by rIFN-y. MATERIALS AND METHODS Mice
Eight- to 12-week-old BALB/c female mice were used. The
116
117
Inhibition of LAK activity by rIFN-y BALB/c mice were maintained at West Seneca Laboratory, Roswell Park Memorial Institute. Histocompatibility test by skin grafting was performed routinely to ascertain the maintenance of the line.
applied to the column and incubated at 37°, 5% Co2, for 45 min. Non-adherent cells were collected by washing the column with 30 ml of prewarmed medium (I drop/second), centrifuged (200 g, 10 min) and then resuspended in RPMI-1640-10% FCS.
Target cells YAC-l, a NK-sensitive cell line of A/Sn mouse origin, was obtained from the American Type Culture Collection (Rockville, MD). JC tumour line was established from a spontaneously developed mammary adenocarcinoma in an aged virgin female BALB/c mouse (Capone et al., 1987). JC is a NKresistant cell line, but susceptible to LAK cell lysis (Chao & Chu, 1989). These cell lines were maintained in RPMI-1640 with 10% fetal calf serum (FCS) (Gibco, Grand Island, NY).
In vitro cytotoxic assay NK and LAK cell activities were measured by using a standard 4-hr 5"Cr-release microcytotoxicity assay. Target cells (2 x 106) were harvested and washed twice, and then incubated in 50 pCi of 51Cr (Na5lCrO4, 402-61-615 67 mCi/mg, New England Nuclear, Boston, MA) for 2 hr at 370 in a moist atmosphere with 5% CO2. The labelled target cells were washed three times with RPMI-1640-10% FCS. The effector cells and 5'Cr-labelled target cells (5 x 103) were then placed in 96-well U-bottomed microtitre plates (Falcon, Oxnard, CA) with various effector to target cell ratios in a final volume of 0-2 ml in RPMI-1640-10% FCS. After a 4-hr incubation period at 370 in a humidified 5% CO2 atmosphere, supernatant was collected and counted in a gamma-counter (Packard Multi-Prias, Downers Grove, IL) for experimental release (ER). For the spontaneous release (SR) control, a sample of labelled target cells was resuspended in RPMI-1640-10% FCS alone. Total release (TR) of 51Cr from target cells was obtained by mixing 5 x I03 labelled target cells in 0-1 ml of RPMI-1640-10% FCS with 0-1 ml 10% SDS. The spontaneous release was generally 10-20% of the total release. Percentage specific release for 5'Cr release was calculated by using the following equation:
Recombinant IL-2 and IFN-y Human rIL-2, produced from Escherichia coli, was kindly supplied by Cetus Corporation (Emeryville, CA), with a specific activity of 18 x 106 IU/mg. Endotoxin level was less than 0-01 ng/ml. The purity was >98% by sodium dodecyl sulphatepolyacrylamide gel electrophoresis (SDS-PAGE) (Rosenberg et al., 1983; Wang, Lu & Mark, 1984). Mouse rIFN-y was provided by Schering Corporation, Kenilworth, NJ, as part of the American Cancer Society's programme on IFN, with a specific activity of 106 U/mg. The purity was >98% SDS-PAGE. Generation of LAK cells The spleens from normal BALB/c mice were harvested asceptically and minced in a stainless steel mesh to produce a single cell suspension. The erythrocytes were lysed osmotically with Trisammonium chloride buffer, 0-17 M, pH 7-2. The remaining splenocytes were either washed twice with RPMI-1640, or passed through a nylon-wool column before washed, and then incubated at 370 in complete medium containing rIL-2, under a moist atmosphere with 5% CO2, in culture flasks, at a cell concentration of 1-25 x 106/ml. Complete medium contained RPMI-1640 with 10% FCS, 0-1 mm non-essential amino acids (Gibco), 0-1 mm sodium pyruvate (Gibco), 5 x I0-I M 2mercaptoethanol (Sigma, St Louis, MO), 0-3 mg/ml L-glutamine (Gibco), 100 pg/ml streptomycin (Gibco) and 100 U/ml penicillin (Gibco). LAK cells were harvested after culture for various periods of time, washed twice, and the viable cells were counted and resuspended in RPMI-1640-10% FCS for the standard 4 hr 5"Cr-release assay. Generation of LAK cells in the presence of rIFN-y The splenocytes were prepared as described above. rIFN-y was added to the complete medium containing rIL-2. The concentrations of rIFN-y were adjusted from 0-01 to 1000 U/ml, and were evaluated either alone or with rIL-2 at 60, 600 and 3000 IU/ ml. After incubation for various periods of time, the activated splenocytes were harvested, washed and used as effector cells in 5'Cr-release assay.
Nylon-wool treatment Approximately 0-6 g of nylon-wool (Polysciences Inc., Warrington, PA) was packed into a syringe (10 ml) and incubated with RPMI-1640-10% FCS for 1 hr at 37°. After washing the column with 10 ml ofprewarmed RPMI-1640-10% FCS, 2 ml of whole splenocytes (107/ml) in RPMI-1640-10% FCS were
percentage specific release (% cytotoicity)=
ER-SRX 100 TR-SR x10 Each cytotoxic assay was carried out in triplicate and the results were expressed as lytic units. Lytic units were calculated from linear regression plotting from various effector to target cell ratios, by a MacIntosh SE computer, as described elsewhere (Pross et al., 1981). One lytic unit was defined as the number of effector cells required to cause 30% specific 51Cr release from 104 target cells and was expressed as LU/ 107 effector cells. Statistical analysis was performed with Student's t-test. RESULTS Effect of rIFN-y on rIL-2 induced murine LAK cell activity Two target cell lines, YAC- 1, a commonly known NK-sensitive cell line of A/Sn mouse origin, and JC, a NK-resistant but LAKsensitive spontaneously developed mammary adenocarcinoma cell line of BALB/c mouse origin, were used as the indicators of cytotoxicity as measured by 5"Cr-release assay. Freshly prepared normal BALB/c mouse splenocytes showed only marginal, if any, NK cell activity in the standard 4 hr 51Cr-release assay. The culture of the normal BALB/c mouse splenocytes in complete medium in the presence of rIFN-y, 10, 100 and 1000 U/ml, for 3 days produced no change in cytotoxicity against either YAC-l or JC cells (data not shown). In concert with the use of rIL-2, rIFN-y was found to demonstrate no augmentation on LAK cell activity. On the contrary, the addition of rIFN-y to rIL-2 containing complete medium decreased LAK cell activity. This inhibitory effect of rIFN-y on the LAK cell activities directed against both YAC-l
118
T.- Y. Chao, H. Ohnishi & T. M. Chu 60 50 C) -i 40
iz rlL-2 alone * rlL-2+ IFN --y
120 100 C
0.00
150 125 100 .2 75 aJ 50 25
80 0
c
30 J- 20 10
60
m
42 O._
___o
140
>1
L 0MI 0-1 10 100 IFN-T (U/ml) I
0
YAC-1
2
3
4
60 50 40 30 20 10 0
JC
|,LI ~~ 2
3
4
Culture (days)
Figure 1. The dose-dependent suppressive effect of rIFN-y on LAK cell activity. LAK cell activity was generated from co-culturing BALB/c mouse splenocytes (1-25 x 106/ml) with rIL-2 (3000 IU/ml) for 3 days. rIFN-y at the concentrations shown was added to complete medium at the beginning of rIL-2 culture. LAK cell activity was determined by the standard 4 hr 51Cr-release assay with the use of YAC- 1 and JC tumour cells as the targets, and converted to LU. Lytic units ofcytotoxic activity were calculated from linear regression plotted from various effector: target cell ratios (100: 1, 50: 1, 25: 1 and 12: 1). One lytic unit was defined as the number of effector cells required to cause 30% specific Cr release from 104 target cells. Each point represents the mean of triplicate assays. 5
and JC tumour cells was found to be dose-dependent (Fig. 1). As shown in Fig. I, a concentration as low as 1 U/ml was effective in significantly inhibiting rIL-2 (3000 IU/ml)-induced LAK cell activity directed against JC tumour cells (mean + SEM, 45 5+5-9 versus 31-6+ 3 1, PJ
25 0
L, rIL-2 Iday 2day 3doy alone With IFN-y
Figure 3. The length of rIFN-y incubation affecting LAK activity. LAK cell activity was generated from co-culturing BALB/c mouse splenocytes (1-25 x 106/ml) with rIL-2 (3000 IU/ml) for 3 days, and was determined by the standard 4 hr 5'Cr-release assay with the use of JC cells as the target and expressed as lytic units, as described in Fig. 1. rIFN-y (100 IU/ml) was added to the rIL-2-containing complete medium, 1, 2 and 3 days prior to the 5"Cr-release assay. Each block and bar represent the mean+SEM of triplicate assays. A statistically significant difference in LAK cell activity was found between the splenocytes cultured in rIL-2 alone and rIL-2 plus rIFN-y for 2 and 3 days (P= 0-02 and 0 005, respectively).
medium 24 hr prior to the 5'Cr-release assay produced no significant inhibitory effect on LAK cell activity (51-5+6-8, P=0-7). Effect of rIFN-y on murine LAK cell activity generated from nylon-wool non-adherent splenocytes The normal mouse splenocytes in single cell suspension were passed through a nylon-wool column, and the non-adherent cells were collected and cultured in complete medium with rIL-2 and rIFN-y for 3 days. As shown in Fig. 4, after incubation with rIL-2 (3000 IU/ml) and rIFN-y (100 U/ml) for 3 days, the LAK
Inhibition of LAK activity by rIFN-y rIL-2 alone * rlL-2+ IFN-y 250 S
200
c u
150
3
100 50
4\
e
&
Figure 4. The effect of nylon-wool treatment on LAK cell activity generated with rIL-2 and rIL-2 plus rIFN-y. LAK cell activity was generated by co-culturing either whole spleen cells (1-25 x 106/ml) or nylon-wool non-adherent spleen cells (1-25 x 106/ml) with rIL-2 (3000 IU/ml), and rIL-2 plus rIFN-y (100 U/ml) for 3 days. The cytotoxicity was determined by the standard 4 hr 51Cr-release assay using JC cells as the targets and expressed as lytic units, as described in Fig. 1. Each block and bar represent the mean + SEM of triplicate assays. The difference in LAK cell activity between cultures with rIL-2 and rIL-2 plus rIFN-y was statistically significant in whole splenocytes (YAC-1, P=0009; JC, P= 0 02), and not significant in nylon-wool-treated non-adherent splenocytes (YAC-1, P=0-85; JC P=0 13).
cell activity directed against both YAC-1 and JC cells as the targets and generated from the whole splenocytes was significantly less than that after incubation with rIL-2 alone (YAC-1 cells, 39 3+3 9 versus 140-4+13-3, P=0 009; JC cells, 13-9+2-1 versus 45-4+5.9, P=0-02), confirming the results shown in Figs 2 and 3. Also shown is that rIFN-y exhibited no effect on the rIL-2-induced LAK cell activity generated from the nylon-wool-treated non-adherent 'macrophage-free or poor' splenocytes, since a similar cytotoxicity was obtained from the culture incubated with rIL-2 alone (YAC-1 cells, 155-6+28-2 versus 152-3+28-5, P=0-85; JC cells, 97-3+18-8 versus 77-3 + 10-3, P=0-13).
DISCUSSION In the present report, the effect of rIFN-y on murine LAK cell activity was investigated using a NK-sensitive cell line (YAC- 1) and a NK-resistant yet LAK-sensitive cell line (JC) as the targets. The data indicated that rIFN-y alone was incapable of enhancing the cytotoxicity of normal BALB/c mouse splenocytes. However, when used in combination with rIL-2, rIFN-y was found to suppress the rIL-2-induced LAK activity. Additional experiments using nylon-wool to remove the macrophages (and B lymphocytes) from the whole splenocytes revealed that the 'down-regulation' of LAK cell activity by rIFN-y was completely abrogated when the 'macrophage-poor or free' splenocyte preparations were used as the effector cells. The underlying reason for the 'down-regulation' of LAK cell activity by rIFN-y is not appreciated at this time, but may be explained by the following rationale. rIFN-y has been postulated to act as a mediator of rIL-2 on the augmentation of NK activity (Sverdersky et al., 1984; Giovarelli et al., 1988). In the human system a synergism of rIFN-y and rIL-2 has been
119
reported by some (Sverdersky et al., 1984; Itoh et al., 1985), but questioned by others (Hoyer et al., 1986). In the mouse system, it is not entirely clear or conclusive whether or not rIFN-y could augment NK activity (Ortaldo, 1987). In the BALB/c mouse system, as used in this investigation, it has been shown that the major precursors of LAK cells are of NK nature, i.e. expressing asialo GM antigen (Chao, Ohnishi & Chu, 1989). In the present study rIFN-y alone neither enhanced murine NK activity nor induced LAK activity. On the contrary, rIFN-y was found to exhibit a suppressive effect on the induction of 3-day cultured LAK cell activity at a concentration as low as 1 U/ml. The kinetic study further revealed that the suppressive effect of rIFN-y on LAK cell activity was influenced by the timing of rIFN-y and rIL-2 incubation. The inhibition on the 3-daycultured LAK cell activity could only be detected by the addition of rIFN-y to rIL-2-containing medium at the early phase (48 hr or more) of the incubation. An incubation of rIFN-y in rIL-2-containing medium for 24 hr produced no decrease in LAK cell activity. This finding suggested that the effect of rIFN-y on LAK cell activity is very probably by way of the precursor cells. One biological function played by rIFN-y is its ability to activate the macrophages, which in turn secrete numerous soluble products regulating positively or negatively the immune function (Boraschi et al., 1985; Johnston, 1988). Prostaglandin is one such secreted mediator that is involved in the inhibition of T-cell activation (Chouaib etal., 1985). In vivo experiments have indicated that prostaglandin inhibitor could magnify the therapeutic effect of IL-2 on murine tumour models (Lala & Parher, 1988). Therefore, one likely explanation for the data on the reduction of LAK cell activity by rIFN-y would be based on the notion that rIFN-y would enhance the secretion of prostaglandin by the macrophages that finally depress the LAK cell activity. To test the notion that rIFN-y affected the LAK cell activity indirectly by way of the macrophages within the splenocyte population, nylon-wool non-adherent splenocytes were incubated with rIL-2 plus rIFN-y versus rIL-2 alone. rIFN-y indeed lost its effect on LAK cell activity generated from the nylonwool non-adherent and 'macrophage-free or poor' splenocyte population. The nylon-wool treatment reduced the macrophages within the splenocyte preparations from approximately 15% to 2% (data not shown). Another possibility which may explain the effect of rIFN-y on the decrease in LAK cell activity is based upon a previous report describing that the combination of prostaglandin and rIFN-y could induce the differentiation of T suppressor cells (Elmasry, Fox & Rich, 1987). It is thus possible in the experiments shown here that differentiated T suppressor cells exhibited suppressive activity on LAK cells after the stimulation by endogenous prostaglandin and exogenous rIFN-y, and that when the macrophages were removed by nylon-wool the differentiation of T suppressor cells ceased. Therefore, the data presented here suggest that during the generation of LAK cell activity from the normal BALB/c mouse splenocytes, it is possible that rIL-2 can also trigger the macrophages to secrete some negative regulatory mediators. rIFN-y would exert its regulatory effects on LAK cell activity indirectly by means of the macrophages. Although the exact mechanisms are yet to be elucidated, including the examination of other cells and mediators, this study reports a 'down-
120
T.- Y. Chao, H. Ohnishi & T. M. Chu
regulation' of LAK cell activity by rIFN-y, and suggests the potential role of the macrophages in the generation of LAK cell activity.
ACKNOWLEDGMENTS This investigation was supported in part by research grants IM-416 awarded by the American Cancer Society and CA-37646 awarded by the National Cancer Institute. We thank Ms J. Ogledzinski for her assistance in manuscript preparation and the Department of Laboratory Animal Resources for providng the animals and their care.
REFERENCES BORASCHI D., CENSINI S., BARTALINI M. & TAGLIABUE A. (1985) Regulation of arachidonic acid metabolism in macrophages by immune and nonimmune interferons. J. Immunol. 135, 502. CAPONE P.M., KADOHAMA N. & CHU T.M. (1987) Immunotherapy in a spontaneously developed murine mammary carcinoma with syngeneic monoclonal antibody. Cancer Immunol. Immunother. 25, 93. CHAO T.Y. & CHU T.M. (1989) Characterization of a new spontaneously developed murine mammary adenocarcinoma in syngeneic BALB/c hosts. In Vitro Cell. Devel. Biol. 25, 621. CHAO T.Y., OHNISHI H. & CHU T.M. (1989) Augmentation of murine lymphokine (rIL-2)-activated killer cell activity by indomethacin. Molec. Biotherap. 1, 318. CHOUAIB S., WELTE K., MERTELSMANN R. & DUPONT B. (1985) Prostaglandin E2 acts at two distinct pathways of T lymphocyte activation: inhibition of interleukin-2 production and down-regulation of transferrin receptor expression. J. Immunol. 135, 1172. ELMASRY, M.N., Fox E.J. & RICH R.R. (1987) Sequential effects of prostaglandins and interferon-y on differentiation of CD8 + suppressor cells. J. Immunol. 139, 688. GIOVARELLI M., SANTONI A., JEMMA C., Musso T., GIUFFRIDA A.M., CAVALLO G.I., LANDOLFO S. & FORNi G. (1988) Obligatory role of rIFN-y in induction of lymphokine-activated and T lymphocyte killer activity, but not in boosting of natural cytotoxicity. J. Immunol. 141,2831. HOYER M., MEINEKE T., LEWIS W., ZWILLING B. & RINEHART J. (1986) Characterization and modulation of human lymphokine (interleukin2) activated killer cell induction. Cancer Res. 46, 2834. IBAYASHI Y., HOON D.S.B. & GOLUB S.H. (1987) The regulatory effect of adherent cells on lymphokine-activated killer cells. Cell. Immunol. 110, 365. ITOH K., SHIBA K., SHIMIZU Y., SUZUKI R. & KUMAGAI K. (1985) Generation of activated killer (AK) cells by recombinant interleukin2 (rIL-2) in collaboration with interferon-y (IFN-y). J. Immunol. 134, 3124. JOHNSTON R.B. (1988) Monocytes and macrophages. New Eng. J. Med. 318, 747. KALLAND T., BELFRAGE H., BHILADVALA P. & HEDLUND G. (1987) Analysis of the murine lymphokine-activated killer (LAK) cell phenomenon: dissection ofeffectors and progenitors into NK- and Tlike cells. J. Immunol. 138, 3640. LALA P.K. & PARHAR, R.S. (1988) Cure of Bl6FIO melanoma lung metastasis in mice by chronic indomethacin therapy combined with
repeated rounds of interleukin-2: characteristics of killer cells generated in situ. Cancer Res. 48, 1072. MAZUMDER A. & ROSENBERG S.A. (1984) Successful immunotherapy of natural killer-resistant established pulmonary melanoma metastases by the intravenous adoptive transfer of syngeneic lymphocytes activated in vitro by interleukin-2. J. exp. Med. 159,495. Nii A., SONE S., UTSUGI T., YANAGAWA H. & OGURA T. (1988) Up- and down-regulation of human lymphokine (IL-2)-activated killer cell induction by monocytes, depending on their functional state. Int. J. Cancer, 41, 33. ORTALDO J.R. (1987) Regulation of natural killer activity. Cancer Met. Rev. 6, 637. O-row R.T., STELLER E.P., SUGARBAKER P.H., WESLEY R.A. & ROSENBERG R.A. (1987) Immunotherapy of intraperitoneal cancer with interleukin-2 and lymphokine-activated killer cells reduces tumor load and prolongs survival in murine models. Cell. Immunol. 104, 366. PRoss G.F., BAINES M.G., RUBIN P., SHRAGGE P. & PATrERSON M.S. (1981) Spontaneous human lymphocyte-mediated cytotoxicity against tumor target cells. J. clin. Immunol. 1, 51. RAMSDELL F.J. & GOLUB S.H. (1987) Generation of lymphokineactivated killer cell activity from human thymocytes. J. Immunol. 139, 1446. ROSENBERG S.A., GRIMM E.A., MCCROGAN M., DOYLE M.V., KAWASAKI E., KOTHS K. & MARK D.F. (1983) Biological activity of recombinant interleukin-2 produced in Escherichia coli. Science, 233, 1412. ROSENSTEIN M., YRON I. & KAUFMANN Y. & ROSENBERG S.A. (1984) Lymphokine-activated killer cells: lysis of fresh syngeneic natural killer-resistant murine tumor cells by lymphocytes cultured in interleukin-2. Cancer Res. 44, 1946. SAYERS T.J., MASON A.T. & ORTALDO JR. (1986) Regulation of human natural killer cell activity by interferon-y: lack of a role in interleukin2 mediated augmentation. J. Immunol. 136, 2176. SILVENNOINEN O., VAKKILA J. & HURME M. (1988) Accessory cells, dendritic cells, or monocytes, are required for the lymphokineactivated killer cell induction from resting T cells but not from natural killer cell precursors. J. Immunol. 141, 1404. SUZUKI R., HANDA K., ITOH K. & KUMAGAI K. (1983) Natural killer (NK) cells as a responder to interleukin-2 (IL-2). I. Proliferative response and establishment of cloned cells. J. Immunol. 130, 981. SVEDERSKY L.P., SHEPARD H.M., SPENCER S.A., SHALABY M.R. & PALLADINO M.A. (1984) Augmentation of human natural cellmediated cytotoxicity by recombinant human interleukin-2. J. Immunol. 133, 714. TAKAI N., TANAKA R., YOSHIDA S., HARA N. & SAITO T. (1988) In vivo and in vitro effect of adoptive immunotherapy of experimental murine brain tumor using lymphokine-activated killer cells. Cancer Res. 48, 2047. VUJANOVIC N.L., HERBERMAN R.B. & HISERODT J.C. (1988) Lymphokine-activated killer cells in rats: analysis of tissue and strain distribution, ontogeny, and target specificity. Cancer Res. 48, 878. WANG A., Lu S.D. & MARK D. (1984) Site-specific mutagenesis of the human interleukin-2 gene: structure-function analysis of the cysteine residues. Science, 222, 1431. YANG J.C., MULE J.J. & ROSENBERG S.A. (1986) Murine lymphokineactivated killer (LAK) cells: phenotypic characterization of the precursor and effector cells. J. Immunol. 137, 715.
Indirect inhibition of generation of murine lymphokine-activated killer cell activity in splenocyte cultures by interferon-gamma T.-Y. CHAO, H. OHNISHI & T. M. CHU Department of Diagnostic Immunology Research and Biochemistry, Roswell Park Memorial Institute, New York State Department of Health, Buffalo, New York, U.S.A.
Acceptedfor publication 5 January 1990
SUMMARY (rIFN-y) on murine lymphokine-activated killer interferon-gamma recombinant The effect of mouse (LAK) cell activity was investigated using natural killer (NK)-resistant, spontaneously developed, weakly immunogenic and highly tumourigenic, syngeneic murine mammary adenocarcinoma, JC, mimicking that of human disease, as the target. Murine YAC-1 also was used as a target cell line. rIFN-y, when used in combination with recombinant interleukin-2 (rIL-2), was shown to exhibit a suppressed effect on LAK cell activity generated from BALB/c mouse splenocytes, compared to that with rIL-2 alone. The decrease in LAK cell activity was rIFN-y dose-dependent. Kinetic study revealed that this inhibitory effect was demonstrated only when rIFN-y was added to the medium at the early phase of rIL-2 culture. The inhibitory effect on LAK cell generation by rIFN-y was completely abrogated when the nylon-wool-treated non-adherent 'macrophage-free' splenocytes were incubated with rIL-2 and rIFN-y. These results indicated that the LAK cell activity generated from murine splenocytes cultured with rIL-2 could be depressed by rIFN-y, and that the macrophages may be involved as mediators. INTRODUCTION The induction of lymphokine-activated killer (LAK) cell activity by recombinant interleukin-2 (rIL-2) from lymphoid cells exhibiting cytotoxicity against a broad spectrum of fresh tumour cells and cultured tumour cell lines has been well demonstrated (Rosenstein et al., 1984; Ramsdell & Golub, 1987; Vujanovic, Herberman & Hiserodt, 1988). IL-2-induced LAK cells exhibiting such a non-major histocompatibility complex (MHC)-restricted cytotoxicity have been used successfully in experimental adoptive immunotherapy against animal tumours (Mazumder & Rosenberg, 1984; Ottow et al., 1987; Takai et al., 1988). Although LAK cell activity was generated primarily from natural killer (NK) cells and T cells (Suzuki et al., 1983; Yang, Mule & Rosenberg, 1986; Kalland et al., 1987), other immune cells such as monocytes and macrophages have been reported to exhibit regulatory function on LAK cell activity (Ibayashi, Hoon & Golub, 1987; Nii et al., 1988; Silvennoinen, Vakkila & Hurme, 1988). An alteration of the generation of LAK cell activity is thus possible through the modulation of the interactions among these cells, which may provide insight into the Abbreviations: JC, a spontaneously developed syngeneic murine adenocarcinoma; LAK, lymphokine-activated killer; NK, natural killer; rIFN-y, mouse recombinant interferon-gamma; rIL-2, human recombinant interleukin-2. Correspondence: Dr T. M. Chu, Dept. of Diagnostic Immunology Research and Biochemistry, Roswell Park Memorial Institute, 666 Elm Street, Buffalo, NY 14263, U.S.A.
mammary
mechanisms of action by LAK cells, with the use of other agents. Interferon-gamma (rIFN-y), a potent lymphokine modulating some immune response, has been reported to be able to augment NK activity and to exhibit synergistic effects with rIL-2 on the generation of human LAK cell activity (Svedersky et al., 1984; Itoh et al., 1985). However, a negative effect by rIFN-y on LAK cell activity has also been reported (Hoyer et al., 1986; Sayers, Mason & Ortaldo, 1986; Giovarelli et al., 1988). A spontaneously developed syngeneic murine mammary adenocarcinoma cell line, designated JC, has been established that is resistant to NK cell but susceptible to LAK cell lysis (Capone, Kadohama & Chu, 1987; Chao & Chu, 1989). Because it is weakly immunogenic and highly tumourigenic, the major biologic characteristics mimicking that ofhuman disease, JC is a more suitable model for experimental immunotherapy than other chemically or virally induced mammary tumour models. Using JC and YAC-1 as the targets, the effect of rIFN-y on murine LAK cell activity in vitro has been investigated. The results indicated that rIFN-y was capable of exerting an inhibitory effect on the generation of rIL-2-induced LAK cell activity from BALB/c mouse splenocytes. The suppressed effect was completely abrogated by treatment of whole splenocytes with nylon-wool, suggesting a potential role played by the macrophages in the decrease in LAK cell activity by rIFN-y. MATERIALS AND METHODS Mice
Eight- to 12-week-old BALB/c female mice were used. The
116
117
Inhibition of LAK activity by rIFN-y BALB/c mice were maintained at West Seneca Laboratory, Roswell Park Memorial Institute. Histocompatibility test by skin grafting was performed routinely to ascertain the maintenance of the line.
applied to the column and incubated at 37°, 5% Co2, for 45 min. Non-adherent cells were collected by washing the column with 30 ml of prewarmed medium (I drop/second), centrifuged (200 g, 10 min) and then resuspended in RPMI-1640-10% FCS.
Target cells YAC-l, a NK-sensitive cell line of A/Sn mouse origin, was obtained from the American Type Culture Collection (Rockville, MD). JC tumour line was established from a spontaneously developed mammary adenocarcinoma in an aged virgin female BALB/c mouse (Capone et al., 1987). JC is a NKresistant cell line, but susceptible to LAK cell lysis (Chao & Chu, 1989). These cell lines were maintained in RPMI-1640 with 10% fetal calf serum (FCS) (Gibco, Grand Island, NY).
In vitro cytotoxic assay NK and LAK cell activities were measured by using a standard 4-hr 5"Cr-release microcytotoxicity assay. Target cells (2 x 106) were harvested and washed twice, and then incubated in 50 pCi of 51Cr (Na5lCrO4, 402-61-615 67 mCi/mg, New England Nuclear, Boston, MA) for 2 hr at 370 in a moist atmosphere with 5% CO2. The labelled target cells were washed three times with RPMI-1640-10% FCS. The effector cells and 5'Cr-labelled target cells (5 x 103) were then placed in 96-well U-bottomed microtitre plates (Falcon, Oxnard, CA) with various effector to target cell ratios in a final volume of 0-2 ml in RPMI-1640-10% FCS. After a 4-hr incubation period at 370 in a humidified 5% CO2 atmosphere, supernatant was collected and counted in a gamma-counter (Packard Multi-Prias, Downers Grove, IL) for experimental release (ER). For the spontaneous release (SR) control, a sample of labelled target cells was resuspended in RPMI-1640-10% FCS alone. Total release (TR) of 51Cr from target cells was obtained by mixing 5 x I03 labelled target cells in 0-1 ml of RPMI-1640-10% FCS with 0-1 ml 10% SDS. The spontaneous release was generally 10-20% of the total release. Percentage specific release for 5'Cr release was calculated by using the following equation:
Recombinant IL-2 and IFN-y Human rIL-2, produced from Escherichia coli, was kindly supplied by Cetus Corporation (Emeryville, CA), with a specific activity of 18 x 106 IU/mg. Endotoxin level was less than 0-01 ng/ml. The purity was >98% by sodium dodecyl sulphatepolyacrylamide gel electrophoresis (SDS-PAGE) (Rosenberg et al., 1983; Wang, Lu & Mark, 1984). Mouse rIFN-y was provided by Schering Corporation, Kenilworth, NJ, as part of the American Cancer Society's programme on IFN, with a specific activity of 106 U/mg. The purity was >98% SDS-PAGE. Generation of LAK cells The spleens from normal BALB/c mice were harvested asceptically and minced in a stainless steel mesh to produce a single cell suspension. The erythrocytes were lysed osmotically with Trisammonium chloride buffer, 0-17 M, pH 7-2. The remaining splenocytes were either washed twice with RPMI-1640, or passed through a nylon-wool column before washed, and then incubated at 370 in complete medium containing rIL-2, under a moist atmosphere with 5% CO2, in culture flasks, at a cell concentration of 1-25 x 106/ml. Complete medium contained RPMI-1640 with 10% FCS, 0-1 mm non-essential amino acids (Gibco), 0-1 mm sodium pyruvate (Gibco), 5 x I0-I M 2mercaptoethanol (Sigma, St Louis, MO), 0-3 mg/ml L-glutamine (Gibco), 100 pg/ml streptomycin (Gibco) and 100 U/ml penicillin (Gibco). LAK cells were harvested after culture for various periods of time, washed twice, and the viable cells were counted and resuspended in RPMI-1640-10% FCS for the standard 4 hr 5"Cr-release assay. Generation of LAK cells in the presence of rIFN-y The splenocytes were prepared as described above. rIFN-y was added to the complete medium containing rIL-2. The concentrations of rIFN-y were adjusted from 0-01 to 1000 U/ml, and were evaluated either alone or with rIL-2 at 60, 600 and 3000 IU/ ml. After incubation for various periods of time, the activated splenocytes were harvested, washed and used as effector cells in 5'Cr-release assay.
Nylon-wool treatment Approximately 0-6 g of nylon-wool (Polysciences Inc., Warrington, PA) was packed into a syringe (10 ml) and incubated with RPMI-1640-10% FCS for 1 hr at 37°. After washing the column with 10 ml ofprewarmed RPMI-1640-10% FCS, 2 ml of whole splenocytes (107/ml) in RPMI-1640-10% FCS were
percentage specific release (% cytotoicity)=
ER-SRX 100 TR-SR x10 Each cytotoxic assay was carried out in triplicate and the results were expressed as lytic units. Lytic units were calculated from linear regression plotting from various effector to target cell ratios, by a MacIntosh SE computer, as described elsewhere (Pross et al., 1981). One lytic unit was defined as the number of effector cells required to cause 30% specific 51Cr release from 104 target cells and was expressed as LU/ 107 effector cells. Statistical analysis was performed with Student's t-test. RESULTS Effect of rIFN-y on rIL-2 induced murine LAK cell activity Two target cell lines, YAC- 1, a commonly known NK-sensitive cell line of A/Sn mouse origin, and JC, a NK-resistant but LAKsensitive spontaneously developed mammary adenocarcinoma cell line of BALB/c mouse origin, were used as the indicators of cytotoxicity as measured by 5"Cr-release assay. Freshly prepared normal BALB/c mouse splenocytes showed only marginal, if any, NK cell activity in the standard 4 hr 51Cr-release assay. The culture of the normal BALB/c mouse splenocytes in complete medium in the presence of rIFN-y, 10, 100 and 1000 U/ml, for 3 days produced no change in cytotoxicity against either YAC-l or JC cells (data not shown). In concert with the use of rIL-2, rIFN-y was found to demonstrate no augmentation on LAK cell activity. On the contrary, the addition of rIFN-y to rIL-2 containing complete medium decreased LAK cell activity. This inhibitory effect of rIFN-y on the LAK cell activities directed against both YAC-l
118
T.- Y. Chao, H. Ohnishi & T. M. Chu 60 50 C) -i 40
iz rlL-2 alone * rlL-2+ IFN --y
120 100 C
0.00
150 125 100 .2 75 aJ 50 25
80 0
c
30 J- 20 10
60
m
42 O._
___o
140
>1
L 0MI 0-1 10 100 IFN-T (U/ml) I
0
YAC-1
2
3
4
60 50 40 30 20 10 0
JC
|,LI ~~ 2
3
4
Culture (days)
Figure 1. The dose-dependent suppressive effect of rIFN-y on LAK cell activity. LAK cell activity was generated from co-culturing BALB/c mouse splenocytes (1-25 x 106/ml) with rIL-2 (3000 IU/ml) for 3 days. rIFN-y at the concentrations shown was added to complete medium at the beginning of rIL-2 culture. LAK cell activity was determined by the standard 4 hr 51Cr-release assay with the use of YAC- 1 and JC tumour cells as the targets, and converted to LU. Lytic units ofcytotoxic activity were calculated from linear regression plotted from various effector: target cell ratios (100: 1, 50: 1, 25: 1 and 12: 1). One lytic unit was defined as the number of effector cells required to cause 30% specific Cr release from 104 target cells. Each point represents the mean of triplicate assays. 5
and JC tumour cells was found to be dose-dependent (Fig. 1). As shown in Fig. I, a concentration as low as 1 U/ml was effective in significantly inhibiting rIL-2 (3000 IU/ml)-induced LAK cell activity directed against JC tumour cells (mean + SEM, 45 5+5-9 versus 31-6+ 3 1, PJ
25 0
L, rIL-2 Iday 2day 3doy alone With IFN-y
Figure 3. The length of rIFN-y incubation affecting LAK activity. LAK cell activity was generated from co-culturing BALB/c mouse splenocytes (1-25 x 106/ml) with rIL-2 (3000 IU/ml) for 3 days, and was determined by the standard 4 hr 5'Cr-release assay with the use of JC cells as the target and expressed as lytic units, as described in Fig. 1. rIFN-y (100 IU/ml) was added to the rIL-2-containing complete medium, 1, 2 and 3 days prior to the 5"Cr-release assay. Each block and bar represent the mean+SEM of triplicate assays. A statistically significant difference in LAK cell activity was found between the splenocytes cultured in rIL-2 alone and rIL-2 plus rIFN-y for 2 and 3 days (P= 0-02 and 0 005, respectively).
medium 24 hr prior to the 5'Cr-release assay produced no significant inhibitory effect on LAK cell activity (51-5+6-8, P=0-7). Effect of rIFN-y on murine LAK cell activity generated from nylon-wool non-adherent splenocytes The normal mouse splenocytes in single cell suspension were passed through a nylon-wool column, and the non-adherent cells were collected and cultured in complete medium with rIL-2 and rIFN-y for 3 days. As shown in Fig. 4, after incubation with rIL-2 (3000 IU/ml) and rIFN-y (100 U/ml) for 3 days, the LAK
Inhibition of LAK activity by rIFN-y rIL-2 alone * rlL-2+ IFN-y 250 S
200
c u
150
3
100 50
4\
e
&
Figure 4. The effect of nylon-wool treatment on LAK cell activity generated with rIL-2 and rIL-2 plus rIFN-y. LAK cell activity was generated by co-culturing either whole spleen cells (1-25 x 106/ml) or nylon-wool non-adherent spleen cells (1-25 x 106/ml) with rIL-2 (3000 IU/ml), and rIL-2 plus rIFN-y (100 U/ml) for 3 days. The cytotoxicity was determined by the standard 4 hr 51Cr-release assay using JC cells as the targets and expressed as lytic units, as described in Fig. 1. Each block and bar represent the mean + SEM of triplicate assays. The difference in LAK cell activity between cultures with rIL-2 and rIL-2 plus rIFN-y was statistically significant in whole splenocytes (YAC-1, P=0009; JC, P= 0 02), and not significant in nylon-wool-treated non-adherent splenocytes (YAC-1, P=0-85; JC P=0 13).
cell activity directed against both YAC-1 and JC cells as the targets and generated from the whole splenocytes was significantly less than that after incubation with rIL-2 alone (YAC-1 cells, 39 3+3 9 versus 140-4+13-3, P=0 009; JC cells, 13-9+2-1 versus 45-4+5.9, P=0-02), confirming the results shown in Figs 2 and 3. Also shown is that rIFN-y exhibited no effect on the rIL-2-induced LAK cell activity generated from the nylon-wool-treated non-adherent 'macrophage-free or poor' splenocytes, since a similar cytotoxicity was obtained from the culture incubated with rIL-2 alone (YAC-1 cells, 155-6+28-2 versus 152-3+28-5, P=0-85; JC cells, 97-3+18-8 versus 77-3 + 10-3, P=0-13).
DISCUSSION In the present report, the effect of rIFN-y on murine LAK cell activity was investigated using a NK-sensitive cell line (YAC- 1) and a NK-resistant yet LAK-sensitive cell line (JC) as the targets. The data indicated that rIFN-y alone was incapable of enhancing the cytotoxicity of normal BALB/c mouse splenocytes. However, when used in combination with rIL-2, rIFN-y was found to suppress the rIL-2-induced LAK activity. Additional experiments using nylon-wool to remove the macrophages (and B lymphocytes) from the whole splenocytes revealed that the 'down-regulation' of LAK cell activity by rIFN-y was completely abrogated when the 'macrophage-poor or free' splenocyte preparations were used as the effector cells. The underlying reason for the 'down-regulation' of LAK cell activity by rIFN-y is not appreciated at this time, but may be explained by the following rationale. rIFN-y has been postulated to act as a mediator of rIL-2 on the augmentation of NK activity (Sverdersky et al., 1984; Giovarelli et al., 1988). In the human system a synergism of rIFN-y and rIL-2 has been
119
reported by some (Sverdersky et al., 1984; Itoh et al., 1985), but questioned by others (Hoyer et al., 1986). In the mouse system, it is not entirely clear or conclusive whether or not rIFN-y could augment NK activity (Ortaldo, 1987). In the BALB/c mouse system, as used in this investigation, it has been shown that the major precursors of LAK cells are of NK nature, i.e. expressing asialo GM antigen (Chao, Ohnishi & Chu, 1989). In the present study rIFN-y alone neither enhanced murine NK activity nor induced LAK activity. On the contrary, rIFN-y was found to exhibit a suppressive effect on the induction of 3-day cultured LAK cell activity at a concentration as low as 1 U/ml. The kinetic study further revealed that the suppressive effect of rIFN-y on LAK cell activity was influenced by the timing of rIFN-y and rIL-2 incubation. The inhibition on the 3-daycultured LAK cell activity could only be detected by the addition of rIFN-y to rIL-2-containing medium at the early phase (48 hr or more) of the incubation. An incubation of rIFN-y in rIL-2-containing medium for 24 hr produced no decrease in LAK cell activity. This finding suggested that the effect of rIFN-y on LAK cell activity is very probably by way of the precursor cells. One biological function played by rIFN-y is its ability to activate the macrophages, which in turn secrete numerous soluble products regulating positively or negatively the immune function (Boraschi et al., 1985; Johnston, 1988). Prostaglandin is one such secreted mediator that is involved in the inhibition of T-cell activation (Chouaib etal., 1985). In vivo experiments have indicated that prostaglandin inhibitor could magnify the therapeutic effect of IL-2 on murine tumour models (Lala & Parher, 1988). Therefore, one likely explanation for the data on the reduction of LAK cell activity by rIFN-y would be based on the notion that rIFN-y would enhance the secretion of prostaglandin by the macrophages that finally depress the LAK cell activity. To test the notion that rIFN-y affected the LAK cell activity indirectly by way of the macrophages within the splenocyte population, nylon-wool non-adherent splenocytes were incubated with rIL-2 plus rIFN-y versus rIL-2 alone. rIFN-y indeed lost its effect on LAK cell activity generated from the nylonwool non-adherent and 'macrophage-free or poor' splenocyte population. The nylon-wool treatment reduced the macrophages within the splenocyte preparations from approximately 15% to 2% (data not shown). Another possibility which may explain the effect of rIFN-y on the decrease in LAK cell activity is based upon a previous report describing that the combination of prostaglandin and rIFN-y could induce the differentiation of T suppressor cells (Elmasry, Fox & Rich, 1987). It is thus possible in the experiments shown here that differentiated T suppressor cells exhibited suppressive activity on LAK cells after the stimulation by endogenous prostaglandin and exogenous rIFN-y, and that when the macrophages were removed by nylon-wool the differentiation of T suppressor cells ceased. Therefore, the data presented here suggest that during the generation of LAK cell activity from the normal BALB/c mouse splenocytes, it is possible that rIL-2 can also trigger the macrophages to secrete some negative regulatory mediators. rIFN-y would exert its regulatory effects on LAK cell activity indirectly by means of the macrophages. Although the exact mechanisms are yet to be elucidated, including the examination of other cells and mediators, this study reports a 'down-
120
T.- Y. Chao, H. Ohnishi & T. M. Chu
regulation' of LAK cell activity by rIFN-y, and suggests the potential role of the macrophages in the generation of LAK cell activity.
ACKNOWLEDGMENTS This investigation was supported in part by research grants IM-416 awarded by the American Cancer Society and CA-37646 awarded by the National Cancer Institute. We thank Ms J. Ogledzinski for her assistance in manuscript preparation and the Department of Laboratory Animal Resources for providng the animals and their care.
REFERENCES BORASCHI D., CENSINI S., BARTALINI M. & TAGLIABUE A. (1985) Regulation of arachidonic acid metabolism in macrophages by immune and nonimmune interferons. J. Immunol. 135, 502. CAPONE P.M., KADOHAMA N. & CHU T.M. (1987) Immunotherapy in a spontaneously developed murine mammary carcinoma with syngeneic monoclonal antibody. Cancer Immunol. Immunother. 25, 93. CHAO T.Y. & CHU T.M. (1989) Characterization of a new spontaneously developed murine mammary adenocarcinoma in syngeneic BALB/c hosts. In Vitro Cell. Devel. Biol. 25, 621. CHAO T.Y., OHNISHI H. & CHU T.M. (1989) Augmentation of murine lymphokine (rIL-2)-activated killer cell activity by indomethacin. Molec. Biotherap. 1, 318. CHOUAIB S., WELTE K., MERTELSMANN R. & DUPONT B. (1985) Prostaglandin E2 acts at two distinct pathways of T lymphocyte activation: inhibition of interleukin-2 production and down-regulation of transferrin receptor expression. J. Immunol. 135, 1172. ELMASRY, M.N., Fox E.J. & RICH R.R. (1987) Sequential effects of prostaglandins and interferon-y on differentiation of CD8 + suppressor cells. J. Immunol. 139, 688. GIOVARELLI M., SANTONI A., JEMMA C., Musso T., GIUFFRIDA A.M., CAVALLO G.I., LANDOLFO S. & FORNi G. (1988) Obligatory role of rIFN-y in induction of lymphokine-activated and T lymphocyte killer activity, but not in boosting of natural cytotoxicity. J. Immunol. 141,2831. HOYER M., MEINEKE T., LEWIS W., ZWILLING B. & RINEHART J. (1986) Characterization and modulation of human lymphokine (interleukin2) activated killer cell induction. Cancer Res. 46, 2834. IBAYASHI Y., HOON D.S.B. & GOLUB S.H. (1987) The regulatory effect of adherent cells on lymphokine-activated killer cells. Cell. Immunol. 110, 365. ITOH K., SHIBA K., SHIMIZU Y., SUZUKI R. & KUMAGAI K. (1985) Generation of activated killer (AK) cells by recombinant interleukin2 (rIL-2) in collaboration with interferon-y (IFN-y). J. Immunol. 134, 3124. JOHNSTON R.B. (1988) Monocytes and macrophages. New Eng. J. Med. 318, 747. KALLAND T., BELFRAGE H., BHILADVALA P. & HEDLUND G. (1987) Analysis of the murine lymphokine-activated killer (LAK) cell phenomenon: dissection ofeffectors and progenitors into NK- and Tlike cells. J. Immunol. 138, 3640. LALA P.K. & PARHAR, R.S. (1988) Cure of Bl6FIO melanoma lung metastasis in mice by chronic indomethacin therapy combined with
repeated rounds of interleukin-2: characteristics of killer cells generated in situ. Cancer Res. 48, 1072. MAZUMDER A. & ROSENBERG S.A. (1984) Successful immunotherapy of natural killer-resistant established pulmonary melanoma metastases by the intravenous adoptive transfer of syngeneic lymphocytes activated in vitro by interleukin-2. J. exp. Med. 159,495. Nii A., SONE S., UTSUGI T., YANAGAWA H. & OGURA T. (1988) Up- and down-regulation of human lymphokine (IL-2)-activated killer cell induction by monocytes, depending on their functional state. Int. J. Cancer, 41, 33. ORTALDO J.R. (1987) Regulation of natural killer activity. Cancer Met. Rev. 6, 637. O-row R.T., STELLER E.P., SUGARBAKER P.H., WESLEY R.A. & ROSENBERG R.A. (1987) Immunotherapy of intraperitoneal cancer with interleukin-2 and lymphokine-activated killer cells reduces tumor load and prolongs survival in murine models. Cell. Immunol. 104, 366. PRoss G.F., BAINES M.G., RUBIN P., SHRAGGE P. & PATrERSON M.S. (1981) Spontaneous human lymphocyte-mediated cytotoxicity against tumor target cells. J. clin. Immunol. 1, 51. RAMSDELL F.J. & GOLUB S.H. (1987) Generation of lymphokineactivated killer cell activity from human thymocytes. J. Immunol. 139, 1446. ROSENBERG S.A., GRIMM E.A., MCCROGAN M., DOYLE M.V., KAWASAKI E., KOTHS K. & MARK D.F. (1983) Biological activity of recombinant interleukin-2 produced in Escherichia coli. Science, 233, 1412. ROSENSTEIN M., YRON I. & KAUFMANN Y. & ROSENBERG S.A. (1984) Lymphokine-activated killer cells: lysis of fresh syngeneic natural killer-resistant murine tumor cells by lymphocytes cultured in interleukin-2. Cancer Res. 44, 1946. SAYERS T.J., MASON A.T. & ORTALDO JR. (1986) Regulation of human natural killer cell activity by interferon-y: lack of a role in interleukin2 mediated augmentation. J. Immunol. 136, 2176. SILVENNOINEN O., VAKKILA J. & HURME M. (1988) Accessory cells, dendritic cells, or monocytes, are required for the lymphokineactivated killer cell induction from resting T cells but not from natural killer cell precursors. J. Immunol. 141, 1404. SUZUKI R., HANDA K., ITOH K. & KUMAGAI K. (1983) Natural killer (NK) cells as a responder to interleukin-2 (IL-2). I. Proliferative response and establishment of cloned cells. J. Immunol. 130, 981. SVEDERSKY L.P., SHEPARD H.M., SPENCER S.A., SHALABY M.R. & PALLADINO M.A. (1984) Augmentation of human natural cellmediated cytotoxicity by recombinant human interleukin-2. J. Immunol. 133, 714. TAKAI N., TANAKA R., YOSHIDA S., HARA N. & SAITO T. (1988) In vivo and in vitro effect of adoptive immunotherapy of experimental murine brain tumor using lymphokine-activated killer cells. Cancer Res. 48, 2047. VUJANOVIC N.L., HERBERMAN R.B. & HISERODT J.C. (1988) Lymphokine-activated killer cells in rats: analysis of tissue and strain distribution, ontogeny, and target specificity. Cancer Res. 48, 878. WANG A., Lu S.D. & MARK D. (1984) Site-specific mutagenesis of the human interleukin-2 gene: structure-function analysis of the cysteine residues. Science, 222, 1431. YANG J.C., MULE J.J. & ROSENBERG S.A. (1986) Murine lymphokineactivated killer (LAK) cells: phenotypic characterization of the precursor and effector cells. J. Immunol. 137, 715.
