Immunopharmacology, 21 (1991) 199-210 0162-3109/91/$03.50 © 1991 Elsevier Science Publishers B.V.
199
IMPHAR 00545
Drug-mediated increase of susceptibility of human lung cancer to NK or LAK effector cells Stefania D'Atri 1, Stefano Marini 2, Lucio Tentori 2, M a r i a Tricarico and E n z o B o n m a s s a r 2
I,
M a r i a Pia Fuggetta 1
llnstitute of Experimental Medicine, CNR, Rome and 2Department of Experimental Medicine and Biochemical Sciences, 11 University of Rome, Rome, Italy (Received 21 August 1990; accepted 25 February 1991)
Abstract: Previous studies in murine models have shown that in vivo or in vitro treatment of tumor cells with mutagenic triazene compounds (TZC) lead to the appearance of novel drug-mediated tumor antigens (DMTA) capable of eliciting graft resistance in syngeneic hosts. This phenomenon, defined as 'chemical xenogenization' (CX), could be of potential value for immunochemotherapy of human neoplasias. It was also shown that TZC modulate NK sensitivity of murine tumor cells. Therefore, experiments were conducted to evaluate whether susceptibility of a human lung adenocarcinoma cell line (H 125) to natural cytotoxic effectors could be affected by treatment with an in vitro active TZC. The results showed that drug treatment of H125 line leads to heritable increase of susceptibility to NK and LAK cells. Moreover, increased binding between effector and drug-treated target cells was observed. Additional studies on HLA antigens showed that changes in HLA-ABC molecule expression were probably not involved in TZC-induced increase of NK/LAK susceptibility. These results suggest that TZC treatment of a human tumor could result in increased expression of membrane structures recognized by natural cytotoxic effector cells. Further studies are required to explore whether these changes are generated by mutational events correlated with TZC-induced CX of human cancer cells. Key words:
Chemical xenogenization; LAK susceptibility; NK susceptibility; Triazene compounds
Correspondence: Stefania D'Atri, Department of Experimental Medicine and Biochemical Sciences, II University of Rome, Via O. Raimondo, 00173 Rome, Italy. Abbreviations: CM, complete medium; cpm, counts per minute; CX, chemical xenogenization; DMTA, drug-mediated tumor antigens; DTIC, 5-(3,3-dimethyl-l-triazeno) imidazole-4-carboxamide; E, effector; EDTA, ethylenediamine-tetraacetic acid; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; IL-1, interleukin-1; IL-2, interleukin-2; KC, killed cells; LAK, IL-2 activated killer cells; LUn, lytic unit n; mAb, monoclonal antibody; MFV, mean fluorescence value; MHC, major histocompatibility complex; MNC, mononuclear cells; MTBA, p-(3-methyl1-triazeno)benzoic acid potassium salt; NCMC, natural cellmediated cytotoxicity; NK, natural killer; NKCF, NK cyto-
Introduction
A large number of experimental studies demonstrated that in vivo (Bonmassar et al., 1970; Bonmassar et al., 1972; Bonmassar et al., 1979) or in vitro (Contessa etal., 1981) treatment of murine lymphomas with the antitumor drug 5- (3,3- dimethyl- 1-triazeno)imidazole-4-carbox-
toxic factors; NKTS, NK target structures; PBS, phosphate buffered saline; T, target; TNF, tumor necrosis factor; TZC, triazene compounds.
200 amide (DTIC), or other triazene compounds (TZC) (Nardelli et al., 1984), leads to the appearance of novel drug-mediated tumor antigens (DMTA). These antigens are capable of eliciting T cell-mediated proliferative and cytotoxic responses in vitro (Romani et al., 1979), specific transplantation resistance (Riccardi et al., 1978), and specific tolerance (Houchens et al., 1976) in syngeneic hosts. Moreover recent studies pointed out that a D M T A expressed by a DTIC-treated murine leukemia, is a surface glycoprotein of approximately 80 kDa, possibly related to retroviral gp70 molecules (Grohman et al., 1990). This phenomenon, described as 'chemical xenogenization' (CX) (Puccetti et al., 1985), similar to that obtained with other methylating mutagenic agents (Boon and Kellerman, 1977; Boon, 1983) could lead to new perspectives in cancer immunochemotherapy. On these bases it was of primary interest to investigate whether TZC-mediated CX could occur in human neoplasias. Previous studies in murine models (Romani et al., 1983) showed that treatment of tumor cells with TZC modulates their susceptibility to N K effector cells. Therefore experiments were carried out on a human cell line (i.e. lung adenocarcinoma H 125), to test whether its susceptibility to natural cell-mediated cytotoxicity (NCMC) could be affected by treatment with an in vitro active TZC, i.e. the potassium salt ofp-(3-methyl-l-triazeno) benzoic acid (MTBA) (Fioretti et al., 1983). Parental H125 cells, and the MTBA-treated subline H125/M6 (i.e. H125 cells subjected to 6 cycles of drug treatment) were comparatively tested for their susceptibility to NCMC, using 2 types of effector cells, i.e. fresh mononuclear cells from peripheral blood, or IL-2 activated killer cells (LAK cells) (Grimm et al., 1982; Grimm et al., 1983; Ortaldo et al., 1986; Phillips and Lanier, 1986; Hersey and Bolhuis, 1987). Furthermore it was investigated whether MTBA treatment could induce changes in the expression of HLA antigens, since an inverse relationship between susceptibility to N C M C and the extent of M H C antigen expression has been pro-
posed by a number of investigators (Ljunggren and Karre, 1985; Harel-Bellan et al., 1986; Karre et al., 1986; Storkus et al., 1987; Quillet et al., 1988). The results presented here show that MTBA treatment of H 125 line increased its susceptibility to N C M C and binding to effector lymphocytes. Moreover drug treatment produced a limited decrease of HLA-class I antigen expression. On the other hand, marked reduction of the density of class I histocompatibility molecules induced by citric acid (Sugawara et al., 1987) in H125 cells did not result in increased NCMC susceptibility of the tumor. These data suggest that MTBAmediated increase of H125 sensitivity to natural effector ceils is presumably the result of increased density or availability of target structures recognized by NK or LAK cells, not correlated with changes in HLA antigen expression. Methods Cell lines All cell lines were maintained in RPMI 1640 medium (GIBCO, Paisley, Scotland, U.K.) supplemented with 10~o heat inactivated fetal calf serum (FCS, GIBCO) and 2 mM L-glutamine (Flow Laboratories, McLean, VA, U.S.A.) (hereafter referred to as 'Complete Medium', CM). The erythroteukemic cell line K562 (Lozzio and Lozzio, 1979) and the Burkitt lymphoma cell line Daudi (Klein et al., 1968) were grown in suspension culture and subcultured three times weekly. Cell line H125, a human lung adenocarcinoma line growing in monolayer, kindly provided by Dr. J.D. Minna (National Cancer Institute, NIH, Navy Medical Oncology Branch, Bethesda, MD, U.S.A.), drug-treated H125/M6 subline and cell clones were also grown in CM and subcultured twice a week. In vitro cloning of tumor cells Parental or TZC-treated line were cloned by limiting dilution in flat-bottom 96-well microtiter plates (Falcon, Becton Dickinson Labware, Oxnard, CA, U.S.A.) at 0.5 cells/well in CM.
201 Reagents p-(3-Methyl- 1-triazeno)benzoic acid potassium salt (MTBA), was generously provided by Prof. L. Lassiani (Institute of Pharmaceutical Chemistry, University of Trieste, Trieste, Italy). The drug was dissolved in ice-cold RPMI 1640 protected from light, sterilized through 0.22 #m size filter (Millipore, Molsheim, France) and added to cell suspension for in vitro treatment. All procedures were performed within 1 min in order to avoid drug inactivation, being the half-life of MTBA in cold medium less than 9 min (Lassiani L., personal communication). Recombinant human interleukin-2 (rlL-2) was kindly provided by Dr. D. Criscuolo (Hoffmann La Roche, Milan, Italy). Purified anti-HLA-ABC (IgGaa) monoclonal antibody (mAb) from W6/32 hybridoma (Barnstaple et al., 1978; Muller et al., 1983) and fluorescein isothiocyanate (FITC)conjugated F(ab')2 rabbit anti-mouse IgG were obtained from Dako (Dakopatts, Copenhagen, Denmark). FITC-conjugated anti-HLA-DR from L243 hybridoma (Lampson et al., 1980; Brodsky, 1984) and FITC-labeled mouse IgG2a+b were obtained from Becton Dickinson (Mountain View, CA, U.S.A.). In vitro treatment of tumor cells with M T B A To obtain MTBA-treated sublines, tumor cells were removed from continuous culture using phosphate buffered saline containing 0.02~ EDTA (Sigma, St. Louis, MO, U.S.A.) centrifuged and resuspended at a final concentration of 2 x 10 6 cells/ml in 4 ml of CM containing 100/~g/ml of MTBA. The mixture was incubated for 1 h at 37 °C in a Dubnoff metabolic shaker. The cells were then washed twice in RPMI 1640 medium and maintained in continuous culture. In vitro treatment with citric acid A significant reduction of HLA class I antigens on H125 cell surface was obtained as described by Sugawara et al., (1987). Briefly, citric acidN a z H P O 4 buffer, pH 3.0, was prepared by mixing citric acid 263mM with equal volume of Na2HPO4 123 mM containing 1~o of bovine
serum albumin (Sigma). The buffer solution (0.5 ml) was added to cell pellet (5 × 106 cells) and the mixture was incubated for 2 min at 4 ° C. Cells were then washed extensively and processed for FACS analysis or cytotoxicity assay. Immunofluorescence staining and cytofluorimetric analysis The expression of HLA-classI (ABC) and class II (DR) antigens was assayed by direct (HLA-DR) or indirect (HLA-ABC) immunofluorescence. Tumor cells were removed from continuous culture and washed twice in phosphate buffered saline supplemented with 0.1~o bovine serum albumin and 0.02~o sodium azide (Sigma) (PBS-A). Cells (1 × 106) were then suspended in 50/~1 of PBS-A containing: (a) FITCconjugated F(ab')2 rabbit anti-mouse IgG or FITC-labeled m o u s e I g G z a + b alone (negative controls); (b) FITC-labeled anti-HLA-DR mAb; and (c)anti-HLA-ABC mAb. Samples were incubated for 30 min at 0 °C and then washed twice in PBS-A. The second step of staining (HLA-ABC) was carried out by incubating cells with FITC-conjugated F(ab')2 rabbit anti-mouse IgG for 45 min at 0 ° C, and then washing twice in PBS-A. The labeled samples were assayed with a FACS Analyzer (Becton Dickinson FACS Systems) with the appropriate combinations of filters, in order to measure emission from fluorescein conjugate. Data were collected on 2 x 104 viable cells as determined by side light scatter. Data analysis was performed by using a 'Consort 30' hardware (Hewlett Packard, Fort Collins, U.S.A.) and the software supplied with the same instrument. Mononuclear cell preparation Peripheral blood mononuclear cells (MNC) were separated from buffy coats obtained from healthy donors on a Ficoll-Hypaque gradient (Boyum, 1968). The MNC collected from the interface were washed twice in RPMI 1640 medium, suspended in CM and used as effector cells in cytotoxicity assays.
202 For binding studies, M N C were deprived of adherent cells on plastic dishes and passed through nylon-wool columns.
0.1 ml of supernatant was harvested from each well, and counted in a gamma scintillation counter (Auto-gamma 800C, Packard Instruments, Downers Grove, IL, U.S.A.).
Generation of IL-2 activated killer cells (LAK) Peripheral M N C were suspended at the final concentration of 1 × 106cells/ml in R P M I 1640 supplemented with 10~o human AB serum (Flow Lab.) and 2 mM L-glutamine, alone or in the presence of 100 U/ml human rIL-2. M N C were incubated for 72 h at 37 °C and then recovered, washed twice, and used in cytotoxicity assays.
Labeling of target cells for cytotoxicity assay Aliquots of cell lines growing in suspension culture were removed from continuous culture, centrifuged, resuspended in 0.1 ml of FCS and incubated in the presence of 100/~Ci of Na2 51CRO4 (Amersham International plc, Amersham, U.K.) for 1 h at 37 °C. To remove H125 tumor cells from continuous culture, culture medium was discarded and 3 ml of phosphate buffered saline containing 0.02~o EDTA were added to the adherent cells. After incubation of 5-10 rain at 37 ° C, the cells were recovered, washed in R P M I 1640 medium and labeled as described above. After incubation, the cells were extensively washed in R P M I 1640 medium and resuspended in CM for cytotoxicity assays.
Cytotoxicity assay The cytotoxic activity of M N C was determined using a 5~Cr-release assay. Effector cells in 0.1 ml of CM were plated in quadruplicate in U-bottom 96-microtiter plates (Greiner C.A., and Sohne, Nurtingen, F.R.G.) by making serial two-fold dilutions, starting at a concentration of 2 x 106cells/ml. Labeled target cells (2 x 103) were then added in a volume of 0.1 ml to give a final volume of 0.2 ml and an effector (E) to target (T) cell ratio ranging from 100 : 1 to 12.5 : 1. The plates were then centrifuged at 80 x g for 5 min and incubated for 4 h at 37 °C in a 5~o CO2 humidified atmosphere. After incubation the plates were centrifuged at 250 x g for 10 min and
CoM competition assay The ability of cold (i.e. unlabeled) tumor cells to compete with 5~Cr-labeled cells in binding effector M N C was assessed in a modification of the 5~Crrelease assay. Graded numbers (0.25-2 x 105) of effector M N C suspended in 0.1 ml of CM were plated in U-bottom 96-well microtiter plates. A mixed suspension of cold competitor cells (1 x 104) and labeled target cells (2 z 103) in 0.1 ml of CM were then added to each well. The E : T ratio did not differ from that of the standard 5~Cr-release assay and the competitor to target ratio was 5:1. The plates were then handled as described above.
Calculation of percentage of specific target ceil lysis The percentage of specific lysis was calculated as follows:
% specific lysis
Test cpm - Autologous cpm Total cpm x 0.5
× 100
where 'Test cpm' is the mean cpm relative to 51Cr released in presence of effector cells; 'Autologous cpm' is the mean cpm relative to 5~Cr released by target cells incubated with unlabeled autologous cells in place of effector cells, and 'Total cpm' is the mean cpm of 2 × 103 51Cr labeled target cells (i.e. total radioactivity incorporated by target cells).
Evaluation of cytotoxic activity of effector cells in terms of "killed cells' (KC) Aim of data processing for cytotoxic assays is to express the overall cytotoxic activity of effector cells using a single value instead of the percentages of specific lysis at different E : T cell ratios. Many investigators utilize the lytic unit (LU) value, as the amount of effector cells
203 required to obtain n~o lysis. This value is extrapolated from the best-fit curve obtained by plotting the percentage of specific lysis (n ~o) versus lnx, where x is the number of effector cells per well (Thorn and Herney, 1976). However L U does not appear to be convenient when the cytolytic activity of different effector cells is extremely variable. In this case L U should be selected using different n values according to the level of cytotoxicity of different and relatively homogenous series ofeffector cells. Since L U values calculated on different n values are not directly comparable, in the present investigation the overall cytotoxic activity of effector cells was calculated using a different data processing technique valid for any level of cytolytic effects. Therefore the data were expressed in terms of number of target cells lysed (i.e. 'killed') by a fixed number of effector cells, as follows:
KC(m) -
mxTxn% En × 100
where KC(m) is the number of target cells killed by 'm' effector cells; 'T' is the total number of target cells present in each well; 'En' is the number of effector cells present in each well at the selected E : T cell ratio, and 'n ~o' is the specific lysis produced theoretically by En effector cells. The 'n %' value is extrapolated from the best-fit curve obtained by plotting the different percentages of specific lysis vs. In of the number of effector cells/well. In the present study KC(m) was calculated for m = 106cells (i.e. KC(106)), and E n = 2 × 105 cells (i.e. at E : T cell ratio of 100: 1).
Effector-target cell binding assay Nylon wool-passed M N C were suspended at a concentration of 2 x 10 6 cells/ml, and 0.1 ml was mixed with an equal volume of target cells (2 × 106 cells/ml) in U-bottom 96-well microtiter plates. Plates were centrifuged at 80 × g for 5 min and incubated for 30 min at 4 °C. The number of M N C binding to target cells was determined in
quadruplicate by microscopic observation in a hemocytometer (Sugawara et al., 1989). M N C were distinguished from tumor cells on size basis, target cells diameter being roughly five-fold that of MNC.
Statistical analysis Differences in cytolytic effects produced by effector cells in various experimental conditions were evaluated taking into account the percentage of specific cytotoxicity at all E : T ratios. Therefore, p values were calculated using co-variance analysis performed on the regression of the percentage of specific 51Cr-release over the In of the number of effector cells/well. All data relative to cellmediated cytolysis are expressed in terms of mean KC(10 6) values without conventional standard error or standard deviation of the mean. Actually no statistical analysis can be performed using these parameters which are not suitable for covariance analysis of regression lines. Differences in effector-target cell binding were evaluated using Student's t-test analysis.
Results
Susceptibility of parental H125 and MTBA-treated H125/M6 tumor lines to cytotoxicity mediated by NK or L A K cells Parental and MTBA-treated lines were tested for their susceptibility to N C M C using allogeneic M N C obtained from healthy donors. The results of a representative experiment with N K effector cells are illustrated in Fig. 1. The data, expressed in terms of percentage of specific lysis at different E : T ratios, and in terms of KC(106) at the E : T ratio of 100:1, show that MTBA-treated H125/M6 cells were significantly more susceptible than parental line to NK-mediated cytotoxicity. Similar results were obtained when the two lines were used as targets of LAK cells (Fig. 2). To analyze whether the difference in sensitivity to N K or LAK cells might be due to different effector-target binding, cold competition and
204 40-
C4': C0 P
199
IMPHAR 00545
Drug-mediated increase of susceptibility of human lung cancer to NK or LAK effector cells Stefania D'Atri 1, Stefano Marini 2, Lucio Tentori 2, M a r i a Tricarico and E n z o B o n m a s s a r 2
I,
M a r i a Pia Fuggetta 1
llnstitute of Experimental Medicine, CNR, Rome and 2Department of Experimental Medicine and Biochemical Sciences, 11 University of Rome, Rome, Italy (Received 21 August 1990; accepted 25 February 1991)
Abstract: Previous studies in murine models have shown that in vivo or in vitro treatment of tumor cells with mutagenic triazene compounds (TZC) lead to the appearance of novel drug-mediated tumor antigens (DMTA) capable of eliciting graft resistance in syngeneic hosts. This phenomenon, defined as 'chemical xenogenization' (CX), could be of potential value for immunochemotherapy of human neoplasias. It was also shown that TZC modulate NK sensitivity of murine tumor cells. Therefore, experiments were conducted to evaluate whether susceptibility of a human lung adenocarcinoma cell line (H 125) to natural cytotoxic effectors could be affected by treatment with an in vitro active TZC. The results showed that drug treatment of H125 line leads to heritable increase of susceptibility to NK and LAK cells. Moreover, increased binding between effector and drug-treated target cells was observed. Additional studies on HLA antigens showed that changes in HLA-ABC molecule expression were probably not involved in TZC-induced increase of NK/LAK susceptibility. These results suggest that TZC treatment of a human tumor could result in increased expression of membrane structures recognized by natural cytotoxic effector cells. Further studies are required to explore whether these changes are generated by mutational events correlated with TZC-induced CX of human cancer cells. Key words:
Chemical xenogenization; LAK susceptibility; NK susceptibility; Triazene compounds
Correspondence: Stefania D'Atri, Department of Experimental Medicine and Biochemical Sciences, II University of Rome, Via O. Raimondo, 00173 Rome, Italy. Abbreviations: CM, complete medium; cpm, counts per minute; CX, chemical xenogenization; DMTA, drug-mediated tumor antigens; DTIC, 5-(3,3-dimethyl-l-triazeno) imidazole-4-carboxamide; E, effector; EDTA, ethylenediamine-tetraacetic acid; FCS, fetal calf serum; FITC, fluorescein isothiocyanate; IL-1, interleukin-1; IL-2, interleukin-2; KC, killed cells; LAK, IL-2 activated killer cells; LUn, lytic unit n; mAb, monoclonal antibody; MFV, mean fluorescence value; MHC, major histocompatibility complex; MNC, mononuclear cells; MTBA, p-(3-methyl1-triazeno)benzoic acid potassium salt; NCMC, natural cellmediated cytotoxicity; NK, natural killer; NKCF, NK cyto-
Introduction
A large number of experimental studies demonstrated that in vivo (Bonmassar et al., 1970; Bonmassar et al., 1972; Bonmassar et al., 1979) or in vitro (Contessa etal., 1981) treatment of murine lymphomas with the antitumor drug 5- (3,3- dimethyl- 1-triazeno)imidazole-4-carbox-
toxic factors; NKTS, NK target structures; PBS, phosphate buffered saline; T, target; TNF, tumor necrosis factor; TZC, triazene compounds.
200 amide (DTIC), or other triazene compounds (TZC) (Nardelli et al., 1984), leads to the appearance of novel drug-mediated tumor antigens (DMTA). These antigens are capable of eliciting T cell-mediated proliferative and cytotoxic responses in vitro (Romani et al., 1979), specific transplantation resistance (Riccardi et al., 1978), and specific tolerance (Houchens et al., 1976) in syngeneic hosts. Moreover recent studies pointed out that a D M T A expressed by a DTIC-treated murine leukemia, is a surface glycoprotein of approximately 80 kDa, possibly related to retroviral gp70 molecules (Grohman et al., 1990). This phenomenon, described as 'chemical xenogenization' (CX) (Puccetti et al., 1985), similar to that obtained with other methylating mutagenic agents (Boon and Kellerman, 1977; Boon, 1983) could lead to new perspectives in cancer immunochemotherapy. On these bases it was of primary interest to investigate whether TZC-mediated CX could occur in human neoplasias. Previous studies in murine models (Romani et al., 1983) showed that treatment of tumor cells with TZC modulates their susceptibility to N K effector cells. Therefore experiments were carried out on a human cell line (i.e. lung adenocarcinoma H 125), to test whether its susceptibility to natural cell-mediated cytotoxicity (NCMC) could be affected by treatment with an in vitro active TZC, i.e. the potassium salt ofp-(3-methyl-l-triazeno) benzoic acid (MTBA) (Fioretti et al., 1983). Parental H125 cells, and the MTBA-treated subline H125/M6 (i.e. H125 cells subjected to 6 cycles of drug treatment) were comparatively tested for their susceptibility to NCMC, using 2 types of effector cells, i.e. fresh mononuclear cells from peripheral blood, or IL-2 activated killer cells (LAK cells) (Grimm et al., 1982; Grimm et al., 1983; Ortaldo et al., 1986; Phillips and Lanier, 1986; Hersey and Bolhuis, 1987). Furthermore it was investigated whether MTBA treatment could induce changes in the expression of HLA antigens, since an inverse relationship between susceptibility to N C M C and the extent of M H C antigen expression has been pro-
posed by a number of investigators (Ljunggren and Karre, 1985; Harel-Bellan et al., 1986; Karre et al., 1986; Storkus et al., 1987; Quillet et al., 1988). The results presented here show that MTBA treatment of H 125 line increased its susceptibility to N C M C and binding to effector lymphocytes. Moreover drug treatment produced a limited decrease of HLA-class I antigen expression. On the other hand, marked reduction of the density of class I histocompatibility molecules induced by citric acid (Sugawara et al., 1987) in H125 cells did not result in increased NCMC susceptibility of the tumor. These data suggest that MTBAmediated increase of H125 sensitivity to natural effector ceils is presumably the result of increased density or availability of target structures recognized by NK or LAK cells, not correlated with changes in HLA antigen expression. Methods Cell lines All cell lines were maintained in RPMI 1640 medium (GIBCO, Paisley, Scotland, U.K.) supplemented with 10~o heat inactivated fetal calf serum (FCS, GIBCO) and 2 mM L-glutamine (Flow Laboratories, McLean, VA, U.S.A.) (hereafter referred to as 'Complete Medium', CM). The erythroteukemic cell line K562 (Lozzio and Lozzio, 1979) and the Burkitt lymphoma cell line Daudi (Klein et al., 1968) were grown in suspension culture and subcultured three times weekly. Cell line H125, a human lung adenocarcinoma line growing in monolayer, kindly provided by Dr. J.D. Minna (National Cancer Institute, NIH, Navy Medical Oncology Branch, Bethesda, MD, U.S.A.), drug-treated H125/M6 subline and cell clones were also grown in CM and subcultured twice a week. In vitro cloning of tumor cells Parental or TZC-treated line were cloned by limiting dilution in flat-bottom 96-well microtiter plates (Falcon, Becton Dickinson Labware, Oxnard, CA, U.S.A.) at 0.5 cells/well in CM.
201 Reagents p-(3-Methyl- 1-triazeno)benzoic acid potassium salt (MTBA), was generously provided by Prof. L. Lassiani (Institute of Pharmaceutical Chemistry, University of Trieste, Trieste, Italy). The drug was dissolved in ice-cold RPMI 1640 protected from light, sterilized through 0.22 #m size filter (Millipore, Molsheim, France) and added to cell suspension for in vitro treatment. All procedures were performed within 1 min in order to avoid drug inactivation, being the half-life of MTBA in cold medium less than 9 min (Lassiani L., personal communication). Recombinant human interleukin-2 (rlL-2) was kindly provided by Dr. D. Criscuolo (Hoffmann La Roche, Milan, Italy). Purified anti-HLA-ABC (IgGaa) monoclonal antibody (mAb) from W6/32 hybridoma (Barnstaple et al., 1978; Muller et al., 1983) and fluorescein isothiocyanate (FITC)conjugated F(ab')2 rabbit anti-mouse IgG were obtained from Dako (Dakopatts, Copenhagen, Denmark). FITC-conjugated anti-HLA-DR from L243 hybridoma (Lampson et al., 1980; Brodsky, 1984) and FITC-labeled mouse IgG2a+b were obtained from Becton Dickinson (Mountain View, CA, U.S.A.). In vitro treatment of tumor cells with M T B A To obtain MTBA-treated sublines, tumor cells were removed from continuous culture using phosphate buffered saline containing 0.02~ EDTA (Sigma, St. Louis, MO, U.S.A.) centrifuged and resuspended at a final concentration of 2 x 10 6 cells/ml in 4 ml of CM containing 100/~g/ml of MTBA. The mixture was incubated for 1 h at 37 °C in a Dubnoff metabolic shaker. The cells were then washed twice in RPMI 1640 medium and maintained in continuous culture. In vitro treatment with citric acid A significant reduction of HLA class I antigens on H125 cell surface was obtained as described by Sugawara et al., (1987). Briefly, citric acidN a z H P O 4 buffer, pH 3.0, was prepared by mixing citric acid 263mM with equal volume of Na2HPO4 123 mM containing 1~o of bovine
serum albumin (Sigma). The buffer solution (0.5 ml) was added to cell pellet (5 × 106 cells) and the mixture was incubated for 2 min at 4 ° C. Cells were then washed extensively and processed for FACS analysis or cytotoxicity assay. Immunofluorescence staining and cytofluorimetric analysis The expression of HLA-classI (ABC) and class II (DR) antigens was assayed by direct (HLA-DR) or indirect (HLA-ABC) immunofluorescence. Tumor cells were removed from continuous culture and washed twice in phosphate buffered saline supplemented with 0.1~o bovine serum albumin and 0.02~o sodium azide (Sigma) (PBS-A). Cells (1 × 106) were then suspended in 50/~1 of PBS-A containing: (a) FITCconjugated F(ab')2 rabbit anti-mouse IgG or FITC-labeled m o u s e I g G z a + b alone (negative controls); (b) FITC-labeled anti-HLA-DR mAb; and (c)anti-HLA-ABC mAb. Samples were incubated for 30 min at 0 °C and then washed twice in PBS-A. The second step of staining (HLA-ABC) was carried out by incubating cells with FITC-conjugated F(ab')2 rabbit anti-mouse IgG for 45 min at 0 ° C, and then washing twice in PBS-A. The labeled samples were assayed with a FACS Analyzer (Becton Dickinson FACS Systems) with the appropriate combinations of filters, in order to measure emission from fluorescein conjugate. Data were collected on 2 x 104 viable cells as determined by side light scatter. Data analysis was performed by using a 'Consort 30' hardware (Hewlett Packard, Fort Collins, U.S.A.) and the software supplied with the same instrument. Mononuclear cell preparation Peripheral blood mononuclear cells (MNC) were separated from buffy coats obtained from healthy donors on a Ficoll-Hypaque gradient (Boyum, 1968). The MNC collected from the interface were washed twice in RPMI 1640 medium, suspended in CM and used as effector cells in cytotoxicity assays.
202 For binding studies, M N C were deprived of adherent cells on plastic dishes and passed through nylon-wool columns.
0.1 ml of supernatant was harvested from each well, and counted in a gamma scintillation counter (Auto-gamma 800C, Packard Instruments, Downers Grove, IL, U.S.A.).
Generation of IL-2 activated killer cells (LAK) Peripheral M N C were suspended at the final concentration of 1 × 106cells/ml in R P M I 1640 supplemented with 10~o human AB serum (Flow Lab.) and 2 mM L-glutamine, alone or in the presence of 100 U/ml human rIL-2. M N C were incubated for 72 h at 37 °C and then recovered, washed twice, and used in cytotoxicity assays.
Labeling of target cells for cytotoxicity assay Aliquots of cell lines growing in suspension culture were removed from continuous culture, centrifuged, resuspended in 0.1 ml of FCS and incubated in the presence of 100/~Ci of Na2 51CRO4 (Amersham International plc, Amersham, U.K.) for 1 h at 37 °C. To remove H125 tumor cells from continuous culture, culture medium was discarded and 3 ml of phosphate buffered saline containing 0.02~o EDTA were added to the adherent cells. After incubation of 5-10 rain at 37 ° C, the cells were recovered, washed in R P M I 1640 medium and labeled as described above. After incubation, the cells were extensively washed in R P M I 1640 medium and resuspended in CM for cytotoxicity assays.
Cytotoxicity assay The cytotoxic activity of M N C was determined using a 5~Cr-release assay. Effector cells in 0.1 ml of CM were plated in quadruplicate in U-bottom 96-microtiter plates (Greiner C.A., and Sohne, Nurtingen, F.R.G.) by making serial two-fold dilutions, starting at a concentration of 2 x 106cells/ml. Labeled target cells (2 x 103) were then added in a volume of 0.1 ml to give a final volume of 0.2 ml and an effector (E) to target (T) cell ratio ranging from 100 : 1 to 12.5 : 1. The plates were then centrifuged at 80 x g for 5 min and incubated for 4 h at 37 °C in a 5~o CO2 humidified atmosphere. After incubation the plates were centrifuged at 250 x g for 10 min and
CoM competition assay The ability of cold (i.e. unlabeled) tumor cells to compete with 5~Cr-labeled cells in binding effector M N C was assessed in a modification of the 5~Crrelease assay. Graded numbers (0.25-2 x 105) of effector M N C suspended in 0.1 ml of CM were plated in U-bottom 96-well microtiter plates. A mixed suspension of cold competitor cells (1 x 104) and labeled target cells (2 z 103) in 0.1 ml of CM were then added to each well. The E : T ratio did not differ from that of the standard 5~Cr-release assay and the competitor to target ratio was 5:1. The plates were then handled as described above.
Calculation of percentage of specific target ceil lysis The percentage of specific lysis was calculated as follows:
% specific lysis
Test cpm - Autologous cpm Total cpm x 0.5
× 100
where 'Test cpm' is the mean cpm relative to 51Cr released in presence of effector cells; 'Autologous cpm' is the mean cpm relative to 5~Cr released by target cells incubated with unlabeled autologous cells in place of effector cells, and 'Total cpm' is the mean cpm of 2 × 103 51Cr labeled target cells (i.e. total radioactivity incorporated by target cells).
Evaluation of cytotoxic activity of effector cells in terms of "killed cells' (KC) Aim of data processing for cytotoxic assays is to express the overall cytotoxic activity of effector cells using a single value instead of the percentages of specific lysis at different E : T cell ratios. Many investigators utilize the lytic unit (LU) value, as the amount of effector cells
203 required to obtain n~o lysis. This value is extrapolated from the best-fit curve obtained by plotting the percentage of specific lysis (n ~o) versus lnx, where x is the number of effector cells per well (Thorn and Herney, 1976). However L U does not appear to be convenient when the cytolytic activity of different effector cells is extremely variable. In this case L U should be selected using different n values according to the level of cytotoxicity of different and relatively homogenous series ofeffector cells. Since L U values calculated on different n values are not directly comparable, in the present investigation the overall cytotoxic activity of effector cells was calculated using a different data processing technique valid for any level of cytolytic effects. Therefore the data were expressed in terms of number of target cells lysed (i.e. 'killed') by a fixed number of effector cells, as follows:
KC(m) -
mxTxn% En × 100
where KC(m) is the number of target cells killed by 'm' effector cells; 'T' is the total number of target cells present in each well; 'En' is the number of effector cells present in each well at the selected E : T cell ratio, and 'n ~o' is the specific lysis produced theoretically by En effector cells. The 'n %' value is extrapolated from the best-fit curve obtained by plotting the different percentages of specific lysis vs. In of the number of effector cells/well. In the present study KC(m) was calculated for m = 106cells (i.e. KC(106)), and E n = 2 × 105 cells (i.e. at E : T cell ratio of 100: 1).
Effector-target cell binding assay Nylon wool-passed M N C were suspended at a concentration of 2 x 10 6 cells/ml, and 0.1 ml was mixed with an equal volume of target cells (2 × 106 cells/ml) in U-bottom 96-well microtiter plates. Plates were centrifuged at 80 × g for 5 min and incubated for 30 min at 4 °C. The number of M N C binding to target cells was determined in
quadruplicate by microscopic observation in a hemocytometer (Sugawara et al., 1989). M N C were distinguished from tumor cells on size basis, target cells diameter being roughly five-fold that of MNC.
Statistical analysis Differences in cytolytic effects produced by effector cells in various experimental conditions were evaluated taking into account the percentage of specific cytotoxicity at all E : T ratios. Therefore, p values were calculated using co-variance analysis performed on the regression of the percentage of specific 51Cr-release over the In of the number of effector cells/well. All data relative to cellmediated cytolysis are expressed in terms of mean KC(10 6) values without conventional standard error or standard deviation of the mean. Actually no statistical analysis can be performed using these parameters which are not suitable for covariance analysis of regression lines. Differences in effector-target cell binding were evaluated using Student's t-test analysis.
Results
Susceptibility of parental H125 and MTBA-treated H125/M6 tumor lines to cytotoxicity mediated by NK or L A K cells Parental and MTBA-treated lines were tested for their susceptibility to N C M C using allogeneic M N C obtained from healthy donors. The results of a representative experiment with N K effector cells are illustrated in Fig. 1. The data, expressed in terms of percentage of specific lysis at different E : T ratios, and in terms of KC(106) at the E : T ratio of 100:1, show that MTBA-treated H125/M6 cells were significantly more susceptible than parental line to NK-mediated cytotoxicity. Similar results were obtained when the two lines were used as targets of LAK cells (Fig. 2). To analyze whether the difference in sensitivity to N K or LAK cells might be due to different effector-target binding, cold competition and
204 40-
C4': C0 P
