Immunobiol., vol. 185, pp. 63-77 (1992)
Department of Pathology, University of Southern California, Los Angeles, California, USA
Resistance of Different Tumor Cells to Lysis by Lymphokine Activated Killer Cells Can Be Mediated by Distinct Mechanisms PAMELA BEAN and AMITABHA MAZUMDER 1 Received September 24, 1991 . Accepted in Revised Form February 14, 1992
Abstract Lymphokine activated killer (LAK) cells have been shown to exert a potent cytotoxic effect on many histologically different tumors and virally infected targets. Most normal cells but very few tumors have proven resistant to LAK lysis. The availability of two LAK resistant tumors, P815r, a murine mastocytoma, and SNUC-1, a human colon carcinoma, allowed us to study the phenomenon of LAK lysis. We examined the role of surface molecules on targets, which mediate binding to LAK cells, by cold target competition experiments and lectin dependent cellular cytotoxicity assays.The results showed that in the murine system, P815r cells do not compete for lysis of the LAK sensitive target B16 whereas other LAK sensitive murine targets compete. Alternatively, in the human system, SNUC-1 cells compete for lysis of the LAK sensitive target SNUC-4 as do other LAK sensitive human tumor cells. Furthermore, inducing binding of target and effector cells with lectin reverted the resistance of P815r but not SNUC-1 targets to lysis by LAK cells. These results imply that distinct stages of the lytic pathway might be involved in the resitance of different tumors to killing by LAK cells. The murine cell line is resistant to lysis because it cannot bind LAK cells. The human target, which does bind LAK, was insensitive to the effects of tumor necrosis factor alpha (TNF-a), a Iymphokine released by LAK effectors and possibly involved in their lysis. Resistance to TNF-a was not mediated by the presence of endogenous short-lived proteins in the SNUC-1 targets. The elucidation of mechanisms of resistance may provide a tool to improve current protocols of adoptive immunotherapy as well as insights as to how tumor cells are or are not killed by LAK effectors.
Introduction Culture of normal non-immune lymphocyte populations with interleukin-2 (IL-2) results in the generation of lymphokine activated killer (LAK) cells cytotoxic for a wide variety of histologically distinct tumors (1, 2). Previous reports in the literature have indicated that there is no apparent specificity for lysis of the autologous tumor because allogeneic as well as xenogeneic fresh or cultured tumors are lysed (3). Moreover, virus infected 1 Present address: Division of Medical Oncology, Georgetown University School of Medicine, Washington, D.C. 20007, USA
This work was supported by a grant from the Hastings Foundation
64 . P.
BEAN
and A.
MAZUMDER
targets (4) and single cell suspensions of placental and fetal tissues are also sensitive to LAK effectors (2). Interestingly, we have found murine (P815r) and human (SNUC-l) tumor cell lines which are not killed by LAK cells and their mechanisms of resistance constitute the subject of this study. The first step in LAK mediated cytolysis involves recognition and binding of target and effector cells. Blocking cell surface epitopes with antibodies (5, 6) or digesting cell membrane molecules with proteases abolishes LAK lysis (7). We therefore investigated the binding capacities of LAK resistant targets by means of in vitro competition experiments and lectin-dependent cellular cytotoxicity (LDCC) assays. For the in vitro competition assays we assumed that if cell membrane molecules present in LAK sensitive tumor cells were indeed responsible for their lysis by LAK effectors, the presence of unlabeled LAK resistant targets (presumably lacking ligand epitopes for LAK cells) should not interfere with killing of chromated LAK sensitive targets. Alternatively, in the lectin-dependent cellular cytotoxicity assay, the effector and target cells are brought into contact through the use of lectins. If the resistance of the targets is due to their incapacity to bind LAK cells, the presence of the lectin Concanavalin A (ConA) should overcome lack of lysis. We found that the resistant human and murine tumor cell lines possessed different abilities to compete in these assays. The murine cell line could not bind LAK cells while the human cell line did bind LAK cells and therefore is deficient in a different step(s) of the lytic pathway. It has been suggested that not only cell-to-cell contact but also the release of cytotoxic factors may be involved in the cytolytic event mediated by LAK cells (8, 9). LAK cells release the cytokines interferon gamma (IFN -y) and tumor necrosis factor alpha (TNF-a) when stimulated with some types of tumor cells (9,10). However, TNF-a is not equally cytotoxic to all tumor cell and certain kinds of tumor cells (and most normal cells) are known to be resistant to TNF-a (11). The mechanism(s) involved in the acquisition of TNF-a resistance is still controversial, but inhibition of RNA or protein synthesis makes normal cells and TNF-a-resistant tumor cells become sensitive to TNF (12). Thus, the defense mechanism of the cell most likely involves endogenous protective proteins (13). Based on these findings, we tested the sensitivity of SNUC-l cells to TNF-a and analyzed the effects of actinomycin D upon this sensitivity. Resistance to LAK cells has been suggested as a means whereby tumors may escape immune destruction and a greater understanding of sensitivity and resistance to this phenomena might have implications in improving immunotherapy with LAK cells. Materials and Methods Cells and culture media The murine tumor target cells employed in this study were B16 melanoma, obtained from the tumor repository at the Frederick Cancer Research Facility, YAC-l lymphoma, PS1S
Differential Tumor Resistance to Activated Killer Cells . 65 mastocytoma and WEHr 164, a methylcholanthrene-induced fibrosarcoma, were obtained from the American Type Culture Collection (ATCe). MCA methylcholanthrene induced sarcoma was obtained from NIH. The P815r cells represent a subclone of the parental P815 cell line which displayed resistance to lysis by LAK cells. WEHI164 is highly sensitive to human TNF-a and LAK. The human target cells used were: peripheral blood leukocytes (PBL) derived from normal donors; SNUC-4 and SNUC-1, both colon tumors (kindly provided by Dr. SOO-SANGJOHN, Keimyung University School of Medicine, Taegu, South Korea); and clone #83, derived from the undifferentiated lymphoma line JD40 (14) (kindly supplied by Dr. CARLO GAMBACORTIPASSERINI, Department of Human Oncology, University of Wisconsin, Madison, WI, USA). As previously reported (15), SNUC-1 cells are resistant to lysis by LAK cells while SNUC-4 cells are sensitive. All cell lines except MCA were grown to log phase (3-5 x 106 cells/ml) in complete media (CM) and were stored as cryopreserved cells to provide a standard stock for assay. The preparation of CM and the passage of MCA cells in vivo have been described (1). Generation of murine and human LAK cells Generation of murine LAK and NK cells was performed as described (1). Spleens from C57B!/6 mice 8 weeks or older were used in all experiments. Human LAK cells were prepared by thawing stocks of peripheral blood lymphocytes (PBL) obtained by leukapheresis of blood derived from normal donors. Cell suspensions of 106 PBL/ml were incubated for 5 days in 1000 units/ml of recombinant interleukin-2 (rIL-2) in CM. Cytotoxicity assays Four hour chromium release assays have been described elsewhere (16). For assay of TNF-a cytotoxicity,S x 10 3 51Cr-labeled targets were placed in each well of a 96-well plate and incubated with different concentrations of TNF-a (0 to 104 units/ml). The percentages of cytotoxicity were calculated as in a standard cytotoxicity assay (16). In studies that measured tumor cell sensitivity to TNF-a after inhibition of RNA synthesis, chromated targets were incubated for 6 h with 0.2 or 1.0 fAg/ml of Actinomycin D (Sigma, St. Louis, MO, USA) in the presence of the same concentrations of TNF-a described above. Results are expressed as the mean of triplicate measurements. Cold target competition assays Cold target competition assays using 51Cr-labeled murine B16 target cells were performed as follows: 2.5 x 105 LAK effector cells in 100 fAl CM were plated in each well of a round bottom microplate with various numbers of unlabeled murine target cells: 105,5 x 10\ 2.5 X 104 and 5 X 103 cells per well, representing cold:hot (C:H) ratios of 20, 10, 5 and 1 unlabeled (cold) competitor per each labeled (hot) target cell, respectively. After 30 minutes, wells were inoculated with 5 x 103 slCr-labeled B16 cells and the plate was incubated at 37°C,S % CO 2 for 3 h. The percentage of lysis induced by LAK cells against labeled B16 in the presence or absence of unlabeled targets (B16, YAC-1, P815 and MCA) was then determined. In the human system, the same procedure was followed using SNUC-4 cells as chromated targets. The unlabeled competitors were either SNUC-4, SNUC-l or PBL targets. The percent inhibition was calculated as: [1 - (% lysis in the absence of competitor!% lysis in the presence of competitor)] x 100. Lectin dependent cellular cytotoxicity assay (LDCC) LDCC was done by incubating chromated targets with a 50: 1 ratio of LAK effectors to target cells (E:T) in the presence of different concentrations (2.5 and 5 fAg/ml) of the lectin Concanavalin A (Sigma). Determinations of tumor cell lysis were done as in a standard cytotoxicity assay (16).
66 . P. BEAN and A. MAZUMDER
Binding of Fluorescein Isothiocyanate (FITC) labeled Con A to tumor cells As previously described (17), cultured cells (SNUC-1, clone 83 and P815r) resuspended at 106 cells/ml were layered onto sterile coverslips and incubated for 48 h at 37°C, 5 % CO 2 in CM. The coverslips were rinsed once in PBS and the cells were fixed by soaking in 3.7 % formaldehyde for 5 minutes. FITC conjugated Con A (100 f.tl) at concentrations of 0.5,5.0 and 20.0 f.tg/ml was then added. Control cells were incubated with PBS alone. The coverslips were incubated at room temperature in a humidified chamber for 15 minutes and were then rinsed in PBS before mounting onto glass slides with 80 % glycerol.
Results
Patterns of cytotoxicity mediated by LAK cells against a panel of tumor targets Three and five days of lymphokine activation were chosen in the murine and human systems, respectively, as they showed the highest LAK activity. The results of cytotoxicity by LAK cells against four tumor cell lines are shown in Table 1. Spleen LAK cells killed the B16 cell line with an average efficiency of 32 % (over three separate experiments, each done in triplicate) at a 50:1 E:T cell ratio. In contrast, under the same conditions, the levels of cytotoxicity exerted by LAK cells against P815r targets averaged 5.9 % lysis in three separate experiments. As a background, lysis of P815r targets by unstimulated splenocytes reached an average of 2.5 % at the 50:1 E:T ratio in 2 experiments (data not shown). In our laboratory, killing of this subclone of P815 has been achieved, but only with antibody against major histocompatibility complex antigen plus complement. Murine spleen LAK cells killed YA C-1 and M CA targets with an average efficiency of 68 % and 49 % respectively at the 50:1 E:T ratio in four experiments (data not shown). The effects of seven day human LAK cells on the human colon tumors SNUC-4 and SNUC-1 are presented in Table 1. The SNUC-4 cell line is killed by LAK cells with an average efficiency of 33 % lysis while the cytotoxicity exerted against SNUC-1 averaged 4.5 % at the same 50:1 E:T ratio. In three separate experiments unstimulated PBLs (NK) displayed a similar (3.5 %) cytotoxic effect on SNUC-1 targets (data not shown). The effects of human LAK on resting PBL, a population of normal cells which are not killed by LAK cells, averaged 1.7 % lysis in three experiments.
Competition for lysis of labeled B16 targets by spleen LAK cells, in the presence of unlabeled murine tumor cells Cold YAC-1 cells greatly decreased (78 % inhibition) B16 lysis while unlabeled P815r induced only 16 % inhibition at the 10:1 C:H ratio (Fig. 1a). As expected, cold B16 targets acted as competitors for their own lysis by inhibiting 75 % of the effect exerted by LAK cells alone, i.e., in the absence of unlabeled targets. MCA and YAC-1 cells induced 67 % and 91 % inhibition of B16lysis, respectively, at the 10: 1 C:H ratio (Fig. 1b). At an equivalent ratio, P815r cells inhibited B16 killing by 25 %. Overall, the
Differential Tumor Resistance to Activated Killer Cells . 67
percentages of inhibition exerted by cold P815r cells on chromated B16 targets averaged 15 % in four experiments and they differed from those found for the other three unlabeled targets by more than three standard deviations. Thus, P815r cells did not compete for lysis of B 16 targets as did the LAK sensitive YAC-1 and MCA tumor cell lines.
Competition for lysis of labeled SNUC-4 cells by human LAK cells in the presence of unlabeled human cells The lysis of SNUC-4 cells was unaffected by resting PBL (Fig.2). In contrast, LAK resistant SNUC-1 tumor cells inhibited 65 % of SNUC-4 cells lysis at 20: 1 C:H ratio. The inhibition exerted by cold SNUC-1 cells was very similar to the one displayed by unlabeled SNUC-4 (self) targets (70 % inhibition at an equivalent C:H ratio). The percentages of inhibition exerted by cold SNUC-1 cells over chromated SNUC-4 targets averaged 66 % in three experiments and differed from those obtained for unlabeled SNUC-4 by less than one standard deviation. Thus, lysis of LAK sensitive SNUC-4 cells was inhibited by self and by LAK resistant SNUC-1 targets, but not by normal PBL.
Effects of lectin in LAK mediated cytotoxicity of LAK resistant targets The addition of 5.0 flg/ml Con A increased LAK mediated lysis of P815r targets from 1.8 % to 17.8 % and from 0.8 % to 16.7 % respectively at the Table 1. Patterns of cytotoxicity mediated by LAK cells against a panel of targets a Target
% Cytotoxicity
E:T Exp. 1
Exp.2
Exp. 3
B16
50:1 10:1 1: 1
39.5 ± 2.9 19.5±0.9 3.7 ± 0.9
26.3 ± 2.8 8.0 ± 1.5 0.7 ± 0.7
30.5 ± 1.7 11.7± 1.1 2.7 ± 0.7
P815r
50:1 10: 1 1: 1
5.4 ± 1.8 -4.0 ± 1.2 -2.1 ± 0.9
4.7 ± 1.4 0.7 ± 0.2 0.1 ± 0.5
7.6 ± 2.3 6.4 ± 2.2 -0.9 ± 2.8
SNUC-4
50:1 10: 1 1: 1
25.6 ± 2.6 12.1 ± 2.8 3.0 ± 0.4
41.2±1.5 11.6 ± 1.5 2.3 ± 0.6
30.7 ± 1.3 10.1 ±0.6 3.4 ± 1.0
SNUC-1
50:1 10: 1 1: 1
4.6 ± 1.1 3.9 ± 0.8 3.9 ± 0.8
2.5 ± 0.2 0.2 ± 0.4 1.8 ± 0.3
6.5 ± 0.5 3.2 ± 2.6 1.5 ± 1.4
PBL
50:1 10: 1 1: 1
4.4 ± 2.2 1.5 ± 7.4 3.7 ± 1.5
1.5 ± 2.9 -1.8 ± 0.4 -3.3 ± 0.2
-0.8 ± 0.5 -2.8 ± 0.5 -0.7 ± 1.1
a
Results are reported as the mean ± SEM of the triplicate measurements. Statistical analysis was performed by the student's t test. A conservative cut-off number for a one-sided 0.05 level test for 2 means, each based on 3 observations.
68 . P.
BEAN
and A.
MA ZUMDER
P815r YAC-1 816
50 % lysis of 816 targets b
---
unlabeled competitors
60
40
41--- •
.. --at
30 20 10
o
o
5
-'
15
10
unlabeled targets per labeled 816 cell unlabeled compet itors
:~:' I:: I YAC-1
% lysis of
816 targets b
... - - .
30
10 --1-_ _ _ _ _ - - - '
5
10
15
unlabeled targets per labeled 816 cell
Figure 1. Competition for lysis of labeled B16 targets by spleen LAK cells in the presence of unlabeled murine tumor cells.' a In each experiment, the percentages of cytotoxicity exerted by LAK cells against the targets used as cold competitors have been indicated in the «y" axis. bStandard error was < 10 % of the mean in all experiments.
50:1 E:T ratio in the two experiments (Fig. 3). As a control, lysis of B16 cells was 24.5 % in the absence and 54.8 % in the presence of 5 !J,g/ml of lectin (Fig. 3). These experiments showed that P815r cells can become sensitive to lysis when allowed to contact LAK effectors through the use of lectins. In contrast, when the same experiment was performed with the human system, LAK resistance of SNUC-1 targets was unaffected by the presence of the lectin. In two experiments, SNUC-11ysis was 4.3 % and 5.9 % in the absence of Con A, compared to 2.8 % and 6.6 %, respectively, in the presence of 5 !J,g/ml of Con A (Fig. 4). Moreover, lysis remained at an average 7.1 % when using 15 !J,g/ml of lectin in three experiments. Clone 83 (14) was chosen as a positive control in these assays because it is relatively resistant to LAK in the absence of lectin (8.4 % and 18.2 % lysis in 2 experiments). However, the presence of 2.5 !J,g/ml Con A reverted this
Differential Tumor Resistance to Activated Killer Cells . 69 unlabeled competitors
% lysis of SNUC-4 targets b
60
PBl
50
SNUC-1 SNUC-4
...............
-
40~
30
.• ...':-::w._
..................
20
........... ...-
-
-~
•. •.••
..... ::-;
':"':-.":":" ."",:",:~.---
10
°0~l-------L 5 -------1~0------~ 15------~ 20~----~ 25 unlabeled targets per labeled SNUC- 4 cell
Figure 2. Competition for lysis of labeled SNUC-4 cells by LAK cells in the presence of unlabeled human cells. a a In each experiment, the percentages of cytotoxicity exerted by LAK cells against the targets used as cold competitors have been indicated in the "Y>' axis. b Standard error was < 10% of the mean in all experiments.
resistance by increasing lysis to 72.7 % and 82.3 %, respectively (Fig. 4). These experiments demonstrate that SNUC-l cells resist killing by LAK cells even after inducing binding of target and effector cells with lectin.
Binding of Fluorescein Isothiocyanate (FITC) labeled Concanavalin A to tumor cells P815r cells and clone 83 were used as a positive control for lectin binding. Incubation of P815r cells with different concentrations of FITC labeled 60
~EXP.1 1~ ExP. -~
816
50 % lysis of chromated ta rgets ·
2
40 30
20 10
o
""
o
.....
-
.....
-------
... ....- ..... 2
3
4
5
6
7
8
9
10
Con A (ug/mU
Figure 3. Effects of lectin in LAK mediated cytotoxicity of LAK resistant (P815r) and sensitive (B16) murine targets." a Results are presented as percentages of cytotoxicity exerted by LAK cells on chroma ted P815r and B16 targets in the presence of Can A. b Standard error was < 10 % of the mean of all triplicate measurements.
70 . P.
BEAN
and A.
MAZUMDER
100 80
% lysis of chromated
targets b
i i- l EXP 1
--
____ .. Clone 83
. Exp. 2
- - -:
"...----~ Clone 83
60
40 20 SNUC-1
0 ~~S=~~~~-~-~-~-=~~~~~S~N~U~-L 1 ~____ o 2 3 4 5 6 7 8
L -_ _
~ 9
__
~ 10
Con A (ug/mU Figure 4. Effects of lectin in LAK mediated cytotoxicity of LAK resistant human targets." Results are presented as percentages of cytotoxicity exerted by LAK cells on chromated SNUC-l or JD40 (clone 83) cells in the presence of Con A. b Standard error was < 10% of the mean of all triplicate measurements. a
Con A resulted in different intensities of cell labeling (Fig. 5). With 20.0 /-lgl ml of FITC lectin (Fig. 5c) the label was brighter than when using 5.0 /-lg/ml labeled Con A (Fig.5b). As a control, P815r cells incubated in PBS alone showed no label (Fig. Sa). The same results were obtained with clone 83 (data not shown). In equivalent experiments using SNUC-l cells, incubation of these targets in PBS alone showed no labeling (Fig.6a) whereas incubation with 5 /-lg/ml of fluorescent lectin showed labeling (Fig.6b). Effects of TNF-a on LAK sensitive and resistant targets Cytotoxicity of WEHI 164, our control for lysis, increased steadily up to 78.6 % when using concentrations of TNF-a ranging from 10 to 10 4 U/ml in an 18 h assay. In the same experiment, however, lysis of SNUC-l cells by TNF-a reached a maximum of 4.5 % at the highest (10 4 U/ml) concentration of cytokine tested (Exp. 1, Fig. 7). Incubation of SNUC-4 cells for 18 h with increasing concentrations of TNF-a proportionally increased target cell lysis (Exp. 2, Fig. 7); using 103 U/ml of lymphokine we found 50.5 % SNUC-4 lysis. In a similar experiment, SNUC-l tumor cells resisted the effects of this cytokine at all concentrations tested. The highest effect was 7.3 % lysis at 103 U/ml of cytokine. Thus, these experiments show that the SNUC-l tumor is resistant to TNF-a while this lymphokine readily induces the lysis of SNUC-4 and WEHI 164 target cells. Effect of Actinomycin D on susceptibility of SNUC-J cells to TNF In order to study the role of possible endogenous protective proteins in the acquisition of TNF-a resistance we treated SNUC-l cells for 6 h with 1 /-lg/ml actinomycin D (AD) in the presence of varying concentrations of
Differential Tumor Resistance to Activated Killer Cells . 71
a
c
b
Figure 5. Binding of P815 r cells to Fluorescein Isothiocyanate labeled Concanavalin A. Cells were incubated for 15 minutes with a ) phosphate-buffered saline alone, b) 5.0 ~g/m l, and c) 20.0 I1g/ml fluorescent lectin.
a
b
Figure 6. Binding of SNUC-l cells to Fluorescein Isothiocyanate labeled Concanavalin A. Cells were incubated for 15 minutes with a) phosphate-buffered saline alone, and b) 5.0 I1g/ml flu orescent lectin.
72 . P.
BEAN
and A.
MAZUMDER
80
WEHI164
60
% target cell Iys is b
SNUC-4
40
IEXp.11
Exp. 2 - - -
20
[TNF) units/m L
Figure 7. Effects of tumor necrosis factor on LAK sensitive (WEHr 164/13) and resistant (SNUC-l) targets. a a Results are presented as percentages of cytotoxicity exerted by TNF on chromated targets in 18 h assays. b Standard error was < 10% of the mean of triplicate measurements.
TNF (10 to 10 4 U/ml). The WEHI 164 cell line was used as a control. The titration curve for WEHI 164 shows that cell lysis increased proportionally with increasing TNF-a concentrations (Fig. 8). Simultaneous treatment of WEHI 164 cells with AD (0.2 [tg/ml and 1 [tg/ml) in the presence of 104 UI ml TNF-a increased sensitivity to this cytokine from 21.5 to 75.6 and 85.3 % lysis, respectively. There were no significant differences between the effects exerted by TNF on WEHI 164 at the two concentrations of protein inhibitor tested (0.2 [tg/ml and 1.0 [tg/ml). In contrast, under the same
100 90
@r- .-.-.
80
.~f/fl - - - - - -@- - - -
70
% target cell lysis C
_ . @ . _ . _ . _ . Jj?J WEH I1 64
__ @
WEH I 164
/;
60
.~
50
';
. ,/'
.I!J
40 30
/,/'
20 10
" - . -
o~
0
...
I"!"-----'"""~)-~------
Department of Pathology, University of Southern California, Los Angeles, California, USA
Resistance of Different Tumor Cells to Lysis by Lymphokine Activated Killer Cells Can Be Mediated by Distinct Mechanisms PAMELA BEAN and AMITABHA MAZUMDER 1 Received September 24, 1991 . Accepted in Revised Form February 14, 1992
Abstract Lymphokine activated killer (LAK) cells have been shown to exert a potent cytotoxic effect on many histologically different tumors and virally infected targets. Most normal cells but very few tumors have proven resistant to LAK lysis. The availability of two LAK resistant tumors, P815r, a murine mastocytoma, and SNUC-1, a human colon carcinoma, allowed us to study the phenomenon of LAK lysis. We examined the role of surface molecules on targets, which mediate binding to LAK cells, by cold target competition experiments and lectin dependent cellular cytotoxicity assays.The results showed that in the murine system, P815r cells do not compete for lysis of the LAK sensitive target B16 whereas other LAK sensitive murine targets compete. Alternatively, in the human system, SNUC-1 cells compete for lysis of the LAK sensitive target SNUC-4 as do other LAK sensitive human tumor cells. Furthermore, inducing binding of target and effector cells with lectin reverted the resistance of P815r but not SNUC-1 targets to lysis by LAK cells. These results imply that distinct stages of the lytic pathway might be involved in the resitance of different tumors to killing by LAK cells. The murine cell line is resistant to lysis because it cannot bind LAK cells. The human target, which does bind LAK, was insensitive to the effects of tumor necrosis factor alpha (TNF-a), a Iymphokine released by LAK effectors and possibly involved in their lysis. Resistance to TNF-a was not mediated by the presence of endogenous short-lived proteins in the SNUC-1 targets. The elucidation of mechanisms of resistance may provide a tool to improve current protocols of adoptive immunotherapy as well as insights as to how tumor cells are or are not killed by LAK effectors.
Introduction Culture of normal non-immune lymphocyte populations with interleukin-2 (IL-2) results in the generation of lymphokine activated killer (LAK) cells cytotoxic for a wide variety of histologically distinct tumors (1, 2). Previous reports in the literature have indicated that there is no apparent specificity for lysis of the autologous tumor because allogeneic as well as xenogeneic fresh or cultured tumors are lysed (3). Moreover, virus infected 1 Present address: Division of Medical Oncology, Georgetown University School of Medicine, Washington, D.C. 20007, USA
This work was supported by a grant from the Hastings Foundation
64 . P.
BEAN
and A.
MAZUMDER
targets (4) and single cell suspensions of placental and fetal tissues are also sensitive to LAK effectors (2). Interestingly, we have found murine (P815r) and human (SNUC-l) tumor cell lines which are not killed by LAK cells and their mechanisms of resistance constitute the subject of this study. The first step in LAK mediated cytolysis involves recognition and binding of target and effector cells. Blocking cell surface epitopes with antibodies (5, 6) or digesting cell membrane molecules with proteases abolishes LAK lysis (7). We therefore investigated the binding capacities of LAK resistant targets by means of in vitro competition experiments and lectin-dependent cellular cytotoxicity (LDCC) assays. For the in vitro competition assays we assumed that if cell membrane molecules present in LAK sensitive tumor cells were indeed responsible for their lysis by LAK effectors, the presence of unlabeled LAK resistant targets (presumably lacking ligand epitopes for LAK cells) should not interfere with killing of chromated LAK sensitive targets. Alternatively, in the lectin-dependent cellular cytotoxicity assay, the effector and target cells are brought into contact through the use of lectins. If the resistance of the targets is due to their incapacity to bind LAK cells, the presence of the lectin Concanavalin A (ConA) should overcome lack of lysis. We found that the resistant human and murine tumor cell lines possessed different abilities to compete in these assays. The murine cell line could not bind LAK cells while the human cell line did bind LAK cells and therefore is deficient in a different step(s) of the lytic pathway. It has been suggested that not only cell-to-cell contact but also the release of cytotoxic factors may be involved in the cytolytic event mediated by LAK cells (8, 9). LAK cells release the cytokines interferon gamma (IFN -y) and tumor necrosis factor alpha (TNF-a) when stimulated with some types of tumor cells (9,10). However, TNF-a is not equally cytotoxic to all tumor cell and certain kinds of tumor cells (and most normal cells) are known to be resistant to TNF-a (11). The mechanism(s) involved in the acquisition of TNF-a resistance is still controversial, but inhibition of RNA or protein synthesis makes normal cells and TNF-a-resistant tumor cells become sensitive to TNF (12). Thus, the defense mechanism of the cell most likely involves endogenous protective proteins (13). Based on these findings, we tested the sensitivity of SNUC-l cells to TNF-a and analyzed the effects of actinomycin D upon this sensitivity. Resistance to LAK cells has been suggested as a means whereby tumors may escape immune destruction and a greater understanding of sensitivity and resistance to this phenomena might have implications in improving immunotherapy with LAK cells. Materials and Methods Cells and culture media The murine tumor target cells employed in this study were B16 melanoma, obtained from the tumor repository at the Frederick Cancer Research Facility, YAC-l lymphoma, PS1S
Differential Tumor Resistance to Activated Killer Cells . 65 mastocytoma and WEHr 164, a methylcholanthrene-induced fibrosarcoma, were obtained from the American Type Culture Collection (ATCe). MCA methylcholanthrene induced sarcoma was obtained from NIH. The P815r cells represent a subclone of the parental P815 cell line which displayed resistance to lysis by LAK cells. WEHI164 is highly sensitive to human TNF-a and LAK. The human target cells used were: peripheral blood leukocytes (PBL) derived from normal donors; SNUC-4 and SNUC-1, both colon tumors (kindly provided by Dr. SOO-SANGJOHN, Keimyung University School of Medicine, Taegu, South Korea); and clone #83, derived from the undifferentiated lymphoma line JD40 (14) (kindly supplied by Dr. CARLO GAMBACORTIPASSERINI, Department of Human Oncology, University of Wisconsin, Madison, WI, USA). As previously reported (15), SNUC-1 cells are resistant to lysis by LAK cells while SNUC-4 cells are sensitive. All cell lines except MCA were grown to log phase (3-5 x 106 cells/ml) in complete media (CM) and were stored as cryopreserved cells to provide a standard stock for assay. The preparation of CM and the passage of MCA cells in vivo have been described (1). Generation of murine and human LAK cells Generation of murine LAK and NK cells was performed as described (1). Spleens from C57B!/6 mice 8 weeks or older were used in all experiments. Human LAK cells were prepared by thawing stocks of peripheral blood lymphocytes (PBL) obtained by leukapheresis of blood derived from normal donors. Cell suspensions of 106 PBL/ml were incubated for 5 days in 1000 units/ml of recombinant interleukin-2 (rIL-2) in CM. Cytotoxicity assays Four hour chromium release assays have been described elsewhere (16). For assay of TNF-a cytotoxicity,S x 10 3 51Cr-labeled targets were placed in each well of a 96-well plate and incubated with different concentrations of TNF-a (0 to 104 units/ml). The percentages of cytotoxicity were calculated as in a standard cytotoxicity assay (16). In studies that measured tumor cell sensitivity to TNF-a after inhibition of RNA synthesis, chromated targets were incubated for 6 h with 0.2 or 1.0 fAg/ml of Actinomycin D (Sigma, St. Louis, MO, USA) in the presence of the same concentrations of TNF-a described above. Results are expressed as the mean of triplicate measurements. Cold target competition assays Cold target competition assays using 51Cr-labeled murine B16 target cells were performed as follows: 2.5 x 105 LAK effector cells in 100 fAl CM were plated in each well of a round bottom microplate with various numbers of unlabeled murine target cells: 105,5 x 10\ 2.5 X 104 and 5 X 103 cells per well, representing cold:hot (C:H) ratios of 20, 10, 5 and 1 unlabeled (cold) competitor per each labeled (hot) target cell, respectively. After 30 minutes, wells were inoculated with 5 x 103 slCr-labeled B16 cells and the plate was incubated at 37°C,S % CO 2 for 3 h. The percentage of lysis induced by LAK cells against labeled B16 in the presence or absence of unlabeled targets (B16, YAC-1, P815 and MCA) was then determined. In the human system, the same procedure was followed using SNUC-4 cells as chromated targets. The unlabeled competitors were either SNUC-4, SNUC-l or PBL targets. The percent inhibition was calculated as: [1 - (% lysis in the absence of competitor!% lysis in the presence of competitor)] x 100. Lectin dependent cellular cytotoxicity assay (LDCC) LDCC was done by incubating chromated targets with a 50: 1 ratio of LAK effectors to target cells (E:T) in the presence of different concentrations (2.5 and 5 fAg/ml) of the lectin Concanavalin A (Sigma). Determinations of tumor cell lysis were done as in a standard cytotoxicity assay (16).
66 . P. BEAN and A. MAZUMDER
Binding of Fluorescein Isothiocyanate (FITC) labeled Con A to tumor cells As previously described (17), cultured cells (SNUC-1, clone 83 and P815r) resuspended at 106 cells/ml were layered onto sterile coverslips and incubated for 48 h at 37°C, 5 % CO 2 in CM. The coverslips were rinsed once in PBS and the cells were fixed by soaking in 3.7 % formaldehyde for 5 minutes. FITC conjugated Con A (100 f.tl) at concentrations of 0.5,5.0 and 20.0 f.tg/ml was then added. Control cells were incubated with PBS alone. The coverslips were incubated at room temperature in a humidified chamber for 15 minutes and were then rinsed in PBS before mounting onto glass slides with 80 % glycerol.
Results
Patterns of cytotoxicity mediated by LAK cells against a panel of tumor targets Three and five days of lymphokine activation were chosen in the murine and human systems, respectively, as they showed the highest LAK activity. The results of cytotoxicity by LAK cells against four tumor cell lines are shown in Table 1. Spleen LAK cells killed the B16 cell line with an average efficiency of 32 % (over three separate experiments, each done in triplicate) at a 50:1 E:T cell ratio. In contrast, under the same conditions, the levels of cytotoxicity exerted by LAK cells against P815r targets averaged 5.9 % lysis in three separate experiments. As a background, lysis of P815r targets by unstimulated splenocytes reached an average of 2.5 % at the 50:1 E:T ratio in 2 experiments (data not shown). In our laboratory, killing of this subclone of P815 has been achieved, but only with antibody against major histocompatibility complex antigen plus complement. Murine spleen LAK cells killed YA C-1 and M CA targets with an average efficiency of 68 % and 49 % respectively at the 50:1 E:T ratio in four experiments (data not shown). The effects of seven day human LAK cells on the human colon tumors SNUC-4 and SNUC-1 are presented in Table 1. The SNUC-4 cell line is killed by LAK cells with an average efficiency of 33 % lysis while the cytotoxicity exerted against SNUC-1 averaged 4.5 % at the same 50:1 E:T ratio. In three separate experiments unstimulated PBLs (NK) displayed a similar (3.5 %) cytotoxic effect on SNUC-1 targets (data not shown). The effects of human LAK on resting PBL, a population of normal cells which are not killed by LAK cells, averaged 1.7 % lysis in three experiments.
Competition for lysis of labeled B16 targets by spleen LAK cells, in the presence of unlabeled murine tumor cells Cold YAC-1 cells greatly decreased (78 % inhibition) B16 lysis while unlabeled P815r induced only 16 % inhibition at the 10:1 C:H ratio (Fig. 1a). As expected, cold B16 targets acted as competitors for their own lysis by inhibiting 75 % of the effect exerted by LAK cells alone, i.e., in the absence of unlabeled targets. MCA and YAC-1 cells induced 67 % and 91 % inhibition of B16lysis, respectively, at the 10: 1 C:H ratio (Fig. 1b). At an equivalent ratio, P815r cells inhibited B16 killing by 25 %. Overall, the
Differential Tumor Resistance to Activated Killer Cells . 67
percentages of inhibition exerted by cold P815r cells on chromated B16 targets averaged 15 % in four experiments and they differed from those found for the other three unlabeled targets by more than three standard deviations. Thus, P815r cells did not compete for lysis of B 16 targets as did the LAK sensitive YAC-1 and MCA tumor cell lines.
Competition for lysis of labeled SNUC-4 cells by human LAK cells in the presence of unlabeled human cells The lysis of SNUC-4 cells was unaffected by resting PBL (Fig.2). In contrast, LAK resistant SNUC-1 tumor cells inhibited 65 % of SNUC-4 cells lysis at 20: 1 C:H ratio. The inhibition exerted by cold SNUC-1 cells was very similar to the one displayed by unlabeled SNUC-4 (self) targets (70 % inhibition at an equivalent C:H ratio). The percentages of inhibition exerted by cold SNUC-1 cells over chromated SNUC-4 targets averaged 66 % in three experiments and differed from those obtained for unlabeled SNUC-4 by less than one standard deviation. Thus, lysis of LAK sensitive SNUC-4 cells was inhibited by self and by LAK resistant SNUC-1 targets, but not by normal PBL.
Effects of lectin in LAK mediated cytotoxicity of LAK resistant targets The addition of 5.0 flg/ml Con A increased LAK mediated lysis of P815r targets from 1.8 % to 17.8 % and from 0.8 % to 16.7 % respectively at the Table 1. Patterns of cytotoxicity mediated by LAK cells against a panel of targets a Target
% Cytotoxicity
E:T Exp. 1
Exp.2
Exp. 3
B16
50:1 10:1 1: 1
39.5 ± 2.9 19.5±0.9 3.7 ± 0.9
26.3 ± 2.8 8.0 ± 1.5 0.7 ± 0.7
30.5 ± 1.7 11.7± 1.1 2.7 ± 0.7
P815r
50:1 10: 1 1: 1
5.4 ± 1.8 -4.0 ± 1.2 -2.1 ± 0.9
4.7 ± 1.4 0.7 ± 0.2 0.1 ± 0.5
7.6 ± 2.3 6.4 ± 2.2 -0.9 ± 2.8
SNUC-4
50:1 10: 1 1: 1
25.6 ± 2.6 12.1 ± 2.8 3.0 ± 0.4
41.2±1.5 11.6 ± 1.5 2.3 ± 0.6
30.7 ± 1.3 10.1 ±0.6 3.4 ± 1.0
SNUC-1
50:1 10: 1 1: 1
4.6 ± 1.1 3.9 ± 0.8 3.9 ± 0.8
2.5 ± 0.2 0.2 ± 0.4 1.8 ± 0.3
6.5 ± 0.5 3.2 ± 2.6 1.5 ± 1.4
PBL
50:1 10: 1 1: 1
4.4 ± 2.2 1.5 ± 7.4 3.7 ± 1.5
1.5 ± 2.9 -1.8 ± 0.4 -3.3 ± 0.2
-0.8 ± 0.5 -2.8 ± 0.5 -0.7 ± 1.1
a
Results are reported as the mean ± SEM of the triplicate measurements. Statistical analysis was performed by the student's t test. A conservative cut-off number for a one-sided 0.05 level test for 2 means, each based on 3 observations.
68 . P.
BEAN
and A.
MA ZUMDER
P815r YAC-1 816
50 % lysis of 816 targets b
---
unlabeled competitors
60
40
41--- •
.. --at
30 20 10
o
o
5
-'
15
10
unlabeled targets per labeled 816 cell unlabeled compet itors
:~:' I:: I YAC-1
% lysis of
816 targets b
... - - .
30
10 --1-_ _ _ _ _ - - - '
5
10
15
unlabeled targets per labeled 816 cell
Figure 1. Competition for lysis of labeled B16 targets by spleen LAK cells in the presence of unlabeled murine tumor cells.' a In each experiment, the percentages of cytotoxicity exerted by LAK cells against the targets used as cold competitors have been indicated in the «y" axis. bStandard error was < 10 % of the mean in all experiments.
50:1 E:T ratio in the two experiments (Fig. 3). As a control, lysis of B16 cells was 24.5 % in the absence and 54.8 % in the presence of 5 !J,g/ml of lectin (Fig. 3). These experiments showed that P815r cells can become sensitive to lysis when allowed to contact LAK effectors through the use of lectins. In contrast, when the same experiment was performed with the human system, LAK resistance of SNUC-1 targets was unaffected by the presence of the lectin. In two experiments, SNUC-11ysis was 4.3 % and 5.9 % in the absence of Con A, compared to 2.8 % and 6.6 %, respectively, in the presence of 5 !J,g/ml of Con A (Fig. 4). Moreover, lysis remained at an average 7.1 % when using 15 !J,g/ml of lectin in three experiments. Clone 83 (14) was chosen as a positive control in these assays because it is relatively resistant to LAK in the absence of lectin (8.4 % and 18.2 % lysis in 2 experiments). However, the presence of 2.5 !J,g/ml Con A reverted this
Differential Tumor Resistance to Activated Killer Cells . 69 unlabeled competitors
% lysis of SNUC-4 targets b
60
PBl
50
SNUC-1 SNUC-4
...............
-
40~
30
.• ...':-::w._
..................
20
........... ...-
-
-~
•. •.••
..... ::-;
':"':-.":":" ."",:",:~.---
10
°0~l-------L 5 -------1~0------~ 15------~ 20~----~ 25 unlabeled targets per labeled SNUC- 4 cell
Figure 2. Competition for lysis of labeled SNUC-4 cells by LAK cells in the presence of unlabeled human cells. a a In each experiment, the percentages of cytotoxicity exerted by LAK cells against the targets used as cold competitors have been indicated in the "Y>' axis. b Standard error was < 10% of the mean in all experiments.
resistance by increasing lysis to 72.7 % and 82.3 %, respectively (Fig. 4). These experiments demonstrate that SNUC-l cells resist killing by LAK cells even after inducing binding of target and effector cells with lectin.
Binding of Fluorescein Isothiocyanate (FITC) labeled Concanavalin A to tumor cells P815r cells and clone 83 were used as a positive control for lectin binding. Incubation of P815r cells with different concentrations of FITC labeled 60
~EXP.1 1~ ExP. -~
816
50 % lysis of chromated ta rgets ·
2
40 30
20 10
o
""
o
.....
-
.....
-------
... ....- ..... 2
3
4
5
6
7
8
9
10
Con A (ug/mU
Figure 3. Effects of lectin in LAK mediated cytotoxicity of LAK resistant (P815r) and sensitive (B16) murine targets." a Results are presented as percentages of cytotoxicity exerted by LAK cells on chroma ted P815r and B16 targets in the presence of Can A. b Standard error was < 10 % of the mean of all triplicate measurements.
70 . P.
BEAN
and A.
MAZUMDER
100 80
% lysis of chromated
targets b
i i- l EXP 1
--
____ .. Clone 83
. Exp. 2
- - -:
"...----~ Clone 83
60
40 20 SNUC-1
0 ~~S=~~~~-~-~-~-=~~~~~S~N~U~-L 1 ~____ o 2 3 4 5 6 7 8
L -_ _
~ 9
__
~ 10
Con A (ug/mU Figure 4. Effects of lectin in LAK mediated cytotoxicity of LAK resistant human targets." Results are presented as percentages of cytotoxicity exerted by LAK cells on chromated SNUC-l or JD40 (clone 83) cells in the presence of Con A. b Standard error was < 10% of the mean of all triplicate measurements. a
Con A resulted in different intensities of cell labeling (Fig. 5). With 20.0 /-lgl ml of FITC lectin (Fig. 5c) the label was brighter than when using 5.0 /-lg/ml labeled Con A (Fig.5b). As a control, P815r cells incubated in PBS alone showed no label (Fig. Sa). The same results were obtained with clone 83 (data not shown). In equivalent experiments using SNUC-l cells, incubation of these targets in PBS alone showed no labeling (Fig.6a) whereas incubation with 5 /-lg/ml of fluorescent lectin showed labeling (Fig.6b). Effects of TNF-a on LAK sensitive and resistant targets Cytotoxicity of WEHI 164, our control for lysis, increased steadily up to 78.6 % when using concentrations of TNF-a ranging from 10 to 10 4 U/ml in an 18 h assay. In the same experiment, however, lysis of SNUC-l cells by TNF-a reached a maximum of 4.5 % at the highest (10 4 U/ml) concentration of cytokine tested (Exp. 1, Fig. 7). Incubation of SNUC-4 cells for 18 h with increasing concentrations of TNF-a proportionally increased target cell lysis (Exp. 2, Fig. 7); using 103 U/ml of lymphokine we found 50.5 % SNUC-4 lysis. In a similar experiment, SNUC-l tumor cells resisted the effects of this cytokine at all concentrations tested. The highest effect was 7.3 % lysis at 103 U/ml of cytokine. Thus, these experiments show that the SNUC-l tumor is resistant to TNF-a while this lymphokine readily induces the lysis of SNUC-4 and WEHI 164 target cells. Effect of Actinomycin D on susceptibility of SNUC-J cells to TNF In order to study the role of possible endogenous protective proteins in the acquisition of TNF-a resistance we treated SNUC-l cells for 6 h with 1 /-lg/ml actinomycin D (AD) in the presence of varying concentrations of
Differential Tumor Resistance to Activated Killer Cells . 71
a
c
b
Figure 5. Binding of P815 r cells to Fluorescein Isothiocyanate labeled Concanavalin A. Cells were incubated for 15 minutes with a ) phosphate-buffered saline alone, b) 5.0 ~g/m l, and c) 20.0 I1g/ml fluorescent lectin.
a
b
Figure 6. Binding of SNUC-l cells to Fluorescein Isothiocyanate labeled Concanavalin A. Cells were incubated for 15 minutes with a) phosphate-buffered saline alone, and b) 5.0 I1g/ml flu orescent lectin.
72 . P.
BEAN
and A.
MAZUMDER
80
WEHI164
60
% target cell Iys is b
SNUC-4
40
IEXp.11
Exp. 2 - - -
20
[TNF) units/m L
Figure 7. Effects of tumor necrosis factor on LAK sensitive (WEHr 164/13) and resistant (SNUC-l) targets. a a Results are presented as percentages of cytotoxicity exerted by TNF on chromated targets in 18 h assays. b Standard error was < 10% of the mean of triplicate measurements.
TNF (10 to 10 4 U/ml). The WEHI 164 cell line was used as a control. The titration curve for WEHI 164 shows that cell lysis increased proportionally with increasing TNF-a concentrations (Fig. 8). Simultaneous treatment of WEHI 164 cells with AD (0.2 [tg/ml and 1 [tg/ml) in the presence of 104 UI ml TNF-a increased sensitivity to this cytokine from 21.5 to 75.6 and 85.3 % lysis, respectively. There were no significant differences between the effects exerted by TNF on WEHI 164 at the two concentrations of protein inhibitor tested (0.2 [tg/ml and 1.0 [tg/ml). In contrast, under the same
100 90
@r- .-.-.
80
.~f/fl - - - - - -@- - - -
70
% target cell lysis C
_ . @ . _ . _ . _ . Jj?J WEH I1 64
__ @
WEH I 164
/;
60
.~
50
';
. ,/'
.I!J
40 30
/,/'
20 10
" - . -
o~
0
...
I"!"-----'"""~)-~------
