~

Cancer Immunol Immunother(1992) 34:321-328

ancer mmunolggy mmunothërapy

© Springer-Verlag 1992

Hierarchy of in vitro sensitivity and resistance of tumor cells to cytotoxic effector cells, cytokines, drugs and toxins* Jeffrey T. Safrit and Benjamin Bonavida Department of Microbiologyand Immunology,UCLA School of Medicine and the Jonsson ComprehensiveCancer Center, University of California, Los Angeles, USA Received 16 May 1991/Accepted 24 October 1991

Summary. Drug resistance of tumor cells has led to the development of other therapeutic modalities including biological response modifiers, lymphokine-activated killer cells (LAK), and cytokines alone and in combination. The premise of these alternative modalities is that drug resistance can be overcome by other cytotoxic agents or cytotoxic effector cells. However, the relationship between tumor cell sensitivity to these different agents and the cytotoxicity caused by drugs is not known or weil understood. Thus, understanding the relationship between these different systems of tumor cell cytotoxicity is essential for optimal therapeutic intervention. To this end, we compared the tumor cell cytotoxicity mediated by recombinant tumor necrosis factor (rTNF), cytotoxic effector cells (natural killer cells, monocytes, L A K cells), chemotherapeutic drugs, and microbial toxins. Human tumor cell lines sensitive and resistant to rTNF or drugs were used to evaluate the effectiveness of the other cytotoxic modalities. Sensitivity was considered as tumor cell cytotoxicity above 15% while resistance refers to that below 10%. Cell lines tested consisted of several histological types such as brain, lung, colon and ovarian tumors. In our experiments, cell lines made resistant to rTNF by coculture were also relatively resistant to unactivated monocytes and their supernatants. These lines were sensitive to all other methods tested including activated monocytes, natural killer and L A K cells, drugs, and toxins. The tumor lines naturally resistant to rTNF were found to have various degrees of sensitivity and resistance to these other systems. Upon the analysis of our data, a pattern emerged that suggested a hierarchy of sensitivity and resistance of the tumor cells to the cytotoxic mechanisms explored. From a majority of cell lines re* Supported in part by grant CA43 121 from the Department of Health and Human services, NIH, and NRSA clinical and fundamental immunology training grant A107 126, NIH (J. S.), and in part by a grant from the Concern Foundation, Los Angeles and a gift from the Boiron Foundation Offprint requests to: B. Bonavida, Department of Microbiology and Immunology, UCLA School of Medicine, University of California, CA 90 024, USA

sistant to rTNF to a minority of lines resistant to LAK, we found an interesting gradation of sensitivity and/or resistance to the other cytotoxic modalities employed. The hypothesis of an underlying common mechanism of action within these systems is discussed.

Key words: Cytotoxicity - Resistance - Sensitivity - Chemotherapy - Immune effector cells

Introduction The mechanism of tumor cell killing by host systems as well as by chemotherapy both in vivo and in vitro is not yet completely understood and has been an intriguing question for many years. One major problem with chemotherapy in general is the development of resistance to the cytotoxic agent used, resulting in the search for new methods of tumor cell killing. Overcoming this resistance will most likely require a combination of different modalities, including enhancing the host's own immune response against tumors and infections. For several years, certain biological response modifiers have been prepared with this aim in mind. The advent of biological response modifiers was postulated to enhance the host immune response and to mediate antitumor effectors resulting in tumor regression. For example, the interferons and interleukins can enhance the cytotoxic response to tumor cells [ 12, 16] and at times even cause tumor cell damage directly [32]. Also, the discovery and purification of tumor necrosis factor (TNF) as a cytotoxic molecule helped shed some light onto the immune systems' own defenses [7]. Still T N F ' s mechanism of killing has yet to be identified, and its use in clinical trials has been hampered by its tissue toxicity [26]. Further, several new therapies with either activated cells, such as lymphokine-activated killer (LAK) cells [12], and tumor-infiltrating lymphocytes [23], or "magic bullets" such as anti-

322

body-toxin conjugates [1] have been prepared to attack tumor cells directly and effect their destruction. These various approaches to boost the immune response or to target toxins to tumor cells are currently being experimentally and clinically investigated. A major premise of these interventions is the assumption that the tumor cells now refractory to chemotherapy or radiation remain sensitive to immunocytotoxic effector systems and or toxins. Alternatively, there exsisted the possibility that the development of resistance to chemotherapeutic drugs would also result in resistance to cytotoxic effector cells, cytokines and microbial toxins and vice versa. Thus the present study was designed to investigate whether there exists any correlation between the sensitivity and resistance of tumor cells to chemotherapeutic drugs, immune effector cells or factors, and bacterial toxins.

Materials and methods Cell lines and media. The human histiocytic cell line U937 and its rTNF-resistant counterpart, U9TR, were maintained as previously described [30]. The human ovarian carcinoma cell lines, 222, PA1, OVC-3, and SKOV3, were maintained in RPMI-1640 medium supplemented with non-essential amino acids, glutamine, antibiotics, and 10% fetal bovine serum. A rTNF-resistant variant of 222 was created in the laboratory by coculturing the line in stepwise concentrations of rTNF for several weeks. The ovarian line A2780 and its drug-resistant sublines AD 10 (Adriamycin-resistant), C30 (cisplatin-resistant) as well as OVC-8 were obtained from Dr. Robert Ozols, Philadelphia. The human brain cell line U251 as well as the human lung lines 322-P75 and 226-P59 and the melanoma RP were generous gifts from Dr. E. Grimm of the M. D. Anderson Cancer Center. The breast line B 1 and the colon line UCLASO-C2 as well as the melanoma UCLA-SO-M14 were generously supplied by Dr. S. Golub at this institute. All the above lines as well as the human B-cell line Raji were maintained as above. All lines were grown at 37°C in 5% CO2. Reagents. The recombinant TNF and interleukin-2 (IL-2) were generously provided by SmithKline-French, Philadelphia and Genentech, respectively. The interferons (IFN) Y and o~ were generously supplied by Genentech and Hofman La-Roche respectively. Antiserum directed against rTNF was raised in rabbits by intramuscular injection of 50 gg rTNF in complete Freund's adjuvant. The rabbits were boosted after 3 weeks, and 7 days later were bled by venous puncture to test for neutralizing activity. The rabbits, care was in accordance with UCLA guidelines. Adriamycin was purchased from Sigma Chemical Co. cis-Diaminedichloroplatinum (cisplatin) was purchased from Bristol-Myers Co. Diptheria toxin and Pseudomonas toxin were generous gifts from Dr. Bernadine Wisnieski.

Cytotoxic effector cells. Human peripheral blood lymphocytes (PBL) and monocytes were obtained as previously described [30]. Briefly, whole human blood was fractionated on Ficoll/Hypaque and the resulting cells were divided into plastic-adherent monocytes and plastic and nylonwool-nonadherent lymphocytes. Lymphokine-activated killer cells (LAK) were generated by culturing PBL in the presence of 500 U/ml rIL-2 for 5 - 7 days. Cytotoxicity assay. The 51Cr-release assay was used to determine the extent of tumor cell cytotoxicity in both the cell-mediated and cytokineor drug-mediated systems. Bfiefly, tumor cells at (1-3) x 106 cells/ml were incubated with 0.1 ml Na251CRO4 (Amersham, 1 mCi/ml) in a total volume of 1 ml RPMI-1640 medium with 20% fetal calf serum at 37°C for 1 h. Cells were then washed three times and resuspended at 1 x 105 cells/ml in culture medium. A sample of 0.1 m151Cr-labelled tumor cells at 1 × 105 cells/ml was added to 96-well microtiter plates and 0.1 ml

45 40

7

35-

o

0

25

~

Ei:]

20~D

15-

lo. 5

•

o 0.1

.

m~__--m

w

;

m

;,, c) 3O x o o 2O

~

.Sug/rnl .05ug/ml

N 10-

0 U937

PA1

222

U9TR 222TR SKOV3 RA~

Fig. 5A, B. Cytotoxicityof variousconcentrationsof Adriamycin(A), or cisplatinum (B). Tumorcells were incubatedin the presence of the drug at 37°C for 20 h, supernatantstested for radioactivity and percentage cytotoxicity determinedas for rTNF

U251 and 226-P59. Thus it appears that the mechanism of resistance to the cytotoxic effects of rTNF may differ depending on the cell line used. It is also clear that conferring rTNF resistance on a tumor cell does not make it equally resistant to other cytotoxic agents.

Hierarchy of sensitivity and resistance of tumor cells to various cytotoxic agents Upon analysis of the above data, using a large number of histologically diverse tumor lines, a pattern emerges that can be seen in Table 1. Thus, + symbolizes sensitivity to the above agent with cytotoxicity of 15% or greater, while - symbolizes resistance and cytotoxic values below 10%. Borderline sensitivity and/or resistance is symbolized by _ . Thus, the lysis of tumor cells by cytotoxic modalities in vitro can be arranged according to a hierarchy of sensitivity and resistance. The relative ease of obtaining lines resistant to rTNF as compared to the difficulty in finding LAK-resistant lines is striking. We hypothesize that this pattern suggests an underlying mechanism of cytotoxicity that the various agents have in common and perhaps a shared mechanism of resistance by the tumor cell lines.

326

A 70 60 ù_o 50 40 . u

5O

x o

i

3o i 202 I

2: U957

PAI

222

U9TR

222TR SKOV5

RAJ!

Fig. 6. Cytotoxicity of various concentrations of diptheria toxin (Dtx, A), or Pseudomonas toxin (Ptx, B). Tumor cells were incubated in the presence of the toxin at 37 ° C for 20 h, supernatants tested for radioactivity, and percentage cytotoxicity determined as for rTNF

Discussion

Development of tumor cell resistance in cancer patients treated with chemotherapeutic drugs has led to the introduction of other therapeutic modalities including biological response modifiers, LAK cells, cytokines, etc. However, the relationship between sensitivity and resistance to one agent when compared to another is not known. Thus, a comparison of tumor cell sensitivity and resistance to the cytotoxic effects of various agents currently used in cancer therapy was undertaken. In the present study, the analysis of our data (examining a wide variety of tumors of different histological origins) reveals a pattern that suggests that there may be a hierarchy of sensitivity and resistance to these various toxic modalities. The hierarchy demonstrates, for example, that tumor cells sensitive to rTNF or NK cells are most likely to be sensitive to monocytes, activated NK cells, cytotoxic drugs, toxins, and LAK cells. On the other hand, cells that are resistant to LAK cells or bacterial toxins will most likely be resistant to drugs, macrophages, NK cells, and TNF. These findings suggest that the expression of resistance to a particular agent may be linked with the simultaneously coexpressed resistance to an unrelated agent. Further, the existence of cell lines with resistance to drugs, toxins, and cross-resistance to immune effector cells suggests that there may be a common pathway for the expression of sensitivity and resistance. Several mechanisms have been reported to underlie the expression of tumor cell

resistance to certain agents. Some of these mechanisms may be shared and may explain in part the existence of the hierarchy. Alternatively, separate mechanisms and pathways could be involved that may complement each other. We therefore compared the postulated mechanisms that have been reported for the cytotoxic agents used in our study to determine the underlying basis of the hierarchy. TNF is an immunologically important protein, originally identified in the serum of animals injected with endotoxin, that could mediate the necrosis of transplanted tumors in vivo [7]. Since the cloning and expression of human TNF in E. coli [20, 28], sufficient quantities of rTNF have become available for clinical trials in the treatment of cancer. The mechanism of TNF cytotoxicity, while not yet fully understood, has been linked to both the generation of free radicals [17] and the induction of apoptosis or programmed cell death in the tumor cell [25]. The cytotoxic mechanisms employed by PBL/NK cells are still the subject of debate, with NKCF Natural killer cytotoxic factor [2], TNF [ 19], and perforins [ 10] receiving some of the recent attention. However, in a short-term assay, NK cells kill target cells by an apoptosis-mediated mechanism [11]. IFNo~-activated NK cells may use the above as well as the lytic potential of their secretory granules to kill tumor cells [4]. Monocytes may kill tumor cells in vivo and in vitro by a number of different mechanisms, including the use of membrane-bound TNF or other membrane proteins, secreted TNF and/or reactive oxygen species [9, 14, 18, 15]. Clearly, as with soluble TNF, macrophage-mediated cytotoxicity also involves apoptosis [9]. Activated monocytes, on the other hand, seem to have at least two independent cytotoxic pathways to lyse tumor cells. In addition to the aforementioned TNF, IFNT-activated monocytes produce superoxide radicals, which have also been shown to be involved in the lysis perpetrated by certain chemotherapeutic drugs [5]. In comparing the tumor cell lysis caused by the effector systems previously described (rTNF, monocytes and NK cells with or without IFN(z), one possible correlation may be the need for the presence of a receptor or receptors for lysis to occur. The lack of TNF receptor expression by the tumor cell may not be related to resistance to rTNF, or other effector systems, but its presence is positively correlated with sensitivity. The cytotoxic activity of certain chemotherapeutic drugs, as mentioned above, has been linked to the production of superoxide radicals, since agents that block intercellular detoxification schemes can enhance their cytotoxicity [27]. The generation of free radicals has been shown to be involved in DNA strand breaks that eventually lead to the depletion of cellular stores of NAD and ATP and therefore programmed cell death [6]. Likewise, the cytotoxicity mediated by certain bacterial toxins such as diptheria and Pseudomonas toxins may be the result of programmed cell death due to the ADP-ribosylation of elongation factor 2 and the subsequent loss of NAD [8]. The generation of free radicals has also been linked to TNF cytotoxicity as freeradical scavengers can inhibit lysis [17]. Also, increased levels of manganous superoxide dismutase, a potent free-

327 radical scavenger, have been shown to be involved in cellular mechanisms of resistance to TNF [29]. Therefore, because of the similarities noted above, we compared T N F and activated m o n o c y t e cytotoxicity to that mediated by several chemotherapeutic drugs and bacterial toxins. Out findings with both the rTNF-sensitive lines and those lines made resistant by coculture indicate an equal sensitivity to both agents. In fact, similar dose/response curves were generated and both drugs and toxins lysed all the cell lines except for SKOV3, Raji, U251, and 226-P59 (excluding the culture-generated drug-resistant lines). As for these drug-resistant lines such as the multidrug-resistant gp 170 + AD10, our results indicate this cell line is more sensitive to rTNF than is its drug-sensitive parental line A2780, as seen in previous data with leukemia and m y e l o m a cell lines [24]. Thus if a c o m m o n pathway is shared by the lytic mechanisms of TNF, monocytes, cytotoxic drugs and toxins, it is possible that the drugs and toxins override the T N F resistance at a step beyond the block. Alternatively, many different pathways o f cytotoxicity could be linked only at small intersections thus allowing the blockage o f one pathway without affecting others. Recent experiments using T N F in combination with these drugs have shown that T N F synergizes with the drugs in m a n y o f the cell lines [3]. This pattern o f tumor cell cytotoxicity fits extremely well into our hypothetical sequential pathway and lends credence to the idea o f a c o m m o n pathway o f lysis between T N F and other modalities. At the opposite end o f the hierarchy from T N F are lymphokine-activated killer cells, which have been shown to lyse a number of different tumor cell types [12, 22]. W h e n tested against the rTNF-sensitive and -resistant lines, we found L A K cells capable o f lysing all but one o f the lines, even the highly resistant S K O V 3 , although quantitative differences were noticed. The cytotoxic capacity of L A K cells represents a further enhancement of the lytic mechanism with the possible involvement of pore-forming perforins [13]. The examination of the mechanisms o f cytotoxicity employed by the various agents tested reveals a limited number o f effects on the biological processes o f tumor cells. Monocytes and N K cells use T N F [21, 9, 14], which m a y mediate its cytotoxicity by the formation o f superoxide radicals [ 17], and/or the induction of apoptosis. Superoxide radical formation and programmed cell death are likewise linked to the cytotoxicity caused by chemotherapeutic drugs and bacterial toxins [5, 8]. Thus, it seems that only a few mechanisms of cytotoxicity can be used to explain the response of tumor cells to these agents. In actuality, these mechanisms m a y not act alone but rather in concert with one another, eventually to cause cell death. Therefore, a hierarchy o f cytotoxic mechanisms, f r o m a singular causality to multiple interactions, m a y be a reasonable explanation o f the gradation o f tumor cell sensitivity and resistance to various cytotoxic agents. Our results suggest that a hierarchy of sensitivity of tumor cells to this variety of cytotoxic modalities exists. It is possible that the dissection o f this hierarchy m a y help establish c o m m o n pathways by which resistance to all these agents operate. Because o f this development o f resistance of tumors to chemotherapy in vivo, our results indi-

cate that determining the resistance pattern of the tumor cells prior to application of other cytotoxic systems may be of extreme importance. In any event, the p h e n o m e n o n of cross-resistance or sensitivity to therapies other than cytotoxic drugs must not be overlooked.

Acknowledgements.We would like to acknowledge Dr. Robert Ozols for providing us with cell lines A2780, A2780-ADI0, A2780-C30, and OVC-8; Dr. Sidney Golub for providing us with cell lines B1, UCLASO-C2, and UCLA-SO-M 14; and Dr. Elizabeth Grimm for providing us with cell lines U251, 322-P75,226-P59, and RP. We also acknowledge Dr. Y. Nio for establishing the cell line 222.

References 1. Barton RL, Briggs SL, Koppel GA (1991) Monoclonal antibody drug targeting. Drugs News and Perspectives 4: 73- 88 2. Bonavida B, Wright SC (1987) Multistage model of NK cell mediated cytotoxicity involving NKCF as soluable cytotoxic mediators. Adv Can Res 49: 169-187 3. Bonavida B, Tsuchitani T, Safrit J, Zighelboim J (1989) Mechanism of target cell lysis by cytotoxic cells, factors, and drugs. In: Tapiero H, Robert J, and Lampidis TJ (eds) Anti cancer drugs. Colloque INSERM. Libbey Eurotext, London, pp 113 - 128 4. Brunda MJ, Davatelis V (1985) Augmentation ofNK cell activity by recombinant IL-2 and recombinant IFNs. In: Herberman R (ed) Mechanisms of cytotoxicity by NK cells. Academic Press, New York, pp 397-407 5. Capel ID, Leach D, Dorrel HM (1983) Vitamin E retards the lipoperoxidation resulting from anticancer drug administration. Anticancer Res 3:59-62 6. Carson DA, Seto S, Wasson DB, Carrera CJ (1986) DNA strand breaks, NAD metabolism, and programmed cell death. Exp Cell Res 164:273-281 7. Carswell EA, Old LJ, Kassel R, Green S, Fiore N, Williams B (1975) An endotoxin-induced serum factor that causes necrosis of tumors. Proc Natl Acad Sci USA 72:3666-3670 8. Chang MP, Bramhall J, Graves S, Bonavida B, Wisnieski B (1989) Internucleosomal DNA cleavage precedes diptheria toxin-induced cytolysis. J Biol Chem 264:15 261 - 15 267 9. Decker T, Lohmann-Matthes M, Gifford GE (1987) Cell associated tumor necrosis factor (TNF) as a killing mechanism of activated cytotoxic macrophages. J Immunol 138:957-962 10. Dennert G, Podack E (1983) Cytolysis by H-2 specific T killer cells. J Exp Med 157: 1483-1493 11. Duke RC, Cohen J, Chervenak R (1986) Differences in target cell DNA fragmentation induced by mouse CTL and NK cells. J Immunol 137: 1442-1448 12. Grimm EA, Mazumder A, Zhang HZ, Rosenberg SA (1982) Lymphokine activated killer cell phenomenon. Lysis of NK resistant fresh and solid tumor cells by IL-2 activated autologous human PBL. J Exp Med 155: 1823-1841 13. Hiserodt JC, Chambers WH (1985) Role of soluble cytotoxic factors in lymphkine activated killer cell mediated cytotoxicity. Ann NY Acad Sci 532:395-404 14. Ichinose Y, Bakouche O, Tsao JY, Fidler IJ (1988) Tumor necrosis factor and IL-1 associated with plasma membranes of activated human monocytes lyse monokine-sensitive but not monokine-resistant tumor cells whereas viable activated monocytes lyse both. J Immunol 141:512-518 15. Kriegler M, Perez C, DeFay K, Albert I, Lu SD (1988) A novel form of TNF/Cachectin is a cell surface cytotoxic transmembrane protein: ramifications for the complex physiology of TNF. Cel153: 45- 53 16. Lichtenstein A, Berek JS, Bast RC, Spina C, Hacker N, Knapp RC, Zighelboim J (1984) Activation of peritoneal lymphocyte cytotoxicity in patients with ovarian cancer by intraperitoneal treatment with Cornebacteriumparvum. J Biol Response Modif 3:371-376

328 17. Matthews N, Neale ML, Jackson SK, Stark JM (1987) Tumor cell killing by tumor necrosis factor: inhibition by anaerobic conditions, free radical scavengers, and inhibitors of arachitanate metabolism. Immunology 62:153 - 155 18. Nathan CF, Silverstein SC, Bruckner LH, Cohn ZA (1979) Extracellular cytolysis by activated macrophages and granulocytes: II. Hydrogen peroxides as mediators of cytotoxicity. J Exp Med 149: 100-113 19. Paya CV, Kenmotsu N, Schoon RA, Liebson PJ (1988) Tumor necrosis factor and lymphotoxin secretion by human NK cells leads to antiviral cytotoxicity. J Immunol 141: 1989-1995 20. Pennica DGE, Haflick JS, Sesburg PH, Denynck R, Palladino MA, Kohr WJ, Aggarwaal BB, Goeddel DV (1984) Human tumor necrosis factor: precursor, structure, expression and homology to lymphotoxin. Nature 312: 724-728 21. Philip R, Epstein LB (1986) Tumor necrosis factor as immunomodulator and mediator of monocyte cytotoxicity induced by itself, y-interferon and interleukin-1. Nature 323: 8 6 - 89 22. Rosenberg SA, Lotze MJ (1986) Cancer immunotherapy using IL-2 and IL-2 activated lymphocytes. Annu Rev Immunol 4 : 6 8 1 - 7 0 9 23. Rosenberg SA, Spiess P, Lafreniere R (1986) A new approach to the adoptive immunotherapy of cancer with tumor infiltrating lymphocytes. Science 233: 1318- 1320 24. Salmon SE, Soehnlen B, Dalton WS, Meltzer P, Scuderi P (1989) Effects of tumor necrosis factor on sensitive and multidrug resistant human leukemia and myeloma cell lines. Blood 74:1723 - 1727

25. Schmid DS, Honung R, McGrath KM, Paul N, Ruddle NH (1987) Target cell DNA fragmentation is mediated by lymphotoxin and TNF. Lymphokine Res 6 : 1 9 5 - 2 0 2 26. Sherman ML, Spriggs DR, Arthur KA, Imamura K, Frei E, Kufe DW (1988) Recombinant human tumor necrosis factor administered as a five day continuous infusion in cancer patients: phase 1 toxicity and effects on lipid metabolism. J Clin Oncol 6: 3 4 4 - 350 27. Suzukate K, Vistica BP, Vistica DT (1982) Reduction in glutathione content of L-PAM-resistant L1210 cells confers ding sensitivity. Biochem Pharmacol 31: 121 - 124 28. Wang AM, Creasey AA, Ladner MB, Lin LS, Strickle J, Vanarsdell JN, Yamamoto R, Mark DF (1985) Molecular cloning of the complementary DNA for tumor necrosis factor. Science 228:149 - 154 29. Wong GHW, Elwell JH, Oberley LW, Goeddel DV (1989) Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor. Cel158:923 - 931 30. Wright SC, Bonavida B (1987) Studies on the mechanism of natural killer cell mediated cytotoxicity: IX. Functional comparison of human natural killer cytotoxic factors with recombinant lymphotoxin and tumor necrosis factor. J Immunol 138:1791 - 1797 31. Yoshie O, Tada K, Ishida N (1986) Binding and crosslinking of 125I-labelled recombinant human tumor necrosis factor to cell surface receptors. J Biochem (Tokyo) 100:531 - 591 32. Zighelboim J, Nio Y, Berek JS, Bonavida B (1988) Immunologic control of ovarian cancer. Nat Immun Cell Growth Regul 7: 216-225