Mechanism of interferon gamma-induced protection of human gliosarcoma cells from lymphokine-activated killer lysis: division of lymphokine-activated killer cells into natural killer- and T-like cells.

AUTHOR(S): Kondo, Seiji, M.D.; Miyatake, Shinichi, M.D.; Kikuchi, Haruhiko, M.D.; Oda, Yoshifumi, M.D.; Iwasaki, Koh-ichi, M.D.; Ohyama, Kenji, M.D.; Namba, Yuziro, M.D. Department of Neurosurgery, Faculty of Medicine (SK, SM, HK, YO, KI, KO), and Department of Pathology, Institute for Virus Research (YN), Kyoto University, Kyoto, Japan Neurosurgery 31; 534-540, 1992 ABSTRACT: THE MECHANISM BY which interferon gamma (IFN-γ) decreases the susceptibility of the established cultured gliosarcoma line GI-1 to lymphokine-activated killer (LAK) lysis was analyzed. The results of monolayer depletion and lectin-dependent cellular cytotoxicity assays by LAK cells revealed that the resistance to LAK lysis of IFN-γ-treated GI-1 cells is manifested at the stage of LAK cell target recognition alone. We have also divided LAK cells into populations of phenotypically natural killer (NK)- and T-like cells with monoclonal antibodies and complement, respectively. We have used these cells to examine the mechanism of IFN-γinduced protection of GI-1 cells from LAK lysis in cold target inhibition, monolayer depletion, and direct binding assays. The results revealed that NKlike cells do not recognize IFN-γ-treated GI-1 cells as efficiently as they do untreated targets, whereas Tlike cells show the opposite tendency. In conclusion, we have demonstrated that the IFN-γ induced protection of tumor cells from LAK lysis is predominantly regulated by the target recognition of NK-like cells. On the other hand, IFN-γ-treated tumor cells may bind to T-like cells but fail to trigger them to initiate further stages for lysis as effectively as NKlike cells.

experiments to be the only exogenous factor necessary for the generation of mature NK cells from bone marrow precursors (16). On the other hand, interferon (IFN) gamma has already been shown to decrease the susceptibility of a variety of human and murine tumor cells (11,12,19,32,34) to NK lysis both in vivo (9) and in vitro (22); the mechanism for this effect, however, remains controversial (3,8). In light of the similarities between LAK and NK cells, the effect of IFN-γ on the NK susceptibility of tumor cells has prompted some investigators to study the effect of IFN-γ on the sensitivity of target cells to LAK lysis. IFN-γ has recently been reported to decrease the susceptibility of tumor cells to lysis by LAK cells as well as by NK cells (5). However, the mechanism underlying this phenomenon is still unclear. We have shown previously that IFN-γ treatment decreased the susceptibility of GI-1 cells to LAK lysis and increased their expression of human lymphocyte antigen Class I molecules in an inverse fashion (21). In this study, we have investigated further the mechanism by which IFN-γ decreases GI-1 susceptibility to LAK lysis in monolayer depletion and lectin-dependent cellular cytotoxicity (LDCC) assays. It has also been shown that there are two major populations of LAK cells (2,7,17,25,31). We have therefore also attempted to determine which population (NK- or T-like cells) is involved in the change in susceptibility of GI-1 cells after IFN-γ treatment in cold target inhibition, monolayer depletion, and direct binding assays.

KEY WORDS: Interferon gamma; Lymphokineactivated killer cell; NK-like cell; T-like cell

PATIENTS AND METHODS Target tumor cells and interferon gamma treatment The GI-1 cells were derived from a 61-year-old man with gliosarcoma, and they were maintained in tissue culture for more than 5 years (20). These cells were maintained in Dulbecco's modified minimal essential medium (Nissui, Tokyo, Japan), supplemented with 10% heat-inactivated fetal calf serum (FCS; GIBCO Laboratories, Grand Island, NY) and antibiotics. Half-saturated GI-1 cells were treated with recombinant IFN-γ (TRP-2) for 18 hours at a concentration of 1000 U/mL. Control cultures were set up without treatment with IFN-γ. Cells were then detached with 0.05% trypsin and 0.01% ethylenediamine tetraacetic acid in phosphatebuffered saline for use as target cells in the cytotoxicity assays.

Lymphokine-activated killer (LAK) cells are composed of heterogeneous populations of cells induced only by interleukin (IL)-2 and express anomalous killing potential for natural killer (NK)resistant as well as NK-sensitive target cells (10). It is generally believed that the cytotoxic functions against target cells of LAK cells as well as NK cells are mediated in a major histocompatibility complexnonrestricted manner (14,30,35). There are also some reports that LAK and NK cells also share similar regulatory pathways; IL-2, which induces LAK activity, enhances the proliferation and cytotoxicity of NK cells and has also been shown in in vitro

Preparation of lymphokine-activated killer cells The peripheral blood lymphocytes of healthy donors were isolated by Ficoll-Paque (Pharmacia Fine Chemicals AB, Uppsala, Sweden) density gradient centrifugation. The isolated lymphocytes were washed three times with phosphate-buffered saline and were cultured with RPMI 1640 supplemented with 10% heat-inactivated FCS, 1 mmol/L sodium pyruvate, 1 mmol/L glutamine, antibiotics, and recombinant IL-2 (TGP-3, 2 units/ml) for 3 days at 37°C in a 5% CO2 atmosphere. After culture, the cells were washed twice before being used in the cytotoxicity assays.

Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/3/534/2752153 by Frankfurt Univesity Library user on 03 March 2018

Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.

Neurosurgery 1992-98 September 1992, Volume 31, Number 3 534 Mechanism of Interferon Gamma-Induced Protection of Human Gliosarcoma Cells from Lymphokine-Activated Killer Lysis: Division of Lymphokine-Activated Killer Cells into Natural Killer- and T-Like Cells Experimental Study

Cytotoxicity assay A 51Cr release assay was used for the determination of cell-mediated cytotoxic activity. The method and formula for calculating specific cytolysis have been described previously (20). Cytotoxic activity results in this article are represented as the percentage of cytolysis at effector/target cell ratios indicated by 4hour incubation analysis. Spontaneous release was always less than 20% of the maximal release. Lytic unit values per 106 cells, which were the reciprocals of the lymphocyte numbers required for 50% killing of 104 target cells, were also calculated (26). Statistical analysis for lytic unit determination was performed by the paired t test. Monolayer depletion and lectin-dependent cellular cytotoxicity assays A monolayer depletion assay was performed by a modification of the method of De Fries and Golub (5). GI-1 adherent cells were grown as monolayers in a 25cm2 tissue culture flask (Type 25100; Corning, NY) with or without the addition of 1000 U/mL of IFN-γ for 18 h. The IFN-γ-containing medium was then removed, and washed LAK cells, NK-like cells, and Tlike cells were added to target cell monolayers in 5 mL of RPMI 1640 supplemented with 10% FCS. Effector cells were allowed to adhere for 1 hour at 37°C in a 5% CO2 atmosphere; then, nonadherent cells were gently removed from the flask. Nonadherent effector cells were then washed twice and tested for their ability to lyse untreated target cells.


For the LDCC assay, a 51Cr release assay was performed with the addition of 2 mL of phytohemagglutinin per microgram (PHA-P; Pharmacia LKB Biotechnology) during the 4-hour assay period (6). Cold target inhibition and direct binding assays Untreated GI-1 cells were labeled with 51Cr and dispensed into 96-well, U-shaped microtiter plates (Type 25850; Corning) at 5 × 103 cells/well in a 50-µL volume. Unlabeled GI-1 cells, both IFN-γ treated and IFN-γ untreated, were added to the wells at unlabeled/labeled target cell ratios ranging from 2.5:1 to 20:1. NK- and T-like cells were then added at an effector-to-labeled target cell ratio of 40:1, and the lysis of labeled target cells was determined after the 4hour incubation. For the direct binding assay, washed LAK cells, NK-like cells, and T-like cells were labeled with fluorescein diacetate (Polysciences, Inc., PA) by a modification of the method of D'Amore and Golub (4). Fluorescein diacetate-labeled effector cells were then mixed with washed target cells at an effector-target cell ratio of 1:1 in a total volume of 0.2 mL of RPMI 1640 without FCS. Effector-target cell mixtures were incubated for 10 minutes in a 37°C water bath and then centrifuged for 10 minutes at 200 g. After the pellets had been resuspended gently and a sample of the cell suspension had been transferred to a hemocytometer, conjugates were counted under a fluorescence microscope. The percentage of conjugates was calculated by determining the number of effector cells bound to target cell per total number of effector cells counted. Statistical analysis for percent binding ± standard deviation was performed by use of the paired t test. RESULTS Interferon gamma-induced protection of GI-1 cells from lymphokine-activated killer lysis To determine the optimal amount of IFN-γ necessary to decrease the susceptibility to LAK lysis, we treated GI-1 cells with IFN-γ in a dose-dependent manner as previously described (21). Furthermore, to determine the time required for IFN-γ-induced protection to develop, GI-1 cells were treated with 1000 U of IFN-γ per milliliter for periods of 12, 18, and 24 hours before their susceptibility to LAK lysis was tested (data not shown). We found that the treatment of GI-1 cells with IFN-γ at 1000 U/mL for 18 hours resulted in a marked and consistent decrease in LAK susceptibility. GI-1 cells were therefore treated as indicated in this study and tested for susceptibility to LAK lysis by allogeneic LAK cells (Fig. 1). A significant decrease in susceptibility to LAK lysis was observed after the treatment of GI-1 cells with IFN-γ (P < 0.01). Similar results were obtained with U 251-MG cells (derived from a malignant glioma) as the target cells (data not shown). Monolayer depletion and lectin-dependent cellular cytoxicity assays by lymphokine-activated killer cells To determine whether or nor IFN-γ reduces the


Division of lymphokine-activated killer cells into natural killer- and T-like cells Cell division in the cytotoxicity assays was performed by a modification of the method of Reinhertz et al. (28). All division experiments were performed in a cytotoxicity medium composed of RPMI 1640, 5% heat-inactivated FCS, and 25 mmol/L 4-(2-hydroxyethyl)-1-piperazine ethanesulfonic acid buffer. Murine monoclonal antibodies, OKT3 (Pharmacia LKB Biotechnology, Piscataway, NJ) for CD3 and Leu-11b (BectonDickinson, CA) for CD16, and lyophilized rabbit complement-MA (Cedarlane, Canada) were used for the division experiments. In brief, the lymphocyte suspension was adjusted to a concentration of 107 cells/mL in the cytotoxicity medium. The appropriately diluted antibodies (1:10 for OKT3 and 1:100 for Leu-11b) were added in the same volume to the cell suspension, the mixtures were incubated at 20°C for 1 hour, and reconstituted rabbit complement was then added to the suspension at the same volume. Incubation was continued for 1 more hour at 37°C. At the end of this time, cells reactive with monoclonal antibodies were lysed. After being washed with the medium, viable cells were resuspended at the desired concentration. It was confirmed that cell division experiments were performed successfully enough if less than 2% of the residual cells after antibody and complement treatments were recognized by immunofluorescence microscopic observation with the responsible antibodies.

Monolayer depletion assays by natural killer-like and T-like cells The division of LAK cells into NK- and T-like cell populations was carried out in order to determine which population was involved in the change of susceptibility that occurred in GI-1 cells after IFN-γ treatment. Figure 4A demonstrates that NK-like cells divided by anti-CD3 antibody and complement did not recognize IFN-γ-treated GI-1 cells as efficiently as untreated targets (P < 0.05). On the other hand, we found that T-like cells divided by anti-CD16 antibody and complement had the opposite tendency (P < 0.01; Fig. 4B). This difference between NK- and T-like cells in monolayer depletion assays was reproducible, and the results shown herein are representative of those of three similar experiments. Cold target inhibition assays by natural killer-like and T-like cells As a second method for comparing the ability of IFN-γ-treated and IFN-γ-untreated GI-1 cells to be recognized by either NK- or T-like cells, we performed cold target inhibition experiments. Graded amounts of unlabeled inhibitors (either IFN-γ-treated or IFN-γ-untreated GI-1 cells) were added to a fixed number of labeled GI-1 cells not treated with IFN-γ, and the lysis by effectors was determined. As is shown in Fig. 5A, IFN-γ-treated targets clearly inhibit less effectively than do untreated cells for the NKlike cell lysis of untreated targets. On the other hand, Fig. 5B demonstrates that there was no significant difference in inhibition effect between IFN-γ-treated


and IFN-γ-untreated cells for lysis by T-like cells. The reasons for the discrepancy in results between monolayer depletion and cold target inhibition assays for lysis by T-like cells lysis are not clear. Direct binding assay To determine directly the ability of IFN-γ-treated GI-1 cells to be recognized by NK- or T-like cells, we compared IFN-γ-treated and IFN-γ-untreated targets in a direct binding assay in which conjugate formation between GI-1 cells and effectors was determined visually. Representative results are presented in Table 1. We found a consistent reduction in the number of conjugates formed between IFN-γtreated targets and whole LAK or NK-like cells as opposed to those with untreated targets (P < 0.05 and P < 0.01, respectively). Consistent with our findings with monolayer depletion and cold target inhibition assays, these results clearly suggest that IFN-γ treatment of GI-1 cells may reduce their ability to be recognized by both whole LAK and NK-like cells. On the other hand, we found an increase in the number of T-like cell conjugates formed with IFN-γ-treated targets (P < 0.05). This results indicates that IFN-γ treatment increases the binding ability of GI-1 cells to undergo T-like cell lysis. Consequently, we found opposite phenomena occurring with the use of NKand T-like cells. DISCUSSION It is well established that LAK cells as well as NK cells mediate their cytotoxic functions against tumor cells in a major histocompatibility complexnonrestricted manner and can lyse tumor cells in an antigen-nonspecific manner regardless of whether the tumors are NK sensitive or NK resistant. In addition, IFN-γ has been recently reported to decrease the susceptibility of tumor cells to lysis by either LAK or NK cells (5,12). Indeed, we have shown recently that IFN-γ treatment decreased the susceptibility of GI-1 cells to LAK lysis and increased their expression of human lymphocyte antigen Class I molecules in opposing fashions (21), whereas other investigators have reported that IFN-γ can modulate the sensitivity of tumors to LAK lysis by major histocompatibility complex-independent pathways as well (27,29). The mechanism underlying this phenomenon remains controversial (33). The IFN-γ induced protection of target cells from LAK lysis can be the consequence of a variety of changes in the process of cell-mediated cytotoxicity; these changes can be divided into four stages (target recognition, triggering and releasing of cytotoxic factors, and lysing) (13). De Fries and Golub (5) have already reported that the reduction in lysis of IFN-γ-treated target cells is evident immediately after the contact of effector and target cells in kinetic assays. Therefore, to determine whether the increased resistance to lysis in the IFN-γ treated targets occurs at the level of recognition or at a postbinding stage of LAK cells, we have first determined the ability of IFN-γ treated GI-1 cells to be recognized by LAK cells by monolayer depletion and LDCC assays. The results we obtained in monolayer depletion assays demonstrate that IFN-γ-treated monolayer


ability of GI-1 cells to be recognized by LAK cells, we performed a monolayer depletion assay with adherent GI-1 cells. LAK cells were added to either control or IFN-γ-treated GI-1 cell monolayers and incubated for 1 hour to allow the binding of effectors to targets. The effect of monolayer cells in depleting LAK cells by target cell binding was then assessed by the gentle removal of nonadherent effectors and the testing of them for their ability to lyse untreated GI-1 cells. As Fig. 2 demonstrates, IFN-γ-treated monolayer cells were without question less effective than untreated monolayer cells in depleting LAK cells (P < 0.05). Furthermore, to determine whether or not any factor other than target recognition by LAK cells influences the susceptibility of IFN-γtreated GI-1 cells, we determined the difference in susceptibility between IFN-γ-treated and -untreated GI-1 cells to lysis by LDCC. If the resistance to lysis shown by IFN-γ-treated GI-1 cells is manifested at the level of target recognition alone, then the susceptibility of target cells in the presence of lectin should not be affected. As is shown in Fig. 3, we found that there was no difference in susceptibility to LDCC lysis between IFN-γ-treated and IFN-γuntreated GI-1 cells. We also obtained similar results with a LDCC assay with concanavalin A (10 µg/mL; Sigma Chemical Co., St. Louis, MO; data not shown) (5) . These results together suggest that the resistance to lysis shown by IFN-γ-treated tumor cells is manifested at the stage of effector-target recognition alone.


however, in analyzing the literature concerning T-like cells, at least two points can be made: 1) LAK cells prepared from short-term culture are predominantly NK-like cells (37), whereas T-like cells are still too immature for the releasing of cytolytic factors; and 2) T-like cells require larger amounts of IL-2 (1) or a combination of biological agents (i.e., OKT3, tumor necrosis factor [TNF] and IL-2), which yields significant increases in both LAK activity and proliferation (36). In conclusion, we have demonstrated that IFN-γinduced protection of tumor cells from LAK lysis is predominantly regulated by the target recognition of NK-like cells. Moreover, if there are any factors triggering T-like cells to initiate further stages for lysis, the present LAK phenomenon may present an interesting contrast to their behavior. Therefore, further analysis will be necessary to clearly understand the inverse phenomenon shown for T-like cells. From a practical point of view, a high concentration of IFN-γ may be present in vivo, systematically or at the tumor site, during an antitumor immune response. Moreover, the question of how IFN-γ might modulate host-tumor interaction would depend upon the native or immunotherapyinduced response to the tumor and upon which is the dominant population of LAK cells. Therefore, the results obtained in this study may serve as a basis for refining the therapeutic protocols used for LAK immunotherapy. ACKNOWLEDGMENTS We thank Takeda Chemical Industries, Ltd., and Rosch Japan for generously providing recombinant IL-2 (TGP-3) and recombinant IFN-γ (TRP-2). We also thank Dr. Tomokazu Aoki and Mr. Tatsuo Sakagoshi for the preparation of the manuscript. This work was supported in part by a grant in aid for cancer research (2-14) from the Ministry of Health and Welfare of Japan. Received, October 30, 1991. Accepted, March 31, 1992. Reprint requests: Seiji Kondo, M.D., Department of Neurosurgery, Faculty of Medicine, Kyoto University, 54 Kawahara-cho, Sho-goin, Sakyo-ku, Kyoto 606, Japan. REFERENCES: (1-38) 1.

2.

3.

Ballas ZK, Rasmussen W, Van Otegham JK: Lymphokine-activated killer (LAK) cells. II. Delineation of distinct murine LAK-precursor subpopulations. J Immunol 138:1647-1652, 1987. Bukowski JF, Yang H, Welsh RM: Antiviral effects of lymphokine- activated killer cells: Characterization of effector cells mediating prophylaxis. J Virol 62:3642-3648, 1988. Chervenak R, Wolcott RM: Target cell expression of MHC antigens is not (always) a turn-off signal to natural killer cells. J Immunol 140:3712-3716, 1988.


cells failed to deplete LAK cells as effectively as did untreated cells. Moreover, we found that there was no difference in susceptibility to lysis by LDCC between IFN-γ-treated and IFN-γ-untreated GI-1 cells. This result suggests that the change in IFN-γ-treated targets is manifested at the stage of target recognition alone, because the susceptibility of target to LDCC was not affected (5,38). Taken together, these results imply that IFN-γ may reduce the LAK recognition of tumor cells. The negative selection of LAK cells with antibodies and complement or the positive selection by fluorescence-activated cell sorting has clearly shown that LAK cells consist of at least two major populations, the relative contribution of which is dependent upon the donor organ and type of target cells examined (15,23,24). The surface phenotype of one population is CD16+ and CD3- (NK-like cells), whereas the other is CD16- and CD3+ (T-like cells), not only in the precursor but also in the effector phase. We based our approach on the hypotheses that LAK cells are heterogenous and are composed of two major populations and that a large part of the LAK phenomenon that occurs after IFN-γ treatment is induced by NK- and T-like cells. To test these hypotheses, we thought it reasonable to attempt to determine which population was involved in the change of susceptibility that occurred in GI-1 cells after IFN-γ treatment by using monolayer depletion, cold target inhibition, and direct binding assays. In monolayer depletion and direct binding assays with NK-like cells, IFN-γ-treated tumor cells failed to be recognized and bound as effectively as untreated targets. On the other hand, interestingly, we observed the opposite tendency with respect to T-like cells. The similarity of phenomena between LAK and NKlike cell results thus supports the hypothesis that NKlike cells have stronger cytotoxic activity than do Tlike cells and consist of the major portion of LAK cells prepared from short-term culture (15,23). The discrepancy between our data and those of other authors, such as De Fries and Golub (5), may be principally the result of differences in the effector populations used in the experiments. In contrast, the finding of inverse phenomena for NK- and T-like cells may coincide with those of reports that NK- and T-like cells are, respectively, similar to NK and T cells phenotypically and that high Class I expression can potentiate an in vivo response dominated mainly by T cells (38), whereas low Class I expression can favor an NK-mediated response (18). Our finding using cold target inhibition by NK-like cells also supports the hypothesis that IFN-γ may reduce NK-like cell recognition of tumor cells. However, we have observed no change in experiments with inhibition by T-like cells. Additionally, we have previously shown that T-like cells exhibited no change in their lytic activity for IFN-γ-treated and IFN-γ-untreated targets. In agreement with our findings for monolayer depletion and direct binding assays, these results suggest that IFN-γ-treated targets bind to T-like cells but fail to trigger them to initiate further stages for lysis. The reasons for these T-like cell findings are not evident;

5.

6.

7.

8.

9.

10.

11.

12.

13.

14. 15.

D'Amore PJ, Golub SH: Detection of double target binders at the single cell level: An evaluation of the specificity of NK and MLCinduced NK-like cells. Cell Immunol 71:6679, 1982. De Fries RU, Golub SH: Characteristics and mechanism of IFN- gamma-induced protection of human tumor cells from lysis by lymphokine-activated killer cells. J Immunol 140:3686-3693, 1988. De Vries JE, Spits H: Cloned human cytotoxic T lymphocyte (CTL) lines reactive with autologous melanoma cells. I. In vitro generation, isolation, and analysis to phenotype and specificity. J Immunol 132:510519, 1984. Ferrini S, Miescher S, Zocchi R, Von Fliedner V, Moretta A: Phenotypic and functional characterization of recombinant interleukin 2 (rIL-2)-induced activated killer cells: Analysis at the population and clonal levels. J Immunol 138:1297-1302, 1987. Gorelik E, Gunji Y, Herberman RB: H-2 antigen expression and sensitivity of BL6 melanoma cells to natural killer cell cytotoxicity. J Immunol 140:2096-2102, 1988. Greenberg AH, Miller V, Pohajdak B, Jablonski T: The in vivo effects of IFN suppression of the NK target structure, in Herberman RB (ed): NK Cells and Other Natural Effector Cells. New York, Academic Press, 1982, pp 1387-1393. Grimm EA, Mazumder A, Zhang HZ, Rosenberg SA: Lymphokine- activated killer cell phenomenon. Lysis of natural killerresistant fresh solid tumor cells by interleukin 2 activate autologous human peripheral blood lymphocytes. J Exp Med 155:1823-1841, 1982. Grönberg A, Kiessling R, Masucci G, Guervara LA, Eriksson E: Gamma-interferon (IFN-γ) produced during effector and target interactions renders target cells susceptible to NK cell-mediated lysis. Int J Cancer 32:609616, 1983. Grönberg A, Ferm M, Tsai L, Kiessling R: Interferon is able to reduce tumor cell susceptibility to human lymphokine-activated killer (LAK) cells. Cell Immunol 118:10-21, 1989. Henkart PA, Yue CC, Yang J, Rosenberg SA: Cytolytic and biochemical properties of cytoplasmic granules of murine lymphokineactivated killer cells. J Immunol 137:26112617, 1986. Herberman RB, Ortaldo JR: Natural killer cells: Their role in defenses against disease. Science 214:24-30, 1981. Itoh K, Tilden AB, Kumagai K, Balch CM: Leu-11+ lymphocytes with natural killer (NK) activity are precursors of recombinant interleukin 2 (rIL-2)-induced activated killer (AK) cells. J Immunol 134:802-807, 1985.


16. 17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

28.

29.

Kalland T: Generation of natural killer cells from bone marrow precursors in vitro. Immunology 57:493-498, 1986. Kalland T, Belfrage H, Bhiladvala P, Hedlung G: Analysis of the murine lymphokineactivated killer (LAK) cell phenomenon: Dissection of effectors and progenitors into NK- and T-like cells. J Immunol 138:36403645, 1987. Karre K, Ljunggren HG, Piontek G, Kiessling R: Selective rejection of H-2 deficient lymphoma variants suggests alternative immune defence strategy. Nature (Lond) 319:675-678, 1986. Marini S, Guadagni F, Bonmassar E, Potenza P, Guiliani A: Influence of interferon on the functional expression of natural killer target structures of murine lymphoma cells. Cell Immunol 102:113-125, 1986. Miyatake S, Kikuchi K, Iwasaki K, Yamashita J, Li ZY, Namba Y, Hanaoka M: Specific cytotoxic activity of T lymphocyte clones derived from a patient with gliosarcoma. J Neurosurg 69:751-759, 1988. Miyatake S, Kikuchi H, Oda Y, Nishioka T, Takahashi J, Kondoh S, Matsumoto M, Yamasaki T, Iwasaki K, Aoki T, Kasakura S, Namba Y: Decreased susceptibility of lined human gliosarcoma cells to lymphokineactivated killer cell cytolysis by γ-Interferon treatment. Cancer Res 50:596-600, 1990. Moore M, White WJ, Potter MR: Modulation of target cell susceptibility to human natural killer cells by interferon. Int J Cancer 25:565572, 1980. Ortaldo JR, Mason A, Overton R: Lymphokine-activated killer (LAK) cells: Analysis of progenitors and effectors. J Exp Med 164:1193-1205, 1986. Ortaldo JR: Regulation of natural killer activity. Cancer Metastasis Rev 5:637-657, 1987. Phillips JH, Lanier LL: Dissection of the lymphokine-activated killer phenomenon: Relative contribution of peripheral blood natural killer cells and T lymphocytes to cytolysis. J Exp Med 164:1193-1205, 1986. Pross HF, Baines MG, Rubin P: Spontaneous human lymphocyte- mediated cytotoxicity against tumor target cells. IX. The quantitation of natural killer cell activity. J Clin Immunol 1:51-63, 1981. Ramani P, Balkwill FR: Enhanced metastases of a mouse carcinoma after in vitro treatment with murine interferon gamma. Int J Cancer 40:830-834, 1987. Reinhertz EL, Kung PC, Goldstein G, Levey RH, Schlossman SF: Discrete stages of human intrathymic differentiation: Analysis of normal thymocytes and leukemic lymphoblasts of T-cell lineage. Proc Natl Acad Sci USA 77:1588-1592, 1980. Reiter Z, Fischer DG, Rubinstein M: The protective effect of interferon against natural


4.

31.

32.

33.

34.

35. 36.

37.

38.

COMMENTS In this work, Kondo et al. attempted to evaluate the mechanism by which interferon gamma decreases tumor susceptibility to lymphokine-activated killer cell lysis. They chose the GI-1 cell line as an in vitro model. Their results are interpreted as showing that interferon gamma reduces natural killer-like cell recognition of the tumor but fails to inhibit T-like cellinduced lysis. On the basis of these findings, they imply in their discussion that current failures of


therapy, which include the use of lymphokineactivated killer cells, may be based on the fact that lymphokine-activated killer cell production favors natural killer-like cells that are less efficacious in cell killing. Certainly, in vitro studies such as this are warranted to understand the general failure of the use of lymphokine-activated killer cells to induce tumor lysis. They should also point out that although initially promising, lymphokine-activated killer cellbased therapies may not be the optimal mode by which the immune system can be used for cancer control. Strong emphasis should be placed on alternative approaches, for instance, the development of methods recognizing unique tumor molecules. Certainly, the data supporting the use of immunotoxins and antibodies linked to various radioactive compounds warrant further development. Jeffrey J. Olson Atlanta, Georgia This is an interesting article trying to elucidate the effect of interferon gamma (IFN-γ) on lymphokineactivated killer cell lysis in vitro. The information presented by this work is not entirely new, as is admitted by the authors themselves. In order to investigate a protective effect of IFN-γ a little further, the authors tried to determine at which stage of effector-target cell interaction this effect occurs by a monoloyer depletion assay and lectin-dependent cellular cytotoxicity. There are two major problems in this investigation. One is the borderline effect induced by IFN-γ on the cytotoxicity curves, although very high doses of IFN-γ were used (can such high doses occur at the tumor site in vivo?). In all experiments, the E:T ratios needed to kill 50% of the target cells differ by less than one third of a log (i.e., less than one titration step). Even if such differences are shown to be statistically significant, it is possible that the GI-1 cell line is not very sensitive to lymphokine-activated killer cell lysis or is relatively insensitive to the IFN-γ "protection," or is both, thus providing a suboptimal assay system with which to test the hypothesis. In addition, the dose of interleukin-2 used to generate activity is extremely low (2 U/mL). The application of lectin-dependent cellular cytotoxicity on a population including both activated natural killer and T lymphocytes is problematic. The lytic activities induced in this test are a mixture of lymphokine-activated killer cell activity, CD3dependent killing, and to some extent, CD2dependent killing (by /δ T-lymphocytes), implying different mechanisms and different effectors. Thus, no clear conclusion can be drawn from this experiment regarding recognition or triggering. Despite these criticisms this is a very good piece of experimental work. It clearly shows how complex cellular cytotoxicity is and should be a warning to anyone embarking in clinical immunotherapy of gliomas. Vladimir Von Fliedner


30.

killing activity is mediated via the expression of class I MHC antigens. Immunol Lett 17:323-328, 1988. Shiloni E, Eisenthal A, Sachs D, Rosenberg SA: Antibody-dependent cellular cytotoxicity mediated by murine lymphocytes activated in recombinant interleukin 2. J Immunol 138:1992-1998, 1987. Tilden AB, Itoh K, Balch CM: Human lymphokine-activated killer (LAK) cells: identification of two types of effector cells. J Immunol 138:1068-1073, 1987. Trinchieri G, Granato D, Perussia B: Interferon-induced resistance of fibroblasts to cytolysis mediated by natural killer cells: Specificity and mechanism. J Immunol 126:335-340, 1981. Tsai L, Ohlen C, Ljunggren H-G, Karre K, Hansson M, Kiessling R: Effect of IFNtreatment and in vivo passage of murine tumor cell lines on their sensitivity to lymphokine-activated killer (LAK) cell lysis in vitro; association with H-2 expression on the target cells. Int J Cancer 44:669-674, 1989. Welsh RM, Karre K, Hansson M, Kunkel LA, Kiessling RW: Interferon-mediated protection of normal and tumor target cells against lysis by mouse natural killer cells. J Immunol 126:219-225, 1981. Welsh RM: Natural killer cells and interferon. CRC Crit Rev Immunol 5:55-93, 1985. Yang SC, Owen-Schaub LB, Roth JA, Grimm EA: Characterization of OKT3-initiated lymphokine-activated effectors expanded with interleukin 2 and tumor necrosis factor α Cancer Res 50:3526- 3532, 1990. Zanovello P, Rosato A, Bronte V, Cerundolo V, Treves S, Virgilio FDi, Pozzan T, Biasi G, Collavo D: Interaction of lymphokineactivated killer cells with susceptible targets does not induce second messenger generation and cytolytic granule exocytosis. J Exp Med 170:665-677, 1989. Zinkernagel RM, Doherty PC: MHC-restricted cytotoxic T-cells; studies on the biological role of the polymorphic major transplantation antigens determining T-cell restrictionspecificity, function and responsiveness. Adv Immunol 27:51-177, 1979.

Nicolas de Tribolet Lausanne, Switzerland

Redistribution of this article permitted only in accordance with the publisher’s copyright provisions. Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/3/534/2752153 by Frankfurt Univesity Library user on 03 March 2018

Figure 2. Depletion of LAK cells on IFN-γ-treated or IFN-γ-untreated target GI-1 cell monolayers. Solid line, lysis of target cells by LAK cells that were not depleted on target cell monolayers (control), LU = 2.83 ± 0.57; dashed line, lysis by LAK cells depleted on monolayers of target cells treated with 1000 U IFNγ per milliliter for 18 hours, LU = 1.04 ± 0.15; doubledashed line, lysis by LAK cells depleted on monolayers of untreated target cells, LU = 0.68 ± 0.14. Solid line versus dashed line, P < 0.01. Dashed line versus double-dashed line , P < 0.05.



Figure 1. Decreased susceptibility of target GI-1 cells to LAK cells after IFN-γ treatment. GI-1 cells were treated with 1000 U of IFN-γ per milliliter for 18 hours before the cytotoxic assay. The assay was performed with an incubation time of 4 hours. Lysis at each effector/target cell ratio is given as the mean of triplicate wells. Solid line, no treatment, lytic units (LU) = 2.34 ± 0.08 (SD): dashed line, IFN-γ 1000 U/mL, LU = 1.81 ± 0.09. Solid line versus dashed line, P < 0.01.



Figure 3. Lysis of IFN-γ-treated or IFN-γ-untreated GI-1 cells by LAK cells in the presence of PHA-P. Solid line, lysis of untreated target cells without PHAP (control), LU = 9.68 ± 1.05; dashed line, LDCC lysis of IFN-γ-untreated cells with 2 µg of PHA-P per milliliter, LU = 56.48 ± 21.00; double-dashed line , lysis of 1000 U of IFN-γ per milliliter for 18 hours, treated cells without PHA-P, LU = 5.95 ± 0.22; double-dotted line , LDCC lysis of 1000 U of IFN-γ per milliliter for 18 hours, treated cells with 2 µg of PHA-P per milliliter, LU = 32.54 ± 10.62. Dashed line versus double-dashed line , results not significant.



Figure 4. Depletion of NK- (A) and T-like cells (B) on IFN-γ-treated or IFN-γ-untreated target GI-1 cell monolayers. LAK cells were divided with anti-CD3 or anti-CD16 monoclonal antibodies and complement just before a monolayer depletion assay. A, NK-like cells, divided by anti-CD3 monoclonal antibody and complement. Solid line, lysis by NK-like cells that were not depleted on target cell monolayers (control), LU = 3.49 ± 0.84; dashed line, lysis by NK-like cells depleted on monolayers of target cells treated with IFN-γ (1000 U/mL) for 18 hours, LU = 1.82 ± 0.42; double-dashed line , lysis by NK-like cells depleted on untreated monolayers, LU = 0.98 ± 0.08. Solid line versus dashed line, P < 0.05; dashed line versus double-dashed line , P < 0.05. B, T-like cells, divided by anti-CD16 monoclonal antibody and complement. Solid line, control, LU = 2.37 ± 0.91; dashed line, lysis by T-like cells depleted on monolayers of target cells treated with IFN-γ (1000 U/mL) for 18 hours, LU = 0.96 ± 0.09; double-dashed line , lysis by T-like cells depleted on untreated monolayers, LU = 1.56 ± 0.24. Solid line versus double-dashed line , results not significant. Double-dashed line versus dashed line, P < 0.01.

Table 1. Direct Binding of Effectors to Target GI-1 Cells: IFN-γ-treated and IFN-γ-untreated Targets



Figure 5. Cold target inhibition assays by NK- (A) and T-like cells (B). Untreated GI-1 cells were used as labeled target cells. A, solid line, inhibition of lysis of untreated GI-1 cells by NK-like cells; dashed line, inhibition of lysis of GI-1 cells treated with IFN-γ (1000 U/mL) for 18 hours by NK-like cells. B, solid line, inhibition of lysis of untreated GI-1 cells by Tlike cells; dashed line, inhibition of lysis of GI-1 cells treated with IFN-γ (1000 U/mL) for 18 hours by T-like cells.

Human Adaptive Natural Killer Cells: Beyond NKG2C.

Tissue-resident natural killer cells.

Natural killer cells and natural killer T cells in Lyme arthritis.

Lysis of myotubes by alloreactive cytotoxic T cells and natural killer cells. Relevance to myoblast transplantation.

Insights into the paracrine effects of uterine natural killer cells.

Neutrophils induced licensing of natural killer cells.

Development of invariant natural killer T cells.

Isolation of murine natural killer cells.

Subsets of human natural killer cells and their regulatory effects.

Natural Killer Cells for Therapy of Leukemia.

Activation of Natural Killer Cells by Probiotics.

Natural killer cells and tumor metastasis.

Inborn errors of the development of human natural killer cells.

Fibronectin facilitates the migration of human natural killer cells.

Effects of opioid therapy on human natural killer cells.

Natural killer cells: a TACTILE restraint.

Natural killer cells modulation in hematological malignancies.

Natural killer cells in cutaneous malignant melanoma.

Expression of Fc mu receptors on human natural killer cells.

Are natural killer cells superior CAR drivers?

Adipose invariant natural killer T cells.

HPV vaccine stimulates cytotoxic activity of killer dendritic cells and natural killer cells against HPV-positive tumour cells.

Immune surveillance of unhealthy cells by natural killer cells.

Natural killer T cells: Lyme scaled back.