Int. Archs Allergy appl. Immun. 52: 347-354 (1976)
Characterization of K-Cell Activity by Use of Depletion Experiments1 D. Kraft2, D. Franks and R. R. A. Coombs Immunology Division, Department of Pathology, University of Cambridge, Cambridge
Abstract. Normal human blood lymphocytes with affinity for ox red cells sensitized with IgG antibody, for normal or papain-treated sheep red cells and for ox red cells coated with mouse complement were used for rosette formation and the rosetting cells separated by density gradient centrifugation on Ficoll/Hypaque. The non-rosetting cells at the inter face were collected and compared with cell suspensions before treatment for direct and antibody-dependent cytotoxicity of human target cells. Depletion of Fc-receptor-bearing lymphocytes strongly decreased antibody-dependent cell-mediated cytotoxicity; reduction in the number (or depletion) of T cells and cells with C'3 receptor had no effect, showing the same or enhanced K-cell activity. It is concluded that one type of K or killer cell has Fc receptors but lacks C'3 receptors.
It seems possible that the reason for the present differences of opinion about the na ture of the effector cell in antibody-de pendent cell-mediated cytotoxicity is that each investigator uses a different system (different target cells, different antibody and different preparations of effector cells), and that there are, in fact, several different types 1 This study was supported by a grant from the Austrian Research Council (Project No. 2885). 2 Recipient of a fellowship from the Austrian Ministry of Science and Research. Received: May 18, 1976.
of effector cell, even within one species, each with different immunobiological prop erties, and each manifesting a different im munological function [7-9, 17, 19, 23], The purpose of the present study was to determine the nature of the human effector cell which kills human target cells coated with human antibody, and by fractionation and separation of lymphocytes on a density interface after rosetting, to find out whether effector cell activity is associated with T cells or with cells possessing receptors for either Fc or the third component of com plement (C'3). Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Introduction
348
Materials and Methods Lymphocyte Preparation Lymphoid cells were prepared from human peripheral blood as previously reported [5]. For all tests the same normal donor was used. Briefly, defibrinated blood was diluted by mixing 6 vol with 2 vol of phosphate-buffered saline (PBS) and 2.66 vol of 1% methylcellulose (Koch-Light, vis cosity 17-23 cP at 20 °C) in PBS. After sedimen tation of the red cells for 30 min at 37 °C, the su pernatant containing the white cells was trans ferred to new tubes, centrifuged and washed twice in Medium 199 supplemented with 10°/o fetal calf serum, previously heated to destroy complement. Phagocytic cells were removed by adding 200 mg carbonyliron SF (Fe2 [CO]q) (GAF, UK) to 5 ml cell suspension (2-3X10® cells/ml), with shaking every 10 min during incubation at 37 °C for 30 min and using a powerful magnet to remove all cells which had ingested carbonyliron particles. The supernatants were adjusted to a final concen tration of 2X10® cells/ml. All cells were prepared in Sterilin UGB size tubes (90X24 mm) and there fore only non-adherent cells were recovered for rosette formation and cytotoxicity experiments. Indicator Red Cells for Rosette Formation Fc-rosetting reaction: ox red blood cells stored in ADC were washed 3 times in PBS and a 2% suspension was made. The ox cells were sensitized with rabbit anti-ox antiserum at a subagglutinat ing dose for 45 min at room temperature [10]. After sensitization, the cells were washed twice in PBS and a 0.8% suspension was prepared in PBS. Sheep-red-cell-rosetting reaction (spontaneous rosetting): sheep red cells (SRBC) stored in ACD were washed 3 times with PBS, then used as a 1% suspension in PBS. Furthermore, for depletion ex periments 1% papain-treated SRBC which give more stable rosettes [26] were used in 2 of 3 ex periments. C'3-rosetting reaction: a 2% suspension of ox red blood cells was incubated with rabbit IgM anti-ox antibody for 15 min at 4 °C in PBS [10]. The sensitized cells were washed twice in PBS and re suspended in complement fixation test diluent (CFT; Oxoid) as a 1.0% suspension. An equal vol ume of diluted mouse serum (1/10) from C'5deficient AKR-mice was added to the IgM-coated
ox red cells and then incubated for 15 min at 37 °C. The cells were then washed twice with PBS and a 0.8% suspension was made in PBS. The complement source was absorbed with ox red cells and the activity titrated enumerating the number of C'3 rosettes in a given lymphocyte suspension using increasing amounts of complement. For control, suspensions of ox red cells were used af ter incubation with heat-inactivated serum or in PBS only. Finally, all red cell preparations were checked by rabbit anti-mouse C'3 serum (a gift from Dr. H. Waldmann) in microtitration plates. Rosette formation: for testing, 1 drop of indi cator cells and 1 drop of lymphocytes were mixed in a silicone-coated glass tube (50X4 mm) and the lymphocyte indicator red cell mixture was centri fuged immediately at 200 g for 3 min at room temperature. After leaving the tube at 4 °C for 1 h, the pellet was resuspended, mounted on a glass-slide and stained with toluidine blue. 300 lymphocytes from each preparation were counted and the percentage of rosetting lymphocytes (i.e., lymphocytes binding at least 3 or more indicator cells) was recorded; all tests were performed in duplicates, sometimes in 3 or 4 replicates. Depletion experiments: 5 ml of lymphocyte suspension (2X10® cells/ml) were mixed in Sterilin UGB size tubes (90X24 mm) with the same vol ume of indicator cells (prepared under sterile con ditions) as 0.8% IgG or IgM-C'3 ox red cells and 1% normal or papain-treated SRBC suspensions and the cells immediately spun down at 200 g for 3 min at room temperature. The pellets were then kept at 4 °C for 1 h and afterwards carefully re suspended by hand. For separation of non-rosetting lymphocytes, the resuspended suspensions were layered on 10 ml of sterile Ficoll/Hypaque solution [4], After continuous centrifugation for 10 min at 200 g, 10 min at 300 g and 10 min at 400 g, the layer at the interface was collected, washed 3 times with Medium 199 supplemented with 10% fetal calf serum (FCS) and the cell up take adjusted to 2X10® cells/ml with the same me dium. The cells were then used in rosetting experi ments (i.e., re-rosetting) and cytotoxicity assays. Lymphocytes treated in the same manner, using unsensitized ox red cells or collected after Ficoll/ Hypaque treatment without red cells served as controls.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Kraft/Franks/Coombs
Characterization of K-Cell Activity
interaction was found in the analysis of variance, a standard error of the mean was calculated from the residual mean square of the analysis of vari ance, and used to test whether means were signifi cantly different by Duncan’s [2] multiple-range test using the improved tables calculated by Har ter [12]. In all experiments, this was done on the mean of the logarithms of replicate tubes; in the text and tables, an indication is given of the ioge least significant difference and what differences in isotope release this corresponds to. Rosette counts were analyzed by analysis of variance on the orig inal counts, and differences between means of the counts were tested by Duncan’s multiple-range test.
Results Experimental Procedures The use of an anti-HL-A serum dilution of 1:500, lymphocyte to target cell ratios of 40:1 and 10:1 and an incubation time of 14 h gave sufficient lysis for significant re sults and therefore these conditions were chosen for all experiments. Washing of 51Crlabelled and antibody-coated target cells be fore the cytotoxicity assay in 2 experiments resulted in a decreased radio-isotope labell ing of the target cells, but gave no evidence for any inhibitory effect by residual HL-A antibodies on the particular effector lym phocytes used in these experiments. (This may not be true with other effector lympho cytes.) Thereafter all tests were performed without washing the target cells after addi tion of anti-HL-A serum. Rosette Formation In 5 experiments investigating the peri pheral blood of one donor, Fc-receptorbearing cells ranged between 27 and 39% (mean 33) of the total counted lymphocytes as shown by rosetting with IgG-sensitized
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Target cells: Detroit 6 (Det 6), a continuous human cell line was maintained in monolayer cul ture on glass in Medium 199 with 10% FCS. This cell line was obtained originally from the Ameri can Type Culture Collection (ATCC CCL3). Fetal calf serum: this was checked by immunoelectrophoresis to ensure that it had not been contaminated with post-suckling calf serum. For use in cytotoxicity assays, it was heated at 56 °C for 2 h before mixing with Medium 199. Antiserum: several human sera containing anti-HL-A antibodies were kindly given to us by Dr. Valerie Joysey. After screening for the ability to induce antibody-dependent cell-mediated cytotox icity, one serum was selected for use in all experi ments. The specificity of the antibodies in this serum is unknown. Cytotoxicity assay: the cytotoxicity in vitro of human lymphocytes for 51Cr-labelled Det-6 cells with and without anti-HL-A antibodies was tested in a microassay [3, 5], Target cells were labelled with 5,Cr-chromate (Na251C r0 4, 0.5-1.0 mCi/ml (The Radiochemical Centre, Amersham, England) [4] and made up to a concentration of 5X104 cells/ml in Medium 199 with 10% FCS. To half of the cell suspensions an appropriate amount of an tiserum was added (i.e., a dilution of 1:500 was found to be sufficient) and the mixture allowed to stand 30 min at 22 °C. 100 u\ sensitized or nonsensitized target cells (5X103 cells) and 100^1 effector cells (2X105 or 5X104 cells) for a final effector to target cell ratio of 40:1 and 10:1 were mixed in 6X37 mm sterile plastic tubes and incu bated at 37 °C in a sealed box gassed with 5% C 0 2 for 14 h. At the end of the incubation, the cells were resuspended to ensure dispersal of the released isotope, the cells then centrifuged down (600 g for 2 min) and 100 «1 supernatant removed to another tube. Both tubes were counted in an LKB-Wallac gamma-counter with a 5-cm sodium iodide crystal. Controls for each experiment in cluded incubation of sensitized or unsensitized tar get cells in medium only. The cytotoxicity was measured as percent isotope release from SICr-labelled target cells. Each experimental group was performed in quadruplicate. Statistical analysis: all cytotoxicity experi ments were subjected to analysis of variance after logarithmic (log0) transformation of the percen tage of isotope released. If a significant effect or
349
350
Kraft/Franks/Coombs
Table I. Cytotoxic activity of lymphocytes before and after depletion of SRBC-rosette-forming cells (T lympho cytes), Fc-receptor-bearing cells and complement-rcceptor-bearing cells Expe-
Preparation
Rosettes, %
rlments
T cells Fc+
51Cr release, % C3+
direct non-specific 10:1 medium control
40:1
10:1 medium control
40 35* 44*
33
82 46* 85
68 40* 78*
37
88 44* 83
1
Untreated Fc-depleted T-depleted
63 90* 20*
35 1 69*
47 37* 48
2
Untreated Fc-depleted T-depleted
65 84* nd
33 1* nd
46 39* 48
3
Untreated Fc-depleted C3-depleted
58 84* 64*
31 1* 35
4
Untreated T-depleted
58 5*
39 81*
5
Untreated C3-depleted
62 59*
27 29
20 nd l*
22 1*
ADCC
40:1
36
37
56 49* 54
53 48* 52
49
77 53* 80
68 50* 68
48
39 40
33 36
28
71 82*
57 75*
28
42 45
35 39*
30
61 71*
51 57*
30
* Rosette frequency or 51Cr release significantly reduced or increased as compared with untreated cells Summary of statistical analysis Expe Rosette counts riments ANOVA multiple-range test treatment x rosette least significant corresponds interaction
1 2 3 4 5
F
d.f.
(5% level)
991 262 275 416 21
2,12 1,4 3,8 1,4 2,10
16 46 13 25 9
Cytotoxicity ANOVA multiple-range test treatment x target ,oge ,east corresponds to interaction ■_____ .
of 300 cells, % F 5 15 4 8 3
45 27 21 43 17
d.f.
difference (5% level)
40% 60% 80%
8, 54 6,36 8, 54 12,63 10, 66
0.06-0.07 0.08-0.09 0.07-0.08 0.06-0.07 0.07-0.08
3 4 3 3 3
4 6 5 4 5
6 8 8 6 6
ox red blood cells. In the same experiments using SRBC for spontaneous-rosetting tests, the values ranged between 58 and 65%> (mean 61). Testing for C'3-receptor-bearing
cells performed with ox red cells carrying rabbit anti-ox-IgM antibody and mouse C'3 values of 20 and 22% (mean 21) were found in 2 experiments.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Values in this table have been rounded; tests of significance were done as described in the text, and least significant differences are given only as a guide. F values correspond to a probability of less than 0.1 % in all experiments.
Characterization of K-Cell Activity
Cytotoxicity of Different Lymphoid Cell Preparations In 5 experiments, the isotope released from non-sensitized target cells by lympho cytes were between 7 and 14°/o (above the medium controls) at a 40:1 effector to target cell ratio and amounts of half as high were obtained at a 10:1 ratio. Addition of antiHL-A antibody to the target cells increased these levels to between 29 and 51% above the medium controls (for the 40:1 ratio) and between 20 and 32% (for the 10:1 ratio). Cell suspensions depleted of Fc-receptorbearing lymphocytes (table I) showed a strong reduction in both spontaneous and antibody-mediated cytotoxicity; with less than 1% of Fc-rosette-forming cells in 3 ex periments (and with T-cell counts of 84, 84 and 90%) the remaining isotope release in the antibody-dependent cellular cytotox icity (ADCC) was not significantly higher
than that ascribable to the direct cytotoxic activity. Cell preparations depleted of T cells (ta ble 1) had the same or increased antibodydependent cytotoxic potential: whereas re duction of SRBC-rosette-forming cells from 62 to 20% made no significant difference, the reduction from 58 to 5% T cells was fol lowed by an increase in the antibody-dependent cell-mediated cytotoxic 51Cr re lease. After depletion of cells with receptors for mouse complement (table I) the remain ing cell preparations containing less than 1% C'3-receptor-bearing cells (but with high numbers of Fc-receptor-positive cells) exhibited increased antibody-dependent kill ing activity. However, these depleted sus pensions showed no marked difference in direct nonspecific lysis of target cells. In these latter experiments, the percentages of T cells varied only slightly from the original levels.
Discussion The percentages of rosette-forming peri pheral blood lymphocytes were found to dif fer only slightly in 5 experiments, but were in the normal range with mean levels of 33 for Fc rosettes, 61 for SRBC rosettes and 21 for C3 rosettes. The last was higher than previously reported [10]; the reason for this may be the use of mouse serum (in the place of human) as source of C'3. Using lympho cyte preparations after removing the phago cytic and plastic-adherent cells, the activi ties in the antibody-dependent cellular cyto toxicity model as shown by 5ICr release of target cells were 19-42% above the direct cytotoxicity levels after 14 h incubation
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Depletion Experiments Using rosette formation by mixing large quantities (5 ml) of lymphoid cells with the same volume of sensitized ox RBC and nor mal SRBC respectively, the non-rosette-for ming cells were separated by centrifugation on Ficoll/Hypaque and were concentrated at the interface between the layers of Medium 199/PBS and Ficoll/Hypaque. Re-rosetting of these cell preparations showed considera bly decreased numbers of Fc- or C'3receptor-bearing cells (table I) but gave evidence of insufficient T-cell depletion in 1 experiment (table I; experiment 1); in two further experiments, papain-treated SRBC were used to obtain stronger rosettes and after Ficoll-Hypaque centrifugation a reduction in the number of E-rosette-form ing cells was achieved to a level of 5% (ta ble I; experiment 4).
351
Kraft/Franks/Coombs
352
exclude the possibility that there are some cytotoxic effector cells which have C'3 re ceptors as well as Fc receptors [14, 15, 24], Removal of one type of cytotoxic cells could enrich another cell population with the same or higher cytotoxic activity. For example, this was reported by Papamichael and Temple [21] who separated glassadherent from non-adherent monocytes and found both preparations active in the antibody-dependent cellular cytotoxicity model. In general, there is evidence for the inde pendence of some functions of the Fc and C'3 receptors of human lymphocytes and for the heterogeneity of K cells (which are lymphoid cells with Fc receptors and killing potential in the antibody-dependent cellular cytotoxicity model) in different situations [1, 6, 13, 18-20, 23, 25]; but the possibility remains that one cell type could be more important than others in relation to in vivo situations like immunological tumour re gression and transplantation rejection.
References 1 Butterworth, A. E.; Sturrock, R. F., and Houba, V.: Antibody-dependent cell-mediated damage to schistosomula in vitro. Nature, Lond. 252: 503-505 (1974). 2 Duncan, P. B.: Multiple range and multiple F-tests. Biometrics 11: 1-12 (1955). 3 Eremin, A.; Ashby, J., and Franks, D.: Killer cell (K) activity in human normal lymph node, regional tumour lymph node and inflammato ry lymph node. II. Int. Archs Allergy appl. Immun (to be submitted). 4 Franks, D.; Kelly, M. M., and Sanderson, C. J.: Inhibition of xenogeneic cell-mediated cyto toxicity in the rat by antiglobulin sera. Int. Archs Allergy appl. Immun. 48: 621-631 (1975). 5 Franks, D.; Tayabali, B., and O’Brien, P.: Competitive inhibition of lymphoid cell me
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
time. A one-step separation procedure for depleting Fc-receptor- and C'3-receptorbearing and SRBC-rosetting lymphocytes was performed. The method is based on the fact that rosette-forming lymphocytes will sediment and the non-rosette-forming lym phocytes float on gradient centrifugation on Ficoll/Hypaque [22]. Modifications were applied using papain-treated SRBC for T-cell-specific rosettes [26] and with slowvelocity centrifugation to prevent disruption of the rosettes on the dense surface of the Ficoll/Hypaque layer. The efficiency of de pleting Fc- and C'3-receptor-bearing lym phocytes was >99% . T-cell depletion was less successful, but even the partially deplet ed samples gave sufficient information. The ADCC corresponding to each preparation of non-rosetting cells showed markedly de creased levels after depletion of Fc-rosetteforming lymphocytes which is in agreement with results of other workers using the same technique [11, 27] or columns coated with aggregated IgG or immune complexes [16, 27], After T-cell depletion, information was obtained that T cells have no influence on K-cell activity and the results correlated well with data of experiments using T-cellenriched [23] or T-cell-depleted fractions [24, 27]. However, after depletion of C'3rosette-forming lymphocytes, we found in creased levels of antibody-dependent cellu lar cytotoxicity which indicates lymphoid K cells with Fc receptor, but lacking C'3 re ceptor. In contrast to this, Perlmann et al. [23] found no killing activity of cell prepa rations depleted of C'3-rosette-forming lym phocytes. A possible explanation for this discrepancy may be that different target cells (Perlmann et al. used chicken red cells) are preferentially lyzed by different effector cells. On the other hand, our results do not
6
7
8
9
10
11
12
13
14
15
diated cytotoxicity analysed by double labell ing with 5,Cr-chromate and S8'nTc-pertechnetate (in preparation). Gale, R. P. and Zighelboim, J.: Polymorphon uclear leukocytes in antibody-dependent cellu lar cytotoxicity. J. Immun. 114: 1047-1051 (1975). Greenberg, A. H.; Shen, L., and Roitt, I. M.: Characterization of the antibody-dependent cy totoxic cell. A non-phagocytic monocyte? Clin, exp. Immunol. 15: 251-259 (1973). Greenberg, A. H.; Shen, L., and Medley, G.: Characteristics of the effector cells mediating cytotoxicity against antibody-coated target cells. I. Phagocytic and non-phagocytic effector cell activity against erythrocyte and tumour target cells in a 51Cr release cytotoxicity assay and [125I] IUdR growth inhibition assay. Immunolgy 29: 719-729 (1975). Greenberg, A. H.; Shen, L.; Walker, L.; Arnaiz-Villena, A., and Roitt, I. M.: Characteris tics of the effector cells mediating cytotoxicity against antibody-coated target cells. II. The mouse nonadherent K cell. Eur. J. Immunol. 5: 475-480 (1975). Haegert, D. G.; Hallberg, T., and Coombs, R. R. A.: B and T lymphocyte subpopulations in human peripheral blood. Int. Archs Allergy appl. Immun. 46: 525-538 (1974). Hallberg, T.: In vitro cytotoxicity of human lymphocytes: reduction of the lymphocyte cy totoxicity induced by antibodies or phytohem agglutination by removing immune-complex binding cells. Scand. J. Immunol. 3: 645-654 (1974). Harter, M. L.: Critical tables for Duncan’s new multiple range test. Biometrics 16: 671-685 (1960). Holm, G.: Lysis of antibody-treated human erythrocytes by human leukocytes and macro phages in tissue culture. Int. Archs Allergy appl. Immun. 43: 671-682 (1972). Holm, G.; Bjorkholm, M.; Mellstedt, H., and Johansson, B.: Cytotoxic activity of lympho cytes from patients with Hodgkin’s disease. Clin. exp. Immunol. 21: 376-383 (1975). Huber, H. and Holm, G.: Surface receptors of mononuclear phagocytes: effect of immune complexes on in vitro function in human mono cytes; in van Furth Mononuclear phagocytes in
353
16
17
18
19
20
21
22
23
24
25 26
immunity, infection and pathology, pp. 291301 (Blackwell, Oxford 1975). Isturiz, M. A. de; Bracco, M. M. de, and Manni, J. A.: Reaction of human lymphocytes with aggregated IgG columns. Effect on antibodydependent cytotoxicity and EAC rosette for mation. Cell. Immunol. 16: 82-91 (1975). Kovithavongs, T.; Rice, G.; Thong, K. L., and Dossetor, J. B.: Effector cell activity in anti body mediated cell dependent immune lysis. II. Evidence for different populations of effector cells for different targets. Cell. Immunol. 18: 167-175 (1975). MacDermott, R. P.; Chess, L., and Schlossman, St. F.: Immunologic functions of isolated human lymphocyte subpopulations. V. Isola tion and functional analysis of a surface Ig negative, E rosette negative subset. Clin. Im munol. Immunopathol. 4: 415-424 (1975). MacDonald, H. R.; Bonnard, G. D.; Sordat, B., and Zawodnik, S. A.: Antibody-dependent cell-mediated cytotoxicity: heterogeneity of effector cells in human peripheral blood. Scand. J. Immunol. 4: 487-497 (1975). MacLennan, I. C. M.: Comments on antibodydependent, cell (K) mediated cytotoxicity; in Advances in the bio-sciences. 12th ScheringSymposium on Immunopathology, pp. 87-91 (Pergamon Press, Vieweg 1974). Papamichael, M. and Temple, A.: Characteri zation of the human effector cell causing anti body-mediated cytotoxicity. Clin. exp. Immu nol. 20: 459-467 (1975). Parish, C. R.: Separation and functional analy sis of subpopulations of lymphocytes bearing complement and Fc receptors. Transplantn Rev. 2J: 98-120 (1975). Perlmann, P.; Perlmann, H.; Larsson, A., and Wahlin, B.: Antibody dependent cytotoxic effector lymphocytes (K cells) in human blood. J. reticuloendoth. Soc. 17: 241-250 (1975). Perlmann, P.; Perlmann, H., and MüllerEberhard, H. J.: Cytotoxic lymphocytic cells with complement receptors in human blood. Induction of cytolysis by IgG antibody but not by target cell-bound C3. J. exp. Med. 141: 287-2% (1975). Roitt, I. M.; Essential immunology, p. 136 (Blackwell, Oxford 1974). Wilson, A. B.; Haegert, D. G., and Coombs,
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Characterization of K-Cell Activity
R. R. A.: Increased sensitivity of the rosette forming reaction of human T lymphocytes with sheep erythrocytes afforded by papain treatment of the sheep cells. Clin. exp. Immu nol. 22: 177-182 (1975). 27 Wislpff, F.; Frpland, S. S., and Michaelsen, T. E.: Antibody-dependent cytotoxicity mediated by human Fc-receptor-bearing cells lacking
Kraft/Franks/Coombs
markers for B- and T-lymphocytes. Int. Archs Allergy appl. Immun. 47: 139-154 (1974).
Correspondence to: Dr. D. Kraft, Institute of General and Experimental Pathology, University of Vienna, Wahringerstrasse 13, A-1090 Vienna (Austria)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
354
Characterization of K-Cell Activity by Use of Depletion Experiments1 D. Kraft2, D. Franks and R. R. A. Coombs Immunology Division, Department of Pathology, University of Cambridge, Cambridge
Abstract. Normal human blood lymphocytes with affinity for ox red cells sensitized with IgG antibody, for normal or papain-treated sheep red cells and for ox red cells coated with mouse complement were used for rosette formation and the rosetting cells separated by density gradient centrifugation on Ficoll/Hypaque. The non-rosetting cells at the inter face were collected and compared with cell suspensions before treatment for direct and antibody-dependent cytotoxicity of human target cells. Depletion of Fc-receptor-bearing lymphocytes strongly decreased antibody-dependent cell-mediated cytotoxicity; reduction in the number (or depletion) of T cells and cells with C'3 receptor had no effect, showing the same or enhanced K-cell activity. It is concluded that one type of K or killer cell has Fc receptors but lacks C'3 receptors.
It seems possible that the reason for the present differences of opinion about the na ture of the effector cell in antibody-de pendent cell-mediated cytotoxicity is that each investigator uses a different system (different target cells, different antibody and different preparations of effector cells), and that there are, in fact, several different types 1 This study was supported by a grant from the Austrian Research Council (Project No. 2885). 2 Recipient of a fellowship from the Austrian Ministry of Science and Research. Received: May 18, 1976.
of effector cell, even within one species, each with different immunobiological prop erties, and each manifesting a different im munological function [7-9, 17, 19, 23], The purpose of the present study was to determine the nature of the human effector cell which kills human target cells coated with human antibody, and by fractionation and separation of lymphocytes on a density interface after rosetting, to find out whether effector cell activity is associated with T cells or with cells possessing receptors for either Fc or the third component of com plement (C'3). Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Introduction
348
Materials and Methods Lymphocyte Preparation Lymphoid cells were prepared from human peripheral blood as previously reported [5]. For all tests the same normal donor was used. Briefly, defibrinated blood was diluted by mixing 6 vol with 2 vol of phosphate-buffered saline (PBS) and 2.66 vol of 1% methylcellulose (Koch-Light, vis cosity 17-23 cP at 20 °C) in PBS. After sedimen tation of the red cells for 30 min at 37 °C, the su pernatant containing the white cells was trans ferred to new tubes, centrifuged and washed twice in Medium 199 supplemented with 10°/o fetal calf serum, previously heated to destroy complement. Phagocytic cells were removed by adding 200 mg carbonyliron SF (Fe2 [CO]q) (GAF, UK) to 5 ml cell suspension (2-3X10® cells/ml), with shaking every 10 min during incubation at 37 °C for 30 min and using a powerful magnet to remove all cells which had ingested carbonyliron particles. The supernatants were adjusted to a final concen tration of 2X10® cells/ml. All cells were prepared in Sterilin UGB size tubes (90X24 mm) and there fore only non-adherent cells were recovered for rosette formation and cytotoxicity experiments. Indicator Red Cells for Rosette Formation Fc-rosetting reaction: ox red blood cells stored in ADC were washed 3 times in PBS and a 2% suspension was made. The ox cells were sensitized with rabbit anti-ox antiserum at a subagglutinat ing dose for 45 min at room temperature [10]. After sensitization, the cells were washed twice in PBS and a 0.8% suspension was prepared in PBS. Sheep-red-cell-rosetting reaction (spontaneous rosetting): sheep red cells (SRBC) stored in ACD were washed 3 times with PBS, then used as a 1% suspension in PBS. Furthermore, for depletion ex periments 1% papain-treated SRBC which give more stable rosettes [26] were used in 2 of 3 ex periments. C'3-rosetting reaction: a 2% suspension of ox red blood cells was incubated with rabbit IgM anti-ox antibody for 15 min at 4 °C in PBS [10]. The sensitized cells were washed twice in PBS and re suspended in complement fixation test diluent (CFT; Oxoid) as a 1.0% suspension. An equal vol ume of diluted mouse serum (1/10) from C'5deficient AKR-mice was added to the IgM-coated
ox red cells and then incubated for 15 min at 37 °C. The cells were then washed twice with PBS and a 0.8% suspension was made in PBS. The complement source was absorbed with ox red cells and the activity titrated enumerating the number of C'3 rosettes in a given lymphocyte suspension using increasing amounts of complement. For control, suspensions of ox red cells were used af ter incubation with heat-inactivated serum or in PBS only. Finally, all red cell preparations were checked by rabbit anti-mouse C'3 serum (a gift from Dr. H. Waldmann) in microtitration plates. Rosette formation: for testing, 1 drop of indi cator cells and 1 drop of lymphocytes were mixed in a silicone-coated glass tube (50X4 mm) and the lymphocyte indicator red cell mixture was centri fuged immediately at 200 g for 3 min at room temperature. After leaving the tube at 4 °C for 1 h, the pellet was resuspended, mounted on a glass-slide and stained with toluidine blue. 300 lymphocytes from each preparation were counted and the percentage of rosetting lymphocytes (i.e., lymphocytes binding at least 3 or more indicator cells) was recorded; all tests were performed in duplicates, sometimes in 3 or 4 replicates. Depletion experiments: 5 ml of lymphocyte suspension (2X10® cells/ml) were mixed in Sterilin UGB size tubes (90X24 mm) with the same vol ume of indicator cells (prepared under sterile con ditions) as 0.8% IgG or IgM-C'3 ox red cells and 1% normal or papain-treated SRBC suspensions and the cells immediately spun down at 200 g for 3 min at room temperature. The pellets were then kept at 4 °C for 1 h and afterwards carefully re suspended by hand. For separation of non-rosetting lymphocytes, the resuspended suspensions were layered on 10 ml of sterile Ficoll/Hypaque solution [4], After continuous centrifugation for 10 min at 200 g, 10 min at 300 g and 10 min at 400 g, the layer at the interface was collected, washed 3 times with Medium 199 supplemented with 10% fetal calf serum (FCS) and the cell up take adjusted to 2X10® cells/ml with the same me dium. The cells were then used in rosetting experi ments (i.e., re-rosetting) and cytotoxicity assays. Lymphocytes treated in the same manner, using unsensitized ox red cells or collected after Ficoll/ Hypaque treatment without red cells served as controls.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Kraft/Franks/Coombs
Characterization of K-Cell Activity
interaction was found in the analysis of variance, a standard error of the mean was calculated from the residual mean square of the analysis of vari ance, and used to test whether means were signifi cantly different by Duncan’s [2] multiple-range test using the improved tables calculated by Har ter [12]. In all experiments, this was done on the mean of the logarithms of replicate tubes; in the text and tables, an indication is given of the ioge least significant difference and what differences in isotope release this corresponds to. Rosette counts were analyzed by analysis of variance on the orig inal counts, and differences between means of the counts were tested by Duncan’s multiple-range test.
Results Experimental Procedures The use of an anti-HL-A serum dilution of 1:500, lymphocyte to target cell ratios of 40:1 and 10:1 and an incubation time of 14 h gave sufficient lysis for significant re sults and therefore these conditions were chosen for all experiments. Washing of 51Crlabelled and antibody-coated target cells be fore the cytotoxicity assay in 2 experiments resulted in a decreased radio-isotope labell ing of the target cells, but gave no evidence for any inhibitory effect by residual HL-A antibodies on the particular effector lym phocytes used in these experiments. (This may not be true with other effector lympho cytes.) Thereafter all tests were performed without washing the target cells after addi tion of anti-HL-A serum. Rosette Formation In 5 experiments investigating the peri pheral blood of one donor, Fc-receptorbearing cells ranged between 27 and 39% (mean 33) of the total counted lymphocytes as shown by rosetting with IgG-sensitized
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Target cells: Detroit 6 (Det 6), a continuous human cell line was maintained in monolayer cul ture on glass in Medium 199 with 10% FCS. This cell line was obtained originally from the Ameri can Type Culture Collection (ATCC CCL3). Fetal calf serum: this was checked by immunoelectrophoresis to ensure that it had not been contaminated with post-suckling calf serum. For use in cytotoxicity assays, it was heated at 56 °C for 2 h before mixing with Medium 199. Antiserum: several human sera containing anti-HL-A antibodies were kindly given to us by Dr. Valerie Joysey. After screening for the ability to induce antibody-dependent cell-mediated cytotox icity, one serum was selected for use in all experi ments. The specificity of the antibodies in this serum is unknown. Cytotoxicity assay: the cytotoxicity in vitro of human lymphocytes for 51Cr-labelled Det-6 cells with and without anti-HL-A antibodies was tested in a microassay [3, 5], Target cells were labelled with 5,Cr-chromate (Na251C r0 4, 0.5-1.0 mCi/ml (The Radiochemical Centre, Amersham, England) [4] and made up to a concentration of 5X104 cells/ml in Medium 199 with 10% FCS. To half of the cell suspensions an appropriate amount of an tiserum was added (i.e., a dilution of 1:500 was found to be sufficient) and the mixture allowed to stand 30 min at 22 °C. 100 u\ sensitized or nonsensitized target cells (5X103 cells) and 100^1 effector cells (2X105 or 5X104 cells) for a final effector to target cell ratio of 40:1 and 10:1 were mixed in 6X37 mm sterile plastic tubes and incu bated at 37 °C in a sealed box gassed with 5% C 0 2 for 14 h. At the end of the incubation, the cells were resuspended to ensure dispersal of the released isotope, the cells then centrifuged down (600 g for 2 min) and 100 «1 supernatant removed to another tube. Both tubes were counted in an LKB-Wallac gamma-counter with a 5-cm sodium iodide crystal. Controls for each experiment in cluded incubation of sensitized or unsensitized tar get cells in medium only. The cytotoxicity was measured as percent isotope release from SICr-labelled target cells. Each experimental group was performed in quadruplicate. Statistical analysis: all cytotoxicity experi ments were subjected to analysis of variance after logarithmic (log0) transformation of the percen tage of isotope released. If a significant effect or
349
350
Kraft/Franks/Coombs
Table I. Cytotoxic activity of lymphocytes before and after depletion of SRBC-rosette-forming cells (T lympho cytes), Fc-receptor-bearing cells and complement-rcceptor-bearing cells Expe-
Preparation
Rosettes, %
rlments
T cells Fc+
51Cr release, % C3+
direct non-specific 10:1 medium control
40:1
10:1 medium control
40 35* 44*
33
82 46* 85
68 40* 78*
37
88 44* 83
1
Untreated Fc-depleted T-depleted
63 90* 20*
35 1 69*
47 37* 48
2
Untreated Fc-depleted T-depleted
65 84* nd
33 1* nd
46 39* 48
3
Untreated Fc-depleted C3-depleted
58 84* 64*
31 1* 35
4
Untreated T-depleted
58 5*
39 81*
5
Untreated C3-depleted
62 59*
27 29
20 nd l*
22 1*
ADCC
40:1
36
37
56 49* 54
53 48* 52
49
77 53* 80
68 50* 68
48
39 40
33 36
28
71 82*
57 75*
28
42 45
35 39*
30
61 71*
51 57*
30
* Rosette frequency or 51Cr release significantly reduced or increased as compared with untreated cells Summary of statistical analysis Expe Rosette counts riments ANOVA multiple-range test treatment x rosette least significant corresponds interaction
1 2 3 4 5
F
d.f.
(5% level)
991 262 275 416 21
2,12 1,4 3,8 1,4 2,10
16 46 13 25 9
Cytotoxicity ANOVA multiple-range test treatment x target ,oge ,east corresponds to interaction ■_____ .
of 300 cells, % F 5 15 4 8 3
45 27 21 43 17
d.f.
difference (5% level)
40% 60% 80%
8, 54 6,36 8, 54 12,63 10, 66
0.06-0.07 0.08-0.09 0.07-0.08 0.06-0.07 0.07-0.08
3 4 3 3 3
4 6 5 4 5
6 8 8 6 6
ox red blood cells. In the same experiments using SRBC for spontaneous-rosetting tests, the values ranged between 58 and 65%> (mean 61). Testing for C'3-receptor-bearing
cells performed with ox red cells carrying rabbit anti-ox-IgM antibody and mouse C'3 values of 20 and 22% (mean 21) were found in 2 experiments.
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Values in this table have been rounded; tests of significance were done as described in the text, and least significant differences are given only as a guide. F values correspond to a probability of less than 0.1 % in all experiments.
Characterization of K-Cell Activity
Cytotoxicity of Different Lymphoid Cell Preparations In 5 experiments, the isotope released from non-sensitized target cells by lympho cytes were between 7 and 14°/o (above the medium controls) at a 40:1 effector to target cell ratio and amounts of half as high were obtained at a 10:1 ratio. Addition of antiHL-A antibody to the target cells increased these levels to between 29 and 51% above the medium controls (for the 40:1 ratio) and between 20 and 32% (for the 10:1 ratio). Cell suspensions depleted of Fc-receptorbearing lymphocytes (table I) showed a strong reduction in both spontaneous and antibody-mediated cytotoxicity; with less than 1% of Fc-rosette-forming cells in 3 ex periments (and with T-cell counts of 84, 84 and 90%) the remaining isotope release in the antibody-dependent cellular cytotox icity (ADCC) was not significantly higher
than that ascribable to the direct cytotoxic activity. Cell preparations depleted of T cells (ta ble 1) had the same or increased antibodydependent cytotoxic potential: whereas re duction of SRBC-rosette-forming cells from 62 to 20% made no significant difference, the reduction from 58 to 5% T cells was fol lowed by an increase in the antibody-dependent cell-mediated cytotoxic 51Cr re lease. After depletion of cells with receptors for mouse complement (table I) the remain ing cell preparations containing less than 1% C'3-receptor-bearing cells (but with high numbers of Fc-receptor-positive cells) exhibited increased antibody-dependent kill ing activity. However, these depleted sus pensions showed no marked difference in direct nonspecific lysis of target cells. In these latter experiments, the percentages of T cells varied only slightly from the original levels.
Discussion The percentages of rosette-forming peri pheral blood lymphocytes were found to dif fer only slightly in 5 experiments, but were in the normal range with mean levels of 33 for Fc rosettes, 61 for SRBC rosettes and 21 for C3 rosettes. The last was higher than previously reported [10]; the reason for this may be the use of mouse serum (in the place of human) as source of C'3. Using lympho cyte preparations after removing the phago cytic and plastic-adherent cells, the activi ties in the antibody-dependent cellular cyto toxicity model as shown by 5ICr release of target cells were 19-42% above the direct cytotoxicity levels after 14 h incubation
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Depletion Experiments Using rosette formation by mixing large quantities (5 ml) of lymphoid cells with the same volume of sensitized ox RBC and nor mal SRBC respectively, the non-rosette-for ming cells were separated by centrifugation on Ficoll/Hypaque and were concentrated at the interface between the layers of Medium 199/PBS and Ficoll/Hypaque. Re-rosetting of these cell preparations showed considera bly decreased numbers of Fc- or C'3receptor-bearing cells (table I) but gave evidence of insufficient T-cell depletion in 1 experiment (table I; experiment 1); in two further experiments, papain-treated SRBC were used to obtain stronger rosettes and after Ficoll-Hypaque centrifugation a reduction in the number of E-rosette-form ing cells was achieved to a level of 5% (ta ble I; experiment 4).
351
Kraft/Franks/Coombs
352
exclude the possibility that there are some cytotoxic effector cells which have C'3 re ceptors as well as Fc receptors [14, 15, 24], Removal of one type of cytotoxic cells could enrich another cell population with the same or higher cytotoxic activity. For example, this was reported by Papamichael and Temple [21] who separated glassadherent from non-adherent monocytes and found both preparations active in the antibody-dependent cellular cytotoxicity model. In general, there is evidence for the inde pendence of some functions of the Fc and C'3 receptors of human lymphocytes and for the heterogeneity of K cells (which are lymphoid cells with Fc receptors and killing potential in the antibody-dependent cellular cytotoxicity model) in different situations [1, 6, 13, 18-20, 23, 25]; but the possibility remains that one cell type could be more important than others in relation to in vivo situations like immunological tumour re gression and transplantation rejection.
References 1 Butterworth, A. E.; Sturrock, R. F., and Houba, V.: Antibody-dependent cell-mediated damage to schistosomula in vitro. Nature, Lond. 252: 503-505 (1974). 2 Duncan, P. B.: Multiple range and multiple F-tests. Biometrics 11: 1-12 (1955). 3 Eremin, A.; Ashby, J., and Franks, D.: Killer cell (K) activity in human normal lymph node, regional tumour lymph node and inflammato ry lymph node. II. Int. Archs Allergy appl. Immun (to be submitted). 4 Franks, D.; Kelly, M. M., and Sanderson, C. J.: Inhibition of xenogeneic cell-mediated cyto toxicity in the rat by antiglobulin sera. Int. Archs Allergy appl. Immun. 48: 621-631 (1975). 5 Franks, D.; Tayabali, B., and O’Brien, P.: Competitive inhibition of lymphoid cell me
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
time. A one-step separation procedure for depleting Fc-receptor- and C'3-receptorbearing and SRBC-rosetting lymphocytes was performed. The method is based on the fact that rosette-forming lymphocytes will sediment and the non-rosette-forming lym phocytes float on gradient centrifugation on Ficoll/Hypaque [22]. Modifications were applied using papain-treated SRBC for T-cell-specific rosettes [26] and with slowvelocity centrifugation to prevent disruption of the rosettes on the dense surface of the Ficoll/Hypaque layer. The efficiency of de pleting Fc- and C'3-receptor-bearing lym phocytes was >99% . T-cell depletion was less successful, but even the partially deplet ed samples gave sufficient information. The ADCC corresponding to each preparation of non-rosetting cells showed markedly de creased levels after depletion of Fc-rosetteforming lymphocytes which is in agreement with results of other workers using the same technique [11, 27] or columns coated with aggregated IgG or immune complexes [16, 27], After T-cell depletion, information was obtained that T cells have no influence on K-cell activity and the results correlated well with data of experiments using T-cellenriched [23] or T-cell-depleted fractions [24, 27]. However, after depletion of C'3rosette-forming lymphocytes, we found in creased levels of antibody-dependent cellu lar cytotoxicity which indicates lymphoid K cells with Fc receptor, but lacking C'3 re ceptor. In contrast to this, Perlmann et al. [23] found no killing activity of cell prepa rations depleted of C'3-rosette-forming lym phocytes. A possible explanation for this discrepancy may be that different target cells (Perlmann et al. used chicken red cells) are preferentially lyzed by different effector cells. On the other hand, our results do not
6
7
8
9
10
11
12
13
14
15
diated cytotoxicity analysed by double labell ing with 5,Cr-chromate and S8'nTc-pertechnetate (in preparation). Gale, R. P. and Zighelboim, J.: Polymorphon uclear leukocytes in antibody-dependent cellu lar cytotoxicity. J. Immun. 114: 1047-1051 (1975). Greenberg, A. H.; Shen, L., and Roitt, I. M.: Characterization of the antibody-dependent cy totoxic cell. A non-phagocytic monocyte? Clin, exp. Immunol. 15: 251-259 (1973). Greenberg, A. H.; Shen, L., and Medley, G.: Characteristics of the effector cells mediating cytotoxicity against antibody-coated target cells. I. Phagocytic and non-phagocytic effector cell activity against erythrocyte and tumour target cells in a 51Cr release cytotoxicity assay and [125I] IUdR growth inhibition assay. Immunolgy 29: 719-729 (1975). Greenberg, A. H.; Shen, L.; Walker, L.; Arnaiz-Villena, A., and Roitt, I. M.: Characteris tics of the effector cells mediating cytotoxicity against antibody-coated target cells. II. The mouse nonadherent K cell. Eur. J. Immunol. 5: 475-480 (1975). Haegert, D. G.; Hallberg, T., and Coombs, R. R. A.: B and T lymphocyte subpopulations in human peripheral blood. Int. Archs Allergy appl. Immun. 46: 525-538 (1974). Hallberg, T.: In vitro cytotoxicity of human lymphocytes: reduction of the lymphocyte cy totoxicity induced by antibodies or phytohem agglutination by removing immune-complex binding cells. Scand. J. Immunol. 3: 645-654 (1974). Harter, M. L.: Critical tables for Duncan’s new multiple range test. Biometrics 16: 671-685 (1960). Holm, G.: Lysis of antibody-treated human erythrocytes by human leukocytes and macro phages in tissue culture. Int. Archs Allergy appl. Immun. 43: 671-682 (1972). Holm, G.; Bjorkholm, M.; Mellstedt, H., and Johansson, B.: Cytotoxic activity of lympho cytes from patients with Hodgkin’s disease. Clin. exp. Immunol. 21: 376-383 (1975). Huber, H. and Holm, G.: Surface receptors of mononuclear phagocytes: effect of immune complexes on in vitro function in human mono cytes; in van Furth Mononuclear phagocytes in
353
16
17
18
19
20
21
22
23
24
25 26
immunity, infection and pathology, pp. 291301 (Blackwell, Oxford 1975). Isturiz, M. A. de; Bracco, M. M. de, and Manni, J. A.: Reaction of human lymphocytes with aggregated IgG columns. Effect on antibodydependent cytotoxicity and EAC rosette for mation. Cell. Immunol. 16: 82-91 (1975). Kovithavongs, T.; Rice, G.; Thong, K. L., and Dossetor, J. B.: Effector cell activity in anti body mediated cell dependent immune lysis. II. Evidence for different populations of effector cells for different targets. Cell. Immunol. 18: 167-175 (1975). MacDermott, R. P.; Chess, L., and Schlossman, St. F.: Immunologic functions of isolated human lymphocyte subpopulations. V. Isola tion and functional analysis of a surface Ig negative, E rosette negative subset. Clin. Im munol. Immunopathol. 4: 415-424 (1975). MacDonald, H. R.; Bonnard, G. D.; Sordat, B., and Zawodnik, S. A.: Antibody-dependent cell-mediated cytotoxicity: heterogeneity of effector cells in human peripheral blood. Scand. J. Immunol. 4: 487-497 (1975). MacLennan, I. C. M.: Comments on antibodydependent, cell (K) mediated cytotoxicity; in Advances in the bio-sciences. 12th ScheringSymposium on Immunopathology, pp. 87-91 (Pergamon Press, Vieweg 1974). Papamichael, M. and Temple, A.: Characteri zation of the human effector cell causing anti body-mediated cytotoxicity. Clin. exp. Immu nol. 20: 459-467 (1975). Parish, C. R.: Separation and functional analy sis of subpopulations of lymphocytes bearing complement and Fc receptors. Transplantn Rev. 2J: 98-120 (1975). Perlmann, P.; Perlmann, H.; Larsson, A., and Wahlin, B.: Antibody dependent cytotoxic effector lymphocytes (K cells) in human blood. J. reticuloendoth. Soc. 17: 241-250 (1975). Perlmann, P.; Perlmann, H., and MüllerEberhard, H. J.: Cytotoxic lymphocytic cells with complement receptors in human blood. Induction of cytolysis by IgG antibody but not by target cell-bound C3. J. exp. Med. 141: 287-2% (1975). Roitt, I. M.; Essential immunology, p. 136 (Blackwell, Oxford 1974). Wilson, A. B.; Haegert, D. G., and Coombs,
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
Characterization of K-Cell Activity
R. R. A.: Increased sensitivity of the rosette forming reaction of human T lymphocytes with sheep erythrocytes afforded by papain treatment of the sheep cells. Clin. exp. Immu nol. 22: 177-182 (1975). 27 Wislpff, F.; Frpland, S. S., and Michaelsen, T. E.: Antibody-dependent cytotoxicity mediated by human Fc-receptor-bearing cells lacking
Kraft/Franks/Coombs
markers for B- and T-lymphocytes. Int. Archs Allergy appl. Immun. 47: 139-154 (1974).
Correspondence to: Dr. D. Kraft, Institute of General and Experimental Pathology, University of Vienna, Wahringerstrasse 13, A-1090 Vienna (Austria)
Downloaded by: Univ. of California Santa Barbara 128.111.121.42 - 3/5/2018 6:28:14 PM
354
