R. Kiessling, E. KIein and H. Wigzell
112
R. Kiessling’,
Eva Klein’ and H. Wigzell”
Department of Tumor Bioloqy, Karolinska InAin~te.Stockholm’ and Department of Immunology, Uppsaia University, UppsaIaO
Eur. J. Immunol. 1975.5: 112-117
“Natural” killer cells in the mouse 1. CytOtOXiC Cells with specificity for mouse Moloney leukemia cells. Specificity and distribution according to genotype* In the spleens of young, adult mice there exist naturally occurring killer lymphocytes with specificity for mouse Moloney leukemia cells. The lytic activity was directed against syngeneic or allogeneic Moloney leukemia cells t o a similar extent, but was primarily expressed when tested against in vitro grown leukemia cells. T w o leukemias of non-Moloney origin were resistant and so was the mastocytoma line P8 15. Although killer activity varied between different strains of mice, the specificity of lysis was the same as indicated by competition experiments using unlabeled Moloney or other tumor cells as inhibitors in the cytotoxic assays. Capacity t o compete and sensitivy t o lysis by the killer cells were found t o be highly positively correlated. Analysis of the kinetics of the cytotoxic assay revealed a rapid induction of lysis within one to four hours, arguing against any conventional in vitro induction of immune response. N o evidence was found of soluble factors playing any role in the cytolytic assay.
1. Introduction Cytolytic activity of lymphoid cells from normal, nonimmunized donors has been reported in several in vitro systems both in animal [ 1 -2]* and human [ 3 , 4 ] systems. Some systems would seem t o involve specific recognition in the sense that only certain target cells will be lysed by the effector cells [ 1-41, whereas in other assays n o discrimination in killer activity can be demonstrated**. In a few systems there exist indications that the capacity of normal lymphoid cells t o function as specific killer cells is induced by antibodies in these S.C. normal individuals, thereby recruiting effector cells in antibody-induced, cell-mediated lysis. This would occur either directly by coating the target cells when the test is carried out in the presence of donor serum, or indirectly via arming the effector cells through in vivo coating with antigenantibody complexes [5]. In the majority of tests, however, the exact cause of “normal” cytolytic activity would seem obscure and different effector cells would seem t o be involved in the various tests [ 1, 21. The biological relevance of these normally occurring killer cells is hypothetical and has been assumed to be of importance in autoimmune reactions as well as in the reaction against persistant virus infections [2]. Irrespective of hypothetical biological importance the occurrance of such spontaneous killer cells has serious impacts in experimental assays, especially in tumor immune research, when very weak “immune” reactions are t o be measured that can be easily overshadowed by the existance of varying spontaneous killer activity in the control material. In the present article we report on the existence in normal mice of killer cells with rapid cytolytic, specific activity against in vitro grown mouse Moloney leukemia cells. We
* Gomard, E., Leclerc, J.C. and Levy, J.P., submitted for publication.
** Grant, C.K. and Alexander, P., submitted for publication. [I 8751
* Supported by the Swedish Cancer Society and by NIH contracts NOI-CB-33870 and N01-CB-33859.
Correspondence: Rolf Kiessling, Department of Tumor Biology, Karolinska Institute, S-104 01 Stockholm 60, Sweden Abbreviations: MLV: Moloney leukemia virus
have analyzed the killer activity in relation to genotype of killer or target cells, and report on selective killer deficiencies linked to certain strains of mice. Competition experiments using unlabeled cells as competing targets for killer activity together with labeled targets demonstrate a highly positive correlation between sensitivity t o cytolysis and competing capacity. As will be seen from an accompanying article, the killer cells involved would seem to be neither T nor B lymphocytes as classified by conventional markers but rather belong t o yet a third group of lymphocytes [ 6 ] ,maybe of the kind described by other workers on the basis of morphology [7].
2. Materials and methods 2.1. Tissue culture medium
Medium F 13 (Grand Island Biological Co., Grand Island N.Y.) containing 10 % heat-inactivated fetal calf serum (BioCult, Glasgow, Scotland) with penicillin (75 units/ml) and streptomycin (50 pg/ml) were used throughout the experiments as diluents and tissue culture media.
2.2. Animals Animals of our own inbred strains of mice were used throughout most of the experiments. Sex of effector cell donors was not found to be important in this assay and is not documented. Age of cell donors is indicated in each experiment if not between 1 to 2 months of age. In experiments involving genetically athymic mice, such mice were bought from Dr. C.W. Friis (GI. Bomholtgaard Ltd., Ry, Denmark), and were backcrossed to BALB/c mice on the 6th backcross generation when used. The breeding scheme is described in detail elsewhere [ 81. Mice of the same age and sex, homozygous or heterozygous for the nude gene‘, were used as indicated.
2.3. Tumors Eight different tumor cell lines were used. Five of these were lymphomas induced by inoculation of the Moloney leukemia - virus (MLV) into newborn recipients. The various cell lines are symbolized as follows: the first letter ‘Y’stands for Moloney lymphomas. The second letter in 3 letter symbols and the second and third letter in 4 letter symbols indicate
Eur. J. Immunol. 1975.5: 112-117
Specificity of natural killer cells against Moloney leukemia cells
the genotype of origin. The symbols correspond t o the following strains: ‘A’ = strain A, ‘CA’ = (A x ACA)F,, ‘B’ = = CBA, ‘L‘= Leaden. The final letter distinguishes between different MLV tumors of the same genotype. The final numbers denote different sublines of the same tumor. Three non-MLV tumors were also used. P815-X2, a methylcholanthrene-induced mastocytoma of DBA origin, EL-4, a chemically induced lymphoma of C57BL origin and LAA, a spontaneous lymphoma of A/Sn origin. All tumors were propagated both in vivo in syngeneic mice and in vitro as established cell lines.
113
2.7. Preparation of effector cell suspension Spleen cell suspensions were prepared in complete F-13 medium. Erythrocytes were lysed by treatment for 10 min at 4 “C with Tris buffered 0.75 % ammonium chloride, pH 7.2. If not otherwise stated, this treatment was done with all spleen cell suspensions.
3. Results 3.1. Strain distribution of normally occurring anti-Moloney leukemia killer cell activity
2.4. Target cell labeling
Tumor cells were incubated for 30 min at 37 OC at a concentration of 5 x lo6 - 1 0 x l o 6 cells suspended in 0.5 ml of complete medium t o which 200 pCi of Cr as sodium chromate was added. The cells were washed twice, counted and adjusted t o the desired cell density.
Spleen cells from normal, adult, nonimmunized mice have a certain cytolytic activity against the in vitro line of the A/Sn Moloney leukemia YAC. When analyzing the strain distribution of this “background” cytolytic activity it was found that cells from A/Sn mice were least efficient. Cells of certain other mouse strains were consistently highly cytotoxic as shown in Table 1.
2.5. Cytotoxic test
0.1 ml aliquots of various effector cell concentrations were added t o glass tubes (1 1 x 55 mm) in duplicates. A constant number of 4 x lo4 Cr-labeled target cells in 0.05 ml were added. Target cells were also incubated without lymphocytes in medium only t o estimate the level of spontaneous release of ”Cr. This varied between 20-30 % of total label under the experimental conditions described. The tubes were incubated in a 3 7 OC, 5 % C 0 2 incubator o n a rocking platform for 12- 14 h. Then 1 ml of medium was added t o each tube and the tubes were centrifugated. For each tube total radioactivity and radioactivity of half of the supernatant was measured in a gamma counter. Maximum release was determined by treating the target cells with distilled water overnight, which released 70-80 % of the total ”Cr.
Table 1. Spontaneous cytolytic activity in spleen cells from normal adult mice of different strains tested against YAC in uifro linea)
’’
The % ”Cr release was calculated according to t h e following formula: 2 x (radioactivity of 1/2 of the supernatant) % ” ~ rrelease = x 100 total radioactivity Results are expressed as percent lysis according t o the following formula:
% lysis =
% release in test
-
7% spontaneous release
x 100
% maximum release - % spontaneous release
The standard deviation between the two duplicates never exceeded 5 % lysis.
*
No. of
% lysis at 30/1 ratio
Strain
Exp.
Mean
Range
CBA A/Sn
12 9 2 1 1 2 18 4 3 1 2
63 16 30 31 38 48 37 23 24 29 30
34- 95 5-33 23-36
C57BL C3H ACA BALB/c A x C57BL A x Leaden A x ACA A x C3H A x CBA
46-50
9-70 9-44 5-48 16-44
a ) In each experiment 2-5 spleens are pooled and tested for cytotoxicity. For experimental details see Section 2.
As YAC is of A/Sn origin, we assumed that the low effect of normal A/sn cells might be due t o the lack of isoantigenic difference between effector and target cells. However, low spontaneous activity was also observed when the target cell YBA, another MLV leukemia of CBA origin, was tested against the non-H-2 compatible effector cells from A/Sn animals. Furthermore, effector cells of the syngeneic strain (CBA) were more cytolytic than A/Sn cells as could be seen from the experiment in Table 2. Table 2. Cytolytic activity of A/Sn and CBA spleen cells tested against the Y BA in vifro linea)
2.6. Competition studies Effector cell
1 x 1 O6 effector cells were incubated with various concentrations of “competing” tumor cells in duplicates for 1 h at 37 OC at a total volume of 0.2 ml complete medium. 2 x l o 4
’*Cr-labeled target cells were added u p t o a final volume of 1 ml. Thus a constant effector t o target cell ratio of 50: I was used in all experiments, while the ratio of competing tumor cells t o target cells was titrated out. The rest of the experiment was carried out as described above. All results are expressed as % inhibition at a certain competitor t o target cell ratio according t o the following formula: % lysis without competing cells % lysis with competing cells x 100. %inhibition = % lysis without competing cells
A/Sn % lysis 5011
Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 Mean % lysis
CBA
lo lysis 501 1
0
Mouse I Mouse 2 Mouse 3 Mouse4 Mouse 5
10 29 28
1
Mean Q lysis
18
2 2 3 -2
17
7
a) Spleens from five A/Sn and five CBA mice of the same age and sex were tested individually in the same experiment against the YBA in uitro line.
R. Kiessling, E. Klein and H. Wigzell
114
It would thus seem clear that for yet unknown reasons A/Sn mice have a comparatively low “background” activity of killer cells against these two MLV lymphomas.
3.2. Time kinetic studies on the cytolytic effect o n YAC-1 cells b y normal spleen cells In spite of the fact that the effector cells in our system were taken from normal nonimmunized mice, the onset of the in vitro lytic activity was remarkably fast. As could be seen in the time course experiment in Fig. I , a clear effect was obtained by the CBA spleen cells on the YAC target cells already after 4 h of incubation, with an effector-to-target cell ratio of SO/]. Thus, it seemed that a rapid cell lysis independent of known preimmunization or added antibodies was active in this system.
Eur. 1. Immunol. 1975.5: 112-117 Table 3. Specificity of spontaneous cytolytic activity for Moloney leukemia cellsa)
Exp.
H o u r s of
63 -4
I1
CBA CBA CBA A x BL A x BL AxBL
YAC-1 YCAB-1 P815 YAC-1 YCAB-1 P815
64 8 -2 38
A x BL A x BL AxBL A x BL A x BL A/Sn A/Sn A/Sn A/Sn A/Sn
YAC-1 YAC-301 YCAB-1 YCAB-6 P815 Y AC- 1 Y AC-30 1 YCAB-1 YCAB-6 P815
47
CBA CBA CBA CBA CBA
YAC-1 YCAB-1 YBA YAA EL-4
69
CBA CBA CBA CBA CBA CBA CBA CBA
YAC-1 YBA YCAB-1 YLI YAA LAA P815 E L-4
91 26 42
111
12
lncubotion
Figure 1. Effect of CBA spleen cells on YAC in vivo ( 0 - 0 ) or in vitro (0-0) cell lines as a function of incubation time. Spleens from 4 CBA mice were pooled. Spontaneous release after 12 h of incubation is for the in vifro line 13 70 and for the in vivo line 22 70 of thc total label.
’ib Lysis
(ratio 5011)
YAC-1 YCAB-1 P815
V
8
cells
CBA CBA CBA
1V
4
Target
I
1
1
Effector cells
-2
2 0
25 5 11 0 3 4
-2 2 -11
8 17
0 10
7
6 10
-2 10
a ) Comparison between different in vitro tumor cell lines for cytolytic sensitivity. Five independent experiments are shown. In each experiment the effector cells were derived from 2-5 spleens from CBA, A/Sn or A x C57BL F1-hybrids. The spontaneous release is within the same range for all cell lines compared (20-30 %of total
label). 3.3. Origin of mice In order to rule out the possibility that the activity of the normal spleen cells might depend o n “spread” of leukemia virus in our mouse colony, we also investigated mice from other sources where n o work on mouse leukemia viruses is carried out. We thus tested CBA mice from the Department of Genetics, Stockholm University and also BALB/c mice from a commercial breader in Denmark (Friis, see Section 2.2.). We found, however, that spleen cells taken from mice on the day of arrival t o the department were as efficient as our own mice of the same strain and age. 3.4. Comparative studies between different tumor cell lines
for their sensitivity for lysis by spleen cells from normal animals In order t o establish the nature of this reaction, more tumor cell lines were tested for sensitivity in the present system. Among the 8 cell lines tested significant effects (> 10 % lysis) have only been seen with three MLV cell lines. Two non-MLV lymphomas as well as two induced by MLV were not sensitive (Table 3). It is of interest that the P815 line, known t o be highly sensitive t o lymphocyte lysis, was not lysed.
3.5. Competition studies
The fact that a target cell is found t o be insensitive t o lysis does not exclude that the relevant target structures are present on the surface. Some target cells might be intrinsically more resistant t o lysis than others, and therefore, competition studies were carried out as a more direct way of defining presence or absence of relevant target structures in different cells [ 9 ] . These studies were designed so that different numbers of non-labeled tumor cells were added to a constant number of 5 1 Cr-labeled target cells and effector cells. Thus, a tumor cell with the relevant target structure on its surface may compete for the effector cells, resulting in reduction of lysis of the 5 1 Cr-labeled target cells. Tumor cells with irrelevant surface components would not compete. The results of a typical competition experiment are shown in Fig. 2. Here, as in all experiments, five different ratios of competitorto-target cell were used. In this experiment, among the 6 lines tested only the YAC and the YAA in vitro lines ;howed competition. We express the competition results as % inhibition (formula see Section 2.6.) at a competitorto-target cell ratio of 7/1.
Eur. J. Immunol. 1975.5: 112-117
Specificity of natural killer cells against Moloney leukemia cells
115
Table 5. Detailed analysis of the capacity of various in vitro tumoi Lines to compete in cytolysis against YAP) Competing tumor cell
YLl
93 70 70 87
YCAB-1
33
YAC YAA YBA
EL-4 P815 LAA .Oi
Exp. 1
37 22 n.d.
% inhibitionb) 11 111
59 54 n.d. n.d. -5 3 -8 n.d.
113 125 138 106 11
19 n.d.C) n.d.
1v 108 41 69 39 0
0 4 8
i
00:I
I:(
2:l
8:l
4:l
C o m p e t i t o r : torgel
a ) As effector cells CBA spleen cells at an effector-to-target cell ratio of 50/1 were used. As target cells the YAC-1 in vitro line were used. b) Competing capacity for each tumor cell line is expressed as % inhibition at a competitor-to-target cell ratio of 7/1. c) n.d. = Not done.
16.1
cell r a t i o
Figure 2. Competition experiment. As effector cells CBA spleen cells at an effector-to-target cell ratio of SO/l were used. As target cells the YAC-1 in vitro line were used. Competing cells are as follows: YAC-I in vitro line (*--.), YAC in vivo line (0-o), YAA in vitro line (x-x), YCAB-1 in vitro line (a--a), EL-4 in vitro line (A-A) and P815 in vitro line (0-0).
cluded that whereas none of the non-MLV-induced cell lines compete strongly, 3 out of 4 MLV-induced cell lines are efficient competitors.
3.8. Differences between in vitro and in vivo grown cell lines
3.6. Comparative studies between sensitivity for lysis and competing capacity Table 4 shows an experiment similar t o that described above in which it can be seen that a positive correlation between sensitivity for lysis and capacity t o compete exists. The competition experiments seemed t o be more sensitive in demonstrating the presence of relevant target structures than direct cytolytic tests o n the same cells (see YLI which was significant in competition but insensitive to lysis). As seen in Table 4, both sensitivity to lysis and capacity t o compete seems to be confined t o tumors induced by MLV (exception YCAB). Table 4. Positive correlation between cytolytic sensitivity of target cells and capacity to compete in the cytotoxic assaya)
In vitm line Y AC- 1 YAA YBA YLI YCABl
% lysis (CBA)
46 10
LAA
9 0 1 3
EL-4
5
P815
2
% inhibition
108 47 69 39 0
During these studies we observed that the in vivo grown tumors were considerably less sensitive t o the lytic effect of normal spleen cells than their corresponding in vitro lines. This difference can be clearly seen in Fig. 1, where both the in vitro and the in vivo grown YAC cells were used as target cells. The same difference was also seen when comparing in vivo and in vitro lines of the YBA target cell. As already mentioned, competition studies may be the most sensitive assay t o detect the relevant target structures in the present system. We therefore performed such studies t o compare in vivo with in vitro grown cells. In agreement with the results on cytolytic sensitivity, in vivo grown tumor cells did not compete, or competed t o a lesser extent than did their corresponding in vitro lines, as can be seen from the 3 experiments summarized in Table 6. It thus seems clear that the structure or antigen which the killer cells recognize is less well expressed on the in vivo grown MLV tumor cells.
4. Discussion
8
0 4
a) Experiment comparing the cytolytic sensitivity with competing capacity in different in vitro grown tumor cell Lines. As effector cells in both systems CBA spleen cells at an effector to target
cell ratio of 50/1 were used. As target cells in the competing system the YAC-1 in vitro cell line was used. Competing capacity for each tumor cell line is expressed as % inhibition at a competitor-to-target cell ratio of 7/1.
3.7. Differences in competing capacity between MLV- and non-MLV-induced tumor cells These results prompted us to carry out further competition experiments in order t o see if there was a consistent difference between MLV- and non-MLV-induced tumors. In Table 5 the results of four experiments are shown. It can be con-
In the present article we describe the existance of “naturally” occurring killer cells in mice with selective lytic activity against Moloney leukemia cells. Of the tumor lines tested, only Moloney leukemia cell lines were found t o be sensitive t o the cytotoxic activity of the killer cells. Between the M e loney leukemia lines there existed significant variations in susceptibility t o lysis. Also, in vitro grown cells were found to be more sensitive t o lysis than in vivo grown cells from the same Moloney leukemia cell line. In virro Moloney leukemia cells express more Moloney leukemia-associated antigens than the corresponding cells grown in vivo [ 101 and this might be relevant t o our findings. Susceptibility t o lysis could be shown t o be positively correlated t o the capacity of cells of the same tumor line t o function as competitive inhibitors when added t o the cytolytic test (see Table 4). Any susceptible Moloney leukemia cell line could be used as a selective competitor for lysis of isotope labeled
116
R. Kiessling, E. Klein, H. Pross and H. Wigzell
Eur. J. Immunol. 1975. 5: 112-117
Table 6. Comparison in competing capacity in the cytotoxic assays against YAC using in vivo or in vifro grown Moloney leukemia cell linesa) % inhibitionb)
Competing tumor cell YAC YAA Y BA Y LI
Exp. I in vitm
93 70 80 87
I1
in vivo 27 37 33 43
in vitro
113 125
n.d. 106
in vivo 43 75
n.d. 38
111 in v i m in vivo
1V in vitm
59
-3
87
n.d.c) n.d. n.d.
n.d. n.d. n.d.
n.d. n.d. n.d.
in vivo 13 n.d.
n.d. n.d.
a) As effector cells CBA spleen cells at an effector-to-target cell ratio of 50/1 were used. As target cells the YAC-1 in vifro line were used. b) Competing capacity for each tumor cell line is expressed as % inhibition at a competitor to target cell ratio of 7/1. c) n.d. = Not done.
target cells from another Moloney leukemia line demonstrating that the killer cells, irrespective of strain of origin, had the same specificity for the different Moloney leukemia lines. However, that each tumor line had its own level of susceptibility and each mouse strain had its own level of killer activity as well will be discussed further below.
No evidence for any influence linked t o allogeneic inhibition phenomena [ 1 I ] could be demonstrated in the present system, as the relative killer activity was the same whether tested against syngeneic or allogeneic Moloney leukemia cells. We could thus conclude that the killer cells from these normal mice behave as though specifically immunized against Moloney lymphoma-associated antigens. The in vifro cytolytic activity of the killer cells is very rapid and can be detected within less than 4 h (see Fig. I ) as detected by ”Gr release from the damaged target cells. In fact, when using purified killer cells as obtained by fractionation procedures [6], specific lysis can be found within 1 h after start of incubation. We would thus consider the time periods so short that conventional in vifro induction of immune killer cells is excluded and the present test actually demonstrates the existence of in vivo generated killer cells with cytotoxic potential for Moloney leukemia cells. As the nature of the killer cell is the topic of an accompanying article [6], let us only state that the effector cell has the morphological appearance of a small lymphocyte. However, it can be clearly separated from both B and T lymphocytes by organ distribution data, as well as by surface markers used in fractionation procedures. The relationship t o the lymphocytes described by Stobo and coworkers [7] remains t o be clarified. The killing capacity is seemingly mediated by contact, as all attempts t o demonstrate lytic activity of supernatants from in vitro cultivated killer cells have failed (unpublished observations). Spleen cells contain the highest killer activity of the lymphoid organs [6], and killer activity peaks around the first 3 months of the life of a mouse. A variation in t h e efficiency of killer cells present in the spleen of normal mice of different, inbred strains was observed with A/Sn mice containing the lowest killer activity (see Table I ) . Of t h e other strains tested, CBA mice seemed t o have t h e highest killer activity in their spleens. What is then the cause and origin of the present “spontaneous” killer cells? As already stated, the present killer cells would seem quite distinct from conventional, immune effector cells against Moloney leukemia cells that dominate
the in v i m cytotoxic reactions when immune anti-Moloney effector cells are used [ 121. This in itself makes a discussion as to the cause of the occurrence of the present killer cells difficult, as the present effector cells would seem to be lymphocytes of a new type with yet unknown biological properties. If we assume that the activity of the killer cells is of immunological nature, a simple explanation would be that they are directed against murine leukemia virus antigens, and as such viruses are known to be quite ubiquitous, this would seem a reasonable assumption. Alternatively, the killer activity might not be Moloney leukemia-specific, but rather of an autoimmune type directed against some organ-associated antigen being expressed on the Moloney leukemia cells but not being directly coded for by the Moloney leukemia virus. Such a n argument is partly weakened by the fact that both susceptible Moloney lymphomas and resistant lymphomas such as EL-4 can be shown to carry surface markers such as the theta antigen denoting their common, thymic origin, but only the Moloney leukemia cells are susceptible to lysis. Other workers have reported o n the existence of “spontaneous” killer cell activity against tumor cells in the mouse or rat systems, with two systems detecting anti-lymphoma activity [ 1 ,*I. In one system, preleukemic AKR mice were found to express a n anti-lymphoma specific activity with the effector cells being of probable macrophage type*, whereas in the other test spontaneous cytolytic activity was found in mice against a Rauscher virus-induced leukemia with the effector cells being of lymphocyte nature [ I ] . Interestingly, in the latter assay the effector cells, although of lymphocyte nature, were found to be resistant t o anti-theta serum plus complement treatment in analogy with results in the present system [6]. Received August 5 , 1974.
* Gomard, E., Leclerc, J.C. and Levy, J.P., submitted for publication.
5. References 1 Herberman, R., Nunn, M.E., Lavrin, D.H. and Asofsky, R., J. N u t
Gmcer Insf. 1973.51: 1509. 2 Greenberg, A.H. and Playfair, J.H.L., Clin. Exp. Immunol. 1974. 1 6 : 99.
3 Tagasuki, M., Mickey, M.R. and Terasaki, P., Cuncer Res. 1973. 33: 2898.
Eur. 1. Immunol. 1975.5: 117-121
A new cell type involved as natural killer cells against Moloney leukemia cells
4 Petranyi, G.G., Banczur, M., Onody, C. a n d Holland, S.R.,Lancet 1974. i : 736. 5 Perlmann, P., Perlmann, H. and Biberfeld, P., J. Irnrnunol. 1972. 108: 558. 6 Kiessling, R., Klein, E., Pross, H. and Wigzell, H., Eur. J. Irnrnunol. 1975. 5: 117. 7 Stobo, J.D., Rosenthal, AS. and Paul, W.E., J. Exp. Med. 1973. 138: 71. 8 Rygaard, J., Acta Pathol. Microbiol. Scand. 1969. 7 7 761.
R. Kiessling+, Eva Klein+, H. Pros+ and H. Wigzell+'
Department of Tumor Biology, Karolinska Institute, Stockholm+ and Department of Immunology, Uppsala University, Uppsala'
117
9 Rosenberg, E., McCoy, J., Green, F., Donelly, F., Siwarski, D., Lewine, P. and Herberman, R., J. Nut. Cancer Inst. 1974.52: 345. 10 Cikes, M., Friberg, S., Jr. and Klein, G., J. Naf. Cancer Inst. 1973 . 50: 347. 11 Hellstrom, K.E. and Hellstrom, I., Progr. Exp. Tumor Res. 1967. 9 : 40. 12 Lamon, E.W.,Wigzell, H., Klein, E., Andersson, B. and Skurzak, H., J. Exp. Med. 1973. 137: 1472.
"Natural" killer cells in the mouse II. Cytotoxic cells with specificity for mouse Moloney leukemia Cells. Characteristics O f the killer Cell" Normal mice contain cytolytic cells with specificity for in v i t r o grown mouse Moloney leukemia cells. Such killer cells are most frequent in the spleens; lymph node and bone marrow contain less and thymus virtually n o killer activity. Peak activity is found around one t o three months of age. Spleen cells from genetically athymic mice are as active killer cells as those from normal mice of the same strain. Treatment with anti-theta serum plus complement followed by removal of adherent and surface Ig positive cells by filtration through anti-lg columns will leave between 1-5 % of the original spleen cell population from a normal mouse. These cells have the morphology of small lymphocytes and perhaps contain all of the total original killer activity of the spleen against the Moloney leukemia cells. Such killer enriched cells are devoid of T and B lymphocytesand largely fail t o function in antibody induced, cell-mediated lysis against antibody-coated chicken erythrocytes. It is concluded that the spontaneous selective cytotoxic activity of normal mouse spleen cells against Moloney leukemia cells is exerted by small lymphocytes of yet undefined nature.
1. Introduction
Specific cytolytic or growth inhibitory activity has been ascribed t o various effector cells derived from the reticuloendothelial cell system. Thymus-derived lymphocytes have been shown t o function as highly efficient killer cells [ I ] , which recognize the target via its actively produced, antigenspecific surface receptors [ 21. Macrophages can be rendered selectively cytotoxic via a specific T cell factor [ 3 ] . Monocytesand lymphocytesof yet undefined type have been found capable of cytolitic aggression towards IgC-coated target cells [4]. Immune B lymphocytes were shown t o be necessary in certain growth inhibitory assays in vitro [S]. A variety of cells have thus been shown t o be effective in various in vitro cytolytic or growth inhibitory test systems. We would like t o report another new cell type having selective in vitro cytotoxic behavior with unique functional as well as surface characteristics. This cell is found in normal mice, yet it displays specific cytolytic activity against certain mouse Moloney leukemia lines in v i t r o . [I 8761
* Supported by the Swedish Cancer Society, and by NIH contracts NOI-CB-33870 and NOI-CB-33859.H.P. was supported by a fellowship from the Canadian Medical Research Council. Correspondence: Rolf Kiessling, Department of Tumor Biology, Karolinska Institute, S-104 01 Stockholm 60, Sweden Abbreviations: MLV: Moloney leukemia virus PBS: Phosphate buffered saline CRBC: Chicken red blood cells FCS: Fetal calf serum
It will cause rapid lysis of such leukemic cells in the absence of added antibodies [ 6 ] . As judged by surface markers and other characteristics it cannot be classified as a lymphocyte of B o r T type nor as a monocyte. The distribution pattern with regard t o age and organs will be reported, as well as its lack of known surface markers.
2. Materials and method 2.1. Materials and methods described in the previous paper [6]
These refer to animals and tumors used, to tissue culture medium, target cell labeling, cytotoxic test and preparation of effector cell suspension.
2.2. Assay of antibody-dependent cell-mediated cytotoxicity Fresh chicken red cells, washed 3 times in PBS, pH 7.4 were incubated at 37 OC with 150 pCi "Cr per lo' CRBC in 0.35 ml PBS for 3-4 h. Thereafter they were washed three times in culture medium and adjusted t o a concentration of 105 cells/ml. The cytotoxic test was carried out using Ellermann plastic tubes ( 1 1 mm x 55 mm). Each tube contained 5 x lo4 "Crlabeled CRBC, normal rabbit sera of rabbit anti-chicken sera at a 104-fold dilution and varying amounts of spleen cells in a total of one ml of culture medium (F-13 plus 10 % heatinactivated FCS). The tubes were centrifuged at 2000 x g for
112
R. Kiessling’,
Eva Klein’ and H. Wigzell”
Department of Tumor Bioloqy, Karolinska InAin~te.Stockholm’ and Department of Immunology, Uppsaia University, UppsaIaO
Eur. J. Immunol. 1975.5: 112-117
“Natural” killer cells in the mouse 1. CytOtOXiC Cells with specificity for mouse Moloney leukemia cells. Specificity and distribution according to genotype* In the spleens of young, adult mice there exist naturally occurring killer lymphocytes with specificity for mouse Moloney leukemia cells. The lytic activity was directed against syngeneic or allogeneic Moloney leukemia cells t o a similar extent, but was primarily expressed when tested against in vitro grown leukemia cells. T w o leukemias of non-Moloney origin were resistant and so was the mastocytoma line P8 15. Although killer activity varied between different strains of mice, the specificity of lysis was the same as indicated by competition experiments using unlabeled Moloney or other tumor cells as inhibitors in the cytotoxic assays. Capacity t o compete and sensitivy t o lysis by the killer cells were found t o be highly positively correlated. Analysis of the kinetics of the cytotoxic assay revealed a rapid induction of lysis within one to four hours, arguing against any conventional in vitro induction of immune response. N o evidence was found of soluble factors playing any role in the cytolytic assay.
1. Introduction Cytolytic activity of lymphoid cells from normal, nonimmunized donors has been reported in several in vitro systems both in animal [ 1 -2]* and human [ 3 , 4 ] systems. Some systems would seem t o involve specific recognition in the sense that only certain target cells will be lysed by the effector cells [ 1-41, whereas in other assays n o discrimination in killer activity can be demonstrated**. In a few systems there exist indications that the capacity of normal lymphoid cells t o function as specific killer cells is induced by antibodies in these S.C. normal individuals, thereby recruiting effector cells in antibody-induced, cell-mediated lysis. This would occur either directly by coating the target cells when the test is carried out in the presence of donor serum, or indirectly via arming the effector cells through in vivo coating with antigenantibody complexes [5]. In the majority of tests, however, the exact cause of “normal” cytolytic activity would seem obscure and different effector cells would seem t o be involved in the various tests [ 1, 21. The biological relevance of these normally occurring killer cells is hypothetical and has been assumed to be of importance in autoimmune reactions as well as in the reaction against persistant virus infections [2]. Irrespective of hypothetical biological importance the occurrance of such spontaneous killer cells has serious impacts in experimental assays, especially in tumor immune research, when very weak “immune” reactions are t o be measured that can be easily overshadowed by the existance of varying spontaneous killer activity in the control material. In the present article we report on the existence in normal mice of killer cells with rapid cytolytic, specific activity against in vitro grown mouse Moloney leukemia cells. We
* Gomard, E., Leclerc, J.C. and Levy, J.P., submitted for publication.
** Grant, C.K. and Alexander, P., submitted for publication. [I 8751
* Supported by the Swedish Cancer Society and by NIH contracts NOI-CB-33870 and N01-CB-33859.
Correspondence: Rolf Kiessling, Department of Tumor Biology, Karolinska Institute, S-104 01 Stockholm 60, Sweden Abbreviations: MLV: Moloney leukemia virus
have analyzed the killer activity in relation to genotype of killer or target cells, and report on selective killer deficiencies linked to certain strains of mice. Competition experiments using unlabeled cells as competing targets for killer activity together with labeled targets demonstrate a highly positive correlation between sensitivity t o cytolysis and competing capacity. As will be seen from an accompanying article, the killer cells involved would seem to be neither T nor B lymphocytes as classified by conventional markers but rather belong t o yet a third group of lymphocytes [ 6 ] ,maybe of the kind described by other workers on the basis of morphology [7].
2. Materials and methods 2.1. Tissue culture medium
Medium F 13 (Grand Island Biological Co., Grand Island N.Y.) containing 10 % heat-inactivated fetal calf serum (BioCult, Glasgow, Scotland) with penicillin (75 units/ml) and streptomycin (50 pg/ml) were used throughout the experiments as diluents and tissue culture media.
2.2. Animals Animals of our own inbred strains of mice were used throughout most of the experiments. Sex of effector cell donors was not found to be important in this assay and is not documented. Age of cell donors is indicated in each experiment if not between 1 to 2 months of age. In experiments involving genetically athymic mice, such mice were bought from Dr. C.W. Friis (GI. Bomholtgaard Ltd., Ry, Denmark), and were backcrossed to BALB/c mice on the 6th backcross generation when used. The breeding scheme is described in detail elsewhere [ 81. Mice of the same age and sex, homozygous or heterozygous for the nude gene‘, were used as indicated.
2.3. Tumors Eight different tumor cell lines were used. Five of these were lymphomas induced by inoculation of the Moloney leukemia - virus (MLV) into newborn recipients. The various cell lines are symbolized as follows: the first letter ‘Y’stands for Moloney lymphomas. The second letter in 3 letter symbols and the second and third letter in 4 letter symbols indicate
Eur. J. Immunol. 1975.5: 112-117
Specificity of natural killer cells against Moloney leukemia cells
the genotype of origin. The symbols correspond t o the following strains: ‘A’ = strain A, ‘CA’ = (A x ACA)F,, ‘B’ = = CBA, ‘L‘= Leaden. The final letter distinguishes between different MLV tumors of the same genotype. The final numbers denote different sublines of the same tumor. Three non-MLV tumors were also used. P815-X2, a methylcholanthrene-induced mastocytoma of DBA origin, EL-4, a chemically induced lymphoma of C57BL origin and LAA, a spontaneous lymphoma of A/Sn origin. All tumors were propagated both in vivo in syngeneic mice and in vitro as established cell lines.
113
2.7. Preparation of effector cell suspension Spleen cell suspensions were prepared in complete F-13 medium. Erythrocytes were lysed by treatment for 10 min at 4 “C with Tris buffered 0.75 % ammonium chloride, pH 7.2. If not otherwise stated, this treatment was done with all spleen cell suspensions.
3. Results 3.1. Strain distribution of normally occurring anti-Moloney leukemia killer cell activity
2.4. Target cell labeling
Tumor cells were incubated for 30 min at 37 OC at a concentration of 5 x lo6 - 1 0 x l o 6 cells suspended in 0.5 ml of complete medium t o which 200 pCi of Cr as sodium chromate was added. The cells were washed twice, counted and adjusted t o the desired cell density.
Spleen cells from normal, adult, nonimmunized mice have a certain cytolytic activity against the in vitro line of the A/Sn Moloney leukemia YAC. When analyzing the strain distribution of this “background” cytolytic activity it was found that cells from A/Sn mice were least efficient. Cells of certain other mouse strains were consistently highly cytotoxic as shown in Table 1.
2.5. Cytotoxic test
0.1 ml aliquots of various effector cell concentrations were added t o glass tubes (1 1 x 55 mm) in duplicates. A constant number of 4 x lo4 Cr-labeled target cells in 0.05 ml were added. Target cells were also incubated without lymphocytes in medium only t o estimate the level of spontaneous release of ”Cr. This varied between 20-30 % of total label under the experimental conditions described. The tubes were incubated in a 3 7 OC, 5 % C 0 2 incubator o n a rocking platform for 12- 14 h. Then 1 ml of medium was added t o each tube and the tubes were centrifugated. For each tube total radioactivity and radioactivity of half of the supernatant was measured in a gamma counter. Maximum release was determined by treating the target cells with distilled water overnight, which released 70-80 % of the total ”Cr.
Table 1. Spontaneous cytolytic activity in spleen cells from normal adult mice of different strains tested against YAC in uifro linea)
’’
The % ”Cr release was calculated according to t h e following formula: 2 x (radioactivity of 1/2 of the supernatant) % ” ~ rrelease = x 100 total radioactivity Results are expressed as percent lysis according t o the following formula:
% lysis =
% release in test
-
7% spontaneous release
x 100
% maximum release - % spontaneous release
The standard deviation between the two duplicates never exceeded 5 % lysis.
*
No. of
% lysis at 30/1 ratio
Strain
Exp.
Mean
Range
CBA A/Sn
12 9 2 1 1 2 18 4 3 1 2
63 16 30 31 38 48 37 23 24 29 30
34- 95 5-33 23-36
C57BL C3H ACA BALB/c A x C57BL A x Leaden A x ACA A x C3H A x CBA
46-50
9-70 9-44 5-48 16-44
a ) In each experiment 2-5 spleens are pooled and tested for cytotoxicity. For experimental details see Section 2.
As YAC is of A/Sn origin, we assumed that the low effect of normal A/sn cells might be due t o the lack of isoantigenic difference between effector and target cells. However, low spontaneous activity was also observed when the target cell YBA, another MLV leukemia of CBA origin, was tested against the non-H-2 compatible effector cells from A/Sn animals. Furthermore, effector cells of the syngeneic strain (CBA) were more cytolytic than A/Sn cells as could be seen from the experiment in Table 2. Table 2. Cytolytic activity of A/Sn and CBA spleen cells tested against the Y BA in vifro linea)
2.6. Competition studies Effector cell
1 x 1 O6 effector cells were incubated with various concentrations of “competing” tumor cells in duplicates for 1 h at 37 OC at a total volume of 0.2 ml complete medium. 2 x l o 4
’*Cr-labeled target cells were added u p t o a final volume of 1 ml. Thus a constant effector t o target cell ratio of 50: I was used in all experiments, while the ratio of competing tumor cells t o target cells was titrated out. The rest of the experiment was carried out as described above. All results are expressed as % inhibition at a certain competitor t o target cell ratio according t o the following formula: % lysis without competing cells % lysis with competing cells x 100. %inhibition = % lysis without competing cells
A/Sn % lysis 5011
Mouse 1 Mouse 2 Mouse 3 Mouse 4 Mouse 5 Mean % lysis
CBA
lo lysis 501 1
0
Mouse I Mouse 2 Mouse 3 Mouse4 Mouse 5
10 29 28
1
Mean Q lysis
18
2 2 3 -2
17
7
a) Spleens from five A/Sn and five CBA mice of the same age and sex were tested individually in the same experiment against the YBA in uitro line.
R. Kiessling, E. Klein and H. Wigzell
114
It would thus seem clear that for yet unknown reasons A/Sn mice have a comparatively low “background” activity of killer cells against these two MLV lymphomas.
3.2. Time kinetic studies on the cytolytic effect o n YAC-1 cells b y normal spleen cells In spite of the fact that the effector cells in our system were taken from normal nonimmunized mice, the onset of the in vitro lytic activity was remarkably fast. As could be seen in the time course experiment in Fig. I , a clear effect was obtained by the CBA spleen cells on the YAC target cells already after 4 h of incubation, with an effector-to-target cell ratio of SO/]. Thus, it seemed that a rapid cell lysis independent of known preimmunization or added antibodies was active in this system.
Eur. 1. Immunol. 1975.5: 112-117 Table 3. Specificity of spontaneous cytolytic activity for Moloney leukemia cellsa)
Exp.
H o u r s of
63 -4
I1
CBA CBA CBA A x BL A x BL AxBL
YAC-1 YCAB-1 P815 YAC-1 YCAB-1 P815
64 8 -2 38
A x BL A x BL AxBL A x BL A x BL A/Sn A/Sn A/Sn A/Sn A/Sn
YAC-1 YAC-301 YCAB-1 YCAB-6 P815 Y AC- 1 Y AC-30 1 YCAB-1 YCAB-6 P815
47
CBA CBA CBA CBA CBA
YAC-1 YCAB-1 YBA YAA EL-4
69
CBA CBA CBA CBA CBA CBA CBA CBA
YAC-1 YBA YCAB-1 YLI YAA LAA P815 E L-4
91 26 42
111
12
lncubotion
Figure 1. Effect of CBA spleen cells on YAC in vivo ( 0 - 0 ) or in vitro (0-0) cell lines as a function of incubation time. Spleens from 4 CBA mice were pooled. Spontaneous release after 12 h of incubation is for the in vifro line 13 70 and for the in vivo line 22 70 of thc total label.
’ib Lysis
(ratio 5011)
YAC-1 YCAB-1 P815
V
8
cells
CBA CBA CBA
1V
4
Target
I
1
1
Effector cells
-2
2 0
25 5 11 0 3 4
-2 2 -11
8 17
0 10
7
6 10
-2 10
a ) Comparison between different in vitro tumor cell lines for cytolytic sensitivity. Five independent experiments are shown. In each experiment the effector cells were derived from 2-5 spleens from CBA, A/Sn or A x C57BL F1-hybrids. The spontaneous release is within the same range for all cell lines compared (20-30 %of total
label). 3.3. Origin of mice In order to rule out the possibility that the activity of the normal spleen cells might depend o n “spread” of leukemia virus in our mouse colony, we also investigated mice from other sources where n o work on mouse leukemia viruses is carried out. We thus tested CBA mice from the Department of Genetics, Stockholm University and also BALB/c mice from a commercial breader in Denmark (Friis, see Section 2.2.). We found, however, that spleen cells taken from mice on the day of arrival t o the department were as efficient as our own mice of the same strain and age. 3.4. Comparative studies between different tumor cell lines
for their sensitivity for lysis by spleen cells from normal animals In order t o establish the nature of this reaction, more tumor cell lines were tested for sensitivity in the present system. Among the 8 cell lines tested significant effects (> 10 % lysis) have only been seen with three MLV cell lines. Two non-MLV lymphomas as well as two induced by MLV were not sensitive (Table 3). It is of interest that the P815 line, known t o be highly sensitive t o lymphocyte lysis, was not lysed.
3.5. Competition studies
The fact that a target cell is found t o be insensitive t o lysis does not exclude that the relevant target structures are present on the surface. Some target cells might be intrinsically more resistant t o lysis than others, and therefore, competition studies were carried out as a more direct way of defining presence or absence of relevant target structures in different cells [ 9 ] . These studies were designed so that different numbers of non-labeled tumor cells were added to a constant number of 5 1 Cr-labeled target cells and effector cells. Thus, a tumor cell with the relevant target structure on its surface may compete for the effector cells, resulting in reduction of lysis of the 5 1 Cr-labeled target cells. Tumor cells with irrelevant surface components would not compete. The results of a typical competition experiment are shown in Fig. 2. Here, as in all experiments, five different ratios of competitorto-target cell were used. In this experiment, among the 6 lines tested only the YAC and the YAA in vitro lines ;howed competition. We express the competition results as % inhibition (formula see Section 2.6.) at a competitorto-target cell ratio of 7/1.
Eur. J. Immunol. 1975.5: 112-117
Specificity of natural killer cells against Moloney leukemia cells
115
Table 5. Detailed analysis of the capacity of various in vitro tumoi Lines to compete in cytolysis against YAP) Competing tumor cell
YLl
93 70 70 87
YCAB-1
33
YAC YAA YBA
EL-4 P815 LAA .Oi
Exp. 1
37 22 n.d.
% inhibitionb) 11 111
59 54 n.d. n.d. -5 3 -8 n.d.
113 125 138 106 11
19 n.d.C) n.d.
1v 108 41 69 39 0
0 4 8
i
00:I
I:(
2:l
8:l
4:l
C o m p e t i t o r : torgel
a ) As effector cells CBA spleen cells at an effector-to-target cell ratio of 50/1 were used. As target cells the YAC-1 in vitro line were used. b) Competing capacity for each tumor cell line is expressed as % inhibition at a competitor-to-target cell ratio of 7/1. c) n.d. = Not done.
16.1
cell r a t i o
Figure 2. Competition experiment. As effector cells CBA spleen cells at an effector-to-target cell ratio of SO/l were used. As target cells the YAC-1 in vitro line were used. Competing cells are as follows: YAC-I in vitro line (*--.), YAC in vivo line (0-o), YAA in vitro line (x-x), YCAB-1 in vitro line (a--a), EL-4 in vitro line (A-A) and P815 in vitro line (0-0).
cluded that whereas none of the non-MLV-induced cell lines compete strongly, 3 out of 4 MLV-induced cell lines are efficient competitors.
3.8. Differences between in vitro and in vivo grown cell lines
3.6. Comparative studies between sensitivity for lysis and competing capacity Table 4 shows an experiment similar t o that described above in which it can be seen that a positive correlation between sensitivity for lysis and capacity t o compete exists. The competition experiments seemed t o be more sensitive in demonstrating the presence of relevant target structures than direct cytolytic tests o n the same cells (see YLI which was significant in competition but insensitive to lysis). As seen in Table 4, both sensitivity to lysis and capacity t o compete seems to be confined t o tumors induced by MLV (exception YCAB). Table 4. Positive correlation between cytolytic sensitivity of target cells and capacity to compete in the cytotoxic assaya)
In vitm line Y AC- 1 YAA YBA YLI YCABl
% lysis (CBA)
46 10
LAA
9 0 1 3
EL-4
5
P815
2
% inhibition
108 47 69 39 0
During these studies we observed that the in vivo grown tumors were considerably less sensitive t o the lytic effect of normal spleen cells than their corresponding in vitro lines. This difference can be clearly seen in Fig. 1, where both the in vitro and the in vivo grown YAC cells were used as target cells. The same difference was also seen when comparing in vivo and in vitro lines of the YBA target cell. As already mentioned, competition studies may be the most sensitive assay t o detect the relevant target structures in the present system. We therefore performed such studies t o compare in vivo with in vitro grown cells. In agreement with the results on cytolytic sensitivity, in vivo grown tumor cells did not compete, or competed t o a lesser extent than did their corresponding in vitro lines, as can be seen from the 3 experiments summarized in Table 6. It thus seems clear that the structure or antigen which the killer cells recognize is less well expressed on the in vivo grown MLV tumor cells.
4. Discussion
8
0 4
a) Experiment comparing the cytolytic sensitivity with competing capacity in different in vitro grown tumor cell Lines. As effector cells in both systems CBA spleen cells at an effector to target
cell ratio of 50/1 were used. As target cells in the competing system the YAC-1 in vitro cell line was used. Competing capacity for each tumor cell line is expressed as % inhibition at a competitor-to-target cell ratio of 7/1.
3.7. Differences in competing capacity between MLV- and non-MLV-induced tumor cells These results prompted us to carry out further competition experiments in order t o see if there was a consistent difference between MLV- and non-MLV-induced tumors. In Table 5 the results of four experiments are shown. It can be con-
In the present article we describe the existance of “naturally” occurring killer cells in mice with selective lytic activity against Moloney leukemia cells. Of the tumor lines tested, only Moloney leukemia cell lines were found t o be sensitive t o the cytotoxic activity of the killer cells. Between the M e loney leukemia lines there existed significant variations in susceptibility t o lysis. Also, in vitro grown cells were found to be more sensitive t o lysis than in vivo grown cells from the same Moloney leukemia cell line. In virro Moloney leukemia cells express more Moloney leukemia-associated antigens than the corresponding cells grown in vivo [ 101 and this might be relevant t o our findings. Susceptibility t o lysis could be shown t o be positively correlated t o the capacity of cells of the same tumor line t o function as competitive inhibitors when added t o the cytolytic test (see Table 4). Any susceptible Moloney leukemia cell line could be used as a selective competitor for lysis of isotope labeled
116
R. Kiessling, E. Klein, H. Pross and H. Wigzell
Eur. J. Immunol. 1975. 5: 112-117
Table 6. Comparison in competing capacity in the cytotoxic assays against YAC using in vivo or in vifro grown Moloney leukemia cell linesa) % inhibitionb)
Competing tumor cell YAC YAA Y BA Y LI
Exp. I in vitm
93 70 80 87
I1
in vivo 27 37 33 43
in vitro
113 125
n.d. 106
in vivo 43 75
n.d. 38
111 in v i m in vivo
1V in vitm
59
-3
87
n.d.c) n.d. n.d.
n.d. n.d. n.d.
n.d. n.d. n.d.
in vivo 13 n.d.
n.d. n.d.
a) As effector cells CBA spleen cells at an effector-to-target cell ratio of 50/1 were used. As target cells the YAC-1 in vifro line were used. b) Competing capacity for each tumor cell line is expressed as % inhibition at a competitor to target cell ratio of 7/1. c) n.d. = Not done.
target cells from another Moloney leukemia line demonstrating that the killer cells, irrespective of strain of origin, had the same specificity for the different Moloney leukemia lines. However, that each tumor line had its own level of susceptibility and each mouse strain had its own level of killer activity as well will be discussed further below.
No evidence for any influence linked t o allogeneic inhibition phenomena [ 1 I ] could be demonstrated in the present system, as the relative killer activity was the same whether tested against syngeneic or allogeneic Moloney leukemia cells. We could thus conclude that the killer cells from these normal mice behave as though specifically immunized against Moloney lymphoma-associated antigens. The in vifro cytolytic activity of the killer cells is very rapid and can be detected within less than 4 h (see Fig. I ) as detected by ”Gr release from the damaged target cells. In fact, when using purified killer cells as obtained by fractionation procedures [6], specific lysis can be found within 1 h after start of incubation. We would thus consider the time periods so short that conventional in vifro induction of immune killer cells is excluded and the present test actually demonstrates the existence of in vivo generated killer cells with cytotoxic potential for Moloney leukemia cells. As the nature of the killer cell is the topic of an accompanying article [6], let us only state that the effector cell has the morphological appearance of a small lymphocyte. However, it can be clearly separated from both B and T lymphocytes by organ distribution data, as well as by surface markers used in fractionation procedures. The relationship t o the lymphocytes described by Stobo and coworkers [7] remains t o be clarified. The killing capacity is seemingly mediated by contact, as all attempts t o demonstrate lytic activity of supernatants from in vitro cultivated killer cells have failed (unpublished observations). Spleen cells contain the highest killer activity of the lymphoid organs [6], and killer activity peaks around the first 3 months of the life of a mouse. A variation in t h e efficiency of killer cells present in the spleen of normal mice of different, inbred strains was observed with A/Sn mice containing the lowest killer activity (see Table I ) . Of t h e other strains tested, CBA mice seemed t o have t h e highest killer activity in their spleens. What is then the cause and origin of the present “spontaneous” killer cells? As already stated, the present killer cells would seem quite distinct from conventional, immune effector cells against Moloney leukemia cells that dominate
the in v i m cytotoxic reactions when immune anti-Moloney effector cells are used [ 121. This in itself makes a discussion as to the cause of the occurrence of the present killer cells difficult, as the present effector cells would seem to be lymphocytes of a new type with yet unknown biological properties. If we assume that the activity of the killer cells is of immunological nature, a simple explanation would be that they are directed against murine leukemia virus antigens, and as such viruses are known to be quite ubiquitous, this would seem a reasonable assumption. Alternatively, the killer activity might not be Moloney leukemia-specific, but rather of an autoimmune type directed against some organ-associated antigen being expressed on the Moloney leukemia cells but not being directly coded for by the Moloney leukemia virus. Such a n argument is partly weakened by the fact that both susceptible Moloney lymphomas and resistant lymphomas such as EL-4 can be shown to carry surface markers such as the theta antigen denoting their common, thymic origin, but only the Moloney leukemia cells are susceptible to lysis. Other workers have reported o n the existence of “spontaneous” killer cell activity against tumor cells in the mouse or rat systems, with two systems detecting anti-lymphoma activity [ 1 ,*I. In one system, preleukemic AKR mice were found to express a n anti-lymphoma specific activity with the effector cells being of probable macrophage type*, whereas in the other test spontaneous cytolytic activity was found in mice against a Rauscher virus-induced leukemia with the effector cells being of lymphocyte nature [ I ] . Interestingly, in the latter assay the effector cells, although of lymphocyte nature, were found to be resistant t o anti-theta serum plus complement treatment in analogy with results in the present system [6]. Received August 5 , 1974.
* Gomard, E., Leclerc, J.C. and Levy, J.P., submitted for publication.
5. References 1 Herberman, R., Nunn, M.E., Lavrin, D.H. and Asofsky, R., J. N u t
Gmcer Insf. 1973.51: 1509. 2 Greenberg, A.H. and Playfair, J.H.L., Clin. Exp. Immunol. 1974. 1 6 : 99.
3 Tagasuki, M., Mickey, M.R. and Terasaki, P., Cuncer Res. 1973. 33: 2898.
Eur. 1. Immunol. 1975.5: 117-121
A new cell type involved as natural killer cells against Moloney leukemia cells
4 Petranyi, G.G., Banczur, M., Onody, C. a n d Holland, S.R.,Lancet 1974. i : 736. 5 Perlmann, P., Perlmann, H. and Biberfeld, P., J. Irnrnunol. 1972. 108: 558. 6 Kiessling, R., Klein, E., Pross, H. and Wigzell, H., Eur. J. Irnrnunol. 1975. 5: 117. 7 Stobo, J.D., Rosenthal, AS. and Paul, W.E., J. Exp. Med. 1973. 138: 71. 8 Rygaard, J., Acta Pathol. Microbiol. Scand. 1969. 7 7 761.
R. Kiessling+, Eva Klein+, H. Pros+ and H. Wigzell+'
Department of Tumor Biology, Karolinska Institute, Stockholm+ and Department of Immunology, Uppsala University, Uppsala'
117
9 Rosenberg, E., McCoy, J., Green, F., Donelly, F., Siwarski, D., Lewine, P. and Herberman, R., J. Nut. Cancer Inst. 1974.52: 345. 10 Cikes, M., Friberg, S., Jr. and Klein, G., J. Naf. Cancer Inst. 1973 . 50: 347. 11 Hellstrom, K.E. and Hellstrom, I., Progr. Exp. Tumor Res. 1967. 9 : 40. 12 Lamon, E.W.,Wigzell, H., Klein, E., Andersson, B. and Skurzak, H., J. Exp. Med. 1973. 137: 1472.
"Natural" killer cells in the mouse II. Cytotoxic cells with specificity for mouse Moloney leukemia Cells. Characteristics O f the killer Cell" Normal mice contain cytolytic cells with specificity for in v i t r o grown mouse Moloney leukemia cells. Such killer cells are most frequent in the spleens; lymph node and bone marrow contain less and thymus virtually n o killer activity. Peak activity is found around one t o three months of age. Spleen cells from genetically athymic mice are as active killer cells as those from normal mice of the same strain. Treatment with anti-theta serum plus complement followed by removal of adherent and surface Ig positive cells by filtration through anti-lg columns will leave between 1-5 % of the original spleen cell population from a normal mouse. These cells have the morphology of small lymphocytes and perhaps contain all of the total original killer activity of the spleen against the Moloney leukemia cells. Such killer enriched cells are devoid of T and B lymphocytesand largely fail t o function in antibody induced, cell-mediated lysis against antibody-coated chicken erythrocytes. It is concluded that the spontaneous selective cytotoxic activity of normal mouse spleen cells against Moloney leukemia cells is exerted by small lymphocytes of yet undefined nature.
1. Introduction
Specific cytolytic or growth inhibitory activity has been ascribed t o various effector cells derived from the reticuloendothelial cell system. Thymus-derived lymphocytes have been shown t o function as highly efficient killer cells [ I ] , which recognize the target via its actively produced, antigenspecific surface receptors [ 21. Macrophages can be rendered selectively cytotoxic via a specific T cell factor [ 3 ] . Monocytesand lymphocytesof yet undefined type have been found capable of cytolitic aggression towards IgC-coated target cells [4]. Immune B lymphocytes were shown t o be necessary in certain growth inhibitory assays in vitro [S]. A variety of cells have thus been shown t o be effective in various in vitro cytolytic or growth inhibitory test systems. We would like t o report another new cell type having selective in vitro cytotoxic behavior with unique functional as well as surface characteristics. This cell is found in normal mice, yet it displays specific cytolytic activity against certain mouse Moloney leukemia lines in v i t r o . [I 8761
* Supported by the Swedish Cancer Society, and by NIH contracts NOI-CB-33870 and NOI-CB-33859.H.P. was supported by a fellowship from the Canadian Medical Research Council. Correspondence: Rolf Kiessling, Department of Tumor Biology, Karolinska Institute, S-104 01 Stockholm 60, Sweden Abbreviations: MLV: Moloney leukemia virus PBS: Phosphate buffered saline CRBC: Chicken red blood cells FCS: Fetal calf serum
It will cause rapid lysis of such leukemic cells in the absence of added antibodies [ 6 ] . As judged by surface markers and other characteristics it cannot be classified as a lymphocyte of B o r T type nor as a monocyte. The distribution pattern with regard t o age and organs will be reported, as well as its lack of known surface markers.
2. Materials and method 2.1. Materials and methods described in the previous paper [6]
These refer to animals and tumors used, to tissue culture medium, target cell labeling, cytotoxic test and preparation of effector cell suspension.
2.2. Assay of antibody-dependent cell-mediated cytotoxicity Fresh chicken red cells, washed 3 times in PBS, pH 7.4 were incubated at 37 OC with 150 pCi "Cr per lo' CRBC in 0.35 ml PBS for 3-4 h. Thereafter they were washed three times in culture medium and adjusted t o a concentration of 105 cells/ml. The cytotoxic test was carried out using Ellermann plastic tubes ( 1 1 mm x 55 mm). Each tube contained 5 x lo4 "Crlabeled CRBC, normal rabbit sera of rabbit anti-chicken sera at a 104-fold dilution and varying amounts of spleen cells in a total of one ml of culture medium (F-13 plus 10 % heatinactivated FCS). The tubes were centrifuged at 2000 x g for
