Original Paper International Journal of Cell Cloning 10:190-195 (1992)

Differential Lysis of firnor Target Cells Displayed by Lymphokine Activated Killer (LAK) Cell Clones Pamela Beana, Ravin Agah ', Amitabha Mazumderb "Department of Pathology, University of Southern California, School of Medicine, Los Angeles, California, USA; hDivision of Medical Oncology, Georgetown University, School of Medicine, Washington, D.C., USA Key Words. LAK

Clones

Receptors

Abstract. Lymphokine-activated killer (LAK) cells exhibit major histocompatibility complex (MHC) unrestricted cytolysis against a wide variety of fresh and cultured tumor cells. Because previous work from our laboratory suggested that trypsin treatment of unseparated populations of LAK cells had a differential effect on lysis of different tumors, in this report we analyzed the lytic specificity of LAK cell clones against a panel of three different targets: MCA, B16 and YAC-1. We found that 21 out of the 24 analyzed rnurine spleen and bone marrow clones killed a combination of two, but not all three, of these tumor cells. Determinations of the phenotype of 10 LAK cell clones showed six with rearrangements for the T cell receptor (TCR)P chain gene, suggesting a T cell origin, and four with germ line configurations for the TCRP and 6 chain genes, a result consistent with a non-T cell lineage. This cloning procedure provided an experimental tool to develop new procedures of adaptive immunotherapy.

cells could make important contributions to the improvement of cancer immunotherapy. In fact, assuming that each effector cell in the LAK population is capable of mediating the lysis of not all, but just one or a combination of tumor cells, selecting for the appropriate LAK cells that kill any given tumor could considerably improve cell transfer therapy. In this study, using rIL-2-activated murine spleen and bone marrow cells, we developed a procedure for producing LAK cell clones with an elevated but differential cytotoxic activity against different tumor cells. In order to determine whether the culture conditions used in this study allowed generations of clones derived from both NK and T cell lineages, we analyzed the phenotype of these clones by detection of rearrangements of the T cell receptor (TCR)P and 6 chain genes, expression of specific cell surface markers and patterns of cytotoxicity against a panel of tumor cells.

Introduction Lymphokine-activated killer (LAK) cells are immune cells which, after activation with recombinant interleukin 2 (rIL-2), acquire the capacity to kill a variety of tumor targets in a reaction which is independent of the major histocompatibility complex (MHC) composition of the tumor [l]. The major contributors to the LAK system are activated natural killer (NK) cells and different subpopulations of T cells [2-41. The elucidation of the mechanisms by which LAK cells recognize a broad spectrum of tumor target

Correspondence: Dr. Pamela Bean, Specialty Laboratories, Inc., 221 1 Michigan Avenue, Santa Monica, CA 90404-3900, USA. Received December 19, 1991; accepted for publication March 10, 1992. 0131- 1454/92/$2.00/0 OAlphaMed Press

Materials and Methods Cells and Culture Media The target cells employed in this study were the murine cell lines B16 melanoma [S], MCA methylcholanthrene-induced sarcoma [6] and YAC- 1 lymphoma [ 7 ] . B16 and MCA are resistant to NK lysis while YAC-1 is NK-sensitive. B16 and YAC-I cells were maintained and passaged in complete media (CM), grown to log phase (3 x lo5to 5 x lo5 cells/ml) and then stored as cryopreserved cells to have a standard stock for assay. The composition of CM has been described previously [8]. MCA cells were passaged in vivo by injecting 2 x lo6 cells intradermally into mice. After 2 weeks of active growth, the tumor was extracted and cell suspensions were obtained by treatment with collagenase (0.01 pglml). hyaluronidase (2 x l o 3 pg/rnl) and DNase (2 x pg/ml) in HBSS (Hanks Bal-

LAK Clones Display Differential Tumor Lysis

191 Table 1. Differential tumor cell lysis mediated by LAK cell clones" Groupb

Clone

Origin

Effector: target

% Cytotoxicity against targetsc B16 MCA YAC

spleen spleen spleen

6: 1 6: 1 4: 1

33 20 24

spleen spleen spleen spleen bone marrow

4: 1 4:1 4: 1 4: 1 4: 1

3

15

10

2 2 0 2

22 32 14 32

36

spleen spleen spleen spleen spleen spleen spleen spleen spleen spleen bone marrow bone marrow bone marrow bone marrow

4: 1 4: 1 4:1 4: 1 6: 1 6: 1 4: 1 4: 1 4: 1 6: 1 4: 1 4: 1 4: 1 4: 1

35 18 20 18 10 39 24 21 12 11

2 0 0 0 0 2

19 22

spleen bone marrow

4: 1 4: 1

18 16

16 24 34

26

28 14

31

0 0 0

0 8 0 2 1

55 47 45

25 22 58

35 25 10

58 11 87 27 21 30 19 29

30

23

4

19

6

OLAK cell clones were derived from 3 days rIL-2-activated spleen and bone marrow cells from C57B116 mice. bGroup 1: killed all three targets; Group 2: killed MCA and YAC-I; Group 3: killed B16 and YAC-1; Group 4: killed MCA and B-16. 'Results are expressed as the mean of three measurements of the percentages of 5'Cr released by tumor cells in a 4 h cytotoxicity assay. SE was 110% of the mean value in all experiments reported.

ance Salt Solution, Gibco). Tumor cells were then stored as a cryopreserved stock for assay.

Generation of C57B1/6 Spleen and Bone Marrow LAK Cells Generation of LAK cells was performed by aseptically removing spleens or bone marrow from C57B1/6 mice, 8 weeks or older, as described previously [8]. CytotoxiciQ Assay Four h chromium (5'Cr) release assays were performed as described elsewhere [9].

Cloning of LAK Cells Cloning was done by seeding limiting numbers (1 and 0.1 cell/well) of three days pre-activated C57B1/6 spleen or bone marrow LAK cells in round-bottom microwells containing 5 x l o 3 irradi-

ated (1,500 rads) spleen feeder cells in a final volume of 0.2 ml C M supplemented with 500 U/ml rIL-2. Microcultures were fed weekly with the same amount of feeder cells per well and rIL-2 (500 U/ ml). Every other week the clones were also supplemented with y-interferon (100 U/rnl). After four weeks in culture and three days after the addition of rIL-2, each microwell was assessed for cytolytic activity against 5'Cr-labeled B 16 targets using a replica plating assay. Clones with 210% activity were expanded into larger wells according to their growth rates, and their phenotypes were determined using anti-Thy- 1, anti-asialo G M I and anti-NK 1.1 antibodies. After a further two months of growth, clones were examined for p and 6 TCR gene rearrangements, and their cytolytic activity against B 16, MCA and YAC-I tumor cells was assessed. Poison distribution studies confirmed the clonal identity of these effector cell lines.

192

BeanlAgahlMazumder

provided by D K Mitch Kronenberg, University of California, Los Angeles, California). Hybridization conditions were as previously described [ 111.

Results

Fig. 1. TCRP chain gene rearrangements in LAK cell clones. A)Lane 1, banding pattern in P815 germ line control; lane 2, clone 4; lane 3, clone 9; lanes 4 and 5, clone 19 frozen after two different times in culture; lane 6, clone 21; and lane 7, clone 30. B) Lanes 1 and 6 are the banding pattern in the P8 I5 germ line configuration control. The corresponding banding patterns for clones 2 (lanes 2 and 7). 7 (lanes 3 and 8). 22 (lanes 4 and 9) and 31 (lanes 5 and 10) are presented.

Southern Blots Rearrangements for the TCRP and 6 chain genes in LAK cell clones were determined by Southern blot analysis [ 101. High molecular weight DNA was digested with EcoRl and BamHI (Boehringer Mannheim, Indianapolis, Indiana). The munne mastocytoma line, P815, was used as a control for the germ line configuration of TCR P and 6 chain genes. Digests were then electrophoresed through 0.8% agarose gels and blotted on nylon membranes according to standard procedures [lo]. The blots were probed with lo6 cpm/ml 32P-labeled human C,, DNA (Oncor, Gaithersburg, Maryland) or murine J6, DNA (kindly

Differential Tumor Cell Lysis Mediated by LAK Cell Clones Cell lines were defined as clones only when derived from 96 well plates where Inine wells (10%) of the plate had cell growth. Of 2196 wells seeded, 55 exhibited growth after four weeks in culture, 48 wells from plates seeded at 0.1 and seven wells from plates seeded at one cell/well, respectively. LAK cell clones were first selected for their ability to kill B16 tumor cells ( ~ 1 0 %lysis) and, after propagation, their lytic activity was further tested against B 16, MCA and YAC- 1 targets (Table I). Clones were classified into four groups according to their patterns of lysis. Group 1 consists of clones which killed all three targets with >lo% efficiency. Three out of 18 (17%) spleen clones were cytotoxic for B 16, MCA and YAC-1 targets. None of the six clones derived from bone marrow exhibited this lytic pattern. Group 2 represents clones with activities against MCA and YAC- 1 targets exclusively. Four out of 18 (22%) spleen and one out of six (17%) bone marrow LAK clones lost anti-B16 activity during culture but were cytotoxic for YAC- 1 and MCA cell lines. Group 3, the major subset, contained 10 of 18 (56%) spleen and four of six (67%) bone marrow clones which killed only B16 and YAC-1 targets. These clones did not lyse MCA tumor cells. Finally, two clones representing Group 4, one from spleen and the other from bone marrow, were mainly cytotoxic for MCA and B 16 cells, both NK-resistant targets. TCR Rearrangements in LAK Cell Clones DNA from P815 cells probed with CTP showed 10 kb and 6 kb germ line bands for BamHI digests (Fig. 1A and lB, lane 1) and 10 kb and 1.5 kb germ line bands for the EcoRl enzyme (Fig. lB, lane 6 ) . BamHI digests of clones 4 and 21 (Fig. I A , lanes 2 and 6, respectively) each exhibited one rearranged band different from the germ line (cf. Fig. lA, lane 1). Clone 30 (lane 7) presented only the smaller germ line BamHI DNA fragment. Alternatively, clones 9, 19(a), and 19(b) (lanes 3, 4 and 5, respectively) do not appear to have rearranged the TCRP gene. Clones 19(a) and 19(b) represent the same clone frozen at two different times during proliferation, suggesting that long periods in culture did not induce alterations of the germ line configuration for the TCRP gene.

LAK Clones Display Differential Tumor Lysis

193

BamHI digestion of clones 7, 22 and 31 (Fig. lB, lanes 3, 4 and 5, respectively) showed rearrangement for the TCRP chain gene. Two of these clones (22 and 3 1 ) did not show equivalent changes in the banding pattern for their corresponding EcoRl digests (lanes 9 and 10, respectively). This pattern can be explained by assuming that the EcoRl sites for both germ line bands generated with this enzyme (10 kb and 1.5 kb) lie downstream from the V-D-J segments that rearranged to form the mature TCRP chain gene in these two clones. The only modification expected as a result of these joining events is in the size of the 10 kb BamHI fragment (lanes 3 and 5). Clone 2 (lanes 2 and 7) and clone A3 (blot not shown) exhibited the germ line configuration for the p chain gene. Thus, in all, six of ten LAK cell clones (4, 7, 21, 22, 30, and 31) exhibited a rearranged TCRP chain gene. LAK cell clones which exhibited germ line bands for the P chain gene were examined for rearrangements of the 6 gene. DNA from P815 cells probed with J6, showed an 8 kb germ line band for EcoRl digest and a 10 kb germ line band for the BamHI enzyme (Fig. 2, lanes 1 and 5, respectively). LAK cell clones 2, 19 and A3 exhibited the germ line configuration (Fig. 2). LAK clone 9 also exhibited the germ line banding pattern (data not shown).

1

2

3

Eco RI

4

5

I

6

7

8

Barn H I

Fig. 2. TCRG chain gene configuration in LAK cell clones. Lane 1 is the banding pattern in the P815 germ line configuration control. The corresponding banding patterns for clones 2 (lanes 2 and 6). A, (lanes 3 and 7) and 19 (lanes 4 and 8) are presented.

different cell lineages. The former showed germ line bands for the TCR genes and exhibited only the Thy-1 surface marker, while the latter presented rearrangements for the TCRP gene, presented a deleted TCR6 gene (data not shown) and tested positive for Thy-I and asialo-GM 1 antibodies. Similarly, clones 4 and A3 (group 2) seem to have derived from different progenitors, and, after activation, lyse the same tumors: MCA and YAC- 1. Moreover, clones 7 and 2, with T cell and non-T cell origins, respectively, are both cytotoxic for B16 and YAC-1 but not for the MCA cell line.

Patterns of Cytotoxiciw and Phenotypes of LAK Cell Clones

Discussion

The associations between the cytotoxicity of LAK cell clones against different tumors and the origin of these clones were investigated (Table 11). Clones 19 and 30 (group l), shared the capacity to kill all three tumor cells tested, yet they belong to

The prevalent hypothesis concerning lysis of tumor cells by LAK effectors claims the existence of an unique share epitope in all targets that is recognized equally well by all LAK cells [ 12, 131.

Table 11. Cytotoxicity and phenotypes of representative LAK cell clones

Group"

Clone

1

19 30 4 A, 7 2

2 3

Targets lysed B 16-MCA-YAC B 16-MCA-YAC -MCA-YAC -MCA-YAC B16-YAC B16-YAC

TCR Gene rearrangementb (JG,) (CTP) -

+ + -

+ -

D D D

-

Phenotype' Thy 1 AGM 1

+ + +

ND'

+

NCd

-

+

+

ND -

"Groups defined as in Table I. h(-) = germ line; (+) = rearranged; D = deleted (data not shown) 'ND = Not done. dNC = Not conclusive. 'A specific cell surface antigen was identified by comparing the activity of the LAK cell clones against control (untreated) and antibody-treated effectors. A positive (+) result indicates a decrease of 9 0 % of baseline tumor cell lysis when comparing the percentages of cytotoxicity exerted by each LAK cell clone after treatment with and without antibody plus complement.

Bean/Agah/Mazumder Only one study identified a selective lysis of tumor cells by IL-2 expanded peripheral mononuclear cells [14]. In this report, we studied whether LAK cells do indeed exert a distinct spectrum of reactivity against different tumors by cloning murine LAK cells and determining the lytic potential of each effector cell line against B16, YAC-1 and MCA targets. The results presented in Table I show 21 out of 24 LAK cell clones displaying a selective lytic activity against tumor cells. Only 3 out of 24 clones tested lysed all three targets with similar capacities, and most of the clones killed combinations of only two cell lines. This is in agreement with previous work from our laboratory demonstrating that treatment of whole unselected populations of murine spleen LAK cells with proteolytic enzymes had a differential effect on lysis of tumor targets [ 151. Sensitivity to tumor necrosis factor [ 161, susceptibility of tumor cells to cytostatic drugs [ 171 and the expression of cell adhesion molecules in effectors and targets [I81 have recently been implicated as additional mediators of resistance to LAK lysis. Because finding the presence of both NK and T cell types within these clones would mean that the culture system used was generating heterogeneous effector cell lines, we next determined LAK clone phenotypes. Six clones (4, 7, 21, 22, 30 and 31) exhibited TCR banding patterns pertinent to a T cell lineage and four others (2, 9, 19 and AJ had germ line configurations for both the p and 6 chain genes, suggesting a non-T cell origin. Moreover, the experiments with cell surface markers showed that nine clones examined (7, 19, 16, 4, 30, A12, F3, E l 2 and H1) tested positive for the Thy-1 marker but were heterogeneous with respect to the asialo-GM 1 and NK 1.1 markers (data not shown). Therefore, not only non-T cells but also several subpopulations of T cells were derived with this cloning procedure, a phenomenon which has been reported previously [ 191. Several observations are noteworthy from the correlations established between LAK cell clone cytotoxicity and phenotypes. First, within each group, lysis of the same combination of tumor cells has been achieved by a T cell clone exhibiting a rearranged TCR P chain and deleted 6 chain gene and a non-T cell clone presenting the germ line configuration for both of these genes. Secondly, clone 30 in group 1 and clone 7 in group 3 showed the same pattern of TCR rearrangements and tested positive for Thy- 1, but they differed in the expression of the asialo-GM 1 marker. These two clones most likely represent different subsets of the T cell lineage. Thirdly, clone 19 showed germ line bands for TCRP and 6

194 chain genes, and clone 7 had a rearranged TCRP gene and a deleted TCR6 chain gene; however, they both tested Thy-](+) and asialo-GM1 (-). This is consistent with the observation that NK cells have been found to express some T cell associated markers [20]. Finally, of three clones which lysed B16 and YAC-I (but not the MCA targets), two exhibited Thy-1 but lacked asialo-GM1 and N K 1.1, and the other one tested positive for all three markers (data not shown). In summary, we found that LAK cell clones lysed histologically different tumor cells with different capacities. We have shown that the cloning procedure used in this study generated a substantial number of different types of LAK cell clones, suggesting amplification of a possibly representative sample of the heterogeneous LAK population. Thus, if the cell yield is improved and clone cytotoxicity is monitored routinely, more effective protocols in adoptive immunotherapy could be developed by utilizing a functionally well-characterized group of clones when treating specific tumors. In preliminary experiments, three clones (16,55 and 63) tested showed reduction of pulmonary metastases of the tumor cell line (MCA versus B16) that they lysed in vitro when injected in vivo, whereas two noncytolytic clones showed no activity in vivo.

Acknowledgments This work was supported by PHS grants CA 42962 and C A 01222 and by a grant from the Hastings Foundation. We thank D K B. Yankelevich for stimulating discussions, DK Karen L. McKeown and Dr. Herm Reyes for a critical review of this manuscript, and Nancy E. Campman for typing the manuscript.

References I Rosenstein M, Yron I, Kaufmann Y,Rosengerg SA.

Lymphokine-activated killer cells: lysis of fresh syngeneic natural killer resistant murine tumor cells by lymphocytes cultured in interleukin-2. Cancer Res 1984;44:1946-1953. 2 Gerosa F, Tommasi M, Spiazzi AL, et al. Heterogeneity of lymphokine-activated killer (LAK) populations at the clonal level: both NK and CD3'. CD4-, CD8- clones efficiently mediate tumor cell killing. Clin Immunol Immunopath 1988;49:91-100. 3 Fisch P,Malkovsky M, Braakman E, et al. y/S T cell clones and natural killer cell clones mediate distinct patterns of non-major histocompatibility complex-restricted cytolysis. J Exp Med 1990;171: 1567-1579.

LAK Clones Display Differential Tumor Lysis

195 4 Testa U, Cane A, Montensoro E, et al. Interleukin-2dependent long-term cultures of low-density lymphocytes allow the proliferation of lymphokine-activated killer cells with natural killer, T y/S or TNK phenotype. Cancer Immunol Immunother 1990;3l : l l - 18. 5 Poste G, Doll J, Brown AE, Tzeng J, Zeidman I. Companion of the metastatic properties of B16 melanoma clones isolated from cultured cell lines, subcutaneous tumors and individual lung metastasis. Cancer Res 1982;42:2770-2778.

6 Reznikoff CA, Brankow DW, Heidelberger 0.Establishment and characterization of a cloned line of C3H mouse embryo cells sensitive to postconfluence inhibition of division. Cancer Res 1973;33:3231. 7 Cikes M, Friberg S, Mein G . Progressive loss of H-2 antigen with concomitant increase of cell surface antigen(s) determined by Moloney leukemic virus in cultured murine lymphomas. J Natl Cancer Inst 1973;50:347-362. 8 Girgis E, Agah R. Sherrod A, Mazumder A. Enrichment of interleukin-2 activated bone marrow cells against disseminated neoplasia and comparison with interleukin-2 activated splenocytes. The Cancer Journal 1989;2:331-333.

9 Rosenstein M, Eberlein T, Schwartz S, Rosenberg SA. Simplified techniques for the isolation of alloreactive cell lines and clones with specific cytotoxic or proliferative activity. J Immunol Methods 1983;61:183- 193. 10 Southern EM. Detection of specific sequences among

DNA fragments separated by gel electrophoresis. J Mol Biol 1975;98:503-517. 1I

Meinkoth J, Wahl G. Hybridization of nucleic acids immobilized on solid supports. Anal Biochem 1984; 1381267-284.

12 Rayner AA, Grimm EA, Litze MT, Wilson DJ, Rosenberg SA. Lymphokine-activated killer (LAK) cell phenomenon. IV. Lysis by LAK cell clones of

fresh human tumor cells from autologous and multiple allogeneic tumors. J Natl Cancer Inst 1985; 75167-75. 13 Ferrini S , Miescher S , Zocchi MR, Fliedner V, Moretta A. Phenotypic and functional characterization of recombinant interleukin-2 induced activated killer cells: analysis at the population and clonal levels. J Immunol 1987;138: 1297- 1302. 14 Chen BP, Hank JA, Krauss EE, Sondel PM. Selective lysis of target cells by interleukin-2 expanded peripheral blood mononuclear leukocyte clones. Cell Immunol 1989;118:458-469. 15 Bean P, Agah R, Mazumder A. Heterogeneity of cell surface structures involved in cytotoxicity mediated by lymphokine-activated killer cells. J Biol Response Mod 1990;9:92-97. 16 Bean P, Mazumder A. Resistance of different tumor cells to lysis by lymphokine-activated killer cells can be mediated by distinct mechanisms. Immunobiology 1992; (in press). 17 Rivoltini L, Ganibacorti-Passerini C, Supino R, Parmiani G. Generation and partial characterization of melanoma sublines resistant to lymphokine activated killer (LAK) cells. Relevance to doxirubicin resistance. Int J Cancer 1989;43:880-885. 18 Gambacorti-Passerini C, Supino R, Marinoni A, Phillips J, Sondel PM, Parmiani G. Human lymphoma clones with stable resistance to lymphokineactivated killer cells show decreased LFA-3 and ICAM-I expression and resistance to TNF-a. J Immunol Res 1991;3:102-105. 19 Phillips JH, Lanier L. Dissection of the lymphokineactivated killer phenomenon. J Exp Med 1986; 164: 8 14-825. 20 Grossman Z , Herberman RB. Natural killer cells and their relationship to T cells: hypothesis on the role of T cell receptor gene rearrangement on the course of adaptive differentiation. Cancer Res 1986; 4612651-2658.