The Tumorigenicity of IL-2 Gene-Transfected Murine M-3D Melanoma Cells Is Determined by the Magnitude and Quality of the Host Defense Reaction: NK Cells Play a Major Role

Achim Schneeberger, Frieder Koszik, Walter Schmidt, Raphaela Kutil and Georg Stingl
J Immunol June 1, 1999, 162 (11) 6650-6657;

Abstract

Transfection of a variety of tumor lines with the IL-2 gene strongly reduces their tumorigenic potential when applied to either euthymic or athymic animals. To elucidate the mechanisms underlying this phenomenon, we inoculated IL-2-transfected M-3D melanoma (M-3D-IL-2) cells into DBA/2 mice immunosuppressed by γ-irradiation. Animals thus treated developed pigmented tumors, suggesting that IL-2 transfection of melanoma cells, instead of altering their neoplastic growth properties, renders them capable of evoking a tumoricidal host response. To define the critical effector cell, we injected M-3D-IL-2 and, for control purposes, nontransfected M-3D cells into DBA/2 recipients and analyzed the injection site. We found that 1) IL-2-expressing M-3D cells induce a much stronger inflammatory reaction than wild-type cells, 2) in both instances the infiltrate consists mainly of macrophages (40–60%) and granulocytes (30–40%), and 3) only the infiltrate of M-3D-IL-2 cell deposits contains a minor fraction of NK cells (∼1–2%). When we reconstituted sublethally irradiated animals with various leukocyte subsets, we found that unfractionated as well as macrophage-depleted peritoneal lavage cells but not NK cell-depleted peritoneal lavage cells were able to suppress the growth of IL-2-expressing M-3D cells. In vivo leukocyte depletion experiments showed that the NK cell-depleting asialo-GM1 antiserum, but not anti-macrophage and/or anti-granulocyte reagents, restored the tumorigenicity of M-3D-IL-2 cells. Our results indicate that the inflammatory tissue response evoked by IL-2-transfected cancer cells includes the attraction and/or activation of NK cells and that, in the experimental system used, these cells are critically needed for successfully controlling cancer growth in vivo.

During the last several years, we and others have shown that 1) the genetic modification of cancer cells to secrete IL-2 abrogates their ability to form tumors upon s.c. injection into the syngeneic host and 2) animals that had been inoculated with such cells are able to reject a subsequent challenge with the parental cancer cells (1, 2, 3, 4, 5, 6, 7, 8, 9). There exists good evidence that the latter phenomenon is the consequence of a T cell-mediated anti-tumor immune response including both CD4+ and CD8+ cells (1, 3, 7, 8). Much less understood are the processes that reduce the tumorigenicity of IL-2-expressing cancer cells. It is conceivable that the transfected gene somehow interferes with the intrinsic tumorigenicity of cancer cells and/or that it modifies the complex tumor cell-host interactions. The first possibility is quite unlikely, because our previous experiments demonstrated that IL-2 transfection does not adversely influence the in vitro growth characteristics of M-3 cells (1). Alternatively, the transgene product could either disturb the microenvironment needed for tumor formation (e.g., cell/cell or cell/matrix interactions or neovascularization) or induce a series of tumor-destructive host defense mechanisms. The presence of an inflammatory infiltrate that, in most instances, is dominated by macrophages and granulocytes (4, 10, 11) and directly correlates in intensity with the amount of IL-2 produced by the cancer cells (10) suggests that infiltrating leukocytes suppress the growth of IL-2-expressing cancer cells.

To determine whether this is indeed the case and, if so, to identify the relevant effector cell population(s), we generated a series of IL-2-expressing M-3D clones. These clones did not form tumors in syngeneic animals and could be used to induce a protective and specific antitumor immune response. We reasoned that IL-2-transfected M-3D melanoma (M-3D-IL-2)4 cell-injected animals depleted of and/or reconstituted with different leukocyte subpopulations should allow us to define the critical tumoricidal host cell subset.

Materials and Methods

Animals

DBA/2 (H-2d) and athymic nu/nu BALB/c mice were obtained from Charles River Wiga GmbH (Sulzfeld, Germany). Animals were used at the age of 8–10 wk. Mice that had been immunosuppressed either by sublethal whole-body γ-irradiation or by injection of leukocyte-depleting Abs were kept in a laminar flow hood and were supplied with a pasteurized diet and neomycin sulfate (2 mg/ml; Sigma, St. Louis, MO)-supplemented water ad libitum. For injection of tumor cells, animals were anesthetized by i.p. injection of a 2.5% (v/v) solution (∼350 μl/mouse) of 2,2,2-tribromoethanol (Aldrich Chemical, Milwaukee, WI) in PBS (Avertin). All animal procedures were approved by the Austrian Ministry of Science and Transportation (approval no. GZ 66009/107-I/A/2/93).

Cell lines and culture conditions

Murine M-3 melanoma cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). This cell line is derived from the S91 Cloudman melanoma that spontaneously arose in a DBA/2 mouse (12, 13). These cells were grown in DMEM (Life Technologies, Paisley, Scotland) supplemented with 10% FCS (PAA Laboratories GmbH, Linz, Austria), 10 mM HEPES, 2 mM glutamine, 100 U penicillin/100 μg/ml streptomycin, and 2 mM mercaptoethanol (all Life Technologies), and 0.1 mM nonessential amino acids (JRH Biosciences, Lenexa, KS). We are referring to M-3 cells grown under these conditions as M-3D cells. For the culture of transfected M-3D cells, DMEM was supplemented with 1 mg/ml of geneticin sulfate (G-418, Life Technologies). All M-3D lines and clones used in this study were propagated on culture flasks (Nunc, Roskilde, Denmark) coated with gelatin (Bio-Rad, Richmond, CA; 0.1% (w/v) in Aqua destillata; 24-h incubation at 4°C).

Generation of IL-2 gene-transfected M-3D cells

The Lipofectamin (Life Technologies) method was used to cotransfect M-3D cells with the murine IL-2 gene-containing plasmid pWS2m (1) and the pRSVneo resistance plasmid. The transfection mixture was prepared by adding 25 μl of Lipofectamin to 175 μl of PBS containing 18 μg of linearized (using XM I) pWS2m and 2 μg of pRSVneo DNA. After 15 min of incubation at room temperature, 1.8 ml of DMEM without FCS was added, and this mixture was used to transfect 5 × 105 M-3D cells, which had been plated the day before and were washed once in PBS before transfection was started. After 4 h of transfection, 5 ml of DMEM containing 10% FCS was added and, after 24 h, medium was completely removed and substituted with 10% FCS-containing DMEM supplemented with 3 mg/ml G-418 (Life Technologies). Two days later, the G-418 concentration was reduced to 1 mg/ml. During the selection phase, medium was changed once a week. Limiting dilution of subconfluent bulk cultures was performed to generate IL-2-producing M-3D clones. IL-2 production of the clones obtained was assessed using the CTLL bioassay as described (14, 15).

Irradiation of animals

Immunosuppression of animals was achieved by whole-body irradiation with doses of 3, 4, or 5 Gy using either an x-ray device (Philips RT 305, Philips, Hamburg, Germany; 1.5 Gy/min) or a 137Cs γ-ray source (Gammacell 40, Nordion International, Kanata, Ontario, Canada; 1.2 Gy/min).

Tumorigenicity studies

Injection of IL-2-expressing M-3D clones into euthymic and athymic nude animals.

DBA/2 or BALB/c nu/nu mice (at least five per group) were s.c. inoculated with graded doses (1 × 104, 1 × 105, 5 × 105, 1 × 106, and 2 × 106) of IL-2-producing M-3D clones and, for control purposes, with neomycin resistance gene-transfected or wild-type M-3D cells. In a second set of experiments, DBA/2 mice received mixtures of IL-2-producing (1 × 105) and wild-type M-3D (1 × 105, 2 × 105, and 3 × 105) cells. Animals were inspected daily, and tumor growth was monitored every 5–8 days by measuring the largest diameter and the two largest right-angle diameters to it. Mice were killed for ethical reasons when tumors became ulcerated or tumor diameters exceeded 20 mm.

Injection of IL-2-expressing M-3D clones into sublethally irradiated DBA/2 mice and Winn transfer assay.

Groups of DBA/2 mice were whole-body irradiated with doses of 3, 4, and 5 Gy, respectively. One day later, they were s.c. injected with 1 × 105 IL-2-producing M-3D cells; with a mixture of 1 × 105 wild-type M-3D cells and 1 × 105 IL-2-producing M-3D cells; or, for control purposes, with 1 × 105 wild-type M-3D cells.

In certain experiments, sublethally irradiated animals (5 Gy) were reconstituted with nonstimulated peritoneal lavage cells (PLC) in a Winn transfer setting (1, 16). To this end, mice were injected with 1 × 105 IL-2-producing M-3D cells either alone or mixed with 2 × 106 unfractionated, macrophage-depleted, or NK cell-depleted PLC at 24 h after the irradiation. For control purposes, we administered non-IL-2-producing M-3D cells alone or in combination with PLC. Tumor growth was assessed as described above.

Isolation of PLC

DBA/2 mice were killed by CO2 asphyxiation and used to prepare PLC. Briefly, 5 ml of RPMI 1640/10% FCS were instilled into the peritoneal cavity of the animals, and the abdomen of injected mice was massaged slightly to better mobilize the PLC. Thereafter, the PLC-containing lavage fluid was removed using a syringe equipped with a 21-gauge needle. Cells were put immediately on ice, washed, and counted. Cell yield per mouse was regularly between 2 and 3 × 106 cells. PLC were depleted of macrophages by a 1-h incubation at 37°C in tissue culture flasks (Nunc) containing FCS-supplemented RPMI 1640. Nonadherent cells were recovered by gentle washing with warm FCS-supplemented RPMI 1640 and assessed for contaminating macrophages by FACS analysis. For NK cell depletion, the PLC were treated with asialo-GM1 antiserum (dilution, 1:100) for 30 min at 4°C. Low-Tox-M rabbit complement (Cedarlane Laboratories, Hornby, Ontario, Canada) was then added at a final dilution of 1:10 for 45 min at 37°C. A population of >95% viable cells was recovered after washing and then tested for its ability to lyse the NK cell target YAC-1 in a europium release assay (1, 17).

Immunolabeling studies

Cytofluorometric analysis.

EDTA-detached aliquots of M-3D or M-3D-IL-2 clones were reacted with either relevant (Table I⇓) or isotype-matched irrelevant mAb using either direct or indirect FITC-immunolabeling procedures. PLC were incubated in 5% (v/v) mouse serum (Sigma) to block nonspecific binding and stained with FITC-labeled mAb against various leukocyte differentiation Ags (Table I⇓). Fluorescence parameters of 10,000 living cells were measured using a FACScan (Becton Dickinson, Mountain View, CA) flow cytometer and analyzed with the “Lysys II” software (Becton Dickinson).

View this table:
Table I.

Abs and reagents used in this study

Immunohistologic analysis of inoculation sites.

M-3D or M-3D-IL-2 cells (3 × 105) were inoculated s.c. into DBA/2 mice. Inoculation sites and the surrounding tissue were excised at 2, 4, and 24 h, and 2, 4, 6, 8, and 11 days after injection of the cells. Tissue specimens were snap frozen in Tissue-Tec II (Miles Laboratory, Elkhart, IN) and stored at −20°C. Cryostat sections thereof (5 μm) were acetone fixed, air dried, and processed for immunohistologic single stainings, essentially as described (1). Sections developed in 3-amino-9-ethylcarbazole (Sigma) were counterstained with Harris’ hematoxylin (Merck, West Point, PA). Numbers of positive cells/mm2 of tumor area were evaluated using a semiautomatic image analysis system (Vidas, Elektronik GmbH, Eching, Germany).

Proliferation assays

Cells of the different M-3D cell lines/clones (1 × 104) were incubated in gelatin-coated flat-bottom 96-well plates for various periods of time. During the last 16 h of culture, [3H]thymidine was added to the individual culture wells (37 × 103 Bq ≡ 1 μCi/well). Incorporated radioactivity was measured by liquid scintillation spectroscopy and is expressed as cpm.

In vivo Ab depletion

A rabbit anti-asialo-GM1 (18, 19, 20) Ig preparation (Wako Chemicals, Neuss, Germany) was used to deplete BALB/c nu/nu mice of NK cells. For this purpose, mice were injected i.p. on days −1, +1, 6, 11, 16, 21, 26, and 31 with 200 μl of a 1:10 PBS dilution of the reconstituted anti-asialo-GM1 preparation. The animals were injected with M-3D-IL-2 on day 0 and subsequently monitored for tumor growth. Preliminary experiments had shown that the anti-asialo-GM1 treatment of the animals led to a >90% reduction of capacity of their spleen cells to lyse the NK cell target YAC-1 as compared with spleen cells of nontreated mice or mice injected with a dialyzed 1:10 dilution of normal rabbit serum (ICN Pharmaceuticals, Costa Mesa, CA) in PBS.

Statistical analysis

The Mann-Whitney U test was used for statistical analysis of data. A p value ≤0.05 was considered to be significant.

Results

Characterization of the parental M-3D line and IL-2-producing M-3D clones

To investigate the mechanisms underlying the lack of tumorigenicity of IL-2-producing M-3 cells, we generated a series of M-3D-IL-2 transfectants. The individual clones obtained exhibited different IL-2 production rates (Table II⇓). One low producer (clone 2/5) and one high producer (clone 4/3) were selected and used for most of the experiments.

View this table:
Table II.

IL-2 production rate by selected M-3DmIL-2 clones

Morphologic and phenotypic comparison of the transfected M-3D clones and the M-3D wild-type cells did not reveal any notable differences. All cells were polygonal in shape, adhered well to gelatin-coated culture flasks, expressed low levels of MHC class I (Fig. 1⇓) but no class II molecules, and did not react with Abs against costimulatory molecules CD80 (B7-1) and CD86 (B7-2). All M-3D clones exhibited low levels of CD44 (Pgp-1) and, with the exception of clone 4/3, moderate levels of ICAM-1 (CD54) (Fig. 1⇓). Treatment of cells with IFN-γ (200 U/ml; 24 h) led to a slight up-regulation of class I but could not induce the expression of class II Ags. [3H]Thymidine incorporation assays showed that the various IL-2-transfected M-3D clones, as well as a neomycin resistance gene-transfected M-3D clone, display in vitro growth characteristics similar to the one of the parental M-3D cell line (Fig. 2⇓A).

FIGURE 1.
FIGURE 1.

FACS analysis of M-3D cells and selected IL-2 gene transfectants. In vitro-cultured, EDTA-detached M-3D and clone 1/2, 2/5, and 4/3 cells were incubated with FITC-labeled Abs directed against MHC class I (H-2Dd) or CD54 (ICAM-1) and analyzed by FACS. Clone 1/2, 2/5, and 4/3 cells expressed class I at a level similar to that of cells of the parental line. While 4/3 cells did not react with anti-CD54 Abs, clones 1/2 and 2/5 and the M-3D cells express significant and similar amounts of this molecule.

FIGURE 2.
FIGURE 2.

In vitro and in vivo growth characteristics of parental or mock- or IL-2 gene-transfected M-3D cells. A, Cells (104) of the various IL-2-transfected clones, the neomycin-transfected clone, and the parental M-3D line were cultured in 96-well flat-bottom plates, and their proliferation rates were assessed at the time points indicated in the figure. B, Groups of DBA/2 mice (n = 5) were s.c. implanted with 1 × 105 parental or neo- and IL-2 gene-transfected M-3D cells and thereafter monitored for the appearance and growth of melanomas. Exponentially growing tumors were seen in animals injected with M-3D, M-3D-neo, and 2/5 cells. C, Cells (1 × 105) of the clones 2/5 and 4/3 were s.c. inoculated into BALB/c nu/nu animals. D, DBA/2 mice received graded doses (1 × 104, 1 × 105, 5 × 105, 1 × 106, and 2 × 106) of either M-3D or clone 4/3 cells and were then monitored for tumor formation.

Tumorigenicity of non-IL-2-producing and IL-2-producing M-3D cells

In a first set of experiments, we s.c. injected wild-type M-3D cells and neomycin resistance gene-transfected and various IL-2-producing M-3D clones at a dose of 1 × 105 cells into different groups of mice. In euthymic mice, only one (clone 2/5) of the IL-2-producing M-3D clones developed into neoplastic tumor nodules, whereas wild-type M-3D and neomycin resistance gene-transfected M-3D cells formed exponentially growing cancers (Fig. 2⇑B). In a similar fashion, the low (2/5) but not the high (4/3) IL-2-producing clone was tumorigenic in athymic animals (Fig. 2⇑C). Injection of animals with titrated doses of IL-2-producing vs non-IL-2-producing tumor cells showed that all mice inoculated with wild-type M-3D cells, but none of the 4/3 recipients, developed melanomas (Fig. 2⇑D).

Similar to what we observed with IL-2-transferrinfected M-3 cells (1), these experiments show that IL-2 expression by M-3D clones reduces their tumorigenic potential in a dose-dependent fashion and that T cells are not needed for this phenomenon to occur.

Histologic examination of inoculation sites

As revealed by hematoxylin and eosin staining of M-3D and clone 2/5 inoculation sites (Fig. 3⇓A), melanoma cells began to form aggregates during the first 2 h after injection. After 12 h, they had enlarged to compact nests of seemingly viable tumor cells. Their density reached a plateau 6 days after injection (Fig. 3⇓F), but their growth rate remained almost unchanged during the subsequent weeks, resulting in ≈2 × 2 × 2 cm tumors after 40–50 days. During the first 3 days, the dermis overlying the tumor, the tumor margins and, to a lesser extent, the tumor nests themselves were infiltrated by small numbers of leukocytes. They consisted mainly of RB6-8C5+ granulocytes and F4/80+ macrophages but were virtually devoid of T cells and NK cells. When assessed on day 6 after injection, this leukocytic tumor infiltrate had already disappeared.

FIGURE 3.
FIGURE 3.

Inoculation sites of IL-2-producing M-3D cells are heavily infiltrated by leukocytes. Inoculation sites were biopsied 2 days after the injection of either 3 × 105 clone 2/5 (A) or clone 4/3 cells (B), and sections thereof were stained using the hematoxylin and eosin method. While only a few leukocytes (arrowheads) were present at the inoculation sites of 2/5 cells, inocula of 4/3 cells, as well as the adjacent dermis, were heavily infiltrated by leukocytes. The leukocytes infiltrating into 4/3 inocula were immunophenotyped using Abs directed against CD45 (C), F4/80 cells (D), and NK cells (LGL-1) (E). Magnification: ×120 (A—D); ×360 (E).

Histopathologic analysis of skin sites inoculated with the IL-2-producing clone 4/3 yielded thoroughly different results. Similarly to wild-type and low dose IL-2-producing M-3 cells, 4/3 cells began to form aggregates during the first few hours after injection. Their further expansion was hampered by the early appearance of leukocytes that seemed to originate from dilated dermal vessels above the tumor and, after a rapid numerical increase, soon extended down into the melanoma cell nests. Tumor cells neighboring these leukocytes exhibited signs of injury and occasionally transformed into globular eosinophilic cellular remnants without appreciable nuclei. As a consequence, 4/3 cells began to numerically decline 1 day after inoculation and were no longer detectable within the massive leukocytic infiltrate after an additional 4 days. During the entire observation period, the composition of the infiltrate was similar to that seen with wild-type and low dose IL-2-producing cells; i.e., it was dominated by macrophages (Fig. 3⇑D) and granulocytes. An important difference was the additional presence of a distinct cellular subset reacting with the NK cell-specific Ab LGL-1 (detectable by days 2–3; Fig. 3⇑E) and of a few CD4+ and CD8+ T cells appearing at day 6 after inoculation. Concerning both the quality and the quantity of the inflammatory infiltrate, 4/3 injection sites were indistinguishable from sites injected with other high IL-2-secreting clones (e.g., 1/2, 1/5, and 4/4; data not shown).

IL-2-producing M-3 cells evoke a tumoricidal host response mediated by radiation-sensitive leukocytes

The inverse correlation between the number of 4/3 cells and the infiltrating leukocytes strongly suggested that the latter are involved in the processes that lead to the reduced tumorigenicity of IL-2-secreting melanoma cells. This was further supported by the findings that 1) animals that had been injected with mixtures of 1 × 105 or 2 × 105 M-3D and 1 × 105 clone 4/3 cells did not show tumor formation at the injection site and 2) three of four mice that had been inoculated with 3 × 105 M-3D and 1 × 105 4/3 cells developed melanomas that appeared later and grew slower than those in mice receiving 1 × 105 M-3D cells alone (Fig. 4⇓).

FIGURE 4.
FIGURE 4.

4/3 cells inhibit the in vivo growth of coadminstered wild-type M-3D cells. DBA/2 mice (five per group) were s.c. injected with 1 × 105 M-3D cells, 1 × 105 clone 4/3 cells, or mixtures of M-3D and 4/3 cells. The compositions of these mixtures are indicated in the figure. Thereafter, melanoma growth was monitored regularly by measuring the tumor diameters. Tumor sizes of animals receiving M-3D cells differed significantly from those of mice injected with 4/3 cells (p < 0.001) or the cell mixtures (M-3D:4/3 = 1:1, p < 0.001; M-3D:4/3 = 2:1, p < 0.05; M-3D:4/3 = 3:1, p < 0.05).

We next injected high IL-2-producing M-3D clones (1/2, 1/5, 4/1, 4/3, and 4/4) into either nonirradiated or 5 Gy-irradiated animals. As expected, these clones did not induce the growth of melanomas in nonirradiated mice (Fig. 5⇓A) but led to tumor formation in 5 Gy-irradiated animals (Fig. 5⇓B). After initial growth, these nodules regressed once the immune system had recovered, with the exception of nodules induced by clone 4/3, which continued to grow in most instances (four of seven experiments; Fig. 5⇓B). Tumor formation was not observed when 4/3 cells were injected in 3 or 4 Gy-irradiated recipients either alone or in combination with 1 × 105 wild-type M-3D cells (Fig. 5⇓C).

FIGURE 5.
FIGURE 5.

Effect of immunosuppression of recipient DBA/2 mice on the ability of IL-2-transfected M-3D cells to form melanomas. Cells (1 × 105) of the various clones were s.c. implanted into either nonirradiated (A) or sublethally irradiated (5 Gy; B) mice (five per group). These were inspected regularly, and tumor growth was assessed and recorded. C, Mixtures of clone 4/3 (1 × 105) and M-3D (1 × 105) cells were s.c. inoculated into DBA/2 mice (five per group) previously irradiated with 3, 4, or 5 Gy. Nonirradiated control mice received either M-3D cells or clone 4/3 cells alone or a mixture of M-3D cells and clone 4/3 cells.

NK cells critically contribute to the tumoricidal leukocyte response evoked by IL-2-producing M-3D cells

To identify the leukocytic population(s) that suppress(es) the in vivo growth of IL-2-producing M-3D cells, 4/3 cells were injected together with PLC into sublethally irradiated (5 Gy) animals. As determined by flow cytometric analysis, PLC consist of ∼30–50% macrophages, 40–60% B cells, 5–7% T cells, 1% granulocytes, and ∼1% NK cells. Results obtained showed that, while the s.c. administration of 4/3 cells alone regularly led to tumor growth, the coadministration of PLC prevented the occurrence of this event (Fig. 6⇓B). In sharp contrast, the coadministration of PLC did not interfere with the tumorigenicity of nontransfected M-3D cells (Fig. 6⇓A). Even though macrophages compose the largest portion of the infiltrate, we made the surprising observation that PLC that had been depleted of macrophages by plastic adherence were as effective as unfractionated PLC in inhibiting 4/3 cell growth (Fig. 6⇓D). However, when we subjected PLC to a treatment with the NK cell-reactive anti-asialo-GM1 serum and complement, they lost their ability to suppress the growth of 4/3 cells (Fig. 6⇓D).

FIGURE 6.
FIGURE 6.

PLC inhibit in vivo growth of coadministered 4/3 but not M-3D cells in irradiated hosts. NK cells represent the tumoricidal effector cell population. Shown are two independent experiments; experiment 1 is shown in A and B, and experiment 2 is shown in C and D. A, M-3D cells (1 × 105) were injected into either nonirradiated or irradiated (5 Gy) animals. Two other groups of irradiated animals were inoculated with mixtures consisting of M-3D and PLC (2 × 106) or macrophage-depleted PLC (2 × 106). B, In a parallel experiment, clone 4/3 cells were used instead of the M-3D cells. The groups consisted of five mice each. With untreated PLC, 48% of cells were F4/80 positive; with macrophage-depleted PLC (plastic adherence), ∼1% were F4/80 positive. C, Groups of nonirradiated mice received 1 × 105 M-3D or 4/3 cells while 5 Gy-irradiated DBA/2 mice were s.c. injected with a cell mixture consisting of 1 × 105 4/3 cells and 1 × 105 M-3D cells. D, Groups of 5 Gy-irradiated DBA/2 mice s.c. injected with mixtures consisting of 1 × 105 4/3 cells and 2 × 106 unfractionated or macrophage- or NK cell-depleted PLC. Control experiments had shown that compared with untreated PLC, macrophage-depleted PLC exhibited a more pronounced NK cell activity; by contrast, anti-asialo-GM1 + complement-treated PLC failed to lyse the NK cell target YAC-1.

In a last set of experiments, clone 4/3 cells were s.c. inoculated into nonirradiated BALB/c nu/nu mice treated with the NK cell-reactive asialo-GM1 antiserum. Nude, athymic animals were chosen to exclude a contribution of T cells. Results obtained showed that the s.c. inoculation of 1 × 105 4/3 cells led to the formation of slowly growing melanomas in the NK cell-depleted animals but not in mice treated with normal rabbit serum (Fig. 7⇓). Interestingly, these tumors started to regress when the injection of the asialo-GM1 antiserum was stopped. (Cells, which had been isolated from these tumors and cultured in the presence of G-418, produced IL-2 at a level comparable to that of 4/3 cells not passaged through an animal (data not shown)). These data demonstrate that NK cells, although representing only a minor fraction of the leukocytes infiltrating IL-2-secreting M-3D inocula, are critical for their destruction.

FIGURE 7.
FIGURE 7.

Treatment of recipient BALB/C nu/nu mice with anti-asialo-GM1 restores the tumorigenicity of IL-2-producing M-3D cells. Groups of BALB/C nu/nu mice were repeatedly injected (arrows) either with asialo-GM1 antiserum or, for control purposes, with normal rabbit serum. One day after the first Ab application, mice were implanted with 1 × 105 4/3 cells. Thereafter, they were monitored for the appearance of melanomas, and tumor growth was recorded. Tumor sizes of mice that had received 4/3 cells and asialo-GM1 antiserum differed significantly from the ones that had been injected with 4/3 cells and normal rabbit serum (p = 0.0008).

Discussion

This investigation was undertaken to determine why certain cancer cells lose their tumorigenicity upon IL-2 gene transfection. In accordance with our initial studies, in which we had used adenovirus-enhanced transferrinfection (1) to transiently express the IL-2 gene in M-3 cells, we could now demonstrate that while the s.c. injection of wild-type M-3D cells into syngeneic DBA/2 mice regularly results in tumor growth, M-3D cells stably transfected with the murine IL-2 gene fail to do so. Similar to what was reported by Cavallo et al. (4), one can assume that an IL-2-induced inflammatory reaction is responsible for this reduced tumorigenicity for the following reasons. 1) There exists a direct correlation between IL-2 production rate of the tumor cells and the intensity of the inflammatory infiltrate and, conversely, an indirect relationship between these two parameters and the tumor growth rate in vivo. 2) When injected into immunosuppressed mice, high IL-2-secreting M-3D cells give rise to cancers that begin to regress upon recovery of the host’s immune system.

A first clue concerning the nature of the critical tumoricidal cell population involved came from the immunohistologic comparison of the injection sites of the wild-type and IL-2-secreting M-3 cells. While, in both instances, the inflammatory reaction was dominated by macrophages and granulocytes, the reaction induced by the M-3D-IL-2 cells differed from the one triggered by the M-3D cells not only by its magnitude, but also by the additional presence of NK cells. The role of NK cells as the major effector cell population gained further support from the following experimental results. First, the influx of NK cells precedes the cessation of growth and, finally, demise of IL-2-expressing M-3 melanoma cells. Second, coinjection of NK-depleted, but not of macrophage-depleted, PLC and high-dose IL-2-secreting M-3D cells in immunosuppressed DBA/2 mice results in tumor formation. Third, high doses of IL-2 activate unfractionated PLC, but not NK cell-depleted PLC or granulocytes (either isolated from peripheral blood or generated in vitro from bone marrow cells), to lyse M-3D cells in vitro (A.S. and R.K., unpublished observations). Fourth, the s.c. inoculation of 4/3 cells induced melanomas in nude mice treated with the NK cell-depleting anti-asialo-GM1 serum; these tumors began to regress when the Ab injections were stopped. Finally, we found that neither animals that had been injected with the granulocyte-depleting mAb (RB6-8C5) nor mice that had been treated with the macrophage-inactivating agent silica showed tumor formation upon the s.c. injection of 4/3 melanoma cells (F.K. and A.S., unpublished observations). Together, these findings demonstrate that NK cells are necessary for the tumoricidal host response to occur. Similar results were obtained by some (3, 21, 22) although not all (4, 23) researchers investigating the reduced tumorigenicity of IL-2-transfected cancer cells. More recent reports indicated that NK cells are not only necessary and sufficient for the elimination of IL-2-transfected cancer cells, but also of CD80 (B7-1)-transfected cancer cells (24, 25, 26, 27). It appears that B7-1 acts on NK cells by stimulating their proliferative capacity and lytic effector function via an as yet unknown counterreceptor (28, 29).

Although our findings point to a critical role of NK cells, they do not exclude a significant contribution of other cell types during IL-2-induced tumor cell destruction. In this regard, NK cells could act not only by destroying cancer cells themselves, but also by stimulating the tumoricidal activity of macrophages and granulocytes. Tumor-specific T cells, raised during the initial tumoricidal processes (see below), could represent another synergistically acting effector cell population. This notion is based on our observation that mice that had received inocula of M-3D cells secreting between 1500 and 2000 U/IL-2/24 h (e.g., 2 × 105 1/2 cells or 1.5 × 105 4/3 cells) are protected against a subsequent challenge with wild-type M-3D cells (A. Schneeberger, unpublished observation). In all of these possible scenarios, the NK cell would represent the critical component of the tumoricidal host response.

Concerning the mechanism by which IL-2-secreting M-3D cells attract and activate NK cells, it is conceivable that the s.c. implantation of cancer cells secreting large amounts of IL-2 leads to a localized vascular leak syndrome similar to what has been described as a systemic reaction in high dose IL-2 recipients (30). Besides that, IL-2 can act as a direct chemoattractant not only for eosinophils (31) and activated human T cells (32), but also for large granular lymphocytes, which include activated NK cells (33). Alternatively, the tumor cell-derived IL-2 could activate IL-2 receptor-expressing resident and/or immigrating leukocytes (e.g., mononuclear phagocytes or granulocytes) to secrete NK cell chemoattractants such as RANTES, MIP1α, MIP1β, MCP-1, MCP-2, MCP-3, and IP-10 (34). In fact, expression of the latter has been demonstrated after intradermal application of human rIL-2 into normal-appearing skin of leprosy patients (35).

The factors responsible for the activation of the attracted NK cells have yet to be clarified. IL-2 itself may be important in this regard (36), as could be the chemokines RANTES, MIP1α, MIP1β, MCP-1, MCP-2, MCP-3, and IP-10; (34), the cytokines IFN-γ, IL-12, IL-15, and TNF-α, (37, 38, 39, 40); and the costimulatory molecule CD80 (B7-1) (24, 25, 26, 27, 28, 29) expressed on dendritic cells (DC) and activated macrophages.

The findings reported here have important implications for our understanding of the mechanism(s) by which IL-2-secreting cancer cells induce protective T cell responses. Our data showing that the destruction of the cancer cell depot is essentially completed before the influx of T lymphocytes argues against a direct functional M-3D-IL-2-T cell interaction. Instead, they support the indirect Ag presentation model according to which tumor Ag-bearing, professional APCs initiate the tumor cell-specific T cell response. DCs, but not macrophages, undergo a maturation process upon receipt of activation signals (e.g., ingestion of apoptotic cells (41)), migrate from nonlymphoid to peripheral lymphoid tissues where they stimulate naive T cells and, thus, are most likely involved in this process.

We have preliminary evidence that DC infiltrate M-3D-IL-2 injection sites. Clearly, their numbers do not correlate with the IL-2 production rate of the vaccine. In this regard, it should be remembered that cancer cells secreting medium amounts of IL-2 are often much more efficacious than vaccines secreting high doses of IL-2 (2, 4, 42). It may well be that maximum attraction and activation of NK cells by a high dose IL-2 vaccine leads to such a rapid tumor cell destruction that the resulting tumor cell fragments are scavenged and digested by macrophages before the advent of DC at the injection site.

Comparative assessments of time of appearance, phenotype, and function of both NK cells and DC are needed to define the conditions for an optimal cooperation between the innate and adaptive immune system triggered by IL-2-based cancer vaccines.

Acknowledgments

We thank Drs. F. Melchers and H. Karasuyama for providing the IL-2-encoding plasmid BMGneoIL-2 and Laura A. Stingl for critical review of the manuscript.

Footnotes

  • ↵1 This work was supported by Grant 6/676 from the Industrial Research promotion fund, Vienna, Austria.

  • ↵2 A.S. and F.K. contributed equally to this work.

  • ↵3 Address correspondence and reprint requests to Dr. Georg Stingl, Department of Dermatology, Division of Immunology, Allergy and Infectious Diseases, University of Vienna Medical School, A-1090 Vienna, Austria. E-mail address: georg.stingl{at}akh-wien.ac.at

  • ↵4 Abbreviations used in this paper: M-3D-IL-2, IL-2-transfected M-3D melanoma; PLC, peritoneal lavage cell; DC, dendritic cell.

  • Received November 11, 1998.
  • Accepted March 16, 1999.

References

  1. ↵
    Zatloukal, K., A. Schneeberger, M. Berger, W. Schmidt, F. Koszik, R. Kutil, M. Cotten, E. Wagner, M. Buschle, G. Maass, et al 1995. Elicitation of a systemic and protective anti-melanoma immune response by an IL-2-based vaccine: assessment of critical cellular and molecular parameters. J. Immunol. 154: 3406
    OpenUrlAbstract
  2. ↵
    Schmidt, W., T. Schweighoffer, E. Herbst, G. Maass, M. Berger, F. Schilcher, G. Schaeffner, M. L. Birnstiel. 1995. Cancer vaccines: the interleukin-2 dosage effect. Proc. Natl. Acad. Sci. USA 92: 4711
    OpenUrlAbstract/FREE Full Text
  3. ↵
    Bannerji, R., C. D. Arroyo, C. Cordon-Cardo, E. Gilboa. 1994. The role of IL-2 secreted from genetically modified tumor cells in the establishment of antitumor immunity. J. Immunol. 152: 2324
    OpenUrlAbstract
  4. ↵
    Cavallo, F., M. Giovarelli, A. Gulino, A. Vacca, A. Stoppacciaro, A. Modesti, G. Forni. 1992. Role of neutrophils and CD4+ T lymphocytes in the primary and memory response to nonimmunogenic murine mammary adenocarcinoma made immunogenic by IL-2 gene. J. Immunol. 149: 3627
    OpenUrlAbstract
  5. ↵
    Fearon, E. R., D. M. Pardoll, T. Itaya, P. Golumbek, H. L. Levitsky, J. W. Simons, H. Karasuyama, B. Vogelstein, P. Frost. 1990. Interleukin-2 production by tumor cells bypasses T helper function in the generation of an antitumor response. Cell 60: 397
    OpenUrlCrossRefPubMed
  6. ↵
    Gansbacher, B., K. Zier, B. Daniels, K. Cronin, R. Bannerji, E. Gilboa. 1990. Interleukin 2 gene transfer into tumor cells abrogates tumorigenicity and induces protective immunity. J. Exp. Med. 172: 1217
    OpenUrlAbstract/FREE Full Text
  7. ↵
    Pardoll, D.. 1992. New strategies for active immunotherapy with genetically engineered tumor cells. Curr. Opin. Immunol. 4: 619
    OpenUrlCrossRefPubMed
  8. ↵
    Schneeberger, A., F. Koszik, G. Stingl. 1995. Immunologic host defense in melanoma: delineation of effector mechanisms involved and of strategies for the augmentation of their efficacy. J. Invest. Dermatol. 105: 110
    OpenUrl
  9. ↵
    Zatloukal, K., W. Schmidt, M. Cotten, E. Wagner, G. Stingl, M. L. Birnstiel. 1993. Somatic gene therapy for cancer: the utility of transferrinfection in generating tumor vaccines. Gene 135: 199
    OpenUrlCrossRefPubMed
  10. ↵
    Cavallo, F., F. Di Pierro, M. Giovarelli, A. Gulino, A. Vacca, A. Stoppacciaro, M. Forni, A. Modesti, G. Forni. 1993. Protective and curative potential of vaccination with interleukin-2 gene-transfected cells from a spontaneous mouse mammary adenocarcinoma. Cancer Res. 53: 5067
    OpenUrlAbstract/FREE Full Text
  11. ↵
    Maass, G., W. Schmidt, M. Berger, F. Schilcher, F. Koszik, A. Schneeberger, G. Stingl, M. L. Birnstiel, T. Schweighoffer. 1995. Priming of tumor-specific T cells in the draining lymph nodes after immunization with IL-2-secreting tumor cells: three consecutive stages may be required for successful tumor vaccination. Proc. Natl. Acad. Sci. USA 92: 5540
    OpenUrlAbstract/FREE Full Text
  12. ↵
    Cloudman, A. M.. 1941. The effect of an extrachromosomal influence upon transplanted spontaneous tumors in mice. Science 93: 380
    OpenUrlFREE Full Text
  13. ↵
    Yasumura, Y., A. H. Tashjian, G. H. Sato. 1966. Establishment of four functional, clonal strains of animal cells in culture. Science 154: 1186
    OpenUrlAbstract/FREE Full Text
  14. ↵
    Gillis, S., M. M. Ferm, W. Ou, K. Smith. 1978. T cell growth factor: parameter of production and a quantitative microassay for activity. J. Immunol. 120: 2027
    OpenUrlAbstract/FREE Full Text
  15. ↵
    Bottomly, K., L. S. Davis, and P. E. Lipsky. 1991. Measurement of human and murine IL-2 and IL-4. In Current Protocols in Immunology. J. E. Coligan, A. M. Kruisbeck, D. H. Margulies, E. M. Shevach, and W. Strober, eds. Wiley & Sons, New York, p. 6.3.
  16. ↵
    Winn, H. J.. 1961. Immune mechanisms in homotransplantation. II. Quantitative assay of the immunologic activity of lymphoid cells stimulated by tumor homografts. J. Immunol. 86: 228
  17. ↵
    Blomberg, K., C. Granberg, I. Hemmilä, T. Lövgren. 1986. Europium-labelled target cells in an assay of natural killer cell activity. I. A novel non-reactive method based on time-resolved fluorescence. J. Immunol. Methods 86: 225
    OpenUrlCrossRefPubMed
  18. ↵
    Stein, K. E., G. A. Schwarting, D. M. Marcus. 1978. Glycolipid markers of murine lymphocyte subpopulations. J. Immunol. 120: 676
    OpenUrlAbstract/FREE Full Text
  19. ↵
    Habu, S., H. Fukui, K. Shimamura, M. Kasai, Y. Nagai, K. Okumura, N. Tamaoki. 1981. In vivo effects of anti-asialo-GM1: reduction of NK activity and enhancement of transplanted tumor growth in nude mice. J. Immunol. 127: 34
    OpenUrlAbstract/FREE Full Text
  20. ↵
    Kasai, M., T. Yoneda, S. Habu, Y. Maruyama, K. Okumura, T. Tokunaga. 1981. In vivo effect of anti-asialo GM1 antibody on natural killer cell activity. Nature 291: 334
    OpenUrlCrossRefPubMed
  21. ↵
    Bubenik, J., N. N. Voitenok, J. Kieler, V. S. Prassolow, P. M. Chumakow, D. Bubenikova, J. Simova, T. Jandlova. 1988. Local administration of cells containing an inserted IL-2 gene and producing IL-2 inhibits growth of human tumours in nu/nu mice. Immunol. Lett. 19: 279
    OpenUrlCrossRefPubMed
  22. ↵
    Karp, S. E. A., A. Farber, J. C. Salo, P. Hwu, G. Jaffe, A. L. Asher, E. Shiloni, N. P. Restifo, J. J. Mulé, S. A. Rosenberg. 1993. Cytokine secretion by genetically modified nonimmunogenic murine fibrosarcoma: tumor inhibition by IL-2 but not tumor necrosis factor. J. Immunol. 150: 896
    OpenUrlAbstract
  23. ↵
    Colombo, M. P., A. Modesti, G. Parmiani, G. Forni. 1992. Local cytokine availability elicits tumor rejection and systemic immunity through granulocyte-T-lymphocyte cross-talk. Cancer Res. 52: 4853
    OpenUrlFREE Full Text
  24. ↵
    Cavallo, F., A. Martin-Fontecha, M. Bellone, S. Heltai, E. Gatti, P. Tornaghi, M. Freschi, G. Forni, P. Dellabona, G. Casorati. 1995. Co-expression of B7-1 and ICAM-1 on tumors is required for rejection and the establishment of a memory response. Eur. J. Immunol. 25: 1154
    OpenUrlCrossRefPubMed
  25. ↵
    Wu, T.-C., A. Y. C. Huang, E. M. Jaffee, H. I. Levitsky, D. M. Pardoll. 1995. A reassessment of the role of B7-1 expression in tumor rejection. J. Exp. Med. 182: 1415
    OpenUrlAbstract/FREE Full Text
  26. ↵
    Geldhof, A. B., G. Raes, M. Bakkus, S. Devos, K. Thielemans, P. De Baetselier. 1995. Expression of B7-1 by highly metastatic mouse lymphomas induces optimal natural killer cell-mediated cytotoxicity. Cancer Res. 55: 2730
    OpenUrlAbstract/FREE Full Text
  27. ↵
    Yeh, K.-Y., B. A. Pulaski, M. L. Woods, A. J. McAdam, A. A. Gaspari, J. G. Frelinger, E. M. Lord. 1995. B7-1 enhances natural killer cell-mediated cytotoxicity and inhibits tumor growth of a poorly immunogenic murine carcinoma. Cell. Immunol. 165: 217
    OpenUrlCrossRefPubMed
  28. ↵
    Nandi, D., J. A. Gross, J. P. Allison. 1994. CD28-mediated costimulation is necessary for optimal proliferation of murine NK cells. J. Immunol. 152: 3361
    OpenUrlAbstract/FREE Full Text
  29. ↵
    Chambers, B. J., M. Salcedo, H.-G. Ljunggren. 1996. Triggering of natural killer cells by the costimulatory molecule CD80 (B7-1). Immunity 5: 311
    OpenUrlCrossRefPubMed
  30. ↵
    Cotran, R. S., J. S. Pober, M. A. Gimbrone, T. A. Springer, E. A. Wiebke, A. A. Gaspari, S. A. Rosenberg, M. T. Lotze. 1987. Endothelial activation during interleukin 2 immunotherapy: a possible mechanism for the vascular leak syndrome. J. Immunol. 139: 1883
    OpenUrl
  31. ↵
    Rand, T. H., D. S. Silberstein, H. Kornfeld, P. F. Weller. 1991. Human eosinophils express functional interleukin 2 receptors. J. Clin. Invest. 88: 825
  32. ↵
    Kornfeld, H., J. S. Bergman, D. J. Beer, D. M. Center. 1985. Induction of human T lymphocyte motility by interleukin 2. J. Immunol. 134: 3887
    OpenUrlAbstract/FREE Full Text
  33. ↵
    Natuk, R. J., R. M. Welsh. 1987. Chemotactic effect of human recombinant interleukin 2 on mouse activated large granular lymphocytes. J. Immunol. 139: 2737
    OpenUrlAbstract/FREE Full Text
  34. ↵
    Hedrik, J. A., A. Zlotnik. 1996. Chemokines and lymphocyte biology. Curr. Opin. Immunol. 8: 343
    OpenUrlCrossRefPubMed
  35. ↵
    Kaplan, G., R. Kiessling, S. Teklemariam, G. Hancock, G. Sheftel, K. B. Job, P. Converse, T. H. M. Ottenhoff, M. Becx-Bleumink, M. Dietz, Z. A. Cohn. 1989. The reconstitution of cell-mediated immunity in the cutaneous lesions of lepromatous leprosy by recombinant interleukin 2. J. Exp. Med. 169: 893
    OpenUrlAbstract/FREE Full Text
  36. ↵
    Nagler, A., L. L. Lanier, J. H. Phillips. 1990. Constitutive expression of high affinity interleukin 2 receptors on human CD16 natural killer cells in vivo. J. Exp. Med. 171: 1527
    OpenUrlAbstract/FREE Full Text
  37. ↵
    Gidlund, M., A. Örn, H. Wigzell, A. Senik, I. Gresser. 1978. Enhanced NK cell activity in mice injected with interferon and interferon inducers. Nature 273: 759
    OpenUrlCrossRefPubMed
  38. ↵
    Yokoyama, W. M.. 1995. Right-side-up and up-side-down NK cell receptors. Curr. Biol. 5: 982
    OpenUrlCrossRefPubMed
  39. ↵
    Trinchieri, G.. 1989. Biology of natural killer cells. Adv. Immunol. 47: 187
    OpenUrlCrossRefPubMed
  40. ↵
    Biron, C. A.. 1997. Activation and function of natural killer cell responses during viral infections. Curr. Opin. Immunol. 9: 24
    OpenUrlCrossRefPubMed
  41. ↵
    Albert, M. L., B. Sauter, N. Bhardwaj. 1998. Dendritic cells acquire antigen from apoptotic cells and induce class I-restricted CTLs. Nature 329: 86
    OpenUrl
  42. ↵
    Fakhrai, H., D. L. Shawler, R. Gjerset, R. K. Naviaux, J. Koziol, J. Royston, R. E. Sobol. 1995. Cytokine gene therapy with IL-2-transduced fibroblasts: effects of IL-2 dose on anti-tumor immunity. Hum. Gene Ther. 6: 591
    OpenUrlCrossRefPubMed
  43. Mason, D. W., M. Dallman, A. N. Barclay. 1981. Graft-versus-host disease induces expression of Ia antigen in rat epidermal cells and gut epithelium. Nature 293: 150
    OpenUrlCrossRefPubMed
Back to top

In this issue

The Journal of Immunology: 162 (11)
The Journal of Immunology
Vol. 162, Issue 11
1 Jun 1999
Print
Download PDF
Article Alerts
Email Article
Citation Tools

Jump to section