SR/CR (spontaneous regression/complete resistance) mice resist multiple types of cancer cells injected at numbers that are lethal to wild type (WT) mice. When the anti-tumor response was examined, leukocytes of the innate immune system, including neutrophils (PMN), macrophages and NK cells, infiltrated the tumor site for a multipronged killing response. Each cell type had independent killing activity against the cancer cells. A second aspect of this multipronged response was that cancer cells could be killed either via necrosis in vivo or via apoptosis by purified macrophages. Lymphoid cells displayed perforin (pfp) and granzymes (gzm) as effector molecules, but macrophages produced reactive oxygen species (ROS) and secreted serine proteases to kill the cancer cells. However, SR/CR macrophages did not use the well-studied tumoricidal mechanism of reactive nitrogen species (RNS) production. We previously demonstrated that macrophages tightly bound cancer cells in rosettes, and we show here that macrophages required contact with the target cells in order to unleash their cytotoxic mechanisms. Once SR/CR mice survived challenge with cancer cells, they produced antibodies that recognized the cancer cells. However, the antibodies were not required for killing by SR/CR macrophages through antibody-dependent cell-mediated cytotoxicity (ADCC) and did not enable wild type macrophages to kill target cells. In summary, purified SR/CR macrophages killed cancer cells in a non-ADCC manner via apoptosis induced by ROS and serine proteases.
This article was published in Cancer Immunity, a Cancer Research Institute journal that ceased publication in 2013 and is now provided online in association with Cancer Immunology Research.
SR/CR mice are a model of innate host resistance to multiple transplantable cancer cell lines. The colony was derived from a single BALB/c mouse that survived challenges with sarcoma 180 (S180) cells that were lethal to WT mice (1). The resistance trait was transmitted in a single-locus, dominant, autosomal manner into all mouse strains tested (BALB/c, CAST/Ei, C57BL/6). When SR/CR mice were challenged with cancer cells, leukocytes rapidly infiltrated the tumor site and bound cancer cells tightly to form rosettes. The target cells were rapidly destroyed by the infiltrating leukocytes. SR/CR mice responded to both primary and subsequent cancer challenges with a mixture of multiple subsets of effector leukocytes rather than just a single subset (2). This mixture is composed primarily of macrophages, PMN and NK cells, each having independent killing activity against the target cancer cells. The involvement of multiple subsets of leukocytes implies the likelihood of multiple killing mechanisms. However, the effector mechanisms employed by this model in the specific recognition and destruction of cancer cells have not been previously identified.
Several questions remained unanswered. First, what is the main mechanism of cancer cell death in SR/CR mice (necrosis vs. apoptosis)? Second, do all leukocyte subclasses kill S180 cells with the same mechanism, perhaps the result of a novel mutation, or is the unusual multipronged approach of multiple effector cell types reflected in multiple effector mechanisms? Third, what effector mechanisms are employed in this killing of cancer cells by SR/CR leukocytes? Although the killing mechanisms of this model have not been reported, previous work has focused on mechanisms employed by each of the innate leukocytes. First, the killing mechanisms of NK cells are well defined. NK cells produce and secrete the plasma membrane pore-forming and proteolytic effector molecules pfp (3, 4, 5, 6), gzmA and gzmB (7, 8, 9, 10, 11), resulting in rapid necrosis of the target cell. On the other hand, macrophages have also been reported to effectively target and kill cancer cells through apoptosis primarily through the production of RNS (12, 13, 14, 15, 16, 17) and ROS (18, 19, 20). We sought to answer the question whether SR/CR leukocytes possess a novel killing mechanism stemming from their mutation or whether they utilize known killing mechanisms.
In this study, we show that the primary mechanism of cell death of the S180 cells in young SR/CR mice was necrosis. Pfp, gzmA and gzmB, secreted by lymphoid cells, were detected during cancer cell death in SR/CR mice. However, we have also shown that macrophages constituted a large percentage of the cells infiltrating and binding to S180 cells in vivo (2). Also, macrophages from SR/CR mice protected WT recipient mice from cancer development through adoptive cell transfer. In addition, purified SR/CR macrophages killed cancer cells in vitro. We thus looked specifically at the mechanisms of macrophage-mediated cytotoxicity in vitro. Macrophages killed cancer cells primarily through apoptosis as a result of secretion of ROS and serine proteases. This cytotoxicity completely depended upon contact between effector and target cells and was abolished when contact was blocked. However, the presence of anti-cancer cell antibodies did not stimulate either SR/CR or WT macrophage cytotoxicity through ADCC. The observation that mixed leukocytes killed S180 cells through necrosis in vivo while purified macrophages killed S180 cells through apoptosis in vitro likely arose from suboptimal killing conditions in vitro. The time frame of killing was greatly slowed down in vitro, and the ratio of effector to target cells likely did not represent physiological conditions.
Induction of necrosis in cancer cells through pfp, gzmA and gzmB
First, we set out to identify the principal mechanism of cancer cell death in SR/CR mice. We challenged SR/CR mice with S180 cells and washed the infiltrating leukocytes. We then observed the peritoneal infiltrate with the remaining S180 cells by time-lapse video (Figure 1 and supplementary video clip 1). Leukocytes closely bound the S180 cells, resulting in swelling and lysis of the target cells. Such a mechanism of cell death was consistent with necrosis via pfp and gzmA and B. To determine if pfp and gzmA and B were involved in the cytolysis of S180 cells, we challenged SR/CR mice with 107 S180 cells. We washed the peritoneal cavities after 18 h or 24 h, when most of the S180 cells had ruptured, and analyzed the cell-free portions by Western blots. Pfp and gzmA and B were clearly detected in the extracellular fluid of SR/CR mice in a time-dependent manner, but not in WT mice which showed no lysis of S180 cells (Figure 2, panel A). The abnormal molecular weights of the proteins detected in the Western blots likely reflect protein modifications resulting from the processes of target cell killing and release into the extracellular debris. They are not necessarily the same forms as those secreted by lymphocytes.
We verified the production of pfp and gzmA and B by immunocytochemistry on cells washed from the peritoneal cavity 45 min after challenging SR/CR mice with 106 S180 cells. In frequently challenged, young SR/CR mice, the majority of transplanted cancer cells were ruptured in less than 1 h. Therefore, at 45 min rosettes remained to visualize the interaction between effector and target cells. Pfp and gzmA and B were detected in approximately 10% of the leukocytes in rosettes at this time point. Furthermore, all leukocytes positive for pfp and gzmA and B were consistent with lymphoid morphology (Figure 2, panels B-D). Polarization of these molecules toward the membrane contact sites between target and effector cells was also observed in a fraction of leukocytes (Figure 2, panels B-D). On the other hand, pfp and gzmA and B were not detected in macrophages by immunohistochemistry or by Western blot (data not shown).
Induction of cancer cell apoptosis by purified macrophages
The above results were consistent with lymphocytes utilizing pfp and gzmA and B as effector molecules. However, these results did not explain the effector mechanism of SR/CR macrophages. To determine the mechanism of cancer cell death induced by purified macrophages, we incubated SR/CR macrophages with S180 cells and recorded their appearance using phase contrast, time-lapse video microscopy. Upon incubation with purified macrophages, S180 cells underwent membrane blebbing followed by rupture of the cells, suggesting that the mechanism of cell death may have been apoptosis (Figure 3, panel A, and supplementary video clip 2). To determine if apoptotic pathways were activated in S180 cells, the S180 cells in the in vitro killing assays were stained for cleaved caspase 3. S180 cells that formed rosettes with macrophages were positive for cleaved caspase 3 (Figure 3, panel B). Together, these results indicate that macrophages induced S180 cell death primarily through apoptosis.
Effector mechanisms of SR/CR macrophages
We next determined if the known effector mechanisms of macrophages were involved in the killing of cancer cells, such as RNS production via nitric oxide synthase II (NOSII) and ROS production via the oxidase complex. First, we determined if incubation of macrophages with S180 cells would stimulate RNS production, a necessary event if this molecule was involved in the killing. Additionally, macrophages from SR/CR mice must have significantly higher RNS production than WT macrophages upon stimulation. Our findings indicated that: (i) S180 cells could not stimulate any significant amount of RNS production in either SR/CR or WT macrophages (Figure 4, panel A); (ii) Complete inhibition of RNS production via inhibition of NOSII by NMMA, 1400W and L-NIL did not inhibit the killing activity of macrophages (Figure 4, panels B-C); and (iii) Induction of RNS production by LPS neither enabled the WT macrophages to kill cancer cells, nor did it enhance the killing activity of SR/CR leukocytes. On the contrary, LPS inhibited the killing activity of SR/CR macrophages although it stimulated RNS production (Figure 4, panel D).
Macrophages utilize a second cytotoxic property by producing ROS through stimulation of the oxidase complex (18, 19, 20). We set out to determine if ROS were involved in the killing of cancer cells by SR/CR macrophages through several inhibition strategies. We inhibited ROS through depletion of glucose in the culture medium, diphenylene iodonium (DPI), catalase, superoxide dismutase (SOD) and Brewer thioglycolate medium (TG). All these inhibitory conditions were previously reported to be effective in blocking ROS in macrophages (21, 22). Our results showed that the in vitro killing activity of SR/CR macrophages was inhibited 23.5% ± 27.58% by glucose depletion (P = 0.44), 31.0% ± 7.07% by DPI (P = 0.025), 36.5% ± 0.71% by SOD (P = 0.0002), 42.0% ± 25.46% by catalase (P = 0.145) and up to 82.0% ± 8.40% by TG (P = 0.005) (Figure 5). The statistical significance of the difference in the killing of S180 cells compared to uninhibited control samples was determined by the Student’s t-test.
Macrophages also release a cytolytic factor in response to stimulation with cancer cells. This cytolytic factor acts synergistically with ROS and has been identified as a serine protease (23, 24). To determine the role of serine proteases in macrophage killing, we pre-incubated SR/CR macrophages with a non-specific trypsin inhibitor for serine proteases and found that at concentrations shown to block serine proteases, SR/CR macrophages cytotoxicity was inhibited up to 35.0% ± 4.24% (P = 0.005) (Figure 6). Thus, macrophages apparently employ multiple mechanisms to kill S180 cancer cells. This likely explains the finding that no single inhibitor completely blocks cytotoxicity.
Killing of cancer cells by macrophages requires direct cell-cell contact
We have previously demonstrated that, in the absence of humoral components, leukocytes alone were capable of killing cancer cells in vitro (1). This finding, together with the observation of rosette formation, highlights the question of whether cell-cell contact is required for killing. Since killing of S180 cells takes place effectively when these are mixed directly with SR/CR macrophages, we set out to determine if the killing occurred when target cells were placed in close proximity allowing free exchange of diffusible molecules but without direct physical contact. To do so, we placed the SR/CR or WT macrophages in one side of a Transwell plate and S180 cells in the other. We mixed identical aliquots of SR/CR macrophages and S180 cells and placed them in one chamber as positive controls for killing. All killing activities of SR/CR macrophages, as seen in the positive controls, were completely abolished when direct physical contact between target and effector cells was blocked (Figure 7). This finding indicates that physical contact between SR/CR macrophages and S180 cells is required for the killing of S180 cells in vitro.
Role of ADCC for SR/CR macrophages
The fact that SR/CR mice survived repeated challenges with S180 cells raised the question of the role that humoral immunity plays in the anti-cancer response. It was clear that the in vitro killing of cancer cells did not require the presence of humoral components (1). However, it was not clear whether humoral immunity enhanced the pre-existing cellular immunity via a mechanism such as ADCC. We therefore looked for the presence of Abs against S180 cells in SR/CR mice. The cell-free component of the peritoneal washes from S180-challenged SR/CR mice, unchallenged SR/CR mice and unchallenged WT mice were analyzed by immunocytochemistry for the presence of Abs that could coat the surface of the cancer cells. Peritoneal washes were used in lieu of mouse serum since 2D protein gels have shown that both fractions exhibit identical protein profiles, and washes are easier to obtain without sacrificing mice (25). Peritoneal washes from SR/CR mice repeatedly exposed to S180 cells contained strong reactivity to the surface of S180 cells (Figure 8, panel A). On the contrary, such reactivity was completely absent from peritoneal washes from WT or unchallenged SR/CR mice.
The exact role of anti-S180 Abs in cancer immunity was unknown. Several hypotheses could explain the presence of such Abs in the SR/CR mice. First, the Abs may be able to bind to target cells and induce lysis in the absence of effector leukocytes. Second, the Abs may have no independent activity in killing but may enable leukocytes from both WT and SR/CR mice to recognize and destroy target cells. Third, the Abs may not be sufficient for the induction of killing by WT leukocytes but may enhance killing by SR/CR leukocytes. A role for Abs in macrophage-mediated target cell killing would not be unexpected. Macrophages are known to recognize Ab-coated target cells and initiate cytotoxicity in a process known as ADCC. Fourth, the Abs may be a byproduct of surviving a challenge with S180 cells with little activity in S180 cell killing. To test these possibilities, we pre-labeled S180 cells with anti-S180 Abs from SR/CR mice. We then compared the survival of Ab-labeled S180 cells alone, in the presence of WT macrophages, and in the presence of SR/CR macrophages with the survival of unlabeled S180 cells for each treatment group. Ab-labeling caused no direct lysis of target cells and did not enhance lysis by either SR/CR or WT macrophages (Figure 8, panel B).
SR/CR mice efficiently and specifically killed transplanted mouse cancer cells that were otherwise lethal to WT mice (1). Cancer cells injected into the peritoneal cavities of SR/CR mice were killed at an approximate rate of 106 cells/h (data not shown). This anti-cancer response occurred without detectable side-effects, and the mice remained phenotypically normal. The leukocytes that infiltrated the cancer site included PMN, macrophages and NK cells, and each had independent killing activity against the target cells (2). Such a concerted, multipronged response of the immune system conferred an obvious advantage for immune compromised mice. Depletion of one or two subsets of these infiltrating leukocytes had no apparent effect on the ability of the mice to eradicate cancer cells, but depletion of all subsets rendered the SR/CR mice sensitive to cancer. The multipronged response was also reflected in the observation that cancer cells were killed via cytolysis by lymphocytic effector cells, such as NK cells, and via apoptosis by macrophages. Unlike the lymphocytic effector cells that primarily used pfp and gzmA and B as effector molecules, macrophages produced ROS and serine proteases to kill cancer cells. The previously-known tumoricidal mechanism of macrophages via production of RNS was not involved in the SR/CR killing activity. While all subsets of innate effector cells were active in resistance, humoral immunity appeared to be an inactive bystander. The anti-S180 Abs in the SR/CR mice displayed no activity in either lysing cancer cells directly or facilitating the killing activities of SR/CR macrophages.
Although the phenotype of cancer resistance in SR/CR mice is rare, the identified effector mechanisms employed by leukocytes in killing cancer cells have been previously studied. Lymphocytes are known to kill target cells by secreting cytolytic granules containing pfp and gzmA and B. Phagocytes, including macrophages and PMN, are known to produce RNS and ROS as cytotoxic molecules in killing cancer cells. The primary event of cancer cell killing in SR/CR mice was necrosis, consistent with the rapid clearance of a large cancer cell burden. But clear evidence also indicated that macrophages induced apoptosis in target cells. The double killing mechanisms were consistent with the fact that apoptosis and necrosis often share similar pathways. For instance, NK cells and CTLs release pfp, a potent inducer of necrosis and apoptosis, as well as gzmA and B which activate the caspase cascades resulting in apoptosis. Regardless of the effector mechanism, killing of cancer cells by each leukocyte subset required the formation of close membrane contact sites between effector and target cells. However, these contact sites were normally inaccessible to potential high molecular weight inhibitors, such as Abs to pfp and gzmA and B to block the effects of these proteins. Therefore, effector mechanisms were identified primarily by using low molecular weight specific inhibitors that penetrate into the cells. Using these inhibitors, our results indicated that the production of RNS was not involved in macrophage killing. Furthermore, RNS production was not induced when SR/CR leukocytes were activated by cancer cells. On the contrary, stimulation of RNS production did not facilitate the killing activity of SR/CR leukocytes. These findings were somewhat surprising since RNS have been previously implicated as a major effector mechanism of cancer cell killing by macrophages. Instead, production of ROS was involved, at least partially, in the killing by macrophages.
The above results beg an answer to the question of what makes SR/CR mice unique. The killing of cancer cells in SR/CR mice requires three distinct phases. First, the leukocytes must migrate to the site of cancer cells after sensing their presence. Second, they must recognize the unique properties of the cancer cell surface and make tight contact with it. Third, the effector mechanisms must finally be delivered to target cells. The difference between SR/CR and WT mice seems to lie in one of the first two phases. Upon challenge with cancer cells, WT mice lack leukocyte infiltration and rosette formation. Apparently, the mutation in SR/CR mice renders the leukocytes capable of sensing unique diffusible and surface signals from cancer cells, and of responding to the activation signals by migration and physical contact. Once the first two phases are accomplished, unleashing the pre-existing effector mechanisms for killing seems to ensue by default. Therefore, the mutated gene (or genes) likely determines whether leukocytes interpret the signals from cancer cells as inhibition, as in WT leukocytes, or as activation of migration and target recognition, as in SR/CR leukocytes. Identifying the mutated gene (or genes) will likely explain this unique resistance to cancer through immunity.
Materials and methods
Cell lines and mouse strains
S180 cancer cells were propagated in DMEM with 10% FBS at 37˚C in 5% CO2 or maintained by serial passage through WT mice as ascites tumors. C57BL/6 mice were purchased from The Jackson Laboratory, and BALB/c mice were purchased from Charles River Laboratories. SR/CR mice in BALB/c and C57BL/6 backgrounds were bred at Wake Forest University. All treatments and procedures in mice were conducted according to guidelines approved by the Animal Care and Use Committee of Wake Forest University Health Sciences.
Isolation of leukocytes from SR/CR mice
Infiltration of mixed leukocytes into the peritoneal cavity of SR/CR mice was induced by i.p. challenge with 107 S180 cells. At desired time points, the mice were anesthetized with Avertin (2% tribromoethanol, 2% 2-methyl-2-butanol) and the peritoneal cavities washed with 10 ml PBS (2). Macrophages were obtained from SR/CR mice by i.p. injection of 3 ml 3% TG. After 72 h, the cells infiltrating the peritoneal cavity were retrieved. At this time point, >90% of the infiltrating cells were macrophages as determined by staining with hematoxylin. Macrophages were further purified by adherence to culture dishes in DMEM with 10% FCS at 37˚C for at least 30 minutes. Nonadherent cells were removed by gently rinsing with PBS.
Peritoneal washes from WT or SR/CR mice were harvested at 18 and 24 h after S180 cell injection. The cell-free supernatants were subjected to SDS-PAGE/Western blot analysis using peroxidase-based chemiluminescence detection. The blots were probed using goat anti-pfp, anti-gzmA, or anti-gzmB (Santa Cruz Biotechnology), followed by incubation with horseradish peroxidase-conjugated donkey anti-goat IgG (Jackson ImmunoResearch).
Cytoprep and histology of effector molecules
SR/CR mice were challenged with 106 S180 cells and infiltrating leukocytes were harvested 45 min later. The cells were fixed and permeabilized. Goat anti-pfp, anti-gzmA, and anti-gzmB Abs were used to detect the lytic components from lymphoid cells. The slides were then incubated with rhodamine-conjugated rabbit anti-goat Ab (Jackson ImmunoResearch) and counterstained with DAPI.
Detection of cleaved caspase 3 in target cells
Purified macrophages were co-cultured with S180 cells at a ratio of 10:1 at 39˚C in 5% CO2. At desired time points, samples were incubated with a 1:100 dilution of rabbit anti-cleaved caspase 3 Ab (Cell Signaling Technology) as described previously (26). Cells were then incubated with 25 µg/ml rhodamine-conjugated affinity-purified goat anti-rabbit IgG (Jackson ImmunoResearch) and counterstained with DAPI.
Immunofluorescence was observed with a Zeiss Axioplan 2 fluorescence microscope. Images were captured by a Zeiss Axiocam CCD camera, and composite figures were prepared with Adobe Photoshop.
In vitro killing assays and inhibitors for signaling pathways and effector mechanisms
In vitro immune cell killing of S180 cells was performed as described previously (1). Briefly, macrophages were adhered to a 6-well plate and cultured at 37˚C for 24 h to recover from TG exposure. DiO-labeled S180 cells were added to the purified macrophages at a 1:10 ratio and incubated in DMEM with 10% heat-inactivated fetal calf serum at 39˚C for 24 h. Equal numbers of S180 cells alone were cultured in the same conditions and served as no-killing controls (100% survival). At the desired time points, S180 cells were identified by size, morphology, and fluorescence. Live S180 cells were characterized as negative for Trypan blue staining and positive for DiO fluorescence. Dead S180 cells were identified by positive Trypan blue staining and negative DiO fluorescence. The requirement for cellular contact between immune cells and tumor cells was analyzed by co-incubating S180 cells with WT or SR/CR leukocytes in a Transwell plate, separated by a 0.4 µm pore filter.
RNS production via NOSII in purified macrophages was inhibited by pre-incubation with 5 mM NMMA, 50 µM 1400W or 500 µM L-NIL (Sigma). ROS production and respiratory burst by phagocyte oxidase in purified macrophages was inhibited by co-incubation with either TG (Sigma), 2.8 µM diphenylene iodonium, 2000 U/ml catalase or 0.2 mg/ml superoxide dismutase (kind gifts from Dr. Linda McPhail). Serine proteases in purified macrophages were inhibited with 250, 1000 or 4000 U/ml trypsin inhibitor from bovine pancreas (Sigma).
Production of RNS from NOSII was stimulated and detected according to standard protocols. Briefly, macrophages from SR/CR and WT mice were plated at 2 x 105 per well in a 96-well plate and incubated at 37˚C overnight. Macrophages were then incubated at 39˚C for 24 h with either S180 cells or 10 U/ml IFN-γ and 10 ng/ml LPS as a positive control for RNS production. 50 µl of supernatant was added to each well of a new 96-well plate with 50 µl Greiss reagent and absorbance was read on an ELISA plate reader at 540 nm. IFN-γ, LPS, and Greiss reagent were purchased from Sigma.
Detection of anti-cancer Abs in the cell-free fraction of peritoneal washes of SR/CR mice
S180 cells were centrifuged onto poly-lysine coated dishes and incubated with peritoneal washes from either SR/CR mice that had been previously challenged with S180 cells or from WT controls that had not received S180 cells to test for the presence of anti-cancer Abs. The dishes were washed with cold PBS three times and incubated with 25 µg/ml goat anti-mouse Ab (Jackson ImmunoResearch) in buffer for 30 min on ice. Cells were counterstained with DAPI.
The studies described were supported by grants from the Cancer Research Institute, the National Cancer Institute, and the Charlotte Geyer Foundation.
- Received March 28, 2006.
- Accepted September 28, 2006.
- Copyright © 2006 by Zheng Cui