Anti-tumor effects of human peripheral γδ T cells in a mouse tumor model
Article first published online: 7 MAR 2001
Copyright © 2001 Wiley-Liss, Inc.
International Journal of Cancer
Volume 92, Issue 3, pages 421–425, 1 May 2001
How to Cite
Zheng, B.-J., Chan, K.-W., Im, S., Chua, D., Sham, J. S.T., Tin, P.-C., He, Z.-M. and Ng, M.-H. (2001), Anti-tumor effects of human peripheral γδ T cells in a mouse tumor model. Int. J. Cancer, 92: 421–425. doi: 10.1002/ijc.1198
- Issue published online: 30 MAR 2001
- Article first published online: 7 MAR 2001
- Manuscript Accepted: 4 DEC 2000
- Manuscript Revised: 28 NOV 2000
- Manuscript Received: 23 OCT 2000
- Research Grants Council
- Committee on Conference and Research Grants, University of Hong Kong
- γδ T cells;
- anti-tumor effect;
- mouse tumor model;
- tumor-infiltrating lymphocytes
Peripheral γδ T cells derived from healthy donors were found to exhibit cytotoxicity against a variety of tumor cell lines in vitro, including CNE2, which was established from nasopharyngeal carcinoma (NPC). The anti-tumor effects were further studied in a mouse model. Control nude mice inoculated s.c. with 5 × 106 CNE2 cells regularly developed hypodermal tumors, which progressively increased in size, and animals had a mean survival of 35 ± 3.4 days. Tumor growth was arrested and tumor size was reduced after animals were infused with 5 × 107 γδ T cells derived from a healthy donor. The anti-tumor effects were temporary, however, and tumor growth was resumed after about 1 week in a group of the animals that had been given a single dose of γδ T cells. In another group of animals given 2 doses of γδ cells 1 week apart, resumption of tumor growth was delayed for a further week. Mean survival of the 2 groups was increased to 61 ± 15.7 and 74 ± 12.9 days, respectively. Immunohistology revealed an accumulation of infused cells in tumors attended by focal tumor necrosis in specimens taken 2 days after infusion. Infiltrative cells virtually disappeared from tumor tissues 6 days after infusion, accompanied by increased mitotic indices of tumor cells. These temporal relationships suggested that the accumulation of infused γδ T cells in hypodermal tumors was responsible for the observed anti-tumor effects. © 2001 Wiley-Liss, Inc.
T lymphocytes harboring TCR-αβ or -γδ contribute to host defense against chronic infection and tumors through the discharge of cytolytic activity, secretion of multiple cytokines and the helper function for immunoglobulin production. They perform these functions efficiently through unique pathways, complementing the defense mechanisms, to maintain immunological competence.
γδ T cells account for <10% of the T-cell population in thymus, spleen, lymph nodes and peripheral blood; but they are the predominant, or even exclusive, T-cell type in certain epithelial tissues.1–3 The marked tissue distribution of γδ T cells bearing distinct TCR V gene products is a striking feature in mice. In lymphoid organs and peripheral blood, γδ T cells primarily express Vγ2 or Vγ1 TCR chains. Dendritic epidermal T cells (DETCs) express an identical TCR Vγ3/Vδ1,1 but the γδ T cells in the reproductive epithelium express the Vγ4/Vδ1 TCR.2, 4 Distinct γδ T-cell populations with restricted TCR expression have also been found in the epithelium of other tissues, such as lung, tongue and breast.2, 4 The particular TCR repertoire of these subsets appears to be linked to functional properties and maintenance in different tissues. DETCs, e.g., can produce keratinocyte growth factor (KGF), which could help to maintain the homeostasis of epithelial tissues following perturbation induced by physical, chemical or infectious agents.5 These results suggest that subsets of γδ cells in different anatomical locations may play distinct roles. Different from mice, human γδ cells display significant TCR heterogeneity with extensive insertions and deletions at the re-arranged junctions. The great diversity of γδ cells and wider tissue distribution may underlie their capacity to recognize a broader variety of antigens than the apparently monoclonal γδ cell populations in mice. However, it remains unclear whether subsets of γδ T cells in human peripheral blood have the ability of homing to other “damaged” tissues, including tumors. It has been suggested that activated γδ cells have a surveillance function against tumors. Some subsets of human γδ cells could recognize superantigen complexes on the surface of Burkitt lymphoma (BL) cells,6–8 Epstein-Barr virus (EBV)–transformed B cells9 or autologous tumor cells in acute lymphoblastic leukemia (ALL).10 γδ T-cell lines established from populations of tumor-infiltrating lymphocytes (TILs) of lung cancer11 and renal carcinoma12,13 have been shown to lyse autologous tumor cells. No specific cytolytic activity, however, was observed with γδ TILs isolated from Wilms' tumors, melanomas, sarcomas14 and lung carcinoma.15 Thus, the precise role of γδ TILs against tumors has yet to be determined.
Zheng et al.16 showed that γδ T cells derived by stimulating peripheral blood mononuclear cells (PBMCs) from healthy donors with the irradiated myeloma cell line XG7 exhibited cytotoxic activity against a variety of tumor cell lines, including the nasopharyngeal carcinoma (NPC) cell lines CNE2 and 915. The tumor is especially common among southern Chinese, and the findings were consistent with the belief that γδ T cells may contribute to surveillance against the cancer. We have further studied the anti-tumor effects of γδ T cells in an NPC model produced by inoculating nude mice with the CNE2 cell line. Infusion of γδ T cells temporarily arrested tumor growth, and the inhibition was coincident with an accumulation of infused cells in the tumor, attended by increased tumor necrosis.
MATERIAL AND METHODS
BALB/c nude mice (Hfh11nu/+), designated N38, were obtained from the Australian Nuclear Science and Technology Organisation (Sydney, Australia) in 1987. They were bred under standard pathogen-free conditions in the Laboratory Animal Unit, University of Hong Kong. Female mice, 6 weeks old and weighing 16 to 18 g, were used. Mice were housed in cages under standard conditions with regulated temperature and humidity, fed with pelted food and tap water and cared for according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals.21
NPC cell line CNE2 (a gift from Dr. Y. Zeng, Beijing, China) was propagated in MEM (GIBCO BRL, Gaithersburg, MD) supplemented with 10% heat-inactivated FBS (GIBCO BRL), 0.2 mM L-glutamine and antibiotics (100 U/ml penicillin, 100 μg/ml streptomycin, 20 γg/ml garamycin and 100 U/ml nystatin) at 37°C in a 5% CO2 incubator.
γδ T cells were derived from peripheral blood of healthy donors as described previously.16 In brief, PBMCs from a healthy donor were separated according to a standard Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) procedure. PBMCs (5 × 106) and irradiated (7,000 cGy) XG7 cells (2.5 × 106, a gift from Dr. X.G. Zhang, Suzhou, China) were mixed and incubated in RPMI-1640 medium (GIBCO BRL) supplemented with 15% FCS for 10 days. Irradiated XG7 cells (1.5 × 106) were added to cultures on days 3, 6 and 9, respectively. rIL-2 (100 U/ml; R&D Systems, Minneapolis, MN) was added to cultures on day 10. Cultures were fed every 2 days or as needed thereafter with culture medium containing the same cytokine. Expanded cultures contained >80% of γδ T cells as confirmed by flow cytometry using FITC- or PE-labeled monoclonal antibodies (MAbs) against CD3, CD4, CD8, CD19, CD56, pan-TCR-αβ, pan-TCR-γδ and G1 (negative control).
The cytotoxic activity of γδ T cells against CNE2 target was determined by a standard 4 hr calcein-release assay17,18 in U-bottomed, 96-well microplates. Briefly, triplicate cultures were seeded with graded numbers of effector cells and 5,000 calcein AM (Molecular Probes, Eugene, OR)–labeled target CNE2 cells. Target cells were labeled immediately before use by incubation with 2 μM calcein AM in PBS for 40 min at 37°C. Plates were centrifuged at 100 g for 3 min and incubated for 4 hr. Cytolysis of targets was determined by measuring the fluorescence intensity (FI) of calcein using a fluorometer. Maximum release was estimated by incubating target cells with 5% SDS (Sigma, St. Louis, MO) (total lysis) and spontaneous release, by incubating targets in medium alone (target control). The percentage specific cytolysis was calculated as follows:
In a separate set of experiments, effector γδ T cells were pre-incubated with or without 2 μg/ml of pan-γ/δ MAb (Immunotech, Marseille, France) for 30 min before being mixed with calcein-labeled targets. Target lysis in the presence or absence of γδ antibody was measured as described above.
Each of 16 nude mice was inoculated s.c. with 5 × 106 CNE2 cells that had been washed twice with PBS and resuspended in serum-free RPMI-1640. All of these mice developed hypodermic tumors of 6 to 12 mm after 10 days. They were then evenly divided into 4 groups (T1, C1, T2 and C2) based on tumor size. Groups T1 and T2 were infused i.v. into the tail vein with 5 × 107 γδ T cells. One week later, group T2 received a second infusion with the same dosage of γδ T cells. Groups C1 and C2, which did not receive the infusion of γδ T cells, served as untreated controls. Tumors were measured weekly thereafter and the times of death recorded.
In a separate set of experiments, tumors growing s.c. in nude mice were excised for histological and immunohistochemical studies at 2, 4, 6 and 8 days after infusion of γδ T cells. Tumors were immediately fixed in 10% buffered formaldehyde or frozen in liquid nitrogen and then embedded in paraffin. They were sectioned at 4 μm thickness and mounted on slides. Paraffin sections prepared from formaldehyde-fixed tissue were processed for hematoxylin and eosin (HE) staining according to the standard procedure. After dehydration in graded ETOH, immunoperoxidase staining was performed on paraffin sections obtained from frozen samples using rabbit anti-human CD3, biotinylated goat anti-rabbit Ig and strptABComplex/HRP (Dako, Copenhagen, Denmark) according to the manufacturer's instructions.
Ex vivo expansion of peripheral γδ T cells was effected by stimulating PBMCs from a healthy donor with the irradiated myeloma cell line XG7, as described in Material and Methods. After 3 weeks of culture, the total cell number was increased approximately 600-fold, from 5 × 106 to approximately 3 × 109. Flow-cytometric analysis showed that 82.4% of the cell population were γδ T cells with the phenotype CD3+/TCR-γδ+. The remainder was made up of 16.7% CD3+/TCR-αβ+ T cells. While CD3–/CD19+ B cells and CD3–/CD16+ or /CD56+ NK cells comprised <1% of the total cell population (FACS data not shown). The highly enriched γδ T-cell culture exhibited cytotoxicity against CNE2 cells (Fig. 1). Significant cytotoxicity of at least 20% of specific target lysis was seen at an E:T ratio of 3, and the percent of target lysis increased as the E:T ratio increased. Cytotoxicity was virtually abrogated by the TCR-γδ antibody. Since the HLA types of the target (A26 B15) and effector (A2 B60) were different, the observed cytotoxicity was not MHC-restricted.
The anti-tumor effects of γδ T cells were further studied in nude mice inoculated s.c. with 5 × 106 CNE2 cells/mouse (Fig. 2). Animals developed a hypodermal tumor, which became visible after about 5 days. In control mice (groups C1 and C2), the tumor continued to progress, increasing in size, and all animals died, with a mean survival of 35 ± 3.4 days. In one experiment, test animals (T1) were given a single i.v. infusion of γδ T cells on day 10. Tumor growth was arrested and tumor size reduced, but the anti-tumor effects were temporary; tumor growth resumed after about 1 week, on day 17 (Fig. 3a). In another experiment, test animals (T2) were given a second dose of γδ T cells on day 17 and the above-described anti-tumor effect was sustained for 1 further week before tumor growth was resumed after day 24 (Fig. 3b). Mean survival times of both groups were significantly longer than that of controls (p < 0.03), and that of animals given 2 infusions of γδ T cells was significantly longer than that of animals given 1 dose of γδ T cells (p = 0.02) (Fig. 3c).
The above experiments were repeated and the tumors excised on days 2, 4, 6 and 8 after infusion of γδ T cells for histological and immunohistochemical studies. In the specimens taken 2 days after infusion, there was extensive tumor necrosis (Fig. 4a-i) accompanied by significant accumulation of CD3+ T cells, particularly at the tumor periphery (Fig. 4a-ii). Infiltrating CD3 T cells were essentially absent from specimens taken after 6 days (Fig. 4b-ii), and the tumor exhibited increased mitotic activity (Fig. 4b-i). Similar results were observed with specimens taken 2 and 6 days after the second infusion of γδ T cells (Fig. 4c,d).
In agreement with our previous finding,16 the myeloma cell line XG7 was found to stimulate vigorous and selective ex vivo expansion of peripheral γδ T cells from healthy donors. The resulting bulk cultures were highly enriched with γδ T cells and exhibited vigorous cytotoxicity against the NPC cell line CNE2 in vitro. Since only an insignificant number of CD3–/CD56+ or /CD16+ NK cells were present to confound the CTL assay results, it was reasonable to attribute the observed cytotoxicity against the HLA mismatched tumor targets to γδ T cells. The latter comprised 82.4% of the bulk culture, and this subset of T cells constituted a distinct component of the immune system, which is believed to respond to tissue injuries caused by physical, chemical and biological agents, including tumors.10–13 The other subset of αβ T cells made up the remaining 16.7% of the cell population. These are unlikely to contribute significantly to the observed cytotoxicity against the CNE2 target because, unlike γδ T cells, αβ T cells recognize mainly peptide antigens in association with MHC molecules.19, 20 Moreover, the healthy donor from whom the cells were derived presumably had not been exposed to the tumor previously and, hence, would not be expected to contain immune αβ T cells specific for the tumor. Indeed, cytotoxicity was virtually abrogated by the γδ TCR-specific antibody. It was concluded, therefore, that the αβ T cells and the small percentage of NK cells present in the bulk culture either did not contribute or contributed only minimally to the observed cytotoxicity.
The anti-tumor effect was further evaluated in vivo in a nude mouse tumor model. In control animals, the hypodermal tumor was observed 4 to 6 days after s.c. inoculation with NPC tumor cells. Thereafter, the tumor continued to increase in size, and mean survival was 35 days. One group of test animals (T1) was infused with γδ T cells 10 days after inoculation of tumor cells, when tumors were already well established. The treatment arrested tumor growth. The effect was temporary, however, and tumor growth resumed after 7 days. A further infusion given 7 days later to another group of animals (T2) arrested tumor growth for about 14 days. Mean survival of animals given 1 infusion was increased by almost 2-fold, to over 61 days, and further to 74 days in the second group of animals, which were given 2 infusions of γδ T cells.
In a separate experiment, it was shown that arrest of tumor growth was concurrent with accumulation of infused CD3+ T cells in the tumor attended by the presence of focal tumor necrosis 2 days after infusion. The number of infused cells infiltrating the tumor declined markedly after 4 days, and they virtually disappeared from tumor specimens after 6 days. The limited life span of the infused cells in nude mice may be partly attributed to the absence of the human cytokines required to sustain the viability of these cells. It was significant, nevertheless, that the disappearance of the infused cells was concomitant with increased mitotic indices of the tumor cells and disappearance of tumor necrosis. In the group given 2 infusions 1 week apart, increased accumulation of γδ T cells attended by focal tumor necrosis was observed again 2 days after the second infusion. After 6 days, when the infused cells were absent, the tumor again showed increased mitotic activity.
Taken together, these results indicate an association between the accumulation of infused cells and arrest of tumor growth as well as focal tumor necrosis. Infused cells had a limited life span, and the disappearance of infused cells from the tumor was co-incident with resumption of tumor growth and increased mitotic activity of the tumor. The temporal relationships suggested that the accumulation of infused cells in tumors was responsible for the anti-tumor effects. These effects were temporary and evidenced by the arrest of tumor growth, focal tumor necrosis and prolonged survival. The effector cells responsible for the in vivo anti-tumor effects most probably correspond to those γδ T cells identified earlier in the in vitro CTL assay as being cytotoxic for the tumor. Any contribution of other cell types to the in vivo anti-tumor effects was considered unlikely for similar reasons, based on the earlier attribution of in vitro cytotoxicity exhibited by bulk culture to γδ T cells. In addition, the anti-tumor effects were short-term. Infused cells had a limited life span in mice, presumably in part because of the lack of human cytokines. Consequently, it was considered unlikely that other types of cell, e.g., NK or alloreactive αβ T cells, proliferated in the animals and thereby contributed to the in vivo anti-tumor effect.
In sum, γδ T cells derived from healthy donors exhibited in vitro cytotoxicity against the NPC cell line CNE2 as well as in vivo anti-tumor activity, as determined in a nude mouse hypodermal tumor model produced with the same tumor cell line. γδ T cells constitute a distinct component of the immune system, which is believed to respond to tissue injuries caused by different agents, including tumors. Since the test subject presumably had not been exposed to the tumor before, the immune cells were probably produced in response to previous tissue injuries caused by other agents.
We thank Dr. S.S. Lee, Hong Kong, China for critical reading of the manuscript.
- 21National Institutes of Health, U.S. Guide for the care and use of laboratory animals. Publication 86-23, 1985.