Interleukin-2 blocks the antitumour activity caused by depletion of CD25+ cells in a murine renal adenocarcinoma model

Authors


Takumi Takeuchi, Department of Urology, Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113–8655, Japan.
e-mail: nezu-tky@umin.ac.jp

Abstract

OBJECTIVE

To test the effectiveness of antimouse CD25 monoclonal antibody (mAb) against murine renal adenocarcinoma (RENCA) cells, as immunoregulatory/suppressor cells are known to be involved in tumour development in vivo, but the functions of these cells are not yet clear, and eliminating naive CD25 (interleukin-2 receptor α)-positive T cells elicits potent immune responses to syngeneic tumours in vivo.

MATERIALS AND METHODS

Aliquots of 1 × 104 or 1 × 105 RENCA cells were implanted into the subcapsule of the left kidney of syngeneic male Balb/c mice. Mice were injected with 125 µg of antimouse CD25 mAb to deplete CD25+ cells before RENCA implantation. Then 104 units of recombinant human interleukin-2 (rhIL-2) were subcutaneously injected twice daily for 7 days. Fourteen or 25 days later the tumour size was determined by laparotomy, and cells sorted using two-colour flow cytometry.

RESULTS

Depletion of naive CD25+ cells with anti-CD25 mAb and rhIL-2 administration effectively induced anti-RENCA tumour activity in Balb/c hosts. However, co-administration of anti-CD25 mAb and rhIL-2 abrogated this significant suppression of RENCA tumour growth. RENCA implantation reduced the proportion of CD4+ cells among splenocytes, whereas anti-CD25 mAb treatment increased it. The proportion of CD25+CD8+ cells among splenocytes and that of CD25+ cells among CD8+ cells were markedly reduced by co-administration of anti-CD25 mAb and rhIL-2 with RENCA implantation. Both CD4+ and CD8+ cells were stained around the remnant microscopic RENCA tumour after anti-CD25 mAb treatment.

CONCLUSION

Either depletion of naive CD25+ cells or rhIL-2 administration suppressed RENCA tumour growth in murine hosts. However, co-administration of anti-CD25 mAb and rhIL-2 abrogated this significant suppression of RENCA tumour growth.

Abbreviations
RENCA

murine syngeneic renal cell carcinoma cell

SC

splenocyte

CTL

cytotoxic T lymphocyte

mAb

monoclonal antibody

rhIL-2

recombinant human interleukin-2.

INTRODUCTION

The involvement of immunoregulatory/suppressor cells in tumour development in vivo is well known, but the functions of these cells have not been fully determined. Eliminating naive CD25 (interleukin-2 receptor α)-positive T cells, which constitute 5–10% of peripheral CD4+ T cells in normal naive mice, elicits potent immune responses to syngeneic tumours in vivo, resulting in their eradication, suggesting that CD25+CD4+ immunoregulatory T cells are involved in regulating the growth of those tumours [1]. The responses were reported to be mediated by tumour-specific CD8+ cytotoxic T lymphocytes (CTLs) and tumour-nonspecific CD4-CD8-cytotoxic cells akin to natural killer cells. Furthermore, in vitro culturing of CD25+CD4+ T cell-depleted splenic cell suspensions prepared from normal mice not sensitized to the tumour led to the spontaneous generation of similar CD4–CD8- cytotoxic cells capable of killing a broad spectrum of tumours [2].

Immunological self-tolerance is also in part maintained by this unique CD25+CD4+ naturally anergic/suppressive T cell population [3–5], and its functional abnormalities as well as in vivo blockade of the CTLA-4 [6], B7/CD28 pathway [7] or CD40/CD40L pathway [8] lead to the development of autoimmune disease. These CD25+CD4+ regulatory T cells require activation through their CD3 cell receptor to suppress CD25CD4+T cells in an antigen-nonspecific manner [9].

Human CD25+CD4+ regulatory T cells have been reported to strongly express CTLA4 but have reduced expression of CD40 ligand [10]. In addition, those cells produce interleukin-10, TGF-β, and low levels of interferon-γ, and no interleukin-2 or -4. These CD25+CD4+ regulatory T cells expand in vitro in the presence of interleukin-2 and allogeneic feeder cells, and maintain their suppressive capacity. Foxp3, which encodes a transcription factor that is genetically defective in an autoimmune and inflammatory syndrome in humans and mice, is specifically expressed in naturally arising CD4+ regulatory T cells. Furthermore, retroviral gene transfer of Foxp3 converts naive T cells toward a regulatory T cell phenotype similar to that of naturally occurring CD4+ regulatory T cells. Thus, Foxp3 is a key regulatory gene for the development of regulatory T cells [11].

Interleukin-2 has the potential to suppress some human tumours such as renal cancer in vivo[12]. It is also effective against immunogenic and highly malignant murine syngeneic renal cell carcinoma cells (RENCA) [13]. Nevertheless, in some human renal cancer cases, administering interleukin-2 appears to worsen the prognosis by promoting tumour growth. In those cases, CD25+CD4+ regulatory T cells may proliferate in response to exogenous interleukin-2 and suppress cell-mediated antitumour immunity.

In the present study, we tested the effectiveness of antimouse CD25 monoclonal antibody (mAb), recombinant human interleukin-2 (rhIL-2), and co-administration of both against RENCA cells, especially to delineate whether the combined therapy has additive, none or even detrimental effects.

MATERIALS AND METHODS

Aliquots of 1 × 104 or 1 × 105 RENCA cells suspended in 0.1 mL of RPMI 1640 culture medium or medium alone were implanted into the subcapsule of the left kidney of syngeneic male Balb/c mice (6–8 weeks old, Nisseizai Tokyo, Japan) under ether anaesthesia as described previously [14]. Mice were injected via the penile vein with 125 µg of depleting antimouse CD25 (PC61) or control IgG 24 h before RENCA implantation; 104 units of rhIL-2 (Shionogi Pharmaceutical Company Ltd., Osaka, Japan) or vehicle (PBS) was injected subcutaneously twice daily for 7 days beginning 2 h after RENCA or medium alone (sham-operated) inoculation. Then 14 or 25 days later the tumour size (calculated as 0.4 width2× length) was determined by laparotomy. For flow cytometry, mice were killed 6 days later after treatment with or without 125 µg of anti-CD25 mAb, sufficient to deplete CD25+ cells in vivo[1], and/or rhIL-2 (1 × 104 units twice daily for 6 days) and inoculation with 1 × 105 RENCA cells into the left renal subcapsule, as described above. All comparable experiments were performed at the same time. The principles of laboratory animal care (NIH publication no. 85–23, revised 1985) were followed.

Hybridoma cells producing antimouse CD25 mAb (PC61, a rat IgG1 antibody) were cultured in RPMI 1640 containing 10% fetal calf serum. Culture supernatants were collected and the antibody purified using a mAb TrapTM GII Column kit (Amersham Pharmacia Biotech AB, Uppsala, Sweden) according to the manufacturer's instructions. Fluorescein isothiocyanate-labelled anti-CD25 mAb (PC61, Cedarlane Laboratories Ltd, Ontario, Canada), phycoerythrin-labelled anti-CD4 mAb (CT-CD4, Caltag, Burlingame, CA), and phycoerythrin-labelled anti-CD8 mAb (CT-CD8, Caltag) were used for flow cytometry. Splenocytes (SCs) were prepared as described previously [15] from the spleens of mice. Red blood cells were lysed with ACK lysing buffer (0.15 mol/L NH4Cl, 1.0 mmol/L KHCO3, 0.1 mmol/L Na2EDTA, pH 7.4) for 5 min at room temperature, and washed twice in RPMI 1640. Aliquots of 1 × 106 spleen cells resuspended in RPMI 1640 medium containing 1% BSA and 0.1% sodium azide (staining medium) were incubated with fluorescence-labelled antibodies for 20 min at 4 °C, washed twice in staining medium, and fixed in 1% paraformaldehyde. Two-colour flow cytometry was used (10 000 cells counted) using a FACScan system (Beckton-Dickinson, USA). Flow cytometry data were analysed using Cellquest software version 3.1 (Beckton-Dickinson) with lymphocyte gating. To assess the effectiveness of anti-CD25 mAb (PC61), flow cytometry was used on SCs of mice 24 h after the intravenous administration of 125 µg of the antibody. For histology, haematoxylin and eosin staining, CD4 staining and CD8 staining were used as previously described [14].

The Mann–Whitney U-test was used to compare the tumour volumes, with Fisher's exact test or the chi-square test to compare the tumour acceptance rate. Flow cytometric data were analysed using an unpaired t-test.

RESULTS

Intravenous anti-CD25 mAb (125 µg) effectively depleted CD25+CD4+ T cells among SCs, as shown in Table 1. The CD25+ cell-depleting mAb retarded or completely suppressed the growth of RENCA tumours when 1 × 104 RENCA cells were inoculated into the renal subcapsule (Table 1), and the rate of macroscopic tumour formation was reduced to three of nine, while seven of eight control RENCA hosts formed macroscopic tumours (Table 1). The histology of the kidneys of the inoculated mice, compared with the control (Fig. 1A), showed a very small region of dying RENCA cells (Fig. 1B) and massive lymphocyte infiltration with no remaining viable RENCA cells at the site of implantation (Fig. 1C) when the implanted cells did not form macroscopic tumours. CD4+ and CD8+ cells were sparsely stained, mainly in the peritumoral area in macroscopic RENCA tumours (Fig. 2). In the remnant microscopic tumour in the anti-CD25 antibody-treated group, there was stronger staining of both CD4+ and CD8+ cells just under the tumour (Fig. 2).

Table 1.  Data at 25 (1 × 104/mouse) and 14 days (1 × 105/mouse) after RENCA implantation
VariableControl Iganti-CD25 mAb
  • *

    P < 0.05 (Fisher's exact test);

  • P < 0.01 (Mann–Whitney U-test).

  • P < 0.05 against control Ig + vehicle;

  • P < 0.05 by Mann–Whitney U-test.

N  3  3
CD25+CD4+ cells/SC, %  0.95 (0.05)  0.46 (0.07)
CD25+CD8+ cells/SC, %  0.99 (0.30)  0.39 (0.08)
N  8  9
Macroscopic tumour rate, n  7*  3
Tumour volume, mm3474 (233)  5 (4)
N  7  6
+ vehicle363 (177)103 (102)
N  6  6
+ rhIL-2 72 (62)171 (113)
Figure 1.

A: control RENCA tumour. B, C: kidneys from mice inoculated with RENCA cells following anti-CD25 antibody treatment. D: RENCA tumour in rhIL-2 treated mice, E: RENCA tumour in mice treated with both anti-CD25 antibody and rhIL-2. B: the arrow indicates the locus of dying RENCA cells. C: the arrow indicates massive lymphocyte infiltration. A,D,E: 1 × 105 RENCA cells were inoculated. B,C: 1 × 104 RENCA cells were inoculated. Original × 100.

Figure 2.

Upper panel: CD4+ cell staining. Lower panel: CD8+ cell staining. Left: control RENCA tumour, Right: microscopic RENCA tumour following anti-CD25 mAb administration. Original × 400. T: RENCA tumour. Arrows: representative of CD4/CD8 positive cells.

When 1 × 105 RENCA cells were inoculated, either giving rhIL-2 or depleting CD25+ cells by the mAb significantly suppressed RENCA tumour growth in vivo compared with that in the controls (Table 1). Co-administration of anti-CD25 mAb and rhIL-2 abrogated this significant suppression of RENCA tumour growth induced by either. There was splenomegaly in the rhIL-2-treated groups but not in the others. Most of the mice (more than three-quarters) formed macroscopic tumours in all of the groups with the inoculation of 1 × 105 RENCA cells. RENCA tumours in rhIL-2-treated mice (Fig. 1D) were histologically smaller than those in control mice (Fig. 1A) and mice with co-administration of anti-CD25 antibody and rhIL-2 (Fig. 1E).

Flow cytometry (Figs 3 and 4) showed that RENCA implantation reduced the proportion of CD4+ cells among SCs, whereas anti-CD25 mAb treatment significantly reduced it. Neither rhIL-2 alone nor co-administration of anti-CD25 mAb and rhIL-2 restored the splenic CD4+ cells. The proportion of CD25+CD4+ cells (activated CD4+ cells and/or regenerated/remaining naive immunoregulatory CD25+CD4+ cells) in the spleen at death was decreased by anti-CD25 mAb treatment followed by rhIL-2, but it recovered with after imposing the RENCA burden. The proportion of CD25+ cells among CD4+ cells significantly increased with RENCA burden in the group with co-administered anti-CD25 mAb and rhIL-2. The proportion of CD25+CD8+ cells among SCs and CD25+ cells among CD8+ cells were markedly reduced in the group with co-administration of anti-CD25 mAb and rhIL-2, and RENCA implantation. The proportion of CD25+ cells among CD8+ cells was increased in the rhIL-2 group with RENCA burden compared with the control group with or without RENCA.

Figure 3.

Representative dot plots obtained by flow cytometry, veh, vehicle for rhIL-2 (PBS). A: CD4/CD25. B: CD8/CD25.

Figure 4.

Flow cytometric data (three mice in each group); veh: vehicle for hIL-2 (PBS), RENCA: 1 × 105 RENCA cells. All mice without RENCA inoculation were sham-operated. *P < 0.05 vs control Ig + veh, ♯; P < 0.05 vs control Ig + veh + RENCA; &, $P < 0.05 between groups by unpaired t-test.

DISCUSSION

Depletion of naive CD25+ cells with anti-CD25 mAb effectively induced anti-RENCA tumour activity in Balb/c hosts. There were reductions in both tumour incidence and tumour volume, with a lower tumour burden (1 × 104 cells/mouse), while there was only a reduction in tumour volume with a higher tumour burden (1 × 105 cells/mouse). The effectiveness of antitumour immune cells would be expected to be enhanced when the tumour burden is smaller, considering that tumour growth is determined by the sum of the proliferation of tumour cells and the apoptotic loss of tumour cells induced by antitumour immune cells. Thus, tumour cells can be completely eradicated by antitumour immune cells when the tumour burden is smaller. We implanted more tumour cells (1 × 105) in experiments with co-administration of anti-CD25 mAb and rhIL-2 because we supposed that the more profound immunological effects caused by co-administration might be more discernible with a higher tumour burden.

As previously reported [12], rhIL-2 alone also retarded RENCA tumour growth in vivo; surprisingly, co-administering antimouse CD25 mAb and rhIL-2 (each of which was effective in suppressing RENCA growth) abrogated the significant suppression of RENCA tumour growth. It remains unknown why co-administration failed to induce RENCA tumour suppression, whereas either treatment alone did so. It would not have been surprising if the tumour growth after co-administration had been similar to that of groups with either treatment alone, indicating that there were no additive effects, possibly because the initial depletion of CD25+ cells with the mAb would abrogate the subsequent rhIL-2 effects. Considering the flow cytometric findings, RENCA growth appears to be inversely linked to the proportion of CD4+ cells among SCs. Depletion of naive CD25+ cells in vivo with mAb that suppressed RENCA growth partly restored the CD4+/SC ratio toward the control level. That co-administration of antimouse CD25 mAb and rhIL-2 did not completely restore this ratio may have accounted for the abrogation of the suppression of RENCA growth.

Although CD25+CD4+ cells were detected to some extent even in mice given anti-CD25 mAb on 6 days after administration they might have been re-populated from the thymus, or CD25 might have been newly expressed on T cells via activation by the RENCA cells, rhIL-2 and the depletion of naive CD25+ cells. CD25+CD4+ cells may be naïve and/or activated immunoregulatory T cells that suppress antitumour immunity.

Reportedly, the antitumour activity caused by the deletion of naive CD25+CD4+ immunoregulatory T cells is mediated by CD8+ CTLs as well as natural killer cells/lymphocyte activated killer cells [2]. The reasons why CD25+CD8+ cells, possibly activated CD8+ CTLs, were by far the lowest in the group with co-administration of anti-CD25 mAb and rhIL-2 plus RENCA burden among all the groups remain unknown. The combination of anti-CD25 mAb, rhIL-2 and RENCA cells might have suppressed the division of CD8+ CTLs or induced them to die by apoptosis after activation. Murakami et al.[16] reported that CD25+CD4+ T cells are involved in the IL-2-mediated inhibition of memory CD8+ T cell division and that IL-2 controls the number of memory phenotype CD8+ T cells, at least in part through maintaining the CD25+CD4+ T cell population. Both CD4+ and CD8+ cells were stained around the remnant microscopic RENCA tumour after anti-CD25 mAb treatment, and thus both cells may be involved in tumour suppression.

Our hypothesis that if naive CD25+CD4+ immunoregulatory T cells, which proliferate in response to IL-2 and maintain their immunosuppressive capacity [10], were depleted before rhIL-2 administration, suppression of RENCA tumour growth might be greater, was not true; instead the result was the opposite. In conclusion, either depletion of naive CD25+ cells or rhIL-2 administration suppressed RENCA tumour growth in hosts. However, co-administration of anti-CD25 mAb and rhIL-2 abrogated this significant suppression of RENCA tumour growth in vivo.

ACKNOWLEDGEMENTS

Hybridomas producing antimouse CD25 mAb (PC61) were provided by Dr J. Shimizu (Department of Immunopathology, Tokyo Metropolitan Institute of Gerontology). RENCA cells were provided by Dr R.H. Wiltrout (The National Cancer Institute). We are grateful to Ms. E. Tanaka for assistance with cell culture.

CONFLICT OF INTEREST

None declared.

Ancillary