Cutting Edge: IL-10-Producing CD4⁺ T Cells Mediate Tumor Rejection

Mice

C57BL/6 wild-type mice were obtained from either the National Cancer Institute (Fredrick, MD) or The Jackson Laboratory (Bar Harbor, ME). C57BL/6 mice genetically deficient in either IFN-γ, IL-4, IL-12 p40, IL-10, MHC class II, or β₂-microglobulin were obtained from The Jackson Laboratory. All mice were housed under specific pathogen-free conditions.

Glioma cell lines

Cryopreserved GL261 tumor fragments were obtained from the National Cancer Institute Tumor Repository (Frederick, MD) and maintained in monolayer cultures in RPMI 1640 with 10% FCS and standard supplements. CT-2A cells were obtained from Dr. T. N. Seyfried (Boston College, Boston, MA) and grown in DMEM with 10% FCS and standard supplements. Both cells lines were found to be negative for the presence of indigenous murine viruses (standard panel of 12) by the Mouse Antibody Production test (Charles River Laboratories, Wilmington, MA). Glioma cells were harvested from culture by treatment with trypsin, washed three times, counted by trypan blue exclusion, and resuspended in PBS before s.c. or i.p. inoculation.

Priming and challenge of mice with glial tumor cells

Mice were injected i.p. with three weekly doses of 1 × 10⁶ irradiated GL261 cells (15,000 rad) in 0.1 ml of PBS. Control mice were injected with vehicle alone according to the same schedule. One week following the final inoculum all mice were anesthetized with Avertin, shaved across the lower back, and challenged s.c. with 4 × 10⁶ live GL261 cells in 0.1 ml of PBS. In certain experiments, mice were injected i.p. with rat anti-mouse IL-10 (1 mg each of mAbs SXC-1 and SXC-2 per injection; Ref. 9) or control rat IgG (2 mg; Sigma-Aldrich, St. Louis, MO) on days 0, 3, 6, and 9 post challenge. Tumor size was determined as the product of length and width measured by a blinded examiner using calipers.

Cytokine quantification

Supernatants of splenocyte cultures were collected at 24-h intervals and subjected to a standard sandwich ELISA technique using noncompeting pairs of cytokine-specific mAbs as previously described (10). The lower limit of detection of each assay was 30 pg/ml or less.

RT-PCR

Total RNA was isolated from tumor biopsy specimens using RNAzol RNA isolation solvent (Tel-Test, Friendswood, TX) and transcribed into cDNA using reverse transcriptase. RT-PCR was performed with primers specific for cytokines, β-actin, and the CD3ε chain (BD PharMingen, San Diego, CA).

Mice vaccinated with irradiated GL261 cells are protected from tumor growth

We successfully stimulated protective immunity against the murine glioma, GL261, by priming syngeneic C57BL/6 mice with three i.p. injections of irradiated tumor cells in PBS. The vaccine was similarly effective whether GL261 cells were obtained from cell lines propagated in vitro (Fig. 1⇓A) or from s.c. tumors serially passaged in nu/nu mice (data not shown). One hundred percent of GL261-vaccinated mice rejected tumors within 1 wk of challenge. In contrast, control mice inoculated with either vehicle alone or with irradiated cells from an independently derived C57BL/6 glial tumor cell line (CT-2A) experienced progressive GL261 glioma growth up to the time of sacrifice. Furthermore, mice primed with irradiated GL261 cells were not protected against challenge with live syngeneic B16 melanoma cells or CT-2A cells (data not shown). Collectively these results suggest that vaccination with GL261 cells generates an immune response specific for an Ag expressed by the GL261 tumor, and not an Ag universally expressed on gliomas or transformed cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Vaccination with irradiated tumor cells elicits CD4⁺ T cell-dependent anti-glioma immunity. A, C57BL/6 mice were injected three times (i.p., spaced 7 days apart; 10⁶ cells in 0.2 ml/mouse/injection) with either irradiated GL261 (▵) or CT-2A (□) glioma cells or vehicle alone (•). GL261 and CT-2A cell lines were treated with trypsin, washed, and irradiated (3000 rad) before injection. One week following the final inoculation, all mice were challenged by a s.c. injection of 4 × 10⁶ live GL261 cells. Subsequently, tumor size was measured every other day by a blinded observer using calipers. The results shown are representative of five experiments with 5–10 mice per group. B, C57BL/6 wild-type (squares), β₂-microglobulin-deficient (circles), and MHC class II-deficient (triangles) mice were primed with irradiated GL261 cells (closed symbols) or PBS alone (open symbols) before s.c. challenge with live GL261 cells as described in A.

Accelerated rejection of tumors in vaccinated mice is CD4⁺, but not CD8⁺, T cell dependent

Flow cytometric analysis of the cells contained in GL261 gliomas undergoing rejection revealed infiltration by both CD8⁺ and CD4⁺ T cells, as well as CD45⁺MHC class II⁺ cells that could serve as APCs (our unpublished observations). We next assessed the relative contributions of CD4/CD8 cell subsets to the antitumor response. C57BL/6 wild-type, MHC class II-deficient, and MHC class I (β₂-microglobulin)-deficient mice were primed with irradiated GL261 cells or vehicle alone and subsequently challenged with live GL261 cells. In contrast to their wild-type counterparts, none of the MHC class II-deficient mice mounted a protective anti-glioma response, suggesting that tumor-specific CD4⁺ T cells are required for accelerated glioma rejection (Fig. 1⇑B). In fact, tumor growth rates in GL261-primed as well as mock-primed MHC class II-deficient mice were comparable to those observed in immunodeficient nu/nu and RAG2^−/− mice (our unpublished observations). In contrast, 100% of vaccinated β₂-microglobulin^−/− mice successfully rejected their tumors during a 25-day observation period. Therefore, we concluded that glioma rejection can proceed in the absence of CD8⁺ T cells. Consistent with the results of these in vivo studies, CD8⁺ T cells purified from the spleens of GL261-vaccinated immunocompetent mice failed to lyse glioma targets, as opposed to allogeneic lymphoblasts, in ⁵¹Cr release assays (data not shown).

GL261-primed CD4⁺ T cells selectively produce the cytokine IL-10 upon challenge with glioma cells ex vivo

To investigate the mechanism underlying the anti-glioma effects of CD4⁺ T cells, we harvested spleens from vaccinated mice and measured cytokine production in response to tumor challenge ex vivo. Surprisingly, whole splenocytes as well as highly purified CD4⁺ T cells cocultured with fresh APCs failed to produce cytokines that have been shown to have antitumor effects in other experimental systems, including IFN-γ, IL-4, and TNF-α (data not shown). In contrast, CD4⁺ T cells isolated from the spleens of both primed wild-type or β₂-microglobulin^−/− mice produced large quantities of IL-10 upon stimulation with mitomycin C-treated GL261 cells (Fig. 2⇓A). We did not detect IL-10 in supernatants from GL261 cells cultured either alone or in combination with splenocytes from primed MHC class II-deficient mice (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

CD4⁺ T cells from GL261-vaccinated mice produce IL-10 upon challenge with tumor cells in vitro as well as within the tumor bed in vivo. A, Spleens were harvested from primed or control wild-type or β₂-microglobulin-deficient mice 12 days after s.c. challenge with live GL261 cells (4 × 10⁶ cells/mouse). CD4⁺T cells were purified using R&D enrichment columns (R&D Systems, Minneapolis, MN) and resuspended in tissue culture medium (1 × 10⁶ cells/ml) supplemented with irradiated syngeneic splenocytes (4 × 10⁶ cells/ml). Mitomycin C-treated GL261 cells were added to some of the wells at 5 × 10⁵ cells/ml. Supernatants were collected at 96 h and IL-10 levels were measured by sandwich ELISA. The results shown are representative of three experiments in which spleens were harvested from 5–10 mice per group. B, Tumors were removed from five primed or unvaccinated C57BL/6 mice 11 days following s.c. challenge with live GL261 cells. At this point the tumors were regressing in the primed group while they were continuing to grow in the naive group. Tumor specimens were pooled in each group and used for RNA extraction, purification, and RT-PCR. Messenger RNA specific for an array of cytokines, CD3ε, and β-actin was measured as indicated. The experiment was repeated three times with similar results. C, The experiments described in B were repeated using primed and challenged wild-type, β₂-microglobulin-deficient and MHC class II-deficient C57BL/6 mice. The experiment was repeated twice with similar results.

IL-10 mRNA is up-regulated in tumor beds of vaccinated mice during remission

We next questioned whether IL-10 is expressed within gliomas undergoing rejection in vaccinated hosts. Tumors were removed from GL261- or mock-primed C57BL/6 mice on day 11 following challenge and RNA was isolated for RT-PCR with primers specific for a panel of cytokines and the housekeeping gene, β-actin. CD3ε chain mRNA was also measured in an attempt to assess the extent of T cell infiltration. At the time of harvest, gliomas were beginning to shrink in the vaccinated cohort, whereas they were continuing to grow in the control group. Messenger RNA encoding IL-10, but not IL-4, TGFβ or IFN-γ was up-regulated in the gliomas from vaccinated relative to control mice (Fig. 2⇑B). This was true despite low levels of CD3ε mRNA, suggesting that few infiltrating T cells were responsible for significant intratumor IL-10 expression.

It is likely that the IL-10 is expressed by infiltrating CD4⁺ T cells, because we detected it in the tumor beds of primed wild-type and β₂-microglobulin^−/−, but not MHC class II^−/−, C57BL/6 mice (Fig. 2⇑C). IL-10 mRNA was also not detected in GL261 cells that were freshly harvested from in vitro lines or from tumors grown in immunodeficient mice.

IL-10 plays a physiological role in glioma rejection

To definitively determine which cytokines actively participate in glioma rejection in vaccinated mice, we primed and challenged C57BL/6 mice deficient in either IL-12, IL-4, IFN-γ, or IL-10 with GL261 cells. By day 10 post challenge, tumors were significantly smaller on vaccinated IL-12^−/−, IFN-γ^−/−, and IL-4^−/− mice than the tumors borne by their unvaccinated counterparts (p < 0.005, Student’s t test). In fact, vaccinated IL-12^−/−, IFN-γ^−/−, and IL-4^−/− mice rejected GL261 tumors at a comparable rate to identically treated immunocompetent animals (Table I⇓ and data not shown). By contrast, glioma rejection was significantly delayed in primed IL-10^−/− mice as well as in wild-type mice treated with neutralizing Abs against IL-10.