Adoptive Transfer of Tumor-Specific Th2 Cells Eradicates Tumors by Triggering an In Situ Inflammatory Immune Response

Adoptive cell therapy (ACT) is a personalized immunotherapy for human cancer in which in vitro–grown autologous tumor-specific T cells are infused into patients (1). Animal models have been instrumental for the development of ACT. In the 1960s, initial experiments in rats and mice showed that tumor growth could be inhibited by the transfer of syngeneic lymphocytes from immunized donors (2, 3). Some years later, T cells were identified as being the cell type responsible for successful ACT (4). Eradication of tumors by ACT requires large numbers of transferred tumor-specific T cells. Therefore, a major advance was the development of a protocol to expand autologous tumor-infiltrating lymphocytes (TIL) in vitro by using the cytokine IL2 (5). These early animal studies set the stage for TIL-based ACT in humans (6–8). Objective response rates of 50%–70% have been reported for patients with metastatic melanoma refractory to standard therapies (9). To date, the majority of ACT trials have focused on tumor-specific cytotoxic CD8⁺ T cells, which can kill cancer cells directly. However, when tumor-specific CD8⁺ T-cell clones were used for ACT, only very limited therapeutic effect was observed. The TIL cultures that are used for ACT generally contain both CD4⁺ and CD8⁺ T cells. Therefore, it was proposed that cotransfer of CD4⁺ T cells may be essential to support the persistence and function of CD8⁺ T cells in vivo (8).

The potential of CD4⁺ T cells for ACT is being increasingly appreciated. Early animal studies demonstrated that transfer of tumor-specific CD4⁺ T cells could eradicate large sarcomas (10) and cure disseminated leukemia (11). It was shown that CD4⁺ T cells could cure cancer in the absence of CD8⁺ T cells (12, 13). Recent studies suggested that ACT protocols based on CD4⁺ rather than CD8⁺ T cells may be less complicated and more efficient (14–16). In humans, TIL cultures containing tumor-specific CD4⁺ T cells were successfully generated from 20%–40% of patients with metastatic melanoma and gastrointestinal cancers (17, 18). In pilot trials, ACT based on the transfer of tumor-specific CD4⁺ T cells was reported to induce tumor regression in metastatic melanoma, advanced colorectal cancer, and metastatic cholangiocarcinoma (18–20).

CD4⁺ T cells are very diverse in terms of cytokine production and effector functions. Several subsets have been described, including Th1, Th2, and Th17 cells, and it is not yet known which subset is most efficient for ACT. On the basis of mouse studies, which demonstrated the potential of IFNγ-producing Th1 cells for ACT (21–23), all the currently reported human trials were based on Th1 cell transfer (18–20). However, other Th subsets may potentially be very efficient as well and provide alternative avenues for immunotherapy. A report in mice concluded that Th17 cells were more potent than Th1 cells against melanoma (24). Th2 cells, which secrete IL4, IL5, and IL13 have been largely ignored for ACT and are generally considered to be detrimental for cancer immunotherapy (22, 25, 26). However, there are a few studies indicating that Th2 cells may be very efficient at eradicating cancer upon transfer (21, 27, 28).

This prompted us to reexamine the efficacy of tumor-specific Th2 cells for ACT in a mouse model for myeloma. We took advantage of idiotype (Id)-specific T-cell receptor (TCR)-transgenic mice that were homozygous for the severe combined immunodeficiency (SCID) mutation to prevent rearrangement of endogenous TCR chains (29). In these mice, tumor-specific CD4⁺ T cells recognize an Id peptide from the variable region of the immunoglobulin (Ig) light chain of the MOPC315 myeloma, presented on MHC class II molecules by professional antigen (Ag)-presenting cells (30–32). MOPC315 cells do not express MHC class II molecules, even in the presence of IFNγ (33). Using this well-characterized mouse model, we tested adoptive transfer of Id-specific Th2 cells for the treatment of MHC class II-negative myeloma. We found that the Th2 cells were excellent at eradicating Ag-producing tumor cells and conferred long-lasting immunity. The adoptive transferred Th2 cells were shown to induce a type II inflammation at the tumor site with a massive infiltration of M2-type macrophages producing arginase. In vivo blockade of arginase strongly inhibited myeloma eradication by Th2 cells.

Id-specific TCR-transgenic SCID mice (29) or SCID littermates on BALB/c background were bred in a heterozygous fashion. IFNγ^−/− BALB/c mice (from The Jackson Laboratory) were crossed with Id-specific TCR-transgenic SCID mice to obtain IFNγ^−/− Id-specific TCR-transgenic SCID mice (34). OVA-specific TCR-transgenic BALB/c mice (DO11.10) were obtained from Taconic, and wild-type BALB/c mice were obtained from The Jackson Laboratory. The transplantable BALB/c Id⁺ plasmacytoma MOPC315 (IgA, λ2³¹⁵) was obtained from the ATCC and propagated as in vitro growing cells. The Ag-loss MOPC315 variant (MOPC315.36) was kindly provided by Alexander Marks (University of Toronto, Toronto, Canada). The F9 lymphoma cells are A20 B-lymphoma cells transfected with the Id-containing Ig light chain from MOPC315 (35). All cell lines were authenticated in the laboratory by the means of Ag expression (33). Adult mice were injected subcutaneously (s.c.) in the interscapular region or in the flank, with1–1.6 × 10⁵ MOPC315 or F9 cells suspended in 100-μL PBS (Gibco) or 10⁵ MOPC315 cells suspended in 250-μL growth factor reduced Matrigel (BD Biosciences). Palpation was used to monitor subcutaneous tumor growth over time. Mice with a tumor diameter ≥10 mm were euthanized. In vivo inhibition of arginase was obtained by using Alzet osmotic pumps containing S-(2-boronoethyl)-l-cysteine (BEC, 4 mg/kg/d), a selective inhibitor of arginase I and II. The study was approved by the National Committee for Animal Experiments (Oslo, Norway).

Lymph nodes (LN) and spleens from Id-specific TCR-transgenic SCID mice (IFNγ^+/+ or IFNγ^−/−) were squeezed through a stainless steel sieve (Sigma). Released cells were stimulated with Id-containing 89-107 synthetic peptide (2 μg/mL) and irradiated (2,000 Rads) BALB/c splenocytes (used as Ag-presenting cells), in the presence of 20 U/mL IL2 and 20 U/ml IL4 (BD Biosciences/Peprotech). T-cell cultures were either used for ACT on day 8–10 or restimulated and used for ACT on day +16–20 (1–6 × 10⁶ Th2 cells i.v. per mouse). The adoptively transferred Th2 cells were >95% CD4⁺.

The following commercially available mAbs were used, conjugated with either fluorescein, phycoerythrin, allophycocyanin, or biotin: CD4 (clones RM4.5 or GK1.5), CD11b (M1/70), IFNγ (XMG1.2), IL4 (11B11), IL5 (TRFK5), IL17 (TC11-18H10), MHC class II I-A/I-E (M5/114.15.2), TCRCβ (H57-597), (BD Biosciences); CD11b (3A33), CD62L (MEL-14; Southern Biotechnology). The following mAbs were affinity-purified and, if needed, biotinylated in our laboratory: anti-transgenic-TCR-clonotype (GB113), anti-FcγRII/III (2.4G2; ATCC), anti-Idλ2³¹⁵ (Ab2-1.4), anti-IgCα (8D2), anti-CD4 (GK1.5), M315. Antibodies for IHC are listed in Supplementary Table S1. Serum myeloma protein (M315, IgA-λ2³¹⁵) levels were measured by sandwich ELISA employing Ab2-1.4 (anti-Id) as capture antibody, 8D2 (anti-IgCα) as detection antibody, and purified M315 as standard. Detection limit was 2 ng/mL.

The Matrigel plugs were excised and dissolved with 1 mg/mL collagenase type IV from Clostridium histolyticum and 0.3 mg/mL DNase I from bovine pancreas (both from Sigma), at 37°C for 30 minutes. Single-cell suspension from lymph nodes and Matrigel plugs were obtained by use of a stainless steel sieve. Erythrocytes were lysed with a hypotonic solution containing NH₄Cl, KHCO₃ and EDTA (Sigma). Unspecific binding was blocked by incubation with heat-inactivated (56°C, 30 minutes) 30% normal rat serum in PBS and 100 μg/mL anti-FcγRII/III mAb (clone 2.4G2). Cells were stained for 15 minutes on ice with specific mAbs in PBS supplemented with 0.5% BSA (Biotest). For intracellular cytokine detection, cells were stimulated with phorbol myristate acetate and ionomycin (Sigma) in vitro for 4 hours in cell culture medium supplemented with monensin, prior to staining with Cytofix/Cytoperm Plus reagents (BD Biosciences) and specific mAbs. Quadruple-stained cells were analyzed on aFACSCalibur instrument with CellquestPro (BD Biosciences) and FlowJo (Tree Star, Inc.) softwares. Cell sorting was done on a FACSAria instrument (BD Biosciences).

Cytokine levels in Matrigel supernatants, cell lysates, and serum were measured using single-plex or multiplex bead assays from Bio-Rad Laboratories (Bio-Plex 23-plex, 9-plex, 8-plex, TGFβ 3-plex, and IL25, IL27, ICAM-1 single-plex assay panels), according to the manufacturer's instructions. Samples were analyzed as singlets and standards in duplicates, using a Luminex-100 instrument with Bio-Plex Manager 6.0 software (Bio-Rad Laboratories).

Statistical analysis was performed with the Mann–Whitney test and the log-rank test (survival) using GraphPad Prism 5 software. P < 0.05 was considered significant.

More details on IHC are described in the Supplementary Data.

TCR-transgenic SCID mice were used as source of naïve CD4⁺ T cells recognizing a tumor-specific Id Ag from the MOPC315 myeloma (29). Id-specific Th2 cells were generated by short-term in vitro cultures using Th2-polarizing conditions. To investigate the therapeutic potential of Th2-polarized CD4⁺ T cells, SCID mice were inoculated subcutaneously with MOPC315 myeloma cells. Within the next three days, some of the mice were treated intravenously with Th2 cells, whereas control mice were left untreated (Fig. 1A). Mice that received Id-specific Th2 cells did not develop tumors, while untreated mice succumbed to cancer (Fig. 1B). We envisaged the possibility that Th2-based immunotherapy may be mediated by Th2 cells that had converted in vivo into IFNγ-producing Th1-like cells. To address this issue, we generated IFNγ^−/− Id-specific Th2 cells by using IFNγ^−/− Id-specific TCR-transgenic SCID mice. IFNγ^−/− Id-specific Th2 cells lack the gene coding for IFNγ and thereby cannot produce this cytokine. Transfer of Th2 IFNγ^−/− cells eradicated myeloma in both wild-type and IFNγ^−/− SCID recipient mice, demonstrating that IFNγ was not required for Th2 ACT (Fig. 1C). Intracellular flow cytometry analysis of the tumor-specific Th2 IFNγ^−/− cells confirmed a typical Th2 phenotype with production of IL4 and IL5, but no IFNγ or IL17 (Fig. 1D, top). Additional characterization of cell lysates by Luminex technology revealed that the Th2 cells also produced a vast range of other cytokines, including IL2, IL13, and CCL11/eotaxin-1 (Fig. 1D, bottom).

Additional flow cytometry analysis of the Id-specific IFNγ^−/− Th2 cells revealed a typical effector/memory phenotype with upregulation of CD69, CD71, and CD278 (ICOS), and high levels of CD44, among other molecules (Fig. 2A). Notably, the phenotype of the Id-specific IFNγ^−/− Th2 cells bore many similarities with that of in vivo–differentiated Id-specific Th1 cells able to eradicate MOPC315 (36). The Th2 cell cultures used for ACT contained >95% CD4⁺ T cells (data not shown). Administration of depleting, anti-CD4 mAb into Th2-treated mice abrogated tumor protection, confirming that therapy was dependent on CD4⁺ T cells (Fig. 2B). Transfer of Th2 cell numbers ranging from 1–5 × 10⁶ was efficient at eradicating myeloma in most recipient mice (Supplementary Fig. S1). Collectively, the data show that adoptive transfer of tumor-specific CD4⁺ T cells with a bona fide Th2 phenotype efficiently eradicates myeloma cells. Therapy was successful using IFNγ^−/− SCID mice, thereby excluding any contribution from B cells, cytotoxic CD8⁺ T cells, natural killer (NK) T cells, or IFNγ.

ACT was efficient when the Th2 cells were transferred up to 3 days after tumor challenge (Fig. 3A, left) but was not successful at later time points, presumably because the tumor load was too high (data not shown). Measurement of serum levels of the IgA mAb secreted by MOPC315 (the so-called myeloma protein) confirmed the complete elimination of the tumor cells in most treated mice (Fig. 3A, right). To test the specificity of the antitumor response, we used an Ag-loss variant of MOPC315, which does not produce the IgA containing the tumor-specific Id Ag. Th2 ACT failed to protect mice inoculated with Ag-loss MOPC315, demonstrating that the therapy was Ag-specific (Fig. 3B). In this experiment, we used as untreated controls DO11.10 mice, which are transgenic for a TCR with an irrelevant specificity, recognizing ovalbumin (OVA). DO11.10 mice, which contain high numbers of OVA-specific CD4⁺ T cells, succumbed to cancer upon inoculation with either MOPC315 or Ag-loss MOPC315 (Fig. 3B). To expand our findings to another tumor type, we conducted a tumor challenge experiment with F9 lymphoma cells. F9 cells are A20 lymphoma cells that have been transfected with the tumor-specific Id Ag from MOPC315 (35). ACT with Id-specific Th2 cells mediated protection against F9 cells (Fig. 3C). Thus, Th2 ACT is Ag-specific and efficient in mouse models for both myeloma and lymphoma.

To investigate whether long-term immunity was induced, mice successfully treated by ACT were rechallenged with MOPC315 >40 days after the first tumor cell inoculation. The mice remained fully protected (Fig. 4A). In contrast, mice that were rechallenged with Ag-loss MOPC315 cells succumbed to cancer (Fig. 4B). The long-lasting immunity did not require IFNγ, as IFNγ^−/− mice treated with Th2 IFNγ^−/− cells also were protected against a second tumor challenge (Fig. 4C). Mice from the first rechallenge experiment (shown in Fig. 4A) were euthanized at day +98 after ACT (44 days after rechallenge). In both blood and lymph node, small but distinct Id-specific CD4⁺ T-cell populations were identified, revealing long-term persistence of the transferred Th2 cells (Fig. 4D). In the blood, the transferred Id-specific CD4⁺ T cells did not express the CD62L cell surface marker, whereas two distinct populations of CD62L⁺ and CD62⁻ cells were identified in the lymph node (Fig. 4D). These data indicate the presence of both CD62L⁺ central memory and CD62L⁻ effector memory Id-specific CD4⁺ T cells in the recipient mice 3 months after Th2 transfer. Thus, Th2 ACT provides long-lasting, Ag-specific immunity, which is independent of IFNγ.

To clarify the mechanism whereby Th2 ACT eradicates myeloma, we took advantage of a method that we have previously developed consisting in subcuatneous injection of tumor cells embedded in a Matrigel collagen gel (32). This strategy allows the analysis of the immune cells that are recruited towards the injected tumor cells as well as the quantification of locally secreted cytokines (32, 34). MOPC315 cells embedded in Matrigel were successfully eradicated by transfer of Id-specific Th2 IFNγ^−/− cells (Fig. 5A). We proceeded to analyze by flow cytometry the cells that infiltrated the tumor site. SCID mice were injected subcutaneously with MOPC315 in Matrigel and treated the same day with Id-specific Th2 IFNγ^−/− cells or left untreated. At day +8, the mice were euthanized, and the Matrigel plugs were dissected out for analysis. In Th2-treated mice, the Matrigel plugs contained large amounts of infiltrating cells expressing the myeloid marker CD11b (Fig. 5B). There were significantly more Matrigel-infiltrating CD11b⁺ cells in Th2-treated mice compared with untreated controls (Fig. 5B). Figure 5C shows that Matrigel-infiltrating CD11b⁺ cells expressed both CD206 (mannose receptor) and intermediate levels of surface MHC class II molecules, consistent with an alternatively activated (M2) macrophage phenotype (25, 37). MHC class II levels on macrophages are useful markers of M1/M2 polarization because IFNγ induces strong upregulation of MHC class II molecules on target cells. Therefore, IFNγ-activated M1 macrophages harbor high surface levels of MHC class II. In contrast, M2 macrophages, which are by definition activated by other cytokines than IFNγ, may only have intermediate MHC class II levels. We have previously reported that Id-specific TCR-transgenic SCID mice injected subcutaneously with MOPC315 developed a Th1/M1 immune response with a strong upregulation of MHC class II on Matrigel-infiltrating macrophages (32). With that in mind, we compared surface MHC class II levels on Matrigel-infiltrating macrophages from Th2-treated SCID mice, untreated SCID mice, and Id-specific TCR-transgenic SCID mice, respectively, at day +8 after MOPC315 inoculation and ACT. As expected, Matrigel-infiltrating macrophages from TCR-transgenic SCID had much higher surface MHC class II levels than those from untreated nontransgenic SCID mice (Fig. 5D). Notably, in Th2-treated mice, Matrigel-infiltrating CD11b⁺ cells showed no upregulation of MHC class II consistent with an M2 macrophage polarization (Fig. 5D). Collectively, the data show that Th2 ACT induces a massive recruitment of macrophages with an M2 phenotype at the tumor site.

Tissue section staining was used to visualize the immune response at the tumor site. Mice were injected subcutaneously with MOPC315 in Matrigel and either treated with Th2 ACT or left untreated for comparison. At day +4 after Th2 cell transfer, a strong inflammatory infiltrate was apparent in the dermis in close proximity to the Matrigel plug containing the myeloma cells (Fig. 6A). Such inflammation was absent in untreated mice (Fig. 6B). Fluorescence microscopy was employed to identify the cells involved, using CD3, CD138, and F4/80 as specific markers for T cells, myeloma cells, and macrophages, respectively. Hoechst was added to visualize cell nuclei. At day +4 of ACT, Matrigel plugs contained CD138⁺ myeloma cells grouped in islets, presumably resulting from cell division in situ (Fig. 6C and D). In Th2-treated mice, Matrigel plugs were infiltrated by numerous F4/80⁺ macrophages, some of those making contact with the myeloma cells (Fig. 6C). In contrast, few Matrigel-infiltrating F4/80⁺ macrophages were observed in untreated mice (Fig. 6D). Notably, virtually all nucleated cells in Matrigel plugs could be classified as being either CD138⁺ myeloma cells or F4/80⁺ macrophages. Therefore, these data confirmed that the Matrigel-infiltrating CD11b⁺ cells detected by flow cytometry (Fig. 5B) were indeed F4/80⁺ macrophages. In Th2-treated mice, numerous CD3⁺ T cells were observed surrounding the Matrigel plugs (Fig. 6E). A few T cells had infiltrated the Matrigel plugs (Fig. 6F). No T cells could be detected in untreated SCID mice (data not shown). Within 8 days after Th2 ACT, virtually all myeloma cells were eradicated (Supplementary Fig. S2A and S2B). In contrast, extensive proliferation of tumor cells could be observed in untreated mice (Supplementary Fig. S2C and S2D). Locally secreted cytokines within the Matrigel plugs were quantified at day +7 after cell transfer. In comparison with untreated SCID mice, Th2 ACT was associated with elevated local levels of many cytokines, including the proinflammatory cytokines IL1α, IL1β, and TNFα, as well as the prototypic Th2 cytokines IL4, IL5, and IL13 (Fig. 6G). Furthermore, several chemokines, such as CCL2, CCL3, CCL4, CCL5, CCL11 (eotaxin-1), and CCL20, were also found at higher concentrations in Th2-treated mice (Fig. 6G). In summary, Th2 ACT induced a type II inflammatory reaction at the tumor site with massive recruitment of macrophages and T cells. Complete myeloma eradication was achieved by day +8.

Mast cells and eosinophils are two types of innate immune cells, which are often associated with type II inflammatory immune responses. Therefore, we envisaged the possibility that these cells may be involved in Th2 ACT. Tissue section staining revealed the presence of a few scattered mast cells, identified by expression of mast cell protease-6 (MCP-6), at the periphery of Matrigel plugs containing MOPC315 cells. However, no difference in mast cell density was observed between Th2-treated and untreated mice (Supplementary Fig. S3A and S3B). In the dermis close to the Matrigel plugs, a modest infiltration of eosinophils was observed in Th2-treated mice, but not in control mice, suggesting that eosinophils may contribute to myeloma eradication (Supplementary Fig. S3C–S3F). However, direct killing of myeloma cells by eosinophils is unlikely, because no eosinophil infiltration was seen inside the Matrigel plugs (Supplementary Fig. S3G and S3H). The observed increased local levels of CCL11/eotaxin-1, which is a chemoattractant for eosinophils (Fig. 6G), represent a likely explanation for the presence of eosinophils around Matrigel plugs in Th2-treated mice. Thus, Th2 ACT is associated with a modest recruitment of eosinophils, which may potentially contribute to myeloma eradication, although direct involvement in killing is unlikely.

Arginase expression is a characteristic feature of M2-polarized macrophages (38, 39). Immunofluorescence microscopy revealed that Matrigel-infiltrating macrophages in Th2-treated mice expressed arginase (Fig. 7A), whereas most macrophages (identified as Hoechst⁺, CD138⁻ cells) in the tissue samples from untreated mice stained negatively (Fig. 7B). Double staining confirmed that the arginase-positive cells in the Matrigel plugs were F4/80⁺ macrophages (Fig. 7C). Cell quantification confirmed the dominance of arginase-positive macrophages in the inflammatory response, in comparison with eosinophils and mast cells (Fig. 7D). Previous in vitro studies indicated that arginase produced by activated macrophages may inhibit the growth of tumor cells (40, 41). To investigate a possible role of arginase in Th2 ACT, a selective arginase inhibitor (BEC) was administered to Th2-treated mice. Although in vivo arginase inhibition did not completely abolish myeloma eradication, tumor protection was severely reduced (Fig. 7E). Thus, Th2 ACT is associated with induction of arginase in the macrophages that are recruited towards the tumor cells, and in vivo inhibition of arginase strongly suggests that this enzyme plays a key role in myeloma eradication.

In this report, we show that complete eradication of myeloma and lymphoma cells could be achieved in mice by ACT with tumor-specific Th2 cells. The therapy was Ag-specific, and provided protection against tumor recurrence. The potential of Th2 cells for ACT is supported by a few previous reports (21, 27, 28). Some limitations in these studies may explain why Th2 cells remained neglected in ACT. In one of the report, the sarcoma-specific Ag was not defined (27). The two other studies used OVA as a surrogate tumor-specific Ag, which may seem quite artificial (21, 28). OVA-specific Th2 cells were effective at treating mice with MHC class II-positive lymphomas, but the therapeutic relevance of this finding is questionable as most cancers lack MHC class II (21). In the last report, transferred OVA-specific Th2 cells could reduce the tumor burden of OVA-expressing melanoma, but the survival benefits were unclear (28). Our data demonstrate that bona fide Th2 cells recognizing a natural tumor-specific Ag can be used to eradicate cancer.

Transferred Th2 cells were shown to induce a strong type II inflammatory immune response at the tumor with massive infiltration of macrophages and increased levels of both proinflammatory (IL1α, IL1β, TNFα) and Th2-associated (IL4, IL5, IL13) cytokines. The role of inflammation in cancer is controversial because both tumor-promoting and tumor-suppressive effects of inflammation have been reported. To reconcile the opposite views, we have recently proposed that there are different types of inflammation, and that certain types, such as inflammation driven by tumor-specific Th1 cells, confer protection against cancer (34, 42, 43). In the absence of sufficient numbers of T cells, inflammatory mediators may instead promote tumor growth, for example, by inducing neoangiogenesis (42). Our data strongly suggest that inflammation driven by tumor-specific Th2 cells also protects against cancer.

Th2 cells are potent immune cells that can eliminate large pathogens such as intestinal nematode parasites (44, 45). Th2 cells are not directly cytotoxic but instead mediate their effector functions by releasing cytokines that activate other immune cells. For the clearance of nematodes, Th2 cells produce IL4 and IL13 to activate mast cells, eosinophils, B cells, and macrophages (44, 45). We found that Th2 ACT was associated with increased levels of IL4, IL5, and IL13, at the tumor site. These prototypic Th2 cytokines, most likely produced locally by the transferred Th2 cells, may have a pivotal role in mediating cancer eradication. IL4 has previously been shown to exert strong antitumor effects in experiments with tumor cells engineered to produce IL4 by transfection (46). Evidence was provided for a key role of eosinophils, although a contribution of macrophages could not be excluded (46). In another report, tumor protection conferred by a prophylactic cancer vaccine against the B16 melanoma was shown to depend on both IL4 and IL5 and a role for eosinophils was also suggested (47). Finally, in one of the previous reports of Th2 ACT, therapeutic effect was lost in eotaxin-deficient mice and a role for degranulating eosinophils was proposed (28). Thus, there is experimental support in the literature for antitumor effects of the Th2 cytokines IL4 and IL5.

By using mice with selective immunodeficiencies, we could show that Th2 ACT does not require B cells, NK T cells, CD8⁺ T cells, or IFNγ. Notably, the use of SCID recipient mice bears a resemblance to human patients who are treated with immunodepleting chemotherapy before ACT (8, 9). Myeloma eradication by Th2 cells was associated with a massive infiltration of macrophages and a more modest recruitment of eosinophils. Macrophages were the only immune cells coming in close contact with the cancer cells consistent with a key role of macrophages in tumor eradication. Eosinophils may also play a role, although direct involvement in killing is unlikely because they remained at the tumor periphery. The transferred Th2 cells were shown to induce arginase production by tumor-infiltrating macrophages. Arginase catalyzes the breakdown of the amino acid arginine to ornithine and urea. In vivo blockade of arginase strongly inhibited myeloma eradication by Th2 cells. Antitumor function of arginase produced by activated macrophages has been previously suggested by in vitro studies. For example, macrophages activated with zymosan or lipopolysaccharide (LPS) were cytotoxic to tumor cells in vitro and the effect was associated with arginase activity in macrophages (40). Moreover, cytotoxicity was abrogated by addition of exogenous l-arginine. (40). In another report, IL4-activated macrophages were able to inhibit the in vitro proliferation of B16 melanoma cells and the effect could be blocked by using arginase inhibitors such as BEC or by addition of arginine (41). These in vitro data are consistent with a requirement for arginine for protein synthesis in cancer cells and a detrimental effect of arginine deprivation. Our findings provide support for the in vivo relevance of these previous in vitro observations.

Collectively, our data strongly suggest that transferred Th2 cells collaborate with tumor-infiltrating macrophages in cancer elimination. Like most malignant cells, MOPC315 does not express MHC class II molecules and cannot directly be recognized by Th2 cells (33). Macrophages are professional Ag-presenting cells that may ensure optimal processing and presentation of tumor-derived Ag to Th2 cells. Once a tumor-specific Th2 cell recognizes its cognate Ag presented on MHC class II molecules by macrophages, it will secrete cytokines and thereby trigger arginase production by macrophages, resulting in cancer eradication.

The protective role of tumor-specific Th1 cells against cancer is well accepted (15, 16, 18–20, 32, 34). In contrast, Th2 cells are often referred to as being detrimental and even to promote tumor growth (25, 26). The concept that one class of immune response (Th1) is beneficial while another class (Th2) is harmful against cancer is actually much older than the Th1/Th2 nomenclature. Early animal studies revealed the importance of cell-mediated (Th1) immunity against cancer (2, 3). Already in the 1960s, it was proposed that the occurrence of cancer may be due to an inappropriate humoral (Th2) immune response (reviewed in ref. 26). The association between the presence of Th1 cells in human tumors and good patient prognosis is now well established (48). In contrast, there is very little documentation for the presence of Th2 cells in tumors. According to a recent review, an association between Th2 infiltrate in human tumors and poor patient prognosis is supported by only 4 studies (48). Two other studies showed no association while two reports concluded that intratumoral Th2 cells were in fact associated with good prognosis (48). Thus, there is very little experimental evidence supporting the hypothesis of Th2 cells having a detrimental role in cancer. Our data suggest instead that Th2 cells may be very useful in cancer immunotherapy.

Mirroring the Th1/Th2 paradigm, a M1/M2 nomenclature was proposed for macrophages (39). Macrophages activated by Th1-derived IFNγ are called M1 macrophages, whereas macrophages activated by Th2 cytokines, or associated with Th2 immune responses, are called M2 macrophages (25, 39). Because Th2 cytokines induce the expression of arginase and CD206 (mannose receptor) in macrophages, these two molecules are widely used as markers for M2 macrophages (37, 38). Notably, expression of arginase and CD206 is often observed in tumor-infiltrating macrophages, which are therefore named M2 or M2-like, although Th2 cells are rarely seen in tumors (25, 39, 49). M2 macrophages with arginase activity have been associated with tumor progression (25, 49). Our data strongly suggest that arginase-producing M2 macrophages may also participate in tumor eradication, in the presence of tumor-specific Th2 cells.

Multiple myeloma remains an incurable malignancy and new treatment modalities are needed. ACT is particularly suitable for multiple myeloma because the Id of the secreted myeloma protein represents an easily identifiable tumor-specific Ag. Several investigators have demonstrated that Id-specific Th1 and Th2 cells can be isolated from multiple myeloma patients and expanded in vitro (reviewed in ref. 50). Our data suggest that ACT with tumor-specific Th2 cells may represent a very efficient immunotherapy for multiple myeloma, lymphoma, and possibly other cancers. Notably, Th2 ACT was most efficient when performed 0–3 days after myeloma inoculation, that is, before large tumors had formed. Such an experimental situation resembles a state of minimal residual disease that should be optimal for Th2 ACT.

No potential conflicts of interest were disclosed.

Conception and design: K.B. Lorvik, B. Bogen, A. Corthay

Development of methodology: K.B. Lorvik, C. Hammarstrom, M. Fauskanger, A. Corthay

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): K.B. Lorvik, C. Hammarstrom, M. Fauskanger, O.A.W. Haabeth, M. Zangani, G. Haraldsen, A. Corthay

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): K.B. Lorvik, C. Hammarstrom, M. Fauskanger, O.A.W. Haabeth, M. Zangani, B. Bogen, A. Corthay

Writing, review, and/or revision of the manuscript: K.B. Lorvik, C. Hammarstrom, M. Fauskanger, G. Haraldsen, B. Bogen, A. Corthay

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): M. Zangani, G. Haraldsen, A. Corthay

Study supervision: B. Bogen, A. Corthay

We thank Inger Øynebråten and Michael Bodd for critical reading of the manuscript.

This work was funded by grants from The South-Eastern Norway Regional Health Authority, The Research Council of Norway, the Multiple Myeloma Research Foundation, and the Norwegian Cancer Society.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.