It has been demonstrated that γδ T cells accumulating in early tumor lesions and those purified from spleen cells of tumor-bearing mice attenuate the activity of CTLs and NK cells. We, therefore, investigated whether depletion of γδ T cells from early lesions of tumors results in restoration of CTL and NK cell activities and subsequent regression of tumors. A daunomycin-conjugated anti-γδTCR mAb UC7-13D5 (Dau-UC7) was prepared to efficiently deplete γδ T cells. An in vitro study revealed that Dau-UC7 specifically lysed γδTCR+ cells and effectively inhibited splenic γδ T cells from tumor-bearing mice to produce cytotoxic cell-suppressive factors. Furthermore, intralesional injections of Dau-UC7 at an early stage of tumor development led to augmentation of tumor-specific CTL as well as NK cell activities and to the resultant regression or growth inhibition of the tumors. On analysis of cytokine profile, γδ T cells transcribed mRNAs for IL-10 and TGF-β, but not IL-4 or IFN-γ, suggesting the T regulatory 1-like phenotype. Finally, a blocking study with mAbs showed that the inhibitory action of γδ T cells on CTLs and NK cells was at least partly mediated by IL-10 and TGF-β. These results clearly demonstrated the novel mechanism by which T regulatory 1-like γδ T cells suppress anti-tumor CTL and NK activities by their regulatory cytokines in early tumor formation.
It has been well documented that in tumor-bearing mice, both tumor cells and certain populations of lymphocytes produce suppressive factors that may abrogate tumoricidal immunity, including TGF-β, IL-10, and vascular endothelial growth factor (1, 2, 3, 4). Recently, TGF-β- and IL-10-secreting CD4+ lymphocytes were classified into T regulatory (Tr)41 cells (5), whose suppressive effects on tumor immunity remain to be elucidated. In vivo neutralization of these cytokines or elimination of cells producing these suppressants in tumor-bearing individuals is one theoretical approach to successful tumor immunotherapy. In this strategy, mAbs are a powerful tool and have been used in various ways. For example, the in vivo administration of anti-TGF-β mAb to tumor-bearing mice resulted in the attenuation of tumor cell growth (6, 7). Likewise, depletion with mAb of suppressor T (Ts) cells that depress anti-tumor effector lymphocytes may down-modulate tumor growth. However, there have been few reports about whether the inhibition of tumor cell growth by neutralization or cell elimination is derived from restoration of the functions of cytotoxic effector cells, i.e., CTLs and NK cells. Difficulty in dissecting this possibility is due to the lack of useful methods to distinguish suppressor cells from normal lymphocytes and to deplete the function of those lymphocytes. Only one report has shown that abolishment of CD4+ Ts cells with CD4-specific mAbs in L5178Y lymphoma-bearing mice leads to vigorous generation of tumor-specific CTLs and subsequent regression of the established tumor (8).
On the one hand, mAbs specific for αβTCR or γδTCR are used in vivo for abrogation of the target recognition by T cells because of their ability to render TCR internalized or for specific depletion of T cell subpopulations by apoptosis or Ab-dependent cell-mediated cytotoxicity (8, 9, 10, 11, 12). On the other hand, however, some TCR-specific mAbs have been used in vitro as stimulators of T cells. In normal mice administered a TCR-specific mAb, some TCR-null T cells survive and remain constant in number after a certain period (13). In tumor-bearing mice, Ts cells rather than CTLs or Th cells are predominantly primed and accumulate in the tumor lesions (14, 15). It is possible that treatment of tumor-bearing hosts with intact TCR-specific mAbs, originally intended to eliminate Ts cells, instead enhances Ts cell functions. Therefore, the effect of treatment with anti-TCR mAbs is thought to be ambivalent and thus may lead to misinterpretation of the results. Several studies have demonstrated the therapeutic effectiveness of daunomycin-conjugated Abs on the direct killing of tumor cells (16, 17). Daunomycin is internalized by cells following binding of Abs to the cell surface. Thus, daunomycin conjugates with mAbs specific for TCR on Ts cells may be an efficacious and reliable tool for complete depletion of Ts cells.
γδ T cells infiltrating at an early stage of tumor development of MM2 (mammary tumor cell line), MH124 (hepatoma cell line), and B16 (melanoma cell line) suppress CTL and NK cell activities not only by cell-cell interaction but also by locally producing suppressive factors (18, 19). The percentage of these γδ T cells in tumor cell suspensions peaks on days 5–7 after tumor inoculation and gradually decreases thereafter (19). Therefore, elimination of these cells at the tumor sites may result in tumor regression by relaxing CTLs and NK cells. The purpose of this study was to investigate whether daunomycin-conjugated anti-γδTCR mAb can effectively damage the suppressor functions of γδ T cells, whether intralesional treatment with this conjugates at an early stage can lead to tumor regression by restoring tumor-specific CTL and NK cell functions, and whether these γδ T cells produce regulatory cytokines such as IL-10 and TGF-β. The results show that successful elimination of γδ T cells producing Tr1-type cytokines results in tumor regression by the actions of CTLs and NK cells.
Materials and Methods
Mice and tumor cells
Seven to nine-week-old male C3H/He and C57BL/6 (B6) mice were obtained from Japan SLC (Hamamatsu, Japan). MM2, MH134, B16, and YAC-1 tumor cells were used in this study. MM2, MH134, and B16 were mammary tumor cell lines of C3H/He, a hepatoma cell line of C3H/He, and a melanoma cell line of B6, respectively. MM2 cells were maintained i.p. in C3H/He. MH134 and B16 cells were maintained by culturing them in DMEM (Nissui, Tokyo, Japan) supplemented with 10% FCS. In in vivo experiments, MM2 and MH134 cells (2 × 105 cells/mouse) were inoculated into C3H/He mice i.p. and s.c., respectively. The same number of B16 cells were injected s.c. into B6 mice. YAC-1 cells are an NK-sensitive cell line that was cultured in DMEM supplemented with 10% FCS and used in in vitro assays.
mAbs and chemical substances
Anti-γδTCR mAb (UC7-13D5)-producing hybridoma was a gift from Dr. Bluestone (Chicago University, Chicago, IL). UC7-13D5 and anti-αβTCR mAb (H57-597) were purified from hybridoma culture supernatants by affinity column chromatography with anti-hamster IgG-Sepharose after ammonium sulfate precipitation. Anti-hamster IgG-Sepharose was prepared by covalent coupling of anti-hamster IgG sheep polyclonal Abs (Organon Teknika, West Chester, PA) with cyanogen bromide-activated Sepharose 4B (Pharmacia Biotech, Uppsala, Sweden). Purified forms of anti-CD4 mAb (GK1.5), anti-CD8 mAb (53-6.7), and hamster IgG were obtained from PharMingen (San Diego, CA). Anti-IL-10 and anti-TGF-β1, -2, and -3 neutralizing mAbs, rIL-10, and rTGF-β1 were purchased from Genzyme (Cambridge, MA). Daunomycin was purchased from Calbiochem-Novabiochem (La Jolla, CA). Sodium periodate and sodium borohydride were purchased from Sigma (St. Louis, MO).
Daunomycin (40 mg/ml) was oxidized with 0.1 M sodium periodate (in PBS) at 20°C for 20 min in a dark room (final 1-ml volume). Immediately, 100 μl of 1 M glycerol was added to the mixture to stop the reaction, and incubation proceeded at 20°C for 30 min. Supernatant obtained after centrifugation of the mixture was used as an oxidized daunomycin solution. One milliliter of 5 mg/ml UC7-13D5 or H57-597 mAb or hamster IgG (in 0.15 M potassium carbonate buffer, pH 9.5) and the oxidized daunomycin solution were mixed and incubated at 20°C for 2 h. After centrifugation, the supernatant (2 ml) containing daunomycin-conjugated Ab was reduced by the addition of 0.6 mg of sodium borohydride and incubation at 4°C for 2 h and was subjected to gel filtration (Bio-Gel P-100, Bio-Rad, Hercules, CA) to separate daunomycin-bound Ab from free daunomycin. Conjugate fractions were collected and used in in vitro and in vivo assays. By assessment of 495 nm (daunomycin) and 280 nm (Ab) absorbances, 50 μg of daunomycin-conjugated UC7-13D5 (Dau-UC7), H57-597 (Dau-H57), and hamster IgG (Dau-IgG) were bound with 2 μg of daunomycin.
Preparation of lymphocytes
T cell-enriched fractions of splenocytes from normal and MM2-bearing C3H/He mouse splenocytes were prepared using a generally established method. Spleen cells were hemolyzed with 0.17 M ammonium chloride. After washing three times, the cells were incubated on a plastic dish in RPMI 1640 medium (Nissui) supplemented with 10% FCS at 37°C for 1 h to remove dish adherent cells. The dish-nonadherent cells were collected by gentle shaking and subjected to nylon wool column. The T cell-enriched fraction passed through a nylon wool column was used for αβ or γδ T cell preparation. The enriched T cells were incubated with αβTCR (H57-597)- or γδTCR (UC7-13D5)-specific mAb at 4°C for 30 min. After washing three times, Ab-bound cells were mixed with anti-hamster IgG-conjugated magnetic beads (Dynal, Oslo, Norway) at a ratio of three beads per cell at 4°C for 1 h on a rocking shaker. Anti-hamster IgG-conjugated beads were prepared by coupling the anti-hamster IgG (Organon Teknika) with tosyl-activated magnetic beads (Dynal) according to the Dynal manual. Cells bound with beads were collected with a magnet and cultured overnight in RPMI 1640 medium supplemented with 10% FCS to separate cells from beads. αβ and γδ T cells purified by this manipulation were confirmed to be >96% pure by flow cytometric analysis using FITC-conjugated αβTCR- or γδTCR-specific mAb (PharMingen, San Diego, CA).
Tumor-infiltrating lymphocytes (TILs) were separated from B16 and MH134 lesions. B16 and MH134 tumor cell suspensions were prepared on day 7 after s.c. inoculation (2 × 105/mouse) from 50 lesions. Ten milliliters of B16 and MH134 cell suspensions (1 × 105 cells/ml in PBS supplemented with 10% FCS) applied on 5 ml of Histopaque 1083 (Sigma) were subjected to centrifugation at 1000 × g for 30 min at 20°C. The cells at the interface were collected, washed three times with DMEM, and used as B16 and MH134 TILs. For separation of γδ T cells from MM2-infiltrating lymphocytes, 2 × 105 MM2 cells were inoculated i.p. into C3H/He mice. Seven days after inoculation, ascites containing MM2 and tumor-infiltrating lymphocytes were collected and then diluted with DMEM containing 10% FCS. After washing three times, the cells were suspended in DMEM and subjected to weak centrifugation (500 × g, 10 s, five times). Approximately 90% of large MM2 cells were precipitated by this manipulation. The remaining cells in culture supernatants were used as MM2-tumor infiltrating lymphocytes. γδ T cells were separated from MM2 tumor-infiltrating lymphocytes using magnetic beads as described above. Anti-H-2b and anti-MM2 CTLs were prepared from T cell-enriched splenocytes of C3H/He mice immunized with B6 lymphocytes and MM2 regressor mice, respectively, as described previously (18).
Preparation of lymphocyte culture supernatants
Splenic γδ T cells purified from spleens of normal or MM2-bearing mice and from MM2 tumor-infiltrating lymphocytes were cultured in RPMI 1640 supplemented with 10% FCS and rIL-2 (5 U/ml) at 37°C for 3 days. The expanded cells were recultured in RPMI 1640 supplemented with 10% FCS in a 24-well plate (Corning, Corning, NY; 1 × 106 cells/well) for 24 h. After centrifugation, each culture supernatant was collected and added to the anti-H-2b CTL assays at a 50% volume. Otherwise, to examine the effect of daunomycin conjugates on the ability of γδ T cells to produce suppressive factors, γδ T cells separated from MM2 tumor-infiltrating lymphocytes were cultured in RPMI 1640 supplemented with 10% FCS and rIL-2 (5 U/ml) for 3 days. The expanded cells were incubated with Dau-UC7, Dau-H57, Dau-IgG, UC7-13D5, H57-597, hamster IgG, or daunomycin at varying concentrations at 4°C for 1 h. After washing three times, these treated cells were cultured in RPMI 1640 supplemented with 10% FCS in 24-well plates (1 × 106 cells/well). Culture supernatants obtained by this manipulation were also added to the anti-H-2b CTL assays at a 50% volume.
Cytotoxicity assay and cytotoxicity inhibition assay
To test the cytotoxicity of daunomycin-conjugated Ab, purified αβ and γδ T cells (2 × 104 cells/well) were cultured and expanded in an anti-CD3 mAb-immobilized 24-well culture dish with RPMI 1640 supplemented with 10% FCS and 50 U/ml rIL-2. The propagating cells (5 × 106 cells/ml) were radiolabeled with RPMI 1640 containing 10% FCS and 200 μCi Na[51Cr] (DuPont-New England Nuclear, Boston, MA) for 1 h at 37°C. After washing four times, 51Cr-labeled αβ and γδ T cells were used as target cells for drug-conjugated Abs. Daunomycin-conjugated H57-597, UC7-13D5, or hamster IgG was added at varying concentrations to each well containing target cells (1 × 104 cells/200 μl). The mixtures were incubated at 37°C for 12 h. To examine in vitro CTL induction in TILs, MM2, MH134, and B16, TILs were treated with daunomycin-conjugated mAbs (100 ng/ml daunomycin-5 μg/ml Abs) at 37°C for 4 h and washed three times with DMEM. Dau-Ab-treated TILs were cultured with rIL-2 (100 U/ml)-containing medium at 37°C for 5 days. The expanded cells were subjected to the CTL assays against 51Cr-labeled tumor target cells. For cytotoxicity test of CTLs against tumor cells, MM2, MH134, and YAC-1 cells (5 × 106 cells/ml) were radiolabeled with RPMI 1640 containing 10% FCS and 200 μCi of Na[51Cr] for 1 h at 37°C. Varying numbers of splenic T cells from mice bearing MM2 or those regressing MM2 were mixed with 51Cr-labeled MM2, MH134, or YAC-1 target cells (1 × 104 cells) at a final volume of 200 μl and incubated for 12 h at 37°C at varying ratios. Cytotoxicity inhibition assays were performed to investigate the lytic mechanisms of cytotoxic effector cells. Varying concentrations of anti-αβTCR mAb (H57-597), anti-γδTCR mAb (UC7-13D5), anti-CD4 mAb (GK1.5), or anti-CD8 mAb (53-6.7) was added in cytotoxicity assays of splenic T cells from MM2 regressor mice against MM2 target cells. For cytotoxicity test of anti-H-2b CTLs, varying numbers of anti-H-2b CTLs were mixed with 51Cr-labeled B6 lymphoblasts (1 × 104 cells) for 6 h at 37°C. B6 lymphoblasts were prepared by culturing B6 splenocytes with RPMI 1640 supplemented with 10% FCS and 5 μg/ml Con A (Pharmacia Biotech) at 37°C for 3 days. In all cytotoxicity tests, the radio activities of medium and cells were counted by gamma counter, and the percent specific lysis was calculated as follows: % specific lysis = (cpm experimental release − cpm spontaneous release)/(cpm maximum release − cpm spontaneous release) × 100.
H-2b- or MM2-specific CTLs (2 × 105 cells/well) were incubated in triplicate for 24 h in 96-well plates (Corning Glass Works, Corning, NY) in 100 μl of complete medium. Methyl-tritiated thymidine ([3H]TdR; Amersham, Arlington, IL; 1 μCi/well) was added to the culture 8 h before harvest. The cells were harvested on glass-fiber filters using a cell harvester (Cambridge Technologies, Watertown, MA), and their radio uptake was measured in a scintillation counter. Culture supernatants from γδ T cells, rIL-10, and rTGF-β1 were added to the H-2b- or MM2-specific CTL proliferation assay at varying concentrations.
Administration of daunomycin-conjugated Abs in vivo
MM2 or MH134 cells (2 × 105) were inoculated i.p. or s.c. into C3H/He mice, respectively. On either 3 consecutive days, days 4–6 or days 15–17 after tumor inoculation, 100 μg of daunomycin-conjugated Abs containing 5 ng of daunomycin were injected at a tumor site, and subsequent tumor progression was observed. As a negative control, 2 μg of daunomycin alone was injected at a tumor site.
RT-PCR of cytokine mRNAs
Tumor-infiltrating γδ T cells accumulating on day 7 after MM2 i.p. inoculation and splenocytes from MM2-bearing C3H/He were prepared as described above. Cells containing PBMC were prepared from blood of MM2-bearing and regressor mice by treatment with 0.17 M ammonium chloride. Total RNAs of these cells were extracted with an RNA extraction kit (RNeasy, Qiagen, Hilden, Germany). First-strand cDNA was reverse transcribed using each RNA sample and was amplified by PCR with a RNA PCR kit (GeneAmp RNA PCR Kit, Takara Biomedicals, Osaka, Japan) according to the manufacturer’s directions. All pairs of primers for β-actin, IL-4, IFN-γ (20), IL-10 (21), and TGF-β (22) were used in this PCR study. PCR was run for 35 cycles with a thermal cycler (DNA amplifier, Sanyo Co., Osaka, Japan) as follows: 1 min at 94°C, 1 min at 55°C, and 15 s at 72°C. The PCR products and DNA m.w. marker VI (Boehringer Mannheim, Mannheim, Germany) were loaded in 2% agarose gels and visualized with UV exposure of 1 μg/ml ethidium bromide-staining agarose gel.
Depletion of γδ T cells and abrogation of their functions by Dau-UC7
To prepare the daunomycin-conjugated Abs, the sugar moiety of daunomycin was cleaved with periodate and then coupled with an anti-γδTCR mAb (UC7-13D5), anti-αβTCR mAb (H57-597), or hamster IgG. Following reduction with sodium borohydride, the conjugates were separated from the unbound daunomycin by Bio-Gel P-100 column chromatography. Fractionated eluates were collected, and their absorbances at 280 nm (Abs) and 495 nm (daunomycin) were monitored. Free daunomycin and free hamster IgG were eluted over 20 and in 8–15 fractions, respectively (Fig. 1,D). The daunomycin conjugates had a dual absorbance derived from daunomycin and Ab and were also eluted in nearly the same fractions as free Abs (Fig. 1, A–C), indicating successful conjugation. The conjugates of UC7-13D5, H57-597, and hamster IgG contained daunomycin at a ratio of approximately six daunomycin moieties to one Ab.
To investigate whether these daunomycin-conjugated Abs can specifically destroy corresponding cells, splenic αβTCR+ and γδTCR+ cell populations were separated from normal C3H/He mice and were cultured over a short term under CD3 stimulation. The expanded cells were 51Cr labeled and incubated with daunomycin-conjugated UC7-13D5 (Dau-UC7), H57-597 (Dau-H57) or hamster IgG (Dau-IgG), unconjugated UC7-13D5 (UC7), H57-597 (H57) or hamster IgG (IgG), or daunomycin alone. Following 12-h incubation, αβTCR+ and γδTCR+ cells were killed by Dau-H57 and Dau-UC7, respectively, while neither conjugate exhibited substantial cytotoxicity against the irrelevant T cell population (Fig. 2 A). Dau-IgG, any of the unconjugated Abs (data not shown), or daunomycin did not lyse αβTCR+ or γδTCR+ cell target.
In tumor-bearing mice, γδ T cells suppress NK lineage cells and/or CTLs at least partly by releasing soluble factor(s) (18, 19). To confirm the functional abolishment of γδ T cells by Dau-UC7, we used culture supernatants from γδ T cells by testing their immunosuppressive activity against CTLs. γδ T cells purified from MM2-bearing mice were treated with Dau-UC7, and their culture supernatant was added in assays of cytotoxicity and proliferation of anti-H-2b CTL. γδ T cells separated from splenocytes or tumor-infiltrating lymphocytes of MM2-bearing mice produced suppressive factors against anti-H-2b CTL activity (Fig. 2, B and C). The treatment of these γδ T cells with Dau-UC7 totally abrogated their ability to secrete suppressive factors. In contrast, the γδ T cells exposed to Dau-H57, Dau-IgG, any of the unconjugated Abs, or daunomycin produced these factors unchangedly. Interestingly, the γδ T cells treated with UC7 produced these suppressants more vigorously than untreated cells, suggesting that intact UC7 mAb transmits activation signals into the γδ T cells. These results further demonstrated that daunomycin-conjugated mAbs were able to attenuate the development of specific lymphocytes.
Tumor regression by intralesional injections of Dau-UC7 at an early stage of tumor development
The above finding raised the possibility that intralesional injections of Dau-UC7 restore the function of tumor-specific CTLs and NK cells by depressing γδ T cells and result in subsequent inhibition of tumor development. When Dau-UC7 was intralesionally given on days 4, 5, and 6 after i.p. inoculation of MM2, the MM2 tumor regressed completely (Figs. 3,A and 4), and the MM2 regressor mice subsequently survived for over 10 mo. On the other hand, the growth of MH134 tumor was also suppressed by the administration of Dau-UC7 on days 4, 5, and 6 after s.c. inoculation of tumor cells, although the MH134 tumor gradually developed thereafter, and all of the mice died within 2 mo after the tumor inoculation (Fig. 3,B). In contrast, there were no therapeutic effects when Dau-UC7 was administered intralesionally on days 15, 16, and 17 after MM2 or MH134 inoculation (data not shown). This is consistent with our previous finding that γδ T cells function as suppressor cells in an early tumor formation (19). No attenuation of tumors was found when Dau-IgG or daunomycin alone was used at the same concentration as Dau-UC7 (Figs. 3 and 4). A weak inhibition of MM2 tumor progression was found in mice treated with Dau-H57 on days 4, 5, and 6 after i.p. MM2 inoculation, while such an inhibition of tumor growth was not observed in MH134-bearing mice (Fig. 3,B). Subsequently, the mice died within 25 days following vigorous progression of MM2 (Fig. 3 A). These results raise the possibility that a portion of αβ T cell populations modestly inhibit antitumor cytotoxic cells in early formation of MM2 lesions. This may be in accordance with our previous finding that early-appearing Th2-αβ T cells exhibit suppressive features against NK activities (19).
Results from the in vivo studies indicated that attenuated local cytotoxic cell activities were restored by removing γδ T cells from tumor lesions. Therefore, it was suggested that depletion of such γδ T cells from early tumor lesions results in the vigorous expansion of lesional CTLs in vitro. Certainly, anti-MM2 and -MH134 cytotoxic cells were much more rapidly induced in Dau-UC7-treated TILs than in Dau-hamster IgG-treated or untreated cells (Fig. 5, A and B). Furthermore, similar results were obtained from the study using B16 TILs (Fig. 5,C), suggesting that the attenuated cytotoxic cell activities by γδ T cells are generally seen in early tumor lesions. We further confirmed activation of tumoricidal cytotoxic cells in mice treated with Dau-UC7 by an in vitro study. Splenocytes separated from Dau-UC7-treated MM2 regressor mice lysed MM2, but not MH134 tumor cells, vigorously (Fig. 6,A) in an αβTCR- and CD8-dependent manner (Fig. 6,B), whereas no cytotoxicity was exhibited by splenocytes isolated from untreated MM2-bearing mice and control Dau-IgG-treated MM2-unregressing mice (Fig. 6,A). This indicates that the activity of tumor-specific CTLs was powerfully augmented by removing γδ T cells from tumor-bearing mice. It is notable that splenocytes from MM2 regressor mice were also modestly cytotoxic against NK-sensitive YAC-1 tumor cells (Fig. 6 A). This also suggests that NK cell activity was elevated by depleting γδ T cells. Taken together, these findings indicate that Dau-UC7 exerts its antitumor action by damaging γδ T cells that function as down-regulators against CTLs and NK cells.
Suppressive cytokines produced by MM2-infiltrating γδ T cells
To investigate the cytokine expression pattern of γδ T cells accumulating in tumor lesions, γδ T cells were freshly isolated from 7-day ascites fluid of i.p. inoculated MM2 using γδTCR-specific mAb (UC7-13D5)-conjugated magnetic beads. Their cytokine profile was examined by PCR of cDNA with primers specific for IL-4, IL-10, IFN-γ, and TGF-β. Fig. 7 shows that freshly isolated MM2-infiltrating γδ T cells transcribed IL-10 and TGF-β mRNAs, whereas neither amplified product of IL-4 nor IFN-γ was detected. In addition, when cultured over a short term in the presence of IL-2, these γδ T cells secreted IFN-γ but not IL-4 (data not shown), suggesting the IFN-γ-producing capacity of the γδ T cells. Since it has recently been reported that CD4+ T lymphocytes producing IL-10, TGF-β, and IFN-γ are a novel population, termed Tr1 cells (5), these results suggested that γδ T cells accumulating in MM2 tumor lesions are of the Tr1 type.
To elucidate the participation of these Tr1 cytokines in γδ T cell suppression of CTL and NK activities, culture supernatants of MM2-infiltrating γδ T cells were mixed with IL-10- and TGF-β-specific neutralizing mAbs and added to the culture of the anti-MM2 CTL proliferation assay. The inhibitory effect of the γδ T cell culture supernatant was reduced by the addition of either anti-IL-10 or anti-TGF-β mAb in a dose-dependent manner (Fig. 8,A). CTLs proliferated more vigorously when both IL-10 and TGF-β were neutralized, suggesting an additional or synergistic effect of IL-10 and TGF-β on the suppression of CTLs. To further examine the inhibitory role of IL-10 and TGF-β, splenocytes from MM2-bearing mice were cultured in low dose (10 U/ml)-IL-2-containing medium in the presence of anti-IL-10 and anti-TGF-β mAbs. This neutralization led to augmentation of the lytic activity of cultured cells against MM2 as well as YAC-1 (Fig. 8,B), consistent with the augmentation of NK cytotoxicity in mice that regressed MM2 tumors following injection of Dau-UC7 (see Fig. 6,A). An RT-PCR study revealed that a reduction in IL-10 and TGF-β mRNA transcriptions of PBMC was correlated with regression of MM2 tumor cells in Dau-UC7-treated MM2-bearing mice (Fig. 8,C). The inhibitory effect of TGF-β on the growth of CTLs was further confirmed by the finding that rTGF-β1 suppressed the proliferation of CTLs in response to MM2 in a dose-dependent manner (Fig. 8,D). However, only weak inhibition was obtained with rIL-10, and the effect of the mixture of rIL-10 and rTGF-β was similar to that of TGF-β alone. Since IL-10 neutralization of culture supernatant from γδ T cells abolished its inhibitory activity (see Fig. 8 A), it is possible that another suppressive factor(s) that synergizes with IL-10 is present in γδ T cell culture supernatants. These results demonstrated that Tr1 cytokines produced by γδ T cells participate in the attenuation of CTL and NK activities.
This study demonstrated that TGF-β and IL-10 produced by Tr1-type γδ T cells inhibit CTL generation and NK activity, and that elimination of this type of γδ T cells from tumor-bearing mice with a daunomycin-conjugated specific mAb augments the activity of CTLs and NK cells and subsequent tumor regression. The in vitro study further confirmed the accumulation of γδ T cells functioning as suppressors against tumoricidal lymphocytes in early tumor lesions, as cytotoxic cells were rapidly induced in γδ T cell-depleted TILs of melanoma as well as hepatoma and mammary tumors. Our previous study has shown that Th2-type γδ T cells present at an early stage of B16 melanoma development also exert an inhibitory action on CTLs and NK cells. Those γδ T cells dominantly infiltrate in early tumor lesions, and their number subsequently decreases thereafter (19). Thus, γδ T cells in these two systems function as immunosuppressors against effector T cells by producing Tr1 or Th2 cytokines in innate immune responses (23). It is well known that Th2 cytokines down-regulate Th1 cell functions, resulting in abolishment of effective induction of CTL and NK cells (24). Likewise, Tr1 cells directly attenuate Ag-specific immune responses mediated by Th1 cells (5, 25). In our preliminary study, Th2-type γδ T cells in the B16 system also secrete a great amount of TGF-β (unpublished data), suggesting a similarity between these two types of T cells. Since TGF-β is an autocrine and paracrine inhibitor of CTLs (26, 27), this cytokine seems to be a key factor in the γδ T cell-mediated suppression of antitumor activities. In fact, blocking of TGF-β with mAb abrogated the activity of γδ T cell supernatants, and the function of γδ T cells was replaced by exogenously added TGF-β.
On the other hand, the effect of IL-10 on suppression of CTL generation and NK activities remains controversial. IL-10 suppresses the cytotoxicity of and IFN-γ production by NK cells, and the induction of tumor-specific CTLs (28, 29), whereas IL-10 augments CTL and NK activities synergistically with IL-2 (30, 31, 32). In our study neutralization of IL-10 in γδ T cell culture supernatants resulted in elevation of CTL and NK activities, indicating the down-regulatory role of IL-10. However, the exogenous addition of IL-10 did not suppress the proliferative response of CTLs as did TGF-β. This suggests that as yet unelucidated factors that synergize with IL-10 are required for the suppression. Our previous study has demonstrated that γδ T cells release soluble suppressant(s) acting on the cytolytic effector phase of CTLs, and this soluble factor seems to be different from TGF-β and IL-10 (18). This unidentified factor may participate in the synergistic inhibition with IL-10. Taken together, Tr1 cytokines IL-10 and TGF-β are involved in the mechanism underlying the γδ T cell-mediated inhibition of CTL and NK activities.
Many reports, however, have provided in vitro evidence for a cytotoxic effector role for γδ T cells against tumor cells (33, 34, 35). The vast majority of these observations were obtained from experiments using culture conditions with high doses of IL-2. Since the amounts of Th2 cytokines, including IL-4 and IL-10, are frequently elevated in tumor-bearing mice (1, 36), γδ T cells may overt their suppressive immunoregulatory capacity under such Th2-predominant conditions. In this situation, cultivation of γδ T cells under artificial IL-2-rich conditions possibly converts their function to cytotoxic cells. Our observation that fresh MM2-infiltrating γδ T cells acquire the ability to secrete IFN-γ after short term culture with IL-2 suggests the conversion of immunosuppressive γδ T cells to cytotoxic cells by Th1 cytokines. Alternatively, it is possible that some γδ T cells originally distributed in certain organs, such as skin, liver, intestine, periphery, and reproductive tracts (37) acquire immunosuppressive activity when they accumulate in tumor lesions. Interestingly, a study from another group has shown that extrathymically differentiated γδ T cells may negatively regulate immune reactions, as administration of hepatic γδ T cells leads to unresponsiveness to skin allograft (38). Our preliminary study showed that γδ T cells accumulate markedly in early tumor lesions of athymic nude mice (N. Seo et al., unpublished observation), further providing evidence for the participation of extrathymic γδ T cells in the suppression of tumor immunity. In addition, studies from experimental pregnancy have revealed that the appearance of extrathymic γδ T cells in early decidua of pregnant mice is a crucial event for the maintenance of pregnancy (39, 40). These different lines of studies suggest the role for extrathymic γδ T cells in suppression of cytotoxic cell-mediated immune reactions.
It should be noted that the intralesional administration of Dau-UC7 at an early stage of tumor development also augmented NK activity. The NK-suppressive role of IL-10 and TGF-β has been demonstrated by several groups (28, 29, 41). However, it has also been reported that IL-10 and TGF-β down-modulate MHC class I expression on target cells and render cells NK sensitive (42, 43). We have previously demonstrated the modulation of NK activity by class I molecules on bystander cells in tumor lesions (19). Thus, the activity of NK cells seems to be regulated bivalently by Tr1 cytokines. In another line of studies, it has been reported that NK cells are important in the generation of tumor-specific CTLs, because NK cell-depleted B16 tumor-bearing mice fail to induce CTLs at their tumor sites (44). Thus, γδ T cells in tumor-bearing mice are suggested to suppress both NK cell activity and subsequent generation of anti-tumor CTLs.
Daunomycin-conjugated TCR-specific mAbs were very useful for the elimination of γδ T cells. In contrast, unconjugated mAbs specific for TCR did not inhibit but, rather, enhanced T cell function. The conjugates were prepared by the classical method in this study. It has been documented that the daunomycin conjugates prepared by the dextran bridge method lyse target cells more effectively than those by direct binding method because of the high efficacies of internalization and the release of daunomycin in the cytoplasm of target cells (45, 46). Furthermore, daunomycin bound to dextran without any Abs exhibited more vigorous antitumor cell cytotoxicity than the free form (47). Therefore, daunomycin-anti-TCR Ab conjugates synthesized via the dextran bridge may be a more powerful tool than conjugates constructed by the classical method. On the other hand, since diphtheria toxin-conjugated (48, 49) and ricin-conjugated (50, 51) tumor cell-specific Abs have been used for targeting of tumors, we suggest that Ts cell-specific Abs bound with diphtheria toxin or ricin may also be useful as an effective eliminator of T cells compared with daunomycin conjugates.
In conclusion, elimination of T cells that down-regulate CTLs and/or NK lineage cells is one strategy for tumor immunotherapy. Depletion of the down-regulator such as Th2- and Tr1-type γδ T cells, when the treatment is efficacious but not harmful, may be of clinical importance for the development of an alternative way to enhance immunity against tumor cells.
We thank Keiko Sugaya for technical assistance and Fumiyo Ohmori for preparation of the manuscript.
This work was supported by grants from the Lydia O’Leary Memorial Foundation Fund, the Shiseido Fund, and the Ministry of Health of Japan.
Abbreviations used in this paper: Tr, T regulatory; Ts, suppressor T; Dau, daunomycin; TIL, tumor-infiltrating lymphocyte.