Basophils Can Directly Present or Cross-Present Antigen to CD8 Lymphocytes and Alter CD8 T Cell Differentiation into IL-10-Producing Phenotypes¹

Following Ag stimulation, naive T cells undergo differentiation processes during which they acquire the capacity to produce distinct sets of effector cytokines. Different lineages of CD4 T cell-derived effectors have been extensively studied, including Th1, Th2, Th17, regulatory T, and Tr1 cells (1). Cytokine-derived signals in the course of activation/proliferation are known to be the prime factor that not only induces but also maintains the lineage differentiation (1). On the other hand, our understanding of such a differentiation pathway in CD8 T cell lineage is relatively limited. Earlier studies showed that CD8 T cells, similar to CD4 T cells, can be divided into IFN-γ-producing type 1 (also known as Tc1) and IL-4-producing type 2 (also known as Tc2) effector cells (2, 3). Although factors involved in type 1 differentiation have well been studied, factors associated with type 2 differentiation remain less clear.

We and others have recently reported immunoregulatory roles of basophils in the context of type 2 immunity. Mice deficient in IRF-2 spontaneously develop Th2-type immune responses and increased production of basophils was shown to be correlated with the Th2-prone immunity (4). Basophils were capable of inducing CD4 T cells producing IL-4 both in vitro and in vivo (5). More recently, it was shown that mice immunized with allergens develop allergen-specific Th2-type immune responses and that basophils recruited and activated by the allergens play an important role in directing IL-4 production of the Ag-specific CD4 T cells primarily by providing IL-4 and thymic stromal lymphopoietin (TSLP)⁴ (6). Modulatory functions of basophils were also shown for B cell immunity. In particular, basophils can capture large amounts of Ag through surface FcR-bound Igs, thus enhancing memory B cell responses (7, 8). Again, basophil-derived cytokines, such as IL-4 and IL-6, were shown to mediate the modulatory effects (7). However, whether basophils influence CD8 T cell responses has not previously been tested.

In the current study, we investigated modulatory roles of basophils during CD8 T cell activation. Naive CD8 T cells activated in the presence of basophils efficiently differentiated into IL-10-producing phenotype cells. Basophils were capable of directly presenting Ag to CD8 T cells via the MHC class I (MHC I) molecule. Surprisingly, basophils pulsed with protein Ag efficiently cross-presented Ag to CD8 T cells. IL-4 produced by basophils was necessary but not sufficient for the IL-10 induction of CD8 T cells. Interestingly, IL-6 neutralization significantly reduced the IL-10 production, suggesting a critical role of IL-6. Consistent with this finding, adding both IL-4 and IL-6 into the T cell coculture with APCs in the absence of basophils was sufficient to induce IL-10 production in activated CD8 T cells. However, IL-10-producing CD8 T cells did not display regulatory functions. Collectively, our findings demonstrate that basophils may function as professional APCs that efficiently induce IL-10 production in CD8 T cells and that IL-10 production of CD8 T cells induced by basophils is not necessarily associated with regulatory functions.

C57BL/6, OT-I TCR-transgenic, and IL-4KO B6 mice were purchased from The Jackson Laboratory. All animal experiments were performed according to the Cleveland Clinic Foundation guidelines for laboratory animals and were approved by the Institutional Animal Care and Use Committee.

Basophils were in vivo generated as previously reported (5). In brief, mice were implanted with a miniosmotic pump (Durect) containing 5 μg of IL-3 (PeproTech). Seven days later, the mice were sacrificed and both liver and bone marrow cells were harvested. FcεRI⁺ basophils were positively isolated with FITC-labeled anti-FcεRI Ab (MAR1; eBioscience), followed by anti-FITC microbeads (Miltenyi Biotec), and passed through an LS column (Miltenyi Biotec). Purity after isolation was typically ∼95% FcεRI⁺CD49b⁺ basophils. In some experiments, FcγR^highCD45^int basophils were sorted using a FACSAria cell sorter (9). Sorted basophils were subsequently used in a T cell stimulation experiment.

Lymph node CD8 T cells were isolated as previously reported (5). In brief, lymph node cells were incubated with FITC-labeled anti-CD4 (GK1.5), anti-B220 (RA3-6B2), anti-NK1.1 (PK136), anti-FcγR (93), and anti-MHC class II (MHC II) (M5/114.15.2). FITC-labeled non-CD8 T cells were next incubated with anti-FITC microbeads and passed through an LS column (Miltenyi Biotec). Purity of CD8 T cells after isolation was typically higher than 97% purity. In some experiments, T cells were labeled with CFSE (Molecular Probes). Splenic dendritic cells (DCs) were isolated with anti-CD11c microbeads (Miltenyi Biotec). Unless stated otherwise, 2.5 × 10⁵ T cells were incubated with 0.5 × 10⁵ splenic DCs plus 0.1 μM SIINFEKL peptide in the presence or absence of 1.25 × 10⁵ basophils. In some experiments investigating cross-presentation, basophils were pulsed with 0.5 mg of OVA protein (Sigma-Aldrich) for 5 h in the presence of DMSO or 5 μM MG132 (Calbiochem). Cells were extensively washed with medium and 1.25 × 10⁵ live basophils were recounted using a trypan blue solution before the coculture experiment with 2.5 × 10⁵ OT-I CD8 T cells. To measure cytokine production, cells were stimulated with PMA (10 ng/ml) plus ionomycin (1 μM) for 4 h. Monensin was added into the culture for the last 2 h of stimulation. Cells were then harvested, fixed, permeabilized, and stained for intracellular IFN-γ and IL-10 expression. All of the Abs were purchased from eBioscience: PE-anti-IFN-γ (XMG1.2) and allophycocyanin-anti-IL-10 (JES5-16E3). In some experiments, rIL-4 (a gift from Dr. W. Paul, National Institutes of Health, Bethesda, MD), rIL-6 (PeproTech), neutralizing anti-TSLP Ab (clone 152614; R&D Systems), or anti-IL-6 Ab (clone MP520F3, provided by Dr. S. Stohlman, Cleveland Clinic Foundation, Cleveland, OH) was included in the culture.

A suppression assay (see Fig. 7) was performed as follows. OT-I CD8 T cells were stimulated with splenic DCs in the absence and presence of basophils, which subsequently generated IFN-γ- and IL-10-producing effector cells, respectively. Harvested 5 × 10⁵ effector T cells were then recultured with 1.25 × 10⁵ SIINFEKL peptide-pulsed splenic DCs plus 2.5 × 10⁵ CFSE-labeled naive OT-I CD8 T cells for 3 days. CFSE profile of naive T cells as well as cytokine profiles of effector and naive OT-I T cells were examined as described above. CFSE dilution and cytokine production were determined by FACS analysis using a FACSCalibur cytometer (BD Biosciences) and FlowJo software (Tree Star).

IFN-γ- and IL-10-producing effector OT-I T cells were generated as described above and injected i.v. into B6 mice. Normal splenocytes were labeled with CFSE (low and high). CFSE^high cells were incubated with 1 μM SIINFEKL peptide for 45 min at 37°C. Equal numbers (3 × 10⁶) of peptide-pulsed target (CFSE^high) and control target (CFSE^low) cells were injected i.v. into B6 recipients 2 h after the effector T cell transfer. The recipients were sacrificed 4 h after the transfer. Target cell killing was analyzed from the spleen of the recipients.

A standard two-tailed t test was used for statistical analysis; p values of 0.05 or less were considered significant.

We previously reported that basophils promote differentiation of naive CD4 T cells into IL-4-producing Th2 effector cells in the absence of exogenous IL-4 and that IL-4 secreted from the cocultured basophils is primarily involved in the Th2 differentiation (5). To examine the effects of basophils on CD8 T cell differentiation, OVA-specific OT-I CD8 T cells were stimulated with SIINFEKL peptide and splenic CD11c⁺ DCs in the presence or absence of basophils for 3 days. Basophils used in this study were separated from mice implanted with a miniosmotic pump containing IL-3 using a magnetic column as previously described (5). OT-I CD8 T cells stimulated with peptide and splenic DCs underwent robust proliferation (Fig. 1,A). Activated CD8 T cells mainly expressed intracellular IFN-γ and IL-2 (Fig. 1, B and C, and data not shown). By contrast, the inclusion of basophils into the culture dramatically switched the cytokine profiles of the activated CD8 T cells. As shown in Fig. 1, B and C, intracellular IFN-γ expression was substantially reduced, while they acquired IL-10 expression. Unlike CD4 T cells (5), basophils induced little IL-4 production by the CD8 T cells (data not shown). CD8 T cell proliferation was not affected by the basophils or the IL-10 production (Fig. 1 A). Of note, when the T cells were restimulated with PMA/ionomycin to examine cytokine production, CD8 T cells were not repurified from the culture. Therefore, ionomycin might stimulate “any” remaining cells, in particular basophils, in the culture to degranulate, which then may alter T cell cytokine production profiles. However, when cocultured CD8 T cells were restimulated with plate-bound anti-CD3/28 Abs instead of PMA/ionomycin, comparable IL-10⁺ CD8 T cells were observed (data not shown). Furthermore, it is important to note that at the end of the 3-day culture almost no basophils were found in the culture, probably due to activated CD8 T cell-mediated killing. Therefore, these results suggest that T cell IL-10 production is not caused by some factors degranulated by ionomycin-stimulated basophils but that basophils are likely to exert the modulatory effects on T cells during early T cell activation.

The purity of basophils isolated by typical magnetic separation methods was 93–98%. Therefore, it is possible that the remaining 2–7% nonbasophils may contain APCs and be involved in CD8 T cell activation as well as in IL-10 production. To clarify this possibility, purified basophils were examined for the contamination of DCs or B cells. As shown in Fig. 1 D, contaminating nonbasophils (FcεRI negative) did not contain CD11c- or B220-expressing cells. Moreover, when high-purity basophils obtained by cell sorting (>99%) were used in the coculture, CD8 T cell IL-10 production was not reduced (see below). Therefore, these results strongly suggest that contaminating nonbasophils are not involved in Ag presentation and the resulting CD8 T cell IL-10 production.

It has been known that granulocytes such as mast cells and eosinophils are capable of directly presenting Ag to T cells (10, 11, 12). Therefore, we examined whether basophils function as APCs in the culture and induce CD8 T cell IL-10 production. First, we determined the expression of MHC and costimulatory molecules on the surface of basophils. As shown in Fig. 2,A, FcεRI⁺ basophils used in the experiments expressed high levels of MHC I K^b molecules on the surface, although the level was lower than that of CD11c⁺ splenic DCs. MHC II I-A^b expression was very low on basophils (Fig. 2,A). Basophils also expressed moderate levels of CD86 but not of CD80, while splenic DCs expressed high CD80 and CD86. We also observed similar phenotypes of the primary basophils from naive animals; similar levels of MHC I K^b and CD86 expression were found on FcεRI⁺CD49b⁺ liver basophils (Fig. 2 B).

Whether basophils function as APCs to stimulate and to promote IL-10 production of CD8 T cells was next examined. OT-I CD8 T cells were cultured with splenic DCs or basophils pulsed with SIINFEKL peptide and cytokine production was determined. As shown in Fig. 2,C, OT-I CD8 T cells stimulated by Ag-pulsed DCs mainly produced IFN-γ, whereas those T cells stimulated by Ag-pulsed basophils mainly produced IL-10 and much reduced levels of IFN-γ, which was similar to the coculture experiments described in Fig. 1. Culture of T cells with paraformaldehyde (PFA)-fixed Ag-pulsed basophils (Fig. 2 C) or with basophils and peptide Ag in the presence of anti-IL-4 Ab (data not shown) completely abolished the IL-10 secretion, highlighting the importance of IL-4 secretion from the basophils (see below).

Based on the APC function of basophils shown above, we next examined whether basophils are capable of cross-presenting protein Ag to CD8 T cells. To test this possibility, basophils were pulsed with OVA protein or with medium for 5 h and then cocultured with CFSE-labeled OT-I CD8 T cells for 3 days. As demonstrated in Fig. 3,A, OVA protein-pulsed basophils were fully capable of inducing robust proliferation of OT-I CD8 T cells. Moreover, >10% of CD8 T cells produced IL-10 upon restimulation. However, medium-pulsed control basophils failed to induce any signs of proliferation (Fig. 3 A) and of IL-10 production of cocultured CD8 T cells (data not shown). Interestingly, addition of anti-CD28 Ab to the culture dramatically enhanced IL-10 production (see below for the role of costimulation).

To confirm the cross-presentation of basophils, basophils were pulsed with OVA protein or with medium for 5 h in the presence of the proteasome inhibitor MG132. Proteasome plays critical roles in cross-presentation (13), thus inhibiting proteasome function blocks CD8 T cell activation mediated by Ag cross-presentation (14). As shown in Fig. 3 B, pretreatment of basophils with MG132 blocks CD8 T cell proliferation when basophils were pulsed with OVA protein, while DMSO pretreatment did not affect T cell proliferation. By contrast, basophils pulsed with OVA peptide were fully capable of inducing T cell proliferation even in the presence of MG132. These results strongly support the finding that basophils can cross-present protein Ag to CD8 T cells by a mechanism involving proteasome. Consistent with these results, basophils incubated with fluorescence-labeled OVA protein were highly efficient in capturing the Ag (data not shown). It has previously been reported that human basophils tend to lose the IL-4 release capacity if stored ∼24 h in vitro, implying that the storage conditions may affect basophil survival and functionality (15). Our data demonstrate that a 5 h-incubation at 37°C does not appear to affect Ag presentation and IL-4 release function of mouse basophils. Of note, live basophils were recounted after the 5-h incubation and the same numbers of the basophils were used in the coculture experiments. Taken together, basophils can function as professional APCs that efficiently process protein Ag and cross-present it to CD8 T cells.

IL-4 produced by basophils play a key role in basophil-mediated CD4 T cell Th2 differentiation (4, 5, 6). The contribution of basophil-derived IL-4 during CD8 T cell IL-10 expression was thus examined. First, separating basophils from the cocultured CD8 T cells using a Transwell system induced comparable levels of CD8 T cells producing IL-10, indicating that a soluble factor(s) produced by basophils is involved in the process (Fig. 4,A). Second, CD8 T cell IL-10 expression was absent when PFA-fixed basophils were used in the coculture (Fig. 4,B). Third, adding anti-IL-4 Ab into the coculture completely abolished basophil-mediated CD8 T cell IL-10 production (Fig. 4,C). Consistent with these findings, adding anti-IL-4 Ab during OVA cross-presentation by basophils significantly reduced T cell IL-10 production (Fig. 3 A). Lastly, coculture of IL-4-deficient basophils failed to induce CD8 T cell IL-10 production (data not shown). Therefore, IL-4 produced by basophils seems to be the primary factor that promotes the differentiation of naive CD8 T cells into IL-10-producing effector phenotype cells.

However, the presence of IL-4 was not solely sufficient to mediate IL-10 production in CD8 T cells. As shown in Fig. 4 D, adding rIL-4 into the coculture of CD8 T cells plus APCs without basophils did not induce a significant level of IL-10⁺ CD8 T cells. Therefore, these results suggest that IL-4, although necessary, is not sufficient to induce CD8 T cell IL-10 production and that there appears to be a separate mechanism by which basophils mediate the generation of IL-10-producing CD8 T cells.

As basophils express CD86, we next examined whether CD86-mediated costimulation contributes to the IL-10 secretion of CD8 T cells. Blocking CD86-mediated costimulation by anti-CD86 Ab during coculture did not inhibit IL-10 production (data not shown), suggesting CD86-mediated costimulation is dispensable for the IL-10 production. Interestingly, however, IL-10 production dramatically increased upon adding anti-CD28 Ab in the culture. As shown in Fig. 5,A, intracellular IL-10 expression dramatically increased when OT-I CD8 T cells were stimulated with peptide Ag in combination with basophils in the presence of anti-CD28 Ab. In particular, adding anti-CD28 Ab promoted the generation of CD8 T cells that exclusively produce IL-10 not IFN-γ (Fig. 5 A). Anti-CD28 Ab-mediated IL-10 production of CD8 T cells was completely abrogated by anti-IL-4 Ab, confirming the earlier observation that IL-4 is necessary (data not shown).

In addition to IL-4, basophils are also known to produce other soluble mediators including IL-6 and TSLP (6, 7). Whether they contribute to IL-10 expression was next examined. Basophil-derived TSLP was recently shown to partially induce CD4 T cell Th2 differentiation following papain immunization (6); however, neutralizing TSLP in CD8-basophil coculture had no effect on IL-10 expression (data not shown). Instead, the addition of neutralizing anti-IL-6 Ab significantly reduced IL-10 expression of CD8 T cells (Fig. 5,A), suggesting that IL-6 may be involved in basophil-mediated CD8 T cell IL-10 induction. To confirm this finding, OT-I T cells were stimulated with peptide Ag plus T-depleted splenocytes and different cytokines were added. T cells stimulated by Ag/APCs mainly expressed IFN-γ but not IL-10 (Fig. 5 B). Adding rIL-4 into the culture induced noticeable (<5%) IL-10, while no IL-10 was induced when IL-6 was added in the culture. Strikingly, adding both IL-4 and IL-6 dramatically increased CD8 T cell IL-10 expression even without basophils. These results strongly suggest that both IL-4 and IL-6 play a synergistic role in inducing IL-10 in activated CD8 T cells.

Although ICOS-ICOSL interaction has been suggested to play an important role in CD4 T cell IL-10 production (16), we found that ICOS-deficient CD8 T cells produced high levels of IL-10 when stimulated with basophils with anti-CD3/CD28 Abs, suggesting that it is not involved during basophil-mediated CD8 T cell IL-10 production (data not shown).

Basophils used in this study were isolated from mice implanted with an IL-3-containing miniosmotic pump. Since IL-3 is a well-known factor that not only enhances basophil generation from the precursors (17) but also primes basophils to release more IL-4 in response to FcR cross-linking (18), it is possible that the IL-10-inducing function of basophils may be the result of an IL-3-mediated priming effect. Moreover, isolating basophils using anti-FcεRI might deliver stimulating signals to further enhance cytokine release, which then alter T cell cytokine production. To explore this possibility, we FACS sorted blood basophils from naive mice. The frequency of circulating basophils identified as FcγR^highCD45^int cells in naive mice was ∼0.5% (Fig. 6,A). Importantly, obtained basophils were neither primed by IL-3 nor activated by FcR cross-linking. The cells were then cocultured with OT-I CD8 T cells. Naive basophils were capable of inducing CD8 T cell IL-10 production and, to our surprise, the level of induction was significantly greater than that by basophils from IL-3-treated mice (Fig. 6 B). Basophils isolated from IL-3-treated mice and by CD49b- or FcεRI-mediated magnetic separation did not differ in inducing IL-10 production of cocultured CD8 T cells. It is possible that basophils isolated from IL-3-treated mice might have down-regulated some modulatory functions probably due to “prepriming/activation,” thus less efficient in inducing T cell IL-10 production compared with fresh naive basophils. This finding is consistent with the fact that basophils do not require preactivation to express IL-4 and IL-13 (19).

CD8⁺CD122⁺ regulatory T cells are recently identified naturally occurring CD8 regulatory T cells that exert regulatory roles (20), i.e., suppression of IFN-γ production and proliferation, by mainly producing IL-10 (21, 22). Therefore, it is possible that IL-10-producing CD8 T cells induced by basophils may display regulatory functions. We first examined whether IL-10-producing OT-I T cells display cytotoxic activity. OT-I CD8 T cells were stimulated with peptide Ag plus splenic DCs in the presence or absence of basophils as described above, from which T cells that primarily produce IL-10 or IFN-γ were obtained. Activated T cells were subsequently transferred into mice that had received OVA-pulsed and control target splenocytes that are differentially labeled with CFSE. As shown in Fig. 7,A, we found comparable cytotoxic activity of OT-I effector T cells in vivo regardless of the cytokine production pattern. Therefore, IL-10-producing OT-I T cells seem to be equally cytotoxic as those mainly producing IFN-γ. We next tested whether IL-10- producing CD8 T cells suppress proliferation or IFN-γ production of target CD8 T cells. Effector OT-I cells that produce IL-10 or IFN-γ were cocultured with peptide-pulsed splenic DCs along with CFSE-labeled naive OT-I T cells. CFSE profiles of newly activated OT-I T cells were examined after 3 days of coculture. As seen in Fig. 7 B, the presence or absence of effector CD8 T cells producing different cytokines did not alter proliferation of naive OT-I T cells. Furthermore, IFN-γ production of activated naive OT-I CD8 T cells was not reduced by the presence of IL-10-producing CD8 T cells. IL-10 production of the IL-10-producing OT-I T cells was well preserved upon restimulation without basophils during the secondary coculture condition. Furthermore, IL-10 production of CFSE-labeled OT-I T cells slightly increased when IL-10-producing effector CD8 T cells were cocultured, while such IL-10 induction was not observed when IFN-γ-producing effector cells were included in the culture. Taken together, IL-10-producing CD8 T cells generated by basophils do not appear to be regulatory cells, suggesting that IL-10 production of CD8 T cells is not necessarily associated with regulatory cell phenotypes.

Basophils are increasingly recognized as potent immunomodulatory cells that promote the development of Th2 immunity (9, 23, 24). One key mechanism underlying such modulation involves efficient production of Th2 inducing cytokine IL-4. IL-4 production of basophils was shown to be critical to the induction of CD4 T cell IL-4 production during in vitro stimulation (5, 25) as well as allergen-induced Th2 differentiation in vivo (6). IL-4-independent suppression of CD4 T cell IFN-γ production was also noticed, although the underlying mechanism remains unclear (5). In this report, we examined immunomodulatory roles of basophils in the context of CD8 T cell immunity and found that CD8 T cells stimulated by Ag in the presence of basophils primarily differentiate into IL-10-producing effector T cells. It was also noticed that basophils have the capacity to directly present peptide Ag or to cross-present protein Ag to CD8 T cells. Despite the immunosuppressive roles of IL-10, however, regulatory roles of IL-10-producing CD8 T cells were not observed.

IL-10-producing CD8 T cells have been previously reported. Suzuki and colleagues (20) identified naturally occurring CD8⁺CD122⁺ regulatory T cell subsets that significantly suppress IFN-γ production as well as proliferation of target T cells. These regulatory T cells were shown to primarily secrete IL-10, mediating the suppression (26). Noble et al. (27) also reported IL-10-secreting CD8 T cells that can be induced following stimulation of CD8 T cells in the presence of IL-4 and IL-12. Similar to the CD8⁺CD122⁺ regulatory T cells, IL-4/IL-12-induced IL-10-producing CD8 T cells also displayed significant suppressive activity both in vitro and in vivo. The fact that IL-10-producing CD8 T cells reported in this study failed to show regulatory functions suggests that IL-10 production by CD8 T cells does not always need to be the mediator of the regulation. Consistent with this notion, a considerable regulatory effect of IL-10-deficient CD8⁺CD122⁺ regulatory T cells were noticed (22). Moreover, cytokine-induced IL-10-producing regulatory T cells were capable of suppressing target T cell proliferation following the addition of neutralizing anti-IL-10 Ab (27). It was recently reported that following acute respiratory influenza virus infection, some antiviral effector T cells, mostly CD8 T cells, express IL-10 and that the T cell-derived IL-10 plays a key role in controlling the lung inflammation as well as immune-mediated pathology (28). Blockade of IL-10 led to enhanced pulmonary inflammation and lethal injury (28). Of note, high-dose virus infection significantly decreased IL-10 production and enhanced infiltration of inflammatory cells in the lung (28), suggesting that the acquisition of IL-10-producing capacity might be related to the magnitude of T cell activation. Whether CD8 T cell IL-10 production occurs in vivo during IL-4-abundant environments such as parasite infection and, if so, whether the IL-10 controls the immune responses remain to be examined.

Based on high MHC I expression, it is not surprising to find that basophils can stimulate CD8 T cells when incubated with peptide Ag. Indeed, other granulocytes such as mast cells or eosinophils were also shown to function as APCs and to stimulate naive T cells (11, 29). However, it is unexpected to find that basophils are fully capable of presenting Ag to CD8 T cells via a cross-presentation mechanism. DCs, macrophages, and even B cells were capable of cross-presenting soluble Ag to CD8 T cells in vitro (30, 31, 32), although cross-presentation in vivo seems limited to DCs (32). Our results provide the first evidence showing that basophils are highly efficient in processing soluble protein Ag and cross-presenting it to CD8 T cells in vitro. In addition, basophils were recently demonstrated to present Ag to naive CD4 T cells via surface expression of MHC II molecules (33, 34, 35). Although we observed that basophils express very low levels of MHC II, it is possible that the level of MHC expression may increase under certain circumstances such as parasite infection or protease immunization (33, 34, 35). Examining whether basophils function as “efficient” APCs in vivo will be the matter of great importance. However, the fact that CD8 IL-10 production is achieved at a relatively higher ratio of T cells:basophils (typically 2:1 or 1:1 ratio) compared with that of T cells:DCs (typically 5:1 ratio) implies that the APC function of basophils may be rather inefficient compared with that of professional APCs, such as DCs.

Cellular mechanism(s) leading to IL-10 induction in CD8 T cells by basophils will be an important subject for the future study. IL-4 is an important factor involved in programming activated CD8 T cells for IL-10 production. However, the fact that IL-4 alone induces very low IL-10-producing CD8 T cells when stimulated without basophils implies that basophils may deliver a “non-IL-4” signal that preferentially induces IL-10 production in T cells. Because IL-6 is known to promote CD4 T cell Th2 differentiation and to enhance IL-10 production in CD4 T cells (36) and basophils produce IL-6 (7), it is possible that IL-6 may induce IL-10 expression in CD8 T cells. Indeed, neutralization of IL-6 in the basophil:T cell coculture significantly reduced IL-10 expression. However, it should be noted that rIL-6 alone fails to induce IL-10 by CD8 T cells. Instead, the presence of both IL-4 and IL-6 was sufficient to efficiently induce IL-10 without basophils. Molecular mechanisms involved in IL-4/IL-6-mediated CD8 T cell IL-10 expression will require further investigation. TSLP was necessary for basophil-mediated CD4 Th2 differentiation (6). However, we found no role of TSLP in basophil-mediated IL-10 expression in CD8 T cells. Therefore, basophils may influence CD4/CD8 T cell differentiation via different mechanisms.

Another interesting finding is that CD28-mediated costimulation greatly enhances the generation of IL-10-producing T cells. It is interesting to notice that the enhancing effect of anti-CD28-mediated costimulation is only seen when basophils are used to stimulate CD8 T cells. Indeed, we did not observe significant IL-10 production when T cells were stimulated with DCs, IL-4, and anti-CD28, which is considered “the Th2 (or Tc2-)-inducing culture condition” (37). Of note, basophils do not express CD28. It will be important to examine how CD28-mediated costimulation enhances T cell IL-10 expression.

Finally, IL-10-producing CD8 T cells generated in this study produced very low IL-4. Whether they are “Tc2-like” cells or different phenotype cells (possibly Tc10 cells?) will require closer examination.

In conclusion, our results show that CD8 T cells differentiate into IL-10-producing phenotype cells when stimulated with Ag by basophils or in the presence of basophils. IL-4 produced by basophils is necessary yet not sufficient to induce IL-10 production. This study also provides the first evidence that basophils may function as professional APCs to cross-present Ag to CD8 T cells. Identifying the mechanism of basophil-mediated T cell IL-10 production may unveil novel modulatory functions of basophils during immune responses that enhance basophil generation, such as parasite infection or allergic inflammation.

We thank Dr. Stephen Stohlman (CCF) for providing anti-IL-6 Ab and Jennifer Powers for excellent cell sorting.

The authors have no financial conflict of interest.