Shp1 regulates T cell homeostasis by limiting IL-4 signals

To examine the cell-intrinsic role of Shp1 in T cells, we generated mice homozygous for a floxed allele of Ptpn6 that also expressed CD4-cre. As expected, absence of Shp1 expression was detected in double-positive (DP) thymocytes and their progeny (Fig. 1 A). Shp1^fl/fl CD4-cre mice did not develop overt autoimmunity or inflammation (Fig. 1 B), consistent with prior work showing that the me phenotype is T cell independent (Yu et al., 1996).

Next, we asked if Shp1 plays a role in thymocyte development. Shp1 conditional knockout mice contained normal proportions of double negative (DN1→DN4), DP, and CD4⁺ and CD8⁺ single-positive (SP) thymocyte populations (Fig. 2 A). There also was no change in the absolute number of cells contained within each developmental compartment of the thymus (Fig. 2 B), nor did we detect any differences in the staining of CD4⁺ or CD8⁺ SP thymocytes for the markers CD5, CD69, or CD24 (Fig. 2 C). Together, these data suggest that Shp1 is dispensable for thymocyte development.

Previous studies of mice expressing TCR transgenes identified a role for Shp1 in thymocyte selection (Carter et al., 1999; Johnson et al., 1999; Plas et al., 1999; Zhang et al., 1999). We crossed our T cell–specific Shp1 conditional knockout mice with mice expressing the P14 TCR transgene, an MHC-I restricted TCR composed of Vα2 and Vβ8.1, which recognizes the gp33-41 epitope of lymphocytic choriomeningitis virus in the context of H-2D^b (Pircher et al., 1989). Surprisingly, Shp1-deficient P14 transgenic mice had normal proportions and numbers of thymocyte subsets, including the positively selected CD8⁺ SP population (Fig. 2, D and E). The CD8⁺ SP population also expressed similar levels of CD5, CD69, and Vα2, suggesting that positive selection was unaltered by the absence of Shp1 (Fig. 2 F). Shp1 deficiency also did not influence the selection of the OT-I or OT-II TCR transgenes (unpublished data). To verify that Shp1 does not regulate thymocyte selection, we performed in vitro thymocyte stimulations. Wild-type and Shp1-deficient P14 thymocytes required an equivalent concentration of gp33 to induce proliferation. Furthermore, Shp1 had no impact on the extent of proliferation induced by gp33 (Fig. 2 G). Together, these results suggest that Shp1 is not essential for regulating thymocyte positive selection or TCR signaling threshold.

Next, we examined the phenotype of peripheral T cells from Shp1^fl/fl CD4-Cre mice compared with control Shp1^fl/fl mice. We found that spleens from Shp1^fl/fl CD4-cre mice were enriched for T cells expressing elevated levels of CD44 (Fig. 3, A and B). This increase was observed in the CD4⁺ and the CD8⁺ T cell compartments, although the increase in the CD8⁺ CD44^hi population was more prominent. Shp1-deficient lymph node and blood T cells also displayed a CD44^hi phenotype (unpublished data). In contrast, CD44 levels were normal in mature SP thymocytes, indicating that Shp1-deficient T cells become CD44^hi after thymic egress (Fig. 3 A). A previous study of me mice implicated Shp1 in control of regulatory T cell (T reg cell) development (Carter et al., 2005). However, we found no difference in the frequency of T reg cells in T cell–conditional Shp1 knockout mice (unpublished data).

Elevated CD44 expression is associated with activated and memory T cells. The CD44^hi population in our mice appeared to have a memory phenotype, as they did not express the activation markers CD25 or CD69 (unpublished data). Memory T cells can be divided further into central memory (CD44^hi CD62L^hi CD127^hi CCR7^hi) and effector memory (CD44^hi CD62L^lo CD127^lo CCR7^lo) populations. The CD4⁺ CD44^hi T cell populations from wild-type and T cell conditional Shp1 knockout mice were composed of a mixture of CD62L^hi and CD62L^lo cells (Fig. 3 C). The CD4⁺ CD44^hi populations also displayed heterogeneous expression of CD127 and CCR7. Together, these findings suggest that the CD4⁺ memory phenotype compartment contained a mixture of effector and memory phenotype cells. In contrast, the CD8⁺ CD44^hi populations contained a high proportion of CD62L^hi CD127^hi CCR7^hi cells, suggesting a prominent central memory phenotype.

The expression of these markers was identical in Shp1-deficient memory phenotype cells and the naturally occurring memory phenotype population present in wild-type mice. This finding suggests that the Shp1-deficient CD44^hi population reflects an expansion of normally occurring memory phenotype T cells. Notably, the central memory phenotype of the CD44^hi populations is consistent with the phenotype of naive T cells that have undergone homeostatic expansion (Sprent and Surh, 2011). To test if the memory phenotype population was a consequence of an antigen-specific T cell expansion, we examined the expression of CD44 on P14 TCR transgenic T cells. However, CD44 expression also was elevated in Shp1-deficient P14 T cells, demonstrating that the accumulation of memory phenotype cells is not driven by a response to a specific endogenous antigen (Fig. 3 D).

To test whether Shp1 plays a role in regulating TCR sensitivity, T cells from Shp1fl/fl and Shp1^fl/fl CD4-cre mice were stimulated in vitro with α-CD3 ± α-CD28. Total CD4⁺ T cells from Shp1 conditional knockout animals displayed increased proliferation, compared with wild-type CD4⁺ T cells, when stimulated with α-CD3 or α-CD3 + α-CD28. (Fig. 4 A). Shp1-deficient CD8⁺ T cells stimulated with α-CD3 displayed dramatic hyperproliferation, although this difference was eliminated upon addition of α-CD28. To determine if the enriched memory-like population in Shp1 conditional knockout animals was responsible for the enhanced proliferation, T cells were sorted for CD44, and the CD44^lo population was then stimulated with α-CD3 ± α-CD28 (Fig. 4 B). Notably, the enhanced proliferation of Shp1-deficient (total) T cells was eliminated when the large memory phenotype population was removed. This result indicates that the apparent hyperproliferative response of Shp1-deficient T cells reflects the intrinsically more robust response of memory phenotype cells, rather than the effects of Shp1 on TCR responsiveness per se.

We also examined the capacity of Shp1-deficient T cells to produce cytokines in response to TCR stimulation. When cells were gated based on their CD44 expression, there was no difference in the ability of wild-type or Shp1-knockout T cells to produce IL-2 (Fig. 4 C) or TNF (Fig. 4 D) in response to stimulation with α-CD3 and α-CD28. In contrast to previous studies (Johnson et al., 1999; Fowler et al., 2010), these data suggest that Shp1 does not have a role in regulating TCR sensitivity in peripheral T cells.

We next investigated if Shp1 has a role in controlling the differentiation of CD4⁺ Th cells. Naive CD4⁺ T cells were stimulated in vitro with α-CD3 and α-CD28, followed by 3 d of culture in the presence of IL-2. Upon restimulation, a significantly lower proportion of Shp1-deficient T cells expressed IFN-γ compared with controls (Fig. 5, A and B). Additionally, there was a significant increase in the number of IL-4–producing cells. In contrast, Shp1-deficient T cells stimulated immediately after their isolation exhibited normal production of IL-4 (unpublished data), suggesting that Shp1-deficient T cells are not preprogrammed for IL-4 production. Together, these findings suggest that Shp1 negatively regulates Th2 differentiation in vitro. To examine whether the in vitro Th2 bias is also seen in vivo, we examined serum antibody levels. Indeed, the serum concentration of IgE, the prototypic Th2 antibody isotype, was ∼50-fold higher in knockout mice compared with controls (Fig. 5 C). In contrast, knockout and wild-type mice had similar levels of serum IgG2a, a Th1-driven isotype. Therefore, Shp1 regulates Th2 skewing in vitro and in vivo.

Th2 differentiation is regulated by a transcriptional network that includes IL-4R–induced, pStat6-directed transactivation of the master Th2 transcription factor GATA-3 (Zhou and Ouyang, 2003). To determine if Shp1 regulates signaling downstream of IL-4R in T cells, we stimulated wild-type and Shp1-deficient T cells with IL-4 and measured Stat 6 tyrosyl phosphorylation. Shp1 deficiency did not affect the dose–response curve for IL4-evoked Stat 6 phosphorylation (Fig. 5 D). To measure the kinetics of pStat6 dephosphorylation, T cells were pulsed with IL-4 for 30 min, followed by three washes to remove residual cytokine. Wild-type T cells showed robust Stat6 tyrosyl phosphorylation upon IL4 stimulation, followed by a loss of the pStat6 signal 2 h later. In contrast, Shp1-deficient T cells maintained high levels of Stat6 phosphorylation at 2 and 3 h after cytokine withdrawal (Fig. 5 E), indicating that Shp1 is required for the efficient dephosphorylation of Stat6 after IL-4 stimulation.

We sought to determine if cell-intrinsic or -extrinsic forces were driving the formation of memory phenotype T cells within Shp1 conditional knockout mice. Toward this aim, we generated mixed bone marrow chimeras. Irradiated, congenically marked (CD45.1) hosts were reconstituted with bone marrow cells from wild-type mice (Thy1.1), conditional knockout mice (Shp1^fl/fl CD4cre), or a 1:1 mixture of the two. In mice reconstituted with conditional knockout bone marrow, there was a significant enrichment for both CD4⁺ and CD8⁺ memory phenotype T cells in comparison with mice reconstituted with wild-type bone marrow (Fig. 6, A and B). Mixed bone marrow chimeras contained wild-type T cells with a predominantly naive phenotype and knockout T cells with an enriched memory phenotype population. This finding indicates that Shp1 knockout T cell phenotype is cell intrinsic and not a response to altered cell-extrinsic factors.

Previous studies have linked IL-4 to the abnormal expansion of memory phenotype CD8⁺ T cells (Lee et al., 2011). To test if the enhanced sensitivity of Shp1 deficient T cells to IL-4 was driving the accumulation of memory phenotype cells, we crossed our Shp1 conditional knockout mice to IL-4^−/− mice. The absence of IL-4 had no effect on the frequency of peripheral blood CD44^hi cells in mice expressing Shp1 (Fig. 6, C and D). However, lowering (IL4^+/−) or eliminating (IL4^−/−) IL-4 expression in Shp1 conditional knockout mice resulted in a reduced percentage of CD4⁺ and CD8⁺ CD44^hi cells in blood (Fig. 6 D). The percentage of CD4⁺ CD44^hi T cells returned to wild-type levels in double knockout mice, suggesting that IL-4 is essential for the development of excess CD4⁺ CD44^hi T cells caused by the absence of Shp1. Decreasing IL-4 levels also lowered the proportion of CD8⁺ CD44^hi T cells, although the percentage of these cells consistently remained above wild-type levels in all organs, suggesting that additional factors contribute to CD8⁺ memory phenotype T cell development in the absence of Shp1. Serum IgE was undetectable in double knockout mice (unpublished data), indicating that IL-4 is critical for the elevated IgE levels detected in Shp1 conditional knockout mice.

To ask whether IL-4 also is required to maintain the increased memory phenotype population, mice were treated with a neutralizing α-IL-4 antibody (Ohara and Paul, 1985). Administration of α-IL-4 to Shp1 conditional knockout mice resulted in a significant reduction in the levels of CD4⁺CD44^hi T cells, which reached wild-type levels, and CD8⁺CD44^hi T cells, although this population remained elevated compared with controls (Fig. 6, E and F). These data further suggest that the enriched CD4⁺ memory phenotype population is completely IL-4 dependent, whereas other factors contribute to the expanded CD8⁺ population. In sum, IL-4 is required for both the development and maintenance of the enriched population of memory phenotype T cells found within mice with Shp1-deficient T cells.

The severe and complex phenotype of Shp1 mutant mice has hindered attempts at determining the cell-autonomous role of Shp1 in various hematopoietic cell types. Floxed Shp1 mice provide the best tool available for analyzing the cell-intrinsic consequences of Shp1 deficiency. Therefore, we generated and analyzed mice with Shp1-deficient T cells. These mice lack the overt autoimmunity of me mice, confirming that the absence of Shp1 in T cells alone is insufficient for the development autoimmunity (Shultz, 1988; Yu et al., 1996).

Shp1 has been identified as a negative regulator of T reg cell development in me mice (Carter et al., 2005). In contrast, our results demonstrate that T reg cell development is normal in T cells specifically lacking Shp1. Furthermore, in contrast to the results of studies of me/me- and me^v/me^v-derived T cells, we found that Shp1 deficiency has no effect on sensitivity to TCR stimulation. Several groups have reported enhanced positive selection of me thymocytes expressing transgenic TCRs (Carter et al., 1999; Johnson et al., 1999; Zhang et al., 1999). In contrast, we find no change in the selection of thymocytes expressing either MHCI (P14, OT-I)– or MHCII (OT-II)–restricted transgenic TCRs. The P14, OTI, and OTII transgenes are all “strongly” selected TCRs; consequently, Shp1 deficiency might affect the selection of TCRs with lower affinity (Azzam et al., 1998; Atherly et al., 2006; Hu et al., 2010). However, previous data implicated Shp1 in regulating the selection of the DO11.10 TCR (Carter et al., 1999), which is believed to receive a stronger selecting signal than OT-II (Hu et al., 2010). Together, our findings and the previous reports suggest that the selection defect found in me mice is not the result of a cell-autonomous effect of Shp1 deficiency but rather is a consequence of Shp1 deficiency in another cell type (e.g., thymic DCs) and/or the systemic inflammatory signals in these mice.

Likewise, the reported TCR hypersensitivity of peripheral T cells from me mice (Johnson et al., 1999) also is likely to be the result of systemic inflammation and the enhanced activation state of the cells. Mice expressing a putative dominant-negative (phosphatase inactive) Shp1 allele in T cells also were reported to show enhanced TCR sensitivity for thymocyte selection (Plas et al., 1999; Zhang et al., 1999) and activation (Plas et al., 1996). These mice lack the severe systemic inflammation of me mice yet still display TCR hypersensitivity. The discrepancy between these results and our study might be explained by the ability of the catalytically impaired Shp1 mutant to interfere with the binding of other SH2 domain–containing negative regulators to proteins with ITIMs and immunoreceptor tyrosine-based switch motifs (ITSMs; Lorenz, 2009). Overexpression of dominant-negative Shp1 not only might outcompete wild-type Shp1 for binding to ITIM- and ITSM-containing proteins but also could block the binding of Ship and/or Shp2 to these motifs, thereby inhibiting their roles in antagonizing TCR signaling as well. In Shp1 conditional knockout T cells, competition for ITIM and ITSM binding is, if anything, decreased, allowing other factors to interact with these proteins. Regardless, our results establish that Shp1 is not essential for the negative regulation of TCR signaling. This study underscores the value in separating cell-extrinsic and -intrinsic effects of Shp1 activity.

Memory phenotype T cells, also known as innate T cells, are characterized by their surface memory phenotype and ability to rapidly produce cytokines upon TCR stimulation. Such cells have been hypothesized to play an important role in the early stages of immune responses. IL-4 has been demonstrated to be essential for the accumulation of memory phenotype T cells in several other knockout mice. For example, mice deficient for inducible T cell kinase (Itk) have an accumulation of memory phenotype T cells in the thymus (Atherly et al., 2006), and the development of these CD44^hi thymocytes subsequently was shown to be dependent on the production of IL-4 by NKT cells (Weinreich et al., 2010). Mice deficient for Krüppel-like factor 2 (KLF2) or inhibitor of DNA binding 3 (Id3) also develop prominent CD44^hi populations in the thymus that are dependent on IL-4 (Weinreich et al., 2009, 2010; Verykokakis et al., 2010). Like these memory-like or innate T cells, the Shp1-deficient memory population expresses high levels of CD44 and has an increased capacity to quickly produce cytokines after TCR stimulation. However, the memory-like populations in Itk-, KLF2-, and Id3-deficient mice all arise during thymocyte development, whereas the CD44^hi population in Shp1 conditional knockouts is restricted to the periphery, suggesting that a distinct developmental pathway is involved. Additionally, mixed bone marrow chimera experiments revealed that loss of KLF2 in the T cell lineage results in the development of memory phenotype T cells of both wild-type and cKO origin. The KLF2 phenotype therefore is a result of elevated extracellular IL-4. In contrast, we found that the development memory phenotype cells in Shp1 cKO mice is cell intrinsic; wild-type bystander cells in mixed bone marrow chimeras maintain normal homeostasis. T cell protein tyrosine phosphatase (TCPTP) T cell–conditional knockout mice also accumulate CD4⁺ CD44^hi and CD8⁺ CD44^hi populations in the periphery (Wiede et al., 2011). Unlike Shp1-deficient T cells, however, the T cells in TCPTP conditional knockout mice predominantly have an activated/effector-memory phenotype, and these mice develop systemic inflammation and autoimmunity, including lymphoid infiltrates in the liver and lungs, elevated antinuclear antibodies, and increased germinal center formation. T cells from Shp1 conditional knockout mice had a mixed effector/central memory phenotype, did not infiltrate tertiary tissues, and displayed normal germinal center formation (unpublished data).

We identified IL-4 as a critical factor driving the development and survival of CD4⁺ CD44^hi cells in Shp1^fl/fl CD4-cre mice. Elimination of IL-4 resulted in wild-type levels of CD4⁺ CD44^hi cells in all lymphoid organs examined. IL-4 also promoted the development and survival of CD8⁺CD44^hi cells, as this population was reduced substantially upon elimination of IL-4. In contrast to the normalization of the number of CD4⁺CD44^hi cells, we observed only a partial reduction of CD8⁺ CD44^hi cells in the blood, spleen, and lymph nodes of Shp1^fl/fl CD4-cre IL4^−/− mice. This finding indicates that there are other factors promoting the accumulation of CD8⁺ CD44^hi cells. The identity of these factors and their relative contributions to memory phenotype T cell development in various lymphoid organs remain to be elucidated.

Previous studies of me-derived cells had identified Shp1 as a negative regulator of Th1 (Park et al., 2005; Yu et al., 2005) and Th2 differentiation (Kamata et al., 2003), but it remained unclear if Shp1 has a cell-intrinsic role in regulating CD4⁺ T cell differentiation. Our results show that Shp1 restricts the development of Th2 (but not Th1) cells in a cell-autonomous manner. This finding is congruent with the autoimmune lung disease to which me mice succumb, which is characterized by excessive type 2 inflammation and can be partially limited by the elimination of IL-4, IL-13, or Stat6 (Oh et al., 2009). However, given that CD4⁺ T cells are nonessential for the me lung phenotype (Shultz, 1988; Yu et al., 1996), additional sources of IL-4 and IL-13 must drive lung inflammation in these mice.

A critical step in the differentiation of Th2 cells is the IL-4–mediated activation of Stat 6 (Zhou and Ouyang, 2003), and Shp1 has been reported to antagonize IL-4/Stat 6 signaling. The IL-4Rα chain contains an ITIM that can interact with Shp1, Shp2, and Ship, and mutation of the ITIM results in a hyperproliferative response to IL-4 in a myeloid cell line (Kashiwada et al., 2001). Additionally, B cells harboring the ITIM mutation have impaired dephosphorylation of Stat 6 (Hanson et al., 2003), and various hematopoietic cells types from me mice show enhanced Stat 6 phosphorylation in response to IL-4 or IL-13 (Haque et al., 1998; Hanson et al., 2003). However, Shp1 deficiency has been reported to have no impact on the IL-4–induced phosphorylation of Stat 6 in CD4⁺ and CD8⁺ T cells (Huang et al., 2005). Our results are in direct contrast, and suggest a cell-intrinsic mechanism for the regulation of IL-4 signaling by Shp1. A likely explanation for this discrepancy is that the previous study only examined the induction of Stat 6 phosphorylation, whereas we found that Shp1 primarily controls Stat 6 dephosphorylation. Regulation of Stat 6 phosphorylation likely explains how Shp1 regulates Th2 differentiation.

IL-4 reportedly has diverse effects on peripheral T cells beyond its role in Th2 polarization. For example, IL-4 promotes the survival of resting T cells (Vella et al., 1997). Consistent with Shp1 deficiency having a more dramatic effect on CD8⁺ homeostasis, IL-4 has been demonstrated to have potent mitogenic effects on CD8⁺ T cells (Morris et al., 2009) and T cells expressing a constitutively active form of Stat 6 primarily adopt an activated phenotype (Bruns et al., 2003). Yet whether IL-4 has a positive or negative effect on CD8⁺ activation, cytotoxicity, and memory formation remains controversial (Terabe et al., 2000; Schüler et al., 2001; Carvalho et al., 2002; Bronte et al., 2003). The IL-4 hypersensitivity of Shp1-deficient T cells likely contributes to the development of memory phenotype T cells. However, enhanced IL-4 production by Shp1-deficient T cells is detected only after prolonged stimulation (Fig. 5, A and B) and not directly ex vivo. This finding demonstrates that in Shp1 conditional knockout animals, IL-4 facilitates the accumulation of memory phenotype cells but does not induce IL-4 production. Additionally, our mixed bone marrow chimera experiments demonstrate that IL-4 from Shp1-deficient T cells alone is insufficient to alter the homeostasis of wild-type T cells. Together these data strongly suggest that Shp1-deficient T cells are reacting to homeostatic levels of IL-4, and that cell-intrinsic hypersensitivity to IL-4 signals is crucial for the increase in memory phenotype T cells and Th2 differentiation in the absence of Shp1. In this context, our findings highlight that IL-4, like other IL2-Rγc cytokines, can have powerful regulatory effects on CD4⁺ and CD8⁺ T cell homeostasis and differentiation.

By analyzing T cell–conditional Shp1 knockout mice, we have delineated the cell-autonomous role of Shp1 in T cell development, homeostasis, and activation. What remains to be determined is the precise role Shp1 plays in regulating T cells during various immune responses. We have demonstrated that, in contrast to reports using me mice, Shp1 deficiency in T cells does not have a major impact on thymocyte development. Rather, we have established that Shp1 is an important regulator of peripheral T cell homeostasis. Through regulation of IL-4 signals in T cells, Shp1 limits Th2 differentiation, IgE production, and critically limits the development of memory phenotype T cells.

Shp1 floxed (Shp1^fl/fl; Pao et al., 2007b), P14 (Pircher et al., 1989), and CD4cre (Lee et al., 2001) mice have been previously described. OT-I, CD45.1, Thy1.1, and IL-4 knockout mice (IL-4^−/−) were obtained from The Jackson Laboratory. All mice were maintained on C57BL/6 background. Mice were housed in the Ontario Cancer Institute animal facility in accordance with institutional regulations. Animal protocols were approved by the Ontario Cancer Institute Animal Care Committee. For IL-4 neutralization experiments, 200 µg α-IL-4 IgG1 (11B11; BioXCell) or an isotype control (HRPN; BioXCell) were administered to mice by i.p. injection. Mice were sacrificed for analysis 5 d after injection. Bone marrow chimeras were generated by i.v. injection of 5 × 10⁶ bone marrow cells into irradiated mice. Recipient mice received 900 cGy of radiation delivered by an X-RAD320 (PXi) 2 h before reconstitution. Chimeras were analyzed 3 mo after their generation.

Sorted thymocyte populations were lysed in NP-40 lysis buffer (Roche). Equal amounts of protein were resolved on NuPAGE 4–12% Bis Tris gels (Invitrogen) and transferred onto PVDF membranes using an iBlot (Invitrogen). Membranes were blocked with 5% milk in TBS containing Tween-20 (Sigma-Aldrich), stained with antibodies against Shp1 (EMD Millipore) and actin (Sigma-Aldrich), and subsequently developed using ECL Plus (GE Healthcare).

Sera were isolated from 12–16-wk-old littermate pairs using microtainer serum separator tubes (BD), and duplicate samples were analyzed by ELISA for IgE (BioLegend) and IgG2a (eBioscience), according to the manufacturer’s instructions.

Splenic T cells were isolated using the magnetic Pan T Isolation kit II (Miltenyi Biotec). Where indicated, cells were stained with α-CD44 (eBioscience) in preparation for cell sorting. Thymocytes were stained with α-CD4 (eBioscience) and α-CD8 (BD). Cell sorting was performed on a MoFlo (Beckman Coulter), and the purities of target populations were routinely >90%. For analytical flow cytometry of T cell subsets, cells were stained with antibodies against CD5, CD8α, CD25, CD45.2, CD69, Thy1.1, Thy1.2, TNF, Vα2, and pStat6 (BD), and CCR7, CD4, CD24, CD44, CD62L, CD127, IFN-γ, IL-2, IL-4, and IL-5 (eBioscience). Data were collected on a FACSCalibur or FACSCanto (BD) and analyzed with FlowJo software (Tree Star).

Cells were cultured in RPMI-1640 supplemented with 10% FBS, l-glutamine, β-mercaptoethanol, penicillin, and streptomycin. For thymocyte assays, P14 CD8⁺ SP thymocytes were co-cultured in round-bottom 96-well plates at a 1:10 ratio with irradiated splenocytes from C57BL/6 mice pulsed with gp33 (KAVYNFATM). T cell stimulation assays were performed in flat-bottom 96-well plates containing α-CD3 and α-CD28 (eBioscience), cross-linked with 10 µg/ml α-hamster IgG (Jackson ImmunoResearch Laboratories). For proliferation assays, cells were incubated with 2.5 µM CFSE (Invitrogen) before culture. Cytokine production assays were performed by adding GolgiPlug (BD) to cultures 1 h after stimulation, and then harvesting cells for staining 5 h later. CD4⁺ T cell stimulation assays were performed by culturing naive CD4⁺ T cells with α-CD3 and α-CD28 for 3 d, as above, with the addition of 50 U/ml IL-2 (eBioscience) for an additional 3 d. Cells were then restimulated with PMA and ionomycin (eBioscience). To assess Stat 6 activation, cells were stimulated with IL-4 (eBioscience) in 96-well round-bottom plates under the indicated conditions, before fixation with Lyse/Fix Buffer (BD), permeabilization with PermBuffer III (BD), and analysis of pStat6 by flow cytometry.

We thank Francis Tong, Catherine Kerr, and Kyle Gill for technical assistance, as well as Sarah Q. Crome for critical reading of the manuscript.

This work was supported by grants National Institutes of Health R37 CA49152 and R01 CA 114945 to B.G. Neel , and the Terry Fox Program Project Grant to P.S. Ohashi. B.G. Neel and P.S. Ohashi are Canada Research Chairs, Tier 1. D.J. Johnson was partially funded by a CIHR graduate award. This work was supported in part by a grant from the Ontario Ministry of Health and Long Term Care and the Princess Margaret Cancer Foundation.

The authors have no conflicting financial interests.