Research ArticleTCR Signaling

Inhibition of T cell activation and function by the adaptor protein CIN85

See allHide authors and affiliations

Science Signaling  05 Feb 2019:
Vol. 12, Issue 567, eaav4373
DOI: 10.1126/scisignal.aav4373

Thou shalt not CIN

The ubiquitously expressed adaptor protein CIN85 has distinct roles in different cell types. This adaptor inhibits the activation of the receptor tyrosine kinases EGFR and PDGFR but promotes B cell receptor signaling. To determine its function in T cells, Kong et al. compared the activation of mouse T cells lacking CIN85 or the highly related adaptor CD2AP. Only loss of CIN85 augmented T cell growth and IL-2 production. CIN85 associated with the inhibitory phosphatase Sts-2 and promoted its recruitment to T cell receptor microclusters after activation. These data define a previously unknown inhibitory interaction, which could be targeted to augment T cell function in cancer and immunity.


T cell activation is initiated by signaling molecules downstream of the T cell receptor (TCR) that are organized by adaptor proteins. CIN85 (Cbl-interacting protein of 85 kDa) is one such adaptor protein. Here, we showed that CIN85 limited T cell responses to TCR stimulation. Compared to activated wild-type (WT) T cells, those that lacked CIN85 produced more IL-2 and exhibited greater proliferation. After stimulation of WT T cells with their cognate antigen, CIN85 was recruited to the TCR signaling complex. Early TCR signaling events, such as phosphorylation of ζ-chain–associated protein kinase 70 (Zap70), Src homology 2 (SH2) domain–containing leukocyte protein of 76 kDa (SLP76), and extracellular signal–regulated kinase (Erk), were enhanced in CIN85-deficient T cells. The inhibitory function of CIN85 required the SH3 and PR regions of the adaptor, which associated with the phosphatase suppressor of TCR signaling–2 (Sts-2) after TCR stimulation. Together, our data suggest that CIN85 is recruited to the TCR signaling complex and mediates inhibition of T cell activation through its association with Sts-2.


T cells play a central role in adaptive immune responses, which can include the priming of naïve T cells, effector functions from activated effector T cells, and the recall responses by memory T cells (1). Antigen (Ag)–induced T cell stimulation initiates not only positive signal cascades but also negative signal pathways that attenuate the strength and duration of positive signals to prevent excessive T cell activation. Fine-tuning/balance between such positive and negative signals determines appropriate T cell responses. Adaptor molecules mediate branches in signal cascades and are essential in propagating both positive and negative signals.

CIN85 (Cbl-interacting protein of 85 kDa) is a ubiquitously expressed adaptor protein that has been identified by four groups from different tissues and named differently: CIN85 from human (2), regulator of ubiquitous kinase (Ruk) from rat (3), Src homology 3 (SH3) domain–containing gene expressed in tumorigenic astrocytes (SETA) from rat (4), and SH3 domain kinase binding protein 1 (SH3KBP1) from mouse (5). CIN85 shares ~54% amino acid sequence similarity with CD2-associated protein (CD2AP) (6). Thus, CIN85 and CD2AP constitute the CD2AP/CIN85 family of adaptor proteins.

The CIN85 protein is composed of three SH3 domains at the N terminus followed by a proline-rich (PR) domain, a serine-rich (SR) domain, and a coiled-coil (CC) domain at the C terminus. The CC domain mediates homotypic oligomerization of CIN85 and heterotypic interactions with CD2AP (4, 7, 8). CIN85 has at least 10 isoforms whose expression is regulated in either a tissue- or developmental stage–specific manner (8, 9). T cells exclusively express one CIN85 isoform lacking SH3A (CIN85ΔA) (10).

Functions of CIN85 are well known downstream of receptor tyrosine kinase (RTK) signaling (2). In fibroblasts, it promotes internalization of epidermal growth factor receptor (EGFR), hepatocyte growth factor receptor (HGFR), and platelet-derived growth factor receptor (PDGFR) through a clathrin-dependent mechanism, which is necessary for RTK degradation and signaling termination (1113). CIN85 as an adaptor protein recruits endocytic regulatory proteins, such as E3 ubiquitin ligase Cbl, endophilin, and dynamin, to regulate receptor trafficking and degradation (11, 12, 1416). Furthermore, the other family member, CD2AP, inhibits T cell receptor (TCR) signaling by promoting the endocytosis and degradation of the TCR (17, 18). These data provide a hint that CIN85 may also negatively regulate TCR signaling and activation.

In contrast, CIN85 promotes B cell receptor (BCR) signaling, particularly for nuclear factor κB (NF-κB) activation (10). Because B and T cells have similar Ag receptor signaling properties and both express the same CIN85 isoform, it seems reasonable to speculate that CIN85 might have a positive regulatory function in T cells. The physiological function of CIN85 in T cells is still unknown, and there are controversial findings for its function depending on the cell types studied; therefore, we investigated the precise function of CIN85 in T cells.

Some studies suggest that CIN85 and CD2AP can compensate for each other in terms of their expression and functions. For example, CIN85 compensates for the loss of CD2AP in CD2AP-deficient [knockout (KO)] podocytes (19, 20), and the function of CIN85 in B cell activation is partially replaced by CD2AP (21). Moreover, CD2AP regulates posttranslational modification of CIN85 (22). Therefore, it was also important to determine the functional relationship between CIN85 and CD2AP and whether these adaptors had distinct or redundant roles in T cells.

In this study, we generated a T cell–specific CIN85-KO mouse to analyze the function of CIN85 in T cells and its relationship with CD2AP. We found that CIN85-KO, but not CD2AP-KO, T cells exhibited hyperresponses to TCR stimulation. Further, we demonstrated that CIN85 was recruited to the TCR signaling clusters after stimulation and inhibited T cell activation by associating with the phosphatase suppressor of TCR signaling–2 (Sts-2). Our results suggest that CIN85 facilitates the recruitment and association of Sts-2 clusters to TCR clusters during initial TCR activation, which dampens T cell activation.


T cell–specific loss of CIN85 enhanced thymocyte maturation

To understand T cell–specific functions of CIN85, we generated T cell–specific CIN85–conditional KO (cKO) mice. CIN85fl/y mice (10) were crossed with CD4-cre transgenic (Tg) mice, and crossbreeding of CIN85 heterozygotes produced offspring according to the expected Mendelian ratio, which indicated that this deletion did not result in obvious developmental defects. Immunoblot analysis using CIN85 antibody (Ab) confirmed the loss of CIN85 protein in both CD4 and CD8 CIN85-KO T cells (Fig. 1A).

Fig. 1 Increase of SP thymocytes and CD44hieffector/memory T cells in CIN85-cKO mice.

(A) Western blot for CIN85 protein in lysates of WT and CIN85-KO T cells. Blots are representative of three independent experiments. (B) Total cellularity of thymus, spleen, and LNs in WT and CIN85-cKO littermates. Data are means ± SD of seven experiments. (C and D) Flow cytometry analysis of thymic T cell subsets in CIN85-cKO mice stained for the indicated markers. Dot plots including the percentage of cells within the given gate (C) are representative of seven experiments. Quantification of % CD4 and CD8 SP thymocytes, the ratio of CD24lo (mature) to CD24hi (immature) SP thymocytes, and % of CD69hi cells (positively selected) among CD3ε+ cells (D) are means ± SD from all experiments. (E and F) Flow cytometry analysis of splenic memory T cell subsets in WT and CIN85-cKO littermates. Dot plots including the percentage of cells within the given gate (E) are representative of five experiments. Quantification of % CD4 and CD8 memory-type T cells (CD62Llo CD44hi) (F) is the mean ± SD from all experiments. *P < 0.05 and **P < 0.01, Student’s t test.

When we examined T cell development in CIN85-cKO mice, we found that there was no alteration in the total cellularity of thymus, spleen, or lymph nodes (LNs) compared to control littermates (Fig. 1B). Examination of thymic development by flow cytometry revealed a moderate increment in the frequency of both CD4 and CD8 single-positive (SP) cells and a slight decrease in CD4 CD8 double-positive (DP) cells in thymocytes from CIN85-cKO when compared to wild-type (WT) controls. (Fig. 1, C and D). Note that CD4-cre–mediated CIN85 depletion took place in the DP stage and did not occur in the double-negative (DN) stage. CD24 expression in SP cells decreases along with cell maturation. Surface staining of CD24 and TCRβ on SP cells of CIN85-cKO thymocytes showed that the maturation of both CD4 and CD8SP cells was notably enhanced. We found that the ratio of mature (CD24lo) to immature (CD24hi) SP cells was substantially increased in CIN85-cKO mice as compared to WT mice (Fig. 1, C and D). These data imply that CIN85 augments with thymocyte maturation. In addition, when we examined thymocyte development using CD3 and CD69 markers, we found that the number of CD3lo CD69lo cells (DN and preselection DP cells) and CD3int CD69hi cells (thymocytes undergoing selection) was unchanged in CIN85-cKO mice, whereas the CD3hi CD69hi (post-positive selection) and CD3hi CD69lo thymocytes (mature SP) were increased compared to WT (Fig. 1, C and D). CD5 expression at the DP stage is proportional to the TCR signal strength during positive selection, and it plays a role in thymocyte selection (23); however, CD5 expression on CIN85-cKO SP and DP cells was normal, which suggests that there was no difference in positive selection of CIN85-cKO and WT mice (fig. S1A). Overall, no obvious effect of CIN85 deficiency in thymic selection was observed. Furthermore, the expression of Bcl-2 and Bcl-xL in CIN85-cKO thymocytes was not altered compared to WT (fig. S1A); therefore, the increase of CIN85-cKO SP cells is not due to the enhanced expression of these prosurvival molecules. Collectively, these observations suggest that CIN85 plays a role in repressing thymocyte maturation but has no substantial role in selection or survival of thymocytes.

Increased effector/memory-type T cells in CIN85-cKO mice

In CIN85-cKO mice, although the frequency and cellularity of T and B cells in peripheral lymphoid organs were normal, we found that the proportion of total CD4+ and CD8+ T cells and effector/memory T cells was different from WT controls. We observed an increased frequency of CD8 T cells and a decreased frequency of CD4 T cells in the spleen of CIN85-cKO mice as compared to WT mice (fig. S1D). In addition, CIN85-cKO mice had an increased frequency of CD62Llo CD44hi effector/memory-type T cells in spleen than WT mice (Fig. 1, E and F). These data suggest that CIN85 plays a role in maintaining the homeostasis of CD4/CD8 T cells and effector/memory T cells.

Studies on DT40 B cells suggest that CIN85 promotes BCR signaling and its loss could be partially replaced by the CD2AP adaptor protein (21). However, this was not the case for T cells. We found that CIN85 expression was not altered in CD2AP-KO T cells (fig. S1B), which indicated that loss of CD2AP did not stimulate compensatory CIN85 expression in T cells. When we compared T cell development in CIN85-cKO, CD2AP-cKO, and CIN85/CD2AP–double cKO (dcKO) mice, we observed distinct patterns of thymocyte development. Although we observed moderate increases in the frequency of SP thymocytes, the proportion of CD8+ T cells, and the proportion of effector/memory-type T cells in the periphery in CIN85-cKO and dcKO mice, we found no effect on thymocyte development or peripheral CD4+ T cell function in CD2AP-cKO mice (fig. S1, C to E). These data suggest that, in T cells, CIN85 and CD2AP are not functionally redundant, because only CIN85 augments T cell development.

Hyperresponsiveness of CIN85-KO T cells to TCR stimulation

CIN85-cKO mice had increased effector/memory T cells, suggesting that CIN85-KO T cells may have enhanced T cell signaling and function. We examined T cell activation after TCR stimulation of WT and CIN85-KO T cells and found that CIN85-KO T cells produced more interleukin-2 (IL-2) and proliferated more than WT T cells (Fig. 2, A and B, and fig. S2, A and B). These data indicate that CIN85-KO T cells are hyperresponsive to TCR stimulation. CD8 T cells were especially affected, and KO CD8+ T cells produced about fourfold more IL-2 than WT CD8+ T cells. We further analyzed Ag-specific responses using AND-Tg CIN85-cKO mice, which express a TCR specific for a moth cytochrome c (MCC) peptide. Similar to Ab cross-linking, peptide + Ag-presenting cell (APC) stimulation stimulated stronger IL-2 production and enhanced cell growth in CIN85-KO T cells compared to WT T cells (Fig. 2, C and D). In contrast to CIN85-KO T cells, CD2AP-KO and WT T cells produced comparable amounts of IL-2 after TCR stimulation by either Ab cross-linking or Ag stimulation (Fig. 2, A to D, and fig. S2, A and B). The enhanced IL-2 production by CIN85-KO T cells was also detected after secondary stimulation (fig. S2C), indicating that CIN85 would also function in the effector phase of immune responses. To confirm that these hyperresponses were not caused by activation-induced cell death, we analyzed cell viability. Naïve T cells were stimulated with CD3 and CD28 monoclonal Abs (mAbs) for 1 day and stained with propidium iodide (PI) and CD25 mAb (fig. S2D). The results showed that most of the cells were activated and PI-positive cells were very rare, although observed at similar frequency in WT, CD2AP-KO, and CIN85-KO T cell cultures.

Fig. 2 CIN85-KO T cells showed hyperresponsiveness to TCR stimulation.

(A and B) Naïve (CD44lo CD62Lhi) CD8 T cells isolated from WT, CD2AP-cKO, and CIN85-cKO mice were stimulated as indicated with Abs against CD3ε and CD28 or with phorbol 12-myristate 13-acetate (PMA) plus ionomycin. Supernatants were analyzed by enzyme-linked immunosorbent assay (ELISA) for IL-2 production after 24 hours (A), and proliferation was assessed at 46 hours (B). Data are means ± SD pooled from four independent experiments performed in triplicate. OD450, optical density at 450 nm. (C and D) Naïve CD4 AND-Tg T cells were stimulated with MCC peptide–pulsed irradiated splenocytes from B10.BR mice, and IL-2 production (C) and proliferation (D) were measured as in (A) and (B). Data are means ± SD pooled from three independent experiments performed in triplicate. (E) Flow cytometry analysis of CD4 WT/CIN85-KO T cells cultured for 6 days under skewing conditions for TH0 (neutral), TH1, and TH2 cells and then restimulated for 6 hours with Ab against CD3ε. Dot plots including the % of cytokine-positive cells are representative of four independent experiments. *P < 0.05 and **P < 0.01, Student’s t test.

The hyperresponsiveness of CIN85-KO T cells could affect the differentiation of effector T cells. To examine this possibility, CD4 naïve T cells were stimulated and differentiated into T helper 1 (TH1), TH2, or TH17 cells under appropriate skewing conditions and then restimulated and analyzed for cytokine production. As expected, few T cells produced cytokines in the control TH0 cells. In T cells under TH1-skewed conditions, we found that a higher proportion of CIN85-KO T cells produced interferon-γ (IFN-γ) than WT T cells. In contrast, we found that WT and CIN85-KO T cells generated a comparable frequency of IL-4+ and IL-17+ T cells when cultured under TH2- and TH17-skewed conditions, respectively (Fig. 2E and fig. S2E). Thus, in addition to enhanced IL-2 production and cell growth, TH1 differentiation was also enhanced in the absence of CIN85.

CIN85 inhibits TCR signaling

Because CIN85-KO T cells are hyperresponsive to TCR stimulation, we analyzed TCR signaling intermediates. We found that the phosphorylation of ζ-chain–associated protein kinase 70 (Zap70), which is one of the earliest events after TCR stimulation, was more pronounced in CIN85-KO T cells than in WT T cells (Fig. 3, A to D, and fig. S3A). In addition, phosphorylation of other TCR signaling molecules, SLP76 (SH2 domain–containing leukocyte protein of 76 kDa), PLCγ (phospholipase Cγ), and Erk (extracellular signal–regulated kinase), was also significantly enhanced upon TCR stimulation when compared to WT T cells (Fig. 3, A to D, and fig. S3A). The total protein abundance of these molecules was not altered. In contrast, the phosphorylation of CD3ζ and Lck (lymphocyte-specific protein tyrosine kinase), the most proximal phosphorylation events in TCR signaling, was not affected by the loss of CIN85 (Fig. 3, C and D, and fig. S3A). Collectively, these data demonstrated that TCR signaling molecules downstream of Zap70 were enhanced in CIN85-KO T cells after TCR activation, which implies that CIN85 inhibits proximal TCR signaling.

Fig. 3 CIN85 suppressed phosphorylation of early TCR downstream signal molecules.

(A and B) Western blot for phosphorylated TCR signaling molecules in lysates of WT and CIN85 KO T cells after stimulation with Abs against CD3ε and CD28 for the indicated times. Blots (A) are representative of three to seven independent experiments. Quantified data (B) are means ± SEM of all experiments. a.u., arbitrary units. (C and D) Flow cytometry analysis of phosphorylated TCR signaling molecules in WT and CIN85-KO CD8 T cells after stimulation with Abs against CD3ε and CD28 for the indicated times. Histograms of staining in WT (black) and CIN85-KO (red) T cells at each time point (C) are representative of four to eight independent experiments. Quantified values are means ± SEM from all experiments. *P < 0.05 and **P < 0.01, Student’s t test.

No detectable contribution of CIN85 in TCR internalization

CIN85 participates in Cbl-mediated internalization/degradation of RTKs (1116). CIN85 forms a ternary complex with Cbl and endophilin, which mediates ligand-dependent internalization of RTKs (11, 12). Because our data showed that CIN85 inhibits T cell signaling and function, it was possible that CIN85 promoted TCR internalization/degradation, similar to its function in receptor internalization in nonimmune cells. Thus, we investigated whether TCR internalization after TCR stimulation was affected by CIN85 deficiency. We confirmed that the surface TCR abundance before stimulation was equivalent in WT and CIN85-KO T cells and found that about 10 to 50% of the surface TCR in both WT and CIN85-KO CD8 T cells was internalized during 5 to 30 min after stimulation by anti-TCRβ cross-linking (fig. S3B). When AND-Tg T cells were stimulated with Ag peptide and APC, the surface TCR abundance was similarly decreased in WT and CIN85-KO T cells. We observed maximum internalization of 25% of surface TCRs at 90 min (fig. S3C). When we analyzed the efficiency of receptor degradation in cells treated with cycloheximide for 1 hour and then stimulated with Abs against CD3ε/CD28, we found that the amount of CD3ζ detected by Western blot was similar in WT and CIN85-KO T cells (fig. S3D). These results suggest that CIN85 is dispensable for TCR internalization and degradation and that the inhibitory function of CIN85 in T cell activation is not due to its effect on TCR internalization and degradation.

SH3 and/or PR regions are responsible for the inhibitory function of CIN85

To identify the molecular basis of CIN85-mediated inhibition of T cell signaling, we assessed CIN85 structure-function relationships using truncation mutants lacking different domains: SH3B, SH3C, PR, SR, and CC, respectively (Fig. 4A). These internal ribosomal entry site (IRES)–green fluorescent protein (GFP) fusion constructs, which also had a FLAG-tag at the N terminus, were retrovirally transfected into CIN85-KO T cells, and IL-2 production was then measured after TCR stimulation (Fig. 4B). We confirmed that CIN85-KO T cells (empty vector control) had an enhanced IL-2 response to TCR stimulation, and this enhancement was rescued to the level of WT T cells by expression of either full-length CIN85 or the CIN85ΔSR mutant. However, other truncated mutants lacking either SH3 (ΔSH3B and ΔSH3BC) or PR (ΔPR) regions failed to correct this overproduction of IL-2 by CIN85-KO T cells (Fig. 4B). These results strongly suggested that CIN85 inhibited TCR signaling through both SH3 and PR regions and that these regions are required for its inhibitory function.

Fig. 4 Critical domains of CIN85 for its inhibitory function and its association of Sts-2 upon TCR stimulation.

(A) Schematic of WT CIN85, which contains two SH3 (SH3B and SH3C) domains, a PR domain, a SR domain, and a CC domain, and the indicated deletion mutant constructs, which include an N-terminal FLAG-tag. (B) ELISA analysis of IL-2 production by WT or CIN85-KO T cells retrovirally transduced with CIN85 mutants or GFP vector control and restimulated with Abs against CD3 and CD28, as indicated. Data are means ± SD from four experiments performed in triplicate. (C) Coimmunoprecipitation analysis of proteins associated with CIN85 in lysates of 2D12 T cell hybridoma cells transduced with FLAG-CIN85 and stimulated with Abs against CD3ε and CD28. Blots (left) are representative of three independent experiments. Quantified band intensity values (right) are means ± SD from all experiments. (D and E) Coimmunoprecipitation analysis of CIN85 association with Sts-2 in lysates of 2D12 T cell hybridoma (D) and mouse CD8 T cells (E) transduced with the indicated FLAG-tagged CIN85 mutant and stimulated with Abs against CD3ε and CD28. Blots are representative of at least two independent experiments. *P < 0.05 and **P < 0.01, Student’s t test.

CIN85 inhibits TCR signaling by recruiting the phosphatase Sts-2

We identified CIN85-interacting partners to further clarify the mechanism of CIN85-mediated inhibition of T cell activation. Because of our functional reconstitution results (Fig. 4B), we investigated the proteins that bind to WT and CIN85ΔSR but not to the ΔSH3B, ΔSH3BC, and ΔPR mutants by immunoprecipitation and liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis. CIN85ΔPR was used for comparison because it did not rescue the hyperresponsiveness of CIN85-KO T cells, and nontransfected CIN85-KO T cells were used as the background control. Among the list of associated proteins detected, we selected the CIN85ΔPR- and CIN85-interacting proteins with quantity values at least twofold higher than those of the background control, and each value was normalized to that of CIN85 in each sample. When we compared the ratio of the peptide abundance found in cells that expressed CIN85 to cells that expressed CIN85ΔPR, we could identify proteins with stronger binding to CIN85 than CIN85ΔPR (ratio > 1) and vice versa (ratio < 1). The analysis identified known CIN85-binding proteins, including Cbl, CapZ, and CD2AP in both unstimulated and stimulated T cells (fig. S4, A and B). In addition, we observed CIN85 interaction with a novel potential candidate Sts-2 in stimulated T cells (fig. S4B). Sts-2 is a cytoplasmic phosphatase that binds to Cbl and inhibits Zap70 activation and TCR signaling (24). MS analysis revealed that Sts-2 associates with CIN85 but not with CIN85ΔPR and that this association occurred only after T cell stimulation.

The TCR stimulation–induced association of CIN85 with Sts-2 was further confirmed by biochemical analysis. 2D12 T hybridoma cells were transfected with FLAG-tagged CIN85 and stimulated with Abs against CD3 and CD28, and FLAG or His (control) immunoprecipitates were analyzed by Western blot (Fig. 4C). We found that cCbl and Sts-2 associated with CIN85 in a TCR activation–dependent manner. Because Sts-1 has high amino acid sequence homology (75%) to Sts-2 and much stronger phosphatase activity than Sts-2 (2427), we also tested the association of Sts-1 with CIN85. However, we could not detect any specific association between CIN85 and Sts-1 in the pull-down analysis (Fig. 4C). These results confirmed the results of MS analysis—that CIN85 associates with Sts-2 but not Sts-1.

We determined which domains of CIN85 were necessary for its association with Sts-2 in 2D12 cells transfected with the four CIN85-truncated mutants, i.e., ΔSH3B, ΔSH3BC, ΔPR, and ΔSR. Although Sts-2 coimmunoprecipitated with CIN85 or CIN85ΔSR after TCR stimulation, we found no association with the other CIN85 mutants, ΔSH3B, ΔSH3BC, or ΔPR (Fig. 4D). Control immunoprecipitation confirmed the presence of each truncated or WT CIN85. Furthermore, immunoprecipitation of the lysates from normal T cells transduced with these same constructs confirmed that CIN85, but not CIN85ΔPR, associated with endogenous Sts-2 in T cells after TCR stimulation (Fig. 4E). These results indicated that CIN85 associates with Sts-2 through the SH3 and PR domains.

CIN85 colocalizes in TCR microclusters and enhances recruitment of Sts-2 to dampen TCR activation

After Ag recognition, TCR microclusters (MCs) form at the interface between T cell and APC, which contain TCR, kinases, and adaptors to initiate T cell activation signals (2831). CIN85 may function as an adaptor protein to create nodes for the recruitment of signaling proteins into the TCR-MCs to direct the signals toward inactivation. The spatial and temporal dynamics of CIN85 were analyzed in CIN85-GFP–transduced AND-Tg T cells on a supported planar bilayer containing glycosylphosphatidylinositol (GPI)–anchored I-Ek pulsed with MCC peptide and intercellular adhesion molecule–1 (ICAM-1). The images of CIN85-GFP and TCR stained with a DyLight 549–conjugated Fab fragment specific for TCRβ (H57) were collected by total internal reflection fluorescence (TIRF) microscopy. TCR-MCs were generated at the contact region during the early stage of activation and then translocated to the center to form the cSMAC (28, 32). CIN85-GFP formed clusters colocalizing with TCR-MCs at the early stage of TCR stimulation, but later, they did not accumulate into the cSMAC and most of them disappeared (Fig. 5A), as do other signaling molecules such as Zap70 and SLP76 (28). As early as 2 min after stimulation, almost all of the CIN85 clusters colocalized with TCR-MCs (Fig. 5A, right bars). Similarly, dynamic movement of Sts-2–GFP was analyzed. Sts-2 formed clusters after TCR-MCs were generated, which were partly (about 30% of TCR-MCs and 60% of Sts-2 clusters) colocalized with TCR-MCs, and then translocated into the cSMAC (Fig. 5B). Simultaneous analysis of CIN85-GFP and Sts-2–Halo revealed that around 85% of Sts-2–Halo clusters were colocalized with CIN85-GFP clusters in the early phase of activation (Fig. 5C, left). Meanwhile, CIN85-GFP and Sts-2–Halo were not colocalized in the later phase (Fig. 5C, right). These data suggested that CIN85 transiently recruits and colocalizes with Sts-2 during early activation.

Fig. 5 CIN85 formed clusters at TCR MCs and enhanced recruitment of Sts-2 clusters to dampen TCR activation signals.

(A and B) TIRF microscopy analysis of AND-Tg T cells transduced with CIN85-GFP (A) or Sts-2–GFP (B) and stimulated on a supported lipid bilayer containing I-Ek with MCC peptide. Images (left) of GFP-tagged protein and TCR at early (top, 1 to 2 min) and later (bottom, 8 to 12 min) times after membrane interaction are representative of 24 cells from four independent experiments (A) and 21 cells from three independent experiments (B). Quantification of GFP and TCR colocalization clusters per cell at 2 min (right) are pooled from all experiments. (C) TIRF microscopy analysis of AND-Tg T cells cotransfected with CIN85-GFP and Sts-2–Halo tag and stimulated on an MCC–I-Ek–containing supported lipid bilayer at the indicated times. Images (top) and fluorescent intensity profiles between white arrows in the merge images (bottom) are representative of 17 cells from three independent experiments. Quantification of GFP and Halo colocalized clusters per cell at 2 min (middle) are pooled from all experiments. (D) TIRF microscopy analysis of CIN85-cKO/AND-Tg T cells transduced with Sts-2–GFP stimulated on an MCC–I-Ek–containing supported lipid bilayer early after membrane interaction. Images (top) are representative of three independent experiments. Quantification of Sts-2–GFP and TCR colocalized clusters and the ratio of colocalized clusters among Sts-2 clusters (bottom) are means ± SEM of 18 to 24 cells from all experiments. (E) ELISA analysis of IL-2 production by WT, Sts-2–cKO, CIN85-cKO, and CIN85/Sts-2–dKO T cells at 24 hours after stimulation as indicated with Abs against CD3ε and CD28. Data are means ± SD pooled from four independent experiments. Scale bars, 2.5 μm. *P < 0.05, Student’s t test.

To further understand the dynamic relationship between CIN85 and Sts-2, Sts-2–GFP was expressed in CIN85-KO T cells and the dynamics were similarly analyzed. Sts-2 clusters were generated similarly in the absence of CIN85 (Fig. 5D, top). Statistical analysis revealed that the number of TCR-MCs and Sts-2 clusters was similar in WT and CIN85-KO T cells, whereas the Sts-2 clusters that colocalized with TCR-MCs were significantly reduced in CIN85-KO T cells (Fig. 5D, bottom). These data suggested that CIN85 facilitates the recruitment of Sts-2 clusters to TCR-MCs.

Because Cbl is a known binding partner of CIN85, we also analyzed the dynamics of cCbl. cCbl-GFP generated MCs upon TCR stimulation that colocalized with TCR-MCs and Sts-2–GFP clusters (fig. S5A). We confirmed the molecular assembly of CIN85-cCbl complexes by immunoprecipitation for cCbl-GFP in the lysate of 2D12 cells. Immunoprecipitation also showed that cCbl-GFP associated with endogenous CIN85 and Sts-2 in a TCR stimulation–dependent manner. These data are consistent with the results of the FLAG-CIN85 immunoprecipitation MS analysis (fig. S4B) and suggest that Cbl is included in the transient CIN85–Sts-2 complex in TCR-MCs.

Last, we performed functional analysis of CIN85-cKO, Sts-2–KO, and CIN85/Sts-2–dKO (double knockout) T cells. Cytokine production by naïve CD4 and CD8 T cells from each mouse was measured (Fig. 5E). Especially in naïve CD8 T cells, the amount of IL-2 produced by Sts-2–KO T cells was greater than that produced by CIN85-KO T cells. However, dKO T cells produced the most IL-2. However, during secondary activation of effector T cells, the loss of CIN85 promoted greater IL-2 production than the loss of Sts-2 in CD4 T cells (fig. S5C), whereas dKO effector T cells still produced the most IL-2, similar to naïve T cells (Fig. 5E). Collectively, these data indicated that CIN85 and Sts-2 cooperatively reduced IL-2 production after T cell activation.


The present study aimed to elucidate the role of CIN85 in T cell activation and functions. Despite extensive studies on CIN85 over a decade (25), its function remains nebulous because of its various properties in different cell types. CIN85 hinders receptor kinase endocytosis in nonimmune cells, whereas it also promoted NF-κB activation in B cells (10, 33). Our analysis using T cell–specific CIN85-KO mice unexpectedly revealed that CIN85 restricts TCR signaling and T cell function. CIN85 deficiency induced hyperresponsiveness of peripheral T cells in terms of cell growth and IL-2 production, as well as enhanced TH1 differentiation and increased numbers of effector/memory T cells. CIN85-KO T cells showed hyperresponses in early TCR signaling, i.e., enhanced phosphorylation of Zap70, SLP76, and other downstream molecules such as PLCγ and Erk; however, CD3ζ and Lck were unaffected, which indicates that CIN85 inhibits events downstream of Zap70 in TCR signaling.

CIN85 and CD2AP have strong amino acid sequence homology and similar structural features. However, it was unclear whether CIN85 has any physiological relationship with CD2AP in T cell function. Because both CIN85 and CD2AP bind to Cbl and the CIN85/CD2AP-Cbl complex restricts EGFR signaling through receptor endocytosis, it was proposed that CIN85 and CD2AP might have similar functions in T cells. CD2AP deficiency causes hyperresponsiveness to Ag stimulation and augmented cell growth by preventing cSMAC formation necessary for TCR degradation (18). Accordingly, we postulated that CIN85-KO and CD2AP-KO T cells would have a similar phenotype. In this regard, it was interesting to know whether CIN85 and CD2AP are compensatory in their T cell functions. However, in our analysis, CD2AP-KO T cells were not hyperresponsive to TCR stimulation, in contrast to CIN85-KO T cells. Furthermore, simultaneous deletion of CIN85 and CD2AP in T cells did not cause additional difference in T cell function or development compared to those seen in CIN85-cKO T cells. Most T cell development and function is normal in CD2AP-KO mice, which is consistent with our results, and the development of a small subset of follicular helper T cells after infection is increased in CD2AP-KO mice (34). However, our analysis of TCR signaling clusters after stimulation on a planar bilayer revealed that there was no difference between WT and CD2AP-KO T cells in the generation of TCR-MCs and the cSMAC. Therefore, unlike previous reports (18, 35), CD2AP plays no obvious role in T cell function or activation, whereas CIN85 serves as a negative regulator of T cell function.

We postulated that the augmented response of CIN85-KO T cells might be attributed to defective TCR endocytosis and degradation, a high survival rate, or impaired inhibition of TCR signaling. Although CIN85 limits receptor internalization in nonimmune cells (1113), we found no defect in either TCR internalization or degradation, nor did we find enhanced activation-induced cell death in CIN85-KO T cells after TCR stimulation. The enhanced phosphorylation of TCR-proximal signaling molecules such as Zap70, SLP76, PLCγ, and Erk in CIN85-KO T cells showed that CIN85 has an inhibitory effect on early TCR signaling. Because CIN85 is an adaptor protein, we postulated that the enhanced phosphorylation could be mediated by recruiting negative effector molecule(s).

CIN85 has both SH3 domains and a PR region that can specifically interact with proteins containing PR regions and SH3 domains, respectively. Analysis of a series of CIN85 mutants revealed that the SH3 and PR regions were critical for the inhibitory function of CIN85 in T cells. The PR region contains multiple PxxP motifs that could possibly bind to internal SH3 domains (36). The fact that the CIN85 mutants lacking both SH3 and PR regions could not rescue the hyperresponsiveness of CIN85-KO T cells might be due to the structural conformation of CIN85 with its inter- or intramolecular contacts between SH3 and PR regions (36).

We identified several molecules that bind to CIN85 in T cells by MS analysis. Among these candidates, the phosphatase Sts-2 was particularly interesting because of its specific binding to CIN85 upon TCR stimulation and its inhibitory function. Sts-2 associated more with WT CIN85 than the CIN85 mutant lacking the PR region upon TCR stimulation, and the association was activation dependent. Although many proteins have been found to bind with CIN85 in various cell types, Sts-2 had not been reported. This may be because Sts-2 is predominantly expressed in T cells but not in Hela cells (37) or B cells (38).

Sts-2 limits TCR signaling (24), and Sts-2 contains single-nucleotide polymorphisms that are highly associated with several autoimmune diseases such as celiac disease, rheumatoid arthritis, and diabetes (3945). There are many functional similarities between CIN85-KO and Sts-2–KO T cells, such as hyperphosphorylation of Zap70, increased TH1 differentiation and numbers of effector T cells (24), and elevated cytokine responses. In addition, both CIN85 and Sts-2 can bind Cbl (46). These similarities, together with the specific binding of Sts-2 with CIN85 after TCR stimulation, implied that these two molecules may function cooperatively in downstream TCR signaling. CIN85/Sts-2–dKO T cells produced more IL-2 than CIN85-KO and Sts-2–KO T cells after activation, which indicated that both synergistic cooperation of CIN85 and Sts-2 as well as CIN85-independent Sts-2 functions inhibit T cell activation.

Sts-2 associated with CIN85 through the SH3 and PR regions of CIN85, and its binding was essential for inhibitory functions. CIN85 constitutively binds cCbl through its SH3 domain (11), and our results in T cells showed that the association of cCbl and CIN85 was increased by TCR activation and reduced by loss of the CIN85 PR domain, which similarly decreased association with Sts-2. We think that all these possible binding partners may form a molecular complex; CIN85 may directly bind both cCbl and Sts-2, cCbl may mediate Sts-2 binding to CIN85, or Sts-2 may mediate cCbl binding to CIN85. Imaging analysis showed that CIN85, cCbl, and Sts-2 accumulated together within TCR-MCs during early T cell activation. Therefore, these results suggest that a negative feedback mechanism through the association of CIN85, Sts-2, and Cbl may generate an “inhibitory complex” that regulates the inhibitory function in T cell activation.

The cellular function of CIN85 appears to be cell type–specific, probably depending on the binding partners present in different cells. For example, CIN85 binds to SLP65/BLNK expressed in B cells, but not to SLP76 expressed in T cells (21). This difference may result in the functional differences of CIN85 between B and T cells. Such differences have also been observed in humans. where CIN85-deficient patients have defects in B cell activation and Ab production but not in T cells (47). Cell type–specific function was also revealed in thymocyte development, where CIN85 associates with pre-TCRα to mediate signals (48).

Specific blockade of the inhibitors of TCR signaling could induce enhanced T cell immunity, as in the case of Csk (49). Our finding suggests that CIN85 restricts T cell function only after TCR stimulation and not during the resting stage. Therefore, targeting CIN85 or its associated proteins like Sts-2 may be an untapped therapeutic strategy to promote T cell immunity in diseases like cancer.


Mice and cells

CIN85-floxed and CD2AP-KO mice on the C57BL/6 (B6) background were generated by T.K. (RIKEN, Japan). CIN85-floxed mice have been previously described (10), and CD2AP-floxed mice were designed with the LoxP sites flanking the same sites of cd2ap as shown in (35). To generate T cell–specific CIN85/CD2AP-KO mice, CD4-cre-Tg mice were crossed with the flox-CIN85/CD2AP mice. Sts-2–KO mice were from N.C. and were crossed with CIN85-floxed mice. C57BL/6, B10.A, and B10.BR mice were purchased from CLEA Japan Inc. All mice were maintained under specific pathogen–free conditions, and all the animal experiments were conducted in accordance with RIKEN Animal Research Committee–approved protocols. The 2D12 T cell hybridoma cell line expressing the AND-Tg TCR was established by fusing AND-Tg CD4+ T cells (specific for MCC 88-103 on major histocompatibility complex class II I-Ek) (5052).

Flow cytometry

Cell sorting and analysis were performed using FACSAria and FACSCalibur (BD Biosciences), respectively, and the data were analyzed using FlowJo software (Tree Star). Fluorophore Abs against CD4 (RM4-5), CD8a (53-6.7), CD24 (M1/69), TCRβ (H57-597), CD3ε (145-2C11), CD69 (H1.2F3), CD5 (53-7.3), CD45R (RA3-6B2), CD62L (MEL-14), CD44 (IM7), CD11b (M1/70), CD11c (N418), NK1.1 (PK136), TCRγδ (GL3), CD25 (PC61), Ly-6G (Gr-1), and CD49b (DX5) were purchased from BD Biosciences, BioLegend, or eBioscience. Intracellular staining with Abs against Bcl-2 and Bcl-xL was conducted following the manufacturer’s protocol (BD Pharmingen).

ELISA and proliferation assay

T cells were stimulated with plate-bound Abs against CD3ε (145-2C11) and CD28 (PV-1) for 24 hours; then, the culture supernatant was analyzed for IL-2 production by ELISA (BD Biosciences). Cell proliferation was measured using a Cell Counting Kit-8 (CCK-8) assay kit (Dojindo Laboratories) after 48 hours of T cell stimulation.

Helper T cell differentiation and intracellular staining of cytokines

Naïve (CD4+ CD62Lhi CD44lo CD25 NK1.1) T cells were collected from spleens and LNs and sorted using FACSAria. For both TH1 and TH2 cells, T cells were stimulated with plate-bound anti-CD3ε (10 μg/ml) (145-2C11) and anti-CD28 (10 μg/ml) (PV-1) Abs, whereas TH17 cells were stimulated with plate-coated 2C11 (10 μg/ml) and anti-CD28 (5 μg/ml) (clone 37.51) Abs. The medium was supplemented with IL-12 (10 ng/ml; PeproTech) and anti–IL-4 Ab (10 μg/ml) for TH1 cells; IL-4 (10 ng/ml; PeproTech) for TH2 cells; and transforming growth factor–β (TGFβ) (3 ng/ml; R&D Systems), IL-6 (20 ng/ml; PeproTech), IL-1β (20 ng/ml; PeproTech), anti–IL-4 (10 μg/ml), and anti–IFN-γ Abs (10 μg/ml) for TH17 cells.

On day 7, the cells were restimulated with plate-bound 2C11 and PV-1 Abs in the presence of 2 μM monensin (Wako, Osaka, Japan) for 6 hours. The restimulated cells were fixed with 4% paraformaldehyde (PFA), permeabilized with 0.5% Triton X-100 permeabilizing buffer, and then stained with fluorophore Abs specific to cytokines, i.e., anti-mouse IL-4, IFN-γ (XMG1.2), and IL-17A (TC11-18H10). Abs were purchased from BD Biosciences and eBioscience. Cells were analyzed using FACSCalibur.

Western blot

T cells were left either unstimulated or stimulated by 2C11 and PV-1 Abs followed by cross-linking with goat anti-hamster (GAH) immunoglobulin G (IgG) Ab at 37°C for the indicated times and lysed with 1% NP-40– or Triton X-100–containing lysis buffer. The cell lysates were immunoprecipitated or boiled at 95°C for 5 min in SDS sample buffer. The samples were separated by SDS–polyacrylamide gel electrophoresis, and the transferred membrane was blotted by reacting with the indicated Abs and developed with enhanced chemiluminescence assay according to the manufacturer’s procedure (Pierce). For Western blotting, anti-CIN85 Ab was developed by K.K. (RIKEN, Japan). The anti-CD2AP (Santa Cruz Biotechnology, sc-9137), actin, FLAG (M1) (Sigma-Aldrich, F3040), phospho-Zap70 (Cell Signaling Technology, 2701s), phospho-PLCγ (Cell Signaling Technology, 2821s), phospho-Erk (Cell Signaling Technology, 9101s), SLP76 (Cell Signaling Technology, 4958), phospho-SLP76 (GeneTex, 300076), Zap70 (BD Biosciences, 610239), PLCγ (BD Biosciences, 610027), Erk (Promega, V114A), Sts-2 (Proteintech, 15823-1-AP), Sts-1 (Proteintech, 19563-1-AP), and cCbl (Cell Signaling Technology, 8447) Abs were purchased.

Intracellular staining of phospho-proteins

CD4+ or CD8+ naïve (CD44lo) T cells were stimulated by TCR cross-linking, fixed with 2% PFA, permeabilized with pre-cold methanol, blocked with 1% bovine serum albumin in phosphate-buffered saline (PBS), and stained with fluorophore-conjugated Abs specific for phospho-proteins. The Abs used in this study are as follows: anti–phospho-CD247 (BD Phosflow, 558489), phospho-Lck (BD Phosflow, 558577), phospho-Zap70 (BD Phosflow, 557817), phospho-SLP76 (BD Phosflow, 558438), phospho-Erk (BD Phosflow, 561991), phospho–c-Jun N-terminal kinase (JNK) (BD Phosflow, 562481), and goat anti-rabbit IgG (H+L) secondary Ab (Thermo Fisher Scientific, 65-6120). Cells were analyzed using FACSCalibur.

TCR internalization assay

Whole splenocytes were stained with phycoerythrin (PE)–conjugated anti-TCRβ (PE-H57) on ice and were stimulated with prewarmed GAH Ab (5 μg/ml) at 37°C for the indicated duration. For Ag stimulation, B cells were purified from B10.BR mice by using MicroBeads [Mouse CD45R (B220), Miltenyi Biotec] and activated with lipopolysaccharide (LPS) (1 μg/ml) for 24 hours with MCC peptide on the day before the assay. CD4+ AND-Tg T cells prestained with PE-H57 were stimulated with the LPS-activated B cells at a 1:1 ratio at 37°C and stopped by adding ice-cold PBS on each time point. The sample was divided into two parts; one part was subjected to evaluate total TCR, and the other part was subjected to evaluate the internalized TCR by removing surface Ab with acid treatment. The proportion of internalized TCR to the total TCR was calculated as follows: TCR internalization (%) = [(mean fluorescence intensity (MFI) of acid-treated sample at each time − MFI of acid-treated sample at time 0)/(MFI of the non–acid-treated sample at each time)] × 100.

Plasmid construction and transduction

CIN85 complementary DNA was obtained by polymerase chain reaction (PCR) and subcloned into the pMX-IRES-hCD8 retroviral expression vector. GFP was added to the C terminus of CIN85 and subcloned into the pMX retroviral vector (provided by T. Kitamura, Tokyo University, Japan). The constructs of mutant CIN85 (CIN85ΔSH3B, CIN85ΔSH3BC, CIN85ΔPR, and CIN85ΔSR) were created by PCR. The mutant CIN85 constructs were cloned into pMX-FLAG-IRES-GFP to generate in-frame fusion with FLAG peptide (1 kDa) at the N terminus.

For retrovirus infection, the expression vectors were transfected into Platinum-E packaging cells (provided by T. Kitamura, Tokyo University) with Lipofectamine 2000 (Invitrogen) and VSV-G (vesicular stomatitis virus glycoprotein) retroviral vector. Viral supernatants were collected and concentrated by centrifugation at 8000g (4°C) overnight. One day after T cells or AND-Tg T cells were stimulated with plate-bound anti-CD3ε (10 μg/ml) and anti-CD28 (10 μg/ml) Abs or irradiated B10.BR splenocytes with 3 μM MCC mutant (K99A) peptide, the cells were resuspended into the concentrated retroviral supernatants in the presence of polybrene (10 μg/ml) (Sigma). T cell hybridomas were retrovirally transfected similarly without stimulation. Transfectants were sorted by fluorescence-activated cell sorting on the basis of expression of the selection markers.


The 2D12 T cell hybridomas expressing FLAG-CIN85 or its mutants with and without TCR stimulation (cross-linking with anti-CD3ε, anti-CD28, and GAH Abs) were lysed and treated with anti–FLAG-M2 Ab-conjugated magnetic beads (Sigma-Aldrich, M8823). Anti–His-tag Ab-conjugated magnetic beads (Sigma-Aldrich, H9914) were used as the negative control. For immunoblotting, the precipitated proteins were released from the beads by boiling in SDS sample buffer. For MS analysis, the precipitated proteins were released from the magnetic beads using the elution buffer by rotating for 20 min at room temperature. The Sts-2–GFP– or cCbl-GFP–expressing 2D12 T cell hybridomas were activated and lysed as mentioned above, and the immunoprecipitations were done using anti-GFP Ab-conjugated magnet beads (MBL, D153-11).

MS analysis

Proteins in the eluate were digested by trypsin/Lys-C mix for 18 hours at 37°C. Peptides were injected onto a PicoFrit emitter (New Objective, MA, USA) and then separated with an Eksigent ekspert nanoLC 400 HPLC system. Peptides eluting from the column were analyzed on a TripleTOF 5600+ mass spectrometer for both shotgun-MS and SWATH-MS analyses (53, 54). All shotgun-MS files were searched against mouse UniProt Swiss-Prot database (September 2015 release) using ProteinPilot software v.4.5 with the Paragon algorithm for protein identification (55). The SWATH-MS data files were then annotated using the expanded protein/peptide library generated above using PeakView v.2.2. Proteins in the protein/peptide library were quantified from summed peak areas (56).

Supported planar bilayer system and imaging analysis

GPI-anchored proteins of mouse I-Ek and ICAM-1 were transfected into and purified from Chinese hamster ovary or baby hamster kidney cells and then incorporated into dioleoylphosphatidylcholine liposomes (DOPS; Avanti Polar Lipids) (57). Planar bilayer formation was carried out on a glass slip with 0.16-mm thickness. Planar bilayers were then loaded with MCC peptide in citrate buffer (pH 4.5) overnight at 37°C, followed by blocking with 5% nonfat dried milk in PBS at 37°C. Imaging analysis on planar bilayers was performed on an IX81 microscope (Olympus, Japan) with a TIRF objective lens (UAPON 100×/NA1.49OTIRF, 1.49 numerical aperture) (Olympus, Japan) with dual excitation lasers: Ar 488-nm laser and DPSS 561-nm laser, connected to a charge-coupled device camera (Hamamatsu Photonics), were used. Images were taken every 10 s.


Fig. S1. Subpopulations of thymocytes and splenocytes in CIN85 and/or CD2AP-cKO mice.

Fig. S2. CIN85-KO T cells showed hyperresponsiveness to TCR stimulation.

Fig. S3. CIN85 is not involved in TCR internalization.

Fig. S4. Analysis of CIN85-binding molecules by MS.

Fig. S5. Sts-2 partially colocalized with TCR-MCs.


Acknowledgments: We are grateful to T.K. and N.C. for providing KO mice; T. Kitamura for vectors and cells; T. Imanishi and M. E. S. Badr for technical advice; M. Unno and W. Kobayashi for technical support; and M. Yoshioka and H. Yamaguchi for secretarial assistance. Funding: This work was supported by a Grant-in-Aid for Scientific Research from Japan Society for the Promotion of Science KAKENHI (S 24229004 to T.S.). Author contributions: T.S., M.S.K., and A.H.-T. designed experiments. M.S.K., A.H.-T., M.S., Y.K., and R.O. performed experiments. T.Y., O.O., and K.K.P. provided help/advice. K.K., T.K., and N.C. provided materials. T.S., M.S.K., and A.H.-T. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The raw data files of the LC-MS/MS analyses have been deposited in the ProteomeXchange Consortium through the jPOST partner repository ( with the dataset identifier PXD012369. A material transfer agreement (MTA) was required to obtain the CIN85-KO and CD2AP-KO mice from T.K. and the Sts-2–KO mice from N.C. An MTA was also required to obtain the 2D12 hybridoma T cells from T.S. All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.

Stay Connected to Science Signaling

Navigate This Article