T cell receptor–induced phosphoinositide-3-kinase p110δ activity is required for T cell localization to antigenic tissue in mice

The establishment of T cell–mediated inflammation requires the migration of primed T lymphocytes from the blood stream and their retention in antigenic sites. While naive T lymphocyte recirculation in the lymph and blood is constitutively regulated and occurs in the absence of inflammation, the recruitment of primed T cells to nonlymphoid tissue and their retention at the site are enhanced by various inflammatory signals, including TCR engagement by antigen-displaying endothelium and resident antigen-presenting cells. In this study, we investigated whether signals downstream of TCR ligation mediated by the phosphoinositide-3-kinase (PI3K) subunit p110δ contributed to the regulation of these events. T lymphocytes from mice expressing catalytically inactive p110δ displayed normal constitutive trafficking and migratory responses to nonspecific stimuli. However, these cells lost susceptibility to TCR-induced migration and failed to localize efficiently to antigenic tissue. Importantly, we showed that antigen-induced T cell trafficking and subsequent inflammation was abrogated by selective pharmacological inhibition of PI3K p110δ activity. These observations suggest that pharmacological targeting of p110δ activity is a viable strategy for the therapy of T cell–mediated pathology.

Coordinated migratory events guiding immune cell compartmentalization are essential to the development of effective and targeted immunity. Specific adhesive interactions and the responsiveness to chemotactic factors regulate cell-cell interactions leading to T cell activation, the provision of T cell help to B cells, and T cell access to target tissues. Naive T cells are programmed to recirculate predominantly in secondary lymphoid tissue. By contrast, primed T lymphocytes selectively localize in nonlymphoid antigenic sites upon detection of antigen (1). Antigen-presenting DCs help program primed T cells to establish specific molecular interactions that allow their access to various organs, such as the skin and the gut (1). Although these mechanisms provide a level of specificity by restricting the area to which primed T cells can recirculate, the extent of the vascular route that these cells must take to reach their target and the size of the anatomical site that they must patrol suggest that other mechanisms must be in place for T cells to locate and enter target tissues. We and others have shown that antigen recognition on the endothelium and parenchymal antigen-presenting cells contributes to the recruitment and retention of antigen-specific T cells into target tissue where antigen can be found. For example, we have reported that cognate recognition of MHC-peptides expressed by human ECs, under conditions that do not lead to T cell division or anergy (2), can enhance transendothelial migration of human T cells (3). Moreover, using a mouse model, we showed that TCR engagement by the endothelium leads to T cell transendothelial migration and tissue infiltration in vivo (4). Similarly, islet-specific homing by insulin-specific H2-K^d–restricted CD8⁺ T cells was abrogated in mice lacking MHC class I expression or in mice displaying impaired insulin peptide presentation by local endothelium due to deficient insulin secretion, highlighting the ability of ECs to cross-present tissue antigens (5). Interestingly, the induction of TCR-mediated signals leading to either T cell division or anergy abrogates T cell migration, suggesting that qualitatively different signals may determine the outcome of TCR triggering on T cell migration. Although it is well established that TCR-derived signals can regulate T lymphocyte motility (6, 7) and immobilize migrating T cells (8), the molecular basis of these effects is still unclear. Class IA phosphoinositide 3-kinases (PI3Ks) are a family of p85/p110 heterodimeric lipid kinases that generate second messenger signals (e.g., PIP3) downstream of tyrosine kinases, thereby controlling various cell functions, including motility (9). The expression of the PI3K p110δ subunit is mostly restricted to hematopoietic cells (9). The capacity of the TCR to induce PI3K signaling is dependent on the p110δ subunit and has previously been shown to promote proliferation, differentiation, and regulation of T helper cells (10–12). In addition to antigen receptor signal transduction, PI3K activity has been associated to a variety of biological functions in both T cells and other leukocytes (13), including the regulation of their motility (14). In addition to signaling mediated by dedicator of cytokinesis 2 (DOCK2) (15), both class IB and class IA PI3K have been involved in lymphocyte responses to chemokines (14), although their relative contribution appears to be different in distinct lymphocyte subsets. Chemokine-induced migration of B cells in vitro mostly relies upon p110δ activity, and it is not significantly affected by p110γ deficiency (16). In contrast, in vitro migration of p110γ-deficient T cells to CCL19, CXCL12, and CCL21 is mildly reduced, while p110δ-mediated activity appears to be dispensable (16) and CXCL12-induced PI3K activity is entirely dependent on p110γ and not p110δ (12). Accordingly, PI3Kγ-deficient T but not B cells show a mild reduction in their homing efficiency to secondary lymphoid tissue (15) or on migration velocity within lymph nodes (17). T cells lacking both DOCK2 and PI3Kγ show a virtually abolished migration to chemokines in vitro and have a further reduced homing capacity toward secondary lymphoid organs (15). Recent data have indicated a role for the class IA regulatory subunit p85 in sustaining spontaneous motility of both T and B lymphocytes in the lymph node in vivo (18); however, Asperti-Boursin and colleagues reported that wortmannin, a potent PI3K inhibitor, had no effect on constitutive CCL19- and CCL21-dependent migration in lymph nodes (19). Thus, under some circumstances, PI3Ks may affect chemokine-dependent migration of primary mouse T cells; however, DOCK2 appears in most cases to be the main signal transducer employed by chemokine receptors.

The contribution of PI3Ks to pathological effector T cell migration has recently been explored. In a model of lupus-like disease induced in p65PI3K-transgenic mice, in which class IA PI3K activity is elevated, genetic deletion of the class IB isoform p110γ did not prevent tissue invasion by pathogenic CD4⁺ T cells, although it ameliorated disease by reducing their survival (20). However, whether tissue infiltration was a direct result of PI3K-dependent trafficking or a more general loss of self tolerance in this model is not clear.

In this study we have investigated the contribution of TCR- and CD28-induced PI3K p110δ–mediated signals to the regulation of memory T cell trafficking to target tissue. Our observations suggest that PI3K p110δ is an essential mediator of primed T cell migration to antigenic sites and highlight its potential as a pharmacologic target for the control of T cell-mediated inflammation.

PI3K p110δ activity is not required for antigen-independent memory T cell transendothelial migration, chemotaxis, and constitutive trafficking. Before investigating the effect of PI3K-mediated signals on TCR-driven T cell trafficking, the ability p110δ^D910A T cells to undergo antigen-independent migration in steady state and in response to inflammatory stimuli was analyzed. Memory HY-specific H2-A^b–restricted CD4⁺ T cells were generated by immunization of female C57BL/6 WT and p110δ^D910A mice with syngeneic WT male splenocytes followed by in vitro expansion. The 2 T cell populations displayed similar phenotypes, as assessed 7 days after restimulation with male splenocytes and recombinant IL-2 (rIL-2) (Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI33267DS1). Following further resting in medium alone and subsequent antibody stimulation with anti-CD3 and anti-CD28, HY-specific WT and p110δ^D910A CD4⁺ T cells equally upregulated surface expression of CD25, CD69, and the chemokine receptor CXCR4 (Supplemental Figure 2, A–D). However, fewer p110δ^D910A CD4⁺ T cells were induced to produce IFN-γ (Supplemental Figure 2E), consistent with previous results (21). HY-specific WT and p110δ^D910A CD4⁺ T cells migrated equally through resting (Figure 1A) and TNF-α–treated (data not shown) syngeneic, female-derived EC monolayers and displayed similar migration speed over plastic-bound recombinant ICAM-1 (Figure 1B) in vitro. Thus, p110δ does not appear to regulate basal motility in absence of cognate antigen.

Figure 1

PI3K P110δ does not contribute to constitutive T cell migration. (A) HY-specific H2-A^b–restricted WT and P110δ^D910A T cells (5 × 10⁵/well) were seeded onto untreated syngeneic female-derived EC monolayers grown on 3-μm-pore Transwells, and T cell migration was monitored as described in Methods. Results are expressed as a percentage of migrated T cells at the given time points and reported as the average of 3 experiments of identical design. Error bars are shown. (B) HY-specific H2-A^b–restricted WT and P110δ^D910A T cells (5 × 10⁵/well) were seeded onto ICAM-1–coated (2 μg/ml) 35-mm dishes, and their migration was observed for 25–30 minutes by time-lapse microscopy. A representative example of a series of 3 independent experiments with similar design is shown. Mean migration speed ± SEM is shown. (C and D) PKH26-labeled HY-specific H2-A^b–restricted CD4⁺ WT and P110δ^D910A T cells were injected i.v. into syngeneic female mice. T cell localization in the indicated tissues was assessed 24 hours later by wide-field fluorescence microscopy. To minimize the effect of arbitrary choice of field, tissue infiltration was quantified by randomly selecting ten ×10-magnified fields from tissue samples from at least 3 animals and assessing the number of fluorescent cells in each field. Each panel shows a representative tissue image. The mean T cell infiltration ± SD observed in samples from at least 3 animals is shown in D.

To investigate whether constitutive memory T cell trafficking was affected by lack of p110δ activity in physiological settings, WT female mice were injected with either PKH26-labeled WT or p110δ^D910A T cells (10⁷/mouse), and their localization in several tissues was analyzed 16–24 hours later by wide-field microscopy. Both WT and p110δ^D910A T cells recirculated normally and could be detected in the spleen, lung, liver, and kidney of recipient mice in similar numbers (Figure 1, C and D). Notably, WT and p110δ^D910A T cells displayed similar patterns of distribution within the various organs and localized in the liver, kidney, and heart in a scattered pattern, while they clustered in restricted areas in lymph nodes. These observations suggest that p110δ signaling is not involved in constitutive memory T cell trafficking. The inefficient localization of both WT and p110δ^D910A T cells to the lymph nodes as compared to the spleen was likely due to lack of expression of CD62L and CCR7 (Supplemental Figure 1; ref. 22).

As PI3K activity is associated with chemokine-induced T cell migration (14), we also compared the recruitment of WT and p110δ^D910A T cells in response to chemokines. WT and p110δ^D910A T cell migration was equally enhanced by the chemokines CXCL10, CXCL12, and CCL5 in vitro (Figure 2, A–F). As both WT and p110δ^D910A T cells were found to express CXCR4 only following antibody-mediated stimulation with plastic-bound anti-CD3 (1 μg/ml) and anti-CD28 (5 μg/ml) (Supplemental Figure 2), in the experiments assessing migration to CXCL12, T cells were used following exposure to these antibodies for 72 hours in vitro. In these experiments, the increase in migration in response to CXCL12 was quite low, possibly due to the low expression of CXCR4, but p110δ and WT T cells responded similarly.

Figure 2

PI3K p110δ is not required for chemokine-induced T cell migration. (A and B) HY-specific H2-A^b–restricted WT and P110δ^D910A CD4⁺ T cells (5 × 10⁵/well) were exposed to CXCL10 (300 ng/ml; filled symbols, A and D), CXCL12 (50 ng/ml; filled symbols, B and E), or CCL5 (100 ng/ml; filled symbols, C and F) through 5-μm-pore Transwells. In the experiments analyzing migratory responses to CXCL12, T cells were stimulated with plastic-bound anti-CD3 (1 μg/ml) and anti-CD28 (5 μg/ml) for 72 hours to induce CXCR4 expression. Spontaneous migration in chemotaxis medium (RPMI 0.5% FCS) alone was also measured (open symbols). The number of migrated T cells was monitored at the indicated time points. Results are expressed as the percentage of input T lymphocytes that had migrated through the filters at any given time point and represent the mean of at least 3 independent experiments ± SEM. *P < 0.05 at all time points except 2 hours (B, C, and E) and 24 hours. (G–J) C57BL/6 female mice received an i.p. injection of CXCL10 (1,200 ng). One hour later, PKH26-labeled HY-specific H2-A^b–restricted WT and P110δ^D910A T cells (5 × 10⁵/well) T cells were injected i.v. After 6 hours, localization of PKH26-labeled T cells in the peritoneal cavity was assessed by flow cytometry. The panels (representative of 1 experiment) show the number of PKH26-labeled T cells in the CD4-gated T cell population. The mean T cell numbers ± SEM observed in samples from at least 3 animals are summarized in K. *P < 0.02.

As chemokine-induced T cell adhesion and subsequent transendothelial migration is regulated by shear flow (23), the potential contribution of PI3K p110δ activity to T cell migratory response to chemokines in vivo was investigated. HY-specific PKH26-labeled CD4⁺ WT or p110δ^D910A T cells (10⁷/mouse) were injected i.v. into syngeneic female mice that had received an i.p. injection of CXCL10 (1,200 ng/mouse) 1 hour earlier. CXCL10 was used in these experiments because of the high level of expression of its counterreceptor CXCR3 by both WT and p110δ^D910A T cells (Supplemental Figure 1) and because CXCL10 elicited strong chemotactic responses in vitro. Control mice received an equal volume of saline solution i.p. T cell migration to the peritoneal cavity was assessed 6 hours later by flow cytometric analysis of the peritoneal lavage. As shown in Figure 2, G–K, WT and p110δ^D910A CD4⁺ T cells localized to the peritoneal cavity in response to chemokine challenge with similar efficiency. In line with previous observations (15, 16), these data suggest that PI3K p110δ activity is not required for constitutive or chemokine-induced T cell migration and that the machinery required to leave the circulation and cross the endothelial barrier in response to a chemokine gradient is intact in p110δ^D910A T cells.

PI3K p110δ activity is essential for antigen-dependent T cell recruitment. We and others have previously reported that antigen displayed by the endothelium contributes to the recruitment of specific T cells into the antigenic tissue. As PI3K activity is regulated by TCR signaling (9) and has the potential to regulate cytoskeletal rearrangements (24), we sought to investigate whether its activity contributes to antigen-dependent T cell recruitment. To explore this, the effect on migration after cognate recognition of syngeneic WT male- or female-derived endothelial cells by HY-specific WT and p110δ^D910A T cells on their migration was first compared using an in vitro transendothelial migration assay that we have previously described (25). T cells were purified on a Ficoll gradient and seeded onto IFN-γ–treated EC monolayers (derived from female or male WT mice) grown on a Transwell insert, and the number of transmigrated T cells recovered in the bottom well was monitored hourly. As shown in Figure 3, A–C, WT HY-specific T cells displayed a much greater rate of migration through MHC class II H2-A^b–expressing male-derived ECs (Figure 3A). By contrast, migration of p110δ^D910A T cells through male-derived EC monolayers was not enhanced when compared with the migration through female-derived ECs (Figure 3B). The male and female EC monolayers mediated comparable levels of migration of third-party control MHC class I–restricted CD8⁺ T cells (HY-specific K^k-restricted C6 T cells; Figure 3C).

Figure 3

PI3K P110δ is instrumental to antigen-induced T cell recruitment. (A–C) HY-specific H2-A^b–restricted WT and p110δ^D910A CD4⁺ T cells or HY-specific H2-K^k–restricted CD8⁺ C6 T cells were seeded onto IFN-γ–treated female- and male-derived syngeneic EC monolayers grown on 3-μm-pore Transwells. T cell migration was assessed as described in Figure 1. Error bars for 3 averaged experiments are shown. *P < 0.05 versus migration through female ECs at all time points except 2 hours. (D and E) Male and female C57BL/6 mice were injected i.p. with 600 U IFN-γ or PBS. Two days later, mice received an i.v. injection of PKH26-labeled HY-specific CD4⁺ T cells (10⁷/mouse). The presence of labeled cells in the peritoneal membrane (D) and cavity (E) was assessed 24 hours later by wide-field fluorescence microscopy and flow cytometry, respectively. A cluster of 3 PKH26-labeled cells is visible in the membrane of male recipient of WT HY-specific T cells. Due to the presence of an autofluorescent population of non-T cells often detected in FL-2 (also in control mice that received IFN-γ but no T cells; far-left dot plot, “saline”), cells were double stained with an APC-conjugated anti-CD4 antibody following harvesting, and the percentage of PKH26 (FL-2)–labeled T cells gated in the CD4⁺ T cell population is shown in the dot plots and the graph representing cumulative data from at least 3 animals. The mean ± SEM observed in samples from at least 3 animals is shown on the right. *P < 0.04 versus female mice.

To assess the physiological relevance of the dependence of p110δ on antigen-mediated transendothelial migration, we compared the recruitment of HY-specific CD4⁺ WT and p110δ^D910A T cells from the circulation into the peritoneum of male and female WT mice. In this model, HY-specific T cells localize in the peritoneal membrane and cavity of male but not female mice following induction of local MHC molecule upregulation (and subsequently antigen presentation) by i.p. injection of IFN-γ (4). PKH26-labeled WT or p110δ^D910A H2-A^b–restricted CD4⁺ T cells (10⁷/mouse) were injected i.v. into male or female WT C57BL/6 mice that had received an i.p. injection of IFN-γ (600 U) 48 hours earlier. Significant amounts of WT T cells were detected in the peritoneal membrane (Figure 3D) and cavity (Figure 3E) of male mice. By contrast, HY-specific p110δ^D910A T cells failed to localize to the peritoneal membrane and were not found in the peritoneal lavage of IFN-γ–treated male mice. As expected, neither WT nor p110δ^D910A T cells migrated the peritoneal membrane or cavity of IFN-γ–treated female mice. Similar observations were made when TCR-dependent recruitment of HY-specific CD8⁺ T cells was analyzed, suggesting that PI3K p110δ activity is required for antigen-driven localization of both CD4⁺ and CD8⁺ T cells (Supplemental Figure 3).

TCR-induced PI3K recruitment selectively mediates antigen-dependent T cell localization. We have previously shown that the PI3K binding motif in CD28 is necessary for the efficient localization of CD8⁺ memory T cells to nonlymphoid tissue (22). However, this is also required for Grb2 and Gads binding, and hence a requirement of PI3K had been implied but not proven. Constitutive low expression of CD80 molecules by murine endothelium provided an opportunity to discriminate the relative contribution of TCR- and CD28-induced PI3K activity in antigen-dependent T cell localization. To test the role of PI3K specifically in the context of CD28 signaling, we used HY-specific H2-A^b–restricted CD4⁺ CD28^Y170F T cells (which carry a mutation in the cytoplasmic tail of CD28 that abrogates PI3K recruitment without leading to defects in clonal expansion; ref. 26) and compared their responses to those by p110δ^D910A T cells, where PI3K signaling is uncoupled from both TCR and CD28 signaling.

First, localization of HY-specific WT, p110δ^D910A, and CD28^Y170F CD4⁺ T cells to the peritoneal membrane and cavity of male mice was compared by the antigen-dependent peritoneal recruitment model described above. PKH26-labeled HY-specific CD4⁺ T cells derived from WT, CD28^Y170F, or p110δ^D910A T cells (10⁷/mouse) were injected i.v. into male C57BL/6 mice that had received an i.p. injection of IFN-γ (600 U) 72 hours earlier. Infiltration of the peritoneal membrane by labeled T cells and enrichment of HY-specific T cells in the peritoneal lavage were assessed 24 hours later as described for the experiments in Figure 3, D and E. WT T cells were readily identified from the peritoneal membrane and in the peritoneal lavage, demonstrating efficient recruitment from the circulation. The recruitment of HY-specific CD28^Y170F CD4⁺ T cells to the peritoneal cavity of male mice was moderately impaired, whereas the recruitment of p110δ^D910A T cells was virtually undetectable (Figure 4, A and B). In line with our previous findings, these observations suggest that CD28- and TCR-mediated PI3K activation can independently regulate T cell migration.

Figure 4

PI3K p110δ activation by TCR but not CD28 is required for antigen-dependent T cell recruitment. (A and B) H2-A^b–restricted HY-specific WT (black bars), CD28^Y170F (white bars), and p110δ^D910A (light gray bars) CD4⁺ T cells were injected i.v. into male mice (10⁷/mouse) that had received an i.p. injection of IFN-γ (600 U) 48 hours earlier. The following day, mice were sacrificed, and the presence of fluorescently labeled cells in the peritoneal cavity (A) and membrane (B) was assessed by flow cytometry and wide-field fluorescence microscopy, respectively. To facilitate visualization by flow cytometry, cells were double stained with an APC-conjugated anti-CD4 antibody following harvesting. The mean T cell numbers ± SEM observed in samples from at least 3 animals are shown. A: *P < 0.04, CD^28Y170F versus WT T cells; **P < 0.001, p110δ^D910A versus WT T cells; P < 0.005, p110δ^D910A versus CD^28Y170F T cells. B: *P < 0.03, CD^28Y170F versus WT T cells; **P < 0.002, p110δ^D910A versus WT T cells; P < 0.007, p110δ^D910A versus CD^28Y170F T cells. (C and D) HY-specific WT or p110δ^D910A CD4⁺ T cells that had either undergone antibody-mediated CD28 ligation (30 minutes at 37°C, PKH26-labeled) or had been pretreated with an antibody isotype control (IsC) (CFSE-labeled) were injected i.v. (10⁷/mouse) into male mice that had received an i.p. injection of IFN-γ 48 hours earlier. The presence of fluorescently labeled cells in the peritoneal cavity (C) and membrane (D) was assessed 24 hours later as described for A and B. The mean T cell number ± SEM observed in samples from at least 3 animals is shown. C: *P < 0.02 versus WT + IsC; D: *P < 0.02 versus WT + IsC.

We have previously shown that anti-CD28 can stimulate trafficking of T cells into nonlymphoid tissues and that this was dependent on the tyrosine within a YMNM motif in the cytoplasmic domain of CD28 (13). As this tyrosine can bind the SH2 domains of Grb2 and Gads in addition to the SH2 domains of p85, we sought to independently determine whether PI3K p110δ is required for CD28-dependent trafficking. To this end, HY-specific CD4⁺ WT or p110δ^D910A T cells (10⁷/mouse) that had either undergone CD28 ligation (PKH26-labeled) or had been pretreated with an isotype control antibody (CFSE-labeled) were coinjected i.v. into male C57BL/6 mice that had received an i.p. injection of IFN-γ (600 U) 72 hours earlier, as previously described (13). Tissue infiltration by labeled T cells and enrichment of HY-specific T cells in the peritoneal lavage were assessed 24 hours later as described above and in Methods. CD28 ligation significantly enhanced the recruitment of WT T cells, which were detected in the peritoneal membrane and lavage (Figure 4, C and D). By contrast, peritoneal recruitment of HY-specific p110δ^D910A T cells was not reconstituted by CD28 triggering. Isotype control antibody–treated and untreated T cells displayed similar localization (data not shown). These results show that CD28 through its YMNM motif signals via p110δ to promote tissue infiltration.

PI3K p110δ activity is instrumental in retention of antigen-specific T cells into antigenic tissue. Once T cells have reached the antigenic sites, they are retained in the tissue by mechanisms involving antigen presentation by resident antigen-presenting cells (27). To investigate the contribution of TCR-induced PI3K p110δ activity on this effect, we used an established model of antigen-dependent tissue infiltration (22, 25). In this model, tissue infiltration is strictly dependent on recognition of HY antigen within the peritoneal tissue. As in the model described above, male or female C57BL/6 mice were injected with 600 U IFN-γ. Two days later, PKH26-labeled, HY-specific WT or p110δ^D910A T cells (3 × 10⁶ cells/mouse) were injected directly into the peritoneum. The following day mice were sacrificed, and T cells that had infiltrated the peritoneal membrane or remained in the peritoneal lavage were detected by wide-field fluorescence microscopy and flow cytometry, respectively. HY-specific WT T cells efficiently localized into the peritoneal membrane of male recipients (Figure 5, A and B) and were depleted from the peritoneal cavity (Figure 5, C and D), whereas only a few p110δ^D910A T cells could be detected in the peritoneal membrane (Figure 5, A and B) and were instead readily recovered from the peritoneal lavage of male recipients (Figure 5, C and D).

Figure 5

PI3K p110δ is required for tissue infiltration by antigen-specific T cells. Female and male C57BL/6 mice were treated i.p. with 600 U IFN-γ. After 48 hours, 3 × 10⁶ PKH26-labeled WT or p110δ^D910A HY-specific H2-A^b–restricted CD4⁺ T cells (10⁷/mouse) were injected i.p. The presence of labeled T cells in the peritoneal membrane and lavage was analyzed after 24 hours by wide-field fluorescence microscopy (A and B) and flow cytometry (C and D), respectively. In addition, the presence of labeled cells in the spleen was quantified by wide-field fluorescence microscopy (E). Original magnification, ×10. Tissue infiltration was quantified by randomly selecting ten ×10-magnified fields and assessing the number of fluorescent cells in each field. The mean values ± SD observed in samples from at least 3 animals are summarized in B and E (infiltration of the peritoneal membrane and spleen, respectively) and D (cells retrieved in the peritoneal lavage). B and D: *P < 0.004 versus p110δ^D910A; E: *P < 0.02 versus p110δ^D910A.

As expected, HY-specific WT and p110δ^D910A T cells were not detected in the peritoneal membrane but were readily recovered from the peritoneal cavity of female recipients. In female recipient mice, the few T cells that had entered the peritoneum also entered the general circulation as evidenced by their recovery from the spleens (Figure 5E). By contrast, in male recipients, few WT T cells could be recovered from the spleens, reflecting their retention in the peritoneal wall. In contrast, p110δ^D910A T cells were readily detected in the spleens of male recipient mice as well. We conclude that p110δ controls both the recruitment and retention of T cells in nonlymphoid tissues without affecting passive entry into the general circulation. These findings raise the possibility that p110δ-mediated signals may be required for the establishment of antigen-dependent stable interactions with conventional antigen-presenting cells.

Pharmacological inhibition of PI3K p110δ abrogates antigen-dependent T cell recruitment and retention into antigenic tissue. Having established its role in mediating TCR-dependent T cell recruitment, we sought to investigate the effectiveness of pharmacological inhibition of p110δ activity on this effect. The IC87114 compound has been reported to selectively inhibit p110δ at doses up to 5–10 μM both in vitro and in vivo (28). In all the experiments described below, control T cells were treated with PBS containing vehicle (1% DMSO).

First we carried out studies to establish whether IC87114 might affect the viability of ECs or their ability to mediate migration of T cells. IC87114-treated ECs did not display increased cell death or permeability following exposure to up to 5 μM of the inhibitor for 30 minutes to 24 hours (data not shown). In addition, compared with untreated endothelium, female-derived, IFN-γ–treated ECs were equally capable of mediating migration of HY-specific T cells following exposure of ECs to 5 μM IC87114, suggesting that EC function was not adversely affected by inhibition of PI3K p110δ activity (data not shown). Nonetheless, in order to exclude effects that IC87114 might have on other cell types, T cells were treated with the inhibitor and subsequently washed prior to use. In addition, exposure of T cells to the inhibitor (5 μM) for up to 48 hours did not affect their viability (data not shown).

Migration of untreated and IC87114-treated HY-specific T cells through IFN-γ–treated EC monolayers derived from C57BL/6 male or female mice and grown on a Transwell insert in vitro was compared. As shown in Figure 6, A and B, inhibition of PI3K p110δ activity prevented antigen-induced enhancement of HY-specific T cell migration, while it did not affect T cell migration through female-derived ECs.

Figure 6

Pharmacologic inhibition of PI3K p110δ prevents antigen-dependent T cell localization to antigenic sites. HY-specific H2-A^b–restricted CD4⁺ T cells were treated with 5 μM IC87114 or with medium containing the vehicle (1% DMSO) for 30 minutes at room temperature and subsequently washed before use. (A and B) Untreated (filled symbols) and treated (open symbols) T cells (5 × 10⁵/well) were seeded onto syngeneic female- or male-derived IFN-γ–treated EC monolayers. T cell migration was monitored as described in Figure 1. The percentage of migrated T cells at the given time points is reported as the average of 3 experiments of identical design. *P < 0.05, except at 2 and 4 hours. (C) PKH26-labeled HY-specific untreated (black bars) or IC87114-treated (white bars) T cells were injected i.v. into syngeneic female mice. Infiltration of the indicated tissues was assessed 24 hours later as described in Figure 1, B and C. The mean T cell infiltration ± SD observed in samples from at least 3 animals is shown. (D–G) Female and male C57BL/6 mice were treated i.p. with 600 U IFN-γ. After 48 hours, 10⁷ labeled untreated (black bars) or IC87114-treated (white bars) HY-specific T cells were injected i.v. (10⁷ T cells/mouse, D and E) or i.p. (3 × 10⁶ T cells/mouse, F and G). The presence of labeled T cells in the peritoneal membrane and cavity was analyzed after 24 hours as described in the legends to Figures 3 and 5, respectively. The mean values ± SD observed in samples from at least 3 animals are shown. *P < 0.02 (D), *P < 0.05 (E), *P < 0.02 (F), *P < 0.03 (G) versus untreated T cells.

To assess the effect of pharmacological inhibition of PI3K p110δ^D910A on constitutive T cell trafficking in vivo, we compared the recruitment of HY-specific untreated and IC87114-treated T cells from the circulation into various tissues as described in Figure 1. PKH26-labeled untreated or IC87114-treated (5 μM, 1 hour at 37°C) H2-A^b–restricted CD4⁺ T cells (10⁷/mouse) were injected i.v. into C57BL/6 WT female mice. Similar numbers of T cells localized in lymphoid or different nonlymphoid tissues, suggesting that pharmacological inhibition of PI3K p110δ does not affect constitutive memory T cell trafficking (Figure 6C).

Next, to assess the effect of pharmacological inhibition of PI3K p110δ^D910A on antigen-dependent T cell localization to antigenic sites in vivo, we compared the recruitment of HY-specific untreated and IC87114-treated T cells from the circulation into the peritoneum of male and female mice as described above. PKH26-labeled untreated or IC87114-treated (5 μM, 1 hour at 37°C, CFSE-labeled) H2-A^b–restricted CD4⁺ T cells (10⁷/mouse), were injected i.v. into male or female C57BL/6 mice that had received an i.p. injection of IFN-γ (600 U) 48 hours earlier to induce local HY antigen presentation in male recipients. As shown in Figure 6, D and E, significant numbers of untreated T cells were detected in the peritoneal membrane and cavity of male mice. By contrast, HY-specific IC87114-inhibited T cells failed to localize to the peritoneal membrane and lavage of male mice. As expected, neither untreated nor IC87114-treated T cells localized in the peritoneal membrane and lavage of IFN-γ–treated female mice.

Finally, the effect of pharmacological inhibition of p110δ activity on tissue infiltration by HY-specific CD4⁺ T cells was investigated by experiments similar to those described in Figure 5. T cells were treated with IC87114 prior to i.p. injection into male and female C57BL/6 mice that had received an i.p. injection of IFN-γ (600 U) 48 hours earlier. As shown in Figure 6, F and G, pharmacologic inactivation of PI3K p110δ dramatically inhibited tissue infiltration by HY-specific T cells in male mice, which remained localized in the peritoneal cavity.

Taken together, these observations suggest that selective pharmacological inhibition of PI3K p110δ activity is effective in preventing the recruitment of antigen-specific T cells and their localization into target tissue.

Genetic and pharmacological inactivation of PI3K p110δ activity prevents HY-specific T cell infiltration of HY-mismatched skin grafts. Having established that p110δ-mediated signals are required for antigen-dependent T cell recruitment into — and infiltration of — antigenic nonlymphoid tissue, we investigated the effect of PI3K p110δ inactivation on the development of HY-specific T cell–mediated inflammation in HY-mismatched skin grafts.

First, we assessed the effect that antigen display exerts on T cell localization in this model. C57BL/6 female mice received 3 skin grafts: one from a syngeneic female mouse, one from a syngeneic male mouse (usually rejected between days 19 and 31 in WT female mice), and the third from a CBA/Ca male mouse (usually rejected by day 12). Four to 9 days following grafting, mice were injected i.v. with PKH26-labeled HY-specific CD4⁺ T cells (10⁷/mouse), and T cell infiltration in the grafts was assessed the following day (after sacrifice). It has to be noted that CBA/Ca-derived grafts bear MHC molecules of the H2-K haplotype, while the HY-specific T cells used here are H2-A^b restricted. As shown in Figure 7, A–D, HY-specific T cells selectively infiltrated antigenic (male) grafts, suggesting that antigen location plays a major role in determining primed T cell localization. Graft infiltration was assessed 24 and 48 hours after T cell transfer with similar results (data not shown).

Figure 7

Lack of PI3K p110δ activity prevents T cell localization into HY-mismatched skin grafts. C57BL/6 female mice received 3 skin grafts: one from a syngeneic male mouse (A), one from a syngeneic female mouse (B), and the third from a CBA/Ca male mouse (C). Four days following grafting, HY-specific A^b-restricted WT (PKH26-labeled) CD4⁺ T cells (10⁷/mouse), were injected i.v. Graft infiltration by labeled T cells was assessed 24 hours later by wide-field fluorescence microscopy, as described in the legend to Figure 1, C and D. Original magnification, ×10. The mean values ± SD observed in samples from at least 3 animals are summarized in D. *P < 0.03 versus female and CBA/Ca graft infiltration. Parallel experiments were run in which C57BL/6 female mice engrafted with C57BL/6 male-derived skin 4 days earlier were coinjected with either WT (PKH26-labeled, red) or p110δ^D910A (CFSE-labeled, green) HY-specific CD4⁺ T cells (E and F) or with untreated (PKH26-labeled, red) or IC87114-treated (CFSE-labeled, green) HY-specific CD4⁺ T cells (G and H). Grafts were removed 24 hours later, and tissue infiltration was quantified by randomly selecting ten ×10-magnified fields from tissue samples from at least 3 animals and assessing the number of fluorescent cells in each field. The few CFSE-labeled cells are indicated by an arrowhead. Nuclei are stained by DAPI (blue). A–C show representative ×10-magnified tissue images, while the images shown in E and G were taken at a ×20 magnification. The mean T cell infiltration ± SD observed in samples from at least 3 animals is shown; *P < 0.03 (F); *P < 0.01 (H).

Next, we investigated the effect of p110δ inactivation on the infiltration of HY-mismatched skin grafts by HY-specific H2-A^b–restricted CD4⁺ T cells. C57BL/6 female mice that had received syngeneic male mouse skin graft were coinjected i.v. with PKH26-labeled, HY-specific WT CD4⁺ T cells (10⁷/mouse) together with either CFSE-labeled HY-specific p110δ^D910A T cells (Figure 7, E and F) or CFSE-labeled IC87114-pretreated (5 μM, 1 hour at 37°C; Figure 7, G and H) WT T cells. Genetic (Figure 7, E and F) or pharmacological (Figure 7, G and H) inhibition of PI3K p110δ strongly inhibited male skin graft infiltration by HY-specific CD4⁺ T cells (the very few green-labeled cells are indicated by arrows in the figure). No differences were observed in T cell infiltration of female-derived skin grafts, which was virtually absent (data not shown).

TCR-derived signals can determine the fate of T cells in many ways. Antigen stimulates T cells to divide and differentiate or, alternatively, to become anergic. TCR-derived signals are also instrumental for directing T cells to target tissue where antigen can be found. The outcome of TCR signaling is also influenced by other receptor systems. For example, the delivery of costimulatory signals determines whether T cells expand or become unresponsive, whereas TCR triggering in the presence of certain cytokines induces the development of different T helper cell subsets. Thus, extrinsic factors, such as the microenvironment in which antigen presentation takes place and the type of antigen-presenting cell, help influence the outcome of TCR signaling. Here we present evidence that in addition to regulating T cell proliferation and differentiation, PI3K p110δ activity is required for TCR-dependent localization of T cells to tissues that play host to cognate antigens.

Although PI3Ks have been suggested to play a key role in chemotaxis of Dictyostelium and neutrophils, PI3Ks are not major regulators of chemotaxis in T cells (15–17). Where PI3K do play a role, it is the class IB PI3K p110γ that is required (~10% of the total response) (15). The main regulator of chemotaxis in T cells is DOCK2, which appears, to a large extent, to regulate migration independently of PI3Ks (15).

In line with previous observations (16), our data show that primed T cell lacking PI3K p110δ display intact constitutive trafficking and chemotactic responses both in vitro and in vivo; however, they fail to localize to antigenic tissue, suggesting that PI3K p110δ is required for T cell migration to antigenic sites independently of chemokine signaling. This effect does not appear to be subset restricted, as both CD4⁺ and CD8⁺ memory T cells display similar requirements for p110δ activity to be recruited in an antigen-dependent manner. In this context, class I PI3K regulatory subunit p85 has been shown to contribute to spontaneous lymphocyte motility in the lymph node (18). As PI3Ks have been implicated in signaling pathways leading to integrin activation, it is possible that p85 activity may be required for efficient cell-cell interactions in the lymph nodes. Alternatively, as it has been suggested that PI3K requirement for lymphocyte migration may depend on their differentiation as well as the microenvironment in which T cell locomotion takes place (14), it is possible that the requirement for class I activity may differ between naive T cells in the LN and differentiated T cells migrating to nonlymphoid tissue (as those used in the present study).

Although our data do not support a role for PI3K p110δ activity in TCR-independent chemokine-induced responses by memory T cells, TCR-driven T cell trafficking has previously been shown to be enhanced by the concomitant delivery of chemokine-mediated signals (5, 29). The complete abrogation of TCR-dependent T cell migration in the absence of PI3K p110δ activity as observed in this study suggests that this activity may precede and may be required for the subsequent contribution of chemokine-initiated signaling pathways (mediated by either by DOCK2 or p110γ). In this context, a direct cross-talk between TCR- (mediated by zeta-associated protein 70 [ZAP-70]) and chemokine receptor–mediated signaling has recently been shown to be essential for chemokine-induced T cell migration (30).

We have recently shown that CD28-mediated signals are required for primed T cells to access nonlymphoid tissue (22). As murine endothelium expresses CD80 (but not CD86) molecules (31), we sought to discriminate the relative contribution of TCR- and CD28-dependent PI3K activation to the regulation of primed T cell recruitment. In line with our previous observation, TCR- and CD28-mediated effects on T cell migration can act independently. However, our data suggest that both receptors rely on P110δ activity to exert their effects. Therefore, it is likely that inhibition of PI3K P110δ activity would prevent antigen-dependent recruitment of T cells by human endothelium, which does not express either CD80 or CD86 molecules.

Finally, our results show that selective PI3K p110δ inhibitors have the potential to reduce access and retention of specific T cells into different target tissues. Small molecule inhibitors against p110δ are currently being developed with the aim of alleviating autoimmune and inflammatory diseases (32). Our observation that pharmacological p110δ inhibition prevents TCR-dependent T cell recruitment to skin grafts may be relevant to autoimmune diseases such as psoriasis and paves the way for a wider application of this strategy for a variety of T cell–mediated pathologies, including autoimmunity and transplantation. These inhibitors may prevent antigen-dependent T cell migration and subsequently cell-cell interactions without inducing overt immune suppression, thus representing a novel and desirable target for the regulation of undesired T cell inflammation.

View Supplemental data

This work was supported by the British Heart Foundation (BHF PG/05/136/19997). We are grateful to Liz Simpson (Imperial College London) and Alan Salama (Imperial College London) for critical review of the manuscript.

Address correspondence to: Federica M. Marelli-Berg, Department of Immunology, Imperial College London, Hammersmith Hospital Campus, Du Cane Road, London W12 0NN, United Kingdom. Phone: 44-0-2083831704; Fax: 44-0-2083832788; E-mail: f.marelli@imperial.ac.uk.

Nonstandard abbreviations used: DOCK2, dedicator of cytokinesis 2; PI3K, phosphoinositide-3-kinase; rIL-2, recombinant IL-2.

Conflict of interest: Klaus Okkenhaug is a paid consultant of PIramed Inc., Slough, United Kingdom. The other authors have declared that no conflict of interest exists.

Reference information: J. Clin. Invest.118:1154–1164 (2008). doi:10.1172/JCI33267

Anne J. Ridley’s present address is: King’s College London, Randall Division of Cell and Molecular Biophysics, London, United Kingdom.

1. Mora, J.R., von Andrian, U.H. 2006. T-cell homing specificity and plasticity: new concepts and future challenges. Trends Immunol. 27:235-243.
   View this article via: CrossRef PubMed Google Scholar
2. Marelli-Berg, F.M., Frasca, L., Imami, N., Lombardi, G., Lechler, R.I. 1999. Lack of T cell proliferation without induction of nonresponsiveness after antigen presentation by endothelial cells. Transplantation. 68:280-287.
   View this article via: CrossRef PubMed Google Scholar
3. Marelli-Berg, F.M., Frasca, L., Weng, L., Lombardi, G., Lechler, R.I. 1999. Antigen recognition influences transendothelial migration of CD4+ T cells. J. Immunol. 162:696-703.
   View this article via: PubMed Google Scholar
4. Marelli-Berg, F.M., et al. 2004. Cognate recognition of the endothelium induces HY-specific CD8+ T-lymphocyte transendothelial migration (diapedesis) in vivo. Blood. 103:3111-3116.
   View this article via: CrossRef PubMed Google Scholar
5. Savinov, A.Y., Wong, F.S., Stonebraker, A.C., Chervonsky, A.V. 2003. Presentation of antigen by endothelial cells and chemoattraction are required for homing of insulin-specific CD8+ T cells. J. Exp. Med. 197:643-656.
   View this article via: CrossRef PubMed Google Scholar
6. van Kooyk, Y., et al. 1989. Enhancement of LFA-1-mediated cell adhesion by triggering through CD2 or CD3 on T lymphocytes. Nature. 342:811-813.
   View this article via: CrossRef PubMed Google Scholar
7. Dustin, M.L., Springer, T.A. 1989. T-cell receptor cross-linking transiently stimulates adhesiveness through LFA-1. Nature. 341:619-624.
   View this article via: CrossRef PubMed Google Scholar
8. Dustin, M.L., Bromley, S.K., Kan, Z., Peterson, D.A., Unanue, E.R. 1997. Antigen receptor engagement delivers a stop signal to migrating T lymphocytes. Proc. Natl. Acad. Sci. U. S. A. 94:3909-3913.
   View this article via: CrossRef PubMed Google Scholar
9. Okkenhaug, K., Vanhaesebroeck, B. 2003. PI3K in lymphocyte development, differentiation and activation. Nat. Rev. Immunol. 3:317-330.
   View this article via: CrossRef PubMed Google Scholar
10. Okkenhaug, K., et al. 2002. Impaired B and T cell antigen receptor signaling in p110delta PI 3-kinase mutant mice. Science. 297:1031-1034.
    View this article via: PubMed Google Scholar
11. Patton, D.T., et al. 2006. Cutting edge: the phosphoinositide 3-kinase p110 delta is critical for the function of CD4+CD25+Foxp3+ regulatory T cells. J. Immunol. 177:6598-6602.
    View this article via: PubMed Google Scholar
12. Garcon, F., et al. 2007. CD28 provides T cell costimulation and enhances PI3K activity at the immune synapse independently of its capacity to interact with the p85/p110 heterodimer. Blood. 111:1464-1471.
    View this article via: CrossRef PubMed Google Scholar
13. Deane, J.A., Fruman, D.A. 2004. Phosphoinositide 3-kinase: diverse roles in immune cell activation. Annu. Rev. Immunol. 22:563-598.
    View this article via: CrossRef PubMed Google Scholar
14. Ward, S.G. 2006. T lymphocytes on the move: chemokines, PI 3-kinase and beyond. Trends Immunol. 27:80-87.
    View this article via: CrossRef PubMed Google Scholar
15. Nombela-Arrieta, C., et al. 2004. Differential requirements for DOCK2 and phosphoinositide-3-kinase gamma during T and B lymphocyte homing. Immunity. 21:429-441.
    View this article via: CrossRef PubMed Google Scholar
16. Reif, K., et al. 2004. Cutting edge: differential roles for phosphoinositide 3-kinases, p110gamma and p110delta, in lymphocyte chemotaxis and homing. J. Immunol. 173:2236-2240.
    View this article via: PubMed Google Scholar
17. Nombela-Arrieta, C., et al. 2007. A central role for DOCK2 during interstitial lymphocyte motility and sphingosine-1-phosphate-mediated egress. J. Exp. Med. 204:497-510.
    View this article via: CrossRef PubMed Google Scholar
18. Matheu, M.P., Deane, J.A., Parker, I., Fruman, D.A., Cahalan, M.D. 2007. Class IA phosphoinositide 3-kinase modulates basal lymphocyte motility in the lymph node. J. Immunol. 179:2261-2269.
    View this article via: PubMed Google Scholar
19. Asperti-Boursin, F., Real, E., Bismuth, G., Trautmann, A., Donnadieu, E. 2007. CCR7 ligands control basal T cell motility within lymph node slices in a phosphoinositide 3-kinase-independent manner. J. Exp. Med. 204:1167-1179.
    View this article via: CrossRef PubMed Google Scholar
20. Barber, D.F., et al. 2006. Class IB-phosphatidylinositol 3-kinase (PI3K) deficiency ameliorates IA-PI3K-induced systemic lupus but not T cell invasion. J. Immunol. 176:589-593.
    View this article via: PubMed Google Scholar
21. Okkenhaug, K., et al. 2006. The p110delta isoform of phosphoinositide 3-kinase controls clonal expansion and differentiation of Th cells. J. Immunol. 177:5122-5128.
    View this article via: PubMed Google Scholar
22. Mirenda, V., et al. 2007. Physiological and aberrant regulation of memory T cell trafficking by the costimulatory molecule CD28. Blood. 109:2968-2977.
    View this article via: PubMed Google Scholar
23. Cinamon, G., Shinder, V., Alon, R. 2001. Shear forces promote lymphocyte migration across vascular endothelium bearing apical chemokines. Nat. Immunol. 2:515-522.
    View this article via: CrossRef PubMed Google Scholar
24. Acuto, O., Cantrell, D. 2000. T cell activation and the cytoskeleton. Annu. Rev. Immunol. 18:165-184.
    View this article via: CrossRef PubMed Google Scholar
25. James, M.J., et al. 2003. Anergic T cells exert antigen-independent inhibition of cell-cell interactions via chemokine metabolism. Blood. 102:2173-2179.
    View this article via: CrossRef PubMed Google Scholar
26. Okkenhaug, K., et al. 2001. A point mutation in CD28 distinguishes proliferative signals from survival signals. Nat. Immunol. 2:325-332.
    View this article via: CrossRef PubMed Google Scholar
27. Kawai, T., Shimauchi, H., Eastcott, J.W., Smith, D.J., Taubman, M.A. 1998. Antigen direction of specific T-cell clones into gingival tissues. Immunology. 93:11-18.
    View this article via: CrossRef PubMed Google Scholar
28. Ali, K., et al. 2004. Essential role for the p110delta phosphoinositide 3-kinase in the allergic response. Nature. 431:1007-1011.
    View this article via: CrossRef PubMed Google Scholar
29. Svensson, M., et al. 2002. CCL25 mediates the localization of recently activated CD8{alpha}β+ lymphocytes to the small-intestinal mucosa. J. Clin. Invest. 110:1113-1121.
    View this article via: JCI CrossRef PubMed Google Scholar
30. Kumar, A., et al. 2006. CXCR4 physically associates with the T cell receptor to signal in T cells. Immunity. 25:213-224.
    View this article via: CrossRef PubMed Google Scholar
31. Marelli-Berg, F.M., Peek, E., Lidington, E.A., Stauss, H.J., Lechler, R.I. 2000. Isolation of endothelial cells from murine tissue. J. Immunol. Methods. 244:205-215.
    View this article via: CrossRef PubMed Google Scholar
32. Rommel, C., Camps, M., Ji, H. 2007. PI3K delta and PI3K gamma: partners in crime in inflammation in rheumatoid arthritis and beyond? Nat. Rev. Immunol. 7:191-201.
    View this article via: CrossRef PubMed Google Scholar
33. Hebeis, B.J., Vigorito, E., Turner, M. 2004. The p110delta subunit of phosphoinositide 3-kinase is required for the lipopolysaccharide response of mouse B cells. Biochem. Soc. Trans. 32:789-791.
    View this article via: CrossRef PubMed Google Scholar
34. Tarakhovsky, A., et al. 1995. Defective antigen receptor-mediated proliferation of B and T cells in the absence of Vav. Nature. 374:467-470.
    View this article via: CrossRef PubMed Google Scholar
35. Scott, D., et al. 2000. Dendritic cells permit identification of genes encoding MHC class II-restricted epitopes of transplantation antigens. Immunity. 12:711-720.
    View this article via: CrossRef PubMed Google Scholar
36. Millrain, M., et al. 2001. Examination of HY response: T cell expansion, immunodominance, and cross-priming revealed by HY tetramer analysis. J. Immunol. 167:3756-3764.
    View this article via: PubMed Google Scholar
37. Wells, C.M., Ridley, A.J. 2005. Analysis of cell migration using the Dunn chemotaxis chamber and time-lapse microscopy. Methods Mol. Biol. 294:31-41.
    View this article via: PubMed Google Scholar
38. Billingham, R., Medawar, P.B. 1951. The technique of free skin grafting in mammals. J. Exp. Biol. 28:385-402.