Little is known about the signaling that occurs in an APC during contact with a T cell. In this article we report the concentration of the signaling lipid phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2) at the APC side of the immunological synapse. In both human and mouse cells, a PI(4,5)P2-specific fluorescent reporter, PH-GFP (where PH is pleckstrin homology), detected an Ag-dependent enrichment of PI(4,5)P2 at the synapse between Ag-specific T cells and APC. When PIP(4,5)P2 was sequestered by a high concentration of PH-GFP reporter, cells were less susceptible to CTL-mediated lysis than control cells. These findings suggest a new regulatory target for modulating immune function that may be exploited for immune escape by pathogens and tumors.
Engagement of T cells and APCs is highly orchestrated. Cognate cells must find each other in an Ag-specific manner and, upon recognition, form conjugates that remain stable for minutes to hours. In vitro, these contacts have been shown to involve rearrangements of receptors and ligands into a highly organized immunological synapse (IS)4 (1). Many molecules that localize to the synapse play important signaling roles for activation and effector function (2, 3, 4).
Despite a wealth of study into the effector cell side of the immunological synapse, there has been little progress in understanding changes in organization and signaling on the APC side upon T cell engagement. APC molecules that localize to the synapse are T cell ligands; their redistribution into central and proximal supramolecular complexes (5) seems to be driven and regulated by the T cell (6). A few studies demonstrate a role for actin organization on the APC side of the synapse (7, 8). One recent paper showed a scaffolding protein of neuronal synapses localizing to the immunological synapse (9). Beyond these findings, little more is known about how the APC side of the synapse is organized and about what, if any, signaling occurs there.
This report shows that phosphatidylinositol 4,5-bisphosphate (PI(4,5)P2), an actin-modulating and signaling lipid (10), concentrates at the immunological synapse in APCs. Through its interactions with many regulators of actin polymerization, including Cdc42, neural Wiskott-Aldrich syndrome protein, and gelsolin (11), it maintains cell shape and membrane integrity (12). In addition, PI(4,5)P2 regulates dynamic processes such as cell movement, cell reshaping, phagocytosis, and vesicle traffic (13, 14, 15). PI(4,5)P2 anchors cytosolic factors to the membrane; its own localization within the membrane has been shown to be cholesterol dependent (12, 16). Upon cleavage by phospholipases, PI(4,5)P2 cleavage products can activate calcium channels and other signaling pathways (17).
To our knowledge, this is the first report of a signaling lipid being concentrated at the IS on the APC side. Our data show that this concentration of PI(4,5)P2 plays a role in function and suggest that other signaling pathways may be triggered at the synapse in the APC.
Materials and Methods
Cell lines, constructs, and treatments
JY HLA-A2,B7 human B-cells and T2-Kb cells are described elsewhere (18). T2-Kb cells were pulsed with the SIY peptide (SIYRYYGL) for 1–2 h at 37°C. Allogenic, HLA-A2-negative T cells were provided by the Johns Hopkins Medical Institutions Laboratory of Immunogenetics, Baltimore, MD. Naive alloreactive T cells were activated by coculture with irradiated JY stimulator cells for 4 days (18). 2C T cells (19) were provided by J. Schneck (Department of Pathology, Johns Hopkins Medical Institutions) and activated by irradiated splenocytes from BALB/c mice. Expression constructs for pleckstrin homology (PH)-GFP and PH-GFP R40L were provided by T. Balla (National Institutes of Health, Bethesda, MD) (20). T2 cells were transfected, transiently or stably, by electroporation with ∼10 μg of DNA. Stable cell lines were generated by drug selection (300 μg/ml G418) and flow cytometric sorting.
Death of APCs was measured in terms of annexin V and 7-aminoactinomycin D labeling. T2-Kb cells, pulsed with SIY peptide (19), were mixed with activated 2C T cells at an E:T ratio of 5:1 for 2 h followed by annexin V staining for apoptosis. FACS analysis was conducted on a BD FACSCalibur flow cytometer. Anti-mouse CD8 (BD Biosciences), anti-mouse H2-Kb 20.8.4s (21), and anti-HLA Ke2 were used for staining (22).
For MHC crosslinking experiments, ionized glass coverslips were coated with the mAbs anti-Kb, cloneY3 (23), or anti-CD59, clone MEM43 (24). T2-Kb APCs were allowed to settle and adhere to the coverslip and then fixed after 30 min with 4% paraformaldehyde. Alternatively, anti-MHC beads were constructed using 5-μm magnetic protein A beads (Dynal) conjugated with mAb anti-Kb (20.8.4s). For MHC capping experiments, cells were labeled with 20.8.4-Cy3 at 37°C for 15 min and then washed and fixed as described above.
All imaging was done on a Zeiss LSM 510-meta confocal microscope using the appropriate filter settings and laser lines. For fixed conjugates, T2-Kb APCs and 2C-activated T cells or beads were incubated for 30 min and then fixed in 4% paraformaldehyde. Cells were washed and mounted in microslides (Vitrotubes). For live cell imaging, alloreactive T cells were mixed with JY B cells, lightly centrifuged, and mounted in PBS plus 1% FBS in microslides. The stage was warmed to 37°C. Images were analyzed using Image Examiner (Zeiss) and ImageJ. Quantification and plots were generated using Excel (Microsoft) and Prism 4.0 (GraphPad).
Two APC cell lines, JY (HLA-A2) and T2-Kb (H2-Kb), were used to investigate the response of APCs to T cell engagement. The APCs were transfected with PH-GFP, a fusion protein that binds PI(4,5)P2 and reports its cellular localization. In APCs expressing PH-GFP, a characteristic uniform membrane localization was observed (Fig. 1⇓A) as previous reports have shown (25). Cells expressing the mutant variant PH-GFP R40L showed a diffuse cytoplasmic distribution (Fig. 1⇓B). When effector T cells were mixed with Ag-specific APCs, PI(4,5)P2 concentrated at the interface between the APC and the T cell (Fig. 1⇓, C and D). There was a 30–50% increase in intensity at the interface as compared with the rest of the APC cell membrane (Fig. 1⇓, E and F). This enrichment was seen for both live and fixed cell conjugates. PH-GFP R40L did not concentrate at the IS. Because these sites of contact have been shown by electron microscopy to have single membrane thickness (26), it is unlikely that membrane ruffling could explain the increase in signal, but rather argues for an active recruitment of PI(4,5)P2.
To see whether MHC engagement alone was sufficient to induce PIP2 redistribution, coverslips were coated with either anti-MHC Abs or control Ab (anti-CD59) and allowed to conjugate with APCs. We did not observe a change in PH-GFP surface distribution when MHC I molecules or CD59 molecules (controls) were bound by Ab on coverslips or beads (supplemental Fig. S1).5 Capping MHC with soluble Ab also did not induce a consistent redistribution of PH-GFP.
The accumulation of PI(4,5)P2 at the immunological synapse suggested a functional response by APCs to T cell engagement. Live cell imaging was used to observe the dynamics of PI(4,5)P2 lipids and cell lysis. During live cell imaging of CTL-APC conjugates, in some cases, T cells were in contact with multiple target cells, each expressing PH-GFP at a different level. Time lapse imaging of these aggregates showed that while T cells scanned APCs without bias, they preferentially lysed cells expressing lower levels of PH-GFP; APCs that expressed higher levels of PH-GFP were more resistant to T cell lysis. Two sets of time lapse images are shown in Fig. 2⇓ (and in supplemental movie 1).
PH-GFP is frequently used as a reporter for PI(4,5)P2 localization, but at high expression levels it will compete with endogenous molecules that bind PI(4,5)P2 (25) resulting in a PI(4,5)P2 hypomorphic cell. This suggested that the resistance of PH-GFP-high cells is due to sequestering of PI(4,5)P2. To test this idea, we measured the ability of 2C CTLs to kill peptide-loaded T2Kb targets expressing PH-GFP or the R40L variant. Apoptosis was markedly reduced in PH-GFP-expressing APCs as compared with cells expressing the PH-GFP R40L (Fig. 3⇓). These results confirmed that blocking PI(4,5)P2 in APCs inhibited effector T cell function. A similar reduction was found for alloreactive T cell-mediated lysis of JY targets expressing PH-GFP, as measured by a chromium release assay (data not shown).
Sequestering PI(4,5)P2 by expression of PH-GFP did not reduce surface MHC class I levels (data not shown). MHC surface stability, measured in terms of MHC surface half-life after brefeldin A treatment, was also unaffected. This indicates that sequestering PI(4,5)P2 in APCs affects a pathway to cell death that is downstream of MHC display.
APCs concentrate PI(4,5)P2 in their contacts with T cells. Crosslinking MHC on the plasma membrane was insufficient to induce this concentration of PI(4,5)P2. This suggests that a more complex engagement is needed to generate signals from PI(4,5)P2. It is known that capping MHC molecules can induce calcium signals (27, 28), but only after several hours. Crosslinking of MHC did not induce local concentration of PIP2 in the same way as T cell engagement.
Because PIP(4, 5)P2 diffuses rapidly, it is likely that local synthesis is not sufficient to maintain its localization; rather, exchange with PIP2-binding proteins could retain these lipids at the synapse (29). Recruitment of PIP2-binding proteins may play a role in MHC I-dependent signal transduction, but further work will be required to elucidate their identities.
It is unclear how blocking MHC I signaling with PH-GFP affects CTL-mediated apoptosis. It is possible that PI(4,5)P2 sequestration affects T cell degranulation or the susceptibility of target cells to lysis. The accumulation of PI(4,5)P2 at the IS may be required for the uptake of cytolytic granules, because PI(4,5)P2 plays an important role in various pathways of endocytosis and phagocytosis at the plasma membrane (30, 31). It is thought that granzyme B function requires internalization and acidification before activation (32). It may be that PH-GFP expression inhibits a PI(4,5)P2-dependent endocytic pathway, such as the clathrin-coated trafficking pathway. By blocking granule uptake, the cells are protected from apoptosis.
APC responses to CTL contact are poorly defined. Various viral mechanisms exist to evade MHC class I presentation (33). PI(4,5)P2 inhibition represents a novel form of CTL-mediated inhibition. This strategy inhibits lysis without compromising class I presentation, which might activate NK cell recognition. A similar mechanism may be used by intracellular pathogens seeking to prevent apoptosis of their host cells. Furthermore, it may be fruitful to investigate the role of natural and synthetic phosphatidylinositol inhibitors, which are used to inhibit tumor growth (34), to see whether they have unexpected effects on cell-mediated immunity.
We thank Jonathan Schneck and his laboratory for 2C T cells, reagents, and critical discussions of the work.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was funded by National Institutes of Health Grant AI-14584 (to M.E.) and National Institutes of Health Training Grant 5T32 AI-07247 (to S.R.S.).
↵2 D.R.F. and S.R.S. contributed equally to this work.
↵3 Address correspondence and reprint requests to Dr. David Robert Fooksman at the current address: Skirball Institute, New York University School of Medicine, 540 First Avenue, New York, NY 10016. E-mail address:
↵4 Abbreviations used in this paper: IS, immunological synapse; PH, pleckstrin homology; PI(4,5)P2, phosphatidylinositol 4,5-bisphosphate.
↵5 The online version of this article contains supplemental material.
- Received June 4, 2008.
- Accepted March 4, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.