Abstract
CTLs can acquire MHC class I-peptide complexes from their target cells, whereas CD4+ T cells obtain MHC class II-peptide complexes from APCs in a TCR-specific manner. As a consequence, Ag-specific CTL can kill each other (fratricide) or CD4+ T cells become APCs themselves. The purpose of the acquisition is not fully understood and may be either inhibition or prolongation of an immunological response. In this study, we demonstrate that human CD4+ Th cells are able to capture membrane fragments from APC during the process of immunological synapse formation. The fragments contain not only MHC class II-peptide complexes but also MHC class I-peptide complexes, rendering these cells susceptible to CTL killing in an Ag-specific manner. The control of CD4+ Th cells by Ag-specific CTL, therefore, maybe another mechanism to regulate CD4+ T cell expansion in normal immune responses or cause immunopathoglogy during the course of viral infections such as HIV.
An immunological synapse forms within minutes of T cell and APC association; TCR and MHC proteins cluster, along with an assortment of adhesion molecules and activation coreceptors, causing sustained T cell signaling and cell activation (1, 2). Th lymphocytes recognize antigenic peptides presented by MHC class II molecules on the surface of APC. Similarly, CTL recognize peptides displayed by MHC class I molecules. Interestingly, during this process the T cells are able to acquire molecules from the APC, via a process termed membrane biting or trogocytosis (3, 4).
Various activated immunological cells acquire molecules from their targets. The capture of proteins from APC has previously been described for B cells (5, 6), γδ T cells, (7), NK cells (8), and dendritic cells readily release molecules, specifically MHC membrane components, for uptake by other dendritic cells (9). Most well described, however, is the transfer of APC-derived molecules to T cells; early studies revealed MHC class II molecules on mouse T cells, which do not normally express these proteins, after contact with APC (10, 11, 12, 13). Recent studies reveal that T cells can act as APCs after capturing MHC molecules (14, 15, 16, 17, 18, 19); however, the mechanism of capture and physiological functions are not yet elucidated. Several reports have indicated that the T cells can present the captured APC membrane components to other T cells, with varying outcomes. T cells are not professional APCs (20), and the lack of costimulatory molecules could lead to anergy in the triggered T cells (21) or specific cytotoxicity (15), but others have reported that costimulatory molecules can be coacquired and full effector CTL responses can be triggered by Th cells displaying APC components (22). The process of acquisition could have other immunological functions, it may aid the disengagement of T cells and APC postactivation (23) or drive affinity maturation of T cells (24).
In this study, we show the transfer of MHC class II-peptide complexes to specific human Th cells, the transfer is dependent on HLA type and appropriate peptide expression; as trogocytosis occurs, other molecules are coacquired by the T cells, significantly MHC class I-peptide complexes, which may result in an important immunoregulatory process.
Materials and Methods
Cells
293T cells were maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 1% penicillin-streptamycin (50 IU/ml and 50 μg/ml), and 1% glutamine. The T cell clones (HA1.7, KB-D1, and CMV-CTL) were expanded by addition of mixed (three donors) irradiated PBMC and PHA (5 μg/ml). T cells were cultured in RPMI 1640 supplemented with 10% human serum (R10), 1% penicillin-streptomycin (50 IU/ml and 50 μg/ml), and 1% glutamine. From 2 to 3 days poststimulation, 100 U/ml IL-2 were added, and medium was regularly changed. T cells were used for assays 9–12 days poststimulation and restimulated every 12–21 days. HA1.7 CD4+ T cells recognize the influenza hemagglutinin (HA)3 peptide (residues 307–318, PKYVKQNTLKLA) on HLA-DR1 molecules and were provided by Jonathan Lamb (25, 26). KB-D1 CD4+ T cell clones recognize an acetylcholine receptor (AchR) α peptide (residues 146–160, KLGTWTYDGSVVAIN) presented by HLA-DR52a and were provided by Professor Nick Willcox (Neuroscience, Weatherall Institute of Molecular Medicine, Oxford University, Oxford, U.K.). The CD8+ T cell clone CMV-CTL recognizes a pp65 CMV peptide (residues 495–503, NLVPMVATV) presented by HLA-A2 and was provided by Tao Dong and Tumelo Mashishi (Human Immunology Unit, Weatherall Institute of Molecular Medicine, Oxford University, Oxford, U.K.).
Construction of HLA constructs and peptides
DR α-chain was cloned into the pcDNA3.1 hygromycin expression vector (Invitrogen Life Technologies) and the DR β-chain (DRB1*0101) β-chain was cloned into the pEGFP-N1 expression vector (Clontech). HA307–318 was covalently tagged to the β-chain N terminus via a 12-aa Gly-Gly-Ser repeat, and a leader sequence preceded the peptide. GFP was tagged to the β-chain via the C-terminal cytoplasmic tail. Alternatively, the HA307–318 peptide was replaced by the acetylcholine receptor (AchR) α peptide (AchRα146–160). Full length human A2 was linked to GFP using the pEGFP-N1 vector. The pp65 CMV495–503 peptide was synthesized by Kati Di Gleria (Human Immunology Unit, Weatherall Institute of Molecular Medicine, Oxford, U.K.) and stored at 10 mM in 100% DMSO at −80°C.
Transfections and stable cell lines
From 1 to 2 μg of DNA were transfected per well (0.5 × 106 cells), using calcium phosphate precipitation (27). Cells were transfected at 50–60% confluency, and proteins were expressed for 2 days posttransfection. Stable cell lines were generated as follows. 293T cells were transiently transfected by calcium phosphate precipitation, and proteins were expressed for 2 days; then cells were split to 50% confluency, and 1 mg/ml G418 and 1 mg/ml hygromycin were added. After several weeks of selection, GFP-positive cells were single-cell selected by FACS, and clones were grown. HLA-DR expression was confirmed by FACS staining.
PKH-26 cell labeling
PKH-26 (Sigma-Aldrich) is a red dye that nonspecifically labels lipid cell membranes. 293T cells (2.3 × 106), or T cell clones, were washed in R10, and cells were resuspended in 250 μl of diluent C. PKH-26 dye was diluted to ×2 concentration in 250 μl of diluent C and then added to the cell suspension and gently mixed. Cells were incubated for 5 min at 25°C with gentle mixing, and then an equal volume of FCS was added to aid cell recovery and to stop staining. Cells were then washed four times in large volumes of R10.
FACS biting assay
293T cells (either transiently transfected or stably expressing MHC class I- or class II-peptide complexes or stained with PKH-26) were mixed with T cells (1:1 ratio) and were gently centrifuged for 1 min to increase cell interactions. Cells were then incubated at 37°C with 5% CO2 for the time indicated. Next, cells were incubated in 1 ml of 0.5 mM EDTA-PBS for 10 min at 4°C to disrupt cell-cell interactions. Cells were fixed in 2% formaldehyde-PBS for 10 min and stained with Abs toward CD4-CD8 and CD3 to identify T cell populations. FACS was performed using a BD Biosciences FACSCalibur.
Immunofluorescence
293T cells expressing wild-type (wt) DR1-HA-GFP molecules were incubated with T cell clones for 1 h at 37°C and then stained with first layer Ab UCHT-1 (CD3 unconjugated; BD Biosciences) and second layer Ab anti-mouse-Alexa Fluor 568 (Molecular Probes). Cells were fixed for 10 min in 2% formaldehyde-PBS and briefly incubated with Vectorshield-4′,6′-diamidino-2-phenylindole (Vector Laboratories). Cells were placed on glass slides and viewed by oil immersion fluorescence microscopy (×100).
Biting and killing assay
293T cells stably expressing wt DR1-HA-GFP were incubated with or without 10 μM CMV495–503 for 1.5 h at 37°C. In parallel, HA1.7 or KB-D1 T cell clones were labeled with 10 μM or 0.5 μM PKH-26 red lipid membrane dye, respectively, as described previously. Cells were washed extensively, and then HA1.7 and KB-D1 cells were mixed, at a 1:1 ratio, before addition of CMV495–503 ± wt DR1-HA-GFP expressing 293T cells. Interactions were allowed for 1.5 h at 37°C; then T cells were isolated using a CD4+ isolation kit (Dynal Biotech). Briefly, cells were washed and resuspended in 2% FCS-PBS, and then for each 1-ml sample 107 anti-CD4− beads were added, using no more than 0.5 × 107 cells/ml. Cells and beads were incubated for 20 min at room temperature with gentle shaking and then isolated using a Dynal-MCP magnet. Supernatant was discarded, and rosetted cells were washed in 2% FCS-PBS; then cells were resuspended in R10. We added 10 μl of Detachabead reagent per 100-μl sample and incubated the mixture for 45 min at 4°C with rotation. Beads were extracted with Dynal-MCP; then cells were washed four times in R10. T cells were then used as targets for killing by an A2-CMV-specific CTL clone. CTL were incubated with the Th cells at a 10:1 ratio for 0 or 6 h and then fixed in 2% formaldehyde-PBS followed by FACS acquisition.
Results
wt DR1-HA-GFP molecules can activate HA1.7 T cell clones
To investigate the transfer of MHC class II molecules to Th cells, MHC class II α and β-chain constructs were made, and the β-chain was tagged to GFP. Constructs containing HLA-DRA1*0101 and HLA-DRB1*0101 were generated (Fig. 1,A). Constructs were cotransfected into 293T cells using calcium phosphate precipitation, and expression was verified by FACS (Fig. 1,B). 293T cells expressing wt DR1-HA-GFP molecules are able to induce IFN-γ production by HA1.7 Th cells, confirming appropriate presentation of the class II molecules and HA peptide (Fig. 1,C). Cells with lone α- or β-chain transfection, or cotransfection of class II molecules that lack the HA peptide (DR1-GFP) or contain the irrelevant AchR α peptide (DR1-α-GFP), are unable to activate HA1.7 T cells (Fig. 1 C).
wt DR1-HA-GFP transfer to T cells during T cell activation
A FACS-based assay to follow the transfer of wt DR1-HA-GFP molecules to the restricted HA1.7 T cell was devised. Briefly, 293T cells expressing wt DR1-HA-GFP complexes act as APCs, APC to T cell contact was encouraged by brief and gentle centrifugation, cells were incubated to allow biting, and then cell-cell interactions were disrupted by incubation with EDTA. Cells were fixed and stained for CD3 and CD4, and FACS acquisition performed (Fig. 2). The T cell population lies in a distinct region to the 293T cells on forward light scatter vs side light scatter plots (Fig. 2,A). 293T cells do not express CD3 or CD4 but do express GFP (wt DR1-HA-GFP class II molecules; Fig. 2,A). T cells are CD3+ and CD4+ (Fig. 2,A), and this population does not express GFP molecules (Fig. 2,B, left). However, for example, when wt DR1-HA-GFP-expressing 293T cells and HA1.7 T cells are incubated and biting is allowed to proceed for 1 h, the T cells acquire the wt DR1-HA-GFP molecules (Fig. 2 B, right).
HA1.7 T cells are only able to acquire DR1 molecules that express the HA peptide, and if the β-chain is expressed alone without α-chain, there is very little transfer (Fig. 3,A). Approximately 80% of the HA1.7 T cells acquire wt DR1-HA-GFP molecules with an increase in MFI of 15–20 units (data not shown). During a 6-h time course, the HA1.7 T cell clones, but not the non-DR1-HA-restricted Th clone KB-D1, acquire wt DR1-HA-GFP molecules from 293T cells (Fig. 3,B). Incorporation is rapid, with 80% of the cells becoming GFP+ within 30 min of cell-cell contact. The KB-D1 cell line is similarly able to acquire HLA-DR52a molecules expressing the AchR α peptide, but not wt DR1-HA-GFP molecules (data not shown). wt DR1-HA-GFP molecules incorporated into HA1.7 T cells were visualized by immunofluorescence microscopy (Fig. 3,C). Biting proceeded for 1 h, the samples were incubated with 0.5 mM EDTA-PBS postbiting and stained for CD3 molecules as described (Fig. 3,C, left), and small clusters of GFP could be visualized in the HA1.7 Th cell. Cells not incubated with 0.5 mM EDTA-PBS can remain conjugated (Fig. 3 C, right), the wt DR1-HA-GFP molecules cluster at the synapse and class II molecules can be seen becoming incorporated into the T cell membrane. There was no detectable transfer of CD3 to the 293T cell, suggesting that membrane biting is unidirectional.
Bystander molecule transfer
The fate of other molecules in the APC membrane was next investigated. Microscope images (Fig. 3,C) of T cells acquiring molecules from the 293T cell suggested that MHC class II molecules are diffusing into the Th cell membrane. The precise mechanism was not determined; however, it is possible that the membranes have fused (28) or that large quantities of exosomes are being released by the 293T cell and specifically absorbed by the T cell (14). Either process indicates that other molecules could also be transferred to the T cell. First, the fate of lipids was examined. The red dye PKH-26 nonspecifically stains membrane lipid bilayers; this was used to label 293T cells expressing wt DR1-HA-GFP molecules. Biting was allowed to proceed for 1 h, and then wt DR1-HA-GFP and red dye incorporation into the HA1.7 T cells was investigated. Both the class II molecules and the PKH-26 dye transferred after 1 h of biting (Fig. 4,A). Bystander MHC class I molecules on the surface of the 293T cell are also transferred to HA1.7 T cells when wt DR1-HA-GFP molecules are acquired. 293T cells expressing A2 molecules linked to GFP with or without DR1-HA (lacking GFP tag) were used as presenting cells to HA1.7 T cells, only in the presence of DR1-HA molecules were the class I A2 complexes transferred (Fig. 4 B).
Acquired class I-peptide complexes render HA1.7 Th cells susceptible to CTL killing
To test whether Th cell acquisition of MHC class I-peptide molecules makes the cell a target for specific CTL killing, a FACS-based assay was developed (Fig. 5 A). A CTL clone that recognizes a CMV495–503 peptide presented by HLA-A2 was used in this assay. The CD4+ T cell clone HA1.7 acquires class II and class I molecules from 293T cells, making them a target for CTL, and an irrelevant CD4+ T cell clone (KB-D1) is introduced to control for nonspecific biting and killing. 293T cells naturally express high levels of A2 (data not shown); therefore, the CMV495–503 peptide can be efficiently presented. Pulsed peptides can be released from the 293T cells and taken up by other cells (producing false targets); however, FACS staining confirmed that neither of the CD4+ T cell clones expresses HLA-A2 (data not shown). Therefore, the CMV-CTL clone could only kill Th cells that have acquired the A2-CMV peptide complexes through bystander acquisition as a consequence of class II membrane biting.
293T cells stably expressing wt DR1-HA-GFP (Fig. 5,B) were pulsed with the A2-restricted CMV peptide, or left unpulsed as a control. To ensure that the killing is specifically due to bystander class I-peptide acquisition and presentation, a control T cell population was introduced which should not acquire these molecules, the DR52a-restricted CD4+ T cell clone KB-D1. To distinguish between the HA1.7 and KB-D1 populations, they were labeled with the red lipid dye PKH-26 at different levels of fluorescent intensity; HA1.7 cells were labeled with 10 μM PKH-26 and KB-D1 cells were labeled with 0.5 μM PKH-26, which gave a 1-log difference in intensity when detected by FACS (Fig. 5 C).
293T cells stably expressing DR1-HA-GFP ± CMV peptide were incubated with mixed labeled HA1.7 and KB-D1 for 1.5 h. Only the HA1.7 clone acquired DR1-HA-GFP molecules (Fig. 5,D), and the acquisition was identical with or without CMV495–503-pulsed peptide (data not shown). Th cells were then isolated using anti-CD4-coated Detachabeads (Dynal Biotech). This effectively removes the A2-CMV/DR1-HA-GFP+ 293T cells from the assay, which would otherwise compete for the killing activity of the CMV-specific CTL. Most 293T cells remained in the supernatant (data not shown). The resulting HA1.7 and KB-D1 cells were then incubated with A2-CMV-specific CTL, that were neither GFP+ nor PKH-26+ (Fig. 5,E) for 6 h. Before the biting or beading, neither HA1.7 nor irrelevant KB-D1 T cells expressed GFP (Fig. 6,A, top); in this example, CTL have been added to the reaction after fixing to demonstrate the location of the cells before biting occurs. Post-biting and beading, the HA1.7 Th cells clearly acquired wt DR1-HA-GFP, whereas the KB-D1 cells had not (Fig. 6,A, bottom). After incubation with the CMV-specific CTL for 6 h, there were less HA1.7 cells remaining that acquired wt DR1-HA-GFP and A2 molecules from 293T target cells that were preincubated with CMV peptide than those that acquired molecules from 293T cells lacking CMV peptide (Fig. 6,A, bottom). The percentage of HA1.7 T cells killed by the CTL was calculated with and without the addition of the CMV495–503 peptide on the original wt DR1-HA-GFP presenting 293T cell. Killing was rapid, with maximal killing occurring within 6 h of CTL addition, and death was >3 times more when A2-CMV peptide complexes were acquired as compared with A2 complexes lacking CMV peptide (Fig. 6 B). The bystander acquisition of class I-peptide complexes when Th cells bite class II-peptide-restricted complexes resulted in specific CTL-mediated cell death.
Discussion
Ag-specific Th cells are able to acquire specific restricted MHC class II-peptide complexes from APC. Consequently, the T cells acquire membrane lipids and MHC class I-peptide complexes from the same APC. Here, we show that the acquisition of the bystander MHC class I-peptide complexes by Th cells causes them to become targets for specific CTL killing. In vivo, the process could have immunoregulatory properties, leading to the down-regulation of the immune response preventing dangerous immunopathology.
The functional significance of membrane biting by T lymphocytes in vivo remains to be elucidated, and in vitro experiments yield conflicting results. CTLs acquire class I MHC-peptide complexes (16), rendering them susceptible to both fratricide (15) and increased CTL activation depending on the quantity of Ag acquired (29). Full effector responses, such as cytokine release and CTL proliferation, would also require the transfer of costimulatory molecules by trogocytosis. Several molecules are know to transfer; CD54 and CD80 (B7-1) are acquired by naive T cells, which can then act as APCs, and by memory T cells, which undergo apoptosis (18, 22, 23). Th cells can also acquire MHC class II molecules (19, 21), and acquired MHC class II molecules associate with the TCR, signaling molecules and the cytoskeleton, with the MHC correctly orientated on the T cell surface to allow Ag presentation (30, 31). However, the subsequent presentation of the acquired class II-peptides by the T cells is inefficient, due to poor expression of costimulatory molecules. T cells triggered by these molecules on T cells become anergic (21) or trigger increased apoptosis and hyporesponsiveness (17).
In the system reported here, the 293T fibroblast cell line does not express classical coreceptor molecules; however, neither the memory CTL response nor CTL killing require coreceptor expression for function, and memory CTL lose their CD28 expression (32). The MHC class I and II molecules are transferred to the Th cells and become susceptible to specific CTL killing. The passive coacquisition of MHC class I molecules was less than the active class II transfer (Fig. 4 B), but even the transfer of a small number of molecules would be sufficient for CTL recognition and killing (33).
Recent reports have suggested that APC membrane acquisition by T cells may occur only when high levels of Ag are expressed, such as when high levels of virus are detected in acute viral infection. The concomitant acquisition of MHC class I-peptides when MHC class II complexes transfer could prevent exponential T cell growth, and an overreactive Th cell response that could lead to dangerous immunopathology. In essence, the CD4+ T cell becomes a target for killing by the cells it is helping activate.
Taken together, our study provides another mechanism of how Ag-specific CD4+ T cells can be regulated and may have a potential implication for the dangerous immunopathoglogy during the course of viral infection such as HIV.
Acknowledgment
We thank Dr Simon Brackenridge for critical reading of the manuscript.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
Abbreviations used in this paper: HA, hemagglutinin; AchR, acetylcholine receptor; wt, wild type.