The frequency of circulating alloreactive human memory T cells correlates with allograft rejection. Memory T cells may be divided into effector memory (TEM) and central memory (TCM) cell subsets, but their specific roles in allograft rejection are unknown. We report that CD4+ TEM (CD45RO+CCR7CD62L) can be adoptively transferred readily into C.B-17 SCID/bg mice and mediate the destruction of human endothelial cells (EC) in vascularized human skin grafts allogeneic to the T cell donor. In contrast, CD4+ TCM (CD45RO+CCR7+CD62L+) are inefficiently transferred and do not mediate EC injury. In vitro, CD4+ TEM secrete more IFN-γ within 48 h in response to allogeneic ECs than do TCM. In contrast, TEM and TCM secrete comparable amounts of IFN-γ in response to allogeneic monocytes (Mo). In the same cultures, both TEM and TCM produce IL-2 and proliferate in response to IFN-γ-treated allogeneic human EC or Mo, but TCM respond more vigorously in both assays. Blockade of LFA-3 strongly inhibits both IL-2 and IFN-γ secretion by CD4+ TEM cultured with allogeneic EC but only minimally inhibits responses to allogeneic Mo. Blockade of CD80 and CD86 strongly inhibits IL-2 but not IFN-γ production by in response to allogeneic EC or Mo. Transduction of EC to express B7-2 enhances allogeneic TEM production of IL-2 but not IFN-γ. We conclude that human CD4+ TEM directly recognize and respond to allogeneic EC in vitro by secreting IFN-γ and that this response depends on CD2 but not CD28. Consistent with EC activation of effector functions, human CD4+ TEM can mediate allogeneic EC injury in vivo.

A cardinal feature of the adaptive immune response is memory, i.e., a second immune response to a given Ag that has been encountered previously is faster and more vigorous compared with the initial (primary) immune response. When compared with a naive response, a memory response is characterized by both a higher frequency of Ag-specific T cells and by a series of changes in the individual T cells that respond to that Ag (1). The latter changes include an altered requirement for costimulators and a difference in the expression of receptors involved in homing. Memory T cells also have a special relationship to vascular endothelial cells (EC)2 in that resting memory T cells can be activated by alloantigens presented by EC but not by other stromal or parenchymal cells in vitro (2).

Memory T cells may be divided into at least two subsets based on their homing characteristics and effector functions (3). Central memory T cells (TCM) express chemokine receptors (e.g., CCR7) that bind lymph node-derived chemokines, such as secondary lymphoid-tissue chemokine (SLC), and adhesion molecules (e.g., CD62L) that preferentially interact with ligands expressed by high endothelial venules of lymph nodes (3). Consequently, TCM recirculate preferentially to the T cell areas of lymphoid organs where they conduct surveillance for their specific Ag displayed by professional APCs such as myeloid dendritic cells (4). Here, resting TCM are activated to differentiate into effector cells that only then can migrate to peripheral tissues, a process that may require several days. In contrast, effector memory T cells (TEM) preferentially migrate to inflamed peripheral tissues where they can rapidly express effector molecules such as IFN-γ (5). This migration pattern of TEM is mediated by the expression of chemokine receptors (e.g., CXCR3) that respond to inflammatory chemokines such as CXCL9, 10, and 11 and by the expression of adhesion molecules (e.g., LFA-1 or VLA-4) whose ligands (ICAM-1 and VCAM-1, respectively) are preferentially expressed by cytokine-activated peripheral vascular EC (6).

The memory response has been reported to be less dependent on costimulation from CD28, which binds B7-1 and B-2 (also known as CD80 and CD86, respectively), than is a primary immune response (7). An alternative source of costimulation in humans involves CD2. Although, CD2 can bind to CD48 in rodents, it is preferentially engaged in humans by LFA-3 (CD58), a molecule missing from the rodent genome (8). Human EC express LFA-3 but generally not B7-1 or B-2, and the LFA-3-CD2 pathway seems to be particularly important in providing costimulation during human allogeneic responses to this cell type (9). However, Ab-blocking experiments have suggested that CD28 may be involved in the response to EC, the source of the ligand being provided by the T cells themselves (10).

It has become increasingly evident that T cell memory may pose a significant challenge to clinical transplantation. Memory T cells specific for and expanded by microbial Ag, like naive T cells, may cross-react with allogeneic MHC-peptide complexes with very high frequency (11). Recent evidence has suggested that the number of cross-reactive antidonor memory T cells correlates with increasing rejection rates and may prevent the ability to achieve tolerance (12). These phenomena had largely been overlooked in the past because alloreactive memory T cell populations are not typically present in the young rodents commonly used to study transplantation. However, the adoptive transfer of memory T cells in mice can rapidly reject a cardiac graft in the absence of secondary lymphoid organs, whereas the transfer of naive T cells cannot (13). The subset of memory T cells responsible for this phenomenon is not clear. In some mouse models CD8+ TCM are principally responsible for memory cell-mediated rejection (11), whereas other models support a role for CD4+ TEM cells (14). Our previous report (15) using adoptive transfer into immunodeficient mice demonstrated that unfractionated human memory T cells can reject skin allografts, but isolated memory T cell subsets were not examined.

We report here that in CB.17 SCID/bg mice bearing a vascularized human skin graft, adoptively transferred human CD4+ TEM can damage graft EC. We also show that both CD4+ TCM and TEM subsets from human peripheral blood can produce IL-2 and proliferate in response to allogeneic EC in vitro, but TEM more readily produce IFN-γ. In contrast, allogeneic monocytes (Mo) stimulate comparable levels of IFN-γ for both subsets. Furthermore, IFN-γ produced by EC-activated TEM requires CD2 but not CD28 signals. These observations may be important for the design of new therapies to address the problem of immunological memory in transplant rejection.

All human cells and tissues were obtained under protocols approved by the Yale Human Investigations Committee (New Haven, CT). PBMCs were isolated by density gradient centrifugation of leukapheresis products from healthy adult volunteers by using a lymphocyte separation medium (Invitrogen Life Technologies). Isolated cells were stored in 10% DMSO and 90% FBS at −196°C and were thawed and washed before use (16). CD4+ T cells were isolated from PBMC by positive selection using BD IMag CD4+ magnetic particles (BD Biosciences). The selected population obtained by this procedure was routinely >95% CD4+ by flow cytometry (data not shown). The isolated CD4+ T cells were then stained with mouse Abs to mouse anti-human CD62L (BD Biosciences), mouse anti-human CD45RO (BD Biosciences), and mouse anti-human CCR7 (R&D Systems) and subjected to FACS sorting on a FACSAria (BD Biosciences) to isolate CD4+ TCM (CD45RO+CCR7+CD62L+) from CD4+ TEM (CD45RO+CCR7CD62L).

Supernatants were collected from cocultures of 1 × 106 T subset cells with 5 × 104 allogeneic HUVEC or 5 × 104 allogeneic Mo after 24 h of coculture. The samples were then assessed using an ELISA kit for IFN-γ (BioSource International) or IL-2 (eBioscience). Both ELISAs were performed as described by the respective manufacturers. Where indicated, inhibitory anti-human LFA-3 Ab or CTLA-4Ig were incorporated into the assay at the indicated final concentrations.

T cell proliferation was measured by [3H]thymidine incorporation. In brief, HUVEC were cultured in 96-well U-bottom plates (Falcon) as described previously (17) and confluent monolayers were then pretreated with 50 ng of IFN-γ (BioSource International) before coculture with CD4+ T cells. All HUVEC were then treated with mitomycin C (50 μg/ml in PBS for 30 min; Sigma-Aldrich) before coculture to prevent their proliferation as described previously (18). Twenty-four hours before each indicated time point, 1 μCi of [3H]thymidine (Amersham Biosciences) was added to each well. Plates were frozen, thawed, and then harvested on a 96-well harvester (Tomtec) and counted on a MicroBeta scintillation counter (Wallac). The mean of 16 replicates was calculated and the mean [3H]thymidine incorporation into wells containing only EC was subtracted.

C.B-17 SCID/beige female mice (Taconic Farms) were used at 5–8 wk of age. All protocols involving animals were approved by the Yale Animal Care and Use Committee (New Haven, CT). The animals were housed individually in microisolator cages and were fed autoclaved food and water. Before each experiment, serum mouse IgG levels were determined by sandwich ELISA using reagents from Cappel as previously described (19). SCID/beige animals were considered “leaky” at IgG levels of > 1 μg/ml and were excluded from experimental use.

Human skin was obtained from cadaveric donors through the Yale University Skin Bank (New Haven, CT) under a protocol approved by the Yale Human Investigations Committee. Human skin was orthotopically transplanted to SCID/beige mice as previously described (19). In brief, 0.5-mm-thick sheets were divided into 1-cm2 pieces, kept at 4°C in RPMI 1640 medium (Invitrogen Life Technologies), and fixed onto similarly sized defects on the dorsum of C.B-17 SCID/beige recipients using staples (3M). The resultant surface area of healed grafts was kept constant between animals when possible. The skin reproducibly grafted with a >95% success rate and was allowed to heal for 4 wk before the graft was manipulated. Rare animals that did not successfully engraft were excluded from the experimental groups before treatments.

CD4+CD45RO+CD62LCCR7 (TEM) or CD4+CD45RO+CD62L+CCR7+ (TCM) T cells were isolated as described above and the resulting cells were >95% pure upon reanalysis. PMBC-derived TCM and TEM cells (1 × 106) were adoptively transferred by tail vein injection into SCID/beige mice bearing human skin from a donor allograft 4 wk after skin engraftment and 2 wk following splenectomy. None of the animals demonstrated signs of graft-vs-host disease.

Circulating human T cells were evaluated by flow cytometry as previously described (16). In brief, heparinized retro-orbital venous samples were obtained 14 days after the adoptive transfer of human T cells, and the erythrocytes were lysed. The leukocytes were incubated with FITC-conjugated mouse anti-human CD3 (Immunotech) and PE-conjugated rat anti-mouse CD45 (Sigma-Aldrich) mAbs. Samples were then analyzed using a FACSort cytometer(BD Biosciences) for the number of human CD3+ cell normalized to the number of mouse CD45+ cells expressed as a percentage.

Human skin grafts, harvested at the indicated times, were processed for paraffin-embedded or frozen sections as previously described (16). The degree of graft microvascular damage was evaluated from H&E-stained sections by a dermatopathologist (J.M.M.) blinded to treatment protocols as previously described (20). In brief, the percentage of dermal vessels showing injury, defined as EC loss or sloughing, and thrombosis were assessed from an average of three high-power (×200) fields using the following semiquantitative grading scale: grade 0, all vessels patent and uninvolved; grade 1, <25% of vessels show injury; grade 2,∼50% of vessels show injury; and grade 3, >75% of vessels show injury.

Statistical differences between groups with respect to T cell infiltrates were evaluated using a two-tailed t test. Statistical differences between the pathology scores were evaluated by using a nonparametric analysis Mann-Whitney U test. [3H]Thymidine incorporation was analyzed using a two-tailed t test.

Cryopreserved skin samples, generated as described above or stored from previously described experiments in which animals were inoculated with unfractionated memory T cells (15), were analyzed by real-time quantitative RT-PCR for transcripts encoding CD3ε, CCR7, and CD62L as previously described (15). Primers used for this analysis are in Table I. These same markers were also evaluated in purified TCM and TEM.

Table I.

PCR primers used in this studya

TargetPrimers (5′→3′)
CCR7 TGGCTCTCCTTGTCATTTTC 
 GGAGCACAAAGACTCGAACA 
  
CD3ε GGCAAAGGGGACAAAACAAG 
 CTTTCCGGATGGGCTCATAG 
  
CD62L ACTAAAGGCAGCCCTCTGTT 
 GTACCCCACATCACAGTTGC 
TargetPrimers (5′→3′)
CCR7 TGGCTCTCCTTGTCATTTTC 
 GGAGCACAAAGACTCGAACA 
  
CD3ε GGCAAAGGGGACAAAACAAG 
 CTTTCCGGATGGGCTCATAG 
  
CD62L ACTAAAGGCAGCCCTCTGTT 
 GTACCCCACATCACAGTTGC 
a

Each primer is listed 5′→3′. The forward primer is above and the reverse primer is below. In each case, the identity of the amplified fragment was confirmed by sequencing.

Sallusto and colleagues (3) were the first to observe that the circulating human memory T cell pool may be divided into TCM and TEM subsets. In their schema, TCM, which are CCR7+, preferentially migrate into secondary lymphoid organs and must be activated to acquire effector functions such as IFN-γ production or cytolytic activity. However, TEM, which lack expression of CCR7 and respond to CXCR3-binding chemokines (6), can rapidly express effector function and have a migratory pattern that favors extralymphoid sites of inflammation. In addition to chemokine receptor differences, recirculation patterns of T cells are also influenced by adhesion molecules. CD62L is particularly important for the entry of TCM into lymph nodes. The CCR7+ and CCR7 memory T cell subsets contain a mixture of CD62L+ and CD62L cells (3). Some investigators have separated memory T cells into TCM and TEM based solely upon CD62L expression (21). We sorted CD4+CD45RO+ T cells into CCR7+CD62L+ and CCR7CD62L subsets because both CCR7 and CD62L are required for TCM recirculation and neither are needed for recruitment to peripheral inflammatory sites (Fig. 1) (22).

FIGURE 1.

FACSort of human CD4+ memory into TCM and TEM subsets. CD4+ T cells were isolated using magnetic beads before being stained with Abs to CD45RO, CD62L, and CCR7. Cells were then sorted into TCM and TEM by collecting only the gated regions. Left hand panels are presort and right hand panels show analysis of sorted populations.

FIGURE 1.

FACSort of human CD4+ memory into TCM and TEM subsets. CD4+ T cells were isolated using magnetic beads before being stained with Abs to CD45RO, CD62L, and CCR7. Cells were then sorted into TCM and TEM by collecting only the gated regions. Left hand panels are presort and right hand panels show analysis of sorted populations.

Close modal

To address the question of which human memory CD4+ T cell subset can mediate allograft rejection, we adoptively transferred TEM or TCM into C.B-17 SCID/bg hosts bearing allogeneic skin grafts. FACS analysis of mouse blood at 14 days postinoculation of T cells revealed that TEM reached between 1 and 3% of the circulating mononuclear cells, whereas TCM were not detectable in the circulation. Skin grafts were harvested and analyzed at 3 wk postinoculation of T cells. We found in three independent experiments that CD4+ TEM (Fig. 2, C and D), but not CD4+ TCM (Fig. 2, A and B), reproducibly infiltrated grafts and destroyed human EC-lined microvessels. When scored by a dermatopathologist blinded to the experimental conditions, mice receiving adoptively transferred CD4+ TEM showed significantly more infiltration and endothelial damage (p < 10−4) compared with mice receiving CD4+ TCM (Fig. 2,E). These data establish that human TEM alone are capable of mediating allograft rejection. To further examine the question of whether TCM may have a role in the memory T cell response in our adoptive transfer models, we also examined mRNA levels of CCR7 and CD62L (normalized to CD3ε) in grafts that have been infiltrated either by whole memory T cells or by TEM alone. These transcripts are, as expected, much more abundant in isolated CD4+ TCM than in isolated CD4+ TEM (Fig. 2,F). We find very low levels of CCR7 and CD62L transcripts in grafts that are rejected by TEM, and these levels are equally low in grafts rejected by whole memory T cells (Fig. 2 G). The simplest interpretation of these data is that CD4+ TCM do not contribute to the rejection responses observed in our model. However, due to the poor efficiency of adoptive transfer of TCM in our model, this does not rule out a role for TCM in other settings, e.g., when professional APCs or human lymphoid organs are present.

FIGURE 2.

Effector memory T cells alone can mediate graft injury. A–D, Grafts from mice that received either TCM (magnification: ×20 for A and ×40 for C) or TEM (magnification: ×20 for B and ×40 for D) for 14 days were harvested and stained with H&E (n = 12 per group). Intact vessels are indicated with black arrows (A and B) and damaged vessels are indicated with white arrows (C and D). E, Grafts were then scored blindly by a dermatopathologist (∗∗, p < 10−4). Data were pooled from three independent experiments. F and G, Isolated TCM (CM) or TEM (EM) were analyzed for transcripts encoding CCR7 or CD62L, markers of TCM, normalized to the pan T cell marker CD3ε (F). The same markers were then used to probe skin grafts that had been infiltrated by human T cell in animals inoculated with unfractionated T memory cells or purified TEM (G). Note that there is no evidence of increased markers for TCM in the infiltrating population even when TCM were present in the inoculum. Transcript data represents pooled data from three independent experiments.

FIGURE 2.

Effector memory T cells alone can mediate graft injury. A–D, Grafts from mice that received either TCM (magnification: ×20 for A and ×40 for C) or TEM (magnification: ×20 for B and ×40 for D) for 14 days were harvested and stained with H&E (n = 12 per group). Intact vessels are indicated with black arrows (A and B) and damaged vessels are indicated with white arrows (C and D). E, Grafts were then scored blindly by a dermatopathologist (∗∗, p < 10−4). Data were pooled from three independent experiments. F and G, Isolated TCM (CM) or TEM (EM) were analyzed for transcripts encoding CCR7 or CD62L, markers of TCM, normalized to the pan T cell marker CD3ε (F). The same markers were then used to probe skin grafts that had been infiltrated by human T cell in animals inoculated with unfractionated T memory cells or purified TEM (G). Note that there is no evidence of increased markers for TCM in the infiltrating population even when TCM were present in the inoculum. Transcript data represents pooled data from three independent experiments.

Close modal

Our in vivo results are consistent with the hypothesis that TEM effector functions are directly activated by allogeneic EC within the graft. To test this idea, we compared the abilities of allogeneic EC and Mo to activate isolated CD4+ TEM and TCM subsets in vitro. CD4+ TCM produced more IL-2 (Fig. 3,B) and proliferated more vigorously than did CD4+ TEM when cultured with either allogeneic EC or Mo (Fig. 3, C and D), but both subsets responded to both stimulator cell types. Mo were stronger than EC at stimulating these responses. TEM made more IFN-γ than did TCM when cultured with allogeneic EC. In contrast, TCM produced as much or more IFN-γ than did TEM when cultured with allogeneic Mo (Fig. 3 A). These data suggest that either allogeneic EC provide a signal that can enhance IFN-γ production by TEM or that EC lack a signal, provided by Mo, that is necessary for stimulating optimal IFN-γ production by TCM.

FIGURE 3.

TEM exhibit greater IFN-γ production and less IL-2 production than TCM in response to EC but not to Mo. A and B, TCM (CM) or TEM (EM) were cultured with either allogeneic EC or Mo and supernatant was collected at 24 h. Supernatants were then assessed for IFN-γ (A) or IL-2 (B) using ELISA. C and D, T cell subsets were also assessed for proliferation at day 5, 6, and 7 using [3H]thymidine incorporation when cultured with EC (C) or Mo (D). One of five experiments with similar results is shown.

FIGURE 3.

TEM exhibit greater IFN-γ production and less IL-2 production than TCM in response to EC but not to Mo. A and B, TCM (CM) or TEM (EM) were cultured with either allogeneic EC or Mo and supernatant was collected at 24 h. Supernatants were then assessed for IFN-γ (A) or IL-2 (B) using ELISA. C and D, T cell subsets were also assessed for proliferation at day 5, 6, and 7 using [3H]thymidine incorporation when cultured with EC (C) or Mo (D). One of five experiments with similar results is shown.

Close modal

We next examined whether differences in the varying responses of TCM and TEM to EC vs Mo could be attributed to differences in the use of costimulatory pathways. We found that CD4+ TCM express significantly more CD27 and somewhat more CD28 compared with CD4+ TEM (Fig. 4,A). Expression of CD27 on TEM was heterogeneous. Specifically, we observed a small population displaying an intermediate level of CD27 expression and a much larger population of TEM expressing only low levels of CD27. At the same time, we found similar levels of expression of CD2 on TCM and TEM (Fig. 4,A). We also compared the levels of the ligands for these costimulators on EC vs Mo. As shown in Fig. 4 B, HUVEC do not express CD70 or B7 molecules at rest nor do they increase CD70 or B7 molecules with IFN-γ treatment or in coculture with CD4+ T cells. In contrast, adherent Mo express low levels of both CD70 and B7 molecules. Both EC and Mo express LFA-3, the principal ligands for CD2 with Mo expressing significantly more LFA-3 than resting EC or EC either treated with IFN-γ or cocultured with CD4+ T cells.

FIGURE 4.

Expression of costimulators and receptors by T cells and APCs. A, CD4+CD45RO+ T cells were isolated from PBMCs using magnetic beads and subsequently stained with CCR7 and CD62L. The TCM CCR7+CD62L+ (gray-filled histogram) population was gated on and compared with the TEM CCR7CD62L (black line) population. The corresponding mean fluorescence intensities for TCM (gray text) and TEM (black text) are at the top of the histogram. B, HUVEC or negatively isolated Mo were treated with IFN-γ for 3 days, cocultured with CD4+ T cells, or both and assessed for CD70 expression or binding to CTLA-4Ig by FACS analysis. The black line and number represent the percentage of total cells in that region. One of three experiments with similar results represented.

FIGURE 4.

Expression of costimulators and receptors by T cells and APCs. A, CD4+CD45RO+ T cells were isolated from PBMCs using magnetic beads and subsequently stained with CCR7 and CD62L. The TCM CCR7+CD62L+ (gray-filled histogram) population was gated on and compared with the TEM CCR7CD62L (black line) population. The corresponding mean fluorescence intensities for TCM (gray text) and TEM (black text) are at the top of the histogram. B, HUVEC or negatively isolated Mo were treated with IFN-γ for 3 days, cocultured with CD4+ T cells, or both and assessed for CD70 expression or binding to CTLA-4Ig by FACS analysis. The black line and number represent the percentage of total cells in that region. One of three experiments with similar results represented.

Close modal

To explore whether the differential expression of these costimulatory molecules on the two different subsets is responsible for the difference in their response to EC, we used CTLA4-Ig to block B7 interactions and a blocking Ab to LFA-3 to study the role of CD2. (We did not have access to blocking reagents for CD70.) CTLA-4Ig binds B7-1 and B7-2 with higher affinity than does CD28 and inhibits the binding of native CD28 or CTLA-4 to either of the B7 molecules (23). The addition of CTLA-4Ig has only a minimal effect on IFN-γ production by TCM or TEM with EC (Fig. 5,A, upper panels), but there is a significant reduction in IFN-γ produced when they are cocultured with Mo (Fig. 5,A, lower panels). CTLA-4Ig also reduced IL-2 production by CD4+ TEM and TCM activated by allogeneic Mo (Fig. 5,B, lower panels) or by allogeneic EC (Fig. 5,B, upper panels). Although EC do not express B7 molecules, this finding is consistent with a previous study (10) using whole CD4+ T cell populations in which T cells themselves proved to be the source of B7 ligands. Overall, the effect of CTLA-4Ig is more pronounced on IL-2 production than on IFN-γ production and is more pronounced in cocultures with Mo than in those with EC. The addition of anti-LFA-3 decreased the production of IL-2 by CD4+ TEM and TCM in cocultures with EC (Fig. 5,A, lower panels) significantly, but only modestly with Mo (Fig. 5,B, lower panels). The same Ab inhibited IFN-γ production by CD4+ TEM in cocultures of EC but only weakly affected IFN-γ production in TCM and TEM stimulated by Mo (Fig. 5 A, upper panels). In other words, Mo appear to use B7 ligands for costimulation of both IL-2 and IFN-γ by CD4+ memory T cells whereas EC are much more highly dependent on LFA-3 for costimulation of IFN-γ, especially from CD4+ TEM.

FIGURE 5.

Roles of LFA-3 and B7 in TEM and TCM response to allogeneic EC and Mo. A and B, TCM (CM) and TEM (EM) were cocultured with EC (A) or Mo (B) and blocking Ab to LFA-3 (□) or CTLA-4Ig (▪). ELISAs were performed for IFN-γ (upper panels) or IL-2 at 48h (lower panels). C and D, TCM (CM) and TEM (EM) were also cultured with either B7-2-transduced or EGFP-transduced EC and supernatants were assessed for IFN-γ (C) and IL-2 (D) at 48 h. One of four experiments with similar results is represented.

FIGURE 5.

Roles of LFA-3 and B7 in TEM and TCM response to allogeneic EC and Mo. A and B, TCM (CM) and TEM (EM) were cocultured with EC (A) or Mo (B) and blocking Ab to LFA-3 (□) or CTLA-4Ig (▪). ELISAs were performed for IFN-γ (upper panels) or IL-2 at 48h (lower panels). C and D, TCM (CM) and TEM (EM) were also cultured with either B7-2-transduced or EGFP-transduced EC and supernatants were assessed for IFN-γ (C) and IL-2 (D) at 48 h. One of four experiments with similar results is represented.

Close modal

To test whether the lack of B7-mediated costimulation accounts for the more limited IFN-γ production by TCM compared with TEM cultured with EC, we generated EC that express B7-2 using retroviral transduction (data not shown). When cocultured with these B7-2-expressing EC, TCM produce greater amounts of both IFN-γ (Fig. 5,C) and IL-2 (Fig. 5,D). In contrast, TEM do not produce any additional IFN-γ (Fig. 5,C), although they do produce significantly more IL-2 (Fig. 5 D), similar to the response observed when TEM are cultured with Mo. These data indicate that the absence of B7 ligands on EC may contribute to the lesser capacity of EC to activate TCM to secrete IL-2 and proliferate. However, the lack of the ability of B7.2 transduction to boost IFN-γ production by TEM suggests that Mo must provide alternative signals (other than B7 ligands or LFA-3) to effectively costimulate IFN-γ production. This could be in the form of a secreted cytokine such as IL-12 (24). The strong dependence of the TEM response to EC upon LFA-3 may explain the efficacy of the LFA-3 blockade previously demonstrated in our mouse human skin graft models (20).

In this report we present three new observations about memory T cell subsets and their relationship to EC. First we demonstrate that, following adoptive transfer, CD4+ TEM can mediate graft injury in a human SCID chimera model of skin allograft rejection where the major source of Ag is graft EC. Second, we show that human TCM and TEM subsets, defined by the expression of CCR7 and CD62L, are differentially stimulated by allogeneic EC. Specifically, CD4+ TEM produce more IFN-γ and less IL-2 compared with TCM when cocultured with allogeneic EC. The difference in IFN-γ production is not seen when the same subsets are cocultured with allogeneic Mo. Third, we show that the activation of TEM effector functions is largely dependent on costimulation via LFA-3 and not B7 signals. The capacity of EC to efficiently activate the TEM but not the TCM effector function is consistent with the in vivo observation that TEM alone can mediate allograft rejection.

The current conception of T cell memory is that the memory T cell pool may be divided into TCM and TEM subsets (3). TCM, which are CCR7+, preferentially migrate into secondary lymphoid organs and must be activated to acquire effector functions such as IFN-γ production or cytolytic activity. However, TEM, which lack expression of CCR7 and respond to CXCR3-binding chemokines (6), can immediately express effector function and have a migratory pattern that favors extralymphoid sites of inflammation. This distinction was first reported by Sallusto and colleagues (3). In their experiments, they used CCR7 alone to separate memory T cells into their different subsets and they report, as we have also found, that CCR7+ and CCR7 memory cells contain a mixture of CD62L+ and CD62L cells (3). To avoid this heterogeneity, we chose to isolate CD4+CD45RO+ memory cells into CCR7+CD62L+ double positive and CCR7CD62L double negative cells. We made this choice based on evidence that both CCR7 and CD62L are required for TCM recirculation and that neither is needed for recruitment to peripheral inflammatory sites (22, 25, 26, 27). The populations that are CCR7+ but CD62L or CCR7 but CD62L+ were not analyzed (28, 29). It is not clear what these intermediate populations may have contributed to the various responses reported by Sallusto et al. (3).

It has recently become evident that the presence of memory T cells presents a significant challenge to long-term graft survival. The presence of alloreactive CD4+ memory T cells correlates with increased episodes of acute rejection and diminished graft function (30, 31). However, the role of memory T cells in allograft rejection has not been well studied due to the lack of appropriate in vivo models. Mouse models have been described that use polyclonal alloreactive memory T cells (14) or those that have been generated from viral infection (11), but these still lack the diversity of the human memory T cell pool. Our study using human memory T cells fractionated into their different subsets shows a role for CD4+ TEM in allograft rejection, but not for TCM. One key difference between our human-mouse chimera and the current mouse models is that human T cells cannot effectively circulate to mouse secondary lymphoid organs, providing an isolated model of peripheral rejection only. This lack of the ability to circulate in the secondary lymphoid organs may explain why, in contrast to experiments with mice, TCM appear unable to mediate graft rejection. TCM may require signals found in the secondary lymphoid organs whereas TEM do not (32). This difference in the activation potential of EC vs Mo to activate TCM in vitro is abolished with sufficient B7 signals (i.e., from monocytes or transduced EC), which supports the idea that TCM may need to circulate through the secondary lymphoid organs and encounter B7-bearing APC to attain full functional capacity. Another consideration is that EC injury in our human skin graft model inoculated with CD4+ TEM occurred in the absence of CD8+ T cells and macrophages. This finding is consistent with a previous report from our laboratory showing that unfractionated CD4+ T cells can also rejection skin grafts in this model. Interestingly, in the absence of CD8+ T cells, CD4+ T cells acquired perforin expression (33). Rejection by alloreactive CD4+ TEM may explain why it is much more difficult to suppress rejection in humans, where these cells are typically found in the circulation, than in mice, where they are not. The behavior of TEM may also contribute to the difficulty in tolerizing humans to allografts. This concern is suggested by the observation that mouse CD8+ TEM are less inhibitable than TCM by tryptophan catabolism (34), a mechanism of peripheral tolerance. The role of such mechanisms in our model is not known.

In vivo, studies using mice have demonstrated that a large number of CD4+ memory T cells can be found in extralymphoid sites and that there is a functional difference between memory T cells residing in the lymphoid tissue compared with those of the extralymphoid tissues (5, 35). However, human studies have not been able to confirm the relationship between immediate effector function and the potential to migrate to extralymphoid sites (as defined by the lack of expression of CCR7) (4, 36). In fact, these studies suggest that the majority of cytokine-producing memory T cells express CCR7 and that immediate effector function did not correlate with CCR7 or CD62L expression (37, 38, 39). Our data show that human memory T cell subsets do show such differences when responding to allogeneic EC but not to allogeneic Mo. Because ECs are likely to play a role in presenting Ag (or alloantigen) to circulating TEM, the ability of human EC to elicit efficient effector function is likely to be of physiological significance. Sallusto et al. (3 initially demonstrated that dendritic cells pulsed with the bacterial superantigen toxic shock syndrome toxin 1 (TSST-1) stimulate CD45RACCR7 cells to produce IFN-γ and other effector cytokines, whereas CD45RACCR7+ T cells produce only IL-2. Our results differ in that we find that CD4+CD45RO+CCR7+CD62L+ TCM can produce significant amounts of IFN-γ in addition to IL-2 when stimulated by allogeneic Mo but not by allogeneic EC. However, our results agree that CD4+CD45RO+CCR7CD62L TEM are inefficient producers of IL-2. One reason for the observed difference in IFN-γ production may be that we examined our cocultures for cytokine production at 48 h, which is a later time point than that used by Sallusto et al (3). Thus, IFN-γ production by the TCM in our cocultures with Mo may reflect the early differentiation of TCM into effector cells. Because these cells produce increased levels of IL-2 and proliferate, TCM in cocultures with EC may also be undergoing some degree of differentiation because they also produce IFN-γ, although significantly less than in cocultures with Mo.

Memory T cells have different costimulatory requirements for activation when compared with their naive counterparts. Early work in mice (40) and humans (41) demonstrated that memory T cells can be activated independently of CD28, and more recently it has been appreciated that there may be a role for memory-specific costimulators in memory T cell activation (42, 43). In accordance with other studies that have shown decreased CD27 and CD28 in the TEM population (44, 45), we have also found decreases in CD27 and CD28 in the TEM populations used in our experiments. We have not explored this further because EC lack CD70, the only known ligand for CD27. We also find slightly different expression levels of CD28 and CD2 on TEM. To explore whether the differential expression of these costimulatory molecules on the two different subsets is responsible for the difference in their responses to EC, we used B7.2 transductants as well as CTLA4-Ig to block the B7 interaction and a blocking Ab to LFA-3 to study the role of CD2. Stimulation by EC transduced with B7.2 increases IL-2 by TCM and TEM, making EC behave more like Mo, and also increases IFN-γ production by TCM but not TEM. As expected, CTLA-4Ig had little effect on the CD4+ TEM response to EC, which lack B7 molecules. Unexpectedly, CTLA-4Ig did inhibit IL-2 synthesis by CD4+ TCM stimulated by allogeneic EC. As noted earlier, this finding is consistent with one previous report (10) using unfractionated CD4+ T cells in which the T cells themselves appeared to be the source of B7 ligands. Inhibition of human CD4+ TCM and TEM IL-2 production by B7 blockade is consistent with data showing that IL-2 but not IFN-γ production by mouse CD4+ TCM and TEM depends in part on CD28 (21). These data support that idea that CD28/B7 interactions may play an important role in memory T cell activation. Whereas a blockade of B7 signals differentially affects the response to EC vs Mo and IL-2 vs IFN-γ, a blockade of LFA-3 is broadly inhibitory of costimulation by EC, consistent with previous reports (24, 46).

In conclusion, our new results support the hypothesis that alloreactive human CD4+ TEM can participate in allograft rejection by direct recognition and injury of graft EC. Furthermore, a blockade of B7 signals is not likely to be as effective as a blockade of LFA-3 for the amelioration of this effect.

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.

2

Abbreviations used in this paper: EC, endothelial cell; TCM, central memory T cell; TEM, effector memory T cell; Mo, monocyte.

1
Sprent, J., C. D. Surh.
2002
. T cell memory.
Annu. Rev. Immunol.
20
:
551
-579.
2
Murray, A. G., P. Libby, J. S. Pober.
1995
. Human vascular smooth muscle cells poorly costimulate and actively inhibit allogeneic CD4+ T cell proliferation in vitro.
J. Immunol.
154
:
151
-161.
3
Sallusto, F., D. Lenig, R. Forster, M. Lipp, A. Lanzavecchia.
1999
. Two subsets of memory T lymphocytes with distinct homing potentials and effector functions.
Nature
401
:
708
-712.
4
Campbell, J. J., K. E. Murphy, E. J. Kunkel, C. E. Brightling, D. Soler, Z. Shen, J. Boisvert, H. B. Greenberg, M. A. Vierra, S. B. Goodman, et al
2001
. CCR7 expression and memory T cell diversity in humans.
J. Immunol.
166
:
877
-884.
5
Masopust, D., V. Vezys, A. L. Marzo, L. Lefrancois.
2001
. Preferential localization of effector memory cells in nonlymphoid tissue.
Science
291
:
2413
-2417.
6
Manes, T. D., J. S. Pober, M. S. Kluger.
2006
. Endothelial cell-T lymphocyte interactions, IP-10 stimulates rapid transendothelial migration of human effort but not central memory CD4+ T cells: requirements for shear stress and adhesion molecules.
Transplantation
82
:
S9
-S14.
7
Fontenot, A. P., L. Gharavi, S. R. Bennett, S. J. Canavera, L. S. Newman, B. L. Kotzin.
2003
. CD28 costimulation independence of target organ versus circulating memory Ag-specific CD4+ T cells.
J. Clin. Invest.
112
:
776
-784.
8
Davis, S. J., P. A. van der Merwe.
1996
. The structure and ligand interactions of CD2: implications for T cell function.
Immunol. Today
17
:
177
-187.
9
Hughes, C. C., C. O. Savage, J. S. Pober.
1990
. Endothelial cells augment T cell interleukin 2 production by a contact-dependent mechanism involving CD2/LFA-3 interaction.
J. Exp. Med.
171
:
1453
-1467.
10
Denton, M. D., C. S. Geehan, S. I. Alexander, M. H. Sayegh, D. M. Briscoe.
1999
. Endothelial cells modify the costimulatory capacity of transmigrating leukocytes and promote CD28-mediated CD4+ T cell alloactivation.
J. Exp. Med.
190
:
555
-566.
11
Adams, A. B., M. A. Williams, T. R. Jones, N. Shirasugi, M. M. Durham, S. M. Kaech, E. J. Wherry, T. Onami, J. G. Lanier, K. E. Kokko, et al
2003
. Heterologous immunity provides a potent barrier to transplantation tolerance.
J. Clin. Invest.
111
:
1887
-1895.
12
Valujskikh, A., B. Pantenburg, P. S. Heeger.
2002
. Primed allospecific T cells prevent the effects of costimulatory blockade on prolonged cardiac allograft survival in mice.
Am. J. Transplant.
2
:
501
-509.
13
Chalasani, G., Z. Dai, B. T. Konieczny, F. K. Baddoura, F. G. Lakkis.
2002
. Recall and propagation of allospecific memory T cells independent of secondary lymphoid organs.
Proc. Natl. Acad. Sci. USA
99
:
6175
-6180.
14
Tang, A. L., A. W. Bingaman, E. A. Kadavil, D. B. Leeser, D. L. Farber.
2006
. Generation and functional capacity of polyclonal alloAg-specific memory CD4 T cells.
Am. J. Transplant.
6
:
1275
-1284.
15
Shiao, S. L., J. M. McNiff, J. S. Pober.
2005
. Memory T cells and their costimulators in human allograft injury.
J. Immunol.
175
:
4886
-4896.
16
Kirkiles-Smith, N. C., K. Mahboubi, J. Plescia, J. M. McNiff, J. Karras, J. S. Schechner, D. C. Altieri, J. S. Pober.
2004
. IL-11 protects human microvascular endothelium from alloinjury in vivo by induction of survivin expression.
J. Immunol.
172
:
1391
-1396.
17
Kluger, M. S., D. R. Johnson, J. S. Pober.
1997
. Mechanism of sustained E-selectin expression in cultured human dermal microvascular endothelial cells.
J. Immunol.
158
:
887
-896.
18
Epperson, D. E., J. S. Pober.
1994
. Ag-presenting function of human endothelial cells: direct activation of resting CD8 T cells.
J. Immunol.
153
:
5402
-5412.
19
Murray, A. G., P. Petzelbauer, C. C. Hughes, J. Costa, P. Askenase, J. S. Pober.
1994
. Human T cell-mediated destruction of allogeneic dermal microvessels in a severe combined immunodeficient mouse.
Proc. Natl. Acad. Sci. USA
91
:
9146
-9150.
20
Sultan, P., J. S. Schechner, J. M. McNiff, P. S. Hochman, C. C. Hughes, M. I. Lorber, P. W. Askenase, J. S. Pober.
1997
. Blockade of CD2-LFA-3 interactions protects human skin allografts in immunodeficient mouse/human chimeras.
Nat. Biotechnol.
15
:
759
-762.
21
Ndejembi, M. P., J. R. Teijaro, D. S. Patke, A. W. Bingaman, M. R. Chandok, A. Azimzadeh, S. G. Nadler, D. L. Farber.
2006
. Control of memory CD4 T cell recall by the CD28/B7 costimulatory pathway.
J. Immunol.
177
:
7698
-7706.
22
Warnock, R. A., S. Askari, E. C. Butcher, U. H. von Andrian.
1998
. Molecular mechanisms of lymphocyte homing to peripheral lymph nodes.
J. Exp. Med.
187
:
205
-216.
23
Lee, K. M., E. Chuang, M. Griffin, R. Khattri, D. K. Hong, W. Zhang, D. Straus, L. E. Samelson, C. B. Thompson, J. A. Bluestone.
1998
. Molecular basis of T cell inactivation by CTLA-4.
Science
282
:
2263
-2266.
24
Ma, W., J. S. Pober.
1998
. Human endothelial cells effectively costimulate cytokine production by, but not differentiation of, naive CD4+ T cells.
J. Immunol.
161
:
2158
-2167.
25
Potsch, C., D. Vohringer, H. Pircher.
1999
. Distinct migration patterns of naive and effector CD8 T cells in the spleen: correlation with CCR7 receptor expression and chemokine reactivity.
Eur. J. Immunol.
29
:
3562
-3570.
26
Sallusto, F., E. Kremmer, B. Palermo, A. Hoy, P. Ponath, S. Qin, R. Forster, M. Lipp, A. Lanzavecchia.
1999
. Switch in chemokine receptor expression upon TCR stimulation reveals novel homing potential for recently activated T cells.
Eur. J. Immunol.
29
:
2037
-2045.
27
Gallatin, W. M., I. L. Weissman, E. C. Butcher.
1983
. A cell surface molecule involved in organ-specific homing of lymphocytes.
Nature
304
:
30
-34.
28
Wherry, E. J., V. Teichgraber, T. C. Becker, D. Masopust, S. M. Kaech, R. Antia, U. H. von Andrian, R. Ahmed.
2003
. Lineage relationship and protective immunity of memory CD8 T cell subsets.
Nat. Immunol.
4
:
225
-234.
29
Marzo, A. L., K. D. Klonowski, A. Le Bon, P. Borrow, D. F. Tough, L. Lefrancois.
2005
. Initial T cell frequency dictates memory CD8+ T cell lineage commitment.
Nat. Immunol.
6
:
793
-799.
30
Augustine, J. J., D. S. Siu, M. J. Clemente, J. A. Schulak, P. S. Heeger, D. E. Hricik.
2005
. Pretransplant IFN-γ ELISPOTs are associated with post-transplant renal function in African American renal transplant recipients.
Am. J. Transplant.
5
:
1971
-1975.
31
Heeger, P. S., N. S. Greenspan, S. Kuhlenschmidt, C. Dejelo, D. E. Hricik, J. A. Schulak, M. Tary-Lehmann.
1999
. Pretransplant frequency of donor-specific, IFN-γ-producing lymphocytes is a manifestation of immunologic memory and correlates with the risk of posttransplant rejection episodes.
J. Immunol.
163
:
2267
-2275.
32
Marzo, A. L., H. Yagita, L. Lefrancois.
2007
. Cutting edge: migration to nonlymphoid tissues results in functional conversion of central to effector memory CD8 T cells.
J. Immunol.
179
:
36
-40.
33
Murray, A. G., J. S. Schechner, D. E. Epperson, P. Sultan, J. M. McNiff, C. C. Hughes, M. I. Lorber, P. W. Askenase, J. S. Pober.
1998
. Dermal microvascular injury in the human peripheral blood lymphocyte reconstituted-severe combined immunodeficient (HuPBL-SCID) mouse/skin allograft model is T cell mediated and inhibited by a combination of cyclosporine and rapamycin.
Am. J. Pathol.
153
:
627
-638.
34
Liu, Z., H. Dai, N. Wan, T. Wang, S. Bertera, M. Trucco, Z. Dai.
2007
. Suppression of memory CD8 T cell generation and function by tryptophan catabolism.
J. Immunol.
178
:
4260
-4266.
35
Reinhardt, R. L., A. Khoruts, R. Merica, T. Zell, M. K. Jenkins.
2001
. Visualizing the generation of memory CD4 T cells in the whole body.
Nature
410
:
101
-105.
36
Fontenot, A. P., B. E. Palmer, A. K. Sullivan, F. G. Joslin, C. C. Wilson, L. A. Maier, L. S. Newman, B. L. Kotzin.
2005
. Frequency of beryllium-specific, central memory CD4+ T cells in blood determines proliferative response.
J. Clin. Invest.
115
:
2886
-2893.
37
Debes, G. F., U. E. Hopken, A. Hamann.
2002
. In vivo differentiated cytokine-producing CD4+ T cells express functional CCR7.
J. Immunol.
168
:
5441
-5447.
38
Kim, C. H., L. Rott, E. J. Kunkel, M. C. Genovese, D. P. Andrew, L. Wu, E. C. Butcher.
2001
. Rules of chemokine receptor association with T cell polarization in vivo.
J. Clin. Invest.
108
:
1331
-1339.
39
Unsoeld, H., S. Krautwald, D. Voehringer, U. Kunzendorf, H. Pircher.
2002
. Cutting edge: CCR7+ and CCR7 memory T cells do not differ in immediate effector cell function.
J. Immunol.
169
:
638
-641.
40
London, C. A., M. P. Lodge, A. K. Abbas.
2000
. Functional responses and costimulator dependence of memory CD4+ T cells.
J. Immunol.
164
:
265
-272.
41
Lovett-Racke, A. E., J. L. Trotter, J. Lauber, P. J. Perrin, C. H. June, M. K. Racke.
1998
. Decreased dependence of myelin basic protein-reactive T cells on CD28-mediated costimulation in multiple sclerosis patients: a marker of activated/memory T cells.
J. Clin. Invest.
101
:
725
-730.
42
Greenwald, R. J., G. J. Freeman, A. H. Sharpe.
2005
. The B7 family revisited.
Annu. Rev. Immunol.
23
:
515
-548.
43
Watts, T. H..
2005
. TNF/TNFR family members in costimulation of T cell responses.
Annu. Rev. Immunol.
23
:
23
-68.
44
Fritsch, R. D., X. Shen, G. P. Sims, K. S. Hathcock, R. J. Hodes, P. E. Lipsky.
2005
. Stepwise differentiation of CD4 memory T cells defined by expression of CCR7 and CD27.
J. Immunol.
175
:
6489
-6497.
45
Amyes, E., C. Hatton, D. Montamat-Sicotte, N. Gudgeon, A. B. Rickinson, A. J. McMichael, M. F. Callan.
2003
. Characterization of the CD4+ T cell response to Epstein-Barr virus during primary and persistent infection.
J. Exp. Med.
198
:
903
-911.
46
Savage, C. O., C. C. Hughes, B. W. McIntyre, J. K. Picard, J. S. Pober.
1993
. Human CD4+ T cells proliferate to HLADR+ allogeneic vascular endothelium: identification of accessory interactions.
Transplantation
56
:
128
-134.