| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print January 13, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Medicine and
Celiac Disease Center, College of Physicians and Surgeons of Columbia University, New York, New York; and
Department of Medicine, Chaim Sheba Medical Center and Tel Aviv University, Tel Hashomer, Israel
2Correspondence: The Sheba Cancer Research Center, Chaim Sheba Medical Center, Tel Hashomer 52621, Israel. E-mail: itamar.goldstein{at}sheba.health.gov.il
ABSTRACT
Celiac disease (CD) is a chronic, immune-mediated disorder of the gut, driven by T cells reacting locally to a distinct antigen, gliadin. Thus, CD offers the opportunity to study the T cell memory response to gliadin and whether gut tropism and T helper cell type 1 (Th1) polarization, which characterize the effector phase, are preserved in the memory progeny. It is notable that previous studies yielded conflicting results as to the presence of gliadin-specific memory CD4+ T cells in the peripheral blood of CD patients. However, we used a different and highly sensitive approach based on fluorescein-derived label dilution, whereby the memory cells are identified operationally by their greater capacity to proliferate upon re-encounter with antigen. Thus, using flow cytometry, we could resolve multiple successive generations as well as immunophenotype the dividing cells. Here, we show that the peripheral blood lymphocyte of some CD patients on a gliadin-free diet, but not healthy donors, contains a detectable population of CD4+ memory T cells specific for deamidated gliadin. Moreover, these gliadin-specific memory T cells are marked by a distinctive phenotype: They express high levels of the gut-homing ß7 integrins and primarily produce interferon-
and tumor necrosis factor 
. We conclude that memory for gliadin-derived antigens within the circulating CD4+ T cells is linked with gut tropism as well as Th1 polarization.
Key Words: human • cell trafficking • adhesion molecules
INTRODUCTION
Circulating, antigen-experienced T cells differ in the expression of homing and chemokine receptors. This heterogeneous expression determines the tissue tropism of the various populations of circulating effector and memory T cells [1
2
3
]. However, the factors that determine the particular tissue affinity of T cells are still being explored. It has been postulated that the particular expression of a set of tissue-homing molecules is dependent on the site of initial antigen encounter. Recent studies in mice show that the tissue source of the dendritic cells (DC) used to prime in vitro, naïve T cells determines the ensuing homing phenotype of the activated T cells. For example, T cell activation by gut-derived DC induces a T cell population with high expression of the ß7 integrins [4
]. It is postulated that the 4ß7 integrin is instrumental for the homing of lymphocytes to the Peyer’s patches via binding to its ligand, mucosal addressin cell adhesion molecule-1 (MAdCAM-1) [5
, 6
]. Moreover, in humans, oral immunization induces a larger population of antigen-reactive T cells expressing gut-homing molecules, including ß7 integrin, when compared with parenteral immunization [7
, 8
]. However, it is unknown whether this tissue tropism is also maintained in the memory lineage; i.e., are human memory T cells programmed to "remember" the antigen-priming locale?
In this regard, the established role of continued oral exposure to a well-defined antigen, as the primary cause of an immune-mediated disease, makes celiac disease (CD) an ideal model for the study of memory immune responses in humans. CD is an immune disorder of hypersensitivity to ingested gluten proteins, mostly gliadin [9 ], and is manifested as a chronic, small bowel inflammation. Moreover, it is postulated to be mediated by pathogenic human leukocyte antigen (HLA)-DQ2/8-restricted CD4+ T cells, specific for various gliadin-derived peptides [10 ]. In addition, there is ample evidence that the deamidation of gliadin by the ubiquitous enzyme, tissue transglutaminase (TTG), apart from rendering it more water-soluble, enhances its immunogenicity. This is probably a result of the observation that the negatively charged, deamidated gliadin peptides bind more avidly to the antigen-presenting molecules HLA-DQ2 and -DQ8 [11 12 13 ]. It is also postulated that T cells isolated from inflamed celiac guts are predominantly T helper cell type 1 (Th1)-polarized and that Th1 cytokines are implicated in epithelial cell activation and increased turnover [14 , 15 ].
It is notable that previous studies focusing on the peripheral blood lymphocytes (PBLs) of celiac patients, rather than on the lamina propria lymphocytes (LPL), and using conventional techniques (e.g., thymidine incorporation and cytokine secretion assays) have not consistently demonstrated gliadin-specific memory responses [16 ].
In this context, in a recent work, we used a highly sensitive, function-based approach to identify CD4+ memory T cells; they were distinguished operationally by their greater capacity to proliferate upon re-encounter of antigen. More precisely, we adapted the vital dye carboxyfluorescein diacetate succinimidyl ester (CFSE) dilution technique [17 ] to identify the memory T cells that have divided rapidly following triggering. This approach allows not only to detect the presence of antigen-specific memory T cells in the circulation but also to further analyze, at the single-cell level, the unique phenotype of the dividing memory cells.
In this work, we used this experimental system to identify comprehensively and characterize the circulating, gliadin-specific CD4+ memory T cells in celiac patients. In addition, we asked whether they remember the site and milieu of original priming. Here, we show for the first time that when detected in CD patients, the circulating memory T cells for gliadin tightly maintain a phenotype similar to the celiac gut-derived effector CD4+ T cells: ß7hi Th1-polarized and tumor necrosis factor (TNF-)+ cells.
MATERIALS AND METHODS
Study population
The study included 11 patients with CD, diagnosed according to accepted criteria, as determined by a thorough interview with one of the investigators (P. H. R. Green), positive serology results (antigliadin, antiendomysial, or anti-TTG), and a history of diagnostic small bowel biopsy results. The clinical characteristics of the study patients are summarized in Table 1
. The control group consisted of 10 non-HLA-DQ-matched, healthy blood donors. The Columbia University Medical Center Institutional Review Board (New York, NY) approved the study, and all subjects gave a written, informed consent.
|
All cell cultures were maintained in a humidified 37°C, 5% CO2 incubator, and the cultures were supplemented on Day 5 with 20 U/ml recombinant interleukin (IL)-2 (Hoffman-La Roche Inc., Nutley, NJ). Cultures were harvested on Day 8 for proliferation, cytokine secretion, and surface molecule expression assays. In addition, in some experiments, PBLs were restimulated on Day 14 with gTG, gliadin, or medium alone with the addition of 2 x 106 cells/well irradiated, autologous PBLs (250 Gy).
Isolation of small bowel LPL
The LPL were isolated from duodenal mucosal biopsies of volunteer patients, which were obtained during endoscopic or surgical procedures. Briefly, mucosal tissue samples were first washed twice in PBS to eliminate contaminating particles. They were then incubated with constant shaking for 10 min in 37°C in PBS with 1 mM EDTA and 10 µg/ml DNase I (Sigma-Aldrich). This step was repeated twice, and the cells collected from the supernatant after this phase (Fraction A) were mostly intraepithelial lymphocytes, and >90% of them were CD8+ on cytofluoroscopic analysis. Subsequently, the samples were cut into minute pieces, which were incubated for 2 h at 37°C on a shaker in RPMI with 100 U/ml penicillin, 100 µg/ml streptomycin, 1 mg/ml collagenase A (Sigma-Aldrich), and 10 µg/ml of DNase I. The supernatant obtained after this phase (Fraction B) contained mostly LPL and consisted of 
40% CD4+ cells, 40% CD8+, and 20% other cell types.
Antibodies
The following fluorochrome-conjugated, murine anti-human monoclonal antibodies (mAb) were all purchased from PharMingen (San Diego, CA): CD4 phycoerythrin (PE)-Cy5, CD8-PE, CD45RO-PE, CD45RA-fluorescein isothiocyanate, very late antigen (VLA)-4 (CD49d)-PE, the integrin ß7-PE, anti-interferon- (IFN-) PE, anti-TNF- PE, and anti-IL-4 PE. The appropriate isotype control fluorochrome-conjugated mouse mAb were also obtained from PharMingen.
Fluorescein-activated cell sorter analysis
The immunostaining and subsequent flow cytometric analysis of human lymphocytes were carried out as described previously [18
]. Briefly, cells were first treated with human immunoglobulin (Sigma-Aldrich) to reduce nonspecific antigen binding to Fc receptors and subsequently, were incubated with saturating concentrations of the indicated fluorescein-conjugated mAb for 15 min at 26°C. Cells were then washed twice in PBS, and fluorescence intensity was measured on a FACScan cytofluorograph and analyzed using Cellquest software (Becton Dickinson Immunocytometry Systems, San Jose, CA). Viable lymphocytes were defined by their forward-scatter/side-scatter characteristics or by propidium iodide exclusion.
Intracellular cytokine-staining assays
T cells were activated with 20 ng/ml phorbol 12,13-dibutyrate and 0.8 µM ionomycin (Sigma-Aldrich) in the presence of 2 µg/ml monensin (GolgiStop, PharMingen) for 5 h at 37°C. Following this brief activation, cells were harvested, fixed, and permeabilized using the Cytofix/Cytoperm Plus kit and then immunostained for intracellular IFN- or IL-4 production, all in accordance with the manufacturer’s recommendations (PharMingen). All samples were then analyzed on a FACScan cytofluorograph.
Analyzing T cell proliferation using the CFSE dilution method
The flow cytometric-based analysis of T cell proliferation by serial halving of the fluorescence intensity of the vital dye CFSE was used. Briefly, T cells at 107/ml were suspended in 1 ml RPMI and then pulsed with 4 µM CFSE (Molecular Probes Inc., Eugene, OR) for 15 min at 37°C with constant shaking. Subsequently, the CFSE was quenched with 50% human AB serum for 1 min, and the cells were washed twice with large volumes of RPMI medium. The cells were then triggered with a polyclonal stimulus, recall antigens, or medium alone as described above, plated, and cultured for 8 days. The cells were harvested on Day 8 and analyzed for CFSE content (generation number), various surface markers, and when indicated, intracellular cytokines. As some dye is lost from the parental generation, and some T cells can proliferate slowly in response to soluble factors in the medium (bystander effect), we considered in the data analysis only those CD4+ T cells that have undergone more than two cellular divisions as antigen-responsive. Moreover, a stringent compensation protocol for interdetector spillage of fluorescence was used during the sample acquisition to correct for the significant spillage of CFSE into the FL-2 channel.
Statistical analysis
The statistical analysis of the various samples was performed by the two-tailed paired Student’s t-test or the Mann-Whitney test as appropriate. In all statistical tests, a P value less than 0.05 was considered significant.
RESULTS
Detection of circulating gTG-specific memory CD4+ T cells in the peripheral blood of celiac patients
CD is driven by CD4+ effector and memory-type T cell clones in the gut, which respond locally to gliadin and particularly, after its in situ deamidation, by TTG [11
, 12
]. It is notable that the current view is that memory T cells specific for gTG are absent in the circulation of celiac patients, who adhere strictly to a gliadin-free diet [16
]. However, in the same studies that motivated this notion, it was shown that gTG-specific T cells re-emerge rapidly in the circulation of CD patients, following a short course of gliadin challenge. We hypothesized that these observations reflect the low precursor frequency (below assay detection) prior to the gliadin challenge and not necessarily a complete lack of circulating memory cells. Therefore, we first asked whether when using the highly sensitive CFSE-dilution assay, we can detect gTG-specific memory T cells in the peripheral blood of celiac patients on a gliadin-free diet (see Table 1
for patients’ details).
As described previously [19 ], the memory CD4+ T cells were defined operationally as the population of cells that had undergone at least three successive cellular divisions during the first 8 days following antigen triggering. Thus, PBLs from celiac patients (n=11) and from healthy controls (n=10) were obtained, labeled with CFSE, and stimulated with gliadin alone, TTG alone, gTG, TT (a prototypical recall antigen), or media alone. The positive control consisted of polyclonal stimulation with the superantigen TSST-1. At the end of culture (Day 8), the cells were harvested, and the CD4+ T cells were analyzed by flow cytometry for cell proliferation. As every T cell that divides n times will generate 2n daughter cells, the proportion of CFSE low cells at the time of analysis exponentially reflects the initial precursor frequency of antigen-specific T cells in vivo. In addition, the data were usually compiled from >1 x 10 4 independent, live-gate, single-cell events.
We found in six out of the 11 celiac patients tested, a significant memory response to gTG, defined as more than a twofold increase in the proliferation to gTG compared with media. Thus, as shown in Figure 1a
, in a typical responder, following stimulation with gTG, we detected a significant increase (30-fold) in the percentage of dividing CD4+ T cells compared with media (6% vs. 0.2%, respectively). Moreover, following stimulation with gliadin or TTG alone, we did not detect such a significant increase compared with the background proliferation. Of note, the response to gTG, in this particular experiment, was of the same order of magnitude as the memory response to TT. In comparison, TSST-1 induced a marked, proliferative response of CD4+ cells, as it triggers most Vß2+ T cell clones. In contrast to the findings in the celiac population, in all the healthy individuals tested (n=10), we did not detect a significant response to gliadin, TTG, or gTG compared with the media control (Fig. 1b)
. As expected, the response to TT or TSST-1 was comparable with the response observed in the patient population (Fig. 1a
and 1b)
.
|
It is notable that the analysis of the combined data (Fig. 2a ) from all patients and healthy controls demonstrates that the mean proliferation after one round of gTG stimulation in the celiac patients’ group was significantly higher than the proliferation to media alone (4.8±2.6% vs. 1.2±0.6%, P<0.05, Mann-Whitney test). Moreover, it was significantly higher than the gTG-induced proliferation in the healthy donors’ group (1.4±0.7%, P<0.05). In addition, the proliferation to gTG, in five out of the six patients with detectable memory responses, was significantly higher than the proliferation to native gliadin or TTG alone (data not shown). In contrast, as seen in Figure 2b , the responses to the recall antigen TT (5.3±2.9% vs. 9.4±4.5%, P=0.08) and to TSST (59.5±11.4% vs. 53±6.1%, P=0.38) were quite similar in the patients and the healthy donors.
|
High ß7 integrin expression marks gut LPL and a subset of circulating antigen-experienced CD45RA– CD4+ T cells
As discussed in Introduction, the function of the 4ß7 integrin, through its ligand MAdCAM-1, is known to be important for the migration of T cells into the gut [5
]. As 4ß7 mediates adhesion to MAdCAM-1, but 4ß1 does not, we used the expression of ß7 integrin as a marker for T cells with a gut-trafficking potential.
Thus, it was of interest to examine the pattern of expression of the ß7 chain in the PBLs of celiac patients and normal donors compared with gut-derived lymphocytes. We find that freshly isolated, circulating CD4+ T cells (in all samples tested) consist of three distinct populations with respect to ß7 expression: negative, low/intermediate, and high (Fig. 3a ). It is interesting that these three populations are also distinct with respect to their coexpression of CD45RA. More precisely, the ß7 low/intermediate cells were primarily CD45RA+ (i.e., naïve) cells, whereas the CD45RA– (i.e., primed) T cell compartment exclusively contained a small subset of ß7 high (ß7hi) cells as well as a much larger population of ß7-negative cells (Fig. 3a) . Moreover, no significant differences were observed for the pattern of expression between the patient and control population.
|
Taken together, the data show that the expression of high ß7 integrins is a feature of most CD4+ LPL and of a small, distinct population of antigen-experienced CD45RA– T cells in freshly isolated PBLs.
The circulating gTG-specific CD4+ memory T cells in celiac patients are enriched for the ß7hi phenotype
The above observation supports the notion that in vivo, the ß7hi phenotype marks gut LPL as well as gut-derived (specialized), circulating effector memory T cells. Thus, we hypothesized that the circulating, gliadin-specific CD4+ T cells also express high levels of the ß7-integrin chain and represent such a gut-specialized ß7hi subset.
To address this hypothesis, samples from the highly reactive celiac patients (n=6) were triggered with gTG, TT, or media alone. The rationale for using TT as a control in these experiments was that this prototypical recall antigen, routinely introduced by subcutaneous (s.c.) injection unlike the ingested gliadin, will allow us to better analyze the correlation between ß7-integrin expression and the original site of antigen encounter.
We found that the dividing population of gTG-specific CD4+ memory T cells was highly enriched for ß7hi cells (70% of the responding cells). In contrast, in the TT-specific CD4+ memory T cells, only 34% of the cells expressed this molecule (Fig. 4
). Moreover, the level of ß7 expression, as assessed by the mean fluoresence intensity, was much higher in the dividing, gTG-specific memory T cells when compared with the TT-specific memory T cells (427 and 155, respectively).
|
|
and TNF-, the hallmark cytokines of tissue inflammation.
Therefore, PBLs from celiac patients were obtained, stained with CFSE (as above), and stimulated with gTG or media alone. The Th1/Th2 profile of the rapidly dividing cells was determined on Day 8, using a short activation with phorbol 12-myristate 13-acetate (PMA)/ionomycin in the presence of a Golgi transport inhibitor and followed by intracellular staining for IFN- or IL-4 expression. We found in all patients, where a significant memory response was detected, that the memory CD4+ T cells, which have divided at least three times following gTG stimulation, are highly enriched for Th1 cells, with over 50% of the cells producing high levels of IFN- (Fig. 6
). In contrast, less than 10% of responding cells produced IL-4. Moreover, in parallel experiments, we also measured by enzyme-linked immunosorbent assay, the levels of IFN- in the supernatants (Day 6) of the gTG-stimulated cultures and unstimulated cultures, and found a significant increase in IFN- levels but not IL-4 in the gTG-stimulated cultures (data not shown).
|
is known to mediate tissue inflammation in many inflammatory disorders, including Crohn’s disease, we asked whether the dividing population of memory T cells produces TNF-. As predicted, we found that a substantial fraction (60%) of the cells dividing in response to gTG produced high levels of TNF-. These data suggest that the human CD4+ memory lineage for gliadin-derived antigens is committed to the proinflammatory Th1 phenotype upon re-encounter of gTG.
DISCUSSION
In this paper, using a function-based approach to characterize human memory T cells, we demonstrate for the first time that the PBLs of a large percentage of CD patients contain a discrete population of gTG-specific CD4+ memory T cells, marked by the expression of high levels of the ß7 integrin and Th1 predominance.
The generation of immunological memory is a complex process. Antigen presentation to naïve T cells is postulated to occur in secondary lymphoid organs and is accomplished mainly by professional antigen-presenting cells (APC), which have migrated there from the tissue sites, where antigen encounter had taken place. Stimulated, naïve T cells expand and differentiate to effector cells, which recirculate through the blood and home into sites of inflammation in the nonlymphoid tissues. Some of these cells subsequently assume a resting state with specific antigenic memory, defined as the capacity to readily mount an immune reaction upon subsequent re-encounter with antigen [21 , 22 ]. However, the study of true memory T cells, particularly in humans, has been difficult, chiefly as a result of the inability to distinguish unambiguously among recently activated effector cells, memory precursor, and true memory T cells.
Indeed, previous studies in CD have yielded variable results with respect to the detection of memory responses to gliadin-derived antigens in the PBLs of patients. For example, O’Keeffe et al. [23
] detected a proliferative response as well as IFN- secretion following stimulation with whole gliadin in untreated and dieting CD patients. In contrast, others found gliadin-induced cytokine secretion only in the PBLs of patients rechallenged with wheat but not in newly diagnosed patients or dieting patients [11
]. In yet another study, the researchers found circulating, gluten-responsive T cells, which were not restricted by HLA-DQ [24
]. These and other conflicting studies have led some authorities to believe that gliadin-specific, memory T cells with a pathogenic role are practically absent from the peripheral blood [16
].
In this context, we have described previously a function-based approach to more precisely identify true memory cells [19 ]. Briefly, we used the CFSE-dilution technique to identify the memory population of cells as those that have rapidly divided following stimulation with recall antigens. Thus, our first aim was to ask whether using this newer approach, we could detect gTG-specific memory CD4+ T cells in celiac patients treated with a gliadin-free diet. In addition, we used TTG-deamidated gliadin as the prime antigen and not a particular peptide sequence; we were aware that some authors had described various immunodominant, "canonical" epitopes [11 ]. However, others have shown that in children, T cell responses are directed toward multiple gliadin epitopes [25 ]. In addition, it is our experience that responses to whole TT are more consistent than responses to various "promiscuous" TT peptides [26 ]. Moreover, it is well-established that the different APC populations present in fresh PBLs are highly efficient in processing and presenting whole protein antigens [27 , 28 ].
We found that six out of the 11 celiac patients tested had a detectable population of gTG-specific memory T cells in their peripheral blood, distinguished as the cells that expanded promptly after stimulation with this antigen. In contrast, such a response was not present in any of the healthy controls investigated (Figs. 1 and 2) . Of note, as a single T cell that divides n times will generate 2n daughter cells, the proportion of CFSE low cells at the time of analysis correlates exponentially with the precursor frequency of the antigen-specific T cells in PBLs. Indeed, this proportion was different for each donor. For example, five of the 11 of patients had no significant response above background. One possible explanation may be the "temporal" decay in antigen memory, as all patients (but one) were on a gluten-free diet for a prolonged period before study entry. It can be envisaged that a strict diet can lead to reduction in the precursor frequency of gTG memory T cell in the circulation. For example, the one patient who admitted to "rare" dietary indiscretion had a positive response.
Of note, the above findings challenge the current dogma that gTG-specific CD4+ T cells are not present in the circulation of celiac patients, unless they "spill over" from the gut tissue during gliadin challenge. It is argued that the CFSE-dilution system has a superior sensitivity to detect memory responses with a minute precursor frequency [29
]. This could partly stem from the fact that each digital data graph is compiled from the analysis of 1 x 104-independent, single-cell events. Indeed, using this system, recall responses to whole TT protein can be detected consistently, even years after the last immunization [19
, 30
].
The issue of the limitations of assay sensitivity is emphasized further by the findings in a patient classified as negative, per the first round of antigenic stimulation, where we could actually detect a significant, specific response upon restimulation with gTG (Fig. 1c) . In addition, restimulation resulted in a greater response to gliadin compared with media but still significantly smaller than the response to gTG. In this regard, it has been shown previously that nondeamidated gliadin antigens can induce T cell activation as well [25 ]. It is possible that such a scenario is observed here, whereby the repeated stimulation reveals the capacity of the nondeamidated epitopes to induce detectable proliferation ex vivo.
Of note, HLA-DQ2 and -DQ8 are postulated to be necessary for the generation of the T cell-dependent, antigliadin responses. In this study, we used a control population of healthy, non-HLA-matched individuals. Thus, it can be argued that some of the difference between the patient population (usually positive for one of these DQ genes) and the control population is a result of the different frequencies for these gene products between the two groups. However, we reason that a specific, proliferative response in some of the healthy HLA DQ2+ donors would also indicate the presence of such memory T cells. Moreover, such an observation would not necessarily reduce the significance of a positive response in a celiac patient. Future studies may answer whether gTG memory T cells are also found in some healthy HLA DQ2+ donors and address their particular characteristics. There is some evidence that when such T cells are found in healthy individuals, unlike celiac patients’ T cells, they do not secrete IFN- upon antigen challenge [23
].
Next, it was of interest to examine whether augmented expression of the ß7 integrin is characteristic of these gTG-specific memory T cells. We reasoned that as the function of the 4ß7 integrin, via adhesion to MAdCAM-1, is essential for the migration of T cells into the gut [5
], the memory lineage for oral antigens may have been programmed to maintain (or rapidly express) higher levels of the ß7 integrins. Moreover, it is postulated that the ability of memory T cells to provide effective immune surveillance in different organs is dependent on their capacity to migrate into specific tissues and be retained there, which is in turn dependent on the expression of particular adhesion molecules and homing receptors that dictate the particular tissue tropism [1
, 2
]. Thus, we used high expression of the ß7 integrin molecule as a surrogate marker for enhanced gut-trafficking capacity.
Indeed, we found that the memory progeny of gTG-specific CD4+ T cells comprised mostly of cells with the ß7hi phenotype (Fig. 4) . Moreover, to confirm that this phenomenon was specific for ingested antigens (i.e., gliadin), we used TT, a typical recall antigen routinely introduced by s.c. injection, as an important control. The rationale was that the coanalysis of the response to this antigen will enable us to better answer whether high ß7 expression in activated, dividing CD4+ memory T cells depends on the gut being the original site of antigen priming, or alternatively, it is a general feature of all activated CD4+ memory T cells. As predicted, we found that high ß7 expression was a more distinctive characteristic of the gTG response.
To further address the possibility that the increased expression of ß7 on T cells responding to gTG may merely reflect a difference in activation, we also examined the patterns of expression of other postactivation-dependent molecules, e.g., VLA-4 and CD45RO. As predicted, the dividing CD4+ T cells in the TT- and gTG-stimulated cultures uniformly expressed high levels of these two postactivation markers (Fig. 5) .
It is interesting that in recent work, Anderson et al. [31
] found that predepletion of circulating ß7+ T cells practically eliminates the gliadin TTG-specific response, as assessed by the IFN- enzyme-linked immunospot. In contrast, this depletion had only a minimal effect on the responses to purified protein derivative or TT. In addition, they did not detect significant responses in HLA-DQ2-matched, healthy controls nor a positive effect for a double-dose of HLA-DQ2. It is important that these responses were detected in PBLs isolated 6–7 days following gluten challenge, which probably represent recently activated ß7+ effector-type cells rather than "true", resting memory T cells.
The precise mechanism by which the gut environment favors the induction of high ß7 expression is unknown. In this context, recent experimental observations suggest that the microenvironment in the Peyer’s patches, including interactions with local DC, may influence the homing preference of effector T cells [4
, 32
]. In addition, in a recent study, it was shown that effector T cells can express a gut or a skin-homing phenotype, depending on the site of initial antigen encounter [33
]. It was also demonstrated that following intestinal infection with Rotavirus or Salmonella typhi, the T cells activated by these antigens are programmed for gut trafficking [8
, 34
]. Furthermore, immunization with oral keyhole limpet hemocyanin (KLH) induces antigen-experienced T cells with higher 4ß7 expression compared with parenteral administration of KLH [7
]. These works in effector-type T cells, together with our work in human memory T cells, support the notion that conditions unique to the site of initial antigen encounter dictate the expression of certain homing receptors. It is notable that the data presented in our study are the first to demonstrate that this concept can be extended to include the memory lineage of CD4+ T cells implicated in a human inflammatory disorder.
To further assess the functional relevance of these circulating, gTG-experienced CD4+ cells, we examined their capacity to differentiate into memory-effector Th1 cells after antigen re-encounter. In view of the known Th1 polarization of gut-derived, gliadin-reactive T cells, our findings that the circulating, gTG-specific memory cells primarily secrete IFN- and TNF- (Fig. 6)
lend support to the relevance of these Th1 memory cells to the inflammatory process in the gut. In addition, we reason that these observations suggest the existence of a stringent commitment to the Th1 phenotype, which is maintained throughout the lifespan of the CD4+ T cell, from effector to long-lived memory T cell. These finding may lend support to the notion that these memory T cells were precommitted, probably through epigenetic mechanisms, to the Th1 phenotype.
In conclusion, these data point to a broader concept of immunological memory: The memory T cell remembers not only the antigen but also the site and cytokine milieu, where it was originally primed. We envision that future studies adopting the function-based, CFSE-dilution assay will produce more insight into the unique biology of gliadin-specific (and other gut-derived) memory T cells.
ACKNOWLEDGEMENTS
This work was supported in part by National Institutes of Health Grant U19 AI 46132 to L. C.
FOOTNOTES
1 Current address: Dept. of Medicine/Gastroentrology, Chaim Sheba Medical Center, Tel Hashomer 52621, Israel. 
Received July 17, 2005; revised August 23, 2005; accepted November 21, 2005.
REFERENCES
4ß7 integrin mediates lymphocyte binding to the mucosal vascular addressin MAdCAM-1 Cell 74,185-195[CrossRef][Medline] Gut 37,766-7761ß1 integrin, VLA-1, marks a distinct subset of human CD4+ memory T cells J. Clin. Invest. 112,1444-1454[CrossRef][Medline]4ß7, a gut homing integrin, by circulating and mucosal T cells in colonic mucosal inflammation Gut 40,241-246This article has been cited by other articles:
|
|
 |
|
  C. P. Wong, R. Stevens, B. Long, L. Li, Y. Wang, M. A. Wallet, K. S. Goudy, J. A. Frelinger, and R. Tisch Identical beta Cell-Specific CD8+ T Cell Clonotypes Typically Reside in Both Peripheral Blood Lymphocyte and Pancreatic Islets J. Immunol., February 1, 2007; 178(3): 1388 - 1395. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
