| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
INSERM U 395, CHU Purpan and Paul Sabatier University, BP3028 31024 Toulouse Cedex 3, France; and
INSERM U 387 Laboratoire d’Immunologie, Hopital de Sainte-Marguerite BP 29, 13274 Marseille Cedex 09, France
1Correspondence: INSERM U 395, CHU Purpan, BP3028 31024 Toulouse Cedex 3, France. E-mail. Denis.Hudrisier{at}toulouse.inserm.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
Key Words: T cell activation • T cell receptor • adhesion • biomembranes
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
. TCR triggering results in a sustained intracellular signaling that ultimately leads to T cell functions such as proliferation, cytokine production, or target cell killing. Cytotoxic T lymphocytes (CTL) expressing the CD8 coreceptor recognize a peptide bound to MHC class I molecules whereas helper T cells (Th) expressing the CD4 coreceptor do so with a peptide associated to MHC class II molecules. Much interest has focused on membrane-related events occurring when lymphocytes meet their target cells. The interface between lymphocytes and targets is termed ‘immunological synapse’ (IS) and has been observed for T (2
, 3)
, NK (4)
, and B (5)
cells. The T cell IS is thought to be where proximal TCR signaling events are initiated (6)
. Many different molecules on APC and T cell sides are involved in formation of the IS and participate in the increase of the overall avidity of cellular adhesion and in T cell activation (Fig. 1
). This is true for the CD4 or CD8 coreceptors, receptor–ligand pairs such as CD28/CD80 (7)
, and adhesion molecules such as LFA-1/ICAM-1, whose affinity has been shown to increase upon stimulation (8)
. The T cell cytoskeleton has been involved in the formation of the IS and of a scaffold of actin filament allowing TCR signaling to proceed (6
, 9
, 10)
. Molecules promoting T cell activation appear to be recruited within the IS by direct or indirect association to T cell cytoskeleton (10)
or to cholesterol-rich lipid microdomains (11)
.
|
Upon stimulation by a variety of ligands (peptide–MHC complexes, superantigens, anti-TCR antibody) the TCR is rapidly down-modulated (12)
. It is still unclear whether this process is directly linked to T cell activation or not (13)
. TCR internalization is followed by a degradation process that affects the TCR itself (14
, 15)
and TCR signaling molecules (16)
.
Receptor internalization is often accompanied by receptor-mediated ligand internalization; this is a common feature of ligand-RTK systems (17)
. In most cases, RTK ligands consist of soluble molecules; in some cases, membrane-bound ligands have been characterized (18
19
20)
. Ligand capture occurred in these conditions as well (21
, 22)
.
Internalization of soluble ligands by RTK of immunological interest [i.e., B cell receptor (BCR), complement receptors (CR), or Fc fragment receptors (FCR)] has been reported (23)
. More recently, internalization of membrane-bound antigens by the BCR (5
, 24)
and other immune RTK was demonstrated at biochemical and/or morphological levels. Such was the case for the TCR (25
26
27
28
29
30)
or NK cell receptors (31
32
33)
whose ligand is membrane bound. From the literature on receptor-mediated ligand internalization by nonimmunological RTK, one can easily conceive that soluble ligands or antigens become cointernalized with their high-affinity receptors such as the BCR, CR, or FcR. How ligand internalization proceeds when the ligand is membrane bound, especially when it has a low-affinity receptor (such as the TCR), is less straightforward and requires careful examination.
Ligand recognition by T cells initiates the TCR signaling cascade that ultimately leads to different lymphocyte functions such as proliferation, cytokine production, and (in the case of CTL) target cell killing. Those functions are triggered rapidly and efficiently by minute amounts of antigen (34)
. Despite intense investigation, it is largely unknown how such an efficient T cell activation can be generated on weak TCR–ligand interactions and low doses of ligand (35)
. The predominant model proposes that many TCR are serially engaged by a single peptide–MHC complex (36)
, the fast off-rates of TCR–ligand interaction being instrumental (37)
. Any engaged TCR is thought to internalize and then be degraded, thus limiting excessive T cell activation (38)
. The role of TCR internalization and of TCR-mediated ligand internalization with regard to T cell activation or inactivation is still hypothetical and will be discussed below.
|
RECENT RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
reported the transfer of antigenic material from macrophages to lymphocytes after a contact of 24–72 h. Later experiments by Lee and Paraskevas (40)
suggested that T cells could incorporate Fc receptors and MHC molecules released by macrophages and acquire the capacity to bind immunoglobulin-coated particles. The presence of unexpected MHC class II expression on lymphocytes was detected after allogeneic stimulation or in radiation bone marrow chimera (41
42
43)
. In light of the increasing knowledge on molecular aspects of T cell activation, these issues have been reexamined from a biochemical, morphological, and functional point of view.
The acquisition of ligand from APC by CD4+ T cells was indirectly observed. Indeed, Arnold et al. showed that CD4 + T cells first exposed to APC were able to mediate the proliferation of CD4+ T cells specific for the same antigen in a secondary culture (44)
. This process required the presence of antigen on the APC and could be blocked by antibodies directed against the restricting MHC class II molecules. Uptake of irrelevant antigens by T cells was observed when the specific antigen was present on the same APC. Simultaneously, ligand acquisition by CD8+ CTL was demonstrated. Indeed, a CTL clone was shown to acquire MHC class I molecules fused to green fluorescent protein expressed by Drosophila cells in an antigen-dependent manner (26)
. This acquisition was 1) antigen dose dependent, 2) in parallel with TCR down-modulation, 3) clearly blocked by anti-class I or clonotypic anti-TCR antibodies, and 4) followed by ligand internalization (26)
. Acquisition of the antigenic peptide itself by T cells was first shown in indirect experiments. APC loaded with the antigen were incubated with T cells specific for that antigen. Those T cells served as stimulators for T cells sharing the same antigen specificity in a secondary culture (26
, 44)
. In the case of CD4+ T cells, antigen acquisition by T cells was evidenced by the proliferative response in the secondary culture (44)
. In the case of CTL, antigen acquisition was assessed by the lysis of primary CTL by fresh CTL endowed with the same specificity, a process called ‘fratricide killing’ (26)
. Evidence that T cells capture the antigenic peptide was provided in direct experiments where the acquisition of a specific fluorescently labeled (but not an unrelated fluorescent) peptide by CTL was shown by flow cytometry (27)
.
Sprent et al. suggested that the capture process is a result of TCR stimulation rather than linked to T cell activation (29)
, because T cell activation occurs with plastic-bound TCR ligands or anti-TCR antibody that were assumed to represent nonextractable ligands. Such stimuli do not represent physiological TCR stimuli, and information drawn from experiments involving these ligands needs to be considered carefully. In the case of the BCR, an efficient extraction of such immobilized ligand was reported even for weak BCR–ligand affinity (24)
.
Therefore, though it is clear that antigen acquisition results directly from TCR–ligand interaction, it was not known whether TCR signaling was required. Using a combination of approaches, the importance of TCR signaling in the capture process was demonstrated (27)
. First, the capture process performed by CTL is dramatically inhibited by anti-CD8 antibodies that block the contribution of CD8 to TCR signaling or by an inhibitor of src kinases, the first to be involved in TCR signaling. Second, those reagents blocked the capture process even when added after CTL–target cell conjugate formation, indicating that a sustained TCR signaling was required. Third, using a battery of mutated antigens endowed with different CTL activation properties, a strong correlation between their capacity to 1) stimulate CTL activation and 2) promote target cell molecules acquisition by CTL was observed. In another independent study, the capture process was shown to be sensitive to inhibitors of the cytoskeleton (29)
just like T cell activation (38)
.
Note that an immediate consequence of specific TCR stimulation is to tighten the molecular association between T lymphocytes and antigen-presenting cells. Thus, when specific and nonspecific conjugates between cytotoxic T lymphocytes and target cells were compared, it was found that the presence of specific TCR ligand on target cells resulted in a fourfold increase of the contact area, together with a concentration of several molecular species in this contact area (45)
.
In addition to MHC molecules, costimulatory molecules such as B7–1 (CD80) and B7–2 (CD86) were extracted from APC in a CD28-dependent manner (28)
. When CD28-B7 interaction is possible (i.e., when APC express B7 molecules and T cells express CD28), CD28-mediated acquisition largely masks TCR-mediated acquisition of APC molecules. CD4+ and CD8+ T cells have been shown to perform this capture process (28
, 29)
. Naive and activated T cells are both able to acquire APC molecules even though quantitative and possibly qualitative differences exist between the capture process performed by these cells (28
, 29)
. Indeed, activated T cells are much more efficient in capturing APC-derived molecules than resting T cells.
Brezinshek et al. (46)
reported that activated T lymphocytes acquired endothelial cell surface determinants (essentially adhesion molecules such as CD31, CD49d, CD54, or CD62E) during trans-endothelial migration. Three interesting features were 1) the transfer of a lipophilic dye, 2) much less transfer efficiency with resting T lymphocytes compared with activated T cells, and 3) failure to inhibit the transfer of an adhesion molecule such as ICAM-1 (CD54) when its receptor was blocked with suitable antibodies. Thus, the acquisition process seems to involve similar requirements independent of the triggering event: i.e., antigen recognition or trans-endothelial migration.
APC–T cell contact is critical for the capture performed by resting cells, whereas an APC–T cell contact-independent mechanism seems to contribute, although moderately, to the overall capture performed by activated T cells (29)
. APC previously fixed or sonicated do not donate molecules to T cells (28
, 29)
. This capture occurs in vitro and in vivo (28
, 30)
.
Apart from the mechanism of transfer itself, little is known about the biological consequences of antigen capture by T cells. The capacity of T cells to act as APC is still controversial. In some instances, T cell–T cell antigen presentation has been shown to support the induction of cytotoxic effector functions (47)
. In other cases, induction of tolerance by T cell deletion or anergy was obtained (30
, 48
, 49)
. In CTL, the fratricide killing mechanism has been proposed to account for the state of T cell exhaustion that results from the exposure of T cells to high doses of antigen (26)
. More recently it has been proposed that fratricide killing could be the mechanism by which peripheral CD4-CD8-T cells could exert their regulatory function (49)
.
|
CONTROVERSY/UNANSWERED QUESTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
Whereas the requirements of the capture process are relatively well defined (see above), the mechanism of transfer remains controversial. Using APC that were either surface biotinylated or labeled with fluorescent lipids, it was shown that besides specific ligands, several bystander APC molecules were acquired by CD4+ (25
, 30)
or CD8+ (27)
T cells upon specific antigenic stimulation. How could different APC molecules be acquired by T lymphocytes? One may consider three possibilities (see Fig. 2
):
|
Transfer of cleavage fragments of membrane molecules
Shedding of cell membrane constituents by living cells seems to be a universal phenomenon (50)
. GPI-anchored membrane molecules may be released after local phospholipase activation, and membrane proteins may be cleaved by activation of a suitable protease (see ref 51
for a review). Membrane-bound RTK ligands (20)
, including MHC molecules [52
), do not escape this rule. B7.2 molecules have been shown to exist as soluble molecules in vivo (53)
. Multiple experiments suggest that release could be regulated by at least three different mechanisms. First, lymphocytes might induce the release of membrane molecules by a neighboring cell through cytokines such as IFN-
or TNF-
(54)
. This might appear as a contact phenomenon if the stimulus is expressed as a membrane-bound form. The dependence of the proteolytic cleavage of cell membrane molecules on signaling events has recently been documented. Thus, it was demonstrated that L-selectin cleavage was hampered when calmodulin was bound to the cytoplasmic tail of the molecule (55)
. Second, cross-linking of cell surface molecules may induce the cleavage of the same (56)
or other (57)
membrane molecules as a result of signaling events. This finding may be relevant to the interaction between T lymphocytes and APC. Indeed, nearly all accessory cell molecules involved in interaction with T lymphocytes may be endowed with signaling properties, including MHC molecules (58
, 59)
and ICAM-1 (60)
. Therefore, it is not surprising that APC were often activated after interaction with T lymphocytes, as revealed by transient increases of cytosolic calcium level (61)
. Third, trans-proteolysis of membrane glycoproteins by surface proteases is a conceivable mechanism.
However, although capture of shed molecular fragments might constitute the simplest way for a cell to acquire foreign constituents, it seems likely that intact molecular components may be more readily embedded in membranes with functional conformation than soluble fragments. First, soluble MHC molecules have been shown to trigger the apoptosis of T cells in vivo (62)
. Although proteolysis has not been examined directly, alternative mechanisms involving membrane-bound APC molecules have been demonstrated in the capture process by T cells (25
, 27)
.
Transfer of membrane material through shed vesicles
It is well known that living cells constitutively shed membrane fragments whose composition may differ slightly from that of bulk membranes (63)
A possible use of this phenomenon might be to allow cells to lose potentially harmful components such as complement factors (64)
. Vesicle shedding may be regulated; thus, monocyte vesiculation was reported to follow exposure to bacterial lipopolysaccharide (65)
.
This process likely is heavily dependent on interactions between the plasma membrane and underlying cytoskeleton: local disruption of the cytoskeleton is known to result in membrane blebbing. Microvesicle release by activated platelets probably involves calpain action, with rupture of membrane-associated cytoskeleton and dissociation of membrane/cytoskeleton attachment (66)
. Recent biophysical evidence (reviewed by in ref 67
) sheds new light on the importance of membrane/cytoskeleton interaction. The use of so-called ‘optical tweezers’ allowed precise study of the force generated by small membrane tethers obtained by pulling microbeads bound to the cell surfaces with a force of a few pico newtons (pN). It was possible to estimate intrinsic plasma membrane tension and energy of adhesion to the cytoskeleton; the latter parameter accounted for 
75% of the apparent membrane tension. Thus, transient disruption of cortical microfilaments as a consequence of cytosolic calcium increase (68)
or weakening of membrane/cytoskeleton interaction with a second messenger such as phosphatidyl inositol 4,5 bisphosphate (69)
might result in local release of the plasma membrane, favoring vesicle formation.
Mannie et al. (25)
reported the purification of large vesicles continuously produced by APC. Once incubated with these vesicles, T cells were shown to acquire MHC class II molecules (if APC were pulsed with the cognate antigen before preparation of vesicles) (25)
. Vesicles could also convey APC molecules when T cells and APC were separated by a semipermeable membrane (29)
. Being enriched in MHC, costimulatory molecules, heat shock proteins, and adhesion molecules, exosomes (50–90 nm vesicles released by a variety of cells after direct fusion of the external membrane of the endosome with plasma membrane) may be good candidates (70)
. However, the phenomenon of fratricide killing argues against a major contribution of vesicles in the capture process as performed by CTL. When APC were pulsed with two different antigens and incubated with CTL specific for one of them, those CTL inherited mainly, if not exclusively, the specific antigen, as demonstrated by functional assays (26)
. When a similar experiment was performed by Mannie et al., CD4+ T cells also apparently inherited most antigens initially present on the APC and not preferentially the antigen for which they were specific (44)
. It is therefore possible that different mechanisms are involved in these two situations. Other differences between these studies concern the kinetics of the capture process [Arnold et al. (44)
reported a much slower process than others (26
, 27
, 29)
] and its effectiveness in the presence of inhibitory antibodies such as anti-CD4 and anti-LFA-1 (25
, 30
, 44)
or not (27)
. The propensity of the APC used in these different systems to produce vesicles might condition the manner whereby T cells acquire the antigen. Alternatively, the long incubation periods required for the experiments performed by Mannie et al. could favor cell death and the release of particles. It is unlikely that a difference between CD4+ and CD8+ T cells is involved because Hwang et al. used both types of cells in their studies and observed similar requirements (28
, 29)
.
Direct transfer of intact molecules between apposed membranes
As a direct result of thermodynamic fluctuations, membrane constituents may be spontaneously released with low frequency into the extracellular medium. They could then be captured by a neighboring membrane. This phenomenon may be especially efficient with small molecules: thus, the transfer of a fatty acid (9-(3-pyrenyl) nonanoic acid) between liposomes was reported to occur with a time scale of a second or less (71)
. In another study, Brecher et al. (72)
concluded that long chain fatty acids were readily exchanged between liposomes provided the bilayer was in liquid state, as is the case in cell plasma membranes under physiological conditions. It is thus not surprising that GPI-linked proteins were readily transferred from erythrocytes to endothelial cells during adhesion (73)
.
The possibility that membrane molecules might be uprooted from the plasma membrane may at first seem a more remote possibility on a thermodynamic basis. As early as 1978, Bell (74)
used simple physical reasoning to estimate that a force of a few tens of pN might be sufficient to extract a standard glycoprotein from the plasma membrane. More recently, Evans et al. (75)
used micromanipulation to separate erythrocytes bound through a minimal number of membrane carbohydrate groups: based on ultrasensitive fluorescence detection techniques, they concluded that separation might involve the transfer of isolated molecules from one cell to the other. Finally, Dwir et al. (76)
provided strong evidence suggesting that leukocyte L-selectin molecules could withstand a traction of several tens of pN for a fraction of a second when they were normally interacting with cytoskeletal elements. But detachment might occur within a millisecond when cytoskeletal interactions are prevented by deleting the cytoplasmic tail. A filopodium of 0.1 µm thickness should be able to exert traction on the order of 100 pN, thus allowing uprooting of weakly bound ligands on a contacting surface (77)
. Thus, though membrane lipids or GPI-anchored molecules are expected to spontaneously exchange between tightly bound membranes, intrinsic glycoproteins should, in principle, be extractable by moderate active traction through membrane receptors if they are not strongly bound to cytoskeletal elements.
Thus, APC membrane fragments might be torn by T cells during cell–cell adhesion (78)
. This can be deduced from the qualitatively important capture that requires direct cell–cell contact (29)
as well as from the impressive correlation between the efficiency of the capture process and the strength of TCR stimulation (27)
. The exchange of T cell- and target cell-derived proteins has been documented (79
80
81)
. These transfers were mediated by direct cell–cell contact, did not require specific target cell–T cell interaction, and were bidirectional. In contrast, capture processes observed more recently were shown to require specific APC–T cell interaction and were unidirectional, no molecule derived from T cell being observed onto APC (27
28
29)
. Since the former capture process was observed using very sensitive techniques, it is likely that the nonspecific and bidirectional capture process represent the background exchange of molecules involved in all cell–cell contact. Besides this basal exchange, a specific stimulation would trigger cellular machinery that might represent the driving force for the specific and unidirectional capture of molecules. This is probably the situation observed in the T cell-mediated capture of APC molecules. It will be worthwhile to understand the role of membrane tension in both the specificity and the polarization of the capture process. For example how the lipid composition of the membrane (raft vs. nonraft), protein–protein interaction, the intensity and the origin of shear forces exerted at the APC–T cell contact area are involved is an interesting question. Inhibition of the capture process by molecules preventing T cell activation has been well documented. It will be interesting to search for molecules that may selectively prevent the detachment of APC membrane fragments and test their effect on T cell activation.
The second important question concerns the fate of the acquired material. The phenomenon of fratricide killing suggests that, once acquired, the ligand is presented properly at the CTL surface. It has been shown that at least some acquired molecules—MHC class I, B7.1, and B7.2—were internalized (26
, 28)
. How could this apparently paradoxical situation be explained? The antigenic peptide available for TCR recognition at the surface of CTL could originate from a minute fraction of acquired molecules not submitted to the internalization process. This could explain why, despite the well-known sensitivity of the cytotoxic function of CTL (one single peptide–MHC being eventually sufficient to promote target cell killing) (82)
, fratricide killing is remarkably inefficient and occurs only at extremely high doses of antigen (26)
. The inefficiency of fratricide killing resulting from TCR-mediated ligand acquisition contrasts with the efficiency of CD4+ T cell activation resulting from BCR-mediated acquisition of the antigen by B cells (23)
. One explanation could be that in the case of B cells, the antigen needs to be internalized and degraded within the B cell before being presented again to CD4+ T cells in the form of MHC class II-associated peptides. Thus, internalization and degradation of the antigen are instrumental in this system. In the case of T cells, however, one expects that ligand internalization and degradation result in antigen loss and no representation at the CTL surface. Alternatively, part of the internalized material could recycle at the CTL surface. In an elegant study, Vignali et al. (14)
recently demonstrated that the process of ligand-induced TCR down-modulation is based on an inefficient recycling and active TCR degradation rather than on increased endocytosis. Moreover, MHC class I molecules themselves are rapidly degraded and poorly recycled after endocytosis (14)
.
The most important question concerns the biological significance of the capture process. In the case of the BCR or FcR, cointernalization of the antigen has a clear biological significance (23)
. Once internalized, the ligand enters the antigen processing pathways, leading to generation of short antigenic peptides that associate intracellularly to MHC class II molecules, then as peptide–MHC complexes migrating to the B cell surface, where they are recognized by CD4 T cells (23
, 83)
. These molecular events ultimately result in the T cell–B cell cooperation necessary for initiation of the B cell response in secondary lymphoid organs resulting in antibody production. In contrast to that clear situation, the role(s) of TCR-mediated ligand internalization is(are) still hypothetical (Fig. 3
).
|
Ligand-receptor internalization is generally considered an extinction mechanisms for cell signaling (17)
. Indeed, because ligand density on the target cell surface and receptor expression on T cells are both decreased, cells are less susceptible to activation. This process corresponds to the phenomenon of receptor desensitization observed for many different ligand-receptor systems, particularly for ligand–TCR interaction (84)
. This hypothesis stems from two observations. First, T cells previously exposed to high ligand densities on APC exhibited a decreased response to a second rapid restimulation by the ligand (84)
. Second, T cells inefficient at internalizing the TCR (because of the expression of the dominant negative form of the Rab5GTPase involved in endocytosis) were hyper-responsive to ligand stimulation (85)
.
The striking correlation between kidnapping of target cell membrane and target cell lysis suggests that the release of membranes could be the result of the damages imposed on target cells (27)
. In fact, the perforin and granzymes released by CTL are known to cause the apoptosis of target cells. Apoptosis is known to provoke membrane blebbing (86
, 87)
; these blebs could convey the molecules acquired by T cells. However, it has been shown that perforin-deficient and perforin-competent CTL have a similar ability to capture target cell membrane fragments (27)
. CD4+ T cells normally devoid of cytotoxic function capture target cell molecules (25
, 28
29
30
, 44)
.
By providing prolonged exposure to TCR ligands and other molecules involved in T cell activation, the APC membranes could favor serial TCR triggering (12)
and thus participate positively in intracellular signaling. Whether the capture of soluble or membrane-bound ligands by RTK could participate in intracellular signaling, however, is still a matter of debate (21
, 88
, 89)
. Many investigators have searched for (and seldom found) signaling capabilities of internalized ligand-receptor pairs. Such is the case for the EGFR/EGF, which, once internalized, continues to signal in the p21ras pathway (90
, 91)
. In the case of the TCR, no data are available to support this concept. Nevertheless, it is well known that treatments preventing extracellular ligand–TCR rebinding, such as physical APC/T cell conjugate disruption or blocking anti–ligand or anti-TCR antibodies, result in rapid and massive inhibition of T cell-mediated cytokine production (38)
. These results indicate that if the ligand/TCR complex, once internalized, retains certain signaling capacities, these capacities are not sufficient to maintain T cell activation. However, this possibility has not been evaluated carefully. Theoretically, such a situation would benefit the CTL, which, in contrast to CD4+ T cells stays a short time in contact with APC (<5 min). That would allow CTL to remain minimally activated while recycling from one target to another, thus permitting TCR signaling to be sustained for a long period (>45 min), a condition necessary for cytokines production (92)
.
Finally, it has been proposed that ligand internalization could have feedback inhibitory effects on the immune response such as tolerance induction by lymphocyte ‘exhaustion’ via deletion (fratricide killing) or anergy (26
, 28
, 30)
. Different CTL functions are triggered by different concentrations of antigen (93)
. We recently observed that capture of antigen, in contrast to the capture of membranes, occurs only at relatively high concentrations (27)
. At these high doses of antigen, TCR occupancy results in TCR down-modulation, but fratricide killing requires even higher concentrations (26)
. Since it is an inefficient process, it is questionable whether sufficiently high doses of antigen can be reached in infectious situations for fratricide killing to play a role in the exhaustion of the CTL response.
|
PROSPECTS AND PREDICTIONS |
|---|
|
TOPABSTRACTINTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
imposed by the spatially and temporarily limited reorganization of lipids and cytoskeleton on the T cell side of the T cell/APC contact area (6
, 11)
, a portion of the target cell membrane is torn. Indeed, the attachment between T lymphocytes and antigen-bearing cells was found to be strong enough to rupture the target cell membrane when conjugates were disrupted by mechanical force (94)
. One can easily imagine that the target cell side of the IS becomes weak and is more prone to detach. Indeed, MHC molecules have been proposed to concentrate into rafts on the APC (95)
. Because these domains have a higher fusion temperature that makes them more resistant to mechanical stress or detergent extraction (96)
, domains containing the concentrated antigen could detach from the remaining part of the membrane. The driving force needed to detach these fragments could come from TCR signaling, particularly the recruitment of a dense network of actin cytoskeleton on the T cell side of the synapse (6)
. Recent experiments performed with BCR indeed support the hypothesis that it is at the level of the IS that capture occurs (5
, 97)
.
Once transferred by the T cell, APC molecules could participate in several T cell functions. That intercellular molecule transfer may be of functional significance is illustrated by the phenomenon of trans-signaling (98)
. When the IL-6Ralpha chain of the IL-6 receptor is shed, it can bind soluble IL-6, and the complex may activate cells expressing the gp130 unit of the receptor, thus bestowing on them the capacity to respond to IL-6.
Released membrane constituents may have a biological function; thus, soluble adhesion molecules may exert a competitive effect. The presence of selectins and ICAM in normal plasma may inhibit leukocyte endothelium interaction (99)
; soluble forms of adhesion molecules were shown to stimulate receptor-bearing cells (100)
. Similarly, soluble cytokine receptors may either competitively inhibit cytokine function or conversely reduce blood cytokine clearance, resulting in more durable persistence and action. Finally, shedding may be a general mechanism of eliminating potentially harmful membrane components (64)
.
It is known that T lymphocytes have a capacity to recruit other T lymphocytes and act as helper cells for the generation of effector T cell (such as CTL). The presence of antigenic material on these helper T cells might provide a simple way of imparting some specificity on help delivery, which might be particularly significant when there is a need to expand T cell clones specific for a relatively rare antigen.
An important question is whether the capture of bystander molecules could play a physio-pathological role in certain infectious situations. Cellular receptors could be inherited by T lymphocytes while they capture membrane fragments of virally infected target cells. Once acquired, such a receptor could insert properly into the T cell membrane (just like peptide–MHC when recognized in a fratricide fashion) and confer at least transiently novel properties on T cells. Thus, cellular receptors once acquired by T cell might function as a decoy receptor or allow T cells to respond to nonphysiological stimuli. As receptors for pathogen entry into cells, acquired receptors might confer on T cell sensitivity to pathogen infection.
It appears that the intercellular transfer of molecules during cell–cell interaction is a general process. It is possible that certain cell types such as T lymphocytes have exploited this process to greater benefit than other cells in order to favor the productive encounter of rare antigens. When high doses of antigen are present, the capture process could participate in the physiological or (as proposed above) pathological down-regulation of the immune response.
Note added in proof:
After submission of this paper, the formation of membrane bridges at the CTL/target cell synapse was observed by electron microscopy: Stinchcombe, J. C., Bossi, G., Booth, S., and Griffiths, G. M. (2001) The immunological synapse of CTL contains a secretory domain and membrane bridges. In Immunity, Vol.15, pp. 751–761. The unidirectional passage of target cell molecules onto CTL through the synapse was reported in the same study.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRECENT RESULTSCONTROVERSY/UNANSWERED QUESTIONSPROSPECTS AND PREDICTIONSREFERENCES |
|---|
