| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print January 13, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Cell Biology and Kaplan Cancer Center, New York University School of Medicine, New York
1Correspondence: Department of Cell Biology, New York University School of Medicine, 550 First Avenue, New York, NY 10016. E-mail: freya01{at}med.nyu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
15 years, antitumor immunoglobulin and T cells have been used to identify tumor antigens, which in turn, have served as the basis for therapeutic vaccine trials [1
, 2
]. However, experimental cancer vaccines, although in some patients result in elimination of large tumor burdens, have a low frequency of long-term cancer remission in most patients, ca. <5% [2
]. Therefore, as tumors express antigens that distinguish themselves from nontransformed cells in immunological terms (i.e., elicit immune responses to growth of primary tumor and can target tumor cells in vivo), and tumor vaccines prime unsuccessful antitumor immune responses in patients, it is likely that growth of cancer induces immune tolerance to tumor cells. Although there are several types of T cell tolerance, mature, antigen-specific CD8+ T cells isolated from tumors are lytic-defective, implying that the tumor microenvironment inactivates the antitumor effector phase. The nature of the functional local tolerance to antitumor immune response is the subject of this review.
Key Words: T cells • cytolysis • tumor
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
(IFN-) upon recognition of cognate tumor in vitro, supporting the notion that the presence of extant tumor in situ does not restrict the ability to prime T cell immune response by vaccination [4
]. The lack of systemic immunesuppression is also inferred from the observation that tumor-bearing mice or patients can be immunized against a variety of nontumor antigens [5
]. However, in the unmanipulated patient and in patients receiving immunotherapy, in spite of evident antitumor immune response, tumors are not eliminated, implying tolerance. Several potential mechanisms of tumor escape from immune response have been proposed, which are not mutually exclusive (Fig. 1
), and include a low frequency of high-avidity antitumor T cells (likely as a result of thymic selection [6
]), inefficient priming of antitumor T cells {perhaps as a result of the close similarity of most tumor antigens to "self" [7
], inhibition of priming function of dendritic cells (DC) [8
], or the absence of an acute phase and attendant "danger" signals in the early stage of tumorigenesis [9
]}, enhanced expression on tumor cells of inhibitors of complement function (which blocks complement-mediated lysis in spite of production of antitumor Ig [10
]), suppression of antitumor T cells {by CD4+CD25+ regulatory T (Treg) cells [11
], CD8+ Treg cells [12
], or tumor-derived factors [13
]}, and various strategies of tumor evasion from the antitumor T cell effector phase [14
].
|
The finding that tumor-specific T cells isolated from tumors are lytic-defective, and systemic T cell responses to nontumor antigens are normal also strongly imply tumor-specific, functional inactivation of antitumor T cells [7 ]. Furthermore, using cloned patient CD8+ antitumor T cells as the basis of in vitro gene expression assays, there have been several human tumor antigens identified, which are peptides whose expression results from aberrant transcription or translation events (or even post-translational [21 ]) uniquely in tumor cells [22 23 24 25 26 27 28 29 30 ]. The amino acid sequences of these antigens are entirely novel and are thus not self, as the epitopes are not expressed in normal cells. T cells recognizing this class of tumor antigen are probably high-avidity, as not being reactive with self, they are unlikely to be subject to negative selection. However, as the T cells used for antigen identification were obtained from patients, those T cells are, like those reactive with model foreign antigens expressed in murine tumor cells (e.g., ß-galactosidase, influenza hemagglutinin, ovalbumin, or lymphocytic choriomeningitis virus gp33), unable to eliminate tumors in situ.
The findings that tumor-infiltrating lymphocytes (TIL) are lytic-defective in situ [31 ] (see below), and freshly isolated TIL are transiently lytic-defective in vitro [18 , 19 ] suggest that the tumor microenvironment provides a functional "barrier", such that CD8+ T cells cannot respond after recognition of cognate antigen expressed by tumor cells in situ.
|
TUMOR ESCAPE FROM T CELL KILLING |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
Although tumors in patients with a variety of cancer types have been shown to contain TIL [20
, 36
37
38
], infiltration of melanoma has been analyzed extensively for unmanipulated patients [39
, 40
] and after experimental immunotherapy [41
]. A high percentage of primary melanoma contains T cells, even if analyzed at early stages of tumor development [42
]. This observation argues that primary tumor is antigenic, and antitumor T cells can home to tumor, at least for the case of human melanoma. [The presence of T cells within tumor tissue does not prove antigen specificity of TIL, as the inflammatory environment in tumors may be sufficient to recruit immune cells. However, following vaccination, tumor epitope-specific T cells were recovered from TIL, and recall antigen-specific T cells (influenza) were not detected, suggesting that peripheral blood lymphocyte did not contaminate TIL.] In addition, the percentage of TIL in tumor increases and antigen-specific clonotypes become identifiable upon vaccination [17
, 40
]. However, as mentioned above, increased frequency of tumor antigen-specific T cells in tumor issue (or in the circulation), which upon isolation can secrete IFN-, usually does not correlate with tumor regression [4
, 43
, 44
].
The abundance of TIL in human tumors is not known sufficiently to judge whether infiltration is "robust"; however, in murine model systems, it is generally accepted that the number of antitumor TIL is low, even if the tumor is modified to express a model, nonself antigen. The number of antigen-specific TIL can be significantly greater in transgenic mice expressing a TCR for a model or cognate tumor antigen compared with wild-type mice, implying that the precursor frequency and activation of antitumor T cells in wild-type mice are modest. For tumors expressing a model antigen, the low percentage of TIL is likely not a result of thymic deletion or low antigen levels. TIL abundance may be characteristic of a given tumor, as for example, murine B16 melanoma is poorly infiltrated, even when modified to express ovalbumin [45 ]. Therefore, the extent of T cell infiltration in tumors is probably a result of multiple factors. Clearly though, whatever the level of TIL (even if substantial as for TCR transgenic mice), they are ineffective in eliminating primary tumors.
|
ROLE OF TUMOR BURDEN IN RESISTANCE TO T CELL-MEDIATED KILLING |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
|
WHY CAN "EARLY-STAGE" TUMOR BUT NOT "LATE-STAGE" TUMOR RESPOND TO VACCINATION OR ADOPTIVE TRANSFER? |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
As mentioned above, in some models (and in human cancers [20 , 48 , 49 ]), tumor infiltration by mononuclear host cells is appreciable [50 ], implying recruitment of antigen-specific T cells. In other models, CD8+ TIL infiltration is modest (ca., 1% of cells obtained from enzymatically digested tumor tissue [18 , 19 , 45 , 51 ]). Adoptive transfer also results in appearance of T cells in tumor tissue [32 ], although sometimes penetration of tumor margins is modest [50 , 52 53 54 ], perhaps reflecting the level of cognate antigen expressed, as suggested by Schreiber and colleagues [55 56 57 58 ]. In regards to this last point, Hanson et al. [50 ] made an interesting experiment, wherein they adoptively transferred antitumor, transgenic TCR T cells into mice bearing a "large" tumor (7 days of growth) and a "small" tumor (3 days of tumor growth). The T cells did not affect the growth of the larger tumor, but the smaller tumor was eliminated. This experiment crisply showed that antitumor T cells were not tolerized systemically in that the smaller tumor was cleared, and the larger tumor was resistant to killing. It is unknown why the larger tumor escaped killing but possibly may reflect poor penetration of T cells into the tumor. As the level of cognate antigen expressed is presumably equivalent between the two temporally distinct tumors, the notion that antigen expression levels impact on T cell recognition of tumor cannot explain why the large tumor was resistant to killing. A more likely reason is that the microenvironment of the large tumor is different from that of the small tumor or that the tumor itself changes during its growth, becoming resistant to T cell killing (see below).
In regards to the issue of whether antitumor T cells are localized in tumor, although there are examples of significant and poor infiltration [59 ], this issue has not been analyzed systematically, and a consensus position cannot be achieved. For example, even within an individual tumor, there can be heterogeneity of infiltration in terms of relative abundance of mononuclear cells as well as the precise pattern of localization (interstitial or parenchymal). After review of the literature, this subject can be generalized: tumor growth primes antitumor T cells which are recruited to a varying extent into tumors, and the number of tumor cells does not impact upon tumor accessibility by antitumor T cells.
Tumors may actively induce TIL apoptosis in situ [60 ]. If the ability to induce TIL apoptosis is acquired as a function of tumor growth, this mechanism may contribute to the failure of adoptively transferred antitumor T cells to eliminate tumor. The observation that some tumors express TRAIL or FasL is compatible with the notion of tumor cell-induced TIL deletion [61 ]. Such a notion is supported by data showing that some tumors are only modestly infiltrated by TIL but does not explain how some tumors can contain appreciable TIL (even if not robustly infiltrated [62 , 63 ]) or the existence of tumors, where the level of apoptosis in TIL is low [63 ]. Similarly, this possibility does not explain reports that patients have preternaturally high levels of circulating apoptotic (or proapoptotic) T cells, as the level of apoptotic T cells far exceeds the possible number of tumor antigen-specific T cells, and it is unlikely that the whole body complement of T cells circulates through tumor tissue, wherein they could potentially be induced to apoptosis by contact with tumor cells [64 65 66 ]. In addition, as recognition of cognate antigen by T cells often results in AICD, some percentage of apoptotic TIL in situ may reflect authentic TCR-mediated interaction of TIL with antigens expressed on the tumor cell. This explanation for TIL apoptosis has been validated definitively in human melanoma [67 ]. Nevertheless, in spite of poor mechanistic understanding of this phenomenon, it remains a possibility that certain tumors (e.g., head and neck and renal cell carcinomas [68 , 69 ]) secrete a factor/s (possibly FasL), which sensitizes T cells to AICD [69 ].
However, the issue of tumor expression of FasL is contentious, and several labs were unable to replicate the findings made by others [67 , 70 71 72 ]. There are at least two points confounding this issue: the questionable specificity of certain anti-FasL antibody reagents used to analyze primary tumor or cell lines and the possible contamination of primary tumor samples by infiltrating immune cells, which may contribute mRNA-encoding FasL, thus providing a "false positive" conclusion to reverse transcriptase-polymerase chain reaction analysis [73 ]. The problems with some anti-FasL antibodies include the possibility that tumor cells may express proteins, which are not FasL but contain an epitope recognized by the antibody [74 ]. Our lab has not examined FasL expression directly in tumors, although we have noted that CD8+ TIL are FasL+, indirectly supporting the notion that contaminating lymphocytes in tumor preparations are likely to provide FasL mRNA, thus contaminating "tumor" RNA analysis [51 ]. The observation that activated T cells (including antitumor T cells) are FasL+ is supported by others [67 ].
It is important that several papers purporting to demonstrate apoptotic TIL in tumor samples in fact do not definitively make that analysis, which requires double-staining of tumor samples with an anti-T cell reagent plus a marker of apoptosis, deoxyuridine triphosphate nick-end labeling (TUNEL), or for activated caspase [63 , 75 76 77 78 79 80 81 ]. In the small number of papers that performed double-staining of tumor samples, although identification of TUNEL+ cells is relatively unambiguous, definitive identification of TUNEL+ cells as T cells is complicated by the modest quality of T cell labeling in tumor tissue, the use of antibody reagents reactive with T cells and non-T cells (anti-CD45, for example), as well as by the potential for subjective interpretation of staining, which is attendant to immunocytochemistry. In addition, if tumor cells express FasL (postulated to be responsible for inducing apoptosis of TIL), then the second tumor innocula in the experiment of Hanson et al. [50 ] would be predicted to cause the death of adoptively transferred T cells and result in the growth of the tumor. However, the smaller tumor is killed in that experiment (and the first/larger tumor innocula grows), arguing against a role for tumor expression of FasL in tumor escape. (It is a formal possibility that tumor expression of FasL is delayed in transplantable tumors until the injected cells have grown in situ for several days, although this experiment has not been reported.) In addition, the possibility that putative, tumor-expressed FasL is released from tumor cells, and the soluble FasL causes T cell apoptosis needs to be considered in light of the finding that soluble FasL is significantly less efficient at triggering Fas than the membrane form [82 ]. Thus, the premise motivating the study of FasL expression in tumors is itself not firmly established, as data in support of the general notion of active tumor-induced TIL apoptosis are not compelling (see below), although it remains a formal possibility that some tumor types use such a mechanism of tolerance [63 ], and others may not.
|
INHIBITION OF TIL SIGNAL TRANSDUCTION BY TUMOR |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
|
We have recently shown that when stimulated by cognate tumor cells in vitro, TIL are unable to transmit TCR-mediated signals distal from p56lck activation, thus calcium flux, protein kinase C mobilization, and extracellular signal-regulated kinase activation are blocked [19 ], therein accounting for lytic deficiency, cell-cycle arrest, and defective cytokine secretion, i.e., the full panoply of TIL defects [20 , 38 ]. As mentioned previously, the proximal TIL signaling deficit is transient: In one well-studied model, signaling is restored after purification and in vitro culture, for as little as 2 h (N. Monu, unpublished observations). The rapid reversibility of lytic dysfunction is distinct from other described defects in CD8+ T cells, for example, deficient cell maturation (as has been shown for CD8+ T cells in human immunodeficiency virus infection [83 ], perhaps as a result of deficient IL-12 required for expression of granzyme B, as recently shown by Curtsinger et al. [84 ]) or hyporesponsiveness as a result of imprinting (as has been shown for lamina propria T cells [85 ]). The nature of the initiator that induces the transient proximal signaling block in TIL is described below.
Using a tumor model, which expresses a model antigen (adenovirus E1a), the Toes laboratory [86
] has reported that memory CD8+ T cells induced by peptide immunization are nonlytic in situ but recover lytic function in the absence of antigen, a characteristic similar to TIL. As TIL have been shown to be memory/effector cells on the basis of cell-surface phenotype, having mature (granzyme B+) granules and rapid kinetics of cytokine production upon antigen stimulation in vitro [18
, 51
], the observations of Toes and colleagues [86
] may be relevant to TIL unresponsiveness. However, in distinction to the phenotype of TIL, T cells in that model produce IFN- ex vivo, showing that cells are mature and antigen-specific, but the mechanism underlying the lytic dysfunction was not described, so it is unclear how the phenomenon of antigen-induced T cell unresponsiveness described by Toes and colleagues [86
] relates mechanistically to TIL lytic dysfunction.
The role of CD4+CD25+ Treg cells on effector-phase function
Another characteristic of tumors that may impact upon TIL function is the activity of CD4+CD25+ Treg cells, which can be isolated from a variety of tumors. In animal models, antibody depletion of Treg prior to tumor transplantation results in development of antigen-specific CTL and tumor immunity [87
]. Significant infiltration of tumor by Treg has been observed in murine models and humans, implying a role in tumor escape from immune-mediated elimination [88
89
90
]. Recently Zou [8
] reported that Treg isolated from ovarian carcinoma ascites could inhibit proliferation and CTL function of antigen-specific T cells in vitro (although the effect of Treg on signaling capability of responder CTL was not analyzed). The implication of this important work is that tumor antigen-specific Treg cells are elicited during tumor growth and that the priming and effector phases of antitumor immune response can be dampened, thought to contribute to tumor escape. The functional phenotype of CD8+ TIL (cell-cycle arrest and defective cytolysis and cytokine secretion) is in keeping with that induced by Treg coculture in vitro, inferring a causal association.
Another recent report showed in a murine model that Treg preferentially accumulated in tumors as a function of time of tumor growth [91
]. Tumor tissue was characterized by an anti-inflammatory environment (high levels of IL-4, IL-5, and IL-10 and low levels of IFN-, TNF, and IL-6), coincident with nonproliferating CD8+ T cells. Depletion of CD4+ T cells in vivo caused reversal of the cytokine expression pattern and rejection of late-stage tumors. Treatment of mice with anti-IL-10 or anti-transforming growth factor-ß (TGF-ß) resulted in inhibition of tumor growth, implicating these two cytokines in inhibition of tumor killing. Similar to the findings of Zou [8
], this compelling study strongly implies that CD4+CD25+ Treg restrict the effector phase of CD8+ T cell-mediated tumor immunity. Formal proof of a direct role for Treg in inhibition of antitumor CTL function in vivo awaits comparison of the specific signaling lesion in CD8+ TIL in situ with that induced by Treg in vitro.
The role of soluble factors on effector-phase function
A large variety of soluble factors is present in the tumor microenvironment, including those that can promote tumor growth, chemotaxis of host cells, activation of immune cells, and inhibition of immune cell function [92
]. As tumors are complex, adaptive systems, it is unclear how any given factor(s) which may have immunomodulatory ability, impact on TIL function in situ, especially as some factors can be both inhibitory and stimulatory, depending on the concentration and the context (e.g., IL-10 [93
]). When assayed in isolation in vitro, several factors [e.g., TGF-ß, IL-10, IL-6, prostaglandin E2, nitric oxide (NO)] have been shown to inhibit T cell functions.
Definitive assessment of the contribution of any single factor on defective TIL function in vivo is harder to establish, as in "knockout" mice, compensatory responses may obscure analysis, but one promising avenue of investigation is to determine the type of cell implicated in a particular inhibitory phenotype and then to identify the mediator of inhibition. In this regard, inhibition of CD8+ TIL effector-phase function in situ has been studied recently by the Bronte lab [31 ], who showed that TIL in prostate carcinoma fail to polarize cytolytic granules to the T cell membrane, a phenotype of purified TIL in conjugates with cognate tumor cells in vitro that we have previously shown to reflect TIL lytic dysfunction [18 , 19 ]. Treatment of tumor explants in vitro with a combination of arginase-2 and NO synthase (NOS) inhibitors reversed the TIL lytic granule localization defect coincident with increased tumor cell apoptosis, implying recovery of T cell cytolytic function. In addition, TIL in explants accumulated nitrotyrosinylated proteins, the levels of which were decreased upon treatment with inhibitors, and nitrosylation of tyrosine residues in key signaling molecules can restrict the activation of T cells [94 ]. Although specific, nitrosylated components of the proximal TCR signaling pathway were not identified in human prostate TIL, which would solidify the findings, the results of Bronte and colleagues [31 ] suggest that the activity of L-arginine-metabolizing enzymes is increased in the primary tumor, which negatively impacts on TIL signaling and therein prevents lytic function. The cellular source of inducible NOS and arginase in the work of Bronte et al. [31 ] was shown by immunocytochemistry to be the carcinoma cells, in keeping with the suggestion that enhanced arginase activity might be essential for elevated polyamine production, which is associated with tumor growth [95 ]. In support of this notion, several other human tumor cells have been shown to express elevated arginase [96 ] (but see below). In addition, increased arginase activity in tumors may suppress NO production by tumor-infiltrating macrophages, which in isolation, can have antitumor effects, thus potentially providing an additional level of tumor escape.
In a similar approach, Chen and colleagues [97 ] and Yu and colleagues [91 ] showed that CD4+CD25+ Treg could inhibit the in vitro, cytolytic function of antitumor T cells, which was mediated by TGF-ß and TGF-ß and IL-10, respectively, in that expression of a dominant-negative TGF-ß receptor in Treg or inclusion of blocking antibodies could prevent the inhibitory effects of coculture with CD4+CD25+ Treg. The signaling ability of responder T cells was not described in those publications.
The role of myeloid suppressor cells on effector-phase function
Myeloid cells which can inhibit T cell responses, are known to accumulate in tumors, and are termed myeloid-derived suppressor cells (MSC) [98
, 99
], which are heterogeneous (containing DC, macrophage, granulocytes, and myeloid cells), are usually incompletely differentiated and can be found in peripheral lymphoid organs, blood, or tumors [100
, 101
]. MSC, purified from tumor tissue or spleens of tumor-bearing mice, although likely containing several different cell types, are defective in T cell priming, which may restrict induction of tumoricidal immune responses [98
, 99
]. The systemic levels of MSC appear to be related to tumor burden, as upon tumor resection, MSC numbers decline dramatically [102
].
MSC can potentially impact on T cell function in a variety of ways: inhibition of antigen-dependent cytokine secretion in vitro [103
], induction of apoptosis in activated T cells in vitro [104
], secretion of a variety of factors having immunomodulatory properties (e.g., H2O2, TNF, NO, TGF-ß), as well as production of enzymes that modulate amino acid metabolism [indoleamine 2,3-dioxygenase (IDO) and arginase], which have been shown to be associated with peripheral tolerance [98
, 99
, 105
, 106
]. MSC, isolated from tumors, produce high levels of arginase activity (ca., 20x10–5 ug urea/cell/6 h, compared with 0.5x10–5 ug urea/cell/6 h for MCA38 tumor cells, N. Monu, unpublished observations), which may induce local depletion of arginine, possibly leading to cytostasis of TIL [107
]. However, if local, inadequate supply of arginine (or tryptophan) causes TIL unresponsiveness, it is unclear how other proximal cells (tumor and stromal cells) appear to be unaffected by deprivation of an essential amino acid. In regards to a potential effect of MSC production of arginase upon TIL function, Ochoa and colleagues [107
] have shown that a subpopulation of tumor MSC produces high levels of arginase and not H2O2 or IDO, which inhibits proliferation of non-TIL T cells in vitro. (In addition, tumor MSC were shown to produce high levels of the cationic amino acid transporter, which may explain why MSC can metabolize arginine at high levels in an arginine-deficient environment.) Loss of cell-surface CD3
and TCR
was observed, coincident with the proliferation defect, suggesting that arginine depletion caused the proliferation deficiency via down-regulation of key components of the proximal TCR signaling machinery. A causal relation between MSC production of arginase and antitumor T cell dysfunction was implied further by biochemical inhibition of arginase in vivo, which resulted in diminished tumor growth rate.
Tumor MSC can also produce significant levels of NO in vitro after stimulation by IFN- and/or TNF [104
], which in combination with enhanced arginase activity, may lead to release of highly reactive peroxynitrites that can inhibit T cell signal transduction by covalently modifying susceptible tyrosine residues in TIL signaling enzymes [108
], as was intimated by the work of Bronte and colleagues. Furthermore, MSC release of NO has been shown to induce apoptosis of activated T cells in vitro [104
]. As infiltration of tumors with MSC is a common feature of many tumor types, and MSC produce a variety of pharmacologically active substances [98
, 99
], collectively, the abundant literature about MSC suggests that these cells likely play a significant role in alteration of antitumor immune responses dependent on inhibition of TIL signal transduction, leading to loss of function and/or apoptosis.
However, the inhibitory effects on TIL of tumor cells (Bronte [108
]) or MSC (Gabrilovich [103
] and Ochoa [107
] labs) have been demonstrated for those T cell functions which require several days to be manifest in vitro proliferation or induction of apoptosis. In contrast, the lytic dysfunction of CD8+ TIL can be reversed quickly following purification, implying a different and more rapid kinetics of induction [18
, 19
]. We have recently developed an assay which permits simultaneous analysis of lytic function and signaling in freshly purified (nonlytic) or purified (lytic) TIL (Fig. 3
). TIL are purified, briefly cultured in vitro (6 h, which restores lytic function), and then cocultured, in contact or separated by a porous membrane with purified cells obtained from the primary tumor, tumor-associated macrophages, or the tumor cell line.
|
In other experiments, we have shown that when conjugated with cognate tumor cells in vitro, signal transduction in nonlytic TIL is blocked, such that p56lck is inactivated rapidly (phosphorylation of Y394 is inhibited, unpublished observations), -associated protein 70 is not activated [19
], and TCR is phosphorylated only transiently (D. Schaer, unpublished observations). In sum, the phenotype of nonlytic TIL appears to result from tumor-induced blockade of proximal TCR-mediated signaling, which prevents effector-phase function. The biochemical manifestation of defective proximal TCR signaling in nonlytic TIL is coincident with colocalization with p56lck of one phosphatase whose substrate is p56lck, SH-2-containing tyrosine phosphatase-1 (Shp-1) [19
]. In addition, Shp-1 in nonlytic TIL is tyrosine-phosphorylated (implying its activation) but is not in lytic TIL (N. Monu, unpublished observations). Those observations collectively are compatible with the notion that TCR signaling in nonlytic TIL is inhibited as a result of the activity of an inhibitory phosphatase, possibly Shp-1, and that p56lck function in lytic TIL is not down-regulated; thus, signaling and lytic function are intact.
Inhibitory phosphatases are recruited from the cytoplasm to their membrane-associated substrates by interaction with adaptor proteins, which are localized in the plasma membrane. Adaptor proteins require phosphorylation on specialized cytoplasmic motifs (immunocreceptor tyrosine-based inhibitory motif) to attract and activate the phosphatase [109 ]. Although we have yet to eliminate the possibility of involvement of other candidate phosphatases in addition to Shp-1 (e.g., proline, glutamic acid, serine, threonoine domain-enriched tyrosine phosphatase) or to identify definitively the specific adaptor responsible for TIL signaling inhibition, other labs have implicated involvement of a variety of different adaptor proteins in regulation of TCR-mediated signaling [110 111 112 113 114 115 116 ]. We are, at present, refining our knowledge of the role of Shp-1 adaptor proteins and expression of their counter ligands on MCA38 tumor cells in the induction of defective signaling in nonlytic TIL.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
A role for MSC or CD4+CD25+ Treg in dampening the expansion of tumor-specific T cells or inhibition of de novo priming of additional antitumor T cell clones is plausible and compatible with the observation that growth of tumor inhibits priming. However, patient antitumor T cells can be primed by vaccination, which at first consideration, may reflect the possible absence of MSC or Treg from the site of priming upon vaccination. However, MSC and Treg are isolable from LN and spleens of tumor-bearing animals, arguing that the scenario is unlikely. Instead, perhaps successful vaccination overcomes or can resist the potential inhibitory effects on priming of MSC and/or Treg. A role for tumor MSC or CD4+CD25+ Treg in inhibition of the TIL effector phase is certainly possible, and definitive proof awaits analysis of TIL signaling defects induced by these potent regulatory cell types. As antigen-specific antitumor T cells are commonly found in tumors, whatever restricts priming in the unmanipulated host (such as weak danger signals?) can be at least partially overcome. Ultimately, both facets of tumor-induced immune response restrictions, priming and effector-phase functions, will need to be ameliorated before immunotherapy of cancer can succeed.
Received November 3, 2005; revised December 9, 2005; accepted December 12, 2005.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONTUMOR ESCAPE FROM T...ROLE OF TUMOR BURDEN...WHY CAN "EARLY-STAGE" TUMOR...INHIBITION OF TIL SIGNAL...CONCLUSIONREFERENCES |
|---|
secretion by circulating CD8 T lymphocytes: implications of a novel approach for T cell monitoring in infectious and malignant diseases J. Immunol. 166,7634-7640
ß T cells J. Immunol. 167,6165-6170 chain expression and apoptosis in CD3+ peripheral blood T lymphocytes of patients with melanoma Clin. Cancer Res. 7,947s-957s[Medline]
