| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print June 3, 2003
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Department of Internal Medicine, Division of Hematology and Oncology, Laboratory of Molecular Cytology, University of Innsbruck Medical School, Austria; and
Tyrolean Cancer Research Institute, Innsbruck, Austria
1Correspondence: Laboratory of Molecular Cytology, Department of Internal Medicine, Division of Hematology and Oncology, Innsbruck University Hospital, Anichstrasse 35, A-6020 Innsbruck. E-mail: richard.greil{at}uibk.ac.at
 |
ABSTRACT |
|---|
Key Words: lymphoma • tumor counterattack • novel therapeutic strategies
|
INTRODUCTION |
|---|
, perforin, and granzyme. The pivotal role of the Fas/FasL system in regulating normal B and T cell function [1
], suppression of autoimmunity [2
], control of infection [3
], and immune surveillance [4
] has been extensively analyzed. As a result of its dual role, namely, in self-control of T cell expansion and in killing of virally infected or neoplastically transformed target cells, alterations of the Fas/FasL system are at the interface among autoimmunity, immune evasion of cancer cells, and immune deficiency, which are all well-known features of neoplastic diseases, in particular, of the lympho-hematopoietic system. TRAIL has been less-extensively investigated than FasL, but its preferential cytotoxicity against neoplastic cells and its low toxicity in normal tissues warrant further studies on its function and therapeutic potential in cancer. It is the aim of this review to focus on the role of Fas and TRAIL-R signaling in normal lymphopoiesis and their significance in the pathophysiology of lymphomas. In addition, current concepts regarding the therapeutic targeting of these two death receptors will be discussed.
|
STRUCTURE AND FUNCTION OF Fas DEATH RECEPTOR/LIGAND SYSTEM |
|---|
|
In conclusion, the specific composition of the humoral micromilieu and/or the triggering of specific membrane receptors by components of the cellular microenvironment modify a relationship between FasL+ effector and Fas+ target cells. These interactions are operative in normal and neoplastic hematopoiesis. However, in the latter, the oncogenetic transformation process may functionally alter molecules representing physiological and essential constitutents of an intact Fas signaling pathway and thus induce intrinsic resistance to Fas.
|
CONSTITUTION AND EXPRESSION OF TRAIL AND ITS RECEPTORS |
|---|
|
B pathway but is unable to transduce proapoptotic signals, as it lacks a part of the cytoplasmic death domain [44
] (reviewed in ref. [35
]). Finally, TRAIL may bind to osteoprotegerin [45
], a soluble TNF-R homologue that inhibits osteoclastogenesis by binding to the osteoprotegerin ligand. As a result, bone density increases in vivo [45
]. By this interaction, osteoprotegerin inhibits TRAIL-induced cell death in the lymphatic cell system and enables TRAIL to interfere with bone formation [45
].
Early investigations explained the different sensitivity of normal versus neoplastic cells and the differential sensitivity of various tumor types to TRAIL signaling by the various patterns of expression of full-length versus decoy receptors with or without activation of the antiapoptotic NF-B pathway [40
]. The critical review of data mentioned above and presented in Table 2
sheds some doubt on such a simplified view for the following reasons: There is a broad overlap in expression profiles of pro- and antiapoptotic receptors (see Table 2
), and the correlation between their expression levels and sensitivity of cells to TRAIL is weak in normal dendritic cells (DCs) [54
], tumor cell lines [35
], and many native human tumor cells (Table 2) . Rather, receptor/DISC composition, equipment with caspases and their inhibitors, as well as external influences of the microenvironment cooperate in regulating TRAIL sensitivity in the various tissues.
|
REGULATION OF NORMAL LYMPHOPOIESIS BY DEATH RECEPTOR SYSTEMS |
|---|
In contrast to the Fas system, upon activation, thymocytes up-regulate TRAIL and TRAIL-R1 and -R2, thus sensitizing thymocytes to TRAIL. Although suggestive of a contribution of the TRAIL system to antigen selection of thymocytes, this could not be proved, leaving open the physiological role of TRAIL circuits in the thymus. TRAIL is also expressed on activated but not resting murine CD4+ and CD8+ T cells [62 ], but the specific receptor is not increased, which enables peripheral T cells to resist autocrine suicide via TRAIL under conditions of activation [63 ]. Although FasL-bearing T cells are also involved in the control of the immune system stimulated for expansion during inappropriate antigen selection, the central physiological targets for FasL- and TRAIL-bearing effector cells are virally infected and cancer cells.
B cell development
The Fas/FasL system is a central player in the negative-selection process of normal B cells, inducing antigen-specific B cell responses and inhibiting clonal expansion of autoreactive cells. lpr and gld mice are characterized by expanded B cells, increased pool of abnormal T cells, lymphadenopathy, splenomegaly, and production of double-stranded DNA antibodies [2
]. Fas plays an important role in the antibody-selection process during germinal center passage of B cells. Within this specific environment, preformed oligomeric, Fas-associated DISC complexes are expressed in B cells, but the interruption of the death pathway by their association with the inhibitory FLIP molecule rescues germinal center B cells from their rapid demise [64
]. Outside the germinal center, FLIP rapidly dissociates from the preformed DISC and allows death to occur, unless its continuous association is triggered by two different stimuli, namely ligation of the B cell receptor and contact with CD40L+ T cells or contact with follicular DCs (reviewed in ref. [65
]). The rapidity of the Fas/FLIP interaction therefore seems to serve as an emergency switch for apoptosis during the selection process. It prevents the survival of low-affinity B or plasma cells but in case of Fas resistance, allows prolonged recruitment of germinal center cells into the B cell memory pool [66
]. Although up-regulation of Bcl-2 or BcL-XL also confers resistance to Fas signaling in B cells, their expression after CD40L binding takes longer, suggesting that they play a role during memory or plasma cell differentiation rather than during the antigen-selection process.
After their maturation to plasma cells, B cells may still be partly under the control of the Fas/FasL system, as at least in B-CLL, hypogammaglobulinemia results from binding of tumor-associated mFasL to Fas-sensitive plasma cells [67 ] (Fig. 1 ). In contrast, within the mucosa-associated lymphoid tissue (MALT) of the gastrointestinal tract, plasma cells express FasL, pointing to their role not only as target but also as active effector cells of the immune system [74 ]. The expression of TRAIL has been investigated in the murine system, where B220+ mouse B cells are intrinsically TRAIL+. The similar expression pattern of FasL and TRAIL in peripheral B and T cells points to a complementary effector function in the regulation of these cells.
|
|
THE ROLE OF DEATH RECEPTORS IN DC FUNCTION AND DEVELOPMENT |
|---|
|
DEATH RECEPTOR/DEATH RECEPTOR LIGAND SYSTEMS IN LYMPHATIC NEOPLASIAS |
|---|
B cell malignancies in Fas/FasL-deficient mice
Although fas-deficient mice usually show only background incidence for the development of B cell lymphomas, the prolonged observation of C3H-gld and BALB-gld mice revealed a 28% and 57% lymphoma rate, respectively [85
]. Transfection of FLIP, an inhibitor of caspase-8, into murine A 20 lymphoma cells and their subsequent injection into an immunocompetent-recipient mice strain led to the development of rapidly growing tumors resisting control and rejection via death receptor ligands of T cells [86
]. The consequences of an altered Fas signaling pathway are aggravated by further T cell deficiencies [87
]. These data established inhibitors of death-receptor signaling as tumor-progression factors.
Cancer incidence under Fas-deficient conditions in humans
Autoimmune lymphoproliferative syndromes (ALPS) are inherited human diseases with increased incidence of autoimmunity and neoplasias observed also in gld and lpr mice. The ALPS phenotype is associated with inherited mutations in the Fas (type Ia) or the FasL gene (type Ib) [88
]. In inbred lpr and gld mice, essentially the same disease results from homozygous, recessive mutations in the genes encoding Fas and Fas ligand, respectively. However, in ALPS type Ia, individuals are more often heterozygous for mutant Fas alleles. By an unknown mechanism, the Fas protein encoded by the mutant allele dominantly interferes with the function of the wild-type protein [89
], causing a severe defect in apoptosis [90
]. The localization of mutations/deletions within the death domain of the Fas gene varies widely (between nucleotide positions 915 and 1123) in the cases reported by Martin et al. [89
] and so does the clinical spectrum of the disease manifestation. Although lymphadenopathy and splenomegaly are rather constant features, the number of cell types binding autoantibodies and the severity of their destruction differ substantially [89
].
In ALPS type II, it is assumed that a similar, although more severe, clinical phenotype is caused by an undefined, inherited gene defect in the absence of mutations in Fas or FasL genes. Finally, ALPS type III is a more heterogeneous group, as yet without any defined genetic cause. An increased incidence in lymphoma development has been described in one family with inborn genomic fas mutations [91 ] and ALPS [92 ]. Delayed onset and manifestation of the neoplastic disease at a mean age of 28 years [93 ] suggest a need for acquisition of further oncogenetic events particularly in B cells. A recent survey of families with ALPS revealed a significantly increased risk of Hodgkin’s (up to 51-fold) and Non-Hodgkin’s lymphoma (NHL; up to 14-fold) and a broad range of histological types of NHL including one case of T-NHL [93 ]. In all patients, tumor cells retained the heterozygous fas mutations and demonstrated severe disturbances in Fas-mediated killing. Female patients without fas mutations but severe disturbances in Fas signaling and/or ceramide-induced T cell death (autoimmune lymphoproliferative disease or type III disease) displayed an increased incidence of (standardized incidence ratio=1.87) and mortality (standardized mortality ratio=2.75) from cancer, but there was no particular preponderance of lymphomas [94 ].
The FasL counterattack hypothesis
Neoplastic T cell lymphomas are capable of expressing FasL and of killing tumor-specific, Fas-sensitive, cytotoxic T cells as a result of their lineage derivation [95
]. However, a broad range of tumor types acquire FasL expression [96
97
98
] and the ability to escape the immune system, not only by modification of Fas sensitivity but also by defeating sensitive immune-effector cells (Fig. 1)
. The Fas counterattack hypothesis has been challenged, mostly on technical grounds [99
]. Other reasons for dissatisfaction with the hypothesis are observations that transfection and overexpression of allogeneic or xenogeneic FasL in grafted pancreatic islets do not kill autoreative T cells in diabetic animals but rather attract granulocytes, which destroy the graft [100
], and that transfection of tumor cells with FasL can induce tumor regression by tumor cell suicide [101
] or by the induction of an inflammatory response mediated by granulocytes and monocytes [101
]. There is, however, an appreciable and growing body of evidence supporting the Fas counterattack hypothesis, particularly based on observations made in native human tumor material without genetic modifications of the system [4
, 80
].
|
THE ROLE OF TRAIL IN TUMOR CONTROL |
|---|
-dependent pathway, and this establishes protection against metastatic infiltration [104
]. (ii) DCs may be crucial for establishing antitumor immunity. Surprisingly, DCs may not only serve as antigen-presenting machinery but turn into effector cells with killing capacity against papilloma virus-infected, preneoplastic epithelial cells [105
]. IFN- or IFN- stimulation of CD11c+ blood DCs and IFN-ß activation of monocyte-derived [106
] or CD34+ cell-derived [107
] DCs lead to expression of TRAIL and turn DCs into cytotoxic effector cells against tumor cells [106
, 107
]. The killing capacity of DCs via TRAIL is absent in blood precursor (pre) DCs but is acquired by them at the CD11-positive stage [81
]. By killing tumor cells via TRAIL, DCs may subsequently serve a second function by taking up apoptotic bodies of tumor cells, processing, presenting them to CD8+ cytotoxic T cells, and thus, inducing a broader antitumor immunity. (iii) Certain T cells might use TRAIL to kill tumor cells. In fact, CD4+CD8dim+ T cells isolated from cutaneous T cell lymphomas are cytotoxic against autologous tumor cells via a TRAIL-dependent mechanism [108
]. Also, certain CD4+ cells express TRAIL after stimulation with the immune modulator -galactosylceramide and thus become capable of killing acute myeloid leukemia (AML) cells [109
]. Inversely, disturbances in the TRAIL/TRAIL-R system may favor the expansion of the malignant clone: (i) Mutation of TRAIL-R1 and -R2 renders neoplastic B cells insensitive to TRAIL and occurs in 6.8% of 117 NHL investigated [110 ]. (ii) Neoplastic cells may aberrantly acquire the capacity to produce osteoprotegerin [111 ]. Whether this phenomenon interferes with the TRAIL-mediated control of tumor is not yet clear. (iii) Like FasL, neoplastic lymphatic cells may express TRAIL, allowing them to kill T cells and suppress the immune system [112 ]. The counterattack hypothesis about TRAIL, however, is less well investigated than about FasL.
|
DEATH RECEPTOR SYSTEMS IN SPECIFIC LYMPHATIC NEOPLASIAS |
|---|
T-ALL
Despite expression of Fas in the majority of primary T-ALL cases, these tumor cells are constitutively resistant to Fas-induced apoptosis [114
]. This is in contrast with tumor cells of patients with adult T cell leukemia, which expressed Fas and died after Fas ligation [115
]. Mutations within Fas, detected in rare cases of both types of native T cell malignancies, may account for Fas resistance [115
, 116
].
Inactivation of the Fas signaling pathway may occur more often in T cell leukemias than alterations in the structure of Fas and may contribute to Fas resistance. A comparison of neoplastic lymphoblasts of pediatric patients with ALL at diagnosis and at the first relapse showed a down-regulation of Fas and caspase-3, as well as an up-regulation of Bcl-2 expression in the latter [117 ]. Deregulation of the transaldolase 1 stem cell leukemia expression commonly occurs in T-ALL [118 ] and when overexpressed, contributes to the development of T cell malignancies [119 ] and to the inhibition of apoptosis induced by Fas or cytotoxic agents [120 ].
Peripheral T/NK cell neoplasias
Peripheral nasal T/NK cell lymphomas coexpress Fas, FasL, and perforin/granzyme [121
]. This is in good correlation with the observation that in NK cell lymphoma, mFasL was constitutively expressed, and sFasL was shed, whereas NK cells of healthy volunteers expressed FasL only upon activation [122
]. As the nonneoplastic cells within these tumors were often Fas+, neoplastic cells might cause apoptosis and necrosis of normal tissue constituents considered typical of this disease. In fact, apoptotic cells typically surrounded the proliferating tumor cells [121
]. Increased levels of sFasL observed during active phases of the disease in NK cell lymphomas usually correlated with signs of liver damage, declined during successful radiotherapy, and recurred during relapses of disease [123
]. In CD3+ large, granular lymphocytic leukemias, tumor cells expressed Fas and FasL to a similar extent as normal, activated CD3+ cells [124
] but resisted Fas-induced cell death in most cases. Fas resistance was not a result of mutations but rather associated with a failure to respond to CD3 cross-linking and a more aggressive course of the disease. In cutaneous T cell lymphomas, Fas expression was observed in all of the favorable CD30+ cases but only in one-third of the tumor stages of Mycosis fungoides and in only 22% of CD30- cases [125
]. Down-regulation of Fas paralleled disease progression from plaques to tumor, and FasL was present on tumor cells in 82% of cases. These findings favor a role of membrane-bound or soluble FasL in tissue destruction and morbidity and point to a sequence of transformation in which a loss of Fas sensitivity and up-regulation of FasL occur on neoplastic T cells. Indeed, T cell lymphomas and syngeneic tumor-specific T cells may coexpress Fas and FasL, giving rise to a bidirectional interaction and killing in both cell types [95
]. Up-regulation of FasL and intrinsic protection of tumor cells against Fas signaling may therefore confer a survival advantage on tumor cells by enabling them to escape from and to defeat specific cytotoxic T cells.
B cell neoplasias
Numerous observations have implicated the Fas/FasL system in the regulation of B cell homeostasis in many ways. Immunohistochemical analyses of the normal B cell compartments demonstrate positivity of secondary germinal centres of lymph nodes for Fas [126
, 127
] and caspase-3 [128
] but no Bcl-2 expression [129
]. Thus, the physiological B cell-immune reaction should still be under the control of Fas. An intact Fas signaling pathway may therefore be considered as tumor-suppressive.
|
MECHANISMS OF Fas RESISTANCE IN SPECIFIC B CELL NEOPLASIAS |
|---|
|
Mutation of the receptor
Mutations within the fas gene were identified in 11% of 150 cases of NHL examined [139
]. Most of the mutations found were believed to interrupt a functional signaling cascade along the Fas pathway. The highest frequency of fas mutations was seen in MALT–NHL, and almost all patients with mutations showed extranodal disease (Table 3)
. Strikingly, most patients with fas mutations showed clinical features of autoimmunity consistent with the expectations from murine lpr models and from patients with ALPS who showed an accumulation of CD4-/CD8- T and autoreactive B cells and features of autoimmunity [150
]. Mutations within the cytoplasmic region of the fas gene responsible for successful signal transduction were also shown to occur in 10% of myeloma patients [144
] but were quite rare in other hematopoietic tumors [134
].
Intracellular equipment with constituents of death receptor signaling
Other strategies of B cell neoplasias to evade lethal Fas signals are related to alterations in constitutents of the Fas signaling cascade. Oligonucleotide microarray analyses of mantle cell lymphoma show a more than tenfold down-regulation of the FADD adaptor molecule [151
]. The involvement of members of the Bcl-2 family in the signaling pathway of Fas may vary depending on the cellular background examined (i.e., type I vs. type II cells; ref. [9
]). Bcl-2 is expressed at high levels in myeloma cells, but unlike Bax [148
], its expression does not correlate with Fas sensitivity [148
, 152
]. BcL-XL levels were increased in myeloma cells [153
], and IL-6-induced up-regulation contributed to resistance to apoptosis [154
]. In the vast majority of Hodgkin cases [127
], the neoplastic Reed-Sternberg (RS) cells express Fas [70
, 155
] (Table 3)
. The intact equipment with the constituents of the apoptotic machinery and histological finding of signs of apoptosis in some RS cells point to Fas sensitivity at least in some cases. The survival of most RS cells may be a result of the down-regulation of caspase-3 and -8 [156
] or an altered cell-cycle behavior [157
] with a short pause in the usually Fas-sensitive, late G1 or G1/S transition phase [158
].
Soluble factors and cytokines
Investigations about the role of cytokines on Fas sensitivity, partially preserved in myeloma, pointed to the presence of protective factors in sera of myeloma patients and normal individuals [143
], and paracrine IL-6 was identified [159
, 160
] as one of these cytokines. Also, autocrine IL-6 proved protective against anti-Fas mAb-induced apoptosis [161
]. Paracrine and autocrine IL-15 [161
] also counteracted Fas signaling and protected myeloma cells from Fas-induced apoptosis. As IL-15 is usually considered as a factor that promotes growth, differentiation, and activation of T and NK cells acting against tumor cells, these data suggest that acquisition of a T cell growth and survival factor by tumor cells helps them to evade the immune system. Different results were reported for therapeutically applied cytokines such as IFN-. By interference at the level of protein kinase C, this cytokine mediated resistance to Fas-induced cell death when applied simultaneously or before cross-linking of Fas [162
], whereas prolonged incubation with IFN- [162
] or IFN- [163
] induced Fas up-regulation and rendered myeloma cells susceptible to Fas-mediated apoptosis [163
]. These results suggest that the protective versus survival effects of cytokines on the modification of Fas/FasL interaction depend on the concentration, duration of stimulation, and thus, on the humoral microenvironment in vivo. In addition, some myeloma cell lines are sensitive to cross-linking of Fas by anti-Fas mAb but resist apoptosis by FasL+ allogeneic T cell lines [97
]. This points to a specific self-protection during cell/cell contact, which at least in part could be a result of or augmented by the release and autocrine action of protective cytokines.
|
THERAPEUTIC INTERVENTIONS TARGETING THE DEATH RECEPTOR PATHWAYS |
|---|
Interventions targeting Fas or TRAIL-R signaling cascades of tumor cells
Attempts to increase the sensitivity of tumor cells along the Fas pathway may in principle aim at two different aspects: namely resensitization to FasL+ or TRAIL+ effector cells and simultaneously, to cytotoxic drugs, as the signaling cascades used are partially overlapping. Briefly, the strategies mentioned below might be followed.
Up-regulation of Fas on tumor cells
Previous experiments have demonstrated that the degree of Fas expression is directly correlated with [148
] or even represents the major determinant of Fas sensitivity in myeloma cells [166
], and a number of diverse approaches have successfully been applied to increase the sensitivity in many tumor types [163
, 167
168
169
]. At least in myeloma, this sensitization was probably a result of simultaneous up-regulation of proapoptotic molecules such as Bax [148
]. Although up-regulation of Fas via IL-12 sensitized CLL cells to FasL-bearing, autologous T cells [170
], up-regulation of Fas was not uniformly associated with sensitization. More complex maneuvers including stimulation of CD40 were required for sensitization of CLL cells to Fas up-regulated via IL-15 [171
]. In addition, it has to be stressed that the in vivo effects may be much more complex, as a given cytokine such as IFN- had completely opposite effects on myeloma cell sensitivity to Fas in the short and long run [162
, 163
]. Moreover, the sequence of cytokine signaling may be important to predict their effect on Fas-induced cell death. Indeed, auto- or paracrine IL-6, while inhibiting Fas-induced myeloma cell death, significantly increased Fas expression and Fas sensitivity of these cells following their treatment with IFN- [172
]. Similarly, the effect of a given cytokine may be completely opposite, depending on the tumor-specific maturation stage. IL-15 might up-regulate Fas and sensitize cells to Fas-induced apoptosis when acting in concert with CD40L in CLL [171
] and at the same time, protect myeloma cells from Fas-induced cell death [161
]. Therefore, knowledge of disease-specific tumor biology must guide attempts to up-regulate Fas signaling by using cytokines such as IL-12 [170
, 173
] and antibodies counteracting cytokines such as IL-6, IL-15 [161
], or IL-4 and/or histone deacetylase inhibitors [174
].
Resensitization to Fas signaling by CD40 stimulation
Given the paramount importance of the CD40L/CD40 system for the induction of specific T and B cell responses of the FasL/Fas system for limiting inappropriate stimulation and the expansion of autoreactive T cells, it is not surprising that cells carrying components of these two systems interact within the normal germinal center and that signaling events along these two receptor systems are interdependent or crossing. During the transit through the germinal center, B cells develop profound changes in apoptosis regulation [175
] and so do their relevant neoplastic counterparts. It is therefore not surprising that interactions between the CD40/CD40L and the Fas/FasL system might be heterogenous according to the relevant differentiation stage of neoplastic B cells, and this might impact on the capacity of the T cell system to control the tumor. For example, CLL cells usually express extremely low amounts of Fas, and CD40L stimulation of cultured B-CLL cells leads to up-regulation of Fas, but tumor cells continue to resist death as a result of a simultaneous increase in the expression of FLIP [176
] or survivin [177
]. Autologous serum of patients rescues cultured CLL cells from death, an effect, which is in part a result of sCD40L and can be counteracted by anti-CD40L antibodies [178
]. The antiapoptotic effect of CD40 signaling may partially be explained by its potential to stimulate secretion of growth-promoting and apoptosis-inhibiting cytokines such as IL-6, as demonstrated in myeloma [179
]. Besides protection from Fas-induced cell death, CD40 signals inhibited induction of cell death by fludarabine in CLL cells [180
] and by doxorubicin in NHL cell lines [181
]. In addition, a proliferative potential of CD40L has been demonstrated in mantle cell lymphomas [182
]. These results should discourage the use of CD40 stimulation as a therapeutic tool in attempts to resensitize cells to Fas. However, mantle cell lymphoma, CLL and ALL cells, cell lines [183
], and native cells of B cell NHL [184
] as well as the majority of Burkitt lymphoma cell lines [167
] not only up-regulated Fas but also became sensitive to FasL when stimulated by CD40L. Exploitation of this mechanism with a therapeutic intent showed efficacy in elegant xenograft models of high-grade lymphomas [185
, 186
] using oral CD40L gene therapy or agonistic mAb.
These somewhat contradictory results might be explained by the fact that in vivo effects of CD40L are of a pleiotropic nature. In fact, the induction of a specific and successful immune response to CLL cells might be initiated by a transfection of CD40L into CLL cells, which are then retransferred into the bloodstream of the patient [187 ]. The increase in immunogenicity might in part be a result of induction of members of the cytokine network such as IL-15, which like CD40L, up-regulates Fas and when acting in concert with CD40L, sensitizes these cells to Fas [171 ]. In addition, up-regulation of CD80 and CD86 on antigen-presenting cells may induce effector mechanisms of T cells against the neoplastic clone, which might not be observed when signaling interferences between Fas and CD40 are studied only in vitro.
Although some of these results seem encouraging, the pleiotropic effects of the CD40/CD40L interaction and the significant differences observed between high-grade lymphoma models (beneficial) versus subtypes of low-grade lymphoproliferative diseases (reviewed in ref. [188 ]) suggest that caution must be exercised in early in vivo trials of CD40L therapy with or without attempts at Fas-resensitization [189 ]. These should be accompanied by rigorous examination of the biological basis of the results observed. Keeping these caveats in mind, a phase I clinical trial with recombinant CD40L (Avrend®, Immunex, Seattle, WA) was performed only on patients with intermediate and high-grade NHL. It demonstrated an 11% response rate in this disease when used as a single agent [190 ]. No evidence of disease acceleration was observed during treatment [189 ]. These results encourage and justify use of this drug as a monotherapy in high-grade NHL in phase II trials and in combination with death receptor agonists in preclinical experiments. In low-grade NHL and CLL, the complex induction of sensitivity against FasL+ effector cells may be more reasonable than direct use of CD40L.
Direct intervention by death receptor agonists and modification of the intracytoplasmatic signaling machinery
In principle, therapeutic exploitation of signaling along the Fas and TRAIL pathway may be accomplished by several means, namely, by the use of recombinant ligands of the receptor or by designing specific agonistic antibodies (Fig. 2
). Both approaches are currently being investigated. The choice of an adequate technical tool as well as the attempt to specifically address the Fas or the TRAIL pathway may be determined by side-effects and toxicity profiles of the tools and the relevant degree of efficacy in a specific lymphatic neoplasia.
|
R)-bearing effector cells. Normal chondrocytes remained unaffected [194
]. Given the different roles of Fas and TRAIL in the regulation of physiologic B and T cell expansion and of their aberration in lymphoma genesis, this approach deserves further preclinical development.
Recombinant human (rhu) TRAIL.
TRAIL is a promising candidate for therapy of some hematopoietic tumors, as its agonistic receptors DR4 and DR5 are expressed in a wide variety of tumor cells, rendering them sensitive to its proapoptotic action, and normal tissues, in particular, hematopoiesis, are usually protected, despite the somewhat controversial results in the erythroid lineage [38
, 76
, 77
, 195
] (Table 1) . Furthermore, Apo2L/TRAIL effects can be augmented by chemotherapeutic agents and ionizing irradiation [38
, 195
196
197
] without increasing toxicity on normal hematopoiesis [77
]. Synergism with cytotoxic agents may also be indicated by the observation that treatment of AML M3 cells with retinoic acid induced TRAIL expression, which was involved in paracrine killing of leukemic cells and in elimination of retinoic acid-resistant leukemic cells [198
]. Similarly, IFN--induced autocrine suicide of lymphoma cells proceeded via TRAIL [50
] (Table 2)
. Myeloma seems to be a candidate disease for treatment with TRAIL, particularly as the high intrinsic expression levels of Bcl-2 and BcL-X do not interfere with sensitivity. In fact, ectopic Bcl-2 overexpression failed to block TRAIL-induced apoptosis in cell lines [199
]. TRAIL sensitivity was further enhanced by treatment with IFN- and -ß [52
], NF-B, or proteasome inhibitors and doxorubicin [53
]. These substances are part of clinical practice or tested in early clinical trials (see below and Table 4
). Results were confirmed in native material and in xenotransplanted nu/xid/bg mice [53
]. TRAIL has also been tested in vitro in a limited range of other hematological neoplasias, and sensitivity of tumor cells has been reported to a lower degree (Table 2)
.
|
Agonistic anti-TRAIL-R antibodies.
Two agonistic mAb have been developed, which selectively bind to TRAIL-R1 and TRAIL-R2 on hematopoietic and solid tumors (Table 4) . In contrast to rhu TRAIL and the physiological ligand to TRAIL-R1 and -R2, the agonistic anti-TRAIL-R1 mAb does not bind to DcR1 and DcR2 or the soluble receptor osteoprotegerin, all of which act as decoy receptors. In addition, the antibody has a significantly longer serum half-life as compared with the natural ligand. These features make agonistic antibodies an interesting and attractive tool for further development and in fact, patients are being recruited for phase I trials with agonistic anti-TRAIL-R1 mAb (Table 4)
, and agonistic anti-TRAIL-R2 mAb is under advanced preclinical investigation (Table 4)
.
Use of FasL and TRAIL for graft-versus-tumor (GvT) effects.
The potency of the immune-effector mechanisms is apparent in the clinical manifestation of allogeneic hematopoietic cell transplantation. The success of allogeneic transplantation depends on a delicate balance among host-mediated graft rejection, the desired GvT effect, and the detrimental GvHD induced by donor lymphocytes. The kind of effector mechanisms involved in the interaction between host’s and donor’s immune cells seems to depend on whether human leukocyte antigen-matched, major histocompatibility complex (MHC)-disparate, or minor histocompatibility antigen-mismatched allogeneic transplantations are performed, as well as on the degree of host T cell depletion caused by the conditioning regimen. Host CD8+ and NK cells serve as a barrier against the donor hematopoiesis and the transplanted lymphocytes [210
]. Experiments with perforin or perforin/FasL double-deficient mice demonstrate that donor-mediated cytotoxicity has a role to play in enabling engraftment [211
]. Under nonmyeloablative-conditioning regimens, the expansion of donor CD4+ cells in the early post-transplantation period crucially relies on their FasL expression to defend themselves against the host-versus-graft effect [212
]. In addition, viral transfection of donor lin- BM cells with FasL protected the donor graft from acute rejection without acute hepatotoxicity, myelosuppression, and immunosuppression [213
]. Transplantation of FasL-defective donor T cells leads to a mixed chimerism in the murine recipient [214
], pointing to the contribution of the FasL system for elimination of residual host hematopoiesis surviving total body irradiation and for the successful expansion of donor hematopoiesis. In contrast, FasL expression on donor effector cells may also be deleterious, as Fas expression on hematopoietic cells occurs within 2 weeks after the induction of acute GvHD, probably via stimulation by IL-12 and IFN-, and CD4+ cytolytic T lymphocytes mediate BM failure in a FasL-dependent manner [215
]. Furthermore, post-transplant lymphoid hypoplasia and B cell dysfunction are linked to an intact FasL [214
]. Inactivation of FasL on donor lymphocytes by mutation [216
] or antagonistic mAb [217
] led to reduced cytotoxicity against the host liver and skin. Although the contribution of the FasL system to the extent, pattern, and morbidity/mortality of GvHD varied in the different models examined, transplantation of FasL-mutated donor lymphocytes was associated with prolonged life of the allogeneic host in at least one report [218
].
The GvT effect has usually been attributed to CD8+ cells and their equipment with perforin and granzyme, as CD8+ cells were thought to act primarily through granule-dependent mechanisms, and perforin-/- mice are characterized by profound disturbances in the immune surveillance of tumors [219 ]. Recent investigations, however, suggest a more complex mechanism of action. In a strictly tumor antigen-dependent, syngeneic murine model, GvT effects were mediated by CD4+ and CD8+ cells, but the former acted primarily via FasL and secondarily via perforin, and the situation was the reverse in CD8+ cells [220 ]. These analyses demonstrate that there is a redundancy in the effector mechanisms, as they were not restricted to certain types of T cells but rather preferentially used by them. The existence of individual MHC class I- or II-restricted tumor antigens can be postula
TOP
ABSTRACT
INTRODUCTION
