| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print March 30, 2006
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Pathology, Division of General Pathology, University of Verona, Italy
1Correspondence: Department Pathology, Division of General Pathology, Strada Le Grazie 4, 37134 Verona, Italy. E-mail: marco.cassatella{at}univr.it
ABSTRACT
Contrary to their traditional characterization as terminally differentiated effectors of inflammation, neutrophils are remarkably versatile cells. Indeed, their capacity to change phenotype under specific circumstances, their active involvement in the regulation and resolution of inflammation, their response to a wide variety of cytokines and chemotactic molecules, and their regulatory role in angiogenesis and tumor fate have made it clear that they represent far more than "short-lived cells devoid of transcriptional activities, that only release preformed mediators and kill pathogens". The multiple and amazing functional capacities of this cell type are also illustrated by the fact that the neutrophil may function as an important source of cytokines, at levels comparable with and in some cases, higher than those made by other leukocytes. To date, the families of cytokines, which in vitro or in vivo have been convincingly reported as being produced by neutrophils, include proinflammatory/anti-inflammatory cytokines, immunoregulatory cytokines, chemokines, angiogenic/fibrogenic factors, and tumor necrosis factor (TNF) superfamily members. The latter molecules are multifaceted cytokines whose integrated actions not only influence the development, homeostasis, and adaptive responses of many cells and tissue types but are also implicated in the antitumoral response. The recent findings that neutrophils produce in a finely regulated manner a TNF superfamily member involved in tumor cell killing and autoimmunity, namely TNF-related apoptosis-inducing ligand, open an additional perspective to exploit neutrophils for novel roles in anticancer responses and modulation of autoimmune diseases.
Key Words: IFN • cancer • inflammation
INTRODUCTION
Neutrophils represent the first line of defense and play an active role in inflammatory response, acting as a powerful defensive system against invading bacteria [1
]. After challenge by various stimuli, neutrophils have the capacity to release lytic enzymes with potent antimicrobial potential and to generate reactive oxygen intermediates (ROI), which are essential for pathogen killing [1
]. Neutrophils can also produce, upon appropriate stimulation in vitro and in vivo, a variety of proteins, including cytokines, chemotactic molecules, and other mediators that are involved in their effector functions [2
]. This latter function is currently the subject of a new wave of enthusiastic research, which highlights that neutrophils may also act as key regulators of the in vivo cross-talk among immune, endothelial, stromal, and parenchymal cells [2
]. Furthermore, the intriguing finding of major histocompatibility complex class II molecule expression by activated neutrophils has extended the range of their potential immunological functions, suggesting that these cells may even play an unexpected role in immunity and autoimmune diseases, other than in inflammation or infections [3
]. In addition, the discovery that neutrophils produce cytokines and chemokines has led to further speculation about their involvement in the antitumoral response. Granulocytes are an uncommon component of human and animal tumors. In animal models, they may sometimes favor malignant growth and progression [4
]. Nevertheless, recent studies suggest that neutrophils are active in immunosurveillance against several tumors [5
]. Experimental studies of tumor cure and prevention have in fact suggested that engagement of neutrophil functions is crucial for the establishment of an effective antitumoral immune response and immune memory reaction. Cytotoxic mediators of tumor and endothelial cell killing produced by neutrophils include tumor necrosis factor 
(TNF-) [6
], defensins [7
], proteases (such as elastase and cathepsin G) [8
], and ROI, such as nitric oxide, hydrogen peroxide, and hypochlorous acid [9
]. Conversely, neutrophils may act "antiangiogenically" by releasing CXC chemokine ligand 9 (CXCL9), CXCL10, and CXCL11 {on TNF- and interferon-
(IFN-) [10
]}. In addition, it has been observed recently that they generate a bioactive, angiostatin-like fragment, inhibiting basic fibroblast growth factor plus vascular endothelial growth factor-induced endothelial cell proliferation in response to inflammatory stimuli [11
]. In this context, the recent identification of TNF-related apoptosis-inducing ligand (TRAIL/APO-2L) as an additional, potent antitumoral mediator produced by neutrophils [12
13
14
15
] has extended the list of effective weapons, which neutrophils can use against cancer.
TRAIL is a TNF family member, which has received considerable attention, as many cancer cell types have been shown to be sensitive to its capacity to induce apoptosis, and most normal cells appear to be resistant to this action [16
, 17
]. TRAIL is a transmembrane (type II) glycoprotein (mTRAIL), which induces apoptosis through engagement of death receptors [18
]. Like Fas ligand (FasL)/APO-1L and TNF-, TRAIL also exists physiologically in a biologically active, soluble, homotrimeric form (sTRAIL). TRAIL is unusual as compared with the other cytokines of the TNF family, as in humans, it interacts with a complex system of receptors. The latter consists of two proapoptotic death receptors (TRAIL-R1/DR4 and TRAIL-R2/DR5) and three decoy receptors, which are devoid of functional death domains (TRAIL-R3/DcR1/TRID and TRAIL-R4/DcR2/TRUNDD) or produced as secreted protein (osteoprotegerin) [19
, 20
]. Decoy receptors may therefore act as a negative regulator of cell death. The reasons for the existence of multiple TRAIL receptors are unclear, but it is likely that the balance of the expression levels between the death and the decoy receptors is an important factor determining the apoptotic effect of TRAIL. In this regard, preclinical studies in mice and nonhuman primates have shown the potential use of recombinant (r)sTRAIL and agonistic anti-DR5 or -DR4 antibodies for cancer therapy [21
]. Furthermore, mice genetically deficient of the TRAIL gene exhibit increased susceptibility to experimental and spontaneous tumor [22
], suggesting an important role of endogenous TRAIL in tumor surveillance.
Recent work suggests that TRAIL also has different physiological functions, which are not limited to the killing of transformed cells. TRAIL has consistently been shown to inhibit autoimmune disease in a number of animal models. Studies using TRAIL-deficient mice, sTRAIL receptor (sDR5), or neutralizing anti-mTRAIL antibodies to block TRAIL function or TRAIL adenovirus have shown that TRAIL inhibits autoimmune arthritis [23 , 24 ], diabetes induced by cyclophosphamide and streptozatocin [23 , 25 ], diabetes development in nonobese diabetic mice [26 ], and experimental autoimmune encephalomyelitis (EAE) [27 , 28 ]. In these studies, TRAIL was suggested to play a variety of effects ranging from inhibiting cytokine and antibody production to inhibiting inflammation, cell cycle progression, and proliferation of autoreactive T cells. It is interesting that recent findings have shown that TRAIL can also accelerate the course of autoimmune diseases under particular settings, particularly in EAE [29 ]. A dual role for TRAIL has also been proposed in rheumatoid arthritis (RA), where TRAIL was shown to play a role in stimulating proliferation and apoptosis of fibroblast-like synoviocytes [30 ]. However, although numerous studies have examined the role of TRAIL in autoimmune disease in experimental models of animals, less is known about the role of TRAIL in autoimmune disease in human patients.
TRAIL is known to be produced by different cells of the immune system, including CD4+ T lymphocytes upon treatment with anti-CD3 or monocytes, natural killer cells, or monocyte-derived or plasmacytoid dendritic cells, for instance, following treatment with type I or type II IFNs [31
32
33
34
35
36
]. Consistent with these findings, TRAIL seems to play an important role in surveillance by cells of the immune system against viral infections and malignant transformation of host cells [37
]. As IFNs are now clinically used for the treatment of several cancers, the IFN-inducible expression of TRAIL on these immune effector cells may be partly involved in the anti-tumor effect of IFNs [21
]. It is interesting that in certain tumor cell lines, TRAIL expression can be induced on the tumor cells themselves by agents such as IFN-, all-trans retinoic acid, or histone deacetylase inhibitors, which may result in TRAIL-mediated tumor cell apoptosis in a paracrine manner [38
, 39
]
REGULATION OF NEUTROPHIL-DERIVED TRAIL
The capacity of human neutrophils to accumulate TRAIL mRNA transcripts, express mTRAIL, or release sTRAIL has been investigated by different groups subsequently to the identification of other cells of the immune system as TRAIL-producing cells. This is, however, not surprising, as any potential involvement of neutrophils in physiopathological conditions, which at first sight, appear distantly related to the inflammatory process, usually receives little attention by specialists in the field. In this regard, the first indication that neutrophils express TRAIL mRNA has been provided by Renshaw et al. [40
] in the context of studies that were specifically aimed at clarifying whether neutrophils respond or not to exogenous TRAIL. These authors showed that freshly isolated neutrophils display no surface expression of TRAIL and have a low but constitutive expression of TRAIL mRNA [40
]. They also verified whether surface TRAIL could be expressed in neutrophils treated with IFN-, lipopolysaccharide (LPS), or TNF-, but their hypothesis was disproven [40
]. Surface TRAIL expression was also studied in neutrophils purified from joint aspirates of patients with RA [40
]. Again, no differences were assessable when comparing these cells with peripheral blood neutrophils [40
]. During the course of experiments performed in my laboratory in 1996, we had also observed that IFN--activated neutrophils express high levels of TRAIL mRNA but not the mTRAIL protein. To reconcile the apparent discrepancy between the elevated TRAIL mRNA accumulation and the lack of mTRAIL expression, we suspected that TRAIL could be expressed effectively by neutrophils, but only transiently, because of its rapid cleavage by the action of membrane-bound proteases. This hypothesis was consistent with the description of protease activities able to cleave TRAIL from the cell surface of transformed B and T cells [41
] and with similar mechanisms described for other death receptor ligands [42
]. However, in neutrophils cultured with or without IFN- or IFN- in the presence of the general cysteine protease inhibitor E64, which was used by Mariani and Krammer [41
] to validate their findings, mTRAIL was never detected (Cristina Tecchio and M. A. Cassatella, unpublished observations). Only a few years later, however, we could proceed further on our investigations concerning neutrophil-derived TRAIL, when more useful and appropriate reagents became available, in particular, a commercial enzyme-linked immunosorbent assay (ELISA), which was specific and sensitive enough to detect sTRAIL. This ELISA together with Northern blotting analysis enabled us to demonstrate clearly that peripheral blood neutrophils, although constitutively expressing and releasing, respectively, low levels of TRAIL transcripts and antigenic sTRAIL (
100 pg/ml/5x106 neutrophils/24 h), dramatically up-regulate TRAIL mRNA and secrete elevated amounts (600 pg/ml/5x106 neutrophils/24 h) of functionally active sTRAIL in response to therapeutic concentrations of IFN- or IFN- [12
]. IFN-ß was subsequently found to be as potent as IFN- [43
], in line with the observations made in human monocytes [44
45
46
]. In contrast, classical neutrophil agonists such as formyl-methionyl-leucyl-phenylalanine (fMLP), LPS, granulocyte-colony stimulating factor (G-CSF), and granulocyte macrophage (GM)-CSF failed to up-regulate TRAIL gene expression and consequently, sTRAIL production [12
]. It is interesting that by using antibodies different from those used in 1996, we confirmed that mTRAIL is not expressed on the surface of freshly isolated neutrophils and increases only weakly in response to IFN- or IFN- in a small percentage of donors [12
, 43
]. Conversely, surface TRAIL expression was always induced by IFN- in monocytes, confirming the observations made by Griffith and colleagues [32
]. After detecting the presence of the sTRAIL by ELISA in supernatants of IFN-activated neutrophils, we were interested in a better characterization of the aggregation state of extracellular sTRAIL. For this purpose, concentrated supernatants were electrophoresed under reducing conditions followed by immunoblot analysis. We could show that sTRAIL was detectable as two major bands of 24 and 48 kDa, likely representing the monomeric and dimeric form of the protein, respectively [12
]. We then performed Western blot analysis, under nonreducing conditions or under native conditions, which further confirmed the similarity of the migration patterns displayed by a human bioactive rsTRAIL (rhsTRAIL) and the sTRAIL present in neutrophil and peripheral blood mononuclear cell (PBMC)-derived supernatants [12
]. Collectively, our experiments not only confirmed and extended the information previously reported by Renshaw et al. [40
] but also revealed that sTRAIL released by IFN--activated leukocytes is characterized by an aggregation state, which preserves its capacity to exert biological activities [12
].
Our observations were corroborated by Koga and co-workers [14
], who reported that TRAIL mRNA is constitutively expressed in resting neutrophils and augments after stimulation with IFN- and at lower extent, IFN- [by real-time polymerase chain reaction (PCR)]. On the contrary, TRAIL mRNA was shown to be absent in resting CD4+, CD8+, CD14+, and CD19+ cells. Other cytokines, including TNF-, G-CSF, GM-CSF, and transforming growth factor-ß, were found to be unable to induce TRAIL mRNA expression in neutrophils or PBMC, and IFN- and LPS were found effective in PBMC. Koga and co-workers [14
] detected sTRAIL in culture media of neutrophils stimulated with IFN- for 4 h (400 pg/ml/107 neutrophils/4 h). However, in contrast with the results described previously [12
, 40
], the authors detected, in resting neutrophils of a few individuals, variable but low levels of mTRAIL, which could be even slightly augmented in response to IFN- [14
].
The capacity of human neutrophils to serve as a cellular source of sTRAIL was documented further by Kamohara and colleagues [13
]. This group reported a constitutive and abundant expression of TRAIL mRNA in resting neutrophils, unaffected by IL-1 and GM-CSF but enhanced by IFN- in a time- and dose-dependent manner. The authors also showed that neutrophils incubated with LPS or TNF- for 6 h contain much lower levels of TRAIL mRNA than untreated cells and that the addition of TNF- to IFN-, GM-CSF, or IFN- plus GM-CSF markedly decreases the levels of TRAIL mRNA expression [13
]. Contrary to other reports [12
, 14
, 40
], Kamohara and colleagues [13
] also detected high levels of mTRAIL on most freshly isolated neutrophils but not in cells exposed for 24 h to medium alone, IFN-, or TNF-, in which mTRAIL was shown to dramatically decrease and become almost undetectable. The latter observations are also in contrast with those recently published by Lum et al. [47
], who found that overnight-cultured neutrophils express significant levels of mTRAIL, which can further increase in response to CXC chemokine receptor 4 (CXCR4) ligands, including CXCL12/stromal cell-derived factor-1 (SDF-1) and human immunodeficiency virus (HIV) gp120. Unfortunately, these same authors did not mention if freshly isolated neutrophils express surface TRAIL or if CXCL12/SDF-1 and HIV gp120 induce the release of sTRAIL. Recently, a strong mTRAIL expression has been observed in eosinophils and neutrophils isolated from patients affected by atopic dermatitis but only exceptionally in control individuals [48
]. At present, all the divergent data concerning mTRAIL expression in resting or activated neutrophils remain to be clarified. At least in part, they could be explained by the use of anti-TRAIL antibodies obtained from different sources. Such a hypothesis, however, is no longer valid in the case of the opposing results described by myself and co-workers [12
, 43
] and Kamohara and colleagues [13
], as the same brand of commercial antibodies was used.
Kamohara and colleagues [13
] examined TRAIL protein expression by immunoprecipitaion followed by immunoblotting in neutrophil lysates and neutrophil-derived supernatants to understand why, in their experiments, mTRAIL was lost upon culture. A single protein complex of unspecified molecular weight was immunoprecipitated and was easily detectable in lysates of freshly isolated neutrophils. This complex seemed to decrease greatly within 6 h of incubation in complete medium with or without IFN- and almost disappeared in the presence of TNF- or TNF- plus IFN- [13
]. The same TRAIL immunoprecipitate was present in supernatants of 6 h-cultured neutrophils and at higher levels, in those from IFN--treated cells but was almost absent in supernatants from TNF-- or TNF- plus IFN--treated neutrophils. Based on these findings, the authors speculated that the TRAIL protein is synthesized by resting neutrophils and after being displayed on the membrane, is shed with the time into the supernatants by the activities of matrix metalloproteinases (MMPs) [13
]. If true, this theory does not, however, explain why sTRAIL was absent in supernatants from TNF--treated neutrophils [13
]. In any case, as the addition of the MMP inhibitor KB8301 to neutrophil cultures only increased the level of mTRAIL expression [13
] slightly, the "shedding" hypothesis of TRAIL remains an open issue.
Kemp and colleagues [49
] recently provided evidence for the necessity of protease activities to optimize sTRAIL release. These researchers, on the basis of their previous findings that neutrophils isolated from the urine of patients undergoing Mycobacterium bovis bacillus Calmette Guerin (BCG) immunotherapy express mTRAIL [15
], decided to further investigate the direct response of neutrophils to BCG in terms of TRAIL expression and regulation. In their experiments, they observed a slight but reproducible increase in surface-bound TRAIL on neutrophils stimulated with BCG for 4 h and a nearly six-fold enhancement of released sTRAIL levels as compared with unstimulated neutrophils (600 vs. 100 pg/ml/3x106 neutrophils/24 h). As neither inhibition of protein synthesis nor inhibition of RNA synthesis suppressed the release of sTRAIL by BCG-stimulated neutrophils, the authors concluded that resting neutrophils possess intracellular stores of prefabricated TRAIL, which can be secreted quickly upon appropriate stimulation [49
]. In support of their hypothesis, Kemp and colleagues [49
] detected appreciable levels of intracellular TRAIL, which did not change between freshly isolated and 4 h-cultured neutrophils. This intracellular TRAIL was shown to substantially decrease after BCG stimulation [49
]. When measured by ELISA, however, the content of cell-associated TRAIL in untreated neutrophils (corresponding to 2 ng/ml/3x106 neutrophils) declined in a time-dependent manner [49
]. Such decrease was attributed to a proteolytic degradation of TRAIL occurring within the cultured neutrophils, as there was no concomitant increase of sTRAIL levels in the corresponding supernatants and as it was prevented if unstimulated neutrophils were treated with the irreversible protease inhibitor diisopropylfluorophosphate (DFP) [49
]. Furthermore, less sTRAIL was detected in the supernatants of neutrophils stimulated with BCG in the presence of DFP, therefore suggesting that putative protease activities are needed for maximal sTRAIL release after BCG stimulation. At present, the neutrophil protease(s) responsible for cleaving TRAIL within the cells have not been identified yet. It is surprising that in this same study, IFN- or IFN-, at up to 1000 U/ml and over a 24-h time-course period, was found neither to increase mTRAIL expression nor sTRAIL release [49
]. These data, conflicting with several studies published previously [12
13
14
, 43
] and with the well-established ability of IFNs to up-regulate TRAIL expression in other myeloid and lymphoid cells [31
32
33
34
35
36
], were explained by donor variability, antibodies used, neutrophil isolation techniques, or even cell concentration used [49
].
In summary, in spite of some issue(s) that need to be resolved, there is no doubt that neutrophils can produce and release sTRAIL upon appropriate stimulatory conditions. It is remarkable that the data collected to date clearly point to the existence of at least two mechanisms controlling sTRAIL release in neutrophils. One of these two, which occurs in the case of cells treated with type I or type II IFN [12 13 14 , 43 ], determines TRAIL release only after its gene activation and de novo synthesis (Fig. 1A ). The second mechanism, which occurs, for instance, in response to BCG and irradiated M. tuberculosis [49 ], regulates TRAIL release via the discharge of a small TRAIL pool that appears to be constitutively present in freshly isolated neutrophils (Fig. 1B) .
|
Given the ability of neutrophils to release sTRAIL, the most obvious question was to verify whether neutrophil-derived TRAIL might retain any of the biological activities known for this cytokine. For this purpose, we initially examined whether endogenous TRAIL could be involved in modulating apoptosis in IFN-activated neutrophils, in light of the well-described role of other members/receptors of the TNF ligand family, such as TNF- and Fas death receptor, on apoptosis modulation [50
, 51
]. To address this issue, it was necessary to concomitantly investigate the relative expression of death receptors (TRAIL-R1 and TRAIL-R2) versus decoy receptors (TRAIL-R3 and TRAIL-R4) and osteoprotegerin, as they participate to regulate TRAIL sensitivity. Prior to our study, other groups had already examined the pattern of TRAIL receptor expression and the sensitivity of neutrophils to exogenous rTRAIL as well but had reached inconsistent conclusions. For instance, Daigle and Simon [52
] detected the expression of all four TRAIL receptors at mRNA level by reverse transcription (RT)-PCR but only TRAIL-R1, -R3, and -R4 at the protein level in freshly isolated neutrophils. By contrast, Renshaw et al. [40
] detected the expression of only TRAIL-R2 and -R3 at mRNA and protein levels but no TRAIL-R1 and TRAIL-R4. These different results may be explained by the fact that different antibodies were used and by the possibility that the use of RT-PCR may result in the amplification of mRNA species from small numbers of contaminating cells. With regard to a potential TRAIL-mediated proapoptotic effect, Daigle and Simon [52
] reported that 100 ng/ml rTRAIL, in combination with an enhancer that causes oligomerization of sTRAIL and enhances its apoptotic effect, did not increase the rate of neutrophil death after 40 h of incubation but partially inhibited the survival signals provided by GM-CSF, G-CSF, or IFN- [52
]. By contrast, Renshaw et al. [40
] found that neutrophil apoptosis could be accelerated by exposure to a leucine zipper-tagged form of TRAIL (LZ-TRAIL), which mimics cell-surface TRAIL, and that such effect was abrogated completely by the presence of blocking antibodies to TRAIL-R2 but not to TRAIL-R1. It is interesting that inhibition of TRAIL binding to TRAIL-R3 did not alter TRAIL sensitivity of neutrophils [40
]. Furthermore, experiments using TRAIL-R2-blocking antibodies or a TRAIL-R1:Fc fusion protein excluded any autocrine or paracrine effect of TRAIL upon constitutive neutrophil apoptosis [40
].
In agreement with the data reported by Renshaw et al. [40
] and subsequently by Kamohara et al [13
] and Hasegawa et al [53
], our flow cytometric analyses of TRAIL receptors in freshly isolated neutrophils confirmed a low expression of TRAIL-R2, a high expression of TRAIL-R3, and the absence of TRAIL-R1 and TRAIL-R4 [12
]. It is interesting that neutrophils isolated from the bone marrow of chronic myeloid leukemia (CML) patients also displayed a similar pattern of TRAIL receptors at mRNA levels [54
]. In our hands, increasing concentrations of rhsTRAIL His-tag CC mutant, a multimeric rsTRAIL characterized by a biological activity higher than LZ-TRAIL, slightly augmented the spontaneous apoptosis of neutrophils occurring in a 24-h culture period [12
]. The latter, however, was prevented by IFN-, which by itself, significantly reduced neutrophil apoptosis, as also shown by Wang et al. [55
]. Therefore, release of sTRAIL by IFN-activated neutrophils did not correlate with an increased apoptosis rate of the same cells. Also Kamohara et al. [13
] formulated the hypothesis that neutrophil-generated TRAIL might play a role in mediating their own apoptosis. It is interesting that they found that the addition of anti-TRAIL-neutralizing antibodies to neutrophils incubated in complete or TNF-/IFN--containing medium decreased their apoptotic rate significantly but only if measured after 3 and 6 h of culture and not after 24 h [13
]. As TNF- but not IFN- rapidly down-regulated the expression of TRAIL-R3 at mRNA and protein level, these researchers concluded that endogenous TRAIL is involved in modulating the early stage of neutrophil apoptosis, particularly in response to TNF-, which concomitantly was found to also modulate TRAIL receptors [13
].
A novel, potential role of TRAIL in regulating neutrophil apoptosis was also recently offered by Lum et al. [47
], who proposed that TRAIL might be involved in eliminating senescent neutrophils. Senescence affects those neutrophils that do not infiltrate the inflammatory tissue and therefore return to specific organs (such as the liver, the spleen, and the lungs), in which they are subsequently eliminated. Senescent neutrophils are known to express high levels of CXCR4, which is instead low/absent in immature neutrophils. Therefore, senescent neutrophils respond to the chemotactic effects of SDF-1/CXCL12 [56
]. Consequently, senescent neutrophils migrate to areas of high SDF-1/CXCL12 production, including the bone marrow, the spleen, and the liver [57
], where they undergo an as-yet undefined apoptotic pathway. It is interesting that Lum et al. [47
] had observed previously that treatment of T cells with the CXCR4 ligand, HIV gp120, can induce an acquired sensitivity to TRAIL-mediated killing. This observation led the authors to question whether SDF-1/CXCL12 or other CXCR4 ligands might also alter TRAIL and TRAIL receptor regulation in senescent neutrophils and in turn, TRAIL-mediated apoptosis. This hypothesis turned out to be correct [47
]. In fact, and as already discussed above, these researchers found that neutrophils treated with HIV gp120 or SDF-1/CXCL12 display not only an increased expression of TRAIL ligand but also of all four TRAIL receptors [47
]. Because of these findings, they pretreated neutrophils with SDF-1/CXCL12, HIV gp120, or vehicle control prior to LZ-TRAIL and found that the cells incubated with HIV gp120 or SDF-1/CXCL12 became highly sensitive to TRAIL-mediated apoptosis, whereas those treated with vehicle control were poorly susceptible to TRAIL [47
]. These observations led the authors to reason that in SDF-1/CXCL12-treated neutrophils, the combination of enhanced susceptibility to, together with the enhanced production of, TRAIL might result in an autocrine or paracrine TRAIL-mediated apoptosis. To formally test this hypothesis, Lum et al. [47
] treated neutrophils with SDF-1/CXCL12 or HIV gp120 for up to 3 days in the continuous presence of inhibitory, anti-TRAIL receptor, or anti-FasL antibodies. They found that although the number of apoptotic neutrophils in control treatments was not different from cells treated in the presence of anti-TRAIL-R2 and anti-FasL in cultures treated with anti-TRAIL-R1, death was reduced remarkably (from 82% to 23% at Day 3). They therefore concluded that TRAIL-R1, but not TRAIL-R2, is required for neutrophil death signaling following activation with SDF-1/CXCL12 or HIV gp120 and moreover, suggested that the TRAIL/TRAIL-R1 interaction is supplied in a paracrine manner [47
]. In this scenario, the acquisition of high CXCR4 expression by aged neutrophils would represent a prerequisite to acquire responsiveness to SDF-1/CXCL12, which might lead to their elimination via chemotaxis to the bone marrow, mTRAIL production, and acquisition of TRAIL sensitivity through TRAIL-R1 and not through TRAIL-R2, as it occurs in nonsenescent cells [40
]. In vivo experiments have supported this model, as mice receiving antagonistic, anti-TRAIL antibodies showed an inhibiton of neutrophil accumulation within the bone marrow, whereas animals receiving rTRAIL had reduced numbers of bone marrow neutrophils [47
].
A potential, clinical implication of neutrophil apoptosis mediated by TRAIL was raised recently by the studies of Matsuyama et al. [58 ] and Lub-de Hooge et al. [59 ]. These authors strongly suggested that TRAIL can accelerate the apoptosis of neutrophils in systemic lupus erythematosus (SLE) patients. Accordingly, it was found that the levels of serum sTRAIL correlated significantly with neutrophil numbers in SLE patients. Apoptotic cell percentage induced by TRAIL in vitro was significantly higher in neutrophils isolated from SLE patients with neutropenia than in SLE patients without neutropenia, strongly suggesting that in vivo, TRAIL can accelerate the apoptosis of neutrophils in SLE patients [58 ]. It is interesting that TRAIL-R3 expression levels were significantly lower in neutrophils of SLE patients with neutropenia than in patients without neutropenia or healthy donors [58 ], a factor that at least in these patients, could contribute to render neutrophils more susceptible to TRAIL-induced apoptosis. Besides, glucocorticoid therapy up-regulated the expression level of TRAIL-R3 in neutrophils of SLE patients, in parallel with the neutrophil count [58 ]. It is striking that Matsuyama and co-workers [58 ] also found that in SLE patients, T cells display elevated amounts of mTRAIL on their surface and that in vitro, these T cells kill autologous neutrophils in a TRAIL-dependent manner. All in all, these results not only have provided novel insights into the molecular pathogenesis of SLE neutropenia, but they have also highlighted that unique mechanisms may contribute to increased neutrophil apoptosis in specific diseases, some of them being TRAIL-dependent. Nonetheless, further studies are necessary to clarify the molecular pathogenesis of SLE neutropenia.
ANTITUMOR EFFECTS OF NEUTROPHIL-DERIVED TRAIL
The second, more obvious, approach to verify whether neutrophil-derived TRAIL might have biological functions was to test whether supernatants from activated neutrophils (or neutrophils expressing mTRAIL) could exert cytotoxic activities toward TRAIL-sensitive cells. In this regard, we could show that sTRAIL, present in supernatants conditioned by IFN--stimulated neutrophils, exerted remarkable, proapoptotic activities toward TRAIL-sensitive (Jurkat J32 clone and MEG-01) but not toward TRAIL-resistant (KU-812) leukemic cell lines [12
]. These activities were neutralized partially by different TRAIL inhibitors, including anti-TRAIL-neutralizing antibodies or TRAIL-R1/Fc and TRAIL-R2/Fc chimeras [12
]. In similar manner, Koga and colleagues [14
] confirmed a cytotoxic effect of neutrophil-derived TRAIL in a coculture system of neutrophils with Jurkat cells in the presence of IFN-. Such cytotoxicity was partially suppressed by the addition of anti-TRAIL antibodies, indicating that other mechanisms were involved in the direct cell killing induced by IFN--stimulated neutrophils. Additional data supporting a TRAIL-mediated, antitumoral activity of human neutrophils in vitro and possibly in vivo have been provided by Griffith’s group [15
, 49
] in two studies aimed at clarifying the mechanisms mediating the immunostimulatory effect of BCG in bladder cancer. In this tumor, a proinflammatory, T helper cell type 1 (Th1) cytokine (e.g., IL-2, IL-12, IFN-) response, which is often associated with a more favorable response [60
], is typically initiated following BCG instillation [61
]. It is interesting that if BCG is combined with IFN-2B, the Th1 cytokine response appears to be potentiated further [62
]. After BCG administration, there is a strong, local immune response characterized by an early granulocytic influx into the bladder wall, followed by mononuclear cells, of which CD4+ and CD8+ cells predominate [63
]. T cells are known to be essential for the BCG-mediated antitumor activity [64
], although the contribution of the granulocyte infiltrate has been largely neglected. In their studies, Kemp and colleagues [15
] were able to demonstrate that intravesical administration of BCG is followed by an increase of urinary levels of IFN- and subsequently, of TRAIL. Urinary TRAIL was characterized further and found to correspond to the cleaved sTRAIL form and retain cytotoxic activities toward RT-4 bladder tumor cell targets [15
]. It is remarkable that the authors also noticed that patients better responding to BCG therapy had significantly higher urine TRAIL levels as compared with nonresponders [15
]. As only neutrophils and not the other leukocytes present in voided urine after BCG instillation expressed mTRAIL, Kemp and colleagues [49
] speculated that the neutrophil response to BCG in vivo could contribute to TRAIL production/release and in turn, to the antitumor activity seen in their clinical setting through TRAIL itself. Once they had established that neutrophils stimulated in vitro with BCG effectively express and produce TRAIL [49
], the authors sought to determine whether the small amount of TRAIL present on the cell surface of BCG-stimulated neutrophils conferred them a cell-mediated, tumoricidal activity or if the TRAIL released following BCG stimulation retained cytotoxic functions. Their experiments demonstrated that when BCG-stimulated neutrophils were used in coculture as effectors against 51Cr-labeled RT-4 bladder tumor cell targets, a small but reproducible amount of target cell lysis occurred [49
]. Similarly, when RT-4 cells were incubated with the culture supernatants harvested from BCG-stimulated neutrophils but not from unstimulated neutrophils, the RT-4 cells underwent TRAIL-induced killing [49
]. The cell death induced by the BCG-stimulated, neutrophil-conditioned medium was determined to be apoptotic, as demonstrated by the activation of caspase-8 and by the protective effect of z-Val-Ala-Asp-fluoromethylketone [49
]. In this regard, we obtained similar results using supernatants of IFN-treated neutrophils toward Jurkat cells (Veronica Huber, Licia Rivolfini, and Marco A. Cassatella., unpublished observations). Collectively, these results led the authors to favor the hypothesis that the sTRAIL released by IFN- and/or BCG-activated neutrophils or even their mTRAIL not only exerts antitumor effects in vitro but likely in vivo too.
The potential antitumor significance of neutrophil-derived TRAIL has been evaluated in additional clinical settings. For instance, given the previous use of IFN- in the treatment of cancer diseases such as CML [65
], we investigated the release of sTRAIL by leukocytes isolated from CML patients [12
]. We could show that CML neutrophils and mononuclear cells, upon incubation with therapeutic doses of IFN-, release sTRAIL into the extracellular environment as efficiently as normal leukocytes [12
]. Immature CML myeloid elements, isolated from CML bone marrows, were also responsive to IFN- in terms of sTRAIL release [12
]. In addition, we observed that the pattern of mTRAIL and TRAIL receptor expression in freshly isolated CML neutrophils was substantially similar to those of healthy donors. All of these findings were confirmed recently by Liu et al. [54
], who also reported that the mRNA ratio (R1+R2)/R3 of TRAIL receptors in neutrophils isolated from the bone marrow of five new diagnosed CML patients increased after in vitro stimulation with IFN-. These results imply that IFN- may render these cells more susceptible to TRAIL-induced apoptosis, similarly to what occurs in SLE patients with neutropenia [58
, 59
]. In our experiments, CML neutrophils displayed a constitutive, apoptotic rate lower than normal neutrophils but remained susceptible to the proapoptotic effect of rhsTRAIL His-tag CC mutant [12
]. It is regretful that we could not obtain sera of CML patients subjected to IFN- administration as a result of the introduction of STI571 as treatment of choice in CML. However, we had access to blood of stage IV metastatic melanoma patients undergoing IFN- therapy. In these latter samples, we detected serum levels of sTRAIL, which were dramatically more elevated than those measured prior to cytokine injection, as well as in healthy donors [12
]. Consistent with an IFN--mediated induction of TRAIL de novo synthesis, the accumulation of leukocyte-associated TRAIL in these samples dramatically augmented following IFN- exposure, implying that peripheral blood leukocytes contribute greatly to the sTRAIL release in vivo [12
]. Based on these findings, we reasoned that in the chronic phase of CML, which is a disease clinically characterized by a high number of circulating neutrophils belonging to the CML clone, administration of IFN- might induce a massive release of biological, active sTRAIL into the serum [12
]. In summary, these data suggest that the release of sTRAIL by IFN--treated peripheral blood cells represents a novel mechanism, whereby therapeutic concentrations of IFN- might exert immunomodulatory and antitumor activities. They also suggest that TRAIL alone, or in combination with other drugs such as IFN-, which sensitizes a variety of solid tumor cells to TRAIL-induced apoptosis [66
67
68
], may constitute good candidate(s) for alternative anticancer therapy in many tumors.
TRAIL ACCUMULATION AND RELEASE IN NEUTROPHILS EXPOSED TO TYPE I AND TYPE II IFNs
Although the molecular mechanisms regulating TRAIL gene expression in human neutrophils are poorly characterized, recent studies have uncovered that sTRAIL release can be amplified greatly under certain conditions. After measurement of total TRAIL production by cultured neutrophils, i.e., released sTRAIL in parallel with cell-associated TRAIL, my colleagues and I [43
] and others [14
] noticed that only a minor fraction (ranging from 10% to 20%) of the total TRAIL, newly synthesized by IFN--, IFN-ß-, or IFN--activated neutrophils, was released into the extracellular milieu, the rest remaining cell-associated. Depending on the type of IFN used, total TRAIL de novo synthesis increased approximately four- to sevenfold (up to 6 ng/ml/5x106 cells/24 h) relative to neutrophils cultured in medium only (1 ng/ml/5x106 cells/24 h) [43
]. We were initially puzzled by these findings, and reminiscent of former observations made about the mechanisms controlling the expression of B lymphocyte stimulator in neutrophils [69
], we wondered whether the intracellular pool of TRAIL retained in IFN-treated cells could be mobilized eventually outside and/or onto the cell surface. Our subsequent experiments proved this to be the case, as we could demonstrate that intracellular TRAIL can move rapidly to the cell surface (already within 20 min) and/or be secreted into the extracellular medium if IFN-treated neutrophils were exposed to TNF-, LPS, and fMLP [43
]. Other proinflammatory mediators, such as CXCL8/IL-8, insoluble IC, and the heat shock protein Gp96, could also effectively induce sTRAIL release from IFN-treated neutrophils [43
]. Of note, all these various proinflammatory mediators functioned only as secretagogue-like agonists, in that they failed to augment TRAIL mRNA expression or TRAIL de novo synthesis in IFN-treated or freshly isolated neutrophils. In support of these results, immunoblots of concentrated supernatants electrophoresed under reducing conditions allowed us to detect a more abundant expression of sTRAIL in samples harvested from 24 h IFN--treated neutrophils stimulated with TNF- or fMLP than from neutrophils treated with IFN- only [43
]. Accordingly, supernatants from IFN-treated neutrophils additionally stimulated with proinflammatory mediators elicited a more effective apoptosis of Jurkat cells than supernatants harvested from neutrophils activated with IFN alone. Although stimulation of IFN-treated neutrophils with TNF-, LPS, and fMLP transiently enhanced the levels of mTRAIL too [43
], we did not explore whether this newly expressed mTRAIL could exert killing activities against TRAIL-sensitive cell targets or autocrine/paracrine proapoptotic effects. We preferred instead to identify the intracellular stores/organelles that contain TRAIL in IFN-treated neutrophils. By subcellular fractionation studies of neutrophil-derived nitrogen cavitates centrifuged on a three-layer, discontinuous Percoll density gradient [70
], we were able to localize TRAIL to the secretory vesicles/light membrane fraction [43
]. As secretory vesicles are notoriously, easily degranulated, such TRAIL localization thus favors its rapid secretion, exactly as we described [71
].
The capacity of IFNs to prime neutrophils for enhanced TRAIL release in response to proinflammatory mediators has also been reported by Kemp and colleagues [49
]. They showed that incubation of neutrophils with IFN- for 20 h, other than increasing by 90-fold the levels of TRAIL mRNA (as determined by real-time PCR), results in a roughly twofold enhancement of the intracellular TRAIL levels compared with unprimed cells [49
]. Under these conditions, release of sTRAIL after stimulation with BCG, Pam3CSK (a specific Toll-like receptor 2 agonist), or LPS was found to be at least three times more elevated than in unprimed neutrophils, consistent with the findings of my co-workers and me [43
]. In the same work, the authors show that BCG directly induces sTRAIL release from preformed stores present in freshly isolated neutrophils (see above), and using a similar, subcellular fractionation approach as we did, they localized TRAIL in the plasma membrane-enriched/secretory vesicles fraction and azurophilic granules [49
]. Therefore, it appears that although TRAIL accumulates in secretory vesicles or azurophilic granules in resting neutrophils [49
], its intracellular localization becomes more selective and is strictly in secretory vesicles/light membrane fraction in IFN-activated neutrophils [43
].
Taken together, the above data reveal that the release of sTRAIL by human neutrophils can be controlled by an additional mechanism, that is, by the effect of proinflammatory mediators, including bacteria and their products, which mobilize the intracellular TRAIL stores, specifically accumulating in IFN-treated neutrophils (Fig. 1A) . This phenomenon might serve to greatly amplify the extracellular release and in turn, the effect of neutrophil-derived sTRAIL, especially in inflammatory conditions or antitumor responses, for instance, in response to BCG therapy [49 ]. In this latter setting, the "priming"/stimulating combination of IFN and BCG provides additional support for the correlation between the high urinary levels of proinflammatory cytokines, including IFN, high urinary sTRAIL levels, and responsiveness to BCG therapy [49 ].
CONCLUDING REMARKS
Collectively, the findings that were summarized in this article provide unequivocal evidence about the capacity of neutrophils to produce biologically active TRAIL in vitro or in vivo. Nonetheless, additional work is necessary to better comprehend the biological significance of these findings in health and disease. It will be important to dissect the role of neutrophil-derived TRAIL in those tumors that are heavily infiltrated by granulocytes, for instance, naturally developed or obtained by cytokine gene transfer strategies [5 ]. The success of BCG therapy of bladder cancer associated with the increased TRAIL expression on local neutrophils [49 ] not only testifies the relevance of neutrophil-derived TRAIL but also serves as a basis to further investigate the potential antitumoral effects of the TRAIL-neutrophils network in other forms of immunotherapies. Another important, open problem that is necessary to explore is whether neutrophil-derived TRAIL plays a role in conditioning antimicrobial or inflammatory responses. This would clarify the biological meaning of those observations that have identified the existence of conditions that amplify the secretion of TRAIL by neutrophils [43 , 49 ]. On the same line, it would be extremely important to also characterize the role of neutrophil-derived TRAIL in autoimmune diseases such as RA, in light of the growing awareness that infiltrating neutrophils certainly play an important role in the molecular pathology of this disease [72 ]. In this regard, the recent findings that TRAIL induces the expression of IL-1 receptor antagonist [48 ] might point surprisingly to an anti-inflammatory effect of TRAIL. Finally, other fascinating aspects concerning the mechanisms regulating TRAIL gene expression and TRAIL metabolism in neutrophils need to be elucidated urgently: Is TRAIL mRNA expression regulated at a transcriptional or post-transcriptional level? Which are the signaling pathways leading to TRAIL mRNA expression or TRAIL release in response to IFN and/or BCG? How is TRAIL released from neutrophils? Where is TRAIL intracellularly processed in IFN-treated neutrophils? Which is/are the protease(s) responsible for cleaving TRAIL within the cell?
The discovery that neutrophils can produce TRAIL represents a further advancement into the knowledge regarding the capacity of neutrophils to produce cytokines [2 ]. A deeper insight into the biological role of neutrophil-derived TRAIL in vivo, other than in vitro, will tell us whether novel therapies aimed at inhibiting/potentiating this novel neutrophil function may represent a useful approach to the treatment of inflammatory, immune, or neoplastic disorders.
ACKNOWLEDGEMENTS
This work was supported by grants from Ministero dell’Istruzione, dell’Università e della Ricerca (PRIN 2003, FIRB, and 60%), Fondazione Cassa di Risparmio di Verona-Vicenza-Ancona e Belluno, and Associazione Italiana per la Ricerca sul Cancro (AIRC). I thank Veronica Huber, Cristina Tecchio, Martin Pelletier, and Patrick P. McDonald for helpful discussion.
Received October 4, 2005; revised February 9, 2006; accepted March 2, 2006.
REFERENCES
J. Exp. Med. 178,151-161-stimulated neutrophils and monocytes release a soluble form of TNF-related apoptosis-inducing ligand (TRAIL/Apo-2 ligand) displaying apoptotic activity on leukemic cells Blood 103,3837-3844-stimulated monocyte cytotoxicity J. Exp. Med. 192,1373-1380 induces TNF-related apoptosis-inducing ligand/Apo-2L-mediated antitumor activity by human monocytes following CpG oligodeoxynucleotide stimulation J. Immunol. 171,212-218 in human myeloma cells FEBS Lett. 579,6217-6222[Medline]: requirement for TNFR55 and TNFR75 for induction of apoptosis in vitro Blood 90,2772-2783
, and NF-
B signaling pathways J. Immunol. 171,1035-1041 2B enhances Th1 cytokine responses in bladder cancer patients receiving Mycobacterium bovis bacillus Calmette-Guerin immunotherapy J. Immunol. 162,2399-2405 versus interferon- and low-dose arabinosyl cytosine in chronic myeloid leukemia Blood 99,1527-1535 in A549 lung epithelial cells: enhancement of caspase-8 expression through IFN-response element Cytokine 20,283-288[CrossRef][Medline] sensitizes human hepatoma cells to TRAIL-induced apoptosis through DR5 upregulation and NF- B inactivation Oncogene 22,1653-1662[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  A. Polenghi, F. Bossi, F. Fischetti, P. Durigutto, A. Cabrelle, N. Tamassia, M. A. Cassatella, C. Montecucco, F. Tedesco, and M. de Bernard The Neutrophil-Activating Protein of Helicobacter pylori Crosses Endothelia to Promote Neutrophil Adhesion In Vivo J. Immunol., February 1, 2007; 178(3): 1312 - 1320. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||
