| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
* Human Anatomy and
Histology Sections, Department of Experimental Medicine, University of Genova, Genova, Italy;
Pathologic Anatomy and Unit of Haematopathology, Institute of Haematology and Clinical Oncology "L. & A. Seràgnoli", Bologna University, Bologna, Italy;
Laboratory of Cellular Biotechnology, Institute of Chemical and Biological Process Science, Federal Institute of Technology, Lausanne, Switzerland;
¶ National Institute for Cancer Research, Genova, Italy; and
|| DIBIT, San Raffaele Scientific Institute, San Raffaele University, Milan, Italy
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In mice, CMV infection is kept under check by the Ly49H-activating receptor expressed on NK cells that is specific for the viral protein m157 (8, 9). However, NK cells do not appear to control entirely the productive phase of murine CMV (mCMV) infection as NK cell depletion is not sufficient to sustain active infection. Instead, a crucial role for the maintenance of systemic mCMV latency appears to be exerted by CD8+ T lymphocytes (10). It is not clear whether T cell control is Ag specific as, despite the high coding ability of the hCMV genome, only few hCMV antigenic peptides are immunodominant and are presented to CD8+ T cells by MHC class I (11). This suggests that a non-MHC-restricted T cell control of hCMV infection is operational.
UL18, a surface glycoprotein coded by the hCMV genome (12), is a viral homologue of MHC class I (13). However, its function is unknown. It has been proposed that UL18 engages the inhibitory NK cell receptor CD94/NKG2A, thus providing an escape mechanism to infected cells (14). However, this finding has been questioned and the result has been attributed to expression of HLA-E, the ligand of CD94/NKG2A (15, 16), on cells transfected with UL18 (3, 17). So far, no evidence of molecular interactions between UL18 and CD94 has been provided. Another study suggests an activating role for UL18 on cells of the immune system rather than an inhibitory one, as cells transfected with the UL18 gene are susceptible to NK cell-mediated lysis in vitro (17).
Expression cloning of the receptor for UL18 has revealed a novel molecule of the Ig superfamily, i.e., CD85j/LIR-1/ILT2 that binds to MHC class I (18, 19). CD85j is a transmembrane molecule with four cytoplasmic immunoreceptor tyrosine-based inhibition motifs (ITIMs) that mediate transduction of inhibitory signals (20). It is detected on the surface of a proportion of T cells (19, 20) and in the cytoplasm of all T lymphocytes (21). Cytoplasmic expression of CD85j in all CD4+ T cells has been recently confirmed (22). CD85j is functional in all T cells, independently of its surface expression (21), and it down-regulates Ag-specific functions of T lymphocytes (21, 23). Cross-linking of the receptor by specific mAbs exerts a dual effect on cytokine production. IL-10 and TGF-
that inhibit immune responses are increased, whereas cytokines supporting T cell activation such as IL-2, IFN-
, and IL-13 are reduced (24). In addition, in Jurkat cells transfected with CD85j, TCR/CD85j cross-linking induces a phosphorylation of CD85j by the p56lck kinase (25). Remarkably, CD85j is not phosphorylated following cross-linking of this receptor alone (25). This suggests that, in the absence of an activating signal able to recruit p56lck, CD85j does not transduce inhibitory signals.
Expression of UL18 was investigated in this study and it was detected in tissues from immunocompromised patients with productive hCMV infection, as well as in fibroblasts infected in vitro. Therefore, we searched for a role of UL18 in hCMV infection. We found that binding of UL18 to T lymphocytes via CD85j, in the absence of CD3/TCR involvement, leads to activation and not inhibition of T cells and to the subsequent lysis of hCMV-infected cells. In vitro expanded CD8+ T cell clones as well as resting or activated peripheral blood CD8+ T cells lysed efficiently hCMV-infected and UL18+ cells, irrespective of their Ag-specificity, in a non-MHC-restricted fashion. Lysis results from an interaction between monomorphic proteins, namely CD85j on T lymphocytes and UL18 expressed by infected cells. These data indicate that an invariant receptor expressed on CD8+ T cells mediates the specific recognition of an infectious Ag component, and provide a novel example that extends the characteristics of innate immunity to cells that are mainly involved in adaptive immunity.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following Abs were used: anti-CD4, anti-CD8, anti-CD16, anti-CD19, anti-CD31, anti-CD44, anti-CD54, anti-CD58, anti-CD69, anti-IL-2, anti-IFN- mAb (BD PharMingen, Hamburg, Germany), anti-CD3 (clone OKT3; American Type Culture Collection (ATCC), Manassas, VA), anti-MHC class I mAb (clone A2, produced in our laboratory), anti-CD85j clone HP-F1 (kindly provided by M. Lopez-Botet, Servicio de Immunologia, Hospital Universitario de la Princesa, Madrid, Spain), anti-hCMV UL44 (DAKO, Milan, Italy), and anti-UL18 (clone 10C7; ATCC).
CD8+ T cell clones were produced by PHA stimulation and by specific EBV-Ag selection as described previously (21, 23, 24). HeLa and Jurkat cells were provided by ATCC. Human fibroblasts (hF) were derived from bone marrow, and stromal cells were obtained from iliac crest marrow aspirates from healthy donors enrolled for bone marrow transplantation at the S. Martino Hospital and the G. Gaslini Pediatric Hospital (Genoa, Italy). Donor age ranged between 3 and 50 years and informed consent was obtained; institutional ethical committees approved all procedures. Bone marrow and stromal cell cultures were performed as described (26). Mononuclear cells were plated at 2–5 x 106/100 mm dish in Coon’s modified Ham’s F-12 medium supplemented with 10% FCS. Half of the plates were cultured in the presence of 1 ng/ml human recombinant fibroblast growth factor-2 (Austral Biologicals, San Ramon, CA). The medium was replaced after 3 days and then twice a week.
CD8+ resting T lymphocytes were obtained from PBMC by negative immunomagnetic cell sorting (MACS; Miltenyi Biotec, Auburn, CA), with depletion of CD4+, CD19+, and CD16+ cells. Enrichment was tested by FACS analysis (>95% CD8+) and cells were plated in complete medium supplemented with human AB serum alone (resting CD8+ T cells) or with the addition of recombinant human (rh)IL-2 (100 U/ml) and of immobilized OKT3 mAb (activated CD8+ T cells). After 48 h, cells were harvested and used in a cytotoxicity assay (see below).
The hCMV strain AD169 was purchased from ATCC. 
UL18 and UL16 hCMV viral mutants and vaccinia virus (VV2 and VV103) containing genes coding for 2-microglobulin and UL18, respectively, were selected as described (27) and kindly provided by H. Browne (Division of Virology, Department of Pathology, University of Cambridge, Cambridge, U.K.).
Virus infection
In all of the experiments, fibroblasts were infected at a multiplicity of infection of 10. After adsorption of the virus for 1 h at 37°C, the inoculum was removed and fresh medium was added. In some experiments, HeLa cells were infected with vaccinia virus at a multiplicity of infection of 5. After adsorption of the virus for 1 h at 37°C, the inoculum was removed and fresh medium was added. Infected HeLa cells were used for immunostaining and cytotoxicity experiments 2 days after infection.
Immunohistochemical localization of UL18
hCMV-infected cells in two interstitial pneumonia autopsy specimens and one colon biopsy were analyzed for the expression of UL44 and UL18 by immunohistochemistry. One lung specimen was obtained from a patient (age 20 years) who underwent bone marrow transplantation for the treatment of T cell lymphoma. The second lung specimen was derived from a spontaneous abortion of a 25-wk fetus affected by hCMV infection. The colon biopsy was from a patient (age 53 years) who underwent bone marrow transplantation for the treatment of renal cancer and was affected by ulcerative hCMV infection of the bowel.
Abs for immunohistochemical analyses were anti-UL44 (DAKO) and anti-UL18 (clone 10C7; ATCC). Tissue fragments were fixed with 2% paraformaldehyde and paraffin embedded. Five micrometer thick sections were treated in a microwave oven four times with citrate buffer (pH 6) for 5 min at 960 W. Sections were saturated with 10% BSA in PBS with 0.1% Triton X-100 and incubated overnight at 4°C in a humidified chamber with the specific primary Ab. The reaction was developed after addition of a secondary goat anti-mouse antiserum (Southern Biotechnology Associates, Birmingham, AL) according to the alkaline phosphatase-anti-alkaline phosphatase technique, and stained with Fast Red TR (DAKO) (28). Controls were provided by samples incubated with isotype-matched mAb (anti-CD31) + goat anti-mouse antiserum or with the secondary reagent alone (see Results).
Cytotoxicity assays
Cytotoxicity tests were performed as described (21, 23), using as targets hF, the murine mastocytoma cell line P815, and HeLa cells uninfected or after viral infection. Target cells were labeled for 1 h with 51Cr (Amersham Biosciences, Buckinghamshire, U.K.) in 96-well plates. CD8+ T cell clones or freshly isolated CD8+ T lymphocytes were the effector cells at an E:T ratio of 10:1 (with some exceptions, as indicated).
Biochemical analyses
Cell surface proteins were labeled with NHS-LC-biotin (Pierce, Helsingborg, Sweden) as described (29), using cells growing in suspension or adherent (30). Metabolic labeling was performed by incubating cells for 1 h in Met-Cys free culture medium followed by 1 h in [35S]Met-Cys 3.7 MBq/ml (Amersham Biosciences). Cells were lysed with RIPA lysis buffer (10 mM NaH2PO4, 1 mM EDTA, 1 mM EGTA, 1 mM NaF, with 150 mM NaCl, 1% Triton X-100, protease inhibitors) (complete mini protease inhibitory cocktail; Roche, Mannheim, Germany) and nuclei were discarded by centrifugation at 400 x g. After two 30-min preclearing cycles with GammaBind protein G-Sepharose beads (Amersham Biosciences), specific absorption was performed by incubating cell extracts with anti-UL18 mAb, HP-F1 mAb, or UL18Fc protein (10 µg/ml respectively) and 20 µl protein G-Sepharose beads. Incubation was extended from 3 h to overnight at 4°C with rotation. Sepharose was thoroughly washed with lysis buffer, and bound material was eluted with SDS-PAGE sample buffer. Immunoprecipitated proteins were fractionated by SDS-electrophoresis in 10% polyacrylamide gels, unless otherwise indicated, under reducing conditions. Biotinylated samples were analyzed by Western blot using HRP-conjugated streptavidin, according to standard procedures. Polyacrylamide gels with metabolically labeled samples were soaked in Amplify (Amersham Biosciences), dried, and autoradiographed. CD85j phosphorylation on tyrosines was triggered and evaluated as described (25).
Protein identification by matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) peptide mass fingerprinting was performed as follows. Bands of interest were excised from the gel, cut in 
1 x 1 mm pieces, transferred to a 0.5-ml Eppendorf tube, rinsed with water and reduced, alkylated, and digested overnight with bovine trypsin as described elsewhere (31). Briefly, 1 µl of the digestion supernatant containing the generated tryptic peptides was loaded onto the MALDI target using the dried droplet technique and 
-cyano-4-hydroxycinnamic acid as matrix. MALDI mass spectroscopy measurements were performed on a Voyager-DE STR (Applied Biosystems, Framingham, MA) TOF mass spectrometer and processed via the Data Explorer software (32). Proteins were unambiguously identified by searching against a comprehensive nonredundant sequence database using the program ProFound (27).
Transfectants and Fc fusion protein
The region of ILT2 gene coding for the CD85j receptor was amplified by RT-PCR, cloned in the pRc/CMV vector (Invitrogen, San Diego, CA), and sequenced. This construct (pCMV/ILT2) was subsequently mutated to obtain an ILT2 mutant lacking all four ITIM motifs. To this end, a stop codon was generated by mutation of the TAT triplet coding for Y533 to TAA (pCMV/ILT2-STOP) using the QuickChange XL Site-Directed Mutagenesis kit (Stratagene, La Jolla, CA). The construct was sequenced to assess the presence of the mutation. The constructs obtained and the empty vector (pCMV) were transfected into Jurkat cells (J-pCMV) by electroporation using a multiporator (Eppendorf, Milan, Italy). Stable transfectants were obtained by selection with 500 µg/ml Geneticin (Invitrogen). Jurkat cells transfected with pCMV/ILT2 (J-ILT2) and with pCMV/ILT2-STOP (J-STOP) were cloned and a positive clone for each transfectant was selected.
The UL18Fc recombinant protein was obtained by cloning the hCMV UL18 gene-coding region, from AA.16 to AA.318, into the pcDNAspIg2.0 vector that contains the CD33 leader sequence and the genomic Fc portion (CH2 and CH3) of human IgG1. Transient expression of UL18Fc in suspension-adapted HEK293-EBV-encoded nuclear Ag cells was determined as described (33). The UL18Fc protein was purified from cell-free supernatants by chromatography on Streamline protein A (Amersham Biosciences).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
To investigate the role of UL18 in hCMV infection, two lung autopsy specimens from patients with interstitial pneumonia and lethal productive hCMV infection were analyzed for the expression of this protein in vivo. Both samples displayed multiple foci of cytomegalic cells that stained strongly for the hCMV Ag UL44, an early DNA-binding protein (data not shown), and for UL18 (Fig. 1, a–c). In addition, UL18 was detected in a colon biopsy specimen from a patient with ulcerative intestinal infection after allogeneic bone marrow transplantation (Fig. 1d). Of note, in this patient, in addition to cytomegalic interstitial cells present in the submucosa, UL18 was also expressed in the cytoplasm of endothelial cells (Fig. 1, e and f). An isotype matched mAb, anti-CD31, used as control yielded a negative result. These observations indicate that cells infected by wild-type hCMV express UL18 in vivo.
|
|
Infected fibroblasts were the target cells in cytotoxicity assays using allogeneic (Fig. 2a) or autologous (Fig. 2b) CD8+ T cell clones as effectors. The specificity of the clones is unrelated to hCMV Ags as clones AK2 and BE1, derived from two distinct donors, were specific for autologous B-EBV cell lines, whereas clones PG 1.14 and PG 10B5 were obtained following PHA stimulation. In both instances, T cell clones lysed fibroblasts significantly, starting 4 days after infection, and the peak of lysis was concomitant with the highest level of UL18 surface expression, 6 days after infection. No lysis occurred when noninfected fibroblasts were the target cells.
To assess the ability of cytolytic T lymphocytes to lyse additional allogeneic hCMV-infected fibroblasts, clones BE1 and AK2 were used as effectors in a cytotoxicity assay against allogeneic hF, from several donors, 6 days after hCMV infection. Comparable levels of cytotoxicity were observed consistently when fibroblasts were infected with hCMV (data not shown). These data indicate that target cell lysis is non-MHC-restricted.
Next, the ability of 49 cytotoxic T cell clones, derived from 12 donors, to lyse allogeneic fibroblasts 6 days after hCMV infection was evaluated. Possibly none of these clones was specific for hCMV Ags, as 39 of them were produced following PHA-stimulation and 10 clones were specific for autologous B-EBV cells. This is reinforced by the finding that no lysis occurs on day 2 postinfection. Lysis of fibroblasts, 6 days after infection, was significant (>20%) for 41 T cell clones in comparison with lysis of noninfected fibroblasts (Fig. 2c). Fibroblasts were derived from seven donors and all exhibited similar levels of susceptibility to lysis.
To determine whether this was due to interaction between CD85j and UL18, mAb specific for these molecules were included in the cytolytic assay (Fig. 2). In the presence of either mAb, 6 day-infected fibroblasts were not lysed. Inhibition did not occur significantly before day 6 (see Discussion). Moreover, an irrelevant isotype-matched mAb (anti-CD44) had no effect on cytotoxicity.
The level of hCMV-infected target cell lysis is neither related with surface expression of CD85j detected using the HP-F1 mAb, nor with the cytotoxic function of the clones determined in a redirected killing assay in the presence of anti-CD3 mAb (Table I). Clones AK5 and CO1 that lyse <20% hCMV-infected targets exhibited a surface expression of CD85j and a redirected cytotoxicity comparable to that of clones lysing >20% hCMV-infected cells (Table I).
|
To further define a role of UL18 for the susceptibility to lysis of hCMV-infected cells, vaccinia viruses containing UL18 (VV103) or 2-microblobulin (VV2) genes, respectively, were used to infect HeLa cells. Expression of UL18 in VV103-infected cells was assessed by surface staining (Fig. 3a) and by immunoprecipitation with the anti-UL18 mAb 10C7 of either metabolically labeled (Fig. 3b) or biotin surface-labeled cells (Fig. 3c). As expected, cells infected with VV2 failed to express UL18. Vaccinia-infected HeLa cells were also tested in a cytolytic assay using 20 CD8+ T cell clones randomly selected among the 41 clones that were strongly cytotoxic for hCMV-infected fibroblasts. All clones lysed significantly HeLa cells expressing UL18, but not VV2-infected cells, and lysis was inhibited by anti-UL18 or anti-CD85j mAb (Fig. 3d).
|
2- (Fig. 3a) or of noninfected HeLa cells. Thus, although it has been suggested that killing of hCMV-infected fibroblasts by NK cells is, at least in part, related to increased expression of CD54 (17), lysis in our model is attributable exclusively to UL18. Additional evidence for the ability of UL18 to induce T cell-mediated lysis of hCMV-infected cells has been provided by the use of a virus deleted for the UL18 gene (UL18), or for another hCMV gene (UL16). Fibroblasts infected with UL16 or AD169 virus were susceptible to lysis whereas cells infected with UL18 were resistant. Coinfection of fibroblasts with UL18 and vaccinia virus VV103 containing the UL18 gene re-established the susceptibility of target cells to lysis (Fig. 3e). Resting and activated PBMC-derived CD8+ T lymphocytes lyse UL18+ HeLa cells
To assess the role of resting CTL for the control of hCMV infection, CD8+ T lymphocytes were purified from peripheral blood cells by subtraction of CD4+, CD16+, and CD19+ lymphocytes. The remaining cells were >95% CD3+CD8+ and were used as effector cells in cytolytic assays where the targets were uninfected HeLa cells or cells infected with vaccinia viruses VV103 or VV2 containing the UL18 or 2-microglobulin genes, respectively. In this assay, resting CD8+ lymphocytes lysed VV103-infected HeLa cells but failed to kill uninfected or VV2-infected targets. The level of lysis was increased when CD8+ T lymphocytes were cultured for 48 h with rhIL-2 or activated by immobilized anti-CD3 mAb (Table II). Clones RP1.19 and BE1 were used in these experiments as positive controls.
|
As CD85j is the only human receptor known to recognize UL18, the HP-F1 mAb was included in all cytotoxicity assays. As mentioned above, inhibition of UL18-expressing cell lysis by the HP-F1 mAb indicates an involvement of this receptor. To assess whether UL18 could recognize molecules other than CD85j on T cells, a soluble recombinant UL18Fc protein was produced and used to precipitate receptors from human T lymphocytes. HP-F1 and UL18Fc immunoprecipitates from biotin surface-labeled T cell clones and CD85j-transfected Jurkat cells were compared. As shown in Fig. 4a, the two reagents yielded an identical pattern. Moreover, sequential immunoprecipitation performed on clone BE1 demonstrated that no HP-F1 reactive molecules were left after UL18Fc immunoprecipitation and vice versa (Fig. 4c). These data suggest that UL18Fc and the HP-F1 mAb challenge the same molecular species on the surface of T cells.
|
The identity of the 120-kDa and of the 95-kDa molecular species was then investigated by MALDI-TOF peptide mass mapping. Both bands identified the ILT2 protein, with 7 peptides covering 13% of the sequence with a mass accuracy within 40 ppm for the upper band and 15 peptides covering 29% of the sequence with a mass accuracy within 40 ppm for the lower band. Therefore, the 95-kDa band is the ILT2 protein, possibly lacking part or all of the ITIMs-bearing region. In addition, we analyzed the HP-F1 reactive molecule expressed by Jurkat cells transfected with a CD85j mutant without ITIMs (J-STOP) as control. It yielded a 100-kDa protein (Fig. 4a), not phosphorylated on tyrosines following pervanadate treatment of the cells (Fig. 4b).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
60% of the individuals (5). Consequently, in addition to mechanisms that allow evasion from immune recognition, hCMV must have developed a counterbalance strategy that provides control of the infection aimed at preventing host death. A relevant point in our work is the demonstration that UL18 expression leads to lysis of infected cells and may thus be responsible for the control of infection in vivo. This would explain why wild-type viruses express UL18, although this protein is not directly involved in viral replication and infectivity in vitro (27). We suggest that the control of infection occurs via interactions between CD85j/LIR-1/ILT2 on cytolytic T cells and UL18 expressed by target cells that leads to lysis of hCMV-infected cells. Because CD85j and UL18 are monomorphic proteins, lysis does not require involvement of TCR and MHC-restriction. The large majority of cytolytic T cell clones display this unrestricted function. Resting and activated CD8+ T lymphocytes derived from peripheral blood are also able to lyse hCMV-infected cells. Altogether, the data suggest that, in vivo, the number of T lymphocytes is the major control mechanism for the onset of productive infection.
It is of note that the immune system of the patients studied for the in vivo expression of UL18 was strongly impaired. Two adults had undergone bone marrow transplantation less than a month earlier and were severely immunocompromised. Accordingly, the 25-wk-old fetus had an immune system not yet fully developed and thus unable to control hCMV infection. To reinforce this contention, immunohistochemical analyses of autopsy or biopsy specimens from these patients for the expression of CD3 failed to detect infiltrating T lymphocytes (data not shown). A recent report on allogeneic peripheral blood stem cell transplant patients, with or without hCMV viremia or disease, further supports the relevance of T cell numbers. The study shows that viremic patients who had progressed to hCMV disease displayed, on day 60 after transplantation, CD8+ T cell counts significantly lower than those of nonviremic patients or of viremic patients who had not progressed to disease (35).
Activation via CD3/TCR and addition of rhIL-2 enhanced the ability of T cells to kill infected cells. Therefore, a role could be also envisaged for CD4+ T lymphocytes, as they could help CD8+ T cells by producing IL-2. Our preliminary results suggest that CD4+ T cells are activated via CD85j/UL18 and that this interaction leads to IL-2 production.
A novel mechanism of recognition and lysis of hCMV-infected cells by cytolytic T lymphocytes, mediated by two monomorphic receptors, is demonstrated in the present study. Usually, T cells recognize antigenic peptides associated with MHC molecules through clonally distributed receptors. Ag experience results in expansion of specific T cells and generation of immunological memory, a distinct feature of adaptive immunity. Our data indicate that an invariant receptor expressed on CD8+ T cells is capable of specific recognition of a viral component, and provide a novel example that extends the characteristics of innate immunity to cells that are mainly involved in adaptive immunity. One mechanism mediated by monomorphic proteins expressed by T lymphocytes and target cells, respectively, has been described previously (36). In hCMV-infected cells, the NKG2D-activating receptor present on most CD8 T cells recognizes the stress-induced class I-like molecule MIC; this provides a costimulus that overcomes viral interference with MHC class I Ag presentation (36).
The non-MHC-restricted T cell-mediated cytotoxic function described in this study reaches its peak on day 6 of infection. This has not been observed in several reports where hCMV-specific T cell clones have been investigated. In these studies, however, fibroblasts were analyzed on day 2 after infection (36, 37). Lysis that we show 6 days after infection is not due to a cytopathic effect of hCMV, as fibroblast viability is >85%, and it is inhibited by anti-CD85j (HP-F1) or anti-UL18 (10C7) mAb. More importantly, infection with an hCMV strain deleted for the UL18 gene does not confer to cell targets a susceptibility to lysis by CD8+ T lymphocytes. Altogether, our data indicate a specific role of UL18 for the lysis of hCMV-infected fibroblasts.
Interaction between CD85j and UL18 might not be the only non-MHC-restricted mechanism involved in hCMV-infected cell lysis. Lysis inhibition by addition of HP-F1 or anti-UL18 mAb has been particularly effective on day 6 of infection and only a weak inhibition by these mAb was observed on days 4 and 5. It has been suggested that lysis of hCMV-infected fibroblasts by NK cells, in addition to a direct involvement of UL18, is at least in part due to enhanced expression of CD54 (17). Up-regulation of CD58 during infection by some hCMV strains has been considered as a possible mechanism of NK cell-mediated lysis (38). Although increased expression of CD54 and CD58 was not confirmed in another study (39), an explanation could be provided by the expression kinetics of these two molecules, that is highest between day 3 and 5 postinfection (33).
Lysis of fibroblasts 6 days after infection is due to expression of UL18 as shown by its abrogation in the presence of anti-UL18 mAb or when UL18 hCMV is used to infect target cells. Thus, in the early phase of infection, when inhibition by anti-CD85j or anti-UL18 mAb is incomplete, CD54 and CD58 could be operational in T cell-mediated lysis.
It remains to be determined how CD85j, despite its inhibitory properties, can mediate activating signals. In this context, it is of note that UL18 displays a much higher affinity for CD85j than MHC class I (40, 41); thus, it might compete with CD85j-MHC class I interactions that yields inhibitory signals. This paradox has already been described for other receptors. Indeed, the stem cell factor receptor, the erythropoietin receptor, the -chain of the IL-3 receptor, the NKp44 receptor, and the CD66a adhesion molecule are involved in cell activation despite possessing ITIM sequences in their cytoplasmic regions (42, 43, 44, 45, 46). Two possible nonmutually exclusive models can be considered. The activation may depend on a truncated form of CD85j lacking the cytoplasmic tail and thus becoming unable to transduce inhibitory signals via ITIMs. Accordingly, an alternatively spliced form of CD85j, with a truncated cytoplasmic tail lacking ITIMs (ILT2c), has been described (20). It has also been described recently that, in a murine model, CMV infection is kept under control by an activation receptor of the Ly49 gene family that encodes for both activating and inhibitory NK cell receptors (8, 9). In the second model, according to the observation that CD85j is not phosphorylated by cross-linking of the receptor alone (25), it is possible that, in the absence of an activation signal able to recruit p56lck, CD85j does not transduce an inhibitory signal. In this context, it is relevant that UL18 is recognized by CD85j independently of TCR.
In conclusion, our data suggest that the strategy used in vivo by the hCMV to ensure survival of the host and, hence, its spreading in the population, is mediated by activation signals triggered by the infected cells when their surface receptor UL18 binds to CD85j on T lymphocytes. As both receptors are monomorphic, one could envisage immunotherapeutic interventions against productive hCMV infection using infusions of allogeneic T lymphocytes from MHC class I-matched donors. To avoid graft-vs-host disease, donor lymphocytes could be genetically manipulated to express drug-inducible suicide genes (47).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 D.S. and F.G. contributed equally to this work.
3 Address correspondence and reprint requests to Dr. Ermanno Ciccone, Department of Experimental Medicine, Human Anatomy Section, University of Genova, Via De Toni 14, 16132 Genova, Italy. E-mail address: cicc{at}unige.it
4 Abbreviations used in this paper: hCMV, human CMV; hF, human fibroblasts; ITIM, immunoreceptor tyrosine-based inhibition motif; MALDI-TOF, matrix-assisted laser desorption ionization-time of flight; mCMV, murine CMV; rh, recombinant human.
Received for publication December 15, 2003. Accepted for publication February 23, 2004.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 microglobulin. Nature 347:770.[Medline]
T cells by NKG2D via engagement by MIC induced on virus-infected cells. Nat. Immunol. 3:255.
-chain and down-regulates IL-3-induced tyrosine phosphorylation and mitogenesis. Mol. Cell. Biol. 13:7577.This article has been cited by other articles:
|
|
 |
|
  M. Occhino, F. Ghiotto, S. Soro, M. Mortarino, S. Bosi, M. Maffei, S. Bruno, M. Nardini, M. Figini, A. Tramontano, et al. Dissecting the Structural Determinants of the Interaction between the Human Cytomegalovirus UL18 Protein and the CD85j Immune Receptor J. Immunol., January 15, 2008; 180(2): 957 - 968. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Maffei, F. Ghiotto, M. Occhino, M. Bono, A. De Santanna, L. Battini, G. L. Gusella, F. Fais, S. Bruno, and E. Ciccone Human Cytomegalovirus Regulates Surface Expression of the Viral Protein UL18 by Means of Two Motifs Present in the Cytoplasmic Tail J. Immunol., January 15, 2008; 180(2): 969 - 979. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. S. Wagner, L. Walther-Jallow, E. Buentke, H.-G. Ljunggren, A. Achour, and B. J. Chambers Human cytomegalovirus-derived protein UL18 alters the phenotype and function of monocyte-derived dendritic cells J. Leukoc. Biol., January 1, 2008; 83(1): 56 - 63. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. Wagner, H.-G. Ljunggren, and A. Achour Immune Modulation by the Human Cytomegalovirus-Encoded Molecule UL18, a Mystery Yet to Be Solved J. Immunol., January 1, 2008; 180(1): 19 - 24. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. S. Wagner, G. C. Riise, T. Bergstrom, K. Karre, E. Carbone, and L. Berg Increased Expression of Leukocyte Ig-Like Receptor-1 and Activating Role of UL18 in the Response to Cytomegalovirus Infection J. Immunol., March 15, 2007; 178(6): 3536 - 3543. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Iannello, O. Debbeche, E. Martin, L. H. Attalah, S. Samarani, and A. Ahmad Viral strategies for evading antiviral cellular immune responses of the host J. Leukoc. Biol., January 1, 2006; 79(1): 16 - 35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Griffin, E. C. Y. Wang, B. P. McSharry, C. Rickards, H. Browne, G. W. G. Wilkinson, and P. Tomasec Characterization of a highly glycosylated form of the human cytomegalovirus HLA class I homologue gpUL18 J. Gen. Virol., November 1, 2005; 86(11): 2999 - 3008. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. E. Kirwan and D. N. Burshtyn Killer Cell Ig-Like Receptor-Dependent Signaling by Ig-Like Transcript 2 (ILT2/CD85j/LILRB1/LIR-1) J. Immunol., October 15, 2005; 175(8): 5006 - 5015. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Tenca, A. Merlo, E. Merck, E. E. M. Bates, D. Saverino, R. Simone, D. Zarcone, G. Trinchieri, C. E. Grossi, and E. Ciccone CD85j (Leukocyte Ig-Like Receptor-1/Ig-Like Transcript 2) Inhibits Human Osteoclast-Associated Receptor-Mediated Activation of Human Dendritic Cells J. Immunol., June 1, 2005; 174(11): 6757 - 6763. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Vales-Gomez, M. Shiroishi, K. Maenaka, and H. T. Reyburn Genetic Variability of the Major Histocompatibility Complex Class I Homologue Encoded by Human Cytomegalovirus Leads to Differential Binding to the Inhibitory Receptor ILT2 J. Virol., February 15, 2005; 79(4): 2251 - 2260. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and dashes indicate the upper and lower 95% confidence interval. Wilcoxon test has been used for statistical analysis.
