当前位置: 首页 > 医学版 > 期刊论文 > 基础医学 > 病菌学杂志 > 2005年 > 第9期 > 正文
编号:11200459
Highly Protective In Vivo Function of Cytomegalovi
http://www.100md.com 病菌学杂志 2005年第9期
     Institute for Virology

    Tumor Vaccination Centre at the III. Medical Clinic, Johannes Gutenberg-University, Mainz, Germany

    ABSTRACT

    Reconstitution of antiviral CD8 T cells is essential for controlling cytomegalovirus (CMV) infection after bone marrow transplantation. Accordingly, polyclonal CD8 T cells derived from BALB/c mice infected with murine CMV protect immunocompromised adoptive transfer recipients against CMV disease. The protective population comprises CD8 T cells with T-cell receptors (TCRs) specific for defined and for as-yet-unknown viral epitopes, as well as a majority of nonprotective cells with unrelated specificities. Defined epitopes include IE1/m123 and m164, which are immunodominant in terms of the magnitude of the CD8 T-cell response, and a panel of subordinate epitopes (m04, m18, M45, M83, and M84). While cytolytic T-lymphocyte lines (CTLLs) were shown to be protective regardless of the immunodominance of the respective epitope, the individual contributions of in vivo resident epitope-specific CD8 T cells to the antiviral control awaited investigation. The IE1 peptide 168-YPHFMPTNL-176 is generated from the immediate-early protein 1 (IE1) (pp89/76) of murine CMV and is presented by the major histocompatibility complex class I (MHC-I) molecule Ld. To quantitate its contribution to the protective potential of a CD8-T memory (CD8-TM) cell population, IE1-TCR+ and IE1-TCR– CD8-TM cells were purified by epitope-specific cell sorting with IE1 peptide-loaded MHC-immunoglobulin G1 dimers as ligands of cognate TCRs. Of relevance for clinical approaches to an adoptive cellular immunotherapy, sorted IE1 epitope-specific CD8-TM cells were found to be exceedingly protective upon adoptive transfer. Compared with CTLLs specific for the same epitope and of comparable avidity and TCR ?-chain variable region (V?)-defined polyclonality, sorted CD8-TM cells proved to be superior by more than 2 orders of magnitude.

    INTRODUCTION

    The Ld-restricted immediate-early 1 (IE1) peptide 168-YPHFMPTNL-176 of murine cytomegalovirus (mCMV) was the first antigenic peptide to be identified for a herpesvirus (50). The IE1 protein derived from open reading frame (ORF) m123, an intranuclear phosphoprotein which exists in molecular species of 89 and 76 kDa (32), is expressed in the IE phase of viral gene expression and performs regulatory and transactivating functions (12, 31, 39). It is encoded in transcription unit ie1/3 of which mRNAs specifying proteins IE1 (encoded by exons 2, 3, and 4) and IE3 (encoded by exons 2, 3, and 5) are generated by differential splicing (31, 39). The IE1 protein is processed to yield peptide 168-YPHFMPTNL-176 and an N-terminally elongated precursor 166-DMYPHFMPTNL-176 by the constitutive proteasome and, more efficiently, by the immunoproteasome, with only the precursor being translocated into the lumen of the endoplasmic reticulum for major histocompatibility complex class I (MHC-I) loading (36). N-terminal trimming finally leads to the IE1 peptide presented by the MHC-I molecule Ld (reviewed in reference 45).

    Based on the frequency of IE1 epitope-specific CD8 T cells primed during acute infection and on the establishment of long-term IE1-specific memory, the IE1 peptide was classified as one of just two immunodominant MHC-I-restricted antigenic peptides in the H-2d haplotype (25). Owing to MHC polymorphism, it is evident that immunodominance of peptides and the proteins from which they are derived cannot be extrapolated to other haplotypes in the same animal species. This is all the more true for extrapolation from mCMV to antigenic peptides of human cytomegalovirus (hCMV) presented by HLA molecules. Thus, although an early report by Borysiewicz et al. (4) had indicated the CD8 T-cell immunogenicity of the hCMV IE1 ortholog, the predictive value of the BALB/c mouse model in terms of the role of IE proteins in immunity to CMVs has been debated for a long time, until more recently the IE1 immunogenicity in humans was revisited with unbiased new methodology (10, 14, 33, 34). Most intriguingly, in a comprehensive pangenomic search for antigenic ORFs of hCMV by using overlapping peptides and by covering all major HLA molecules present in the human population, Louis Picker and colleagues identified ORF UL123 encoding the IE1 protein as one of the top three antigen-encoding ORFs of hCMV that are most frequently detected by HLA class I-restricted human CD8 T cells (L. Picker, "T cell recognition of hCMV in natural human infection: pan-genome analysis of immunogenic open reading frames," Instituto Juan March de Estudios e Investigaciones workshop "Immunodominance: the key to understanding and manipulating CD8+ T cell responses to viruses," Madrid, Spain, 2004; L. Picker, personal communication). Thus, in this specific aspect of immunity to CMVs, the BALB/c mouse model matches the situation found in approximately half of all hCMV-infected individuals; from this novel knowledge, the BALB/c mouse model gains increasing importance as an in vivo experimental model for analyzing the contribution of IE1-specific CD8 T cells to the control of CMV disease.

    At first glance, one may argue that this issue is long settled, as several previous approaches, including immunization with vaccinia virus recombinants expressing the IE1 protein (29) or the isolated IE1 peptide (6), immunization with synthetic IE1 peptide (55), genetic immunization with IE1-expressing plasmids (11), and adoptive transfer of IE1 epitope-specific cytolytic T-cell lines (CTLLs) (24, 25), have bona fide demonstrated an in vivo protective antiviral effector function of IE1-specific CD8 T cells. Since polyclonal CD8 T-cell populations isolated from BALB/c mice during acute or latent infection from lymphoid organs (47, 49, 52) or from infiltrates at tissue sites of CMV disease (43) were found to be protective too, it was somehow logical to take it for granted that IE1-specific CD8 T cells are a major component of the in vivo CD8 T-cell response that resolves productive infection and prevents multiple-organ CMV disease (44). However, in a formal scientific sense, such an extrapolation is no proof. So, one important link was actually missing in the chain of evidence, namely, the direct demonstration of a protective antiviral function of ex vivo-isolated IE1-specific CD8 T cells.

    Here, we have combined T cell receptor (TCR)-based epitope-specific cytofluorometric cell sorting and adoptive immunotherapy of CMV infection to compare the protective function of IE1-TCR+ CD8 T cells with that of the remaining IE1-TCR– population comprising CD8 T cells specific for m164, the second immunodominant epitope (25), and CD8 T cells specific for the panel of known subordinate epitopes m04, m18, M45, M83, and M84 (45). The data show for the first time that ex vivo isolated IE1 epitope-specific memory CD8 T cells (CD8-TM cells) are highly protective.

    MATERIALS AND METHODS

    Infection and adoptive cell transfer. For immunological priming of CD8 T-cell donors, 8- to 11-week-old, immunocompetent, female BALB/cJ mice (H-2d haplotype) were infected in the left hind footpad (intraplantar infection) with 105 PFU of cell culture-propagated, sucrose gradient-purified (42), wild-type mCMV (mCMV-WT), strain Smith ATCC VR-194, recently reaccessioned as VR-1399. For adoptive transfer experiments, BALB/cJ mice (see above), referred to as transfer recipients, were immunocompromised by hematoablative irradiation with a single dose of 6 Gy delivered by a 137Cs -ray source. IE1 epitope-specific CD8 T cells, either ex vivo-sorted memory cells or effector cells of a polyclonal CTLL (IE1-CTLL; see below), were transferred 6 h later by intravenous infusion in 0.5 ml of physiological saline. Intraplantar infection was performed 2 h later with 105 PFU of mCMV-WT as previously described (42). BALB/cJ mice were bred and housed under specified pathogen-free conditions in the Central Laboratory Animal Facility of the Johannes Gutenberg-University, Mainz, Germany. Animal experiments were approved by the Ethics Commission, permission no. 177-07/991-35 and 177-07/021-28, according to German federal law.

    Antigenic peptides. Custom peptide synthesis with a purity of >75% was performed by Jerini Peptide Technologies (Berlin, Germany). Peptides of mCMV were as follows: IE1/Ld, 168-YPHFMPTNL-176 (50); m04/Dd, 243-YGPSLYRRF-251 (24); m18/Dd, 346-SGPSRGRII-354 (20); M45/Dd, 507-VGPALGRGL-515 (45); M83/Ld, 761-YPSKEPFNF-769 (22); M84/Kd, 297-AYAGLFTPL-305 (26); and m164/Dd, 257-AGPPRYSRI-265 (25). The Ld-restricted peptide 118-RPQASGVYM-126 derived from the nucleoprotein of lymphocytic choriomeningitis virus (LCMV NP118-126) (57) was used as a control peptide for peptide loading of Ld dimers and for the folding of Ld tetramers.

    MHC tetramers and dimers. Cell labeling with peptide-loaded MHC dimers or tetramers was performed for 45 min at 4°C in the dark.

    (i) Tetramers. Custom synthesis of peptide-folded phycoerythrin (PE)-conjugated MHC-I (Ld) tetramers was performed by ProImmune Limited (Oxford, Great Britain). Tetramers were provided purified by gel chromatography to remove excess free peptide. Tetramers used in this study were PE-LdTetra-IE1mCMV and PE-LdTetra-NPLCMV. As recommended by the manufacturer, tetramers were used to label cells in a saturating dose of 1 to 2 μg per 106 cells in a volume of 50 μl.

    (ii) Dimers. Dimer X I (H-2Ld:Ig), here briefly referred to as a dimer, was purchased from BD Biosciences Pharmingen (catalog no. 550751). Dimers are hybrid molecules of 250 kDa, in which each of the two heavy chains of mouse immunoglobulin G1 (IgG1) carries a covalently linked MHC-I -chain (domains 1 to 3), of Ld in the specific case. The peptide binding pocket formed by domains 1 and 2 of the MHC-I portion of the dimer was loaded with peptide according to the protocol provided by the manufacturer. In brief, 1 mM stock solutions of nonapeptides (1 mg/ml) were prepared in 30% (vol/vol) acetonitrile. Dimers (10 μl = 5 μg) were incubated in an Eppendorf tube for 16 to 24 h at 4°C with the respective peptides. Specifically, dimers were loaded with peptides mCMV IE1 (LdDimer-IE1mCMV) or LCMV NP118-126 (LdDimer-NPLCMV) at a peptide molar excess of 40, because higher doses recommended by the manufacturer for testing (160- and 640-fold molar excess) were found to give no improvement in cytofluorometric staining. It should be noted that the presence of free peptide is becoming a problem in some types of assays. Therefore, after the loading incubation, unbound peptide (1 kDa) was removed by excessive flow dialysis with Dispo-Biodialyzer (Sigma, catalog no. D-9187) with an exclusion size of 50 kDa. The amount of peptide-loaded dimers required for an optimal staining of cells depends on the frequency of cells that express a cognate TCR in a cell population as well as on the number of TCR molecules per cell. Titration of IE1 dimers revealed an optimal staining at 5 μg per 106 cells in the case of a monospecific IE1-CTLL and 1 μg per 106 cells in the case of ex vivo isolated CD8-TM cells from the spleen. For specificity control of the staining, LdDimer-NPLCMV was used at a concentration of 10 μg per 106 cells in both cases. It should be noted that dimers not loaded with exogenous peptide can accommodate multiple endogenous peptides, which led to a somewhat higher background staining in polyspecific CD8 T-cell populations (data not shown).

    Effector cells. Memory cells were pooled from spleens of at least three BALB/cJ mice at 3 months after infection, time points at which productive primary infection was resolved. They were either used for analyses (see below) or served as a source for the generation of IE1-CTLLs.

    (i) Immunomagnetic cell sorting. For subsequent cytofluorometric analysis (with exceptions specified), for intracellular cytokine staining, and for cell sorting, CD8 T cells were enriched by one round of positive immunomagnetic cell sorting with the autoMACS system (Miltenyi Biotec Systems, Bergisch-Gladbach, Germany) following the protocol recommended by the manufacturer. For enzyme-linked immunospot (ELISPOT) assays (see below), two sequential runs of automated magnetic cell sorting (autoMACS sorting) were performed to reach >95% purity. In brief, up to 107 cells were resuspended in 90 μl of running buffer (2 mM EDTA in phosphate-buffered saline containing 0.5% [wt/vol] bovine serum albumin) and mixed with 10 μl of rat anti-mouse CD8a (Ly-2) MicroBeads (Miltenyi catalog no. 130-049-401). After 15 min of incubation at 4°C in the dark, followed by washing and resuspension in running buffer, immunomagnetic sorting was performed by using Possel and Posseld separation programs (Miltenyi) for one- and two-column separation, respectively.

    (ii) Cytofluorometric cell sorting. For the purification of IE1-specific CD8-TM (IE1-TCR+ CD8+), immunomagnetically preenriched CD8 T cells (see above) were saturated with normal goat serum (Jackson ImmunoResearch Laboratories catalog no. 005-000-121) and rat monoclonal antibody (MAb) anti-mouse CD16/CD32 (anti Fc III/II receptor, clone 2.4G2; BD Biosciences Pharmingen catalog no. 553142), and were labeled with fluorescein (isothiocyanate; FITC)-conjugated rat MAb anti-mouse CD8a (clone 53-6.7; BD Biosciences Pharmingen catalog no. 553031) and IE1-peptide-loaded dimer X I H-2Ld:Ig (see above) with PE-conjugated rat MAb anti-mouse IgG1 (clone A85-1; BD Biosciences Pharmingen catalog no. 550083) as a second antibody. The cell sorter used was an EPICS Altra HyperSort (Beckman Coulter, Fullerton, Calif.) equipped with a 100-μm nozzle and operated with EXPO32 acquisition software, version 2.0. The excitation wavelength was 488 nm. Band-pass filters for 525 and 575 nm were used to measure FITC and PE fluorescence, respectively. Cytomics RXP Analysis software, version 1.0, was used for data processing. Sort gates were set on living cells with positive FL-1 (FITC) and high FL-2 (PE) fluorescence, discarding CD8 T cells with low expression of the IE1-TCR. Sorting was performed in AltraSort mode 3, with a flow rate of ca. 5,000 cells/s. Recovered cells were collected in fetal calf serum-saturated polystyrene tubes.

    (iii) Generation of epitope-specific CTLLs and cytolytic assay. IE1-CTLLs were generated by repeated restimulation of memory cells with synthetic IE1 peptide (25). In brief, 1.5 x 107 unseparated memory spleen cells were seeded per 2-ml well (24-well flat-bottom culture plates) in 1.5 ml of clone medium supplemented with synthetic IE1 peptide in a concentration of 10–9 M. At day 4, 100 U of recombinant human interleukin-2 (IL-2) was added to 0.5 ml of fresh clone medium. The next restimulation was performed on day 7 by a 1:2 split of the cultures and a supply of 1 ml of fresh clone medium supplemented with 10–9 M of IE1 peptide, 200 U of recombinant human IL-2, and 8 x 105 -irradiated (90 Gy) P815-B7 cells (2) as stimulator cells. On day 14 and every second week thereafter, further rounds of restimulation were performed after the number of effector cells was adjusted to 5 x 105 per well. In the intervals between the restimulations, cultures were split and refed with the IL-2 medium when required. The experiments described here were performed with IE1-CTLLs propagated for three (adoptive transfer experiments) to not more than six (in vitro experiments) rounds of restimulation and 1 week after the last restimulation, so that stimulator cells were largely absent at the time of harvest. Cytolytic activity was measured by a standard 51Cr release assay with the DBA/2 mouse (H-2d)-derived P815 mastocytoma cells as target cells that were pulsed with 10–8 M IE1 peptide for measuring IE1 epitope-specific cytolytic activity or were left without exogenous peptide for measuring noncognate lysis of the target cells.

    Two-color cytofluorometric analysis of epitope-specific CD8 T cells. Cells of an IE1-CTLL or immunomagnetically enriched CD8 T cells were blocked against Fc receptor binding (as in the cell sorting protocol above) and were labeled with peptide-loaded dimers or peptide-folded PE-conjugated tetramers (see above) for 45 min at 4°C in the dark. In the case of the nonfluorochromated dimers, cells were washed, resuspended in 50 μl of cytofluorometry buffer, and stained for 15 min with PE-conjugated rat MAb anti-mouse IgG1 (see above). In both cases, after a washing step, cells were finally labeled with FITC-conjugated rat MAb anti-mouse CD8a (see above). All labeling procedures were performed on ice to minimize T-cell activation, receptor capping, and receptor internalization. The analysis was performed with a FACSort (Becton Dickinson) using CellQuest 3.3 software for data processing. Fluorescence channel 1 (FL-1) represents fluorescein fluorescence, and FL-2 represents PE fluorescence.

    Intracellular cytokine staining. Memory CD8 T cells derived from spleens of latently infected mice and IE1-CTLLs were prepurified by one round of positive MACS sorting to remove non-CD8 cells or, in the case of IE1-CTLLs, residual stimulator cells and debris. The assay was performed essentially as described in greater detail previously (22, 43) with some modification. Briefly, cells were incubated for 45 min at 4°C with IE1 dimer, tetramer, or 10–8 M peptide; washed; seeded in 0.2-ml microcultures at a concentration of 106 cells per well; and cultured at 37°C for 6 h in the presence of brefeldin A. For quantification of gamma interferon (IFN-)-expressing CD8 T cells, two-color (FL-3 versus FL-1) cytofluorometric analysis was performed with Cy-Chrome-conjugated rat anti-mouse CD8a (clone 53-6.7; BD Biosciences Pharmingen catalog no. 553034) and FITC-conjugated rat anti-mouse IFN- (clone XMG1.2; BD Biosciences Pharmingen catalog no. 554411) or, for isotype control, FITC-conjugated rat IgG1 (clone R3-34; BD Biosciences Pharmingen catalog no. 554682). Gates were set on lymphocytes and positive Cy-Chrome fluorescence (FL-3) to restrict analysis to CD8 T cells.

    Cytofluorometric analysis of V? usage. Unseparated lymphoid cells (primed lymph node cells and naive or memory spleen cells), MACS-purified CD8 T cells, or cells in the course of selection of a CTLL were tested for the expression of the TCR ?-chain variable region (V?). After cells were washed and counted, the sedimented cells were resuspended in Fc receptor blocking solution (1 μg of MAb anti-mouse CD16/CD32 per 100 μl per 106 cells; see above) for 15 min at 4°C and then distributed in Eppendorf tubes in aliquots of 1 x 106 cells each for labeling. Directly fluorochromated antibodies were purchased from BD Biosciences Pharmingen throughout: (i) PE-conjugated anti-mouse TCR ?-chain (catalog no. 553172), (ii) FITC-conjugated anti-mouse CD8a (catalog. no. 553031) or Cy-Chrome-conjugated anti-mouse CD8a (see above), (iii) Cy-Chrome-conjugated anti-mouse CD4 (catalog. no. 553050), (iv) PE-Cy5-conjugated anti-mouse CD3e (catalog. no. 553065), and (v) the mouse V? TCR screening panel (catalog. no. 557004), including FITC-conjugated MAbs directed against mouse V? chains (V?x, where x is 2, 4, 6, 7, 8.1 and 8.2, 8.3, 9, 10b, 13, or 14) that are expressed in BALB/c mice. PE-LdTetra-IE1mCMV was used for the quantitation of cells expressing IE1 epitope-specific TCRs. To determine the percentage of V?x-positive T cells among all CD8 T cells, V? screening panel antibodies were combined with anti-CD8 in two-color (FITC/FL-1 versus Cy-Chrome/FL-3) analyses. To determine the percentage of V?x-positive T cells among all T cells, V? screening panel antibodies were combined with anti-CD3 in two-color (FITC/FL-1 versus PE-Cy5/FL-3) analyses. To determine the percentage of IE1 epitope-specific T cells among all T cells, PE-LdTetra-IE1mCMV was combined with anti-CD3 in a two-color (PE/FL-2 versus PE-Cy5/FL-3) analysis. To determine the percentage of V?x-positive cells among all IE1-specific CD8 T cells, two-column MACS-purified CD8 T cells were stained for two-color (FITC/FL-1 versus PE/FL-2) analysis with the V? screening panel antibodies and PE-LdTetra-IE1mCMV. Finally, to determine the percentage of CD4 and CD8 T cells among all T cells, anti-CD4 and anti-CD8 were combined with anti-TCR ?-chain in a three-color (Cy-Chrome/FL-3 versus FITC/FL-1 versus PE/FL-2) analysis. Two-color (PE-Cy5/FL-3 versus PE/FL-2) control staining for the expression of CD3 and TCR ?-chains (data not shown) revealed the identity of the stained cells so that both markers equally defined the 100% T cells. It should be noted that the seemingly straightforward combination of measuring V?-expressing cells among all TCR ?-chain expressing cells is precluded by antibody interference at the same target molecule. Cells were incubated for 20 min at 4°C, washed, and analyzed with a FACSort (see above).

    ELISPOT assay. The ELISPOT assay for the detection of epitope-specific, IFN--secreting effector cells was performed as previously described (22, 25), with graded effector cell numbers and triplicate cultures per titration step. In essence, CD8 T cells purified by two sequential runs of autoMACS sorting or cells of an IE1-CTLL were incubated for 18 h with P815-B7 cells (2) that were loaded with the indicated peptides at a constant concentration of 10–8 M, which is a saturating concentration for all peptides used here, or with the indicated graded IE1-peptide concentrations in the dose-response assays. Spots were counted, and the frequencies of IFN- secreting, spot-forming cells were calculated by intercept-free linear regression analysis (cell numbers on the abcissa x and triplicate spot numbers on the ordinate y) with Mathematica Statistics LinearRegression software, version 4.2.1 (Wolfram Research, Inc., Champaign, Ill.). The calculation gives the slope a of the regression line (y = ax) and its 95% confidence interval as well as the P value for the null hypothesis of random distribution, which must be <0.01 for accepting a linear function. The most probable number (MPN) of IFN--secreting effector cells per 104 cells in the case of CD8-TM and per 100 cells in the case of IE1-CTLL is then the ordinate coordinate y(MPN) = ax, with x = 10,000 or x = 100, respectively. The 95% confidence interval of the MPN is calculated accordingly with the upper and lower limit values of the slope a.

    Quantitation of infection in host organs. Infectious virus, defined as the number of PFU, was quantitated in organ homogenates by a plaque assay on mouse embryonal fibroblasts by the method of centrifugal enhancement of infectivity (42). The number of infected cells in representative areas (10 and 2 mm2 for liver and adrenal glands, respectively) of 2-μm tissue sections of host organs was determined by IE1 protein (pp89/76)-specific immunohistochemistry (IHC) with the peroxidase-diaminobenzidine-nickel method (42), which results in the highly contrasted black staining of viral intranuclear inclusion bodies.

    RESULTS

    Selection of an IE1 epitope-specific polyclonal CTLL. The principal question of whether IE1 epitope-specific CD8 T cells exert a protective antiviral function in vivo has already been indirectly answered in the positive by adoptive cell transfer of IE1-CTLLs (24, 25). However, since for technical reasons ex vivo IE1 epitope-specific CD8 T cells generated during a natural, polyclonal, and polyspecific immune response to infection were not previously accessible to direct testing, in vivo antiviral efficacies could not be compared. Consequently, the influence of in vitro selection on the antiviral function during the generation of an epitope-specific CTLL was unknown.

    Here, we have undertaken every effort to keep the in vitro selection to an unavoidable minimum to generate a CTLL that is as close to its ex vivo precursors as possible. The criterium for the selection that defines the minimal, i.e., necessary, number of in vitro restimulations was an effector cell population in which all cells expressed an Ld-IE1 peptide-specific TCR, designated IE1-TCR. Criteria for the selection that define the maximal, i.e., still tolerable, number of in vitro restimulations were the maintenance of polyclonality and maintained expression of the coreceptor molecule CD8.

    The TCR ?-chain variable region (V?) expression pattern was used to evaluate gross shifts in clonal composition during CTLL selection. To get comparative figures, we first recorded the V? expression patterns for ex vivo-isolated CD8 T-cell populations of BALB/c mice (Fig. 1). At a glance, the patterns were very similar for CD8 T cells present in the spleens of adult mice naive in terms of immunity to mCMV, in draining popliteal lymph nodes at the peak of an acute immune response on day 8 after intraplantar mCMV infection, and during early memory in the spleens of mice 10 weeks after intraplantar primary infection (Fig. 1A). Thus, apparently, mCMV infection had no significant impact on overall V? usage in the CD8 T-cell pool of BALB/c mice. Since IE1 epitope-specific CD8-TM cells are only a minority constituent of primed CD8 T-cell pools, we next directly compared V? usage by all CD8 T cells and by IE1-specific CD8 T cells derived from spleens in a late memory phase 28 weeks after primary infection (Fig. 1B). Early and late memory V? usage patterns were almost identical. Importantly, the pattern for IE1 epitope-specific CD8 T cells largely mirrored the pattern observed with all CD8 T cells.

    The late memory spleen cell population was the one used as the source for the selection of IE1-CTLL (Fig. 2 ), and the V? expression pattern among all T cells was monitored during six rounds of in vitro restimulation with IE1 peptide (Fig. 2A). In essence, the V? expression pattern was quite similar to the ex vivo pattern of the starting population, with V?8.1-V?8.2 predominating throughout. Importantly, the pattern was fairly robust in the course of the restimulations. Apart from minor, barely significant fluctuation, the important information is that the IE1-CTLLs remained polyclonal during the observation period and differed little from the starting population. It must be emphasized that this relative stability was a feature repeatedly observed for IE1-specific CTLLs, whereas CTLLs of other specificities tended to show more rapid V? selection, though the expanded V? family cells can vary between individual CTLLs of the same epitope specificity (data not shown).

    The polyclonality revealed by the V? expression pattern of the IE1-CTLL does not reflect absence of epitope-specific selection. As shown in Fig. 2B, the proportion of IE1-specific cells among the CD8 T cells increased rapidly, and epitope monospecificity of the still-polyclonal CTLL was reached with the third restimulation and remained stable thereafter. Thus, three restimulations were necessary and sufficient for the generation of a monospecific IE1-CTLL. Combined, these findings show that many different TCRs recognize the Ld-IE1 peptide complex.

    Since the avidity of binding to an MHC-peptide complex is known to be enhanced by the coreceptor molecule CD8, expression of CD8 on a CTLL may have an impact on its antiviral function. As shown in Fig. 2C, CD4 T cells were lost from the starting spleen cell population, and a pure IE1-TCR+ CD8+ cell line was generated after the third restimulation. From then on, however, CD8 expression was gradually lost, yielding a subpopulation within the IE1-CTLL that displayed the phenotype IE1-TCR+ CD8–.

    According to the selection criteria defined above, three restimulations were defined for studies of in vivo antiviral efficacy as the optimal condition for the generation of a polyclonal IE1-TCR+ CD8+ line with only minimally selected V? composition.

    Frequency of IE1-specific memory CD8 T cells: comparison between tetramer and dimer staining. It is a problem inherent in TCR-based epitope-specific cell sorting that cross-linking of TCR molecules by multivalent MHC-peptide reagents leads to signal transduction that can activate effector T cells to instant delivery of effector functions and subsequent apoptosis or can induce anergy in resting T cells. This can potentially counteract the use of the sorted cells in functional assays in vitro, as well as in vivo (1, 35). On the other hand, the affinity of peptide-loaded MHC-I monomers is too low for stable binding to the TCR (3, 5, 38). In accordance with the literature on TCR binding affinities of MHC multimers (for a review, see reference 61), we reasoned that a dimeric tool, such as peptide-loaded MHC-IgG1 hybrids in which each of the two IgG1 heavy chains carries an MHC-I -chain covalently linked to its V region (27, 56), binds less tightly and causes less cross-linking than peptide-folded MHC-I tetramers or higher-order multimers. Therefore, the use of dimers may be a compromise.

    With this rationale in mind and for the specific case of the high-affinity Ld-restricted IE1 peptide, we first compared LdDimer-IE1mCMV and LdTetra-IE1mCMV for staining intensity and for the frequencies of detected IE1-TCR+ CD8+ T cells in a polyclonal but monospecific IE1-CTLL (Fig. 3A), as well as in a polyspecific CD8-TM population derived from the spleens of BALB/c mice during latent mCMV infection (Fig. 3B). For specificity control, we used LdDimer-NPLCMV and LdTetra-NPLCMV that bind only to unrelated TCRs specific for the Ld-restricted NP118-126 epitope of LCMV. At a glance, the dimer was as good as the tetramer in terms of detected frequencies of IE1-TCR+ CD8+ T cells and staining (here, PE/FL-2 fluorescence) intensity. In both instances, staining was highly specific.

    Effect of dimer and tetramer binding on TCR signaling indicated by the expression of IFN-. Intracellular cytokine staining was used to monitor the activation of CD8 T cells by ligation of their TCRs with dimers or tetramers. CD8+ CTLs of the IE1-CTLL, purified by one round of positive MACS sorting, were activated to express IFN- regardless of whether TCR ligation occurred with LdDimer-IE1mCMV or LdTetra-IE1mCMV (Fig. 4A). The level of activation was comparable to that observed after ligation with the physiological cell-bound ligand Ld-IE1mCMV provided by inter-CTL presentation of IE1 peptide added externally at an optimal concentration of 10–8 M. Notably, CD8 molecule cross-linking in the process of MACS sorting already caused an elevated basal level of IFN- expression in the absence of TCR ligation. This was not a constitutive IFN- expression in CTLL, as it was not found in control experiments in which no MACS sorting was performed (data not shown). A corresponding experiment was done for MACS-purified CD8-TM cells (Fig. 4B) with similar results, except that the memory cells did not respond to CD8 molecule cross-linking.

    In conclusion, as predicted, TCR ligation activated the cells to express IFN-. However, these experiments did not reveal a significant difference between dimers and tetramers in this respect. Since the difference in binding affinities between dimeric and higher-order multimeric ligands is not debatable (61), the data likely indicate that both reagents triggered the maximum possible level of signaling. While these results reaffirmed a potential problem of TCR-based cell sorting, they did not give preference to either of the TCR binding reagents.

    Effect of dimer and tetramer binding on the cytolytic activity of IE1-specific CTLs. We next investigated if LdDimer-IE1mCMV and LdTetra-IE1mCMV differ in their influence on the in vitro cytolytic activity of an IE1-CTLL (Fig. 5). Notably, and somewhat unexpectedly in view of the literature (35), neither of these TCR binding and cross-linking reagents inhibited the recognition of target cells that presented the IE1 peptide. This finding implies that a sufficient amount of free TCRs remained available for recognition of the Ld-IE1 peptide complexes on the target cells and that CTL apoptosis was not a critical factor during the 4-h assay period. Though most of the target cell lysis was specific in that it required presentation of the IE1 peptide, some lysis of target cells in absence of added IE1 peptide occurred after TCR-mediated signal transduction elicited by either of the two reagents. Yet, one has to consider that free residual peptide present in the reagents from the dimer-loading or tetramer-folding process and, alternatively or in addition, peptide released from the peptide-MHC complexes can bind to the class I molecules on the target cells and lead to epitope-specific recognition. This was indeed an issue in earlier experiments that had been performed with undialyzed peptide-loaded dimers (not shown). In the experiment with results shown in Fig. 5, however, the supernatant of IE1-CTLs that were stained with dialyzed IE1 peptide-loaded dimers did not contain a sufficient concentration of free IE1 peptide (>10–12 M) to render peptide-unpulsed target cells susceptible to lysis by "untouched" IE1-CTL (data not shown). A second argument considered by us was the possibility that the IgG1 Fc-portions of the TCR-bound MHC-IgG1 hybrids may bind to Fc receptors on the peptide-unpulsed target cells and thereby provide a surrogate MHC-peptide complex, bridging the target cells with the CTL. However, extensive blocking of Fc receptors on the target cells did not inhibit the lysis (results not shown). We therefore concluded that activation of the CTL by TCR-cross-linking leads to some noncognate lytic activity, possibly by "bystander effects" such as release of perforin and granzymes, due to the activation by TCR cross-linking. Regardless, this phenomenon is not a specific problem of the use of dimers, as it occurred in like manner with tetramers (Fig. 5 and reference 35).

    Dimer binding to the IE1-TCR does not interfere with antiviral in vivo function. According to the literature, binding of tetramers to TCRs can inhibit effector cell function (35). This limitation of tetramer technology in the preparative area of application originally gave the impetus for the development of the MHC-streptagII multimer system: briefly, streptamers that allow a reversible epitope-specific staining (35). They can be detached from T cells by D-biotin-facilitated monomerization, followed by rapid dissociation of low-affinity MHC-peptide monomers from the TCRs. This method was shown to preserve the phenotype and functional status of T cells (35). It also allowed effective adoptive transfer of reversibly stained CTLL cells (35) and ex vivo-sorted CD8 T cells (28). Although data also exist on the successful transfer of CD8 T cells sorted after staining with conventional tetramers, these qualitative studies were performed with high cell numbers and did not address the issue of a possible loss of efficacy (40).

    We previously established a murine model of CD8 T-cell-based preemptive cytoimmunotherapy of CMV dissemination and multiple-organ CMV disease in the immunocompromised (total-body -irradiated) host after local subcutaneous infection (45, 52). The in vivo protective antiviral activity of primed polyclonal CD8 T cells or in vitro-selected epitope-specific CTLLs is usually determined between days 11 to 13 after infection and intravenous cell transfer. In this model, infected cells become detectable in tissue sections of key target organs, such as liver and lungs, at around day 6. Thus, transferred cells have time to recover in the host and will still come in time to control organ infection. As shown in Fig. 6 for IE1-CTL labeled with LdDimer-IE1mCMV, stained IE1-TCRs are rapidly downmodulated and disappear from the cell surface after 24 h of in vitro culture. If this can be extrapolated to in vivo conditions, the bound reagent may be less of a problem than was previously surmised.

    In fact, in a pilot experiment (Fig. 7A) involving the adoptive transfer of a constant number of 5 x 105 cells of a polyclonal IE1-CTLL, the control of infection was 2 log10 PFU in the lung and 4 log10 PFU in the spleen, regardless of whether the IE1-CTL were left untouched (group I), labeled with LdDimer-IE1mCMV (group II), or incubated with the epitope-unrelated reagent LdDimer-NPLCMV (group III). In a separate experiment performed with the third-restimulation IE1-CTLL (Fig. 2), graded cell numbers either left untouched or labeled with LdDimer-IE1mCMV were transferred to reveal any differences in the dose-response curves (Fig. 7B). Yet, there was no difference observed in the lung and no significant difference in the spleen, which implies that none of the above-discussed problems of epitope-specific TCR staining are critical in this particular experimental model.

    In conclusion, binding of LdDimer-IE1mCMV to the IE1-TCRs in an IE1-CTLL did not significantly interfere with protective antiviral in vivo function upon adoptive cell transfer.

    Ex vivo sort-purified IE1 epitope-specific CD8 T cells are highly protective The promising protection data obtained with the IE1-CTLL encouraged us to use LdDimer-IE1mCMV for ex vivo cytofluorometric sorting of the 1 to 3% (Fig. 3 and 4) IE1-TCR+ CD8+ memory T cells present in the preenriched CD8 T-cell population derived from pooled spleens of latently infected BALB/c mice. It should be recalled that only 10% of the spleen lymphocytes are CD8 T cells, so that the frequency of IE1 epitope-specific CD8 T cells in the spleen is only 0.1 to 0.3% of all lymphocytes. Owing to this low absolute number, only 4 x 104 cells in a pilot experiment were available for an adoptive transfer in which each of four recipients received 1 x 104 sorted cells. Experience has shown that such a low cell dose is barely protective in the case of IE1-CTLLs (24, 25) or in CTLLs specific for a number of other mCMV epitopes (22-25). This was again confirmed by the experimental results shown in Fig. 7B. Specifically, 104 in vitro-propagated IE1-CTLs led to a control of infection of <1 log10 PFU in the lung and of <2 log10 PFU in the spleen. Based on this published and here reproduced experience with CTLL dose-response curves in the adoptive transfer model, we did not expect much from this pilot experiment and were therefore even more surprised to find a very impressive control of infection of almost 3 log10 PFU in the lung and 5 log10 PFU in the spleen (Fig. 8, left). Moreover, these cells limited the infection in liver and adrenal glands by 3 log10 to below the technical limit of IHC detection of infected cells in tissue sections (Fig. 8, right). The protective impact of this antiviral control in terms of prevention of viral histopathology is illustrated for the same experiment by the corresponding IHC images in Fig. 9. In the "no transfer" control group, extended areas of infection can be seen in whole-organ sections of the adrenal-suprarenal glands (Fig. 9A), and the liver is literally riddled with plaque-like lesions (Fig. 9B1 and B2). By contrast, transfer of 104 sorted IE1 epitope-specific CD8 T cells completely prevented viral histopathology in these two organs (Fig. 9C and 9D).

    Superior antiviral activity of IE1 epitope-specific ex vivo CD8-TM cells is not explained by higher TCR affinity or avidity. The extraordinary antiviral in vivo activity of sorted IE1-specific memory cells compared to polyclonal CTLL specific for the same epitope prompted investigation into the reason. An immediate idea was that propagation in vitro causes a loss of antigen binding avidity. Since clonotypes with high-affinity TCRs are likely to better withstand limited in vitro antigen presentation and to utilize IL-2 more efficiently, we considered it unlikely that in vitro selection favors the expansion of low-affinity clonotypes; rather, the opposite makes sense. However, downmodulation of accessory and coreceptor molecules of the immunological synapse (9) may weaken the target cell recognition. To cover both arguments, we determined by ELISPOT assay the minimal concentration of IE1 peptide that was needed to trigger TCR signaling in IE1 epitope-specific polyclonal CD8-TM ex vivo, as well as in a polyclonal third-restimulation IE1-CTLL derived thereof (Fig. 10). All IE1-specific cells in both populations were triggered to secrete IFN- by IE1 peptide concentrations of down to 10–9 M. At 10–10 M, 50% of the IE1-specific CD8-TM cells failed to respond, while all cells of the IE1-CTLL were triggered. At the limit concentration of 10–12 M, all IE1-specific CD8-TM cells failed, whereas the IE1-CTLL contained a detectable number of cells that were still capable of responding. In conclusion, as the overall TCR affinity-avidity was higher in the IE1-CTLL, this evidently is not the parameter accountable for its lower antiviral in vivo activity.

    Contribution of IE1 epitope-specific CD8 T cells to the protective capacity of a polyspecific CD8-TM population. Referring to the immune response to mCMV in BALB/c mice, two immunodominant antigenic peptides (IE1 and m164) and several subordinate antigenic peptides (m04, m18, M45, M83, and M84) have been identified (for a review, see reference 45). While the two immunodominant peptides account for most of the mCMV-specific CD8 T-cell memory in latently infected mice (25) so that we do not expect the existence of an unidentified third immunodominant peptide, the list of subordinate peptides is likely to be incomplete. In fact, DNA vaccination experiments have recently predicted the existence of further epitopes in protein M84 (63). Of course, the majority of CD8 T cells express TCRs that are unrelated in specificity to the priming viral antigens. The composition of the CD8 T-cell population in the group of memory mice used for a second sorting experiment was tested by an ELISPOT assay (Fig. 11). IE1-specific, IFN--secreting CD8 T cells accounted for 1% of the population, and the frequencies of CD8 T cells specific for the remaining epitopes added up to another 1%. To separate IE1-specific and nonIE1-specific CD8 T cells, sort windows were set on IE1-TCRhigh CD8+ T cells as well as on IE1-TCR– CD8+ T cells (Fig. 12). Reanalysis documented an enrichment from 1.2 to >95% and from 85.5 to >99%, respectively. Finally, the protective antiviral capacity of the two sorted populations was compared by adoptive transfer of graded cell numbers into immunocompromised recipients infected with mCMV-WT (Fig. 13). The data fully confirmed the extraordinary antiviral activity of ex vivo-sorted IE1-specific CD8 T cells indicated by the first sorting experiment (Fig. 8 and 9). Remarkably, 400 cells were sufficient to control infection by 1 log10 PFU in the lungs and 3 log10 PFU in the spleen. Likewise, the number of infected cells was reduced by >2 log10 and 1 log10 in liver and adrenal glands, respectively. In accordance with the presence of CD8-TM cells specific for other protective mCMV epitopes, the IE1-TCR– subset too was capable of controlling the infection. The seemingly lower efficacy per cell is explained by the fact that the majority of cells in this subset are specific for antigens unrelated to mCMV.

    In conclusion, ex vivo sort-purified epitope-specific CD8-TM are exceedingly protective in controlling CMV disease.

    DISCUSSION

    A role for IE antigen-specific CD8 T cells in the control of acute and latent CMV infections was originally proposed by Reddehase and Koszinowski on the occasion of the first description of mCMV IE protein immunodominance in the BALB/c mouse model of CMV infection (48). Since then, many lines of evidence in this model have indicated a protective antiviral function of mCMV IE1 (168-YPHFMPTNL-176) epitope-specific CD8 T cells in acute infection. Furthermore, the accumulation of activated CD62L– IE1-specific effector-memory CD8 T cells (CD8-TEM) in latently infected lungs (21) in which IE1 transcripts are sporadically generated (13, 37) also suggested a guardian function in the maintenance of viral latency (21, 43). Up to now, however, direct evidence for an antiviral function of host-resident IE1-specific CD8 T cells was missing, not least because technologies for a TCR-based ex vivo isolation of epitope-specific T cells were established only recently (35). Here, we used IE1 peptide-loaded MHC-IgG dimers for cytofluorometric sort-purification of ex vivo-derived CD8-TM cells that express a cognate TCR. Adoptive cell transfer of these cells into immunocompromised indicator recipients demonstrated a highly efficient control of mCMV infection and prevention of viral histopathology. This completes the chain of evidence to conclude that IE1 epitope-specific CD8 T cells play a significant role in the control of CMV disease in the BALB/c mouse model.

    Although no animal model can fully match clinical CMV in all its aspects, and even laboratory strains and selected clinical isolates of hCMV will not do so for all human individuals, the BALB/c mouse model is revalued by the increasing evidence that the IE1 protein of hCMV is a major source of CD8 T-cell epitopes for HLA molecules represented in a large proportion of the human population (see the introduction and reference 45 for an overview). However, as shown recently for the example of an immunodominant Db-restricted epitope in mCMV protein M45, the magnitude of an epitope-specific CD8 T-cell response does not always predict a role in antiviral protection (23). So, defining IE1 epitopes and IE1 epitope-specific CD8 T-cell frequencies in humans is no substitute for functional evidence of an antiviral protective activity. To the best of our knowledge, there exists only one report addressing the question of a contribution of IE1-specific CD8 T cells to protection against human CMV disease. Specifically, as reported by Hebart et al. (15) for a limited number of recipients of an allogeneic stem cell transplantation with hCMV reactivation prior to day 100, the median duration of hCMV-DNAemia in patients with reconstitution of IE1-specific and UL83/pp65-specific CD8 T cells was significantly shorter than in patients with reconstitution of UL83/pp65-specific CD8 T cells only.

    The comparison of sorted IE1-TCRhigh CD8+ T cells versus the sum of the remaining IE1-TCR– CD8+ T cells has revealed an enrichment of protective cells in the IE1-TCR+ subset (Fig. 13). However, this cannot be taken to mean that IE1-specific CD8 T cells are superior, as they were enriched to 100%, whereas CD8 T cells specific for the other known epitopes of mCMV (Fig. 11) stayed at 1% among a majority of cells with specificities unrelated to mCMV. As shown in Fig. 13, the dose-response regression lines for the two subsets are not parallel but intersect at high cell numbers. The steeper slope observed for the IE1-TCR– subset indicates a more efficient control at higher cell numbers. This is explained by the fact that in this subset viral epitope-specific cells become limiting at low cell numbers and that one frequent specificity, namely m164, dominates until the low-frequency specificities, namely m18, m04, M83, M45, and M84, come in and contribute to the protection at the high cell numbers. The conclusion that CD8 T cells specific for different epitopes can cooperate for protection is supported by the work of Ye et al. (62) who have documented an improved protection against challenge infection after coimmunization with plasmids expressing mCMV IE1 and M84. Since a cell presenting different viral epitopes cannot be killed twice, cooperative effects are more likely to indicate differential presentation of these epitopes on different cell types or in the kinetics of viral gene expression.

    We have of course considered the approach to directly compare the antiviral function of sorted IE1-TCR+ and m164-TCR+ cells, but this experiment is pending because the Dd dimer is not yet on the market. Nevertheless, a comparison can be made at low cell numbers at which, as discussed above, m164 accounts for the protection by the IE1-TCR– subset. While the enrichment by sorting predicted a 100-fold-higher protective activity of the IE1-TCR+ cell population, the observed difference in protective activity was only 10-fold (Fig. 13). With some caution based on the fact that in this experiment only the IE1-specific cells were touched by TCR binding reagent, these findings suggest a higher efficacy of the Dd-restricted m164 (257-AGPPRYSRI-265) epitope-specific cells. This may relate to the previous finding that this peptide is constitutively presented during the E phase in infected cells (19) in the presence of all three immune evasion proteins of mCMV, namely, m04/gp34, m06/gp48, and m152/gp40 (for reviews, see references 16 and 45), whereas presentation of the IE1 peptide occurs in the E phase only in certain cell types (18) or under conditions of IFN--induced enhancement of processing and presentation (17). As shown previously for an IE1-CTLL, the IE1 epitope-specific antiviral efficacy in an adoptive transfer and coinfection setting was indeed markedly improved against mutant virus mCMV-m152 compared to revertant virus mCMV-m152-rev (51).

    In view of the susceptibility of IE1 peptide presentation to inhibition by the viral immune evasion proteins, the finding that as few as 400 sorted IE1-TCRhigh CD8+ T cells were nonetheless sufficient to control mCMV-WT infection of adoptive transfer recipients by 1 log10 in the lungs, 3 log10 in the spleen, >2 log10 in the liver, and 1 log10 in the adrenal glands is indeed astonishing and documents an outstanding antiviral efficacy of these cells. As shown in Fig. 7, a similar degree of protection required 105 cells of a third-restimulation IE1-CTLL, that is, an 250-fold-higher number of cell culture-propagated CD8 T cells of the very same epitope specificity!

    It must be emphasized that for these experiments we kept the in vitro cultivation and restimulation period to the minimum required for selecting a monospecific but polyclonal IE1-CTLL with a V? composition still closely resembling that of the starting CD8-TM population. Specifically, predominant usage of V?8.1-V?8.2 followed by V?10b and V?8.3 was fairly well conserved during the IE1 epitope-specific in vitro selection (Fig. 1 and 2). It is worth noting that T-cell clone IE1, the prototype clone with which the IE1 epitope was originally defined (46, 50), expressed a V?6 TCR (M. J. Reddehase, unpublished data). Previous studies of the V? expression by IE1 epitope-specific CD8 T cells also indicated a predominant usage of V?8 (54), specifically of V?8.1-V?8.2 (30) and have suggested preferential expansion of V?8.1-V?8.2 cells during an in vivo mCMV infection, leading to an increasing oligoclonality over time (30). The V? usage distribution among polyclonal IE1 epitope-specific late CD8-TM cells observed in our experiments (Fig. 1) is in accordance with the data reported by Karrer et al. (30); yet, our interpretation with regard to the selection of the clonal repertoire during mCMV infection is somewhat different. Most relevant, the V? composition of the IE1 epitope-specific CD8-TM pool eventually mirrored the epitope-independent general V? composition of early and late memory pools and even the homeostatic V? composition of the CD8 T-cell pool in uninfected, unprimed BALB/c mice. It thus appears that V? usage frequencies among IE1 epitope-specific CD8 T cells reflect the genetically determined general frequencies. This becomes particularly evident from the conserved predominance of V?8.1/8.2 usage (Fig. 1 and 2). We thus conclude that the IE1 epitope-specific expansion in vivo as well as during several rounds of restimulation in vitro is broadly polyclonal and in its rate largely independent of the V? family of the TCRs. All in all, V? usage was not likely the determinant of in vivo antiviral efficacy in the experiments reported here.

    From our own work in progress, we are well aware of the fact that the selection conditions in cell culture define the average TCR affinities of the resulting CTLLs and that this may have an impact on the in vivo antiviral efficacy (D. Gillert-Marien, unpublished data). Likewise, the coreceptor molecule CD8 contributes to overall avidity and may enhance antiviral efficacy by stabilizing the interaction between the TCR and the MHC-peptide complex. As shown in Fig. 10, effector cell function in an IE1-CTLL compared with the starting CD8-TM population was triggered by lower peptide concentrations. Thus, TCR affinity for the MHC-peptide complex and/or avidity of the interaction between effector cells and target cells was not the limiting factor.

    In years of experience with generating CTLLs and testing them for in vivo function in adoptive transfer experiments (22, 24, 25), we never observed protection with such low cell numbers as are shown here for the ex vivo-sorted cells. We therefore believe in a more fundamental difference between ex vivo recovered and in vitro selected and propagated epitope-specific CD8 T cells. We have not yet further pursued this phenomenon experimentally in order to provide the final explanation here, and it is evident that this is not a trivial issue to be studied in polyclonal natural immune responses, in which CD8 T cells specific for any particular epitope represent only a minor fraction. Clearly, these questions are much easier to address in TCR transgenic models (60). Obvious ideas include differences in the in vivo proliferative potential and survival time after transfer, as well as differences in lymphoid homing and tissue infiltration properties. Of likely importance is the difference in the differentiation stage of the transferred cells. While the IE1 epitope-specific, spleen-resident early CD8-TM cells tested in our experiments represented CD62Lhi central memory cells (CD8-TCM) and CD62Llo effector-memory cells (CD8-TEM) at a ratio of 1:1, the cells of the IE1-CTLL derived thereof after three rounds of in vitro restimulation and many more cell divisions were highly activated, cytolytically active CD62L– effector cells. As it has been shown recently by Wherry et al. (60) with a TCR transgenic model, CD8-TCM cells, in comparison to CD8-TEM cells, have a greater capacity to persist in vivo and are more efficient in mediating protective immunity because of their increased proliferative potential. It is reasonable to predict that effector cells of a CTLL have even less in vivo proliferative potential than the CD8-TEM cells. This may be the key to understanding the great difference in the protective capacities between sorted IE1 epitope-specific CD8-TM cells and IE1-CTLs observed here.

    Early work with the BALB/c mouse model successfully using polyclonal, acutely primed lymph node CD8 T cells (49, 52, 58) or spleen-resident CD8-TM cells (47) for adoptive transfer laid the theoretical foundations for clinical trials of CD8 T-cell-based preemptive cytoimmunotherapy of CMV disease with virus-specific blood lymphocyte-derived clonal CTLLs (53, 59; for a recent review, see reference 8). However, this therapy was not established as a clinical routine, mainly because of the excessive logistics required for expanding CTLLs to the very high cell numbers needed to obtain a protective effect. Recent progress in generating more efficient cells by reducing the selection period in cell culture has somewhat improved the prospects of adoptive cell transfer as a therapeutic option against CMV disease (41; for a commentary, see reference 7).

    Our data have documented an impressive protective capacity of ex vivo-sorted viral epitope-specific CD8-TM cells, not to compare with the limited protective capacity of short-term CTLLs specific for the same epitope. In this particular model, the nature of the viral epitope and the efficiency of its presentation in infected host tissues were constants in the system, and this helped to reveal more clearly the impact of parameters intrinsic to the transferred cells. To the best of our knowledge, such a quantitative comparison is without precedent in the literature.

    In conclusion, rather than focusing on in vitro expansion to high cell numbers, the pending question for future research is how to preserve TM status (ideally, TCM status) of CD8 T cells for adoptive transfer therapy.

    ACKNOWLEDGMENTS

    Support to R.H. and M.J.R. was provided by Deutsche Forschungsgemeinschaft Collaborative Research grant SFB490, individual projects B6 ("Persistence of murine cytomegalovirus after modulation of the CD8 T-cell immunome") and B1 ("Immunological control of latent cytomegalovirus infection"), respectively. M.J. was supported by grant 70-2427 HuI from the Deutsche Krebshilfe, and J.P. was supported by Deutsche Forschungsgemeinschaft grant SFB432.

    We thank Dirk H. Busch, Technical University of Munich, Munich, Germany, for valuable advice.

    REFERENCES

    Appel, H., N. P. Seth, L. Gauthier, and K. W. Wucherpfennig. 2001. Anergy induction by dimeric TCR ligands. J. Immunol. 166:5279-5285.

    Azuma, M., M. Cayabyab, D. Buck, J. H. Philipps, and L. L. Lanier. 1992. CD28 interaction with B7 costimulates primary allogeneic proliferative responses and cytotoxicity mediated by small, resting T lymphocytes. J. Exp. Med. 175:353-360.

    Boniface, J. J., J. D. Rabinowitz, C. Wulfing, J. Hampl, Z. Reich, J. D. Altman, R. M. Kantor, C. Beeson, H. M. McConnell, and M. M. Davis. 1998. Initiation of signal transduction through the T cell receptor requires the multivalent engagement of peptide/MHC ligands. Immunity 9:459-466. (Erratum, 9:891.)

    Borysiewicz, L. K., J. K. Hickling, S. Graham, J. Sinclair, M. P. Cranage, G. L. Smith, and J. G. P. Sissons. 1988. Human cytomegalovirus-specific cytotoxic T cells. Relative frequency of stage-specific CTL recognizing the 72-kD immediate-early protein and glycoprotein B expressed by recombinant vaccinia viruses. J. Exp. Med. 168:919-931.

    Cochran, J. R., T. O. Cameron, and L. J. Stern. 2000. The relationship of MHC-peptide binding and T cell activation probed using chemically defined MHC class II oligomers. Immunity 12:241-250.

    Del Val, M., H.-J. Schlicht, H. Volkmer, M. Messerle, M. J. Reddehase, and U. H. Koszinowski. 1991. Protection against lethal cytomegalovirus infection by a recombinant vaccine containing a single nonameric T-cell epitope. J. Virol. 65:3641-3646.

    Einsele, H., and K. Hamprecht. 2003. Immunotherapy of cytomegalovirus infection after stem-cell transplantation: a new option? Lancet 362:1343-1344.

    Einsele, H., and H. Hebart. 2004. CMV-specific immunotherapy. Hum. Immunol. 65:558-564.

    Friedl, P., and J. Storim. 2004. Diversity in immune-cell interactions: states and functions of the immunological synapse. Trends Cell Biol. 14:557-567.

    Gibson, L., G. Piccinini, D. Lilleri, M. G. Revello, Z. Wang, S. Markel, D. J. Diamond, and K. Luzuriaga. 2004. Human cytomegalovirus proteins pp65 and immediate early protein 1 are common targets for CD8+ T cell responses in children with congenital or postnatal human cytomegalovirus infection. J. Immunol. 172:2256-2264.

    Gonzales Armas, J. C., C. S. Morello, L. D. Cranmer, and D. H. Spector. 1996. DNA immunization confers protection against murine cytomegalovirus infection. J. Virol. 70:7921-7928.

    Gribaudo, G., L. Riera, D. Lembo, M. de Andrea, M. Gariglio, T. L. Rudge, L. F. Johnson, and S. Landolfo. 2000. Murine cytomegalovirus stimulates cellular thymidylate synthase gene expression in quiescent cells and requires the enzyme for replication. J. Virol. 74:4979-4987.

    Grzimek, N. K. A., D. Dreis, S. Schmalz, and M. J. Reddehase. 2001. Random, asynchronous, and asymmetric transcriptional activity of enhancer-flanking major immediate-early genes ie1/3 and ie2 during murine cytomegalovirus latency in the lungs. J. Virol. 75:2692-2705.

    Gyulai, Z., V. Endresz, K. Burian, S. Pincus, J. Toldy, W. I. Cox, C. Meric, S. Plotkin, E. Gonczol, and K. Berencsi. 2000. Cytotoxic T lymphocyte (CTL) responses to human cytomegalovirus pp65, IE1-exon4, gB, pp150, and pp28 in healthy individuals: reevaluation of prevalence of IE1-specific CTLs. J. Infect. Dis. 181:1537-1546.

    Hebart, H., G. Rauser, S. Stevanovic, C. Haenle, A. K. Nussbaum, C. Meisner, A. L. Bissinger, S. Tenzer, G. Jahn, J. Loeffler, H.-G. Rammensee, H. Schild, and H. Einsele. 2003. A CTL epitope from human cytomegalovirus IE1 defined by combining prediction of HLA binding and proteasomal processing is the target of dominant immune responses in patients after allogeneic stem cell transplantation. Exp. Hematol. 31:966-973.

    Hengel, H., W. Brune, and U. H. Koszinowski. 1998. Immune evasion by cytomegalovirus—survival strategies of a highly adapted opportunist. Trends Microbiol. 6:190-197.

    Hengel, H., P. Lucin, S. Jonjic, T. Ruppert, and U. H. Koszinowski. 1994. Restoration of cytomegalovirus antigen presentation by -interferon combats viral escape. J. Virol. 68:289-297.

    Hengel, H., U. Reusch, G. Geginat, R. Holtappels, T. Ruppert, E. Hellebrand, and U. H. Koszinowski. 2000. Macrophages escape inhibition of major histocompatibility complex class I-dependent antigen presentation by cytomegalovirus. J. Virol. 74:7861-7868.

    Holtappels, R., N. K. A. Grzimek, C. O. Simon, D. Thomas, D. Dreis, and M. J. Reddehase. 2002. Processing and presentation of murine cytomegalovirus pORFm164-derived peptide in fibroblasts in the face of all viral immunosubversive early gene functions. J. Virol. 76:6044-6053.

    Holtappels, R., N. K. A., Grzimek, D. Thomas, and M. J. Reddehase. 2002. Early gene m18, a novel player in the immune response to murine cytomegalovirus. J. Gen. Virol. 83:311-316.

    Holtappels, R., M.-F. Pahl-Seibert, D. Thomas, and M. J. Reddehase. 2000. Enrichment of immediate-early 1 (m123/pp89) peptide-specific CD8 T cells in a pulmonary CD62Llo memory-effector cell pool during latent murine cytomegalovirus infection of the lungs. J. Virol. 74:11495-11503.

    Holtappels, R., J. Podlech, N. K. A. Grzimek, D. Thomas, M.-F. Pahl-Seibert, and M. J. Reddehase. 2001. Experimental preemptive immunotherapy of murine cytomegalovirus disease with CD8 T-cell lines specific for ppM83 and pM84, the two homologs of human cytomegalovirus tegument protein ppUL83 (pp65). J. Virol. 75:6584-6600.

    Holtappels, R., J. Podlech, M.-F. Pahl-Seibert, M. Jülch, D. Thomas, C. O. Simon, M. Wagner, and M. J. Reddehase. 2004. Cytomegalovirus misleads its host by priming of CD8 T cells specific for an epitope not presented in infected tissues. J. Exp. Med. 199:131-136.

    Holtappels, R., D. Thomas, J. Podlech, G. Geginat, H.-P. Steffens, and M. J. Reddehase. 2000. The putative natural killer decoy early gene m04 (gp34) of murine cytomegalovirus encodes an antigenic peptide recognized by protective antiviral CD8 T cells. J. Virol. 74:1871-1884.

    Holtappels, R., D. Thomas, J. Podlech, and M. J. Reddehase. 2002. Two antigenic peptides from genes m123 and m164 of murine cytomegalovirus quantitatively dominate CD8 T-cell memory in the H-2d haplotype. J. Virol. 76:151-164.

    Holtappels, R., D. Thomas, and M. J. Reddehase. 2000. Identification of a Kd-restricted antigenic peptide encoded by murine cytomegalovirus early gene M84. J. Gen. Virol. 81:3037-3042.

    Hugues, S., L. Malherbe, C. Filippi, and N. Glaichenhaus. 2002. Generation and use of alternative multimers of peptide/MHC complexes. J. Immunol. Methods 268:83-92.

    Huster, K. M., V. Busch, M. Schiemann, K. Linkemann, K. M. Kerksiek, H. Wagner, and D. H. Busch. 2004. Selective expression of IL-7 receptor on memory T cells identifies early CD40L-dependent generation of distinct CD8+ memory T cell subsets. Proc. Natl. Acad. Sci. USA 101:5610-5615.

    Jonjic, S., M. del Val, G. M. Keil, M. J. Reddehase, and U. H. Koszinowski. 1988. A nonstructural viral protein expressed by a recombinant vaccinia virus protects against lethal cytomegalovirus infection. J. Virol. 62:1653-1658.

    Karrer, U., S. Sierro, M. Wagner, A. Oxenius, H. Hengel, U. H. Koszinowski, R. E. Phillips, and P. Klenerman. 2003. Memory inflation: continuous accumulation of antiviral CD8+ T cells over time. J. Immunol. 170:2022-2029.

    Keil, G. M., A. Ebeling-Keil, and U. H. Koszinowski. 1987. Sequence and structural organization of murine cytomegalovirus immediate-early gene 1. J. Virol. 61:1901-1908.

    Keil, G. M., M. R. Fibi, and U. H. Koszinowski. 1985. Characterization of the major immediate-early polypeptides encoded by murine cytomegalovirus. J. Virol. 54:422-428.

    Kern, F., I. P. Surel, N. Faulhaber, C. Frommel, J. Schneider-Mergener, C. Schoenemann, P. Reinke, and H. D. Volk. Target structures of the CD8+ T-cell response to human cytomegalovirus: the 72-kilodalton major immediate-early protein revisited. J. Virol. 73:179-184.

    Khan, N., M. Cobbold, R. Keenan, and P. A. Moss. 2002. Comparative analysis of CD8+ T-cell responses against human cytomegalovirus proteins pp65 and immediate-early 1 shows similarities in precursor frequency, oligoclonality, and phenotype. J. Infect. Dis. 185:1025-1034.

    Knabel, M., T. J. Franz, M. Schiemann, A. Wulf, B. Villmow, B. Schmidt, H. Bernhard, H. Wagner, and D. H. Busch. 2002. Reversible MHC multimer staining for functional isolation of T-cell populations and effective adoptive transfer. Nat. Med. 8:631-637.

    Knuehl, C., P. Spee, T. Ruppert, U. Kuckelkorn, P. Henklein, J. Neefjes, and P. M. Kloetzel. 2001. The murine cytomegalovirus pp89 immunodominant H-2Ld epitope is generated and translocated into the endoplasmic reticulum as an 11-mer precursor peptide. J. Immunol. 167:1515-1521.

    Kurz, S. K., M. Rapp, H.-P. Steffens, N. K. A. Grzimek, S. Schmalz, and M. J. Reddehase. 1999. Focal transcriptional activity of murine cytomegalovirus during latency in the lungs. J. Virol. 73:482-494.

    Matsui, K., J. J. Boniface, P. Steffner, P. A. Reay, and M. M. Davis. 1994. Kinetics of T-cell receptor binding to peptide/I-Ek complexes: correlation of the dissociation rate with T-cell responsiveness. Proc. Natl. Acad. Sci. USA 91:12862-12866.

    Messerle, M., B. Bühler, G. M. Keil, and U. H. Koszinowski. 1992. Structural organization, expression, and functional characterization of the murine cytomegalovirus immediate-early gene 3. J. Virol. 66:27-36.

    Messingham, K. A., V. P. Badovinac, and J. T. Harty. 2003. Deficient anti-listerial immunity in the absence of perforin can be restored by increasing memory CD8+ T cell numbers. J. Immunol. 171:4254-4262.

    Peggs, K. S., S. Verfuerth, A. Pizzey, N. Khan, M. Gulver, P. A. Moss, and S. Mackinnon. 2003. Adoptive cellular therapy for early cytomegalovirus infection after allogeneic stem-cell transplantation with virus-specific T-cell lines. Lancet 362:1375-1377.

    Podlech, J., R. Holtappels, N. K. A. Grzimek, and M. J. Reddehase. 2002. Animal models: murine cytomegalovirus, p. 493-525. In S. H. E. Kaufman and D. Kabelitz (ed.), Methods in microbiology. Immunology of infection, 2nd ed., vol. 32. Academic Press, London, United Kingdom.

    Podlech, J., R. Holtappels, M.-F. Pahl-Seibert, H.-P. Steffens, and M. J. Reddehase. 2000. Murine model of interstitial cytomegalovirus pneumonia in syngeneic bone marrow transplantation: persistence of protective pulmonary CD8-T-cell infiltrates after clearance of acute infection. J. Virol. 74:7496-7507.

    Podlech, J., R. Holtappels, N. Wirtz, H.-P. Steffens, and M. J. Reddehase. 1998. Reconstitution of CD8 T cells is essential for the prevention of multiple-organ cytomegalovirus histopathology after bone marrow transplantation. J. Gen. Virol. 79:2099-2104.

    Reddehase, M. J. 2002. Antigens and immunoevasins: opponents in cytomegalovirus immune surveillance. Nat. Rev. Immunol. 2:831-844.

    Reddehase, M. J., H.-J. Bühring, and U. H. Koszinowski. 1986. Cloned long-term cytolytic T-lymphocyte line with specificity for an immediate-early membrane antigen of murine cytomegalovirus. J. Virol. 57:408-412.

    Reddehase, M. J., S. Jonjíc, F. Weiland, W. Mutter, and U. H. Koszinowski. 1988. Adoptive immunotherapy of murine cytomegalovirus adrenalitis in the immunocompromised host: CD4-helper-independent antiviral function of CD8-positive memory T lymphocytes derived from latently infected donors. J. Virol. 62:1061-1065.

    Reddehase, M. J., and U. H. Koszinowski. 1984. Significance of herpesvirus immediate early gene expression in cellular immunity to cytomegalovirus infection. Nature (London) 312:369-371.

    Reddehase, M. J., W. Mutter, and U. H. Koszinowski. 1987. In vivo application of recombinant interleukin 2 in the immunotherapy of established cytomegalovirus infection. J. Exp. Med. 165:650-656.

    Reddehase, M. J., J. B. Rothbard, and U. H. Koszinowski. 1989. A pentapeptide as minimal antigenic determinant for MHC class I-restricted T lymphocytes. Nature (London) 337:651-653.

    Reddehase, M. J., C. O. Simon, J. Podlech, and R. Holtappels. 2004. Stalemating a clever opportunist: lessons from murine cytomegalovirus. Hum. Immunol. 65:446-455.

    Reddehase, M. J., F. Weiland, K. Münch, S. Jonjíc, A. Lüske, and U. H. Koszinowski. 1985. Interstitial murine cytomegalovirus pneumonia after irradiation: characterization of cells that limit viral replication during established infection of the lungs. J. Virol. 55:264-273.

    Riddell, S. R., K. S. Watanabe, J. M. Goodrich, C. R. Li, M. E. Agha, and P. D. Greenberg. 1992. Restoration of viral immunity in immunodeficient humans by the adoptive transfer of T cell clones. Science 257:238-241.

    Rodewald, H. R., U. H. Koszinowski, K. Eichmann, and I. Melchers. 1989. Predominant utilization of V beta 8+ T cell receptor genes in the H-2Ld-restricted cytotoxic T cell response against the immediate-early protein pp89 of the murine cytomegalovirus. J. Immunol. 143:4238-4243.

    Scalzo, A. A., S. L. Elliot, J. Cox, J. Gardner, D. J. Moss, and A. Suhrbier. 1995. Induction of protective cytotoxic T cells to murine cytomegalovirus by using a nonapeptide and a human-compatible adjuvant (Montanide ISA 720). J. Virol. 69:1306-1309.

    Schneck, J. P. 2000. Monitoring antigen-specific T cells using MHC-Ig dimers. Immunol. Investig. 29:163-169.

    Schulz, M., P. Aichele, R. Schneider, T. H. Hansen, R. M. Zinkernagel, and H. Hengartner. 1991. Major histocompatibility complex binding and T cell recognition of viral nonapeptide containing a minimal tetrapeptide. Eur. J. Immunol. 21:1181-1185.

    Steffens, H.-P., S. Kurz, R. Holtappels, and M. J. Reddehase. 1998. Preemptive CD8-T-cell immunotherapy of acute cytomegalovirus infection prevents lethal disease, limits the burden of latent viral genome, and reduces the risk of virus recurrence. J. Virol. 72:1797-1804.

    Walter, E. A., P. D. Greenberg, M. J. Gilbert, R. J. Finch, K. S. Watanabe, E. D. Thomas, and S. R. Riddell. 1995. Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by transfer of T-cell clones from the donor. N. Engl. J. Med. 333:1038-1044.

    Wherry, E. J., V. Teichgr?ber, T. C. Becker, D. Masopust, S. M. Kaech, R. Antia, U. H. von Andrian, and R. Ahmed. 2003. Lineage relationship and protective immunity of memory CD8 T cell subsets. Nat. Immunol. 4:225-234.

    Xu, X. N., and G. R. Screaton. 2002. MHC/peptide tetramer-based studies of T cell function. J. Immunol. Methods 268:21-28.

    Ye, M., C. S. Morello, and D. H. Spector. 2002. Strong CD8 T-cell responses following coimmunization with plasmids expressing the dominant pp89 and subdominant M84 antigens of murine cytomegalovirus correlate with long-term protection against subsequent viral challenge. J. Virol. 76:2100-2112.

    Ye, M., C. S. Morello, and D. H. Spector. 2004. Multiple epitopes in the murine cytomegalovirus early gene product M84 are efficiently presented in infected primary macrophages and contribute to strong CD8+ T-lymphocyte responses and protection following DNA immunization. J. Virol. 78:11233-11245.(Marcus-Folker Pahl-Seiber)