|Home | About | Journals | Submit | Contact Us | Français|
Human cytomegalovirus (HCMV) can cause life‐threatening disease in infected hosts. Immunization with human leukocyte antigen (HLA)‐restricted immunodominant synthetic peptides and adoptive transfer of epitope‐specific T cells have been envisaged to generate or boost HCMV‐specific cellular immunity, thereby preventing HCMV infection or reactivation. However, induction or expansion of T cells effective against HCMV are limited by the need of utilizing peptides with defined HLA restrictions. We took advantage of a combination of seven predictive algorithms to identify immunogenic peptides of potential use in the prevention or treatment of HCMV infection or reactivation. Here we describe a pp65‐derived peptide (pp65340–355, RQYDPVAALFFFDIDL: RQY16‐mer), characterized by peculiar features. First, RQY‐16mer is able to stimulate HCMV pp65 specific responses in both CD4+ and CD8+ T cells, restricted by a wide range of HLA class I and II determinants. Second, RQY‐16mer is able to induce an unusually wide range of effector functions in CD4+ T cells, including proliferation, killing of autologous HCMV‐infected target cells and cytokine production. Third, and most importantly, the RQY‐16mer is able to stimulate CD4+ and CD8+ T‐cell responses in pharmacologically immunosuppressed patients. These data suggest that a single reagent might qualify as synthetic immunogen for potentially large populations exposed to HCMV infection or reactivation.
Primary infection by human cytomegalovirus (HCMV) is usually mild or asymptomatic and it is effectively controlled by T‐cell‐mediated immune response in healthy individuals . However, congenital HCMV infection in newborns  and viral reactivation in pharmacologically immunosuppressed patients transplanted with solid organs  or haematopoietic stem cells (HSCT)  are frequently associated with significant morbidity and, eventually, mortality, until the immune system is effectively competent or completely reconstituted . Thus, women in childbearing age and transplant recipients may represent candidate target populations for HCMV active protection.
T‐cell‐mediated immune response against immunogenic viral targets is of paramount importance due the potential capacity of epitope‐specific cytotoxic T lymphocytes (CTLs) to control viral fitness. The administration of major histocompatibility complex (MHC)‐restricted immunodominant synthetic peptides to generate CTL immunity preventing the reactivation of HCMV infection represents a promising approach already tested in clinical trials . Interestingly, vaccination of HSCT donors expressing specific HLA determinants might also be important for the generation of HCMV‐specific T cells of potential use in adoptive immunotherapy. In this case, healthy virus carriers might represent an important reservoir of antigen‐specific CD4+ and CD8+ T cells.
Although much emphasis has been placed on the role of MHC class I restricted CD8+ T cells in the recognition of HCMV‐infected cells, there is increasing evidence that CD4+ T cells also have a crucial role in the control of HCMV infection .
The cytotoxic potential of CD4+ T cells specific for viral antigens has been repeatedly documented [8, 9]. Furthermore, transfer of HCMV‐specific CD4+ T cells into HSCT has also been found to promote the expansion of HCMV‐specific CD8+ CTLs  likely due to the helper function exerted by antigen expanded CD4+ T‐cell populations .
The concomitant use of MHC class I and II restricted peptides would be highly desirable to achieve effective immune responses . However, limitations inherent in MHC diversity and epitope hierarchies within HLA presentation  suggest that this approach would require the administration of a panel of immunogenic HLA‐binding peptides particularly in case of unknown typing or when rare HLA specificities should restrict immune responses. Ideally, more than one immunogenic reagent should be used [14, 15] for each HLA determinant of interest.
Indeed, a large number of immunodominant HLA class I and II restricted HCMV‐derived peptides have been identified . However, only a minority of them has been used in clinical procedures due to the necessity of tailoring immunogens according to specific HLA typing .
To address these issues, we attempted the identification of highly immunogenic sequences within HCMV pp65 possibly encompassing both class I and II restricted epitopes using a combination of predictive algorithms. Here we report the functional characterization of a highly immunogenic pp65‐derived peptide (RQYDPVAALFFFDIDL; hereafter referred as RQY‐16mer peptide) capable of inducing class I and II restricted cytotoxic and lymphoproliferative responses across a wide range of HLA specificities.
Peptides derived from the immunodominant HCMV 65‐kD matrix phosphoprotein (pp65; human herpesvirus 5 laboratory strain AD169) were used in this study. Six major algorithms, PAProC (http://www.paproc.de/), MAPPP (http://www.mpiib‐berlin.mpg.de/MAPPP/), NetChop (http://www.cbs.dtu.dk/services/NetChop/), Bimas (http://www‐bimas.cit.nih.gov/molbio/), SYFPEITHI (http://www.syfpeithi.de) and IEDB (http://www.immuneepitope.org/home.do), were utilized to select candidate immunodominant epitopes within pp65. Furthermore, when indicated, NetMHCIIpan (http://www.cbs.dtu.dk/services/NetMHCpan/) algorithm was also used. Peptides were synthesized by Princeton Biomolecules (Langhorne, PA, USA) with purity ranging from 90% to 100% as analysed by high‐performance liquid chromatography, dissolved in 100% Dimethyl Sulfoxide (DMSO) and stored at–70°C until use.
MAPPP, PAProc and NetChop algorithms were initially used to screen pp65 sequence. Eight peptides, pp6549−70, pp65113−127, pp65128−144, pp65186−208, pp65293−305, pp65319−330, pp65340−355 and pp65493−515 displayed the highest score of potential immunogenicity, thus extending previous results . These sequences were further analysed for proteasome cleavage sites, TAP transport and MHC binding as evaluated by IEDB, Bimas and SYFPEITHI algorithms. Based on these sets of data, we selected pp65 sequences nesting high numbers of 9mer‐peptides restricted by the most frequently represented HLA class I specificities included in the algorithms. Pp65340−355 (RQY‐16mer) turned out to be the sequence containing the highest numbers of potentially immunogenic HLA class I restricted 9mer‐peptides. In particular, six of eight 9mers within RQY‐16mer encompassed HLA‐A1, A2, A24, A30, A32 and A68 epitope restrictions (n= 6) and HLA‐B7, B15, B18, B27, B35, B40, B44, B51, B53, B57 and B58 epitope‐restrictions (n= 6) for a total of 17 valuable HLA class I associations out of the 41 covered by the programs (42%).
Fourteen HCMV‐seropositive and two HCMV‐seronegative donors were age, race and sex randomly selected and enrolled in the study upon informed consent (Blutspendezentrum Universitätsspital, Basel, Switzerland). Five umbilical cord blood samples were collected at University Hospital in Basel after either vaginal or caesarean delivery. Ten patients receiving kidney transplant and post‐transplantation care were accrued at the Department of Transplantation Immunology and Nephrology, University Hospital in Basel according to a protocol approved by the internal review board (299/06). All patients were treated with an immunosuppressive regimen consisting of mycophenolate, cyclosporine and prednisone, and one underwent prophylactic antiviral treatment, as detailed in ‘Results’ section. The presence of anti‐HCMV antibodies in the serum (anti‐pp65 IgG titre) was analysed with passive latex agglutination (CMVSCAN kit, Becton Dickinson Microbiology System, Cockeysville, MD, USA or AxSymTM assay, Abbott, Sligo, Ireland; http://www.abbott.com). Quantitative real time PCR (qrt‐PCR) assays detecting HCMV pp65 DNA in CD14+ cells of healthy donors were performed by using 5′‐GTCAGCGTTCGTGTTTCCCA‐3′ direct primer, 5′‐GGGACACAACACCGTAAAGC‐3′ reverse primer and 5′FAM‐CCCGCAACCCGCAACCCTTCATG‐3′TAMRA fluorogenic probe . Regarding transplanted patients, HCMV‐replication was quantified after DNA extraction from ethylenediaminetetraacetic acid (EDTA)‐anticoagulated whole blood (Magnapure TM, Hoffman‐LaRoche, Basel, Switzerland; http://www.roche.com) by using 5′‐TTTTTTCTAGGCGCTTCCGA‐3′ and 5′‐ACACTGCGGCTTTGTATTCTTTATC‐3′ primers and 5′FAM‐AGGCGAAGCCGGCGACGA‐3′TAMRA probe at the laboratory of Transplantation Virology, Institute for Medical Microbiology, University of Basel, Switzerland, by taking advantage of previously published amplification protocols .
A complete, high resolution, HLA genotyping (HLA‐A*, HLA‐B* and HLA‐DRB1*) was performed by PCR with allele‐specific sequencing primers (PCR SSP Protrans, Ketsch am Rhein, Germany) according to the manufacturer's specifications.
Peripheral blood mononuclear cells (PBMCs) were separated by Ficoll‐Hypaque density gradient centrifugation. T lymphocyte subpopulations (CD8+, CD4+) were purified by magnetic cell separation (Miltenyi Biotech, Bergisch Gladbach, Germany) according to producers’ protocols. Cells were then harvested in Roswell Park Memorial Institute (RPMI) medium supplemented with 100 μg/ml Kanamycin, 10 mM Hepes, 1 mM sodium pyruvate, 1 mM Glutamax and non‐essential amino acids (all from GIBCO Paisley, Scotland), hereafter referred to as complete medium supplemented with 5% human serum (Blutspendezentrum Universitätsspital, Basel, Switzerland). Both, CD8+ and CD4+ purified T cells were subsequently plated in complete medium with 5% human serum in 24‐well plates at a final concentration of 1 × 106 cells/ml and were co‐cultured (37°C, 5% CO2 atmosphere) with irradiated (750 sec. in a gamma ray irradiator equipped with a 137Cs radiation source emitting 100 rad/min.) autologous mature dendritic cells (mDCs) (2 × 105/ml; see below for DCs generation) previously pulsed for 2 hrs with peptides at a final concentration of 10 μg/ml either for priming or for re‐stimulation rounds. Recombinant human (rh) interleukin (IL)‐2 (Hoffmann‐LaRoche, Basel, Switzerland) was added to the cultures at 1 ng/ml, 1 ng/ml and 5 ng/ml, on days 3, 7 and 10, respectively, and cells were restimulated with specific peptides in the presence of irradiated mDCs, as detailed above, on day 7 of culture. In the indicated experiments (see below) control cultures were performed as specified above but in the absence of antigenic peptides.
For DCs generation, isolated CD14+ (Miltenyi Biotech) were cultured for 5 to 7 days in RPMI complete medium supplemented with 10% foetal calf serum (FCS) (GIBCO Paisley, Scotland), 0.004%β‐mercaptoethanol, rhIL‐4 (1000 UI/ml) and recombinant human granulocyte macrophage‐colony stimulating factor (GM‐CSF; 50 ng/ml) to generate immature DCs (iDCs). Maturation of iDCs was induced by exposure to lipopolysaccaride (Abortus Aequi, Sigma‐Aldrich, St. Louis, MO, USA) at a concentration of 1 μg/ml.
Immature DCs were infected with cell‐free endotheliotrophic (HUVEC) HCMV VR1814 strain at a MOI of 10 and incubated in complete medium supplemented with 10% FCS for 24 hrs in round‐bottom 96‐well plates at 37°C in a 5% CO2 atmosphere . As control, iDCs were mock infected with the clarified medium of uninfected HUVEC cultures. Infection grade was measured by both immunofluorescent detection of pp72 and qrt‐PCR analysis of pp65 gene expression, as described above (see section ‘Donor selection, cord blood collection, patient accrual, HCMV serology and HLA genotyping’). Immunofluorescence analysis was performed on paraformaldehyde‐fixed DCs using allophycocyanine‐conjugated anti‐CD1a (BD Pharmingen, San Diego, CA, USA), or on acetone: methanol (1:1)‐fixed iDCs using a goat anti‐IE72 monoclonal antibodies (mAb) followed by Alexa Fluor‐555‐conjugated chicken anti‐rabbit antibodies (Molecular Probes, Eugene, OR, USA). Fluorescence was visualized using a 100× Plan Neofluar oil immersion objective (NA 1.3) mounted on a Zeiss Axiovert 100 confocal microscope (Jeha, Germany).
Epstein‐Barr virus (EBV)‐transformed B‐cell lines were generated from each donor by infecting 1 × 106 freshly isolated PBMCs with Epstein‐Barr virus containing supernatants from B95.8 cell line. A total of 5 μg/ml cyclosporin A (Novartis, Basel, Switzerland) was added to prevent T‐cell response and infected cells were cultured in complete medium with 10% FCS in the presence of rhIL‐6 until complete expansion. Cells were then harvested and used for experimental procedures or aliquoted and stored in liquid nitrogen for further use.
Peptide‐induced cytokine gene expression was investigated as follows. On day 14 of culture (see above), CD4+ and CD8+ T cells expanded in vitro in the presence or absence of antigenic peptide(s) were plated in U‐bottom 96‐well plates with irradiated autologous mDCs at 10:1 ratio (2 × 105 CD8+ or CD4+ and 2 × 104 mDCs per well). At day 15, after an overnight resting, cells were either peptide‐stimulated (1 μg/ml) or left unstimulated. Three hours after cells were harvested for total cellular RNA extraction (RNeasy® Mini Kit Protocol, Qiagen, Basel, Switzerland) and cDNA synthesis (Invitrogen, Carlsbad, CA, USA). Cytokine mRNA transcript amplification was performed as previously described  by an ABI prism 7500 FAST sequence detection system using TaqMan® Universal PCR Master Mix Reagents Kit (Applied Biosystems, Rotkreuz, Switzerland) and sets of primers and probes from cytokine genes (interferon [IFN]‐γ, tumour necrosis factor [TNF]‐α, IL‐2, IL‐10) already extensively utilized . CD8+, CD4+, CD14+ and β‐actin were used as endogenous reference genes . Normalized data were subsequently evaluated as relative quantification. The 2−ΔΔCt method (ΔΔCT=[CT, cytokine – CT,β‐actin]induction–[CT, cytokine – CT,β‐actin]baseline, where CT is the mean cycle times of the triplicate well readings) was utilized to compute the fold change of cytokine gene expression after peptide induction relative to baseline (unstimulated cells), normalized to endogenous reference genes . When indicated, absolute numbers of cytokine encoding gene copies were calculated as previously described .
Peptide expanded CD4+ T lymphocytes (5 × 105) were rested overnight in a 14‐ml polypropylene tube (Becton Dickinson, Franklin Lakes, NJ, USA) and then stimulated with peptide‐pulsed (10 μg/ml) mDCs at a final concentration of 5 × 104 per each tube. One hour after cell activation, 10 μg/ml of Brefeldin A (Sigma, Saint Louis, MI, USA) was added. After a 5‐hr additional incubation, cells were transferred to 5‐ml round‐bottom tubes (Becton Dickinson) and cell incubation was stopped by washing cells in 2 ml cold phosphate‐buffered saline (PBS) for 5 min. Pellets were re‐suspended in 1 ml PBS containing 1 mM EDTA and 0.5% FCS and cells were extracellularly stained with 10 μl of fluorescent mAb recognizing CD3, CD4 and CD8 (BD Bioscience, San Jose, CA, USA) for 15 min. at 4°C in the dark. Cell fixation was performed with 2 ml of BD FACS Lysis Solution (BD Bioscience) according to the producers’ instructions. Cellular permeabilization was performed by re‐suspending cells in 500 μl of FACS Permeabilizing Solution 2 (BD Bioscience) at a 1:10 dilution in Diethyl Pyrocarbonate (DEPC) water at room temperature for 10 min. For intracellular staining, cells were stained with either 10 μl of human anti‐IFN‐γ‐FITC, TNF‐α‐Fluorescein Isothiocyanate (FITC), IL‐2‐FITC mAbs or mouse IgG1 isotype‐FITC (BD Bioscience) and incubated for 30 min. at 4°C in the dark. Samples were analysed on a FACSCalibur flow‐cytometer equipped with Cellquest software (Becton Dickinson, San Jose, CA, USA).
Isolated CD4+ T cells from freshly separated PBMCs were cultured in complete medium with 5% human serum in flat‐bottom 96‐well plates (Becton Dickinson, Le Pont de Claix, France) at 2 × 105 cells/well and stimulated with peptide‐pulsed mDCs. On day 5, 3H‐thymidine (Amersham, Little Chalfont, UK) was added at 1 μCi per well. After an 18‐hr incubation, cells were harvested and tracer incorporation was measured by beta counting. Data were expressed as 3H‐thymidine incorporation in counts per minutes (cpm).
Caspase‐3 release was carried out by using PhiPhiLux® (Oncoimmunin Inc, Gaithersburg, MD, USA) as follows. Infected iDCs were suspended in complete medium with 10% FCS at 1 × 106/ml. Cells were then incubated at 37°C, for 1 hr in the presence of 3 M specific dye from CyToxiLux OncoImmunin (CTO™; Molecular Probes) and peptides (1 μg). The cells were then washed once and re‐suspended in complete medium with 10% FCS at 1 × 106/ml. For each condition, 2 × 106 of 2‐week peptide stimulated CD4+ effector‐cells were co‐cultured with 2 × 105 CTO stained target cells (infected or control iDCs) in round‐bottom 96‐well plates (Becton Dickinson) for 2 hrs at 37°C. The supernatant was then removed and the cells were incubated in 75 μl/well of the indicated caspase‐3 substrate (10 M, OncoImmunin, Gaithersburg) for 30 min. at 37°C and washed twice with PBS . An extra staining for CD1a was performed to identify iDCs. Cells were then analysed by flow cytometry (FACSCalibur; Becton Dickinson) using Cell Quest software (Becton Dickinson) and gated for CD1a expression.
Cytotoxic activity of peptide expanded CD8+ T cells was tested on day 14 by 4 hrs chromium release assays using, as targets, autologous EBV‐transformed B cells previously labelled with 51Cr (50 μCi of 51Cr for 1 hr at 37°C) and pulsed for 2 hrs with cognate or control peptides at a concentration of 2.5 μg/ml. Specific lysis of target cells from triplicate wells was calculated according to the standard formula: 100 × ([cpm experimental release – cpm spontaneous release]/[cpm maximal release – cpm spontaneous release]). Results are reported as mean of delta specific lysis (ΔSL; experimental positive release – experimental negative release) .
HLA‐A*0201 and HLA‐A*2402 MCH Pro5™ PE‐labelled pentamers for pp65340−348 (RQY‐9mer) and pp65341−349 (QYD‐9mer) (Proimmune, Oxford, UK) were used for surface staining of cells under investigation. Cells were first stained with 1 μl MHC PE‐labelled pentamer for 15 min. at 4°C in the dark. Subsequently, they were washed once and stained with 5 μl of CD8‐FITC (BD Bioscience) for 30 min. at 4°C in the dark. After staining, samples were analysed on a FACSCalibur flow‐cytometer equipped with Cellquest software (Becton Dickinson).
Immature DCs were fixed as described by Shimonkevitz et al.. Briefly, cells were re‐suspended in PBS and fixed by addition of glutaraldehyde (final concentration 0.05%) for 30 sec. at room temperature. The reaction was stopped by adding L‐glycine (final concentration 0.2 M) for 45 sec. and by three washes with PBS.
For the measurement of antigen‐specific T‐cell activation, 5 × 104 fixed or unfixed autologous iDCs were co‐cultured in U‐bottom 96‐well plates with 2 × 105 antigen specific in vitro expanded T cells for 3 hrs in the presence of the indicated peptides. T‐cell reactivation was evaluated by IFN‐γ gene expression.
Statistical analysis was performed with Prism 4 (GraphPad) software. Data were reported as mean ± S.D. or as median and ranges where appropriate. Categorical markers were analysed by Pearson's chi‐square test. A two‐tailed paired t‐test was used to calculate P‐values. Due to variability related to heterogeneous HLA restriction, unpaired t‐tests and F tests were used to compare variance when appropriate. Differences were considered significant at P < 0.05, CI 95%.
In order to define the overall immunogenicity of RQY‐16mer, we evaluated its capability to elicit CD4+ and/or CD8+ immune responses in T cells. CD4+ and CD8+ T cells from either HCMV‐seropositive or seronegative donors (Table 1A) were 2‐week in vitro stimulated using RQY‐16mer peptide‐loaded or untreated autologous mDCs. A 3‐hr assay of RQY‐16mer mediated induction of IFN‐γ, TNF‐α, IL‐2 and IL‐10 cytokine gene expression was then performed on expanded cells. A cut‐off level of responsiveness (= 1.9 folds) was calculated based on the responses to 3 hrs stimulation of 2‐week cultures from seropositive donors conducted in the absence of antigenic peptide.
RQY‐16mer induced the activation of T cells pre‐cultured in the presence of the peptide from all HCMV‐seropositive donors, as detectable by a significant increase in pro‐inflammatory cytokine gene expression as compared to unstimulated controls, irrespective of their HLA specificities (Fig. 1A). In particular, IFN‐γ gene expression was increased 6.27 ± 2.96 folds in RQY‐16mer stimulated CD4+ T cells, as compared to control cultures. TNF‐α and IL‐2 gene expression were also induced accordingly (6.83 ± 4.67 folds and 8.141 ± 0.85 folds, respectively). In contrast, no concomitant induction of IL‐10 gene expression was detectable (0.83 ± 0.44 folds; IFN‐γversus IL‐10 gene expression P= 0.001; TNF‐αversus IL‐10 gene expression P= 0.02; IL‐2 versus IL‐10 gene expression P= 0.0001).
Most intriguingly, expression of IFN‐γ, TNF‐α and IL‐2 genes was also induced by RQY‐16mer stimulation in purified CD8+ T cells putatively responding only to shorter HLA class I restricted peptides. Indeed, the expression of these genes was increased 5.11 ± 2.94 folds, 6.99 ± 2.45 folds and 5.62 ± 0.95 folds, respectively in RQY‐16mer stimulated CD8+ T cells from HCMV‐seropositive donors, in comparison to unstimulated controls. Again, no induction of IL‐10 gene expression was observed (1.28 ± 0.61 folds; IFN‐γversus IL‐10 gene expression P= 0.02; TNF‐αversus IL‐10 gene expression P= 0.004; IL‐2 versus IL‐10 gene expression P= 0.009, Fig. 1A).
Notably, cytokine gene expression in HCMV‐seronegative donors was not significantly increased following RQY‐16mer stimulation of CD4+ or CD8+ T cells, in comparison with unstimulated cells, and its variations did not reach the calculated cut‐off level of 1.9 folds.
Importantly, a highly significant correlation between the extent of the expression of IFN‐γ and TNF‐α genes in CD4+ or CD8+ T cells upon RQY‐16mer peptide induction and the amount of pp65 DNA transcripts in CD14+ monocytes from the same donors was also observed (CD4+/IFN‐γ: r= 0.797, P= 0.0004; CD4+/TNF‐α: r= 0.710, P= 0.003; CD8+/IFN‐γ: r= 0.734, P= 0.002; CD8+/TNF‐α: r= 0.696, P= 0.004; Fig. 1B–E).
To further support the notion of an antigen‐specific responsiveness to RQY‐16mer, unrelated to unspecific mitogenic effects, five cord blood specimens expressing a variety of HLA class I and II molecules (Table 1B) were collected for ex vivo IFN‐γ gene expression upon peptide stimulation. Four specimens were from HLA‐A*0201 donors. RQY‐16mer failed to induce significant increases in IFN‐γ gene expression in these samples.
In contrast, positive control phytoemagglutinin (PHA) induced a strong cytokine gene expression (P= 0.0001; Fig. 1F). As expectable in cord blood T cells, control virus derived HLA‐A*0201 restricted peptides selected among latent (CMV pp65495−503, BKV LTag579−587, EBV LMP1159−167) and non‐latent viruses (Flu M158−66) were also unable to induce cytokine gene expression.
Cytokine gene expression data indicated that RQY‐16mer peptide was able to efficiently stimulate CD8+ T cells, in spite of its relatively large size. We explored in more detail the ability of each of the eight 9mer‐peptides tiled at one amino acid pace within RQY‐16mer peptide to recall epitope‐specific immune responses from CD8+ T cells previously expanded upon RQY‐16mer induction.
Representative results from three donors (D9, D10 and D11), expressing HLA class I alleles accounting for >60% of the alleles expressed in our study cohort and detectable with high frequency in different populations are reported (Fig. 2A). At least four 9mer‐peptides were found to be able to stimulate significant IFN‐γ gene expression in cells from these donors. Interestingly, the 9mer peptides inducing IFN‐γ gene expression in cells from specific donors were frequently predictable based on the IEDB algorithm. For instance, according to the IEDB MHC binding score (IC50), peptide RQY‐9mer could be restricted by several alleles including, i.e. HLA‐A*0201, A*3001, A*3201 HLA‐B*0702, B*1501, B*2701, B*4001. Indeed, among the donors tested in the experiments, as reported in Fig. 2, this peptide induced significant IFN‐γ gene expression in cells from the two donors (D9 and D10) bearing one or more of these specificities, but its immunogenic activity was barely detectable in the donor (D11) lacking the expression of these alleles (Fig. 2A).
Table 2 cumulatively reports predicted HLA restricting determinants and specific scores of 9mer peptides encompassed by RQY‐16mer. Numbers of donors expressing defined HLA alleles and showing evidence of specific response are also displayed. Notably, considering the redundancy of the restriction of the presentation of some of the peptides under investigation, we are unable to unequivocally attribute responsiveness to restriction by individual allelic products. Nevertheless, it is of interest that at least six of eight peptides nested within the RQY‐16mer sequence were able to induce specific responses. Furthermore, two peptides (RQY‐ and ALF‐9mers) fulfilled immunodominance criteria, as defined by Nastke et al. , in that they were able to stimulate specific responses in at least six of eight donors (75%) expressing HLA‐A*0201. The previously described immunodominant QYD‐9mer  was also found to be able to stimulate specific responses in three of three HLA‐A*2401 donors, thereby confirming the integrity of our experimental approach.
These findings were supported by cytotoxicity assays performed by using RQY‐16mer peptide expanded CTLs from the three representative donors under investigation as effectors and autologous EBV‐transformed B cell loaded with specific HLA class I restricted matching peptides as targets. Significant levels of specific lysis at 10:1 E/T ratio (ΔSL10:1) could be observed by pulsing target cells with 9mer‐peptides capable of inducing IFN‐γ gene expression (Fig. 2B).
To further characterize the CD8+ T‐cell stimulating capacity of RQY‐16mer, multimeric reagents inclusive of HLA‐A*0201 or HLA‐A*2402 and immunodominant RQY‐ or QYD‐9mers, respectively, were generated. These tools allowed to ex vivo stain sizeable percentages of CD8+ T cells from HCMV‐seropositive donors expressing specific HLA alleles (Fig. 2C and F). T cells from the two donors (D9 and D10) were co‐cultured with autologous mDCs in the presence of RQY‐16mer or RQY‐ or QYD‐9mers peptides. Multimer stainings and specific cytotoxic activities were then evaluated. Predictably, RQY‐ or QYD‐9mers were highly effective in expanding specific CD8+ T cells endowed with cytotoxic capacity (Fig. 2D and G). Most interestingly, RQY‐16mer was also effective in expanding CD8+ T cells with similar phenotypic and functional features (Fig. 2E and H).
We hypothesized that RQY‐16mer could be fully internalized into DCs, processed through an exogenous pathway and cross presented in the context of matching HLA class I determinants. To test this hypothesis, iDCs from three HCMV‐seropositive donors bearing different HLA antigens were glutharaldheyde‐fixed and used to stimulate T‐cell responses in the presence of RQY‐16mer or selected 9mer‐peptides. As control, live iDCs were used. We found that the ability of in vitro expanded CD4+ or CD8+ T cells to respond to RQY‐16mer peptide stimulation by IFN‐γ gene expression was significantly jeopardized when fixed iDCs were used as antigen presenting cells, as compared to live cells (CD4+/IFN‐γ: 1.65 ± 0.21 versus 5.08 ± 1.58‐fold increase, P= 0.048; CD8+/IFN‐γ: 1.34 ± 0.25 versus 6.28 ± 1.47‐fold increase, P= 0.046; Fig. 3A). In contrast, fixed iDCs fully retained their capability to activate CD8+ T cells in the presence of 9mer‐peptides, comparably to their unfixed counterparts (Fig. 3B) .
Due to the active role possibly played by CD4+ T cells in controlling viral infections and considering the high immunogenicity of the RQY‐16mer, the activity of specific CD4+ T cells was also tested at protein level. Intracellular cytokine staining (ICS) was performed in CD4+ T cells cultured for 2 weeks in the presence or absence of antigenic peptide and reactivated for 6 hrs by RQY‐16mer loaded autologous mDCs. Seven HCMV‐seropositive donors were selected to represent 90% the HLA‐DRB1* specificities detectable in our group of donors (Table 1A). The two HCMV‐seronegative donors were also included as negative controls.
Cut‐off values (0.18%) were calculated based on the stainings observed in cultures conducted for 2 weeks in the presence of mDCs and IL‐2 but without antigenic stimulation. Percentages of IFN‐γ, TNF‐α and IL‐2 positive cells in 6‐hr assays performed in the presence of antigenic peptide significantly exceeding cut‐off values were observed in antigen pre‐stimulated cultures from each HCMV‐seropositive donor tested irrespective of their specific HLA DRB1* association (IFN‐γ: 4.76 ± 10.18%, median 0.84%, range 0.24–27.8%; TNF‐α: 5.69 ± 13.02, median 0.95%, range 0.22–35.2%; IL‐2: 3.11 ± 7.14, median 0.39%, range 0.23–19.3%, Fig. 4A). In contrast, percentages of ICS cytokine positive cells in 6‐hr assays performed in the absence of antigenic peptide, were equal to or lower than cut‐off values (0.18 ± 0.03% for IFN‐γ, 0.08 ± 0.04% for TNF‐α and 0.08 ± 0.04% for IL‐2). Notably, RQY‐16mer stimulation of cells from HCMV‐seronegative donors did not result in the production of cytokines (<0.05% in all cases). It is of note that analysis of RQY‐16mer by NetMHCIIpan algorithm revealed the existence within this peptide of a number of core sequences capable of binding the DRB1* allelic products expressed by the donors under investigation, as reported in Fig. 4B.
To confirm the ability of RQY‐16mer peptide to activate CD4+ T cells and to demonstrate their possible direct role in the control of HCMV infection, we investigated their antigen‐specific proliferation and cytotoxic activity.
Freshly obtained CD4+ T cells from HCMV‐seropositive donors (n= 4) accounting for a high percentage of the HLA‐DRB1* alleles expressed in our cohort (75%) were co‐cultured with autologous mDCs in the presence or absence of the specific peptide for 7 days. A 0.5 μg/ml concentration of RQY‐16mer sufficed to stimulate a significant proliferation of specific CD4+ T cells in all donors tested, thus confirming its ability to reactivate and expand immune responses from HCMV‐seropositive individual irrespective of their HLA‐DRB1* specificities. It is remarkable that core sequences within RQY‐16mer capable of binding each of these class II determinants were revealed by analysis with NetMHCIIpan algorithm. In contrast no response was induced upon stimulation of cells from a control HCMV‐seronegative donor DRB1*07,14 (P= 0.4) (Fig. 5A).
The cytotoxic capacity of RQY‐16mer peptide stimulated CD4+ T cells was then evaluated by caspase release assays. Immature DCs were infected with HCMV as previously described . CD4+ T cells from HCMV‐seropositive donors bearing different HLA class II antigens (n= 3) were expanded in vitro for 2 weeks by RQY‐16mer peptide stimulation. They were then incubated for 4 hrs with HCMV‐infected iDCs at 10:1 E/T ratio. Cultures were then harvested and stained for intracellular caspase‐3 release and for surface expression of CD1a as marker for DC identification.
Caspase‐3 production in the absence of expanded T cells was negligible in HCMV‐infected iDCs (positive cells = 0.02 ± 0.01%, mean fluorescence intensity [MFI]= 3, left histogram, Fig 5B) and comparable to that of untreated iDCs (data not shown). Co‐culture in the presence of RQY‐16mer peptide expanded CD4+ T cells induced caspase‐3 production in mock‐infected autologous iDC targets, likely due to unspecific stimulation of cytototoxic activity by the IL‐2 used in the T‐cell expansion protocol, and, possibly, at least in part, to their expression of endogenous HCMV genes, including pp65  (positive cells = 19.72 + 7.56%, MFI = 57, mid histogram, Fig. 5B). Superinfection of target cells with HCMV significantly increased their sensitivity to the cytotoxic activity of RQY‐16mer stimulated CD4+ T, resulting in a 60% MFI increase upon caspase‐3 staining (positive cells = 59.55 + 4.01%, MFI = 280, right histogram, Fig. 5B). Figure 5C reports cumulative data obtained in three independent experiments (mock‐infected iDCs fluorescence index [FI]= 3.33 ± 1.52 versus HCMV‐infected iDCs FI = 77.26 ± 7.39, P= 0.0073).
HCMV infection is frequently reactivated in pharmacologically immunosuppressed transplanted patients. Thus, the capability of specific antigenic peptides to stimulate immune responsiveness in these conditions is of utmost clinical relevance. We tested the ability of RQY‐16mer of reactivating a peptide‐specific immune response in isolated T cells from kidney transplanted patients (n= 10; Table 3A). All patients were treated with an immunosuppressive regimen consisting of mycophenolate, cyclosporine and prednisone and tested 183 ± 102 (range 83–374) days after transplantation. One HCMV‐seronegative patient (P3) transplanted with a kidney from a HCMV‐seropositive donor and thereby at high risk of HCMV infection underwent a prophylactic regimen (VALgcv) for the first three months after transplant.
T cells from these patients were cultured in the presence of antigenic peptide for 2 weeks and then tested by ICS as detailed in the ‘Materials and methods’ section. IFN‐γ and IL‐2 protein production upon 6 hrs RQY‐16mer stimulation was significantly increased in comparison with unstimulated cells in CD4+ T lymphocytes from four (P1,2,7,10) and five (P1,2,7,8,10) of the eight HCMV‐seropositive patients, respectively. However, no responses were observed in either patient P3 undergoing primary infection under antiviral prophylactic treatment or in the HCMV‐seronegative patient P5 (Table 3A). Responsiveness to class I restricted epitopes encompassed by RQY‐16mer was also investigated in CD8+ T cells from four patients. Interestingly, in three of them, P4, P6 and P9, showing no CD4+ T‐cell responsiveness, 9mer peptides nested within RQY‐16mer were able to elicit IFN‐γ production in CD8+ T cells. In P10 CD8+ T‐cell responsiveness to nested 9mers matched the vigorous response of CD4+ lymphocytes to RQY‐16mer (Table 3B).
A number of different approaches are currently being investigated to prevent HCMV infection or its reactivation. They include active immunization with live attenuated virus, subunit vaccines, recombinant viral vectors, neutralizing immunoglobulins, synthetic peptides or DNA vaccines . Furthermore, adoptive transfer of HCMV‐specific T cells expanded by using specific and unspecific HLA matching peptides has also been utilized .
In this study we addressed the identification of HCMV pp65‐derived antigenic peptides of potential use across a wide range of HLA specificities. Given the decisive role of helper T cells in the optimal expansion of CTL, and, in particular, in the anti‐HCMV‐specific immune response, the characterization of both class I and II restricted epitopes was envisaged. Clearly, single peptides gathering the ability to stimulate HLA class I and II restricted T cells beyond the limits of foretold MHC‐peptide binding would represent reagents of potentially high clinical relevance.
Algorithms are commonly used to predict HLA‐peptide affinity and they usually provide preliminary insights for the identification of novel antigenic epitopes . Binding to HLA gene products represents a fundamental prerequisite of peptide immunogenicity [37, 38, 39]. However, studies based on linear regression analysis [40, 41] have demonstrated that the immunogenicity of a number of peptides predicted to avidly bind HLA‐restricted molecules is not always validated by in vitro or in vivo studies.
In this work we demonstrate that concurrent use of different epitope prediction algorithms is not only highly informative regarding the likelihood of individual peptides to be processed and presented within specific HLA restrictions but it is also able to predict with low margin of error the immunogenic potential of specific sequences.
The novel HCMV‐derived immunogenic peptide described here is characterized by a potentially high clinical relevance due to a number of important features.
First, RQY‐16mer is able to stimulate HCMV pp65 specific responses in both CD4+ and CD8+ T cells, restricted by a wide range of HLA class I and II determinants. This responsiveness is not related to intrinsic stimulatory activity because it does not extend to cord blood lymphocytes. Thus, a single reagent might qualify as synthetic immunogen for potentially large populations exposed to HCMV infection or reactivation. A large number of pp65‐derived class I restricted antigenic peptides have been described  and key‐target class II restricted peptides have also been identified . However, the number of peptides capable of simultaneously stimulating CD4+ and CD8+ T cells is very limited . Notably, the latter reagents usually induce CTL responses restricted by one HLA class I determinant only. Although they have provided proof of principle of the possibility of activating CD8+ T cells by using ‘long’ peptides, they hardly qualify for a wider clinical use.
Interestingly, RQY‐16mer displays a relatively hydrophobic pattern and its capability to stimulate either CD4+ or CD8+ T‐cell responsiveness requires processing by live antigen presenting cells. These data suggest that its promiscuity could be ascribed to internalization into iDCs, peptide trimming by immunoproteasome activity and processing of shorter sequences swapped into different HLA class I and II clefts to better fit multiple associations, as predicted by algorithm binding scores for RQY‐16mer relevant peptide‐cores. The sequence might thus retain the power of inducing specific immune responses restricted by a wide range of HLA allelic products.
Intriguingly, RQY‐16mer has previously been studied by other groups [30, 43], but its high immunogenicity was not fully disclosed. It is worth noting that in these studies either total PBMC were used as responder cells [30, 44] or T‐cell stimulation was achieved by using TNF‐α matured DC or transfected B cells as APC [43, 45]. Dendritic cells matured through TLR‐4 triggering used in our study are characterized by a significantly higher APC capacity. In particular, they produce IL‐12, while TNF‐α matured DCs fail to do so .
A second important feature of the RQY‐16mer is represented by its capability to induce an unusually wide range of effector functions in CD4+ T cells. As expectable, specific CD4+ T cells are capable of producing IFN‐γ and TNF‐α and of proliferating in response to RQY‐16mer. Most interestingly, however, they are also able to elicit a significant cytotoxic activity against virus‐infected autologous target cells. CD4+ HCMV‐specific CTLs have been typically detected during early phases of HCMV infection . However, their elicitation upon in vitro stimulation by antigenic peptides represents a relatively infrequent event .
Third, and most importantly, the RQY‐16mer is able to stimulate CD4+ and CD8+ T‐cell responses in pharmacologically immunosuppressed patients. These data suggest that this reagent could be advantageously utilized to prevent or treat HCMV infection or reactivation in a population of patients at high risk of developing disease.
Finally, we are confident that the concurrent use of different predictive algorithms, exemplified in this study, could be of particular relevance for the identification of immunogenic sequences within candidate viral and tumour antigens, irrespective of discrete MHC associations [48, 49, 50, 51]. Universal peptides eventually encapsulated into virosome formulations [52, 53] or encoded by viral vectors [52, 54] might represent key‐reagents for the generation of powerful vaccine preparations aimed at strengthening cellular immune responses. Databases inclusive of their sequences might help unravelling specific motives associated with the ability to promiscuously activate T cells across a wide variety of MHC class I and II restricting determinants.
This work was partially supported by a grant from Swiss National Funds for Scientific Research to Giulio C Spagnoli. Cell‐free endotheliotrophic (HUVEC) HCMV VR1814 strain and the goat anti‐IE72 mAb were provided by G. Gerna (Policlinico San Matteo, Pavia, Italy). Recombinant human IL‐4 and IL‐6 were a gift of A. Lanzavecchia (Institute for Research in Biomedicine, Bellinzona, Switzerland). Recombinant human GM‐CSF was a gift from the Laboratorio Pablo Cassara’, Buenos Aires, Argentina.