Antigenic peptides encompassing HLA-A*0201 restricted human cytomegalovirus (HCMV) pp65_495–503, Epstein-Barr virus (EBV) BMLF-1_259–267, EBV LMP-2_426–434 and influenza matrix (IM) _58–66 virus derived epitopes [17,18] used to assess specific T cell responses were obtained from Neosystem (Strasbourg, France). Corresponding peptide specific PE labelled HLA-A*0201 multimers were from Proimmune (Abingdon, UK). Hepatitis B virus (HBV) (Engerix, Glaxo Smith Kline, Münchenbuchsee, Switzerland) and influenza (Inflexal, Berna Biotech, Bern, Switzerland) commercial vaccine preparations were used to monitor T-cell responses to vaccination.

PBMC were isolated from peripheral blood of healthy donors by Ficoll gradient centrifugation. When indicated, specific PBMC subpopulations were purified by magnetic cell separation (Miltenyi Biotech, Bergisch Gladbach, Germany) according to producers' protocols. Cells were then cultured in RPMI 1640 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), thereafter referred to as complete medium, and 5% (v/v) human serum (Blutspendezentrum, University Hospital Basel, Switzerland).

For proliferation assays cells were cultured in presence of antigenic preparations or in the absence of stimuli in 96-well flat bottom tissue culture plates (Becton Dickinson, Le Pont de Claix, France), at 2 × 10⁵ cells per well, in triplicates. On day six, cultures were pulsed with 1 μCi per well of [³H] thymidine (Amersham, Little Chalfont, UK) for 18 h and then harvested. Tracer incorporation was measured by β-counting.

PBMC were phenotyped by staining with FITC- or PE-conjugated mouse monoclonal antibodies (mAb) to human CD8 and CD4 (Becton Dickinson, San Diego, CA). CD8+ lymphocytes bearing specific T cell receptors were identified by staining with HLA-A0201 multimers containing the desired peptide (Proimmune, Oxford, UK) [19]. Data were reported as total number of MHC-multimer+/CD8+ cells obtained from volumes of WB equal to those utilized for RNA extraction.

Antibody response to HBs Ag was evaluated by ELISA assays (Architect System, Abbott, Sligo, Ireland) in sera from naïve or vaccinated donors.

Elispot assays for the enumeration of IFN-γ or IL-2 producing cells were performed as described previously [20].

Appropriate concentrations of specific antigens, in the form of peptides or commercial vaccine preparations (see above) were added to 0.3 ml of heparinized peripheral blood in 2 ml tubes. Samples were then centrifuged for ten seconds in a minifuge to bring cells in close contact and incubated for 5 h or 16 h, for peptide or vaccine preparations, respectively, at 37°. Three volumes of RNAlater (Ambion, no. AM7020, Austin TX) were then added to stabilize RNA. The mixture was then either stored at different temperatures (see below) or treated immediately for RNA extraction. Sterile hoods, incubators or ≤-20°C refrigerators were not required.

Total cellular RNA was extracted by using Ribo Pure-Blood kit (Ambion Inc., no. AM1928, Austin, TX, USA) and eluted in 75 μl of elution buffer. Reverse transcription was done with 11 μl of total RNA by priming it with 1 μl (200 μg/ml) of Oligo dT (Roche Diagnostics, Mannheim, Germany) at 65°C for 10 minutes and quick chilling on ice. This mixture was supplemented with 1 μl 10 mM dNTP mix, 4 μl 5× first-strand buffer, 2 μl 0.1 M DTT and 1 μl (200 units) M-MLV reverse transcriptase (all by Invitrogen Ltd., Paisley, UK) and incubated at 37°C for 1 hour. Two μl of cDNA were used for each PCR amplification by "real time" technology (7300 Real Time PCR system, Applied Biosystems, Rotkreuz, Switzerland) according to manufacturer's recommendation in the presence of primers and probes specific for genes encoding IFN-γ, IL-2, IL-6, IL-10 and TNF-α as already described [21] or MIP-1β (Assays-on-demand, Applied Biosystems, Rotkreuz, Switzerland). Antigen driven cytokine gene expression (triplicate average) was normalized to the detection level of the internal control β-actin house-keeping gene (Pre-developed assays, PDAR, Applied Biosystems, Rotkreuz, Switzerland). Expression data were calculated, as referred to β-actin gene expression in each sample, by using the 2−ΔCt method [22]. For all genes analysed, dynamic linear range of expression-detection was consistent at least up to Ct value of 35 which was therefore considered as the cut-off of significant values. A threshold of 2-fold increase in specific gene expression over control values was considered as cut-off for the definition of positive responses.

All statistical analyses were performed by using SPSS 15.0 software for Windows (SPSS Inc. Chicago, IL, USA). Correlations between the expression of different cytokine genes and MHC-multimer staining or antibody titres were evaluated by the Kendall's tau correlation coefficient (r) and data were considered statistically significant in the presence of p < 0.05. The significance of differential gene expression in paired samples at different days after influenza vaccination was analyzed by the non parametric Wilcoxon signed rank test.

In initial studies we addressed the possibility of using qRT-PCR technology coupled with RNA extraction from WB samples to magnify antigen specific immune responses from low numbers of T cells. To provide reliable quantitative assessments, we spiked cells from a HLA-A0201 restricted CD8+ CTL clone recognizing gp100_280–288 melanoma associated epitope in allogenic WB from a HLA-A0201+ healthy donor and we incubated the mixture for 5 hours in the presence of a 10 μg/ml final concentration of specific or control (Melan-A/MART-1_27–35) peptide. Total cellular RNA was then extracted, reverse transcribed and amplified in the presence of primers and probes specific for β-actin house keeping gene and genes encoding different cytokines.

Expression of IFN-γ and IL-2 genes was significantly (p < 0.05) increased in cultures performed in the presence of specific, as compared to control peptides (figure 1, panel A), thus ruling out the possibility of a prevailing allospecific responsiveness from host WB T cells. In line with these data, the extent of the increased expression of these genes was strictly dependent on the number of spiked gp100_280–288 specific CTL. Most importantly, these results indicate that specific antigen stimulation provides an activation signal detectable 4.8-fold and 2-fold above background for IFN-γ and IL-2, respectively, in WB down to a minimum concentration ≤50 CTL in a 300 μl sample, thus suggesting that qRT-PCR monitoring of antigen specific immune responses in WB is feasible with a sensitivity comparable to that of qRT-PCR monitoring in ficoll isolated PBMC [8].

RNA stability might of decisive relevance in the performance of qRT-PCR and critically affect immune monitoring methods based on the analysis of cytokine gene expression, particularly in the context of field studies. Thus, we stored antigen stimulated WB samples supplemented with "RNAlater" at room temperature, at 4°C or at -20°C for one week prior to gene expression analysis. We observed that no significant variations of specific signal were detectable in whole blood samples stored in the different conditions under investigation (data not shown).

Based on these studies we attempted the detection of cellular immune responses directed against HLA-class I restricted epitopes derived from viral antigens. WB from two different HLA-A0201+ seropositive donors was incubated in the presence of HCMV pp65_495–503, or EBV LMP-2_426–434 and EBV MLF-1_259–267 virus derived epitopes at a 10 μg/ml final concentration. A well characterized HLA-A0201 restricted influenza matrix (IM) _58–66 peptide was also used at the same concentration. Moreover, in order to further support the specificity of the WB assays, we comparatively evaluated in the same amounts of WB multimer staining and cytokine gene expression upon peptide stimulation.

Data from the two donors are reported in figure 2, panels A and B. In both cases a highly significant correlation was observed between the level of IL-2 and IFN-γ gene expression (r = 0.854 p = 0.001 and r = 0.629, p = 0.012 respectively) induced by HCMV pp65_495–503, EBV BMLF-1_259–267, EBV LMP-2_426–434 and IM_58–66 HLA-A0201 restricted peptides and the numbers of CD8+ T cells stained by specific multimers in the same amount of WB (300 μl). Notably, a 4.5-fold increase in IFN-γ gene expression in IM_58–66 stimulated, as compared to control WB from the donor depicted in figure 2 panel A, was observed in the presence of only 41 CD8+ cells staining positive for the specific multimer, thus confirming spiking data (see above).

Data regarding cytokine gene expression in WB from spiking experiments or upon stimulation with peptides derived from viral antigens suggested the feasibility of a sensitive WB monitoring of cellular immune responses. Validation of this technology, however, requires comparison with known clinical end points or accepted surrogate markers. Thus, we comparatively analyzed cytokine gene expression induced in WB by hepatitis B virus surface antigen (HBsAg) and specific antibody titers in healthy donors (n = 29 for a total of n = 39 samples) vaccinated against Hepatitis B virus. Samples from naïve, seronegative donors were also studied (n = 9). WB specimens were cultured o/n in the presence of a commercial vaccine preparation (see "materials and methods") diluted to a final HBsAg concentration of 2 μg/ml. We found a highly significant correlation between antigen stimulated expression of IL-2 gene as detectable by the WB assay and specific antibody titers (r = 0.50, p = 0.000009) (figure 3, panel A) in donors vaccinated between six months and twenty years before the tests. Expression of IFN-γ and TNF-α genes was also significantly, albeit not as strikingly, correlated with specific antibody titers (r = 0.29, p = 0.012 and r = 0.28 p = 0.013, respectively) (figure 3, panels B and C). HBsAg induced IL-2 gene expression was also highly significantly correlated with IFN-γ and TNF-α gene expression (r = 0.50, p = 0.0000085 and r = 0.44 p = 0.0001, respectively). Confirmative tests performed on purified T cells showed that the expression of these cytokine genes was mainly due to CD4+ T cell activation (data not shown).

In contrast, expression of IL-6 or IL-10 genes upon HBsAg stimulation was modest and neither correlated with specific antibody titers nor with each other, nor with IL-2, IFN-γ and TNF-α gene expression (figure 3, panels D and E).

Comparative assays were performed with samples from two seropositive and one seronegative donor. HBsAg was able to induce IL-2 and IFN-γ gene expression in cells from seropositive donors only. However, no antigen specific lymphoproliferation or cytokine production, as detectable by ELISPOT could be observed in any of the donors under investigation, suggesting that WB qRT-PCR monitoring may be endowed with a higher sensitivity, as compared to these techniques.

In an effort to identify additional markers of cellular immune response in WB correlating with HBsAg specific antibody titers, expression of chemokine genes was also evaluated. We observed that HBsAg stimulated expression of MIP-1β (CCL4) gene was highly significantly correlated with specific antibody titers (r = 0.39, p = 0.0006) (figure 3, panel F). Notably, the extents of IL-2 and MIP-1β gene expression induced by HBsAg were also highly significantly correlated with each other (r = 0.48, p = 0.00002). Furthermore, MIP-1β gene expression was highly significantly correlated with IFN-γ and TNF-α gene expression as well (r = 0.37, p = 0.001 and r = 0.48, p = 0.00001, respectively, figure 3, panels G-I). Thus, WB monitoring technique helped defining a novel gene expression profile significantly correlated with protection against HB infection.

These results stemmed from experiments performed at single time points. In order to further validate the WB protocol proposed here, a prospective longitudinal study aimed at the monitoring of cellular immune response to influenza virus specific vaccination (winter 2007) was then performed. WB from healthy donors (n = 8) obtained prior to influenza vaccination and at different time points, 2–16 weeks after it, was cultured o/n, as detailed above, in the presence or absence of a commercial vaccine preparation (see "materials and methods") diluted to a final concentration of influenza hemagglutinin of 0.6 μg/ml. Total cellular RNA was then extracted and reverse transcribed and cytokine gene transcripts were amplified in the presence of specific primers and probes. Interestingly, significant increases in antigen specific IFN-γ gene expression as compared to pre-immunization values were detectable at two and four weeks after vaccination (p = 0.04 and p = 0.01, respectively), declining thereafter. At four months after vaccination, however, levels of antigen stimulated IFN-γ gene expression were back to pre-vaccination values. Notably, antigen specific IL-2 gene expression displayed a trend (p = 0.05) towards increased values in the weeks following vaccination and still showed a significant (p = 0.04) responsiveness at four months after administration of the vaccine (figure 4, panels A-B).