Ex Vivo IFN-γ Secretion by Circulating CD8 T Lymphocytes: Implications of a Novel Approach for T Cell Monitoring in Infectious and Malignant Diseases

Cells

Twenty-eight healthy blood donors registered at the blood transfusion center in Lausanne, Switzerland, and three melanoma patients with advanced tumor stage (LAU 156, LAU 337, and LAU 465) were included in this study. All individuals were selected on the basis of HLA-A2 expression as assessed by flow cytometry of PBMC stained with allele-specific mAb BB7.2 (19). Cord blood lymphocytes from healthy newborns were obtained at the Maternity Service, University Hospital (Lausanne, Switzerland). PBMC and cord blood cells were prepared by standard Ficoll-Paque (Amersham Pharmacia Biotech, Uppsala, Sweden), cryopreserved in RPMI 1640/40% FCS/10% DMSO, and stored in liquid nitrogen.

Tetramers and mAbs

Synthesis of PE- and allophycocyanin-labeled HLA-A2/peptide tetrameric complexes was performed as previously described (20, 21). One peptide was the Melan-A_26–35 A27L analog (ELAGIGILTV), which has a higher binding stability to HLA-A*0201 and a higher T cell antigenicity and immunogenicity than the natural Melan-A decapeptide EAAGIGILTV or the nonapeptide AAGIGILTV (22). Other peptides were the Flu-MA_58–66 (GILGFVFTL) peptide from the influenza matrix protein (23), and the pp65_495–503 (NLVPMVATV) peptide from the CMV structural protein pp65 (24).

Synthetic peptides

Peptides were synthesized by standard solid-phase chemistry on a multiple peptide synthesizer (Applied Biosystems, Foster City, CA) by using F-moc for transient NH₂-terminal protection and were analyzed by mass spectrometry. All peptides were >90% pure as indicated by analytical HPLC. Lyophilized peptides were diluted in DMSO and stored at −20°C.

FACS staining and phenotype analysis

CD8⁺ lymphocytes were purified from PBMC by magnetic cell sorting using a MiniMACS device (Miltenyi Biotec, Bergisch Gladbach, Germany). To assess the frequency of Ag-specific CD8⁺ lymphocytes, cells were stained with the appropriate PE- and APC-labeled tetramers at room temperature for 1 h before incubating with CD8^PerCP mAb (Becton Dickinson, San Jose, CA). The observed frequencies of A2/tetramer⁺ cells in CD8⁺ cells were identical using either PE- or APC-labeled complexes. For surface phenotype analysis, cells were incubated with 1) PE- or APC-labeled tetramers for 60 min at room temperature; 2) anti-CCR7 rat IgG mAb 3D12 (25) (kindly provided by M. Lipp and R. Forster, Max Delbrück Institute, Berlin, Germany) for 20 min at 4°C and washed; 3) goat anti-rat IgG (H + L)^FITC (Southern Biotechnology Associates, Birmingham, AL) for 20 min at 4°C and washed; and 4) anti-CD8^PerCP mAb along with either anti-CD28^PE (Becton Dickinson), anti-CD27^PE (PharMingen, San Jose, CA), or anti-CD45RA^APC (Caltag, Burlingame, CA) mAbs for 20 min at 4°C. Cells were then washed once and immediately analyzed on a FACSCalibur (Becton Dickinson) using CellQuest software.

IFN-γ secretion assay

Measurement of IFN-γ secretion by Ag-specific cells was performed when the frequency ex vivo of A2/tetramer⁺ cells was ≥0.04% of CD8⁺ T cells and when sufficient cell numbers were available. This was done in replicates by two independent examiners. Fresh or previously frozen PBMC (1–10 × 10⁶ cells) were resuspended in culture medium consisting of IMDM supplemented with 0.55 mM Arg, 0.24 mM Asn, 1.5 mM Gln, and 8% pooled human A⁺ serum (complete medium). Cells were incubated for 4 h either with 1 μg/ml cognate peptide or Pol_476–484 (ILKEPVHGV) peptide from the reverse transcriptase of HIV-1 (negative control; all subjects in this study were HIV-1 seronegative). Cell surface detection of IFN-γ-secreted molecules was performed as follows: cells were labeled for 5 min at 4°C with an IFN-γ-specific high-affinity capture matrix (Miltenyi Biotec), i.e., a bispecific Ab-Ab conjugate directed against CD45 and IFN-γ. Afterward, cells were transferred into 37°C warm medium for 45 min to permit secretion of IFN-γ, washed, and stained for 30 min at 4°C with anti-CD8^FITC mAb (Becton Dickinson), PE-labeled anti-IFN-γ mAb (Miltenyi Biotec), and APC-labeled tetrameric complexes. Cells to be activated were stained with tetramers for 30 min at 37°C before activation. Otherwise, TCR down-regulation induced by the stimulation (26) greatly reduces the tetramer staining (15, 18). Half of each cell sample was then immediately analyzed by flow cytometry (nonpurified fraction), while the remaining cells were washed and magnetically labeled for 15 min at 4°C with anti-PE Ab microbeads (Miltenyi Biotec). PE-labeled IFN-γ⁺ cells were enriched in two rounds of positive selection by magnetic cell sorting (purified fraction) and immediately analyzed by flow cytometry. The vast majority (90–98%) of positively sorted cells was generally composed of dead cells, as measured by staining with 1 μg/ml propidium iodide (data not shown). However, selective analysis of small living lymphocytes by forward and side scatter gating and exclusion of propidium iodide binding particles allowed a specific detection of CD8⁺ T lymphocytes.

IFN-γ Cytospot assay

Measurement of intracellular IFN-γ production by Ag-specific lymphocytes was performed in replicates by two independent examiners. Thawed PBMC (1–10 × 10⁶) were resuspended in complete medium and incubated for 4 h either with 1 μg/ml cognate peptide or irrelevant HIV-1 Pol_476–484 peptide. After 1 h, 10 μg/ml brefeldin A (Sigma, St. Louis, MO) was added. After 3 additional hours, cells were incubated with 0.8 ml Orthopermeafix (Ortho Diagnostic Systems, Raritan, NJ) for 40 min at room temperature, washed, and stained for 40 min at 4°C with PE-labeled tetramers along with anti-IFN-γ^FITC (PharMingen) and CD8^PerCP (Becton Dickinson). Cells to be activated were stained with PE-labeled tetramers for 30 min at 37°C before activation to reduce TCR down-regulation. Cells were immediately analyzed on a FACSCalibur (Becton Dickinson).

IFN-γ ELISPOT assay

The IFN-γ ELISPOT assay was performed in replicates by two independent examiners, as previously described (4). PBMC (1–10 × 10⁶) were thawed and cultured overnight in 6 ml of complete medium to ensure good viability. ELISPOT plates (Millipore, Bedford, MA) were coated with Ab to human IFN-γ (Diaclone, Besançon, France) and washed six times. One microgram per milliliter peptide and 1.66 × 10⁵ PBMC/well in 100 μl of complete medium were added and incubated for 20 h at 37°C. Assays were performed in six replicates either with the cognate peptide or the irrelevant HIV-1 Pol_476–484 peptide. Cells were removed and plates were developed with a second Ab to human IFN-γ (biotinylated) and streptavidin-alkaline phosphatase (Diaclone). The deduced frequency of peptide-specific CD8⁺ cells was calculated by subtracting the mean number of nonspecific IFN-γ spots in the control sample from the mean number of specific IFN-γ spots in the peptide-stimulated sample.

Frequency and phenotype analysis ex vivo of A2/Melan-A-, A2/Flu-MA-, and A2/CMV-specific CD8⁺ T cells in HLA-A2 healthy individuals

We investigated the frequency ex vivo of circulating CD8⁺ T lymphocytes specifically recognizing Melan-A_26–35, Flu-MA_58–66, or CMV pp65_495–503 peptides in a group of HLA-A2 healthy individuals. Significant numbers (4) of A2/Melan-A⁺, A2/Flu-MA⁺, and A2/CMV⁺ cells were detected in 18 of 25 (72%), 25 of 26 (96%), and 13 of 26 (50%) individuals with a mean frequency of 0.06, 0.13, and 3.2% of CD8⁺ T cells, respectively (data not shown). As illustrated in Fig. 1⇓, A2/Melan-A⁺, A2/Flu-MA⁺, and A2/CMV⁺ CD8⁺ T cells were evaluated for CD45RA and CCR7 expression (27). A2/Melan-A⁺ cells homogeneously presented a CD45RA^highCCR7⁺ phenotype in four of four individuals analyzed (Table I⇓). Independent analyses performed on other HLA-A2 individuals revealed that A2/Melan-A⁺ cells were also CD45RA^highCD45RO⁻CCR7⁺CD27⁺CD28⁺CD57⁻CD62 L⁺ phenotype (data not shown), indicating that Melan-A tetramer⁺ cells exhibit a naive phenotype. In marked contrast, A2/Flu-MA⁺ cells often segregated into two populations, namely CD45RA^lowCCR7⁺ (central memory) and CD45RA^lowCCR7⁻ (effector memory) cells (Table I⇓). A2/Flu-MA⁺ cells were also mainly CD45RO⁺CD27⁺ CD28⁺ (data not shown). As an exception, 90% of A2/Flu-MA⁺ cells from HD 7833 were CD45RA^highCCR7⁻ (Table I⇓) and mostly CD27⁺CD28⁻ (data not shown). Finally, A2/CMV⁺ cells presented various CD45RA phenotypes among individuals but were consistently CCR7⁻ (Table I⇓). A detailed analysis of A2/CMV⁺ cells from HD 2709, 2713, 4474, 7833, and 7847 revealed that ∼65% were CD28⁻ and that ∼50% were CD27⁻ (data not shown). Together, the pattern of cell surface marker expression is compatible with an effector phenotype.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Surface phenotype ex vivo of circulating Melan-A-, Flu-MA-, and CMV-specific CD8⁺ T cells from three representative individuals. CD8⁺ lymphocyte populations were highly purified from PBMC (>98%), stained with the appropriate PE-labeled tetramers and anti-CD8^PerCP mAb, along with anti-CD45RA^APC mAb or rat anti-CCR7 mAb labeled with goat anti-rat IgG^FITC, and then analyzed by flow cytometry. Dot plots are shown for gated CD8⁺ cells. Numbers in the upper quadrants, Percentages of A2/tetramer⁺ cells with the corresponding phenotype.

Cytokine secretion by single Ag-specific cells

The capacity to secrete effector cytokines such as IFN-γ following short-term stimulation is a typical feature of Ag-experienced T cells. We thus assessed IFN-γ secretion by Melan-A-, Flu-MA-, and CMV-specific CD8⁺ T cells by using the IFN-γ secretion assay (9, 10, 11). IFN-γ production was observed in a sizable fraction of ex vivo CD8⁺ T cells (mean, 0.3% of CD8⁺ T cells; SD, 0.3%, data not shown), preventing the analysis of rare Ag-specific populations (as low as 0.04% of CD8⁺ lymphocytes). We surmised that integrating the tetramer technology in this assay would allow us to limit the analysis to Ag-specific A2/tetramer⁺ populations, thus overcoming the high IFN-γ background (Fig. 2⇓A). Using this approach, A2/Melan-A tetramer⁺ cells did not detectably secrete IFN-γ either upon exposure to the control HIV-1 peptide (mean, 0.4% of A2/tetramer⁺ cells; SD, 0.5%) or upon stimulation with Melan-A peptide (mean, 0.3%; SD, 0.5%) in seven of seven individuals analyzed (Table I⇑). In contrast, subpopulations of A2/Flu-MA⁺ cells (mean, 25%; SD, 23%) secreted IFN-γ upon stimulation with Flu-MA peptide in six of seven individuals, whereas no significant secretion (mean, 0.2%; SD, 0.6%) was noted upon exposure to irrelevant peptide (Table I⇑). Again, the Flu-MA-specific cells from HD 7833 behave exceptionally since they did not secrete IFN-γ (Table I⇑). Large numbers of A2/CMV⁺ cells (mean, 56%; SD, 20%) secreted IFN-γ upon stimulation with CMV peptide in nine of nine individuals, whereas no or relatively low proportions of IFN-γ-secreting cells (mean, 1.0%; SD, 1.0%) were observed upon exposure to irrelevant peptide (Table I⇑).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

IFN-γ secretion based detection and enrichment of Melan-A-, Flu-MA-, and CMV-specific CD8⁺ T cells from three representative individuals. A, PBMC were stimulated for 4 h with 1 μg/ml irrelevant HIV-1 peptide or cognate peptide, labeled as described in Materials and Methods, and then analyzed by flow cytometry. Dot plots show staining with PE-conjugated anti-IFN-γ mAb and appropriate APC-labeled A2/tetramers. Numbers in the upper quadrants, Percentages of A2/tetramer⁺ cells with the corresponding phenotype. B, PE-labeled IFN-γ⁺ cells were enriched in two rounds of positive selection by magnetic cell sorting and analyzed by flow cytometry. Numbers in the upper quadrants, Percentages of IFN-γ⁺ A2/tetramer⁺ cells within purified PBMC. All dot plots are shown for gated CD8⁺, propidium iodide-negative cells.

Cytospot and ELISPOT assays

IFN-γ production by Melan-A-, Flu-MA-, and CMV-specific CD8⁺ T cells was independently assessed by intracellular staining of tetramer-labeled cells (Cytospot assay) (15, 16, 17, 18) and by ELISPOT assays (3, 4, 5, 6). As presented in Fig. 3⇓, the frequencies of IFN-γ⁺ cells within A2/tetramer⁺ lymphocytes measured with the secretion assay correlated with the corresponding frequencies measured by Cytospot (p < 0.0001) and ELISPOT assays (p < 0.001). In accordance with the IFN-γ secretion assay, A2/Melan-A tetramer⁺ cells did not detectably produce IFN-γ upon stimulation with Melan-A peptide in seven of seven individuals in both Cytospot and ELISPOT assays. Furthermore, significant fractions of A2/Flu-MA⁺ cells were IFN-γ⁺ upon stimulation with Flu-MA peptide in six of seven individuals. As already observed when using the IFN-γ secretion assay, Flu-MA-specific cells from HD 7833 did not produce IFN-γ in response to challenge with Flu-MA peptide. Likewise, large numbers of A2/CMV⁺ cells were IFN-γ⁺ upon stimulation with the CMV peptide in seven of seven individuals. Notably, no significant IFN-γ production by A2/Melan-A⁺, A2/Flu-MA⁺, and A2/CMV⁺ CD8⁺ T cells was detectable by intracellular staining upon exposure with irrelevant HIV-1 peptide (percentage IFN-γ⁺ cells in A2/tetramer⁺ cells: mean, 0.6%; SD, 0.4%).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Correlation of A2/tetramer⁺ T cell frequencies producing IFN-γ obtained from secretion assay (x-axis) vs Cytospot assay (y-axis, left) or ELISPOT assay (y-axis, right). For secretion or Cytospot assays, PBMC were incubated for 4 h with 1 μg/ml cognate peptide, labeled as described in Materials and Methods, and then analyzed by flow cytometry. Values are the percentage of IFN-γ⁺ cells in gated A2/tetramer⁺ CD8⁺ cells. For ELISPOT assays, PBMC were incubated for 20 h with 1 μg/ml cognate peptide and treated as described in Materials and Methods. The number of IFN-γ-specific spots in PBMC were related to the frequency of A2/tetramer⁺ CD8⁺ cells in PBMC independently measured by flow cytometry. This enables us to calculate the frequency of IFN-γ⁺ cells within peptide-specific CD8⁺ cells. In the Cytospot assay, peptide-specific CD8⁺ T cells did not detectably produce IFN-γ upon stimulation with irrelevant HIV-1 peptide (percentage of IFN-γ⁺ cells in A2/tetramer⁺ cells: Mean, 0.6%; SD. 0.4%). Correlation of IFN-γ production obtained from the different assays was measured by linear regression analysis.

Increasing the sensitivity for the detection of IFN-γ secretion by Ag-specific cells

The IFN-γ secretion assay allows the selective enrichment of PE-labeled IFN-γ⁺ cells using anti-PE magnetic microbeads (10, 11). PBMC challenged with either irrelevant HIV-1 peptide or cognate peptide were stained with anti-IFN-γ^PE mAb and subsequently isolated by magnetic sorting with anti-PE microbeads (Fig. 2⇑B). As shown in Fig. 4⇓, purification of PE⁺ cells upon challenge with cognate peptide (D) confirmed the results obtained with nonpurified lymphocytes (B). Indeed, we could not purify IFN-γ⁺ Melan-A-specific cells, whereas significant numbers of A2/Flu-MA⁺ and A2/CMV⁺ IFN-γ⁺ cells were sorted in six of seven and nine of nine individuals, respectively. As an exception, A2/Flu-MA⁺ IFN-γ⁺ cells were not detected in PBMC from HD 7833, confirming that A2/Flu-MA⁺ cells remained IFN-γ⁻ in this individual. In marked contrast, purification of PE⁺ cells upon exposure to irrelevant HIV-1 peptide (C) brought additional information, as compared with the results completed with nonpurified lymphocytes (A). Indeed, enrichment of PE⁺ cells allowed the detection of A2/CMV⁺ IFN-γ⁺ cells in high numbers in two individuals (HD 2713 and 7524) and to a lesser extent in three others (HD 4470, 7517, and 7833), whereas these cells were not or only poorly detectable within nonpurified lymphocytes and when using ELISPOT and Cytospot assays. Enrichment of PE-labeled IFN-γ⁺ cells, however, did not allow the detection of A2/Melan-A⁺ and A2/Flu-MA⁺ IFN-γ⁺ cells in the same experimental settings.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

IFN-γ secretion-based detection and enrichment of Melan-A-, Flu-MA-, and CMV-specific CD8⁺ T cells. PBMC were incubated for 4 h with irrelevant HIV-1 peptide (A and C) or cognate peptide (B and D), labeled as described in Materials and Methods, and then directly analyzed by flow cytometry (A and B) or magnetically enriched for PE-labeled IFN-γ⁺ cells (C and D). Numbers in the top panels, Percentages of IFN-γ⁺ cells in gated A2/tetramer⁺ lymphocytes. Numbers in the bottom panels, Percentages of IFN-γ⁺ A2/tetramer⁺ cells within purified PBMC. Percentages of IFN-γ-secreting cells found for the different Ag-specific populations were compared using the Wilcoxon test (NS, p ≥ 0.05).

To rule out unspecific labeling of the used mAbs, PE sorting on CFSE-labeled cord blood cells was performed. As expected for naive T lymphocytes, CD3⁺ cord blood cells remained IFN-γ^– and were not purified in two independent experiments (Fig. 5⇓). To test the possibility that a fraction of A2/tetramer⁺ cells may also catch IFN-γ molecules secreted by unspecific cells (bystander effect), we diluted ∼10% naive CFSE-labeled cord blood cells in ∼90% PBMC (HD 7833), known to contain elevated numbers (1%) of CD3⁺ IFN-γ⁺ cells ex vivo. IFN-γ secretion assay upon challenge with irrelevant HIV-1 peptide and subsequent purification of PE⁺ cells allowed the enrichment of IFN-γ⁺ PBMC, but not of IFN-γ⁺ cord blood cells in two independent experiments (Fig. 5⇓). In contrast, cord blood cells were efficiently labeled when adding exogenous IFN-γ, ruling out lack of binding of the bispecific reagent to cord blood lymphocytes. Furthermore, PE⁺ PBMC (HD 7833) were not purified when labeling with IFN-γ catch reagent was omitted (data not shown). These results excluded a bystander IFN-γ⁺ staining and confirmed the high specificity of the reagents used.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Specificity of the IFN-γ secretion assay. Cord blood lymphocytes were labeled with CFSE and incubated for 4 h with irrelevant HIV-1 peptide either alone (left), in the presence of PBMC (middle), or in the presence of PBMC plus exogenous IFN-γ (right). The IFN-γ secretion assay was investigated either directly (top) or after enrichment of PE-labeled IFN-γ⁺ cells (bottom). Numbers, Percentages of cells in the respective quadrants. Dot plots are shown for gated CD3⁺, propidium iodide-negative cells.

Analysis of Melan-A-specific CD8⁺ T cells in melanoma patients

Others and we have observed that tumor-specific CD8⁺ T cells in cancer patients may expand in vivo and develop an Ag-experienced phenotype. To rule out functional unresponsiveness (17), we assessed the capacity of Melan-A-specific CD8⁺ T cells to secrete IFN-γ by using the mentioned IFN-γ secretion assay in three melanoma patients with different clinical conditions. Patient LAU 465 did not receive any therapy at the time of blood collection; patient LAU 156 was treated with IFN-α and developed vitiligo 5 years before blood collection; patient LAU 337 received repeated vaccination with Melan-A_26–35 peptide plus SBA-S2 adjuvant and concomitantly developed a systemic CD8⁺ T cell response against Melan-A_26–35 (18, 28). A2/Melan-A⁺CD8⁺ cells were mostly CD45RA^lowCD28⁻ in patient LAU 156, CD45RA^lowCD28⁺ in LAU 465, and CD45RA^lowCD28⁻CCR7⁻ in LAU 337. As shown in Fig. 6⇓A, large amounts of A2/Melan-A⁺ cells (30–80%) secreted IFN-γ upon stimulation with Melan-A peptide in three of three melanoma patients. Notably, purification of PE⁺ PBMC from LAU 337 allowed the detection of A2/Melan-A⁺ IFN-γ⁺ CD8⁺ cells even without in vitro stimulation (Fig. 6⇓B). This demonstrates that a fraction of Melan-A-specific cells in this individual secreted IFN-γ molecules ex vivo, similar to the ex vivo activity apparent for A2/CMV⁺ cells in some healthy individuals.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

IFN-γ secretion-based detection and enrichment of Melan-A-specific CD8⁺ T cells in patients with malignant melanoma. A, PBMC from patients LAU 465, LAU 156, and LAU 337 were stimulated for 4 h with 1 μg/ml irrelevant HIV-1 peptide or cognate peptide, labeled as described in Materials and Methods, and then analyzed by flow cytometry. Dot plots show staining with PE-conjugated anti-IFN-γ mAb and APC-labeled A2/Melan-A tetramers. Numbers in the upper quadrants, Percentages of A2/tetramer⁺ cells with the corresponding phenotype. B, PE-labeled IFN-γ⁺ cells from patient LAU 337 were enriched in two rounds of positive selection by magnetic cell sorting and analyzed by flow cytometry. Numbers in the upper quadrants, percentages of IFN-γ⁺ A2/tetramer⁺ cells within purified PBMC. All dot plots are shown for gated CD8⁺, propidium iodide-negative cells.