Mixed functional characteristics correlating with TCR-ligand koff-rate of MHC-tetramer reactive T cells within the naive T-cell repertoire



The low frequency of antigen-specific naïve T cells has challenged numerous laboratories to develop various techniques to study the naïve T-cell repertoire. Here, we combine the generation of naïve repertoire-derived antigen-specific T-cell lines based on MHC-tetramer staining and magnetic-bead enrichment with in-depth functional assessment of the isolated T cells. Cytomegalovirus (CMV) specific T-cell lines were generated from seronegative individuals. Generated T-cell lines consisted of a variety of immunodominant CMV-epitope-specific oligoclonal T-cell populations restricted to various HLA-molecules (HLA-A1, A2, B7, B8, and B40), and the functional and structural avidity of the CMV-specific T cells was studied. Although all CMV-specific T cells were isolated based on their reactivity toward a specific peptide-MHC complex, we observed a large variation in the functional avidity of the MHC-tetramer positive T-cell populations, which correlated with the structural avidity measured by the recently developed Streptamer koff-rate assay. Our data demonstrate that MHC-tetramer staining is not always predictive for specific T-cell reactivity, and challenge the sole use of MHC-tetramers as an indication of the peripheral T-cell repertoire, independent of the analysis of functional activity or structural avidity parameters.


The almost unlimited size of the T-cell repertoire enables the adaptive immune system to protect the individual against a wide range of pathogens. T cells can recognize pathogen-derived antigens when presented on the cell membrane by the major histocompatibility complex (MHC) via their unique T-cell receptor. An adequate T-cell receptor repertoire size is crucial for the immune system to continuously respond to new threats. The immense size of the T-cell receptor repertoire is accomplished by somatic recombination of the genes encoding the two different T-cell receptor heterodimer chains and is shaped by central tolerance to an estimated range of 106–107 unique T-cell receptors [1]. After exposure to novel pathogen, immune surveying naïve T cells that encounter their cognate antigen become activated, initiate a robust immune response and differentiate into memory T cells. Precursor frequencies of antigen-specific naïve T cells are reported to influence both primary and memory T-cell responses to infection [2-4].

The frequencies of distinct naïve T-cell clones are extremely low and these low numbers have complicated the possibilities to study the naïve T-cell repertoire [5, 6]. Quantification of preexisting antigen-specific T-cell numbers have been established by several functional assays such as limiting dilution analysis and repeated in vitro stimulations with antigen-loaded DCs [7-11]. In addition, MHC class I tetramers have been used to identify antigen-specific CD8+ T cells in the memory and naïve repertoire [12-15]. However, the fact that functional CD8+ T cells responding to infection can generally be identified by MHC-tetramer staining within the memory T-cell repertoire does not imply that T cells identified in the naïve repertoire by MHC-tetramer staining are also generally antigen responsive. As we previously reported that MHC-tetramer positive T cells within in vitro generated primary immune responses could demonstrate impaired T-cell reactivity [16, 17], we aim to address this issue. Our observations may put the MHC-tetramer-based quantification of naïve antigen-specific T cells into a different perspective. Functional characterization of antigen-specific naïve T cells may provide insight into the natural avidity range of T cells present within a MHC-tetramer positive T-cell repertoire in the absence of in vivo antigen exposure and subsequent selection of high-avidity T-cell clones. In addition, it may help to understand intrinsic differences observed between T cells derived from the memory or the naïve compartment [18, 19].

In this report, we present a method for the generation of naïve repertoire-derived antigen-specific T-cell lines based on the combined use of MHC-tetramer staining and magnetic-bead enrichment. This method was used for the generation of functional cytomegalovirus (CMV) specific T-cell lines from seronegative individuals within 4 weeks. We studied the functional characteristics of MHC-tetramer positive CMV-specific T cells derived from the naïve compartment. Although the enriched T cells were isolated based on their reactivity toward specific MHC-tetramers, we observed a large avidity variation within the isolated oligoclonal naïve T-cell populations. Part of the CMV-specific T cells derived from the naïve repertoire exhibited equal antigen-specific functionality as well as high MHC-tetramer staining as their memory response derived counterparts. However, MHC-tetramer staining was not always predictive for antigen reactivity of isolated T cells, since also functional nonreacting T cells appeared MHC-tetramer positive. Our data indicate that when MHC-tetramers are used to isolate antigen-specific T cells in an unbiased T-cell repertoire, both high- and low-avidity T cells may be isolated including T cells that react to peptide in the context of MHC-tetramers but are functionally nonreacting.


Isolation of CMV-specific T cells by MHC-tetramer pulldown from seronegative individuals

Primary CMV-specific T-cell lines were generated from peripheral blood mononuclear cells (PBMCs) of CMV-unexposed healthy donors as defined by CMV serology and confirmed by absence of MHC-tetramer positive T cells. Four CMV-unexposed donors were selected for in-duplicate generation of primary CMV-specific T-cell lines, by incubating 100^106 PBMCs with a set of CMV-specific MHC-tetramers, followed by enrichment of MHC-tetramer positive cells on a magnetic column. The set of MHC-tetramers covered ten immunodominant CMV epitopes distributed over the five most common HLA alleles among Dutch populations (HLA-A*0101, A*0201, B*0702, B*0801 and B*4001). The set of MHC-tetramers used for pull down was adjusted to the HLA-typing of each donor. After isolation of the MHC-tetramer-enriched fraction, T cells were expanded using anti-CD3/CD28 beads in the presence of irradiated autologous feeder cells, IL-2 and IL-15 for 14 days, and analyzed by multiparametric flow cytometry [20].

After MHC-tetramer enrichment and subsequent expansion of all in-duplicate generated T-cell lines, we were able to detect MHC-tetramer positive CD8+ T cells specific for eight of ten CMV-epitopes in one or more of the CMV-unexposed individuals covering all five selected HLA alleles (Table 1). In all the cases, MHC-tetramer positive T-cell frequencies were low and varied between 0.01 and 2.3% of total CD8+ T cells. A representative flow cytometric analysis of one T-cell line with the full panel of 11 different two-color coded MHC-tetramers is shown in Fig. 1, in which five CMV-specific populations of the eight CMV MHC-tetramer populations searched for were observed. In addition, a positive control hepatitis B virus specific T-cell population was detected.

Table 1. CMV-specific CD8+ T cells after first and second MHC-tetramer pull down
DonorHLAExp.First pull downSecond Pull down
  1. Frequencies indicate percentage of total CD8+ T cells.

AA*01 A*02 B*07 B*27S10.04000.080.0409.20.0400.40.10
BA*01 A*02 B*07 B*08S10.02000.02000.03010.40.10000110.4
CA*02 A*03 B*07 B*40S100.03002.300000440
DA*01 A*02 B*07 B*08S10.300.10.20.0100.20.014402.
Figure 1.

Dual-color MHC-tetramer analysis of generated cell lines. CMV-specific T cells were detected by dual-color MHC-tetramer screening after pull down and in vitro expansion. All dot plots are shown with biexponential axes and display fluorescence intensity for the indicated fluorochromes at the top and side of the plot matrix. Non-MHC-tetramer-specific CD8+ T cells are indicated in black. (A) Schematic overview of dual-color MHC-tetramer screen covering 11 different CMV and hepatitis B virus (HBV) epitopes. Each peptide-MHC complex was encoded by a unique combination of fluorochromes to screen for recognition of all epitopes in a single staining. (B) Representative example of MHC-tetramer screen of the pull down S1-derived T-cell line from donor B with a panel of 11 different dual-color MHC-tetramer combinations. Shown are total CD8+ T cells. Specific MHC-tetramer positive T cells will appear double positive for the respective peptide-HLA-fluorochrome combination. To determine the efficiency of the pull down protocol, we included in some experiments a human HBV derived epitope. Detected are six dual-labeled CMV- and HBV-specific T-cell populations: IE1-QIK (red), IE1-ELR (pink), pp50-VTE (dark blue), pp65-YSE (green), pp65-NLV (blue), and HBV-FLP (orange). Indicated percentages are frequencies of gated double MHC-tetramer positive T-cell populations of the total CD8+ T cells.

After 2 weeks of expansion, additional MHC-tetramer isolations were performed to increase the frequencies of MHC-tetramer positive T cells. For each donor, with the exception of donor B, the cell line with the most abundant MHC-tetramer positive T-cell populations was selected for a second pull down. For all donors, repeated pull down using the identical set of MHC-tetramers, as was used for the initial isolation, resulted in increased frequencies of MHC-tetramer positive T cells (Table 1). CMV-specific MHC-tetramer positive T-cell frequencies increased to 9.74, 21.9, 44.0, and 51.4% of total CD8+ T cells for donor A, B, C, and D, respectively. Testing with individual MHC-tetramers revealed that 10 of 14 previously detected populations were increased in frequency after an extra round of MHC-tetramer pull down. The frequency of one population remained similar (donor D; B7-pp65-TPR), and three populations decreased or became undetectable (donor B, C, and D; A2-pp65-NLV). Four MHC-tetramer positive populations were newly discovered after the second pull down (donor A; A1-p50-VTE, donor B; A1-pp50-VTE and B8-IE1-ELR, and donor D; B7-pp65-RPH). Although the prevalence of some populations remained low after repeated MHC-tetramer isolation, most frequencies allowed functional assessment.

Precursor frequencies of MHC-tetramer positive T cells affect the enriched T-cell line composition

Although the composition of the in-duplicate generated T-cell lines was relatively similar for each CMV-unexposed donor, the absence of MHC-tetramer positive T cells specific for some well-described immunodominant epitopes was remarkable. To investigate whether the composition of the generated cell lines was affected by differences in precursor frequencies of antigen-specific T cells, unavailable donor TCR repertoire or reflected low reproducibility of the MHC-tetramer pull down method, we performed a comparative analysis of ten pull down experiments, all starting with 100^106 PBMCs from another CMV-unexposed donor (donor E). Consistent with the previous results, MHC-tetramer positive T-cell populations were detected in each of the ten generated cell lines (Table 2). A1-pp65-YSE MHC-tetramer positive T cells were detected in each cell line, whereas A1-pp50-VTE, A2-pp65-NLV, or A2-IE1-VLE MHC-tetramer positive T cells were detected in one or two of the ten T-cell lines. Importantly, the composition of the T-cell lines cannot be considered as direct calculations of antigen-specific precursor frequencies as they are analyzed after in vitro expansion and its accuracy may be affected by selective expansion of specific T-cell clones. However, for A1-pp65-YSE specific T cells, a minimal number of one cell per 100^106 PBMCs, corresponding with 14 × 106 CD8+ T cells stained positive with the specific MHC-tetramer. The precursor frequencies of T cells specific for any of the other three CMV epitopes tested in donor E may be estimated to be at least fivefold lower. Based on these data we speculate that the composition of the T-cell lines, enriched from a fixed PBMCs number, is affected by variable precursor frequencies.

Table 2. CMV-specific CD8+ T cells frequencies after ten parallel MHC-tetramer pull down experiments
DonorHLAExperimentFirst pull down
  1. Frequencies indicate percentage of total CD8+ T cells.

EA*01 A*02 B*51 B*57S10.480.1600

T-cell lines show CMV-specific IFN-γ production, but MHC-tetramer positive T cells respond differently

To assess the functionality of the enriched MHC-tetramer positive T-cell populations, we stimulated the T-cell lines from the four donors with peptide-loaded target cells and measured overnight IFN-γ production. As may be expected from the low frequencies of MHC-tetramer positive T cells in the cell lines generated after one round of pull down, no or only marginal IFN-γ production was observed when these cell lines were stimulated with the different peptides for donors A, B, and C (Fig. 2A, B, C). Only the enriched S1 T-cell line derived from donor D demonstrates a low IFN-γ response directed against the A1-pp65-YSE peptide (Fig. 2D). Increase in the frequency of the MHC-tetramer positive T-cell populations by a second pull down did result in a peptide-specific IFN-γ production for three of the four tested cell lines derived from donor A, C, and D. However, not all T-cell populations present in these T-cell lines demonstrated to produce IFN-γ upon peptide-specific stimulation. The T-cell line derived from donor A, C, and D that were enriched by second pull down harboring four, one, and seven different CMV-specific T-cell populations, appeared only reactive to one peptide: A1-pp65-YSE, B40-pp65-HER, and A1-pp65-YSE, respectively. Although for donor A and C the observed single peptide reactivity could be explained by the absence of other high frequent MHC-tetramer positive T-cell populations, donor D obtained four populations with frequencies above 2% of total CD8+ T cells but appeared only reactive to A1-pp65-YSE stimulation. In addition, both the high frequent A1-pp65-YSE (10.4%) and B8-IE1-QIK (11%) specific T-cell populations in donor B appeared unable to produce IFN-γ upon specific peptide stimulation.

Figure 2.

Stimulation assay MHC-tetramer enriched T-cell lines. Enriched CMV-specific T-cell lines derived from (A) donor A, (B) donor B, (C) donor C, and (D) donor D were stimulated with single HLA-molecule transduced K562 target cells loaded with specific peptide after one and two rounds of MHC-tetramer pull down. T-cell reactivity was analyzed by measuring the secreted IFN-γ concentration in a standard ELISA. Data are shown as mean/median ± SD/SEM of three replicates and are from one experiment representative of two performed.

To investigate whether the lack of IFN-γ production by some of the MHC-tetramer positive T-cell populations was due to their low frequencies or whether these T cells were incapable to produce IFN-γ upon peptide-specific stimulation, we generated single MHC-tetramer specific T-cell lines. For this purpose, different donor E derived T-cell lines were used for a second enrichment with single MHC-tetramers. This resulted in the generation of seven T-cell lines specific for three CMV-epitopes with frequencies varying from 5.1 to 24.6% of total MHC-tetramer positive T cells. T cells specific for the A2-IE1-VLE epitope were lost after isolation. Subsequently, the single MHC-tetramer-enriched T-cell lines were tested in a peptide titration assay to assess their peptide specificity and avidity (Fig. 3A). Four of the seven enriched T-cell lines demonstrated a clear sigmoid shape antigen-dose response with variable half-maximum response (EC50). The enriched T-cell lines S10 A1-pp50-VTE (13.2%) and S1 A1-pp65-YSE (16.1%) demonstrated EC50 values between 1 and 5 ng/mL, comparable with those of three high-avidity memory control clones, T-cell lines S8 A1-pp65-YSE (18.5%) and S1 A1-pp50-VTE (24.6%) demonstrated EC50 values of approximately 300 ng/mL. No peptide-specific IFN-γ production was observed for S6 A1-pp65-YSE (14.0%), S6 A2-pp65-NLV (5.2%), and S10 A2-pp65-NLV (5.1%) even when these T-cell lines were exposed to very high peptide concentrations. Stimulation with a panel of CMV-protein-transduced (TD) target cells demonstrated that all T-cell lines that demonstrated peptide-induced IFN-γ production were capable of recognizing endogenously presented peptide, irrespective of variable EC50 values (Fig. 3B-D). In addition, we screened the reactive and functional nonreactive MHC-tetramer positive T-cell lines for their capacity to produce 17 other cytokines, chemokines, and growth factors besides IFN-γ upon activation (data not shown). Unlike the reactive T-cell lines, the functional nonreactive T-cell lines were not capable of producing other cytokines. These data indicate that functional heterogeneity exists among the different MHC-tetramer positive T-cell lines generated from an identical donor, and that this does not correlate with the frequencies of the specific T-cell populations.

Figure 3.

Avidity analysis of MHC-tetramer positive T-cell lines. Single MHC-tetramer positive T-cell lines demonstrated a wide range of peptide sensitivity. Single MHC-tetramer positive T-cell lines were generated from donor E and stimulated with (A) peptide-pulsed or (B) pp50, (C, D) pp65 CMV protein transduced (TD) HLA-A*0101+ and HLA-A*0201+ EBV-LCL cells. Specific CMV-peptide was titrated in threefold dilution steps starting from 10 μg/mL. T-cell reactivity was analyzed by measuring the overnight IFN-γ secretion in a standard ELISA. For each epitope, a control clone derived from the memory compartment of a CMV experienced donor was used. Data are shown from one experiment.

To assess whether all MHC-tetramer positive T cells within a single T-cell line contributed equally to the CMV-specific response, the T-cell lines were specifically antigen activated and the individual T cells within the T-cell lines were screened for CMV-specific T-cell activation by measuring the efficiency of TCR internalization and induction of the activation marker CD137. The results shown in Fig. 4 demonstrate that variable parts of the total MHC-tetramer positive T-cell population revealed activation-induced expression of CD137 and decreased MHC-tetramer staining intensity. For the S1 A1-pp65-YSE T-cell line containing 18.2% MHC-tetramer positive T cells, only 7.0 and 9.2% of CD8+ T cells, corresponding with 38–51% of total MHC-tetramer positive cells, demonstrated activation-induced CD137 expression when stimulated with peptide-loaded or CMV-protein TD target cells, respectively, and clearly downregulated their TCR as measured by decreased MHC-tetramer staining intensity. Comparable results were obtained with the S1 A1-pp50-VTE T-cell line containing 44.4% MHC-tetramer positive T cells as 15.9 and 36.5% of CD8+ T cells, corresponding with 36–82% of total MHC-tetramer positive cells, demonstrated activation-induced CD137 expression when stimulated with either peptide-loaded or CMV-protein TD stimulator cells, respectively. Furthermore, for the S6 A2-pp65-NLV T-cell line that previously showed no peptide-specific IFN-γ production, only marginal T-cell activation could be observed when stimulated with CMV-specific antigen. Irrespective of their specificity, all tested T-cell lines demonstrated strong activation by αCD3/28 beads, indicating their intrinsic capacity to become activated via TCR-mediated stimulation. Activated T cells demonstrated a clear downregulation of MHC-tetramer staining after overnight stimulation. These data indicate that a substantial part of the MHC-tetramer positive T cells fail to respond to specific antigen stimulation.

Figure 4.

Flow cytometric analysis of stimulated CMV-specific T-cell lines. To demonstrate antigen specificity and functionality of the donor E derived MHC-tetramer positive T-cell lines, the cell lines were co-cultured with peptide-pulsed (1 μg/mL), CMV-protein transduced (TD) HLA-A*0101+ and HLA-A*0201+ EBV-LCL or αCD3/28 stimulation beads. TCR internalization and activation-induced CD137 expression was analyzed by flow cytometry after 18 h. Data are shown for the representative cell lines S1-A1-pp50-VTE, S1-A1-pp65-YSE, and S6-A2-pp65-NLV. All dot plots are shown with biexponential axes and display fluorescence intensity for the indicated MHC-tetramer or cell markers. Most EBV-LCLs are deleted by CD19+ backgating. However, some EBV-LCL contamination remains and is indicated as CD8-black dots. Non-MHC-tetramer positive CD8+ T cells are indicated in black. T cells expressing CD137 expression after stimulation are indicated in blue. MHC-tetramer positive T cells that do not express CD137 are indicated in red. Frequencies of MHC-tetramer+ or CD137+ T cells are indicated and are percentages of total CD8+ T cells. Data shown are representative of three experiments performed.

Functional heterogeneity by variable TCR-ligand koff-rate of MHC-tetramer reactivity T-cell clones

To study the functional heterogeneity observed among the isolate MHC-tetramer positive T cells, we analyzed the TCR Vβ-usage of the MHC-tetramer positive T-cell lines derived from donor E by staining the cells with MHC-tetramer in combination with a Vβ-antibody kit. At least six different TCRs could be identified within the three enriched A1-pp65-YSE MHC-tetramer positive T-cell lines S1, S6, and S8. For A1-pp50-VTE and A2-pp65-NLV MHC-tetramer positive T cells, three and two different TCR Vβ-usage were detected, respectively. T-cell clones representing the oligoclonal TCR spectrum were generated by single cell sorting and analyzed for their MHC-tetramer specificity. For the six A1-pp65-YSE-specific T-cell clones, variable MHC-tetramer reactivity was demonstrated when stained with a fixed MHC-tetramer concentration (Fig. 5A). The reactivity range varied between the S1.5.Y TCR VβND T-cell clone demonstrating high MHC-tetramer staining relatively similar to that of the YSE memory control clone and the S6.12.Y TCR Vβ22 T-cell clone demonstrating a 2.5-fold lower MHC-tetramer reactivity. As a control for nonspecific MHC-tetramer staining, a VTE memory control clone was stained with the A1-pp65-YSE-specific MHC-tetramer.

Figure 5.

MHC-tetramer reactivity and structural TCR-binding avidity of A1-PP65-YSE-specific T-cell clones. Clonally expanded donor E derived A1-pp65-YSE-specific T-cell clones were compared with a memory control clone and demonstrated a wide range of functional heterogeneity. (A) Flow cytometric analysis of six representative A1-pp65-YSE-specific T-cell clones expressing different TCRs is shown. For flow cytometry, T-cell clones were mixed 1:5 with CD4+ T cells prior to incubation with PE-coupled MHC-tetramer to prevent aggregate formation. Included are a high-avidity A1-pp65-YSE memory and a nonspecific A1-pp50-VTE memory control clone. The MFI of the CD4-MHC-tetramer+ population is indicated in the upper right quadrant. (B) A1-pp65-YSE-specific T-cell clones were stimulated with HLA-A*0101 and HLA-A*0201 positive EBV-LCLs pulsed with a titrated concentration of A1-pp65-YSE peptide starting with a concentration of 3 μM. As a control for high-avidity recognition of endogenously presented peptide, T cells were stimulated with pp65 protein transduced target cells. T-cell reactivity was analyzed by measuring the IFN-γ production after overnight stimulation in a standard ELISA. Data are shown as mean ± SD/SEM of three replicates and are from one experiment representative of two performed. (C) Relationship between MHC-tetramer reactivity and T-cell avidity. High-avidity T-cell clones were defined by their capacity to recognize endogenously presented peptide. Low-avidity T-cell clones were defined by their capacity only to respond to high peptide concentrations and nonfunctional T-cell clones demonstrated no IFN-γ production when stimulated with specific peptide. Data shown are from one representative experiment out of two experiments performed. (D) koff-rate assay of five representative T-cell clones. The t1/2 of 12–24 single cells of each T-cell clone stained with the HLA*0101/PP65-YSE Streptamer is shown. The values above symbols indicate the mean t1/2 with standard deviation (± SD) for each measurement. The MFI of the tetramer staining of these T-cell clones was S1.4: MFI 27.500, S1.15: MFI 31.200, S6.20: MFI 13.000, S6.30: MFI 24.060, and S6.42: MFI 32.450. Data shown are from one experiment.

Subsequently, expanding T-cell clones were analyzed for antigen reactivity by stimulating the clones with titrated concentration of peptide. For the A1-pp65-YSE-specific T-cell clones, a wide avidity range was demonstrated. Two representative T-cell clones expressing two different TCRs (S1.11.Y TCR Vβ4 and S1.5.Y TCR VβND) exhibited a high-avidity recognition pattern similar to that of the memory control clone, one T-cell clone (S6.4.Y TCR Vβ17) demonstrated only marginal IFN-γ production when stimulated with very high peptide concentrations, and three T-cell clones (S6.25.Y TCR VβND, S6.6.Y TCR Vβ2, and S6.12.Y TCR Vβ22) demonstrated no peptide recognition (Fig. 5B). No different cytokine profiles were observed when these T-cell clones were screened for their capacity to produce other cytokines besides IFN-γ (data not shown). All the high-avidity T-cell clones were capable of recognizing endogenously presented peptide. Similar results were obtained for A1-pp50-VTE-specific T-cell clones expressing different TCRs as two of three clones demonstrated high-avidity peptide recognition and one clone demonstrated no peptide reactivity. No IFN-γ production was observed for T-cell clones specific for A2-pp65-NLV (data not shown).

To analyze whether the observed MHC-tetramer reactivity correlated with the detected T-cell avidity, we compared the results of both assays obtained for various A1-pp65-YSE-specific T-cell clones expressing different TCRs (Fig. 5C). Although the high-avidity TCR VβND expressing specific T-cell clones demonstrated relatively similar MHC-tetramer reactivity compared to YSE memory control clone, the high-avidity TCR Vβ4 expressing T-cell clones demonstrated a significant twofold lower MHC-tetramer reactivity (YSE M: p = 0.003, VβND: p = 0.004, as determined by the unpaired t-test; p < 0.05). The MHC-tetramer reactivity of the TCR Vβ4 expressing T-cell clones was relatively similar to that of the nonfunctional TCR Vβ2, Vβ22, and part of the S6-derived VβND T-cell clones and even significantly lower than that of the low-avidity TCR Vβ17 expressing T-cell clone (p = 0.008), demonstrating that functionality of T cells does not correlate with the intensity of MHC-tetramer staining.

To prove that the observed functional heterogeneity was not due to intrinsic differences between the TCR-bearing T cells such as their naïve or memory phenotype, or coreceptor and adhesion molecule expression differences, we performed a Streptamer-based koff-rate assay in which the dissociation kinetics of monomeric peptide-MHC molecules from different T-cell clones was measured [21] (Fig. 5D). Representative nonfunctional T-cell clones (S6.30.Y VβND, S6.42.Y VβND, and S6.20.Y Vβ22) exhibited very fast t1/2 values, whereas the functional T cells clones S1.4.Y VβND and S1.15.Y VβND bound the peptide-MHC complex for longer periods. Although all these T-cell clones exhibited positive MHC tetramer reactivity (S6.30: MFI 24.060, S6.42: MFI 32.450, S6.20: MFI 13.000, S1.4: MFI 27.500, S1.15: MFI 31.200), they demonstrated differential functional avidity that correlated with the koff-rates of monomeric peptide-MHC complexes as a readout for structural TCR binding strength.

In addition, to confirm that the incapacity of T cells to functionally react to cell surface presented peptide-MHC complexes despite their ability to bind to MHC-tetramer and capacity to produce IFN-γ upon nonspecific αCD3/28 is due to a low-affinity TCR and not caused by a different differentiation state of the tested T cells, we introduced a high-affinity pp65-NLV-specific TCR into the nonresponsive A1-pp65-YSE-specific Vβ22 expressing T-cell clone by retroviral transduction [22]. Upon specific stimulation of the introduced TCR with peptide-pulsed and CMV-protein TD target cells, the TCR TD T cells were able to produce significant amounts of IFN-γ, which was not observed by stimulating the endogenous A1-pp65-YSE TCR (Fig. 6A). Introduction of the high-affinity pp65-NLV-specific TCR into the functional A1-pp50-VTE-specific TCR Vβ17 expressing T-cell clone resulted in IFN-γ production after stimulation of both the introduced and endogenous TCR (Fig. 6B). These results indicate that these functionally nonresponsive T cells are intrinsically capable of reacting to peptide-MHC complexes when a high-affinity TCR is introduced.

Figure 6.

Stimulation assay TCR transduced (TD) T-cell clones. After clonal expansion, a high-affinity pp65-NLV-specific TCR was introduced in (A) low-avidity T-cell clone YSE-Vβ22 and (B) high-avidity clone VTE-Vβ17. TCR-TD T-cell clones were stimulated with CMV pp50 or pp65 TD HLA-A1 and -A2 positive EBV-LCL or peptide-pulsed (1 μg/mL) K562 TD with HLA-A1 and -A2 for 18 h. Endogenous and introduced TCR-reactivity was analyzed by cytokine secretion in a standard IFN-γ ELISA. As a control for overall IFN-γ production, T cells were nonspecifically stimulated with αCD3/28 beads. Data are shown from one experiment representative of two performed.


In this study, CMV-specific T cells were isolated from five healthy CMV-unexposed individuals. High-avidity CMV-specific T-cell reactivity was observed for T-cell lines generated from four of five individuals when stimulated with endogenous antigen. Based on specific TCR Vβ-chain expression, MHC-tetramer positive T-cell clones representing the oligoclonal TCR spectrum were generated by single cell sorting. The clonally expanded T-cell clones demonstrated a wide avidity range for their specific antigens. As the selected PBMCs donors were CMV unexposed, no previous clonal selection has taken place among T-cell clones competing for cognate antigen, narrowing the range of CMV-specific T cells toward a few immunodominant clones that can efficiently clear infection [23]. The in vitro generated T-cell lines were solely selected based on TCR-MHC-tetramer with no subsequent selection for CMV reactivity. As the TCR diversity in the naïve repertoire has been reported to be at least 100-fold higher compared to the memory T-cell repertoire and high-avidity T cells are selectively enriched in the memory subset, a substantial part of the isolated T cells will be of low avidity [5, 11]. The composition of the generated T-cell lines most likely reflected the broad MHC-tetramer positive T-cell repertoire before antigen-driven T-cell selection.

It has been reported that frequencies of antigen-specific naïve T cells affect the response kinetics to primary infection and regulate immunodominance [2, 24-27]. Although we assessed T-cell precursor frequencies by indirect quantification, we demonstrate that the number of isolated MHC-tetramer positive T cells was not predictive for the number of functional CMV-specific T cells. The isolation of high-avidity T cells was not restricted to the most frequently detected A1-pp65-YSE-specific T cells, and a substantial part of the less frequently detected A1-pp50-VTE-specific T cells demonstrated high CMV functional reactivity. The precursor frequency of A2-pp65-NLV-specific naïve T cells has been described by direct MHC-tetramer analysis to be 1 in 600 thousand (0.6 × 10−6) CD8+ T cells and conserved among individuals [14]. We estimated the precursor frequency even 55-fold lower to be 1 in 33 million CD8+ T cells and we only isolated nonfunctional T cells. Our data indicate that irrespective to quantitative measurements, we need to assess the quality of MHC-tetramer positive T cells to be able to predict response kinetics of the naïve T-cell repertoire to primary infections.

To study the impact of variable precursor frequencies of CMV-specific naïve T cells on the shaping of a protective memory repertoire, we analyzed 40 seropositive donors by MHC-tetramer staining for the prevalence of CMV-specific T cells (Table 3). The frequent detection of A1-pp65-YSE-specific memory T cells in 14 of the 17 individuals was in line with the frequent detection of high-avidity T cells specific for this epitope in seronegative donors. In contrast, no correlation was observed for the detection of T cells specific for A2-pp65-NLV, as a memory response specific for this epitope was frequently observed in 20 of the 24 individuals but no functional T cells were isolated from seronegative donors. Apparently, high-avidity T cells restricted to these immunodominant epitopes are available in the majority of individual T-cell repertoires. However, the success rate for finding these T cells in the naïve repertoire may be affected by their low precursor frequencies. The incapacity to isolate functional A2-pp65-NLV from the naïve repertoire was also reported by Hanley et al. [28]. These data indicate that the observed precursor frequencies of high-avidity CD8+ T cells within an unprimed naïve repertoire may not directly reflect the shape of the late-phase memory repertoire after in vivo antigen exposure. In vivo priming by APCs and subsequent cytokine-driven expansion of specific T-cell clones may result in uneven clonal expansion of different T-cell clones.

Table 3. Direct MHC-tetramer analysis of CMV-specific CD8+ T cells in 40 seropositive donorsa
  1. a

    Frequencies indicate percentage of total CD8+ T cells.

MHC-tetramer positive16/1714/1720/245/2513/1615/1515/195/19
Average frequency0.320.

The frequent isolation of low-avidity T cells with high MHC-tetramer staining addresses the question whether a correlation between TCR-MHC-tetramer reactivity and the avidity of a specific T-cell exists and whether MHC-tetramer staining can function as a predictive value for T-cell functionality. Although MHC-tetramer reactivity of a given TCR may be predictive for its functionality when memory T cells are compared, the observation that A1-pp65-YSE- and A1-pp50-VTE-specific T-cell clones demonstrated no strict correlation between MHC-tetramer reactivity and T-cell avidity suggests that this may be different for an unbiased CMV-inexperienced T-cell repertoire. We demonstrated that high-avidity CMV-specific T cells derived from the naïve T-cell repertoire showed relatively similar MHC-tetramer reactivity as high-avidity T cells obtained from a memory response. In addition, low or nonfunctional T-cell clones could demonstrate comparable or even increased MHC-tetramer staining when compared to high-avidity T-cell clones with the same specificity and obtained from the same individual. Nonresponsiveness of MHC-tetramer positive T-cell clones has also been reported by others for tetramer positive T-cell-derived from the naïve repertoire [24, 29]. The discrepancy between MHC-tetramer reactivity and T-cell functionality may be explained by the staining with multimerized MHC-peptide complexes. Multimerization of MHC-peptide complexes alter the TCR-MHC-peptide dissociation on- and off-rate kinetics and may result in increased binding avidity of the multimerized MHC-peptide complex to surface TCR [30]. By measuring the TCR binding strength to monomeric peptide-MHC complexes, we demonstrated that the dissociation kinetics correlated with the observed functional avidity of the different T-cell clones, and that this was independent of the intensity of MHC-tetramer reactivity.

In conclusion, the method described in this study may find application in studying the naïve T-cell repertoire before antigen-driven T-cell selection. Our data put a critical note to the direct monitoring of the shape of the peripheral naïve T-cell repertoire independent of the analysis of functional activity or structural avidity parameters.

Materials and methods

Culture conditions and cells

Peripheral blood was obtained from different individuals with informed consent (Sanquin Reagents, Amsterdam, The Netherlands). All experiments were approved by the local medical ethics committees. Blood samples were HLA-typed by high-resolution genomic DNA-typing and serological immunoassay was performed using the AxSYM microparticle enzyme immunoassay (Abbott, Abbott Park, IL, USA) for the detection of anti-CMV IgG/IgM antibodies. PBMCs were isolated by Ficoll gradient separation and cryopreserved for further use. Donor samples were included when CMV serology was negative and when the frequencies of MHC-tetramer positive T cells appeared less than 0.01% of total CD8+ T cells. T cells were cultured in Iscove's modified Dulbecco's media (IMDM; Lonza, Basel, Switzerland) containing 5% human serum, 5% fetal bovine serum (FBS), 2 mM l-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin (Invitrogen, Carlsbad, CA, USA) supplemented with 100 IU/mL IL-2 (Proleukin; Chiron, Amsterdam, The Netherlands). Feeder fractions were irradiated with 50 Gy. Stable Epstein-Barr virus transformed lymphoblastoid B-cell lines (EBV-LCL) and phytohemagglutinine (PHA) blasts were generated using standard procedures.

Generation of peptide-MHC complexes

All peptides were synthesized in-house using standard Fmoc chemisty. Recombinant HLA-A*0201 heavy chain and human β2m light chain were in-house produced in Escherichia coli. MHC class I refolding was performed as previously described with minor modifications [31]. MHC class I complexes were purified by gel-filtration HPLC in phosphate buffered saline (PBS) and stored at 4°C.

Isolation of CMV-specific T cells by MHC-tetramer pull down

Prior to isolation, PBMCs samples were stained with PE-coupled MHC-tetramers for 1 h at 4°C. Subsequently, cells were washed and incubated with anti-PE antibody coated magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). Cells were than isolated by magnetic-activated cell sorting (MACS) using an LS column, following the manufacturer's protocol (Miltenyi). Eluted cells were cultured per 5000 cells with 5^104 irradiated (50 Gy) autologous feeder cells in the presence of 100 IU/mL IL-2, 5 ng/mL IL-15 (Biosource), and 11 000 anti-CD3/CD28 Dynabeads (Invitrogen) for polyclonal stimulation in 96-well plates. Cultures were refreshed at least twice a week. After 2 weeks, cell cultures were analyzed for peptide-specific T-cell populations by MHC-tetramer flow cytometry. Subsequently, MHC-tetramer pull down and expansion procedure was repeated or MHC-tetramer positive T-cell populations were sorted on a FACSAria (BD Biosciences, San Diego, USA) into 96-well plates containing 1 × 105 irradiated feeder cells supplemented with 0.5 μg/mL PHA.

Flow cytometry and T-cell staining

Data acquisition was performed on an LSR-II flow cytometer (BD) with FacsDiva software using the following 11-color instrument settings: 488 nm laser: PI: 685LP, 695/40; PE: 550LP, 575/26; FITC: 505LP, 530/30; SSC: 488/10. 633 nm laser: Alexa700: 685LP, 730/45; APC: 660/20. 405 nm laser: QD800: 770LP, 800/30; QD705: 680LP, 710/50; QD655: 635LP, 660/40; QD605: 595LP, 650/12. 355 nm laser: QD585: 575LP, 585/15; QD565: 545LP: 560/20. Approximately, 1 × 106 PBMCs were stained for each analysis with a final concentration of 2 μg/mL per MHC-tetramer in 100 μL PBS with 5% v/v pasteurized plasma protein solution (GPO). Next, antibody mixtures consisting of CD8-Alexa700 (Caltag-Medsystems, Buckingham, UK), CD4-, CD14-, CD19-FITC, or CD137-allophycocyanin (BD) were added and cells were incubated for 30 min at 4°C. Prior to flow cytometry, cells were washed twice. For dual-encoding MHC-tetramer staining, a set of fluorescently labeled MHC-tetramers was generated in which each specific peptide-MHC complex was encoded by a unique combination of two fluorochromes to screen for recognition of all CMV-epitopes in a single sample. For assessment of TCR-Vβ-chain usage, the TCR-Vβ repertoire kit was used (Beckman Coulter, Takeley, UK) according to the manufacturer's protocol.

Functional analysis

For analysis of IFN-γ production, 5000 T cells were co-cultured with 25 000 target cells loaded with different concentrations of CMV, control peptides, or with artificial Ag-presenting beads. To test the functional activity of the T cells against target cells presenting endogenously processed antigen, we used K562 TD with single HLA molecules or EBV-LCL TD with a retroviral vector containing a CMV-pp65/-pp50 or -IE1 expression construct as target cells [32]. Peptide loading was performed by incubating target cells for 1 h in 96-well plates at 37°C and 5% CO2 in IMDM containing 2% FBS and cells were washed twice before use. After 24 h, supernatants were harvested, and the concentration of IFN-γ was measured by an ELISA (Sanquin Reagents). For FACS analysis of activation-induced TCR internalization, 10 000 effector cells were incubated with 50 000 target cells.

TCR gene transfer

The construction of the retroviral vector encoding the TCR chains of the CMV-pp65-specific T-cell clone has been described earlier [33]. The high-affinity CMV-TCR-AV18 T-cell receptor was TD in the selected T-cell clones 2 days after stimulation as previously described. Marker gene eGFP and NGF-R double-positive T cells were subsequently sorted [34].

koff-rate assays

The dissociation kinetics of monomeric pMHC molecules bound to cell surface expressed TCRs was determined using the Streptamer-based koff-rate assay that was recently described [21]. In short, T-cell clones were incubated with Strep-tactin allophycocyanin (APC) and HLA*0101/PP65-YSE Atto565 double-labeled Streptamers for 45 min on ice. To prevent internalization of MHC molecules, Streptamer-stained cells were constitutively kept at 4°C. Fluorescence images were taken on a Leica SP5 confocal laser scanning microscope before and every 10 s after the addition of d-biotin until complete dissociation of the MHCs.


We thank Menno van der Hoorn, Guido de Roo, and Patrick van der Holst for flow cytometric cell sorting and MHC-tetramer analysis of CMV seropositive donor cohort. This work was supported by the Dutch Cancer Society grant number 07-3825 and SFB TR36 (TP-B10/13).

Conflict of interest

The authors declare no financial or commercial conflict of interest.