Bone marrow mononuclear cells were washed with phosphate-buffered saline (PBS) and stained with the amine reactive viability dye ViViD (Invitrogen/Molecular Probes) as described previously (19). After a further wash, cells were stained at 37 °C for 20 minutes with pretitered APC-conjugated CMV pp65_495–503/HLA A*0201 tetramer mutated in the α3 domain of the heavy chain (D227K/T228A) to abrogate CD8 binding (26); such CD8-null tetramers enable the selective identification of high avidity cognate CD8⁺ T cells (28,37,38). Cells were subsequently stained for surface markers prior to immediate acquisition as described in Figure 1; αCD14 and αCD19 were conjugated to Pacific Blue to fluoresce in the same channel as ViViD. The top row shows a "traditional" gating strategy used with six-parameter flow cytometers, in which lymphocyte populations are identified in a forward scatter-area (FSC-A) versus side scatter-area (SSC-A) plot; tetramer⁺ cells are then displayed in a bivariate plot with CD8 after gating to identify T cells on the basis of CD3 expression. The bottom row shows a polychromatic gating strategy to eliminate aberrant binding events. Singlet cells are first selected in a FSC-A versus FSC-height plot and then CD3⁺ T cells are identified against the "dump" channel to exclude cells that could bind tetramer and mAbs non-specifically (dead/dying cells with compromised membranes, monocytes and B-cells). Fluorochrome aggregates are then excluded from the analysis before gating small lymphocytes on the basis of standard light scatter properties. Finally, tetramer⁺ cells are displayed versus CD8 as shown in the top row. Both rows show the same dataset. The "true" tetramer⁺ CD8⁺ T cell population is cleanly identified with the polychromatic gating strategy (0.62%); the poorly resolved tetramer⁺ CD8⁺ population (1.57%) identified with the traditional gating strategy contains a high proportion of non-specific "background" binding events that are eliminated by inclusion of the "dump" channel (right panel, top row; tetramer⁺ events are shown in blue overlaid on the total population in a bivariate CD3 versus CD14/CD19/ViViD plot). It should be noted that in some cases it appears that activated CD8⁺ T cells can associate with CD14 within the small lymphocyte gate and that, consequently, some antigen-specific events might be excluded from the analysis inadvertently by the inclusion of αCD14 in the dump channel (Betts & Price, unpublished observation).

PBMC were stained with ViViD, pre-titered pHLA A*0201 tetramers as indicated and mAbs specific for cell surface markers as described in the legend to Figure 2; the αCD8 mAbs used in these distinct experiments were conjugated to Pacific Orange (CMV pp65_495–503), Alexa 594-PE (EBV BMLF1_259–267) and Cy7-APC (PR1; proteinase-3, residues 169–177). Expression of CD27 and CD45RO was examined in the bright tetramer⁺ CD8⁺ T cell population and in the "shoulder" that comprises tetramer^dim/neg CD8⁺ T cells. The vast majority of the bright tetramer⁺ CD8⁺ T cells express markers consistent with a memory phenotype (various colors; middle panels), whereas the tetramer^dim/neg CD8⁺ T cells follow a phenotypic distribution similar to that of the overall CD8⁺ T cell population in the sample (red; right panels). Thus, the presence of substantial numbers of naïve CD8⁺ T cells helps to resolve non-specific background staining; this can be useful when staining patterns lack clarity, for example with low avidity cognate CD8⁺ T cell populations that tend to separate poorly. The incorporation of additional memory markers can also be helpful, especially with antigen-specific CD8⁺ T cell populations that tend to exhibit atypical phenotypic characteristics (45).

PBMC were stained with ViViD, pre-titered CMV pp65_495–503/HLA A*0201 tetramers as indicated and mAbs specific for cell surface markers as described in the legend to Figure 2. The D227K/T228A α3 domain mutation abrogates CD8 binding (26), thereby enabling the selective identification of high avidity cognate CD8⁺ T cell populations (28,37,38); the Q115E α2 domain mutation increases CD8 binding from K_D ≈ 120 µM to K_D ≈ 85 mM (24,40) and enables the identification of low avidity cognate CD8⁺ T cells that lie below the threshold of detection with wildtype tetramers under identical conditions (41). In these experiments, tetrameric preparations were standardized for concentration and volume, such that conditions differed solely with respect to the nature of the pHLA A*0201/CD8 interaction; furthermore, all tetramers were prepared afresh from pMHCI monomers frozen at −80 °C to minimize confounding effects related to differential protein stability (48). The titrations shown range from 5 to 0.05 µg pMHCI with respect to the monomeric component in minimal staining volume; median fluorescence intensities (MFIs) of the tetramer⁺ populations are displayed in the right panel. Several effects are apparent. First, signal intensity improves as the affinity of the pMHCI/CD8 interaction increases; this is more apparent at lower pMHCI tetramer concentrations. Second, an "avidity separation" effect is visible within the cognate CD8⁺ T cell population; thus, at more limiting concentrations of wildtype pMHCI tetramer, two distinct populations emerge that presumably represent CD8⁺ T cells with differential avidities for antigen as described previously (28). This segregation is more apparent at higher concentrations with the D227K/T228A pMHCI tetramer as intrinsic avidities for antigen are exposed more clearly in the absence of CD8-mediated compensatory effects (28). In contrast, the low avidity CD8⁺ T cells are enveloped within the overall cognate population in the presence of enhanced CD8 binding; this is consistent with the observation that the Q115E mutation preferentially enhances pMHCI tetramer binding, and hence signal amplification, in the case of lower avidity CD8⁺ T cells (41).

Peripheral blood mononuclear cells (PBMC) were incubated with pre-titered allophycocyanin (APC)-conjugated cytomegalovirus (CMV) pp65_495–503/human leukocyte antigen (HLA) A*0201 tetramer at 37 °C for 20 minutes and then stimulated for 6 hours with the corresponding cognate peptide at 2 µg/ml in the presence of costimulatory monoclonal antibodies (mAbs; αCD28 and αCD49d, each at 1 µg/ml), brefeldin A (10 µg/ml), monensin (0.7 µg/ml) and pre-titered Alexa 680-conjugated αCD107a to detect degranulation as described previously (42). After stimulation, the cells were washed, and then stained sequentially for surface and intracellular markers prior to acquisition as described previously with minor modifications (28). A minimum of 200,000 live cells, identified within the "dump negative" population as depicted for a different dataset in Figure 2, were acquired on a modified LSRII flow cytometer (BD BioSciences) and analyzed with FlowJo software (TreeStar Inc.). The top row shows CD8⁺ T cells expressing cognate TCRs specific for the HLA A*0201-restricted CMV pp65_495–503 epitope that were identified with the corresponding pMHCI tetramer (5.68% of CD8⁺ T cells); degranulation (CD107a mobilization) and cytokine production (IFNγ and TNFα) were then assessed in parallel (tetramer⁺ events are shown in red). Functional CD8⁺ T cells that do not stain with the CMV pp65_495–503/ HLA A*0201 tetramer are observed; this could reflect either non-specific background activation in vivo or antigen-specific activation of CD8⁺ T cells ex vivo that fail to bind cognate tetramer (13–15). Discrimination between these possibilities is difficult due to the fact that the appropriate negative control cannot be conducted because pMHCI tetramers are not biologically inert. Thus, cognate CD8⁺ T cell activation in the absence of subsequent peptide stimulation can be induced by: (i) multimeric TCR engagement by pMHCI tetramers in isolation (26); and, (ii) peptide representation by endogenous MHCI molecules (43,44). The bottom row shows the proportion of non-functional tetramer⁺ cells by sequential gating (from left to right) on CD8⁺ T cells that lack TNFα and IFNγ production and CD107a mobilization in response to antigenic stimulation as described above; in this experiment, 1.30% of CD8⁺ T cells were tetramer⁺ and functionally inert within the limitations of the measured parameters. Atypically, a substantial proportion of cognate CD8⁺ T cells in this experiment failed to produce either TNFα or IFNγ, yet mobilized CD107a; this is consistent with the differential triggering sensitivities of these functions and probably reflects suboptimal stimulation due to pMHCI tetramer-induced down-regulation of cell surface TCRs that might be stimulated more effectively by membrane-bound antigen.

PBMC were stained with BiViD or ViViD, pre-titered Epstein-Barr virus (EBV) BMLF1_259–267/HLA A*0201 (top row) or preferentially expressed antigen on melanomas (PRAME)_100–108/HLA A*0201 (bottom row) tetramers and mAbs specific for cell surface markers as described in the legend to Figure 2. Antigen-specific CD8⁺ T cells identified with phycoerythrin (PE)-conjugated or APC-conjugated pHLA A*0201 tetramers are shown in the left and middle panels respectively. As shown, background staining can be more pronounced with PE-conjugated pMHCI tetramers due, in part, to greater spreading error and the tendency for heterogeneity within such preparations (17,20). Random background originating from individual tetramers representing a given antigen specificity can be discerned by staining with a mixture of differentially conjugated tetramers together (right panel). Cognate CD8⁺ T cells will capture pMHCI tetramers from solution regardless of the associated fluorochrome; singly stained cells likely represent aberrant binding events. In these experiments, pre-mixing of tetramer solutions prior to staining minimizes the bias that can arise from sequential additions due to rapid on-rates for binding (26). Nevertheless, it remains possible that some cognate CD8⁺ T cells will stain with only one of the corresponding pMHCI tetramers simply by chance, although this is unlikely given that many fluorochromes must be taken up for an individual cell to be visible by flow cytometry. Furthermore, doubly stained cognate CD8⁺ T cell populations exhibit reduced individual fluorescence intensities due to competition between tetramers for a finite number of surface TCRs. Thus, this technique is best considered as confirmatory for situations in which low frequency antigen-specific CD8⁺ T cells are difficult to discriminate reliably; this is especially pertinent when traditional gating trees are used (Figure 2).

PBMC were stained with ViViD, washed and then incubated with PBS alone or with PBS containing 100 nM dasatinib for 30 minutes at 37 °C. Subsequently, cells were stained at 4 °C for 20 minutes with pre-titered PE-conjugated CMV pp65_495–503/HLA A*0201 tetramer; dasatinib was maintained in the medium at a concentration of 100nM throughout the staining procedure in the relevant experimental tubes. After a futher wash, cells were stained with mAbs specific for cell surface markers and then with GriViD, an amine reactive viability dye that is spectrally distinct from ViViD. Cells were acquired and analyzed as described in the legends to Figure 1 and Figure 2. The left panels show that tetramer staining is substantially enhanced in the presence of dasatinib; specifically, both greater signal intensity and improved resolution of intermediate binding events are observed. The right panels show the percentage of tetramer⁺ cells that die during the staining procedure. It is established that pMHCI multimer-induced signaling can trigger cell death (17,26,46,47), and it appears that dasatinib can mitigate these effects. In the experiment shown, few cells die during the assay; however, the extent of procedural cell death varies widely in different circumstances and can exert a substantial effect on the results. Similar data were obtained with dasatinib experiments conducted at 37 °C using pMHCI tetramers representing several different specificities (Lissina et al., submitted).