Without CD8 expression, only the two highest affinity TCRs (19LF6 and 16LD6) showed significant tetramer staining (Fig. 2B and Supporting Information Fig.1C and D). The co-expression of CD8 significantly enhanced the mean fluorescence intensity (MFI) of tetramer staining for all T cells (Fig. 2B and Supporting Information NVP-BGJ398 cell line Fig. 1C). The tetramer MFI increased

with the TCR affinity by SPR (Fig. 2C); the increase was most significant from the lowest to the second lowest affinity TCRs (W2C8 with a KD ∼100 μM and L2G2 with a KD ∼60 μM). This observation is similar to our previous study performed using primary mouse CD8+ cells [36] and to other studies [8]. Similar to 3D TCR affinity, tetramer staining had no statistically significant correlation with TCR function (R2 = 0.46, p = 0.14, Fig. 2D). Furthermore, the off-rates of tetramer dissociation from hybridoma cells measured by the tetramer decay assay [5, 24] (Supporting Information

Fig. 1D and E) did not correlate with TCR functional activity (R2 = 0.046, p = 0.68, Supporting Information Fig. 1F). A possible reason for the lack of correlation between 3D kinetic parameters measured by SPR and T-cell functional activities could be that the soluble αβ TCR in SPR measurement no longer connects with the cellular environment and hence misses its regulation or constraints [30]. Indeed, recent studies on several mouse TCR systems [26-28, 33] suggest that 2D TCR–pMHC kinetic measurements, which are performed in the native membrane environment, show better Dimethyl sulfoxide correlation with T-cell responsiveness. However, human Romidepsin self-antigen-specific TCR systems have not been investigated. Furthermore, the previous 2D TCR–pMHC kinetic measurements varied the pMHC as opposed to the TCR. Therefore, we asked whether 2D measurements would better correlate the kinetics with responsiveness in our

system. Using the micropipette adhesion frequency assay [37], we first measured the 2D TCR–pMHC interaction using CD8− hybridoma cells. Despite the slow 3D off-rates for some of the TCRs [36], the adhesion frequency (Pa) versus contact time (tc) curves had already reached plateaus at the shortest tc (0.1 s) for all six TCRs (Fig. 3A and Supporting Information Fig. 2A–E). The lack of a gradual transient phase in the binding curves indicates that the 2D off-rates are too fast to be measured by the micropipette system due to its limited temporal resolution (∼0.2 s). Using Eq. (1) (see Materials and methods), we calculated the effective affinities for the panel of TCRs from the plateau Pa levels (Fig. 3C). These 2D affinities showed a positive correlation (R2 = 0.75; p = 0.025) with, but a two-log broader range than their 3D counterparts (Supporting Information Fig. 3A). Because of the fast TCR–pMHC dissociation, we used the thermal fluctuation assay [38] to determine the off-rates (Supporting Information Fig. 4).