Abstract
Abstract
Using a similar strategy that successfully identified PR1 as a leukemia-associated antigen (LAA), we identified two homologous HLA-A2-restricted peptides from cyclin E1 (CCNE1) and cyclin E2 (CCNE2) that could be used to elicit peptide-specific CTL from healthy donors in vitro. Two homologous nonameric peptides from CCNE1 (CCNE1144–152) and CCNE2 (CCNE2144–152), which differ by a single amino acid at position 7, have equal binding affinity for HLA-A2 and each elicited peptide-specific CTL with equal efficiency, as measured by specific lysis of T2 cells pulsed with either peptide (CCNE1 59.7% vs CCNE2 72.6% specific lysis, respectively, at E: T 10:1). TCR-Vβ spectratype analysis showed CCNE1-CTL clones to be derived from 3 Vβ families, while CCNE2-CTL clones were derived from a single Vβ family. The CCNE1-CTL and the CCNE2-CTL bound to each of the CCNE1/A2 and CCNE2/A2 tetramers, but staining intensity was greater for the CCNE1-CTL, suggesting greater TCR avidity of the CCNE1-CTL for both peptides. Because each clone cross-recognized the other homologous peptide, we hypothesized that each clone would efficiently kill leukemia that over-expressed either or both CCNE1 and CCNE2 proteins. FACsorted high avidity CTL showed higher specific lysis of peptide-pulsed T2 than did low avidity CTL (38.8% vs 31.9% specific lysis, respectively, at E: T 10:1, p = 0.02). The fluorescence decay of tetramer dissociation (ln (peptide/HLA-A2 tetramer)) over time was linear for each clone, suggesting that avidity was proportional to TCR affinity and tetramer dissociation t1/2 was determined based on first order kinetics. CCNE1-CTL had higher affinity for CCNE1144–152/HLA-A2 (CCNE1/A2, t1/2=84.5min; CCNE2/A2, t1/2=25.3min) and preferentially killed CCNE1144–152-pulsed T2 cells (CCNE1, 56.9% vs CCNE2, 38%, respectively, at E: T 10:1, p = 0.004). Interestingly, CCNE2-CTL also had higher TCR affinity for CCNE1144–152/HLA-A2 (CCNE1/A2, t1/2=29.5min; CCNE2/A2, t1/2=10.7min), but showed only slightly higher specific lysis of CCNE1144–152-pulsed T2 cells (CCNE1 = 49.3% vs CCNE2 = 44.2% specific lysis, respectively, at E: T 10:1, p = 0.33). Each clone specifically lysed HLA-A2+ T-ALL leukemia cells in proportion to both CCNE1 and CCNE2 protein overexpression assessed by Western blot (CCNE1-CTL, R2=0.89; CCNE2-CTL, R2=0.88). In contrast, healthy HLA-A2+ BM cells, which do not overexpress CCNE1 or CCNE2, and control HLA-A2− CML cells that overexpress both proteins, were not lysed. Both the high and low affinity clones showed equal lysis of T-ALL cells that expressed large amounts of each protein (specific lysis = 24.3% by CCNE1-CTL, vs lysis = 23.8% by CCNE2-CTL, at E: T 10:1). However, high affinity CCNE1-CTL killed T-ALL cells significantly better than low affinity CCNE2-CTL (16.8% vs 6.6% lysis, respectively, at E: T 10:1, p =0.02) when the T-ALL expressed a 2.5-fold lower amount of both CCNE1 and CCNE2 proteins. We conclude that the CCNE1 and CCNE2 homologous self-peptides are lymphoid leukemia-associated antigens. Furthermore, while the higher TCR affinity of CCNE1-CTL suggests that the CCNE1 peptide is the more dominant epitope, ultimate target susceptibility is enhanced due to degeneracy of the resulting CTL clones against homologous peptide epitopes.