Cancer Immune Therapy Edited by G. Stuhler and P. Walden ...

8 1 Search for Universal Tumor-Associated T Cell Epitopes
good correlation between peptide prediction and MHC affinity. For other classically
discovered antigens, such as MART-1/Melan A or gp100, not all well-characterized
peptide epitopes receive high predictive binding scores, either because the peptides
are of low affinity or because the algorithms fail to pick out high-affinity peptides
with unusual motifs. The ability to alter certain peptide sequences to increase peptide
affinity for MHC without affect T cell specificity has been used both for enhanced
immune assessment [24, 27] and formulation of vaccines [8].
1.6
Identification of the Telomerase Reverse Transcriptase (hTERT) as a Widely Expressed
Tumor-Associated Antigen
For a prototype of a universal tumor-associated antigens, we recently evaluated the
immunogenic profile of hTERT by epitope deduction [28]. hTERT is expressed in
more than 85% of all tumors, but rarely in normal cells [29, 30]. In tumors that express
telomerase activity, hTERTappears to be found in nearly all tumor cells within
a given lesion. Based on in situ hybridization studies, hTERT expression appears to
increase as tumors progress from carcinoma in situ to primary tumors to metastatic
tumors [31]. hTERT is the most widely expressed tumor antigen yet described.
Evidence that hTERTcan function as a tumor-associated antigen comes from epitope
deduction: peptides derived from hTERT were shown to be naturally processed by tumor
cells, presented in an MHC class I-restricted fashion and function as a target for
antigen-specific CTL [28, 32]. The first such peptide described, I540 (ILAKFLHWL),
is found within the middle of the deduced amino acid sequence of hTERT, roughly 70
amino acids to the N-terminus of the first reverse transcriptase motif. The I540 peptide
is restricted to the MHC class Iallele HLA-A2, the most frequently expressed
HLA allele found among nearly 50 % of Caucasian, Asians and Hispanics, and 33%
of African-Americans. This peptide was initially deduced from the sequence of
hTERT based on computer-assisted analysis of MHC-binding motifs, and was subsequently
shown experimentally to bind strongly to HLA-A2 [28]. CTL generated using
the hTERT peptide in an ex vivo expansion system were specific for the I540 hTERT
peptide in more than 70% of normal donors tested. In vitro, these CTL lines killed a
wide range of hTERT + tumor cell lines and primary tumors in a peptide-specific,
MHC-restricted fashion. Importantly, CTL were shown to lyse U2OS cells retrovirally
infected with full-length hTERT but not U2OS cells infected with vector alone. The
HLA-A2 + sarcoma cell line U2OS is one of the rare human cancer cell lines that does
not express hTERT. Monoclonal antibodies against HLA-A2 blocked CTL lysis of
hTERT-infected U2OS cells. These observations strongly suggested that the I540 peptide
from hTERT is naturally processed by tumor cells and is presented in the context
of MHC class Ion the cell surface where it can be recognized by specific T cells [28].
Parallel work in murine systems [33] and other human in vitro systems [32, 34, 35]
has corroborated these findings. Additional hTERT-derived epitopes have been identified
that are restricted to other common HLA alleles, including HLA-A3 [36] and
HLA-A24 [37]. Table 1.3 describes currently described hTERT CTL epitopes.