Breaking Immune Tolerance to Granule Proteases with Full-Length Antigen Vaccine in Humanized Transgenic Mice Reveals Alternative Antigen Processing and Immunodominance Heirarchy Applicable to Clinical Immunotherapy.

Abstract 2054

Poster Board II-31

The serine proteinases, human neutrophil elastase (HNE) and proteinase 3 (PR3) are degradative enzymes stored in cytoplasmic azurophilic granules of neutrophils. Both proteins are aberrantly expressed in human myeloid leukemias. Over-expression of PR3 is important in maintaining a leukemia phenotype, while HNE may suppress hematopoietic progenitors. PR3 and HNE contain a conserved nonameric HLA-A*02 restricted T-cell epitope called PR1. The presence of PR1-specific CTL has been correlated with molecular remission in CML. A PR1 nonameric peptide vaccine has been evaluated clinically, with promising results. A limitation of the PR1 peptide vaccine is its restriction to patients with the HLA-A*02 allele. A vaccine comprising a larger portion of the PR3 and/or HNE protein and capable of inducing broader T-cell responses restricted by a wider range of HLA alleles is a potential alternative. We have generated numerous recombinant vaccinia viruses (rVV) expressing forms of PR3 and HNE and have evaluated CTL responses in an HLA-A*02 transgenic mouse model (Tg-A2): (1) rVV-EGFP-PR3, contains a gene encoding EGFP fused to human PR3 lacking the N-terminal signal peptide, while retaining both pro-dipeptide and C-terminal propeptide sequences; (2) EGFP fused to C-terminal truncated PR3 (rVV-EGFP-PR3-T) deleted of PR2 CTL epitope, but maintaining PR1 and; (3) rVV-EGFP-HNE. Tg-A2 were immunized with rVV-EGFP-PR3, and 2 weeks later immune splenocytes were expanded by in vitrostimulation (IVS) with syngeneic naïve mouse splenocytes loaded with an overlapping PR3 peptide library. Intracellular cytokine (ICC) assays using a PR3 peptide library or PR1 9-mer epitope peptide demonstrated a robust PR3-specific CD8+ T-cell response, surprisingly without a detectable response to the PR1 peptide. However, PR2, a decamer, is immunodominant in mice following this immunization. This result was unexpected, since the PR2 epitope was identified by computer prediction as a potential HLA-A*02-restricted epitope within PR3 but, unlike the HLA-A*02-restricted 9-mer epitope PR1, has not been extensively studied as a T-cell epitope in humans. We tested whether the immunodominant PR2-specific response in rVV-EGFP-PR3 immunized mice is also found when a naturally-processed PR3 protein is used as the stimulating antigen, expressed from the cell line K562-A2-PR3, a derivative of K562-A2 endogenously expressing human PR3. Splenocytes from mice immunized with rVV-EGFP-PR3 were expanded by co-culture with irradiated K562-A2-PR3 cells, then tested in ICC assays using PR1 and PR2 CTL epitope peptides. We observed a robust response to PR2 but not PR1. While the PR1 epitope is absolutely conserved between human and murine versions of PR3 and HNE, the PR2 epitope is unique to human PR3, diverging from murine PR3 by 3 amino acids, and by 5 amino acid from human and murine HNE. We speculated that Tg-A2 mice may be tolerant to PR1 due to the presence of this epitope within murine PR3 and HNE, which may account for dominance of PR2-specific responses. To investigate the possibility of inducing PR1-specific CTL in Tg-A2, mice were immunized with rVV-EGFP-HNE, and we detected PR1-specific responses by IVS followed by ICC with the PR1 peptide. These results indicate that tolerance to PR1 can likely be broken by immunization with rVV expressing HNE protein that contains the PR1 epitope. These observations can be explained by either the PR2>PR1 competition hypothesis, or by different amino acid sequences flanking the PR1 epitope in PR3 and HNE proteins affecting processing of the protein to generate the PR1 peptide. To test this question, we used rVV-PR3-T to immunize Tg-A2 mice, then conducted IVS followed by ICC with the PR1 peptide to detect PR1-specific CTL. Results showed that immunization with rVV expressing truncated PR3 reliably induced PR1-specific CTL, unlike immunization with rVV expressing the full-length PR3 antigen. We concluded that PR2 epitope prevented the recognition of PR1. Therefore in Tg-A2 mice, vaccination with poxvirus vaccines expressing PR3 will induce an HLA-A2-restricted CTL response almost entirely focused on PR2, whereas immunization with rVV-HNE will induce PR1-specific CTL. If this were replicated in humans, the majority of PR1-specific CTL detected in CML patients could be derived from HNE rather than PR3 antigen presentation, and that PR2-specific CTL derived from PR3 should be equally frequent as PR1 and likely contribute to clinical GVL and tumor regression.