Abstract
Syngeneic hematopoietic stem cell transplantation (HSCT) can produce long-term disease-free survival in patients with hematologic malignancies including multiple myeloma (MM). Since donor and host are genotypically identical in this setting, allo-immunity is absent and curative responses may result from donor-derived immune responses against tumor-associated antigens. Our previous studies suggest that B cell responses play a role in successful graft-versus-tumor responses, and their identification may accelerate the discovery of targets of coordinated T cell immunity. To identify graft-versus- myeloma target antigens, we performed a detailed analysis of the humoral immune response in an HLA-A2+ patient with MM who underwent myeloablative syngeneic HSCT and achieved molecular remission now lasting more than 2 years. We probed high-density protein microarrays consisting of ~8000 expressed open reading frames with plasma obtained before and at 3, 6, 12, and 16 months after HSCT. Bioinformatic analysis identified 6 unique candidate antigens that elicited significantly increased post- HSCT antibody reactivity compared to pre-HSCT plasma, plasma from the patient’s stem cell donor, and plasma from 2 age-, sex-, and parity-matched normal controls. Two patterns of antibody reactivity were observed: early responses directed at DAPK2, PIM1, and PRKCB1 peaking at 3 months post-HSCT; and late responses to C1orf116, PDGFRB, and RELA proteins arising at 1 year post-HSCT. By antigen-specific ELISA, positive antibody responses against 4 antigens (DAPK2, PDGFRB, PIM1, PRKCB1) were detected in patients with untreated MM (1/10, 1/10, 1/10, and 2/10, respectively), MM patients with durable responses after autologous transplant (1/10, 1/10, 1/10, and 2/10) and patients with monoclonal gammopathy of unknown significance (3/10, 1/10, 0/10, and 1/10, respectively), but not in 10 normal donors. Moreover, 4 of 6 patients achieving durable remission after allo-HSCT each demonstrated antibody responses against 1 to 3 of the 4 candidate antigens. These antibody responses developed in temporal association with clinical responses. Gene expression analysis using Affymetrix U133Plus 2.0 microarrays revealed high expression of 3 of 6 antigens (PIM1, PRKCB1, RELA) in CD138+-selected MM bone marrow samples (n=152). Using quantitative real-time PCR with gene-specific primers and probes, all 6 antigens had detectable expression in MM bone marrow (>95% tumor). Notably, DAPK2 and PIM1 show higher expression in MM bone marrow compared to normal PBMC (p=−0.008, and 0.056, respectively; exact Wilcoxon rank sum test). To determine whether patients develop specific T cell responses against these antigens in vivo, fresh patient PBMC were stimulated ex vivo with a series of peptides derived from PIM1 and DAPK2 that were consensually predicted to be strong HLA-A2 binders by the IEDB class I, NetMHC, and MHC-I peptide energy binding prediction servers. One peptide, DAPK2156–164 (MLLDKNIPI) elicited strong interferon-gamma secretion after two stimulations consistent with a recall response to this antigen. Ongoing work will further characterize T cell responses against DAPK2 in other serologically reactive post-HSCT MM patients, and will assess T cell responses against PIM1. We conclude that antigens identified by serologic screening after syngeneic HSCT are common myeloma-associated targets in vivo and can elicit coordinated T cell responses. The outlined studies will elucidate the potential of these promising antigens as novel immunogens for targeted immunotherapy.
Erythrocyte-targeted immunomodulatory antigens enabled by in vivo selection of D-peptides