Proteomics To Identify Novel Immune-Targeted CML Stem Cell Antigens.

Eradication of residual disease remains a critical problem in patients with chronic myeloid leukemia (CML). Donor lymphocyte infusion (DLI) for the treatment of relapsed CML after transplantation is a robust example of the graft-versus-leukemia (GvL) effect, producing complete durable remission in 75% of patients. Since CML originates as a clonal stem cell disorder, we hypothesize that the curative effect of DLI results from immunologic targeting of the self-renewing CML progenitor cell. We have focused on B cell responses as a critical element in effective tumor immunity after observing significant B cell lymphocytosis in 7 responders of DLI at 6 and 9 months following DLI compared to pre-treatment levels (p=0.03; p=0.04, respectively), which was not seen in 5 DLI nonresponders. Moreover, some DLI responders showed marrow plasmacytosis at the time of cytogenetic response. To identify targets of this visible B cell response, we used post-DLI plasma as a source of immunoglobulin to probe two complementary immunoproteomic platforms and identify antigens eliciting increased reactivity compared to pre-DLI plasma. In the first approach, plasma from 7 individual responders was screened for binding to proteins expressed from a CML cDNA phage library (SEREX). In the second approach, plasma from 3 of the 7 responders was screened for binding to proteins expressed from approximately 5000 open reading frames spotted on a protein microarray (ProtoArray, Invitrogen). Screening by SEREX yielded 24 and by ProtoArray 58 candidate antigens. Analysis of datasets generated from normal and malignant myeloid cells on HG-Focus and HG-U133A Affymetrix microarrays confirmed that most candidate targets were expressed in CML progenitor cells. Of the antigens represented on the microarrays, 83% (10 of 12) and 64% (18 of 28) of the SEREX- and ProtoArray-identified antigens, respectively, were expressed in CML CD34+ cells. EXOSC5 was expressed on CML CD34+ cells and was identified by both platforms. Four of 16 ProtoArray-identified antigens (RAB38, TBCE, DUSP12, RPS6KC1) were expressed at higher levels in CML compared to normal CD34+ and mature myeloid lineage cells (p<0.001; p<0.002, respectively). SEREX screening identified exclusively intracellular antigens, whereas ProtoArray analysis yielded four surface-expressed proteins (TGFBR2, TNFRSF19, ACVR2B, and RET). Our studies have thus uncovered a broad panel of novel antigens expressed in CML progenitor cells that are targeted by the polyclonal B cell response associated with GvL. Compared to SEREX, the ProtoArray is a less disease-specific platform that can enable rapid identification of many antigens with compelling attributes for vaccine development in CML, including cell-surface expression and differential expression between normal and tumor stem cells. We are incorporating the entire panel of CML-associated stem cell antigens onto a custom protein array. A CML antigen array provides a powerful new tool for monitoring antigen-specific immunity in the context of immunotherapies designed to eliminate persistent residual disease in CML.