Abstract
Introduction:Programmed death-ligand 1 (PD-L1) contributes to tumor escape from immune surveillance by binding to programmed death-1 (PD-1), a negative regulator of T-cell responses. PD-L1 is expressed by both tumor cells and immune cells in the tumor microenvironment. In contrast, the role of PD-L2 in tumor immunity is unclear. To better understand the contribution of PD-L1/L2 to immune escape in marrow-based hematologic malignancies, we used multi-color flow cytometry and immunohistochemistry (IHC) to characterize cell-specific PD-L1/PD-L2 expression in patients with multiple myeloma (MM), myelodysplastic syndrome (MDS) and acute myelogenous leukemia (AML).
Results:PD-L1 was detectable (> 2% positive cells) in 100% of patients, with distinct disease-specific patterns. PD-L1 was expressed most widely in MM, predominantly by malignant plasma cells (median, 95% of plasma cells; n = 5) and lymphocytes (median, 12% of marrow lymphocytes; n = 5). In contrast, in MDS and AML, PD-L1 was more commonly expressed by non-tumor hematopoietic cells: MDS median, 36% of CD34− myeloid precursors and 47% of lymphocytes (n = 13) vs 12% of CD34+ myeloid blasts; AML median, 19% of CD34− myeloid precursors and 26% of lymphocytes (n = 7) vs 16% of CD34+ myeloid blasts were PD-L1+. A higher proportion of CD8+ T cells expressed PD-L1 compared with CD4+ T cells in all 3 malignancies: the CD8:CD4 ratio for PD-L1+ T cells was 3.02 in MM (n = 11), 1.92 in MDS (n = 10), and 1.29 in AML (n = 7). PD-L2 expression was largely absent in AML and MDS (< 2% of CD34+ blasts or lymphocytes expressed PD-L2 (n = 13 MDS and n = 7 AML) but was expressed in a subset of patients with MM on plasma cells (median, 18%; n = 5) but not lymphocytes (< 2%). Across all indications on both tumor cells and lymphocytes, PD-L1 was expressed by a larger fraction of cells than PD-L2 (AML: CD34+ blasts, P < .01 and lymphocytes, P = .005 [n = 7]; MDS: CD34+ blasts, P < .01 and lymphocytes, P = .0001 [n = 13]; MM: plasma cells, P = .03 and lymphocytes, P = .04 [n = 5]). These results were confirmed in a distinct set of 16 cases of primary AML analyzed independently, with detectable PD-L1 expression on myeloid blasts in 14 of 16 cases (88%) without co-expression of PD-L2 (0 of 16 cases). Comparisons of flow cytometry and IHC revealed that some IHC methods may underestimate the prevalence of PD-L1/PD-L2 in marrow-based hematologic malignancies, and results comparing different methods for detecting PD-L1/PD-L2 in the bone marrow will be presented.
Conclusions:PD-L1 is highly prevalent in MM, MDS and AML, with significant expression by non-tumor hematopoietic cells, particularly CD8+ T cells. PD-L2 expression was largely absent in myeloid diseases but detectable in MM. Interestingly, PD-L1 expression was most common on tumor cells in MM and on non-tumor hematopoietic cells in MDS, whereas expression on non-tumor and tumor cells in AML was comparable. These data support clinical development of anti-PD-L1/PD-1 therapies in MM, MDS and AML. Future analyses will determine whether different patterns of PD-L1 expression are associated with clinical efficacy.
Authors M Dail, L Yang, S Rodig, and J Venstrom contributed equally to this work.
Dail:Genentech, Inc.: Employment. Green:Genentech, Inc.: Employment. Ma:Genentech, Inc.: Employment. Robert:Genentech, Inc.: Employment. Kadel:Genentech, Inc.: Employment. Koeppen:Roche: Employment, Equity Ownership. Adamkewicz:Genentech, Inc.: Employment. Byon:Genentech, Inc.: Employment. Woodard:Genentech, Inc.: Employment. Rodig:Bristol-Myers Squibb: Honoraria, Research Funding; Perkin Elmer: Membership on an entity's Board of Directors or advisory committees. Venstrom:Genentech: Employment.
Author notes
Asterisk with author names denotes non-ASH members.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal