Genetic Subtypes of Acute Myeloid Leukemia Are Associated with Distinctive Features of the Bone Marrow Microenvironment
Veronica E Klepeis, Lawrence R Zukerberg, Robert P Hasserjian. Massachusetts General Hospital, Boston, MA
Background: The bone marrow (BM) microenvironment, comprising a complex network of stromal cells, has been shown to influence the homing and growth of normal and leukemic stem cells in mouse models. While the WHO categorizes acute myeloid leukemia (AML) based on specific cytogenetic changes and mutational status, a potential relationship between AML genetic subtypes and the BM microenvironment has not been explored. Our goal was to identify features of the BM microenvironment that may be associated with specific AML subtypes and to further our understanding of the tumor-stroma interaction in AML.
Design: We collected 64 cases of newly diagnosed AML, including 16 with mutated NPM1 (8 with FLT3 mutation), 14 not otherwise specified (NOS) with normal karyotype (4 with FLT3 mutation), 20 with myelodysplasia-related changes (AML-MRC), 6 with MLL rearrangement, and 8 with inv(16) or t(8;21). We performed blinded morphologic assessment of BM core biopsies for cellularity, reticulin grade, amount of extracellular material, number of histiocytes (with or without apoptotic activity), and megakaryocyte number. Immunohistochemistry was performed on 38 cases using CD31 to score for vascular density and β-catenin, smooth muscle actin (SMA), and CD10 to score peritrabecular and perivascular stromal cells. The findings were correlated with AML subtypes, clinical parameters, and overall survival.
Results: We identified two distinct microenvironmental patterns: cases with prominent histiocytic and apoptotic activity with sparse perivascular and peritrabecular stromal cells (Hist) and cases with increased peritrabecular and vascular CD10, SMA, and β-catenin positive stromal cells (Trab/Vasc). AML with mutated NPM1 was associated with the Hist pattern (p=0.02), while AML-MRC was associated with the Trab/Vasc pattern (p=0.01). AML-NOS, AML with inv(16)/t(8;21), and AML with MLL showed intermediate features. There were no significant differences in reticulin staining, extracellular material or megakaryocyte number among the AML subtypes. The Trab/Vasc pattern also correlated with lower BM cellularity (p<0.001) and lower white blood count (p<0.001), while the Hist pattern was associated with FAB M4/M5 morphology (p=0.008). There was no significant correlation between histologic pattern and patient survival.
Conclusions: Our findings suggest that heterogeneous molecular mechanisms in AML pathogenesis may also uniquely influence the BM microenvironment. Further studies are needed to examine the relationship of these microenvironmental associations to responses to specific AML therapies.
Wednesday, March 6, 2013 9:30 AM
Poster Session V # 229, Wednesday Morning