Abstract
The study of normal or malignant hematopoiesis requires the analysis of heterogeneous cell populations using multiple morphological and molecular criteria. Flow cytometry has the capacity to provide multi-parameter information of large hematopoietic cell populations, utilizing various combinations of over 200 molecular markers (CDs). However, current flow cytometry analyses are based on serial gating of two-parametric scatter plots - a process that is inherently incapable to discriminate all subgroups of cells in the data. Here we studied the cellular diversity of normal and leukemic bone marrow samples using multi-dimensional cluster analysis of six-parametric flow cytometry data (CD13, CD33, CD34, CD45, forward scatter and side scatter), focusing mainly on the myeloid lineage. Twenty-three subclasses of cells were identified within normal bone marrow samples, many of them inseparable even when examined at all possible two-parametric scatter plots. Seven subclasses of cells expressing the immature CD34 marker were identified. The multi-dimensional analysis could identify the hematopoietic progenitors, and readily distinguish this population from other types of immature cells. We further addressed the compositional variability of bone marrow samples by designing a classifier that assigns cell samples to subclasses, based on robust six-dimensional characteristics. We then applied this classifier to assess the population diversity in nineteen samples of acute myelomonocytic leukemia (AML M4). The myeloid classifier analyzed simultaneously normal cellular populations (e.g. lymphocytes, garnulocytes), and the diverse populations of AML cells. Most of the AML M4 samples contained an immature population assigned to a myeloid immature subclass distinct from the normal hematopoietic progenitors. Nearly 50% of the AML M4 samples contained a combination of immature myeloid cells and a significant population of immature monocytes. The presence of immature monocytic sub-population correlated with lower leukemia cell burden, and higher platelet counts. In summary, our data indicate that AML M4 may be separated into two distinct types, one restricted to an immature myeloid state, and the other with limited monocytic differentiation capacity. The sub-population composition of the leukemia may affect normal hematopoiesis. Multi-dimentional analysis using the myeloid classifier provides a novel method to accurately define both the increasing numbers of hematopoietic cellular markers and sample heterogeneity in AML. This method should further expand our ability to study normal hematopoiesis, and to identify and monitor hematopoietic diseases.
Author notes
Corresponding author
