Leukemia Cell Infiltration Causes Defective Erythropoiesis Partially through MIP-1alpha

Leukemia often results in severe anemia, which may significantly contributes to mortality and morbidity of the patients. However, the mechanisms underlying the insufficient erythropoiesis in leukemia have been poorly understood. In this study, with our recently established non-irradiated MLL-AF9 acute myeloid leukemia (AML) murine model (Cheng H et al, Blood 2015), we observed a significant decrease in hemoglobin and red blood cells (RBCs) of Peripheral blood (PB) in the leukemic mice (n=6 per group, p=0.0122 vs p=0.0003). The absolute numbers of the erythroblasts at different stages (Pro Es, Ery.A, Ery.B, Ery.C) in bone marrow (BM) were also reduced. Consistently, by gene set enrichment analysis (GSEA) of microarray data of LKS⁺ cells (GSE52506 ) from leukemic mice, we found significant down-regulation of erythroid differentiation related genes such as GATA1, FOG-1, LMO2 and KLF1. These genes were more significantly inhibited in megakaryocytic-erythroid progenitors (MEPs) and Pro Es from the leukemic mice. Notably, the MEPs were the most reduced subset among all the committed hematopoietic progenitor cells (HPCs) during leukemia progression (90% decrease compared to control, p = 0.0007). MEPs were gradually accumulated in the G0 phase (from 22% to 70%, p<0.001). In contrast, erythroblasts (Pro Es, Ery.A, Ery.B) were more cycling (G1/S/G2/M) in leukemic mice and the proportions of Annexin V⁺ cells in erythroblasts but not in MEPs were also increased during leukemia development. Colony-forming cell (CFC) assays revealed that BM plasma of leukemia mice exerted an inhibitory effect on both BFU-Es and CFU-Es of BM mononuclear cells (BMMNCs) but not on other types of colonies (40% decrease for CFU-Es, 60% decrease for BFU-Es, p<0.001). Consistently, BM plasma of AML patients could also reduce the yield of BFU-Es and CFU-Es from CD34⁺ cord blood cells (n=7, p=0.006).

To determine which cytokines may be responsible for the inhibitory effect, we collected serum and BM plasma from control and leukemic mice for cytokine array analysis. Among the elevated cytokines, MIP-1alpha was previously reported to be up-regulated in leukemic stem cells and its higher expression was found in the majority of patients with leukemia and a subset of patients with lymphoma and myeloma according to the Oncomine data set. We also confirmed it in a cohort of untreated AML patients (n=32). Importantly, AML patients with higher expression of MIP-1alpha showed reduced survival time (median=13.08 months) compared with the patients with lower expression (median=25.86 months) based on the leukemia-gene-atlas (LGA) analysis (n=72). By the CFC assay and single cell culture with different subsets of hematopoietic stem cells (HSCs) and HPCs, MIP-1alpha was able to largely mimic the inhibitory effect on the erythroid differentiation at both stem cell and progenitor cell levels. Mechanistically, we observed higher expression of MIP-1alpha receptor CCR1 in HSCs, MEPs and erythroblasts than CCR5. Administration of CCR1 antagonist, BX471 could partially recover the yield of erythroid colonies after treatment of MIP-1alpha or leukemia BM plasma. An increase of phosphorylation of p38 (phos p38) and resulted down-regulation of GATA1 after MIP-1alpha treatment were documented by Western blots and immunostaining. In summary, our results demonstrate that leukemic cell infiltration causes severe inhibition of erythropoiesis largely at different erythroid precursor levels and this inhibitory effect is at least partially medicated by elevated MIP-1alpha level via CCR1-p38 activation in the leukemic microenvironment.