Abnormal mobilization of CD34+ hematopoietic stem/progenitor cells in peripheral blood (PB) characterizes patients (pts) with Primary Myelofibrosis (PMF), and is attributed to both disruption of CXCR4/SDF-1 axis by a bone marrow (BM) proteolytic environment (Xu M, Blood 2005) and reduced CXCR4 expression on CD34+ cells (Rosti V, BCMD 2007) due to a transcriptional defect (Guglielmelli P, Stem Cells 2007). CXCR4 expression on CD34+ cells was inversely related to their number in PB and associated with advanced disease. We aimed at addressing potential mechanisms of reduced CXCR4 transcription in PMF CD34+ cells. To rule out cytokine-induced transcriptional regulation, CD34+ cells from controls and PMF pts were exposed to a range of SDF-1, G-CSF, TGF-beta and Interferon-gamma concentrations, and CXCR4 protein and mRNA level were concurrently measured by flow cytometry and quantitative PCR (RTQ-PCR), respectively. SDF-1 at 500 ng/mL for 24 hrs reduced CXCR4+/CD34+ cells to less than 10% basal value while CXCR4 mRNA level was unchanged, suggesting ligand-induced receptor internalization but no effect on gene transcription. Other cytokines tested were ineffective. Since abnormal methylation of CXCR4 has been implicated in reduced CXCR4 expression in pancreatic cancer cells (Sato N, Canc Biol Ther 2005), we evaluated methylation pattern of five CpG islands in CXCR4 promoter in PMF CD34+ cells and in HEL, K562 and HL-60 cell lines as control, using both methylation-specific PCR and bisulphite sequencing. A correlation was found between CXCR4 expression and extent of CpG island methylation in cell lines: K562 had 8±6% CXCR4+ cells and all five CpG islands were methylated while in HL-60, that showed bright CXCR4 expression in 90% of the cells, the promoter was in a completely un-methylated state; in HEL cell line, that expressed CXCR4 in 37±10% of cells, partial methylation of CpG island 1 and 5 was observed. On the other hand, CXCR4 promoter CpG island 1 hypermethylation was found in CD34+ cells of 15 PMF pts, and in a minority of them also of CpG island 4 or 5. The effect of hypomethylating agent 5-azacytydine (5-AZA; 1 uM up to 72 hrs) was evaluated in CD34+ cells from 10 PMF pts; starting from 12 hr incubation, significantly increased proportion (from 3- to greater than 10-fold basal level) of CD34+ cells expressed CXCR4, and a three- to five-fold induction of CXCR4 RNA by RTQ-PCR was detectable at 8–12hr. Similar effects were observed after 5-AZA treatment of HEL, while in K562 cells changes in CXCR4 expression were minimal. Concurrent incubation with 5-AZA and HDAC inhibitor SAHA induced effects similar to hypomethylating agent alone. In both PMF CD34+ and HEL cells treated with 5-AZA, hypermethylation of CpG 1 was almost completely reverted at 24–48 hr. Finally, 5-AZA treated CD34+ cells showed enhanced migration in response to SDF-1. These data point to abnormal methylation of CXCR4 promoter as a mechanims for reduced CXCR4 expression in PMF CD34+ cells, possibly contributing to their constitutive migration in PB. Additionally, owing demonstration of an effect of 5-AZA combined with HDAC inhibitor TSA on the proportion of JAK2617V>F PMF mutant clones in culture (Shi J, Canc Res 2007), they represent a proof-of-principle for exploring use of hypomethylating agents in PMF.

Author notes

Disclosure: No relevant conflicts of interest to declare