Because primary myelofibrosis (PMF) originates at the level of the pluripotent hematopoietic stem cell (HSC), we examined the effects of various therapeutic agents on the in vitro and in vivo behavior of PMF CD34(+) cells. Treatment of PMF CD34(+) cells with chromatin-modifying agents (CMAs) but not hydroxyurea, Janus kinase 2 (JAK2) inhibitors, or low doses of interferon-α led to the generation of greater numbers of CD34(+) chemokine (C-X-C motif) receptor (CXCR)4(+) cells, which were capable of migrating in response to chemokine (C-X-C motif) ligand (CXCL)12 and resulted in a reduction in the proportion of hematopoietic progenitor cells (HPCs) that were JAK2V617F(+). Furthermore, sequential treatment of PMF CD34(+) cells but not normal CD34(+) cells with decitabine (5-aza-2'-deoxycytidine [5azaD]), followed by suberoylanilide hydroxamic acid (SAHA; 5azaD/SAHA), or trichostatin A (5azaD/TSA) resulted in a higher degree of apoptosis. Two to 6 months after the transplantation of CMAs treated JAK2V617F(+) PMF CD34(+) cells into nonobese diabetic/severe combined immunodeficient (SCID)/IL-2Rγ(null) mice, the percentage of JAK2V617F/JAK2(total) in human CD45(+) marrow cells was dramatically reduced. These findings suggest that both PMF HPCs, short-term and long-term SCID repopulating cells (SRCs), are JAK2V617F(+) and that JAK2V617F(+) HPCs and SRCs can be eliminated by sequential treatment with CMAs. Sequential treatment with CMAs, therefore, represents a possible effective means of treating PMF at the level of the malignant SRC.