Deregulation of pro-inflammation signals plays an important role in MDS and AML. We have previously demonstrated that the histone H3 demethylase JMJD3 (KDM6B), a key regulator of inflammatory genes, is significantly overexpressed in the bone marrow progenitor cell population of patients with MDS (Wei et al, 2013). GSK-J4, a novel selective inhibitor against JMJD3, has been shown to modulate pro-inflammatory signals and affect cell growth in T-ALL and glioma. We therefore sought to evaluate the effect of GSK-J4 in MDS and AML.

First, we observed that GSK-J4 significantly reduces the survival of MDS and AML cell lines (MDS-L and TF1) by 70% at concentrations 1-3 uM. These doses have no obvious cytotoxic effect on primary normal peripheral blood cells. The repression of survival was associated with induction of apoptosis, cell cycle blockade, and inhibition of p38MAPK activation. Second, we evaluated the impact of GSK-J4 on histone H3 methylation and found that the compound did not affect overall histone H3K4 and K27 methylation, which is consistent with previous findings (Ntziachristos et al, 2014).

To pinpoint functionally relevant targets of JMJD3 inhibition in AML/MDS, we sought to identify disease-relevant inflammatory cytokines because JMJD3 binds and modulates promoters of many pro-inflammatory genes, particularly cytokines. We analyzed 38 inflammatory cytokines in bone marrow plasma specimens of patients with MDS (N=37) using a Luminex-based cytokine array, which revealed that levels of the cytokines CCL2 (also called MCP-1) and CCL11 (also called eotaxin-1) are significantly elevated in patients compared to healthy controls by 700 and 330 fold respectively (p=0.002 and 0.04, respectively). Both CCL2 and 11 are c-c motif chemokines whose encoding genes are located in the human chromosome 17q cytokine gene cluster. Furthermore, compared to pre-treatment samples, CCL2 was significantly elevated in paired plasma samples that were collected after hypomethylating agent (HMA) treatment (2.5 fold increase, N=11 pairs, p=0.04). Luminex results were confirmed by cytokine-specific ELISA. Previous studies have shown that JMJD3 binds to CCL2 gene promoter in mouse macrophage cells (De Santa et al, 2009). Together these results implicate that CCL2 may be a relevant pro-inflammatory signal modulated by JMJD3 in MDS/ AML.

We therefore sought to evaluate impact of JMJD3 inhibition on CCL2 production by MDS/AML cells. In MDS-L and TF1 cells, GSK-J4 significantly repressed RNA expression and cytokine release of CCL2. Furthermore, in HMA-resistant MDS-L and TF1 derivative cell lines, productions of CCL2 are elevated by 1.8 and 2.5 fold respectively compared to their parental lines. Treatment of HMA-resistant lines with GSK-J4 improved the anti-leukemia effect of decitabine and had a synergistic effect with HMA.

In summary, we have demonstrated that pharmacological inhibition of JMJD3 by GSK-J4 has anti-leukemia effects in MDS/AML cells and has the potential to improve hypomethylation-based therapy in this disease. A possible mechanism for this effect is through the modulation of the inflammatory cytokine CCL2. To further evaluate this hypothesis, we will characterize the effect of JMJD3 and its inhibition on histone methylation of the CCL2 promoter in future work.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution