Awakening your innate power to combat AML

In this issue of Blood,Felipe Fumero et al show that activating the histone demethylase plant homeodomain finger protein 8 (PHF8; KDM7B) can trigger cell-intrinsic immune responses and specifically target acute myeloid leukemia (AML) cells.¹ 

Epigenetic modifying enzymes that are recurrently mutated in AML have become attractive targets because of the reversible nature of epigenetic modifications and relative feasibility of developing small-molecule inhibitors that target their catalytic domains or structurally rigid motifs.² In this regard, the histone lysine demethylase KDM family has been shown to play critical roles in leukemogenesis,³ with inhibitors of lysine-specific demethylase 1 (LSD1/KDM1A) already in clinical trials,⁴ and preclinical studies ongoing for other KDMs.³ KDMs also have been implicated in the response to different therapies, such as all-trans retinoic acid (ATRA)–induced differentiation.³ PHF8 is a key molecule for ATRA treatment response in acute promyelocytic leukemia (APL).⁵ Although ATRA has not been effective in non-APL AML, treatment combining ATRA and the LSD1 inhibitor tranylcypromine (TCP) induced differentiation of non-APL AML and reduced the engraftment of AML in a xenograft mouse model.⁶ This suggests that combinatorial treatment may have activity in treatment of other forms of AML.

In this study, Felipe Fumero et al identified specific phosphorylated residues of endogenous PHF8 by ATRA using a deep phosphorylated proteome analysis of AML on ATRA treatment. PHF8 exhibited antileukemic activity through both its demethylase activity and phosphorylation status. Using RNA sequencing, chromatin immunoprecipitation sequencing, and phosphorylated proteome analysis, the authors showed that phosphorylated PHF8 directly activates cytosolic RNA sensors, in particular, the TRIM25–RIG-I–IFIT5 axis. This, in turn, triggers an interferon (IFN)-I response that leads to differentiation and apoptosis. By analyzing >200 primary AML patient bone marrow samples, they found that half of AML samples had high PHF8 levels with IFN signaling signature at the protein level. This had no association with the specific AML subtype. Importantly, this high PHF8/IFN signaling signature was unique to AML patient samples and not found in healthy controls, making PHF8 an attractive target for AML therapy.

Considering that dimethylation of histone H3 at lysine 9 (H3K9me2) suppresses the expression of IFN and IFN-stimulated genes,⁷ the fact that a KDM family member, such as PHF8, can regulate IFN response is not surprising. Indeed, it has previously been reported that Kdm4d activates type I IFN response in mouse embryonic fibroblasts.⁸ However, in this current study, the authors showed that ATRA-induced phosphorylation of PHF8 upregulates the entire machinery involved in IFN-induced differentiation and apoptosis at the protein level. Moreover, they were able to find that proteins involved in differentiation and apoptosis network were phosphorylated at specific sites in AML cell lines, indicating that those proteins are functional. Although the phosphorylated proteomic data do not show the entire cascade sequence of IFN-I response-driven differentiation and apoptosis, combined ATRA and protein phosphatase inhibitor okadaic acid (OKA) treatment decreased clonogenic growth of primary AML patient samples, suggesting clinical application with potential “mode-of-action” information. Furthermore, this study provides potentially new ATRA-based therapy strategies in AML, which was previously limited to APL, through preventing PHF8 dephosphorylation by OKA, which has been shown to inhibit PHF8 dephosphorylation.⁵ Proteome analysis of >200 primary AML bone marrow samples corroborates the broad therapeutic applicability of targeting PHF8 with ATRA/OKA treatment.

To further extend the utility of ATRA/OKA treatment in AML therapy, the next question is elucidating the molecular basis for high PHF8 expression in certain patients with AML. Given that high PHF8/IFN-I signature is absent in healthy donor CD34⁺ cells, can we specifically increase PHF8 levels in AML cells in patients? It will also be important to test whether maintaining phosphorylation of PHF8 through ATRA/OKA in normal tissues has any adverse effects in vivo, as this can influence its potential use in the clinic. Another important point to consider is the specificity of OKA in vivo. It will be worth noting as well that PHF8 functions in a context-dependent manner in activating immune responses, and that PHF8 can function in a demethylase-independent manner.^9,10 It will be interesting to see whether targeting PHF8 can be more broadly used in other cancers (eg, solid tumors or other hematological malignancies), as shown in colorectal cancer in a mouse model.⁹ Altogether, this study opens up the possibility of targeting PHF8 as a new immunotherapy modality.

Conflict-of-interest disclosure: Y.K. declares no competing financial interests.