Histone deacetylase inhibitors modulate hormesis in leukemic cells with mutant FMS-like tyrosine kinase-3

TO THE EDITOR: The FMS-like tyrosine kinase-3 (FLT3) is mutated in ~30% of acute myeloid leukemia (AML) patients. Common FLT3 mutations are internal tandem duplications (FLT3-ITD) and point mutations in its c-terminal tyrosine kinase domain (FLT3-TKD). The resulting active FLT3 receptors promote cell proliferation and resistance to programmed cell death (apoptosis) [1]. The poor prognosis of patients with FLT3-ITD has spurred an intensive search for FLT3 inhibitors (FLT3i) [2]. These include the nanomolar FLT3i quizartinib, which has produced promising bene ﬁ ts in AML patients with FLT3-ITD [3]. TKD mutations in FLT3-ITD arise during the therapy of patients with quizartinib and confer drug resistance. The broad-range tyrosine kinase inhibitors midostaurin and gilteritinib are used with standard chemotherapy to treat AML. Their pleiotropic actions frequently evoke hematological toxicity [1, 2]. Marbotinib is a hybrid FLT3i containing elements of quizartinib and bis(1H-indol-2-yl)methanone. Marbotinib speci ﬁ - cally inhibits FLT3-ITD and FLT3-TKD [4, 5]. Epigenetic modi ﬁ ers

Epigenetic modifiers of the histone deacetylase (HDAC) family control the development and survival of blood cells.Compared to normal cells, certain leukemic cell types frequently have aberrant expression levels and activities of epigenetic modifiers that belong to the histone deacetylase (HDAC) family.Accordingly, inhibitors of zinc-dependent HDACs (HDACi) are approved drugs for the treatment of subtypes of leukemia and lymphoma.These agents are vorinostat, romidepsin, belinostat, and chidamide.Further HDACi are tested clinically (www.clinicaltrials.org),such as the benzamide entinostat (MS-275) [6,7].HDACi decrease FLT3-ITD through ubiquitin-dependent proteasomal degradation and apoptosis-associated caspase activation.Combinations of HDACi and FLT3i synergistically kill FLT3-ITD-positive cells through accelerated elimination of FLT3-ITD and DNA replication stress/ DNA damage induction.HDAC1, HDAC2, and particularly HDAC3 maintain the stability of FLT3-ITD [8].A cellular self-digestion process termed autophagy also eliminates FLT3-ITD [9].Nonetheless, autophagy can promote FLT3i resistance [10,11].HDACs modulate autophagy in leukemic cells [6,12], but it is unknown if HDACi modulate FLT3-ITD through autophagy and if this has biological relevance.
Despite initial hope in HDACi, their limited clinical efficacy as monotherapies and toxicological issues make additional research necessary.This includes the development of nanomolar HDACi with a low impact on normal cells and the identification of effective drug combinations [7,13].KH16 and KH29 are recently developed HDACi, which we have named yanostat-1/yanostat-2 [12,14].We reveal that their inhibitory profiles are superior to those of SAHA and MS-275 in AML cells carrying FLT3-ITD.Low doses of HDACi cause hormesis effects through FLT3-ITD.Specific inhibition of FLT3-ITD with the nanomolar FLT3i marbotinib and quizartinib abrogates undesired hormesis and is synergistically lethal in combination with nanomolar doses of KH16.
Human embryonic kidney and retinal pigment epithelial (RPE1) cells do not undergo apoptosis when they are exposed to KH16 and KH29 [12,14].We extended these analyses.We incubated MV4-11 cells, RPE1 cells, and human keratinocytes (HaCaT cells) with rounded IC 50 concentrations of KH16, KH29, and SAHA for 24-72 h.Flow cytometry showed that the HDACi induced apoptosis significantly in MV4-11 cells but not in RPE1 and HaCaT cells (Fig. S2).To additionally consider the responses of normal human blood cells to KH16 and KH29, we isolated peripheral blood mononuclear cells (PBMCs) from healthy donors and treated the cells with rounded IC 50 concentrations of KH16, KH29, MS-275, and SAHA for 24 h.None of these HDACi compromised the viability of PBMCs (Fig. 1E).We further distinguished the main leukocyte populations by employing antibodies for lineage-specific surface markers.We found that HDACi did not induce apoptosis in B-cells, T-cells, monocytes, dendritic cells, natural killer (NK) cells, and polymorphonuclear neutrophils (PMNs) (Fig. 1F).Thus, despite being potent inducers of apoptosis in leukemic cells with FLT3-ITD, KH16 and KH29 do not kill normal cells.Immunoblot analyses with lysates from MV4-11 and MOLM-13 cells that were incubated with equimolar concentrations of KH16, KH29, and MS-275 verified that 25 nM of KH16 and KH29 induced an accumulation of hyperacetylated histone H3 (Fig. 1G, H).This effect increased dose-dependently.Up to 100 nM MS-275 did not induce an accumulation of acetylated histone H3.
We confirmed that KH16 and KH29 induced apoptosis by immunoblots assessing activation of caspase-3 and cleavage of its target PARP1.Moreover, we detected a dose-dependent induction of pro-apoptotic BIM and a reduction of anti-apoptotic BCL-XL.KH16 and KH29 caused these molecular alterations more effectively than MS-275.Whereas 25 nM KH16 and KH29 induced the autophagy protein p62, 100 nM KH16 and KH29 decreased p62.25-100 nM MS-275 stabilized p62 (Fig. 1G, H).
Unexpectedly, 25-50 nM KH16 and KH29, and 25-100 nM MS-275, evoked an accumulation of total and phosphorylated FLT3-ITD (Fig. 1G, H).At 100 nM concentrations of KH16 and KH29, this effect was lost.This held for the phosphorylation of the kinases ERK1/ERK2, the transcription factor STAT5, and the levels of the DNA repair protein RAD51.Doses of 100 nM KH16 and KH29 attenuated these tumor-promoting targets of FLT3-ITD (Fig. 1G, H).
We aimed to define the biological importance of the accumulation of p-FLT3-ITD and FLT3-ITD in response to low doses of KH16 (Fig. 1G, H).We hypothesized that this increase in FLT3-ITD attenuated pro-apoptotic effects of HDACi.We experimentally addressed this with a nanomolar dose of marbotinib.
We found that 25 nM KH16 and 2 nM marbotinib combined favorably against MV4-11 cells (Fig. 2A).We also noted this effect using 25 nM KH16 and 2 nM quizartinib.Compared to KH16 plus FLT3i, combinations of 25 nM SAHA and the FLT3i were significantly less effective apoptosis inducers.In such combinations, increasing the dose of SAHA to 650 nM gave similar results as 25 nM KH16 (Fig. 2B).According to these results, combinations of KH16 plus marbotinib or quizartinib are low nanomolar inducers of apoptosis in FLT3-ITD-positive leukemic cell populations.
We analyzed how the combination of 25 nM KH16 and 2 nM marbotinib affected PBMCs and subpopulations therein.We noted that these drugs alone and in combination did not significantly induce apoptosis of B-cells, T-cells, monocytes, dendritic cells, NK cells, and PMNs (Fig. 2E).These data suggest that these inhibitors selectively kill leukemic cells carrying mutant FLT3 and spare normal blood cells.
In conclusion, the epigenetic drugs KH16 and KH29 are innovative lead compounds that deserve further consideration as treatment options for leukemia.One needs to consider the threshold concentrations at which HDACi evoke caspasedependent apoptosis and abrogate cytoprotective autophagy.This finding stresses the need for HDACi that are effective at low nanomolar concentrations.Specific FLT3 kinase inhibitors disable undesired hormesis effects that HDACi cause.This allows such drug combinations to favorably combine against leukemic cells that carry the clinically unfavorable marker FLT3-ITD.

Table 1 .
KH16 and KH29 are more potent inducers of apoptosis than MS-275 and SAHA in FLT3-ITD-positive leukemic cells.