Histone deacetylase inhibitor AR-42 and achiral analogues kill malaria parasites in vitro and in mice

Abstract

Malaria is caused by infection with Plasmodium parasites and results in significant health and economic impacts. Malaria eradication is hampered by parasite resistance to current drugs and the lack of a widely effective vaccine. Compounds that target epigenetic regulatory proteins, such as histone deacetylases (HDACs), may lead to new therapeutic agents with a different mechanism of action, thereby avoiding resistance mechanisms to current antimalarial drugs. The anticancer HDAC inhibitor AR-42, as its racemate (rac-AR-42), and 36 analogues were investigated for in vitro activity against P. falciparum. Rac-AR-42 and selected compounds were assessed for cytotoxicity against human cells, histone hyperacetylation, human HDAC1 inhibition and oral activity in a murine malaria model. Rac-AR-42 was tested for ex vivo asexual and in vitro exoerythrocytic stage activity against P. berghei murine malaria parasites. Rac-AR-42 and 13 achiral analogues were potent inhibitors of asexual intraerythrocytic stage P. falciparum 3D7 growth in vitro (IC₅₀ 5-50 nM), with four of these compounds having >50-fold selectivity for P. falciparum versus human cells (selectivity index 56-118). Rac-AR-42 induced in situ hyperacetylation of P. falciparum histone H4, consistent with PfHDAC(s) inhibition. Furthermore, rac-AR-42 potently inhibited P. berghei infected erythrocyte growth ex vivo (IC₅₀ 40 nM) and P. berghei exoerythrocytic forms in hepatocytes (IC₅₀ 1 nM). Oral administration of rac-AR-42 and two achiral analogues inhibited P. berghei growth in mice, with rac-AR-42 (50 mg/kg/day single dose for four days) curing all infections. These findings demonstrate curative properties for HDAC inhibitors in the oral treatment of experimental mouse malaria.

Keywords: AR-42; HDAC inhibitor; Malaria; P. falciparum.

Graphical abstract

Fig. 1

Plot of P. falciparum 3D7 (Pf3D7) IC₅₀(Log10 scale) versus calculated cLog P for selected compounds. The compounds are grouped based on chemical structure; cyclopentane compounds 13 and 19–21 (black triangles); N-hydroxy-4-(1-phenylcyclohexane-1-carboxamido)benzamide 14 (black star); tetrahydropyran derivatives 15 and 17 (black squares), piperidine analogues 16 and 18 (black circles); 4-(1-(3-fluoro-4-(trifluoromethyl)phenyl)cyclopentane-1-carboxamido)-N-hydroxybenzamide 22 (open square); and naphthalene compounds 23 and 24 (open circles).

Fig. 2

(A) Homology model of PfHDAC1 overlayed with the human HDAC1 (5ICN.pdb) crystal structure. The backbones of the two structures are shown, with identical residues coloured grey and non-identical hHDAC1 residues in orange. (B) PfHDAC1 homology model showing binding site for the inhibitor SAHA (green sticks). PfHDAC1 residues within 6 Å of the ligand SAHA are almost identical with hHDAC1, except for PfHDAC1 Val175 which is Ile in hHDAC1 (orange sticks and label), neither of these residues contact the ligand or the binding site surface. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 3

PfHDAC1 with Gold docked ligands 19 (JT21b, green) and 20 (JT83, cyan). PfHDAC1 is shown in grey. Binding site residues are shown as sticks (carbon – grey, oxygen – red, nitrogen – blue). Ala95 and Thr96 are highlighted in magenta. H-bonds to nearby residues are shown as yellow dotted lines and interactions with Zn atom are shown as grey dotted lines. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4

Plot of recombinant human HDAC1 pIC₅₀versus P. falciparum 3D7 (Pf3D7) pIC₅₀shows that potent human HDAC1 inhibitors were often effective antiplasmodial agents in vitro. Pearson correlation coefficient, r = 0.5461 and ρ = 0.0005. Outliers include compound 34 (square), that killed malaria parasites effectively (Pf3D7 IC₅₀ 21 nM; Table 1) but had modest inhibitory activity against hHDAC1 (IC₅₀ 774 nM; Table 1) and compounds 23 and 29 (triangles) that were the most potent inhibitors of hHDAC1 (IC₅₀ 0.7 nM and 0.9 nM, respectively; Table 1).

Fig. 5

In vivo activity of orally administered AR-42 and analogues in P. berghei infected mice. P. berghei infected BALB/c mice (two mice/group) were treated orally with 25 mg/kg AR-42 (1), JT21b (19), JT94 (34), JT83 (20) (open circles; dashed lines; A-D, respectively), or vehicle only (black squares; solid lines), twice daily (black arrows) for four days beginning 2 h post infection. Mice treated with 19 and 20 were not administered the final dose due to weight loss. Mean % parasitemia for individual mice is shown.

Fig. 6

In vivo activity of orally administered AR-42 (1) in P. berghei infected mice. In one experiment (A-C), P. berghei infected BALB/c mice (six mice/group) were treated orally with 25 mg/kg vorinostat (A; as a control) or 25 mg/kg rac-AR-42 (1) once daily (B) or twice daily (C) for four days. In a second experiment (D-F), P. berghei infected mice were treated for four days with 10 mg/kg chloroquine once daily (D; as a control) or 50 mg/kg 1 once daily (E) or 25 mg/kg 1 twice daily (F). Treated mice are shown with cross symbol and dashed grey lines. In each case, control mice (A-F; diamonds, black line) received vehicle only. Mean % parasitemia for individual mice is shown.

Fig. 7

Hyperacetylation of P. falciparum histone H4. Protein lysates were prepared from synchronous trophozoite-stage P. falciparum 3D7 parasites treated for 3 h with vehicle control (C-3; 0.1% DMSO), 5x IC₅₀ compound 1, 5x IC₅₀ chloroquine (CQ; negative control) or 5x IC₅₀ vorinostat (HDAC inhibitor control), followed by Western blot analysis. (A) Top panels show membrane probed with anti-(tetra)acetyl-H4 primary antibody (1:2000 dilution) and anti-rabbit IRDye® 680RD secondary antibody (1:10,000 dilution) and imaged using Odyssey® Fc Imaging System (Li-Cor Biosciences). Bottom panels show the same membranes stained with REVERT™ Total Protein Stain prior to Western blot. (B) Bar graph showing mean density (±SD) relative to the C-3 DMSO control (set to 1.0) of two independent assays. Signal corresponding to H4 (~11 kDa) is shown in grey bars and total signal which includes H4 and cross-reactive bands corresponding to H2B/H2Bv (~13–14 kDa) and H2A.Z (~16 kDa) is shown in black bars.