Identification of novel quinazoline derivatives as potent antiplasmodial agents

Article abstract of DOI:10.1016/j.ejmech.2018.10.041

Despite the recent reductions in the global burden of malaria, this disease remains a devastating cause of death in tropical and subtropical regions. As there is no broadly effective vaccine for malaria, prevention and treatment still rely on chemotherapy. Unfortunately, emerging resistance to the gold standard artemisinin combination therapies means that new drugs with novel modes of action are urgently needed. In this context, Plasmodium histone modifying enzymes have emerged as potential drug targets, prompting us to develop and optimize compounds directed against such epigenetic targets. A panel of 51 compounds designed to target different epigenetic enzymes were screened for activity against Plasmodium falciparum parasites. Based on in vitro activity against drug susceptible and drug-resistant P. falciparum lines, selectivity index criterion and favorable pharmacokinetic properties, four compounds, one HDAC inhibitor (1) and three DNMT inhibitors (37, 43 and 45), were selected for preclinical studies in a mouse model of malaria. In vivo data showed that 37, 43 and 45 exhibited oral efficacy in the mouse model of Plasmodium berghei infection. These compounds represent promising starting points for the development of novel antimalarial drugs.

Full text of DOI:10.1016/j.ejmech.2018.10.041

Accepted Manuscript
Identification of novel quinazoline derivatives as potent antiplasmodial agents
Anne Bouchut, Dante Rotili, Christine Pierrot, Sergio Valente, Sophia Lafitte, Johan
Schultz, Urban Hoglund, Roberta Mazzone, Alessia Lucidi, Giancarlo Fabrizi, Dany
Pechalrieu, Paola B. Arimondo, Tina S. Skinner-Adams, Ming Jang Chua, Kathy T.
Andrews, Antonello Mai, Jamal Khalife
PII:
S0223-5234(18)30910-3
10.1016/j.ejmech.2018.10.041
EJMECH 10827
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 16 August 2018
Revised Date: 15 October 2018
Accepted Date: 16 October 2018
Please cite this article as: A. Bouchut, D. Rotili, C. Pierrot, S. Valente, S. Lafitte, J. Schultz, U.
Hoglund, R. Mazzone, A. Lucidi, G. Fabrizi, D. Pechalrieu, P.B. Arimondo, T.S. Skinner-Adams,
M.J. Chua, K.T. Andrews, A. Mai, J. Khalife, Identification of novel quinazoline derivatives as potent
antiplasmodial agents, European Journal of Medicinal Chemistry (2018), doi: https://doi.org/10.1016/
j.ejmech.2018.10.041.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract

ACCEPTED MANUSCRIPT
Identification of Novel Quinazoline Derivatives as Potent Antiplasmodial
Agents
Anne Bouchuta¶, Dante Rotilib¶, Christine Pierrot a, Sergio Valenteb, Sophia Lafittea, Johan
Schultz c, Urban Hoglundd, Roberta Mazzoneb, Alessia Lucidi b, Giancarlo Fabrizib, Dany
Pechalrieue, Paola B. Arimondo e, Tina S. Skinner-Adams f, Ming Jang Chua f, Kathy T.
Andrews f, Antonello Mai b,g,*, Jamal Khalife a,*
aCIIL U1019-CNRS UMR 8204, Université de Lille, Institut Pasteur de Lille, 1 rue du Professeur
Calmette, 59019 Lille, France
bDepartment of Drug Chemistry and Technologies, Sapienza University of Rome, 00185 Rome,
Italy
cKancera AB, Banvaktsvägen 22, SE-17148, Solna, Sweden
dAdlego Biomedical AB, P.O. Box 42, SE-751 03 Uppsala, Sweden
eCNRS FRE3600, ETaC, IBCG, 31062 Toulouse cedex 09, France
fTropical Parasitology Lab, Griffith Institute for Drug Discovery, Don Young Road, Griffith
University, Nathan, Queensland, Australia 4111
gPasteur Institute, Cenci-Bolognetti Foundation, Sapienza University of Rome, 00185 Rome,
Italy.
Corresponding Author
J.K.: phone, +33 320877968; fax, +33 320877888; E-mail, jamal.khalife@pasteur-lille.fr
A.M.: phone, +39 0649913392; fax, +39 0649693268; E-mail, antonello.mai@uniroma1.it
1

ACCEPTED MANUSCRIPT
Author Contributions
¶A.B. and D.R. contributed equally to this work.
Abbreviations
DNMT, DNA methyltransferase; DOT1L, disruptor of telomeric silencing 1-like; EZH2,
enhancer of zeste homolog 2; HAT, histone acetyltransferase, HMEs, histone modifying
enzymes; HMT, histone methyltransferase; HPβCD, (2-hydroxypropyl)-beta-cyclodextrin; Jmj-
C, Jumonji-containing demethylases; KDMs, histone demethylases; LSD1, lysine-specific
histone demethylase 1A; PARP, poly(ADP-ribose) polymerase; PRMT, protein arginine
methyltransferase; WHO, world health organization.
2

ACCEPTED MANUSCRIPT
Abstract
Despite the recent reductions in the global burden of malaria, this disease remains a devastating
cause of death in tropical and subtropical regions. As there is no broadly effective vaccine for
malaria, prevention and treatment still rely on chemotherapy. Unfortunately, emerging resistance
to the gold standard artemisinin combination therapies means that new drugs with novel modes of
action are urgently needed. In this context, Plasmodium histone modifying enzymes have
emerged as potential drug targets, prompting us to develop and optimize compounds directed
against such epigenetic targets. A panel of 51 compounds designed to target different epigenetic
enzymes were screened for activity against Plasmodium falciparum parasites. Based on in vitro
activity against drug susceptible and drug-resistant P. falciparum lines, selectivity index criterion
and favorable pharmacokinetic properties, four compounds, one HDAC inhibitor (1) and three
DNMT inhibitors (37, 43 and 45), were selected for preclinical studies in a mouse model of
malaria. In vivo data showed that 37, 43 and 45 exhibited oral efficacy in the mouse model of
Plasmodium berghei infection. These compounds represent promising starting points for the
development of novel antimalarial drugs.
Keywords: antimalarial agents; histone deacetylase inhibitors; DNA methyltransferase
inhibitors; PK studies; P. berghei mouse model.
3

ACCEPTED MANUSCRIPT
1. Introduction
Plasmodium falciparum is the most virulent of the malaria parasites that infect humans and is
responsible for most of the malaria-related deaths. While the global burden of malaria decreased,
according to the latest WHO estimates released, there were still 212 million cases of malaria in
2015 and 429,000 deaths [1]. Although substantial and tangible improvements are being made in
the search of an effective vaccine [2-5], antimalarial drugs remain one of the main strategies to
prevent and treat malaria. Of concern in this regard is that resistance and/or delayed parasite
clearance to most available drugs, including artemisinin-based combination therapies, have
appeared [6,7]. Consequently, it is imperative to discover new antimalarial agents that are able to
prevent or treat malaria via different mechanisms. As most epigenetic enzymes play essential
roles in proliferation and differentiation of eukaryotic cells, Plasmodium histone modifying
enzymes (HMEs) have been proposed as potential drug targets [8-14].
In the P. falciparum genome, five histone deacetylase enzymes (HDACs), 10 histone
acetylation enzymes (HATs), 13 potential histone methylation enzymes (histone
methyltransferases, HMTs, including 10 lysine methyltransferases and three arginine
methyltransferases) and three histone demethylase enzymes (KDMs), have been identified
(reviewed in [15]). Half of the lysine methyltransferases and one HDAC (PfHDAC3/PfHda2)
have been shown to be refractory to genetic disruption, suggesting an essential and important role
in Plasmodium biology [16,17]. Investigation of HMEs as promising drug targets in Plasmodium
showed that apicidin, curcumin and various hydroxamate derivatives were potent inhibitors of
parasite growth in vitro [13,16,18-22]. Further work identified the small molecule BIX01294
(18), an inhibitor of the H3K9 methyltransferases G9a/GLP [23], as a potent antiplasmodial
against blood parasites in vitro and in vivo [24,25]. The clinically approved HDAC-targeting
4

ACCEPTED MANUSCRIPT
cancer drugs vorinostat, romidepsin, belinostat and panobinostat [26] have also been shown to
cause hyperacetylation of parasite histones and to inhibit the growth of multiple P. falciparum
lines in vitro. Epigenetic control also includes DNA methylation that has been extensively
studied as a stable epigenetic marker [27]. In mammalian cells, the role of C5 DNA
methyltransferases (DNMTs) is essential for cell development and survival [28-33]. Hence
DNMTs have been investigated as cancer drug targets and two inhibitors, azacitidine and
decitabine, have been approved by the Food and Drug Administration and the European
Medicines Agency for use against myelodysplastic syndrome, acute myeloid leukemia and
chronic myelomonocytic leukemia [34,35]. In P. falciparum, a functional DNMT
(PF3D7_0727300) has been partially characterized, with peak of expression in early/late
trophozoites (PlasmoDB [36]). More recent genome-wide mapping showed the methylation of P.
falciparum DNA, the hypomethylation of core promoters and a sharp methylation at exon-intron
boundaries and nucleosomes [37].
The above observations underscore the importance of histone and DNA modifying
enzymes to Plasmodium parasite survival and support the concept that these enzymes could
represent drug targets worthy of further exploration. In this study, compounds initially designed
to target different epigenetic enzyme families in eukaryotes, such as histone deacetylases
(HDACs), histone acetyltransferases (HATs), histone methyltransferases (HMTs), histone
demethylases (KDMs), DNA methyltransferases (DNMTs) and poly(ADP-ribose) polymerases
(PARPs), were assessed for growth inhibitory activity against Plasmodium 3D7 parasites in vitro.
After a first screening of epi-drugs, a focused screening has been performed with analogues of
the first hit compounds. The best-scoring compounds from the two screenings were then assayed
against the multi-drug resistant P. falciparum lines W2 and Dd2 to determine their potency, and
against mammalian cells to assess their cytotoxicity and selectivity indexes. From these results,
5

ACCEPTED MANUSCRIPT
four compounds were selected to perform pharmacokinetic (PK) analysis and in vivo studies on
P. berghei-infected mice.
2. Results
2.1. Chemistry
Among the epi-drugs used for the first screening (Table 1) only 6, a N-hydroxy-3-pyridin-2-
ylacrylamide designed as a HDAC inhibitor, was not already described. For the synthesis of 6, 2-
bromopyridin-5-amine was converted into the corresponding amide intermediate 52 through
reaction with 1-naphthylacetyl chloride in dry dichloromethane in the presence of triethylamine.
Afterwards, 52 was dissolved in dry N,N-dimethylformamide and treated with tetra-n-
butylammonium iodide, sodium acetate trihydrate, butyl acrylate and, under nitrogen atmosphere,
triphenylphosphine and palladium acetate (Scheme 1A). The resulting mixture was heated in a
sealed tube at 140 °C providing the butyl acrylate 53, which was converted into the
corresponding carboxylic acid 54 through basic hydrolysis (lithium hydroxide in
tetrahydrofuran/water) and finally into the hydroxamate 6 by i) activation with ethyl
chloroformate and triethylamine in dry tetrahydrofuran at 0 °C, ii) nucleophilic displacement
with O-(2-methoxy-2-propyl)hydroxylamine in dry tetrahydrofuran at 0 °C, and iii) cleavage with
Amberlyst 15 ion-exchange resin in methanol (Scheme 1A).
Three analogues of 1 (30-32) and two analogues of 6 (34, 35) were synthesized and tested in
the focused screening (Table 2), to improve their antimalarial activity. The synthetic route
followed for the preparation of compounds 30-32 is depicted in Scheme 1B. The known 2-
mercapto-6-substituted-pyrimidin-4(3H)-ones 55 [38] and 56 [39] were treated with the
appropriate commercially available ethyl ω-bromoalkanoate in presence of anhydrous potassium
6

ACCEPTED MANUSCRIPT
carbonate to afford the ethyl esters 57-59, which were then hydrolyzed to the related carboxylic
acids 60-62 by the means of potassium hydroxide in ethanol at room temperature (rt), and then
converted into the hydroxamates 30-32 (Scheme 1B) following the procedure used for the
synthesis of 6.
The syntheses of 34 and 35 were accomplished starting from methyl 5-aminopicolinate
acylated with the appropriate acyl chlorides to give the intermediate amides 63 and 64. After
hydrolysis of 63 and 64 under basic condition to the corresponding carboxylic acids 65 and 66,
the compounds were converted into the hydroxamates 34 and 35 by the usual way (Scheme 1C).
Scheme 1. Synthesis of the HDACi 6, 30-32, 34, 35. Reagents and conditions: (a) 1-
naphthylacetyl chloride, dry DCM, triethylamine, rt, 2 h; (b) triphenylphosphine, palladium
acetate, tetra-n-butylammonium iodide, sodium acetate trihydrate, butyl acrylate, dry DMF, 140
°C, sealed tube, overnight; (c) lithium hydroxide, ethanol, rt, overnight; (d) 1) ethyl
chloroformate, triethylamine, dry THF, 0 °C to rt, 10 min; 2) NH2OC(CH3)2OCH3, dry THF, 0
°C to rt, 15 min to 3 h; 3) Amberlist 15, methanol, rt, 1-2 h; (e) appropriate ethyl ω-
7

ACCEPTED MANUSCRIPT
bromoalkanoate, anhydrous potassium carbonate, dry DMF, rt, 1 h; (f) 2N potassium hydroxide,
ethanol, rt, 18 h; (g) appropriate acyl chloride, dry DCM, triethylamine, rt, 2 h.
Since the quinazoline 28 [40] was identified in the first screen as a hit compound against
P. falciparum 3D7, some analogues (36-48) were synthesized and tested as potential DNMT
inhibitors and antimalarials (see Table 4 for structures). Scheme 2 shows the general synthetic
pathway followed for the preparation of the final derivatives 36-45. The known 2,4-
dichloroquinazolines 67 [41], 68 and 69 [42] were subjected to a displacement reaction at the C4
position with the appropriate commercial amines, generating the 2-chloro-4-aminosubstituted
quinazoline intermediates 70 and 71 [43], 72-74, 75 [44], 76 and 77 [45]. The 2-
chloroquinazoline 70, dissolved in a mixture of dry THF and dry methanol, was treated with 4N
hydrochloric acid in dioxane providing the 2-chloro-N-(piperidin-4-yl)quinazolin-4-amine
hydrochloride 78, which underwent an alkylation reaction with commercial 4-methoxyphenethyl
bromide in the presence of potassium carbonate and sodium iodide in dry DMF at rt to give the
intermediate 79. Further C2-chloro displacement on 79 with 4-fluorobenzylamine in a sealed tube
at 110 °C afforded the final compound 36 (Scheme 2A). C2-Chloro displacement performed on
the 2-chloro-6,7-dimethoxyquinazoline 71 with 4-fluorobenzylamine in iso-propyl alcohol at 130
°C under microwave irradiation afforded 37. Compound 37 was then converted into 80 via BOC-
deprotection performed with 4N hydrochloric acid in dioxane and transformed into 38 by
alkylation with 4-methoxyphenethyl bromide in the presence of potassium carbonate and sodium
iodide in dry DMF at rt (Scheme 2A). The final compounds 39-44 were obtained by reacting the
4-amino-2-chloroquinazolines 72-77 with N-phenylpiperazine in a sealed tube at 110 °C (Scheme
2A). The 7-hydroxyquinazoline 45 was prepared through debenzylation reaction of 43 with
trifluoroacetic acid under reflux conditions (Scheme 2A).
8

ACCEPTED MANUSCRIPT
Final compounds 46-48 were prepared as reported in Scheme 2B. The 2-phenylpiperazine-6,7-
substituted-quinazolinones 82, 83 and 84 [46] were synthesized via cyclization reaction between
the appropriate commercial ethyl/methyl anthranilates and the N-phenylpiperazine-1-carbonitrile
81 [45,46], in the presence of dry sodium hydride in xylene at 140 °C under nitrogen atmosphere.
The reaction between 82-84 and 1,2,4-triazole in the presence of phosphorous oxychloride and
triethylamine produced the triazolyl-quinazolines 85-87 [46] which reacted with 4-amino-1-
benzylpiperidine in dry dioxane at 110 °C to provide the final compounds 46-48.
9

ACCEPTED MANUSCRIPT
Scheme 2. Synthesis of the new quinazoline derivatives 36-48. Reagents and conditions: (a) tert-
butyl 4-aminopiperidine-1-carboxylate, triethylamine, dry THF, 0 °C to rt, 4-72 h; (b) 4N
hydrochloric acid in dioxane, dry THF/dry methanol 1:1, 0 °C to rt, 7-25 h; (c) 4-
methoxyphenetyl bromide, dry potassium carbonate, sodium iodide, dry DMF, rt, 19-27 h; (d) 4-
fluorobenzylamine, iso-amyl alcohol, 110 °C, sealed tube, 8 h; (e) 4-fluorobenzylamine, iso-
propyl alcohol, microwave, 130 °C, 5 h; (f) appropriate amine, triethylamine, dry THF, 0 °C to rt,
4-6.5 h; (g) N-phenylpiperazine, iso-amyl alcohol, 110 °C, sealed tube, 4 h; (h) trifluoroacetic
acid, 0 °C to 115 °C, 35 min; (i) dry sodium hydride, xylene, N2, 140° C, 3-4 h; (j) (1)
phosphorus oxychloride, triethylamine, 1,2,4-triazole, dry acetonitrile, 0° C (40 min) then rt (30
min); (2) addition of a solution of quinazolinone in dry acetonitrile (or dry chloroform), N2, rt,
overnight then reflux, 2-5 h; (k) 4-amino-1-benzylpiperidine, N,N-diiso-propylethylamine, dry
dioxane,110° C, 20-24 h.
General procedures for the syntheses of the HDACi 6, 30-32, 34, 35 and of the DNMT3a
inhibitors 36-48, as well as their chemical, physical and spectral (1H- and 13C-NMR) data are
reported in Supplementary material. Chemical, physical and spectral (1H-NMR) data of the
intermediate unknown (i.e. previously not characterized) compounds 52-54, 57-66, 72-74, 78-80,
10

ACCEPTED MANUSCRIPT
82, 83, 85, and 86 are reported in Table S1 in Supplementary material. Elemental analyses for
compounds 6, 30-32, 34, 35, 36-48 are reported in Table S2 in Supplementary material.
2.2. Biochemistry
The N-hydroxy-3-(5-(2-(naphthalen-1-yl)acetamido)pyridin-2-yl)acrylamide 6 as well as the
analogues of 1, 30-32, and the analogues of 6, 34 and 35, were tested against human HDAC1,
HDAC4 and HDAC6 (representative of class I, class IIa, and class IIb HDACs, respectively) to
determine their potency and selectivity. Trichostatin A (TSA) and TMP269 [47] were used as
reference drugs.
Data in Table S3 (Supplementary material) showed that all the new synthesized HDACi
displayed submicromolar inhibition against HDAC1 and nanomolar inhibition against HDAC6,
while they were much less potent (if at all) against HDAC4.
The new quinazoline derivatives 36-48, analogues of the DNMT3a-selective 28 [40], were
tested against active human recombinant DNMT1, DNMT3a and G9a to determine their half
maximal effective concentrations (EC50). Data presented in Table S4 (Supplementary material)
showed that the tested compounds 36-43, 45-47 were more potent (EC50 between 1.6 and 8.7
µM) against DNMT3a than against the other tested methyltransferases, with the different
chemical substitutions at the C2, C4, C6, and C7 quinazoline positions having only modulatory
effects on the potency of derivatives. Compounds 44 and 48 were less potent against DNMT3a
(% inhibition at 10 µM <50%), thus they were not tested against DNMT1 and G9a.
2.3. Phenotypic screening of epigenetic modulators against asexual intraerythrocytic stage of P.
falciparum 3D7 parasites
11

ACCEPTED MANUSCRIPT
The antiplasmodial activities of the first 29 compounds targeting different classes of
epigenetic targets were screened for activity against drug-sensitive P. falciparum 3D7 parasites at
the single concentration of 10 µM. As shown in Table 1, 22 compounds inhibited parasite growth
by at least 50% and seven compounds inhibited parasite growth by less than 25%. The 22
compounds giving >50% inhibition were then tested in dose response assays to determine their
IC50 values (50% or half maximal inhibitory concentration). The most potent compounds (IC50
<50 nM) were two hydroxamate-based HDAC inhibitors, 1 (IC50 = 4.0 nM) and 6 (IC50 = 10.2
nM), both with selectivity for class I/IIb human HDACs, and the G9a/GPL inhibitor 18 (IC50 =
20.8 nM), for which similar activity has been previously reported [24]. Submicromolar inhibition
of Plasmodium growth was also displayed by 19 [48] (UNC0638, IC50 = 67.9 nM), a G9a/GLP
inhibitor related to 18, and by the recently reported DNMT3a inhibitor 28 [40] (IC50 = 325.5 nM)
(Table 1).
Table 1
In vitro antiplasmodial activity of compounds against P. falciparum 3D7-infected erythrocytes.
% inhibition
Epi-targeta
IC50 (nM)b
Compound
at 10 µM
HDAC
97.0
4.0 ± 0.3
1 [49]
class I/IIb HDACi
97.1
>1 µM
2 [50]
context-selective class I
12

ACCEPTED MANUSCRIPT
HDACi
50.9
>1 µM
>1 µM
3 [51]
HDAC1-3 inhibitor
89.6
100
4 [52]
HDAC4-6,8 inhibitor
5
422.4 ± 19.4
[53]
HDAC6-selective inhibitor
96.8
98.0
10.2 ± 2.6
6
class I/IIb HDACi
244.0 ± 52.2
7 [54]
pan-HDACi
2.8
NDc
8 [55]
HDAC8 inhibitor
Sirtuin
80.5
>1 µM
9 [56, 57]
Sirt1/2i
13

ACCEPTED MANUSCRIPT
72.3
>1 µM
10 [58, 59]
Sirt1/2i
18.7
70.7
ND
11 [60]
Sirt2-selective inhibitor
>1 µM
12 [61]
Sirt6 activator
99.5
97.2
671.2 ± 15.5
13 [61]
Sirt6 activator
>1 µM
14 [62, 63]
Sirt1 activator
HAT
12.8
ND
15 [64, 65]
p300 inhibitor
14

ACCEPTED MANUSCRIPT
92.7
>1 µM
>1 µM
16 [66, 67]
p300 inhibitor/PCAF activator
60.1
100
17 [68]
p300/CBP inhibitor
HMT
20.8 ± 4.9
18 [23-25]
G9a/GLP inhibitor
100
67.9 ± 0.1
19 [48]
G9a/GLP inhibitor
O
HN
HN
O
N
N
14.7
ND
ND
20 [69]
EZH2 inhibitor
9.7
21 [70]
DOT1L inhibitor
15

ACCEPTED MANUSCRIPT
100
> 1µM
22 [71, 72]
PRMT4/EZH2/HAT inhibitor
(multi-target)
99.5
97.5
951.9 ± 86.1
23 [71, 72]
PRMT1 inhibitor
KDM
>1 µM
24 [73]
LSD1 inhibitor
64.6
25.4
>1 µM
25 [73]
LSD1 inhibitor
ND
26 [74]
Jmj-C inhibitor
100
>1 µM
27 [75]
pan (LSD1+Jmj-C) KDM
inhibitor
16

ACCEPTED MANUSCRIPT
DNMT
PARP
100
325.5 ± 22.1
28 [40]
DNMT3a inhibitor
2
ND
29 [76]
PARP1 inhibitor
b
aEpi-target, epigenetic modulator target as identified for human cells/targets. IC50 values were
c
not determined when the percentage of inhibition at 10 µM was lower than 50%. ND, Not
Determined.
2.4. Antiplasmodial activity of analogues derived from active compounds
In order to optimize the potency of active compounds, 22 further analogues of previous hit
compounds were tested. The four analogues of 1 had different C6-uracil substitutions [2-naphthyl
(31, 32) or 3-chlorophenyl (33) instead of the 4-biphenyl group], and/or different lengths of
spacers between the sulphur atom and the hydroxamide [five (30, 32) instead of four methylene
groups]. The two analogues of 6 had their N-hydroxyacrylamide group replaced with a N-
hydroxycarboxamide group. Compound 34 also had its 1-naphthyl moiety substituted with a
phenyl ring. All 1 and 6 analogues showed selectivity for human class I/IIb HDACs analogously
to their prototypes (Table S3). In addition, 13 DNMT inhibitors analogues of 28 were prepared
and tested, with changes at quinazoline C2 (p-fluorobenzylamine or phenylpiperazine), C4 (N-
arylalkylsubstituted 4-aminopiperazine, N-tert-butoxycarbonyl 4-aminopiperazine, 2-(1-
naphthylamino)ethylamine), C6 (H, methoxy group), and/or C7 (H, methoxy, benzyloxy group)
17

ACCEPTED MANUSCRIPT
position. Finally, as 27 gave 100% inhibition at 10 µM in the primary screen (Table 1), three pan-
KDM inhibitors (analogues of 27) were also included. These 22 analogues were first screened at
10 µM against P. falciparum 3D7 and, with the exception of 49, all displayed 96-100% inhibition
of parasite growth. Compound 49, a pan-KDM inhibitor, was excluded from further analysis. For
the remaining 21 analogues, IC50 values were then determined as previously described (Table 2).
Among the HDACi examined, the analogues of 1 (compounds 30-33) exhibited IC50 values <100
nM while the analogues of 6 (compounds 34, 35) were much less potent (IC50 >1 µM).
Nevertheless, none of the compound among the two groups showed IC50 values lower than the
corresponding prototypes (Table 2). Among the tested DNMT inhibitors, all 13 analogues of 28
showed lower IC50 values than the prototype, with 36-38, 43 and 45 having a ~10-32-fold lower
IC50 than 28 (Table 2, Fig. 1). In the case of the analogues of the pan-KDM inhibitor 27, all tested
analogues displayed increased IC50 values (>1µM). Based on data of both screenings, we selected
the two HDACi 1 and 6, the G9a inhibitors 18 and 19, and the DNMT inhibitors 37, 38, 43, 45
for further investigation.
Table 2
In vitro antiplasmodial activity of hits (Pf3D7 IC50 <1 µM) and analogues against P. falciparum
3D7-infected erythrocytes.
%
IC50 (nM)a
Compound
inhibition
at 10 µM
HDACi (analogues of 1)
class I/IIb HDACi
96.1
6.0 ± 2.1
1 [49]
18

ACCEPTED MANUSCRIPT
97.4
50. 9 ± 5.0
30
31
32
97.9
97.4
98.0
25.2 ± 5.0
79.4 ± 0.7
88.6 ± 2.2
33
6
HDACi (analogues of 6)
97.9
97.8
21.1 ± 2.4
>1 µM
class I/IIb HDACi
34
97.9
100
>1 µM
35
28 [40]
36
DNMT3a inhibitors
325.5 ± 22.1
(analogues of 28)
100
33.6 ± 8.7
19

ACCEPTED MANUSCRIPT
100
18.7 ± 0.9
37
38
39
100
100
18.0 ± 5.4
138.2 ± 22.1
100
100
98.9 ± 13.7
40
75.1 ± 6.2
41
42
100
100
37.8 ± 6.3
34.0 ± 19.0
20

ACCEPTED MANUSCRIPT
43
99.6
56.2 ± 0.3
10.1 ± 5.1
52.5 ± 6.7
83.0 ± 16.3
66.9 ± 6.9
44
45
46
47
100
100
100
100
48
pan-KDM inhibitors
98.9
45.3
775.8 ± 293
(analogues of 27)
27 [75]
NDb
49 [75]
21

ACCEPTED MANUSCRIPT
97.8
>1 µM
>1 µM
50 [75]
51 [75]
96.7
aIC50 values were not determined when the percentage of inhibition at 10 µM was lower than
50%. bND, not determined.
Fig. 1. Hit compounds and analogues selected for further investigation. Each histogram shows
one candidate and its analogues (X axis). The parent hit compound of each series is shown in
black, the analogues are in grey. Results are shown as mean ± SD for 2 independent assays.
2.5. Antiplasmodial activity against multi-drug resistant P. falciparum lines W2 and Dd2
Next, compounds were assessed for growth inhibitory activity against the multidrug-
resistant P. falciparum lines W2 and Dd2 (resistant to chloroquine, quinine, pyrimethamine and
sulfadoxine) [78,79] and the activities were compared to those obtained against the drug-sensitive
22

ACCEPTED MANUSCRIPT
3D7 line. Under the experimental conditions of the present study, chloroquine had IC50 values of
21.5 nM (± 1.6) and 290.0 nM (± 6) for 3D7 and W2 respectively, consistent with previous
reports [80-82]. Of the eight compounds spanning HDAC, G9a, and DNMT inhibitor classes,
activity against W2 was found to be similar to that against 3D7, with IC50 values <50 nM for all
tested compounds (Table 3). Interestingly, 19 was more potent against the multidrug-resistant W2
than against the wild-type 3D7 line. Similar IC50 values to 3D7 were obtained for all compounds
when tested against Dd2 using the [3H]-hypoxanthine uptake assay method (Table 3). Together,
these data show that these eight compounds have early lead (IC50 <100 nM) profiles [83] with
respect to their potency, and demonstrate a lack of cross-resistance to multi-drug resistant P.
falciparum lines.
Table 3
Comparative in vitro activity (IC50) of selected compounds against drug-sensitive P. falciparum
line 3D7 versus multidrug-resistant P. falciparum lines Dd2 and W2.
P. falciparum IC50 (nM)
Epi-targeta
Compound
3D7 (n=3)
4.2 ± 0.4
Dd2 (n=3)
4.0 ± 1.1
W2 (n=2)
6.6 ± 0.1
HDAC
HDAC
G9a
1
13.3 ± 5.6
23.7 ± 6.1
68.7 ± 1.4
18.9 ± 0.9
18.0 ± 5.4
34.0 ± 18.6
13.5 ± 6.9
9.7 ± 1.6
15.3 ± 0.8
16.7 ± 8.7
38.5 ± 4.0
43.6 ± 10.4
25.8 ± 5.3
39.9 ± 9.9
6
18
19
37
38
43
G9a
16.8 ± 6.3
18.5 ± 3.8
6.1 ± 3.7
DNMT3a
DNMT3a
DNMT3a
81.3 ± 44.3
23

ACCEPTED MANUSCRIPT
DNMT3a
10.1 ± 5.1
42.6 ± 12.0
18.6 ± 8.4
45
aEpi-target, epigenetic modulator target as identified for human cells/targets.
2.6 Plasmodium specific selectivity versus primary activated cells (splenic murine cells) and a
eukaryotic cell line (HFF)
To determine the selectivity Index (SI) of the eight selected compounds (Table 3),
cytotoxicity was assessed on murine splenic primary cells as well as human foreskin fibroblasts
(HFF) and IC50 values were compared to those obtained for P. falciparum growth inhibition. As
shown in Table 4, all compounds showed SI >100 against the parasites versus HFFs, except for
19 which displayed the lowest selectivity (SI = 33). The SI against primary murine cells was
between 44 and >500, with the exception of the HDACi 6 (SI = 9) and the G9a inhibitors 18 and
19 (SI = 23 and 21, respectively).
Table 4
Compound selectivity for parasite versus mammalian cells.
Mammalian cells
Pf3D7
Epi-targeta
Compound
Splenic cells
HFF
IC50 (nM)
SIb
IC50 (nM)
4.2 ± 0.4
IC50 (nM)
785 ± 78
SI
783
>750
211
33
HDAC
HDAC
G9a
186
9
3,295 ± 247
>10,000c
1
6
13.3 ± 5.6
23.7 ± 6.172
68.7 ± 1.4
120 ± 42
555 ± 177
1,440 ± 141
23
21
4,997 ± 355
2,286 ± 659
18
19
G9a
24

ACCEPTED MANUSCRIPT
DNMT3a
DNMT3a
DNMT3a
DNMT3a
18.9 ± 0.9
18.0 ± 5.4
34.0 ± 18.6
10.1 ± 5.1
>10,000
800 ± 141
>10,000
>500
44
7,170 ± 84
3,215 ± 389
5,629 ± 27
7,210 ± 14
377
178
166
712
37
38
43
45
>295
255
2,585 ± 346
aEpi-target, epigenetic modulator target as identified for human cells/targets. bSI, selectivity
index: mammalian cell IC50/P. falciparum IC50. cIC50 values could not be precisely determined
(>10,000 nM) when the highest concentration used was not toxic for mammalian cells.
2.7. Pharmacokinetic analysis
Four compounds were further selected for PK studies: the HDAC inhibitor 1 and the
DNMT inhibitors 37, 43 and 45. A (2-hydroxypropyl)-β-cyclodextrin (HPβCD)-based
formulation, suitable for both intravenous (iv) and per os (p.o.; oral) administration, was used for
PK studies. For each compound, two mice were administered with a single dose (50 mg/kg for 1,
43 and 45, and 10 mg/kg for 37), either iv or p.o. by gavage. Plasma samples were analysed and
concentrations are summarized in Table 5. Compound 1 administered at a 50 mg/kg dose was
well tolerated. The 15 min iv samples showed high (>10 µM) 1 plasma concentration, but
decreasing rapidly with time. The p.o. and iv plasma samples at 1 h post administration showed
that the concentration of 1 was between 0.5 and 2.2 µM and after 3 h between 43 and 300 nM.
The 50 mg/kg dose of 43 was well tolerated when administered p.o. The same dose administered
iv was fatal for the mice. The two mice administered with 43 p.o. showed plasma concentrations
below 1 µM at all-time points. The 50 mg/kg dose of 45 was well tolerated when administered
p.o. However, the same dose administered iv led to severe acute toxicity and death so that plasma
samples could not be collected. The two mice administered with 45 p.o. showed plasma
25

ACCEPTED MANUSCRIPT
concentrations of 1-5 µM at 30 and 60 min after administration. After 180 min, the plasma
concentration was 0.28-0.5 µM. The poor solubility of 37 only allowed a relatively low
concentration formulation (1 mg/mL) to be prepared. The 10 mg/kg dose of 37 was well tolerated
when administered p.o. The two mice treated with 37 iv showed plasma concentrations of 1-5
µM at the 15 and 60 min, and ~ 0.5 µM after 180 min. The two mice treated with 37 p.o. showed
very low plasma concentrations at all-time points.
Table 5
Pharmacokinetic parameters of the four selected compounds administered intravenously or orally
to mice
Administration route
iv
per os
Compound/Dose
Concentration (µM)
Mouse 1 Mouse 2
Concentration (µM)
Time
(min)
15
Time
(min)
30
Mouse 3
2.43
Mouse 4
0.67
15.2
2.16
0.044
3.41
1.11
0.22
10.7
1.67
0.043
5.11
0.93
0.46
1
60
60
0.90
0.48
50 mg/kg
180
15
180
30
0.30
0.070
0.12
0.13
37
60
60
0.28
0.10
10 mg/kg
180
15
180
30
0.03
0.03
0.38
0.94
43
60
Toxic
60
0.48
0.70
50 mg/kg
180
180
0.39
0.28
26

ACCEPTED MANUSCRIPT
15
60
30
4.57
3.52
0.67
3.43
1.04
0.29
45
Toxic
60
50 mg/kg
180
180
2.8. In vivo antimalarial activity in P. berghei infected mice
For the in vivo experiments, the doses of the compounds were defined on the basis of their high
potency, low toxicity and solubility for oral administration that provide plasma concentrations
comparable to the IC50s obtained in vitro. The oral efficacy of the four selected compounds was
examined using the Peters 4-day suppressive test [84]. Groups of six BALB/c mice were infected
intraperitoneally (ip) with 106 P. berghei ANKA parasites and treated as depicted in Fig. 2A. On
day 4 post infection (pi), the peripheral blood parasitemia of all mice in the treated groups was
compared to mice in vehicle control groups. None of the mice administered with 1 p.o. at 50
mg/kg, showed a significant reduction in parasitemia as compared to control mice (Fig. 2B, Exp.
1, p=0.229). As PK studies revealed that concentrations of 1 are higher in the plasma after 15 and
60 min when administered iv (≥10.7 and ≥1.7 µM, respectively) compared to p.o. (≥0.7 and ≥0.5
µM, respectively, Table 5) the efficacy of 1 was also assessed using a combination of oral and iv
administration. However, this combined administration route (twice a day, one p.o. and one iv),
was not effective on blood stage infection (Fig. 2B; Exp. 2).
Oral efficacy of 37 was evaluated at 20 mg/kg because quality control studies revealed that
37 presented some chemical instability and poor solubility at higher concentrations. At day 4 pi,
parasites were undetectable in the 37-treated mice in two independent experiments whereas the
vehicle control group showed a parasitemia of ~6% and ~4% in experiment 1 and 2, respectively
(Fig. 2C). However, parasites were detected in treated animals from day 6-7 pi (Supplementary
27

ACCEPTED MANUSCRIPT
material, Fig. S1A, Exp 1), suggesting that while the 4-day treatment was able to delay the onset
of patent parasitemia, it was not sufficient to cure mice. This is supported by data showing that an
additional treatment at day 5 pi extended the suppressive effect (Fig. S1A, Exp 2).
When 43 was administered at 50 mg/kg p.o. twice a day for four days, it was too toxic (four
mice died, the two remaining mice showed no parasites at day 4 pi but one more mouse died at
day 10 pi (data not shown)). When 43 was tested at 25 mg/kg p.o. twice a day for four days, no
adverse symptoms were observed and peripheral blood parasitemia was not detectable in the 43-
treated mice in two independent experiments compared to the vehicle-treated mice (parasitemia
of ~6% and ~4% on day 4 pi in experiment 1 and 2, respectively; Fig. 2D). Moreover,
parasitemia remained sub-patent until 14 days pi in experiment 1 and was ~2% in experiment 2
(Fig. S1C). When 43 was evaluated at 10 mg/kg p.o., mice treated twice a day for four days had
detectable parasites on day 4 pi, however parasitemia was significantly lower (p<0.002)
compared to the control group (Fig. S1D).
With respect to 45, its oral administration at 50 mg/kg twice a day for four days reduced
blood parasites at day 4 pi in two independent experiments (Fig. 2E). Indeed, infections were
sub-patent in mice treated with 45 until day 8 pi and parasitemia remained lower than 10% until
the end of each experiment (vehicle control mice reached a parasitemia of ~58% and 31% in
experiment 1 and 2, respectively; Fig. S1B).
28