Antiparasitic activity of new gibbilimbol analogues and SAR analysis through efficiency and statistical methods

Article abstract of DOI:10.1016/j.ejps.2018.06.023

Chagas' disease and leishmaniasis are parasitic infections enrolled among the neglected tropical diseases, which urge for new treatments. In the search for new chemical entities as prototypes, gibbilimbols A/B have shown antiparasitic activity against Trypanosoma cruzi and Leishmania infantum, and then a set of analogues (LINS03 series) of this natural product were synthesized and evaluated in vitro against the parasites. In the present paper we reported five new compounds with activity against these protozoan parasites, and quite low cytotoxicity. Moreover, the interference of plasma membrane permeability of these analogues were also evaluated. We found that [(4-methoxyphenyl)methyl]octylamine (4) was noteworthy due to its high activity against the amastigote form of both parasites (IC₅₀ 1.3–5.8 μM) and good selectivity index. In order to unveil the SAR for this chemotype, we also presented a group efficiency analysis and PCA and HCA study, which indicated that the methoxyl provides good activity with lower cytotoxicity to mammalian cells. The results from SAR analyses suggest different mechanisms of action between the neutral and basic compounds. In summary, the analogues represent important activity against these parasites and must be prototypes for further exploitation.

Full text of DOI:10.1016/j.ejps.2018.06.023

Accepted Manuscript
Antiparasitic activity of new gibbilimbol analogues and SAR
analysis through efficiency and statistical methods
Marina T. Varela, Maiara M. Romaneli, Marta L. Lima, Samanta
E.T. Borborema, Andre G. Tempone, João P.S. Fernandes
PII:
DOI:
Reference:
S0928-0987(18)30279-3
doi:10.1016/j.ejps.2018.06.023
PHASCI 4565
To appear in:
European Journal of Pharmaceutical Sciences
Received date:
Revised date:
Accepted date:
27 March 2018
12 June 2018
19 June 2018
Please cite this article as: Marina T. Varela, Maiara M. Romaneli, Marta L. Lima, Samanta
E.T. Borborema, Andre G. Tempone, João P.S. Fernandes , Antiparasitic activity of new
gibbilimbol analogues and SAR analysis through efficiency and statistical methods. Phasci
(2018), doi:10.1016/j.ejps.2018.06.023
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
Antiparasitic Activity of New Gibbilimbol Analogues and SAR Analysis Through
Efficiency and Statistical Methods
Marina T. Varela,1 Maiara M. Romaneli,2 Marta L. Lima,2,3 Samanta E. T. Borborema,2
Andre G. Tempone,2 and João P. S. Fernandes*,1
1Departamento de Ciências Farmacêuticas, Universidade Federal de São Paulo, Rua
São Nicolau 210, 09913-030 Diadema-SP, Brazil
2Centro de Parasitologia e Micologia, Instituto Adolfo Lutz, Av. Dr. Arnaldo 355, 01246-
000 São Paulo-SP, Brazil
3Instituto de Medicina Tropical de São Paulo, Universidade de São Paulo, Av. Dr.
Enéas Carvalho de Aguiar 470, 05403-000 São Paulo-SP, Brazil
ABSTRACT: Chagas’ disease and leishmaniasis are parasitic infections enrolled
among the neglected tropical diseases, which urge for new treatments. In the search
for new chemical entities as prototypes, gibbilimbols A/B have shown antiparasitic
activity against Trypanosoma cruzi and Leishmania infantum, and then a set of
analogues (LINS03 series) of this natural product were synthesized and evaluated in
vitro against the parasites. In the present paper we reported five new compounds with
activity against these protozoan parasites, and quite low cytotoxicity. Moreover, the
interference of plasma membrane permeability of these analogues were also
evaluated. We found that [(4‐methoxyphenyl)methyl]octylamine (4) was noteworthy due
to its high activity against the amastigote form of both parasites (IC50 1.3-5.8 µM) and
good selectivity index. In order to unveil the SAR for this chemotype, we also presented
a group efficiency analysis and PCA and HCA study, which indicated that the methoxyl
provides good activity with lower cytotoxicity to mammalian cells. The results from SAR
analyses suggest different mechanisms of action between the neutral and basic
compounds. In summary, the analogues represent important activity against these
parasites and must be prototypes for further exploitation.
KEYWORDS: Anti-trypanosoma, Group efficiency, Leishmanicide, Natural product
analogues, SAR analysis.

ACCEPTED MANUSCRIPT
Antiparasitic Activity of New Gibbilimbol Analogues and SAR Analysis Through
Efficiency and Statistical Methods
1. Introduction
Neglected tropical diseases are a group of 17 diseases that afflict mostly tropical and
subtropical countries and affect over 1 billion people, especially in developing and
underdeveloped countries (WHO, 2017). Among them, American trypanosomiasis (also
known as Chagas’ disease) and visceral leishmaniasis, caused by the protozoa
Trypanosoma cruzi and Leishmania infantum, respectively, are noteworthy. These
diseases affect millions of people worldwide and have high morbidity and mortality
rates, making them a real public health problem in such countries. The drugs currently
available for the management and treatment of these diseases (benznidazole,
nifurtimox, pentavalent antimonials, amphotericin B and miltefosine) are highly toxic,
mostly effective only during the acute phase and are losing efficacy due to parasite
resistance (Nunes et al., 2013; Mcgwire and Satoskar, 2014; Utzinger et al., 2012).
Even though the necessity of new drugs to treat these parasitic infections is clear, the
pharmaceutical industry does not have economical interest to invest in research and
development and then the hard work is mainly done by the academy (Trouiller et al.,
2002).
There has been a rise in the field of natural product research and its use as prototypes
for the development of novel bioactive compounds (Harvey et al., 2015; Atanasov et
al., 2015). Regarding this, our group had inspiration from some phenolic compounds,
gibbilimbols A and B (Figure 1), isolated from the leaves of Piper malacophyllum which
showed considerable biological activity against both parasites and were unique
regarding their simple structure that could be further explored (De Oliveira et al., 2012).
The first series of proposed analogues (Varela et al., 2016) with 6 different functional
groups (ester, amide, imine, amine, α,β-unsaturated ketone and its corresponding
alcohol) in the side chain of the phenol ring were tested. An improvement in biological
activity was observed, especially for the amine derivative. However, all these
compounds were even more cytotoxic to mammalian cells than the natural prototypes.
Going further, a novel series, lacking the phenol group, was designed (Varela et al.,
2017) and also showed considerable activity against the parasites but with better
cytotoxicity profile, suggesting the phenolic hydroxyl might be related to the toxicity.
Moreover, we also found that a selected compound does not affect the membrane
integrity as verified in the natural prototype gibbilimbol B (De Oliveira et al., 2012). In
summary, the first compounds led us to believe that the substitution of the phenol for a
phenyl or a methoxyphenyl moiety reduced the cytotoxicity maintaining some of the
antiparasitic activity in most of the tested molecules. Considering this, the present work
further explores these modifications and their impact on biological activity as well as a
structure-activity relationship (SAR) study and assessment of their effect in a
membrane disruption assay using the fluorescent probe Sytox green.
In continuation on group’s effort to discover novel antiparasitic compounds active
against both infective and proliferative forms of T. cruzi and L. infantum, five new
analogues (LINS03 series) were prepared and tested to assess this information as well

ACCEPTED MANUSCRIPT
as their selectivity towards parasites over mammalian cells. Previous results (Varela et
al., 2016, 2017) showed that the presence of a secondary amine in the side chain
along with a phenyl group instead of phenol was associated with the high activity and
lower absolute cytotoxicity, however the selectivity towards the parasites went down in
such case (IC50 13.3 μM and selectivity index 5.5). This new set of molecules (Figure 1)
maintained the same modifications on the side chain that resulted in enhancement of
activity in comparison to the natural prototypes, but with other substitution pattern on
the aromatic ring.
gibbilimbol B (18)
LINS03011 (14)
gibbilimbol A (17)
LINS03003 (8)
LINS03014 (1)
LINS03016 (3)
LINS03015 (2)
LINS03017 (4)
LINS03018 (5)
Figure 1: Gibbilimbols A and B, lead synthetic analogues and new series of LINS03
compounds 1-5.
2. Material and Methods
2.1. Chemicals and equipment
All the reagents and solvents were acquired from commercial sources in adequate
purity to be used without further purification. The progress of the reactions was
monitored through TLC in silica gel plates with fluorescence indicator and visualized in
UV at 254 nm. The hydrogen (1H) and carbon (13C) NMR data were acquired on a
Bruker Advance 300, operating at 300 MHz (1H) and 75 MHz (13C) in CDCl3. Chemical
shifts (δ, in ppm) were measured using TMS as internal standard. Coupling constants
are reported in units of hertz (Hz), if applicable. The purity was checked through GC-
MS analysis performed in a Shimadzu GC-2010 coupled to low-resolution mass
spectrometer (LRMS) GCMS-QP2010+, using helium as carrier gas in silica capillary

ACCEPTED MANUSCRIPT
column. Compounds with >95% content were considered pure to biological testing. The
LRMS were obtained through electron impact ionization (70 eV). The ion-radical and its
fragments are given in mass/charge ratio (m/z).
2.2. Synthesis of compound 1
1.5 mmol of benzoyl chloride were dissolved in 15 mL of CH2Cl2 in a round bottom
flask. To this solution, 1 mmol of 1-octanol and 1.5 mmol of trimethylamine (TEA) were
added. The reaction was carried out at room temperature overnight and the solution
was washed with 3x15 mL of aqueous NaHCO3 and 2x15 mL of brine. The crude
product was purified by flash column chromatography in silica gel, using hexane:AcOEt
(3:1) as eluent.
1
Octyl benzoate (1). Colorless oil; yield 71% (0.166 g). H NMR (300 MHz, CDCl3): δ =
8.05 (d, J = 7.2 Hz, 2H), 7.59-7.52 (m, 1H), 7.47-7.40 (m, 2H), 4.32 (t, J = 6.7 Hz, 2H),
1.82-1.71 (m, 2H), 1.48-1.20 (m, 10H), 0.89 (t, J = 6.6 Hz, 3H). 13C NMR (75 MHz,
CDCl3): δ = 166.70, 132.77, 130.57, 129.53, 128.31, 65.16, 31.80, 29.26, 29.20, 28.74,
26.06, 22.64, 14.09. GC tR = 13.56 min. LRMS (EI) m/z (rel. int. %) 234 [M]+ (1), 123
(100), 105 (80), 84 (15), 77 (37), 55 (10), 41 (10). These data are in accordance to
previous literature report (Barry et al., 1985).
2.3. General procedure for the preparation of compounds 2 and 3
In a round bottom flask, 1 mmol of 4-methoxybenzoic acid (2) or vanillic acid (3) and
1.2 mmol of 1-octanol were dissolved in 10 mL of toluene, and then 0.1 mL of
concentrated sulfuric acid was added dropwise. The mixture was heated to reflux for 4
hours, using a Dean-Stark apparatus to remove the water via azeotropic distillation.
The mixture was washed with 3x15 mL of aqueous NaHCO3 and 2x10 mL of distilled
water and dried with anhydrous Na2SO4. The solvent was evaporated, and the crude
products were purified through column chromatography in silica gel, using
hexane:AcOEt (2:1) as eluent.
1
Octyl 4-methoxybenzoate (2). Yellowish oil; yield 13% (0.034 g). H NMR (300 MHz,
CDCl3): δ = 8.05-7.99 (m, 2H), 6.97-6.90 (m, 2H), 4.30 (t, J = 6.6 Hz, 2H), 3.88 (s, 3H),
1.83-1.71 (m, 2H), 1.50-1.23 (m, 10H), 0.91 (t, J = 6.8 Hz, 3H). 13C NMR (75 MHz,
CDCl3): δ = 166.47, 163.24, 131.54, 123.01, 113.55, 64.86, 55.42, 31.81, 29.27, 29.22,
28.79, 26.07, 22.66, 14.10. GC tR = 16.60 min. LRMS (EI) m/z (rel. int. %) 264 [M]+
(9), 153 (20), 152 (100), 135 (50), 107 (8), 77 (10). These data are in accordance to
previous literature report (Barry et al., 1985).
1
Octyl 4-hydroxy-3-methoxybenzoate (3). Yellowish oil; yield 56% (0.157 g). H NMR
(300 MHz, CDCl3): δ = 7.66 (dd, J = 8.3, 1.8 Hz, 1H), 7.58 (d, J = 1.8 Hz, 1H), 6.96 (d,
J = 8.3 Hz, 1H), 6.03 (br.s, 1H), 4.31 (t, J = 6.7 Hz, 2H), 3.97 (s, 3H), 1.83-1.72 (m,
2H), 1.50-1.22 (m, 10H), 0.91 (t, J = 6.9 Hz, 3H). 13C NMR (75 MHz, CDCl3): δ =
166.50, 149.90, 146.15, 124.08, 122.70, 114.00, 111.72, 65.02, 56.10, 31.80, 29.27,
29.22, 28.79, 26.06, 22.66, 14.10. GC tR = 17.91 min. LRMS (EI) m/z (rel. int. %) 280

ACCEPTED MANUSCRIPT
[M]+ (17), 169 (15), 168 (100), 151 (30), 123 (10), 41 (5). These data are in
accordance to previous literature report (Pearl and Beyer, 1951).
2.4. Synthesis of compound 4
In a flat-bottom flask, 5 mmol of anisaldehyde and 5 mmol of 1-octylamine were added
into 5 mL of tetrahydrofuran. The solution was reacted in an ultrasonic bath for 1 hour,
and then dried with anhydrous Na2SO4, filtered and the solvent removed under reduced
pressure. The product was purified by flash column chromatography in silica gel, using
hexane:AcOEt (5:1) as eluent. The corresponding imine 13 was used directly in the
next step and was in accordance with previous report from our group (Varela et al.,
2017). Afterwards 10 mmol of NaBH4 was dissolved into 10 mL of methanol. The
solution was stirred for 1 hour, and then was added to a solution of corresponding
imine (3 mmol) in 10 mL of methanol. The reaction was carried out for 4 h at 25 °C, and
then concentrated under reduced pressure. The residue was taken up with 10 mL ethyl
acetate, and washed with 3x10 mL aqueous NaHCO3. The organic layer was dried with
anhydrous Na2SO4, filtered and the solvent removed under reduced pressure to give 4
with adequate purity.
1
[(4‐methoxyphenyl)methyl]octylamine (4). Colorless oil; yield 87% (0.651 g). H NMR
(300 MHz, CDCl3): δ = 7.26-7.20 (m, 2H), 6.83-6.88 (m, 2H), 3.79 (s, 3H), 3.72 (s, 2H),
2.60 (t, J = 7.2 Hz, 2H), 1.67 (br.s, 1H), 1.57-1.46 (m, 2H), 1.39-1.20 (m, 10H), 0.88 (t,
J = 6.7 Hz, 3H). 13C NMR (75 MHz, CDCl3): δ = 158.60, 132.50, 129.36, 113.77, 55.27,
53.41, 49.36, 31.84, 30.01, 29.54, 29.28, 27.39, 22.67, 14.11. GC tR = 15.79 min.
LRMS (EI) m/z (rel. int. %) 249 [M]+ (5), 190 (5), 150 (15), 137 (10), 121 (100), 77 (5).
These data are in accordance to previous literature report (Castillo et al., 2016).
2.5. Synthesis of compound 5
In a round bottom flask 5 mmol of 4-methoxybenzoic acid was dissolved 10 mL of
toluene and 10 mmol of thionyl chloride was added slowly to the mixture. After 2 hours
under reflux the solvent was evaporated and the mixture was cooled and used without
further purification in the next step. The crude mixture was dissolved in
dichloromethane and added to a previously prepared solution of 2 mL of triethylamine
and 5 mmol of 1-octylamine in 2 mL of the same solvent. The reaction proceeded by 18
h at 60 °C and then the mixture was washed 3x10 mL of 0.1 M hydrochloric acid, 3x10
mL of sat. NaHCO3 and 2x10 mL of brine. The organic layer was dried with anhydrous
Na2SO4, filtered and the solvent removed under reduced pressure. The product was
crystallized from MeOH/water to give 5 in adequate purity.
Octyl 4-methoxybenzylamide (5). Yellowish solid; yield 39% (0.205 g); mp 67-68 °C. 1H
NMR (300 MHz, CDCl3): δ = 7.74-7.70 (m, 2H), 6.95-6.89 (m, 2H), 6.02 (br.s, 1H), 3.85
(s, 3H), 3.50-3.39 (m, 2H), 1.66-1.55 (m, 2H), 1.41-1.22 (m, 10H), 0.88 (t, J = 6.8 Hz,
3H). 13C NMR (75 MHz, CDCl3): δ = 167.01, 162.05, 128.58, 127.20, 113.73, 55.40,
40.07, 31.80, 29.77, 29.31, 29.22, 27.04, 22.65, 14.09. GC tR = 19.56 min. LRMS (EI)

ACCEPTED MANUSCRIPT
m/z (rel. int. %) 263 [M]+ (12), 206 (10), 164 (17), 151 (15), 135 (100), 107 (10), 77
(12). These data are in accordance to previous literature report (Ali et al., 2014).
2.6. Biological assays
Golden hamsters and BALB/c mice were obtained from the animal breeding facility at
the Adolfo Lutz Institute-SP, Brazil. The animals were maintained in sterilized cages
under a controlled environment, and received water and food ad libitum. Animal
procedures were performed with the approval of the Research Ethics Commission, in
agreement with the Guide for the Care and Use of Laboratory Animals from the
National Academy of Sciences.
2.7. Parasites and mammalian cell maintenance
Leishmania (L.) infantum (MHOM/BR/1972/LD) was maintained in Golden hamsters
(Mesocricetus auratus) up to approximately 60-70 days post-infection. Promastigotes
were maintained in M-199 medium supplemented with 10% fetal calf serum and 0.25%
hemin at 24 °C. Amastigotes were obtained from the spleen of previously infected
hamsters and purified by differential centrifugation. Macrophages were collected from
the peritoneal cavity of BALB/c mice by washing them with RPMI-1640 medium
supplemented with 10% fetal calf serum, and were maintained at 37 °C in a 5% CO2-
humidified incubator. Trypomastigotes of T. cruzi (Y strain) were maintained in Rhesus-
monkey kidney cells (LLC-MK2 - ATCC CCL 7), cultivated in RPMI-1640 medium
supplemented with 2% fetal calf serum at 37 °C in 5% CO2-humidified incubator. The
murine conjunctive cells (NCTC clone 929, ATCC) were maintained in RPMI-1640
supplemented with 10 % FBS at 37 °C in a humidified atmosphere containing 5% CO2.
NCTC - clone 929 cells were cultivated in M-199 medium, supplemented with 10% fetal
calf serum at 37 °C in 5% CO2 humidified incubator.
2.8. Determination of the 50% inhibitory concentration (IC50)
Leishmania infantum. Promastigotes in late growth-phase (non-stationary at 3x107
cells/mL, passage 5) were counted in a hemocytometer chamber and seeded at 1x106
cells/well, with a final volume of 150 µL. The compounds were dissolved in DMSO and
diluted in M-199 medium in 96-well microplates, with the highest concentration of 100
µM for 48 h at 24 °C. The parasite viability was determined using the MTT colorimetric
assay. The optical density was read at 570 nm (FilterMax F5 Multi-Mode Microplate
Reader, Molecular Devices) using control wells without drugs (100% viability) and
without cells (blank). The control group consisted of promastigotes incubated with 0.5%
DMSO. Miltefosine was used as a standard drug. Compounds were tested to the
highest concentration of 100 µM and were reported as not active (>100 µM) when the
IC50 value was above this concentration (Reimão et al., 2008).
Amastigotes. Peritoneal macrophages were collected from the peritoneal cavity of
BALB/c mice, and the macrophages were seeded at 1x105 cells/well for 24 h in a 16-

ACCEPTED MANUSCRIPT
well slides (NUNC, ThermoFisher Scientific). Amastigotes were prepared as described
previously in a 1:10 ratio of macrophages to amastigotes and incubated in RPMI-1640
for 24 h at 37 °C 5% CO2-humidified incubator. The compounds were incubated with
infected macrophages for 72 h. Miltefosine was used as a positive control.
Subsequently, the cells were fixed with MeOH, stained with Giemsa and observed
using a light microscope. The parasite burden was determined by the number of
infected macrophages out of 400 cells (Reimão et al., 2008).
Trypanosoma cruzi. Trypomastigotes were counted in a hemocytometer chamber and
seeded at 1x106 cells/well in 96-well microplates. The compounds were diluted in
RPMI-1640 medium and incubated for 24 h at 37 °C in a 5% CO2-humidified incubator.
The parasite viability was determined using the resazurin (0.011% in PBS). The optical
density was read at 570 nm using control wells without drugs (100% viability) and
without cells (blank). The control group consisted of trypomastigotes incubated with
0.5% DMSO. Benznidazole was used as a positive control. Compounds were tested to
the highest concentration of 100 µM and were reported as not active (>100 µM) when
the IC50 value was above this concentration (Martins et al., 2016).
Intracellular amastigotes. Peritoneal macrophages were dispensed in 16-well chamber
slides (NUNC, Thermo, USA) and maintained for 24 h in the same medium at 37 °C in
a 5% CO2 humidified incubator for attachment. Non-adherent cells were removed by
two step washings with medium. After 24 h, these cells were infected with 1x106 culture
trypomastigote forms for 4 h. Subsequently, infected cells were incubated with the
compounds for 48 h. Finally, the slides were fixed with methanol, stained with Giemsa,
and observed in light microscopy. The parasite load was defined by counting 400
macrophages/well by evaluating the number of infected macrophages. Benznidazole
was used as the standard drug (Martins et al., 2016).
2.9. Cytotoxicity in mammalian cells
NCTC cells were counted in a hemocytometer chamber, seeded at 6 x 104/well and
incubated in highest concentrations to 200 μM for 72 h at 37 °C in a 5% CO2-humidified
incubator. The cell viability was determined using the MTT assay (Reimão et al., 2012).
Miltefosine was used as a positive control. The selectivity index (SI) was determined
using the relationship, CC50 against NCTC/IC50 against parasites.
2.10. Evaluation of plasma membrane permeability
Late growth-phase trypomastigotes of T. cruzi (2 x 106/well) were washed and
incubated in the dark with 1 μM Sytox Green probe (Molecular Probes) in HANKS'
balanced salts solution (HBSS; Sigma-Aldrich) supplemented with 10 mM D-Glucose
(HBSS + Glu). The compounds was added (t = 0) at IC50 and IC90 and fluorescence
was measured every 10 minutes for up to 1 h. The maximum permeabilization was
obtained with 0.5% Triton X-100. Fluorescence intensity was determined using a
fluorimetric microplate reader (FilterMax F5 Multi-Mode Microplate Reader, Molecular
Devices) with excitation and emission wavelengths of 485 and 520 nm, respectively.

ACCEPTED MANUSCRIPT
The following internal controls were used in the evaluation: 1) the background
fluorescence of the compound at the respective wavelengths, 2) the possible
interference of DMSO. Samples were tested in duplicates (Luque-Ortega and Rivas,
2010).
2.11. Membrane disruption assessment by Sytox Green model
The compounds 1-5 were studied using the fluorescent probe Sytox Green in late
growth-phase (non-stationary) of trypomastigotes of T. cruzi (Figure 2). Sytox Green
(Molecular Probes) is a high-affinity nucleic acid stain that penetrates cells with
compromised plasma membranes and enhances its fluorescence by more than 500-
fold upon nucleic acid binding. The parasite cells were washed in phosphate-buffered
saline, seeded at 2x106 cells/well and incubated for 15 min with 1 µM of Sytox Green in
M199 medium without phenol red, supplemented with fetal bovine serum and hemin.
The compounds 1-5 were added at the previously determined IC50 and the
fluorescence was measured every 10 min for a total period of 60 min. Maximum
fluorescence was determined by the addition of 0.5% of the non-ionic surfactant Triton
X-100 (positive control). The control group consisted of trypomastigotes incubated with
0.5% MeOH. Fluorescence was measured using a microplate reader (FilterMax F5,
Molecular Devices) with excitation and emission wavelengths of 485 and 520 nm,
respectively (De Oliveira et al., 2012; Varela et al., 2017).
2.12. Group efficiency analysis
Group efficiency (GE) values were calculated using the equation 1:
GE = (1.37xΔpIC50)/ΔHA
(1)
where ΔpIC50 is the variation of the negative logarithm of IC50 value in molar between
the substituted and non-substituted compound and ΔHA is the variation of heavy (non-
hydrogen) atoms in the specific substituent (Verdonk and Rees, 2008).
2.13. Descriptor calculations
Lipinski’s rule-of-five (RoF) descriptors calculated 1-octanol/water coefficient (logP),
molecular mass (MW), hydrogen bond donors (HBD) and acceptors (HBA) were
calculated using the online platform Molinspiration (Molinspiration Property
Calculation Service, 2015). Other parameters such as rotatable bonds (RB),
distribution coefficient in different pHs (logD), water solubility (logS) in pH 7.4,
hydrophilic-lipophilic balance (HLB), molecular surface area (MSA), van der Waals
volume, polarizability, total dipole moment, refractivity, dreiding and MMFF94 energies
were calculated using Marvin Sketch 16.1.8.0 (ChemAxon Inc., 2016) using the default
configuration in the software (Table 4 and Supplementary Information). The values for
the Hammett’s σ and Hansch’s π descriptors (Supplementary Information) were
obtained from Kubinyi (1993).

ACCEPTED MANUSCRIPT
2.14. Principal component and hierarchical cluster analyses
The exploratory principal component (PCA) and hierarchical cluster (HCA) analyses
were carried out using the descriptors presented in Table 4 using the PAST 3.17
software (Hammer et al., 2001). The descriptor data were pretreated by autoscaling
procedure, where each value was subtracted by the column’s mean and divided by the
standard deviation (Vendrame et al., 2002; Van der Berg et al., 2006). The PCA
eigenvectors (principal components or PCs) were obtained by using the variance-
covariance matrix. The obtained component loadings are presented in Table 5, and the
scatter plot of the PCA data is presented in Figure 3. The HCA were performed using
Ward’s Method and the Euclidean similarity index. The obtained dendrogram plot is
presented in Figure 4.
3. Results and Discussion
3.1. Chemistry
The compounds were synthesized following the same methodology previously
described (Varela et al., 2016, 2017). The ester 1 was prepared from the reaction
between benzoyl chloride and 1-octanol in the presence of TEA. Esters 2 and 3 were
prepared following the classical reaction of Fischer’s esterification of 4-methoxybenzoic
or vanillic acid, respectively, and 1-octanol, using concentrated sulfuric acid as catalyst
and toluene as solvent. Dean-Stark apparatus was used for the azeotropic distillation of
the water during the reaction. Amine 4 was obtained through reduction of the
corresponding imine previously reported, using NaBH4 as a reductive agent. Amide 5
was prepared in a two-step procedure starting from the conversion of 4-
methoxybenzoic acid to its acyl chloride with thionyl chloride. The following step was
done by the addition of a 1-octylamine and TEA in the reaction vessel without prior
purification. The compounds were obtained with variable yields but through easy
purification and sufficient quantity for spectroscopic characterization and biological
assessment.
3.2.
Biological evaluation
The antiparasitic activity of these 5 new analogues was determined in vitro for both
intracellular and extracellular forms of T. cruzi and intracellular form of L. infantum and
NCTC cells to determine toxicity for mammalian cells, following the same methodology
as described in Varela et al. (2017). Results are shown in Table 1.

ACCEPTED MANUSCRIPT
Table 1. Antiparasitic activity of compounds against trypomastigotes and amastigotes
of T. cruzi and promastigotes and amastigotes of L. infantum.
Functional
group
R
IC50 ±SD, µM
Functional
Group
CC50 ±SD, µM
Code
1
R
NCTC
TCT
TCA
LIP
LIA
47.4
±17.8
35.8
±15.7
H
ND
>100
>200
>200
27.1
±10.9
20.1
±2.6
2
4-OMe
ND
>100
>100
4.4
±1.5
6.8
±9.5
123.6
±1.4
3-OMe-4-OH
4-OMe
3
4
ND
ND
4.2
±2.8
1.33
±0.3
5.8
±3.5
31.9
±0.1
11.4
±5.8
5.23
±0.6
8.7
±5.0
91.5
±34.8
5
4-OMe
4-OH
ND
11.8
±3.2
22.5
±6.3
82.7
±5.2
82.8
±7.3
6[a]
>100
>100
6.1
±2.4
49.5
±10.1
7[a]
8[a]
9[a]
4-OH
4-OH
4-OH
>100
>100
17.0
±0.4
5.5
±1.7
28.6
±6.2
1.8
±0.7
23.5
±5.9
4.5
±0.8
130.3
±18.4
>100
>100
>100
26.5
±1.3
45.1
±11.3
45.4
±14.7
10[a]
4-OH
>100
>100
>100
>100
19.9
±3.8
28.3
±1.8
27.8
±1.3
3.0
99.2
±9.1
11[a]
12[b]
13[b]
14[b]
4-OH
H
>100
>100
>100
50.0
±18.3
40.4
±15.8
34.9
±2.8
16.7
±3.5
>200
>200
4-OMe
H
>100
20.7
±0.9
13.3
±1.4
73.5
±1.9
±0.06
15[b]
H
>100
>100
>100
>100
>100
>200
>200
26.2
±6.1
16[b]
4-OMe
>100
>100
75.3
±9.2
>100
>100
94.9
±8.9
17[a]
18[a]
4-OH
4-OH
alkene
alkene
>100
>100
>100
>100
>200
>200
IC50: Inhibitory Concentration 50%. CC50: Cytotoxicity concentration 50%. CI 95%:
confidence interval 95%. ND: not determined/not tested; TCT: T. cruzi trypomastigotes;
TCA: T. cruzi amastigotes; LIP: Leishmania infantum promastigotes; LIA: Leishmania
[a]
[b]
infantum amastigotes. Available from Varela et al. (2016); Available from Varela et
al. (2017)
The new set of molecules has shown SI higher than those for the former molecules.
The most active compounds against trypomastigote forms (infective – extracellular) of

ACCEPTED MANUSCRIPT
T. cruzi were 3 and 4 with IC50 values of 4.4 and 4.2 μM. Compound 3 exhibited a
selectivity index (SI) of 28, comparable to the obtained for its phenolic derivative, 9
(IC50 4.5 μM and SI 29), meaning that the addition of the methoxyl group in the 3-
position of the aromatic ring did not affect the activity against this form or mammalian
cells. Amide 5 was also active (IC50 11.4 μM) but not as selective.
In fact, promising compounds should be active against the intracellular forms of both
parasites, since these are the main forms present in the human host and accordingly in
the chronic phase of the disease. Considering the activity against the T. cruzi
amastigotes, compounds 3, 4 and 5 showed interesting activity, which is comparable to
the pharmacological standard, benznidazole (~5 μM) (Varela et al., 2017). The
accepted criterion for a good lead compound for Chagas’ disease is IC50 <10 μM
(Katsuno et al., 2015), and these compounds presented IC50 values of 6.8, 1.3 and 5.2
μM, respectively. Comparing the esters, it is possible to clearly see that the 3-methoxy-
4-hydroxybenzyl groups had an impact on the biological activity leading to compound 3
that is ~6-fold more active than the unsubstituted analogue 1 and ~3-fold more active
than the methoxy compound 2. In addition, the phenolic analogue 9 was not active in
the amastigote form and was more cytotoxic indeed (Varela et al., 2016).
The amide group also led to improvement of the activity, since when comparing the
analogues 2 and 5, classical bioisosters, it is possible to verify an improvement in the
activity in a 4-fold order. However, this role only applies to the methoxy-substituted
analogues. When comparing the isosterism between the pairs 1/15 and 9/10, this was
not observed. Possibly the presence of the methoxy group changed completely the role
in such compounds, leading to different interactions with the target. In addition, the
amides also led to more cytotoxic activity. Nonetheless, compound 5 showed
considerable selectivity towards the parasites (SI = 17.5). The amine 4 is the most
active compound against T. cruzi amastigotes (IC50 1.3 μM) and has the highest
selectivity index of the series for this form of the parasite (SI = 24). We had already
verified that the presence of the amine group in the lateral chain leads to highly active
compounds in every forms of the parasite (Varela et al., 2016, 2017). When compared
to the others analogues previously tested, the compound 4 is 10-fold more active than
the phenyl derivative 14 (Varela et al., 2017) and 5-fold more active than the phenol
derivative 8 (Varela et al., 2016). Compound 4 also showed the highest selectivity over
mammalian cells.
Regarding the activity against the amastigote form of the L. infantum parasite, only
compounds 4 and 5 were significantly active, exhibiting IC50 values of 5.8 and 8.7 μM,
respectively. Once again, the amine derivative 4 was noteworthy due to its high activity
against this parasite, which is also comparable to miltefosine (~7 μM) (Varela et al.,
2017), albeit some cytotoxicity remained with this molecule (CC50 31 μM). After the
results obtained in this and previous reports, we should also attribute some toxicity to
this molecular moiety. Maybe the basic nitrogen is leading to some toxicity that must be
further investigated. Compound 5 is the first amide derivative that showed important
activity against this form of the parasite. On the other hand, compound 5 also led to
some toxicity to mammalian NCTC cells (CC50 91 μM), although this compound is also
quite selective towards L. infantum (SI = 10.5). Considering the nitrogen containing
molecules, the most active analogue tested so far against L. infantum amastigotes was
8, with an IC50 of 1.8 μM and selectivity index of 13 (Varela et al., 2016).

ACCEPTED MANUSCRIPT
3.3.
Plasma Membrane Permeability
Table 2. Relative permeabilization of the cell membrane of T. cruzi trypomastigotes
promoted by the compounds 1-5, calculated after 60 min of exposition. Triton-X100
was the positive control (100%).
Compound
Permeabilization (relative to control)
1
2
3
4
5
8% ± 0.4
3% ± 2.0
11% ± 6.0
21% ± 3.0
1.9% ± 0.8
Figure 2. Determination of membrane permeabilization using Sytox Green assay in T.
cruzi trypomastigotes. Statistical significance was compared to negative (untreated)
control (*P <0.05; ***P <0.0001).
In a previous report from De Oliveira et al. (2012) it was demonstrated the ability of the
natural product gibbilimbol B to increase the membrane permeability using the Sytox
Green model in L. infantum. This fluorescent probe is not able to permeate the cell
membrane if the plasma membrane is intact. Gibbilimbol B increased the membrane
permeability to the probe, suggesting that this molecule could present membrane
disruption activity. The authors discussed that the surfactant-like characteristic of the
molecule (polar head represented by the hydroxyl and aside chain apolar tail) should
favor its insertion in the membrane phospholipids and interact with the membrane
ergosterol in a similar manner than amphotericin does. This could also explain the
selectivity of the alkylphenol towards the parasite. In a more recent work (Varela et al.,
2017), our group also tested analogue 14 in this model and we found that no significant
membrane permeation occurred, suggesting that the synthetic analogues would

ACCEPTED MANUSCRIPT
present a different mechanism of action from gibbilimbols or at least 14 do not have a
membrane disruption activity.
Herein, we tested all the compounds presented in the T. cruzi Sytox Green model. In
Figure 2, it is possible to observe that only compounds 3 and 4 showed low membrane
disruption activity after 60 minutes, and that the ester derivatives 1 and 2 and amide
derivative 5 did not influence the membrane permeability. Triton X-100, used as
positive control, led to important probe penetration after 10 minutes. Although
compounds 3 and 4 had some residual activity on the membrane permeation, they
present considerable selectivity. This residual activity can also explain their higher
cytotoxicity, since both mammalian and parasitic cells suffer the impact of membrane
disruption, even though they do not have exactly the same structural composition. The
results suggest also that the absence of the unsaturation present in gibbilimbol B may
play a role in this activity, since none of the synthetic analogues tested so far present
this effect. Moreover, considering that compound 4 had slightly increased the
membrane permeability in comparison to its non-methoxylated analogue 14 (Varela et
al., 2017), it is also suggested that the presence of the polar (head) group should also
have influence on such mechanism.
3.4.
Group efficiency analysis
In order to have a better insight into the structure-activity relationship of this set of
molecules, statistical approaches should be considered and were performed. Since
there is no identified target or mechanism of action to these molecules, this approach
should shed light in the important structural features of these compounds. Ligand
efficiency is a metric approach to normalize the affinity of a group of molecules for the
same target (Verdonk and Rees, 2008) and it is useful in the optimization of lead
compounds during drug development. Group efficiency follows that same principle but
instead of comparing the entire molecule, it evaluates the impact of certain groups on
the activity according to the number of additional heavy-atoms (non-hydrogen atoms)
added to a parental structure (Schultes et al., 2010).
To make this analysis, we assumed that the compounds with the same functional
group should act over the same common target. Combining the results observed in this
study with the previously obtained information of activity and cytotoxicity of other
analogues with the same functional groups and different ring substituents the group
efficiency of each modification can be calculated (Verdonk and Rees, 2008).
Since all of the compounds tested thus far have analogues with a phenyl ring, those
were chosen to be the parent (non-substituted) molecule. The only exception was
compound 3 that had two modifications (3-methoxy-4-hydroxy substitution pattern on
the aromatic ring) and the chosen parent molecule was the 4-phenol derivative 9,
considering the addition of the methoxy group. It is important to stress that this is a
hypothesis based on the premise that these similar molecules would interact with the
same molecular target and/or have the same mechanism of action. The data for the
calculated group efficiencies are shown in Table 3.

ACCEPTED MANUSCRIPT
Table 3. Group efficiency calculated for part of the LINS03 series. Bold values
represent high efficiency of the group.
Substituents
Group Efficiency
Functional
Group
4-R
OH
3-R
H
TCT
1.401
0.166
0.007[a]
0.790
0.646
0.117
0.475
ND
TCA
LIP
ND
LIA
NCTC
ND
ND
0.172
ND
ND
ND
Ester
OMe
OH
H
ND
ND
OMe
H
ND
ND
0.020[a]
0.882
0.550
0.678
0.248
ND
OH
ND
0.474
ND
ND
Amide
Amine
Imine
OMe
OH
H
0.878
0.525
0.685
ND
0.726
1.325
0.315
ND
H
-1.340
ND
OMe
OH
H
H
0.210
0.005
OMe
H
ND
ND
ND
ND
ND: not determined. TCT: trypomastigotes of T. cruzi; TCA: amastigotes of T. cruzi;
LIP: promastigotes of L. infantum; LIA: amastigotes of L. infantum; NCTC: mammalian
cells. [a]Calculated comparing the variability between 3 and its phenolic derivative 9.
The interpretation of these results depends on the molecular weight of the compounds
evaluated. For compounds with a molecular weight <300 Da the calculated GE must be
at least 0.52 for each heavy atom in order to maintain the ligand efficiency. This means
that the addition of a group is only considered efficient if it can actually cause an
improvement in potency or affinity when compared to another molecule without that
group (Hopkins et al., 2014). The GE values obtained here agree with our previous
discussion of the results. For example, the addition of a 4-OH group in the phenyl ring
(9) led to a considerable improvement in the activity of the parental molecule 3 against
the trypomastigote form of T. cruzi, giving a GE value of 1.401. In counterpart, the
addition of the methoxy group in the 3-position did not improve the activity as well,
leading to a GE value of only 0.007. Thus, the insertion of this group does not mean
real improvement in the activity.
In addition, we can combine the observations of GE values regarding the improvement
in biological activity and cytotoxicity to choose the best compound’s profile and find the
more selective compounds. Considering this, it is possible to verify that the substitution
profile in the amine series should play the role in the safety of those compounds. The
addition of the 4-OH group led to an important improvement in the activity against T.
cruzi amastigotes (GE = 0.525), but its insertion also increased the cytotoxicity even
more (GE = 0.678). Although compound 4 is even more active than the phenol
derivative against T. cruzi amastigotes, the insertion of the 4-methoxy group led to a
slightly higher GE value (0.685). However, this group did not increase the cytotoxicity
as much (GE = 0.248). On the other hand, the insertion of the phenolic 4-OH improved
very much the activity against L. infantum amastigotes (GE = 1.325) whereas for the L.
infantum promastigotes it led to lower activity in comparison to the phenyl analogue
(GE= - 1.34). These results suggest that this compound may behave as an outlier
regarding antipromastigote activity, or also the presence of the OH group led to a host
cell-dependent mechanism of action, which not happen in the extracellular form.
Moreover, the addition of the methoxy group did not show the same behavior as seen
for the amastigotes of T. cruzi, showing favorable contribution in the amide analogue 5
(GE = 0.726) but not as favorable for the amine analogue 4 (GE = 0.315).

ACCEPTED MANUSCRIPT
3.5.
Drug likeness assessment
Since our study is based on a phenotypic assay, some other characteristics can be
assessed, independently of the target or mechanism of action. Lipinski’s rule of five is a
filter used in the pharmaceutical industry during drug discovery, especially high
throughput screening (HTS), to discard molecules that probably will not present
adequate oral bioavailability or skin penetration (Lipinski et al., 2001). To have good
bioavailability a molecule should have a favorable hydrophilic-lipophilic balance to allow
good membrane diffusion, and adequate water solubility, making sure that the drug can
be dissolved from the pharmaceutical form in order to be absorbed (Bartzatt, 2017).
Lipinski’s rule of five (RoF) states that molecules with poor oral bioavailability and/or
skin permeation usually present more than 1 of these characteristics:(1) more than 5
HBD; (2) more than 10 HBA; (3) MW over than 500; (4) logP higher than 5 (Lipinski et
al., 2001). It is important to stress that even though this rule serves as a guide for oral
bioavailability it must be used carefully since many promising compounds may be
wrongly discarded and some successful market drugs do not fulfill these criteria
(Giménez et al., 2010).
Table 4. Drug-likeness assessment and additional physicochemical and steric
properties calculated.
RoF
violations
Comp. LogP
MW
HBD HBA
RB
HLB
logD5.0
logS7.4
MSA
1
2
5.57
5.62
4.91
5.01
4.65
4.81
5.00
4.48
5.09
4.11
4.75
5.23
5.29
4.96
4.59
5.35
5.87
5.87
234.34
264.37
280.36
249.40
263.38
246.35
248.37
235.37
250.34
249.35
233.35
217.36
247.38
219.37
233.35
260.38
232.37
232.37
0
0
1
1
1
1
2
2
1
2
1
0
0
1
1
0
1
1
2
3
4
2
3
2
2
2
3
3
2
1
2
1
2
2
1
1
9
10
10
10
9
3.15
4.78
6.26
8.82
5.01
2.53
3.02
8.75
4.74
4.99
3.34
1.72
3.43
7.14
3.43
2.60
1.37
1.37
5.08
4.92
4.62
1.26
3.99
5.09
4.67
1.11
4.78
3.85
1.39
1.65
1.53
1.41
4.15
5.23
5.82
5.82
-5.05
-5.06
-4.63
-2.04
-4.92
-4.71
-4.50
-2.11
-4.60
-4.46
-3.26
-3.02
-3.30
-1.93
-4.91
-5.13
-5.76
-5.76
425.99
474.29
485.03
481.37
475.80
422.33
432.51
444.84
437.57
439.28
424.74
413.44
461.71
433.45
428.17
459.81
425.57
425.89
1
1
0
1
0
0
0
0
1
0
0
1
1
0
0
1
1
1
3
4
5
6
8
7
8
8
9
9
9
10
11
12
13
14
15
16
17
18
8
8
8
9
9
8
9
8
8
Calculated Lipinski’s RoF parameters and other descriptors are presented in Table 4
and in Supplementary Information. Most of the compounds tested so far did not violate
any of the Lipinski’s RoF criteria, even though some compounds presented logP values
over 5. None of the compounds violated the RB threshold (<10) value, which is another

ACCEPTED MANUSCRIPT
drug-like descriptor (Veber et al., 2002). The selected logD pH’s justify to evaluate the
behavior of the compounds in several biological compartments such as stomach (pH
1.5) and duodenum (pH 6.5), blood (pH 7.4) and in the parasitophorous vacuoles (pH
5.0), where the amastigote forms of the parasites are found. The drug-likeness
assessment of the compounds suggests that the synthetic analogues present better
water solubility and improved drug-like characteristics, since they are in general less
lipophilic than the natural prototypes. This suggests that the presence of the functional
groups in the LINS03 molecules would help to achieve better pharmacokinetic profile
than the gibbilimbols and also present some metabolic softness, avoiding toxicity in
vivo due to excessive lipophilicity. With this regard, the amine-containing compounds
must be highlighted. The presence of the amine group favors the ionization in
physiological pH, leading to lowered lipophilicity as can be noted by the logD values.
3.6.
HCA and PCA
In order to explain the results obtained for biological activity and cytotoxicity the
physicochemical parameters calculated and described in Table 4 were used to perform
statistical analysis through PCA and HCA. As recommended in literature (Vendrame et
al., 2002), a preliminary exploitation of the descriptors was carried out and the most
important (logP, HBA, HLB, logD5.0 and MSA) were used to obtain the final PCA and
HCA results. The other descriptors did not contribute in a significant manner to the
results, and then were excluded from the final results. PCA is an exploratory metric that
allows the analysis of a set of correlated and dependent variables and gathers the most
important information into new orthogonal variables known as PCs (Vendrame et al.,
2002). The number of PCs must be equal or lower than the number of original
variables. From the loading values, important descriptors can be identified and used to
define the SAR for a set of compounds. HCA used the same final descriptors to create
a dendrogram that separates the analogues according to their similarities. This analysis
is complementary to PCA and if the results are similar and lead to the same pattern of
separation this means that the descriptors are suitable for describing the differences
and similarities of the molecules in the series (Ferreira et al., 2014). The separation in
clusters means that molecules that are in the same cluster have high similarity than
those that are more distant and/or in other clusters (Vendrame et al., 2002; Bartzatt,
2017). The results (qualitative in nature) are graphically presented in a dendrogram,
which is obtained through the calculation of Euclidean distances between the samples
and grouped by similarity indexes (or distances) (Vendrame et al., 2002; Hammer et
al., 2001). The Euclidean distance is measured by using the Ward’s method (Ward Jr,
1963) with the same calculated descriptors used in PCA. In other words, it considers
the similarity distance between all selected descriptors for each pair of molecules. The
more different the molecules’ descriptors, the higher the distance values. Using two
different statistical methods makes the evidence of the SAR data more reliable if both
gave similar answers.
The data presented in Table 4 was pretreated by autoscaling the values of the
descriptors. After autoscaling, each descriptor can be analyzed on the base of
correlations, which avoid the spans of the variables in different ranges that could
influence in the weight of each independent variable on the result of the analysis (Van

ACCEPTED MANUSCRIPT
der Berg et al., 2006). Figures 3 and 4 shows the PCA scatter plot result and the
obtained HCA dendrogram, respectively. The compounds were also classified into high
activity (IC50 <20 μM) and low activity (IC50 >20 μM) categories against T. cruzi
amastigotes for discussion purposes, however this same criteria was applied to the
activity against L. infantum amastigotes.
Figure 3. Scatter plot showing the distribution of the compounds around the axis of
principal components 1 (horizontal) and 2 (vertical). The analogues represented by a
white circle (inside the dashed lines) present higher activity (IC50 lower than 20 µM)
against the amastigote form of T. cruzi. The black dots represent compounds with
lower activity. Cluster A is comprised by the basic compounds, while cluster B is
comprised by the neutral analogues.
The process of the selection of the optimum descriptors in PCA was as follows. There
were several calculated descriptors provided by Marvin software (see Supplementary
Information) for each molecule, and then PCA was carried out. Since no satisfactory
discrimination of the molecules between high activity and low activity was obtained, we
started to discard one variable each time (the one which showed lower weight in the
results) and applying PCA again with the remaining variables. If the quality of the
discrimination improved, we continued discarding descriptors until the optimum
classification is obtained (Vendrame et al., 2002). The best separation was attained by
using only five descriptors. The correlation of such descriptors with the activity seems
to be clear. The amastigotes are intracellular forms of the parasite that is found in the
parasitophorous vacuole where the pH is more acidic (~5.0), hence the importance of
the logD5.0. The HLB correlates to the capacity of the molecule being water soluble and
lipophilic at the same time, which determines whether or not it can be absorbed in the
organism. On a scale of 0-20, the higher the value the more a molecule is water
soluble (Griffin, 1949). The lipophilicity of the molecules described by logP, HLB and
logD5.0 shows the importance in entering in the human and parasitic cells to act.
Moreover, the modifications in the aromatic ring and functional groups change the
number of HBA (number of HBD were not as significant according to PCA) and could
be an additional interaction point with the biological target. Finally, the size of the

ACCEPTED MANUSCRIPT
molecule, represented by the MSA, determines the fit with the biological target and is
also associated with the molecular modifications, especially in the aromatic substituent.
PCA analysis showed that the two main principal components, PC1 and PC2, explains
48.8% and 28.4%, respectively, of the total variance of the model. Table 5 shows the
loadings of each descriptor in the principal components and the loading plot (figure 3)
helps the interpretation of the distribution of the analogues in the scatter plot presented
in Figure 4. The loading values indicate that compounds that have positive scores of
PC1 and PC2 are the ones with the most HBAs and bigger MSA; compounds with
positive PC1 and negative PC2 scores have higher values of HLB; and positive PC2
and negative PC1 scores are the most lipophilic and non-ionizable (neutral). The
scatter plot clearly shows two different clusters of high activity compounds, which are in
the right side of the scatter plot (more positive PC1 scores). This means that more
hydrophilic (loaded by HLB and HBA) and bigger compounds (loaded by HBA and
MSA) are more active than the others. The analogues that present an ionizable basic-
nitrogen atom such as in imine or amine form showed negative values for the PC2,
suggesting that more hydrosoluble molecules (loaded mainly by HLB) should present
higher activity. However, amines (4, 8, 14) have more positive PC1 score and are more
active than the imines (11, 12, 13). Therefore, this shows that the ionization capability
led to higher HLB values for the amines than for the imines, even with similar logP and
logD5.0 values. This defines that the balance between hydrosolubility and lipophilicity
determines the activity of such basic compounds.
Table 5. Loadings of the descriptors in each PC.
Descriptors
logP
PC1
PC2
-0.43572
0.4353
0.27683
0.52756
-0.16998
0.66512
0.41684
HBA
HLB
0.56152
-0.30797
0.45882
logD5.0
MSA
It is interesting to stress that both HBA and MSA present high positive contribution for
both PCs. The neutral high activity analogues 2, 3 and 5 have positive scores for both
PC1 and PC2 and are those with the biggest MSA and HBA values. This suggests that
these analogues are more hydrophilic than the other neutral ones, explaining their
separation from the others by PC1. In addition, this also suggests that among the
neutral functional groups the methoxylated carboxylic acid derivatives (esters and
amides) tend to be more active, since they present higher volume (given by the
aromatic ring substituents which lead to higher MSA) and hydrophilicity (given by the
presence of the oxygen which is an HBA). It can be noted that the phenolic amide
(compound 10) is near to the high activity compounds in the scatter plot. This suggests
that only compounds with high (neutral compounds) or low (amines) PC2 scores
should be more active, and that PC2 scores close to zero (compound 10) indicate low
activity.

ACCEPTED MANUSCRIPT
As discussed before, the presence of a methoxy group in the structure seems to bring
better activities and selectivity. This is corroborated by the higher HBA and MSA values
provided by this group. Possibly, HBA is correlated to the high activity, while MSA is
correlated to high SI. For example, methoxylated compounds 2, 4 and 5 are more
active than their non-substituted analogues 1, 14 and 15, and are less cytotoxic than
their phenolic analogues 9, 8 and 10. Compound 3 present both methoxy and hydroxy
groups, and then similar toxicity to the phenolic ester 9 and high activity is observed. In
summary, the clustering obtained with the PCA can also explain the selectivity provided
by the methoxy group, even in more cytotoxic basic molecules.
Regarding the activity against L. infantum amastigotes, the clustering observed in
Figure 3A can also be applied. The basic compounds 4, 8 and 14 were the compounds
with high activity against L. infantum amastigotes, suggesting that the roles discussed
by this group are the same from the activity against T. cruzi amastigotes. The ionization
capability may also play the role in the accumulation in the parasitophorous vacuoles of
both parasites. However, the neutral compounds (with exception of compound 5) were
not active against L. infantum amastigotes. This can also suggest that ionizable
molecules could act by a different mechanism of action from those neutral. Possibly
these compounds penetrate the host cell in the molecular form, and when they reach
the acidic pH in the parasitophorous vacuole, they became ionized and accumulate in
the amastigote surrounding environment, leading to high activity against this form. This
is not possible with the neutral compounds, explaining why compounds 2 and 3
showed low activity against this parasite. This also corroborates with the previous
discussion about the high activity that such compounds present against amastigotes
and low activity against promastigotes.
The importance of ionization and hydrophilicity can also explain the activity of
compound 5 against L. infantum amastigotes. As can be seen in Table 4, compound 5
is more hydrophilic (represented by logP value) and more ionizable (represented by
logD5.0) than compounds 2 and 3. This states the importance of these characteristics in
the design of novel molecules. For example, amide 5 is 4-fold more potent than its
bioisostere 2, which could be possibly associated to the presence of a nitrogen atom in
the structure, leading to a slightly lower logP and logD5.0 values, and higher
accumulation in the parasitophorous vacuoles.

ACCEPTED MANUSCRIPT
Figure 4. Dendrogram obtained from the HCA suggesting that two different groups (A),
represented by high and low activity compounds in the series are clearly separated.
Further separation (B) into basic-nitrogen containing and neutral analogues is also
observed. The high activity compounds are highlighted inside the dashed rectangle.
The HCA results gave a similar separation as the observed in the PCA. The series (A)
is divided into two different groups (Figure 4): high (on the left) and low (on the right)
activity compounds. The clustering B is separated into molecules containing a basic
atom of nitrogen 4, 8 and 14 (on the left) and neutral molecules 2, 3 and 5 (on the
right). Clustering B can also explain the cytotoxicity and the activity against L. infantum
amastigotes, since compounds 4, 8 and 14 are more cytotoxic and more active against
L. infantum amastigotes than 2, 3 and 5. In summary, it can be observed almost the
same clustering than the obtained in PCA. Since the same results were observed with
different methodologies, the obtained SAR can be considered reliable and not a
product from chance.
3.7.
Conclusion
In summary, the results from the SAR analysis through GE metrics and PCA/HCA give
a clear indication that the methoxylated compounds, especially amine 4 are the most
promising lead compounds in the LINS03 series to design improved compounds
against T. cruzi and L. infantum with lower toxicity to mammalian cells and higher
selectivity towards the parasites.
Abbreviations
CC50: Cytotoxic concentration 50%; DMSO: Dimethyl sulfoxyde; GE: Group efficiency;
HA: Heavy atoms; HBA: hydrogen-bond acceptor; HBD: hydrogen-bond donor; HCA:
Hierarchical cluster analysis; HLB: Hydrophilic-lipophilic balance; HTS: High throughput
screening; IC50: Inhibitory concentration 50%; LIA: Leishmania infantum amastigotes;
LIP: Leishmania infantum promastigotes; logD: distribution coefficient in the given pH;

ACCEPTED MANUSCRIPT
logP: partition coefficient n-octanol:water; logS7.4: water solubility in pH 7.4; LRMS, low
resolution mass spectroscopy; MSA: molecular surface area; MTT: (3-(4,5-
Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; MW: molecular mass; PC:
Principal component; PCA: Principal component analysis; RB: Rotatable bonds; RoF:
Rule of five; SAR: Structure-activity relationship; SI: Selectivity index; TCA:
Trypanosoma cruzi amastigotes; TCT: Trypanosoma cruzi trypomastigotes; TEA:
N,N,N-Triethylamine.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgements
The authors are grateful to Fundação de Amparo à Pesquisa do Estado de São Paulo -
FAPESP [Grants 2016/25028-3; 2016/00195-4; 2015/23403-9] and to Conselho
Nacional de Desenvolvimento Científico e Tecnológico - CNPq [Grants 455411/2014-0;
457099/2014-3] for the financial support to this work.
References
Ali, M. A.; Siddiki, S. M. A. H.; Kon, K.; Shimizu, K., 2014. Fe3+-exchanged clay
catalyzed transamidation of amides with amines under solvent-free condition.
Tetrahedron Lett., 55, 1316-1319.
Atanasov, A. G.; Waltenberger, B.; Pferschy-Wensig, E.; Linder, T.; Wawrosch, C.;
Uhrin, P.; Temml, V.; Wang, L.; Schwaiger, S.; Heiss, E. H.; Rollinger, J. M.; Schuster,
D.; Breuss, J. M.;Bochkov, V.; Mihovilovic, M. D.; Kopp, B.; Bauer, R.; Dirsch, V. M.;
Stuppner, H., 2015. Discovery and resupply of pharmacologically active plantderived
natural products: A review. Biotechnol. Adv., 33, 1582-1614.
Barry, J.; Bram, G.; Decodts, G.; Loupy, A.; Orange, C.; Petit, A.; Sansoulet, J., 1985.
Solid-liquid phase-transfer catalysts without added solvent. A simple, efficient, and
inexpensive synthesis of aromatic carboxylic esters by alkylation of potassium
carboxylates. Synthesis, 1985, 40-45.
Bartzatt, R., 2017. Relative Properties and Relative Potency of Various Hydrazide
Compounds That Inhibit Growth of Mycobacterium. IOSR J. Pharm. Biol. Sci., 12, 144-
150.
Castillo, J. C.; Orrego-Hernández, J.; Portilla, J., 2016. Cs2CO3‐Promoted
Direct N‐Alkylation: Highly Chemoselective Synthesis of N‐Alkylated Benzylamines and
Anilines. Eur. J. Org. Chem., 2016, 3824-3835.
ChemAxon Inc., MarvinBeans 16.8.1.0, 2016. http://www.chemaxon.com.

ACCEPTED MANUSCRIPT
De Oliveira, A.; Mesquita, J. T.; Tempone, A. G.; Lago, J. H. G.; Guimarães, E. F.;
Kato, M. J., 2012. Leishmanicidal activity of an alkenylphenol from Piper
malacophyllum is related to plasma membrane disruption. Exp. Parasitol., 132, 383-
387.
Ferreira, J. E. V.; Costa, C. H. S.; Miranda, R. M.; Figueiredo, A. F.; Ginarte, Y. M. A.,
2014. An interdisciplinary study on monosubstituted benzene involving computation,
statistics and chemistry. Educ. Quím., 25, 418-424.
Giménez, B. G.; Santos, M. S.; Ferrarini, M.; Fernandes, J. P. S., 2010. Evaluation of
blockbuster drugs under the Rule-of-five. Pharmazie, 65, 148-152.
Griffin, W. C., 1949. Classification of Surface-Active Agents by 'HLB’. J. Soc. Cosmet.
Chem., 1, 311-326.
Hammer, Ø.; Harper, D. A. T.; Ryan, P. D., 2001. PAST: Paleontological statistics
software package for education and data analysis. Paleontological Electronica, 4, 1-9.
Harvey, A. L.; Edrada-Ebel, R.; Quinn, R. J., 2015. The re-emergence of natural
products for drug discovery in the genomics era. Nat. Rev. Drug Dis., 14, 111-119.
Hopkins, A. L.; Keserü, G. M.; Leeson, P. D.; Rees, D. C.; Reynolds, C. H., 2014. The
role of ligand efficiency metrics in drug discovery. Nat. Rev. Drug Discov., 13, 105-121.
Katsuno, K.; Burrows, J. N.; Duncan, K.; van Huijsduijnen, R. H.; Kaneko, T.; Kita, K.;
Mowbray, C. E.; Schmatz, D.; Warner, P.; Slingsby, B. T., 2015. Hit and lead criteria in
drug discovery for infectious diseases of the developing world. Nat. Rev. Drug Discov.,
14, 751-758.
Kubinyi, H., 1993. QSAR: Hansch analysis and related approaches. VCH, Weinheim.
Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J., 2001. Experimental and
computational approaches to estimate solubility and permeability in drug discovery and
development settings. Adv. Drug Deliv. Rev., 46, 3-26.
Luque-Ortega, J. R.; Rivas, L., 2010. Characterization of the leishmanicidal activity of
antimicrobial peptides. Methods Mol. Biol., 618, 393-420.
Martins, L. F.; Mesquita, J. T.; Pinto, E. G.; Costa-Silva, T. A.; Borborema, S. E. T.;
Galisteo Junior A. J.; Neves, B. J.; Andrade, C. H.; Al Shuhaib, Z.; Bennett, E. L.;
Black, G. P.; Harper, P. M.; Evans, D. M.; Fituri, H. S.; Leyland, J. P.; Martin, C.;
Roberts, T. D.; Thornhill, A. J.; Vale, S. A.; Howard-Jones, A.; Thomas, D. A.; Williams,
H. L.; Overman, L. E.; Berlinck, R. G. S.; Murphy, P. J.; Tempone, A. G., 2016.
Analogues of marine guanidine alkaloids are in vitro effective against Trypanosoma
cruzi and selectively eliminate Leishmania (L.) infantum intracellular amastigotes. J.
Nat. Prod., 79, 2202-2210.
Mcgwire, B. S.; Satoskar, A. R., 2014. Leishmaniasis: clinical syndromes and
treatment. Q. J. Med., 107, 7-14.
Molinspiration Cheminformatics, Calculation of molecular properties and bioactivity
scores, 2015. http://www.molinspiration.com/cgi-bin/properties.

ACCEPTED MANUSCRIPT
Nunes, M. C. P.; Dones, W.; Morillo, C. A.; Encina, J. J.; Ribeiro, A. L., 2013. Chagas
Disease An Overview of Clinical and Epidemiological Aspects. J. Am. Col. Cardiol., 62,
767-776.
Pearl, I. A.; Beyer, D. L., 1951. Reactions of vanillin and its derived compounds. XIII. A
new synthesis of vanillic acid esters. J. Am. Chem. Soc., 73, 4091-4093.
Reimão, J. Q.; Colombo, F. A.; Pereira-Chioccola, V. L.; Tempone, A. G., 2012.
Effectiveness of liposomal buparvaquone in an experimental hamster model of
Leishmania (L.) infantum chagasi. Exp. Parasitol., 130, 195-199.
Reimão, J. Q.; Migotto, A. E.; Kossuga, M. H.; Berlinck, R. G.; Tempone, A. G., 2008.
Antiprotozoan activity of Brazilian marine cnidarian extracts and of a modified steroid
from the octocoral Carijoa riisei. Parasitol. Res., 103, 1445-1450.
Schultes, S.; Graaf, C.; Haaksma, E. E. J.; Esch, I. J. P.; Leurs, R.; Kramer, O., 2010.
Ligand efficiency as a guide in fragment hit selection and optimization. Drug Discov.
Today Technol., 7, e147-202.
Trouiller, P.; Olliaro, P.; Torreele, E.; Orbisnki, J.; Laing, R.; Ford, N., 2002. Drug
development for neglected diseases: a deficient market and a public-health policy
failure. Lancet, 359, 2188-2194.
Utzinger, J.; Becker, S. L.; Knopp, S.; Blum, J.; Neumayr, A. L.; Keiser, J.; Hatz, C. F.,
2012. Neglected tropical diseases: diagnosis, clinical management, treatment and
control. Swiss Med Wkly., 142, w13727.
Van den Berg, R. A.; Hoefsloot, H. C. J.; Westerhuis, J. A.; Smilde, A. K.; van der Werf,
M. J., 2006. Centering, scaling, and transformations: improving the biological
information content of metabolomics data. BMC Genomics, 7, 142.
Varela, M. T.; Dias, R. Z.; Martins, L. F.; Ferreira, D. D.; Tempone, A. G.; Ueno, A. K.;
Lago, J. H. G.; Fernandes, J. P. S., 2016. Gibbilimbol analogues as antiparasitic
agents: Synthesis and biological activity against Trypanosoma cruzi and Leishmania
(L.) infantum. Bioorg. Med. Chem. Lett., 26, 1180-1183.
Varela, M. T.; Lima, M. L.; Galuppo, M. K.; Tempone, A. G.; Oliveira, A. de; Lago, J. H.
G.; Fernandes, J. P. S., 2017. New alkenyl derivative from Piper malacophyllum and
analogues: Antiparasitic activity against Trypanosoma cruzi and Leishmania infantum.
Chem. Biol. Drug. Des., 90, 1007-1011.
Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.,
2002. Molecular properties that influence the oral bioavailability of drug candidates. J.
Med. Chem., 45, 2615-2623.
Vendrame, R.; Ferreira, M. M. C.; Collins, C. H.; Takahata, Y., 2002. Structure-activity
relationships (SAR) of contraceptive progestogens studied with four different methods
using calculated physicochemical parameters. J. Mol. Graph. Model., 20, 345-358.
Verdonk, M. L.; Rees, D. C., 2008. Group Efficiency: A Guideline for Hits-to-Leads
Chemistry. ChemMedChem, 3, 1179-1180.

ACCEPTED MANUSCRIPT
Ward Jr, J. H., 1963. Hierarchical grouping to optimize an objective function. J. Am.
Stat. Assoc., 58, 236-244.
World
Health
Organization.
Neglected
Tropical
Diseases,
2017.
http://www.who.int/neglected_diseases/diseases/en/ (Accessed November 26, 2017).

ACCEPTED MANUSCRIPT
Antiparasitic Activity of New Gibbilimbol Analogues and SAR Analysis Through
Efficiency and Statistical Methods
Marina T. Varela,1 Maiara M. Romaneli,2 Marta L. Lima,2,3 Samanta E. T. Borborema,2
Andre G. Tempone,2 and João P. S. Fernandes*,1
1Departamento de Ciências Farmacêuticas, Universidade Federal de São Paulo, Rua
São Nicolau 210, 09913-030 Diadema-SP, Brazil
2Centro de Parasitologia e Micologia, Instituto Adolfo Lutz, Av. Dr. Arnaldo 355, 01246-
000 São Paulo-SP, Brazil
3Instituto de Medicina Tropical de São Paulo, Universidade de São Paulo, Av. Dr.
Enéas Carvalho de Aguiar 470, 05403-000 São Paulo-SP, Brazil
GRAPHICAL ABSTRACT