New Pd-Fe ferrocenyl antiparasitic compounds with bioactive 8-hydroxyquinoline ligands: a comparative study with their Pt-Fe analogues

Article abstract of DOI:10.1039/d0dt03963b

In the search for a more effective chemotherapy for the treatment of Human African Trypanosomiasis, a disease caused by the parasiteTrypanosoma brucei, the development of ferrocenyl compounds has arisen as a promising strategy. In this work, five new Pd-Fe heterobimetallic [Pd^II(L)(dppf)](PF₆) compounds, including 8-hydroxyquinolyl derivatives HL1-HL5 as bioactive ligands and dppf = 1,1′-bis(diphenylphosphino)ferrocene as the organometallic co-ligand, were synthesized and fully characterized in the solid state and in solution. Molecular structures of three compounds were solved by single crystal X-ray diffraction methods. The compounds displayed submicromolar or micromolar IC₅₀values against bloodstreamT. brucei(IC₅₀: 0.33-1.2 μM), and good selectivity towards the pathogen (SI: 4-102) with respect to mammalian macrophages (cell line J774). The new Pd complexes proved to be 2-fold to 45-fold more potent than the drug nifurtimox but most of them are less active than their Pt analogues. Potential molecular targets were studied. The complexes interact with DNA but they do not alter the intracellular thiol-redox homeostasis of the parasite. In order to understand and predict the main structural determinants on the anti-T. bruceiactivity, a search of quantitative structure-activity relationships (QSAR) was performed including all the [M(L)(dppf)](PF₆) complexes, where M = Pd(ii) or Pt(ii), currently and previously developed by us. The correlation obtained shows the relevance of the electronic effects, the lipophilicity and the type of metal. According to the QSAR study, compounds with electron-withdrawing ligands, higher lipophilicity and harboring Pt would result in higherT. bruceicytotoxicity. From the whole series of [M(L)(dppf)](PF₆) compounds developed, where M = Pt(ii) or Pd(ii) and HL = 8-hydroxyquinolyl derivatives, Pt-dppf-L4 (IC₅₀= 0.14 μM, SI = 48) was selected to perform an exploratory pre-clinical study in infected mice. This hit compound lacks acute toxicity when applied to animals in the dose/regimen described and exerts an anti-proliferative effect on parasites, which extends animal survival but is not curative.

Full text of DOI:10.1039/d0dt03963b

Dalton
Transactions
View Article Online
View Journal | View Issue
PAPER
New Pd–Fe ferrocenyl antiparasitic compounds
with bioactive 8-hydroxyquinoline ligands: a
comparative study with their Pt–Fe analogues†
Cite this: Dalton Trans., 2021, 50,
1651
b
Feriannys Rivas, a Andrea Medeiros, b,c Cristina Quiroga, b Diego Benítez,
d
Marcelo Comini, b Esteban Rodríguez-Arce, a Ignacio Machado,
Hugo Cerecetto e and Dinorah Gambino
*
a
In the search for a more effective chemotherapy for the treatment of Human African Trypanosomiasis, a
disease caused by the parasite Trypanosoma brucei, the development of ferrocenyl compounds has
arisen as a promising strategy. In this work, five new Pd–Fe heterobimetallic [PdII(L)(dppf)](PF6) com-
pounds, including 8-hydroxyquinolyl derivatives HL1–HL5 as bioactive ligands and dppf = 1,1’-bis(diphenyl-
phosphino)ferrocene as the organometallic co-ligand, were synthesized and fully characterized in the
solid state and in solution. Molecular structures of three compounds were solved by single crystal X-ray
diffraction methods. The compounds displayed submicromolar or micromolar IC50 values against blood-
stream T. brucei (IC50: 0.33–1.2 μM), and good selectivity towards the pathogen (SI: 4–102) with respect
to mammalian macrophages (cell line J774). The new Pd complexes proved to be 2-fold to 45-fold more
potent than the drug nifurtimox but most of them are less active than their Pt analogues. Potential mole-
cular targets were studied. The complexes interact with DNA but they do not alter the intracellular thiol-
redox homeostasis of the parasite. In order to understand and predict the main structural determinants
on the anti-T. brucei activity, a search of quantitative structure–activity relationships (QSAR) was per-
formed including all the [M(L)(dppf)](PF6) complexes, where M = Pd(II) or Pt(II), currently and previously
developed by us. The correlation obtained shows the relevance of the electronic effects, the lipophilicity
and the type of metal. According to the QSAR study, compounds with electron-withdrawing ligands,
higher lipophilicity and harboring Pt would result in higher T. brucei cytotoxicity. From the whole series of
[M(L)(dppf)](PF6) compounds developed, where M = Pt(II) or Pd(II) and HL = 8-hydroxyquinolyl derivatives,
Pt-dppf-L4 (IC50 = 0.14 µM, SI = 48) was selected to perform an exploratory pre-clinical study in infected
mice. This hit compound lacks acute toxicity when applied to animals in the dose/regimen described and
exerts an anti-proliferative effect on parasites, which extends animal survival but is not curative.
Received 18th November 2020,
Accepted 24th December 2020
DOI: 10.1039/d0dt03963b
rsc.li/dalton
Introduction
The diseases produced by trypanosomatid parasites, such as
American trypanosomiasis (Chagas disease), human African
aÁrea Química Inorgánica, Programa de Posgrados, Facultad de Química,
Universidad de la República, Gral. Flores 2124, 11800 Montevideo, Uruguay.
E-mail: dgambino@fq.edu.uy; Fax: +5982-9241906; Tel: +5982-9249739
bGroup Redox Biology of Trypanosomes, Institut Pasteur de Montevideo, Mataojo
2020, CP 11400, Montevideo, Uruguay
trypanosomiasis (sleeping sickness, HAT) and Leishmaniasis,
constitute an urgent health problem gathered in the poorest
regions of the world.1–3 HAT is a resurgent endemic disease in
the sub-Saharan regions of Africa affecting mainly people
living in rural areas. The disease is caused by two subspecies
of the protozoa Trypanosoma brucei (T. brucei), Trypanosoma
brucei gambiense and Trypanosoma brucei rhodesiense. The
parasite is transmitted by the bite of a tsetse fly. Ninety
percent of the cases correspond to chronic-like infections
caused by T. b. gambiense whereas T. b. rhodesiense is respon-
sible for the most virulent form of the disease, both being
fatal if left untreated.
cDepartamento de Bioquímica, Facultad de Medicina, Universidad de la República,
Gral. Flores 2125, 11800 Montevideo, Uruguay
dÁrea Química Analítica, Facultad de Química, Universidad de la República, Gral.
Flores 2124, 11800 Montevideo, Uruguay
eÁrea de Radiofarmacia, CIN and Grupo de Química Orgánica Medicinal,
IQB-Facultad de Ciencias, Universidad de la República, Montevideo, Uruguay
†Electronic supplementary information (ESI) available. CCDC 1963631, 1963632
and 1963633. For ESI and crystallographic data in CIF or other electronic format
see DOI: 10.1039/d0dt03963b
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1651

View Article Online
Paper
Dalton Transactions
All the diseases caused by trypanosomatids are classified as pounds bearing improved biological properties and multiple
neglected tropical diseases (NTDs) by the World Health parasite target mechanisms of action. Using this strategy,
Organization (WHO) due to low historical interest by the several series of structurally related compounds have been
pharmaceutical industry. With the exception of eflornithine developed and their physicochemical and biological properties
and fexinidazole,4,5 current chemotherapy still relies on quite and probable mechanism of action have been studied in
old and toxic drugs that require quite long and complicated depth. Many of these molecules represent new hit or lead com-
therapeutic regimes. Additionally, the development of drug re- pounds suitable for further drug developments. This approach
sistance has been observed.6–8
is in line with the current interest of medicinal inorganic
Medicinal inorganic chemistry has emerged as an interest- chemistry in enhancing the activity of drugs by conjugation to
ing alternative for the development of novel metal-based drug organometallic fragments.27
candidates against NTDs. Initially, drug development was cen-
Following this strategy, four series of ferrocenyl heterobime-
tered on classic coordination compounds but currently the tallic Pd–Fe and Pt–Fe derivatives, showing activity on trypano-
design of prospective organometallic drugs is showing promis- somatid parasites, were recently developed. These series
ing results.9–19 In this regard, some research groups are cur- include a bidentate bioactive ligand belonging to one of
rently focused on the development of ferrocenyl derivatives as four different families of structurally related ligands.16–19
antiparasitic agents.11,12,16–21 Ferrocene is a low size organo- Instead of including the ferrocene moiety in the structure of
metallic fragment with a “sandwich” type structure that shows an organic drug, as usually done by other research groups, the
high lipophilicity, conferring a good capacity to penetrate cell new Pd–Fe and Pt–Fe compounds were rationally designed
membranes. From a chemical point of view, it is stable in air including the ferrocene fragment as part of the coligand 1,1′-
and in aqueous solution and a variety of derivatives can be bis(diphenylphosphino)ferrocene (dppf). Overall, these pre-
obtained. These characteristics make it very attractive for the vious studies demonstrated that the inclusion of the ferrocenyl
rational design of medicinal chemistry compounds.22–24 fragment as the dppf coligand improves their biological
Ferroquine, the ferrocene analogue of the antimalarial drug properties.16–19
chloroquine, represents the first example of an antiparasitic
ferrocene-based molecule that is currently undergoing a phase (dppf)](PF6) compounds, where HL = HL1–HL5 are 8-hydroxy-
II clinical trial.25,26
quinoline derivatives shown in Fig. 1. The compounds dis-
In particular, one of these four families includes five [PtII(L)
Taking these pieces of evidence into account, our group is played good activity against bloodstream T. brucei, with IC50
currently focused on the rational design of ferrocenyl deriva- values in the submicromolar range (IC50: 0.14–0.93 μM), good
tives as prospective agents against trypanosomatid parasites. selectivity towards the pathogen and adequate lipophilicity.19
In the last two decades, we contributed in demonstrating that
The current work describes the synthesis, characterization,
the strategy of hybridization, in a single molecule, of bioactive and biological profiling against T. brucei of the Pd analogues
moieties, “metal ion or organometallic core + bioactive organic [PdII(L)(dppf)](PF6) (Fig. 1). Furthermore, a preliminary evalu-
ligand”, leads in many cases to antiparasitic metal-based com- ation of the therapeutic potential of one of the most potent
Fig. 1 Structure of the 8-hydroxyquinoline derivatives, HL, and of the new heterobimetallic 1,1’-bis(dipheny1phosphino)ferrocene (dppf) derivatives,
[PdII(L)(dppf)](PF6).
1652 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
and selective candidates from the Pt–Pd series was conducted C, 54.71; H, 3.91; N, 1.55. FTIR (KBr/cm−1): 1501 (vs) ν(CvN),
in a murine infection model of acute African trypanosomiasis. 1316 (m) ν(C–N), 1110 (m) ν(C–O).
DNA and parasite redox homeostasis were investigated as
[PdII(L2)(dppf)](PF6), Pd-dppf-L2. Yield: 24 mg, 41%. Anal.
potential molecular targets of the compounds. Having a sig- calc. for C43H33F6FeN2O3P3Pd: C, 51.91; H, 3.34; N, 2.82.
nificant number of structurally related [M(L)(dppf)](PF6), Found: C, 51.33; H, 3.50; N, 2.61. FTIR (KBr/cm−1): 1508(s)/
where M = Pt(II) or Pd(II) showing activity on T. brucei and 1291 (vs) ν(N–O)as/s, 1508(s) ν(CvN), 1311 (vw) ν(C–N), 1113
including three different families of bioactive ligands HL,17–19 (vw) ν(C–O).
a QSAR study was also performed. The effect of lipophilicity,
[PdII(L3)(dppf)](PF6), Pd-dppf-L3. Yield: 22 mg, 37%. Anal.
electronic effect and nature of the metal center were analyzed. calc. for C43H32Cl2F6FeNOP3Pd: C, 50.69; H, 3.17; N, 1.37.
These types of studies, although relevant to unraveling the Found: C, 50.42; H, 3.23; N, 1.36.
structural factors affecting biological activity and to guide the
rational design of bioactive compounds, are not commonly (w) ν(C–O).
FTIR (KBr/cm−1): 1492 (s) ν(CvN), 1286 (vw) ν(C–N), 1116
performed in inorganic medicinal chemistry due to the lack of
a large series of really structurally related compounds.28–31
[PdII(L4)(dppf)](PF6), Pd-dppf-L4. Yield: 20 mg, 30%. Anal.
calc. for C43H32ClF6FeINOP3Pd: C, 46.52; H, 2.91; N, 1.26.
Found: C, 46.30; H, 2.94; N, 1.24.
FTIR (KBr/cm−1): 1486 (m) ν(CvN), 1311 (vw) ν(C–N), 1098
(m) ν(C–O).
Experimental section
General considerations
[PdII(L5)(dppf)](PF6), Pd-dppf-L5. Yield: 18 mg, 26%. Anal.
calc. for C43H32F6FeI2NOP3Pd: C, 42.98; H, 2.68; N, 1.17.
Found: C, 43.04; H, 2.77; N, 1.20.
All common laboratory chemicals and 8-hydroxyquinoline
derivatives were purchased from commercial sources and used
without further purification. The [PdCl2(dppf)]·CHCl3 precursor
was synthesized according to a previously reported procedure by
heating for 30 min an equimolar mixture of [PdCl2(dmso)2] and
dppf in CHCl3.32 The [PtII(L4)(dppf)](PF6) compound was syn-
thesized and characterized as previously reported.19
FTIR (KBr/cm−1): 1480 (m) ν(CvN), 1307 (vw) ν(C–N), 1096
(m) ν(C–O).
X-ray diffraction study of [PdII(L)(dppf)](PF6) compounds
Single crystals suitable for structural X-ray diffraction studies
of three of the compounds, namely [PdII(L1)(dppf)](PF6),
[PdII(L3)(dppf)](PF6) and [PdII(L4)(dppf)](PF6), were obtained
by slow diffusion of hexane into dichloromethane solutions of
the complexes. Measurements were performed on a Bruker D8
Venture diffractometer with multilayer mirror monochromated
MoKα (λ = 0.71073 Å) radiation for the [PdII(L1)(dppf)](PF6)
compound and monochromated CuKα (λ = 1.54178 Å) radi-
ation for [PdII(L3)(dppf)](PF6) and [PdII(L4)(dppf)](PF6) com-
pounds. X-ray diffraction intensities were collected with an
APEX 2 v2014.5-0 (Bruker AXS) and integrated with an APEX3
v2017.3-0 (Bruker AXS). These intensities were scaled and the
data were corrected empirically for absorption (employing the
multi-scan method) with the SADABS v2016/2 (SADABS-2016/
2-Bruker AXS) program.33 The structures were solved by a
hybrid method with SHELXT and the molecular model devel-
Chemical and physical measurements
C, H and N analyses were performed with a Thermo Scientific
Flash 2000 elemental analyzer. FTIR spectra (4000–400 cm−1) of
the complexes and the 8-hydroxyquinoline ligands were
measured using KBr pellets with a Shimadzu IR Prestige-21
instrument. 1H NMR spectra were recorded in DMSO-d6 at
25 °C on a Bruker Avance NEO DPX-400 instrument at 400 MHz.
Homo-nuclear (COSY) correlation experiments were carried out
with the same instrument. Tetramethylsilane was used as the
internal standard and chemical shifts are reported in ppm.
Synthesis of [PdII(L)(dppf)](PF6) compounds, HL = HL1–HL5
The [PdII(L)(dppf)](PF6) derivatives were synthesized by the fol- oped by alternated cycles of Fourier methods and full-matrix
lowing procedure: 50 mg of the [PdIICl2(dppf)]·CHCl3 least-squares refinement with Olex2.34 The crystal structure of
(0.059 mmol) precursor were suspended in 5 mL of methanol. [PdII(L1)(dppf)](PF6) contains solvent accessible voids of
An equimolar amount of ligand suspended in 5 mL of metha- 521.6 Å3 filled with disordered solvent molecules. Their contri-
nol including 20 µL of triethylamine (0.142 mmol) was added. bution to the structure factor was secured by back-Fourier
The mixture was kept at room temperature under stirring for transformation using the SQUEEZE routine of OLEX 2 result-
4–5 h for [PdII(L4)(dppf)](PF6), under stirring for 15–24 h for ing in 132.3 electrons per unit cell in the complex. The hydro-
[PdII(L1)(dppf)](PF6), [PdII(L3)(dppf)](PF6) and [PdII(L5)(dppf)] gen atoms were positioned on a stereo-chemical basis and
(PF6) and under reflux for 5 h for [PdII(L2)(dppf)](PF6). The refined with the riding model. Crystal data, data collection pro-
solution was evaporated up to 5 mL and filtered. NaPF6 in cedure, structure determination methods and refinement
excess (30 mg, 0.177 mmol) was added to the solution and the results are summarized in Table S1.†
mixture was stirred for 15 min. The compounds were isolated
by filtration as brown-red solids and were recrystallized from a
Biological studies
dichloromethane solution by diffusion with hexane.
Stability study of [M(L4)(dppf)](PF6) compounds, M = Pd(II)
[PdII(L1)(dppf)](PF6), Pd-dppf-L1. Yield: 21 mg, 35%. Anal. and Pt(II). To get an initial insight into the stability of the com-
calc. for C43H34F6FeNOP3Pd: C, 54.37; H, 3.61; N, 1.47. Found: plexes in the culture medium (pH 7.4) used for the evaluation
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1653

View Article Online
Paper
Dalton Transactions
against T. brucei, the 31P-NMR spectra of [Pt(L4)(dppf)](PF6), reference value of parasites cultured in the presence of DMSO
the most active Pt compound selected to perform an in vivo 1% (v/v). The data were processed and analyzed with the
study, were measured during 48 h at room temperature in a Accuri C6 software.
mixture of DMSO-10 mM Tris-HCl buffer solution pH 7.4 (3 : 1)
at 162 MHz.
Cytotoxicity assays on murine macrophages. Mouse macro-
phages from the cell line J774 (ATCC® TIB-67™) were culti-
Furthermore, the chemical stability of the compounds vated under a humidified 5% CO2/95% air atmosphere at
[M(L4)(dppf)](PF6) (M-dppf-L4), where M = Pd(II) and Pt(II) was 37 °C in Dulbecco’s modified Eagle’s medium (DMEM) sup-
studied in solution, using as solvents DMSO and plemented with 10% (v/v) FBS, 10 U mL−1 penicillin and 10 μg
DMSO-10 mM Tris-HCl buffer solution pH 7.4 (4 : 1), by the mL−1 streptomycin. The experimental protocol for the determi-
HPLC-DAD technique. Ligand HL4, dppf and precursors were nation of IC50 values was essentially the same as that pre-
evaluated under the same conditions. For this purpose, a viously described, except that 200 μL of a cell suspension at 6 ×
Shimadzu Prominence LC-20AT liquid chromatograph was 104 cells per mL was added per well in a 96-well culture
used, coupled to a Shimadzu asSPD-M20A diode array detec- plate.37 The cytotoxicity of the compounds was evaluated at
tor. An Agilent ZORBAX Eclipse Plus C18 column of 4.6 × five different concentrations (0.01–100 μM) in triplicate and
100 mm and 5 µm, thermostated at 25 °C, was used as the using the WST-1 reagent (Roche) (2-(4-Iodophenyl)-3-(4-nitro-
stationary phase. A 40 mM phosphate buffer solution pH 7.4 phenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium sodium salt). The
(A) and MeOH (B) were used as mobile phases at a constant control treatment included cells cultured in the presence of
flow of 1 mL min−1, according to the gradient shown in DMSO 1% (v/v). Absorbance at 450 nm, corresponding to the
Table S2.†
formazan dye produced by metabolically active cells, was
For the preparation of the solutions, 2.5 mg of the com- measured with an EL 800 microplate reader. IC50 values were
pound were dissolved in 2.5 mL of solvent. An injection obtained from drug-response curves as described above for the
volume per run of 50 μL was established. The wavelength of T. brucei assay and the associated errors are expressed as S.D.
measurement was 280 nm, obtaining a retention time (RT) of The software OriginPro 8.6 was used for data analysis.
approximately 24 min for the complexes in both solvents.
Therapeutic efficacy study in a mouse infection model of
Stability was studied in a period of one week, the solutions [PtII(L4)(dppf)](PF6). The bloodstream form of T. b. brucei
being measured at t = 0, t = 1 day and t = 7 days, under the expressing a red-shifted luciferase (T. brucei LUC) was main-
same conditions.35
tained in the exponential growth phase and tested for luci-
Viability and redox assays for bloodstream T. b. brucei. The ferase activity, before being harvested for animal experiments,
anti-trypanosomal and redox activity of the compounds was as previously described.38
evaluated against bloodstream T. b. brucei (monomorphic
The experimental procedures described below were
strain 427, cell line 449-hGrxroGFP2) as previously described.36 approved by the Institut Pasteur de Montevideo Animal Care
For assessing cell viability, 200 µL of a cell suspension contain- Committee (Protocol 011-18) and complied with the National
ing 5 × 105 parasites per mL in the exponential growth phase law for Laboratory Animal Experimentation (Law nr. 18.611)
were seeded per well in a 96-well culture plate and 2 µL of each and the international FELASA guidelines for the use of labora-
compound prepared at different concentrations in 100% v/v tory animals. Female Balb/cJ mice (6–8 weeks-old) were housed
DMSO were added. For the screening and IC50 assays, final on individual ventilated cages in an isolator system.
concentrations of 5 µM and between 0.001 and 50 µM, respect-
In order to establish the dosing regimen of the candidate
ively, were tested and for all conditions the concentration of compound, the acute toxicity of [PtII(L4)(dppf)](PF6), code Pt-
DMSO was ≤1% (v/v). Treatment with nifurtimox (10–15 µM) dppf-L4, was tested in non-infected animals that were adminis-
was included as the positive drug control. Parasite viability tered with different concentrations of the compound on a
after 24 h incubation was assayed from triplicate biological daily basis.
samples, using propidium iodide (PI, final concentration 2 μg
mL−1 BD Accuri C6 (BD were infected with a single intraperitoneal injection of 1 × 103
as the exclusion dye and
Biosciences) or CyAn™ADP (DakoCytomation) flow cytometer bloodstream parasites suspended in 0.3 mL of fresh culture
with settings previously described.36,37
medium and drug treatment was started 48 hours later. The
For the therapeutic efficacy studies, three mice per group
)
a
The IC50 values were obtained from dose–response curves, following groups were studied: (A) treatment with Pt-dppf-L4
fitted to a sigmoidal Boltzmann equation (errors calculated 10.5 mg kg−1 day−1 during 4 days, (B) treatment with a single
using errors’ propagation) or extrapolated from non-linear dose of the drug diminazene aceturate (DAC) at 40 mg kg−1
fitting plots. The errors are expressed as one standard devi- and (C) treatment with vehicle alone (DMSO) 2.75 g kg−1 day−1
ation (S.D., estimated as σn−1). The software OriginPro 8.6 was during 4 days.
used for data analysis.
At different time points post-infection, animals were injected
The fluorescence intensity of the redox biosensor was intraperitoneally with 200 µL of D-luciferin (15 mg mL−1 in
measured in the same samples used for determining cell viabi- PBS), anesthetized with 2.5% isoflurane and imaged for bio-
lity using an BD Accuri™ C6 flow cytometer with the following luminescence using an In vivo Xtreme II equipment (Bruker)
laser/filter pairs: Ex 488 nm/Em 530 nm. The mean fluo- and a cycle of 30 s shutter on and 30 s shutter off for a total of
rescence of the samples was normalized to the fluorescence 10 min. The images were analyzed using FIJI free-software.39
1654 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
For all assays described above, mice’ health status was carried out in 10 mM Tris-HCl buffer solution at pH 7.4.
monitored daily and animals below the healthy cut-off score Millipore® water was used for the preparation of all aqueous
were euthanized for ethical reasons. The long-rank test was solutions. Fluorescence measurements were carried out on
used for statistical comparison of survival curves.
individually prepared samples to ensure the same pre-incu-
Lipophilicity studies. Reversed-phase TLC experiments were bation time for all samples in each assay. Due to the low solu-
performed on precoated TLC plates SIL RP-18 W/UV254 and bility of the complexes in aqueous media, DMSO was used to
eluted with MeOH : DMF : H2O (65 : 5 : 30 v/v/v). Stock solu- prepare concentrated stock solutions followed by appropriate
tions were prepared in pure dichloromethane (Aldrich) prior to dilution to obtain the targeted concentration and the same
use. The plates were developed in a closed chromatographic content of DMSO (5% v/v) in the final samples. DNA stock
tank and dried and the spots were detected under UV light. solutions were prepared by hydrating ctDNA in 10 mM Tris-
The Rf values were averaged from two to three determinations, HCl buffer solution (1 mg mL−1, ∼2 mM nuc−1) by gently stir-
and converted to RM via the relationship: RM = log10[(1/Rf) − ring the solution during 2–3 days. This solution was stored at
1].40,41
Quantitative structure–activity relationship studies (QSAR)
4 °C (in the fridge) in-between measurements. The concen-
tration of the stock solution was determined by UV spectro-
Data set. The data set used for the QSAR studies contains photometry using the molar absorption coefficient ε (260 nm)
eighteen [M(L)(dppf)](PF6) compounds, where M = Pt(II) or Pd = 6600 M−1 cm−1 nuc−1 42
(II), HL = 8-hydroxyquinolyl-, tropolonyl-, hinokitiolyl-, and An EB 5 mM solution was prepared in 10 mM Tris-HCl
5-nitrofurylthiosemicarbazonyl-ligands (Fig. 2) with different buffer solution. ctDNA was pre-incubated with EB at 4 °C for
profiles of anti-T. brucei activities.17–19
24 h. Samples were prepared with a total concentration of DNA
.
The dependent variable was the log10 of the IC50 values and of EB of 20 μM nuc−1 and 10 μM, respectively, varying the
against T. brucei, log10(IC50). Experimental physicochemical total complex concentration from 15 to 110 μM. They were
descriptors, related to lipophilicity (RM, see above), electronic incubated at 37 °C for 30 min. Samples with complex alone,
(1H NMR chemical shifts) and steric/topological properties samples with complex and DNA and samples with complex
(indicator variable), were used for the analyses as independent and EB but no DNA were used as blanks. Fluorescence spectra
variables. The indicator variable, IVPd, that denotes the pres- were recorded from 540 nm to 750 nm at an excitation wave-
ence of Pd in some complexes, was defined as the categorical length of 510 nm on an Infinite M200 PRO TECAN plate
descriptor taking value 1 when Pd is present or value 0 when spectrofluorimeter.
Pt is present.
The relationships between the selected descriptors and the
activities were analyzed by the use of simple and multiple
Results and discussion
linear regressions. In the equations, n represents the number
2
of data points, radj is the adjusted correlation coefficient, p is Five new Pd(II) organometallic heterobimetallic compounds
the p-value, and the F value is related to the F-statistic analysis with 8-hydroxyquinoline derivatives [PdII(L)(dppf)](PF6), where
(Fischer test).
HL = HL1–HL5 (Fig. 1), were obtained with high purities and
DNA interaction by fluorescence studies. Experiments for good yields and they were exhaustively characterized in the
competitive binding to calf thymus DNA (ctDNA, SIGMA, Type solid state and in solution. Elemental analyses agreed with the
I, No. D-1501) with ethidium bromide (EB, SIGMA) were proposed formula.
Fig. 2 Compounds used for the QSAR studies.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1655

View Article Online
Paper
Dalton Transactions
Characterization of the complexes in the solid state
IR spectroscopic characterization. All complexes showed a
similar FTIR absorption band pattern (Fig. S1†). Spectra for
analogous Pd and Pt compounds were almost identical in
agreement with isostructurality observed in their crystal struc-
tures.19 Based on our previous experience on the vibrational
behavior of HL ligands and dppf and their platinum com-
plexes and on previous assignments performed on metal com-
pounds containing these ligands, most relevant vibration
bands were tentatively assigned (Tables S3–S5†).18,19
The infrared spectra of the 8-hydroxyquinoline derivatives
HL1–HL5 show three main vibration bands: ν(O–H), ν(CvN)
and ν(C–O). A strong ν(CvN) band around 1508–1483 cm−1 is
observed. After complexation this band is slightly displaced to
lower energies, as previously reported for related complexes.
The band assigned to the ν(C–O) vibration located displaced to
higher energies after complex formation, indicating coordi-
nation through the oxygen-phenolate atom.43,44 As expected,
the medium intensity ν(O–H) vibration band around
3200–3400 cm−1 is absent in the five complexes due to depro-
tonation of the ligands. FTIR results show that the 8-hydroxy-
quinoline moiety is in all cases coordinated to Pd by the phe-
nolic oxygen and the quinoline nitrogen.
Pd-dppf-L2 showed strong bands around 1500 and
1290 cm−1 assigned to ν(N–O)as and ν(N–O)s vibrations,
respectively. Five bands corresponding to the dppf moiety were
identified as well as the strong stretching ν(P–F) and bending
−
δ(F–P–F) bands of the PF6 counterion at around 840 and
560 cm−1 45,46
.
Crystal structure of the [PdII(L)(dppf)](PF6) compounds (HL
= HL1, HL3 and HL4). Fig. 3 shows the ORTEP drawings of the
three [PdII(L)(dppf)](PF6) complexes. Intra-molecular bond
lengths and angles around metal ions are given in Table 1.
The three compounds crystallize in the monoclinic P21/c
space group. The molecular structures are closely related to
each other with the Pd(II) ion at the center of a nearly square
planar P2NO environment. The metal ion is cis-coordinated to
the 8-hydroxyquinoline moiety acting as a bidentate ligand
through the oxygen and nitrogen atoms with Pd–O distances
in the range of 2.030–2.046 Å and Pd–N distances in the range
of 2.135–2.142 Å. The [PdII(L)(dppf)](PF6) structures show Pd–
O and Pd–N distances in the same range observed for their
Pt(II) analogues previously reported and for other complexes
that include these ligands.19,47–50
The other cis-positions are occupied by a bis(diphenylpho-
sphino) ferrocenyl moiety also acting as a bidentate ligand
through the phosphorus atoms with Pd–P distances in the
range of 2.272–2.282 Å. The Pd–P distances lie also in the
typical ranges observed in other metal complexes with a
planar environment around Pd(II) and with the dppf acting as
a cis-bidentate ligand.18,19
Fig. 3 Drawing of the palladium(II) complexes: (a) [PdII(L1)(dppf)](PF6),
(b) [PdII(L3)(dppf)](PF6) and (c) [PdII(L4)(dppf)](PF6), in the solid state
showing the atomic displacement ellipsoids at the 50% probability level.
For clarity the PF6 counter-ion has been omitted. Red: oxygen, blue:
nitrogen, light purple: phosphorus, light green: chlorine, deep purple:
iodine, dark green: iron, grey: carbon, white: hydrogen.
−
The coordination around the Fe(II) ion in the ferrocenyl
moiety shows the expected Archimedean pentagonal anti-prism
conformation, with the Cp rings staggered to each other. The
Fe–C distances range from 2.010(3)–2.075(4) Å in [PdII(L1)
1656 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Table 1 Bond lengths (Å) and angles (°) around the metal center in
[PdII(L)(dppf)](PF6) compounds
The signals corresponding to the PPh2 rings were observed as
multiplets between 7.4 and 8.2 ppm.
Signals of the bioactive 8-hydroxyquinoline derivatives
showed displacement in chemical shift values in all cases after
coordination (see Table S6†).54 In general, the signals of the
quinoline ring were displaced towards high field, especially H4
(and H7, as applicable) located at the mesomeric acceptor-
position to the coordinated groups. A similar behavior has
been previously observed for Pt analogs.19 The exception was
H2 of Pd-dppf-L2, which is displaced towards a lower field, as
for its Pt analogue.
In the region of 3.5–5.5 ppm, signals of the protons of the
Cp rings of the dppf ligand were observed. For the free dppf,
two proton signals at 3.94 and 4.27 ppm have been reported.55
These signals correspond to equivalent Ha-Hb and Hc-Hd
protons of both Cp rings (figure a in Table S6†). In the case of
the [PtII(L)(dppf)](PF6) compounds, a four signal pattern was
observed.19 Nevertheless, an irregular signal pattern was
observed in the spectra of the analogous Pd compounds
(except for the compound Pd-dppf-L1), not seen in our pre-
vious reports. This behavior has been previously observed for
other M-dppf complexes with bidentate ligands, and it could
be related to complex fluxionality in solution, as it has been
described before.53
[PdII(L1)(dppf)]
(PF6)
[PdII(L3)(dppf)]
(PF6)
[PdII(L4)(dppf)]
(PF6)
Bond lengths (Å)
Pd–O1
2.030 (3)
2.136 (3)
2.277 (6)
2.282 (5)
2.046 (2)
2.135 (3)
2.272 (5)
2.273 (7)
2.038 (4)
2.142 (5)
2.277 (4)
2.275 (7)
Pd–N1
Pd–P1
Pd–P2
Bond angles (°)
O1–Pd–P1
P2–Pd–P1
P2–Pd–O1
N1–Pd–P1
N1–Pd–O1
N1–Pd–P2
86.39 (9)
96.98 (3)
173.5 (1)
167.79 (9)
81.47 (9)
95.28 (9)
88.58 (6)
96.16 (3)
175.04 (7)
169.01 (7)
80.47 (9)
94.76 (7)
88.42 (9)
96.13 (5)
175.31 (9)
168.81 (9)
80.51 (9)
95.08 (9)
(dppf)](PF6), 2.002(3)–2.060(3) Å in [PdII(L3)(dppf)](PF6) and
1.995(6)–2.068(7) in [PdII(L4)(dppf)](PF6). Phosphorus–C(Ph)
bond lengths are in the 1.812(4)–18.27(6) Å range, while P–C
(Cp) lengths in the range of 1.776(6)–1.802(4) Å.
The crystallographic structural data have been deposited at
the Cambridge Crystallographic Data Centre (CCDC). Any
request to the CCCD for this material should quote the full lit-
erature citation and the reference number CCDC 1963631 for
[PdII(L1)(dppf)](PF6), CCDC 1963633 for [PdII(L3)(dppf)](PF6)
and CCDC 1963632 for [PdII(L4)(dppf)](PF6).†
Results of the biological studies
The activity of the new Pd compounds was evaluated against
Characterization of the complexes in solution
bloodstream T. b. brucei and murine macrophages, selected as
1H NMR spectra of all complexes were recorded at room temp- the mammalian cell model.
erature in DMSO-d6 solution. Two-dimensional COSY NMR
To obtain an initial insight into the stability of the com-
experiments aided in the assignment of the signals. The plexes in the culture medium (pH 7.4) used for the biological
results of the experiments for the ligands and the complexes assays, the 31P-NMR spectra of the most active Pt compound
are shown in Table S6.† Fig. S2† depicts the spectra. All com- selected to perform an in vivo study, [Pt(L4)(dppf)](PF6), were
plexes showed a similar pattern of signals corresponding to measured during 48 h at room temperature in a mixture of
the protons of the 8-hydroxyquinoline derivatives and the dppf DMSO-10 mM Tris-HCl buffer solution pH 7.4. The spectrum
ligand. Broad signals were observed in the spectra of the com- showed two signals at 15.7 and 11.5 ppm (JP–P = 23.6 Hz)
plexes with heavier substituents at the positions 5 and 7 of the which correspond to the two not equivalent phosphorus atoms
8-hydroxyquinoline ring. This behavior is probably related to of the dppf moiety. The septuplet at −144 ppm (JP–F = 7.12 Hz)
some extent to geometry deformation in solution around the corresponds to the phosphorus atom of the PF6− counterion.56
metal center from square planar to tetrahedral that generates No new signals were detected, providing evidence of the stabi-
some paramagnetic behavior.51,52 This effect could be gener- lity of the complex in the selected medium in a 48 h period
ated by electronic and steric effects of the halide substituents (Fig. 4). In particular, the signal of the free dppf ligand was
in the 8-hydroxyquinoline ring and/or fluxionality of the dppf not detected.
moiety in solution.53 Given the structural similarity of all com-
Furthermore, HPLC studies were performed for M-dppf-L4
plexes and the higher spectra definition of those with lighter compounds, where M = Pd(II) or Pt(II), in DMSO solvent and
substituents (Pd-dppf-L1 and Pd-dppf-L2), a tentative assign- DMSO/10 mM Tris-HCl buffer solution pH 7.4 (4 : 1), in order
ment of the signals for the compounds with heavier substitu- to evaluate their stability in a medium similar to the ones used
ents (Pd-dppf-L3, Pd-dppf-L4 and Pd-dppf-L5) was made.
in the biological assays. Initially, compounds, precursors and
The signal corresponding to the –OH group proton was not ligands were evaluated using only DMSO as the solvent.
observed in the spectra of the compounds Pd-dppf-L1 and Pd- Chromatograms of the M-dppf-L4 compounds were obtained
dppf-L2, confirming the deprotonation of the oxygen atom of at t = 0, t = 24 h and t = 7 days. New signals were not observed
the ligands for their coordination to the metal center. In the during this period. Retention time values at t = 0 are shown in
region of 6–10 ppm, signals of the protons of the L ligands Table S7.†
and the phosphine (PPh2) substituents of dppf were observed.
Retention time values for M-dppf-L4 compounds were
In general, signals from the protons of the quinoline-derived approximately 24 min, whereas values for the precursors were
ligands do not overlap with the signals from the PPh2 rings. approximately 22 min. This indicates the higher lipophilic
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1657

View Article Online
Paper
Dalton Transactions
Fig. 4 31P-NMR spectra of Pt-dppf-L4 in DMSO-d6/10 mM Tris-HCl buffer solution pH 7.4 at 0 and 48 h. 31P-NMR spectrum of dppf is included for
comparison.
character of M-dppf-L4 in comparison with the corresponding activity against T. brucei with IC50 values in the micromolar or
precursors. This also applies to ligand HL4. In the case of submicromolar range. It is worth noting that all Pd complexes
dppf, the most lipophilic compound, its RT is higher than the proved to be 2-fold to 45-fold more potent than Nifurtimox.
other values as expected.
In addition, M-dppf-L4 compounds were dissolved in
DMSO/10 mM Tris-HCl buffer solution pH 7.4 (4 : 1) and evalu-
ated at t = 0 and t = 24 h, which complies with the exposure
Table 2 Cytotoxicity (IC50) and selectivity index (SI) of the new [PdII(L)
(dppf)](PF6) compounds against the infective stage of T. brucei and
J774 murine macrophages
time used in the biological tests. When comparing the reten-
tion times of the compounds in DMSO and in DMSO/Buffer,
they do not vary significantly (Tables S7 and S8†).
T. brucei-GFP
IC50 SD (µM)
J774
IC50 SD (µM)
Compound
SIa
dppfb
HL1b
HL2b
HL3b
HL4b
HL5b
>10
45
6
<4
1.3
38.9
8.4
16.2
16.4
A single signal was observed in the chromatograms corres-
ponding to compounds M-dppf-L4 (Fig. S3 and S4†). The
retention times of these signals do not coincide with either the
respective precursors or the HL4 and dppf ligands. The unal-
tered shape of the peak over time, as well as the non-signifi-
cant variation of the associated area, reflects the stability of
the compounds in the studied medium. In summary, the
chemical stability of both compounds in the mentioned media
was confirmed for a period of 7 days.
12.4 0.8
0.8 0.3
2.5 0.2
2.4 0.4
3.0 0.2
16.6 1.5
31.1 1.2
21.0 0.8
38.8 6.7
49.4 9.1
Pd-dppf-L1
Pd-dppf-L2
Pd-dppf-L3
Pd-dppf-L4
Pd-dppf-L5
0.9 0.2
0.33 0.09
4.5 0.1
4.8 0.5
8.5 0.5
33.8 1.7
19.8 2.4
30.0 3.3
45 11
9.4
102.4
4.4
6.3
7.2
7
4
Pt-dppf-L1b
Pt-dppf-L2b
Pt-dppf-L3b
Pt-dppf-L4b
Pt-dppf-L5b
Nifurtimox
Suramin
0.3 0.1
0.93 0.03
0.22 0.01
0.14 0.05
0.22 0.04
15 3b
3.4 0.4
25.8 1.5
3.4 0.2
6.7 0.4
6.4 0.3
150 5c
ND
11.3
27.7
15.5
47.8
29.1
10b
Anti-T. brucei activity and selectivity
The antitrypanosomal activity of the new Pd-ferrocenyl deriva-
tives was evaluated against the bloodstream form of T. brucei
(Table 2).
We previously reported the activity against bloodstream
T. brucei of the free HL ligands, with IC50 values mostly in the
micromolar range and significantly lower than that of the
reference antitrypanosomal drug nifurtimox (IC50 = 15 μM,
same technique).19 All the Pd-dppf-L compounds showed
78d
ND
[PtII(L)(dppf)](PF6) complexes, HL ligands, dppf and nifurtimox are
included for comparison. Duration of treatment 24 h. a SI, selectivity
index: IC50 J774 murine macrophages/IC50 T. brucei. b Data from ref.
19. c Data from ref. 57. d Data from ref. 36; ND, not determined.
1658 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
The complexation of the bioactive HL ligands to the {Pd-dppf} free ligands with complexes and complexes among them, all
moiety increased significantly the anti-trypanosomal activity in determined under identical experimental conditions. As
the case of Pd-dppf-L1 and Pd-dppf-L2, exhibiting IC50 values expected, the lipophilicity increased from the free ligand HL to
in the submicromolar range. HL1 and HL2 ligands increased its metal compounds due to the inclusion of the {Pd(dppf)}
their activities by 14- and 2.4-fold when coordinated to Pd, moiety (higher the RM value, higher the lipophilicity) (RM
respectively. In general, the Pd compounds are less active than values: HL1 = −0.43, HL2 = −0.43, HL3 = 0.35, HL4 = 0.51, HL5
their Pt analogues. In order to gain insight into the biological = 0.65) (Table 3).19
selectivity of the five Pd compounds, their cytotoxicity against
macrophages (J774 murine cell line) was tested (Table 2). The
Pd compounds showed similar cytotoxicity in the selected
mammalian cell model to that of the free ligands HL.
Importantly, except for Pd-dppf-L2, all Pd-based compounds
have similar SI values to Nifurtimox. Compared to the Pt ana-
logues and with the exception of Pd-dppf-L2 (SI = 102), the Pd
compounds proved to be less selective towards the parasite (SI
≤ 9 for Pd compounds vs. SI > 10 for Pt compounds (Table 2).
Pd-dppf-L2 proved to be the most selective compound of both
series, with anti-T. brucei activity (IC50 = 0.33 µM) slightly lower
than that of Pt-dppf-L4 (IC50 = 0.14 µM).
QSAR, quantitative structure–activity relationship studies
In order to understand and predict the main structural deter-
minants on the anti-Trypanosoma brucei activity, a search of
quantitative structure–activity relationships (QSAR) was
performed.
Initially eighteen compounds, i.e. [M(L)(dppf)](PF6), where
M = Pt(II) or Pd(II), HL = hydroxyquinolyl-, tropolonyl-, hinokitio-
lyl-, and 5-nitrofurylthiosemicarbazonyl-ligands (Fig. 2), were
used in the QSAR studies. The dependent variable was the log10
of the IC50 values on T. brucei, log10(IC50). As the independent
variables, we selected different physicochemical parameters that
covered lipophilic, electronic, and steric/topological properties,
according to the extrathermodynamic approach (Hansch-ana-
Lipophilicity
Lipophilicity is a very important physicochemical property for lysis).58 For the lipophilic descriptor we used RM values,59 as the
1
prospective drugs that determines biological behaviour, particu- electronic parameter we selected some signals of the H NMR
larly transmembrane transport and interaction with biological spectra of the studied compounds,60 and finally for the steric/
receptors. Therefore, it is usually determined and the eventual topological, and also as the electronic variable we defined a cat-
existence of a correlation with biological activity is checked.
egorical descriptor. The lipophilicity of the studied compounds
The new palladium compounds as well as the previously was determined experimentally, in terms of RM (see above)
reported Pt(II) analogues and a series of PdII-dppf-L and PtII- (Table 3). On the other hand, due to the structural variety of the
dppf-L compounds with other bidentate bioactive ligands HL studied complexes, i.e. two different metals and three different
do not show adequate solubility in water to allow for the deter- families of ligands, we selected the common cyclopentadienyl-
mination of Log P values in the conventional way. Therefore, motif to analyze the compounds’ electronic effect contributions
1
their lipophilicity was experimentally determined using in the displayed activities. Specifically, we observed that the H
reversed-phase TLC, where the stationary phase (precoated NMR signals of the cyclopentadienyl-framework were highly
TLC-C18 plates) may be considered to simulate lipids of bio- dependent on the coordination and nature of the compound.
logical membranes or receptors, and the mobile phase Consequently, as a descriptor we used Δδ defined as the largest
(MeOH : DMF : H2O (65 : 5 : 30 v/v/v)) resembles the aqueous difference between δ of cyclopentadienyl-framework protons in
biological milieu.17–19 This mobile phase allows differentiating the complexes and δ of cyclopentadienyl-framework protons
complexes and free ligands of the three different series without coordination (Table 3). From the original twenty two
M-dppf-L′ developed by us, where M = Pd(II) or Pt(II) and L′: tro- structurally related compounds four could not be included in
polonyl, 5-nitrofurylthiosemicarbazonyl or 8-hydroxyquinolinyl the QSAR study (Pd-dppf-L2, Pd-dppf-L3, Pd-dppf-L4 and Pd-
derivatives (Fig. 5).17–19
dppf-L5) since this Δδ value could not be unequivocally deter-
The most lipophilic compounds run less in this TLC system mined for them due to the complexity of the spectra. Similarly,
and, therefore, the method allows comparing lipophilicities of the structural variety of the studied complexes led us to define
Fig. 5 Ligands of the previously reported [M(L’)(dppf)](PF6) compounds.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1659

View Article Online
Paper
Dalton Transactions
Table 3 Lipophilicity (Rf, RM) of the new [PdII(L)(dppf)](PF6) compounds, their [PtII(L)(dppf)](PF6) analogues and the previously reported [M(L’)(dppf)]
(PF6) compounds, where M = Pd(II) or Pt(II) and HL’ = HTrop, HHino, HTN, HTM, HTE or HTF (Fig. 5). The IC50 value against the infective stage of
T. brucei is included. RM = log10 [(1/Rf) − 1] and Δδ values (defined in the next section)
T. brucei GFP
IC50 SD/µM
T. brucei GFP
IC50 SD/µM
Compound
Rf
RM
Δδd
Compound
Rf
RM
0.84
Δδd
Pd-dppf-L1
Pd-dppf-L2
Pd-dppf-L3
Pd-dppf-L4
Pd-dppf-L5
Pd-dppf-Hinoa
Pd-dppf-Tropa
Pd-dppf-TFb
Pd-dppf-TEb
Pd-dppf-TMb
Pd-dppf-TNb
0.15
0.06
0.12
0.07
0.04
0.09
0.11
0.33
0.41
0.46
0.48
0.76
1.21
0.88
1.15
1.34
1.02
0.93
0.31
0.16
0.07
0.03
1.13
—
—
—
—
0.80
0.56
0.82
0.83
0.85
0.83
0.9 0.2
0.33 0.09
4.5 0.1
4.8 0.5
Pt-dppf-L1c
Pt-dppf-L2c
Pt-dppf-L3c
Pt-dppf-L4c
Pt-dppf-L5c
Pt-dppf-Hinoa
Pt-dppf-Tropa
Pt-dppf-TFb
Pt-dppf-TEb
Pt-dppf-TMb
Pt-dppf-TNb
0.13
0.10
0.11
0.08
0.08
0.10
0.15
0.31
0.41
0.44
0.51
1.24
1.31
1.33
1.12
1.12
0.73
0.47
0.79
0.81
0.77
0.82
0.3 0.1
0.97
0.91
1.04
1.07
0.95
0.76
0.35
0.16
0.10
−0.02
0.93 0.03
0.22 0.01
0.14 0.05
0.22 0.04
4.5 0.1
7
4
1.2 0.1
1.3 0.1
1.56 0.04
0.98 0.06
0.93 0.03
0.90 0.05
2.1 0.1
1.01 0.04
0.52 0.03
0.60 0.03
0.77 0.02
a Data from ref. 17. b Data from ref. 18. c Data from ref. 19. d Δδ: δdppf in complex − δfree dppf.
an indicator variable, IVPd, that adopts value 1 for Pd complexes
or otherwise value 0 for Pt complexes. Besides, the correlation
matrix for the used physicochemical descriptors was performed,
and no cross-correlations between the independent variables
were obtained. Thus, these parameters are orthogonal, allowing
its safe use in the multilinear regression analyses.61,62
In all the cases, non-statistical significant correlations were
obtained when each of the descriptors was studied as a unique
independent variable with the complete population of com-
pounds. Consequently, multivariable regressions between the
biological activities and the physicochemical descriptors were
studied. Considering the eighteen studied compounds and the
variables RM, Δδ and IVPd or the variables RM2, Δδ, and IVPd
2
statistically poor correlations were obtained, radj = 0.23887
and 0.21535, respectively. Both generated equations deficiently
predicted the activities of derivatives Pt-dppf-L2 and Pt-dppf-
Hino. However, when these Pt-complexes were treated as out-
liers, a better correlation was obtained (eqn (1), Fig. 6A).
log10ðIC50Þ ¼ 0:611ð+0:186Þ ꢀ 0:942ð+0:201ÞΔδ
2
ꢀ 0:112ð+0:103ÞRM þ 0:252ð+0:086ÞIVPd ð1Þ
2
radj ¼ 0:7577; FF ¼ 16:6; p ¼ 0:00014; n ¼ 16
This correlation shows the relevance of the electronic prop-
erty, the lipophilicity and the type of metal. Compounds with
electron-withdrawing ligands, with greater lipophilicity and
harboring Pt would result in higher T. brucei cytotoxicity.
When the associated errors of eqn (1) independent vari-
ables were examined one could observe that lipophilicity
descriptor showed the highest value. Consequently, consider-
ing the sixteen complexes of eqn (1) we re-analyzed the
relationship without the RM variable (eqn (2), Fig. 6B).
log10ðIC50Þ ¼ 0:628ð+0:187Þ ꢀ 1:202ð+0:189ÞΔδ
þ 0:257ð+0:087ÞIVPd
ð2Þ
2
radj ¼ 0:7573; FF ¼ 24:2; p ¼ 0:00004; n ¼ 16
Fig. 6 Actual and predicted log10(IC50) regression employing equations
1 (A) and 2 (B). Note: In red outliers for the initial analyses with the com-
plete population of studied compounds.
2
In eqn (2), radj was maintained; however, F and p values
and variable errors were improved.
1660 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Eqn (2) showed that the electronic property and the type of derivatives with voluminous and electron density
metal are the most relevant descriptors to predict anti-T. brucei attracting substituents; in contrast, Pd compounds including
activity.
the hydroxyquinoline derivatives lacking substituents or with
resonance donor substituents showed lower quenching
(Table S9†).
DNA interaction studies
The results suggest the interaction of the compounds with
the DNA, under the conditions tested. The technique does not
allow determining the kind of interaction established. It only
shows that due to conformational changes generated by the
DNA interaction of the complexes EB is released from the
adduct.
The mechanisms involved in the fluorescence quenching
process for the complexes can be accessed by a Stern–Volmer
analysis according to
DNA has shown to be a possible molecular target of the whole
series of M-dppf compounds.17–19
Ethidium bromide (EB) was employed to investigate the
interaction of the new Pd compounds with DNA. Under the
selected experimental conditions EB shows an emission
maximum at 601 nm for an excitation wavelength of 510 nm.
Intercalation of the dye into double stranded DNA induces a
high increase of the fluorescence quantum yield and a shift of
the emission maximum.63 Under the experimental conditions
of the current experiments, the emission maximum of the gen-
erated {DNA-EB} adduct is located at 594 nm. At the selected
excitation wavelength DNA, HL1–HL5 ligands and the Pd-dppf-
L complexes are non-fluorescent at pH = 7.4 10 mM Tris-HCl
buffer solution /5% DMSO, and no fluorescence emission
results from the direct interaction of the complexes or the
ligands with DNA. The complexes showed negligible absorp-
tion at 594 nm in the concentration range used. The titration
of the {DNA-EB} adduct with the new compounds led to the
displacement of EB from the adduct, which is revealed by a
concentration dependent decrease of the fluorescence inten-
sity when the concentration of the complex (quencher)
increases (Fig. 7). Although the extent of the quenching is of a
similar order for the five complexes, a higher quenching is
observed for the Pd compounds including hydroxyquinoline
IF0
IF
¼ 1 þ KSV½Qꢁ
ð3Þ
where IF0 and IF are the emission fluorescence intensity of the
{DNA-EB} adduct in the absence and in the presence of the
complex, KSV the Stern–Volmer constant, and [Q] the concen-
tration of the quencher (in this case, the complexes).64 Eqn
(3) predicts a linear plot for IF0/IF for a homogeneously emit-
ting system. Stern–Volmer analysis of the Pd-dppf-L com-
pounds showed a linear IF0/IF plot and the KSV constant was
calculated for each system (Table S9†). Although KSV informs
about the quenching produced by the compounds, it is a
useful parameter that indirectly indicates the affinity of the
compounds for DNA. Obtained KSV values for the five com-
plexes are of a similar order to those previously reported by
us for other complexes showing antitrypanosomal activity,18,65
but lower than those previously reported for other complexes
with typical planar DNA intercalating ligands.66 It is worth
noting that it has been previously stated that none of the free
ligands affect the emission of the {DNA-EB} adduct, except
for HL2 that generates a slight decrease of fluorescence
(quenching 16% (110 µM); log KSV = 3.0).19 Nevertheless, the
KSV value found for its Pd-dppf-L2 complex as well as for the
previously reported Pt analogue was higher than that found
for the free ligand. Except for the HL2 complexes, the Pd-
dppf-L compounds show higher Log (KSV
) than the Pt
analogues.
The results show that in every case the {Pd-dppf} complexes
of the HL ligands generate an interaction with the {DNA-EB}
adduct, which was absent in the case of the free HL ligands.
The whole set of results suggests that DNA could be one of the
molecular targets of these complexes in the parasite.
Nevertheless, it should not be the main target since Pd com-
pounds show higher KSV values than the Pt analogues but the
Pt compounds show in general higher activities than the Pd
ones.
Fig. 7 Titration of the {DNA-Ethidium bromide} adduct with the Pd-
dppf-L complexes. Relative fluorescence intensity (%) at λem = 594 nm
of the {DNA-EB} adduct (CDNA = 20 μM, CEB = 10 μM, samples prepared
in 5% DMSO/10 mM Tris-HCl buffer medium) with increasing concen-
trations of the five metal complexes upon 30 min incubation at 37 °C.
Squares: Pd-dppf-L1; circles: Pd-dppf-L2; up triangles: Pd-dppf-L3; dia-
monds: Pd-dppf-L4; down triangles: Pd-dppf-L5.
Evaluation of intracellular thiol-redox balance
Ferrocenes are known to undergo one electron oxidation,
which in biological systems may lead to the formation of rad-
icals and the concomitant oxidation of macromolecules, with
the thiol groups of proteins and low molecular metabolites
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1661

View Article Online
Paper
Dalton Transactions
(i.e. glutathione) being important nucleophilic targets. Thus, a days) or a single dose of the drug Berenil (DAC 40 mg kg−1). As
potential involvement of the compounds in the generation of shown in Fig. 8A, relative parasite burden was monitored non-
intracellular redox unbalance was investigated using a redox- invasively by in vivo imaging of anesthetized mice at different
reporter cell line of bloodstream T. brucei that was exposed for time points. Given the small number of animals only a quali-
24 h to the compounds.67 The redox biosensor is genetically tative analysis of data is performed. Imaging at day 2, readily
encoded and consists of the oxidoreductase glutaredoxin prior to treatment initiation, confirms that all animals were
(hGrx) fused to the redox sensitive green fluorescent protein successfully infected. Bioluminescence at day 6 and 9 post-
roGFP2. In this configuration, the biosensor is able to infection (which corresponds to 24 h and 96 h post-finaliza-
detect subtle changes in the glutathione/glutathione tion of treatments) shows that animals treated with Pt-dppf-L4
disulfide (GSH/GSSG) and trypanothione/trypanothione di- present a comparatively lower signal area and intensity than
sulfide (T(SH)2/TS2) ratios, which directly relate to changes in those that received vehicle alone (DMSO). It is worth noting
fluorescence intensity (FI).36,68 For instance: high FI at 488 nm that at day 9 bioluminescence in the Pt-dppf-L4-treated
= high GSH/GSSG or high TSH2/TS2 ratio = reducing intra- animals was still confined to the abdominal region (similar to
cellular milieu, while low FI = low GSH/GSSG or low TSH2/TS2 day 6) whereas in mice from the vehicle group the signal
ratio = oxidizing intracellular milieu. This redox reporter cell spread to the thoracic region. In line with previous obser-
line proved useful to investigate the mode of action of vations relying on determination of parasitemia by microscopy
different
compounds,36,69,70
including
Pd-ferrocenyl counting,72,73 from day 11 on, the bioluminescence signal
increased and spread all over the animals body from both
complexes.18
The redox analysis was performed in the PI-negative cell groups. As reported earlier,74 treatment with a single dose of
population (i.e. viable cells) treated at a compound concen- the control drug DAC proved very effective in controlling and
tration closest to its IC50 value. Under this assay condition and eliminating parasites from mice as evidenced by the lack of
compared to the DMSO control, only the HL2 ligand induced a the bioluminescence signal from day 6 post-infection and
minor (16%) but statistically significant decrease in the fluo- until the end of the experiment.
rescence intensity of the biosensor, which is indicative of the
generation of a minor intracellular oxidizing milieu (see ESI
Fig. S5†). This suggests that the mechanism of parasite killing
by most of the Pd-based complexes, at variance with some
from the Pt-containing ones,18 does not involve the generation
of intracellular oxidative stress.
Therapeutic efficacy test
From the set of compounds tested here and in a previous
study,19 Pd-dppf-L2 (IC50 = 0.33 µM, SI = 102) and Pt-dppf-L4
(IC50 = 0.14 µM, SI = 48) were found to be the most active and
selective against T. brucei and were considered as candidates
for preliminary therapeutic efficacy assays. However, due to the
low solubility of Pd-dppf-L2, the in vivo studies were conducted
with Pt-dppf-L4. T. brucei is mainly an extracellular parasite
that, during the acute stage of the infection, dwells in the
hemolymphatic system and interstitial tissue of the host.
Thus, assuming a distribution of the compound confined to
the extracellular fluids and the reported extracellular fluid
volume of female mice determined by bioimpedance (4.8 mL;
ref. 71), by administering 10.5 mg kg−1 × animal, the concen-
tration of Pt-dppf-L4 should be nearly 260-fold the IC50 value
determined in vitro.
Fig. 8 Therapeutic efficacy studies in an acute murine model for
Human African Trypanosomiasis (HAT). (A) In vivo imaging of mice
infected with T. brucei LUC. The bioluminescence signal was acquired
with animals in the ventral position and using a Xtreme II equipment
First, the toxicity of Pt-dppf-L4 administered intraperitone-
ally at 10.5 mg kg−1 daily for four days was tested in non-
infected mice. There were no clinical signs of acute
toxicity (diarrhea, hyper/hypo-ventilation, -excitation, ataxia,
(Bruker). The mean body signal was recorded for each mouse. At day 2,
suppression of body weight, fur loss, etc.) in the animals immediately after imaging, the animals were injected intraperitoneally
with Pt-dppf-L4 (10.5 mg kg−1 × 4 days), vehicle (DMSO 2.75 g kg−1 × 4
under study.
days) or a single dose of diminazene aceturate, DAC (40 mg kg−1). The
Next, mice (n = 3 per group) were infected with a biolumi-
nescent cell line of T. brucei and treatment was initiated at day
2 post-infection with an intraperitoneal administration of Pt-
scale bar for radiance (right side) indicates signal intensity expressed as
counts per s × 103. (B) Kaplan–Meier plots for each animal group.
Comparison of the vehicle vs. Pt-dppf-L4 group using the Log-rank
dppf-L4 (10.5 mg kg−1 × 4 days), vehicle (DMSO 2.75 g kg−1 × 4 (Mantel-Cox) test yielded a p value = 0.1087.
1662 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
Although survival analysis (Fig. 8B) shows a 5-day extension pound displayed anti-T. brucei activity in vivo, which, though
in the mean survival time (t50%) between the DMSO (t50% = 12 not curative, contributed to extending mice survival. Further
days) and Pt-dppf-L4 (t50% = 17 days) groups, due to the low studies should aim at improving the solubility and/or bio-
animal number this difference was not statistically significant availability of the compound, as well as its testing in trypano-
(p value = 0.1087).
somatid species that display intracellular stages in mammalian
Although preliminary, the in vivo studies show that Pt-dppf- host.
L4 lacks acute toxicity when applied to animals in the dose/
regimen described above, and suggest that the compound is
capable of exerting an anti-proliferative effect in parasites that
Author contributions
extends animal survival but is not curative. Taking into
account that this compound holds one of the highest lipophili- Feriannys Rivas: Investigation, Formal Analysis; Andrea
city indexes from the Pt series, its lack of in vivo efficacy might Medeiros:
Methodology,
Investigation,
Writing
and
be in part ascribed to a low biodistribution of the compound Supervision (in vitro T. brucei and in vivo tests); Cristina
in extracellular fluids. As stated above, African trypanosomes Quiroga: Investigation and Formal Analysis (in vivo study);
are predominantly extracellular parasites. However, additional Diego Benítez: Investigation and Formal Analysis (in vivo
factors may explain the lack of in vivo activity, e.g. compound study); Marcelo Comini: Methodology, Conceptualization,
metabolization and binding to plasma proteins. Future studies Writing, Supervision (in vitro T. brucei and in vivo tests);
will address compound formulation, different administration Esteban Rodríguez-Arce: Investigation, Formal Analysis;
routes and combination therapy, in order to improve the Ignacio Machado: Investigation, Formal Analysis (HPLC
in vivo bio-activity.
studies); Hugo Cerecetto: Writing, Investigation and
Methodology (QSAR study); Dinorah Gambino: Writing,
Supervision, Project administration, Conceptualization.
Conclusions
A series of five new [Pd(L)(dppf)](PF6) compounds, with HL =
8-hydroxyquinoline derivatives and dppf = 1,1′-bis(dipheny1-
Conflicts of interest
phosphino)ferrocene, were synthesized and fully characterized There are no conflicts of interest to declare.
in the solid state and in solution. These new Pd(II)-compounds
showed good activity against bloodstream T. brucei, with IC50
values in the micromolar or submicromolar range, and good
selectivity towards the pathogen. All Pd complexes proved to
Acknowledgements
be 2-fold to 45-fold more potent than the drug nifurtimox. The F. R. acknowledges the support of the Agencia Nacional de
complexation of the bioactive HL ligands to the {Pd-dppf} Investigación e Innovación (ANII, Uruguay) through the grants
moiety increased significantly the anti-trypanosomal activity in POS_NAC_2016_1_129899 and POS_NAC_2018_1_151206. CQ,
the case of Pd-dppf-L1 and Pd-dppf-L2 (14- and 2.4-fold, AM and MAC acknowledge the support of CSIC (Comisión
respectively) that led to submicromolar anti-T. brucei activity. Sectorial de Investigaciones Científicas, Universidad de la
Nevertheless, the Pd compounds are less active than their pre- República Uruguay). HC, DG, IM, DB, AM and MAC belong to
viously reported Pt analogues. The Pd compounds showed the Uruguayan National System of Researchers (SNI).
similar cytotoxicity in the selected mammalian cell model to
that of the free ligands HL. Except for Pd-dppf-L2, all Pd-based
compounds have similar SI values to nifurtimox but are mostly
less selective towards the parasite than the Pt analogues.
References
Nevertheless, Pd-dppf-L2 showed an enhanced potency and
selectivity against T. brucei, with respect to its Pt analogue,
thus deserving further investigation studies.
In regard to the probable mechanism of action, the com-
pounds do not alter the intracellular thiol-redox homeostasis
of the parasite. The interaction of the complexes with DNA
suggested that this biomolecule could be a potential target.
According to the QSAR study, compounds with electron-
withdrawing ligands, higher lipophilicity and harboring Pt
would result in higher T. brucei cytotoxicity.
1 M. De Rycker, B. Baragaña, S. L. Duce and I. H. Gilbert,
Nature, 2018, 559, 498–506; P. J. Hotez, D. Molyneux,
A. Fenwick, J. Kumaresan, S. Ehrlich Sachs, J. D. Sachs and
L. Savioli, N. Engl. J. Med., 2007, 357, 1018–1027.
2 S. P. S. Rao, M. P. Barrett, G. Dranoff, C. J. Faraday,
C. R. Gimpelewicz, A. Hailu, C. L. Jones, J. M. Kelly,
J. K. Lazdins-Helds, P. Mäser, J. Mengel, J. C. Mottram,
C. E. Mowbray, D. L. Sacks, P. Scott, G. F. Späth,
R. L. Tarleton, J. M. Spector and T. T. Diagana, ACS Infect.
Dis., 2019, 5, 152–157.
From the whole series of [M(L)(dppf)](PF6) compounds
developed, where M = Pt(II) or Pd(II) and HL = 8-hydroxyquino-
lyl derivatives (Fig. 1), Pt-dppf-L4 was selected to perform an
exploratory pre-clinical study in infected mice. The hit com-
3 A. S. Nagle, S. Khare, A. B. Kumar, F. Supek, A. Buchynskyy,
C. J. N. Mathison, N. K. Chennamaneni, N. Pendem,
F. S. Buckner, M. H. Gelb and V. Molteni, Chem. Rev., 2014,
114, 11305–11347.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1663

View Article Online
Paper
Dalton Transactions
4 C. J. Bacchi, H. C. Nathan, S. H. Hutner, P. P. McCann and 28 R. G. Kenny and C. J. Marmion, Chem. Rev., 2019, 119,
A. Sjoerdsma, Science, 1980, 210, 332–334.
5 E. D. Deeks, Drugs, 2019, 79, 215–220.
6 http://www.who.int/trypanosomiasis_african/visited2020.
7 M. Berninger, I. Schmidt, A. Ponte-Sucre and U. Holzgrabe,
MedChemComm, 2017, 8, 1872–1890.
1058–1137.
29 E. L. Cedillo-Gutiérrez, L. F. Hernández-Ayala, C. Torres-
Gutiérrez, M. Reina, M. Flores-Alamo, J. C. Carrero,
V. M. Ugalde-Saldívar and L. Ruiz-Azuara, Eur. J. Med.
Chem.,
2020,
189,
112084,
DOI:
10.1016/j.
8 V. Delespaux and H. P. de Koning, Drug Resist. Updates,
2007, 10, 30–50.
9 M. Navarro, C. Gabbiani, L. Messori and D. Gambino, Drug
Discovery Today, 2010, 15, 1070–1078.
ejmech.2020.112084.
30 S. M. Meier-Menches, C. Gerner, W. Berger, C. G. Hartinger
and B. K. Keppler, Chem. Soc. Rev., 2018, 47, 909–
928.
10 D. Gambino and L. Otero, Inorg. Chim. Acta, 2012, 393, 31 M. Fernández, J. Varela, I. Correia, E. Birriel, J. Castiglioni,
103–114.
V. Moreno, J. Costa Pessoa, H. Cerecetto, M. González and
D. Gambino, Dalton Trans., 2013, 42, 11900–11911.
32 T. A. K. Al-Allaf, H. Schmidt, K. Merzweiler, C. Wagner and
D. Steinborn, J. Organomet. Chem., 2003, 678, 48–55.
33 L. Krause, R. Herbst-Irmer, G. M. Sheldrick and D. Stalke,
J. Appl. Crystallogr., 2015, 48, 3–10.
34 (a) G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Adv.,
2015, 71, 3–8; (b) O. V. Dolomanov, L. J. Bourhis,
R. J. Gildea, J. A. K. Howard and H. Puschmann, J. Appl.
Crystallogr., 2009, 42, 339–341.
11 D. Gambino and L. Otero, Inorg. Chim. Acta, 2018, 472, 58–
75.
12 Y. C. Ong, S. Roy, P. C. Andrews and G. Gasser, Chem. Rev.,
2019, 119, 730–796.
13 D. Gambino and L. Otero, Met. Ions Life Sci., 2019, 19, 331–
358.
14 R. W. Brown and C. J. T. Hyland, MedChemComm, 2015, 6,
1230–1243.
15 M. Ravera, E. Moreno-Viguri, R. Paucar, S. Pérez-Silanes
and E. Gabano, Eur. J. Med. Chem., 2018, 155, 459–482.
16 E. Rodríguez Arce, M. F. Mosquillo, L. Pérez-Díaz,
35 S. Keller, Y. C. Ong, Y. Lin, K. Cariou and G. Gasser,
J. Organomet. Chem., 2020, 906, 121059.
G. A. Echeverría, O. E. Piro, A. Merlino, E. L. Coitiño, 36 J. Franco, F. Sardi, L. Szilágyi, K. E. Kövér, K. Fehér and
C. Maríngolo Ribeiro, C. Q. F. Leite, F. R. Pavan, L. Otero
and D. Gambino, Dalton Trans., 2015, 44, 14453–14464.
M. A. Comini, Int. J. Parasitol.: Drugs Drug Resist., 2017, 7,
303–313.
17 F. Rivas, A. Medeiros, E. Rodríguez Arce, M. Comini, 37 B. Demoro, C. Sarniguet, R. Sánchez-Delgado, M. Rossi,
C. M. Ribeiro, F. R. Pavan and D. Gambino, J. Inorg.
Biochem., 2018, 187, 73–84.
18 E. Rodríguez Arce, E. Putzu, M. Lapier, J. Diego Maya,
D. Liebowitz, F. Caruso, C. Olea-Azar, V. Moreno,
A. Medeiros, M. A. Comini, L. Otero and D. Gambino,
Dalton Trans., 2012, 41, 1534–1543.
C. Olea Azar, G. A. Echeverría, O. E. Piro, A. Medeiros, 38 D. Benítez, E. Dibello, M. Bonilla and M. A. Comini, Drug
F. Sardi, M. Comini, G. Risi, G. Salinas, I. Correia, J. Costa Dev. Res., 2020, DOI: 10.1002/ddr.21634.
Pessoa, L. Otero and D. Gambino, Dalton Trans., 2019, 48, 39 J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig,
7644–7658.
M. Longair, T. Pietzsch, S. Preibisch, C. Rueden,
S. Saalfeld, B. Schmid, J. Tinevez, D. J. White,
V. Hartenstein, K. Eliceiri, P. Tomancak and A. Cardona,
Nat. Methods, 2012, 9, 676–682.
19 F. Rivas, A. Medeiros, M. Comini, L. Suescun, E. Rodríguez
Arce, M. Martins, T. Pinheiro, F. Marques and D. Gambino,
J.
Inorg.
Biochem.,
2019,
DOI:
10.1016/j.
jinorgbio.2019.110779.
40 C. Hansch and A. Leo, The hydrophobic parameter:
measurement and calculation, in Exploring QSAR.
Fundamentals and Applications in Chemistry and Biology,
American Chemical Society Ed, Washington, 1995, pp.
97–124.
20 Y. Lin, Y. C. Ong, S. Keller, J. Karges, R. Bouchene,
E. Manoury, O. Blacque, J. Müller, N. Anghel, A. Hemphill,
C. Häberli, A. C. Taki, K. Cariou, J. Keiser and G. Gasser,
Dalton Trans., 2020, 49, 6616–6626.
21 Y. Huentupil, P. Chung, N. A. Novoa, R. Arancibia, 41 H. Cerecetto, R. Di Maio, M. González, M. Risso, P. Saenz,
P. Roussel, J. Oyarzo, A. H. Klahn, C. Silva, C. Calvis,
R. B. Ramon and C. López, Dalton Trans., 2020, 49, 12249–
G. Seoane, A. Denicola, G. Peluffo, C. Quijano and C. Olea-
Azar, J. Med. Chem., 1999, 42, 1941–1950.
12265.
42 J. K. Barton, J. M. Goldberg, Ch. V. Kumar and N. J. Turro,
J. Am. Chem. Soc., 1986, 108, 2081–2088.
22 G. Gasser, I. Ott and N. Metzler-Nolte, J. Med. Chem., 2011,
54, 3–25.
43 K. S. O. Ferraz, D. C. Reis, J. G. Da Silva, E. M. Souza-
Fagundes, E. J. Baran and H. Beraldo, Polyhedron, 2013, 63,
28–35.
23 S. S. Braga and A. M. S. Silva, Organometallics, 2013, 32,
5626–5639.
24 M. Patra and G. Gasser, Nat. Rev. Chem., 2017, 1, 0066.
25 C. Biot and D. Dive, Top. Organomet. Chem., 2010, 32, 155–
193.
44 I. Potočňák and P. Vranec, Monatsh. Chem., 2012, 143, 217–
226.
45 A. M. Heyns, Spectrochim. Acta, Part A, 1977, 33, 315–322.
26 F. Dubar, J. Khalife, J. Brocard, D. Dive and C. Biot, 46 J. Kljun, I. E. León, Š. Peršič, J. F. Cadavid-Vargas,
Molecules, 2008, 13, 2900–2907.
27 P. Chellan and P. J. Sadler, Chem. – Eur. J., 2020, 26, 1–14.
S. B. Etcheverry, W. He, Y. Bai and I. Turel, J. Inorg.
Biochem., 2018, 186, 187–196.
1664 | Dalton Trans., 2021, 50, 1651–1665
This journal is © The Royal Society of Chemistry 2021

View Article Online
Dalton Transactions
Paper
47 Z. F. Chen, Y. Q. Gu, X. Y. Song, Y. C. Liu, Y. Peng and 61 R. H. Myers, Classical and Modern Regression with
H. Liang, Eur. J. Med. Chem., 2013, 59, 194–202. Application, P.W.S. Publishers, Boston, 1986.
48 Y. M. Hu, J. Qu, Y. L. Yi, H. L. Gao and J. Z. Cui, J. Coord. 62 N. R. Draper and H. Smith, Applied Regression Analysis,
Chem., 2013, 66, 18–27. Wiley, New York, 2nd edn, 1981.
49 J. Kljun, I. E. León, Š. Peršič, J. F. Cadavid Vargas, 63 J. B. Lepecq and C. Paoletti, J. Mol. Biol., 1967, 27,
S. B. Etcheverry, W. He, Y. Bai and I. Turel, J. Inorg.
Biochem., 2018, 186, 187–196.
50 C. Martín Santos, S. Cabrera, C. Ríos-Luci, J. M. Padrón,
87–106.
64 J. R. Lakowicz, Principles of Fluorescence Spectroscopy,
Springer Science, New York, 3rd edn, ch. 8, 2006.
I. López Solera, A. G. Quiroga, M. A. Medrano, C. Navarro- 65 M. Cipriani, J. Toloza, L. Bradford, E. Putzu, M. Vieites,
Ranninger and J. Alemán, Dalton Trans., 2013, 42, 13343–
13348.
51 T. Matsuoka, Z. Li, M. Ikeshita, S. Kawamorita and
T. Naota, J. Mol. Struct., 2018, 1165, 217–222.
E. Curbelo, A. I. Tomaz, B. Garat, J. Guerrero,
J. S. Gancheff, J. D. Maya, C. Olea Azar, D. Gambino and
L. Otero, Eur. J. Inorg. Chem., 2014, 27, 4677–4689.
66 B. Y. Wu, L. H. Gao, Z. M. Duan and K. Z. Wang, J. Inorg.
Biochem., 2005, 99, 1685–1691.
52 D.
Maganas,
A.
Grigoropoulos,
S.
Staniland,
S. D. Chatziefthimiou, A. Harrison, N. Robertson, 67 M. Gutscher, A. L. Pauleau, L. Marty, T. Brach,
P. Kyritsis and F. Neese, Inorg. Chem., 2010, 49, 5079–5093.
53 X. L. Lu, S. Y. Ng, J. J. Vittal, G. K. Tan, L. Y. Goh and
T. S. A. Hor, J. Organomet. Chem., 2003, 688, 100–111.
54 O. Sigouin and A. L. Beauchamp, Can. J. Chem., 2005, 83,
460–470.
G. H. Wabnitz, Y. Samstag, A. J. Meyer and T. P. Dick, Nat.
Methods, 2008, 5, 553–559.
68 S. Ebersoll, M. Bogacz, L. M. Günter, T. P. Dick and
R. L. Krauth-Siegel, eLife, 2020, 9, e53227.
69 J. Franco, A. Medeiros, D. Benítez, K. Perelmuter, G. Serra,
M. A. Comini and L. Scarone, Eur. J. Med. Chem., 2017, 126,
776–788.
55 T. Hayashi, M. Konishi, Y. Kobori, M. Kumada, T. Higuchi
and K. Hirotsu, J. Am. Chem. Soc., 1984, 106, 158–163.
56 P. Appelt, J. P. da Silva, O. Fuganti, L. E. Aquino, 70 J. Franco, L. Scarone and M. A. Comini, Eur. J. Med. Chem.,
B. Sandrino, K. Wohnrath, V. A. Santos, M. A. Cunha, 2020, 189, 112043.
A. Veiga, F. S. Murakami, D. F. Back and M. P. de Araujo, 71 M. E. Chapman, L. Hu, C. F. Plato and D. E. Kohan,
J. Organomet. Chem., 2017, 846, 326–334.
57 D. Benítez, A. Medeiros, L. Fiestas, E. A. Panozzo-Zenere,
Am. J. Physiol.: Renal, Fluid Electrolyte Physiol., 2010, 299,
F280–F283, DOI: 10.1152/ajprenal.00113.2010.
F. Maiwald, K. C. Prousis, M. Roussaki, T. Calogeropoulou, 72 C. Hiller, A. Nissen, D. Benítez, M. A. Comini and
A. Detsi, T. Jaeger, J. Šarlauskas, L. P. Mašič, C. Kunick, R. L. Krauth-Siegel, PLoS Pathog., 2014, 10, e1004075.
G. R. Labadie, L. Flohé and M. A. Comini, PLoS Neglected 73 S. Ebersoll, B. Musunda, T. Schmenger, N. Dirdjaja,
Trop. Dis., 2016, 10, e0004617.
M. Bonilla, B. Manta, K. Ulrich, M. A. Comini and
R. L. Krauth-Siegel, Redox Biol., 2018, 15, 532–547.
74 R. Brun and R. Kaminsky, Animal models of acute (first
stage) sleeping sickness, in Handbook of Animal Models of
58 C. Hansch, M. Streich, F. Geiger, R. M. Muir, P. P. Maloney
and T. Fujita, J. Am. Chem. Soc., 1963, 85, 2817–2824.
59 M. Dabrowska, L. Komsta, J. Krzek and K. Kokoszka,
Biomed. Chromatogr., 2015, 29, 1759–1768.
Infection,
Experimental
Models
in
Antimicrobial
60 Y. Xia and H. Zhang, Environ. Res., 2019, 30, 477–490.
Chemotherapy, 1999, pp. 789–793.
This journal is © The Royal Society of Chemistry 2021
Dalton Trans., 2021, 50, 1651–1665 | 1665