DNDI-6174, a preclinical candidate for visceral leishmaniasis that targets the cytochrome bc1 complex

New drugs for visceral leishmaniasis that are safe, low cost and adapted to the field are urgently required. Despite concerted efforts over the last several years, the number of new chemical entities with novel mechanisms of action that are suitable for clinical development remains low. Here, we describe the development of DNDI-6174, an inhibitor of Leishmania cytochrome b that originated from a phenotypically-identified pyrrolopyrimidine series. This compound fulfills all Target Candidate Profile criteria required for progression into preclinical development. In addition to good metabolic stability and pharmacokinetic properties, DNDI-6174 demonstrates potent in vitro activity against a variety of Leishmania species and can reduce parasite burden in animal models of infection, with potential for sterile cure. No significant flags were identified in preliminary safety studies, including an exploratory 14-day toxicology study in the rat. DNDI-6174 represents the first cytochrome b inhibitor to enter preclinical development for visceral leishmaniasis.


Fig. S2
Binding mode of DNDI-6174 in the Qi sites of mutated versions of cytochrome b (A-E).Ser35Asn and Ser206Asn mutations identified in cell line Res 1 are reported in magenta and green.Mutation Asp231Glu identified in Res 2 is represented in orange.The Ser207Pro mutation identified in Res 3 is in light blue.Gly31Ala (Res 4) and Ser206Asn (Res 5) are reported in wheat and green, respectively.The mutation of 206 from Ser to Asn dislodges the ligand from its binding site, breaking a critical H-bond with Asp231 and thus contributing to the decrease in ligand binding stability compared to that seen with the wild type enzyme.The Ser35Asn mutation that accompanies Ser206Asn is not located in the binding site, and its role in drug resistance seems to be the result of an indirect interaction between the residue and the ligand.Our MD analysis strongly suggests that the longer and more flexible Glu231 side chain bends away from the ligand.This new conformation of the Glu231 side chain results in a complete loss of interactions with the ligand, reflected in a 164fold reduction in DNDI-6174 potency compared to the wild-type (Table 2).Our data indicates that the resistance observed in the Gly31Ala mutated results in steric clashes.The methyl group of the Ala side chain displaces the conserved water molecule bridging the interaction between the ligand and Phe34 and disrupts the interaction with Asp231 by clashes with the NH2 of the 2aminopyrrolopyrimidine core.

Fig. S9
Degradation profiles for DNDI-6174 incubated with human liver microsomes in the absence and presence of inhibitors specific for individual CYP isoforms.
There was no significant difference in the degradation slopes (α = 0.05) in the absence and presence of inhibitor for any isoform.

Fig. S10
Metabolite formation profiles for DNDI-6174 metabolites formed following incubation with human liver microsomes in the absence and presence of inhibitors specific for individual CYP isoforms.

Fig. S11
Experimental plasma PK data following (A) twice-daily oral administration (doses given at 6 and 24 h) for 5 days in mice and (B) once-daily oral administration for 5 days in hamsters.Note that at the highest dose of 25 (mice) and 47.3 (hamsters) mg/kg, only a single dose was administered.Symbols represent the measured data (mean ± SD, n=3) and lines represent the best-fit of the data using a one compartment body model.

Fig. S12
Plasma AUC24 ss and Cmax for hamsters (qd dosing for 5 days) and mice (bid dosing at 8 and 24 h for 5 days) based on compartmental fits of the experimental data (black symbols ± SE).Experimental data on day 1 and day 5 of dosing are shown with the blue and red symbols, respectively.

Fig. S13
Simulated repeat dose plasma profiles for DNDI-6174 in mice and hamsters following once or twice-daily oral administration (8 and 24 h) for 5 days as used in the efficacy studies.

Fig. S14
Simulated human plasma concentration vs time profiles to achieve a cumulative plasma AUC of between 60 and 460 µg.h/mL.Profiles were simulated using GastroPlus and the parameters shown in Table S16.

Table S1
In vivo efficacy of DNDI-6174 and positive controls (AmBisome or miltefosine) in (A) the acute mouse model and (B) the chronic hamster model.Data for organ burden are expressed as a % of the vehicle control in the same experiment (mean n=5 ± SEM).

Table S2
Results for the promastigote transformation assay where organs from hamsters treated with DNDI-6174, miltefosine or vehicle control were cultured in vitro and monitored for the emergence of viable parasites (promastigotes).Results represent an arbitrary parasite score attributed (+, ++ or +++) based on visual inspection of parasite density in the positive wells.A score of "-"is attributed in the absence of parasite.Scores are reported individually, 7 days post autopsy, from the three target organs (liver, spleen and bone-marrow).

Table S3
Summary of read counts and coverage for whole genome sequencing of DNDI-6174-resistant clones.

Table S4
Summary of non-synonymous SNPs identified in whole genome sequencing of DNDI-6174-resistant parasites.

Table S6
In vitro intrinsic clearance (CLint, ± standard error of estimate) of DNDI-6174 following incubation with liver microsomes (2 independent experiments) and cryopreserved hepatocytes (single experiment) and predicted in vivo plasma clearance.The measured in vivo plasma clearance from Table S7 is shown for reference.

Table S7
Intravenous and oral plasma pharmacokinetic properties of DNDI-6174 in mice, rats, and dogs following single dose administration.

Table S8
Oral plasma exposure of DNDI-6174 in rats following a single dose (mean n=3 ± S.D.).

Table S9
Oral plasma exposure of DNDI-6174 in dogs following a single dose (mean n=3 ± S.D.) Table S10 Oral plasma exposure of DNDI-6174 in mice following twice daily (at 6 and 24 h) dosing for 5 days (mean n=3 ± S.D.).
Table S11 Oral plasma exposure of DNDI-6174 in hamsters following once daily dosing for 5 days (mean of n=3 ± S.D.).
Table S12 Summary of fitted plasma compartmental parameters for DNDI-6174 following single oral dosing to mice and hamsters.
Table S13 Pharmacodynamic data (liver burden, mean n=5 ± SEM) for mice infected with L. infantum or L. donovani and treated with DNDI-6174.For the liver burden data, the SEM is shown in parentheses and for the plasma PK parameters, the unbound values are shown in parentheses.Pharmacokinetic parameters are from the fitted analysis as described above.
Table S14 Pharmacodynamic data (liver burden, mean n=5 ± SEM, Table S1) for DNDI-6174 in hamsters infected with L. infantum.Pharmacokinetic parameters are from the fitted analysis as described above.Unbound PK parameters are shown in parentheses.
Table S15 Best fit parameters for the data shown in Figure 4 obtained using a 4-parameter logistic function.Values in parentheses represent the standard errors of the fitted parameters.
Table S16 Input parameters for GastroPlus simulations.

Table S17
Early in silico and in vitro cardiotoxicity assessment of DNDI-6174.
Table S19 DNDI-6174 profile in a panel mammalian receptors, enzymes and ion channels.
Table S20 Assessment of human complex III activity and mitochondrial toxicity.
Table S21 Cytochrome P450 inhibition by DNDI-6174.with Asp231 and thus contributing to the decrease in ligand binding stability compared to that seen with the wild type enzyme.The Ser35Asn mutation that accompanies Ser206Asn is not located in the binding site, and its role in drug resistance seems to be the result of an indirect interaction between the residue and the ligand.Table S13 Pharmacodynamic data (liver burden, mean n=5 ± SEM) for mice infected with L. infantum or L. donovani and treated with DNDI-6174.For the liver burden data, the SEM is shown in parentheses and for the plasma PK parameters, the unbound values are shown in parentheses.
Pharmacokinetic parameters are from the fitted analysis as described above.
Fig. S3 (A) The Ligand Root Mean Square Fluctuation (RMSF) for each atom of DNDI-6174 in cytochrome b (wild-type) measuring the changes in the position of ligand atoms during the 100 ns MD simulation (atom number as reported in the chemical representation on the right).These studies indicate that benzodioxole moiety of DNDI-6174 is the portion of the molecule with the highest flexibility during binding.(B) A schematic of detailed ligand atom interactions with the amino acid residues of cytochrome b. (C) Root Mean Square Deviation (RMSD) plot of DNDI-6174 (right Y-axismagenta curve) and wild-type cytochrome b (left Yaxisblue curve).The plot indicates the stability of the protein and the ligand during the simulation.The overall stability of the docking binding pose is highlighted by the protein and ligand RMSD over this simulation.(D) Protein-Ligand Contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.
Fig. S4 (A) The Ligand Root Mean Square Fluctuation (RMSF) of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b (Ser35Asn/Ser206Asn) from Res 1. Ligand flexibility in this mutated enzyme increased considerably, even for the pyrrolopyrimidine scaffold that was particularly tightly bound in the wild-type enzyme.RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) Root Mean Square Deviation (RMSD) plot of DNDI-6174 and cytochrome b. (D) Protein-Ligand Contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the Fig. S5 (A) RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 5 (Ser206Asn).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.See Fig. S4 legend for details of the impact of this mutation on ligand stability.Fig. S6 (A) RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 2 (Asp231Glu).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.Our MD analysis strongly suggests that the longer and more flexible Glu231 side chain bends away from the ligand.This new conformation of the Glu231 side chain results in a complete loss of interactions with the ligand, reflected in a 164fold reduction in DNDI-6174 potency compared to the wild-type (Table2).

Fig. S7 (
Fig. S7 (A) RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 3 (Ser207Pro).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.The Ser207Pro mutation impacts ligand binding by changing the morphology of the binding site.The mutation causes the rearrangement of secondary structure elements, ultimately disrupting the key H-bonds interactions between the Asp231 side chain and the 2-amino group of DNDI-6174.

Fig. S8
Fig. S8 RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 4 (Gly31Ala).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.Our data indicates that the resistance observed in the Gly31Ala mutated results in steric clashes.The methyl group of the Ala side chain displaces the conserved water molecule bridging the interaction between the ligand and Phe34 and disrupts the interaction with Asp231 by clashes with the NH2 of the 2aminopyrrolopyrimidine core.

Fig. S2
Fig. S2 Binding mode of DNDI-6174 in the Qi sites of mutated versions of cytochrome b (A-

Fig
Fig. S4 (A) The Ligand Root Mean Square Fluctuation (RMSF) of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b (Ser35Asn/Ser206Asn) from Res 1. Ligand flexibility in this mutated enzyme increased considerably, even for the pyrrolopyrimidine scaffold that was particularly tightly bound in the wild-type enzyme.RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) Root Mean Square Deviation (RMSD) plot of DNDI-6174 and cytochrome b. (D) Protein-Ligand Contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.The mutation of 206 from Ser to Asn dislodges the ligand from its binding site, breaking a critical H-bond

Fig
Fig. S5 (A) RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 5 (Ser206Asn).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.See Fig. S4 legend for details of the impact of this mutation on ligand stability.

Fig
Fig. S6 (A) RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 2 (Asp231Glu).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.Our MD analysis strongly suggests that the longer and more flexible Glu231 side chain bends away from the ligand.This new conformation of the Glu231 side chain results in a complete loss of interactions with the ligand, reflected in a 164-fold reduction in DNDI-6174 potency compared to the wild-type (Table2).

Fig. S8
Fig. S8 RMSF of DNDI-6174 during the 100 ns MD simulation in the mutated cytochrome b from Res 4 (Gly31Ala).RMSF is reported by atom number as reported in the chemical representation on the right.(B) A schematic of detailed ligand atom interactions with the protein residues.(C) RMSD plot of DNDI-6174 and cytochrome b. (D) Protein-ligand contacts plot.The possible interactions are categorized into four types: hydrogen bonds, hydrophobic, ionic and water bridges.The stacked bar charts are normalized over the course of the trajectory.Our data indicates that the resistance observed in the Gly31Ala mutated results in steric clashes.The methyl group of the Ala side chain displaces the conserved water molecule bridging the interaction between the ligand and Phe34 and disrupts the interaction with Asp231 by clashes with the NH2 of the 2-aminopyrrolopyrimidine core.

Fig. S9
Fig. S9 Degradation profiles for DNDI-6174 incubated with human liver microsomes in the

Fig. S11
Fig. S11 Experimental plasma PK data following (A) twice-daily oral administration (doses

Fig. S12
Fig. S12 Plasma AUC24 ss and Cmax for hamsters (qd dosing for 5 days) and mice (bid dosing

Fig. S14 3M
Fig. S14 Simulated human plasma concentration vs time profiles to achieve a cumulative

Table S1
In vivo efficacy of DNDI-6174 and positive controls (AmBisome or miltefosine) in (A) the acute mouse model and (B) the chronic hamster model.Data for organ burden are expressed as a % of the vehicle control in the same experiment (mean n=5 ± SEM).

Table S2
Results for the promastigote back-transformation assay where organs from hamsters treated with DNDI-6174, miltefosine or vehicle control were cultured in vitro and monitored for the emergence of viable promastigotes.Results represent an arbitrary parasite score attributed (+, ++ or +++) based on visual inspection of parasite density in the positive wells.A score of "-"is attributed in the absence of parasite.Scores are reported individually, 7 days post autopsy, from the three target organs (liver, spleen and bone-marrow).

Table S3
Summary of read counts and coverage for whole genome sequencing of DNDI-6174-resistant clones.

Table S4
Summary of non-synonymous SNPs identified in whole genome sequencing of DNDI-6174-resistant parasites

Table S6
In vitro intrinsic clearance (CLint, ± standard error of estimate) of DNDI-6174 following incubation with liver microsomes (2 independent experiments) and cryopreserved hepatocytes (single experiment) and predicted in vivo plasma clearance.The measured in vivo plasma clearance from TableS7is shown for reference.Table S7Intravenous and oral plasma pharmacokinetic properties of DNDI-6174 in mice, rats, and dogs following single dose administration a Sparse sampling with n=2 mice per time point; data are based on the mean data b Monash University; nominal dose (n=3 rats) c WuXi AppTec; nominal dose (n=3 rats, n=2 dogs, n=3 hamsters)

Table S8
Oral plasma exposure of DNDI-6174 in rats following a single dose (mean n=3 ±

Table S9
Oral plasma exposure of DNDI-6174 in dogs following a single dose (mean n=3 ±

Table S10
Oral plasma exposure of DNDI-6174 in mice following twice daily (at 6 and 24 h)

Table S11
a WuXi AppTec; nominal dose b Single dose on day 1 only c Sparse sampling with n=3 samples per time point

Table S12
Summary of fitted plasma compartmental parameters for DNDI-6174 following single oral dosing to mice and hamsters.

Table S14
Pharmacodynamic data (liver burden, mean n=5 ± SEM (TableS2)) for DNDI-6174 in hamsters infected with L. infantum.Pharmacokinetic parameters are from the fitted analysis as described above.Unbound PK parameters are shown in parentheses.
*n=4 for each dose (outliers based on Grubbs' test excluded from means and SEM)

Table S15
Best fit parameters for the data shown in Figure4obtained using a 4-parameter logistic function.Values in parentheses represent the standard errors of the fitted parameters.

Table S16
Input parameters for GastroPlus simulationsTable S18 Complete in vitro cytotoxicity profiling of DNDI-6174Each value is the mean of at least two independent assays, conducted in duplicate at each concentration.

Table S19
DNDI-6174 profile in a panel mammalian receptors, enzymes and ion channelsTable S20Assessment of human complex III activity and mitochondrial toxicity.Human complex III EC50 values are the mean of at least two biological replicates consisting of at least two technical replicates (n ≥ 2).MEC= Minimum Effective Concentration, was defined as the concentration that significantly crosses vehicle control threshold.