In silico screening of chloroquine analogues for compounds with more affinity for the Plasmodium falciparum chloroquine transporter as potential antimalarial drugs

Introduction

Malaria is an acute febrile infection caused by Plasmodium parasites. The parasite is transmitted through the bite of an infected female Anopheles mosquito when it is feeding on human blood. Five species of plasmodium parasites exist of which Plasmodium falciparum and Plasmodium vivax have been shown to cause the greatest morbidity and mortality in humans. Malaria cases in the African, South East Asian and Eastern Mediterranean region sare commonly caused by the P. falciparum species while the P. vivax species has been implicated mostly with malarial cases from the American region (Ocan et al., 2019).

Methods

Identification of compounds similar to chloroquine was done through in silico search where a ligand-based virtual screening for chloroquine analogues was carried out on the ZINC database (ZINC15). The search was done by feeding the canonical smiles of chloroquine into SwissSimilarity and running it. The search is based on finding compounds with similar pharmacophore, molecular fingerprint and shape of the parent drug.

Compounds from the search results with a similarity index of 95% were isolated and then drawn using Pubchem sketcher V2.4. Three-dimensional models of the compounds were generated through Avogadro version 1.1.0 after which they were prepared through optimization in Avogadro and minimization in Chimera version 1.14c. The 3D structure of the pfCRT protein was retrieved from the Protein Databank-101 (accession number: 2B9L https://www.rcsb.org/) and standardized using Chimera 1.14c. Finally, surface binding analysis of the prepared compounds was carried out by docking them with the standardized pfCRT protein. The docking scores showing the binding strength of the compounds with the protein were generated and compared with that of chloroquine.

In addition, prediction of the pharmacokinetic profile of the selected compounds was done using the web tool SWISSADME and also prediction of other likely binding sites of the drug within the body was carried out using the web tool SwissTargetPrediction.

Results

Discussion

P. falciparum resistance to chloroquine is not attributed to drug modification or inactivation, but rather due to increase in efflux of the drug from the digestive vacuole of the parasite (Chinappi et al., 2010; Sidhu et al., 2002). This decreases the concentration of the drug within the vacuole such that it becomes ineffective in completely inhibiting incorporation of heme to hemozoin. The mutant pfCRT protein also binds verapamil and desipramine resulting in reversing of the resistance induced by the mutant pfCRT (Bellanca et al., 2014; Lakshmanan et al., 2005). However, the challenge of using these chemo-sensitizers is that a higher concentration is required to achieve optimal reversal effect and this higher concentration is associated with more adverse effects, seemingly because they are already established drugs acting on different receptors for different functions (Martin et al., 1987).

Since the mutant pfCRT resistance can be reversed by binding to other compounds, we hypothesized that developing a chloroquine analogue with a higher binding affinity for the protein could be co-administered with the parent compound itself (chloroquine) (Kim et al., 2019; Patel and Roy, 2021). This in turn will lead to competitive racing for the binding site and subsequent elimination from the vacuole. The chloroquine analogue with a high affinity and binding capacity for protein will competitively inhibit binding of chloroquine onto the protein, decreasing its efflux from the vacuole. As such, the concentration of chloroquine will be maintained and even elevated, ensuring maximal efficacy. Furthermore, the fact that these analogues are similarly related to chloroquine means they will also prevent incorporation of heme to hemozoin, adding to the efficacy of the parent compound.

Twenty compounds out of the ZINC database search results had a chloroquine similarity index of > 95%. Twelve of these compounds showed a higher binding affinity to the mutant pfCRT protein. Four of these twelve compounds had binding affinity less than -8.0 compared to -7.0 of chloroquine with the compound ZINC01596768 having the greatest binding strength of -8.3. This strong interaction between the protein and the analogues could be useful in ensuring competitive inhibition of pfCRT binding. The other analogues were ZINC38050614, ZINC38050617, and ZINC38050615 with binding interaction strengths of -8.0, -8.2 and -8.2 respectively.

The pharmacokinetic profile prediction showed all the twelve compounds followed Lipinski rules, had no alerts on the PAINS criteria and in terms of lead likeness, they had violated the rule of XLOGP3 > 3.5 (Daina et al., 2017). They also had a high GI absorption, and thus can be formulated as oral drugs. However, three of these twelve compounds (ZINC38050614, ZINC38050617, and ZINC38050615) are predicted to be substrates of P-glycoprotein and thus, even though they had a high GI absorption, their bioavailability could be reduced due to efflux by the glycoprotein (Kwon et al., 2004). All the compounds were predicted to be permeant to the blood brain barrier and could potentially cause CNS side effects. When it comes to CYP450 inhibition, all the twelve compounds inhibited the enzymes CYP1A2 and CYP26. In contrast to this, all the twelve compounds did not inhibit CYP1A2 but showed mixed results when it comes to inhibition of CYP2C19. Compounds ZINC01873617 and ZINC96331701 did inhibit CYP3A4 while the rest of the compounds did not inhibit the enzyme. The Lipophilicity log P value between 1 and 4 usually provides more optimal physiochemical properties (Benet et al., 2016). Of the twelve selected compounds, compound ZINC01596768, which had the highest docking score, had a log P of 4.15 which, is slightly above the optimal range.

The synthetic accessibility score for the twelve compounds were all below 3.07, nearing one and as such were easy to synthesize in the laboratory (Daina et al., 2017). Other target prediction by the SwissTargetPrediction tool showed that the compounds interacted less with receptors or enzymes within the body. The compound ZINC01596768 with the highest docking score showed a probability of more than 19% of binding to histamine N-methyl-transferase, histamine 3 receptor and Quinone reductase 2 enzyme. The remaining three compounds with docking score less than -8.0 only interacted with histamine 3 receptor at a probability of more than 19%.

Conclusions

Thus, compound ZINC01596768 could potentially be co-formulated with chloroquine in the treatment of malaria in chloroquine- resistant cases.

Data availability