Transcriptomics-Guided In Silico Drug Repurposing: Identifying New Candidates with Dual-Stage Antiplasmodial Activity

Malaria, a parasitic disease primarily caused by Plasmodium falciparum and P. vivax, remains a significant global health concern, with an estimated 247 million cases and 619,000 deaths in 2021. The parasite's complex life cycle, involving both human and mosquito hosts, necessitates effective treatment strategies. Current antimalarial drugs face increasing resistance, highlighting the urgent need for new therapies. This study employs a drug repurposing approach, leveraging transcriptomics data and computational methods to identify potential new antimalarial candidates. Drug repurposing, identifying new uses for existing drugs, significantly reduces the time and cost associated with traditional drug discovery. Chemogenomics, integrating chemical and biological data, further enhances this approach. This research specifically focused on identifying highly expressed genes in all life stages of P. falciparum to uncover potential drug targets and repurpose existing drugs with activity against these targets. The researchers developed a bioinformatics workflow combining transcriptomic analysis, drug target database searches, and machine learning predictions to identify and experimentally validate potential antimalarial candidates.

The introduction extensively reviews the burden of malaria, the parasite's life cycle, existing treatment challenges (drug resistance), and the rationale for drug repurposing and chemogenomics approaches. It cites several relevant articles on antimalarial drug development, drug repurposing strategies, and chemogenomic applications in drug discovery. The literature supports the critical need for new antimalarial drugs and the potential of computational approaches to accelerate this process.

The study employed a multi-step bioinformatics workflow. First, they identified highly expressed genes (≥80% expression) across all life stages of P. falciparum (asexual blood, gametocyte, liver, and sexual stages in the mosquito vector) using transcriptomic data from publicly available databases (Lopez-Barragan, Bunnik, Zanghi, Lasonder, Cubi). This resulted in 674 overlapping genes, including 409 essential genes based on PlasmoDB data. These genes were then searched in the Therapeutic Target Database (TTD) and DrugBank to identify potential drug targets and associated bioactive compounds. The researchers found 70 homologous targets linked to 75 bioactive compounds. The Chemical Checker tool was then utilized to expand the list by identifying compounds with similar bioactivity profiles to the initial 75 compounds, expanding the pool to 1632 potential compounds. Previously developed machine learning (QSAR) models were employed to predict the activity of these compounds against different P. falciparum life stages (asexual blood stages – 3D7 and W2 strains, ookinete, gametocyte, and liver stages). This narrowed down the candidates to 76 compounds, from which two, NVP_HSP990 and silvestrol aglycone, were selected for experimental validation based on commercial availability. In vitro assays were then conducted to assess the antimalarial activity of these two compounds against chloroquine-sensitive (3D7) and multidrug-resistant (Dd2) P. falciparum strains using SYBR Green-based assays. Cytotoxicity was evaluated in mammalian cells (COS-7 and HepG2) using MTT assays. The transmission-blocking potential of silvestrol aglycone was assessed using an ookinete conversion inhibition assay with P. berghei.

The key findings include the identification of 674 overlapping highly expressed genes across multiple P. falciparum life stages, of which 409 were essential. This led to the identification of 70 potential drug targets and 75 associated bioactive compounds. The use of Chemical Checker significantly expanded the number of potential candidates. Machine learning models further refined the selection, highlighting 76 compounds with predicted activity against multiple life stages. Experimental validation revealed that both NVP-HSP990 and silvestrol aglycone exhibited potent inhibitory activity against the asexual blood stage of P. falciparum at nanomolar concentrations in both chloroquine-sensitive (3D7) and multidrug-resistant (Dd2) strains. Importantly, silvestrol aglycone displayed a similar EC50 value to established antimalarials (e.g., chloroquine, artesunate) and exhibited low cytotoxicity. Furthermore, silvestrol aglycone showed potent transmission-blocking activity (99.65% inhibition of ookinete conversion at 10 μM). The EC50 values for both compounds against different P. falciparum strains, cytotoxicity in mammalian cells, and the transmission-blocking activity of silvestrol aglycone are detailed in Table 1 and supplementary figures.

The study successfully combined transcriptomics, drug repurposing, and machine learning to identify potential antimalarial drug candidates. Silvestrol aglycone emerged as a particularly promising candidate, displaying potent activity against both the asexual blood stage and the sexual stage (transmission-blocking), coupled with low cytotoxicity. This dual-action potential is crucial for malaria control, targeting both disease progression and transmission. The findings support the viability of the employed in silico drug repurposing strategy and its potential for accelerating antimalarial drug discovery. The integration of multiple datasets and computational tools enhanced the efficiency of the drug discovery process.

This study demonstrates the power of a transcriptomics-guided, in silico drug repurposing strategy for identifying novel antimalarial candidates. Silvestrol aglycone, identified through this approach, shows significant promise as a dual-acting antimalarial with transmission-blocking properties, warranting further preclinical and clinical investigation. This workflow could be adapted for identifying drugs against other infectious diseases.

The study's limitations include the reliance on publicly available transcriptomic data, which may not fully capture the complexity of P. falciparum gene expression. The in silico predictions were validated in vitro only against two strains of P. falciparum, and further validation with additional strains and in vivo studies are necessary. The use of the murine model P. berghei for transmission-blocking assay may not perfectly reflect the situation in humans. The limited number of compounds experimentally tested restricts the conclusions that can be made regarding the generalizability of the results.