Active Components of Commonly Prescribed Medicines Affect Influenza A Virus-Host Cell Interaction: A Pilot Study

Abstract

Background: Every year, millions of people are hospitalized and thousands die from influenza A virus (FLUAV) infection. Most cases of hospitalizations and death occur among the elderly. Many of these elderly patients are reliant on medical treatment of underlying chronic diseases, such as arthritis, diabetes, and hypertension. We hypothesized that the commonly prescribed medicines for treatment of underlying chronic diseases can affect host responses to FLUAV infection and thus contribute to the morbidity and mortality associated with influenza. Therefore, the aim of this study was to examine whether commonly prescribed medicines could affect host responses to virus infection in vitro. Methods: We first identified 45 active compounds from a list of commonly prescribed medicines. Then, we constructed a drug-target interaction network and identified the potential implication of these interactions for FLUAV-host cell interplay. Finally, we tested the effect of 45 drugs on the viability, transcription, and metabolism of mock- and FLUAV-infected human retinal pigment epithelial (RPE) cells. Results: In silico drug-target interaction analysis revealed that drugs such as atorvastatin, candesartan, and hydroxocobalamin could target and modulate FLUAV-host cell interaction. In vitro experiments showed that at non-cytotoxic concentrations, these compounds affected the transcription and metabolism of FLUAV- and mock-infected cells. Conclusion: Many commonly prescribed drugs were found to modulate FLUAV-host cell interactions in silico and in vitro and could therefore affect their interplay in vivo, thus contributing to the morbidity and mortality of patients with influenza virus infections.

Keywords: commonly prescribed drugs; drug adverse reaction; influenza virus; virus–host interaction.

Figure 1

Active compounds of 45 most prescribed drugs clustered based on their structural similarity, calculated by ECPF4 fingerprints and the Tanimoto coefficient.

Figure 2

Direct and downstream cellular targets of 45 active components of commonly prescribed medicines. Targets associated with FLUAV replication are marked with red dashed circles, and interactions between them and commonly prescribed drugs are highlighted.

Figure 3

Effect of 45 active compounds of commonly prescribed drugs on the viability of mock- and FLUAV-infected RPE cells. CC₅₀ and EC₅₀ values are shown in blue and orange, respectively (mean, n = 3). CC₅₀ = 100 means >100 μM.

Figure 4

Effect of 45 compounds on polyadenylated host RNAs in RPE cells. RPE cells were treated with 10 µM compounds or remained non-treated. After 8 h, total RNA was isolated, and a fraction of polyadenylated RNA was sequenced. A heatmap of 70 most variable mRNAs affected by treatment is shown. Rows represent gene symbols, columns represent treatments with drugs, clustered based on structural similarity. Drugs marked in bold are compounds which were predicted to affect FLUAV–host cell interactions. Each cell is colored according to the log2-transformed expression values of the samples, expressed as fold change relative to the non-treated control.

Figure 5

Effect of 45 compounds on polyadenylated host RNAs in FLUAV-infected RPE cells. RPE cells were treated with 10 µM compounds and infected with FLUAV at moi 1. After 8 h, total RNA was isolated, and a fraction of polyadenylated RNA was sequenced. A heatmap of the most variable genes affected by FLUAV infection is shown (2.5 < log2FC < −2.5). Rows represent gene symbols, columns represent treatments with drugs, clustered based on structural similarity. Drugs marked in bold are compounds which were predicted to affect FLUAV–host cell interactions. Each cell is colored according to the log2-transformed expression values of the samples, expressed as fold change relative to the non-treated mock-infected control.

Figure 6

Effect of 45 compounds on polyadenylated viral RNAs in FLUAV-infected RPE cells. RPE cells were treated with 10 µM compounds or remained non-treated and infected with FLUAV at moi 1. After 8 h, total RNA was isolated, and a fraction of polyadenylated RNA was sequenced. A heatmap of viral RNAs affected by treatment is shown. Rows represent gene symbols, columns represent treatments with drugs, clustered based on structural similarity. Drugs marked in bold are compounds which were predicted to affect FLUAV–host cell interactions. Each cell is colored according to the log2-transformed expression values of the samples, expressed as fold change relative to the non-treated FLUAV-infected control.

Figure 7

Effect of 45 compounds on metabolism of polar molecules in mock-infected RPE cells. The cells were treated with 10 µM compounds or remained non-treated. After 24 h, the media were collected, and polar metabolites were analyzed using LC-MS/MS. A heatmap of 50 most variable metabolites affected by treatment is shown. Rows represent metabolites, columns represent treatments with drugs, clustered based on structural similarity. Drugs marked in bold are compounds which were predicted to affect FLUAV–host cell interactions. Each cell is colored according to the log2-transformed and quantile-normalized values of the samples, expressed as fold change relative to the non-treated control.

Figure 8

Effect of 45 compounds on metabolism of FLUAV-infected RPE cells. The cells were treated with 10 µM compounds or remained non-treated. After 24 h, the media were collected, and polar metabolites were analyzed using LC-MS/MS. A heatmap of 33 most variable metabolites affected by treatment is shown. Rows represent metabolites, columns represent treatments with drugs, clustered based on structural similarity. Drugs marked in bold are compounds which were predicted to affect FLUAV–host cell interactions. Each cell is colored according to the log2-transformed and quantile-normalized values of the samples, expressed as fold change relative to the non-treated mock-infected control.