Polypharmacology of ambroxol in the treatment of COVID-19

Since December 2019, a novel viral pneumonia coronavirus disease 2019 (COVID-19), which was caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), continued to grip the world [1]. As of 9th May 2022, it has spread to more than 200 countries and territories with more than 515 million laboratory-confirmed cases worldwide, resulting in more than 6.3 million deaths since its emergence [2]. The ongoing pandemic has caused widespread economic and social disruption [3]. To prevent further morbidity and death, health organizations over the globe are working to curb the spread of COVID-19 by developing more rapid diagnosis methods for SARS-CoV-2 carriers, as well as efficacious vaccines [4,5].

COVID-19 can be well defined as a disease with two successive phases: an early viral phase characterized by influenza-like upper and lower respiratory tract illness, followed in severe instances by an inflammatory response stage. The latter condition is characterized by inflammation-driven damage to multiorgan system, particularly the lungs, which can result in acute respiratory distress syndrome and life-threatening hypoxia. Serious damage can also occur in other important organs such as the brain and blood vessels [6–8]. While some research indicates that 51.7% of the patients are asymptomatic [9] and most of the infections remain uncontrolled, 13.8% of those infected are critically ill [10].

The World Health Organization recommends systemic corticosteroids, baricitinib, as well as IL-6 receptor blockers for patients with severe or life-threatening COVID-19 [11]. Corticosteroids and long-acting bronchodilators may reduce the replication of coronaviruses, including SARS-CoV-2, and provide a relative 21% reduction in mortality, according to some laboratory evidence [12]. Elevated concentrations of IL-6 are strongly associated with severe COVID-19 outcomes. Thus, IL-6 receptor blockers, such as tocilizumab and sarilumab, can antagonize membrane-bound and soluble forms of the IL-6 receptor, block the cytokine activation, and regulate the immune response upon infection [11]. Despite these therapies, symptomatic treatment approaches remain uncertain. Many patients with severe COVID-19 experience progressive respiratory failure, including bilateral infiltrates and lung edema caused by mucus obstruction [13], which may develop to a cause of death in COVID-19. Given that ambroxol has been used clinically as a first-line drug in expectoration with high dosages of intravenous injection, it is noteworthy that ambroxol plays a vital role in the treatment of severe COVID-19. In addition to its mucoactive function, ambroxol was reported to have anti-inflammatory, antioxidant, antiviral, and antibacterial effects [12]. However, its mechanism, particularly in severe instances of COVID-19, remains unexplained. Here, by examining its pharmacology and mechanism, we investigated the therapeutic application of ambroxol in COVID-19.

The pseudotyped HIV lentivirus with the spike glycoprotein of SARS-CoV-2 that contains a Luc reporter gene was purchased from Sino Biological Inc. (Beijing, China).

HEK293T (ATCC®CRL-3216™) cells stably expressing hACE2 (HEK293T-hACE2) were generated by following a reported protocol [14]. In brief, lentivirus encoding hACE2 (Sino Biological, Beijing, China) were used to transduce HEK293T cells and selected them for 14 days with 2 μg/ml of puromycin (InvivoGen). The puromycin-resistant cells were maintained in Dulbecco’s modified eagle medium (DMEM; Gibco) supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco), 50 U/ml penicillin, 50 μg/ml streptomycin, and 1 μg/ml puromycin. The expression of hACE2 in HEK293T-hACE2 cells was determined by western blot analysis.

The HuProt™ Human Proteome chip, which contains 23156 human full-length proteins, was used in the human proteomic chip screening. Recombinant proteins with GST tags were expressed and purified by the eukaryotic expression system. For chip screening, two technical replicates were set for each protein on the chip. The chip was blocked at room temperature for 1 h with blocking solution (3% BSA in pH 7.4 PBS Buffer). The chip was then rinsed three times by 0.1% PBST and dried by centrifugation (500 g, 3 min). After that, the chip was incubated with ambroxol (100 μM, 1% BSA in PBST), biotinylated ambroxol diluent (10 μM, 1% BSA in PBST), and biotin (100 μM) as control at 37°C for 1 h. The preceding steps were repeated to wash the chip followed by a pure water rinse and centrifugation. Finally, the chip was scanned by the GenePix 4000B chip scanner. GenePix Pro v6.0 software was used to read and analyze the collected data.

A pharmacophore model’s characteristics reflect the mechanism of drug–target interaction. By fitting ambroxol to a panel of pharmacophore models, the Ligand Profiler protocol of Discovery Studio 2021 (DS; BIOVIA-Dassault Systèmes), which is equipped with PharmaDB database, was used to spot the potential targets for ambroxol.

The gene expression data of COVID-19 patients and healthy controls were downloaded from NCBI GEO database (lung, GSE182917; PBMC, GSE171110) [15,16]. All analyses were carried out with R Studio software (www.rstudio.com) in the R 4.1.1 environment. Before analysis, the original expression data were converted to transcriptomes per million reads (TPM) and transformed to log2 (TPM + 1) format. Differentially expression genes (DEGs) matrix was generated by R package limma with the Benjamini–Hochberg Method to preform P-value correction. Genes with both adjusted P-value (adj. P)<0.05 and |log2 FC| ≥ 1.5 were considered significantly differentially expressed.

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were carried out by R package clusterProfiler. Significantly enriched elements were identified as themes with adj. P<0.05. KEGG pathways were shown by R package pathview.

Molecular docking is a common tool for examination protein–small-molecule interactions, to identify receptor-binding sites, and to predict the pharmacological features of prospective small compounds in silico [17]. The molecular docking and visualization between ambroxol and neuropilin-1 (NRP-1; PDB ID: 3i97) [18] was performed following the standard protocol of CDOCKER on DS platform. The 3D structure of human neuropilin-2 (NRP-2) was downloaded from AlphaFold Protein Structure Database (AF-O60462-F1) [19]. In binding free-energy calculation, the Generalized Born model was utilized to account for the influence of the solvent. A 1000-step in situ ligand minimization of the NRP-1-ambroxol complex was performed prior to calculation.

Prior to inoculation with the pseudotyped SARS-CoV-2, the HEK293T-hACE2 cells were treated for 1 h with ambroxol at a concentration gradient ranging from 600 to 1 μM [14]. After 1 h of inoculation in the presence of each drug, the inoculum was removed and fresh medium was added for further culture. The activity of firefly luciferase was measured as a readout of infected cells using the luciferase assay (Promega) for quantitative determination at 48 h post-transduction.

In order to discover potential human protein targets of ambroxol, 23156 proteins were obtained from the HuProt™ Human Proteome Microarray, and 716 human proteins were identified via in vitro screening. Meanwhile, 15056 proteins extracted from the PharmaDB pharmacophore database were analyzed in silico and 314 human proteins were selected (Figure 1A). There were 19 proteins that overlapped between in vitro and in silico analysis, and these proteins were shown in Table 1. The protein–protein interaction (PPI) network of the intersecting proteins was established, revealing that AKT1 and HSP90AB1 are core molecules in the network (Figure 1B). For further efforts, 1012 potential protein targets were obtained by combine the results together, which were then studied by enrichment analysis. GO annotation suggested that the selected potential targets were mainly associated with intracellular signaling transduction functions such as protein serine/threonine kinase activity, protein tyrosine kinase activity, and ubiquitin-like protein ligase binding (Figure 1C). KEGG annotation indicated that the FoxO signaling pathway was most significantly enriched. In addition, many target genes were strongly associated with cancer-related pathways, such as prostate cancer and non-small-cell lung cancer (NSCLC), suggesting that ambroxol might become a reasonable candidate in antitumor treatment (Figure 1D).

By mapping ambroxol’s potential target to Coronavirus disease—COVID-19 pathway in KEGG, we found that ambroxol might interact with the SARS-CoV-2 cell entry receptor, NRP-1, on cellular surface (Figure 2A). We hypothesized that SARS-CoV-2 cell entry process could be impeded by ambroxol through binding to NRP-1. To test the hypothesis, molecular docking was utilized to investigate possible docking conformation between ambroxol and NRP-1. Ambroxol docked into the B1 domain of NRP-1 with a binding energy of −4.4 kcal/mol (Figure 2B,C). We further determined the optimal concentration of ambroxol for inhibiting the entry of SARS-CoV-2 pseudovirus. The concentration for 50% of maximal effect (EC₅₀) of ambroxol to inhibit SARS-CoV-2 from entering cells in 293T tool cells expressing high levels of ACE2 was determined to be 135.6 μM (Figure 2D). Additionally, lung adenocarcinoma cells A549 were employed to stimulate alveolar epithelial cells, and we determined that 163.7 μM ambroxol was adequate to prevent cell entry (Figure 2C). As a consequence, we deduced that ambroxol might bind to NRP-1 on cell surface to block SARS-CoV-2 entry.

NRP-2 shared sequence similarity with NRP-1 (Supplementary Figure S1A), and there was evidence that it played a similar role to NRP-1 in the process of SARS-CoV-2-entering cells [20,21]. Molecular docking for ambroxol and NRP-2 revealed that ambroxol is unable to bind to NRP-2 in a similar conformation due to the absence of NRP-1-like active cavity (red circle in Supplementary Figure S1B, also shown in Supplementary Figure S1C), despite NRP-2 being identical to NRP-1 in key amino acid. This decrease the binding free energy of ambroxol to NRP-2 to −1.6 kcal/mol, leading us to assume that the capacity of ambroxol to bind NRP-2 to block viral entrance into cells is inferior to that of NRP-1.

To evaluate the impact of ambroxol on the PBMC of COVID-19 patients, RNA-seq data from moderate to severe COVID-19 patients and healthy controls were utilized for subsequent analysis. Significant distance difference between the COVID-19 patients and healthy controls was observed via principal coordinates analysis (PCoA) (Figure 3A). The DEGs were further obtained and precented in volcano plot, with 294 significantly up-regulated and 440 down-regulated genes (Figure 3B). GO annotation of the up-regulated genes showed a predominance of associations with immunological processes such as antigen binding, immunoglobulin receptor binding (Figure 3C), and the cell cycle pathway was most significantly enriched in KEGG analysis (Figure 3D). Five pathways intersected between up-regulated pathways and ambroxol’s targets (Figure 3E), which were displayed in Figure 3F with all DEGs. Additionally, GSEA tests were conducted to determine whether the pathways were significantly up-regulated in PBMC (Figure 3G).

As cell cycle and viral carcinogenesis pathways exert high relevance with ambroxol and showed the most significant up-regulation (adj. P<0.01), the complete cell cycle diagram was presented in Figure 4, from which we could formulate a hypothesis on how ambroxol affects the cell cycle of PBMC. Ambroxol might bind to CDK2/4/6 and interfere with their functions, thus prevent cells proceed from phase G to phase S1. Additionally, E2F was likewise inhibited by ambroxol to restrain DNA synthesis, S-phase entry, and mitosis. As a consequence, these effects might antagonize SARS-CoV-2-induced inflammatory cell proliferation in PBMC, hence providing protection. Additionally, several targets including as p53, MAPK, etc., might interact with ambroxol (Supplementary Figure S2). These implied further pharmacological effects on PBMC of COVID-19 patients, which was also consistent with the anti-inflammatory effects suggested in the COVID-19 infection pathway.

Enrichment analysis was performed on the down-regulated genes as well. GO and KEGG annotations revealed that the down-regulated genes were prominent in structural constituent of ribosome and immune receptor function, as well as Ribosome and COVID-19 pathways (Supplementary Figure S3A,B). We also examined the signaling pathways on which ambroxol might affect (Supplementary Figure S3C,E). However, because these signaling pathways were down-regulated in COVID-19 patients and we usually assumed that the binding of ambroxol to its target might hinder the target’s activity, we believed that these pathways have limited relevance for the therapy of moderate-to-severe COVID-19.

Apart from evaluating the polypharmacology of ambroxol in COVID-19 PBMC, we also investigated the polypharmacology of ambroxol in COVID-19 patients’ lung tissue. Using the same methodologies and criteria, we performed PCoA analysis (Figure 5A) and searched for DEGs using RNA-seq data from postmortem lung tissue samples obtained from COVID-19 patients and those who died of nonpulmonary conditions, identifying 125 significantly up-regulated and 1354 down-regulated genes (Figure 5B). GO and KEGG annotation indicated that the up-regulated genes were predominantly associated with extracellular matrix structural constituent and glycosaminoglycan binding, and ECM-receptor interaction and protein digestion and absorption pathway was most significantly enriched (Figure 5C,D). There were five intersecting routes between up-regulated lung processes and ambroxol’s targets, as indicated by the overlapping pathways (Figure 5E,F). GSEA tests were also conducted to confirm the overall level of variation of the intersecting pathways (Figure 5G).

As referred, Bladder cancer pathway has a rather close relevance with ambroxol in COVID-19 treatment. Expansion of this pathway revealed that matrix metalloproteinases (MMPs) not only functioned as the targets of ambroxol but also as the up-regulated gene. Thus, ambroxol’s interference with MMPs might reduce the degradation of extracellular matrix, diminish endothelial cell proliferation, therefore prevent the increase in airway resistance (Figure 6).

GO analysis of down-regulated DEGs was shown in Supplementary Figure S4A. There was no substantially intersected KEGG pathways in lung tissue according to the previous threshold, indicating that ambroxol had limited effect on aggravating signaling pathways in the lung (Supplementary Figure S4B).

Ambroxol [2-amino-3,5-dibromo-N-(trans-4-hydroxycyclohexyl) benzylamine], a synthetic derivative of vasicine, is one of the over-the-counter mucolytic drugs that has been used clinically for the treatment of various respiratory diseases [22]. Because of its ability to promote bronchial secretion and clearance [23], as well as inducing surfactant synthesis from alveolar type 2 cells [24], ambroxol has been used as a mucoactive agent and a secretagogue that aids in the reduction of viscid or excessive secretions in diseases, including chronic bronchitis, chronic obstructive pulmonary disease (COPD), cystic fibrosis (CF), bronchiectasis, and asthma [25]. Therefore, intravenous ambroxol can be utilized to treat severe cases in COVID-19. Furthermore, while ambroxol’s pharmacological effects have been extensively studied, its mechanism remains an open question. The presesnt study shed light on its systematic underlying mechanism, particularly in COVID-19 treatment, and complemented its pharmacological effects. Ambroxol was discovered to produce anti-inflammatory and antiviral effects by interacting with targets including NRP-1, EGFR, etc., in multiple signaling pathways.

NRP-1, known to bind furin-cleaved substrates, serves as a host component that significantly potentiates SARS-CoV-2 infectiousness by facilitating viral entry into cells [26]. Besides human angiotensin-converting enzyme 2 (hACE-2) in the previous study [27], we discovered that ambroxol might interact with NRP-1 on the cellular membrane to prevent SARS-CoV-2 entry. In addition, we found the EC₅₀ for ambroxol was 135.6 μM in 293T cells and 163.7 μM in A549 cells. NRP-1 is highly expressed in endothelial and epithelial cells of the respiratory and olfactory epithelium [28]. Daly et al. discovered that SARS-CoV-2 used the viral C-end rule (CendR) of Spike (S) protein 1 to bind NRP-1 for attachment and entrance into host cells [29]. Thus, inhibiting this interaction may limit then entrance and infectiousness of SARS-CoV-2.

As severe COVID-19 infection is characterized by a significant inflammatory response, inhibiting the classic inflammatory pathway NF-kB/MAPK has become one of the therapeutic development strategies to alleviate local inflammatory response and ameliorate patient symptoms [30]. NF-κB is a crucial regulatory element involved in the immune-inflammatory response. Activation of NF-κB signaling triggers inflammatory cytokines, which also up-regulates the inducible nitric oxide synthase (iNOS) [31]. Previous studies have found a reduction in both protein expression and phosphorylation of NF-κB and MAPKs signaling pathways in treated with ambroxol, thus decreased the synthesis of IL-1β, IL-6, and TNF-α, as well as the production of superoxide anion, hydrogen peroxide, and nitric oxide, and the release of cellular granular enzymes such as lysozyme [31–33]. Moreover, Yamaya et al. observed that ambroxol decreased RV14 infection in part by lowering ICAM-1 and acidic endosomes via suppressing NF-κB activation [34]. It is quite likely that ambroxol diminishes inflammation by blocking the cascade of cytokines via the NF-κB signaling pathway.

We described the possibility that ambroxol interferes with the function of CDK2/4/6 and E2F. As demonstrated by prior research, CDK4/6 may phosphorylate the retinoblastoma protein, RB1, and the RB-like proteins, RBL1 and RBL2, therefore freeing the E2Fs binding to unphosphorylated RB1 and enabling S-phase entrance [35]. As a result, suppression of these genes may antagonize SARS-CoV-2-induced proliferation in PBMC and exerts ambroxol’s pharmacological impact. The pharmacological function of ambroxol can also be found in the lung tissue of severe COVID-19 patients. Components of the extracellular matrix can be degraded by MMPs in respiratory system of severe COVID-19 patients, resulting in airway remodeling, increased resistance, and permanent pulmonary fibrosis [36]. Therefore, the capacity of ambroxol to limit the effect of MMPs has the potential to improve lung function and prevent excessive hypoxia in patients with severe COVID-19.

According to our results, ambroxol shows a high potency of treating COVID-19 along with its polypharmacological benefits such as mucociliary clearance ability, anti-inflammatory effect, and adequate NRP-1-targeting ability. Importantly, it has been demonstrated that ambroxol interacts with many targets in both PBMC and the lung. Still, it is worthwhile to explore if ambroxol has any effect on SARS-CoV-2’s own proteins, such as spike protein, RNA polymerase, and master protease (Mpro), in order to acquire a better understanding of ambroxol’s effectiveness against SARS-CoV-2. There is also the possibility of combining ambroxol with other antiviral drugs and delivery systems to prolong drug maintenance or lessen adverse effects on normal tissues [37,38]. In conclusion, the present study provided evidence for molecular pathways behind ambroxol’s therapeutic utility in COVID-19 prevention and treatment. Ambroxol is a promising therapeutic candidate against SARS-CoV-2.

Data of all screened human potential protein targets underlying Figure 1 are available in Supplementary Table S1. All other data are available from the corresponding author upon reasonable requests.

The authors declare that there are no competing interests associated with the manuscript.

This work was supported by the Peking University Medicine Seed Fund for Interdisciplinary [grant number BMU2022MX017]; the Fundamental Research Funds for the Central Universities [grant numbers BMU2021MX021, BMU2022MX003]; and also supported by Boehringer Ingelheim, the COVID-19 emergency program from Jinhua in China [grant number 2020XG-26]; National Natural Science Foundation of China [grant number 81603119]; and Natural Science Foundation of Beijing Municipality [grant number 7174316].

Ziyuan Wang: Formal Analysis, Investigation, Visualization, Writing—review & editing. Minghui Yang: Data curation, Writing—original draft. Xi Chen: Conceptualization, Investigation, Methodology. Rongxin Xiao: Resources, Validation. Yu Dong: Validation, Writing—review & editing. Ming Chu: Conceptualization, Software, Funding acquisition, Project administration, Writing—review & editing. Guojie Song: Supervision. Yuedan Wang: Supervision, Funding acquisition.

The authors acknowledge Boehringer Ingelheim and Peking University Science Park for their valuable funding support.

COVID-19

coronavirus disease 2019

DEG

differentially expression gene

ECM

extracellular matrix

FBS

fetal bovine serum

GEO

Gene Expression Omnibus

GO

Gene Ontology

GSEA

Gene Set Enrichment Analysis

IL

interleukin

KEGG

Kyoto Encyclopedia of Genes and Genomes

MMP

matrix metalloproteinase

NCBI

National Center for Biotechnology Information

NF-κB

nuclear factor kappa B

NRP

neuropilin

PBMC

peripheral blood mononuclear cell

PBST

phosphate buffer solution containing Tween-20

PCoA

principal coordinates analysis

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

TLR

Toll-like receptor

TPM

transcriptomes per million