Antiviral activity of aspirin against RNA viruses of the respiratory tract-an in vitro study.

Aim Aspirin (acetylsalicylic acid) has been used for more than 115 years in medicine. Research exists to show that aspirin has antiviral effects in vitro, for example, by blocking influenza virus propagation via NF‐κB inhibition when used at high concentrations and short‐term incubation steps. The aim of this study was to confirm the antiviral activity of aspirin against influenza virus and further elucidate the activity of aspirin against other respiratory viruses. Methods Tests to detect antiviral activity were performed using plaque‐reduction assays. Aspirin was administered to the virus‐infected cell cultures one hour after infection. Prior to these assays, the non‐cytotoxic concentrations of aspirin on cells used for propagation of the respective viruses were determined. Results Aspirin was found to be highly effective against influenza A H1N1 virus. The antiviral activity against further respiratory RNA viruses was less distinct. Respiratory syncytial virus was minimally inhibited. However, the activity of aspirin against rhinoviruses was more pronounced. Aspirin demonstrated antiviral activity against all human rhinoviruses (HRV), but the effect on members of the “major group” viruses, namely HRV14 and HRV39, was greater than on those of the “minor group,” HRV1A and HRV2. Conclusions These data demonstrate a specific antiviral activity of aspirin against influenza A virus and HRV. The mode of action against rhinoviruses is still unknown and requires further investigation, as does the possibility of aspirin being effective in vivo to treat the common cold.

inflammation 14 and some modulation of signalling through transcription factor NF-κB [15][16][17] which plays a central role in many biological processes. Although much research has been carried out on the modes of action of this drug, information regarding activity against infectious agents, for example viral pathogens, is sparse. Reports exist describing antiviral activity against hepatitis C virus RNA and protein expression through COX-2 signalling pathways. 18,19 Reactivity against varicella zoster viruses, as well as cytomegalovirus infections, has also been reported. [20][21][22] Furthermore, antiviral activity against influenza A viruses was postulated to be based on the modulation of the NF-κBpathway. 23,24 Nevertheless, reports exist which describe an NF-κBindependent pathway of inhibition of viral replication with particular respect to flaviviruses 25 and hepatitis C virus. 26 Here, we present an in vitro study confirming the antiviral effect of aspirin against influenza viruses and investigating the antiviral activity of the drug on a broad panel of human pathogenic viruses, primarily involved in infections or in "superinfections" of the upper respiratory tract. The dose-dependent antiviral potency of aspirin against a variety of viral candidates was tested in vitro with plaque-reduction assays using a "therapeutic treatment" method by adding serial dilutions of aspirin to the virus-infected cell cultures one hour after infection.
Furthermore, we determined the influence of aspirin against two HRV subtypes from both the "minor group" (HRV1A, HRV2) and the "major group" (HRV14, HRV39).
In assays with adenovirus, a composition of plant-derived substances with the trade name "Sinupret ® " (Bionorica SE, Neumarkt, Germany) was included as an internal standard. 30

| Cytotoxicity tests
Analyses of the in vitro cytotoxicity of aspirin were performed on physiologically active cells using 3-(4,5-Dimethylthiazol-2-yl)-2,5diphenyltetrazoliumbromid (MTT) assays, 31 which are capable of quantifying the activity of mitochondrial enzymes demonstrating a direct correlation between viability and enzyme activity. For the determination of the limits of toxic concentrations, cells were cultivated together with varying dilutions (log 2 dilutions ranging from 10 to 0.01 mmol/L and a final log 10 dilution to 0.001 mmol/L) of aspirin and the ethanol control at 37°C and 5% CO 2 for at least 6 days. Cells cultivated in cell culture media alone were used as an additional control (medium control). Analyses with the MTT assay were performed on days 1, 3 and 6 after the addition of aspirin to the cell cultures.
For this purpose, cells were incubated for 1 hour with an MTT solution and after dissolution of the formazan crystals in DMSO (Sigma-Aldrich). The optical density (OD) of the cell culture supernatants was analysed in a photometer at 570 nm (Spectrafluor Tecan, Crailsheim, Germany). In addition, microscopic analyses of the cell monolayers were performed 6 days after the introduction of aspirin. To be classed as a positive result, the following criteria were applied to these assays: altered cell morphology, generation of intracellular vacuoles and the destruction of cell monolayers.

| Determination of the cytotoxicity
Optical density (OD) values of the non-treated medium controls were defined as 100% viability. The OD values of the supernatants derived from aspirin-treated cell cultures were calculated accordingly. A relative toxicity (% cytotoxicity) was determined using six replicates for each serial dilution. Within the respective dose-response curves, the 50% toxicity concentration (IC 50 ) was defined graphically (Fig. 1A-D).

| Assays for antiviral activity
The antiviral activity of aspirin was measured using plaque-reduction assays in plaque-forming units (PFU) for FluA H1N1, RSV, CA9, HSV-1, HRV1A, HRV2, HRV14 and HRV39 or by analysing CPE for Adeno 5. show the dose-dependent cytotoxicity determined using six replicates for each serial dilution for day 1 (blue), day 3 (red) and day 5 (green) for each of the cell lines. The standard deviations in all assays ranged between approximately 2% and 10%. IC50 values were determined graphically and are presented in Table 1 -  Table S1.

| Calculation of antiviral activity
The quantification of antiviral activity was either carried out by ana-

| Determination of in vitro cytotoxicity of aspirin
Prior to the assays determining the antiviral activity of aspirin, its cytotoxic effect on cells used for the antiviral assays had to be deter- HSV-1 and Adeno 5. This concentration was then followed by five log 2 and a final log 10 dilution of aspirin.

| Determination of the antiviral activity of aspirin
To test whether aspirin has the ability to inhibit virus replication, cells were infected with FluA H1N1, RSV, HSV-1, Adeno 5, CA9 and a variety of rhinovirus strains. A subset of the results of the respective analyses are presented in Fig. 2, showing the dose-dependent antiviral activity against FluA H1N1 ( Fig. 2A). Figure 2B illustrates some examples of original study results, presented graphically in Fig. 2A. With respect to antiviral activity against FluA H1N1, aspirin induced a con-  (Table 1 and Table S2). In Fig. 2C Table 1, presenting the effective concentrations (EC50, mmol/L) of aspirin that led to a 50% reduction in virus replication. Data derived from additional experiments are presented in Table S2.
The positive results from the analyses of the antiviral activity of aspirin against the RNA viruses, FluA H1N1 and CA9, led to the hypothesis that aspirin might have a specific antiviral activity against such viruses. Therefore, this putative antiviral effect was tested on additional non-enveloped RNA viruses: rhinoviruses. The results of these analyses are presented in Fig. 3, showing the dose-dependent antiviral activity against HRV1A (Fig. 3A), HRV2 (Fig. 3B), HRV14 (Fig. 3C) and HRV39 (Fig. 3D) for one (of 3) representative experiment. Data derived from the remaining experiments are presented in Table S2.
The antiviral activity of aspirin against HRV1A and HRV2 resulted in reductions of 48.7% for HRV1A (Fig. 3A) and 66.7% for HRV2 (Fig. 3B) at the highest concentration comparable to reduction levels achieved against CA9 (Fig. 2D). This antiviral activity was dose-dependent, with EC50 values determined for HRV1A and HRV2 at concentrations of ≥1 and 0.69 mmol/L, respectively ( Table 1). The efficacy of aspirin against HRV14 and HRV39 was particularly noteworthy (Fig. 3C,D, respectively). Against both virus strains, the addition of aspirin, at its highest concentration of 1 mmol/L, was shown to lead to a >90% reduction in plaques. This dose-dependent activity ceased at the lowest concentration of 0.01 mmol/L aspirin for HRV14 (Fig. 3C) (Table 1). Overall, these results demonstrated that aspirin exhibited a highly specific dose-dependent antiviral activity against the enterovirus CA9 and against all investigated rhinoviruses. The highest efficacy could be detected against rhinoviruses belonging to the "major group" of the rhinovirus family, namely HRV14 and HRV39.
For a more detailed illustration of the dose-dependent antiviral activity of aspirin against HRV14, photographs of original 6-well plates are displayed in Fig. S2.
The dose-dependent antiviral activity of aspirin against all viruses studied is summarised with EC50 values in Table 1. To enable a better understanding of the antiviral potential of aspirin, this table also contains the IC50 values determining the 50% cytotoxicity value of aspirin with respect to the cell lines used for virus propagation and the antiviral assays.
Further details of the dose-dependent antiviral activity of aspirin, including controls and additional experiments (study [1][2][3], are summarised in Table S2.

| DISCUSSION
In all experiments, aspirin showed a considerable dose-dependent antiviral activity against CA9, HRV1A, HRV2 and substantial activity against FluA H1N1, HRV14 and HRV39. With respect to rhinoviruses in particular, further research is required into the variation in responses observed between "major" and "minor group" strains.
The basic mechanism for this potent antiviral activity against members of the Picornaviridae and Orthomyxoviridae families is still not clear. One might argue that rhinoviruses are known to be extremely pH-sensitive and that the apparent antiviral effect of aspirin may be simply based on a reduction in pH in cell cultures. To counteract this claim, we carefully monitored pH throughout our experiments and used a very stable buffer system with phenol red as a highly sensitive indicator. In addition, this possible acid effect has also been excluded by others studying the antiviral activity of aspirin against, for example varicella zoster virus and human cytomegalovirus. [20][21][22] Another argument against a technical artefact is the fact that, in addition to the pH-sensitive rhinoviruses, the pH-insensitive enterovirus Coxsackie A9 and influenza H1N1 were also affected by aspirin treatment. In vivo experiments with influenza A viruses have suggested the involvement of the NF-κB-pathway 24 and differential regulation of influenza virus RNA synthesis by NF-κB. 33 Whether the NF-κB-pathway is involved in the reactivity against Picornaviridae or whether other cellular mechanisms play a role has yet to be verified.
This remains an important question because Liao et al. 25 were able to demonstrate that blocking of the replication of flaviviruses was independent of blocking NF-κB activation. These authors discussed the involvement of p38 MAP-kinase activation and correlated this in vitro activation with a suppression of flavivirus-induced apoptosis. 25 In the case of hepatitis C virus (HCV), an effect on the cellular receptor claudin-1 is postulated 26 leading to an effect on virus entry. However, this does not appear to be the sole effect of aspirin against HCV as others have described an additional effect on the inducible nitric oxide synthase (iNOS), which seems to be implicated in the antiviral activity of aspirin in HCV-expressing cells. 34 The effect of aspirin on the iNOS expression by downregulating the promoter activity, mRNA and protein expression levels also appears to be only part of the story, but one which correlates with the known COX-2 inhibiting can be made about the fact that, in our experiment, only RNA viruses were the target of the antiviral activity of aspirin. This suggests at least some level of specificity. Further research will be required to confirm this apparent effect, and it should be noted that the results obtained here appear to contrast with the data derived from the antiviral activity studies carried out against DNA viruses: varicella zoster virus and HCMV. [20][21][22] Another important point is the in vivo efficacy of aspirin for the treatment of influenza and other respiratory viruses. One might argue that the effective concentrations (EC50) leading to antiviral activity presented here in in vitro experiments are too high to expect a potential effect in vivo. However, the EC50 values in all cell cultures ranged from 0.69 mmol/L against HRV2 to 0.66 mmol/L for influenza A and to 0.21 and 0.1 for HRV14 and HRV39 infections, respectively. These levels are within the range of plasma levels described for aspirin; anti-inflammatory plasma levels have been described between 1.5 and 2.5 mmol/L for the total and 0.15-0.60 mmol/L for free salicylate. 35 Furthermore, it is interesting to note that toxic concentrations in vivo have been described starting at a 2 mmol/L plasma salicylate level. When we compare these toxicity data with our IC50 data, similarities are obvious, with IC50 data of 3.51 mmol/L and 4.06 mmol/L for Hep-2 and HeLa cells, respectively.
At present, we can only speculate on clinical results, but a comparison of our data with the literature appears to support the possibility of potential antiviral activity against rhinoviruses in vivo. As this question was not the focus of our in vitro studies, additional research is required in this area.