For all experiments, a sample of a production batch (EXCh. 878) of EPs ^® 7630, a dried extract of P. sidoides DC. roots (1:8–10), extraction solvent: ethanol 11% (w/w) was used. 80% of the roots used for the aforementioned production batch were collected from wild plant populations and 20% were harvested from plantations in South Africa. Prior to extraction, the dried plant material was tested in an array of DNA-based and phytochemical methods to confirm the quality and identity of the herbal material. Pharmacognosy was done by the quality control department of Dr. Willmar Schwabe GmbH and Co. KG. Voucher specimens of every lot are deposited in the Department of Pharmacognosy to be retained for 10 years. Chemical fingerprinting of the used EPs^® 7630 batch according to the Consensus statement on the Phytochemical Characterization of Medicinal Plant extracts (Heinrich et al., 2022) by three different methods (NMR, HPLC, GPC) was already carried out in our previous study (Papies et al., 2021).

Umckalin (7-Hydroxy-5,6-dimethoxy-2H-1-benzopyran-2-one, cpd A) (Table 1) and umckalin sulfate (5,6-Dimethoxy-7-(sulfooxy)-2H-1-benzopyran-2-one, cpd B) were isolated as described previously (Schötz et al., 2008). (−)-Epigallocatechin (cpd C) was purchased from Interchim S.A., France. Prodelphinidin B1 (Epigallocatechin-4β→8-gallocatechin, cpd D) and prodelphinidin B4 (Gallocatechin-4α→8-epigallocatechin, cpd E) were isolated by fractionation of the low molecular weight prodelphinidin fraction containing dimers and trimers, as previously described (Schötz and Nöldner, 2007). The single compounds were purified by dissolving the aforementioned dimer/trimer fraction in methanol to a concentration of 12% (w/v) and loading the solution on a Toyopearl HW-40S column (length = 45 cm, diameter = 2.5 cm) preconditioned with methanol, which was previously saturated with N₂ gas. Fractions were eluted from the resin by an isocratic N₂-saturated methanol flow. Fractions were checked by thin layer chromatography and pooled to yield the single compounds. (−)-Epigallocatechin gallate (cpd F) was purchased from TCI Deutschland GmbH, Germany. (+)-Taxifolin (cpd G) was purchased from Merck KGaA, Darmstadt, Germany. The purity of the purchased compounds were taken from the vendors’ specifications and had an >98% HPLC purity. The purity of the isolated substances was checked by ¹H-NMR spectroscopy (Supplementary Figure S1), using the same instrument as described previously (Papies et al., 2021). Apart from residual solvents (water, ethanol), no additional impurities could be detected for umckalin and umckalin sulfate in the NMR spectra. For prodelphinidin B1 and B4, some minor impurities from oligo-/polymeric prodelphinidins were detectable as signal bulges below the sharp signals of the respective pure dimer, in addition to residual solvent (water). Thus, the purity of all substances can be considered suitable for the study.

Test solutions were prepared as follows: EPs^® 7630 was suspended in DMEM for a stock concentration of 2 mg/mL, and serially diluted in assay medium to achieve the required working concentration. All compounds listed in Table 1 were prepared by suspension in DMSO for stock concentrations of 10 mg/mL, and serially diluted in assay medium to achieve the required working concentration. DMSO vehicle controls contained an equivalent amount of DMSO to the amount of DMSO in wells treated with the highest concentration of compound in each assay.

In vivo experiments were performed in the biosafety level three (BSL-3) facility at the Institut für Virologie, Freie Universität Berlin, Germany. Animal work was approved and executed in compliance with all applicable institutional, national and international regulations (Landesamt für Gesundheit und Soziales Berlin, permit number 0086/20).

Syrian hamsters (Mesocricetus auratus; breed RjHan:AURA) were purchased from Janvier Labs at 10 weeks of age. The animals were kept in individually ventilated cages (IVCs) in groups of 1–3 hamsters and had 1 week to get used to the housing conditions. Food and water were offered ad libitum. During the experiment, the cage temperature was constantly between 22°C and 24°C with a relative humidity between 40% and 55%.

Syrian hamsters were randomly assigned into groups of 9 animals (40%–60% female hamsters per group). Intranasal infection with 10⁵ plaque forming units (PFU) of SARS-CoV-2 (BetaCoV/Munich/BavPat1/2020) in 60 µL minimal essential medium (MEM) was performed under general anesthesia. EPs ^® 7630 was applied in strawberry syrup orally at a dose of 50 mg/kg body weight twice daily. One treatment group received the first dosage of EPs ^® 7630 1 day before infection, while the second therapeutic and vehicle treatment group were started on the day of infection. The vehicle group received strawberry syrup without EPs ^® 7630. The treatment group that started on the day of infection additionally received EPs^® 7630 intranasally at 5 mg/mL together with the virus inoculum (60 µL total volume).

The rationale to include an additional intranasal administration of EPs^® 7630 in one of the treatment groups was based on results from our previous study (Papies et al., 2021). In that study fractionation of EPs^® 7630 demonstrated highest antiviral activity in fractions containing oligomeric proanthocyanidins with expected low oral bioavailability. We assumed that local administration at the site of infection to bypass low systemic bioavailability could increase antiviral activity. We decided to include a single intranasal administration concomitantly with the virus inoculum as a first proof-of-principle approach to assess whether topical mucosal administration holds any promise as a future development option.

Infected hamsters were checked twice daily for development of clinical symptoms and body weight loss. Euthanasia was scheduled on day 2, 4 and 7 after infection. Animals were anesthetized with medetomidine (0.15 mg/kg body weight), midazolam (2 mg/kg body weight), and butorphanol (2.5 mg/kg body weight) prior to euthanasia. Lungs, serum, EDTA blood and oropharyngeal swabs were collected to conduct virological and histopathological analysis.

RNA was extracted from oropharyngeal swabs and 25 mg homogenized lung tissue using innuPREP Virus DNA/RNA Kit (Analytic Jena, Jena, Germany) according to the manufacturer’s instructions. NEB Luna universal Probe One-Step RT-qPCR Kit (New England Biolabs, Ipswich, MA, United States) was used to perform qPCR with cycling conditions of 10 min at 55°C for reverse transcription, 3 min at 94°C for activation of the enzyme, and 40 cycles of 15 s at 94°C and 30 s at 58°C on a qTower G3 cycler (Analytic Jena, Jena, Germany) in sealed qPCR 96-well plates. To monitor virus growth, SARS-CoV-2 RNA was quantified in cell culture supernatants by RT-qPCR targeting the SARS-CoV-2 E gene, as described previously (Corman et al., 2020).

To quantify replication-competent infectious virus, titrations were performed from 50 mg lung tissue and oropharyngeal swabs. For sample preparation, swabs were thawed, kept in virus transport medium (PBS with 25 mg/L enrofloxacin and 10 mg/L voriconazole) for 30 min and vortexed 3 times during incubation. The organ samples were homogenized in a bead mill procedure with ceramic beads (Analytic Jena). Thereafter, 10-fold serial dilutions were prepared starting from −1 to −6 and plated on VeroE6 cells grown in 12-well plates. The plates were incubated for 2 h at 37°C and subsequently overlaid with MEM medium containing 1.5% carboxymethylcellulose sodium (Sigma Aldrich, St. Louis, MO, United States). The plates were fixed with 4% PBS-buffered formaldehyde solution 72 h after infection. 0.75% methylene blue was used to visualize and manually count plaques. The assay-specific limit of detection is 10 PFU/50 mg tissue. All titration experiments were performed in duplicate wells. For samples without detectable plaques, a value of 5 PFU, corresponding to half the assay limit of detection was assigned to allow log-transformation of data.

The left lung lobe was prepared for histopathological examination as previously described (Osterrieder et al., 2020). After careful preparation, it was fixed in PBS-buffered 4% formaldehyde solution for 48 h, embedded in paraffin and cut at 2 μm thickness. Subsequently, the slides were stained with hematoxylin and eosin (H&E) as previously published (Bertzbach et al., 2021).

Calu-3 (ATCC HTB-55), VeroFM (ATCC CCL-81), A549-ACE2, A549-ACE2-TMPRSS2 (Widera et al., 2021) were grown in Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% non-essential amino acids, and 1% sodium pyruvate at 37°C and 5% CO₂. VeroE6 (ATCC CRL-1586) and VeroE6-TMPRSS2 (NIBSC 100978) cells were cultured in minimal essential medium (MEM) containing 10% fetal bovine serum, 100 IU/mL penicillin G, and 100 μg/mL streptomycin. For VeroE6-TMPRSS2 cell culture (NIBSC 100978), the medium also contained 1,000 μg/mL geneticin (G418) to select for cells expressing TMPRSS2. The cells were incubated at 37°C and 5% CO₂. All cell lines were cultivated under sterile laboratory conditions and tested for simian virus 5 and mycoplasma contamination as described previously (Biesold et al., 2011).

The SARS-CoV-2 strain Munich/2020/984 was isolated from a respiratory swab obtained from the early 2020 Munich patient cohort (GenBank: MT270101; GISAID: EPI_ISL_406862). The Delta AY.4 variant was isolated from a patient in Cotonou, Benin in July 2021 (GISAID: EPI_ISL_4566935) (Yadouleton et al., 2022). The Omicron BA.2 variant was isolated from a patient in Schleswig-Holstein, Germany in January 2022 (GISAID: EPI_ISL_9553926).

Stocks for animal experimentation were generated on VeroE6-TMPRSS2 cells and titrated on VeroE6 cells. Prior to animal infection, all virus stocks were stored at −80°C.

Stocks for in vitro experimentation were both generated and titrated on VeroE6 cells. For SARS-CoV-2 infection of cell cultures, between 2 × 10⁵ and 3 × 10⁵ cells per mL were seeded in 6-well plates or 24-well plates. After 24 h, cells were infected with SARS-CoV-2 in a serum-free medium. After 1 h, virus dilutions were removed, and the wells were washed twice with PBS and refilled with DMEM (supplemented as described previously). Samples were taken at the indicated time points. The full sequence identity of B.1, Delta AY.4, and Omicron BA.2 SARS-CoV-2 stocks for in vitro experiments was confirmed with NGS and RT-PCR/Sanger sequencing, and can be made available upon request. Lineage assignment was verified with the Pangolin Web Application (v4.2, pangolin-data version v1.18.1.1) (O’Toole et al., 2021).

All virus infection experiments were conducted under biosafety level 3 conditions with enhanced respiratory personal protection equipment.

SARS-CoV-2 plaque-forming units (PFU) were quantified by plaque titration on VeroE6 cells as described before (Dulbecco, 1952; Herzog et al., 2008; Forcic et al., 2010). Briefly, monolayers of VeroE6 cells were seeded in 24-well plates with ∼90% confluency and washed with PBS, incubated with serial dilutions of SARS-CoV-2-containing cell culture supernatants, and overlaid with 1.2% Avicel in DMEM 24 h after seeding. 72 h post-infection, cells were fixed with 6% formalin and visualized by crystal violet staining. The assay-specific limit of detection is 50 PFU/mL. All titration experiments were performed in duplicate wells.

Human nasal and bronchial airway epithelial cells (AEC) MucilAir™ cell cultures containing club, ciliated, and basal cells were purchased from Epithelix Sàrl (Geneva, Switzerland). Both cell cultures were derived from pooled patient material from healthy donors. Cells were cultivated in 24-well plates under air-liquid-interface conditions using transwell^® inserts and predefined serum-free MucilAir™ culture medium obtained from Epithelix at 37°C and 5% CO₂. Mucus was removed by gentle washing of the apical surface with PBS multiple times 2 days before performing experiments to ensure uniform conditions of the mucous surface in each well.

For infection of MucilAir™ cultures, cells were inoculated on the apical side with SARS-CoV-2 B.1, SARS-CoV-2 AY.4, or SARS-CoV-2 BA.2 diluted in MucilAir™ medium using an MOI of 0.005 with and without EPs^® 7630 (100 μg/mL) treatment at 37°C for 2 h. After inoculation, the virus-containing solution was removed and the apical side was washed gently with PBS three times to remove non-attached virus. For sample taking, the apical side of the AEC was incubated with 250 µL MucilAir™ medium for 20 min, which was subsequently removed and frozen at −80°C until analysis. Supernatants were analyzed by plaque assay between 0 and 72 h post-infection.

To assess cytokine levels in primary bronchial AEC (bAEC) supernatant, 25 µL of supernatant were sampled before infection and at 24 h post-infection with SARS-CoV-2. Cytokines were quantified using a Human Cytokine/Chemokine/Growth Factor Panel A 48-Plex Premixed Magnetic Bead Multiplex Assay (Merck Millipore, Burlington, MA, United States), using the Luminex MAGPIX System in 96-well plate format, according to the manufacturer’s instructions. Plate washing steps were performed using the HydroFlex Microplate Washer (Tecan, Männedorf, Switzerland). Calibration and verification checks (Bio-Techne, Minneapolis, MN, United States) were met for all of the analytes.

SARS-CoV-2 spike (S) protein-dependent viral entry was assessed using an established vesicular stomatitis virus (VSV) pseudo-particle (VSVpp) assay as described elsewhere (Kleine-Weber et al., 2019; Hoffmann et al., 2020; Zettl et al., 2020). Briefly, A549-ACE2 cells, A549-ACE2-TMPRSS2 and Calu-3 cells were seeded in DMEM in 96-well plates with a density of 50%–70% 24 h before infection with VSVpp. For pre-treatment, medium was removed 2 h before infection and replaced by fresh DMEM containing EPs^® 7630, compounds listed in Table 1, camostat mesylate, niclosamide, or DMSO as vehicle control for niclosamide at the indicated concentrations. After pre-incubation, medium was removed and fresh DMEM containing VSVpp carrying the SARS-CoV-2 S. S variants included B.1 (encoding the S protein derived from BetaCoV/Munich/BavPat1/2020; GISAID: EPI_ISL_406,862), Delta AY.117 (encoding a protein identical to GenBank: QYN98425.1), Omicron BA.2 (encoding a protein identical to GenBank: UHU97100.1), Alpha B.1.1.7 (encoding the S protein derived from BetaCoV/Baden-Wuerttemberg/ChVir21528/2020; EPI_ISL_754174), and Beta B.1.351 (encoding the S protein derived from Baden-Wuertemberg/ChVir22131/2021; EPI_ISL_862149). All wells were inoculated with the equivalent titer of respective VSVpp, which was previously found to be in the dynamic range of the luciferase assay. 19 amino acids on the C-terminus of the S protein were omitted from all variants to obtain optimal VSVpp incorporation and expression. Plates were centrifuged for 30 min at 4°C and 300 g to achieve synchronized infection. After additional incubation for 90 min at 37°C, 5% CO₂, compound-containing medium was added to the cells. To measure luciferase production, which correlates with successful viral entry, cell lysates were prepared after 24 h using passive lysis buffer (Promega, Madison, WI, United States). Lysates were then transferred to opaque 96-well plates and luminescence was measured in a BioTek multi-well plate reader using Luciferase Assay Substrate (Promega) according to the manufacturer’s recommendations.

The viability of cpd C, cpd F, and cpd G-treated Calu-3 cells was assessed using the CellTiter-Glo 2.0 Cell Viability Assay (Promega) according to the manufacturer’s instructions. Briefly, cells were seeded in 96-well plates and treated with the indicated concentrations of compound. After 48 h, cells were lysed and the luminescence signal was measured using a BioTek multi-well plate reader. Viability was calculated in relation to untreated cells and reported as percent of vehicle control.

For the purpose of determining whether treatments resulted in statistically significant outcomes compared to vehicle-treated controls, two-way ANOVAs were performed with Dunnett’s multiple comparison test (Figure 1, Figure 3, Figure 6, Figure 7, Figure 8 and Supplementary Figure S2). To compare the growth kinetics of multiple SARS-CoV-2 variants, two-way ANOVAs were performed with Tukey’s multiple comparison test (Figures 4B, Figure 5). To obtain a statistically conservative indicator of whether SARS-CoV-2 induced cytokine release was reduced upon EPs^® 7630 treatment, paired t-tests were performed (Figure 4C). Virus replication data was always log10-transformed prior to statistical analysis (Figures 1, Figure 4B, Figure 5).

Our previous data showed antiviral effects of EPs^® 7630 against SARS-CoV-2 (Bavarian strain, B.1) in human lung cells. To assess the effect of EPs^® 7630 on SARS-CoV-2 B.1 infection in vivo, we employed a previously established Syrian hamster model (Osterrieder et al., 2020; Nouailles et al., 2021). Two experimental groups were used. For the first experimental group, a prophylactic pre-treatment strategy was used, such that the hamsters received oral treatment twice daily at a dose of 50 mg/kg body weight EPs ^® 7630, beginning 24 h prior to infection. For the second experimental group, a combined oral plus intranasal therapeutic strategy was used, such that the hamsters received EPs ^® 7630 at 5 mg/kg body weight intranasally once, together with the virus inoculum in addition to oral treatment twice daily at a dose of 50 mg/kg body weight EPs ^® 7630, beginning at the time of infection. The control group was infected and received only vehicle without EPs ^® 7630. A summary of the treatment scheme and sampling times is depicted in Figure 1A.

EPs ^® 7630 showed limited antiviral activity in the hamster model. Swabs were taken to monitor virus replication in the upper respiratory tract. No significant antiviral effects were observed in the upper respiratory tract (Figures 1B,C). Lung tissue was also examined to evaluate SARS-CoV-2 replication in the lower respiratory tract (Figures 1D,E). Both oral only and oral plus i. n. treatment groups exhibited an approximately 10-fold reduction in SARS-CoV-2 PFU at 2 days post-infection (dpi) compared to vehicle-treated controls, however this difference was not apparent at 4 dpi (Figure 1E). Both the oral plus i. n. treatment group and the oral only group exhibited a statistically significant reduction in both SARS-CoV-2 PFU and RNA in the lower respiratory tract at 2 dpi, with slightly more prominent effects seen within the oral plus i. n. group (Figures 1D,E). Overall, the data suggest that EPs^® 7630 delays the onset of virus replication in the lower respiratory tract.

Lung histopathology (hematoxylin-and-eosin-stained, paraffin-embedded left lungs) was performed to determine whether EPs^® 7630 treatment also resulted in a change in COVID-19 pathology (Figure 2). Representative sections of selected observed pathologies (bronchitis and edema) were compared between the three groups (vehicle, EPs^® 7630 p. o., EPs^® 7630 i. n. and p. o.) (Figure 2). Bronchitis and bronchial epithelial cell necrosis were severe in the control group, with a delayed onset of onset of bronchitis in the EPs^® 7630 i. n. and p. o. group (Figure 2A). In comparison to vehicle and EPs^® 7630 p. o. group, perivascular and alveolar edema formation was less apparent in the EPs^® 7630 i. n. and p. o. group (Figure 2B). Regenerative changes with strong pneumocyte type 2 and bronchial epithelial hyperplasia changes were noted in all groups during the late stage of infection.

Specifically, semi-quantitative scoring was performed on representative hematoxylin-and-eosin-stained, paraffin-embedded left lung tissue samples (Figure 3). EPs^® 7630-treated hamsters had significantly lower amounts of lung area exhibiting signs of disease pathology at 4 dpi than vehicle-treated control hamsters (Figure 3A). This difference was no longer apparent at 7 dpi (Figure 3A). EPs^® 7630-treated hamsters also exhibited delayed bronchitis, such that vehicle-treated hamsters had peak bronchitis scores at 2 dpi and EPs^® 7630-treated hamsters exhibited similar peak scores at 4 dpi (Figure 3B). P.o. and i. n. EPs ^® 7630-treated hamsters had a significantly lower pneumonia score at 4 dpi, but comparable scores to controls at 7 dpi, suggesting an EPs^® 7630-mediated delay of pneumonia onset (Figure 3C). Finally, EPs^® 7630-treated hamsters showed markedly reduced lung edema compared to vehicle-treated controls at 4 dpi. This was most apparent for the combined oral plus i. n. treatment group, which continued to exhibit less edema than vehicle-treated controls at 7 dpi (Figure 3D). In summary, EPs^® 7630 treatment, especially when combined with intranasal therapy directly with the SARS-CoV-2 inoculum, resulted in delayed COVID-19 pathology, consistent with immunomodulatory activity early during the course of infection.

Our previous investigation of EPs^® 7630 indicated that it had an antiviral effect in Calu-3 cells (Papies et al., 2021). To obtain more evidence about whether EPs^® 7630 may be a suitable antiviral drug in the context of human SARS-CoV-2 infection, experiments were performed in human bAEC. EPs^® 7630 (100 μg/mL) treatment was performed in an air-liquid interface setup, as depicted in Figure 4A. Plaque assays were performed on supernatants sampled at 0, 16, 24, 48, and 72 h after SARS-CoV-2 infection. EPs^® 7630 treatment was associated with a small decrease in SARS-CoV-2 PFU at 48 h post-infection, yet peak viral titers were similar between treated and untreated controls at 72 h post-infection (Figure 4B). To evaluate whether EPs^® 7630 treatment has potential implications for human disease pathology, cytokines in bAEC supernatant were quantified at 24 h post-infection. EPs^® 7630 treatment caused a statistically significant reduction in inflammatory cytokines, notably CXCL9, CXCL10/IP-10, and TNF-α (Figure 4C). Although variance was greater, the release of chemokines CXCL1 and IL-8, as well as VEGF-A was also considerably reduced due to EPs^® 7630 treatment (Figure 4C). Overall, the bAEC data show that EPs^® 7630 has a limited antiviral effect against SARS-CoV-2 B.1, but a pronounced immunomodulatory effect that may help prevent excessive pro-inflammatory cytokine release and disease pathology.

Based on previous data suggesting that EPs^® 7630 primarily exerts its antiviral effects by limiting SARS-CoV-2 entry (Papies et al., 2021), we hypothesized that EPs^® 7630 may differentially affect SARS-CoV-2 variant Omicron, which apparently exhibits a preference for TMPRSS2-independent late endosomal entry instead of plasma membrane fusion compared to prior variants (Meng et al., 2022; Willett et al., 2022). To evaluate whether EPs^® 7630 differentially affects viruses with distinct entry routes, we performed infection experiments with SARS-CoV-2 variants Omicron BA.2 and Delta AY.4 for comparison. Additionally, we infected nasal AEC (nAEC) and bAEC to model both the upper and lower respiratory tract. Viral PFU were quantified at 0, 16, 24, 48, and 72 h post-infection.

SARS-CoV-2 Delta had slightly higher replication efficiency than SARS-CoV-2 B.1 in both nAEC and bAEC models (Figures 5A,B, B.1 data shown from Figure 4B). In nAECs, EPs^® 7630 treatment significantly reduced SARS-CoV-2 Delta replication at 24 h post-infection, such that SARS-CoV-2 Delta infection with EPs^® 7630 treatment closely resembled the kinetics of infection for SARS-CoV-2 B.1 without EPs^® 7630 treatment (Figure 5A). In bAECs, EPs^® 7630 resulted in an anomalously low SARS-CoV-2 Delta titer at 72 h post-infection, representing an approximately 100-fold reduction in PFU (Figure 5B). Overall, EPs^® 7630 seemed to have a slightly more potent effect against SARS-CoV-2 Delta than against the B.1 strain.

In contrast to SARS-CoV-2 B.1 and SARS-CoV-2 Delta strains, SARS-CoV-2 Omicron exhibited distinct replication kinetics (Figures 5C,D). Whereas previous strains had peak titers at 72 h post-infection, SARS-CoV-2 Omicron replicated more quickly and exhibited a peak titer at 24 h post-infection. EPs^® 7630 treatment resulted in a small decrease in SARS-CoV-2 replication in nAECs at 16 h post-infection, consistent with delayed virus replication (Figure 5C). EPs^® 7630 exhibited a moderate antiviral effect in bAECs against Omicron, resulting in an approximately 100-fold PFU reduction at 24 h post-infection, with reductions also visible at 16 h and 48 h post-infection (Figure 5D, B.1 data shown from Figure 4B). Thus, SARS-CoV-2 Omicron appeared to have increased susceptibility to EPs^® 7630 treatment, compared to either SARS-CoV-2 B.1 or SARS-CoV-2 Delta.

To further clarify a putative mechanism by which EPs^® 7630 inhibits SARS-CoV-2 infection, we investigated whether EPs^® 7630 treatment had differential effects on SARS-CoV-2 S-mediated entry in several cell lines: Calu-3, A549 stably transfected with ACE2, and A549 stably transfected with ACE2 and TMPRSS2 (Widera et al., 2021). Camostat mesylate is a serine protease inhibitor and is known to inhibit TMPRSS2-mediated processing of the SARS-CoV-2 S protein to promote virus entry (Hoffmann et al., 2020). Niclosamide is a protonophore that has been previously shown to inhibit SARS-CoV-2 infection, and was used as a positive control for antiviral activity (Gassen et al., 2021) and pH-dependent endosomal entry inhibition (Jurgeit et al., 2012). A VSVpp entry assay was performed using VSV-G, B.1, Delta, and Omicron S. DMSO, niclosamide, camostat mesylate, or EPs^® 7630 were administered as a pre-treatment for 2 h prior to infection and present in VSVpp inoculum. Readout was performed by quantifying luciferase activity of the cell lysates at 24 h post-infection. As expected, low pH-dependent endosomal entry of VSV-G carrying VSVpp was strongly blocked by niclosamide but moderately affected by camostat mesylate (Figure 6A). Camostat potently inhibited the entry of B.1 and Delta S VSVpp in TMPRSS2-positive cell lines, Calu-3 and A549-ACE2-TMPRSS2. In contrast, camostat mesylate did not have a strong effect in the TMPRSS2-negative cell line, A549-ACE2, exhibiting <25% inhibition (Figures 6B,C). Consistent with other reports that Omicron preferentially enters through endosomal pathway and does not depend on TMPRSS2 (Pia and Rowland-Jones, 2022), we found that TMPRSS2 did not promote SARS-CoV-2 Omicron VSVpp entry, such that the effect of camostat mesylate was comparable between A549-ACE2 and A549-ACE2-TMPRSS2 cell lines (Figure 6D). Whereas 1 μM camostat mesylate was sufficient for >50% reduction of B.1 and Delta S-based VSVpp entry in TMPRSS2-positive cell lines, Omicron S exhibited decreased sensitivity to camostat mesylate, consistent with a low level of TMPRSS2-dependent entry inhibition.

In direct contrast to camostat mesylate, EPs^® 7630 resulted in more potent inhibition of Omicron S-mediated entry and weaker inhibition of B.1 and Delta S-mediated entry. Specifically, 50 μg/mL EPs^® 7630 was sufficient to reduce Omicron S-mediated entry by approximately 75%, whereas B.1 (Figure 6B) and Delta S (Figure 6C) mediated entry was only slightly limited with inhibition in the range of 15%–65%. Although dose-dependent virus entry inhibition by EPs^® 7630 was observed in all cases, the differences between S variants was less apparent at the lower dose of 10 μg/mL, where only a slight 5%–25% reduction in VSVpp entry was observed.

To further investigate which constituents of EPs ^® 7630 might be involved in inhibition of SARS-CoV-2 entry, we performed infection experiments with seven small molecules that are known or putative constituents in EPs ^® 7630 (Table 1). The seven substances were administered together with the viral inocula and evaluated for their antiviral effect on SARS-CoV-2 B.1 in Calu-3 cells at a concentration of 10 μg/mL (Figure 7A). The constituents (−)-epigallocatechin (cpd C) and (+)-taxifolin (cpd G), and the putative constituent, (−)-epigallocatechin gallate (cpd F), resulted in a statistically significant decrease in SARS-CoV-2 PFUs at 24 h post-infection and were thus selected for further investigation. Cpd C, cpd F, and cpd G were tested for dose dependence at concentrations of 0.5, 5.0, and 10.0 μg/mL using the same treatment scheme as described above. Dose-dependent antiviral activity was observed for all three substances. At 10 μg/mL, SARS-CoV-2 PFU exhibited a mean reduction of 29.8% for cpd C, 42.6% for cpd F, and 72.2% for cpd G (Figure 7B) at non-toxic concentrations (Figures 7C,D).

To determine whether the antiviral effects of cpd C, cpd F, or cpd G are based on entry inhibition, a VSVpp entry assay was performed using VSV-G (for comparison, Figure 8A), B.1 (Figure 8B), Delta (Figure 8C), and Omicron (Figure 8D) S proteins. Vehicle control (DMEM), cpd C, cpd F, and cpd G were administered as a pre-treatment for 2 h prior to infection and present in VSVpp inoculum, and readout was performed by quantifying luciferase activity of the cell lysates at 24 h post-infection. For direct comparison, EPs ^® 7630 data from Figure 6 is also shown in this Figure (experiments were done in parallel). All three compounds exhibited dose-dependent entry inhibition. Depending on the cell type, at 100 μg/mL cpd C had a 9.8%–56.1% stronger effect against Omicron S-mediated entry (Figure 8D) compared to B.1 and Delta S proteins (Figures 8B,C). Cpd F was the most potent viral entry inhibitor, as 100 μg/mL resulted in mean reductions of between 66.5% and 98.5%, depending on the variant and cell line. Cpd G showed generally consistent levels of entry inhibition regardless of the S variant, with perhaps slightly greater inhibition of the B.1 S (Figure 8B). Interestingly, at 10 μg/mL cpd G showed more enhanced antiviral activity (see Figures 7A,B) than suggested by its entry inhibition (Figure 8), indicating a possible alternative mechanism of antiviral activity. The effects of 10 μg/mL cpd C, cpd F, and cpd G were also tested against additional SARS-CoV-2 variant S proteins, including VSVpp for Alpha B.1.1.7, Beta B.1.351, and Gamma P.1 (Supplementary Figure S2). At this concentration, Alpha, Beta, and Gamma were all inhibited comparably to other variant strains, Delta and Omicron (Figures 7C,D).