The combined treatment of Molnupiravir and Favipiravir results in a marked potentiation of antiviral efficacy in a SARS-CoV-2 hamster infection model

Favipiravir and Molnupiravir, orally available antivirals, have been reported to exert antiviral activity against SARS-CoV2. In recent days preliminary efficacy data have been reported in COVID-19 patients. We here studied the combined antiviral effect of the drugs in the SARS-CoV2 hamster infection model. We first demonstrate that Molnupiravir can reduce infectious virus titers in lungs of infected animals in a dose-dependent manner by up to 3.5 log10 which is associated with a marked improvement of virus-induced lung pathology. When animals are treated with a combination of suboptimal doses of Molnupiravir and Favipiravir (that each alone result in respectively a 1.3 log10 and 1.1 log10 reduction of infectious virus titers in the lungs), a marked combined potency is observed. Infectious virus titers in the lungs of animals treated with the combo are on average reduced by 4.5 log10 and infectious virus are no longer detected in the lungs of 60% of treated infected animals. Both drugs result in an increased mutation frequency of the remaining viral RNA recovered from the lungs. In the combo-treated hamsters an increased frequency of C-to-T and G-to-A mutations in the viral RNA is observed as compared to the single treatment groups which may explain the pronounced antiviral potency of the combination. Our findings may lay the basis for the design of clinical studies to test the efficacy of the combination of Molnupiravir and Favipiravir in the treatment of COVID-19.

CoV-2. At day four post-infection (pi), the animals were euthanized and lungs were collected for 89 quantification of viral RNA, infectious virus titers and lung histopathology as described before 4 (Fig. 90 1A). Molnupiravir treatment resulted in a dose-dependent reduction in the viral RNA copies per mg of 91 lung tissue with 1.3 (P=0.002), 1.9 (P<0.0001) and a 3.3 log10 (P<0.0001) reduction was noted in the 92 groups that had been treated BID with 75, 150 or 200 mg/kg, respectively (Fig. 1B) 1D). All the doses studied were well tolerated without significant weight 99 loss or any obvious adverse effects (Fig. 1E). and was formulated as 50 or 100 mg/ml (for groups with the highest dose) stocks in a vehicle 178 containing 10%PEG400 and 2.5% Kolliphor-EL in water. Favipiravir was purchased from BOC Sciences 179 (USA) and was formulated as a 50 mg/mL stock in 3% sodium bicarbonate. 180 RT-qPCR was performed on a LightCycler96 platform (Roche) as described before 4 . 199

Histology 204
For histological examination, the lungs were fixed overnight in 4% formaldehyde and embedded in 205 paraffin. Tissue sections (5 μm) were analyzed after staining with hematoxylin and eosin and scored 206 blindly for lung damage by an expert pathologist. The scored parameters, to which a cumulative score 207 of 1 to 3 was attributed, were the following: congestion, intra-alveolar hemorrhagic, apoptotic bodies 208 in bronchus wall, necrotizing bronchiolitis, perivascular edema, bronchopneumonia, perivascular 209 inflammation, peribronchial inflammation and vasculitis. 210

Deep sequencing and analysis of whole genome sequences 211
Genomic sequences from all samples were obtained using SureSelect XT target enrichment and Illumina 212 sequencing.
Reads generated were trimmed with Trim Galore 213 (https://github.com/FelixKrueger/TrimGalore). Duplicated reads were removed using Picard 214 (http://broadinstitute.github.io/picard). Reads from the inoculation sample were mapped to the SARS-215 CoV-2 reference genome (NC_045512) from GenBank using BWA-MEM 14 . The mapping quality was 216 checked using Qualimap and the consensus whole genome sequence was generated using QUASR 15,16 . 217 Reads from the lung samples were mapped to this unique reference sequence. Genomes with less than 218 less than a 100 read depth were excluded. Variants above 1% and with a minimum of 2 supporting 219 reads per strand were identified at sites with a read depth of ≥ 10 using VarScan 17 . 220

Statistics 221
GraphPad Prism (GraphPad Software, Inc.) was used to perform statistical analysis. Statistical 222 significance was determined using the non-parametric Mann Whitney U-test. P-values of ≤0.05 were 223 considered significant. 224

Data Availability 225
All of the data generated or analysed during this study are included in this published article. 226