This is a PDF file of a peer-reviewed paper that has been accepted for publication. Although unedited, the content has been subjected to preliminary formatting. Nature is providing this early version of the typeset paper as a service to our authors and readers. The text and figures will undergo copyediting and a proof review before the paper is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers apply.

Highlights d Spatial proteogenomic single-cell atlas of healthy and obese murine and human liver d Validated flow cytometry and microscopy panels for all hepatic cells d LAMs are differentially located in the lean and obese liver d Evolutionary conserved BMP9/10-ALK1 axis is essential for KC development

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) rapidly spread around the globe after its emergence in Wuhan in December 2019. With no specific therapeutic and prophylactic options available, the virus has infected millions of people of which more than half a million succumbed to the viral disease, COVID-19. The urgent need for an effective treatment together with a lack of small animal infection models has led to clinical trials using repurposed drugs without preclinical evidence of their in vivo efficacy. We established an infection model in Syrian hamsters to evaluate the efficacy of small molecules on both infection and transmission. Treatment of SARS-CoV-2−infected hamsters with a low dose of favipiravir or hydroxychloroquine with(out) azithromycin resulted in, respectively, a mild or no reduction in virus levels. However, high doses of favipiravir significantly reduced infectious virus titers in the lungs and markedly improved lung histopathology. Moreover, a high dose of favipiravir decreased virus transmission by direct contact, whereas hydroxychloroquine failed as prophylaxis. Pharmacokinetic modeling of hydroxychloroquine suggested that the total lung exposure to the drug did not cause the failure. Our data on hydroxychloroquine (together with previous reports in macaques and ferrets) thus provide no scientific basis for the use of this drug in COVID-19 patients. In contrast, the results with favipiravir demonstrate that an antiviral drug at nontoxic doses exhibits a marked protective effect against SARS-CoV-2 in a small animal model. Clinical studies are required to assess whether a similar antiviral effect is achievable in humans without toxic effects.